Search

International audience; Coating complex surfaces by functional amorphous silica films for new applications including energy harvesting and health depends on the operating range and robustness of their deposition process. In this paper, we propose a new kinetic model for the atmospheric pressure chemical vapor deposition of SiO2 films from TEOS/O2/O3 valid in the 150−520 °C temperature range, thus allowing for treating thermally sensitive substrates. For this, we revisit reported chemical schemes in Computational Fluid Dynamics simulations considering original experimental data on the deposition rate of SiO2 films from a hot-wall reactor. The new model takes into account for the first time a thermal dependency of the direct formation of SiO2 from TEOS and O3 and yields excellent agreement in both shape and value between experimental and calculated local deposition rate profiles. The model provides non-measurable information such as local distributions of species concentration and reaction rates, which are valuable for developing optimized CVD reactor designs. Original solutions for the introduction of the reactants are proposed, to uniformly coat complex and/or large parts at a wide temperature range.

International audience; This work presents experimental data on the kinetic of steam gasification of biomass char in a fluidized bed reactor at atmospheric pressure. The char was obtained from fast pyrolysis of cylindrical beech stick in an annex batch fluidized bed reactor at 650 °C. The experiments were performed for temperatures in the range of 700–850 °C and steam partial pressures between 0.1 and 0.7 bars. The results showed that the char steam gasification is done in two steps. First, a thermal degradation of char takes place just after its introduction in the fluidized bed. This step leads to a partial conversion of the reactive solid and a formation of volatile products. Then, the new carbonaceous residue reacts with steam to produce syngas. The effect of hydrogen partial pressure up to 0.25 bars on the kinetic of char gasification was also investigated. The results revealed that the presence of H2 inhibits the reaction of steam gasification. Besides, interactions also occur between char and H2 which lead to the formation of CH4 in the product gas. Moreover, as the hydrogen partial pressure increases, the methane production also increases. Finally, a modelling kinetic study including the inhibiting effect of H2 showed that experimental data can be well-represented by the Shrinking Core Model with an activation energy and a reaction order with respect to steam equal to 123 kJ/mol and 0.62, respectively.

International audience; In this study, 3D numerical simulations using an Eulerian n-fluid approach of a gas–solidfluidized bed composed of very dense particles of tungsten (19,300 kg.m−3) were carriedout to examine the behavior of this suspension, especially the effects of the reduction ofthe fluidization column diameter on the fluidization quality. Tungsten was selected as asurrogate material of U(Mo) (Uranium molybdene) which is of interest for new nuclear fuelswith limited enrichment. Comparisons between experiments and computations for the axialpressure profile of a 5 cm diameter column demonstrate the capability of the mathematicalmodels of the NEPTUNE CFD code to simulate the fluidization of this powder located out-side the classification of Geldart. The numerical results show that the mobility into the bedof these very dense particles is very low. The reduction of the fluidization column diam-eter from 5 cm to 2 cm does not have significant effect on the local solid circulation butstrongly decreases the axial and radial mixing of the particles due to wall-particles frictioneffects. These results confirm and allow to better understand the wall effects experimentally evidenced.

International audience; Nature-inspired chemical engineering (NICE) is promising many benefits in terms of energy consumption, resilience and efficiencyetc.but it struggles to emerge as a leading discipline, chiefly because of the misconception that mimicking Nature is sufficient. It is not, since goals and constrained context are different. Hence, revealing context and understanding the mechanisms of nature-inspiration should be encouraged. In this contribution we revisit the classification of three published mechanisms underlying nature-inspired engineering, namely hierarchical transport network, force balancing and dynamic self-organization, by setting them in a broader framework supported by nonequilibrium thermodynamics, the constructal law and nonlinear control concepts. While the three mechanisms mapping is not complete, the NET and CL joint framework opens also new perspectives. This novel perspective goes over classical chemical engineering where equilibrium based assumptions or linear transport phenomena and control are the ruling mechanisms in process unit design and operation. At small-scale level, NICE processes should sometimes consider advanced thermodynamic concepts to account for fluctuations and boundary effects on local properties. At the process unit level, one should exploit out-of-equilibrium situations with thermodynamic coupling under various dynamical states, be it a stationary state or a self-organized state. Then, nonlinear phenomena, possibly provoked by operating larger driving force to achieve greater dissipative flows, might occur, controllable by using nonlinear control theory. At the plant level, the virtual factory approach relying on servitization and modular equipment proposes a framework for knowledge and information management that could lead to resilient and agile chemical plants, especially biorefineries.

