a Institute of Polymer Science and RIMST, Zhejiang University, Hangzhou 310027, PR Chinab Department of Chemistry, Zhejiang University, Hangzhou 310027, PR China
c Institut fur Technische Chemie, Universit¨at Essen, 45117 Essen, Germany
Received 19 July 2004; accepted 4 October 2004Available online 7 January 2005
urface. Chemical and morphological changes of the membrane surface were characterized in detail by attenuated total reflectransform infrared spectroscopy, X-ray photoelectron spectroscopy, and water contact angles measurements. Results reveallasma treatment time and the adsorbed PNVP amount have remarkable effects on the immobilization degree of PNVP. Pure wngle on the membrane surface decreases with the increase of PNVP immobilization degree, which indicates an enhanced hydro
he modified membranes. Static platelets adhesion experiment on the membrane surface was conducted to characterize the hemof the PNVP-modified PPMM. The statistical amounts of adhered platelets on unit membrane area decrease significantly, whichegree demonstrates that the hemocompatibility of the PNVP-modified membrane has been improved. Finally, permeation fluater and bovine serum albumin solution were measured to evaluate the antifouling property of the PNVP-modified membranesf which have shown an enhancement of antifouling property for the PPMMs. In a word, this pre-adsorption-plasma approach wae facile and useful in improving the hemocompatibility and the antifouling property of the PPMMs.2004 Elsevier B.V. All rights reserved.
For many membrane processes such as ultrafiltration andicrofiltration, flux decline resulted from protein adsorption,
oncentration polarization, pore blocking, and gel layer for-ation, etc., is a repugnant problem, which to a great extentrevents the wide-scale applications of membrane separa-
ion processes in aqueous solution treatment and biosepa-ation. A typical case is the filtration of complex fluids iniotechnology and food industries, from which proteins ad-
sorb onto the surface and deposit within the pores of mbrane, resulting in biofouling and/or flux reduction formembrane[1]. It is reported recently that the adsorptive foing could account for up to 90% of permeability losses[2].Therefore, much attention has been paid in the past 20to find out the mechanism of protein adsorption, andnow known that the electrostatic forces and the hydrophinteractions between certain domains in a protein moleand the hydrophobic membrane surfaces are the maintors [3–5]. Up to now, it has been generally acceptedhydrophilic materials are less sensitive to protein adsorpthan hydrophobic ones, the principle behind which is thadrophilic surface preferentially adsorb water rather thanlutes, leaving the membrane surface with protein-resist
22 Z.-M. Liu et al. / Journal of Membrane Science 249 (2005) 21–31
[2]. Therefore, there is considerable interest in developinghydrophilic membrane materials[6,7]. Unfortunately, mostinherently hydrophilic polymers are not suitable to fabricateinto membranes since such polymers are normally suscep-tible to chemical and thermal impacts in their applications.Furthermore, inherently hydrophilic polymers readily swellby water and thus the corresponding membrane structuresand properties can be changed dramatically. Therefore, inrecent years, there has been much interest in developing sur-face treatment techniques to alter the chemical and physicalproperties of hydrophobic membrane surface[8–24].
