Search

Background: The present study aims to explore the feasibility of using flocculation combined with filtration for microalgae harvesting, i.e., Chlorella vulgaris. This is important because microalgae have small sizes and its broth is stable, which makes it difficult to be harvested. The aforementioned facts cause the harvesting cost to be relatively high and become the bottleneck of microalgae processes. Objective: The objective of this research is to find the relation between microalgae concentration, chitosan dosing as a flocculant, and its filterability on membranes. Methods: Research was performed by first cultivating the microalgae in a lab-scale photobioreactor, followed by jar test, flocculation, and filtration experiment. Jar test flocculation was performed using chitosan and microalgae with different concentrations, by simply mixing it in a 100-mL bottle and analyzing the results with UV Vis Spectroscopy. Filtration experiments were performed using lab-made polyvinylidene fluoride membrane, in a 100-mL dead-end filtration cell and in a 5-L tank for submerged filtration. During both filtration tests, filtration flux and fouling were monitored and compared. Results: Results showed that the chitosan concentration needed as a flocculant depends on the microalgae biomass concentration. For the filterability tests, the results proved that flocculation with chitosan enhanced the filterability of the microalgae broth both in dead-end and submerged filtration mode. For the used biomass concentration of around 400 mg/L, the filterability test showed an optimum concentration of chitosan at 7.5 to 10 mg/L, which resulted in a higher filtration flux and lower irreversible fouling in the dead-end filtration and a higher critical flux in the submerged filtration setup. Conclusion: This increased filterability allowed higher fluxes to be operated, thus resulting in a more efficient harvesting process.

Mixed-matrix membranes (MMMs) have been investigated to render energy-intensive separations more efficiently by combining the selectivity and permeability performance, robustness, and nonaging properties of the filler with the easy processing, handling, and scaling up of the polymer. However, truly combining all in one single material has proven very challenging. In this work, we filled a commercial polyimide with ultrahigh loadings of a high–aspect ratio, CO 2 -philic Na-SSZ-39 zeolite with a three-dimensional channel system that precisely separates gas molecules. By carefully designing both zeolite and MMM synthesis, we created a gas-percolation highway across a flexible and aging-resistant (more than 1 year) membrane. The combination of a CO 2 -CH 4 mixed-gas selectivity of ~423 and a CO 2 permeability of ~8300 Barrer outperformed all existing polymer-based membranes and even most zeolite-only membranes.

Gas-separation membranes are a critical industrial component for a low-carbon and energy-efficient future. As a result, many researchers have been testing membrane materials over the past several decades. Unfortunately, almost all membrane-based testing systems are home-built, and there are no widely accepted material standards or testing protocols in the literature, making it challenging to accurately compare experimental results. In this multi-lab study, ten independent laboratories collected high-pressure pure-gas permeation data for H2, O2, CH4, and N2 in commercial polysulfone (PSf) films. Equipment information, testing procedures, and permeation data from all labs were collected to provide (1) accepted H2, O2, CH4, and N2 permeability values at 35 °C in PSf as a reference standard, (2) statistical analysis of lab-to-lab uncertainties in evaluating permeability, and (3) a list of best practices for sample preparation, equipment set-up, and permeation testing using constant-volume variable-pressure apparatuses. Results summarized in this work provide a reference standard and recommended testing protocols for pure-gas testing of membrane materials.

This work concerns the intensification of membrane processes by coupling the Joule effect with two membrane processes: pervaporation and sweeping gas membrane distillation. For this purpose, conducting metallic hollow fibers impregnated or coated with polydimethyl siloxane were simultaneously used as membrane and heating electric resistance. The application of an electrical potential resulted in an enhancement of 40% of the water vapor permeate flux in sweep gas membrane distillation. However, the flux enhancement is the result not only of the heating on the membrane vicinity but also on the enhancement of the feed temperature. In the case of pervaporation of aqueous ethanol solutions (20%), the direct heating of fibers allowed improving by 100% the ethanol permeate flux while increasing the process selectivity.

