The shrinkage during solidification of aluminium and iron based alloys has been studied experimentally and theoretically. The determined shrinkage behaviour has been used in theoretical evaluation of shrinkage related phenomena during solidification. Air gap formation was experimentally studied in cylindrical moulds. Aluminium based alloys were cast in a cast iron mould while iron based alloys were cast in a water-cooled copper mould. Displacements and temperatures were measured throughout the solidification process. The modelling work shows that the effect of vacancy incorporation during the solidification has to be taken into account in order to accurately describe the shrinkage. Crack formation was studied during continuous casting of steel. A model for prediction of crack locations has been developed and extended to consider non-equilibrium solidification. The model demonstrates that the shrinkage due to vacancy condensation is an important parameter to regard when predicting crack formation. The centreline segregation was studied, where the contributions from thermal and solidification shrinkage were analysed theoretically and compared with experimental findings. In order to compare macrosegregation in continuous casting and ingot casting, ingots cast with the same steel grade was analysed. However, the macrosegregation due to A-segregation is driven by the density difference due to segregation. This is also analysed experimentally as well as theoretically.

Further increase of erbium concentrations in Er-doped amplifiers and lasers is needed for the design of efficient, reliable, compact and cost-effective components for telecommunications and other photonic applications. However, this is hindered by Er concentration dependent loss mechanism known as upconversion. The upconversion arises due to non-radiative energy transfer (ET) interactions (migration and energy-transfer upconversion) among the Er ions exited to the metastable level that is used for amplification. The upconversion deteriorates the conversion efficiency of Er doped gain medium and may even totally quench the gain. The upconversion can be significantly intensified if the Er distribution in glass is non-uniform, which can be minimized by optimizing the fabrication process and the glass composition. The optimization requires detailed characterization techniques capable to distinguish between the effects caused by the uniformly distributed ions (homogeneous upconversion, HUC) and non-homogeneously distributed ions (pair induced quenching, PIQ) The thesis deals with rigorous statistical modeling of the HUC and development of experimental methods that can provide accurate and detailed data about the upconversion, which are needed for the characterization of the upconversion. The presented model interprets the homogenous upconversion as an interplay of ET interactions between randomly distributed Er ions, which is affected by stimulated emission/absorption of the radiation propagating in the medium. The model correspondingly uses the ET interactions parameters as the main modeling parameters. The presented analytical model is verified by Monte-Carlo simulations. It explains strongly non-quadratic character of the upconversion observed in experiments and variety of the associated effects. The model is applicable to the interpretation of the upconversion measurements in various experimental conditions, which facilitates the upconversion characterization. The thesis also presents an advanced experimental method for accurate and detailed characterization of the upconversion in both continues-wave pumping conditions and during the decay of Er population inversion. Using the method the upconversion modeling is experimentally verified by correlating the measurements results with the modeling predictions in the whole range of the practical Er doping levels. This also allows to estimate the parameters for the ET interactions in silica. Finally, it is shown that the presented method can serve as a basis for discrimination of HUC and PIQ effects, which is crucial for optimizing the fabrication process and the glass composition.

SiC MESFETs and GaN HEMTs have an enormous potential in high-power amplifiers at microwave frequencies due to their wide bandgap features of high electric breakdown field strength, high electron saturation velocity and high operating temperature. The high power density combined with the comparably high impedance attainable by these devices also offers new possibilities for wideband power microwave systems. In this thesis, Class C switching response of SiC MESFET in TCAD and two different generations of broadband power amplifiers have been designed, fabricated and characterized. Input and output matching networks and shunt feedback topology based on microstrip and lumped components have been designed, to increase the bandwidth and to improve the stability. The first amplifier is a single stage 26-watt using a SiC MESFET covering the frequency from 200-500 MHz is designed and fabricated. Typical results at 50 V drain bias for the whole band are, 22 dB power gain, 43 dBm output power, minimum power added efficiency at P 1dB is 47 % at 200 MHz and maximum 60 % at 500 MHz and the IMD3 level at 10 dB back-off from P 1dB is below ‑45 dBc. The results at 60 V drain bias at 500 MHz are, 24.9 dB power gain, 44.15 dBm output power (26 W) and 66 % PAE. In the second phase, two power amplifiers at 0.7-1.8 GHz without feed back for SiC MESFET and with feedback for GaN HEMT are designed and fabricated (both these transistors were of 10 W). The measured maximum output power for the SiC amplifier at Vd = 48 V was 41.3 dBm (~13.7 W), with a PAE of 32 % and a power gain above 10 dB. At a drain bias of Vd= 66 V at 700 MHz the Pmax was 42.2 dBm (~16.6 W) with a PAE of 34.4 %. The measured results for GaN amplifier are; maximum output power at Vd = 48 V is 40 dBm (~10 W), with a PAE of 34 % and a power gain above 10 dB. The SiC amplifier gives better results than for GaN amplifier for the same 10 W transistor. A comparison between the physical simulations and measured device characteristics has also been carried out. A novel and efficient way to extend the physical simulations to large signal high frequency domain was developed in our group, is further extended to study the class-C switching response of the devices. By the extended technique the switching losses, power density and PAE in the dynamics of the SiC MESFET transistor at four different frequencies of 500 MHz, 1, 2 and 3 GHz during large signal operation and the source of switching losses in the device structure was investigated. The results obtained at 500 MHz are, PAE of 78.3%, a power density of 2.5 W/mm with a switching loss of 0.69 W/mm. Typical results at 3 GHz are, PAE of 53.4 %, a power density of 1.7 W/mm with a switching loss of 1.52 W/mm. / Report code: LIU-TEK-LIC-2008:32

