Search

International audience; The well-known Villermaux-Dushman system is nowadays widely used for examining the micromixing efficiency either in batch or continuous intensified reactors. However, a bibliographic review shows that kinetic data are too scattered for a reliable determination of the micromixing times. The Dushman reaction kinetics is then reexamined with the use of sulfuric and perchloric acids. The results confirm the fifth-order rate law. More precisely, the I-, H+, and IO3- dependence orders on the rate law are, respectively, 2, 2, and 1, under any condition. To be more consistent with the reactant concentrations used in the Villermaux-Dushman test, we extend their studied range, namely, 1.6 x 10(-3) M

International audience; There are some studies on the synthesis of polyurea microcapsules. However, there is hardly a case where both green solvents and non-toxic isocyanates are used, especially in microfluidics. In this work, an environmentally friendly chemical system of interfacial polymerization (isocyanate: HDB-LV; solvent: octyl salicylate or dibutyl adipate) is tested for the first time to produce polyurea microcapsules. The size of microcapsules is calibrated at 78 μm by microfluidics to quantitatively analyze the relationships among shell thickness, encapsulation efficiency and isocyanate concentrations. The influences of solvent types and reactant concentrations on the shape, morphology and shell thickness of microcapsules are studied. Esters with low water miscibility and low amine concentrations (lower reaction rate) are crucial for the formation of spherical microcapsules. An ester with high water miscibility can diffuse into the continuous phase during encapsulation, which results in broken microcapsules. A high concentration of amine can probably cause cross-linking not only at the interface but also inside the droplet template, which leads to microcapsule deformation. A linear relationship is observed between the shell thickness of microcapsules and the isocyanate concentration. Overall, a high encapsulation efficiency (more than 90%) for octyl salicylate is achieved with polyurea microcapsules.

The common chemical ultraviolet (UV) filters such as octyl salicylate that are used in sunscreen cream may cause irritation and other unknown medical conditions. One possible route to avoid the direct contact of UV filters with human skin is to enclose those filters in microcapsules. In this study, microcapsules with a size calibrated at 78 μm via microfluidics but with different shell thicknesses (0.4 to 2.5 μm) are used for the first time, to study the influence of shell thickness of microcapsules on the UV absorption efficiency. The results show that the shell thickness of microcapsules has no obvious influence on the UV absorption efficiency. Besides, a rough model based on the Beer–Lambert law is used to compare the experimental results.

International audience; Further to the Part I of the present paper, the second Part is concentrated around the VLE modeling of binary mixtures involving the three glycol ethers previously studied experimentally. The authors propose to use the NRTL-PR model for the representation of these non-ideal mixtures. The main difficulties of modelling related to very low vapor pressures and the way of dealing with them are highlighted. The unknown critical parameters for DPM, DPnB and PnB were determined using robust group contribution methods. However, the experimental values of these parameters have never been published before. The main goal of the authors was to obtain the most satisfactory representation of the experimental data provided in the Part I. Some issues that mostly occurred in mixtures involving the PnB as well as in mixtures having very low vapor pressures, were encountered. Nevertheless, we have obtained in general a satisfactory representation of measured points regardless of those issues.

International audience; This work addresses the polygeneration concept integrating concentrating solar power (CSP) with an organic Rankine cycle (ORC) to produce electricity and drinking water by hybrid desalination process combining reverse osmosis (RO) and low-temperature multi-effect distillation (LTMED). Experiments carried out on a bench scale RO pilot led to determine optimal operating parameters as well as options to mitigate the main limiting factors of this technology by hybridizing with LTMED. These data helped to simulate a large scale solar polygeneration plant integrating parabolic trough collectors as CSP technology and a hybrid RO-LTMED system as desalination technology. Various ORC design proposals were simulated and the optimal configuration was pointed out on the basis of thermodynamic criteria (energy efficiency and exergy destruction) and an economic analysis by using two working fluids: an alkane commonly admitted as good candidate and an ester proposed here as green alternative. Results obtained in this work contribute positively to extending the solar polygeneration for desalination and production of energy leading to future sustainable plants.

