Today, with increasing global awareness and regulation of air pollution, interest in the smog-abating property of photocatalytic materials is increasing. Titanium dioxide (TiO2) is the most well known photocatalytic semiconductor and is often considered as one way of solving pollution by a passive but an effective way, particularly to reduce atmospheric nitrogen oxides (NOx=NO+NO2). This relatively new technology is already being used in some of the countries as a construction material, commercially sold as photocatalytic cement, photocatalytic pavement, self-cleaning tiles, and self-cleaning glass. Prior research has examined the photocatalytic properties of TiO2 itself, as well as TiO2-containing cement-based materials. The majority of this effort has been on characterizing and enhancing the photocatalytic efficiency. However, relatively little research was performed to assess the potential impact of the photocatalytic reaction on the "parent" or "host" material. In this research, the focus is on the effect of photocatalysis on the composition, structure, and properties of cementitious materials, which contain titania nanoparticles at early and late ages. Fundamental examinations on the addition of these chemically non-reactive nanoparticles to cement-based materials are performed. The high surface area of nanoparticles could alter early age properties of cementitious materials, such as setting time, dimensional stability, and hydration rate. Various experimental techniques as well as mathematical modeling were used to examine and explain the early age hydration of cementitious materials when TiO2 nanoparticles are present. Further, the effects of the TiO2 on the long term durability of cement-based materials are investigated to demonstrate their suitability for long-term use in the field. The photocatalytic NOx oxidation efficiency and NOx binding capability of TiO2 containing cementitious materials are experimentally investigated. The durability of TiO2-cement is examined by various techniques on samples that went through extensive photocatalysis and environmental exposures. These investigations have led to tentative conclusions on the use of TiO2 nanoparticles in cementitious materials, and suggest avenues for future study.

abstract: Early-age cracks in fresh concrete occur mainly due to high rate of surface evaporation and restraint offered by the contracting solid phase. Available test methods that simulate severe drying conditions, however, were not originally designed to focus on evaporation and transport characteristics of the liquid-gas phases in a hydrating cementitious microstructure. Therefore, these tests lack accurate measurement of the drying rate and data interpretation based on the principles of transport properties is limited. A vacuum-based test method capable of simulating early-age cracks in 2-D cement paste is developed which continuously monitors the weight loss and changes to the surface characteristics. 2-D crack evolution is documented using time-lapse photography. Effects of sample size, w/c ratio, initial curing and fiber content are studied. In the subsequent analysis, the cement paste phase is considered as a porous medium and moisture transport is described based on surface mass transfer and internal moisture transport characteristics. Results indicate that drying occurs in two stages: constant drying rate period (stage I), followed by a falling drying rate period (stage II). Vapor diffusion in stage I and unsaturated flow within porous medium in stage II determine the overall rate of evaporation. The mass loss results are analyzed using diffusion-based models. Results show that moisture diffusivity in stage I is higher than its value in stage II by more than one order of magnitude. The drying model is used in conjunction with a shrinkage model to predict the development of capillary pressures. Similar approach is implemented in drying restrained ring specimens to predict 1-D crack width development. An analytical approach relates diffusion, shrinkage, creep, tensile and fracture properties to interpret the experimental data. Evaporation potential is introduced based on the boundary layer concept, mass transfer, and a driving force consisting of the concentration gradient. Effect of wind velocity is reflected on Reynolds number which affects the boundary layer on sample surface. This parameter along with Schmidt and Sherwood numbers are used for prediction of mass transfer coefficient. Concentration gradient is shown to be a strong function of temperature and relative humidity and used to predict the evaporation potential. Results of modeling efforts are compared with a variety of test results reported in the literature. Diffusivity data and results of 1-D and 2-D image analyses indicate significant effects of fibers on controlling early-age cracks. Presented models are capable of predicting evaporation rates and moisture flow through hydrating cement-based materials during early-age drying and shrinkage conditions. / Dissertation/Thesis / Ph.D. Civil and Environmental Engineering 2011

La carbonatation est une pathologie du béton armé qui peut engendrer la corrosion des armatures et à terme de la fissuration. Dans le cadre de la gestion des déchets radioactifs, les structures et les conteneurs seraient soumis simultanément à un échauffement (exothermie des déchets), au CO2 atmosphérique, ainsi qu'à un séchage important. Afin de rendre compte de leur évolution à l'échelle séculaire, les données actuelles relatives à la carbonatation à température ambiante doivent être complétées, d'une part par une description de la phénoménologie en température, et d'autre part, par la prise en compte de l'impact des transferts hydriques en température (séchage) sur la carbonatation en insaturé. Le travail présenté se focalise sur l’étude de la durabilité de quatre pâtes de ciment différentes dont deux sont directement dérivées des formulations de référence sélectionnées par l’Andra (CEM I et CEM V) et un mélange baspH. Le premier volet est dédié à l'étude des transferts hydriques reposant sur la conduite d'essais de désorption. Il a notamment permis la caractérisation des isothermes de désorption en température (20, 50et 80°C). L'impact thermique modifie les isothermes : la teneur en eau à l’équilibre chute avec la température et le point d'amorçage de la