International audience; Modelling and experimental validation of solvent recovery from an aroma mixture by batch distillation process is presented, with particular emphasis of the effect of the prediction of the physicochemical properties and the phase equilibrium data on the content of the aroma compounds in the distillate cuts. The illustrative case study refers to an industrial batch distillation to recover dimethyl carbonate (DMC) from an extract generated by a solvent extraction process involving variable natural raw materials for tailored perfume and fragrance applications. Due to the high number of aroma compounds in natural extracts, a synthetic mixture containing six target aroma compounds (alpha-pinene, eucalyptol, linalool, cis-3-hexenol, fenchone and benzyl acetate) was mixed with DMC for the modelling and experimental studies of the batch distillation process. The methodology is tested through the simulation of the process with BatchColumn (R) software. As physicochemical properties of the aroma compounds and the vapour liquid equilibrium (VLE) of all involved mixtures are required for simulation study, group contribution methods are used to predict missing properties such as vapour pressure, vaporisation enthalpy and liquid and vapour heat capacities. For VLE calculation, the modified UNIFAC group contribution method is considered and new binary interaction parameters for the main groups COO and CH2O are regressed from experimental data. Simulation results are in good agreement with experiments carried out in a fully automated batch distillation column at 15 kPa and help optimise the industrial process operation. The proposed methodology can be applied for the design of other solvent recovery process in fragrance industry.

International audience; Local mixing dynamics was recently investigated experimentally in Taylor–Couettesingle-phase flow, thanks to simultaneous Particle Image Velocimetry (PIV) and PlanarLaser-Induced Fluorescence (PLIF) techniques. The results highlighted the influence of thesuccessive flow bifurcations and the role of azimuthal wave states on the dispersion of dyeinjected in Taylor–Couette flows.The present work extends this study to two-phase configurations with spherical solidparticles. The respective effect of particle size and concentration on the vortices size andtransition thresholds between the various flow regimes has been examined thanks to flowvisualizations and PIV measurements. These hydrodynamic features have been comple-mented with PLIF experiments, that revealed a drastic enhancement of mixing due to thepresence of particles regardless of the flow regime, highlighting the existence of significantparticle-induced mixing in Taylor–Couette flows.

International audience; This study focuses on the development of a methodological framework for the design of a five-echelon hydrogen supply chain (HSC) (energy source, production, storage, transportation and fuelling station) considering the geographic level of implementation. The formulation based on mixed integer linear programming involves a multi-criteria approach where three objectives have to be optimised simultaneously, i.e., cost, global warming potential and safety risk. The objective is twofold: first, to test the robustness of the method proposed in De-Leon (2014) from a regional to a national geographic scale and, secondly, to examine the consistency of the results. A new phase of data collection and demand scenarios are performed to be adapted to the French case based on the analysis of roadmaps. In this case study, the ArcGIS® spatial tool is used to locate the supply chain elements before and after optimisation. The multi-objective optimisation approach by the ɛ-constraint method is applied, analysed and discussed. Finally, a comparison between the results of different geographic scale cases is carried out.