International audience; An experimental and theoretical study was undertaken for three ether glycols (mixtures of isomers). The present paper, being the first part of a complex study, reports the experimental part of the study using a static apparatus, namely, the vapor-liquid equilibria of dipropylene glycol methyl ether (DPM), dipropylene glycol n-butyl ether (DPnB), and propylene glycol n-butyl ether (PnB), and the vapor-liquid equilibria of their pseudo-binary mixtures composed of (PnB + DPM), (DPM + DPnB), (PnB + DPnB), and the aqueous solutions of each ether. The explored temperature range is between 283.15 and 363.15 K. The experimental points were fitted using the Antoine equation. No literature data was found for comparison. The aim of this paper is not only to provide the obtained experimental data but also to highlight the eventual challenges when working with very low vapor pressure compounds and mixtures. The second part is related to a theoretical study with the development of a thermodynamic model for the representation of the experimental data.

International audience; The CFD simulation of fast reactions in laminar flows can be computationally challenging due to the lack of appropriate sub-grid micromixing models in this flow regime. In this work, simulations of micromixing via the implementation of the competitive-parallel Villermaux/Dushman reactions in a T-micromixer with square bends for Reynolds numbers in the range 60–300 are performed using both a conventional CFD approach and a novel lamellae-based model. In the first, both the hydrodynamics and the concentration fields of the reaction species are determined directly using a finite volume approach. In the second, the hydrodynamic field from the CFD calculations is coupled with a Lagrangian model that is used to perform the chemical reactions indirectly. Both sets of results are compared with previously published experimental data and show very good agreement. The lamellar model has the advantage of being much less computationally intensive than the conventional CFD approach, which requires extremely fine computational grids to resolve sharp concentration gradients. It is a promising solution to model fast chemical reactions in reactors with complex geometries in the laminar regime and for industrial applications.

In cross-flow membrane filtration, fouling results from material deposit which clogs the membrane inner surface. This hinders filtration, which experiences the so-called limiting flux. Among the models proposed by the literature, we retain a simple one: a steady-state reversible fouling is modelled with the use of a single additional parameter, i.e., N d , the ratio of the critical concentration for deposition to the feed concentration at inlet. To focus on fouling, viscous pressure drop and osmotic (counter-)pressure have been chosen low. It results in a minimal model of fouling. Solved thoroughly with the numerical means appropriate to enforce the nonlinear coupling between permeation and concentration polarization, the model delivers novel information. It first shows that permeation is utterly governed by solute transfer, the relevant non-dimensional quantities being hence limited to N d and P e i n , the transverse Péclet number. Furthermore, when the role played by N d and moderate P e i n (say P e i n &lt, 40 ) is investigated, all results can be interpreted with the use of a single non-dimensional parameter, F l , the so-called fouling number, which simply reads F l ≡ P e i n N d - 1 . Now rendered possible, the overall fit of the numerical data allows us to put forward analytical final expressions, which involve all the physical parameters and allow us to retrieve the experimental trends.

International audience; A novel experimental method and its associated model are proposed for the simultaneous determination of membrane solute and solvent permeabilities, which are essential transport parameters of reverse osmosis models used for process simulation. The method utilizes a single bench-scale batch apparatus consisting of two stirred half-cells containing solutions of different concentrations separated by a membrane across which coupled non-steady-state solute diffusion and solvent osmosis take place countercurrently in the absence of transmembrane pressure difference. Results are presented from days-long determinations of the water and sodium chloride permeabilities of Filmtec BW30 and NF270 membrane samples for initial transmembrane salt concentration differences ranging from 1 g L-1 to 35 g L-1. When used as input parameters for the simulation of pilot reverse osmosis desalination tests, the osmotic-diffusive salt permeabilities approximated the experimental rejection rates.

