Günter Borchardt

Mullite based oxidation protection for SiC-C/C composites in air at temperatures up to 1900 K

For an industrial Si-SiC coated C/C material (reference material) the temperature dependence of the linear rate of mass loss is interpreted in the temperature range 773 < T < 1973 K. The Arrhenius plot of the thermogravimetrically determined oxidation rate shows four typical regimes. Only in the temperature range 1323 < T < 1823 K is the oxidation rate close to or lower than the limit for long-term application. Pulsed Laser Deposition (PLD) allows the ablation of nonconductive and high melting targets and the preparation of films with complex composition. High energy impact CO 2 laser pulses (j= 3.10 7 W cm -2 ) lead to melting and evaporation of the target material in a single step. Therefore the flux of the metal components is stoichiometric. Deposited green layers did not show IR peaks typical for mullite. After a short oxidation treatment (15 min at 1673K) the formation of mullite in the coating was completed as was confirmed by IR spectroscopy and XRD investigations. Thin PLD-mullite layers (900 nm) did not markedly improve the oxidation resistance of the reference material in the high temperature range 1473 < T < 1973 K. However, a preoxidation treatment of the substrate material and mullite coatings with a thickness of 2.5 μm improved the oxidation behaviour significantly. Because of SiO 2 formation at the mullite-SiC interface all samples exhibited a mass increase on oxidation. The inward diffusion of oxygen across the outer mullite-containing layer controlled the kinetics of the reaction as was deduced from 18 O diffusivity measurements in PLD mullite layers. The calculated oxidation rates resulting from the diffusion parameters in SiO 2 and mullite are close to the thermogravimetric data. For oxidation durations of three days only amorphous SiO 2 is formed at the mullite-SiC interface.

Oxygen diffusion in yttria stabilised zirconia—experimental results and molecular dynamics calculations

Bulk oxygen self-diffusion in yttria-stabilised zirconia (YSZ) was investigated using tracer diffusion experiments and molecular dynamics (MD) simulation as a function of the yttria content. Experimentally, 18O tracer diffusion was measured as a function of temperature (650–1200 K) and yttria content (8–24 mol% Y2O3) using gas-phase exchange of the stable isotope 18O and SIMS analysis. For a given temperature, the diffusivity was highest for YSZ containing 10 mol% yttria. The activation enthalpy of diffusion was 0.8 to 1.0 eV, independent of the yttria content. The diffusion process was simulated with molecular dynamics using the program DL_POLY and comparing different potential sets. The oxygen diffusion coefficient was found to be of similar magnitude to the experimental value, and also showed similar concentration dependence with a maximum for YSZ containing 10 mol% yttria. The calculated activation enthalpies of oxygen transport are close to the values observed experimentally.

Cation self-diffusion of 44Ca,88Y, and 96Zr in single-crystalline calcia- and yttria-doped zirconia

Self-diffusion of calcium, yttrium, and zirconium in single-crystalline YSZ and CSZ (YSZ: yttria-stabilized zirconia; containing 10 to 32 mol % Y2O3; CSZ: calcia-stabilized zirconia; containing 11 and 17 mol % CaO) was measured at temperatures between 960 and 1700 °C. For zirconium and calcium diffusion, the stable isotopes 44Ca and 96Zr were used as tracers and the samples were analyzed with secondary ion mass spectrometry. In the case of yttrium diffusion, the radioactive tracer 88Y was used and an abrasive sectioning technique was applied. Zirconium bulk diffusion is slower than yttrium and calcium bulk diffusion, and there is a nearly linear correlation of diffusion coefficient with cation radius. In YSZ, zirconium and yttrium bulk diffusivity are maximum for a stabilizer content of 10–11 mol %, while in CSZ both calcium and zirconium tracer diffusion are independent of the calcium content. The activation enthalpy of yttrium stabilizer bulk diffusion (4.2 eV) is, as in CSZ, slightly smaller than for zirconium bulk diffusion (4.5 eV). The yttrium dislocation pipe diffusivity is five to six orders of magnitude faster than the bulk diffusivity, and its activation enthalpy (3.5 eV) is also smaller than that of the bulk diffusion. From the activation enthalpy and from the concentration dependence of the cation bulk diffusion, it is concluded that the cation diffusion occurs either via free vacancies (VZr4′ in YSZ) or via bound vacancies ([VZr4′−2VO2•]x in CSZ).

