14 Externe wissenschaftliche Einrichtungen

Permanent URI for this collectionhttps://elib.uni-stuttgart.de/handle/11682/15

Browse

Search Results

Now showing 1 - 10 of 78
  • Thumbnail Image
    ItemOpen Access
    Elektrische und magnetische Eigenschaften metallreicher Seltenerdmetallhalogenide
    (2004) Ryazanov, Mikhail; Simon Arndt (Prof. Dr.)
    Im Rahmen dieser Arbeit wurden metallreiche Seltenerdmetallhalogenide unterschiedlicher Valenzelektronenkonzentration dargestellt und ihre physikalischen Eigenschaften untersucht. Dafür wurde sowohl der Weg des Einbaus von Interstitialwasserstoffatomen in die Kristallstruktur als auch die Möglichkeit der Kationen- bzw. Anionensubstitution beschritten. Es wurden Hydridiodide und ternaere Iodidtelluride von Y, La und Gd synthetisiert und charakterisiert. Magnetische und elektrische Messungen weisen auf eine erhebliche Änderung der physikalischen Eigenschaften in Abhängigkeit vom Wasserstoffgehalt hin. Es wurden verschiedene auffallende Effekte, wie z.B. Kolossal-Magnetwiderstand sowie auch Spinclusterglas-Verhalten, beobachtet.
  • Thumbnail Image
    ItemOpen Access
    Polymer electrolyte membrane degradation and mobility in fuel cells : a solid-state NMR investigation
    (2010) Ghassemzadeh Khoshkroodi, Lida; Müller, Klaus (Prof. Dr.)
    It is generally believed that fuel cells will play an important role in energy technology already in the near future. Operating polymer electrolyte membrane fuel cells (PEMFCs) at temperatures higher than 100 °C and reduced humidity is anticipated to avoid most of the shortcomings associated with the low-temperature fuel cell operation, such as CO poisoning of the electrode catalysts, slow electrode kinetics of the oxygen reduction reaction and expensive water/thermal management. To date, the operation temperature of PEMFCs is limited to about 90 °C, and this limit is given by the properties of the perfluorosulfonic acid (PFSA) ionomer, Nafion, which is commonly used as a separator material. Apart from the proton conductivity decay at higher temperature and lower humidification, it is also the limited stability of Nafion preventing it from long term operation. Despite the high stability of the PTFE backbone in Nafion, severe deterioration is observed during fuel cell operation. Formation of pinholes and cracks, thinning of the membranes and decrease of ion exchange capacity were reported. The fluorine release indicated that the bond cleavage process takes place under fuel cell operating conditions. Bond cleavage was initially believed to proceed from radical attacks to the carboxyl groups terminating the PTFE backbone of Nafion, and it was claimed to be controlled by the endcapping of the polymer backbone with a CF3 group. However, the release of fluoride was reported even after endcapping of the materials. The observations proved that bond cleavage limits the stability of PFSA membranes, but the elementary reactions and consequences on the membrane microstructure are not fully understood yet. In this work, it has been tried to get new insights into the problems of long term stability of polymer electrolytes for low temperature fuel cells. The aim was to identify the changes in the chemical structure of the membrane after operating in a fuel cell. This understanding is essential for extending the operation limit of PFSA-type membranes by either improving the membrane properties or adjusting the conditions within the running fuel cell. In the present work, therefore the changes taking place in PFSA membranes after applying in-situ and ex-situ aging protocols have been investigated. While the in-situ experiments provide a global picture, the analysis of membranes after ex-situ tests, with various conditions, allows the separation of different types of reactions. In previous studies the degradation changes were mainly monitored by analyzing the released water of the fuel cell or by using the liquid ionomers. In this work with the help of solid-state NMR spectroscopy, the direct study of the chemical structure and dynamics of the polymer membranes before and after the degradation tests became possible. The structural changes in different parts of the PFSA membranes were first inspected after an in-situ aging test. These examined membranes (Nafion and Hyflon Ion) differed by the length of the side chains. The comparison of the solid-state 13C and 19F NMR data of polymers before and after the in-situ degradation test showed that changes can take place not only in the main chain of the polymer, but also within the polymer side chains, as reflected by changes of NMR signals associated with CFSO3, CF3, OCF2 and CF groups. The degree of degradation is found to decrease with increasing membrane thickness while for a given thickness the short side chain polymer, Hyflon Ion, appears to degrade less than Nafion. In order to understand the reason for these observations, a new ex-situ method has been developed to mimic the degradation of polymer electrolyte membranes in PEM fuel cells (caused by the cross-leakage of H2 and O2). In this ex-situ setup, it was possible to expose membranes to flows of different gases with controlled temperature and humidity. H+-form Nafion films with and without electrode layer (Pt) have been treated in the presence of different gases in order to simulate the anode and cathode side of a PEMFC. The changes of the chemical structure occurring during the degradation tests were primarily examined by solid-state 19F NMR spectroscopy. For completion, liquid-state NMR studies and ion exchange capacity measurements were performed. It was found that degradation occurs only when both H2 and O2 are present (condition of gas cross-leakage), and when the membrane is coated with Pt catalyst. The chemical degradation rate is found to be highest for H2-rich mixtures of H2 and O2, which corresponds to the conditions at the anode under OCV. It is further shown that side chain disintegration is very important for chemical degradation, although backbone decomposition also might take place. The fact that in-situ degradation effects were reproduced by the present ex-situ experiments, suggest that membrane degradation in a running fuel cell is mainly the consequence of chemical