condensation capillaire est déplacé. Une phase de modélisation a conjointement été conduite en support aux expérimentations. L'utilisation de l'équation de Clausius-Clapeyron a permis de décrire l'effet de la température sur les isothermes (par la détermination de la chaleur isostérique d’adsorption de chacun des matériaux). Bien que l'impact de la température sur la microstructure des pâtes de ciment soit avéré, la prise en compte du déplacement des équilibres thermodynamiques suffit à restituer cet effet thermique sur les isothermes. La détermination des perméabilités intrinsèques par exploitation des cinétiques de désorption (par analyse inverse) a montré la thermoactivation du transport d'eau. La contribution de l'évolution de la microstructure en température ne peut-être négligée sur la perméabilité des matériaux.Le deuxième volet exploratoire, est consacré à l'étude de la carbonatation en température. Il repose sur la mise en place d'un dispositif de carbonatation spécifique (fonctionnement en température) et la conduite d'essais de carbonatation à HR et température contrôlées. La campagne de caractérisation (DRX et ATG)a conduit à l'obtention de profils de carbonatation caractéristiques de chaque couple (HR, Température).Les évolutions minéralogiques (décomposition des hydrates, distribution polymorphique du carbonate de calcium précipité) mises en évidence en température se rapprochent de celles identifiées à température ambiante. En revanche, il ressort que les conditions environnementales influencent significativement les proportions polymorphiques : plus l'HR est faible, plus les teneurs en phases métastables (vatérite,aragonite) sont élevées. Les réactions de dissolution-précipitation mises en jeu dans la transformation polymorphique (des états métastables vers la calcite) sont inhibées à faible HR, par manque de milieu réactionnel. La cinétique de carbonatation, également impactée par les conditions environnementales, est régie par la concurrence de l'effet thermique sur les transferts hydriques et sur la solubilité rétrograde des réactifs. Les profondeurs carbonatées sont maximales aux points d'amorçage de la condensation capillaire propres aux différents matériaux et à chaque température. Les profondeurs carbonatées augmentent avec la température jusqu'à une température limite, caractéristique de la formulation, au-delà de laquelle la solubilité rétrograde des réactifs deviendrait le facteur limitant.Cette phase de compréhension des mécanismes mis en jeu dans la carbonatation en température et de leur niveau de couplage effectuée, les modèles prédictifs de carbonatation en insaturé pourront être étendus à l'application en température. Les résultats de ce travail fournissent les données d'entrées et de validation nécessaires à la validation des simulations numériques. / Carbonation is the major cause of degradation of reinforced concrete structures. It leads to rebar corrosion and cracking of the concrete cover. In the framework of radioactive waste management, cement-based materials used as building material for structures or containers would be simultaneously submitted to heating (due to the waste thermal output), subsequent drying and atmospheric carbon dioxide. Such environmental conditions are expected to modify the carbonation mechanisms (with respect to temperature). In order to describe their long-term evolution of material, a double approach was developed, combining the description of carbonation and drying for temperatures up to 80°C to complement available data at ambient temperature. The present work focuses on the durability study off our hardened cement pastes; two of them are derived from the reference formulations selected by Andra(CEM I and CEM V) and a low-pH mix. The first experimental campaign focuses on moisture transfer. The effect of temperature on drying is investigated through water vapour desorption experiments. The first desorption isotherms of four hardened cement pastes was characterized at 20, 50 and 80°C. The results show a significant influence of the temperature. For a given relative humidity (RH) the water content equilibrium is always reduced temperature is increased and the starting point of capillary condensation is shifted towards higher RHs. The experimental campaign is complemented through modelling activities. The impact of temperature on the first desorption isotherms is effectively described using the Clausius-Clapeyron equation(characterization of the isosteric heat of adsorption). The intrinsic permeability to water is evaluated through inverse analysis by reprocessing the experimental weight loss of initially saturated samples submitted to constant environmental conditions. The intrinsic permeability appears to increase with temperature in relation to the observed microstructure evolution (porosity coarsening).The environmental conditions impact is studied using preconditioned samples (12 different RHs and 20,50 and 80°C) and accelerated carbonation tests. The latter are performed in a new device allowing accurate control of the environmental conditions as well as the carbon dioxide concentration. The carbonated depths and the mineralogical modifications induced by carbonation are assessed using XRDand TGA for each temperature and RH. Most of the mineralogical modifications notified in temperature(hydrates consumption and nature of crystallographic phase of calcium carbonate) are similar with these identified at ambient temperature. Yet the results show a significant influence of the environmental conditions on calcium carbonate polymorphic abundance: the lower the RH, the more abundant the metastable phases (vaterite and aragonite).The rate of the polymorphic transformation (from the metastable states into calcite by dissolution precipitation)is believed to decrease with RH because of lack of liquid water. A significant influence of the environmental conditions on the carbonation rate is also observed. It depends of the competition between the temperature effect on moisture transfer and retrograde solubility of reactants. Carbonation depths appear to be maximal at the RH-starting point of capillary condensation of each material and temperature. Carbonation depths increase with temperature until a limit of temperature characteristic of the material. Above this temperature, reactants solubility might control the main process.

