Asmaa Selim ,Andras Jozsef Toth ,Daniel Fozer ,Eniko Haaz ,Nóra Valentínyi ,Tibor Nagy ,Orsolya Keri,Lászlo Péter Bakos,Imre Miklós Szilágyi,Peter Mizsey,4,*
1 Department of Chemical and Environmental Process Engineering,Budapest University of Technology and Economics,1111Budapest,Hungary
2 ChemicalEngineering Department,National Research Centre,12611 Giza,Egypt
3 Department of Inorganic and Analytical Chemistry,Budapest University of Technology and Economics,1111Budapest,Hungary
4 Department of Fine Chemicals and Environmental Technology,University of Miskolc,3515 Miskolc,Hungary
ABSTRACT Pervaporation is an important membrane separation method of chemical engineering.In this work,silver-nanoparticles-poly (vinyl alcohol) nanocomposite membranes (AgNPs-PVA) are produced for the sake of improving its potentials for pervaporation of ethanol-water mixture so that the usual opposite trend between membrane selectivity and permeation can be reduced.The nanocomposite membranes are fabricated from an aqueous solution of poly (vinyl alcohol) with silver nanoparticles via the in-situ generation technique in the absence of any reducing agent.Successful generation of the nano size silver is measured by the UV-vis spectrum showing a single peak at 419 nm due to the plasmonic effect of silver nanoparticles.Our nanocomposite AgNPs-PVA membranes are characterized using scanning electron microscope (SEM),Fourier-transform infrared (FT-IR) spectroscopy,X-ray diffraction and thermogravimetric analysis (TGA).The pervaporation tests of our new AgNPs-PVA membranes show good results since at a higher temperature and higher ethanol concentration in the feed,the prepared membranes are highly permeable for the water having stable selectivity values and therefore our membranes show better performance compared to that of the other PVA-based nanocomposite membranes.
Keywords:Silver nanoparticle Poly (vinyl alcohol) nanocomposite membrane Hydrophilicity Swelling degree Pervaporation
Pervaporation is rapidly developing area of chemical engineering.The development includes the production of better and better membranes.A huge area of the application of pervaporation is the bio-fuel production.Bio-ethanol is considered to have huge potential as a green renewable energy source owing to its environmental benefits and its high efficiency.Ethanol was originally dehydrated by distillation separation process.However,high costs,low productivity,and extensive energy consumption which resulted in that distillation became unsuitable for commercial and practical production of highly concentrated bioethanol[1,2].Therefore,sustainable,efficient and economic production of bioethanol has turned to be the challenge.Membrane separation has been used as a convenient,economical,practical and green alternative of the bioethanol purification field.Among membrane separation processes;pervaporation technology (PV) has received much attention due to the advantages attributed to its simple operation condition,low energy consumption,low cost,eco-friendly and higher separation efficiency [3-5].
Pervaporation (PV) has been widely used as liquid separation process,as in the dehydration of valuable organic solvent using hydrophilic membranes,or concentration of organic solvent from an aqueous solution using hydrophobic membranes.Separation of organic-organic mixtures is also possible using organoselective membrane[6-8].The type of the membrane and its intrinsic properties are the core of PV process to acquire high performance.
Poly-vinyl alcohol(PVA)has been widely used for the dehydration of alcohol solutions because of their different advantages including;high hydrophilicity,film-forming ability,resistance to organic pollution,and chemical/thermal stability.However,PVA membranes are in a lack of stability and mechanical strength in aqueous solution due to their swelling phenomena and high crystallinity which eventually leads to decrease water selectivity and the dehydration performance [9-11].
In an attempt to inhibit the PVA membranes swelling in the ethanol dehydration process,various solutions were proposed[12-14].Among those solutions,cross-linking (most efficient)using various crosslinking agents such as glutaraldehyde [15],maleic acid [16]and citric acid [17].The crosslinked reaction consumes some hydroxyl groups in the PVA chain leading to lower hydrophilicity of the membrane that results in higher selectivity towards water molecules and the decrease of the permeation flux.Aspiring to overcome this problem several methods have been proposed such as new crosslinking agents,blending with other polymers [18]and hybrid nanocomposite membranes [19,20].
Nanocomposite membranes gained much attention due to their integrated benefits of the combination of nano-sized inorganic and polymer materials.Various inorganic nanoparticles were embedded in the PVA chain to decrease the membrane crystallinity and consequently promote the flux such as Zeolites [21],silica [19],CNTs [22],graphene and graphene oxide [23,24].
There are two ways to incorporate nanoparticles into a polymer chain;(i) the physical blending (ex situ) of afore synthesized nanoparticles by using appropriate capping agents to stabilize the nanoparticles inside the polymeric matrix;(ii) in situ generation of the nanoparticles in the polymer chain where the polymer chain plays as a stabilizer.