aging. Detecting the degradation for the membranes coated with Pt in the presence of both gases, H2 and O2, points toward the importance of radicals in the degradation process, which in a running fuel cell (in-situ conditions) may only form in the presence of some gas cross-over, allowing H2 and O2 to react at the Pt catalyst of the anode or cathode structure. Since the gas cross-over increases for the thinner Nafion membrane, these results indirectly explain the higher degradation rate of thin Nafion in the in-situ degradation test. The chemical degradation and stability of PFSA membranes against radical attacks was also investigated in a Fenton ex-situ degradation test. Liquid and solid-state NMR as well as ATR-FTIR spectroscopy were applied to the samples before and after the Fenton reaction. A Comparison of the degradation rate of Nafion and Hyflon Ion in the ex-situ Fenton test again proved that the Hyflon Ion membrane is more stable than Nafion. Comparing the degradation rate of the side chain in these two polymers showed that the stability of Hyflon Ion is mainly due to the shortening of the side chain in this polymer. Hence, the absence of one ether group and the tertiary carbon reduces the degradation rate of the side chain and makes this polymer less sensitive to the radical attacks than Nafion. For the performance of a membrane not only the chemical structure but also the polymer dynamics is important. Therefore the molecular mobility of the ionomer was investigated by variable temperature 19F NMR lineshape, T1 and T1ρ relaxation experiments. The decrease of the temperature dependent linewidth was explained by the reduction of static disorder in the Nafion membrane. From the relaxation data there was evidence for structural annealing, which is independent of the chemical degradation. Chemical degradation is considered to reduce the chain flexibility (i.e. the motional amplitudes), which may be explained by chain cross-linking and condensation reaction for the side chains. To overcome the problem of Nafion's low conductivity at temperatures above 100 °C and low relative humidity, also composite membranes were introduced. These membranes consist of Nafion modified by inorganic oxide additives. It has been reported that under dry conditions, these membranes show enhanced water uptake and water diffusion when compared with filler-free Nafion. In order to understand the reason for the better performance of these polymers, the impact of the oxide particles on the polymer dynamics has been investigated. [Nafion/(SiO2)x] composite membranes in the dry and wet state with x ranging from 0 to 15 w/w% were investigated by variable temperature solid-state 19F NMR spectroscopy. 19F T1 and T1ρ relaxation times and NMR lineshapes were analyzed in order to get details about the polymer mobility. It is concluded that solid oxide SiO2 particles play an important role in stabilizing the chemical structure and morphology of the polymer especially in the dry state. The filler particles lead to higher mobility of polymer chains, if the filler content has an optimized value of about 9 w/w%. The results were further supported by comparing the sideband intensity as well as the linewidth in 19F NMR and recording the 19F{1H} CP/MAS NMR spectra. Furthermore, it has been shown that the structure of composite membranes is more stable after dehydration and possible condensation reactions are less likely in these membranes. The presence of filler particles decrease the chance for morphology changes and close packing of polymer chains in the dry state. Also the decrease of ionic exchange capacity after dehydration is less severe for the composite membrane as compared to filler-free Nafion. In conclusion, the present results provide a complete picture of solid membrane before and after degradation and of possible mechanisms for radical formation and radical attacks to the polymer. In addition, it is shown which changes can occur in the morphology of polymer chains in low humidification and high temperature. Some general suggestions for the better performance of polymer electrolyte membrane are therefore: For improving the performance of polymer electrode membrane, the sources for the radical formation in the fuel cell should be controlled. This can be possible to some extend by avoiding the use of iron end plates in the fuel cells. Also the chance for the gas crossover through the membrane should be decreased. Thicker membranes show less gas cross-over. By taking into account the higher resistivity of thicker membranes, an optimized membrane thickness should be selected. Hydrocarbon sulfonated polyetherketones possess narrower hydrophilic channels which significantly reduce electroosmotic drag, water permeation as well as gas cross-over. Also the short side chain perfluorinated polymer, Hyflon Ion, with lower electroosmotic drag of water should possess a reduced gas cross-over though the membrane. The more efficient way for decreasing degradation is to use membranes which are stable against radical attacks. At this point the perfluorinated polymers are still the best available membranes. Endcapping of the backbone in these polymers and decreasing the concentration of reactive end groups like COOH during the polymer manufacturing process can significantly decrease degradation. To minimize degradation of the side chains in perfluorinated polymers, short side chain polymers are suggested because of less reactive groups for the radical attacks and higher concentration of acidic groups. When higher operation temperatures are required, composite Nafion membranes might be used. The higher stability of these membranes makes them advantageous for operating at evaluated temperatures and low relative humidity. The novel results from the present work lead to a better understanding of membrane degradation, which still represents a serious problem for fuel cells under operation conditions, and provide important indications for future developments of membranes with improved performance for alternative energy conversion devices.
  • Thumbnail Image
    ItemOpen Access
    Ab initio thermodynamic study of defective strontium titanate
    (2013) Blokhin, Evgeny; Maier, Joachim (Prof. Dr.)