Fibras vegetais de baixo módulo de elasticidade são conhecidas pela sua capacidade de aumentar a energia absorvida durante o carregamento dos materiais cimentícios, especialmente no estágio pós-fissurado. A utilização de nanofibras celulósicas pode contribuir para a tenacificação de matrizes frágeis, tanto por melhorar o empacotamento das partículas, com o refinamento de poros, quanto pela interceptação de fissuras na escala nanométrica, com a respectiva absorção de energia. A celulose nanofibrilada provém de um recurso natural, abundante e renovável, possui bom desempenho mecânico e superfície específica elevada, o que contribui para melhorar a adesão entre as partículas de cimento. Estes fatores justificam o uso da celulose nanofibrilada e a tornam uma boa alternativa como nanoreforço de materiais cimentícios. Com isso, o objetivo deste trabalho foi a produção de celulose nanofibrilada a partir de polpa organossolve de bambu, definindo a melhor condição para sua produção e posterior utilização como reforço em compósitos híbridos (reforçados na nano e micro escalas) em comparação a compósitos reforçados somente com microfibras (polpa) pelos processos de produção por sucção e prensagem, e extrusão. A celulose nanofibrilada foi produzida utilizando-se polpa não-branqueada e branqueada, por meio de 5, 10, 15 e 20 ciclos de nanofibrilação pelo processo grinding. Foram realizados testes químicos, físicos e mecânicos para definição da condição ótima de nanofibrilação. A celulose nanofibrilada não-branqueada produzida mediante 10 ciclos foi definida como a melhor opção para utilização nos compósitos híbridos, por possuírem maior módulo de elasticidade e, em razão da sua maior estabilidade estrutural, apresentam maior resistência à degradação em meio alcalino. Os compósitos foram submetidos à cura por carbonatação acelerada para mitigação da degradação da fibra pela diminuição do pH da matriz e também para refinamento dos poros. Os compósitos foram submetidos ao teste de envelhecimento acelerado por meio de 200 ciclos de imersão e secagem para análise da sua degradação. Os compósitos híbridos e reforçados somente com polpa aos 28 dias de cura e após o envelhecimento acelerado foram submetidos aos ensaios físicos, mecânicos e microestruturais para acompanhamento do efeito da celulose nanofibrilada nas suas propriedades. Nos compósitos produzidos pelos dois processos aos 28 dias não houve diferença estatística para as propriedades físicas testadas, comparando-se os compósitos híbridos e os reforçados somente com polpa. No processo de sucção e prensagem, embora útil para ajustes na formulação e na cura do compósito híbrido, não se percebeu contribuição estatisticamente significativa da celulose nanofibrilada na formação de pontes de transferência de tensões, e, portanto sem o correspondente aumento na resistência mecânica dos compósitos. Nos compósitos extrudados, a celulose nanofibrilada atuou de modo a melhorar o comportamento mecânico do compósito híbrido em comparação ao compósito sem nanofibras. Esta melhoria pode estar associada à maior adesão entre as nanofibrilas e a matriz cimentícia, o que foi atestado pela análise microestrutural (MEV) dos compósitos. Após o envelhecimento acelerado os compósitos com e sem nanofibras produzidos pelos dois processos não apresentaram redução do desempenho mecânico, o que se atribui à menor alcalinidade provida pela carbonatação acelerada. / Low elastic modulus vegetable fibers are known for their ability to increase the energy absorbed by cement based materials while they are loaded, especially during the post-crack stage. The use of cellulose nanofibers may contribute for toughening of brittle matrices and improving particle packing by both pore refining and crack intercepting at nanoscale, with the corresponding energy absorption. Nanofibrillated cellulose comes from a natural, abundant and renewable resources, it has good mechanical peformance and high specific surface, which contributes to improve the adhesion between the cement particles. These factors justify the use of nanofibrillated cellulose and give rise to an alternative nanoreinforcement for cement based materials. Thus, the aim of this work was the production of the nanofibrillated cellulose from bamboo organosolv pulp, establishing the best condition for its production and subsequent use as reinforcement in hybrid composites (both nano and micro-scale reinforcement) compared to composites reinforced with only microfibers (pulp), produced by the slurry vacuum dewatering followed by pressing and extrusion methods. The nanofibrillated cellulose was produced submitting unbleached and bleached pulps to 5, 10, 15 and 20 nanofibrillated cycles by the grinding method. Chemical, physical and mechanical tests were carried out to define the optimal condition to nanofibrillation. The unbleached nanofibrillated cellulose produced by 10 cycles was defined as the best option to be used in hybrid composites, since their greater modulus of elasticity and, because of their greater structural chemical stability, higher resistance to degradation in alkaline environments. The composites were subjected to accelerated carbonation curing process to mitigate thedegradation of fiber by reducing the matrix pH and also to refine the pores. The composites were subjected to accelerated aging process by means of 200 wet and dry cycles to assess their degradation. The hybrid composites and the composites reinforced only with pulp at 28 days and after accelerated aging were subjected to physico-mechanical and microstructural tests to study the effect of the nanofibrillated cellulose on their properties. There was no difference in the physical properties of the hybrid composites and composites reinforced with only pulp, produced by the two processes at 28 days. For the slurry vacuum dewatering followed by pressing process, although useful for adjustments in the formulation and cure hybrid composite, there was no statistically significant contribution of the nanofibrillated cellulose in the formation of stress transfer bridges, and therefore without a corresponding increase in the mechanical strength of the composites. For the extruded composites, the nanofibrillated cellulose improved the mechanical behavior of the hybrid composite compared to the composite without nanofiber. This improvement may be associated with greater adherence between the nanofibrils and the cement matrix, which was confirmed by microstructural analysis (SEM) of the composites. After accelerated aging, the composites with and without nanofibers produced by the two processes showed no reduction in mechanical performance, which is attributed to the lower alkalinity provided by the accelerated carbonation.

