Ferroelectric materials are characterized by a reversible spontaneous polarization in the absence of an electric field. The characteristic polarization response of a ferroelectric material to an applied electric field enables a binary state device in the form of a thin film ferroelectric capacitor that can be used to store digital information. This provides opportunities for the development of high speed, low cost and low power, nonvolatile memory devices. The development of commercial ferroelectric memory devices has however been hampered by (a) several reliability issues including fatigue, leakage current, aging, time dependent dielectric breakdown, retention and imprint and (b) processing problems including the development of a low temperature thin film deposition process and the development of a patterning technology. Lead zirconate titanate (PZT) is now widely considered as the most promising material for ferroelectric memory applications as a result of its excellent ferroelectric properties and wide operating temperature range. However, it is commonly found that metal electroded-PZT capacitors (e.g., Pt/PZT/Pt) show a loss of switchable polarization with cumulative switching cycles. This phenomenon is known as fatigue and is the one of the critical problems affecting the lifetime of ferroelectric memories. This research is primarily focused on the problem of fatigue. On the basis of a quantitative model, various guidelines to minimize the degradation problem have been derived. The model attributes fatigue to domain pinning by space charge that is caused by defect (e.g. oxygen vacancy) entrapment at various interface sites such as electrode-ferroelectric interface, domain boundaries and grain boundaries. Two different approaches to minimize the problem have been outlined : (a) control of the defect density and (b) control of the interface state. The control of interface state was achieved by replacing the metal electrodes with conducting oxide electrodes such as RuO₂. The oxide electrode/PZT capacitors were characterized for their diffusion barrier properties, perovskite phase formation, interface nature and ferroelectric properties. The results indicate that these oxide electroded PZT films are good candidates for nonvolatile memory applications. However, the leakage current levels at the operating voltages are far higher than the metal counterparts. Simultaneous minimization of fatigue and leakage current in PZT films was achieved by using multilayer metal/conducting oxide electrodes (e.g., Pt/RuO₂). The control of defect density was attained by (a) donor doping to compensate for the oxygen vacancies (e.g, Nb doping in PZT) and (b) utilizing ferroelectric materials that have a low intrinsic defect concentration. As a result of the latter approach, novel ferroelectric materials belonging to the layer-structure family of oxides have been identified as excellent candidates for fatigue free nonvolatile memory applications. Laser ablated SrBi₂(Ta_xNb_{1 - x})₂O₉ (0<x<1) films showed very good hysteresis characteristics (remnant polarization value of 11 µC/cm², coercive field of 60 kV/cm), no fatigue was observed up to 10⁹ switching cycles and very low leakage current densities. Furthermore, the formation and properties of these films were characterized. It was found grain size and orientation played a major role in determining the properties of these films. C-axis oriented films were found to exhibit almost no polarization. An additional objective of this research was to identify an etching technology (process integration issue) for patterning of the ferroelectric capacitors. The etching process should provide high etch rates, good etch anisotropy, high etch selectivity and minimal post etch residues. It has been shown that a reactive ion etch process with CCl₂F₂/O₂ as the etch gas mixture can meet these requirements. A detailed process study has been conducted to determine the mechanism of etching. / Ph. D.

Transition metal carbides (TMCs) are of increasing interest for catalytic processes. Their performance and stability under common oxidative conditions in catalytic reactions are crucial for several applications, including catalysis and electrochemical reactions. In this work, we report a detailed XPS study of the interactions of stoichiometric NbC and TaC surfaces with common oxidizing agents like O2 and H2O, which are important media in many chemical processes. Experimental results showed that NbC reacts with O2 to produce Nb sub-oxrides, while TaC is inert to O2 exposure. TaC surfaces are more sensitive to H2O vapor, with a greater surface oxidation and hydroxylation. Atmospheric oxidation of NbC and TaC was also studied, and results showed that both films oxidized yielding to the formation of Nb2O5 and Ta2O5, hydroxylated/ oxide carbon species, and some adventurous carbon build-up. TMCs are catalytically active in many reactions, especially those involving electrochemical nitrogen reduction reactions (NRR) to ammonia. Experimental and DFT calculations were used to provide insight on how carbide surface structures change electrochemically and how that evolution relates to NRR activity. Results showed that NbC has NRR activity at pH 3.2 after immersion in 0.3 M NaOH, leaving niobium suboxides. However, photoemission data showed that the Nb2O5 overlayer is restored after polarization to -1.3 V vs. Ag/AgCl, inhibiting NRR activity. TaC, on the other hand, is inactive for NRR at potentials more positive than -1.0 V, as NaOH treatment fails to remove the Ta2O5 surface layer induced by ambient exposure. The study also found that the formation and stabilization of intermediate oxidation states on the surface of transition metal ions are crucial for N≡N bond activation and NRR activity.