The occurrence of metal ion exchange in Zn3[Co(CN)6]2 and Cu3[Co(CN)6]2 Prussian blue analogues (Zn-Co and Cu-Co PBAs) was demonstrated for the first time. While Cu(ii) ion exchange easily occurs in Zn-Co PBA, the exchange of Cu(ii) atoms in Cu-Co PBA by Zn(ii) proved to be more difficult. At low to medium Cu(ii) loadings, the catalytic activity of the exchanged PBAs for the A3 coupling reaction of benzaldehyde, piperidine and phenylacetylene was higher than that of the bimetallic PBAs and that of multi metal PBAs of similar composition prepared by co-precipitation. This result showcases the benefits of the ion exchange process as a preparation method of PBA catalysts, since it is believed to lead to the incorporation of the desired metal in a more accessible position for reactant molecules. At higher Cu(ii) loadings, ion exchange with Cu(CH3COO)2·H2O also resulted in co-incorporation of CH3COO-. This incorporation considerably boosted the catalytic activity of the PBAs by providing a basic function that facilitates the C-H activation of phenylacetylene. The most active of the studied PBAs, catallytically outperforms other Cu(ii) based A3 coupling catalysts and completely suppresses the activity for the homocoupling of phenylacetylene, even under oxidative conditions. Furthermore, the basicity of the PBAs was investigated in the nitroaldol (Henry) reaction, where a clear effect of the presence of CH3COO- was observed. The CH3COO- containing PBAs exhibited an activity three times higher than the rest of the PBAs. The presence of the basic CH3COO- groups represents the first case of basic functionalization of PBAs.

Clean water shortage is one of the global major concerns. To overcome this problem, membrane-based desalination process, such as reverse osmosis (RO) have been developed in the recent years. However, this technology operates at a high hydraulic pressure to overcome the seawater osmotic pressure, making this desalination process still relatively expensive. Recently, forward osmosis (FO) process has become increasingly attractive over RO since it is capable to produce clean water by using natural osmotic pressure as the separation driving force provided by draw solution (DS), which eliminates the usage of high hydraulic pressure. DS is a critical material in the FO process, playing a key role in determining the separation performance and energy cost. After the FO process, DS requires a regeneration step to reconcentrate it to its initial state. As a result, the development of effective draw solution is essential for the advancement of FO technology.In this master thesis work, the feasibility of light-responsive TiO2 nanoparticles was investigated as a draw solute in the FO process. Several dispersions of TiO2 nanoparticles in demineralized water (DI) with different concentration were prepared. These dispersions were tested for their FO performance with DI and saline water as feed solutions, respectively. The colloidal stability of TiO2 nanoparticles was achieved by deviation from the isoelectric point of the nanoparticles with a pH adjustment from 6 to 4.5.The effects of TiO2 nanoparticle concentration, ultrasonication time, UV irradiation, primary particle size, and FO membrane orientation were tested using DI water as the feed solution. A concentration of 20 g / L of TiO2 was found as the optimum concentration for FO performance. The generated water flux was not linearly dependent on the nanoparticles concentrations but appeared to be dependent on how well the nanoparticles were dispersed. The effect of UV was investigated by testing the same amount of TiO2 with and without a UV treatment. The UV treated and sonicated before and after UV treatment generated a significantly higher water flux than the pristine one. The effect of the primary size of the nanoparticles was expressed in their response to UV radiation. In comparison to P5 (=5 nm), P15 (=15 nm), and P25 (= 25 nm), P5 showed the highest UV response. The effect of membrane orientation, namely pressure retarded osmosis (PRO) and forward osmosis (FO) mode were tested and the PRO mode resulted in a significantly higher water flux than FO mode due to its low concentration polarization effect, which consequently led to higher driving force than FO mode. The TiO2 nanoparticles draw solutions were also characterized using DLS and FTIR to investigate the effect of UV on the particle size and hydrophilicity respectively. The DLS result showed a larger hydrodynamic diameter after UV treatment comparing to the pristine one but after ultrasonication, the hydrodynamic diameter was significantly decreased. This result confirmed that UV treatment promoted the agglomeration of nanoparticles. Longer UV irradiation time led to a moderately more hydrophilic surface than the shorter period as confirmed from the FTIR result. The highest water flux generated with DI water feed solution by using the optimum concentration and parameters without any functionalization or capping agent was 15.8 LMH and this is quite promising.