Fibras vegetais de baixo módulo de elasticidade são conhecidas pela sua capacidade de aumentar a energia absorvida durante o carregamento dos materiais cimentícios, especialmente no estágio pós-fissurado. A utilização de nanofibras celulósicas pode contribuir para a tenacificação de matrizes frágeis, tanto por melhorar o empacotamento das partículas, com o refinamento de poros, quanto pela interceptação de fissuras na escala nanométrica, com a respectiva absorção de energia. A celulose nanofibrilada provém de um recurso natural, abundante e renovável, possui bom desempenho mecânico e superfície específica elevada, o que contribui para melhorar a adesão entre as partículas de cimento. Estes fatores justificam o uso da celulose nanofibrilada e a tornam uma boa alternativa como nanoreforço de materiais cimentícios. Com isso, o objetivo deste trabalho foi a produção de celulose nanofibrilada a partir de polpa organossolve de bambu, definindo a melhor condição para sua produção e posterior utilização como reforço em compósitos híbridos (reforçados na nano e micro escalas) em comparação a compósitos reforçados somente com microfibras (polpa) pelos processos de produção por sucção e prensagem, e extrusão. A celulose nanofibrilada foi produzida utilizando-se polpa não-branqueada e branqueada, por meio de 5, 10, 15 e 20 ciclos de nanofibrilação pelo processo grinding. Foram realizados testes químicos, físicos e mecânicos para definição da condição ótima de nanofibrilação. A celulose nanofibrilada não-branqueada produzida mediante 10 ciclos foi definida como a melhor opção para utilização nos compósitos híbridos, por possuírem maior módulo de elasticidade e, em razão da sua maior estabilidade estrutural, apresentam maior resistência à degradação em meio alcalino. Os compósitos foram submetidos à cura por carbonatação acelerada para mitigação da degradação da fibra pela diminuição do pH da matriz e também para refinamento dos poros. Os compósitos foram submetidos ao teste de envelhecimento acelerado por meio de 200 ciclos de imersão e secagem para análise da sua degradação. Os compósitos híbridos e reforçados somente com polpa aos 28 dias de cura e após o envelhecimento acelerado foram submetidos aos ensaios físicos, mecânicos e microestruturais para acompanhamento do efeito da celulose nanofibrilada nas suas propriedades. Nos compósitos produzidos pelos dois processos aos 28 dias não houve diferença estatística para as propriedades físicas testadas, comparando-se os compósitos híbridos e os reforçados somente com polpa. No processo de sucção e prensagem, embora útil para ajustes na formulação e na cura do compósito híbrido, não se percebeu contribuição estatisticamente significativa da celulose nanofibrilada na formação de pontes de transferência de tensões, e, portanto sem o correspondente aumento na resistência mecânica dos compósitos. Nos compósitos extrudados, a celulose nanofibrilada atuou de modo a melhorar o comportamento mecânico do compósito híbrido em comparação ao compósito sem nanofibras. Esta melhoria pode estar associada à maior adesão entre as nanofibrilas e a matriz cimentícia, o que foi atestado pela análise microestrutural (MEV) dos compósitos. Após o envelhecimento acelerado os compósitos com e sem nanofibras produzidos pelos dois processos não apresentaram redução do desempenho mecânico, o que se atribui à menor alcalinidade provida pela carbonatação acelerada. / Low elastic modulus vegetable fibers are known for their ability to increase the energy absorbed by cement based materials while they are loaded, especially during the post-crack stage. The use of cellulose nanofibers may contribute for toughening of brittle matrices and improving particle packing by both pore refining and crack intercepting at nanoscale, with the corresponding energy absorption. Nanofibrillated cellulose comes from a natural, abundant and renewable resources, it has good mechanical peformance and high specific surface, which contributes to improve the adhesion between the cement particles. These factors justify the use of nanofibrillated cellulose and give rise to an alternative nanoreinforcement for cement based materials. Thus, the aim of this work was the production of the nanofibrillated cellulose from bamboo organosolv pulp, establishing the best condition for its production and subsequent use as reinforcement in hybrid composites (both nano and micro-scale reinforcement) compared to composites reinforced with only microfibers (pulp), produced by the slurry vacuum dewatering followed by pressing and extrusion methods. The nanofibrillated cellulose was produced submitting unbleached and bleached pulps to 5, 10, 15 and 20 nanofibrillated cycles by the grinding method. Chemical, physical and mechanical tests were carried out to define the optimal condition to nanofibrillation. The unbleached nanofibrillated cellulose produced by 10 cycles was defined as the best option to be used in hybrid composites, since their greater modulus of elasticity and, because of their greater structural chemical stability, higher resistance to degradation in alkaline environments. The composites were subjected to accelerated carbonation curing process to mitigate thedegradation of fiber by reducing the matrix pH and also to refine the pores. The composites were subjected to accelerated aging process by means of 200 wet and dry cycles to assess their degradation. The hybrid composites and the composites reinforced only with pulp at 28 days and after accelerated aging were subjected to physico-mechanical and microstructural tests to study the effect of the nanofibrillated cellulose on their properties. There was no difference in the physical properties of the hybrid composites and composites reinforced with only pulp, produced by the two processes at 28 days. For the slurry vacuum dewatering followed by pressing process, although useful for adjustments in the formulation and cure hybrid composite, there was no statistically significant contribution of the nanofibrillated cellulose in the formation of stress transfer bridges, and therefore without a corresponding increase in the mechanical strength of the composites. For the extruded composites, the nanofibrillated cellulose improved the mechanical behavior of the hybrid composite compared to the composite without nanofiber. This improvement may be associated with greater adherence between the nanofibrils and the cement matrix, which was confirmed by microstructural analysis (SEM) of the composites. After accelerated aging, the composites with and without nanofibers produced by the two processes showed no reduction in mechanical performance, which is attributed to the lower alkalinity provided by the accelerated carbonation.

