Search

This study aims to prepare the sodium bentonite (Bent-Na) and the iron-pillared bentonite (PILC-Fe) and to evaluate its adsorption efficacy against Coomassie blue from aqueous solution. Results demonstrated that the PILC-Fe sample presents an increase in the basal distance d001 from 13.66 to 16.93 A. The presence of the Fe–O and FeOOH functional group band on PILC-Fe was confirmed by FTIR analysis. A clear increase has been observed of the specific surface area from 45.86 to 125.30 m2/g for Bent-Raw and PILC-Fe, respectively, and this increase is greater than 63%. Adsorption of CBB dye into Bent-Na and PILC-Fe was fast, and their adsorption equilibriums were reached within 15 and 5 min, respectively. The maximum adsorption capacity for CBB dye is 9.125 mg/g for PILC-Fe, whereas Bent-Na presents a lower adsorption capacity of around 6.848 mg/g. The maximum CBB dye adsorption rate was attained at pH 4.68 for both bentonites. It is observed a difference in the removal efficiency implying that the adsorption of CBB molecule dye was affected by the structure and functional groups of dyes. It is noticed that the adsorption process for CBB dye into Bent-Na and PILC-Fe was better fitted with the linear and nonlinear Langmuir model. The CBB dye kinetic adsorption is better fitted by the pseudo-second-order model. Thermodynamic study indicated the feasibility and spontaneity of the CCB dye adsorption process that was exothermic and physical in nature. A comprehensive adsorption and mechanism interaction study of CBB molecular structure onto PILC-Fe was appropriately discussed.

This contribution is a comparison study between synthetic hydroxyapatite (Sy-HAP) and commercial hydroxyapatite (C-HAP) for the removal of Pb2+ and Cd2+ ions present in wastewater from industrial effluents. The obtained results show that the equilibrium time required for complete adsorption of Pb2+ and Cd2+ ions on C-HAP and Sy-HAP is 15 min for both. The obtained removal efficiencies for Sy-HAP are 95.52% and 90.91% for Pb2+ and Cd2+ ions, respectively. Whereas, C-Hap presents lower removal efficiencies of 86.53% and 81.43% for Pb2+ and Cd2+ ions, respectively. Maximum adsorption was observed at pH 5; at lower pH levels adsorption was less. The experimental kinetic data fitted with the second order kinetic model. Thermodynamically, the adsorption process was endothermic and spontaneous in nature. Isotherm adsorption studies indicated that Langmuir, Freundlich and Temkin are the most valid models to describe and evaluate the adsorption process. The EDX results also confirmed the presence of lead and cadmium in adsorbents after adsorption. Finally, the HAP porous materials possess great potential for the removal of Pb2+ and Cd2+ ions from aqueous solutions and wastewater from industrial effluents.

Recent advances in applying deep learning to planning have shown that Deep Reactive Policies (DRPs) can be powerful for fast decision-making in complex environments. However, an important limitation of current DRP-based approaches is either the need of optimal planners to be used as ground truth in a supervised learning setting or the sample complexity of high-variance policy gradient estimators, which are particularly troublesome in continuous state-action domains. In order to overcome those limitations, we introduce a framework for training DRPs in continuous stochastic spaces via gradient-based policy search. The general approach is to explicitly encode a parametric policy as a deep neural network, and to formulate the probabilistic planning problem as an optimization task in a stochastic computation graph by exploiting the re-parameterization of the transition probability densities; the optimization is then solved by leveraging gradient descent algorithms that are able to handle non-convex objective functions. We benchmark our approach against stochastic planning domains exhibiting arbitrary differentiable nonlinear transition and cost functions (e.g., Reservoir Control, HVAC and Navigation). Results show that DRPs with more than 125,000 continuous action parameters can be optimized by our approach for problems with 30 state fluents and 30 action fluents on inexpensive hardware under 6 minutes. Also, we observed a speedup of 5 orders of magnitude in the average inference time per decision step of DRPs when compared to other state-of-the-art online gradient-based planners when the same level of solution quality is required.

International audience; In response to the fresh water scarcity, Tunisia is utilizing more and more membrane desalination of unconventional resources, including brackish waters and seawater. The widespread reserves of groundwaters and their low salinity make this resource of special interest. Two predominant ionic compositions have been identified depending on their relative proportion of sulfate to chloride ions. The question arising for the decision-makers concerns the choice of membrane technology and, therefore, of membrane. Two nanofiltration (NF) membranes (NF270 and NF90) and a reverse osmosis (RO) one (BW30) were tested in a desalination study of synthetic feeds reproducing the ionic composition of three representative groundwaters. Sulfate/chloride ratio appears to be the key factor for the membrane choice to obtain good quality drinking water meeting the Tunisian standards. Moreover, validation of two prediction tools was investigated: ROSA, software provided by the membrane manufacturer and Nanoflux ® , software specifically designed for NF. The experimental NF results are well fitted by the Nanoflux ® simulations. We concluded that ROSA cannot generally provide good NF predictions because it does not take into account the electric interactions between membrane and feed.

International audience; This work is aimed at developing a new process to separate alpha-cyclodextrin (CD) from binary alpha-CD/beta-CD mixtures. To achieve this objective, a membrane process was designed with water as solvent and using the selective accommodation of a guest Igepal Co-890 (IG) within the beta-CD cavity. In the first part, host-guest inclusion study suggested the formation of IG(2)/beta-CD complex in presence of IG excess. Moreover, increasing IG/CD molar ratio to 1.5 leads to the complete complexation of beta-CD while alpha-CD remains free. The second part was devoted to the separation of alpha- and beta-CD using UF membranes (GH and GK from GE-Osmonics). The influence of parameters including membrane molecular weight cut-off (MWCO), transmembrane pressure (Delta P), IG/beta-CD and alpha-CD/beta-CD molar ratios, was investigated. MWCO, Delta P, IG/beta-CD and alpha-CD/beta-CD molar ratios respectively of 1000 Da, 2 bar, 1.5 and 11:89 are the optimized values to reach an alpha-CD separation factor as high as 38. The experimental data fits very well with predicted values calculated using the parameters found in the complexation study and rejection of single solutions indicating that selectivity predominantly originates from steric exclusion of species by membrane GH.

