Control of the film stress and optical property has long been considered as an issue in the tunable optical MEMS (Micro-Electro-Mechanical Systems) devices. In this paper, the atmospheric evolution of Titanium Dioxide (TiO2) and Silicon Dioxide (SiO2) thin films for the optical MEMS devices were studied. These films were prepared by ion-assisted e-beam evaporation. It is found that as-deposited SiO2 films exhibit compressive stress; whereas, it is tensile in the TiO2 films under present processing conditions. When annealed at 150 °C, both SiO2 and TiO2 films show slight changes in stress with annealing time. However, increasing the anneal temperature to 250°C caused an apparent change of film stresses with time, in which SiO2 film turns into less compressive and TiO2 film appears to be more tensile. The optical properties after annealing were also investigated by measuring the thickness and the refractive index changes using the spectroscopic ellipsometry technique. At both experimental temperatures, the film thickness increases slightly and the refractive index at 1550 nm decreases a little at the initial annealing stage for SiO2 films. For TiO2 films, it is found that the refractive index increases after annealing at 250°C. This might be caused by the TiO2 film densification process during amorphous-to-crystalline phase transformation. Because most of the significant film evolutions occur during the initial 12 hours of annealing, a practical way of stabilizing the film properties in a MEMS device is to pre-anneal the as-deposited thin films.

This paper describes atomistic determinations of the Young's modulus of free standing thin films, or nanoplates. Using a combination of analytical formulation and molecular statics simulations, we show that the Young's modulus of a nanoplate may either increase or decrease with the thickness. It is the competition of bond saturation and bond loss on surfaces that dictates the increase or decrease. Taking Cu as an example, we demonstrate that the Young's modulus is larger than its bulk counterparts for nanoplates having some surfaces and loading directions, and smaller for others.

The goal of this paper is to investigate the effects of film thickness and the presence of a passivation layer on the mechanical behavior of electroplated Cu thin films. In order to study the effect of passivating layers, freestanding Cu membranes were prepared using standard silicon micromachining techniques. Some of these Cu membranes were passivated by sputter depositing thin Ti films with thicknesses ranging from 20 nm to 50 nm on both sides of the membrane. The effect of film thickness was evaluated by preparing freestanding films with varying thickness but constant microstructure. To that effect, coatings of a given thickness were first vacuum annealed at elevated temperature to stabilize the microstructure. The annealed films were subsequently thinned to various thicknesses by means of chemical mechanical planarization (CMP) and freestanding membranes were prepared both with and without Ti passivation. The stress-strain curves of the freestanding Cu films were evaluated using the bulge test technique. The residual stress and elastic modulus of the film are not affected significantly by the passivation layer. The elastic modulus does not change with film thickness if the microstructure keeps constant. The yield stress increases if the film is passivated. For passivated films, yield stress is proportional to the inverse of the film thickness, which is consistent with the formation of a boundary layer of high dislocation density near the interfaces.

Hardness values of Au/Fused Quartz and SiO2/Al systems, which correspond to the cases of soft film on hard substrate and hard film on soft substrate, were measured using both the nanoindentation and nanoscratch techniques. The effect of substrate interaction on the measurement of hardness using the nanoscratch technique is found to be much less pronounced compared to that of the nanoindentation technique. Such reduction in substrate effect is attributed to the features used in the nanoscratch analysis: (a) direct imaging of residual profile allows for the effect of pile-up/sink-in to be considered, (b) lower normal loads applied as compared to the nanoindentation, (c) effect of elastic recovery of the plastically deformed surfaces is included in the nanoscratch analysis, whereas the nanoindentation analysis is based solely on the load-displacement data. Moreover, experiments with residual scratch depths as shallow as 3 nm are used to estimate hardness of thin films; a promising indication for the use of such technique in the measurements of ultra-thin films.