Researchers found that by using the ex-situ method,the nanoparticles tend to slowly aggregate and become challenging to accomplish uniform apportionment in the polymer matrix.In this context,the in situ method is more advantageous as there is no need for an additional capping agent and the ease to dominate the nanoparticle size accomplishing well dispersion [20,25,26].
Metal nanoparticles e.g.silver or gold nanoparticles gained great consideration due to their optical,electronic,and catalytic properties besides their distinct structure and high porosity.Silver nanoparticles represent an auspicious functional filler because of their unique and tunable properties and their surface Plasmon resonance.Therefore their application is preferred including sensors,signal enhancers in surface-enhanced Raman scattering (SERS)and antimicrobial agent [27].
It is worth mentioning that,the in-situ generation of silver nanoparticles in PVA membrane was reported by Premakshi et al.[28]and Chaudhari et al.[29]for isopropanol and Acetic Acid dehydration,respectively.However,in the first study,the in-situ generation was combined with the addition of reducing agent sodium borohydride and the nanoparticles were generated in PAA matrix before blending with PVA in the second step.
Meanwhile,in this study,in situ generation of silver nanoparticles in PVA matrix is carried out in the absence of reducing agent.In this case,the proposed reaction is based on the hydroxyl groups in the PVA chain which act as chelating sites to interact with the silver ions resulting in the reduction of Ag+.Such chelating forms a stable and closely packed three-dimensional structure nanocomposite [30,31].
The physicochemical properties and the morphology of the AgNPs-PVA nanocomposite membranes are studied consistently using different techniques such as UV-visible,FTIR,TGA,XRD,and SEM.Furthermore,the produced membranes are efficiently utilized for ethanol dehydration.The pervaporation performance is investigated by altering the amount of silver nanoparticles loaded to the PVA nanocomposite membranes.
Additionally,the incorporation of silver nanoparticles with their hydrophilic activity enhances the hydrophilicity of the PVA membranes.The influence of the different operating temperature and feed compositions on the PV performance are also evaluated.
Poly(vinyl alcohol) (C2H4O)x of molecular mass (MW) of 85000-124000 g·mol-1and 99%+hydrolyzed;Glutaraldehyde(GA) (C5H8O2) grade II with 25% in water are purchased from SIGMA-ALDRICH Chemie GmbH (Schnelldorf,Germany).Silver nitrate Powder 99%(MW 196.88)is obtained from Molar chemicals KFT (Budapest,Hungary);Ethanol absolute is bought from VWR Chemicals (Budapest,Hungary).
2.2.1.Pristine membrane
Poly (vinyl alcohol) solution of 5 wt% is prepared by dissolving PVA powder in distilled water at 90°C under stirring condition until a homogeneous solution is formed.The hot solution is filtered by glass disc filter in order to remove the non-solved polymer particles.The solution is then left over a night to release all the air bubbles.For the in-situ crosslinking 1.5 ml of glutaraldehyde(25%)is added at 80°C to the clear homogenous solution and stirred at the same temperature.The crosslinked solution is left overnight to get rid of effervescent air bubbles.The resulting solution is then poured into a petri dish and dried in air for 48 h.The dried membrane is later annealed at 60°C in an oven for 24 h.
2.2.2.Nanocomposite AgNPs-PVA membranes
In the absence of reducing agent,one step polyol process reaction is used to prepare silver nanoparticles-poly (vinyl alcohol)nanocomposite membrane.Different amount of silver nitrate is dissolved in water and added dropwise to equal amounts of PVA 5% solution.The solution is heated for 1 h at 70-80°C under continuous stirring.The resulting solutions have yellow to brownish yellow color according to the different concentration which proves the in-situ generation of the silver nanoparticle AgNPs.To break the aggregated silver nanoparticles forming the solutions are subjected to ultrasonic bath with a fixed frequency of 35 kHz for 45 min.The solutions are cooled down under ambient condition and left overnight to get rid of air bubbles.The solutions are then poured into clean petri dishes.The membranes are dried as aforementioned[30-34].The possible formation reaction of Ag nanoparticles with PVA matrix is simplified schematically shown in Fig.1 based on chelating process proposed earlier by Huang et al.[35]and also by Zidan [36].
To prove the formation of silver nanoparticles in the PVA matrix and characterize them.Ultraviolet-visible spectroscopy absorption spectra of crosslinked PVA and its composite membrane samples are recorded.The membrane samples absorbance are obtained in the range of 300-900 nm using Hewlett-Packard HP 8452A.
The interaction between the polymer matrix and the nanoparticles and the change in the chemical structure resulted from the incorporation of the silver nanoparticles in the polymer matrix is studied using ATR-IR spectroscopy (ATR mode of FTIR,BRUKER Tensor-37,USA).The spectra are in the range of 500-4000 cm-1.In each sample,the average over 16 scans is accomplished.