International audience; Poly(ether imide) (PEI) membrane with enhanced antifouling property was successfully prepared in a mild and simple procedure. The virgin membrane was firstly functionalized with an aqueous solution of diamino-terminated poly(ethylene oxide) block copolymer (PEG-diamine). Glutaraldehyde was used in a second step as a linker to chemically attach additional PEG-diamine to the primary amine groups grafted on PEI membrane surface. Immobilization of PEG segments was confirmed using Fourier transform infrared spectroscopy, X-ray photoelectron spectroscopy, scanning electron microscopy and atomic force microscopy. Ultrafiltration experiments revealed that the enhancement of a PEG coverage on the membrane surface provided superior anti-protein-fouling property. Cycles of protein filtration also demonstrated that the antifouling surface was stable over time and excellent ultrafiltration performance could be maintained without the need of harsh cleansing operation.

International audience; AB(2)-type amphiphilic (HPMC)(2)-PLA copolymers with various hydrophilic block lengths were synthesized using a three step procedure: ring-opening polymerization of L-lactide initiated by propynol, amination reduction of the aldehyde endgroup of HPMC, and thiol-click reaction. The resulted copolymers were characterized by NMR, DOSY-NMR, SEC and FT-IR. The cloud point (CP) was determined by UV-vis spectrometer. Data show that the HPMC block length has little effect on the Cp of the copolymers which is lower than that of HPMC. The self-assembly behavior of the copolymers was investigated from DLS, TEM, and critical micelle concentration (CMC) measurements. Spherical micelles are obtained by self-assembly of copolymers in aqueous solution. The micelle size and the CMC of copolymers increase with increasing HPMC block length. It is concluded that biobased and biodegradable (HPMC) 2-PLA copolymers could be promising as nano-carrier of hydrophobic drugs.

Two important tasks in probabilistic reasoning are the computation of the maximum posterior probability of a given subset of the variables in a Bayesian network (MAP), and the computation of the maximum expected utility of a strategy in an influence diagram (MEU). Both problems are NPPP-hard to solve, and NP-hard to approximate when the treewidth of the underlying graph is bounded. Despite the similarities, researches on both problems have largely been conducted independently, with algorithmic solutions and insights designed for one problem not (trivially) transferable to the other one. In this work, we show constructively that these two problems are equivalent in the sense that any algorithm designed for one problem can be used to solve the other with small overhead. Moreover, the reductions preserve the boundedness of treewidth. Building on the known complexity of MAP on networks whose parameters are imprecisely specified, we show how to use the reductions to characterize the complexity of MEU when the parameters are set-valued. These equivalences extend the toolbox of either problem, and shall foster new insights into their solution. We show a polynomial-time reduction from MAP to MEU that preserves the boundedness of treewidth.We show a polynomial-time reduction from MEU to MAP that also preserves the boundedness of treewidth.We use the reduction from MAP to MEU to characterize the complexity of MEU in credal influence diagrams.Experiments show that reducing MEU to MAP can be an effective alternative for solving MEU problems.

Low molar mass (LMM) biopolymers are highly required to design functional nanomaterials, which mainly find application in biomedical fields. However, the synthesis of LMM polymer is a challenging task. In this work, we report a partial enzymatic depolymerization process which allows to produce a series of LMM hydroxypropylmethyl cellulose (HPMC) polymer, with a weight average molar mass (M w) under and over 10,000 g mol −1 and low dispersity (Ɖ < 1.5). Variation of the starting HPMC grade, reaction time, and enzyme concentration were the key parameters to control the M w and yield of the target molecules. This approach provides a versatile way of producing LMM HPMCs with varying degrees of substitution, and having a single reactive aldehyde function at one chain extremity. LMM HPMC can find for instance application as building blocks for the development of new functional molecular architectures.

We adapt the theory of descriptive complexity to Bayesian networks, to quantify the expressivity of specifications based on predicates and quantifiers. We show that Bayesian network specifications that employ first-order quantification capture the complexity class PP; by allowing quantification over predicates, the resulting Bayesian network specifications capture each class in the hierarchy PP^(NP^...^NP), a result that does not seem to have equivalent in the literature.

International audience; Control of film structures made from a polystyrene-polystyrene sodium sulfonate-polystyrene (PS-PNaSS-PS) copolymer micellar solution is investigated in a THF/water mixture. Four different copolymers (varying molecular weights) are synthesised via RAFT (Reversible Addition Fragmentation chain Transfer) polymerisation. Depending on parameters such as copolymer molecular weight, solvent composition and copolymer concentration, the PS-PNaSS-PS triblock self-assembles into different morphologies in solution and dry state. The effect of each parameter is investigated using characterization techniques such as AFM, TEM, Cryo-TEM, SEM and SAXS. The morphologies obtained for PS-PNaSS-PS are found to be extremely sensitive when the water content of the micellar solution is low. Among the structures observed, a highly ordered nano-porous film is obtained using a PS10k-PNaSS6k-PS10k triblock copolymer solution containing 3.0 wt% of water. This micellar solution is used to prepare a porous membrane for filtration applications. Pure water filtration data suggest a pore size in the range of ultrafiltration, making these membranes attractive for applications in the food industry, for bacteria, virus and protein removal.

Amphiphilic diblock copolymer containing randomly distributed positive and negative charged monomers are synthesized using RAFT polymerization technique to be used as anti-bioadhesion coatings for hydrophobic surfaces. Quaternized 2-(dimethylamino) ethyl methacrylate and potassium 3-sulfopropyl methacrylate (P[qDMAEMA-co-KSPMA]) are randomly polymerized to yield an anti-bioadhesion block which is, in one pot, copolymerized with styrene as an anchoring block. This copolymer has demonstrated high anti-bioadhesion properties to avoid the blood clotting in medical devices through a simple and facile approach to preparation of pseudozwitterionic copolymers.