To obtain a hydrophilic membrane surface with antifoul-ing property, several methods have been investigated, whichcan be divided into two classes: physical and chemical modi-fications. Up to now, it is well known that adsorbing suitablehydrophilic polymers on the membrane surface alleviatesprotein fouling, while grafting the hydrophilic polymers onis expected to provide a much more stable and long-standingsurface layer[8–13,17,18]. For those polymers with func-tional groups in either back-bonds or side chains, numer-ous grafting methods are suitable, while in the case of thosehydrocarbon polymers with no active side chains or end-groups (for example, polyethylene or polypropylene), themembranes have to be activated prior to the graft polymer-ization. To generate active sites on the polymeric substratesf way[b ngi sur-fp oupso mod-i ctsw sur-fT re ei-t thinp e, orp oss-l trate[
ized on poly(ether sulfone) ultrafiltration membrane, whichresulted in higher filtration performance with less total andirreversible fouling [8]. Belfort and coworkers[17] hadalso successfully photochemically modified poly(ether sul-fone) ultrafiltration membrane withN-vinyl-2-pyrrolidoneto increase the surface wettability and decrease the ad-sorptive fouling. On the other hand, Higuchi et al.[26]covalently conjugated PNVP on the surface of polysul-fone membrane with a multiple chemical process. It wasreported that PNVP-modified polysulfone membrane gavelower protein adsorption from a plasma solution and muchsuppressed number of adhered platelets than original polysul-fone and other surface-modified membranes. Most recently,Kang et al.[27] cross-linked PNVP on microporous chlo-rinated poly(vinyl chloride) membranes to improve theirhydraulic permeation behavior. However, among all theseliteratures concerning the graft polymerization ofN-vinyl-2-pyrrolidone and/or the tethering of PNVP on the membranesurface, there is scarce report about the grafting/tetheringof this versatile polymer on the surface of polyolefin mi-crofiltration membranes. For the widely used hydropho-bic membranes such as polypropylene microfiltration mem-brane, one persistent problem causing performance declinehas been “membrane fouling” as mentioned above. Further-more, the poor hydrophilicity and biocompatibility for thist ouss iore-a fore,a findo e-v ility[
fort py-l uswo dr r, theg db ile� graftp mera nm hodw VPo at-m ared( erea es ont mea-s is ag teleta ex-p pati-b edm
or further reactions, plasma treatment is an effective8–15] as well as UV[16–18], �-ray [19,20] and electroneam[21,22] irradiations. In fact, plasma-initiated grafti
s a versatile technique for modifying the membraneace without affecting the bulk properties[8,10–13]. Ex-osing a membrane to plasma will generate active grn the membrane surface, which makes it possible to
fy the surface by graft copolymerization when contaith monomers, this is the so-called plasma initiated
ace grafting and can be performed in various gases[23].he other major processes about plasma technique a
her plasma polymerization on, in which a cross-linkedolymeric layer is deposited on the substrate surfaclasma treatment, in which intensive oxidation or cr
inking is introduced on the surface region of the subs14,24].
As a polymer soluble in both water and organicents, poly(N-vinyl-2-pyrrolidone) (PNVP) has been tocus of numerous applications including additives,etics, coatings and biomedicines. For example, Robind Williams[25] reported that PNVP could be simply aorbed on silica particles to inhibit protein adsorptionven to remove adsorbed proteins from the particleace. This versatile polymer was most recently immized on poly(ethylene terephthalate) (PET) film by Meinht al. [22] to fabricate supported-hydrogels. The princeason for successful PNVP applications is its exceiocompatibility with living tissues and extremely low c
otoxicity. PNVP has also been used to modify theace properties of polymeric membranes. By low-temperalasma treatment,N-vinyl-2-pyrrolidone was graft polyme
ype of membrane limit their further applications in aqueolution treatments, enzyme-immobilized membrane bctors, bioseparations and biomedical devices. Thereseries of work were carried out by our group to
ut a better way to modify the PPMMs with low irrersible adsorptive fouling and enhanced biocompatib28–31].
The goal of this study was to evaluate a new methodhe immobilization of PNVP on the surface of polyproene microfiltration membrane (PPMM). In our previoork [28], the graft polymerization ofN-vinyl-2-pyrrolidonen hydrophobic PPMMs using UV photo-assisted an�-ay preirradiation processes was discussed. Howeverafting polymerizations ofN-vinyl-2-pyrrolidone inducey UV irradiation were not efficient as expected, wh-ray preirradiation processes and plasma-initiatedolymerization always results in undesired homopolynd wastes the starting monomers[8,13,15]. Keep these iind, in the present work, an efficient and facile metas developed, which included the adsorption of PNn the surface of PPMM following with air plasma treent. Attenuated total reflectance Fourier transform infr
FT-IR/ATR) and X-ray photoelectron spectroscopies wdopted to investigate the chemical composition chang
he modified membrane surface. Water contact angleurements were also performed since the wettabilityood state identification of the modified surface. Pladhesion, protein adsorption together with filtrationeriments were conducted to evaluate the hemocomility and the antifouling property of the PNVP-modifiembranes.
Z.-M. Liu et al. / Journal of Membrane Science 249 (2005) 21–31 23
2. Experimental
2.1. Materials and chemicals
PPMM with porosity of 45–50% and an average pore di-ameter of 0.07�m was prepared in our laboratory with themelt-extruded/cold-stretched method[29,30]. To remove anychemicals adsorbed on the membrane surface, nascent mem-brane was presoaked and then washed with acetone, and driedin a vacuum oven at room temperature for 4 h before the initialmass was determined. Poly(N-vinyl-2-pyrrolidone) (PNVP,K-30, the average molecular weight is about 40,000 g mol−1)was of analytical grade purity and purchased from Shang-hai Chemical Agent Co., China. Bovine serum albumin(BSA,Mw = 67,000 Da) was a commercial product of Sino-American Biotechnology Co. and used as received. The sol-vents were all of analytical grade and used without furtherpurification.