A period of cement hydration is one of critical in the life span of concrete structures. One of the reasons of collapse of concrete structural elements during and after construction is the error in determining the concrete characteristics at early age. Recently, non-destructive test emerged as a popular way to evaluate the properties of cement-based materials. This test offers continuous measurements of concrete properties as well as ability to monitor any changes in the current state of structural materials. In the existing research, some of these methods fail to capture well properties of the materials in the plastics stage. A new piezoelectric pulse testing device (Piezoelectric Ring Actuators Technique), (P-RAT ) was initially developed at the University of Sherbrook as a non-destructive test (NDT ) for soil. This technique is considered a completely new, versatile, advanced and accurate. The development of the new technique (P-RAT ) was done on two main bases: the first was the development of piezoelectric ring actuators set-up and the second is the development of the interpretation method. The setup is composed of two main units; emitter and receiver, and is capable of measuring shear and compression wave velocities in specimens. With this technique, many problems of pulse tests, which make interpretation of results difficult and ambiguous, were solved in soil. The P-RAT overcomes wave reflections at boundaries (end-caps and sides), sample disturbance, weak shear coupling between soil and device (interaction) as well as the fixation problems, low resonant frequency and limited input voltage of the existing device. The previous method is exploited forward to measure the hydration properties of cement-based material. To apply this test method, it is necessary to determine how the evolution of shear wave velocity can be related and sensitive to the hydration of cement-based materials. Validation of the P-RAT with four conventional test setups that can be used to monitor early setting and hydration of cement-based materials is carried out. These tests include penetration resistance to monitor initial and final setting respectively, calorimetric to monitor heat of hydration, electrical conductivity to monitor change in continuity of the pore structure and compressive strength at 24 hours. The phase one of this investigation included trial tests to investigate the possibility of employing the original setup used for soil (P-RAT ) to determine setting and hardening properties of cement-based material. Based on the results of the preliminary test, two modifications were conducted to the previous test device to fit with cement based material and to obtain adequate resonant frequency for cement-based materials. These modifications are the design of the container and changing the dimensions of the rings. The resultant version of P-RAT after the modification was referred to be as P-RAT2 . Calibration of the P-RAT2 with water specimen was undertaken using the compression wave velocity and resulted in 99.33% accuracy. One paste mixture was tested three times to determine the experimental error of the P-RAT2 . The repeatability carried out on the P-ART2 proved the ability of this setup to capture accurate results of the shear wave velocity. This relative error is limited to 9 %. A number of series of validation was performed on cement paste and mortar mixtures proportioned with various water cement ratios (w/cm ) as well as chemical admixtures. The w/cm ratio ranged between (0.35 and 0.50). The investigated chemical admixtures comprise of high-range water-reducing agent, viscosity-modifying agent, set-accelerating agent, and set-retarding agent. The presented validations examine the ability of a P-RAT2 to monitor the hydration of the cement-based materials. The hydration is characterized by setting time, heat of hydration, electrical conductivity, and compressive strength at 24 hours. The results obtained using the P-RAT2 was correlated to those obtained using the conventional tests and strength measurement. The results enable to validate the ability of P-RAT2 to accurately detect variations in the hydration of cement-based materials. In addition, the initial and final time of setting can be determined from the derivation of velocity vs. time curve. The results show that conductivity, resistivity, has a bilinear relationship to shear wave velocity. The compressive strength at 24 hours was correlated to both the shear wave velocity and shear modulus obtained using the P-RAT2 . Furthermore, analytical model was derived to estimate the w/cm in mortar mixture by measuring the shear wave velocity (V[subscript s] ) and the corresponding time (t )||Résumé : La période d'hydratation du ciment est l'une des périodes clé du cycle de vie des structures en béton. L'une des raisons de l'effondrement d'éléments structuraux en béton pendant et après la construction peut être attribuée à une détermination des caractéristiques au jeune âge erronée. Depuis quelques années, l’auscultation des structures est devenue une méthode très populaire pour évaluer les propriétés des matériaux cimentaires. Cette méthode permet d'obtenir les propriétés du béton en continue et possibilité un suivi de changements dans I'état des matériaux structuraux. Dans I'état actuel de la recherche dans ce domaine, certaines méthodes ne sont pas adéquates pour bien mesurer les propriétés des matériaux à I'état plastique. Un nouvel appareil d'essai à impulsions piezoélectriques (Piezoelectric Ring Actuators technique), (P-RAT) a initialement été développé à l’Université de Sherbrooke comme technique d'auscultation des sols. Cette technique est considérée complètement nouvelle, polyvalente, évoluée et précise. Le développement de cette nouvelle technique (P-RAT) a été effectué en deux volets : la première sole est le développement du dispositif de vérin de commande annulaire piezoélectrique et le deuxième est le développement d'une méthode d'interprétation. Le dispositif d'essai est composé de deux unités principales, un émetteur et un récepteur. Et permet de mesurer la vitesse de propagation des ondes de cisaillement et de compression. Grâce à cette technique, plusieurs des problèmes associés aux dispositifs d'essais par impulsion des ondes, qui