To improve the mechanical compliance of a piezoelectric generator membrane, alterations in PZT processing including variations in crystallization temperature, cooling rates and precursor chemistry were made to reduce the residual stress. In addition to changing the stresses, these treatments also altered the morphology of the PZT and its modulus. Nanoindentation of these films were made to determine the moduli of the PZT films, which varied from 70GPa to 90GPa. The residual stress of 40:60 2-MOE PZT taken by x-ray diffraction were shown to reduce in residual stress from 210 MPa to 154 MPa with annealing. The static pressure-deflection of membranes with these films showed correlating composite effective stresses of 110MPa and 83MPa. Adding PEG to the precursor solution also lowered the stress of the PZT films. Bulge testing showed that decreasing the crystallization temperature from 700°C to 600°C lowered residual stress from 210 MPa to 154 MPa. Raman spectroscopy showed differences in both stress and structure of PZT deposited on thin support structures (2μm) over bulk wafers (350 μm). The compliance of the generators was also increased by etching Pt in highly stressed regions.

Poly-SiC films were deposited on Si and SiO2 substrates in a high-throughput, low pressure chemical vapor deposition (LPCVD) furnace using dichlorosilane (DCS) and acetylene precursors. The deposition temperature and pressure were fixed at 900°C and 2 Torr, respectively, while the flow rate of DCS was varied between 18 and 54 sccm. Poly-SiC deposition rates on both Si and SiO2 were nearly identical to each other and increased as a function of DCS flow rate. Consistent with both substrate materials, the following observations were made. A slope change of the deposition rate versus DCS flow rate was observed around a DCS flow rate of 35 sccm. Residual stress varied with respect to the deposition rate, with tensile stresses occurring at lower deposition rates and compressive stresses at higher deposition rates. The tensile-to-compressive stress transition corresponded to the slope change of the deposition rate versus DCS flow rate. The surface morphology consisted of pyramidal grains, as observed under an SEM. TEM analysis for poly-SiC films grown on Si substrates showed that microstructural differences exist for poly-SiC films having tensile and compressive stresses.

Characteristics of Cu/C:H films on the PET substrate prepared by ECR-MOCVD coupled with a DC bias under Cu(hfac)2-Ar-H2 was investigated with special attention to process parameters. Our results showed that the Cu/C:H film has strongly adhere with the polymer substrate by chemical bonding and exhibited wide range of electrical conductivity that could be controlled by process parameters. The increase in H2/Ar ratio, microwave power and negative DC bias voltage brought on the higher pulling strength between Cu/C:H films and PET substrate. The effects of surface pretreatments of polymer substrate on pulling strength were insignificant in the range of our experimental range.

The process-induced stress in interconnects within integrated circuits (IC) has a direct influence on the mean time to failure of the devices. Since measurement of stress in individual metallised lines is not possible by existing techniques, another approach has been adopted where a test structure is generated during fabrication based on a micro-rotating cantilever sensor. To support the design, finite element modeling (FEM) has been performed. By comparing the rotation predicted by FEM simulations and that observed experimentally, a clear discrepancy is observed which is critically dependent on the details of the sensor design, the pattern transfer of the lithographic process and on the dry etching processing.

Polycrystalline silicon (polysilicon) single edge-notched fatigue specimens with micrometer-sized dimensions were macromachined and subjected to cyclic loading using an integrated electrostatic actuator. The effects of fatigue were determined by comparing the monotonic bend strength measured after cyclic loading to the monotonic bend strength of specimens that received no cycling. Both strengthening and weakening were observed, depending on the levels of mean stress and fatigue stress amplitude during the cyclic loading. Monotonic loading with similar stress levels prior to bend strength measurements had no effect on measured bend strength. Possible physical mechanisms responsible for this fatigue behavior are discussed.

The present work uses wafer curvature (disk) method to measure the temperature variation of the biaxial modulus and the thermal expansion coefficient TEC from 25°C to 205°C. The following thin films were measured: PolySi, low stress LPCVD SixNy and (Ba0.7, Sr0.3)TiO3 (BST). To improve the precision, perfect circular thin wafers were used: 4 inches circular 280μm thick <111>Si and 450μm thick Fused Silica. The measurements were performed using a commercial Tencor FLX2320 Stress Measurement System. The film thickness was measured with Tencor TF1 thin film optical system.