2.2. Immobilization of PNVP on the membrane surface
The process used in this work is schematically describedin Fig. 1. In our previous work[30], this process was usedto graft a sugar-containing polymer onto the PPMM surface.Similar method was also described by Terlingen et al.[32] toi ecyls owni kedi ons( keno res-s NVPo
A
approa
whereWa is the weight of PPMM after the adsorption ofPNVP andW0 is the weight before that.
The immobilization of PNVP was carried out in a glow-discharge plasma reactor connected with a diffusion pumpand two rotary pumps. The plasma reactor used in this workwas purchased from Beijing KEEN Co., China. Tubular typePyrex reactor (10 cm× 150 cm) was rounded with a pairof copper electrodes. These two electrodes were poweredthrough a matching network by a 13.56 MHz radio-frequency(rf) generator. On the basis of systematic experiments con-sidering membrane etching and immobilization degree (ID)induced by plasma, 30 W was chosen as the applied rf powerfor all the experiments described here. The PPMM with ad-sorbed PNVP was mounted on a poly(vinylidene fluoride)frame which was put in the middle of the plasma cham-ber so that both sides of the membranes were subjected toplasma treatment. The chamber was vacummized to 10 Paand kept constant at this value by a pressure regulator con-nected with the plasma chamber. Then, plasma was gen-erated and the membrane was exposed to the air plasmafor a predetermined period of time. After plasma treatment,the membrane was taken out from the chamber, washedintensively with methanol, aqueous sodium hypochlorite(NaOCl) solution, and de-ionized water, respectively. TheNaOCl solution was adopted here because it was reportedt ly-m oni-c cedp
anes
I
w andW
mmobilize the physically adsorbed layer of sodium dodulfate, a surfactant, onto a polypropylene film. As shn Fig. 1, before air plasma treatment, PPMM was soan PNVP/methanol solutions with different concentrati1–20 wt.%) for 20–24 h. After that, the membrane was taut from the PNVP solutions and dried under reduced pure to constant weight. The adsorption degree (AD) of Pn the membrane was calculated as following:
D = Wa − W0
W0× 100% (1)
Fig. 1. Schematic representative of the
ch used for the surface modification of PPMM.
o be a useful solvent for the removal of NVP homopoers [33]. At last the membrane was washed ultras
ally in de-ionized water for 30 s, then dried under reduressure.
The immobilization degree of PNVP on the membrurface was determined gravimetrically according to:
D = Wt − W0
W0× 100% (2)
hereW0 represents the weight of the nascent PPMM,t is the weight of the PNVP-modified one.
24 Z.-M. Liu et al. / Journal of Membrane Science 249 (2005) 21–31
The grafting yield was designed as:
Yield = immobilized degree
adsorption degree× 100% (3)
2.3. Characterization
To investigate the chemical changes between the origi-nal membrane and the PNVP-modified membranes, Fouriertransform infrared spectroscopy (Bruker Vector 22 FT-IR)with an ATR unit (Attenuated Total Reflection, KRS-5 crys-tal, 45◦) was used. Prior to the measurements, the sampleswere dried under vacuum at room temperature for 36 h inthe presence of P2O5. Data were recorded between 700 and3100 cm−1 with 32 scans.
Spectra of X-ray photoelectron spectroscopy (XPS) wererecorded on a PHI 5000C ESCA system (PHI Co., USA) em-ploying Al K� excitation radiation (1486.6 eV). The X-raysource was run at a power of 250 W (14.0 kV). A pass energyof 93.9 eV was used when obtaining the survey spectra. Priorto the measurements, the samples were dried under vacuumat room temperature for 48 h in the presence of P2O5. Thepressure in the analysis chamber was maintained at 10−6 Paduring measurements. To compensate for surface chargingeffect, all survey spectra were referenced to the C1S hydro-c
EMo or-m ande icro-s was2 asu anes.A sili-cS werem ubblemD riort d in adtC -s ad me-d eter-m thec weref ob-s g ana d sy-r ag-n andu oted,c ter oft ere
averaged to get a reliable value, the standard deviation waswithin ±5%.
