La technologie Spark Plasma Sintering (SPS) permet la conception de matériaux denses avec des microstructures fines. Il s’agit d’une variante du pressage à chaud (HP) qui utilise un courant pulsé pour chauffer la matrice et le matériau. Les phénomènes mis en jeu restent mal compris et sujets à controverse, laissant plusieurs interrogations : - Pourquoi le frittage par SPS apparaît-il plus efficace que les méthodes de frittage sous charge classiques ? Quels sont les mécanismes de densification et de consolidation activés qui déterminent l’élaboration par SPS ? Le passage du courant joue-t-il un rôle dans ces mécanismes et si oui lequel ? Ce travail vise à répondre à ces questions dans le cas de poudres de cuivre sphériques de 10 à 50 µm. Des comparaisons systématiques ont été réalisées avec le pressage à chaud classique, dans des conditions identiques. La cinétique de densification a été étudiée à l’échelle macroscopique et à l’échelle de la microstructure. L’observation de la formation des cous de frittage a été réalisée à partir de fractographies et de sections polies. La densification est assurée par la déformation des particules due à la charge appliquée et à l’augmentation de la température. Aucune différence, ni macroscopique, ni microscopique, n’a été mise en évidence entre l’élaboration par HP et celle par SPS, même lorsque des conditions favorables à la mise en évidence ont été utilisées : couches d’oxyde développée sur les particules, passage du courant forcé dans l’échantillon, fortes intensités appliquées par des « pulses » de courant. Dans les conditions étudiées, il n’apparaît aucun effet spécifique lié au courant. / Spark plasma sintering is a manufacturing process that leads to dense materials with fine microstructures. SPS combines heating and uniaxial load as well as the Hot Pressing (HP) process but the material is heated using a pulsed current. The phenomena occurring during SPS are not fully understood and are still an open point: -Which densification and consolidation mechanisms are involved during SPS? -Why is sintering by SPS more efficient than sintering by traditional ways such as HP? –Does electrical current modify the sintering mechanisms? The aim of this work is to answer these questions in the case of spherical copper powder (from 10 to 50 µm). Comparisons between SPS and HP were performed using the same process conditions. The densification rate was studied macroscopically and microscopically. The evolution of the necks between particles was followed by cross sections and fractography. The densification is realized by plastic deformation due to the applied load and the temperature increase. No difference between SPS and HP was observed although sintering conditions favorable to the occurrence of specific phenomena were applied: oxide layer coating the particles, current forced through the sample, high intensity using a pulsed current. In the studied conditions, no specific effect was observed due to the current presence.

The performance characteristics of pure hydrogen PEMFC (called HFC) stacks made with heterogeneous carbon fiber bipolar plates are studied in this thesis. In addition, the problem that the heterogeneous carbon fiber bipolar plate leaks in the high gas pressure is also solved in this studies so that the new plate can be used to the high current power sources. Because of the gas leakage of the first generation stack at high inlet gas pressure, the fuel supply is insufficient in the high current density. A 4-cell PEMFC stack made with this new bipolar plate is built with weight 370 g and volume 385 cm3 without a fan. The total power out of the 4-cell stack is about 30 W at room temperature. The specific power and volumetric power densities are 81 mW/g and 78 mW/cm3, respectively. The average power density is about 160 mW/cm2, but the power density of a single-cell can reach a value about 220 mW/cm2. The insufficient fuel supply cause that the power density of 4-cell PEMFC stack is lower than single cell, so it is necessary to solve the gas leakage at high pressure. Our experiment found that gas leakage occurs in heterogeneous bipolar plates can be relate to the insufficient or improper hot-pressing temperature, time and pressure while we are making the carbon fiber bunches. So the processes in making new carbon fiber bunches include water expansion, uniform glue adding, high hot-pressing pressure, and using proper temperature and enough solidification time. The airtight of the second generation of heterogeneous carbon fiber bipolar plates improves obviously with the new processes. No leakage occurs for gas pressure under 1atm. We expect that this design can be used to high inlet pressure. It is also quite suitable for various high-power electrical sources.