Fig.1.The possible formation reaction of Ag nanoparticles and their chelation [32].The finally obtained membranes are entitled according to the silver nitrate and silver concentration with respect to PVA listed in Table 1.The thickness of the resulting pristine and AgNPs-PVA nanocomposite membranes is 40-60 μm as measured with a thickness gauge.

Table 1 PVA nanocomposite membranes containing different silver concentration
The X-ray diffraction of Pristine PVA and AgNPs-PVA nanocomposite membranes are recorded using PANanalytical X’Pert Pro MPD X-ray diffractometer using Cu Kα (λ=0.15418 nm) radiation at room temperature in the range of 5°-65° and step time 5.080 s and 0.0083556° step size.
The effect of integrating in-situ silver nanoparticles to the PVA polymer matrix on the thermal stability of the membrane is studied on the pristine and nanocomposite membranes using thermogravimetric analysis.Thermogram curves are obtained using Perkin Elmer TGA-6 Thermogravimetric,in the range of 30-500°C at a heating rate of 10°C·min-1with a sample mass of about 10-12 mg.
The surface morphologies of the pure and nanocomposite PVA membranes are studied with scanning electron microscope (JEOL JSM-5500LV).All the samples are coated with sputtered gold-palladium alloy as a conductive layer for 3 min prior to the analysis.
3.6.1.Effect of silver content
For the amount of water absorption,the dried membranes are massed and immersed in water at room temperature.At every 3 min the membranes are taken out,wiped,massed and immediately returned to the water bottles.The mass of the absorbed water is calculated as follows:

where Wst,Wdis the mass of the membranes at time t and the mass of the dry membrane respectively.
Swelling measurements are completed by immersing the dry massed membranes in different water-ethanol mixtures at 30°C for 24 h to achieve equilibrium.After that the membranes are taken out and dried carefully with tissue paper to remove the solution from the surface and massed as fast as possible and immersed in the mixture solutions again.This process is repeated three times and the results are averaged.The degree of swelling percentage(DS) is calculated by the following equation:

where Mdand Msare the masses of the dry and swelled membrane,respectively.
3.6.2.Effect of the feed composition
In order to study the swelling behavior of the PVA pristine membrane and the AgNPs nanocomposite membranes at different feed composition,solutions in a range of 5 wt%-25 wt% water are used at 30°C.The previously method is followed at the swelling measurements.
Pervaporation dehydration experiments are carried out using a P-28 membrane unit from CM-Celfa Membranetechnik AG (Fig.2)where the flat sheet membrane is placed on a sintered disc which separates the feed and the permeate sides [37].The feed chamber capacity is approximately 0.5 L and the effective area of the membrane in contact with the feed mixture is 28 cm2.The cross flow circulation velocity is maintained at a constant value of approximately 182 L·h-1.The permeate side (downstream side) is connected to a VACUUMBRAND PC2003 VARIO vacuum pump and the pressure is kept at 533 Pa.The constant temperature is kept by a thermostat and controlled by a thermometer at the inlet and outlet of the unit.Previously,the membrane is immersed in a known volume of the feed mixture for 2 h to reach an equilibrium state with the feed component.Hereafter,the feed mixture is circulated until reaching steady state at the required temperature.The permeate is collected in two cold traps connected in series and cooled by immersing them in liquid nitrogen to avoid loss of material and then they are massed using a digital microbalance.
Pervaporation experiments are carried at 40,50 and 60°C and the feed mixture composition is varied between 10 wt% to 25 wt% of ethanol in water.The flux is calculated using the mass of massed permeate,the total time of the process and the effective area of the membrane.

where J is the flux;W is the mass of the collected permeate(g);A is the effective area of the membrane (m2),and t is the permeation time (h).
On the other side,the separation selectivity factor can be calculated by measuring the composition of the water and ethanol both in the feed and permeate by using RA-620 (accuracy ±0.00002,KEM Kyoto Electronics,Japan) and by comparing the result with the standard refractive indices of ethanol solutions.

Fig.2.Schematic illustration of CM-Celfa P-28 pervaporation unit.

where Pw,Pethare the mass fraction of water and ethanol in the permeate mixture and Fw,Fethare the mass fraction of water and ethanol in the feed mixture,respectively.
UV-vis spectroscopy absorption spectra is proved to be quite sensitive to the formation of silver colloids because silver nanoparticles exhibit an intense absorption peak due to the Surface Plasmon excitation.As mentioned before,all the membranes are characterized in the range of 300-900 nm.Fig.3 shows the absorption spectra for the pristine PVA membrane and nanocomposite membranes.It is perceived that except the PVA original membrane(M0),all the other membranes show a peak at around 419 nm which proves the in-situ generation of silver nanoparticles.On the other hand,Plasmon peak at (400 nm-420 nm) is typical for silver nanoparticles [30,38,39].

Fig.3.UV-vis spectra of pristine PVA and AgNPs-PVA nanocomposite membranes with different silver concentration.
Evidently,the absorbance intensity is increasing in conjunction with increasing silver content.The results from the UV-Vis can be well observed from the color difference between the PVA virgin membrane M0 and the nanocomposite membranes,in which the color changed from transparent to light yellow to yellowishbrown with increasing silver content as shown in Fig.4.
The FTIR spectra of the pristine PVA membrane and the different silver nanoparticles composite membranes are performed and the results are shown in Fig.5.The observed spectra are indicating that all the membranes display a characteristic band around 3300 cm-1,attributed to the -OH group stretching [14,40,41].Additionally,a narrow peak band can be observed at about 2950 cm-1due to the C-H stretching vibration and multiple characteristic bands of pure PVA,These are at 1400,1150 and 845 cm-1and belong to CH symmetric bending of CH2,C-O stretching and CH rocking of PVA respectively[14,41-43].It can be distinguished that upon the generation of the silver nanoparticles in the PVA matrix,there are some detectable changes in the intensity of the characteristic band of the -OH group other than the difference in the intensities of some absorption bands and these changes introduce new bands.Due to the interaction between the silver nanoparticles and the -OH group of the PVA,for the M0.5 the intensity of the -OH band shows a noticeable increase compared to M0.
Increasing the silver nanoparticles content in the PVA matrix from M0.5 to M2.5 leads to the increase of the-OH band intensity.This result approves increasing hydrophilicity of the polymeric membrane as a result of the presence of silver nanoparticles[29,44].Furthermore,increasing the intensity of the vibrational bands between 1150 and 650 cm-1shows a notable new band at around 660 cm-1that could belong to the interaction between the Ag and PVA matrix resulting in hydrogen bond [43-45].By the same token,developing AgNPs in the PVA matrix increases the spectrum in the range of 1500 to 1700 cm-1.This could be due to the broken chain of C-H bond in CH2and the new interactions generated in the polymer matrix [46].