An ABA triblock amphiphilic copolymer is synthesized using RAFT chemistry. The self-assembled micelles of this copolymer are then used to prepare nano-organized porous films that could be used as filtration membranes. In this work a novel strategy is developed to build the nanostructures and perform their self-assembly using reversible and non-covalent interactions to create free volume between the micelles, thus giving tuneable porosity to the film. The self-assembly of poly(styrene)-b-poly(phenylboronic acid)-b-poly(styrene) block copolymer, occurs at high concentration through solvent evaporation, which induces a progressive decrease of the inter-micellar distance, and results in the formation of an in situ network of micelles and the final porous film. Subsequent permeability tests were conducted under different stimuli (pH and UV), generating cross-linking and chemical exchange reactions, to ensure the best balance between permeability and mechanical strength. This work highlights an original strategy for pore size control, and provides new insights towards the design of stimuli-responsive materials.

Cationic vectors are ideal candidates for gene delivery thanks to their capability to carry large gene inserts and their scalable production. However, their cationic density gives rise to high cytotoxicity. We present the proper designed core-shell polyplexes made of either poly(ethylene imine) (PEI) or poly(2-dimethylamino ethyl methacrylate) (PDMAEMA) as the core and zwitterionic poly(acrylic acid)-block-poly(sulfobetaine methacrylate) (PAA-b-PSBMA) diblock copolymer as the shell. Gel retardation and ethidium bromide displacement assays were used to determine the PEI/DNA or PDMAEMA/DNA complexation. At neutral pH, the copolymer serves as a protective shell of the complex. As PSBMA is a nonfouling block, the shell reduced the cytotoxicity and enhanced the hemocompatibility (lower hemolysis activity, longer plasma clotting time) of the gene carriers. PAA segments in the copolymer impart pH sensitivity by allowing deshielding of the core in acidic solution. Therefore, the transfection efficiency of polyplexes at pH 6.5 was better than at pH 7.0, from β-galactosidase assay, and for all PAA-b-PSBMA tested. These results were supported by more favorable physicochemical properties in acidic solution (zeta potential, particle size, and interactions between the polymer and DNA). Thus, the results of this study offer a potential route to the development of efficient and nontoxic pH-sensitive gene carriers.

A credal network associates a directed acyclic graph with a collection of sets of probability measures. Usually these probability measures are specified by tables containing probability values. Here we examine the complexity of inference in credal networks when probability measures are specified through formal languages. We focus on logical languages based on propositional logic and on the function-free fragment of first-order logic. We show that sub-Boolean and relational logics lead to interesting complexity results. In short, we explore the relationship between specification language and computational complexity in credal networks. We examine various formal languages for the specification of credal networks.We explore the relationship between specification language and computational complexity of inferences.We derive interesting complexity results for models specified using propositional logic.We discuss diferent semantics for relational credal networks specified using fragments of first-order logic.We derive interesting complexity results for models specified using fragments of first-order logic.

Mixed matrix membranes (MMMs) provide an exciting alternative to traditional membranes due to their favorable properties from both building blocks which are essential for certain separation applications. Block copolymer directed synthesis of MMMs is an innovative approach for the preparation of porous materials. In the current work the syntheses of mixed matrix membranes from a PMAA-b-PMMA block copolymer and functionalized iron oxide magnetic nanoparticles are demonstrated. The block copolymers were synthesized using a RAFT polymerization technique, along with three different types of magnetic nanoparticles with various surface properties. The thin film membranes were prepared by mixing different ratios of diblock copolymer in THF and iron oxide nanoparticles in water followed by tape casting or spin coating. The particles and the membranes were characterized using TEM, DLS, and SEM. The permeation behavior of the membranes was assessed using filtration tests in the presence and absence of a magnetic field.

We adapt the theory of descriptive complexity to Bayesian networks, by investigating how expressive can be specifications based on predicates and quantifiers. We show that Bayesian network specifications that employ first-order quantification capture the complexity class \(\mathsf {PP}\); that is, any phenomenon that can be simulated with a polynomial time probabilistic Turing machine can be also modeled by such a network. We also show that, by allowing quantification over predicates, the resulting Bayesian network specifications capture the complexity class \(\mathsf {PP}^\mathsf {NP}\), a result that does not seem to have equivalent in the literature.

The production of fish protein hydrolysates (FPH) is a promising route to add value to fish by-products due to their potential application as a source of interest peptide fractions. Using Alcalase for hydrolysis of tuna dark muscle by-product, the influence of enzyme/substrate ratio and hydrolysis time on the rate of interest peptide fractions (1–4 kDa) was studied. The rate of this fraction obtained under optimized conditions (temperature 55 °C, pH 8.5, enzyme/substrate ratio of 1% and 60 min of hydrolysis reaction) was 26%. The performance of the combined process associating ultrafiltration and nanofiltration membranes was evaluated for fractionation of produced hydrolysate in order to isolate fractions enriched in peptides of specific molecular weight (MW). The interest peptide fraction (1–4 kDa) was isolated and the positive effect of diafiltration on peptide purification process was underlined. The peptide fractions produced have a high nutritional quality which can be used in human nutrition and they have a potential for applications in aquaculture diets.

Credal networks are graph-based statistical models whose parameters take values in a set, instead of being sharply specified as in traditional statistical models (e.g., Bayesian networks). The computational complexity of inferences on such models depends on the irrelevance/independence concept adopted. In this paper, we study inferential complexity under the concepts of epistemic irrelevance and strong independence. We show that inferences under strong independence are NP-hard even in trees with binary variables except for a single ternary one. We prove that under epistemic irrelevance the polynomial-time complexity of inferences in credal trees is not likely to extend to more general models (e.g., singly connected topologies). These results clearly distinguish networks that admit efficient inferences and those where inferences are most likely hard, and settle several open questions regarding their computational complexity. We show that these results remain valid even if we disallow the use of zero probabilities. We also show that the computation of bounds on the probability of the future state in a hidden Markov model is the same whether we assume epistemic irrelevance or strong independence, and we prove a similar result for inference in naive Bayes structures. These inferential equivalences are important for practitioners, as hidden Markov models and naive Bayes structures are used in real applications of imprecise probability.

