Electronic polymers offer significant advantages towards ubiquitous computing due to their low-cost, flexibility and benign fabrication conditions. In this research, ferroelectric polymers were investigated for usage in non-volatile memory applications. The work is focused on the fabrication and ferroelectricity of Polyvinylidene-trifluoroethylene and Polyamide-11 (Nylon-11) thin films. Polyvinylidene fluoride (PVDF) and its copolymers were the first class of ferroelectric polymers discovered. Although the processes and properties of PVDF and copolymers have been extensively studied, most of the reports have been on polymers in the bulk form. This work focuses on thin films of PVDF-TrFE (75:25) copolymer fabricated by solution spin-casting. Remnant polarization, Pr, of the thin films was measured to be 6 μC/cm 2 with a coercive field, Ec, of 60 MV/m. The thin film properties are highly dependent on the temperature of crystallization and is attributed to the amount of all-trans β-phase and crystallinity. Fatigue, defined as polarization loss with repeated switching, was studied and a model based on space charge formation was proposed as the fatigue mechanism. Space charge formation was proposed to be caused by electrochemical reaction of ions (F-) at electrodes and accumulations of detrapped ions at grain boundaries. Incorporating a F- scavenger and forming small crystallites was both observed to decrease fatigue. Nylon-11 and other odd-nylons are the only other class of polymers that have been reported to exhibit ferroelectric D-E hysteresis. The published work has almost exclusively been reported on melt-quenched and cold-drawn bulk polymers and consequently there is no literature on ferroelectricity in thin film odd-nylons. The present work developed a process for the fabrication of ferroelectric thin films of nylon-11 by spin-casting. Among the solvents tested, only a solution with m-cresol was observed to result in ferroelectricity in spun films and could be correlated to the crystal structure of the films. A polarization response, Pr, of 5μC/cm 2 with a coercive field, Ec, of 50MV/m was observed. The processing conditions and their effect on crystal structure were investigated to achieve optimal polarization response. In conclusion, PVDF-TrFE copolymers were fabricated in thin film form and process conditions developed to improve ferroelectric properties. Nylon-11 thin films were successfully fabricated for the first time with a polarization response equivalent to that in bulk polymers.

No description available.

<p> Resonators are a major component in RF electronic products. They are used in a host of ways to filter radio signals. Modern and Future RF communications have placed high demands on the industry; requiring low power usage, wide array of applications and resistance to noise. </p><p> In this thesis, a discussion of the motivation for RF MEMS filters and basic theory is given with an explanation of the concepts of Q factor, piezoelectricity, acoustics theory, the major types of resonators (SAW, BAW, CMR or LAMB), apodization theory and techniques as well as design, simulation of CMR and BAW devices, testing and process development of aluminum nitride by RF reactive sputtering at RIT. </p><p> Finite element analysis was performed on a number of factors of aluminum nitride contour mode resonators (CMR) from piezoelectric film thickness, to electrode pitch, electrode thickness and electrode configuration; to understand the effects. First order and second order vibration modes were seen including symmetric <i>S</i>₀, <i>S</i>₁ and antisymmetric <i>A</i>₀, <i>A</i>₁ resonant modes in the pizeoacoustic devices and higher. A series of time dependent video simulations of SAW, BAW and LAMB wave resonators were also performed, perhaps the first of their kind. </p><p> The RF reactive sputtering deposition for aluminum nitride was developed at RIT by a fractional factorial experiment with the factors being RF power, nitrogen to argon flow rate ratios, changing the distance of the wafer to the platen from 5 to 4 cm, use of a aluminum, molybdenum or virgin silicon seed layer and chamber pressure. In nearly all cases it was found that an RF power of 1000W is the most important factor contributing to the &lang;002&rang; orientation. The decreasing of the target distance may inhibit a reaction mechanisms in the plasma resulting in a more amorphous deposition. It may be due to the increase in temperature resulting from the higher RF power that promotes the growth of &lang;002&rang; oriented aluminum nitride. A molybdenum seed layer tends to have a stronger &lang;002&rang; peak relative to aluminum and a chamber pressure of 3mT was found to exhibit a deposition that most favors the &lang;002&rang; oriented aluminum nitride. </p><p> It was found that molybdenum is not consumed in a wet etch of KOH. Molybdenum is oxidized during photo resist ashing. The Contact Vias were necessarily over retched in order to ensure complete removal of Al-N over the Bottom Electrode. </p><p> C-V measurements were done on the aluminum nitride to determine its quality, the measured extensional piezoelectric coefficient <i>d</i>₃₃ is -0.000108716 <i>nm/V</i>, which is -0.108716 <i> pm/V</i> lower than 8pm/V typically reported. The lower piezo electric coefficient measured as compared with typical values, may be due to low film density a result of the high power used in the RF reactive sputtering that was used to heat the platen to a high enough temperature to promote the?002?oriented growth of AlN. </p><p> A series of iterations were designed and S11 frequency response measured. The electrode overlap from 25 to 50 to 75&mu;m, it does not appear to have an effect on the resonant frequency, but does increase the amplitude of the response at that die's given frequency. Increasing the anchor width from 5&mu;m to 10&mu;m to 20&mu;m lowers the relative amplitude of the response therefore lowering the Q of the resonator. It may be that the increasingly wide anchor, increases the mechanical resistances within the device and thereby lowers the Q factor of the resonator. Increasing the number of electrodes increases the relative amplitude of the response. Increasing pitch from 5&mu;m to 6&mu;m seems to have a small effect on the resonant frequency of the devices, shifting them from 4.57 to 4.59 GHz. A quality factor was measured, with an anchor width of 5&mu;m, pitch of 5&mu;m, 24 electrodes and an electrode overlap of 75&mu;m had a measured Q value of 98.8.</p>