Fig.4.Photographs of the pristine PVA and AgNPs-PVA nanocomposite membranes.

Fig.5.FTIR spectra of Pristine PVA and AgNPs nanocomposite membranes.

Fig.6.X-ray diffraction patterns of pristine PVA and AgNPs-PVA nanocomposite membranes.
The degree of crystallinity of pristine PVA membrane and AgNPs nanocomposite membranes is examined by XRD analysis and the results are shown in Fig.6.Fig.6 reveals sharp diffraction peak around 19.5 that can be observed for all the membranes which indicates the crystalline phase of PVA chain [47,48].Furthermore,two more peaks at 2θ around 38.1° and 44.1° can be observed which respectively identify the (1 1 1) and (2 0 0) in the facecentered cubic crystallographic planes for the metallic silver-The intensity of those peaks rises,however,concurrently with the silver content [47,49].On the other hand,increasing the silver content in the membranes from M0.5 to M2.5 leads to decrease the intensity of these peaks.These results demonstrate the decrease of the membranes crystallinity contemporaneously with the generation of silver nanoparticles and this behavior is prevalent with increasing silver content in the membrane.Additionally,it points out the successful in-situ production of the silver nanoparticles in the PVA chain via thermal reduction.
Scanning electron microscopy (SEM) pictures (Fig.7) show the surface of the Pristine PVA membrane and the nanocomposite membranes with different silver contents.The photographs reveal doubtlessly the generation of silver nanoparticles in the PVA matrix and equality integration of the silver nanoparticles all through the nanocomposite membranes.It can also be noticed that,at low concentration of AgNO3load,the generated particles are in a uniform degree.While increasing the Ag+content,the number of produced particles are also increasing and lead to the decrease of the particles size,resulting in an irregular degree of the particle size and agglomeration between the nanoparticles as it is shown clearly in pictures M1.5,M2,and M2.5.Nevertheless,AgNPs can be observed on the surface of the membranes at higher concentration of silver nitrate added.Furthermore,the SEM images show no cracks or voids around the AgNPs in the membranes produced.

Fig.7.Scanning electron microscopy (SEM) micrograph for the PVA and AgNPs-PVA nanocomposite membranes.
Hence,introducing nanoparticles can notably impact the thermal stability of PVA nanocomposite membranes[50-52],the thermal stability of the pristine membrane and AgNPs nanocomposite membranes are investigated by thermogravimetric analysis (TGA)in both nitrogen and oxygen atmosphere.
The obtained results from TGA and GTA for the pristine PVA and the nanocomposite membranes are plotted in Fig.8.Evidentially,the mass losses appeared in three stages in both nitrogen and oxygen atmospheres.The first forfeiture is approximately at 10% due to physically absorbed water at around 100°C while the second failure between 200 and 370°C is different depending on the decomposition of the side chain of PVA [37,47].Finally,the third thermal degradation between 380 and 500°C can be attributed to the decomposition of the carbonated residue such as the backbone of PVA and glutaraldehyde that are the same for both media[44,45].The difference can be observed in Fig.8(A)for oxygen and(B) for nitrogen.In the oxygen atmosphere,the curves represent higher slope for the decomposition of the PVA chain and the DTG curves imply broadening in all peaks,also,to shift towards lift and new peaks at the higher concentration membrane.However,under the nitrogen medium,the DTG curves shift to higher temperature with increasing silver concentration.This can be due to partial dehydration of the PVA and the polyene formation then its decomposition and the creation of the macroradicals and the final destruction at around 380°C [50,53].Furthermore,the residual amounts for all membranes in the inert thermal degradation are higher than those of under the oxidative atmosphere.Results obtained for both atmospheres show that the integration of the silver nanoparticles in the PVA membrane improves the thermal stability of the membrane especially at the higher concentration range,see M1.5,M2,and M2.5.
The membrane degree of swelling assumes an essential part of dehydration by pervaporation process.The degree of crystallinity,the diffusion coefficient of the solvent and the rate of diffusion are the most critical factors that cause swelling of any polymer[54].The plots 8 A and B show the increase of both the amount of water adsorbed and the swelling degree concomitantly with increasing the silver nanoparticles content,respectively.This indicates that the in-situ generation of the silver nanoparticles in the PVA matrix increases the affinity of the membrane towards water due to the interaction of -OH groups with the AgNPs and the formation of hydrogen bonds [14].Additionally,it manifests that the crystalline region of the PVA matrix is adequately retracted and the PVA crystallinity is decreased [50].This elicits the upgrade of the membrane hydrophilicity upon the implantation of AgNPs [55].This result is consistent with the analysis of the FTIR study results(Fig.9).