Conference: 9th International Symposium on Mixing in Industrial Processes Location: Birmingham, Angleterre Date: 2017/06/25-28; International audience; Cell growth in a chemostat is a well-documented research topic. How cells uptake the avail-able substrate to gain weight and engage cell division is not generally taken into account inthe modelling bioreactors. In fact, the growth rate is related to a population doubling timewhereas the microorganisms’ growth in mass is due to the mass transfer of substrates fromthe liquid phase to the biotic phase. Clearly, growth in mass precedes growth in number.Similarly, the transport of substrates down to the cell scale precedes the mass transfer. Thisarticle’s main feature is a two-dimensional population balance model that allows to uncou-ple growth in mass and growth in number when the equilibrium between a cell populationand its environment is disrupted. The cell length and the rate of anabolism are chosen asinternal variables. It is proved that the hypothesis “growth in number = growth in mass” isvalid at steady-state or in exponential growth only. The glucose uptake is assumed drivenby two transport systems with a different affinity constant for the substrate. This combina-tion of two regulated uptake systems operating in parallel explains a 3-fold increase in theuptake following a glucose pulse, but can also predict substrate uptake rates higher thanthe maximal batch value as observed in some experiments. These features are obtainedby considering carbon fluxes in the formulation of regulation principles for uptake dynam-ics. The population balance’s implementation in a multi-compartment reactor is a naturalprospective work and allows extensions to industrial processes.

International audience; A three-dimensional Computational Fluid Dynamics model is built for a commercial Atomic Layer Deposition (ALD) reactor, designed to treat large area 20 cm substrates. The model aims to investigate the effect of the reactor geometry and process parameters on the gas flow and temperature fields, and on the species distribution on the heated substrate surface, for the deposition of Al2O3 films from trimethyl aluminum and H2O. The investigation is performed in transient conditions, without considering any surface reaction. A second CFD model is developed for the feeding system of the reactor, in order to calculate the unknown reactant inlet flow rates. The two models are coupled via a computational strategy dictated by the available experimental measurements. Results show that a purging flow entering the reactor through its loading door affects the flow field above the substrate surface and causes non-uniformity in the temperature and reactants concentration on the substrate surface. During the TMA pulse, a recirculation sets in above the substrate surface, leading to a non-uniform distribution of species on the surface.

International audience; In this paper, we revise the structure of the residue curve maps (RCM) theory of simple evaporation from the point of view of Differential Geometry. RCM are broadly used for the qualitative analysis of distillation of multicomponent mixtures within the thermodynamic equilibrium model. Nevertheless, some of their basic properties are still a matter of discussion. For instance, this concerns the connection between RCM and the associated boiling temperature surface and the topological characterization of the distillation boundaries. In this paper we put in evidence the Riemannian metric hidden behind the thermodynamic equilibrium condition written in the form of the van der Waals–Storonkin equation, and we show that the differential equations of residue curves have formal gradient structure. We discuss the first non-trivial consequences ofthis factfor the RCM theory ofternary mixtures.

In chemical engineering applications, it is not uncommon to encounter reactors featuring rotating parts. As these rotating parts are present in order to enhance processes such as chemical reactions and/or ensure homogeneity, it is essential to take them into account to perform predictive numerical simulations. This aspect can be particularly challenging, even more so when complex industrial geometries are to be treated. In this paper a numerical methodology for simulating unsteady granular flow in rotating geometries is presented. The method is based on splitting the domain into static and rotating parts. The information between rotating and static parts is passed by a non-conformal mesh matching technique. The presented methodology is validated numerically by comparing its results with other conventional methods. The method is then applied to an industrial scale problem. The applicability of the method and the way it may be used to investigate complex flow is demonstrated. Therefore this approach enables to consider the full geometry of complex reactors. It opens the door to further investigation, optimization and design of industrial scale chemical processes.