    In the presented thesis the perfect and defective SrTiO3 bulk crystals and their (001) surfaces are considered on ab initio level. Since the experimental study of the complex defective systems is comparatively expensive and difficult, and the computer performance has been greatly increased in the last years, the ab initio modeling became very efficient tool to be applied in this field. Additionally, there is a significant industrial demand for the investigation and improvements of the performance of the perovskite-based electrochemical devices, e.g. solid oxide fuel cells and permeation membranes. Finally, there is a lack of methodological studies on defect thermodynamics. The oxygen vacancies and iron impurities, as well as their complexes, are the characteristic defects in SrTiO3 perovskite material. The understanding of basic defect properties and defect-induced phenomena under realistic external conditions requires calculation of thermodynamic properties (the free energy and entropy effects). Thus, thesis is focused on atomic vibrations, i.e. phonons, which are necessary to go beyond standard 0°K approximation and to provide a link to the thermodynamic properties at finite temperatures. This is necessary for a realistic treatment of electrochemical devices. The chosen ab initio modeling scheme is found to ensure the most accurate description simultaneously for the structural, electron and phonon properties of the perfect SrTiO3. Namely, the splitting of the phonon frequencies due to the antiferrodistortive phase transition at 105°K is confirmed to be very small (2–12 cm-1). The experimental temperature dependence of the SrTiO3 heat capacity is also successfully reproduced. Further, the modeling scheme is applied for thermodynamic treatment of oxygen vacancies and iron impurities in SrTiO3 at finite temperatures. The calculated phonon densities of states and group-theoretical study of the defect-induced phonon frequencies are used for the experiment analysis. Several defect-induced local phonon modes are identified, and the experimental Raman- and IR-spectroscopy data are interpreted. The Jahn-Teller-type local lattice distortion around both Fe4+ impurity and oxygen vacancy VO is shown to result in Raman- and IR-active phonons. In particular, the experimentally observed Raman frequency near 700 cm-1 is shown to arise for both defects due to a local O ion stretching vibration nearby the Jahn-Teller defect. However, an absence of such a frequency in an experimental phonon spectrum is found to be a manifestation of formation of Fe3+–VO complexes with oxygen vacancies in the first coordination sphere of iron impurities. The Gibbs formation energy calculated for the neutral oxygen vacancies in bulk SrTiO3 taking into account the phonon contribution is found to be in excellent agreement with the experiment. The phonon contribution to the formation energy is shown to increase with temperature, to about 5% above 1000°K. The predicted relative stability of several structural complexes of oxygen vacancy and iron impurity in SrTiO3 is confirmed by known experimental measurements. Several structural models of such Fe3+–VO complexes in SrTiO3 are discriminated according to the XANES and EXAFS experiments. On an example of SrO-terminated SrTiO3 ultrathin films, the one-dimensional confinement effect on the vacancy formation energy is found to be inconsiderable at 0°K. The phonon contribution to the Gibbs free energy of VO formation in such ultrathin films at finite temperatures is shown to be minor. This suggests the further account of anharmonic effects is required. The workflow developed in thesis is proposed for the modeling of wide class of defects in non-metallic solids. Several auxiliary computer tools were designed in order to simplify such possible studies.
  • Thumbnail Image
    ItemOpen Access
    Mixed-conducting (Ba,Sr)(Co,Fe,Zn)O3-δ as cathode material for proton-conducting ceramic fuel cells : defect chemistry and oxygen reduction mechanism
    (2014) Pötzsch, Daniel; Maier, Joachim (Prof. Dr.)
    In the present study mixed-conducting solid oxides with perovskite structure are investigated regarding their applicability as oxygen electrode (cathode) in solid oxide fuel cells based on proton-conducting electrolyte membranes. This type of high temperature fuel cells is a promising device as it may allow one to decrease the operating temperature. At reduced temperatures the overpotential of the cathode dominants over all contributions to the performance of the fuel cell. Hence, it is mandatory to systematically and purposefully enhance the catalytic activity of the cathode, for which a fundamental understanding of the electrochemical properties of the materials is key. These properties are inter alia determined by the defect chemistry of the material's mobile charge carriers. In humid, oxidizing atmosphere at elevated temperature three defects have to be considered: Oxygen vacancies, interstitial protonic defects associated with a regular oxygen site and electron holes. A mixed proton/hole conductivity in the cathode is highly desired as to make the whole electrode surface catalytically active for water formation. The bulk thermodynamic and transport behavior regarding three mobile charge carriers are significantly more complex than for systems with only'' two mobile defects. Numerical simulations are necessary to describe and understand the bulk thermodynamic and transport properties. The proton concentration of mixed-conducting perovskites was ex-situ determined by Karl-Fischer titration and thermogravimetry analyzing the mass spectrometer signal, and in-situ by dynamic, thermogravimetric relaxation experiments upon step-wise changes in the water partial pressure. BSFZ was found to be the most promising candidate of the four and, therefore, selected for further investigations. From a thermodynamic point of view two limiting possibilities incorporating protons are identified: Incorporating a water molecule occupying an oxygen vacancy and forming two protonic defects (acid-base thermodynamics) and taking up water releasing simultaneously oxygen, i.e. hydration-deoxygenation, formally equivalent to pure hydrogen incorporation (redox thermodynamics). Depending on temperature, oxygen and water partial pressure any combination of both mechanisms is possible. With the help of numerically simulating the transport behavior at different conditions, measuring the mass relaxation upon water partial pressure changes at two different oxygen partial pressures and determining the thermodynamic properties in dry conditions, the proton concentration could be calculated applying the thermodynamic model. For BSFZ the mass relaxation transients upon pH2O change were measured for two different p2O values. Interestingly, the mechanism of proton uptake was found to change from predominantly acid-base water uptake to predominantly redox hydrogen uptake. The transients could be fitted through the known solution of Fick's second law of one-dimensional diffusion into a plane sheet (with sufficiently fast surface equilibration) obtaining chemical diffusivities. The proton conductivity is calculated using its concentration and diffusivity. The obtained values are up to one and a half orders of magnitude below the proton conductivity of 15% Y-doped BaZrO3 being one of the best known high temperature proton conductors. Nevertheless, even the estimate of the lower limit of proton conductivity in BSFZ is orders of magnitude larger than a required minimum to make the whole electrode surface catalytically active. This is to the best of my knowledge the first study providing quantitative values for the proton conductivity in those mixed-conducting perovskites typically used as cathode material in solid oxide fuel cells. The electrochemical activity of BSFZ and BSCF was investigated by impedance spectroscopy. For microelectrodes the low frequency contribution typically dominates the overall impedance, and its resistance is inversely proportional to the reaction rate determining the overall oxygen to water reaction. The inverse dependence of the surface reaction resistance to the area of the microelectrode confirms that the whole electrode surface is catalytically active. Its dependency on oxygen and water partial pressure provides important information about the oxygen reduction mechanism. The exponents of the oxygen and water partial pressure dependency indicate that molecular oxygen and oxygen vacancies are participating in the rate determining step of the oxygen reduction reaction.