Sr diffusion in undoped and La-doped SrTiO3 single crystals under oxidizing conditions

Strontium titanate SrTiO3(100), (110), and (111) single crystals, undoped or donor doped with up to 1 at% La, were isothermally equilibrated at temperatures between 1523 and 1773 K in synthetic air followed by two different methods of Sr tracer deposition: ion implantation of 87Sr and chemical solution deposition of a thin 86SrTiO3 layer. Subsequently, the samples were diffusion annealed under the same conditions as before. The initial and final depth profiles were measured by SIMS. For strong La-doping both tracer deposition methods yield similar Sr diffusion coefficients, whereas for weak doping the tracer seems to be immobile in the case of ion implantation. The Sr diffusivity does not depend on the crystal orientation, but shows strong dependency on the dopant concentration supporting the defect chemical model that under oxidizing conditions the donor is compensated by Sr vacancies. A comparison with literature data on Sr vacancy, Ti, and La diffusion in this system confirms the concept that all cations move via Sr vacancies. Cation diffusion is several orders of magnitude slower than oxygen diffusion.

Cation transport in yttria stabilized cubic zirconia: 96Zr tracer diffusion in (ZrxY1–x)O2–x/2 single crystals with 0.15⩽x⩽0.48

Abstract For a wide range of stabilizer concentrations in yttria stabilized cubic zirconia (YSZ), Zr diffusion data extracted from published creep data and dislocation loop shrinkage data are discussed together with published Zr tracer diffusion data and our own data on Zr tracer diffusion in order to identify the most probable point defect responsible for Zr diffusion. From this evaluation, complex defects can be ruled out, as the single vacancy, VZr4′, fits best.

Cationic Surface Segregation in Donor-Doped SrTiO3 Under Oxidizing Conditions

The influence of high temperature oxygen annealing on (100) oriented donor-doped SrTiO3 single crystals was studied. Crystalline precipitates were found on the optical scale on surfaces of lanthanum-doped as well as niobium-doped specimens with donor concentrations above 0.5 at.%. The amount of the secondary phase increases with the doping level, oxidation temperature and oxidation time. EDX analyses of the crystallites reveal a SrOx composition.The formation of the observed secondary phase is discussed by means of the defect re-equilibration of the cation sub-lattice. In view of the point defect model for donor-doped perovskites, n-conducting SrTiO3 changes its compensation mechanism during an oxidation treatment from “electronic compensation” (ND= n) to “self-compensation” (ND = 2[VSr″]) by forming cation vacancies. Due to the favored Schottky-type disorder in perovskites, the formation of strontium vacancies is accompanied by a release of strontium from the regular lattice. Since the excess strontium is found to be situated at the surface in form of SrO-rich precipitates only, we propose the formation of strontium vacancies via a surface defect reaction and the chemical diffusion of strontium vacancies from the surface into the crystal as the most probable re-equilibration mechanism for the oxidation treatment of single crystals.The introduced mechanism is in contrast to an established model which proposes the formation of Ruddlesden-Popper intergrowth phases SrO·(SrTiO3)n in the interior of the crystal.