rendent les résultats ambigus et difficiles à interpréter, ont été résolu pour les sols. Le dispositif P-RAT surmonte les problèmes de réflexion des ondes aux limites (embouts et côtés), la perturbation de l’échantillon, le couplage de cisaillement faible entre le sol et l'appareil (interaction) ainsi que les problèmes de fixation, la faible résonnance des fréquences et le voltage d'entrée limité du dispositif. La méthode décrite a été utilisée pour mesurer les propriétés d'hydratation des matériaux cimentaires. Pour pouvoir appliquer cette méthode, il faut déterminer comment l’évolution de la propagation des ondes de cisaillement peut être reliée à l'hydratation des matériaux cimentaires et être sensible à ces dernières. La validation de la méthode P-RAT est réalisée, à l'aide de quatre configurations conventionnelles que l’on peut utiliser pour faire le suivi de la prise et de l'hydratation des matériaux cimentaires. Ces essais consistent à la résistance à la pénétration afin de pouvoir déterminer la prise initiale et finale, la calorimétrie pour suivre l’evolution de la chaleur d'hydratation, la conductivité électrique pour effectuer le suivi de la structure des pores et la résistance à la compression à 24 heures. La phase 1 de l'étude comprend des essais pour évaluer la possibilité d'utiliser la configuration originale utilisée pour les sols (P-RAT) pour déterminer les propriétés de prise et de durcissement des matériaux cimentaires. Selon les résultats des essais préliminaires, deux modifications ont été effectuées à l'appareil original pour permettre son utilisation avec des matériaux cimentaires et pour obtenir une fréquence de résonnance raisonnable sur les matériaux cimentaires. Les modifications effectuées sont la conception du contenant et un changement de la dimension des anneaux. La version modifiée du P-RAT est designée P-RAT2. La calibration du P-RAT2 à l'aide d'échantillon liquide sous propagation d'ondes de compression a été menée, avec des résultats d'une précision de 99,33 %. Un mélange a été testé trois fois pour déterminer l'erreur expérimentale du P-RAT2. La répétitivité des essais sur le PART2 a démontré la capacité de cet appareil à produire des résultats de cisaillement de propagation des ondes de cisaillement très précis. L'erreur relative se limite à 9 %. Une série d'essais de validation a été menée sur des mélanges de pâte de ciment et de mortier de rapport eau/ciment variés (e/c) ainsi qu'avec des adjuvants. Le rapport e/c variait entre 0,35 et 0,50. Les adjuvants utilisés étaient des supers plastiflants (réducteur d'eau), des agents de viscosité, des agents accélérateurs de prise et des agents retardateurs de prise. Les validations présentées ont pour but de valider la capacité du P-RAT2 à suivre l'hydratation des matériaux cimentaires. L'hydratation est caractérisée par le temps de prise, la chaleur d'hydratation, la conductivité électrique et la résistance à la compression à 24 heures. Les résultats obtenus à l'aide du P-RAT2 ont été comparés à ceux obtenus à l'aide d'essais de mesure des caractéristiques physiques et de résistance traditionnels. Ces résultats permettent de valider la capacité du P-RAT2 à détecter avec précision les variations dans l'hydratation des matériaux cimentaires. De plus, le dispositif P-RAT2 peut avoir une correction avec mesure obteniez avec les appareils traditionnels. II est aussi possible de déterminer le temps de prise initial et final à l'aide d'une courbe de propagation vs le temps. Les résultats montrent que la conductivité et la résistivité ont une relation bilinéaire à la propagation des ondes de cisaillement. La résistance à la compression à 24 heures a été comparee à la fois à la propagation des ondes de cisaillement et au module de cisaillement obtenus avec le P-RAT2. De plus, un modèle analytique a été établi pour estimer le rapport e/c dans le mélange de mortier en mesurant la propagation des ondes de cisaillement (V) correspondant au temps (t).

L'objectif de ce travail est d'obtenir, par homogénéisation périodique, des modèles macroscopiques de transfert hydrique dans les milieux poreux partiellement saturés à partir des équations de transfert de l'eau liquide et de vapeur d'eau écrites à une échelle microscopique. La dimensionnalisation des équations fait apparaître naturellement des nombres sans dimension caractérisant les problèmes de transfert hydriques dans les milieux partiellement saturés. Nous nous sommes intéressés à trois différents régimes de transfert (diffusion de vapeur prédominante, couplage diffusion/convection, convection de l'eau liquide prédominante). Pour chaque modèle homogénéisé, nous avons obtenu une expression différente du tenseur de diffusion hydrique homogénéisé. Nous avons ensuite calculé les tenseurs de diffusion hydrique homogénéisés obtenus dans les deux régions hygroscopique et super-hygroscopique, sur des géométries plus ou moins complexes décrivant la microstructure en 2D et 3D. Des comparaisons avec des valeurs expérimentales ont été ensuite effectuées. Pour finir, une résolution numérique de l'équation de transfert hydrique macroscopique homogénéisée a été effectuée en se basant sur les données expérimentales d'un béton BHP. / We propose in this work to construct, by periodic homogenization, macroscopic models of moisture transfer in unsaturated porous media. To do this, the liquid water and water vapor transport equations are averaged from the microscopic scale. The dimensional analysis of transport equations naturally lets appear dimensionless numbers characterizing the moisture transfer in unsaturated porous media. Three different transfer regimes are addressed (predominant water vapor diffusion, coupling diffusion / convection, predominant liquid water convection). For each transfer regime, the associated homogenized moisture diffusion tensor has a different expression. Then, the homogenized moisture diffusion tensors are calculated in both hygroscopic and super-hygroscopic regions on several geometries with varying complexity, describing 2D and 3D microstructures. Comparisons with experimental values are also addressed. Finally, based on experimental data of a BHP concrete, a numerical resolution of the homogenized macroscopic moisture transfer equation is performed.