  • Thumbnail Image
    ItemOpen Access
    Charge carrier formation, mobility and microstructure of sulfonated polyelectrolytes for electrochemical applications
    (2015) Wohlfarth, Andreas; Maier, Joachim (Prof. Dr.)
    Polyelectrolytes are materials consisting of a polymer backbone with covalently attached positively or negatively charge groups including their counterions. Sulfonated polyelectrolytes are a specific class, which is especially interesting for electrochemical application as they can be used to separate the electrodes and mediate the electrochemical reactions taking place at anode and cathode by conducting a specific ion; this ion may be H+ in the case of PEM-fuel cells or Li+ and Na+ in various battery systems. The key challenges for the development of these electrolytes is the combination of good mechanical properties and high ion transport as well as high electrochemical stability. High ionic conductivity of polyelectrolytes depends on the presence of small polar solvents to ensure efficient dissociation and mobility of the counterions. These processes cannot be understood by merely considering electrostatics (i.e. Deby-Hückel approach) such as in Manning counterion condensation theory. Specific interactions between solvent-ion-polymer and the molecular conformations have to be taken into account as well. This is one of the results of the present work using sulfonated polyelectrolytes, different cations and solvents as model systems with a combined approach of experimental techniques and simulations. The dissociation behavior of sulfonated polysulfones was investigated by a combined electrophoretic (E) NMR, pulsed magnetic field gradient (PFG) NMR and conductivity approach. Since the results from the NMR experiments, especially from E-NMR which is by far no standard measurement, are crucial for the key conclusions drawn in this thesis, some critical issues of this technique are studied and discussed in detail. E-NMR is essentially a PFG-NMR experiment with an applied electric field; the applied voltage can reach up to 300 V. Therefore, it was necessary to determine a measurement window in which no decomposition or other interfering effects appeared. In addition, polymers gererally exhibit some polydispersity with the low molecular weight fraction showing a higher diffusion coefficients and drift velocities, which had to be taken into account. By concentrating ionic groups on the polymer, specific polyelectrolyte effects show up. Dissociation is no longer complete, the interaction between ionic charges and the solvent is heavily modified and correlations of ionic motion start to appear. According to a MD-simulation, this very much depends on the polymer conformation and position of the ionic groups as well as the chemical nature of the solvent. Once the density of ionic groups (-SO3H) of polysulfones reaches a point where their average separation is of the order of the Bjerrum length of water, the degree of counterion condensation is shown to depend on details of the molecular structure and the accessible conformations of the polymer chain. In this regime, well-defined ionic aggregates occur, i.e. triple-ions form. The conformational details depend on the degrees of freedom and specific interactions between ions and solvent. When it comes to ion conducting membranes, increasing the ion exchange capacity (decreasing the average separation of ionic groups) is a common measure to increase ionic conductivity. However, the results on dissociation and conductivity of synthesized polysulfones containing octasulfonated units (currently the material with highest known IEC) clearly reveal the limit of this approach. The short separation of ionic charges in such systems at high concentrations additionally leads to electrostatic interactions between neighboring polymer strands. This is the driving force for a nanoscale ordering in polyelectrolyte membranes. Different kinds of solvents, ions and ion exchange capacities directly affect the microstructure formation. Finally, the effects of acid-base interactions between sulfonic acid-based polyelectrolytes and weakly basic modified polymers were investigated as blending of both is a way to form stable membranes for electrochemical applications. Here, the membrane formation process and the resulting properties, in particular proton conductivity, microstructure and mechanical strength have been studied. The developed polymer blends are the first example in which an improvement of mechanical properties not goes along with a significant decrease of proton conductivity. Key to success was to use a hydrophilic polymer with a high IEC and a hydrophobic polymer with a low number of basic groups. In summary, this thesis provides insides into the charge carrier formation process, the transport and microstructure of sulfonated polyelectrolytes by identifying the relevant molecular interactions. Together with the superior mechanical properties of the developed blend membranes, this work significantly contributes to solve the key challenges for electrolytes in electrochemical application.
  • Thumbnail Image
    ItemOpen Access
    Synthesis of new fullerides via the "break-and-seal" approach and their characterization
    (2009) Kozhemyakina, Nina V.; Jansen, Martin (Prof. Dr.)
    The present dissertation deals with the synthesis and characterization of fullerides. For the first time the "break-and-seal" technique was applied for fulleride synthesis. The reaction was performed in a completely all-glass apparatus under vacuum, avoiding the use of glass connections and use of grease. Starting from crown-ethers, potassium metal and C60 fullerene, six new fullerides have been synthesized. The modified "temperature difference method" was successfully used for growing single crystals from solution within a few days. In [K(DB24C8)(DME)]2C60*(DME) the fullerene unit has a charge of 2-. The (C60)2- units are arranged in hexagonal layers parallel to the ab plane, forming distorted trigonal prisms. The fullerene anions and potassium cations develop a pseudobinary topology which is reminiscent of the CdI2 structure type. Bond lengths' distribution in (C60)2- was examined. One orientation of the dianion was found to match perfectly the one predicted by calculations. KC60(THF)5*(THF)2 crystallizes in a structure with fully ordered C60 units. C60- anion-radicals and K+ form ion pairs. The ion pairs form corrugated layers in the ac crystallographic plane, the given compound being an example for a low-dimensional fulleride partial structure. For the compound [K(DB24C8)(THF)]2C60*THF the structure solution was complicated by the disorder of crown-ether and solvent molecules which could not be overcome, although the (C60)2- unit was ordered. In [K(DB24C8)(DME)]C60 the fullerene unit exists as a monomeric anion-radical and in [K(DB24C8)(DME)]2[C60]2 - as a dimer-dianion. The latter compound is an example of rather not many fulleride structures, where C60 exists in the form of dimers. The interfullerene C-C bond length is 1.57(3) Å. In [4{K(DB18C6)(C60-)}(THF)6]*[C60]*(THF)6 at temperatures above 220 K each of the four C60- units exists in form of anion-radicals, and at lower temperatures - as a dimer-dianion, the interfullerene bond being 1.63(0) Å. The dimers are fully ordered. In addition, uncharged disordered C60 molecules are found, what follows from the charge balance. The low-temperature phase is a first example of a fulleride structure where fullerene exists in three different bonding states: anion-radical monomer, dianion-dimer, and a neutral C60. In the dimer, the pentagons adjacent to sp3-hybridized carbon atoms, are in trans-conformation. DFT calculations were performed, and it is now for the first time that a localization of the negative charge on a small fragment of the C60 cage was found out. Knowing this, it becomes conclusive, considering the Coulombic repulsion, that the preferred orientation of two bound C60- units is trans-conformation. Magnetic measurements were performed. The method for fulleride synthesis used in the present work has a big potential for broadening by using different metals (e.g. alkali, alkali-earth), varying the complexing agents (crown-ethers, cryptands), as well as the organic solvent (or solvent mixtures).