The Motzfeldt centre is one of four major alkaline centres belonging to the Igaliko complex of South Greenland. The melts parental to the Motzfeldt centre are interpreted from Hf isotopes to be derived form a common mantle source which experienced subsequent isotopic contamination from older crustal components during the interval between segregation and emplacement. Magmatism within the centre commenced with the emplacement of the Motzfeldt Sø Formation at 1273 ± 8 Ma. This unit is unique within the Motzfeldt intrusion as it is characterised by a high degree of textural and mineralogical variability and hosts localised Nb, Ta, U, Th, Zr and REE mineralisation associated with pyrochlore and late-stage REE bearing carbonate phases. Biotite halogen contents show that in addition to enrichment of incompatible elements the MSF and Motzfeldt centre in general is particularly rich in F. The elevated F content is inferred to have extended the crystallisation interval of the melt and facilitated fractionation down to relatively low temperatures. The unusual enrichment of F and incompatible elements in the MSF is suggested to represent the first and most evolved melts extracted from the top of a stratified storage chamber at depth. The MSF is also characterised by pervasive subsolidus alteration, giving the rock and region a striking brick red colour. Pb-Pb pyrochlore studies indicate that alteration in the formation was effectively synchronous (1267 ± 6 Ma), with the magmatic age of emplacement. Fluid inclusion studies suggest that contemporaneous to the exsolution of juvenile, high salinity, F-rich fluids was the wholesale influx of hydrothermally convected low salinity groundwaters through the formation. The presence of pervasive late-stage hematite and calcite throughout the MSF suggests that the oxidation potential of the bulk fluid increased above the hematite-magnetite buffer during the waning stages of the hydrothermal phase. Mineralisation was promoted by this shift in fluid composition, reducing the complexing potential of fluid ligands and facilitating mineralisation within the high-levels units of the intrusion where alteration is most intense. Economic mineralisation associated with the centre is inferred to be largely sourced from the parental melts, however the role the hydrothermal phase played was particularly important in locally mobilising and concentrating incompatible elements within the high-level units of the formation.

<p>Many materials exhibit interesting and novel properties when prepared as thin films. Thin film metal oxides have had an impact on the technological progress of the microelectronics mainly due to their electrical and optical properties. Since the future goes towards the nanometre scale there is an increasing demand for thin film deposition processes that can produce high quality metal oxide films in this scale with high accuracy.</p><p>This thesis describes atomic layer deposition of Ta₂O₅, HfO₂ and SnO₂ thin films and chemical vapour deposition of SnO₂ thin films. The films have been deposited by employing metal iodides and oxygen as precursors. All these processes have been characterised with regards to important processing parameters. The films themselves have been characterised by standard thin film analysing techniques such as x-ray diffraction, scanning electron microscopy, atomic force microscopy and transmission electron microscopy. The chemical and physical properties have been coupled to critical deposition parameters. Furthermore, additional data in the form of electrical and gas sensing properties important to future applications in the field of microelectronics have been examined.</p><p>The results from the investigated processes have shown the power of the metal iodide based atomic layer deposition (ALD) and chemical vapour deposition (CVD) processes in producing high quality metal oxide thin films. Generally no precursor contaminations have been observed. In contrast to metal chloride based processes the metal iodide processes produces films with a higher degree of crystalline quality when it comes to phase purity, roughness and epitaxy. The use of oxygen as oxidising precursor allowed depositions at higher temperatures than normally employed in water based ALD processes and hence a higher growth rate for epitaxial growth was possible.</p>