Salinity Gradient Power - Reverse Electrodialysis (SGP-RED) is a promising membrane-based technology to harvest the energy of mixing from solutions of different ionic concentration. Unfortunately, currently available commercial ion exchange membranes – being not specifically designed for RED – are far from satisfying the requirements of this operation, especially when operated with hyper-concentrated brines. In this work, novel sulfonated polyethersulfone (sPES) cation exchange membranes (CEM) were prepared by phase inversion method and tested under high salinity gradients. Use of 5 M NaCl electrolyte for immersion precipitation coagulation bath facilitated the self-standing membrane formation as a result of the electrostatic interaction between the fixed charged groups and electrolyte solution. Microscopy results revealed that dense or asymmetric membranes with non-connected pores were formed by solvent evaporation or immersion precipitation, respectively. The membranes were characterized for ion exchange capacity, water uptake, charge density and thickness, and further studied by electrochemical impedance spectroscopy. The obtained properties of the newly developed membranes were subsequently compared to those of commercial CMX (Neosepta, Japan) and Fuji-CEM-Type 1 (Fujifilm, The Netherlands) membranes. The asymmetric membranes resulted in a very low resistance especially for high ionic gradients but relatively low permselectivity, while dense membranes still had a low resistance compared to commercial membranes and exhibited high permselectivity. Interestingly, in terms of power density, lab-made membranes outperformed the commercial benchmarks when tested for RED applications with brackish water (0.1 M NaCl)/hypersaline brine (4 M NaCl) feeds; power density of CMX and Fuji-CEM-Type 1 were 3.23 and 3.77 W/m2, respectively, while power density of asymmetric and dense sPES membranes were 3.64 and 3.92 W/m2, respectively.

In this work have been developed general predictive models for direct contact membrane distillation (DCMD) and sweeping gas membrane distillation (SGMD) as well as a hydrodynamic model for anaerobic membrane bioreactors (AnMBRs) equipped with the induced membrane vibration (MMV) system. The DCMD and SGMD models allow simulating hollow fibre and flat sheet configurations under wide range of process conditions without empirical mass and heat transfer coefficients or laboratory experiments. The models have been validated with experimental and literature data. Indeed, the influence of operating conditions and membrane geometric characteristics on the process performance has been investigated. The model for AnMBRs with MMV studies the effect of the membrane vibration on the hydrodynamics of the AnMBR tank. The parametric study allows knowing, the effects of the vibration frequency and amplitude on the fluid velocity and volume fraction of solids. The conducted studies prove that all the proposed models would be potentially applied for the pre-experimental study, optimization of process conditions, design of membrane modules as well as for the further cost estimation of the processes.; Dans cette thèse ont été développés des modèles mathématiques pour les procédés de distillation membranaire à contact direct (DCMD) et avec balayage gazeux (SGMD) ainsi qu’un modèle sur l’hydrodynamique des bioréacteurs membranaires anaérobiques (AnMBRs) équipés d’un système de vibration membranaire induite (MMV). Les modèles pour la DCMD et la SGMD permettent de simuler le comportement des modules plats ou à fibres creuses sous différentes conditions opératoires, sans avoir recours aux données expérimentales ou à des équations empiriques pour les transferts de masse et de chaleur. Les modèles ont été validés avec des résultats expérimentaux et de la littérature et ont permis de déterminer l'influence de différents paramètres opérationnels et de la géométrie des modules sur les performances des procédés. Le modèle développé pour les AnMBRs équipés du système MMV permet d’étudier l’effet de la vibration membranaire sur l’hydrodynamique du réservoir. L’analyse paramétrique a permis d’étudier l’effet de la fréquence et de l’amplitude des vibrations sur la vitesse du fluide et la fraction volumique des solides dans le réservoir. Dans ce travail il a été démontré que les modèles proposés pourront être potentiellement appliqués à des études expérimentales préliminaires, l’optimisation des conditions opératoires, la conception des modules membranaires ainsi que pour l’estimation des coûts des procédés.