  • Thumbnail Image
    ItemOpen Access
    Darstellung und Charakterisierung neuer Einkomponentenvorläufer und Keramiken im System Si/B/N/C und Festkörper-NMR-Untersuchungen an isotopenmarkiertem SiBN3C
    (2006) Epple, Angelika; Jansen, Martin (Prof. Dr. Dr. h.c.)
    Multinäre amorphe Netzwerke aus den Elementen Si, B, N und C haben aufgrund ihrer Eigenschaften als Hochleistungskeramiken Bedeutung erlangt. Als Darstellungsverfahren hat sich im Gegensatz zu der bei oxidischen Keramiken eingesetzten Pulverroute die Polymerroute ausgehend von geeigneten Einkomponentenvorläufern und Vernetzungsreagenzien bewährt. Mit 1,1,3-Trichlor-perhydro-2,3a,6a,9a-tetraaza-1-sila-3,9b-diboraphenalen (TADB-D), 1,1,3-Trichlor-perhydro-2-methyl-2,3a,6a,9a-tetraaza-1-sila-3,9b-diboraphenalen (DMTA-D), 1,1,3-Trichlor-perhydro-3a,6a,9a-triaza-1-sila-3,9b-diboraphenalen (TSDM-D) und 1,1,3-Trichlor-perhydro-2-methyl-3a,6a,9a-triaza-1-sila-3,9b-diboraphenalen (TSDE-D) wurden neue molekulare Einkomponentenvorläufer für quaternäre Keramiken im System Si/B/N/C dargestellt, die als charakteristisches Strukturmerkmal ein tricyclisches Ringsystem aus den Elementen Si, B, C und N aufweisen. Diese Vorläuferverbindungen wurden im Sinne einer 3+3-Cyclokondensationsreaktion durch Umsetzung von 1N,8N-Dilithio-1,4a,8-triaza-8a-boradekalin mit den Einkomponentenvorläufern Trichlorsilylaminodichlorboran (TADB), Dichlorborylmethyltrichlorsilylamin (DMTA), Trichlorsilyldichlorborylmethan (TSDM) und Trichlorsilylaminodichlorborylethan (TSDE) dargestellt. Für TADB-D ist auch eine Synthese ohne Metallierung direkt aus 1,4a,8-Triaza-8a-boradekalin mit TADB im Sinne einer Dehydrohalogenierung unter der Beteiligung der starken, wenig nucleophilen Amin-Base 1,4-Diazabicyclo[2.2.2]octan (DABCO) möglich. Die Charakterisierung erfolgte mittels Lösungs-NMR-Spektroskopie, Massenspektrometrie (MS) und Einkristallröntgenstrukturanalyse. Die Moleküle kristallisieren in der monoklinen Raumgruppe P21/n (TADB-D) bzw. P21/c (DMTA-D, TSDE-D) sowie Pca21 (TSDM-D). Beide Enantiomere des TSDE-D liegen im Kristall nebeneinander vor. Diese neuen Monosilaborazinderivate ermöglichen eine mehrdimensionale Vernetzung sowohl am endocyclischen Si-Atom als auch an einem endocyclischen B-Atom. Über Polymerisation mit H2NCH3 und Calcinierung bei Temperaturen bis 1400°C ließen sich amorphe Keramiken darstellen, die sich durch ausgezeichnete Beständigkeit gegenüber Kristallisation und Zersetzung bei hohen Temperaturen auszeichnen. Die Dichten der dargestellten amorphen Netzwerke sind mit durchschnittlich 1,73 g/cm3 ausgesprochen gering. Als Grundlage für neue amorphe Keramiken im System Si/B/N/C gelang erstmals die Synthese der Borazinderivate B,B',B''-Tris[trichlorsilylamino]borazin (TSAB) und B,B',B''-Tris[dichlor(methyl)silylamino]borazin (DSAB), deren Borazinringe über Aminobrücken mit chlorfunktionalisierten Siliciumsubstituenten verknüpft sind. Die Darstellung erfolgte über eine Silazanspaltung von Hexamethyldisilazan mit TADB bzw. Dichlor(methyl)silylaminodichlorboran (MADB). Die Moleküle wurden mit Hilfe von Lösungs-NMR-, IR-Spektroskopie und MS charakterisiert. Ihre Vernetzung mit H2NCH3 führte zu festen Polymeren, die durch Pyrolyse unter Argon bis 1400°C zu amorphen Keramiken umgesetzt wurden. Das bei der Darstellung von Hexachlorcyclotrisilazan aus SiCl4 und NH3 ebenfalls isolierbare Octachlorcyclotetrasilazan wurde durch MS und mittels Lösungs-NMR-Spektroskopie nachgewiesen sowie anhand einer Einkristallröntgenstrukturanalyse eindeutig charakterisiert. Es handelt sich hierbei um die erste Cyclotetrasilazanverbindung, in der Si ausschließlich mit Chlorsubstituenten funktionalisiert ist. Die Verbindung kristallisiert in der monoklinen Raumgruppe P21/n. Kohlenstoff im Netzwerk übt einen entscheidenden Einfluss auf die Materialeigenschaften aus, dessen strukturelle Ursache jedoch ungeklärt ist. Unter Verwendung einer Vielzahl moderner Festkörper-NMR-Techniken wie 13C-MAS-NMR, 29Si-13C-REDOR-NMR, 13C-29Si-REDOR-NMR, 15N-13C-REDOR, 11B-13C-REDOR-NMR, 13C-11B-REAPDOR-NMR und 2D-13C-RFDR-NMR wurde das Netzwerk einer quaternären Keramik der ungefähren Zusammensetzung SiBN3C untersucht. Ausgehend von 15N-isotopenreinem NH3 wurde über mehrere Reaktionsschritte TADB dargestellt, mit 13C-isotopenreinem H2NCH3 vernetzt und bei Temperaturen bis 1400°C zur Keramik pyrolysiert. Es erschließt sich ein Netzwerkmodell, in dem sich graphitähnlich gebundene Kohlenstoff-Agglomerate (<1nm) in Teile eines ternären Si3B3N7-Netzwerkes einfügen. Aufgrund der beobachteten 11B-13C-Doppelresonanz aber ausbleibender 13C-29Si-Doppelresonanzeffekte kann man davon ausgehen, dass sich diese Cluster vorrangig in Bereichen mit überwiegend B-N-Bindungen ausbilden. Die chemische Verschiebung des 15N-Signals deutet auf eine C=N-Spezies hin. Ein detektierbarer 15N-13C-REDOR-Effekt bleibt jedoch aus. Darüber hinaus treten in den RAMAN-Spektren, die von den Keramiken aus TADB, TADB-D, DMTA-D, TSDM-D, TSDE-D, TSAB und DSAB aufgenommen wurden, im Bereich 1200-1700 cm-1 zwei Banden (D- und G-Bande) auf, die signifikante Merkmale für ungeordnete, graphitartige Kohlenstoff-Strukturen sind.
  • Thumbnail Image
    ItemOpen Access
    A multiscale study of transport in model systems for proton conducting polybenzimidazole phosphoric acid fuel cell membranes
    (2015) Melchior, Jan-Patrick; Maier, Joachim (Prof. Dr.)