2.4. Platelet adhesion evaluation
Experiments were carried out with fresh platelet-enrichedplasma (PRP) bought from the hospital. Firstly, the mem-brane was placed onto a piece of flat glass; a 20�L PRPwas carefully dropped on the membrane. After incubationfor 30 min at room temperature, the membrane was rinsedseveral times by phosphate-buffer solution (PBS, Na2HPO44.32 g, KH2PO4 1.18 g, de-ionized water 1000 mL, pH 7.4).Adhered platelets on the membrane were fixed with 2.5%glutaraldehyde/PBS solution for 0.5 h, followed by dehydra-tion procedure using a series of ethanol–water mixtures (0,30, 40, 50, 60, 70, 80, 90, 100 vol.% of ethanol) for 30 min,respectively. At least five SEM photographs with magnifica-tion of 1000 were taken randomly from each sample surfaceafter gold sputtering, and the amount of platelets adsorbedon unit membrane surface was counted.
2.5. Filtration experiment
This process was conducted according to the report ofChen and Belfort[8]. The transmembrane pressure was keptc ser-v m-b n ofd very5 suc-c ordeda tionw are-f n atr2 a-t hedq wa-t BSAst anyl , them edw em-b tionso re-p backt wasm
per-f mem-b2 hee -b tioneda
arbon peak at 284.6 eV.Surface morphologies of PPMM were taken through S
r AFM (AFM images were included in supporting infation). Membrane samples were sputtered with gold
xamined using a Stereoscan 260 scanning electron mcope (Cambridge, UK). The accelerating voltage used0 kV. A Seiko instrumental SPA 400 AFM system wsed to examine the surface topography of the membrFM images were acquired in the tapping mode withone tip cantilevers having a force constant of 20 mN cm−1.tatic water contact angles of the membrane surfaceeasured by both sessile drop method and captive bethod at 25◦C with a contact angle goniometer (KRUSSSA10-MK2, Germany) equipped with video camera. P
o the measurements, the membranes were conditioneesiccator for at least one month in the presence of P2O5
o eliminate the effect of polar groups (such as CO andOH) generated by plasma treatment[30]. In a typical ses
ile drop method, a water drop (∼10�L) was added onry membrane sample in air, the image was recorded imiately (less than 3 s) and a water contact angle was dined from the image with the imaging software. For
aptive bubble method, air/water/membrane interfacesormed by immersing small membrane panel in a glasservation cell containing de-ionized water and releasinir bubble beneath the membrane surface with a curveinge. A camera fitted with a video screen provided a mified image of the bubble, which was then recordedsed to calculate the contact angle. Unless otherwise nontact angle measurements were made one month lahe modification[14,30]and at least 10 contact angles w
onstant at 0.1 MPa. In a typical run, the solution reoir was initially filled with de-ionized water, and the merane was precompacted for 30 min during the filtratioe-ionized water, then the water flux was measured emin until the flux remained constant for at least threeessive readings, the average of five readings was recsJ0. Next, the reservoir was emptied and a BSA soluas poured into it. The BSA solution was prepared by c
ully dissolving BSA powder in a phosphate-buffer solutiooom temperature, and the buffer solution (4.56 g NaH2PO4,3.00 g Na2HPO4, 149.76 g NaCl, 4.02 g KCl, de-ionized w
er 1000 ml, pH 6.9) was obtained by dissolving preweiguantities of the salts in the desired amount of de-ionized
er. For each cycle of experiment, the concentration ofolution was kept constant as 1.0 g L−1. The flux of BSA fil-ration (J1) was recorded until the flux did not changeonger, which usually needed at least 40 min. After that
embrane was rinsed with 0.05 N NaOH for 30 min followith 30 min permeation of de-ionized water. Then, the mrane was removed and placed upside down, the filtraf 0.05 N NaOH solution and of de-ionized water wereeated each once. Finally, the membrane was flipped
o its original orientation and the de-ionized water fluxeasured which was designed asJ2.The effect of static BSA adsorption on the permeation
ormances of the membranes was also measured. Therane was previously immersed in 1.0 g L−1 BSA solution for4 h at 30◦C with constant vibration of 100 rpm. After tquilibrium of adsorption, the water flux (J3) of the memrane was measured according to the procedure menbove.