International audience; This work studies how deeply the reagents choice influences micromixing characterisation by the Villermaux-Dushman method, when applying it to a 1 litre stainless steel standard vessel with two baffles, stirred by an inclined blade turbine. For the first time, borate and phosphate buffer are compared on their use in the method. It is observed that triiodide production is higher when borate buffer is used. Moreover, perchloric acid leads to higher triiodide production than sulphuric acid, when injecting the same concentration of both acids. Finally, the influence of the ionic strength is also studied, since there has been a great deal of controversy about it over the last years. The results show that the ionic strength affects triiodide production, although relatively slightly. Advice concerning the choice of the reagents is given in conclusion.

International audience; Membrane solute permeability plays a role in the buildup of concentration polarization in pressure-driven crossflow filtration processes, and thus in the determination of the permeate flux, solute rejection, retentate flux and concentration. We numerically examine reverse-osmosis desalination with membranes of fixed solvent permeability, but of variable selectivity with respect to the solute. The study highlights an intricate coupling between retentate and filtrate properties. In particular, it reveals that, for given values of solute permeability and feed concentration, there is a maximum operating pressure that optimizes solute rejection regardless of the feed salinity. The conditions leading to this and to other peculiar behaviors for permeation fluxes and concentrations are identified.

International audience; An accurate 2-D numerical model that accounts for concentration polarization and osmotic effects is developed for the cross-flow filtration in a membrane channel. Focused on the coupling between laminar hydrodynam-ics and mass transfer, the numerical approach solves the solute conservation equation together with the steady Navier-Stokes equations under the Prandtl approximation, which offers a simplified framework to enforce the non-linear coupling between filtration and concentration polarization at the membrane surface. The present approach is first validated thanks to the comparison with classical exact analytical solutions for hydrodynamics and/or mass transfer, as well as with approximated analytical solutions that attempted at coupling the various phenomena. The effects of the main parameters in cross-flow reverse osmosis (RO) or nanofiltration (NF) (feed concentration, axial flow rate, operating pressure and membrane permeability) on streamlines, velocity profile, longitudinal pressure drop, local permeate flux and solute concentration profile are predicted with the present numerical model, and discussed. With the use of data reported from NF and RO experiments, the Prandtl approximation model is shown to accurately correlate both average permeate flux and local solute concentration over a wide range of operating conditions.

Supercritical water oxidation (SCWO) development is limited by corrosion and salt precipitation. Transpiring wall reactors are emerging to cope with these problems. A reactor has been developed in which the inner porous shell is composed of pure α-alumina, to handle organic effluents generated by nuclear activities. The reactor was not proven to be efficient enough to oxidize salty effluents. However, experimental results concerning the oxidation of a mixture of dodecane and tributyl phosphate, which is used as a model effluent, confirmed the ability of the reactor to treat corrosive wastes. High destruction rates were actually achieved (>98%). Phosphorus was totally recovered in the aqueous effluent as phosphoric acid. No corrosion was noticed in the reactor, except upstream from the waste injector. As expected, the inner alumina tube shielded the pressure vessel from corrosion. The assumed sensibility of alumina to thermal gradients was not a limiting factor of the reactor operation.

Supercritical water oxidation (P>22.1 MPa, T>647 K) is an efficient process to treat hazardous organic compounds: high destruction rates (>99.9%) with no NOx production, rapidity and a good confining of the reaction. Its performances are limited by salt precipitation and corrosion. A new reactor has then been developed to solve these problems. It consists in a concentric double wall reactor in which the corrosive reactants are maintained inside an alumina porous tube whereas pressure resistance is ensured by a stainless steel external vessel. A water flow through the porous pipe prevents sticky solid particles from depositing on the wall. The performances of this reactor were investigated. At 723 K and 25 MPa, the destruction of methanol, used as a model compound, reached 99.9%. Due to high thermal gradients generated by the exothermic reaction, the pipe which plays an important role in the decrease of salt precipitation and corrosion can be broken. Thus, its behaviour must be controlled in-situ. The pressure drop measurement across the porous wall was used to check whether the inner pipe was still intact. In fact, the experimental results show that supercritical water flow through the tube follows the Darcy's law. The experiments also confirm that the porous medium permeability is a characteristic constant of this medium. The permeability value remains equal whichever the fluid used, liquid water or supercritical water. In addition, the pressure drop measurement across the porous wall allows the control of the tube integrity.