    Proton conducting membranes are a key component in modern low (T < 100 °C) and intermediate (T < 180 °C) temperature fuel cells for mobile energy conversion and stationary combined heat and power generators. Finding and evaluating suitable membrane materials with high protonic conductivity in the intermediate temperature range presents a veritable challenge in material science. In this temperature regime neat nominally water free phosphoric acid (H3PO4) was early on identified as conductor with the highest intrinsic proton conductivity. Its high conductivity is due to structure diffusion involving proton transfers between diverse phosphate species. This process contributes up to 97 % to the observed conductivity and makes phosphoric acid an ideal electrolyte for fuel cell applications, which is employed, for example, in the form of phosphoric acid imbibed polybenzimidazole membranes (PBI-PA membranes) in fuel cell technology since the 1990s. However, the membrane’s proton conductivity is decreased in comparison to neat phosphoric acid’s conductivity and the reasons for this stark decrease are not yet fully understood. Indeed, even for neat phosphoric acid only recently a comprehensive rationale of its structure diffusion mechanism on a molecular level was proposed based on ab initio molecular dynamics simulations. This rationale emphasizes the role of frustration in phosphoric acid’s hydrogen bond network caused by the imbalance in the number of proton donors and acceptors. The present thesis provides, for the first time, quantitative experimental insights into how proton conductivity and the underlying conduction mechanisms are affected through changes of hydrogen bond network frustration. Such insights are enabled through the separation of hydrodynamic and structure diffusion in model systems. Model systems with defined composition, furthermore, have been chosen in a way as to discriminate between effects of water and benzimidazole on proton transport in phosphoric acid. In the actual membranes such effects can hardly be separated as the benzimidazole-phosphoric acid ratio is usually insufficiently defined and as the water content under fuel cell operation varies as a function of relative humidity (RH) and—as shown in this work—of the aforementioned benzimidazole content. Hydration isotherms recorded in this study, which encompass the operational temperature and RH range of fuel cells, allow to relate the results obtained from model systems to situations occurring in a running fuel cell. Different experimental techniques are combined to probe proton dynamics on multiple time- and length-scales. These complementary techniques assess dynamic ranges from the subnanosecond regime of quasielastic neutron scattering (QNS) to the millisecond regime of pulsed field gradient nuclear magnetic resonance (PFG-NMR). PFG-NMR is used to measure diffusion coefficients of different nuclei distinguished by their chemical surroundings. Through the analysis of coalescing NMR spectra lifetimes for exchange of nuclei between such chemical surroundings are evaluated also on the millisecond scale. On the nanosecond scale, 1H NMR relaxation measurements are sensitive towards fluctuations involved in hydrodynamic diffusion, while 17O NMR relaxation measurements and QNS are sensitive to the underlying dynamics of proton dislocation in vicinity of the oxygen atoms or transfers inside the hydrogen bond, respectively. In aqueous mixtures four principal transport regimes are identified, i.e., water contents (given as P2O · λH2O) where certain proton diffusion and conduction mechanisms prevail: At high water contents (λ > 14), in the acidic aqueous regime proton transport is essentially that of an acidic aqueous solution. Towards lower water contents (14 > λ > 6), in the viscosity controlled regime, the decrease of conductivity and diffusion is associated with increasing viscosity. With further decreasing water content (λ < 6), in the transition regime, phosphate species progressively aggregate, forming hydrogen bonded structures. The predominant proton transport mechanism changes from hydrodynamic diffusion to structure diffusion. In the decoupling regime, for water contents λ < 3, the molar fractions of condensation products (H4P2O7, etc.) are severely increased and proton transport is further decoupled from the reduced hydrodynamic diffusion. It is found that addition of the same molar amount of either benzimidazole or imidazole to nominally dry phosphoric acid reduces the structure diffusion coefficients by the same degree. That is, the structure diffusion coefficient is associated with the number of additional proton acceptor sites provided by benzimidazole or imidazole. On a related note, this investigation furthermore disproves that structure diffusion rates were also influenced by exchange of protons between benzimidazole and H3PO4 as stated in the literature. The proton exchange in question is determined as to be on the millisecond scale by analysis of coalescing 1H NMR spectra, which is 9 orders of magnitude slower than fast proton exchange between H3PO4 molecules. Proton exchange between benzimidazole and H3PO4, however, affects the decay of the echo intensity in 1H PFG-NMR experiments through which diffusion coefficients are measured. This effect was overlooked in previous literature PFG-NMR measurements on PBI-PA membranes, resulting in deviations between the measured apparent and the actual proton diffusion coefficients. In this work, therefore, a model for intermediate exchange rates is used to obtain 1H PFG-NMR diffusion coefficients of H3PO4; proton exchange rates between (benz)imidazol and phosphoric acid and the benzimidazole diffusion coefficient serve as input parameters for evaluating the echo decay. In order to confirm that the changes in proton dynamics observed in phosphoric acid - benzimidazole mixtures on the millisecond scale are caused by reduced dynamics in the frustrated hydrogen bond network, proton dynamics is also probed on the nanosecond scale. Spatial information is obtained through fitting of the Q-dependence of the quasielastic linebroadening, as obtained from backscattering QNS, to a jump diffusion model. Through this model the diffusion coefficient, lifetime, and “jump length” of a proton involved in structure diffusion are attained. It is found that the activation energies from 17O relaxation rates, backscattering QNS diffusion coefficients, and the PFG-NMR structure diffusion coefficients are virtually identical. Proton transfer inside hydrogen bonds is identified as the rate limiting step of structure diffusion. Addition of benzimidazole neither changes the activation energy of this proton transfer, nor affects proton transport on any scale between the nanosecond and millisecond regime; all influences of the additive on proton transport must occur on lower timescales. Hydration isotherms of phosphoric acid and the phosphoric acid mixtures demonstrate that water uptake is modified by the additive. Benzimidazole reduces the water uptake at fixed RH and the low nominal water content λ of phosphoric acid under fuel cell operational conditions, i.e., without additional humidification, prevails