International audience; Taylor-Couette flows between two concentric cylinders have great potential applications in chemical engineering. They are particularly convenient for two-phase small scale devices enabling solvent extraction operations. An experimental device was designed with this idea in mind. It consists of two concentric cylinders with the inner one rotating and the outer one fixed. Moreover, a pressure driven axial flow can be superimposed. Taylor-Couette flow is known to evolve towards turbulence through a sequence of successive hydrodynamic instabilities. Mixing characterized by an axial dispersion coefficient is extremely sensitive to these flow bifurcations, which may lead to flawed modelling of the coupling between flow and mass transfer. This particular point has been studied using experimental and numerical approaches. Direct numerical simulations (DNS) of the flow have been carried out. The effective diffusion coefficient was estimated using particles tracking in the different Taylor-Couette regimes. Simulation results have been compared with literature data and also with our own experimental results. The experimental study first consists in visualizing the vortices with a small amount of particles (Kalliroscope) added to the fluid. Tracer residence time distribution (RTD) is used to determine dispersion coefficients. Both numerical and experimental results show a significant effect of the flow structure on the axial dispersion.

International audience; We show how reducing pressure can improve the design of a 1.0-1a mixture homogeneous extractive distillation process and we use extractive efficiency indicators to compare the optimality of different designs. The case study concerns the separation of the diisopropyl ether (DIPE)–isopropyl alcohol (IPA) minimum boiling azeotrope with heavy entrainer 2-methoxyethanol. We first explain that the unexpected energy cost OF decrease following an increase of the distillate outputs is due to the interrelation of the two distillate flow rates and purities and the entrainer recycling through mass balance when considering both the extractive distillation column and the entrainer regeneration column. Then, we find that for the studied case a lower pressure reduces the usage of entrainer and increases the relative volatility of DIPE–IPA for the same entrainer content in the extractive column. A 0.4 atm operating pressure is selected to enable the use of cheap cooling water in the condenser. We run an optimization of the entrainer flow rate, both columns reflux ratios, distillates and feed locations by minimizing the total energy consumption per product unit. Double digit savings in energy consumption are achieved while TAC is reduced significantly. An extractive efficiency indicator that describes the ability of the extractive section to discriminate the desired product between the top and the bottom of the extractive section of the extractive section is calculated for comparing and explaining the benefit of lowering pressure on the basis of thermodynamic insight.

International audience; A study on the pressure drop in pulsed extraction columns with internals of immobile discs and rings, usually called Discs and Doughnuts Columns (DDC) is carried out. The local pressure at a desired level of the column is obtained by resolving of turbulent flow model based on Reynolds equations coupled with k-ε model of turbulence. Consequently, the pressure drop for a column stage or for a unit of column length is determined. The results are used for development of correlations for determination of pressure drop as a function of plate free area, interplate distance and pulsation parameters - amplitude and frequency. Good correspondence to experimental data is observed. The developed quantitative relations are useful for non-experimental numerical optimization of stage geometry in view of lesser energy consumption.

International audience; The quasi-emulsion solvent diffusion method of sphericalcrystallization consists in producing in one step crystallization and agglomeration of small crystals in droplets of an emulsion. Additives are generally used to stabilize the emulsion before crystallization. The aim of this study is to investigate the feasibility of sphericalcrystallization without surfactant. Experiments were performed in an automated batch laboratory scale crystallization process to study the influence of the process operating conditions on the structure of the particles obtained. The results clearly show that, for the experiments performed two types of particles are formed: primary spherical particles and secondary agglomerates. The pattern of the primary particles, observed under scanning electron microscopy, suggests that these particles results from a spherulitic crystal growth mechanism inside the droplet. The secondary agglomerates results of the agglomeration of the spherical particles. In addition, a set of experiments were performed with carefully selected solvents to study the influence of the crystallization solution/water interfacial tension, at constant hydrodynamic conditions and supersaturation level. The results of these experiments demonstrate that the interfacial tension is not a key parameter for designing such a process.