This study essentially deals with the description and modeling of transport phenomena in ultrafiltration (UF). In the light of new and published experimental results obtained with solutions of dextran T500, the realism and validity of the most classical models such as the gel-polarization model, the osmotic pressure model, the resistance-in-series model have been first discussed. Experimental results are not in agreement with the first two models. Moreover, despite several modifications, the resistance-in-series model does not predict the permeate flux in a general way. The common failure of these simple models is to average all parameters (permeate flux, solute concentration) over the whole membrane length. As a result, it is impossible to use them to predict the influence of the membrane length. We have then developed a new two-dimensional model based on the numerical resolution of the convective and diffusion equations. Permeate fluxes predicted are in good agreement with experimental results and the respective influence of the velocity, the initial concentration, the transmembrane pressure and the membrane length are very well described.

In a previous paper (Part I), the operating conditions of the iodide–iodate reaction system used to quantify micromixing phenomena in chemical reactors and mixers has been presented. In order to determine the micromixing time through models, the complete kinetics of the reactions must be known. This paper presents the kinetics of the Dushman reaction obtained in operating conditions close to the micromixing test conditions. The results confirm the single fifth-order process widely reported in published work and establish the rate constant as a function of ionic strength ranging between 0 and 2 M.

This paper describes in a very practicable way the operating conditions of the iodide–iodate reaction system used to quantify micromixing phenomena in chemical reactors and mixers. Several operating procedures are given for batch and continuous mixers. Recommendations to adapt these procedures to any specific mixers and complete analysis procedures are also provided. Finally, a simple micromixing model is used to calculate micromixing time tm. It is checked that tm is non-dependent on the reactant concentrations at constant hydrodynamic conditions. In a second following paper (Part II), the reactions kinetics, essential for a knowledge of micromixing time tm, are presented.

The use of the iodide-iodate test reaction was extended to study micromixing in glycerin-water mixtures with viscosities up to 170 mPa s. The properties of the mixtures and the kinetics of the test reaction were determined as a function of viscosity. Experiments were carried out in a stirred tank yielding the micromixedness ratio α = (1−X s ) X s as a function of stirring speed for different viscosities. A micromixing model involving two steps (incorporation of bulk fluid and exchange between smaller eddies) has been used for the interpretation of the experimental data. Further work is required to interpret experiments at higher viscosities where hydrodynamic data are still lacking.

International Conference of the European-Desalination-Society on ă Desalination for the Environment, Clean Water and Energy, Barcelona, ă SPAIN, APR 23-26, 2012; International audience; A mathematical model was developed to predict super saturation along ă reverse osmosis modules (RO) for water desalination. This model is based on conservation principles and chemical equilibrium equations for concentrated solutions. Pitzer's model was used for the activity coefficient calculations. An average rejection rate for each ionic species was also considered. Supersaturations with respect to all calcium carbonate forms and to calcium sulfate are calculated. The model allows assessing when scale is likely to occur along the RO modules. The results for two brackish water qualities and seawater are shown.