There are four big medium density fibreboard (MDF) plants in New Zealand with a total production capacity of close to one million cubic meters per year. A significant quantity of boards (nearly 3% or about 30,000 cubic meters per year) is rejected due to defects such as weak core, low modulus of rupture and elasticity, low internal bonding and delamination. The main cause of these defects, is lack of complete understanding of the inter relationship during the hot-pressing stage between the initial inputs such as temperature, moisture content, platen pressure and its impact on the properties of boards. The best solution is to develop a mathematical model to assist in understanding these relationships and to solve the equations in the model by using advanced software. This will reduce the number of expensive experiments and will enable us to see some of the parameters, which are otherwise difficult to visualise. Several earlier researchers have tried to model hot pressing of wood composites, mostly either for particle board or oriented strand board (OSB), and only a few are for MDF. The type of numerical methods used to solve the model equations and various assumptions, changes from one investigator to the other. The non-availability of source code to convert the mathematical equations into programme, is one of the reasons for this model development. To improve the productivity of MDF plants in New Zealand, there was a need to develop a computer programme which can include all the latest findings and can remove the defects which are present in earlier models. This model attempts a more complete integration than in the previous models of all the components such as heat transfer, moisture movement and vertical density profile formation in a one-dimensional model of hot pressing of MDF. One of the important features added in the heat and mass transfer part of the model is that the equilibrium moisture content (EMC) equation given for solid wood was modified to be applicable for the MDF fibres. In addition, this EMC equation can cover the complete range of hot pressing temperature from 160ºC to 200ºC. The changes in fibre moisture content due to bound water diffusion, which was were earlier neglected, was considered. The resin curing reactions for phenol formaldehyde and urea formaldehyde resins are also incorporated into the model, with the energy and water released during the curing reaction being included in the energy and mass balances. The validation of the heat and mass transfer model was done by comparing the values of core temperature and core pressure from the model and the experiments. The experimental value of core pressure and core temperature is obtained by putting a thermocouple and pressure transducer in the middle of the mat. The experimental core temperature results show qualitative agreement with the predicted results. In the beginning, the core temperatures from both experiment and model overlap each other. In the middle of the press cycle, the experimental core temperature is higher by 10ºC and by the end the difference decreases to 5ºC. The vertical density profile (VDP) is a critical determining factor for the strength and quality of MDF panels. The earlier concept of ratio of modulus of elasticity of the layer to the sum of modulus of elasticity of all the layers in the previous time step, given by Suo and Bowyer (1994), is refined with the latest published findings. The equation given by Carvalho et al. (2001) is used to calculate the MOE of different layers of the mat. The differential equation of a Maxwell element given by Zombori (2001) is used to measure stress, nonlinear strain function and relaxation of fibres. The model gives good agreement of peak and core density at lower platen temperature at 160ºC but with the increase of platen temperature to 198ºC, the rise in peak density is comparatively higher. There is a distinct increase in predicted peak density by 150 kg/m³ in comparison to the experimental result, where the increase is only by 10 kg/m³. There is a large decline (50 kg/m³) in core density in the experimental results in comparison to only a slight decline (13 kg/m³) in the predicted results. The use of Matlab provides a very convenient platform for producing graphical results. The time of computation at present is nearly 20 hrs in a personal computer with Pentium four processor and one GB RAM. The model can predict properties of a pressed board for the standard manufacturing conditions and also the new hot pressing technologies such as the use of steam injection or a cooling zone in the continuous press. A comparative study has been done to show the advantages of using new hot pressing technology. The present model will become an important tool in the hands of wood technologist, process engineers and MDF manufacturing personnel, to better understand the internal processes and to improve production and quality of MDF boards. This theoretical model helped in developing better understanding of internal processes. By using it, we can analyse the impact of platen temperature, moisture content on the core temperature, core pressure and density profile. It gives better insight into the relationship between core pressure and delamination of the board. The model is also able to predict the internal changes in the new hot pressing technologies such as the steam injection pressing and the use of a cooling zone in a continuous press. Using the simulation results, the exact time needed for the complete curing of resin can be calculated and then these results can be applied in the commercial plants. If the pressing time is reduced, then the over all production of both batch press and continuous press will increase. The second part of the project is the development of an empirical model to correlate the physical properties from the MDF board to the mean density. The empirical model is simple and straightforward, and thus can be applied in commercial operation for control and optimization. The empirical model can predict peak density, core density, and modulus of rupture, elasticity and internal bonding within the limits in which those relationships are derived. The model gives good results for thickness ranging from 10 to 13.5 mm and density ranging from 485 kg/m³ to 718 kg/m³.

