Anodisation has been studied for almost eighty years, primary in the field of corrosion science, as a simple and efficient way of producing thick protective oxide coatings on Al, Ti or Zr alloys. Anodisation is an electrochemical oxidation process which relies on the migration of ions across solid films under the action of a large electric field. From the fundamental point of view, many aspects regarding the growth of anodic films have been studied extensively. However, so far, little interest has been devoted to the mechanical aspects involved in the growth process, despite their considerable importance both from an applied as well as from a fundamental point of view. A solid understanding of internal stresses development is indeed crucial in order to guarantee the durability of anodic coatings, their structural and functional properties. In addition, the stress evolution directly reflects the motion of the ions in the film and therefore provides a unique means to investigate in situ the growth mechanisms of anodic films. In this thesis, we have studied the evolution of the internal stresses in anodic TiO2 films in situ during their growth. The stresses have been obtained from changes in the curvature of cantilevered anode samples, measured using a high-resolution multi-beam optical sensor. We demonstrate, for the first time, the capability of this type of curvature sensor for monitoring processes in liquid environments. Experimental data on the internal stresses developing in anodic TiO2 films is provided, and trends regarding the influence of the experimental conditions on the stress evolution are identified. In particular, the evolution of the internal stresses is shown to be strongly correlated with the evolution of the electrochemical variables, which directly demonstrates the interest of curvature measurements as a fundamental technique for investigating the details of the growth process of anodic oxide films. The reversible and irreversible stress contributions associated, respectively, with electrostriction and with growth-related ionic transport have been separated from one another and quantified. A novel constitutive model for the electrostriction stress has been proposed which explicitly takes into account the effect of dielectrostriction.

The reactions of atomically clean, titanium film surfaces with oxygen, deuterium, and water have been investigated. Auger Electron Spectroscopy was utilized to monitor the formation 9f a surface oxide in the case of oxygen exposure, and to characterize the deuteride which formed upon deuterium absorption, and its subsequent oxidation. Quantification of surface oxide stoichiometries was facilitated by novel data acquisition and treatment schemes. The quartz crystal microbalance was used to measure the mass of adsorbed oxygen or deuterium with submonolayer sensitivity. Electron energy loss spectroscopy was sensitive to the presence of Ti⁺³ in the surface oxide. The initial oxidation of the titanium surface was characterized by the dissociative adsorption of three mono1ayers of oxygen atoms at a constant rate. The oxide formed during this reaction stage was a Ti₂0₃/Ti0₂ mixture with a total thickness of 13 A. The rate of oxygen adsorption then decreased such that oxide growth was logarithmic with time. When the oxide had attained a total thickness of 20 A, the initial suboxide was converted to Ti0₂, and subsequent oxide formed was purely Ti0₂. Oxide growth occurred by oxygen anion migration under the influence of an electrostatic field, set up across the oxide layer by electron transfer from the metal to adsorbed oxygen species. The pressure dependence of the oxide growth rate and terminal thickness suggested a constant field growth mechanism. Clean titanium films reacted with deuterium to form a bulk deuteride TiDₓ (x<2). The oxide layer which resulted from oxygen exposure was characterized by the above techniques. Oxide layers greater than 20 A completely inhibited deuterium absorption by prohibiting 02 dissociation, but did not act as a diffusional barrier when additional dissociation sites were provided. Iron adlayers were found to accelerate the D₂ absorption reaction. Removal of the titanium films from the vacuum chamber to an isolable electrochemical reaction chamber, without exposure to the atmosphere, allowed a determination of the effect of the various gas/solid reactions on the subsequent electrochemical oxidation processes.

