Papers by Author: Harald Schmidt

Abstract: Lithium-silicon compounds are promising materials as negative electrodes in Li-ion-batteries. The diffusion of Li in electrode materials is important for charging/discharging rates, maximum specific capacity and possible side reactions. In order to further the development of novel negative electrode materials for lithium-ion batteries, understanding the basic principles of atomic transport is of high importance. Thin Li_xSi films were investigated, which were produced by reactive ion-beam co-sputtering of segmented elemental Li and Si targets. Li tracer self-diffusion experiments were done on Li_xSi|⁶Li_xSi heterostructures and ⁶Li and ⁷Li isotopes depth profiles were analysed by secondary ion mass spectrometry before and after annealing. Diffusivities were extracted by comparing the experimental depth profiles to analytical solutions of the diffusion equation. The diffusivities for low Li concentrations x < 0.1 in Li_xSi follow the Arrhenius law between 140 and 325 °C with an activation energy of 1.4 eV. A trap-limited diffusion mechanism is suggested, comparable to hydrogen diffusion in hydrogenated amorphous silicon. In contrast, for x ≈ 0.4 complete isotope interdiffusion is observed directly after deposition at room temperature. These results indicate a significant acceleration of diffusion with increasing Li content as suggested in literature by theoretical calculations [1].

Abstract: Lithium transport through ultrathin silicon layers can be measured non-destructively by neutron reflectometry (NR) using a multilayer composed of silicon layers embedded between solid state Li reservoirs. An established model system is a multilayer with a repetition of five [Si / ^natLiNbO₃ / Si / ⁶LiNbO₃] units. Two types of Bragg peaks are detectable in reflectivity patterns. These Bragg peaks result from the interference of neutrons reflected at periodic interfaces. One type of Bragg peak originates from the periodicity of the LiNbO₃/Si chemical contrast (first order peak), while the other Bragg peak results from a superstructure with double periodicity. This superstructure may arise from ⁶Li/⁷Li isotope contrast or alternatively from periodic thickness variations, as shown by simulations based on the Parratt algorithm. The intention of the present paper was to elucidate the origin of the second Bragg peak. Experiments done by Secondary Ion Mass Spectrometry (SIMS) isotope sensitive depth profiling showed in a direct way that annealing at 360 °C destroys indeed the ⁶Li/⁷Li contrast, whereas the LiNbO₃/Si chemical contrast remains unchanged. This evidences that the experimentally observed decrease of the second Bragg peak in the reflectivity pattern during annealing is a measure for Li transport through the Si layer.

Abstract: The contributions of vacancies and self-interstitials to silicon (Si) self-diffusion are a matter of debate since many years. These native defects are involved in dopant diffusion and the formation of defect clusters and thus influence many processes that take place during Si single crystal growth and the fabrication of silicon based electronic devices. Considering their relevance it is remarkable that present data about the properties of native point defects in Si are still limited and controversy. This work reports recent results on the properties of native point defects in silicon deduced from self-diffusion experiments below 850°C. The temperature dependence of silicon self-diffusion is accurately described by contributions due to vacancies and self-interstitials assuming temperature dependent vacancy properties. The concept of vacancies whose thermodynamic properties change with temperature solves the inconsistency between self-and dopant diffusion in Si but further experiments are required to verify this concept and to prove its relevance for other material systems.

Abstract: Li diffusion is investigated in Li₂O-deficient, (110) oriented LiNbO₃ single crystals in the temperature range between 523 and 673 K by secondary ion mass spectrometry. A thin layer of ion-beam sputtered isotope enriched ⁶LiNbO₃ was used as a tracer source, which allows one to study pure isotope interdiffusion. The diffusivities coincide with those of (001) oriented single crystals and follow the Arrhenius law with an activation enthalpy of 1.33 eV. The results prove the existence of a three-dimensional diffusion mechanism.

Abstract: In this work we investigated the structural re-organization of thin nanocrystalline Pt films in the temperature range between 250 °C and 400 °C by in-situ XRD, GIXRD and XRR synchrotron experiments. A re-orientation of (111) atomic planes and a relaxation of residual stress occurs. After heating up, Bragg peak fringes can be observed in the diffractograms. They are a direct proof that the Pt films are built of (111) columnar grains which essentially reach the whole film thickness of about 40 nm. During isothermal annealing a relaxation of the dispersion parameter of the atomic planes takes place which is associated with an activation energy of (0.4 ± 0.1) eV.