<p> Piezoelectric voltage transformers (PTs) can be used to transform an input voltage into a different, required output voltage needed in electronic and electro- mechanical systems, among other varied uses. On the macro scale, they have been commercialized in electronics powering consumer laptop liquid crystal displays, and compete with an older, more prevalent technology, inductive electromagnetic volt- age transformers (EMTs). The present work investigates PTs on smaller size scales that are currently in the academic research sphere, with an eye towards applications including micro-robotics and other small-scale electronic and electromechanical sys- tems. PTs and EMTs are compared on the basis of power and energy density, with PTs trending towards higher values of power and energy density, comparatively, indicating their suitability for small-scale systems. Among PT topologies, bulk disc-type PTs, operating in their fundamental radial extension mode, and free-free beam PTs, operating in their fundamental length extensional mode, are good can- didates for microfabrication and are considered here. Analytical modeling based on the Extended Hamilton Method is used to predict device performance and integrate mechanical tethering as a boundary condition. This model differs from previous PT models in that the electric enthalpy is used to derive constituent equations of motion with Hamilton&rsquo;s Method, and therefore this approach is also more generally applica- ble to other piezoelectric systems outside of the present work. Prototype devices are microfabricated using a two mask process consisting of traditional photolithography combined with micropowder blasting, and are tested with various output electri- cal loads. 4mm diameter tethered disc PTs on the order of .002cm</p><p>3 , two orders smaller than the bulk PT literature, had the followingperformance: a prototype with electrode area ratio (input area / output area) = 1 had peak gain of 2.3 (&plusmn; 0.1), efficiency of 33 (&plusmn; 0.1)% and output power density of 51.3 (&plusmn; 4.0)W cm</p><p>-3 (for output power of80 (&plusmn; 6)mW) at 1M? load, for an input voltage range of 3V-6V (&plusmn; one standard deviation). The gain results are similar to those of several much larger bulk devices in the literature, but the efficiencies of the present devices are lower. Rectangular topology, free-free beam devices were also microfabricated across 3 or- ders of scale by volume, with the smallest device on the order of .00002cm</p><p>3 . These devices exhibited higher quality factorsand efficiencies, in some cases, compared to circular devices, but lower peak gain (by roughly 1/2 ). Limitations of the microfab- rication process are determined, and future work is proposed. Overall, the devices fabricated in the present work show promise for integration into small-scale engi- neered systems, but improvements can be made in efficiency, and potentially voltage gain, depending on the application</p><p>