Although a number of definitions of mixing have been proposed in the literature, no single definition accurately and clearly describes the full range of problems in the field of industrial mixing. An alternate approach is proposed which defines segregation as being composed of three separate dimensions. The first dimension is the intensity of segregation quantified by the normalized concentration variance (CoV); the second dimension is the scale of segregation or clustering; and the last dimension is the exposure or the potential to reduce segregation. The first dimension focuses on the instantaneous concentration variance; the second on the instantaneous length scales in the mixing field; and the third on the driving force for change, i.e. the mixing time scale, or the instantaneous rate of reduction in segregation. With these three dimensions in hand, it is possible to speak more clearly about what is meant by the control of segregation in industrial mixing processes. In this paper, the three dimensions of segregation are presented and defined in the context of previous definitions of mixing, and then applied to a range of industrial mixing problems to test their accuracy and robustness.

International audience; This paper is devoted to the analysis of the jet trajectories, obtained using computational fluid dynamics (CFD), at two different scales (laboratory and industrial) with application to quenching of runaway reactions. One of the goals was to describe how the jet penetrates the fluid in the stirred vessel and to build an easy to use correlation for research and industrial purposes. A model of the jet trajectory based on the analogy with a jet in a cross-flow has been used to predict the jet trajectory at the pilot and industrial scales. The correlation, built using a statistical analysis, has shown that the jet in a cross-flow model performs very well to describe the jet trajectories. A very interesting conclusion is that the correlation constants were found to be independent of scale. Finally, the authors proposed a definition of the penetration depth and use it in its dimensionless form to predict how the jet penetrates in the industrial vessel with the current injection conditions.

International audience; Complex industrial processes invest a lot of money in sensors and automation devices to monitor and supervise the process in order to guarantee the production quality and the plant and operators safety. Fault detection is one of the multiple tasks of process monitoring and it critically depends on the sensors that measure the significant process variables. Nevertheless, most of the works on fault detection and diagnosis found in literature emphasis more on developing procedures to perform diagnosis given a set of sensors, and less on determining the actual location of sensors for efficient identification of faults. A methodology based on learning and classification techniques and on the information quantity measured by the Entropy concept, is proposed in order to address the problem of sensor location for fault identification. The proposed methodology has been applied to a continuous intensified reactor, the "Open Plate Reactor (OPR)", developed by Alfa Laval and studied at the Laboratory of Chemical Engineering of Toulouse. The different steps of the methodology are explained through its application to the carrying out of an exothermic reaction.

International audience; The synthesis of catalytic or not composite materials by dry impregnation in fluidized bed is described. This process can be carried out under mild conditions from solutions of organometallic precursors or colloidal solutions of preformed nanoparticles giving rise to reproducible metallic nanoparticles containing composite materials with a high reproducibility. The adequate choice of the reaction conditions makes possible to deposit uniformly the metal precursor within the porous matrix or on the support surface. When the ratio between the drying time and the capillary penetration time (tsec/tcap) is higher than 10, the impregnation under soft drying conditions leads to a homogeneous deposit inside the pores of the particles of support. The efficiency of the metal deposition is close to 100%, and the size of the formed metal nanoparticles is controlled by the pores diameter. Finally, some of the presented composite materials have been tested as catalysts: iron-based materials were used in carbon-nanotubes synthesis, while Pd and Rh composite materials have been investigated in hydrogenation reactions.

The effect of the pumping direction of an axial flow impeller, the feeding rate and the number of feed inlets on the operation of a continuously-fed stirred tank has been studied using computational fluid dynamics (CFD). The flow patterns generated by the up-pumping and down-pumping impeller, under both ‘typical’ and ‘intensified’ operating conditions, are compared. The effect of various tank configurations on the performance of the vessel is assessed by analysing the flow and power numbers, as well as the concentration field of a non-reactive tracer. Furthermore, the inlet feed jets are reduced using traditional jet similarity analysis and are compared with that of a typical round jet. The results show that up-pumping impellers improve circulation in the upper part of the tank and reduce short-circuiting of the feed stream with only a small increase in power consumption. Furthermore, by using multiple feed inlets to increase the total throughput capacity, the amplitude of torque fluctuations is decreased and impeller bypassing is also decreased. The ensemble of conclusions suggest that the throughput capacity and mixing quality of a continuous stirred tank reactor (CSTR) can be improved, without problems of short-circuiting, by employing up-pumping impellers coupled with multiple surface feed points.