Influence diagrams are intuitive and concise representations of structured decision problems. When the problem is non-Markovian, an optimal strategy can be exponentially large in the size of the diagram. We can avoid the inherent intractability by constraining the size of admissible strategies, giving rise to limited memory influence diagrams. A valuable question is then how small do strategies need to be to enable efficient optimal planning. Arguably, the smallest strategies one can conceive simply prescribe an action for each time step, without considering past decisions or observations. Previous work has shown that finding such optimal strategies even for polytree-shaped diagrams with ternary variables and a single value node is NP-hard, but the case of binary variables was left open. In this paper we address such a case, by first noting that optimal strategies can be obtained in polynomial time for polytree-shaped diagrams with binary variables and a single value node. We then show that the same problem is NP-hard if the diagram has multiple value nodes. These two results close the fixed-parameter complexity analysis of optimal strategy selection in influence diagrams parametrized by the shape of the diagram, the number of value nodes and the maximum variable cardinality.

An accurate control of the pore size has been demonstrated with poly(ether imide) (PEI) membranes by using a chemical post-treatment with amino oligomers. Jeffamine M-2070, a mono aminated poly(propylene oxide)- co -poly(ethylene oxide) copolymer (PEG-amine, 2 kDa), and poly(ethyleneimine) (Pei), with molecular weight ranging from 0.8 to 2 kDa, were used as molecular weight cut-off ( MWCO ) modifying agents. PEI membranes having a MWCO of 120±10 kDa (measured by PEG/PEO rejection) and a pure water permeability of 180±30 l h −1 m −2 bar were prepared from a 16 wt% PEI solution in NMP by coagulation into a water bath. Modification with the mono-functionalized PEG-amine leads to a pore opening owing to the dissolution of the overreacted PEI chains. PEGylated UF membranes with increasing MWCO up to 500 kDa were thus prepared by varying the reaction time. Beyond this value, the mechanical strength of self-standing membranes decreased. In contrast, use of a poly-amino reagent prevented the dissolution by the cross-linking of PEI chains. Positively charged membranes with decreasing MWCO down to the NF range were prepared by using Pei oligomers. A plateau was reached after ca 3 h of reaction. Longer reaction times only gave rise to a decline of the permeate rate due to an increase of the dense top layer thickness. A mechanism involving a dynamic rearrangement of the PEI chains was proposed to account for the experimental observations.

International audience; Herein, a novel polymeric porous membrane was developed, without the use of any organic solvent in the initial dope solution and using a biobased polymer derived from cellulose: Hydroxypropylcellulose (HPC). HPC was dissolved in water (20 wt%) and the phase separation was induced by increasing the temperature above the Lower Critical Solution Temperature (LCST) of the polymer solution, around 40 °C in the concentration range concerned in this study. To fix the membrane morphology and to prevent any resolubilization in water during filtration tests, a chemical crosslinking was performed using Glutaraldehyde. The phase diagram of HPC/water system was first studied and not only the cloud point but also the spinodal curves were determined using optical transmission techniques. It was exhibited that HPC phase diagram is very weakly dependent on concentration up to large concentrations and that the metastable region is very small, i.e. the cloud point and the spinodal curves are very close in a large range of concentration. The membrane stability was tested in water and some organic solvents, thus demonstrating the efficiency of the chemical crosslinking during membrane formation. The swelling and mechanical properties of HPC membranes were also investigated depending on the operating condition during membrane formation, showing that the temperature ramp during the membrane formation, from initial to final temperature, have a significant effect on the crosslinking efficiency and hence on the swelling properties. Finally, adding a porogen (PEG200) into the collodion, filtration tests were performed and exhibited that the membrane filtration properties depend on the temperature ramp as well, showing higher water flux for the highest temperature ramp tested.

We examine the complexity of inference in Bayesian networks specified by logical languages. We consider representations that range from fragments of propositional logic to function-free first-order logic with equality; in doing so we cover a variety of plate models and of probabilistic relational models. We study the complexity of inferences when network, query and domain are the input (the inferential and the combined complexity), when the network is fixed and query and domain are the input (the query/data complexity), and when the network and query are fixed and the domain is the input (the domain complexity). We draw connections with probabilistic databases and liftability results, and obtain complexity classes that range from polynomial to exponential levels; we identify new languages with tractable inference, and we relate our results to languages based on plates and probabilistic relational models.

Mixed matrix membranes having combined properties of both polymeric and inorganic materials have become the principal focus on separation technology. This work insights an original strategy of mixed matrix membrane preparation incorporating positively charged inorganic nanoparticles (INPs) with negatively charged polymeric nanoparticles (PNPs) using spin coating technique. The PNPs with different morphologies (spheres, worms, and vesicles) made of poly(methacrylic acid)-b-(methyl methacrylate)) diblock copolymer were synthesized using RAFT dispersion polymerization in ethanol at 70 °C. The inorganic counterpart, iron oxide nanoparticles coated with poly(methacrylic acid)-b-poly(quaternized 2-(dimethylamino)ethyl methacrylate) were synthesized and incorporated into the membrane acting as a bridge between the negatively charged polymeric particles (due to the presence of opposite electrostatic charges). Permeability tests were carried out using the feed of different pH at various pressures.

Probabilistic logic programming combines logic and probability, so as to obtain a rich modeling language. In this work, we extend ProbLog, a popular probabilistic logic programming language, with new constructs that allow the representation of (infinite-horizon) Markov decision processes. This new language can represent relational statements, including symmetric and transitive definitions, an advantage over other planning domain languages such as RDDL. We show how to exploit the logic structure in the language to perform Value Iteration. Preliminary experiments demonstrate the effectiveness of our framework.