Many materials exhibit interesting and novel properties when prepared as thin films. Thin film metal oxides have had an impact on the technological progress of the microelectronics mainly due to their electrical and optical properties. Since the future goes towards the nanometre scale there is an increasing demand for thin film deposition processes that can produce high quality metal oxide films in this scale with high accuracy. This thesis describes atomic layer deposition of Ta2O5, HfO2 and SnO2 thin films and chemical vapour deposition of SnO2 thin films. The films have been deposited by employing metal iodides and oxygen as precursors. All these processes have been characterised with regards to important processing parameters. The films themselves have been characterised by standard thin film analysing techniques such as x-ray diffraction, scanning electron microscopy, atomic force microscopy and transmission electron microscopy. The chemical and physical properties have been coupled to critical deposition parameters. Furthermore, additional data in the form of electrical and gas sensing properties important to future applications in the field of microelectronics have been examined. The results from the investigated processes have shown the power of the metal iodide based atomic layer deposition (ALD) and chemical vapour deposition (CVD) processes in producing high quality metal oxide thin films. Generally no precursor contaminations have been observed. In contrast to metal chloride based processes the metal iodide processes produces films with a higher degree of crystalline quality when it comes to phase purity, roughness and epitaxy. The use of oxygen as oxidising precursor allowed depositions at higher temperatures than normally employed in water based ALD processes and hence a higher growth rate for epitaxial growth was possible.

The availability of high quality thermodynamic data on solid solutions and compounds present in multicomponent systems assists in optimizing processing parameters for synthesis, and in evaluating stability domains and materials compatibility under different conditions. Several oxide systems of technological interest, for which thermodynamic data was either not available or is inconsistent were selected for study. Thermodynamic properties of phases present in the binary systems Nb-O and Ta-O were measured in the temperature range from 1000 to 1300 K using solid state electrochemical cells based on (Y2O3) ThO2 as the electrolyte. Based on these measurements and more recent data on heat capacity and phase transitions reported in the literature, Gibbs energy of formation for NbO, NbO2, NbO2.422, Nb2O5-x and Ta2O5 were reassessed. Significant improvements in the data for NbO2, Nb2O5 and Ta2O5 are suggested. The pseudo binary system MoO2-TiO2 was investigated because of the inconsistency between the phase diagram and thermodynamic properties of the solid solution reported in the literature. Based on new electrochemical measurements, a new improved phase diagram for the system MoO2-TiO2, incorporating recently discovered monoclinic to tetragonal phase transition in MoO2 at 1533 K, is presented. Isothermal section of the phase diagram for the ternary systems Cr-Rh-O and Ta-Rh-O and thermodynamic properties of ternary oxides CrRhO3 and TaRhO4 were measured for the first time in the temperature range from 900 to 1300 K. Phase relations for these systems have been computed as a function of oxygen potential at fixed temperature and as a function of temperature at selected oxygen partial pressures. Metal-spinel-corundum three-phase equilibrium in the Ni-Al-Cr-O system at 1373 K has been explored because of its relevance to high temperature corrosion of super alloys. The Gibbs energy of mixing of spinel solid solution was derived from the tie-line data and is compared with the values calculated from cation distribution models. An oxygen potential diagram is developed for the decomposition of spinel solid solution to nickel and corundum solid solution at 1373 K under reducing conditions. The high temperature thermodynamic properties of the phases present in quaternary systems Ca-Co-Al-O and Ca-Cu-Ti-O have been measured by solid state electrochemical cells based on stabilized zirconia. Gibbs energies of formation of the quaternary oxides Ca3CoAl4O10 in the temperature range from 1150 to1500 K and CaCu3Ti4O12 in the range from 900 to 1350 K are presented. Chemical potential diagrams have been computed for the system Al2O3-CaO-CoO at 1500 K. The oxygen potential corresponding to the decomposition of the complex perovskite CaCu3Ti4O12 (CCTO) has been calculated as a function of temperature from the emf of the cell. The effect of the oxygen partial pressure on the phase relations in the pseudo-ternary system CaO-(CuO/Cu2O)-TiO2 at 1273 K has been evaluated. The phase diagrams are useful for the control of the secondary phases that form during synthesis of CCTO, a material exhibiting colossal dielectric response.