Computer modelling of ion migration in zirconia

Defect structure and migration pathways of cations in cubic zirconia (ZrO2) have been calculated using two computer modelling techniques. The first is based on the Mott–Littleton method, which considers defects to be embedded in an otherwise perfect crystal, and the second is the supercell approach, which allows finite defect concentrations to be modelled. Using the first approach, migration pathways for both intrinsic and dopant cations have been calculated. Activation energies ranging from 3.1 to 5.8 eV have been calculated assuming a vacancy mechanism. For highly charged dopants a curved pathway was found to be favoured over a straight pathway. The effect of stabilizer concentration on the properties of the system investigated has been analysed using the supercell method; 3 × 3 × 3 and 4 × 4 × 4 supercells containing 3–40 mol% calcia (CaO) or yttria (Y2O3) have been constructed assuming a random distribution of both dopant cations and oxygen vacancies. After relaxation the oxygen vacancies were found to be located adjacent to the zirconium cations in the CaO-doped system, while remaining randomly ordered in the Y2O3-doped system. Also cation vacancies were created, and after relaxation they were surrounded in all systems (CaO-stabilized ZrO2 and Y2O3-stabilized ZrO2) on average by 2.7 oxygen vacancies.

Cation tracer diffusion of 138La, 84Sr and 25Mg in polycrystalline La0.9Sr0.1Ga0.9Mg0.1O2.9

Cation tracer diffusion of 138La, 84Sr and 25Mg in polycrystalline samples of doped lanthanum gallate, La0.9Sr0.1Ga0.9Mg0.1O2.9, was investigated by SIMS for temperatures between 900 °C and 1400 °C. It was found that diffusion takes place through the bulk and along the grain boundaries. The bulk diffusion coefficients are similar for all cations with activation energies which are strongly dependent on temperature. At high temperatures, the activation energies are about 4.5 eV, while at low temperatures values of about 1.5 eV are found. These results are explained by a frozen in defect structure at low temperatures. This means that the observed activation energy at low temperatures represents only the migration energy of the different cations while the observed activation energy at high temperatures is the sum of the defect formation energy and the migration energy. The migration energies of all cations are nearly identical, although 138La and 84Sr are occupying A-sites while 25Mg is occupying B-sites in the perovskite-structure. To explain these experimental findings we propose a defect cluster containing cation vacancies of both the A- and the B-sublattice and anion vacancies as well.

Operation limits of langasite high temperature nanobalances

Abstract The high temperature properties of langasite (La 3 Ga 5 SiO 14 ) are presented in order to evaluate its ability to serve as a high temperature nanobalance. Langasite resonators exhibit bulk oscillations at temperatures of up to 900°C. At 800°C, the mass load response for 780 μm thick resonators is approximately −6.5 cm 2 Hz μg −1 . The temperature dependent frequency shift, about −100 Hz K −1 at 600°C, may be effectively compensated by monitoring the difference frequency of closely mounted resonators. As an example, the response of a TiO 2− x coated langasite nanobalance to different oxygen partial pressures at elevated temperatures is presented. The strong frequency shift due to switching from oxidizing to reducing conditions cannot be attributed to mass changes of the sensor film. Mechanical stress caused by changes in the oxygen stoichiometry is the most likely explanation for the frequency changes.

Neutron reflectometry studies on the lithiation of amorphous silicon electrodes in lithium-ion batteries

Neutron reflectometry is used to study in situ the intercalation of lithium into amorphous silicon electrodes. The experiments are done using a closed three-electrode electrochemical cell setup. As a working electrode, an about 40 nm thick amorphous silicon layer is used that is deposited on a 1 cm thick quartz substrate coated with palladium as a current collector. The counter electrode and the reference electrode are made of lithium metal. Propylene carbonate with 1 M LiClO4 is used as an electrolyte. The utility of the cell is demonstrated during neutron reflectometry measurements where Li is intercalated at a constant current of 100 μA (7.8 μA cm(-2)) for different time steps. The results show (a) that the change in Li content in amorphous silicon and the corresponding volume expansion can be monitored, (b) that the formation of the solid electrolyte interphase becomes visible and (c) that an irreversible capacity loss is present.