up to higher relative humidity in phosphoric acid - (benz)imidazole mixtures. It is one of the surprising insights of this thesis that the reduced ionic conductivity associated with low water contents is not detrimental to fuel cell performance. Conductivity is still high enough to avoid unacceptable ohmic losses and proton conductivity is mainly occurring through structure diffusion, i.e., the diffusion of protons is well decoupled from the hydrodynamic background. With increasing water content, however, mainly the increasing hydrodynamic background diffusion increases ionic conductivity. This also means an increase in contribution from rapid H3O+ conductance which is associated with electro-osmotic transport of water to the cathode side. Such accumulation of water will eventually increase leaching of phosphoric acid from the membrane and decreases the conductivity on the water deprived side through enhanced condensation. The low contribution of H3O+ at fuel cell operational conditions and the relatively high contribution of structure diffusion are presumptively the reasons why PBI-PA membrane fuel cells are performing well.
  • Thumbnail Image
    ItemOpen Access
    Proton transport mechanisms of phosphoric acid and related phosphorus oxoacid systems : a first principles molecular dynamics study
    (2012) Vilciauskas, Linas; Maier, Joachim (Prof. Dr.)
    Fundamental understanding of proton transport in hydrogen bonded systems on the molecular level remains a key problem in many areas of science ranging from electrochemical energy conversion to biological systems. Despite the enormous advances in the research of these processes, the ostensibly simplest case, proton transport in homogeneous bulk media at thermodynamic equilibrium, proved to be one of the most challenging and elusive. It is only through enormous theoretical and experimental efforts that clear mechanistic pictures of the transport of excess protonic charge defects in water have emerged. However, water has negligible intrinsic proton conductivity. By contrast, the class of compounds known as phosphorus oxoacids have some of the highest reported proton conductivities. In this work, the molecular level proton transport mechanisms in this family of proton conductors (H3PO4, H3PO3 and H3PO2) and some closely related systems (H3PO4-H2O mixtures) are investigated with the help of ab initio molecular dynamics simulations. In fact, neat liquid phosphoric acid has the highest intrinsic proton conductivity of any known substance. Apart from playing a central role in the structure and function of biological systems, systems containing phosphates/phosphonates are attracting an increasing interest as high-temperature electrolytes for emerging fuel cell applications. The results show that strong, mutually polarizable hydrogen bonds give rise to coupled proton motion and a pronounced protic dielectric response of the medium. This allows for the formation of extended, polarized hydrogen bonded (Grotthuss) chains, never truly observed in bulk hydrogen bonded systems. The results show that, in phosphoric acid such chains containing up to five consecutive hydrogen bonds can form. It is the interplay between these chains and a frustrated (there are more proton donor than acceptor sites) hydrogen bond network, which is found to lead to extremely high proton conductivity in phosphoric acid. This strongly contrasts to water, wherein the anomalously high rate of excess charge transport occurs not through extended chains but rather through local hydrogen bond rearrangements that drive individual proton transfer reactions. The mechanism proposed in this work, suggests that strong hydrogen bonding does not necessarily lead to protonic ordering and slow dynamics of the system, demonstrating that weak solvent coupling and sufficient degree of configurational disorder can lead to fast proton transport. Although, phosphonic and phosphinic acids possess even stronger hydrogen bonds, the stronger dipolar and dynamic backgrounds tend to oppose the formation of extended Grotthuss chains. Moreover, these systems do not have the same intrinsically frustrated hydrogen bond network (there are more proton acceptor than donor sites), thus hindering the solvent reorganization (depolarization). Nevertheless, the results show that the weak hydrogen bonded configurations, although not an intrinsic property of the hydrogen bond network, are still forming in a dynamical sense due to liquid disorder. The latter, together with the formation of polarized chains explain the high charge carrier concentrations and conductivities reported in these materials, especially in H3PO3, where they are only slightly lower than in the case of H3PO4. Apparently, proton transport in phosphoric acid is extremely susceptible to nearly all types of chemical perturbations. Apart from the severe conductivity reduction caused by the addition of bases, even the addition of acids leads to some decrease in conductivity. The only dopant that increases the conductivity of H3PO4 is water which, together with some condensation products is already present even in a nominally dry acid under the conditions of thermodynamic equilibrium. In fact, the severe increase in the conductivity of phosphoric acid upon dilution cannot be explained by simple hydrodynamic diffusion of hydronium ion, indicating that proton structural diffusion plays a major role in these systems as well. The results show that very similar molecular mechanisms are at play in phosphoric acid - water system as in neat oxoacid systems. The properties of hydrogen bonds even in 1:1 H3PO4 - H2O mixture are virtually identical to those of pure H3PO4, generally, showing no resemblance to liquid H2O. It is due to the strong and polarizable acid-water hydrogen bonds, that some degree of cooperativity can still be observed in the proton transport mechanism, although the solvent coupling in this case is much stronger due to the significantly different dielectric nature of the water phase. In addition to some vehicular contribution to proton conductivity, water also has some plasticizing effect, increasing the configurational disorder in the hydrogen bond network, therefore resulting in significantly higher conductivities observed in these systems.