The electrochemical growth of metal-organic framework (MOF) coatings, utilising the anodic dissolution method, has been investigated as a means of preparing MOF coated electrodes for various electrochemical applications. A mechanistic understanding of the formation of the electrode coatings has been further developed. This understanding has been utilised to expand the scope of this technique; to allow for the electrochemical formation of Zn and Co zeoliticimidazolate framework (ZIF) coatings which was hitherto not believed to bepossible. Electrodes coated with Co and Zn ZIFs via this methodology were assessed for their capacitive behaviour and the Co ZIFs exhibited the highest, pure MOF areal capacitance values reported to date. This was attributed to the method of coating formation, which provides well adhered coatings of MOF particles integrated into the electrode surface providing a good electrical connection between the coating and the electrode. Incorporation of GO, via electrophoretic deposition during the coating growth, is shown to improve this capacitance still further. Thecorresponding Zn ZIFs exhibited resistances orders of magnitude higher than their Co analogues; modelling can explain this behaviour with the Co analogue of a given ZIF calculated to have a greater metal contribution to its LUMO leading to a more delocalised electronic structure. Electrodes coated with the Cu MOF HKUST-1 have enabled for the first time the use of MOFs as a template for the electrodeposition of anisotropic metal nanostructures. Such MOF encapsulated metal nanostructures are demonstrated to have applications in surface enhanced Raman spectroscopy (SERS). In addition the same MOF has been discovered to display a redox based hysteresis which allows for the rewritable storage of small amounts of electrically accessible data.

This work aims to investigate the effects of current density, electrolyte temperature and substrate composition on the morphology of porous anodic films formed on AA 2024-T3 alloy in sulphuric acid electrolytes and the factors that determine the transition between linear and sponge-like film porosities. Comparisons were made with pure aluminium. Particular attention is given to understanding the rising voltage that occurs during galvanostatic hard anodizing of the alloy and the role of oxygen in the anodizing process. Conditions were selected to be representative of typical hard and conventional anodizing processes. SEM was employed to observe the film morphology, which was then correlated with the voltage-time responses. The anodic film composition was investigated by TEM/EDX and SEM/EDX to determine the effect of alloy element enrichment and cell diameter on the distribution of copper species in the film. A real-time gravimetric method was developed to measure the rate of oxygen evolution during anodizing and its influence on the anodizing voltage and film morphology. Results showed that hard anodic films on AA 2024-T3 alloy formed at relatively high voltages have linear pores and cells, contrasting with sponge-like porosity under conventional anodizing. The linear porosity is shown to depend on the voltage, with a morphological transition occurring in the range 25 to 30 V, with linear cells promoted by a high current density and/or low electrolyte temperature. As the film thickens with time, pore blockage by oxygen bubbles, impedes oxidation of the alloy leading to current re-distribution and hence localized increases in the current density producing a rise of the anodizing voltage as anodizing proceeds. The rise of the anodizing voltage, which leads to an increasing call diameter and barrier layer thickness, has a minor influence on the rate of oxygen evolution, which typically consumes about 20 % of the applied current density. In contrast, the voltage rise in the presence of sponge-like films is comparatively negligible, which is suggested to be due to easier escape of oxygen from the film. The films comprising linear cells contain more copper than the sponge-like films, with copper being enriched at the cell boundaries. Moreover, a model is proposed to explain the enrichment of copper, suggesting that above a critical cell diameter, an alloy enrichment sufficient for oxidation of the alloying element can be maintained across the alloy/film interface. Below this diameter, the enrichment is less than that necessary for oxidation, and the alloying element is then incorporated into the film at the cell boundaries.

When porous silicon is plated with a catalytic metal, the two materials can act together as a single entity whose electrical properties are sensitive to its environment – the sensing component of an electrochemical gas sensor. Etching pores into silicon is an electrochemical process; and which type of doped silicon used is one of its key parameters. For nearly all reported porous silicon gas sensors, the silicon has been of the p-doped variety – because p-doped porous etching is better understood and the layers that result from it are more predictable – despite n-doped silicon having potentially significant benefits in ease of fabrication and being more conducive to plating by a catalyst. This experiment is an attempt at creating a palladium plated n-doped porous silicon layer, and an examination into what differentiates this fabrication process and the layers that result from the traditional p-doped type. The porous layers to be plated are to be the same and would ideally have properties that are a close approximation to what a functional gas sensor would require. This experiment defined a process that fabricated this “ideal” layer out of N-type, , double polished silicon wafers with a resistance of 20 Ω cm. The wafers were subjected to the anodic etching method with an HF/ethanol mixture as the electrolyte; and only two (of among many) fabrication parameters were varied: HF concentration of the electrolyte and total etching time. We find that a concentration of 12% HF (by volume) and an etching time of 6 hours result in layers most appropriate to carry into plating. The anodization current density is 15 mA cm-2. Deposition of the catalyst, palladium, is done using the electroless method by immersing the porous layer in a .001M PdCl2 aqueous bath. Characterization of this Pd/Porous Silicon layer was done by measuring resistivity by four point probe and imaging through Scanning Electron Microscopy. It was found that layers of a maximum average of 63 ± 6% porosity were created using our fabrication method. There is evidence of palladium deposition, but it is spotty and irregular and is of no improvement despite the n-doping wafer makeup. Resistivity in well-plated regions was measured to be 7-10 Ωcm, while resistivity in regions not well-plated was measured to be 70-140 Ω cm. This is comparable to previous literature values, indicating n-silicon porous silicon can be fabricated and still have potential as a catalytic layer, should metal deposition methods improve.