Alumina-Zirconia thin films were grown on a range of substrates using dual magnetron sputtering. Film growth was achieved at a relatively low temperature of 450 °C and at higher temperatures up to 810 °C. The films were grown on well-defined surfaces such as silicon (100) but also on industrially relevant substrates such as hardmetal (WC-Co). Radio frequency power supplies were used in combination with magnetron sputtering to avoid problems with target arcing. A range of film compositions were possible by varying the power on each target. The influence of sputtering target were investigated, both ceramic oxide targets and metallic targets being used. The phase composition of the as-deposited films was investigated by x-ray diffraction. The pure zirconia films contained the monoclinic zirconia phase, while the pure alumina films appeared either amorphous or contained the gamma-alumina phase. The composite films contained a mixture of amorphous alumina, gamma-alumina and the cubic zirconia phase. In-depth high-resolution electron microscopy studies revealed that the microstructures consisted of phase-separated alumina and zirconia nanocrystals in the case of the nanocomposites. In-situ spectroscopy was also performed to characterise the nature of the bonding within the as-deposited films. The oxygen stoichiometry in the films was investigated as a possible reason for the stabilisation of the cubic zirconia phase in the nanocomposite. Ion beam techniques such as Rutherford backscattering scattering and electron recoil detection analysis were used in these studies. The growth of films with ceramic targets led to films that may be slightly understoichiometric in oxygen, causing the phase stabilisation. The growth of films from metallic targets necessitates oxygen rich plasmas and it is not expected that such films will be oxygen deficient. Initial attempts were also made to characterise the mechanical properties of the new material with nanoindentation. The nanocomposite appeared to have greater resistance to wear than the pure zirconia film. In doing so, some surface interactions and some material interactions have been studied. / Report code: LIU-TEK-LIC-2006:41

Large improvements in cutting tool design and technology, including the application of advanced surface engineering treatments on the cemented carbide insert, have been achieved in the last decades to enhance tool performance. However, the problem of improving the tool body material is not adequately studied. Fatigue is the most common failure mechanism in cutting tool bodies. Rotating tools, tool going in and out of cutting engagement, impose dynamic stresses and require adequate fatigue strength of the tool. Working temperatures of milling cutter bodies in the insert pocket can reach up to 600°C depending on the cutting conditions and material of the workpiece. As a result, steel for this application shall have good hot properties such as high temper resistance and high hot hardness values to avoid plastic deformation in the insert pocket of the cutting tool. Machinability of the steel is also essential, as machining of steel represents a large fraction of the production cost of a milling cutter. This thesis focus on the improvement of the cutting tool performance by the use of steel grades for tool bodies with optimized combination of fatigue strength, machinability and properties at elevated temperatures. The first step was to indentify the certain limit of the sulphur addition for improved machinability which is allowable without reducing the fatigue strength of the milling cutter body below an acceptable level. The combined effect of inclusions, surface condition and geometrical stress concentrator on the fatigue life of the tool steel in smooth specimens and in tool components were studied in bending fatigue. As the fatigue performance of the tools to a large extent depends on the stress relaxation resistance at elevated temperature use, the second step in this research was to investigate the stress relaxation of the commonly used milling cutter body materials and a newly steel developed within the project. Compressive residual stresses were induced by shot peening and their response to mechanical and thermal loading as well as the material substructures and their dislocation characteristics were studied using X-ray diffraction. Softening resistance of two hot work tool steels and a newly developed steel was investigated during high temperature hold times and isothermal fatigue and discussed of with respect to their microstructure. Carbide morphology and precipitation as well as dislocation structure were determined using transmission electron microscopy and X-ray line broadening analysis.