Solvent nanofiltration is a potential technology to recover solvents. For this application, a chemically stable membrane that can endure continuous exposure towards organic solvents is required. This thesis deals with the preparation of chemically stable NF membranes through modification of mesoporous ceramic substrate by means of grafting and studying of their solvent and solute transport properties. In Chapter 1, the background of the grafting technique as well as studies on the SRNF transport behavior found in the literature was presented.In Chapter 2 and 6 of this thesis, mesoporous y-alumina UF membranes were grafted by hydrophobic and hydrophilic organic moieties to decrease the membrane pore diameter of the existing y-alumina UF membrane down to the nanofiltration range. In Chapter 5, the use of coupling agent to couple the grafted moiety forming a polymer network inside the ceramic pores during grafting results in a smaller membrane pore, but at the cost of a lower solvent permeability, when compared with PDMS-grafted alumina membranes where no coupling was applied. In Chapter 6, the grafting performance of γ-Al2O3 powder with various PEG grafting agents having different molecular weights, alkoxy groups, and ureido functionalities were analysed by TGA, 29Si-NMR, FTIR, and BET. The grafting densities are influenced by the molecular weights, the presence of the ureido functionality, and the number of hydrolyzable groups of the grafting agents. The transport behavior of PDMS grafted ceramic membranes and PEG grafted ceramic membranes were studied in Chapter 3, 4, and 6. In Chapter 3, the solvent transport behavior of PDMS grafted ceramic membranes was described by incorporating solvent sorption terms in the Hagen-Pouiseuille equation. A more closed membrane structure is realized when the solvent is strongly sorbed in the grafted moiety. In Chapter 4, the applicability of the existing solute rejection models based on size-exclusion mechanism to describe the solute rejection of membranes towards different types of solvent and solute were assessed. A strong function of rejection behavior with the ratio of the solute diameter versus the membrane pore diameter was observed, indicating that the size-exclusion mechanism may be applicable. Three rejection models based on size-exclusion, namely the Ferry, Verniory, and SHP models were used to predict the rejection of several solutes using pore diameter information from the N2 physisorption measurement when no solvent is present. For dye, PS, and PEG solutes in toluene, the experimental data fall well above the predicted σ for Ferry, Verniory, and SHP model suggesting that the membrane actual pore diameter in the presence of strongly sorbed solvent like toluene is smaller than that when no solvent is present, assuming that there is no important solvent-solute or solute-membrane interaction present in the observed rejection behavior. This may explain the higher rejection of solutes in nonpolar solvents like toluene than that in polar solvents such as isopropanol for PDMS grafted ceramic membranes. In Chapter 6, the permeability behavior of PEG grafted y-alumina membranes with respect to different types of permeating solvent (polar and nonpolar) was studied. A linear relationship between flux and TMP was observed, as was also found for PDMS grafted y-Al2O3 membranes. This indicates the absence of shear-flow induced behaviour in the applied TMP. A higher selectivity of Sudan Black in ethanol than in hexane accompanied by a lower permeability of ethanol than hexane were observed. Here also this phenomenon is explained by the difference in solvent sorption of the grafted moiety for different types of permeating solvents. Finally, the general conclusions and future work are presented in Chapter 7.; La nanofiltration (NF) est un procédé applicable à la récupération des solvants organiques. Une membrane chimiquement stable est alors requise pour résister aux solvants organiques. Cette thèse traite de la préparation de membranes NF chimiquement stables par greffage de substrats céramiques mésoporeux et de l'étude de leurs propriétés de transport des solvants et des solutés. Dans le chapitre 1, l'état de l'art sur les techniques de greffage est présenté ainsi que celui sur le comportement au transport des membranes NF résistantes aux solvants.Dans les chapitres 2 et 6, des membranes d'ultrafiltration en alumine mésoporeuse sont greffées avec des groupements organiques hydrophobes ou hydrophiles. La diminution du diamètre des pores permet ainsi d'accéder à la nanofiltration. Au chapitre 5, un agent couplant est utilisé pour améliorer l'ancrage de ces groupements dans les pores. Ceci