In micro-electro-mechanical system (MEMS) devices that contain multi-layers of metallic, ceramic, and polymeric materials, acquiring a complete understanding of the mechanical properties and interfacial strength are critical in designing the devices. In general, these layered materials are specially fabricated with thickness of each layer on submicron and/or micron scale. Particularly, the properties of these layered materials strongly depend on the fabrication processes and subsequent treatments. In this paper, we study the magnetic metallic films, NiFe (80/20 in weight), for MEMS generator/motor applications. A set of tensile tests was conducted on the plated nickel-iron metallic films of various thicknesses and subjected to various annealing conditions. Distinctive differences in mechanical behaviors between the electroplated films and the corresponding bulk alloys were observed. Furthermore, the NiFe film as-deposited exhibits almost pure elastic deformation and brittle fracture at room temperature; whereas the NiFe film subjected to annealing exhibits strain hardening and ductile fracture when tested at room temperature. X-ray diffraction was used to capture the variation of crystalline structure and the residual stresses in the electroplated films as deposited and after annealing. An atomic force microscope was used to obtain the surface morphology of the films as deposited and after annealing.

Intrinsic stress in coatings is often responsible for its performance. We studied tensile stress in sputter deposited chromium films as a function of film thickness and Ar pressure during deposition. We correlate the stress evolution to the grain growth in the polycrystalline films. Both grain growth and stress evolution obey the same power law dependence on thickness. We conclude that the tensile stress is generated at the grain boundaries. The power law exponent did not depend on pressure of Ar and remained 0.36. However, texture and microstructure in the layers changed when pressure was increased from 2×10-2 to 6×10-2 mbar. Texture switched from 110 to 111 fiber type. Grooves and sharp star-like grains were observed at higher pressure. We explain changes in terms of suppressed shadowing and less surface diffusion.

Atomistic simulations using molecular statics are carried out to study dislocation plasticity in thin metal films attached to stiff substrates. The analysis utilizes a sample two-dimensional crystal, with an embedded initial point defect used for triggering dislocation activities in a controlled manner. The existence of an interface between the film and the substrate is shown to delay plastic yielding and lead to film strengthening. The capability of atoms to slide along the interface plays a crucial role in determining the macroscopic stress-strain response and the microscopic dislocation activities. Within the modeling framework we examine the quantitative interfacial sliding behavior and the resulting dislocation-interface interactions and their consequences.

This paper reports on the use of poly(dimethylsiloxane) (PDMS) thin films to bond pairs of glass slides, borosilicate glass, and silicon substrates, with an emphasis on application for wafer-level three-dimensional (3D) heterogeneous integration technology platforms. PDMS films were spin-cast and cured, surface modified using low power oxygen plasma, and then bonded to various materials. These bonds were destructively tested using a four point bending technique. The critical adhesion energy obtained using surface modified PDMS to bond glass slides is 3.0 J/m2. Adhesion energies obtained using unmodified PDMS are 2.8, 1.8, and 1.6 J/m2 for bonded silicon, glass slides, and borosilicate glass, respectively. Correlation of these results to material surface and interface properties indicates that although PDMS has process advantages as a bonding material for heterogeneous integration, its adhesion strength is lower than that of other dielectric bonding glues.

Coherent Gradient Sensing (CGS) is a full-field optical technique that produces real-time images of macroscopic wafer curvature, which, for thin films, can be related to stress through Stoney's equation. Here we describe the use of CGS as an in situ diagnostic to observe film stress distributions during chemical vapor deposition. The application of this method to measure oxygen diffusion rates in thin film YBa2Cu3O6+δ(YBCO) and stresses in thin film PbxBa1-xTiO3 (PBT) under chemical vapor deposition (CVD) conditions will be discussed.