We use different methods to deposit Ti thin film on the ITO glass substrate. Under the circumstances of using fixed concentration of electrolyte, changing anodic time, and applying voltage, we are able to use anodic method to make ordered nano TiO2 thin film,of which the smallest pore size is 18nm and the thickness is 4500

Conductive anodic filament (CAF) is a failure mode in printed wiring boards (PWBs) which occurs under high humidity and high voltage gradient conditions. The filament, a copper salt, grows from anode to cathode along the epoxy-glass interface. Ready and Turbini (2000) identified this copper salt as the Cu2(OH)3Cl, atacamite compound. This work has investigated the influence of polyethylene glycol (PEG) and polyethylene propylene glycol (PEPG) fluxing agents on the chemical nature of CAF. For coupons processed with PEPG flux, with and without chloride, a copper-chloride containing compound was formed in the polymer matrix. This compound was characterized using x-ray photoelectron spectroscopy (XPS) as CuCl and an electrochemical mechanism for the formation of the chloride-containing CAF has been proposed. For PEG flux, with and without chloride, it has been shown that CAF only formed, but no copper containing compound formed in the matrix. It appears for PEG fluxed coupons, a PEG-Cu-Cl complex forms, binds the available Cu and acts as a barrier to the formation of CuCl in the polymer matrix. Meeker and Lu Valle (1995) have previously proposed that CAF failure is best represented by two competing reactions – the formation of a copper chloride corrosion compound (now identified as Cu2(OH)3Cl) and the formation of innocuous trapped chlorine compounds. Since no evidence of any trapped chloride compounds has been found, we propose that the formation of CAF is best represented by a single non-reversible reaction. For coupons processed with a high bromide-containing flux, bromide containing CAF was created and characterized using transmission electron microscopy (TEM) to be Cu2(OH)3Br. In addition, a copper-containing compound was formed in the polymer matrix and characterized using XPS as CuBr. An electrochemical mechanism for the formation of bromide-containing CAF has been proposed based on the XPS data. .

Conductive anodic filament (CAF) is a failure mode in printed wiring boards (PWBs) which occurs under high humidity and high voltage gradient conditions. The filament, a copper salt, grows from anode to cathode along the epoxy-glass interface. Ready and Turbini (2000) identified this copper salt as the Cu2(OH)3Cl, atacamite compound. This work has investigated the influence of polyethylene glycol (PEG) and polyethylene propylene glycol (PEPG) fluxing agents on the chemical nature of CAF. For coupons processed with PEPG flux, with and without chloride, a copper-chloride containing compound was formed in the polymer matrix. This compound was characterized using x-ray photoelectron spectroscopy (XPS) as CuCl and an electrochemical mechanism for the formation of the chloride-containing CAF has been proposed. For PEG flux, with and without chloride, it has been shown that CAF only formed, but no copper containing compound formed in the matrix. It appears for PEG fluxed coupons, a PEG-Cu-Cl complex forms, binds the available Cu and acts as a barrier to the formation of CuCl in the polymer matrix. Meeker and Lu Valle (1995) have previously proposed that CAF failure is best represented by two competing reactions – the formation of a copper chloride corrosion compound (now identified as Cu2(OH)3Cl) and the formation of innocuous trapped chlorine compounds. Since no evidence of any trapped chloride compounds has been found, we propose that the formation of CAF is best represented by a single non-reversible reaction. For coupons processed with a high bromide-containing flux, bromide containing CAF was created and characterized using transmission electron microscopy (TEM) to be Cu2(OH)3Br. In addition, a copper-containing compound was formed in the polymer matrix and characterized using XPS as CuBr. An electrochemical mechanism for the formation of bromide-containing CAF has been proposed based on the XPS data. .