At 0.18 microns and below minimum device dimensions in Ultra Large Scale Integrated Circuits, signal net parasitic delay amounts to 80% of the overall path delay. This leads to serious problems relating to signal timing, crosstalk, noise and power consumption. Although Copper is being used as an alternative to Aluminum interconnects to reduce the resistive component of the RC delays, finding a suitable material to replace Silicon Dioxide (SiO2) as the interlayer dielectric poses serious challenges. Most of the inorganic candidates are variants of SiO2, while the most prominent among polymeric materials belong to the polyparaxylylene family. The primary disadvantage of polyparaxylylene materials is their low thermal stability. While SiO2 based inorganic films exhibit excellent thermal stability, they offer only incremental improvement in the dielectric constant. The thin film deposition technique for these materials is important as it directly impacts the cost of manufacturing. Chemical Vapor Deposition is known to make high purity, conformal thin films, and is compatible with current silicon manufacturing technology. This research is primarily focused to develop materials which have (i) Low dielectric Constant; (ii) High thermal stability, and to deposit them using Chemical Vapor Deposition technique. The vision was to develop a composite thin film material with the thermal stability of SiO2 and the low dielectric constant of paraxylylenes. The first objective of this research was to develop a technique to deposit SiO2 films at near room temperatures. Thin conformal films of SiO2 were deposited at temperatures around 50°C using Di-acetoxy-di-tertiary-butoxy silane (DADBS) as the precursor. The thermal stability, optical and electrical properties of the codeposited thin films were systematically studied. It was possible to control the composition of these films smoothly and these films were shown to be of nanocomposite type. However, the thermal stability of these nanocomposite thin films was only marginally better than that of paraxylylenes. These films were then heat treated under oxygen to 'burn off' the polymer content. It was shown that annealing these films in oxygen environment leaves porous SiO 2 which exhibits the thermal stability of SiO2 and the porosity results in lower dielectric constant.

Two types of GaAs-based microcoolers have been demonstrated for the first time. One has a 2 μm thermal barrier made from a 100-period Al0.10 Ga0.90As - Al0.20Ga0.80As superlattice. Another has a thermal barrier made from a 5.2 μm Al0.10Ga 0.90As layer. Microcoolers of these two types with a size of 60 μm x 60 μm show maximum cooling temperatures around 0.8°C and 2°C at heatsink temperatures of 25°C and 100°C respectively. The microcoolers have potential to provide in situ cooling for semiconductor devices. As an application of in situ cooling, a unipolar self-cooled laser structure is proposed and has been fabricated. The structure integrates a graded-index (GRIN) separate confining heterostructure (SCH) strained-layer InGaAs quantum well (QW) laser unit and a single-layer AlGaAs cooler using an Esaki tunnel junction as a connector. In this scheme, cooling exists on two sides of the tunnel junction. Broad-area lasers of this structure with a cavity length of 500 μm have an average threshold current density around 212 A/cm 2. A preliminary method is proposed to evaluate the cooling effect of the integrated cooler. With this method, cooling can be estimated from the movement of the spectrum of the laser, excited by pulsed current, as the pulse width varies. Using this method, potential 2∼4°C temperature reduction as a result of the integrated cooler is found in the active layer of the self-cooled laser. Some technological platforms have been built-up to support the investigations on the microcoolers and self-cooled laser structure. Firstly, some primary MOCVD technologies such as controlling compositions, fabricating high quality interfaces, and doping have been developed for the MOCVD system itself and the compound semiconductor group. Secondly, annealing parameters—temperature and time—have been optimized for making ohmic contacts in a homemade carbon-stripe furnace. Thirdly, the SiH4 flow rate for doping the n-side of GaAs tunnel junction has been optimized to obtain a tunnel junction with low zero-bias tunnel resistance. A low zero-bias specific tunnel resistance of 9.59 × 10−5 Ω·cm2 has been achieved, which is the best reported result for the tunnel junction grown at normal growth temperatures. Theoretical evaluation and experimental results indicate that tunneling of the tunnel junction with the n-side doped with optimal SiH4 flow rate is mainly defect-assisted. Finally, a nominal-980-nm In0.2Ga0.8As GRIN-SCH strained-layer QW laser has been grown, fabricated, and characterized.