Fig.8.TGA and GTA for Pristine PVA and AgNPs nanocomposite membranes in O2(A) and N2(B) atmospheres.
Fig.10 demonstrates the influence of the feed composition on the PVA and AgNPs nanocomposite membranes.The results exhibit that the swelling degree is increasing for all membranes with increasing water content in the feed indicating the increment of the membrane affinity towards water[41,56].The obtained results show the strong interaction between the silver nanoparticles and water molecules.
4.7.1.Influence of Ag content
Fig.11 defines the impact of the silver nanoparticles on the total permeation flux and the selectivity of PVA membrane at 10 wt%water in the feed and 40°C operating temperature.The overall flux shows a systematic increase upon the generation of the Ag nanoparticles in the nanocomposite membrane and the trend continues with increasing silver nanoparticle content in the membrane.On the contrary,the selectivity decreases to around its half value consequent to the incorporation of AgNPs.However,at higher concentration of silver nanoparticles,the selectivity shows stable values.This phenomenon is a merit of the augmentation in the permeation flux.It is because of the loading of the silver nanoparticles in the PVA matrix that enhances the affinity of the nanocomposite membranes towards the water with their improved hydrophilicity [57].Furthermore,reducing the crystallinity of the membranes and the relaxation of the amorphous region upon the in-situ generation of the Ag nanoparticles promote the permeation flux of the AgNPs nanocomposite membranes for both water and ethanol while decreasing the separation factor of water [14,25].
Additionally,increasing the silver nanoparticle content can lead to agglomeration which can result in a possible defect in the PVA membrane.Therefore,the further increase of the silver nanoparticle helps to the permeation flux and decreases the separation factor but only until a limit.These results are very well supported with the FTIR,SEM and swelling degree studies.To assess the membrane efficiency and the effect of different silver nanoparticle concentrations on the pervaporation performance of the nanocomposite membranes,the individual fluxes of ethanol and water in a function of the silver nanoparticle content are plotted in Fig.12.
It clearly demonstrates that the water flux and the total flux are very close to each other and increasing synchronously with the silver content.Whereas the ethanol flux is entirely negligible yet and it slightly increases with the higher silver content showing a good selectivity of our membrane.
4.7.2.Effect of feed composition
The water content in the feed mixture shows a significant impact on the membrane performance in the pervaporation process.Thus,pervaporation study at different water contents in the feed for the pristine PVA and AgNPs nanocomposite membranes is carried out and results are plotted in Figs.13 and 14.The dehydration separation processes are accomplished for feed composition ranging from 75 wt%to 90 wt% of ethanol at 40°C.It is found that the total flux is increasing for all the membranes with the increase in the water content of the feed.
Two reasons for this phenomenon are (i) since water is a sufficient plasticizer for the PVA,increasing the water content will boost the interaction between the water molecules and the membrane [15],(ii) higher water concentration will increase the swelling degree of the membrane,which enhances the mobility of the polymer chain and then expands the chain spacing producing free volume hole in the PVA matrix [52,58].These result in facilitating the mass transfer for both the ethanol and water through the membrane and thus increasing the permeation flux while decreasing the separation selectivity of the membrane [59].
4.7.3.Operating temperature effect
The operating temperature has a significant impact on the physical and chemical properties of both the membrane and the feed solution.Thus,to study the effect of temperature on pervaporation process is investigated in the range of 40-60°C at 10 wt% water.The results show that the fluxes for all membranes are enhanced as the temperature increases.Two reasons for this behavior:(i)as the temperature of the feed solution increases the free volume in the polymeric chain is rising and also change in the amorphous region structure can happen [60].This leads to an increase in the penetration of both permeating molecules through the membrane,therefore higher permeation flux will be obtained,(ii)as the pervaporation process is well defined by the solution-diffusion theory then as the temperature increases the transport of driving force increases as well due to the increasing difference between the feed side and the permeate side vapor pressures [18].Thus,because of the aforementioned reasons the membrane selectivity will decrease with increasing permeation flux.
However,in this study that is the case only for the pristine membrane as shown in Figs.15 and 16 (A).While for the AgNPs nanocomposites membrane both selectivity and flux are improved with increasing feed temperature as shown in Fig.16(A) and (B).