Z.-M. Liu et al. / Journal of Membrane Science 249 (2005) 21–31 25
3. Results and discussion
3.1. Immobilization of PNVP on the PPMM surface
Plasma is one of the most widely used techniques for thesurface modification of polymeric membranes, the efficiencyof which can be greatly influenced by such factors as power,treatment time and gas atmosphere. As schematically de-scribed inFig. 1, simply treating the PNVP-presoaked PPMMwith air plasma will readily immobilize PNVP or cross-linkPNVP on the hydrophobic membrane surface. In this case,one of the most effective factors is the plasma treatment time.Therefore, the influence of plasma treatment time on the im-mobilization degree of PNVP was examined and the typicalresults are listed inTable 1. It can be seen that, with nearlysimilar amounts of adsorbed PNVP, the immobilization de-gree increases as the plasma treatment time prolongs from10 to 30 s. While after the plasma treatment time exceeded30 s, the immobilization degree declines slowly. The calcu-lated grafting yield, which is an illustration of the plasmatreatment efficiency, follows the same tendency as immobi-lization degree. These results can be explained that, as theplasma treatment time increases, more and more active sitesare produced both on the membrane surface and the adsorbedPNVP, the combination of these active sites will chemicallyb and,p e ex-t asmae bel fac-t oneda
to the effective depth of plasma treatment. It is well knownthat low-temperature plasma techniques are very surface se-lective, and the effective depth of plasma treatment is usuallyreported to be about 1–1000A [34]. It is obvious that thethickness of the adsorbed PNVP layer increases with the in-crease of the adsorption degree. When the thickness is uponthe effective depth of plasma treatment, the reactive sites inthe atmosphere are difficult to attack directly on the mem-brane surface (or the polypropylene bulk molecules). There-fore, the number of active sites generated on the membranesurface for PNVP immobilization decreases. In that case,most of the adsorbed PNVP only react with each other. Thispart of PNVP can be washed away after plasma radiation,thus the immobilization degree of PNVP on the membranesurface decreases.
In our previous paper[28], PNVP was also tethered ontoPPMM through the UV-assisted or�-ray induced surfacegrafting ofN-vinyl-2-pyrrolidone monomer. Typical resultsfrom the UV-assisted graft polymerization are compared inTable 2with the data obtained from the direct immobilizationmethod described in this work. It can be seen that the plasma-induced immobilization of PNVP is relatively simple, and
ind PNVP on the membrane surface. On the other hlasma treatment would make bond break, which to som
ent leads to the surface deterioration (the so-called pltching) [24], and the polypropylene membrane would
ightened after long-time plasma treatment. These twoors determine the resulted immobilization degree mentibove.
For the pre-adsoption-plasma treatment method discuere, the amount of adsorbed PNVP is another key facelating to the immobilization degree of PNVP. As caneen fromFig. 2, with the increase of the adsorptionree, the immobilization degree increases at first andecreases remarkably. When the adsorption degree is h
han 50 wt.%, however, the immobilization degree of PNeems to be constant. These observations might be as
able 1ffect of plasma treatment time on the immobilization degree of PNVP
a Other plasma treatment conditions: pressure 10 Pa, 50 Hz, 30 W,b PNVP/methanol concentration (wt.%) was 5.00%.c Yield was calculated from AD/ID× 100%.
26 Z.-M. Liu et al. / Journal of Membrane Science 249 (2005) 21–31
Table 2Comparison of two methods used for the surface modification of PPMMwith NVP or PNVP
Concentration of NVPor PNVP (wt.%)
Method ID(wt.%)
Yield(wt.%)
10.36 UV-assisted, 20 min 1.35 13.0320.63 UV-assisted, 20 min 1.68 8.1430.83 UV-assisted, 20 min 1.37 4.4440.94 UV-assisted, 20 min 1.68 4.1018.52 Plasma treatment, 30 s 6.41 34.9323.82 Plasma treatment, 30 s 7.44 31.2629.03 Plasma treatment, 30 s 15.52 53.4637.54 Plasma treatment, 30 s 8.13 21.65
Data for the UV-assisted surface grafting of NVP were cited from[28], andin this situation, the yield was calculated from the weight increase of themembrane divided with the total monomer weight.
the immobilization degree can be easily control through theadjustment of PNVP pre-adsorption amount.