Single phase turbulent flow in a tank stirred with two different axial impellers - a pitched blade turbine (PBT) and a Mixel TT (MTT)- has been studied using Laser Doppler Velocimetry. The effect of the agitator configuration, i.e. up-pumping, down-pumping and reverse rotation, on the turbulent flow field, as well as power, circulation and pumping numbers has been investigated. An agitation index for each configuration was also determined. In the down-pumping mode, the impellers induced one circulation loop and the upper part of the tank was poorly mixed. When up-pumping, two circulation loops are formed, the second in the upper vessel. The PBT pumping upwards was observed to have a lower flow number and to consume more power than when down-pumping, however the agitation index and circulation efficiencies were notably higher. The MTT has been shown to circulate liquid more efficiently in the up-pumping configuration than in the other two modes. Only small effects of the MTT configuration on the power number, flow number and pumping effectiveness have been observed.

The hydrodynamic behaviour of a very dense tungsten powder, 75 μm in median diameter and 19,300 kg/m 3 in grain density, has been studied in a fluidized bed at room temperature using nitrogen and argon as carrier gas. Even if fluidization was achieved, the small bed expansion indicated that it was imperfect. Then, the fluidization was studied at 400 °C in order to investigate the feasibility of coating this powder by Fluidized Bed Chemical Vapour Deposition (FBCVD). In particular, the influence of the H 0 / D ratio (initial fixed bed height to reactor diameter) on the bed thermal behaviour was analyzed. It appeared that at least 1.5 kg of powder (corresponding to a H 0 / D ratio of 1.8) was necessary to obtain an isothermal bed at 400 °C. Finally, first results about alumina coatings on the tungsten powder by FBCVD from aluminium acetylacetonate are detailed. They show that for the quite low temperatures tested, the coatings are uniform on all bed particles and are formed of amorphous carbon containing alumina. This study demonstrates the efficiency to combine fluidization (instead of spouted bed) and CVD to coat such very dense powders.

Photosynthetic microorganisms could serve as valuable compounds, but also for environmental applications. Their production under controlled conditions implies to design specific reactors, named photobioreactors, in which light supply is the main constraint. This paper was devoted to an original external-loop airlift photobioreactor (PBR) with annular light chambers in which a swirling motion was induced. The aim was to characterize this novel geometrical configuration in terms of gas–liquid hydrodynamics, and to test its potentiality for algal cultures. This PBR consisted of two identical columns connected by flanges defining tangential inlets, each column being made of two transparent concentric tubes (6 L in liquid volume, 50 m−1 in specific illuminated area). Firstly, the global flow characteristics (circulation and mixing times) were determined by a tracer method and modelled by an axial dispersed plug flow with complete recirculation (Péclet number). By means of a double optical probe, both local and global time-averaged parameters of the gas phase were measured, namely void fraction, bubble velocity, frequency and size. The gas–liquid mass transfer were also characterized, in tap water and in culture medium, by measuring overall volumetric mass transfer coefficients. In a second time, cultures of the microalga Chlamydomonas reinhardtii were run in batch mode. The variations of biomass concentration and pigment content with time from inoculation were successfully obtained. All these findings highlighted: (i) some significant differences in terms of gas–liquid hydrodynamics between the present PBR and the usual airlift systems, (ii) the interest of this configuration for algal cultures, even if complementary studies and technological improvements are still required for definitively validating its scale-up.