Made available in DSpace on 2014-10-09T12:45:27Z (GMT). No. of bitstreams: 0 / Made available in DSpace on 2014-10-09T13:59:17Z (GMT). No. of bitstreams: 1 07155.pdf: 5285109 bytes, checksum: f414fcad3e4bf9dd06580fa7fa97ce54 (MD5) / Dissertacao (Mestrado) / IPEN/D / Instituto de Pesquisas Energeticas e Nucleares - IPEN/CNEN-SP

Made available in DSpace on 2014-10-09T12:30:00Z (GMT). No. of bitstreams: 0 / Made available in DSpace on 2014-10-09T14:04:31Z (GMT). No. of bitstreams: 1 01046.pdf: 5382760 bytes, checksum: a51a7e739db803d25ebaa7a550e5d8b0 (MD5) / Dissertacao (Mestrado) / IEA/D / University of Illinois Urbana - Champaign, Illinois

Made available in DSpace on 2014-10-09T12:38:20Z (GMT). No. of bitstreams: 0 / Made available in DSpace on 2014-10-09T14:04:43Z (GMT). No. of bitstreams: 1 05682.pdf: 7598679 bytes, checksum: 0a8e36b4d2b6ba00958df86a2415829e (MD5) / Tese (Doutoramento) / IPEN/T / Instituto de Pesquisas Energeticas e Nucleares - IPEN/CNEN-SP

Made available in DSpace on 2014-10-09T12:45:27Z (GMT). No. of bitstreams: 0 / Made available in DSpace on 2014-10-09T13:59:17Z (GMT). No. of bitstreams: 1 07155.pdf: 5285109 bytes, checksum: f414fcad3e4bf9dd06580fa7fa97ce54 (MD5) / Dissertacao (Mestrado) / IPEN/D / Instituto de Pesquisas Energeticas e Nucleares - IPEN/CNEN-SP

Made available in DSpace on 2014-10-09T12:38:20Z (GMT). No. of bitstreams: 0 / Made available in DSpace on 2014-10-09T14:04:43Z (GMT). No. of bitstreams: 1 05682.pdf: 7598679 bytes, checksum: 0a8e36b4d2b6ba00958df86a2415829e (MD5) / Tese (Doutoramento) / IPEN/T / Instituto de Pesquisas Energeticas e Nucleares - IPEN/CNEN-SP

Made available in DSpace on 2014-10-09T12:30:00Z (GMT). No. of bitstreams: 0 / Made available in DSpace on 2014-10-09T14:04:31Z (GMT). No. of bitstreams: 1 01046.pdf: 5382760 bytes, checksum: a51a7e739db803d25ebaa7a550e5d8b0 (MD5) / Dissertacao (Mestrado) / IEA/D / University of Illinois Urbana - Champaign, Illinois