La modélisation numérique du comportement des matériaux cimentaires sous l’action des sollicitations complexes constitue un point de vue alternatif pour identifier et évaluer les mécanismes internes qui ne peuvent être étudiés directement à travers les essais expérimentaux. A cet effet, le développement des outils de modélisation permettant la prise en compte des interactions entre la microstructure hétérogène de la pâte de ciment et le comportement macroscopique des matériaux cimentaires est fortement apprécié. Une telle approche numérique donne une meilleure description des processus physiques et évite la calibration répétitive des paramètres lorsque la microstructure change.Ce travail de thèse a pour objet de mettre au point une approche de modélisation de l’endommagement des matériaux cimentaires compte tenu des mécanismes physiques qui se produisent à l’échelle microscopique. Dans l’approche proposée, les principes de la construction d’une microstructure virtuelle de la pâte de ciment sont présentés et les paramètres micromécaniques des phases cimentaires sont identifiés. La capacité prédictive de l’approche est testée en comparant les résultats numériques aux résultats des essais expérimentaux réalisés dans ce travail et aux résultats tirés de la littérature. L’application de cette approche est ensuite illustrée à travers des simulations de la pâte de ciment sous fluage et cycles de gel-dégel. Cette approche ouvre la voie à une multitude d’applications, comme l’étude de l’effet du retrait, du fluage, des cycles de gel-dégel, de la fissuration thermique, de l’auto-cicatrisation et de la carbonatation sur les propriétés thermomécaniques des matériaux cimentaires. / Numerical modelling of the constitutivebehaviour of cementitious materials exposed to aggressive environment offers an alternative point of view for the identification and assessment of internal mechanisms which cannot be explicitly explored using standard experimental techniques. In this regard, the development of advanced modelling tools that take into account the interactions between the heterogeneous microstructure of cement paste and the macroscopic behaviour cementitious materials is highly valued. Such modelling approaches give a much better description of the physical processes and avoid recurrent parameter calibration when dealing with a different microstructure.The work presented in this PhD thesis proposes a numerical modelling approach of damage in cement based materials taking heed of the physical mechanics that can only be characterized at the microscopic level. In the proposed approach, the principles of constructing a virtual microstructure of cement paste are laid out and the micromechanical parameters of cement phases are identified. The predictive capacity of the micromechanical approach is put to the test by a comparison of numerical results with experimental data determined in the present study and found in the literature. Finally, the power of the approach is illustrated through simulations of creep and freeze-thaw behaviour at the microscopic scale of cement paste.This approach paves the way for a multitude of applications, such as the study of the effect of shrinkage, creep, freeze-thaw cycles, thermal cracking, self-healing and carbonation on the thermomechanical properties of cement-based materials.

La sûreté du stockage des déchets radioactifs repose en grande partie sur la capacité de confinement du conteneur et de l’ouvrage qui lui sont dédiés. Dans le cas des déchets radioactifs de Faible et Moyenne Activité à Vie Courte (FMA-VC), cette propriété de confinement, assurée par des matrices solides à base de matériaux cimentaires, est mesurée sur la base d’un essai de diffusion naturelle, consistant à faire traverser un traceur, de type radioactif, dans un échantillon représentatif, duquel on retire, a posteriori, un coefficient de diffusion. L’évolution de ces matériaux et les améliorations apportées à la fabrication de nouvelles enveloppes de confinement induisent des durées d’essai pouvant atteindre plusieurs années.L’objectif premier de ce travail consiste en la détermination du coefficient de diffusion effectif d’une espèce de référence, qui sera dans notre cas l’eau tritiée, dans un intervalle de temps réduit. L’approche théorique repose sur la compréhension du phénomène de transport d’espèces ioniques en solution soumises à un champ électrique. Sur les bases d’un protocole expérimental clairement établi et de la définition du facteur de formation, caractéristique topologique intrinsèque du réseau poreux, il a été possible de déterminer le coefficient de diffusion effectif de l’eau tritiée d’une gamme de bétons et de mortier d’intérêt, et cela en seulement quelques heures.Dans un second temps, la comparaison de l’essai de migration sous champ électrique constant, développé dans ce travail, avec l’essai classique de diffusion naturelle à l’eau tritiée a mis en exergue deux points cruciaux. La non-prise en compte de la décroissance radioactive du tritium dans l’interprétation de l’essai de diffusion naturelle à l’eau tritiée amène une sous-estimation non-négligeable de la valeur du coefficient de diffusion. La conservation, lors de leur phase de maturation, des matériaux d’essais dans des conditions inadaptées, induit des mesures selon la technique de migration sous champ électrique constant disparates et non répétables.Dans un dernier temps, la validation complète de la technique électrocinétique, sujet initial de ce travail, repose sur la vérification des hypothèses théoriques énoncées au préalable. Le facteur de formation, et a fortiori le coefficient de diffusion effectif, est une grandeur indépendante de la force ionique de la solution porale du matériau considéré, cela pour une gamme de solution courante dans le domaine des matériaux cimentaires. Le facteur de formation s’avère aussi indépendant de l’amplitude du champ électrique appliqué, aussi pour une gamme et des durées d’essai adaptées aux conditions de mesures définies dans le protocole. Enfin, lorsque les valeurs de coefficients de diffusion effectifs obtenues en migration et diffusion naturelle à l’eau tritiée sont comparées sur plusieurs formulations maîtrisées de matériaux cimentaires / The safety and the reliability of a radioactive waste repository rely essentially on the confinement ability of the waste package and the storing structure. In the case of the low-level and intermediate level short-lived radioactive waste, the confinement property, relying on solid matrices made of cement-based materials, is assessed through a natural diffusion