International audience

International audience; Supercritical water oxidation offers a viable alternative treatment to destroy the organic structure of Ion Exchange Resins. In order to design and define appropriate dimensions for the supercritical oxidation reactor, a 2D simulation of the fluid dynamics and heat transfer during the oxidation process has been investigated. The solver used is a commercial code, Fluent® 6.3. The turbulent flow field in the reactor, created by the stirrer is taken into account with a k−ω model and a swirl imposed to the fluid. Particle trajectories are modelled with the Discrete Random Walk Particle Model. For the solubilization of the particles in supercritical water, a mechanism has been proposed and implemented into Fluent® software through the Eddy Dissipation Concept approach, taking into account the identified rate determining species. Simulation results provide results on the flow, temperature fields and oxidation localization inside the reactor. For the reactive particles-supercritical water flow model, the effect of parameters, such as feed flow rates or stirring velocity, can be focussed. Reaction temperature is predicted with deviation lower than 15%. Degradation conversions are in good agreement with experimental ones.

A leaking duct carries a flow that is a characteristic of several applications. Devices for cross-flow microfiltration are composed of a duct, the walls of which are semi-permeable membranes. In subsurface irrigation, the walls of watering pipes can be of porous clay, whereas the watering hoses are riddled with holes or made of porous material, in surface drip irrigation. In these applications, the first approach consists of assuming that the flow concerns a pure fluid (as in a microfiltration system operating at a very low species concentration), and that the wall’s leakage depends only on the local pressure difference between both inner and outer sides of the wall. Actually, only the situation with fixed uniform permeation at the walls is clearly understood. In the case of a 2D plane channel, the configuration is referred to as Berman flow. Berman theory predicts the existence of a critical transverse Reynolds number, denoted R t i s o (with R t i s o ≈ 1.31 ), at which both viscous and inertial effects compensate to maintain the flow isobaric. A recent contribution has extended the Berman theory to non-uniform leakage depending linearly on local pressure, and focused on the situation where R i n , the transverse Reynolds number at the inlet, is smaller than R t i s o . In contrast, the present work investigates the case R i n > R t i s o , for which the leakage is responsible for a pressure enhancement inside the channel. Now, if the transpiration increases with the local pressure, we are faced with a possible runaway in pressure. This phenomenon (generally unwanted in the applications) is here analysed. The particular study of the dead-end channel configuration shows that multiple flows can be found.

International audience; Spent ion exchange resins (IERs) are radioactive process wastes for which there is no satisfactory industrial treatment. Supercritical water oxidation offers a viable treatment alternative to destroy the organic structure of resins, used to remove radioactivity. Up to now, studies carried out in supercritical water for IER destruction showed that degradation rates higher than 99% are difficult to obtain even using a catalyst or a large oxidant excess. In this study, a co-fuel, isopropanol, has been used in order to improve degradation rates by initiating the oxidation reaction and increasing temperature of the reaction medium. Concentrations up to 20 wt% were tested for anionic and cationic resins. Total organic carbon reduction rates higher than 99% were obtained from this process, without the use of a catalyst. The influence of operating parameters such as IERs feed concentration, nature and counterions of exchanged IERs were also studied.

International audience; Spent ion exchange resins are radioactive process wastes for which there is no satisfactory industrial treatment. Supercritical water oxidation could offer a viable treatment alternative to destroy the organic structure of resins and contain radioactivity. IER degradation experiments were carried out in a continuous supercritical water reactor. Total organic carbon degradation rates in the range of 95-98% were obtained depending on operating conditions. GC-MS chromatography analyses were carried out to determine intermediate products formed during the reaction. Around 50 species were identified for cationic and anionic resins. Degradation of polystyrenic structure leads to the formation of low molecular weight compounds. Benzoic acid, phenol and acetic acid are the main compounds. However, other products are detected in appreciable yields such as phenolic species or heterocycles, for anionic IERs degradation. Intermediates produced by intramolecular rearrangements are also obtained. A radical degradation mechanism is proposed for each resin. In this overall mechanism, several hypotheses are foreseen, according to HOOradical dot radical attack sites.

International audience; Concentration polarization affects numerous systems of membrane separation, and combined with osmosis, it can cause substantial reductions in permeation. We establish an exact solution to the conservation law of a solute advected by Berman flow. This flow is characteristic of reverse osmosis or nanofiltration. The resulting concentration polarization is then combined with the osmosis (counter-) effect. For large Péclet number of permeation, it yields a rigorous support to the semi-empirical "film" model, and accounts for the limit flux phenomenon. The main results are summarized in a simple diagram that relates three different Péclet numbers, and show that polarization combined with osmosis can withstand operating pressure almost totally.