Understanding subcritical fracture of low-k dielectric materials and barrier thin films in buffered solutions of different pH value is of both technical and scientific importance. Subcritical delamination of dielectric and metal barrier films from low-k organosilicate glass (OSG) films in pH buffer solutions was studied in this work. Crack path and subcritical fracture behavior of OSG depends on the choice of barrier layers. For the OSG/TaN system, fracture takes place in the OSG layer near the interface, while in OSG/SiNx system, delamination occurs at the interface. Delamination behavior of both systems is well described by a hyperbolic sine model that had been developed previously based on a chemical reaction controlled fracture process at the crack tip. The threshold toughness of both systems decreases linearly with increasing pH value. The slopes of the reaction-controlled regime of the crack velocity curves for both systems are independent of pH as predicted by the model. Near transport-controlled regime behavior was observed in OSG/TaN system.

β-Ta films were prepared in an ultra-high vacuum sputter deposition system, and the stresses that arose during thermal cycles to 750° C were measured using an in situ substrate curvature measurement system. The oxygen content in the environment was controlled during both deposition and thermomechanical testing. In films with no added oxygen, phase transformation from β to α takes place in distinct “jumps.” Addition of controlled amounts of oxygen to the annealing atmosphere led to increases in compressive stress, inhibited the phase transformation, and altered the magnitudes of the “jumps” in stress. The stresses in Ta films were found to be sensitive to oxygen content as well as thermal history.

Laser Induced Delamination is a novel technique aimed at quantification of adhesion of thin polymer films to a metal substrate. In this technique a high power IR pulsed laser beam (maximum 500 mJ in 5 ns) is used to create the primary delaminated area in the form of a blister. The blister is formed due to heating of the metal substrate and because of partial evaporation of the polymer material adjacent to the interface. By varying the power of the laser pulse the gas pressure inside the blister is increased until further delamination takes place. This critical pressure is related to the strength of adhesion of the polymer to the substrate. The shape of the blister (radius and height) is controlled with a stylus profiler and with an optical confocal microscope.

In the present work 30 μm thick PET films on steel substrate were studied. From the shape of the obtained blisters the strength of adhesion of the polymer is estimated. Various models describing blister formation and growth are discussed.

MEMS-scale polycrystalline silicon 2 μm thin film specimens fabricated via the Multi User MEMS Processes (MUMPs) have been employed to obtain the non-uniform nanometric displacement fields in the vicinity of prefabricated circular and elliptical micron-sized perforations. For the hole diameter-to-specimen width ratios considered in this work, and for all practical purposes, the displacement solution for a hole in an infinite plate is applicable. This method requires the ability to reliably and repeatably acquire nanometer level displacements on freestanding thin films. These tensile tests were conducted by a custom microtensile tester with the aid of Atomic Force Microscopy (AFM) and a special gripper that makes use of electrostatically assisted UV adhesion to handle and load miniature MEMS specimens. This non-conventional procedure for material parameter determination relies on Digital Image Correlation (DIC) to compare two AFM images, one before and one after specimen loading, and thus compute the nanometer level displacement fields (<50 nm global displacements) in 15×15-μm2 (or smaller) areas, with better than 2–3 nm resolution in displacements. By posing and solving a non-linear least squares inverse problem where, for known applied loads and measured displacement fields in an infinite plate with either a circular or an elliptical hole, it is possible to recover the elastic modulus (E). The main advantage of this approach is the full utilization of high-resolution displacement measurements over a specific specimen area, using only measurements acquired at one load level. Further statistical measurements of material properties may be obtained at varying load levels.

Nanoindentation experiments were performed on single crystals of (100) Si using a series of triangular pyramidal indenters with centerline-to-face angles in the range 35.3° to 85.0°. The influences of the indenter geometry on cracking and phase transformation during indentation were systematically studied. Although reducing the indenter angle reduces the threshold load for cracking and increases the crack lengths, c, at a given indention load, P, the frequently observed relation between P and c3/2 is maintained for all of the indenters over a wide range of load. Features in the nanoindentation load-displacement curves in conjunction with Raman spectroscopy of the crystalline and amorphous phases in and around the contact impression show that the indenter geometry also plays a role in the phase transformation behavior. Results are discussed in relation to prevailing ideas about indentation cracking and phase transformation in silicon.