An intracranial aneurysm can be a serious, life-threatening condition which may go undetected until the aneurysm ruptures causing hemorrhaging within the brain. The typical treatment method for large aneurysms is by embolization using platinum coils. However, in about 15% of the cases treated by platinum coils, the aneurysm eventually re-opens. The solution to the problem of aneurysm recurrence may be to develop more bio-active materials, including certain polymers, to use as coil implants. In this research, a shape memory polymer (SMP) was investigated as a potential candidate for aneurysm coils. The benefit of a shape memory polymer is that a small diameter fiber can be fed through a micro-catheter and then change its shape into a three-dimensional configuration when heated to body temperature. The SMP was tested to determine its thermo-mechanical properties and the strength of the shape recovery force. In addition, composite specimens containing tantalum filler were produced and tested to determine the mechanical effect of adding this radio-opaque metal. Thermo-mechanical testing showed that the material exhibited a shape recovery force a few degrees above Tg. The effects of the metal filler were small and included depression of Tg and recovery force. SMP coils deployed inside a simulated aneurysm model demonstrated that typical hemodynamic forces would not hinder the shape recovery process. The x-ray absorption capability the tantalum-filled material was characterized using x-ray diffractometry and clinical fluoroscopy. Diffractometry revealed that x-ray absorption increased with tantalum concentration, however, not as the rule of mixtures would predict. Fluoroscopic imaging of the composite coils in a clinical setting verified the radio-opacity of the material.

In der vorliegenden Arbeit wird das Wachstum von Ta- und TaN-Schichten auf Si- und SiO_2-Substraten untersucht Die Schichten werden dabei unter technologienahen Bedingungen mittels Magnetron-Sputtern abgeschieden. Die Untersuchungen erfolgen hauptsächlich mit winkelaufgelöster röntgenstrahlungsangeregter Photoelektronenspektroskopie (ARXPS). Die Analysen erfolgen in situ, ohne Unterbrechung des Ultrahochvakuums, um die Deposite vor Oxidation und Kontamination zu schützen. Zur zerstörungsfreien Tiefenprofilanalyse wird ein Quantifizierungsalgoritmus beschrieben und angewandt. Für die Kombination Ta/Si wird die Bildung einer zunächst unvollständigen TaSi_2-Schicht, danach das Aufwachsen von Ta auf diese Zwischenschicht beobachtet. Für die Kombination Ta/SiO_2 wird eine Reduktion des SiO_2-Substrates bei gleichzeitigem Aufwachsen von Ta-Oxiden beobachtet. Auf dem durchmischten Schichtstapel wächst danach Ta auf. Für die Kombination TaN/Si wird die Bildung einer Si-N-Zwischenschicht bei gleichzeitigem Wachsen einer TaN-Schicht beobachtet. Für die Kombination TaN/SiO_2 wird das Aufwachsen einer TaN-Schicht ohne Ausbilung von Zwischenschichten beobachtet. Das Wachstumsverhalten des Ta/Si-Systems wird zusätzlich mit in situ Rastertunnelmikroskopie und -spektroskopie untersucht. Es wurden Untersuchungen zur thermischen Stabilität von abgeschiedenen Schichten an den Systemen Ta/Si und TaN/SiO_2 durchgeführt. Als mögliche Alternative zur winkelaufgelösten XPS wurden Untersuchungen mittels synchrotronstrahlungsangeregter Photoelektronenspektroskopie bei variierter Anregungsenergie durchgeführt.