A thorough investigation was performed on the behaviour of Ni, Fe and Cu oxides dissolved in cryolite melts, and the solubility of these species was measured as a function of alumina content, NaF/AlF3 molar ratio (CR) and temperature. Predominance area diagrams showing the solid phases containing Ni, Fe and Cu, respectively, as a function of the partial oxygen pressure and the alumina activity at 1020 oC were constructed. These diagrams were based on present emf and solubility measurements. The interpretations of the solubility measurements for the oxides of Ni and Fe gaveconclusive and consistent results. The oxides of Ni and Fe exhibit decreasing solubility with decreasing temperature and with increasing alumina concentration. The Ni(II) concentration decreased from 0.32 wt% in cryolite to 0.003 wt% in alumina-saturated melts, while that of Fe(II) decreased from 4.17 to 0.32 wt% in similar melts. FeO and NiO are stable solid phases at low alumina concentrations, while FeAl2O4 and NiAl2O4 are stable at high concentrations. The alumina concentrations corresponding to the points of coexistence between FeO and FeAl2O4 and between NiO and NiAl2O4 were determined to be 5.03 and 3.0 wt% Al2O3, respectively, corresponding to the following Gibbs energy of formation from the oxide compounds,∆G0fNiAl2o4 = –28.6 ± 2 kJ/mol and ∆G0f FeAl2O4 = –17.6 ± 0.5 kJ/mol. The solubilities of FeAl2O4 and NiAl2O4 as a function of the CR were investigated in alumina-saturated melts at 1020 oC. For both compounds a maximum solubility was found at CR ~5, being 0.008 wt% Ni(II) and 0.62 wt% Fe(II). The results are discussed with respect to the species present in solution. Both Fe(II) and Ni(II) dissolve as fluorides with different numbers of associated “NaF’s”. Ni(II) seems to form Na3NiF5 in melts with molar ratios 2 to 12, while Fe(II) is present as NaFeF3 in acidic (CR 3–10) melts and as Na3FeF5 and probably some Na4FeF6 in basic melts (CR > 3). The solubility of both Cu oxidation states Cu(I) and Cu(II) decreases with decreasing temperature. The solubilities of Cu(I) initially decreased with increasing alumina concentration, showing a minimum at a certain alumina concentration followed by an increase. The solubilities were 0.36 wt% Cu(I) and 0.92 wt% Cu(II) in cryolite, and 0.30wt% Cu(I) and 0.45 wt% Cu(II) in alumina-saturated cryolite at 1020 oC. At 1020 oC the solubilities of Cu2O and CuO were little influenced when changing the CR from 3 to 8 in alumina-saturated melts (~0.30 wt% Cu(I) and ~0.45 wt% Cu(II)), but there was an upward trend for CR < 3. Solubility measurements for CuO in alumina-saturated melts at CR 3.0 to 1.2 clearly showed that the saturation concentration is dependent on both temperature and melt composition. Copper ions in solution show a complex behaviour, since they form fluorides and oxycomplexes simultaneously. The extent of co-existence of Cu(I) and Cu(II) in the same melt is also considerable, and it is depending on the alumina activity in the melt. According to thermodynamics the stable copper oxide phases at high alumina activities are the aluminates CuAlO2 and CuAl2O4. However, no clear changes in the solubilities were found for the points of coexistence between Cu2O and CuAlO2 and CuO and CuAl2O4, respectively, as was the case for Ni(II) and Fe(II). Although there are uncertainties regarding the thermodynamic data available for the formation of copper aluminates, models for the dissolution mechanisms and for the species present in the melt are suggested. Cu(I) seems to form mainly CuF at low alumina contents, while Na5CuO3 dominates at higher alumina concentrations. Likewise, Cu(II) seems to form CuF2, but the concentration of CuF2 decreases with increasing alumina content. The species that gave the best fit for the cupric oxy-complexes was Na16CuO9, and the amount increased with increasing alumina content. Cermet anodes were prepared with a NiFe2O4-based oxide phase mixed with a ~20 wt% copper-rich metal phase. The electrical conductivity for these materials was measured as a function of temperature, showing semiconductor behaviour in the temperature range from room temperature to 1050 oC. The highest electrical conductivity measured was ~30 S/cm at 1000 oC, which is on the low side for use as an anode material for aluminium production. Three cermet anodes were tested by electrolysis for 48 hours. After the experiments the anodes were examined with SEM. There was no metal phase present in the outer 100 µm of the anode, not even pores were observed that could indicate where the metal grains had been. A copper-rich phase was found in one case ~2 mm from the outer surface, and it is believed that copper diffuses out of the anode. The cermet anodes dissolved slowly in the electrolyte during electrolysis. The steady state concentrations of Fe and Cu in the electrolyte were below the saturation concentrations, while the concentration of Ni was 3 - 4 times above saturation. The dissolution of the anode does not fit a first order mass-transport model, but it can probably be explained by a controlled dissolution mechanism with some additional disintegration/spalling of the anode material. Further work is needed to draw a firm conclusion. In general, correct solubility data for the anode constituents are needed to make a proper evaluation of various anode materials. Perhaps the first order mass-transport model agrees for some materials, but based on the present results it seems untenable for cermet materials made of NiFe2O4 with a copper-rich metal phase. The solubilities of the oxides of Ni(II) and Fe(III) are very low for the alumina-saturated melt used during electrolysis, which make them promising candidates for inert anodes. However, if nickel aluminate, which is an insulator, is formed and deposited on the anode surface, it is a cause of concern. Fe(II) aluminate is not expected to form on the anode surface, since Fe(III) is the stable oxidation state in the presence of oxygen gas. However, solid Fe(II) aluminate may be formed in the bulk of the electrolyte where the partial oxygen pressure is lower.