Hidden Markov models (HMMs) are widely used probabilistic models of sequential data. As with other probabilistic models, they require the specification of local conditional probability distributions, whose assessment can be too difficult and error-prone, especially when data are scarce or costly to acquire. The imprecise HMM (iHMM) generalizes HMMs by allowing the quantification to be done by sets of, instead of single, probability distributions. iHMMs have the ability to suspend judgment when there is not enough statistical evidence, and can serve as a sensitivity analysis tool for standard non-stationary HMMs. In this paper, we consider iHMMs under the strong independence interpretation, for which we develop efficient inference algorithms to address standard HMM usage such as the computation of likelihoods and most probable explanations, as well as performing filtering and predictive inference. Experiments with real data show that iHMMs produce more reliable inferences without compromising the computational efficiency.

Limited memory influence diagrams are graph-based models that describe decision problems with limited information such as planning with teams and/or agents with imperfect recall. Solving a (limited memory) influence diagram is an NP-hard problem, often approached through local search. In this work we give a closer look at k-neighborhood local search approaches. We show that finding a k-neighboring strategy that improves on the current solution is W1-hard and hence unlikely to be polynomial-time tractable. We also show that finding a strategy that resembles an optimal strategy (but may have low expected utility) is NP-hard. We then develop fast schema to perform approximate k-local search; experiments show that our methods improve on current local search algorithms both with respect to time and to accuracy. We show that k-local search on the space of strategies is W1-hard (hence unlikely to be polynomial).We devise exact and approximate pruning schemes that allow for faster k-move when performing (stochastic) k-local search.The approximate pruning is shown to be provably good in discarding strategies.Experiments with artificial and realistic problems show that our solutions allow a trade-off in accuracy and runtime.

Credal networks relax the precise probability requirement of Bayesian networks, enabling a richer representation of uncertainty in the form of closed convex sets of probability measures. The increase in expressiveness comes at the expense of higher computational costs. In this paper, we present a new variable elimination algorithm for exactly computing posterior inferences in extensively specified credal networks, which is empirically shown to outperform a state-of-the-art algorithm. The algorithm is then turned into a provably good approximation scheme, that is, a procedure that for any input is guaranteed to return a solution not worse than the optimum by a given factor. Remarkably, we show that when the networks have bounded treewidth and bounded number of states per variable the approximation algorithm runs in time polynomial in the input size and in the inverse of the error factor, thus being the first known fully polynomial-time approximation scheme for inference in credal networks.

A study has been undertaken of an advanced wastewater treatment approach using polymer nanofiltration membranes, in an attempt to obtain water of sufficient quality to allow it to be reused in the same production process or, alternatively, to be discharged without any problems. The study has initially focused on the removal of organic matter (reduction of COD) and the most representative ions present in the wastewater, such as Na+, Mg2+, Cl- y SO42-. In a first part of the study, with a view to optimising the experimental phase, a simulation has been performed of the nanofiltration process using the NanoFlux software. Among other things, the simulation allows the most suitable membranes to be selected as a function of the permeate flow rate and desired level of retention in the substances to be removed. The subsequent experimentation was carried out in a laboratory tangential filtration system that works with flat membranes. It was found that retention values of about 90% were obtained for the studied substances, with a good permeate flow rate, using low operating pressures. These results demonstrate the feasibility of the studied technology and its potential as a treatment for improving ceramic industry wastewater quality.Este estudio ha sido emprendido con el fin de acercar la nanofiltración a través de membranas poliméricas al tratamiento de las aguas residuales industriales de la industria cerámica, esperando obtener un agua con la suficiente calidad como para ser reutilizada en el propio proceso productivo o, alternativamente, poder verterla. El estudio se ha centrado en la eliminación de materia orgánica (reducción de D.Q.O) y algunos iones presentes en las aguas residuales, tales como Na+, Mg2+, Cl- y SO42-. En primer lugar, se ha realizado una simulación del proceso de nanofiltración usando el software NanoFlux. Entre otras cosas, la simulación permite seleccionar las membranas más apropiadas en función del caudal de permeado obtenido y en función del nivel de retención de la sustancia que se desea eliminar. En una segunda parte, se realizaron los ensayos experimentales en un sistema de laboratorio de filtración tangencial que trabaja con membranas planas. La retención obtenida para las sustancias estudiadas es de aproximadamente el 90%. Todo ello, con un buen caudal de permeado y usando bajas presiones de trabajo. Estos resultados demuestran la viabilidad de la tecnología y su potencial como tratamiento para mejorar la calidad de las aguas residuales de esta industria.

Hydrophobic mesostructured organosilica-based thin films with tunable mesoporosity offer promising opportunities for ultra low k applications. They are presently prepared by spin-on solutions and evaporation-induced self-assembly method (EISA) using polystyrene- b -polyethylene oxide block copolymers as structure-directing agents and methyltriethoxysilane (MTES) as organosilica precursor. With PS(1600)- b -PEO(1800) (between brackets is the molecular weights for blocks), different types of ordered mesostructures can be achieved by controlling the PS- b -PEO/MTES weight ratio, i.e. with Fm3m symmetry (evidenced for the first time in a organosilica film) and also with Im3m and P6mm space groups. Mesostructural ordering is characterized by X-ray diffraction (XRD), high-resolution transmission electron microscopy (HRTEM), and atomic force microscopy (AFM). The thin film porosity is analyzed using ellipsometric porosimetry (EP). Porogen removal and crosslinking of the organosilica network are followed by Fourier transform infrared spectroscopy (FTIR) and 29 Si magic angle spinning nuclear magnetic resonance (MAS NMR). The effect of the molecular weight and the hydrophilic lipophilic balance (HLB) of the used block copolymers on the mesostructure ordering is finally investigated using different types of polystyrene- b -polyethylene oxide block copolymers.