réduit cependant la perméabilité aux solvants, en comparaison aux mêmes membranes modifiées avec du polydiméthylsilane (PDMS) mais sans agent couplant. Dans le chapitre 6, la capacité de greffage de poudres d'alumine est mesurée pour des agents de greffage différant par : la masse moléculaire des chaines polyéthylènes glycol (PEG), la nature et le nombre de groupements alcoxy terminaux et la présence ou non de fonctions urée. Ces poudres sont analysés par thermogravimétrie, spectrométrie RMN du 29Si, spectroscopie FTIR, et mesures de surface spécifique. Les densités de greffage estimées varient avec la masse des greffons, la présence de fonctions urée, et le nombre de groupements alcoxy hydrolysables.Le comportement au transport de membranes greffées est étudié dans les chapitres 3, 4 et 6. Dans le chapitre 3, pour des membranes greffées avec du PDMS, ce comportement est décrit en incorporant des termes relatifs à la sorption des solvants dans l'équation Hagen-Poiseuille. Une membrane plus fermée est obtenue lorsque le solvant est fortement adsorbé dans la couche greffée. Dans le chapitre 4, la validité des modèles de rejet de soluté basés sur l'exclusion par la taille est discutée. Une forte influence du diamètre moléculaire du soluté et du rapport de ce diamètre avec celui des pores est observée, indiquant que le mécanisme d'exclusion par la taille est ici vérifié. Trois modèles de rejet sur la base d'exclusion par la taille, à savoir Ferry, Verniory et SHP, sont testés pour prédire, en l'absence de solvant, le rejet des solutés à partir des diamètres de pore mesurés par physisorption de diazote. Pour des colorants et des solutés de type PS ou PEG dans du toluène, les données expérimentales sont bien au-dessus des valeurs prédites par ces modèles. Les résultats suggèrent que le diamètre de pore effectif en présence de solvant fortement adsorbé tel que le toluène est inférieur à celui en l'absence de solvant, une hypothèse étant qu'il n'y a pas d'interactions importantes entre solvant et soluté ou entre le soluté et la surface des pores. Cela peut expliquer un rejet plus élevé des solutés dans des solvants non polaires comme le toluène que dans des solvants polaires tels que l'isopropanol pour les membranes greffées avec du PDMS. Dans le chapitre 6, la perméabilité de membranes greffées avec des PEG est étudiée pour différents solvants (polaires ou non polaires). Une relation linéaire entre le flux et la pression transmembranaire est observée, comme pour les membranes greffées avec du PDMS. Cela indique l'absence de processus induit par des effets de cisaillement dans le fluide en écoulement et variant avec la pression transmembranaire appliquée. Pour le colorant Noir Soudan, une sélectivité supérieure est observée dans l'éthanol que dans l'hexane alors que pour la perméabilité inférieure de l'éthanol est inférieure à celle de l'hexane. Ici aussi, ces phénomènes sont expliqués par la différence de sorption des solvants dans la couche greffée. Les conclusions générales et perspectives de cette étude sont présentées dans le chapitre 7.

The plasticization behavior of pure polymers is well studied in literature. However, there are only few studies on the plasticization behavior of mixed matrix membranes. In Chapter 2 of this thesis, pure and mixed gas plasticization behavior of MMMs prepared from mesoporous Fe(BTC) nanoparticles and the polymer Matrimid® is investigated. All experiments were carried with solution casted dense membranes. Mesoporous Fe(BTC) MOF particles showed reasonably good compatibility with the polymer. Incorporation of Fe(BTC) in Matrimid® resulted in membranes with increased permeability and selectivity. At low pressures of 5 bar the MMMs showed an increase of 60 % in CO2 permeability and a corresponding increase of 29 % in ideal selectivity over pure Matrimid® membranes. It was observed that the presence of Fe(BTC) particles increases the plasticization pressure of Matrimid® based MMMs. Furthermore, this pressure increases more with increasing MOF loading. This delay in plasticization is attributed to the reduced mobility of the polymer chains in the vicinity of the Fe(BTC) particles. Also, at higher Fe(BTC) loadings, the membranes showed more or less constant selectivity over the whole pressure range investigated. Chapter 3 subsequently presented the preparation and plasticization behavior of MMMs based on three distinctively different MOFs (MIL-53(Al) (breathing MOF), ZIF-8 (flexible MOF) and Cu3(BTC)2 (rigid MOF)) dispersed in Matrimid®. The ideal and mixed gas performance of the prepared MMMs was determined and the effect of MOF structure on the plasticization behavior of MMMs was investigated. Among