Fig.9.Effect of silver nanoparticle content on the amount of water absorbed (A) and on the degree of swelling of PVA membrane in 10% water feed solution at 30°C (B).

Fig.10.Influence of feed composition on the swelling degree of PVA and AgNPs nanocomposite membranes at 30°C.
These demonstrate that AgNPs-PVA membranes exceptionally result in high fluxes and good separation being shown by the selectivity factor.At higher operating temperature these membranes overcome the trade-off behavior of the pristine PVA membrane.Improving the selectivity of the membranes at higher operating temperature is possible because of two reasons.First,raising the temperature could enhance the flexibility of the PVA chain,leading to an increase in the chain spacing.The reason is that the presence of ethanol in contact with silver nanoparticles helps to control the size of the nanoparticles and promote their dispersion [61,62].Thus,ethanol can help the silver nanoparticles to diffuse more inside the polymer chain and dissipate the possibly agglomerated nanoparticles on the surface.This phenomenon enhances the hydrophilicity of the membrane and,as a consequence,increases the affinity of the membrane towards the water against the ethanol.
The other possibility is due to the catalytic activity of the silver nanoparticles towards the ethanol.As reported both supported and non-supported silver nanoparticles show an excellent catalytic activity towards ethanol oxidation process [63-65].Therefore,as the temperature increases the catalytic activity of the silver nanoparticles can also improve the oxidizing part of the ethanol in the feed solution.Thus,the amount of ethanol in permeate decreases.However,the latter possibility is ruled out in this study since no other component is recognized.

Fig.11.Variation of total flux and selectivity with different Ag nanoparticles concentration at 10% water in the feed solution at 40°C.

Fig.12.Variation of total flux and individual fluxes of water and ethanol as a function of silver nanoparticles content at 10 wt% of water in the feed at 40°C.

Fig.13.Total Flux of PVA and AgNPs nanocomposite membranes for different feed compositions at 40°C.

Fig.14.Selectivity of PVA and AgNPs nanocomposite membranes for different feed compositions at 40°C.

Fig.15.Influence of feed temperature on the pervaporation performance of pristine PVA membrane at 10 wt% water in the feed.
In order to validate the first hypothesis,the individual fluxes of ethanol and water as a function of the silver nanoparticles content in the range of 40-60°C at 10%water in the feed plotted in Fig.17.It elucidates that the water flux is increasing synchronously with increasing the temperature.Whereas,the ethanol flux is decreasing in conjunction with the temperature rise for all the membranes.This clears up that the increase in the flux is only due to increasing the water permeation flux through the membrane.
Additionally,the pervaporation performances of the highest and lowest silver content nanocomposite membranes are studied with feed solution in a range of 10 wt%-25 wt%water and temperature range of 40-60°C.The values of the separation selectivity are listed in Table 2.The results imply that at 25 wt% water,i.e.,the lowest ethanol concentration the selectivity changes from 33 to 36 for M0.5.At M2.5 it varies between 14 and 23 by increasing the temperature from 40 to 60°C.On the other hand,at the highest ethanol concertation (10 wt% water) in the feed solution,if the temperature is increased from 40 to 60°C it results in higher selectivity of M0.5 membranes by a value of 8:in the same situation it is 27 in the case of M2.5.
These clarify that for different silver contents of the nanocomposite membranes,in the presence of ethanol,increasing the temperature results in higher selectivity of the membrane.However,with the temperature increase,the higher the ethanol concentration in the feed solution the higher the increase in selectivity is.Additionally,at higher silver content increasing the temperature with increasing ethanol concentration leads to a significant improvement of the selectivity value.These results prove that raising the temperature in the presence of ethanol will enhance the dispersion of the agglomerated nanoparticles both on the surface and inside the polymer chain.Thus,it results in the increase of the affinity of the membrane towards the water and its selectivity,too.
Table 3 displays a comparison of the separation performance for ethanol dehydration between AgNPs-PVA membranes produced in this work and PVA-based hybrid membranes found in the literature.It is apparent that the AgNPs-PVA nanocomposite membranes exhibit higher permeation flux as well as a more satisfactory separation factor values than those of the PVA-based hybrid membranes found in the literature.In contrast with other preparation methods,the blending of nanoparticles in the PVA matrix,that is,where AgNP nanoparticles with cross-linking and filling functions are in-situ generated and homogeneously dispersed within the PVA matrix,creates a comparable separation performance for the dehydration of ethanol.