Intersubband transitions in both lattice-matched GaAs/AlGaAs and pseudomorphically strained GaAs/InGaAs/AlGaAs multi-quantum well structures are investigated. Absorption and emission devices were grown by atmospheric organometallic vapor phase epitaxy (OMVPE) and characterized using electroluminescence (EL) and Fourier transform (FT) spectroscopes. Several GaAs/AlGaAs quantum well infrared photodetectors infrared photodetectors (QWIPs) were fabricated and their performance evaluated. The dark currents and responsivities were measured and are found to be comparable to devices grown by advanced molecular beam epitaxial (MBE) methods. Significant improvements in the absorption characteristics were obtained by adopting a localized delta-doping profile in the absorption QWs as opposed to homogeneously doped structures. The measured absorption strength increased significantly in samples which were delta doped with silicon to sheet carrier densities of approximately $1\times10\sp{12}$ cm$\sp{-2}$. The transition linewidths decreased from 40 meV in the uniformly doped sample to 20 meV in the delta-doped devices. In addition, designs for intersubband emitters based on multiple quantum well structures are investigated. The electrical and optical properties of these devices are characterized. While no intersubband emission was observed from these device some possibilities for future designs are discussed. Electrons injected into the excited subbands of SQW structures via a resonant tunneling mechanism was also investigated. The optical properties of light emitted from electron-hole recombination in the QW, characterized by electroluminescence spectroscopy, is correlated with the device electrical characteristics. At 77$\sp\circ$ K both the I-V and the light attributes show strong nonlinear differential behavior similar to NDR effects in resonant tunneling diodes. Semiconductor laser diodes fabricated using these materials exhibit a single mode lasing spectrum and an output power characteristic which is abruptly extinguished as the field-induced resonant injection condition is exceeded.

Stimulated emission, lasing, and related properties of III-V nitride heterostructures are studied. A strain-dependent semi-empirical tight-binding model is developed, using the valence force-field model of Keating, to predict the atomic positions in the strained wurtzite crystal lattice. Predicted deformation potentials and strain-induced exciton splitting are shown to closely match data in the literature. The spectral properties of the edge luminescence from GaN-AlGaN heterostructures is investigated. The existence of stimulated emission is demonstrated and a measurement of the optical gain spectra is reported. In addition, the light emission properties of GaN-AlGaN separate confinement heterostructures is studied. The measured luminescence properties are improved for active region designs with fewer, thicker wells. An analysis of the trend is presented demonstrating recombination at the well-barrier interface as a significant factor. The results also indicate that the quantum wells experience compressive strain from the lattice mismatch with the AlGaN cladding layers. Further experimental results demonstrate that the commonly observed surface stimulated emission is related to in-plane optical gain, and is observed most commonly in samples with rough surface morphology. Photopumping results from GaN-AlGaN laser platelets are presented and discussed. Laser oscillation in GaN-AlGaN separate confinement heterostructures is demonstrated in which the optical cavity is formed unintentionally by parallel cracks in the epilayer. The observed laser modes are broad and shift to shorter wavelengths with increasing pump intensity. An analysis is presented revealing that mode shifting resulting from carrier-induced refractive index changes restricts the observation of laser modes to short optical cavities. GaN-AlGaN Bragg reflectors are investigated through reflectivity modeling and characterization. A transfer-matrix model is developed with an empirical relation for the refractive indices and predictions of the model are compared with data in the literature. Experimental results are then presented and compared with the predictions of the model. The design and characterization of GaN-AlGaN vertical cavity surface emitting lasers is studied. Luminescence spectra are presented from two devices which demonstrate sharp, highly-polarized, regularly spaced modes for pump intensities above a threshold. The spacing of the laser modes is shown to match the mode spacing predicted by the reflectivity model.