the three MOF-MMMs, membranes based on Cu3(BTC)2 showed highest selectivity while ZIF-8 based membranes showed highest permeability. The respective increase in performance of the MMMs is very much dependent on the MOF crystal structure and its interactions with CO2 molecules. Chapter 4 described the preparation of Matrimid® polyimide (PI)/polysulfone (PSF)-blend membranes containing ZIF-8 particles for high pressure gas separation. An optimized PI/PSF blend ratio (3:1) was used and performance and stability of PI/PSF mixed matrix membranes filled with different concentrations of ZIF-8 were investigated. PI and PSF were miscible and provided good compatibility with the ZIF-8 particles, even at high loadings. The PI/PSF-ZIF-8 MMMs showed significant enhancement in CO2 permeability with increased ZIF-8 loading, which was attributed to a moderate increase in sorption capacity and faster diffusion through the ZIF-8 particles. In pure gas measurements, pure PI/PSF blend (3:1) membranes showed a plasticization pressure of ~18 bar while the ZIF-8 MMMs showed a higher plasticization pressures of ~25 bar. Mixed gas measurements of PI/PSF-ZIF-8 MMMs showed suppression of plasticization as confirmed by a constant mixed gas CH4 permeability and a nearly constant selectivity with pressure but the effect was stronger at high ZIF-8 loadings. Gas separation results of the prepared PI/PSF-ZIF-8 MMMs show an increased commercial viability of Matrimid® based membranes and broadened their applicability, especially for high-pressure CO2 gas separations. In Chapter 5, a novel route for the preparation of mixed matrix membranes via a particle fusion approach was introduced. Surface modification of the polymer with 1-(3-aminopropyl)-imidazole led to an excellent ZIF-8-Matrimid® interfacial compatibility. It was possible to successfully prepare MMMs with MOF loadings as high as 30 wt.% without any non-selective defects. Upon increasing the ZIF-8 loading, MMMs showed significantly better performance in the separation of CO2/CH4 mixtures as compared to the native polymer. The CO2 permeability increased up to 200 % combined with a 65 % increase in CO2/CH4 selectivity, compared to the native Matrimid®. Chapter 6 finally discussed the conclusions and directions for future research based on the findings presented in this thesis.; Le comportement plastifiant de polymères purs a été bien étudié dans la littérature. Toutefois, il n'y a eu que peu d'études concernant les membranes à matrices mixtes (MMM). Dans le chapitre 2 de cette thèse, le comportement plastifiant de MMM préparés à partir de nanoparticules mésoporeuses Fe(BTC) et du polymère Matrimid® est étudié avec un gaz pur ou en mélange. Les réseaux métaux-organiques (MOF) sous forme particulaires ont présenté une relativement bonne compatibilité avec le polymère. L'incorporation de Fe(BTC) dans du Matrimid® a permis d'augmenter la perméabilité et la sélectivité des membranes. Pour de faibles pressions de 5 bars, les MMM ont une perméabilité au CO2 de 60% plus grande ainsi qu'une sélectivité de 29% plus grande à comparer à la sélectivité idéale de membranes Matrimid®. Il a été observé que la présence de particules Fe(BTC) retardait l'effet plastifiant vers de plus grandes pressions. De plus, cette pression augmente avec le taux de MOF au sein du matériau. Ce retard est attribué à la mobilité réduite des chaînes polymères dans l'entourage des particules Fe(BTC). Egalement, pour des concentrations en MOF plus élevées, les membranes présentent une sélectivité plus ou moins constante sur toute la gamme de pression étudiée. Le chapitre 3 présente ensuite la préparation et le caractère plastifiant des MMMs basées sur trois types de MOFs (MIL-53(Al) (MOF « repirant »), ZIF-8 (MOF « flexible ») and Cu3(BTC)2 (MOF « rigide »)) dispersés dans le Matrimid®. Les performances en gaz pur ou en mélange ont été étudiées en fonction de la quantité de MOF introduite. Parmi les trois systèmes MOF-MMM, les membranes avec le Cu3(BTC)2 ont présenté la plus haute sélectivité alors que les membranes avec du ZIF-8 ont montré une plus grande perméabilité. Ces améliorations sont essentiellement le fait de la structure cristalline du MOF et de son interaction avec les molécules de CO2. Le chapitre 4 décrit la préparation de membranes à base de mélange Matrimid® polyimide (PI)/polysulfone (PSF) contenant des particules de ZIF-8 pour la séparation gazeuse à haute pression. Un mélange optimisé avec un rapport PI/PSF de 3:1 a été utilisé pour une étude sur la stabilité et la performance de ces MMMs incorporant différentes concentration de ZIF-8. PI et PSF étant miscibles, une bonne compatibilité avec les particules de ZIF-8 