The aim of this thesis was to develop wear resistant physical vapour deposited coatings of TiB2 as well as multilayers of Cr/CrN. The correlation between deposition parameters and fundamental coating properties such as microstructure, composition, residual stress and hardness has been investigated. Finally, the influence of these properties on the coating behaviour in tribological applications has been evaluated. It is shown that the use of electron bombardment of the growing coating during d.c. magnetron sputtering is beneficial for the growth of superhard TiB2 coatings. Furthermore, electron bombardment results in TiB2 coatings with significantly lower residual stresses than coatings deposited using ion bombardment. The low stresses in these coatings open up the possibility to deposit thicker PVD coatings, as confirmed in this thesis. In addition, the use of TiB2 coatings in tribological contacts against aluminium proved to be superior to many other commercial coatings used today, with respect to wear resistance, anti galling properties and a low friction. Finally, a model is proposed which explains the observation that the abrasive wear resistance of multilayered Cr/CrN coatings can outperform that of the individual constituents. The model was found to satisfactory predict experimental data.

Direct semiconductor wafer bonding has emerged as a technology to meet the demand foradditional flexibility in materials integration. The applications are found in microelectronics, optoelectronics and micromechanics. For instance, wafer bonding is used to produce silicon-on-insulator (SOI) wafers. Wafer bonding is also interesting to use for combining dissimilar semiconductors, such as Si and InP, with different dictated optical, electronic and mechanicalproperties. This enables a completely new freedom in the design of components and systems, e.g. for high performance optoelectronic integrated circuits (OEIC). Although wafer bonding has proved to be a useful and versatile tool, the high temperature annealing that is needed to achieve reliable properties sometimes hampers its applicability. Therefore, low temperature wafer bonding procedures may further qualify this technology. In the present thesis, low temperature wafer bonding procedures using oxygen plasma surface activation have been studied. A specially designed fixture was adopted enabling in situ oxygen plasma wafer bonding. Oxygen plasma surface activation was seen to indeed yield high Si-Si bonding-strength at low temperatures. Here, the optimisation of the plasma parameters was shown to be the key to improved results. Furthermore, dependence of wafer bonded Si p-n junctions on the annealing temperature was investigated. InP-to-Si wafer bonding is also presented within this thesis. High temperature annealing was seen to induce severe material degradation. However, using oxygen plasma assisted wafer bonding reliable InP-to-Si integration was achieved already at low temperature, thereby circumventing the problems associated with the lattice and thermal mismatch that exist between these materials. As a result, low temperature InP-based epitaxial-layer transferring to Si could be presented. Finally, high-quality SiO2 insulator on InP and Si was realised at low temperatures. It is concluded that low temperature oxygen plasma assisted wafer bonding is an interesting approach to integrate dissimilar materials, for a wide range of applications.