Predictions made by imprecise-probability models are often indeterminate (that is, set-valued). Measuring the quality of an indeterminate prediction by a single number is important to fairly compare different models, but a principled approach to this problem is currently missing. In this paper we derive, from a set of assumptions, a metric to evaluate the predictions of credal classifiers. These are supervised learning models that issue set-valued predictions. The metric turns out to be made of an objective component, and another that is related to the decision-maker’s degree of risk aversion to the variability of predictions. We discuss when the measure can be rendered independent of such a degree, and provide insights as to how the comparison of classifiers based on the new measure changes with the number of predictions to be made. Finally, we make extensive empirical tests of credal, as well as precise, classifiers by using the new metric. This shows the practical usefulness of the metric, while yielding a first insightful and extensive comparison of credal classifiers.

Homemade Cellulose Acetate nanofiltration membranes elaborated by phase inversion were surface modified by adsorption of alternated layers of sodium alginate (ALG) and chitosan (CHI). The membranes were characterised by Scanning Electron Microscopy (SEM), Atomic Force Microscopy (AFM), Energy Dispersive X-ray analysis (EDX) and by contact angle measurements. Separation performances of the prepared membranes are studied with 2 g/L synthetic aqueous solutions in frontal mode, under 15 bar pressure. Morphological characterization showed a formation of a lamellar structure till 15 bilayers followed by a transition to a granular structure. We observed an increase in water and salt fluxes till 20 bilayers. More adsorbed bilayers decrease fluxes. Retention rates of monovalent salts decrease and remain constant by reaching 15 bilayers. However, we observed a constant increase in the retention rate for salts containing divalent ions. Contact angle values showed a decrease till 15 bilayers followed by an increase till 30 bilayers. The correlation of the morphological characterization, contact angle measurements and the separation performances allowed as a better understanding of all the results. The layer-by-layer surface treatment makes CA membranes less sensitive to protein fouling with 15 adsorbed bilayers.

This research describes a simple approach for attaching poly(ethylene glycol) chains onto the surface of asymmetric ultrafiltration poly(etherimide) (PEI) membranes. Jeffamine ® M-2070, an amino terminated poly(propylene oxide)/poly(ethylene oxide) block copolymer (PEG-amine) was used as a surface modifying agent. PEI membranes prepared by conventional immersion precipitation were treated in an aqueous solution of PEG-amine and characterized by attenuated total reflectance FTIR, X-ray photoelectron spectroscopy, scanning electron microscopy and tensile tests. The effects of the PEG-amine concentration (2 and 10 wt%), modification time (0.5, 1, 3, and 5 h) and treatment temperature on the morphology, hydrophilicity, permeability, and anti-fouling property of the membrane were investigated. The water contact angle measurement and water absorbance ratio demonstrated that the treatment with PEG-amine for only 0.5 h resulted in a better wettability of the modified PEI membrane. An enlargement of the pore size was observed for the longer reaction times and the higher PEG-amine concentration accounted for a breaking of the PEI chains by over grafting. The best compromise in terms of morphology integrity, rejection and permeability performances, was obtained by using 2 wt% PEG-amine and reaction time lower than 5 h. Based on protein filtration experiments, the anti-fouling properties of the obtained membrane were evaluated, and the results showed a large improvement of resistance to protein fouling for the modified PEI membrane with respect to the unmodified one.

Sum-product networks are a relatively new and increasingly popular family of probabilistic graphical models that allow for marginal inference with polynomial effort. They have been shown to achieve state-of-the-art performance in several tasks involving density estimation. Sum-product networks are typically learned from data; as such, inferences produced with them are prone to be unreliable and overconfident when data is scarce. In this work, we develop the credal sum-product networks, a generalization of sum-product networks that uses set-valued parameters. We present algorithms and complexity results for common inference tasks with this class of models. We also present an approach for assessing the reliability of classifications made with sum-product networks. We apply this approach on benchmark classification tasks as well as a new application in predicting the age of stars. Our experiments show that the use of credal sum-product networks allow us to distinguish between reliable and unreliable classifications with higher accuracy than standard approaches based on (precise) probability values.

The present investigation revealed how the morphology and crystalline forms of a poly(vinylidene fluoride) (PVDF) membrane, prepared by using the vapor-induced phase separation (VIPS) method, were affected by the temperature at which PVDF was dissolved to form the casting solution. The results showed that there existed an important transition dissolution temperature, being referred to as the “critical dissolution temperature”, across which the morphology and crystalline forms of membranes drastically changed. With a dissolution temperature above the critical dissolution temperature, the prepared membranes were composed of nodules and the size of nodules decreased as the dissolution temperature decreased. And, with a dissolution temperature below the critical, membranes with lacy (bi-continuous) structure were obtained. In addition, the α-crystalline form of PVDF grew faster than the β-form when the dissolution temperature was below the critical, and the β-form became faster and dominant as the dissolution temperature increased to be above the critical. The existence of the critical dissolution temperature was observed for all the three solvents used in the present study, N-methyl-2-pyrrolidone (NMP), dimethylacetamide (DMAc), and N,M-dimethylformamide (DMF), indicating that the phenomena are general and not limited to a specific solvent. Also, it was observed that the two PVDF/NMP solutions, prepared with dissolution temperatures above and below the critical temperature, responded in different ways when water was added into the solutions. Though both solutions gelled, the solution with higher dissolution temperature started the gelation with a crystallization-initiation gelling process, while the other with a non-crystallization-initiation gelling. We propose that the competition between the two gelling processes play an important role in determining the PVDF membrane morphology and crystalline polymorphs, and the dissolution temperature can change the competition. And the critical dissolution temperature can be interpreted as a dissolution temperature across which the dominant gelling process switched.

A morphological study of polymeric membranes elaborated by VIPS process with the poly(etherimide) (PEI)/1-methyl-2-pyrrolidone (NMP)/water system was conducted using both experimental and numerical approaches. The mass transfer model is non-isothermal and integrated multicomponent diffusivities. The self-diffusion coefficients were estimated from the Vrentas and Duda free-volume theory and various formalisms that make a link between the self- and mutual-diffusion coefficients were tested. At different initial polymer concentration, the Dabral formalism also leaded to the best agreement between membrane morphology (determined by SEM analysis) and the predicted concentration profiles. The influence of the process parameters (RH and temperature) was then investigated on the membrane morphology. Reducing the relative humidity induced a longer delay in term of demixing time, leading to reduce the polymer concentration gradient near the air/solution interface, and therefore to prevent the cell size gradient in this region. Between 25 °C and 40 °C, the influence of the temperature on the concentration profiles was shown to be weaker. Lastly, the asymmetric membrane structures characterized by cell size gradients observed at the lowest polymer concentrations were prevented when doping the initial solution with a small amount of non-solvent. In such cases, the model predictions exhibited that the demixing time and the polymer concentration gradient were reduced, leading to more symmetric membranes.