est observée. Les MMMs PI/PSF-ZIF-8 ont démontré une amélioration significative de la perméabilité en CO2 lors de l'augmentation de la concentration en ZIF-8, ce qui a été attribué à une augmentation modérée de la capacité de sorption et à une diffusion plus rapide au travers des particules de ZIF-8. Lors des mesures en gaz purs, les membranes PI/PSF (3:1) ont présenté une plastification vers 18 bars alors que l'introduction de ZIF-8 repousse cette valeur à 25 bars. En mélange de gaz, les MMMs PI/PSF-ZIF-8 ont abouti à une suppression de la plastification comme l'a confirmé une mesure constante de la perméabilité et de la sélectivité du CH4, et cet effet est plus accentué avec l'augmentation de la concentration en ZIF-8. Les résultats en séparation des gaz avec les MMMs PI/PSF-ZIF-8 montrent une performance supérieure à celle du Matrimid® ce qui laisse augurer un élargissement du spectre d'application de ces membranes, particulièrement pour la séparation du CO2 à haute pression. Dans le chapitre 5, une nouvelle voie de préparation des MMMs via la fusion contrôlée de particules a été introduite. La modification du Matrimid® par du 1-(3-aminopropyl)-imidazole a permis d'améliorer considérablement la compatibilité avec les particules de ZIF-8. Il a ainsi été possible de préparer des MMMs contenant 30% de MOF sans perte de sélectivité. En augmentant la concentration en ZIF-8, les MMMs ont de meilleures performances dans la séparation de mélange CO2/CH4 à comparer au polymère initial. La perméabilité a augmenté de plus de 200% avec une augmentation de 65% de sélectivité pour le mélange CO2/CH4.

Solvent nanofiltration is a potential technology to recover solvents. For this application, a chemically stable membrane that can endure continuous exposure towards organic solvents is required. This thesis deals with the preparation of chemically stable NF membranes through modification of mesoporous ceramic substrate by means of grafting and studying of their solvent and solute transport properties. In Chapter 1, the background of the grafting technique as well as studies on the SRNF transport behavior found in the literature was presented.In Chapter 2 and 6 of this thesis, mesoporous y-alumina UF membranes were grafted by hydrophobic and hydrophilic organic moieties to decrease the membrane pore diameter of the existing y-alumina UF membrane down to the nanofiltration range. In Chapter 5, the use of coupling agent to couple the grafted moiety forming a polymer network inside the ceramic pores during grafting results in a smaller membrane pore, but at the cost of a lower solvent permeability, when compared with PDMS-grafted alumina membranes where no coupling was applied. In Chapter 6, the grafting performance of γ-Al2O3 powder with various PEG grafting agents having different molecular weights, alkoxy groups, and ureido functionalities were analysed by TGA, 29Si-NMR, FTIR, and BET. The grafting densities are influenced by the molecular weights, the presence of the ureido functionality, and the number of hydrolyzable groups of the grafting agents. The transport behavior of PDMS grafted ceramic membranes and PEG grafted ceramic membranes were studied in Chapter 3, 4, and 6. In Chapter 3, the solvent transport behavior of PDMS grafted ceramic membranes was described by incorporating solvent sorption terms in the Hagen-Pouiseuille equation. A more closed membrane structure is realized when the solvent is strongly sorbed in the grafted moiety. In Chapter 4, the applicability of the existing solute rejection models based on size-exclusion mechanism to describe the solute rejection of membranes towards different types of solvent and solute were assessed. A strong function of rejection behavior with the ratio of the solute diameter versus the membrane pore diameter was observed, indicating that the size-exclusion mechanism may be applicable. Three rejection models based on size-exclusion, namely the Ferry, Verniory, and SHP models were used to predict the rejection of several solutes using pore diameter information from the N2 physisorption measurement when no solvent is present. For dye, PS, and PEG solutes in toluene, the experimental data fall well above the predicted σ for Ferry, Verniory, and SHP model suggesting that the membrane actual pore diameter in the presence of strongly sorbed solvent like toluene is smaller than that when no solvent is present, assuming that there is no important solvent-solute or solute-membrane interaction present in the observed rejection behavior. This may explain the higher rejection of solutes in nonpolar solvents like toluene than that in polar solvents such as isopropanol for PDMS grafted ceramic