Boolean random sets are versatile tools to match morphological and topological properties of real structures of materials and particulate systems. Moreover, they can be combined in any number of ways to produce an even wider range of structures that cover a range of scales of microstructures through intersection and union. Based on well-established theory of Boolean random sets, this work provides scientists and engineers with simple and readily applicable results for matching combinations of Boolean random sets to observed microstructures. Once calibrated, such models yield straightforward three-dimensional simulation of materials, a powerful aid for investigating microstructure property relationships. Application of the proposed results to a real case situation yield convincing realisations of the observed microstructure in two and three dimensions.

Agitation of a thixotropic shear-thinning fluid exhibiting a yield stress is investigated both experimentally and via simulations. Steady-state experiments are conducted at three impeller rotation rates (1, 2 and 8 s−1) for a tank stirred with an axial-impeller and flow-field measurements are made using particle image velocimetry (PIV) measurements. Three-dimensional numerical simulations are also performed using the commercial CFD code ANSYS CFX10.0. The viscosity of the suspension is determined experimentally and is modelled using two shear-dependant laws, one of which takes into account the flow instabilities of such fluids at low shear rates. At the highest impeller speed, the flow exhibits the familiar outward pumping action associated with axial-flow impellers. However, as the impeller speed decreases, a cavern is formed around the impeller, the flow generated in the vicinity of the agitator reorganizes and its pumping capacity vanishes. An unusual flow pattern, where the radial velocity dominates, is observed experimentally at the lowest stirring speed. It is found to result from wall slip effects. Using blades with rough surfaces prevents this peculiar behaviour and mainly resolves the discrepancies between the experimental and computational results.

In electrophoretic membrane contactors, a porous membrane is used to put into contact two flowing liquids, between which a mass transfer takes place. The driving force is an electrical field applied in a direction perpendicular to the fluid flow. Such systems should offer the possibility to scale up electrophoresis technologies, since the separation is achieved with respect to the electrophoretic mass transfer of the solute. On the other hand, since a porous membrane is used instead of an ion exchange one, one can expect to open the field of application of electrodialysis to fractionate polyamino acids, peptides or small proteins for instance. This paper gives a review of the different studies devoted to electrophoretic membrane contactors. This review discuss the main results coming form the two different and complementary approaches used, i.e., that based on application on one hand and that based on the investigation of mass transfer on the other hand.

The performance of a down- and up-pumping pitched blade turbine and A315 for gas-liquid dispersion and mass transfer was evaluated and then compared with that of Rushton and Scaba turbines in a small laboratory scale vessel. The results show that when the axial flow impellers are operated in the up-pumping mode, the overall performance is largely improved compared with the down-pumping configuration. Compared with the radial turbines, the up-pumping A315 has a high gas handling capacity, equivalent to the Scaba turbine and is economically much more efficient in terms of mass transfer than both turbines. On the other hand, the uppumping pitched blade turbine is not as well adapted to such applications. Finally, the axial flow impellers in the down-pumping mode have the lowest performance of all the impellers studied, although the A315 is preferred of the pitched blade turbine.

The performance of three sparger diameters (DS = 0.6D, DS = D, DS = 1.6D) in combination with three positions (below, above or level with the impeller) for gas-liquid dispersion and mass transfer were evaluated in the case of the Rushton turbine and the A315 propeller in up- or down-pumping mode. The results show that the best results in terms of gas handling and mass transfer capacities are obtained for all impellers with the sparger placed below it and with a diameter at least equal to the impeller diameter. For the sparger position below the agitator, the kLa values of the Rushton turbine are greater than those of the A315 propeller, whatever the pumping mode. The A315 propeller in up-pumping mode is, however, more economically efficient in terms of mass transfer. In all cases, the up-pumping mode gives better results than the down-pumping one.

International audience; A logical and progressive method is proposed to modify and increase the performance of chemical reactors where the main inner process is mixing. This method includes global analysis based on mixing visualization, followed by local value measurements. This article reports the case of an axisymmetrical tubular jet mixer: turbulent mixing is speeded up by means of a swirling jet.