Group IV transition metal carbides are promising materials for high temperature structural application, due to their unique sets of properties such as high melting temperature, high temperature strength, hardness, elastic modulus, wear and corrosion resistance, metal-like thermal and electrical conductivity and thermal shock resistance. This group includes zir-conium carbide, which, along with its composites, are potential candidates for applications such as nose cones for re-entry vehicle, engines, wear resistant parts and in nuclear fuel cladding. Such structural applications demand high strength material with minimal flaws, in order to achieve the required reliability. Attainment of high strength calls for fully dense material with as small a grain size as possible. Producing fully dense zirconium carbide requires very high temperature, which is a direct consequence of its high melting point. Higher processing temperatures increase grain size, thereby also causing a loss in strength, along with the increased cost. Therefore, there is always a driving force to produce such a material in fully densified form at as low a temperature as possible. There have been a number of studies on processing and densification of zirconium carbide. Pressureless sintering of zirconium carbide requires temperature of 2400oC-3000oC to reach reasonably high density. At such high temperatures, abnormal grain growth limits the final density, as pores get entrapped inside the grains. Hot pressing of zirconium carbide also requires upwards of 2000oC to reach high density and is the primary route to produce densified zirconium carbide product. Reactive hot pressing (RHP), is a relatively new processing approach. Here, the reaction between zirconium and carbon to produce zirconium carbide and the densification of the porous mass, occurs simultaneously. Study on reactive hot pressing of zirconium carbide have shown that, it is possible to achieve very high density at much lower temperatures 1600oC. Clearly, reactive processing is an exciting new technique to process zirconium carbide. However, there has been a lack of studies to understand why it provides better densification than conventional hot pressing. Such understanding is of paramount importance, as it can lead to better optimization of RHP and perhaps even lower the process temperature further. The objective of the present study is to understand the densification process in RHP of zirconium carbide through systematic and carefully designed experiments. A model of reactive hot pressing is also constructed to get more insight into the phenomenon. 0.1 Pressureless Reaction Sintering of Zirconium Car-bide Pressureless reaction sintering (RS) of zirconium carbide is studied to understand the role of stoichiometry and zirconium metal in densification. ZrC of four different stoichiometries are chosen for these sets of experiments which are conducted in vacuum at 1200oC and 1600oC for 1 hour to understand the role of stoichiometry. One sample of pure Zr is also sintered to elucidate the role of zirconium in densification. After reaction sintering, all the samples are characterized by density measurement, x-ray diffraction and microstructure, using scanning electron microscopy. After pressureless sintering at 1600oC, zirconium metal reaches the highest relative density of ~ 95%. Densification decreases monotonically with increasing stoichiometry. Zr+0.5C composition reaches the next best relative density (of 90%), while Zr+0.67C composition shows much lower densification. The other two compositions, Zr+0.8C and Zr+C, in contrast, display de-densification rather than densification. Since the pure zirconium sample reaches high density, it can, in principle, help in densification of the mixed powders before getting fully reacted. Non-stoichiometric carbides also exhibit higher diffusivity of carbon, which aids the densification and the greater the deviation from stoichiometry, the smaller the deleterious effects of de-densification from reaction. This troika of factors is responsible for the substantially better densification in non-stoichiometric carbide, compared to stoichiometric carbide. 0.2 Reactive Hot Pressing of Zirconium and Carbon Reactive hot pressing of zirconium carbide is explored with the emphasis on finding the underlying densification mechanism. The earlier proposed densification mechanism for RHP is the plastic flow of transient non-stoichiometric carbide. To differentiate the effect of transient phases from that of zirconium, RHP is carried out at 800oC. At this low temperature, transient phases cannot take part in plastic flow and subsequent densification. Thus, any densi cation at this temperature can be totally attributed to zirconium and the role of zirconium thus can be separated from that of transient phases. A combination of RHP and RS experiments are carried out at 1200oC to better understand the phenomenon. Again, ZrC carbide of four different stoichiometries are investigated in this RHP study. After RHP at 800oC, all the four different ZrC compositions reached more than 90% RD through plastic flow of the Zr leading to a continuous matrix with embedded graphite particles. Since the reaction remains incomplete at this temperature, it is clear that Zirconium alone is responsible for enabling densification at such a low temperature. It is therefore argued that any unreacted Zr would, at higher temperature, be able to drive densification even more. Thus, zirconium does not only participate in densification; it is a dominant factor enabling low temperature densification. Pressureless reaction sintering at 1200oC following the RHP at 800oC, results in de-densi fication, as the reaction between zirconium and carbon occurs with significant volume shrinkage. Since such shrinkage increases with stoichiometry of the carbide, the higher stoichiometry carbides are more susceptible to de-densification. RHP at 1200oC, mostly completes the reaction, but only ZrC0:5 reaches near theoretical density. Thus, the final density of the fully reacted mixture is arrived at through a combination of processes in which the more stoichiometric carbides suffer from not only the smaller metal content but also a greater volume shrinkage during reaction. Thus, ZrC0:5 reaches 99% RD whereas ZrC reaches only 85% RD. The interplay between these two processes may be controlled by a two step RHP begin-ning at 800oC followed by a ramp up to 1200oC. The higher RD achieved at 800 C results in a higher final density for all the four compositions. Thus, two step RHP is a novel way to get better densification in RHP of zirconium carbide. 