Abstract: Self-diffusion in magnetron sputtered nanocrystalline Fe films was investigated by neutron reflectometry on [natFe(10 nm)/57Fe(5 nm)]20 isotope multilayers between 310 and 510°C. The determined diffusivities corresponding to diffusion length between 0.8 – 2.1 nm are time dependent and decrease by more than two orders of magnitude during isothermal annealing. This behaviour can be attributed due to the annihilation of frozen-in point defects, formed during sputtering. For very long annealing times of more than 8 days the diffusivities above 400°C are in good accordance with the volume diffusivities on single crystals given in the literature. However, at temperatures below 400°C the diffusivities are higher than extrapolated literature data, indicating that defect annihilation is still an ongoing process. Furthermore, a comparison of diffusivities obtained for nanocrystalline Fe films prepared by magnetron sputtering and ion beam sputtering, respectively, is presented and discussed.

Abstract: Studies of self-diffusion in solids are presented, which are based on neutron reflectometry. For the application of this technique the samples under investigation are prepared in form of isotope heterostructures. These are nanometer sized thin films, which are chemically completely homogenous, but isotope modulated. Using this method, diffusion lengths in the order of 1 nm and below can be detected which allows to determine ultra low diffusivities in the order of 10-25 m2/s. For the model system amorphous silicon nitride we demonstrate how the structure of the isotope hetrostructures (triple layers or multilayers) influences the efficiency of diffusivity determination.

Abstract: Literature data on (self-)diffusion in transition metal borides are extremely sparse due to the low atomic mobility of the constituents and due to the fact that for B there exist no suitable radioactive tracers and only two stable isotopes with a high natural abundance of 19 % (10B) and 81 % (11B), respectively. The present paper reviews our experiments on the tracer diffusion of transition metals and boron in TiB2, WB2+x, and (TixWyCrz)B2 which were carried out using stable isotopes and secondary ion mass spectrometry (SIMS). For tracer deposition, ion implantation and magnetron sputtering were used. In order to measure boron diffusion, a specially designed experiment was build up where a TiB2 layer was sputtered on an isotope-enriched Ti11B2 bulk ceramic sample. In addition, first results on chemical interdiffusion in the system (TixWyCrz)B2 will be presented. Here, a method based on magnetron sputtered layers and secondary neutral mass spectrometry (SNMS) was used which allows to determine much lower diffusivities (down to 10-19 m2/s) than the conventional EDX line-scan method on cross-sectional samples.

Abstract: In this work we investigated the mobility of hydrogen in amorphous ceramics with the composition Si13B13C60N13 (AM26C). The material was derived from a pre-ceramic polymer and thermolyzed at 1000 °C. After thermolysis the AM26C ceramics are assumed to be separated in an amorphous SiC phase and an amorphous C(BN)x phase. To measure the diffusivities we used deuterium as a tracer, which was introduced via isotope exchange from the gas phase at temperatures between 700 °C and 1100 °C. Depth profiling was done with secondary ion mass spectrometry (SIMS). The profiles could be fitted with complementary error functions. The diffusivities obey an Arrhenius law. The activation enthalpy is 0.8 eV, the pre-exponential factor is 5×10-12 m2 s-1. These values are close to those found for glassy carbon and thin amorphous C-B-N films as reported in the literature. We therefore conclude that the amorphous C(BN)x phase is the transport path for hydrogen in AM26C ceramics. A direct interstitial diffusion mechanism can account for the activation enthalpy of 0.8 eV. The low value for the pre-exponential factor is attributed to an entropy factor arising from the temperature dependence of the chemical potential of hydrogen.

Abstract: The self-diffusion of nitrogen is studied in amorphous silicon nitride, which is a model system for a covalently bound amorphous solid with a low atomic mobility where reliable diffusion data are still lacking. Comparative experiments on Si14Nx/Si15Nx (x ≈ 1.33) isotope multilayers were carried out with secondary ion mass spectrometry (SIMS) and neutron reflectometry (NR), respectively. It was found that experiments with SIMS are not very well suited for the determination of diffusivities in a broad temperature range. The minimum diffusion length of about 5-10 nm detectable with this method is too large. At high temperatures (> 1200 °C) the amorphous solid crystallizes before any diffusion is measured and at low temperatures (< 1100 °C) the diffusivities are too low to be detected. In contrast, with neutron reflectometry diffusion lengths in the order of 1 nm and diffusivities down to 10-24 m2 s-1 were measured between 950 and 1250 °C. The potential of this method for the determination of ultra slow diffusion processes is discussed.