In today’s communication devices, the need for high performance inductors is increasing as they are extensively used in RF integrated circuits (RFICs). This need is even more pronounced for variable inductors as they are widely required in tunable filters, voltage controlled amplifiers (VCO) and low noise amplifiers (LNA). For RFICs, the main tuning elements are solid state varactors that are used in conjunction with invariable inductors. However, they have limited linearity, high resistive losses, and low self resonant frequencies. This emphasizes the need for developing another tuning element that can be fabricated monolithically with ICs and can offer high range of tuning. Due to the ease of CMOS integration and low cost silicon based IC fabrication, the inductors currently used are a major source of energy loss, therefore driving the overall quality factor and performance of the chip down. During the last decade there has been an increase in research in RF MicroelectroMechanical Systems (RF MEMS) to develop high quality on chip tunable RF components. MEMS capacitors were initially proposed to substitute the existing varactors, however they can not be easily integrated on top of CMOS circuits. RF MEMS variable inductors have recently attracted attention as a better alternative. The research presented here explores using porous anodic alumina (PAA) in CMOS and MEMS fabrication. Due to its low cost and low temperature processing, PAA is an excellent candidate for silicon system integration. At first, PAA is explored as an isolation layer between the inductor and the lossy silicon substrate. Simulations show that although the dielectric constant of the PAA is tunable, the stress produced by the required thicker layers is problematic. Nevertheless, the use of PAA as a MEMS material shows much more promise. Tunable RF MEMS inductors based on bimorph sandwich layer of aluminum PAA and aluminum are fabricated and tested. A tuning range of 31% is achieved for an inductance variation of 5.8 nH to 7.6 nH at 3 GHz. To further improve the Q, bimorph layers of gold and PAA are fabricated on Alumina substrates. A lower tuning range is produced; however the quality factor performance is greatly improved. A peak Q of over 30 with a demonstrated 3% tuning range is presented. Depending on the need for either high performance or tunability, two types of tunable RF MEMS inductors are presented. Although PAA shows promise as a mechanical material for MEMS, the processing parameters (mainly stress and loss tangent) need to be improved if used as an isolation layer. To our knowledge, this is the first time this material has been proposed and successfully used as a structural material for MEMS devices and CMOS processes.

In today’s communication devices, the need for high performance inductors is increasing as they are extensively used in RF integrated circuits (RFICs). This need is even more pronounced for variable inductors as they are widely required in tunable filters, voltage controlled amplifiers (VCO) and low noise amplifiers (LNA). For RFICs, the main tuning elements are solid state varactors that are used in conjunction with invariable inductors. However, they have limited linearity, high resistive losses, and low self resonant frequencies. This emphasizes the need for developing another tuning element that can be fabricated monolithically with ICs and can offer high range of tuning. Due to the ease of CMOS integration and low cost silicon based IC fabrication, the inductors currently used are a major source of energy loss, therefore driving the overall quality factor and performance of the chip down. During the last decade there has been an increase in research in RF MicroelectroMechanical Systems (RF MEMS) to develop high quality on chip tunable RF components. MEMS capacitors were initially proposed to substitute the existing varactors, however they can not be easily integrated on top of CMOS circuits. RF MEMS variable inductors have recently attracted attention as a better alternative. The research presented here explores using porous anodic alumina (PAA) in CMOS and MEMS fabrication. Due to its low cost and low temperature processing, PAA is an excellent candidate for silicon system integration. At first, PAA is explored as an isolation layer between the inductor and the lossy silicon substrate. Simulations show that although the dielectric constant of the PAA is tunable, the stress produced by the required thicker layers is problematic. Nevertheless, the use of PAA as a MEMS material shows much more promise. Tunable RF MEMS inductors based on bimorph sandwich layer of aluminum PAA and aluminum are fabricated and tested. A tuning range of 31% is achieved for an inductance variation of 5.8 nH to 7.6 nH at 3 GHz. To further improve the Q, bimorph layers of gold and PAA are fabricated on Alumina substrates. A lower tuning range is produced; however the quality factor performance is greatly improved. A peak Q of over 30 with a demonstrated 3% tuning range is presented. Depending on the need for either high performance or tunability, two types of tunable RF MEMS inductors are presented. Although PAA shows promise as a mechanical material for MEMS, the processing parameters (mainly stress and loss tangent) need to be improved if used as an isolation layer. To our knowledge, this is the first time this material has been proposed and successfully used as a structural material for MEMS devices and CMOS processes.