Die vorliegende Arbeit hat zum Ziel, ausgehend vom aktuellen Stand der Technik, die Möglichkeiten der optische Glimmentladungsspektroskopie (GD-OES) für Tiefenprofilanalysen dünner und dünnster Schichten (Schichtdicken = 1 bis 100 nm) zu bestimmen und geeignete instrumentelle und methodischen Modifikationen vorzuschlagen, um die Einsatzmöglichkeiten weiter auszubauen. Dies gilt gleichermaßen unter Berücksichtigung der Anforderungen des Einsatzes im Routinebetrieb (geringe Bruttoanalysezeit und Reproduzierbarkeit) sowie in der Erforschung und der Entwicklung dünner Schichten (geringe Nachweisgrenzen, hohe Flexibilität zum Beispiel bei den analysierbaren Elementen oder der Leitfähigkeit der Proben, geringe Matrixeffekte, etc.). Während jeder GD-OES Analyse finden drei räumlich und zeitlich getrennte Teilprozesse statt: (A) durch das Zerstäuben der Oberfläche wird die Probe in der lateralen Ausdehnung des Anodendurchmessers in die Tiefe abgetragen und in die atomaren Bestandteile zerlegt (Sputterprozess); (B) in das Plasmagebiet diffundierte Partikel reagieren mit dem Analysegas (i. d. R. Argon), dadurch werden die Atome (und Ionen) der Probe in angeregte Zustände versetzt, im nachfolgenden Relaxationsschritt emittieren diese unter anderem Photonen einer charakteristischen Wellenlänge, die (C) alle in einem Detektionssystem (Mono- bzw. Polychromator oder CCD-Spektrometer) in ihrer Intensität als Funktion der spektralen Wellenlänge und der Zeit erfasst werden. Ein Vorteil der Methode, die niedrige Analysendauer bedingt durch den vergleichsweise hohen Sputterabtrag bewirkt, dass die Analyse dünner Schichten innerhalb weniger – im Extremfall sogar nur innerhalb von Bruchteilen von – Sekunden stattfindet. Dies lässt die Herausforderungen für die Analyse dünner Schichten verstehen. Der unter anderem von den elektrischen Entladungsbedingungen abhängige Sputterprozess und die komplexen Reaktionen im Plasma müssen möglichst unmittelbar (&amp;lt; 50 ms) nach dem Zündvorgang in einen stabilen Zustand übergehen. Einerseits ist dies instrumentell durch eine Anpassung der Steuer- und Regelungstechnik (z. B. Wahl geeigneter Druckregelventile, -sensoren, etc.) gelungen. Andererseits beeinflussen die unvermeidlichen Kontaminationen [Wasser(filme) und Kohlenwasserstoffe], die in das Plasmagebiet diffundieren, negativ die Stabilität die Entladung. Die Hauptstrategie zur Unterdrückung dieser ‚Dreckeffekte’ sind erfolgreich verschiedene Wege der ex-situ (maximalmögliche Reduzierung der Leckrate, Einsatz von Hochvakuumbauteilen, Einführung von Richtlinien zur Vakuumhygiene) und in-situ Dekontamination (aktive Desorptionsminderung durch ein Vorsputtern mit Si) gewählt worden. Erst in der Summe aller apparativen Verbesserungen ist die Voraussetzung für die Verwendung der Glimmentladungsspektroskopie als zuverlässige Methode der Dünnschichtanalytik gegeben. Für die laborpraktischen Arbeiten wurde während der sukzessiven Optimierung des Vakuumsystems als Nebenergebnis ein anwenderfreundlicher Schnelltest zur Charakterisierung des Geräts für Kurz- und Langzeitvergleiche entwickelt. Dieser wertet die Abpumpkurven bzw. Druckanstiegskurven aus. In Abhängigkeit der Bedürfnisse und dem Aufwand des Anwenders lassen sich interessante Parameter, wie das effektive Saugvermögen, eine Zeitkonstante für die Gasabgabe oder die Leckrate IL bestimmen. Die Bandbreite der untersuchten Proben ist dabei ähnlich unterschiedlich, wie die Fragestellungen: leitfähige und nichtleitfähige Proben; Nachweis und Bestimmung von Matrixelementen, Legierungsbestandteilen oder Spuren; Einfach-, Mehrlagen- und Wechselschichten, Oberflächen- und Zwischenschichten; Adsorbate an Ober- und Grenzflächen, Schichtdickenhomogenität als Teil der Qualitätskontrolle, etc. Ein Teil dieser Schichtsysteme sind in dieser Arbeit ausführlicher diskutiert worden. Das Hartstoffschichtsystem TiN gehört mit den Schichtdicken von 0,5 bis 3 µm zwar eher zu den dicken Schichten, wobei