membranes. In Chapter 6, the permeability behavior of PEG grafted y-alumina membranes with respect to different types of permeating solvent (polar and nonpolar) was studied. A linear relationship between flux and TMP was observed, as was also found for PDMS grafted y-Al2O3 membranes. This indicates the absence of shear-flow induced behaviour in the applied TMP. A higher selectivity of Sudan Black in ethanol than in hexane accompanied by a lower permeability of ethanol than hexane were observed. Here also this phenomenon is explained by the difference in solvent sorption of the grafted moiety for different types of permeating solvents. Finally, the general conclusions and future work are presented in Chapter 7., La nanofiltration (NF) est un procédé applicable à la récupération des solvants organiques. Une membrane chimiquement stable est alors requise pour résister aux solvants organiques. Cette thèse traite de la préparation de membranes NF chimiquement stables par greffage de substrats céramiques mésoporeux et de l'étude de leurs propriétés de transport des solvants et des solutés. Dans le chapitre 1, l'état de l'art sur les techniques de greffage est présenté ainsi que celui sur le comportement au transport des membranes NF résistantes aux solvants.Dans les chapitres 2 et 6, des membranes d'ultrafiltration en alumine mésoporeuse sont greffées avec des groupements organiques hydrophobes ou hydrophiles. La diminution du diamètre des pores permet ainsi d'accéder à la nanofiltration. Au chapitre 5, un agent couplant est utilisé pour améliorer l'ancrage de ces groupements dans les pores. Ceci réduit cependant la perméabilité aux solvants, en comparaison aux mêmes membranes modifiées avec du polydiméthylsilane (PDMS) mais sans agent couplant. Dans le chapitre 6, la capacité de greffage de poudres d'alumine est mesurée pour des agents de greffage différant par : la masse moléculaire des chaines polyéthylènes glycol (PEG), la nature et le nombre de groupements alcoxy terminaux et la présence ou non de fonctions urée. Ces poudres sont analysés par thermogravimétrie, spectrométrie RMN du 29Si, spectroscopie FTIR, et mesures de surface spécifique. Les densités de greffage estimées varient avec la masse des greffons, la présence de fonctions urée, et le nombre de groupements alcoxy hydrolysables.Le comportement au transport de membranes greffées est étudié dans les chapitres 3, 4 et 6. Dans le chapitre 3, pour des membranes greffées avec du PDMS, ce comportement est décrit en incorporant des termes relatifs à la sorption des solvants dans l'équation Hagen-Poiseuille. Une membrane plus fermée est obtenue lorsque le solvant est fortement adsorbé dans la couche greffée. Dans le chapitre 4, la validité des modèles de rejet de soluté basés sur l'exclusion par la taille est discutée. Une forte influence du diamètre moléculaire du soluté et du rapport de ce diamètre avec celui des pores est observée, indiquant que le mécanisme d'exclusion par la taille est ici vérifié. Trois modèles de rejet sur la base d'exclusion par la taille, à savoir Ferry, Verniory et SHP, sont testés pour prédire, en l'absence de solvant, le rejet des solutés à partir des diamètres de pore mesurés par physisorption de diazote. Pour des colorants et des solutés de type PS ou PEG dans du toluène, les données expérimentales sont bien au-dessus des valeurs prédites par ces modèles. Les résultats suggèrent que le diamètre de pore effectif en présence de solvant fortement adsorbé tel que le toluène est inférieur à celui en l'absence de solvant, une hypothèse étant qu'il n'y a pas d'interactions importantes entre solvant et soluté ou entre le soluté et la surface des pores. Cela peut expliquer un rejet plus élevé des solutés dans des solvants non polaires comme le toluène que dans des solvants polaires tels que l'isopropanol pour les membranes greffées avec du PDMS. Dans le chapitre 6, la perméabilité de membranes greffées avec des PEG est étudiée pour différents solvants (polaires ou non polaires). Une relation linéaire entre le flux et la pression transmembranaire est observée, comme pour les membranes greffées avec du PDMS. Cela indique l'absence de processus induit par des effets de cisaillement dans le fluide en écoulement et variant avec la pression transmembranaire appliquée. Pour le colorant Noir Soudan, une sélectivité supérieure est observée dans l'éthanol que dans l'hexane alors que pour la perméabilité inférieure de l'éthanol est inférieure à celle de l'hexane. Ici aussi, ces phénomènes sont expliqués par la différence de sorption des solvants dans la couche greffée. Les conclusions générales et perspectives de cette étude sont présentées dans le chapitre 7.