3.2. Characterization of the PNVP-modified membranes
The membrane surface was characterized by FT-IR/ATRspectroscopy and the typical spectra are depicted inFig. 3.Compared the PNVP-modified membranes with the nascentone, a new peak appears at 1660 cm−1, which is correspond-ing to the typical absorption of carbonyl group in the pyrroli-done ring of PNVP. Another new peak at about 1260 cm−1
is ascribed to the stretching vibration of CN bond. More-over, with the increase of PNVP immobilization degree, theintensity of these peaks become stronger.
To further verify the chemical changes underwent on themembrane surface, X-ray photoelectron spectra (XPS) forboth the PNVP-modified and nascent PPMMs were taken. Itcan be seen from the survey spectrum inFig. 4(a) that, there
F ; (b)6 M.
Fig. 4. Survey XPS spectra of the membranes: (a) nascent PPMM; (b)6.73 wt.% PNVP-modified PPMM.
is one peak at 284.6 eV assigned to C1S for the nascent mem-brane. A very small amount of oxygen at 531.76 eV is alsoevident, which might be due to the surface contaminationof the membrane surface. For the membrane immobilizedwith 6.73 wt.% of PNVP (Fig. 4(b)), a new peak appears at399.56 eV, which is designated to N1Sin the pyrrolidone ring.Furthermore, the intensity of O1S at 531.76 eV increases dra-matically. From the theoretical composition of PNVP, onecan understand that the content of N and O atoms should beequal. However, the results inTable 3reveal that, the amountof O atom is much higher than the N atom for the PNVP-modified membrane. This fact is thought to be due to thegroups containing O element on the membrane surface in-troduced by air plasma treatment. Such groups can also beintroduced on the surface when the membrane was exposedto atmosphere after the plasma treatment.
The surface morphology of different PPMMs studied inthis work was observed with scanning electron microscopyand typical photos are shown inFig. 5. It can be seen thatthere are various micropores distributed on the surface withan average pore diameter of about 0.07�m for the nascentPPMM (Fig. 5(a)). Upon plasma treatment, the pore diam-eters were somewhat enlarged due to the etching effect ofplasma (Fig. 5(b)). With the increase of immobilization de-gree, the PNVP layer spread on the surface and blocked them
3m
ntacta DM)a are
ig. 3. FT-IR/ATR spectra of the membranes: (a) nascent PPMM.73 wt.% PNVP-modified PPMM; (c) 10.12 wt.% PNVP-modified PPM
icropores of the membrane (Fig. 5(c)).
.3. Surface properties of the PNVP-modifiedembranes
Two methods were adopted to measure the water congles of the membranes, namely sessile drop method (Snd the captive bubble method (CBM), and the results
shown inFig. 6. Use of video capture for measuring contactangle of porous materials has been widely used in recent years[12–14,35,36]. Following the reported process, comparisonof these values between samples provides a semi quanti-tative measure of the differences in wettability for porousmembranes and, to a certain extent, removes issues asso-
F0ad
ciated with porous media. For the nascent membrane, thesurface is much hydrophobic, the water contact angle mea-sured by SDM was higher than 110◦. It can be seen fromFig. 6 that the water contact angle declines gradually to 75◦with the increase of PNVP immobilization. This means thatthe membrane surface is somewhat hydrophilic after the im-mobilization of PNVP. As demonstrated by XPS analysis,after the air-plasma treatment, together with the immobiliza-tion of PNVP macromolecules, polar groups containing oxy-gen are also introduced on the membrane surface. Both theimmobilized PNVP and these polar groups can change thewettability of the membrane surface. To eliminate the con-tribution of the small polar groups generated by air-plasmatreatment, all membranes to be measured the contact anglewere conditioned in a desiccator for at least one month. In thatcase, the polar groups generated by the plasma treatment canre-orientate in the surface region and bury in the interior ofthe membrane surface (the so-called “hydrophobic recovery”phenomena of plasma treatment)[14,30]. Therefore, the hy-drophilicity increase for the membranes mentioned above ismainly due to the PNVP chains immobilized on the surface.
On the other hand, the contact angles measured by CBMshow a similar tendency, however, the definite values arelower than those by SDM. It is very interesting that the differ-ence of contact angles measured by SDM and CBM increasesg ◦ ◦
ig. 5. SEM images of PPMMs with different immobilization degrees of (a)wt.%; (b) 6.73 wt.%; and (c) 8.13 wt.%. Membrane samples were preparedt the same plasma treatment time (30 s) with different PNVP adsorptionegrees.