This work describes the surface grafting control of poly(vinylidene fluoride) (PVDF) membrane with poly(ethylene glycol) methacrylate (PEGMA) via three different modification approaches of surface-initiated radical graft copolymerization, including thermal-induced radical polymerization, surface-initiated atom transfer radical polymerization (ATRP), and low pressure plasma-induced graft-polymerization. Two different surface grafting structures of PEGylated layer, brush-like PEGMA and network-like PEGMA, on PVDF membrane surface were achieved in this study. The chemical composition and microstructure of the various surface-modified PEGylated PVDF membranes were characterized by Fourier transform infrared spectroscopy (FT-IR), contact angle, atomic force microscopy (AFM), and X-ray photoelectron spectroscopy (XPS) measurements. Antifouling property of the modified PVDF membranes was evaluated according to the amount of protein adsorption and the filtration test for BSA solution in this study. Results show that the amount of adsorbed proteins on the modified PVDF membranes not only depends on the surface hydrophilicity and hydration capacity but also associates with the surface grafting structures of PEGylated layers on PVDF membrane surface. This study not only introduces different practical modification approaches to achieve a hydrophobic PVDF membrane grafting hydrophilic PEGMA, but also provides a fundamental understanding of various PEGylated grafting structures governing the performance of antifouling properties.

The formation of a surface liquid layer on the top of membrane forming systems made of poly(ether-imide) (PEI) and N-methylpyrrolidone (NMP) was clearly demonstrated during water vapor-induced phase separation (VIPS) through several in situ investigation methods including optical microscopy and dynamic water contact angle measurements for a qualitative approach, and Raman confocal and FTIR microscopy for a quantitative one. A mechanism involving the shrinkage from the polymer-rich phase consecutively to the surface phase separation is proposed to account for the significantly high concentration of PEI in the surface liquid layer. The emergence of a surface liquid layer during the phase separation process is discussed in terms of implications on morphology of membrane fabricated using VIPS and how it contrasts with liquid-induced phase separation.

BACKGROUND: Water sorption by polymer solutions during film fabrication may affect the film morphology and thereby the film properties. RESULTS: Water vapour sorption isotherms at 40 °C of poly(ether imide) (PEI)/N-methyl pyrrolidinone (NMP) solutions were experimentally determined. Considering PEI concentrations ranging from 0 to 20 wt%, results obtained from gravimetric measurements and near-infrared (NIR) spectroscopy, indicative of the total weight uptake and the local water concentration in the bulk solution, respectively, showed that a first stage occurred featuring a quasi-linear time dependence of the weight fraction of absorbed water. Deviations from linearity dependent on the PEI concentration were afterwards observed, this behaviour being more marked for concentrations higher than 14 wt%. To achieve good predictions, the mathematical description of the sorption results must take into consideration both the classical approach using diffusion equations and the possible time dependence of the surface concentration of water at the air-facing interface. CONCLUSION: This paper presents an analysis of water vapour sorption in PEI/NMP solutions combining local and global quantifications of the water content through NIR and gravimetric methods. The development of a mathematical model suggests a possible surface layer formation at the air-facing interface. Copyright © 2008 Society of Chemical Industry

This work aims to evaluate the potential of reverse micelles prepared from amphiphilic polylactide-b-poly(ethylene glycol) (PLA-b-PEG) block copolymers for controlled release of hydrophilic drugs. Different PLA-b-PEG diblock and triblock copolymers were synthesized by ring-opening polymerization of D- or L-lactide in the presence of a PEG macroinitiator. Reverse micelles were prepared by self-assembly of copolymers in a solvent/co-solvent/water system. Toluene was used as solvent, and ethanol as co-solvent to solubilize appropriate amount of water. The resulting nano-sized reverse micelles were able to encapsulate heparin in the hydrophilic core. Dynamic light scattering (DLS) and transmission electron microscopy (TEM) were used to determine the size and morphology of reverse micelles. The results show that reverse micelles are spherical in shape with sizes below 100 nm. Drug loaded reverse micelles were embedded in biocompatible membranes by mixing with 10% PLA solution in toluene with 1:3 volume ratio. In vitro release studies were realized in phosphate buffer saline (PBS) at 37°C. Heparin was almost totally released within 24h. Triblock copolymer reverse micelles exhibited faster drug release than diblock ones probably due to the more compact micelle structure of the latter. Therefore, PLA-b-PEG reverse micelles could be promising for applications as carrier of hydrophilic drugs when embedded in biocompatible membranes.

The very recent developments in preparation of filtration membranes from self-assembled block copolymers (BCPs) are reviewed in this paper. We look into membranes with very sharp pore size distribution and the approaches for manufacture of nanoporous films, including etching and templating, the advantages of the new process based on micelle assembly and phase inversion. The paper is divided in two main sections. In the first part different strategies to prepare membranes from block copolymers are summarized. The second part looks into the different factors affecting the pore formation, morphology and the characteristics of the membranes made from self-assembly of block copolymers.

According to the environmental protection regulations, fluoride concentration is considered as a substance of priority for assessment of drinking water quality to determine their impacts on the environment and public health. Saharan groundwater (Algeria) contains an excess of fluoride ions. Regular consumption of this water by the population of the region may cause endemic fluorosis. To solve this problem, we propose to treat this water by adsorption on different materials, such as activated alumina (AA), sodium clay (SC), and hydroxyapatite (HAP) in order to enhance its quality by reducing its fluoride concentration. The maximum adsorption is achieved with an adsorption capacity of the order of 0.9, 0.667, and 0.370 mg/g and with a percentage of 90, 83.4, and 73.95% for AA, HAP, and SC, respectively. Indeed, the acidity and alkalinity of the medium significantly affect the adsorption of fluoride ions. Results deduced from the curves of adsorption isotherms of fluoride ions showed that the retenti...