International audience; Supercritical water oxidation is an innovative and very efficient process to treat hazardous organic waste. In order to better understand the complex physic phenomena involved in this process, and to design more efficient reactors or to insure future efficient scale-up, a simulation with the Computational Fluid Dynamics software FLUENT was carried out for a simple tubular reactor. The turbulent non-reactive flow is well-represented using the $\kappa-\epsilon$ model. Nevertheless, the $\kappa-\epsilon$ model gives better results when a source term is added to take into account the chemical reaction. Two approaches are used to model the reaction rate: an Arrhenius law and the Eddy Dissipation Concept (EDC) generally used to describe combustion reactions. The results of this simulation using Arrhenius law, are in good agreement with experimental data although a simple thermohydraulic model was used. Moreover, the sensitiveness to the inlet temperature has been demonstrated. It influences the reaction start-up and the shape of the measured wall temperature peak. Equally, the simulated temperature profiles using Eddy Dissipation Concept model are in good agreement with experimental ones. Hence, the two approaches give similar results. Nevertheless, the EDC model predicts more precisely the thermal peak location at the reactor wall.

To carry out controlled precipitation on a practical scale, the use of a two impinging jets mixer is explored in terms of its ability to deliver rapid micromixing. A system of parallel competing reactions developed by Villermaux and co-workers is used to characterise the micromixing level in the device as a function of feed flow rate and of its position in the reactor. The characteristic time for micromixing decreases with the feed flow rate and reaches values as small as 4 ms. Powers dissipated in the device are deduced from the experiments using a simple micromixing model and correlate well with those calculated using computational fluid dynamics (Phoenics). Finally, the energy dissipation rate in impinging jets is compared with the local energy dissipation rate of the impeller. Some rules of thumb for scale up are given.

The present thesis covers the eco-design of a dry cleaning machine by integration of a membrane process for solvent dehydration. This solvent belongs to the family of “green” solutions and has been developed to replace the hazardous perchloroethylene. After washing cycles, the solvent contains fibers, colorants and water. The presence of water represents an issue because relatively high water concentrations (about 10%wt) in solvent can shrink the textiles. Hence, the solvent should imperatively be dehydrated. The current dry cleaning machines are equipped with a simple one-stage distillation. Due to this limitation, the water can not be efficiently separated from the solvent. As an alternative solution, an implementation of the membrane process was proposed.The experimental investigation was carried out for the membrane process. It has proven the feasibility of the solvent dehydration and has demonstrated an extremely high separation and energy efficiency. In order to characterize water-solvent mixture, an in-depth thermodynamic study was performed. The experimental and theoretical results allowed to obtain valuable phase equilibria data that have not previously been documented. These data have served to correctly characterize the membrane transport and to estimate all the required parameters for further dehydration stage design., La présente thèse porte sur l’éco-conception d’une machine de nettoyage à sec par intégration d’un procédé membranaire pour la déshydratation du solvant. Ce solvant, appartenant à la famille des « solvants verts », a été développé afin de remplacer le toxique perchloroéthylène. Après les cycles de lavage, le solvant contient des différentes fibres, colorants ainsi que de l’eau. Lorsque la teneur en eau est relativement grande (aux alentours de 10%wt), le solvant provoque le rétrécissement des textiles et il est primordial de le déshydrater. Les machines actuelles de nettoyage à sec sont équipées par un distillateur batch. Dû à ses limitations, il n’est pas possible de séparer l’eau et le solvant de manière efficace. Dans ce contexte, un procédé membranaire a été proposé en tant qu’une solution alternative.Pour ce procédé membranaire, une étude expérimentale a été réalisée. Lors de cette étude, la faisabilité de la déshydratation du solvant a été prouvée. De plus, une très haute efficacité énergétique et de séparation a été observée. Afin de caractériser le mélange eau-solvant, une étude thermodynamique a été également menée. Les résultats expérimentaux et théoriques ont permis d’obtenir les données importantes des équilibres de phases qui n’ont pas été publiées auparavant. Ces données ont servi pour caractériser le transport membranaire ainsi qu’estimer tous les paramètres indispensables pour la conception de l’étape de déshydratation.