The present work has been carried out in order to investigate Al-Mg-Si alloys that are deformed at moderate temperatures. These temperatures are in the range between 200 C and 300 C. Also some experiments are performed at room temperatures. Two deformation models have been applied in the experiments: material deformation by compression testing and by forward extrusion. The investigated alloys are AA6063, AA6082 and an alloy that is named “Alloy R” in this work. The latter alloy is the industrial alloy AA6082 without the Mn-addition (0.56wt%Mn in the AA6082). The “R” denotes the recrystallized microstructure in the material after hot forming operations. The investigations show the effect of changing the temperature in the given temperature interval on the stress-strain relationship for each alloy. From the compression testing, it is found that none of the alloys AA6063 or Alloy “R” reaches a steady state condition as true strain approaches 0.8 for deformation temperatures between 200 C and 250 C. At compression testing performance at 300 C, the alloy “R” reaches a steady state condition at a true strain equal to 0.4. As true stress-true strain relationship has been investigated for the “Alloy R” and the AA6063 at comparable deformation parameters, it is shown that the alloy “R”, with the highest Si-content, requires the highest true stress for a given true strain value (AA6063: 0.45wt%Si, Alloy “R”: 0.87wt%Si). From the compression testing, the effect of Mn on the material properties in the AA6082-alloy has been determined. For the Alloy “R” and the AA6082, the true stress reached the same value after a certain amount of deformation. As deformation temperature increases, this common value of true stress corresponds to a decrease in true strain. The AA6082 and Alloy “R” are also compared in experiments performed in forward extrusion. One observes that for the same deformation temperature and at identical die diameters, the ram force is identical. It is worth noticing that these alloys did not show the same relationship during the compression testing at low values of true strain (<0.8). On a microscopic scale, one concludes that Mn has no significant effect on the stress-strain relationship for the applied deformation parameters in the forward extrusion equipment. Hardness measurements indicate that the age hardening potential in the extruded test specimen decreases as the deformation temperature increases. The hardness data is similar for both the AA6082 and the Alloy R, thus indicating that the Mn content has no significant effect on the strength of the material. The deformed material has been annealed in order to investigate the recrystallization process in the AA6082 and the Alloy “R”. The recrystallization grain size in the Alloy “R” is significantly larger than in the AA6082 at comparable deformation parameters after annealing at 530 C for 15 minutes. This result is due to the effect of Mn-containing dispersoids in the AA6082. The recrystallization grain size in the Alloy “R” seems to be unaffected by the deformation temperature after annealing for 15 minutes. The observation of the AA6082 is quite different. A small increase in grain size is observed for both reduction ratios as the deformation temperature is elevated from 20C to 200 C and further to 250 C. At extrusion temperatures of 300 C the recrystallization grains are significantly larger. Annealing experiments performed at 430 C on the AA6082 indicates that a change in the deformation temperature from 200 C to 250 C does not affect the amount of stored energy in the material significantly. The Forge2 programme has been used to perform numeric simulations of the forward extrusion experiment. From this the temperature distribution, strain rate variation and true strain development in the test piece had been investigated. As the simulated true strain values are compared to the grain size in the annealed material, the recrystallization grain size is related to the amount of stored energy in the material in a very convincing way. It is also shown that the recrystallization grain diameter is related to the amount stored energy as the grain diameter is investigated in the radial and the extrusion direction separately.

The work presented in this thesis is concerned with the microstructure of zinc oxide varistor materials used in surge protecting devices. This class of material has been characterised with special emphasis on the functional microstructure and the development of the microstructure during sintering. Several different techniques have been used for the analysis, especially scanning electron microscopy (SEM) in combination with electron beam-induced current (EBIC) analysis and in-situ studies of heat-treatment experiments and transmission electron microscopy (TEM) in combination with energy dispersive X-ray spectrometry (EDS) and electron holography. Detailed TEM analyses using primarily centred dark-field imaging of grain boundaries, especially triple and multiple grain junctions, were used to reveal the morphological differences between the various Bi2O3 phases. The triple and multiple grain junctions were found to exhibit distinct differences in morphology, which could be attributed the difference in structure of the crystalline Bi2O3 polymorphs present in the junctions. Electrical measurements were performed on individual ZnO/ZnO grain boundaries using EBIC in the SEM. The EBIC signal was found to depend strongly on the geometric properties of the interface and also on the symmetry of the depletion region at the interface. A symmetric double Schottky barrier was never observed in the experiments, but instead barriers with clear asymmetry in the depletion region. Experimental results together with computer simulations show that reasonably small differences in the deep donor concentrations between grains could be responsible for this effect. Electron holography in the TEM was used to image the electrostatic potential variation across individual ZnO/ZnO interfaces. The sign of the interface charge, the barrier height (about 0.8 eV) and the depletion region width (100 to 150 nm) were determined from holography data. Asymmetries of the depletion region were also found with this technique. The full sintering process of doped ZnO powder granules was studied in-situ in the environmental SEM. The densification and grain growth processes were studied through the sintering cycle. The formation of a functional microstructure in ZnO varistor materials was found to depend strongly on the total pressure.