Desalination Effects And Its Modeling Prediction Of Nanofiltration Membranes Based On The Modification Of Polyacrylonitrile Ultrafiltration Membrane By Titania And Zirconia As Thin-Film Selective Layer Coatings

With an ever-growing population and an increase in their standard of living and needs,as well as the expansion of industrial and agricultural activities,there is still an increasing demand for good quality water around the world.Moreover,worldwide,water scarcity is recognized as not only a present but also a future threat to human activities.To meet this increase in demand,with no loss being tolerable,a water treatment plant,in all its aspects and all its forms,needs to be optimized and sophisticated.One of the major technological challenges nowadays is the development of sustainable processes for water desalination,water reuse,wastewater treatment,and recovery of valuable chemicals from water.Among membrane technologies,the pressure-driven process was reported to be effective for the separation of multivalent ions such as Mg²⁺,Ca²⁺,SO₄^2-,etc.The final goal of this work is to study experimentally and by means of newly developed models the rejection of ions such as Cl^-,Na⁺,SO₄^2-and Mg²⁺contained in groundwater,notably from wells and boreholes in light with synthesized membranes whose properties and structures are well characterized to bring these ions back to levels that comply with current WHO standards.Which is in perfect agreement with the five-year project(2021-2025)of the Chinese government regarding seawater and its desalination.In the present work,nine organic-inorganic nanofiltration(NF)membranes were synthesized through single deposition or co-deposition of nanoparticles(NPs)to conciliate the tremendous advantages of both organic and inorganic membranes.A polyacrylonitrile(PAN)ultrafiltration(UF)membrane has been used as a substrate,titania(Ti O2)and zirconia(Zr O2)nanoparticles(NPs)have been chosen for deposition under the in-situ formation process.Thus,titanium(Ti),zirconium(Zr),and even both nanoparticles(NPs)were successfully seeded on the platform of a polyacrylonitrile(PAN)ultrafiltration(UF)membrane previously coated with bio-glue(a co-deposition of dopamine hydrochloric and bicarbonate buffer or dopamine hydrochloric and tris buffer having undergone pyrocatechol deprotonation).Novel NF membranes were obtained under strict control of the quantity of nanoparticles favoring a defect-free,loose,and strong layer of titanium or zirconium;other membranes were synthesized by a co-deposition of these two nanoparticles.This latter category of membranes,while performing well for multivalent ions rejection,exhibited high water flux release.The tools in vogue,especially field emission scanning electron microscopy(FESEM),energy dispersive spectroscopy(EDS),atomic force microscopy(AFM),and very accessorily Fourier transform infrared spectroscopy/attenuated total reflectance(FTIR/ATR)have made it possible to fully characterize the structure of the new organic-inorganic nanofiltration(NF)membranes.The surface charge properties and dynamic water contact angle of the novel membranes were also measured for each membrane to facilitate more in-depth analyzes and the resulting values were used as input data in algorithms.Firstly,three novel thin-film composites(TFC)nanofiltration membranes have been synthesized.Depending on whether the top layer is made of titania,zirconia,or both,these membranes were called respectively PAN-DT-Ti,PAN-DT-Zr,and PAN-DT-Ti Zr.The three membranes under the optimized conditions(30°C,12h of hydrolysis time,and45 mmol/L)exhibited high rejection and permeation performance.All three membranes demonstrated long-term durability under 120-h testing.PAN-DT-Ti Zr showed the highest rejection(89-95%)for divalent cations with the salts'rejection sequence of Ca Cl2>Mg SO4?Mg Cl2>Na Cl>Na2SO4 while the water flux is not less than 55 L.m^-2.h^-1.Secondly,to understand in depth the tremendous advantage of NPs co-deposition,both dioxide of titanium(Ti O2)and dioxide of zirconium(Zr O2)have been co-deposited on ultrafiltration PAN membrane under strict mass ratio control of these nanoparticles.Accordingly,three other membranes were synthesized,PAN-DT-Ti Zr21,PAN-DT-Ti Zr11,and PAN-DT-Ti Zr12.The PAN-DT-Ti Zr11 membrane exhibited the highest performance rejection towards monovalent ions and the rejection rate for multivalent ions reached 95%.The permeate flux of this organic-inorganic nanofiltration membrane is as high as 58 L.m^-2.h^-1.A 180h continuous exploitation of PAN-DT-Ti Zr11 demonstrated its long-term operability,showed a slight variation of permeate flux less than 1.2%and the salt rejection remained high than 92.5%at the end of the operation.The co-deposition of nanoparticles has been successfully performed in strict compliance with the important insights provided by the tools and stoichiometric conditions.The last three membranes that were prepared displayed properties very similar to those previously mentioned.They include PAN-DC-Ti1,PAN-DC-Ti2,and PAN-DC-Zr and have been used to implement three different models.The filtration velocity approach of GUEROUT-ELFORD-FERRY helped to estimate the average pore size of PAN-DC-Ti2 to rp=0.538 nm.A six-day test carried out on PAN-DC-Ti2 demonstrated its long-term stability and showed a steady-rejection rate of 89.3%of Mg Cl2 salt and permeate flux very close to 56 L.m^-2.h^-1.The PAN-DC-Zr membrane was less performant toward monovalent ions,and its rejection rate for multivalent ions reached 99.3%.The water flux of the PAN-DC-Zr membrane was as high as 58 L.m^-2.h^-1.The most popular model,the Donnan-Steric Pore Model(DSPM),has been extensively studied.First by a reminder of its genesis and the history of the rise of modeling in the NF process,then by a rigorous development of the mathematical aspect of the