ITER (International Tokamak Experimental Reactor) is a fusion machine which should demonstrate scientific and technological feasibility of fusion energy by means of D-T fusion reaction. Therefore, most of the solid radioactive waste produced during operation and dismantling phase (around 34000 tons) will result not only from activation by 14 MeV neutrons, but also from contamination by tritium. One of the main issues in tritiated waste management is the confinement of tritium which presents a good ability to diffusion. One of the solutions is to trap the tritium directly in waste drums. In containers tritium is under gaseous form (HT and T2), tritiated water vapor (HTO and T2O) and organic bounded tritium species (OBT). As an hydrogen isotope, HT and T2 trapping and conversion is possible thanks to a reaction with a mix of metal oxides MnO2 and Ag2O, which can be used for hydrogen hazards mitigation. An experimental study was conducted at the CEA on the study of tritium trapping by a mixture of 90% of manganese oxide and 10% of silver oxide. The tests showed that the addition of Pt and Pd catalysts did not improve the trapping capacity of the powder mixture, such as impregnation of the powder mixture when preparing the mixture, with solutions of KOH or NaOH. Crystal-chemical analysis revealed the formation of a mixed oxide in the preparation of powders, questioning the mechanisms previously established. Two new mechanisms have been proposed and a model on the trapping kinetics was presented. The results of modeling the competition between the trapping phenomenon and the diffusion of tritium through the wall of the waste package showed that the trapper decreased the value of the quantity of tritiated hydrogen degassed from the package, Le tritium, isotope radioactif de l’hydrogène est actuellement utilisé dans le cadre d’études sur la fusion nucléaire. Avec la mise en fonctionnement de la machine ITER (International Tokamak Experimental Reactor) à l’horizon 2020, la quantité de déchets radioactifs tritiés augmentera fortement. De par sa mobilité le tritium est difficile à confiner dans un fût de déchets et les critrères d’acceptabilité de l’ANDRA (Agence Nationale pour la gestion des Déchets Radioactifs) étant stricts pour ce qui concerne le taux de dégazage des colis de déchets tritiés, il est nécessaire de trouver des solutions adaptées pour le stockage de ces déchets radioactifs tritiés. L’une des solutions est de piéger le tritium au sein des fûts de déchets. Ce dernier peut se trouver sous trois formes dans les colis : sous forme d’eau tritiée HTO ou T2O, sous forme gazeuse HT ou T2 ou lié à la matière organique (Organic Bounded Tritium) OBT. Une étude expérimentale a été effectuée au CEA sur l’étude du piégeage du tritium par un mélange de 90% d’oxyde de manganèse et de 10% d’oxyde d’argent. Les tests ont montré que l’ajout de catalyseurs tels Pt et Pd ainsi que l’imprégnation du mélange de poudres lors de la préparation du mélange, par des solutions de KOH ou NaOH n’amélioraient pas la capacité de piégeage du mélange de poudres. Des analyses cristallochimiques ont mis en évidence la formation d’un oxyde mixte lors de la préparation des poudres, remettant en cause les mécanismes établis précédemment. Deux nouveaux mécanismes ont été proposés et un modèle comportemental sur la cinétique de piégeage a été présenté. Les résultats de la modélisation de la compétition entre le phénomène de piégeage et la diffusion du tritium à travers la paroi du colis de déchet ont montré que le piégeur diminuait bien la valeur du flux en hydrogène tritié dégazé du colis.