  • Thumbnail Image
    ItemOpen Access
    Modeling the rational synthesis of magnesium difluoride via the low-temperature atom beam deposition method
    (2013) Neelamraju, Sridhar; Jansen, Martin (Prof. Dr.)
    A given chemical system, in general, will realize many (meta)stable structures, some of which might be observable on an experimentally viable timescale. Some of these polymorphs could have novel properties waiting to be exploited. However, addressing the problem of directing solid state synthesis towards such unknown polymorphs remains a major challenge. The prediction of new compounds using various theoretical methods is not usually followed up by an actual synthesis and planning the synthesis of novel inorganic solids often requires recourse to theoretical methods that can not only predict the thermodynamic stability of possible structure candidates but also model the kinetic behavior of atoms during the experimental synthesis. Here, we strive to fill this gap between knowledge derived from structure prediction methods and performing the actual synthesis of new structures experimentally by using tools available to the theoretical chemist that include energy landscape search algorithms, ab initio spectroscopy calculations and molecular dynamics simulations with the synthesis of MgF2 via the low-temperature atom beam deposition (LT-ABD) method as the model system. Hence, in a first step, in order to understand the possible cluster modifications of MgF2 that can exist in the vapor phase, we perform global optimizations on neutral and charged clusters using Monte-Carlo simulated annealing and find many possible structures. We also explore the energy landscape of (MgF2)3 and (MgF2)4 using the threshold algorithm in order to be able to estimate the stability and dynamics of these clusters. This method allows us to determine not only the stable and metastable isomers but also the barriers separating these isomers and the probability flows among them, yielding estimates of the stability of all the isomers found. We find that there is reasonable qualitative agreement between the ab initio and empirical potential energy landscapes, and important features such as sub-basins and energetic barriers follow similar trends. However, we observe that the energies are systematically different for the less compact clusters, when comparing empirical and ab initio energies. Furthermore, we employ the same procedure to additionally investigate the energy landscape of the tetramer. For this case, however, we use only the empirical potential due to computational limitations. This is followed by the calculation of Raman and IR spectra including the phonon modes and their intensities, for all the clusters found from the above study. We also calculate IR intensities and phonon modes for all bulk polymorphs of MgF2. This way, we provide the synthetic chemist with a means to observe possible (meta)stable phases of this system in both the vapor phase and the deposit while performing a deposition experiment on MgF2. The calculated data are compared with in-situ measurements in the LT-ABD apparatus. The MgF2 vapor and film are characterized via Raman spectroscopy of the MgF2 gas phase species embedded in an Ar-matrix and of the MgF2-films deposited onto a cooled substrate, respectively. We find that, in the vapor phase, there are mostly monomers and dimers of the neutral and charged species present in our experimental setup. Furthermore, the results suggest that in the amorphous bulk MgF2, rutile-like domains are present and MgF2 clusters similar to those in the matrix. Finally, peaks at about 800 cm-1, which are in the same range as the Ag modes of clusters with dangling fluorine atoms connected to three-coordinated Mg atoms, indicate that such dangling bonds are also present in amorphous MgF2 and can be used to track the amorphous to crystalline transition in this system. Finally, we model the growth of solid MgF2 from the gas-phase on an Al2O3 substrate as it occurs in a real LT-ABD experiment, a hypothetical MgF2-anatase substrate and a MgF2-rutile substrate. The process is studied in all its stages, from the dynamics of MgF2 clusters in the gas phase, over their impact on the surface of the cold and hot substrates, and their diffusion on the substrate, to the formation of crystallites. The growth process was analyzed as a function of synthesis parameters including the substrate temperature, deposition rate and types of clusters deposited. Both high and low rates resulted in the formation of amorphous MgF2 deposits. On annealing, we discovered a possible mechanism for the stabilization of the CaCl2-type structure. We find two competing structures in the first few nanoseconds of the deposition related to the CaCl2 and CdI2 structure types and argue that this competition stabilizes the CaCl2-type structure long enough for experimental observations to take place. Furthermore, the atom arrangements found in our simulations are in good agreement with existing experimental observations based on TEM and XRD measurements, for both the amorphous and the partly ordered metastable phase.