test, using a radioactive tracer, from which an effective diffusion coefficient is deduced. The evolution of the materials and more particularly the enhancement of the confinement properties of cement-based materials lead to test duration from a couple of months to a couple of years.The main objective of the present work involves the determination of the effective diffusion coefficient of reference chemical species, in our case the tritiated water, within a shorter time. The theoretical foundation is based on the description of ionic species mass transfer under the effects of an electrical field. With the definitions of a precise experimental protocol and of a formation factor, considered as an intrinsic topological feature of the porous network, it is possible to determine the effective diffusion coefficient of tritiated water for various types of concretes and mortars, and this within a few hours only.The comparison between the developed accelerated test, based on the application of a constant electrical field, and the normed natural diffusion test, using tritiated water, underlined two critical issues. First, omitting the impact of the radioactive decay of tritium during a natural diffusion test, leads to a non-negligible underestimation of the effective diffusion coefficient. Second, maintaining samples in high relative humidity conditions after casting is essential in order to avoid contrasted and unrelated results when performing the electrokinetic tests.Eventually, the validation of the electrokinetics technique, main objective of this work, rests on the assessment of the theoretical hypothesis previously formulated. The formation factor, as well as the effective diffusion coefficient, does not depend on the ionic strength of the material pore solution, this being validated for solutions of different composition encompassing the cement materials pore solution diversity. The formation factor also does not vary when the amplitude of the applied electrical field varies, provided both the test duration and the electrical field amplitude are kept within acceptable boundaries. Finally, the comparison between the values of the effective diffusion coefficient obtained with both the constant field migration test and the natural diffusion techniques, for perfectly conditioned and prepared materials, leads us to invalidate the assumption that the effects of the double electrical layer are negligible

Les structures en béton des centrales nucléaires peuvent être soumises à des contraintes thermo- hydriques modérées, caractérisées par des températures de l’ordre de la centaine de degrés aussi bien en conditions de service qu’accidentelles. Ces contraintes peuvent être à l’origine de désordres importants notamment la fissuration qui a pour effet d’accélérer les transferts hydriques dans la structure. Dans le cadre de l’étude de la durabilité de ces structures, le modèle THMs a été développé au Laboratoire d’Etude du Comportement des Bétons et des Argiles (LECBA) du CEA Saclay pour simuler le comportement du béton face à des sollicitations couplées thermo-hydro-mécaniques. Dans cette thèse on s’est intéressé à l’amélioration dans le modèle THMs d’une part de l’estimation des paramètres mécaniques et hydromécaniques du matériau en conditions partiellement saturées et en présence de fissuration et d’autre part de la description de la fissuration. La première partie a été consacrée à la mise au point d’un modèle basé sur une description multi-échelle de la microstructure des matériaux cimentaires, en partant de l’échelle des principaux hydrates (portlandite, ettringite, C-S-H, etc.) jusqu’à l’échelle macroscopique du matériau fissuré. Les paramètres investigués sont obtenus à chaque échelle de la description par des techniques d’homogénéisation analytiques. Dans la seconde partie on s’est attaché à décrire numériquement de façon précise la fissuration notamment en termes d’ouverture, de localisation et de propagation. Pour cela une méthode de réanalyse éléments finis/éléments discrets a été proposée et validée sur différents cas-test de chargement mécanique. Enfin la procédure a été mise en œuvre dans le cas d’un mur chauffé et une méthode de recalcul de la perméabilité a été proposée permettant de montrer l’intérêt de la prise en compte de l’anisotropie du tenseur de perméabilité lorsqu’on s’intéresse à l’étude des transferts de masse dans une structure en béton fissurée. Mots clés : matériaux cimentaires, homogénéisation, modélisation multi-échelle, microfissures, éléments discrets, éléments finis, chargements thermo-hydro-mécaniques. / The concrete structures of nuclear power plants can be subjected to moderate thermo-hydric loadings characterized by temperatures of the order of hundred of degrees in service conditions as well as in accidental ones. These loadings can be at the origin of important disorders, in particular cracking which accelerate hydric transfers in the structure. In the framework of the study of durability of these structures, a coupled thermo-hydro-mechanical model denoted THMs has been developed at Laboratoire d’Etude du Comportement des Bétons et des Argiles (LECBA) of CEA Saclay in order to perform simulations of the concrete behavior submitted to such loadings. In this work, we focus on the improvement in the model THMs in one hand of the assessment of the mechanical and hydromechanical parameters of the unsaturated microcracked material and in the other hand of the description of cracking in terms of opening and propagation. The first part is devoted to the development of a model based on a multi-scale description of cement-based materials starting from the scale of the main hydrated products (portlandite, ettringite, C-S-H etc.) to the macroscopic scale of the cracked material. The investigated parameters are obtained at each scale of the description by applying analytical homogenization techniques. The second part concerns a fine numerical description of cracking. To this end, we choose to use combined finite element and discrete element methods. This procedure is presented and illustrated through a series of mechanical tests in order to show the feasibility of the method and to proceed to its validation. Finally, we apply the procedure to a heated wall and the proposed method for estimating the permeability shows the interest to take into account an anisotropic permeability tensor when dealing with mass transfers in cracked concrete structures. Keywords : cement-based materials, homogenization, multi-scale modelling, microcracks, discrete elements, finite elements, thermo-hydro-mechanical loadings.