0.3 Hot Pressing of Zirconium Carbide Powders of Different Stoichiometry In the literature, densification in RHP is mostly attributed to the presence of transient non-stoichiometric carbides. To examine this hypothesis, ZrC of three different stoichiometries are prepared and then subjected to hot pressing at the same temperature and pressure as the previous RHP experiments (i.e. 1200oC and 40MPa for 30 min). After the hot pressing experiments, ZrC0:5 composition shows significant densification (95% RD), whereas ZrC0:67 composition shows very limited densification (70% RD) and ZrC composition shows little or no densification (50% RD). Evidently, the transient phase formed with stoichiometry close to ZrC0:5 can certainly contribute substantially to densification. But for the more carbon-rich compositions, the transient phases do not appear to play a significant role and the benefit of RHP, wherein ZrC can reach 90% RD, must come from the contribution of metal plasticity. 0.4 Reactive Hot Pressing of Zirconium and Zirconium Carbide Two limiting factors for densification during RHP are, de-densification (courtesy of the reaction) and the gradual increase in volume fraction of a rigid, non-sintering phase. To investigate the role of these factors further, two compositions of mixed metal and carbide powders, namely Zr+ZrC and 0.5Zr+ZrC, are subjected to RHP. When reaction is complete, the compositions after RHP will correspond to ZrC0:5 and ZrC0:67, respectively, but with the following difference with respect to the metal-carbon mixtures investigated earlier: these new compositions do not experience de-densification due to reaction and they contain significantly more amount of hard phase (53 and 69%) in the starting composition than their zirconium and carbon mixture counterparts i.e. Zr+0.5C and Zr+0.67C (16 and 20%). These two compositions are subjected to the same process schedules, i.e., RHP at 800oC, pressureless reaction sintering at 1200oC following RHP at 800oC and two step 800oC and 1200oC RHP. After 800oC RHP, Zr+ZrC and 0.5Zr+ZrC compositions reach much lower density than Zr+0.5C and Zr+0.67C compositions as a direct consequence of the larger amount of hard phase hindering densification at the lower temperature. After the 1200oC pressureless sintering following the RHP at 1200oC, the RD of Zr+ZrC and 0.5Zr+ZrC compositions increase (which is opposite to the behaviour of Zr+0.5C and Zr+0.67C com-positions) as they do not su er from reaction derived de-densification. After two step RHP, Zr+ZrC and 0.5Zr+ZrC compositions reach a final RD that is higher than the Zr+0.5C and Zr+0.67C compositions, even though after the first RHP at 800oC, they were much less densified. Thus, the absence of de-densification during reaction is able to more than compensate for the increase in hard phase content. 0.5 Reactive Hot Pressing: Low temperature process-ing route Based on the major factors of densification identified earlier, it was investigated whether RHP temperatures could be brought down further while being supplemented by a free sintering step to complete the reaction without de-densification. From a practical standpoint, such a process would allow dense products to be made by hot pressing with low temperature dies and fixtures while carrying out a more economical pressureless sintering at higher temperatures Therefore, Metal-carbide mixtures, Zr+ZrC and Ti+ZrC, are chosen, along with a temperature of 900 C which is above the allotropic phase transformation temperature for Zr around 880oC, thereby utilizing a zirconium phase that is softer than the hexagonal Zr. For completion of reaction, pressureless reaction sintering is done at 1300oC and 1400oC. It is found that after 1400oC reaction sintering, both the compositions reach almost full density and the Ti+ZrC composition also shows a higher hardness (13 vs 10 GPa) than the Zr+ZrC composition, due to the formation of a binary carbide with consequent solid solution hardening. 0.6 Effect of Particle Size on Reactive Hot Pressing During RHP, premature exhaustion of zirconium by reaction can limit densification. One way to have better densification is to slow down the reaction, so that significant amount of densification takes place before the metal zirconium is exhausted. One way to reduce reaction rates is to increase particle size. Larger particles are expected to slow down the reaction without affecting sintering, as densification is controlled by power law creep of Zr which is grain size independent. Because of lack of availability of Zr with different particle sizes, two different graphite particle sizes, i.e. 7-10 m and 50-60 m, were studied and it was shown that after 1200oC RHP, indeed the larger particle size improves densification. 0.7 Modelling of Reactive Hot Pressing Reactive hot pressing is a complicated phenomenon, and to get an insight and also to optimize the parameters, the availability of a computational model is of paramount importance. Keeping that in mind, a model of RHP has been constructed based on four different parts, namely: 1. Densification of zirconium under pressure 2. Reaction of zirconium and carbon 3. The constraint on sintering from a rigid phase and, finally, 4. The volume contraction during reaction. The model uses published data for the 4 steps and shows reasonable qualitative and quantitative agreement with the experimental results. Further experiments are done with the model to optimize the processing parameters. Results from the virtual experiments consolidates our earlier conviction gained from experimental results, by showing zirconium is the principal factor in densification and exhaustion of zirconium coupled with reaction derived de-densification prevent the higher stoichiometric carbide from achieving full densification. It also shows, RHP gives best densification when reaction is 70-80% complete. So two step RHP where the first RHP will only complete the reaction 70-80%, and a final RHP at temperature which will complete the reaction, will possibly be the way to achieve best densification. 0.8 Conclusions The study on RHP of zirconium carbide led to the following conclusions: • Zirconium plays the most crucial role in densification. • Transient phases only play a role when the final stoichiometry of RHPed carbide is close to that of ZrC0:5. • De-densification from reaction prevents higher stoichiometric carbide from reaching full densification. • Two step RHP, with one RHP at lower temperature at which reaction will remain incomplete, and the other at higher temperature to complete the reaction, yields best densification. • For lower stoichiometric carbide (ZrC0:5,ZrC0:67), full densification can be achieved at 1200oC. For higher stoichiometric carbide, even though large amount of densification upward of 90% RD is achieved at 1200oC, full densification will be out of reach. • RHP shows better densification than conventional hot pressing for all stoichiometries.