Thesis (Ph.D.)--University of Illinois at Urbana-Champaign, 2009. / Source: Dissertation Abstracts International, Volume: 70-06, Section: B, page: 3728. Adviser: Kyekyoon Kim. Includes bibliographical references (leaves 111-125) Available on microfilm from Pro Quest Information and Learning.

<p> Thermoelectric devices have been employed for many years as a reliable energy conversion technology for applications ranging from the cooling of sensors or charge coupled devices to the direct conversion of heat into electricity for remote power generation. However, its relatively low conversion efficiency has limited the implementation of thermoelectric materials for large scale cooling and waste heat recovery applications. Recent advances in semiconductor growth technology have enabled the precise and selective engineering of material properties to improve the thermoelectric figure of merit and thus the efficiency of thermoelectric devices. Accurate characterization at the intended operational temperature of novel thermoelectric materials is a crucial component of the optimization process in order to fundamentally understand material behavior and evaluate performance. </p><p> The objective of this work is to provide the tools necessary to characterize high efficiency bulk and thin-film materials for thermoelectric energy conversion. The techniques developed here are not bound to specific material or devices, but can be generalized to any material system. </p><p> Thermoreflectance imaging microscopy has proven to be invaluable for device thermometry owing to its high spatial and temporal resolutions. It has been utilized in this work to create two-dimensional temperature profiles of thermoelectric devices during operation used for performance analysis of novel materials, identification of defects, and visualization of high speed transients in a high-temperature imaging thermostat. We report the development of a high temperature imaging thermostat capable of high speed transient thermoelectric characterization. In addition, we present a noninvasive method for thermoreflectance coefficient calibration ideally suited for vacuum and thus high temperature employment. This is the first analysis of the thermoreflectance coefficient of commonly used metals at high-temperatures. </p><p> High temperature vacuum thermostats are designed and fabricated with optical imaging capability and interchangeable measurement stages for various electrical and thermoelectric characterizations. We demonstrate the simultaneous measurement of in-plane electrical conductivity and Seebeck coefficient of thin-film or bulk thermoelectric materials. Furthermore, we utilize high-speed circuitry to implement the transient Harman technique and directly determine the cross-plane figure of merit of thin film thermoelectric materials at high temperatures. </p><p> Transient measurements on thin film devices are subject to complications from the growth substrate, non-ideal contacts and other detrimental thermal and electrical effects. A strategy is presented for optimizing device geometry to mitigate the impact of these parasitics. This design enabled us to determine the cross-plane thermoelectric material properties in a single high temperature measurement of a 25&mu;m InGaAs thin film with embedded ErAs (0.2%) nanoparticles using the bipolar transient Harman technique in conjunction with thermoreflectance thermal imaging. This approach eliminates discrepancies and potential device degradation from the multiple measurements necessary to obtain individual material parameters. Finite element method simulations are used to analyze non-uniform current and temperature distributions over the device area and determine the three dimensional current path for accurate extraction of material properties from the thermal images. Results match with independent measurements of thermoelectric material properties for the same material composition, validating this approach. </p><p> We apply high magnification thermoreflectance imaging to create temperature maps of vanadium dioxide nanobeams and examine electro-thermal energy conversion along the nanobeam length. The metal to insulator transition of strongly correlated materials is subject to strong lattice coupling which brings about the unique one-dimensional alignment of metal-insulator domains along nanobeams. Many studies have investigated the effects of stress on the metal to insulator transition and hence the phase boundary, but few have directly examined the temperature profile across the metal-insulating interface. Here, thermoreflectance microscopy reveals the underlying behavior of single-crystalline VO₂ nanobeams in the phase coexisting regime. We directly observe highly localized alternating Peltier heating and cooling as well as Joule heating concentrated at the domain interfaces, indicating the significance of the domain walls and band offsets. Moreover, we are able to elucidate strain accumulation along the nanobeam and distinguish between two insulating phases of VO₂ through detection of the opposite polarity of their respective thermoreflectance coefficients.</p>