model than those previously published.An investigation into the potentials and limitations of the DSPM model has inspired the development of new models to fully control the phenomena involved in transport in the active layer of NF membranes.The DSPM model has been therefore developed including dielectric exclusion,solvent viscosity within the pore,chemical potential.All the effects mentioned were included in DSPM in a form that did not increase its complexity of computing and at the same time did not reduce the practical applicability as a predictive tool.The predictions of DSPM have been compared with experimental data,but the prediction results were found very far from experiment ones.Even if the DSPM model has aroused great interest and attraction in the scientific world,by predicting with fairly good precision the performance of NF membranes,there remain important limitations of various nature depicted in this thesis dissertation.We have therefore demonstrated the shortcomings of the DSPM model and the need to develop new models in NF processes.A soft computing model has been applied in a whole new approach to make commonplace the complex and implicit extended Nernst–Planck equations that govern the transport of ions through NF membranes.Euler's numerical method was applied with a small step-size and the results obtained were very interesting.This numerical method has therefore been used to approximate the solution of differential equations that can't be solved by traditional ways,like the ways we use to solve exact,separable,even linear differential equations.Euler numerical method has been found to be very fast and accurate if the guess(the given initial value)is well-chosen on the one hand and a large number of increments is chosen on the other hand.Euler's numerical method has served as a basis to build a more complex computing method and therefore acts as a predictor-corrector method.This method,although simple and soft,has demonstrated in this work its great utility and performance over time.An algorithm that follows in real time the transport of the solute particle in its different positions in the active layer of the NF membrane object of the investigation has been written in an accessible programming language.Finally,the Euler's numerical method corroborated well the experimental findings since the relative error was found to be very low at 0.33%for Cl^-,0.47%for Na⁺,0.16%for SO4^2-,and0.09%for Mg²⁺(RE<<0.1).For more accurate results,a sophisticated computing model consisting of the Runge-Kutta method followed by Richardson extrapolation was applied in this investigation for the first time to solve the extended Nernst–Planck equations,which govern the solute particles'transport across the active layer of PAN-DC-Zr membrane.A smart,adaptive step-size routine is chosen for this simple and robust method,also known as RK4(fourth-order Runge-Kutta).Richardson's extrapolation was then used to get a better approximation of Cl^-,Na⁺,SO₄^2-and Mg²⁺rejection.The relative errors were,respectively,0.06%,0.03%,0.04%and 0.02%for Cl^-,Na⁺,SO₄^2-and Mg²⁺;so,the results of this last model corroborated perfectly the experimental findings.The statistical error analysis was performed through the agreement between experimental and model data while the probability density function curve of the novel organic-inorganic membrane helped to better understand from a physical point of view some typical characteristic parameters of the NF membrane.The PAN-DC-Zr membrane,like the NF membranes,has demonstrated excellent rejection performance against multivalent ions and the ion rejection prediction model based on the Runge-Kutta method has been shown to be effective.In addition,the long-term stability carried out on PAN-DC-Zr reassured that the membrane can be used for a long time before being replaced,especially if the membrane cleaning work consisting of its bombardment of drafts is periodically undertaken.The models are set up and calibrated for predefined input parameters,and they are not subject to change as chemical agents are changed on an experimental scale.Further validation of existing models and the development of better NF models is a requirement for better NF membrane characterization,perfect chemical rejection reassessment,and even more insights on additional experimental available data.As novelty and contribution,an organic-inorganic double-layer structure composite modified polyacrylonitrile nanofiltration membrane with high ion rejection and high pure water flux was prepared by co-deposition of titanium zirconium oxide nanoparticles,and its mechanism of action was explained through various characterization methods.Then,a new method consisting of Euler's numerical method to predict the transport and interception of solute particles in the nanofiltration membrane was established,and a computer program was written to reflect the position of solute particles in the active layer of the nanofiltration membrane in real time.Finally,it is found that the fourth-order Runge-Kutta numerical method and Richardson extrapolation method are combined to extend the Nernst-Planck equation to evaluate the ion transmission and interception of the nanofiltration membrane,and the predicted value is more consistent with the experimental data.Thus,this thesis dissertation is pioneer in developing novel models in a whole new approach for ions rejection reassessment in nanofiltration process.It may provide useful insights to design next-generation NFMs.The practical prediction tools,Euler and Runge-Kutta numerical methods may henceforth help in the choice and calibration of next-generation NF membranes'synthesis while waiting for the rise and expansion of machine learning.Beyond physical models,the rise of machine learning in the prediction of rejection performance of NF membranes is also a horizon to explore.