Fig.16.Effect of operating temperature on different membrane performances at different temperatures for 10 wt% water in feed (A) Flux and (B) Selectivity.

Fig.17.Individual fluxes of ethanol and water as a function of the silver nanoparticle content in the range of 40-60°C at 10 wt% water in the feed.
In this study,high-performance water-selective nanocomposite membranes are prepared with in-situ generation of silver nanoparticles in crosslinked PVA matrix at five different concentrations.The silver nanoparticles prove to act as a hydrophilic filler agent facilitating the ethanol dehydration pervaporation process.

Table 2 Pervaporation separation selectivity of M0.5 and M2.5 at different temperatures and feed compositions

Table 3 Comparison of pervaporation performance of PVA-based membranes for ethanol dehydration
UV-vis,SEM,XRD,and FTIR confirm the formation of the silver nanoparticles and their interaction with the PVA matrix.The silver nanoparticles are well dispersed and generated in the PVA matrix during the preparation of the nanocomposite membranes.The membrane hydrophilicity is significantly improved by the addition of the silver nanoparticles which distinctly appears in the FTIR and swelling measurement analyses.
The dehydration of the water-ethanol mixture is investigated at three different feed compositions and three different temperatures,respectively.The permeation fluxes show simultaneous increase with the higher silver nanoparticle loading level.As it can be expected the separation selectivity slightly decreases.Below 90 wt%ethanol and at 40°C the measured flux and separation factor values of M2.5 are 12.67×10-2kg-1·m-2·h-1and 43.6,respectively,showing better performance than those of the pristine PVA.
However,an anti-trade off phenomenon can be observed between the permeation flux and the separation factor at increasing operating temperature values in the dehydration process by using the AgNPs-PVA membranes.This can be due to the ethanol ability to dissipate the quite probably agglomerated nanoparticles in the PVA matrix and on the membrane surface by increasing the affinity of the membrane towards the water.
The results show also that the ethanol fluxes and selectivity move against each other’s,similar to other membranes.At increasing temperature an improving selectivity of AgNPs-PVA membrane can be observed at lower fluxes.On the contrary,the higher water content in the feed solution results in higher fluxes than other membrane types;nevertheless the selectivity values show stable values at higher AgNPs concentration.This phenomenon means an important improvement compared to the pristine PVA membrane underlining the importance of the application of silver nanoparticles.
Based on our work it can be concluded as well that at a higher temperature and higher ethanol concentration in the feed that the prepared membranes show improved selectivity values towards the water.Our AgNPs-PVA membranes demonstrate better performance compared to those of other PVA-based nanocomposite membranes.Therefore our membrane can be recommended at the development phase of pervaporation and considered at chemical engineering design.
The authors are grateful for the financial support from the Hungarian National Scientific Research Foundations (OTKA) project:Capture of CO2 from biogases and industrial flue gases,project no.112699.This research was supported by the European Union and the Hungarian State,co-financed by the European Regional Development Fund in the framework of the GINOP-2.3.4-15-2016-00004 project,aimed to promote the cooperation between the higher education and the industry.
Chinese Journal of Chemical Engineering2019年7期