besonders der oberflächennahe Bereich (&amp;lt; 100 nm) zuverlässig untersucht wurde (vgl. Kap. 4.1). Mit dem Nachweis und der Quantifizierung von in der Grenzfläche (100 bzw. 1000 nm unter der Oberfläche) zwischen den elektrochemisch (ECD-Cu) und physikalisch (PVD-Cu) abgeschiedenen Cu-Schichten versteckten Adsorbaten bietet GD-OES dem Schichthersteller oder dem Werkstoffwissenschaftler wichtige Informationen, um zum Beispiel gezielt Gefügeänderungen für die Erhöhung der Elektromigrationsresistenz einzustellen. Es wurde einerseits die prinzipielle Machbarkeit und andererseits auch die Grenzen der Methode im Vergleich mit TOF-SIMS gezeigt. Ein weiteres Schichtsystem aus der Mikroelektronik ist im anschließenden Kap. 4.3 Gegenstand der GD-OES Untersuchungen. Dabei wurde nicht nur die Schichtdicken von 10 bis 50 nm dünnen TaN-Barriereschichten, sondern auch die Homogenität der Schichtdicken über einen kompletten 6’’ Wafer bestimmt. Die nachzuweisenden Unterschiede liegen im Bereich von einigen Angström bis zu wenigen Nanometern. Die GD-OES Untersuchungen von TaN zeigen zu Beginn und in der Nähe der Grenzfläche zum Substrat ungewöhnliche Intensitätsverläufe von Ta und N. Erst in Kombination mit anderen oberflächenanalytischen Verfahren (XPS und AES) gelang die Interpretation der Messergebnisse. Aus der Summe aller Argumente wird die Hypothese formuliert, dass sich im Fall des Zerstäubungsprozesses von TaN wegen der großen Unterschiede in den Atommassen ein Vorzugssputtern (engl. preferential sputtering) herausbildet. Bei anderen sputternden Verfahren, z. B. SIMS, ist dieses Phänomen längst bekannt und wurde auch für die Glimmentladungsspektroskopie vermutet. Dies konnte bislang allerdings noch nie beobachtet werden. Rechnungen mit einem Simulationsprogramm für Kollisionsvorgänge aufgrund ballistischer Effekte (TRIDYN) stützen diese Hypothese. Begünstigt wurde die Beobachtung des Vorzugssputterns durch die sauberen Messbedingungen, durch die man in Lage versetzt war, Anfangspeaks klar von Kontaminationspeak zu unterscheiden. Das vorletzte Kapitel 4.4 beschäftigt sich mit Schichten im untersten Nanometerbereich (&amp;lt; 5 nm). Es zeigt sich, dass die wenige Nanometer dicken, natürlichen Oxidschichten deutlich besser analysierbar sind, wenn man die in der Arbeit vorgestellte in-situ Dekontamination durch ein Vorsputtern anwendet. Die GD-OES Untersuchungen an organischen Monolagenschichten in Kap. 4.5 sind Teil einer aktuellen wissenschaftlichen Diskussion innerhalb der weltweiten GD-OES Fachwelt. Die von KENICHI SHIMIZU vorgestellten Ergebnisse konnten am Beispiel von Thioharnstoff mit RF bestätigt und erstmals auch mit einer DC-Entladung gezeigt werden. Das Verfahren der GD-OES kann qualitativ die Existenz von monomolekularen Schichten im Subnanometerbereich nachweisen. Allerdings stellen die Ergebnisse von Substraten mit anderen Molekülen die Interpretation der Intensitäts-Zeitprofile in Frage. Ein anderer Interpretationsansatz wird als Hypothese formuliert, konnte jedoch noch nicht verifiziert werden. Mit den vorgestellten Optimierungen der Messtechnik lassen sich die Möglichkeiten der Anwendung der optischen Glimmentladungsspektroskopie für die Untersuchungen dünner und dünnster Schichten deutlich erweitern.

M. Tech. Metallurgical Engineering. / Aims at improving Cu interconnection problem by homogeneous distribution of ruthenium and tantalum in Cu matrix for excellent interconnection in electronics packaging. The aim will be achieved through the following objectives.Development of appropriate technology for homogenizing submicron metal powders with suitable methods for controlling grain growth during sintering. Study the mechanisms of synergistic incorporation of Ru, and Ta on improving copper interconnection properties. To investigate metallurgical interactions and phenomena occurring during sintering. To investigate specific property and behaviour advantages intrinsic due to the composites and material mix.