Symmetric porous membranes were prepared from concentrated poly(ether-imide) (PEI) solutions using vapor-induced phase separation (VIPS) coupled with a drying pretreatment. Moderately concentrated solutions of PEI in N-methylpyrrolidinone (NMP) (14–16 wt%) were first cast on glass plates and the solvent was then allowed to evaporate under a dry air flow up to the desired concentration (16–38 wt%) before forming the membrane structure by VIPS. The polymer concentration profiles (confocal Raman microscopy) and model predictions were in good agreement to show that the evaporation stage did not induce a polymer gradient concentration with PEI/NMP systems. These results were confirmed by examination of the final membrane morphology (SEM).

Polysulfone hollow fiber membranes were fabricated by a dry/wet spinning process, with N-methyl pyrrolidinone (NMP) and water as solvent and coagulant, respectively. The results indicate that air-gap length and ambient humidity have dramatic effect on the membrane morphology. Macrovoids in the membranes disappeared, reappeared, and redisappeared with increasing air-gap length. Evidence was obtained showing that, because of the high affinity of NMP for water, water vapor was drawn to the dope to induce phase separation, resulting in a nascent structure in the dope before it was immersed in the coagulation bath. We proposed that, shortly after the onset of its phase separation in the air gap, the dope behaved as a transient gel, which can reasonably explain the disappearance and reappearance of macrovoids. On the other hand, when the air gap was long enough to allow phase separation of the entire dope before it reached the coagulation bath, macrovoids redisappeared. In addition, we observed that the air-gap length required for the disappearance and redisappearance of macrovoids decreased with increasing ambient humidity, which can be well reasoned by the decreasing contact time with the humid air needed to bring about phase separation of the dope in the air gap.

Summary: The n-hexylamine-initiated polymerization of Ne-trifluoroacetyl-L-lysine N-carboxyanhydride in N,N-dimethyformamide was studied by nonaqueous capillary electrophoresis. A polypeptide with a broad molecular weight distribution was obtained and side reactions were clearly identified for polymerization at room temperature. The possibility of living polymerization at 0 °C was demonstrated. Synthesis of living polypeptides by primary amine initiated polymerization of NCA at low temperatures.

The formation of polymer films produced by the phase separation process occurring when a cast poly(etherimide)/N-Methyl-2-Pyrrolidone solution was exposed to humid air was studied. It was found that above a relative humidity value of 27%, the films presented a cell-like structure. The size of the cells was shown to decrease when the relative humidity increased. This effect was more pronounced at the film/substrate interface than near the surface. A cell-size gradient from one face of the film to the other was also clearly observed. A phenomenological model has been proposed to explain the morphology obtained by a phase separation induced by the water vapor in the studied system, taking into account thermodynamics and kinetics considerations. In this model, the cell-like structure setting up is shown to result from a nucleation and growth process accompanied by a coalescence coarsening. It was illustrated by a composition path on the ternary phase diagram. It was shown how the relative humidity influenced the film composition leading to the preferential nucleation compared to the growth and coalescence of the cells. Finally, it was found that the cell-size anisotropy resulted in the solvent and non-solvent mass transfers in the film, bringing to the fore the determining role of kinetics.

International audience; This paper presents a new technique for making porous polymer membrane without using organic solvent from a nontoxic and biodegradable polymer, polyvinyl alcohol (PVA). This method consists first in inducing a phase separation in water – 10 wt% PVA solution, by increasing the temperature above the polymer cloud point temperature. A PVA with a degree of hydrolysis of 72% was selected according to its low cloud point temperature (47 1C for 10 wt% polymer) allowing saving energy. The second step of the preparation procedure corresponds to the stabilization of the structure by jointly cross-linking the polymer to stabilize the porous morphology and drying to evaporate the dilute phase. A chemical cross-linking reaction was performed using glutaraldehyde with a molar proportion (GA/monomer alcohol) of 0.04. Results showed that the cross-linking control was the key step to obtain porous membrane morphology. Consequently, the temperature and the catalyst volatility of the cross-linking reaction act directly on the final membrane morphology. Additionally, it was shown that the control of the relative humidity during the elaboration process had an effect on the porosity of the membrane skin layer as depicted by SEM pictures and filtration tests.

Structure formation of poly(ether-imide) films using vapor induced phase separation was studied in relation with the elaboration process. Dense and cellular structure can be obtained depending on the processing parameters. When the solvent evaporation took place under humidified air, kinetic data from gravimetric measurements showed that the overall mass of the film was first increasing before decreasing to a plateau. An analysis in terms of drying rate revealed that mass transfers were occurring as if there was a thin layer of volatile molecule at the top of the forming film. Solvent and non-solvent transfers were demonstrated to be closely dependent on the relative humidity.

The focus of this work is to assess the quality of porous membranes prepared from naturally occurring kaolin clays and to evaluate the performance of tubular ceramic membranes treating integrated raw effluents from seafood industry. This material has been chosen due to its natural abundance, its non-toxicity, low cost and its valuable properties. The preparation and characterization of porous tubular ceramic membranes, using kaolin powder with and without corn starch as poreforming agent, were reported. SEM photographs indicated that the membrane surface was homogeneous. The effects of material compositions, additives and the relatively lower sintering temperature, ranging from 1100° to 1250°C, on porosity, average pore size, pore-size distribution and mechanical strength of membranes have been investigated. A correlation between microstructure and mechanical properties of membranes has been discussed. The performance of the novel ceramic membranes thus obtained was determined by evaluating both the water permeability and rejection. The obtained membrane was used to treat cuttlefish effluents generated from the conditioning seawater product industry which consumes a great amount of water. Cross-flow microfiltration was performed then, in order to reduce the turbidity and chemical oxygen demand (COD).

International audience