F f themt atmentt sure-m
radually from 10 for the nascent membrane to about 40for
ive bubble method. The samples were prepared at the same plasma treime (30 s) with different PNVP adsorption degrees. Prior to the meaents, the membranes were conditioned in a desiccator for 1 month.
a The PNVP-modified PPMMs were prepared at the same plasma treat-ment time (30 s) with different PNVP adsorption degrees.
the membranes immobilized with more than 10 wt.% PNVP.On reason for these results is that the static contact angle mea-sured by SDM is normally close to the dynamic advancingangle while that measured by CBM is similar to the dynamicreceding angle. Another possible explain of these results maybe that the polar parts of pyrrolidone rings interact with waterand this interaction contributes to the decrease of the inter-facial free energy of the membrane surface, even though thematerial itself has a high surface free energy. One shouldkeep in mind that the CBM measurements are performedin an aqueous environment allowing the polymers to absorbwater. This adsorption enhances with the increase of the im-mobilization degree of PNVP, which, in turn, increases thedifference of contact angles measured by SDM and CBM,respectively. The wet surface in CBM measurement cannottherefore be directly compared to the dry surface in SDMmeasurement[37].
The extent of platelet adhesion and the morphology ofthe adhered platelets are considered to be an early indi-cator of thrombogenicity of blood contacting biomaterials.Fig. 7 is the SEM micrographs of the studied membranescontacted with platelets for 30 min. The statistical numbersof platelets adhered on the membrane are listed inTable 4. Forthe nascent PPMM (Fig. 7(a)), a large amount of platelets areadhered on the surface with serious aggregation. Many de-f ervedo withP thef FromT n unitm ase ofP a-t branes ity ofP ofP theb be-h endss ility,h rgy,c eters
Fig. 7. Platelet adhesion on PPMMs with different immobilization degreesof (a) 0 wt.%; (b) 6.41 wt.%; and (c) 8.13 wt.%.
and mechanical characteristics[38]. For the PNVP-modifiedPPMMs, the surface becomes more hydrophilic and rougheras revealed by water contact angle measurement and AFM ob-servation (please see the supporting information). When themembrane surface contacts with platelets, the PNVP chainscan be hydrated, these hydrated chains on the surface haveexerted hydrodynamics and steric hindrance effects to the ap-proaching of platelets, which greatly reduces the adhesion ofplatelets.
3.4. Permeation and antifouling property of themembranes
The filtration results for the nascent PPMM and the PNVP-modified PPMMs with different PNVP immobilization de-gree are shown inFig. 8. Due to the relatively low porosity
ormed platelets, such as pseudopodia, can also be obsn the nascent membrane. After surface modificationNVP (Fig. 7(b) and (c)), the aggregation of platelets and
ormation of pseudopodium are suppressed gradually.able 4we can see that the amount of adhered platelets oembrane surface area decreases sharply with the increNVP immobilization degree. At higher PNVP immobiliz
ion degree, there is scarce platelet adhered on the memurface. These results indicate that the hemocompatibilPMM can be improved obviously by the immobilizationNVP, which may be ascribed to the hydrophilicity andiocompatibility of PNVP. It has been suggested that, theavior of platelet adhesion on polymeric materials deptrongly on the surface characteristics such as wettabydrophilicity/hydrophobicity balance, surface free enehemistry, charge density, roughness, micropore diam
Z.-M. Liu et al. / Journal of Membrane Science 249 (2005) 21–31 29
Fig. 8. Permeation fluxes (a) and flux changes (b) during filtration for thePNVP-modified PPMMs with the immobilization degrees of (1) 0 wt.%; (2)4.23 wt.%; (3) 7.44 wt.%; (4) 10.12 wt.%; (5) 14.18 wt.%.
(45–50% for the nascent PPMM) and hydrophilicity[29],the pure water flux for each PPMM studied is lower than100 mL m−2 h−1. However, it can be seen fromFig. 8(a) thatalmost all modified membranes experience both increases inthe pure water and protein solution fluxes. When filtratedwith a 1.0 g L−1 BSA solution, the membrane has an en-hancing tendency toward antifouling property as the immo-bilization degree of PNVP increases, which is shown by thedecrease of the total flux loss (1− J1/J0), from 69% for thenascent membrane to 34% for the modified membrane im-mobilized with 14.18 wt.% PNVP (Fig. 8(b)). Cleaning themembrane with de-ionized water and NaOH solution was in-tended to remove BSA molecules adsorbed on the membranesurface, the results of washing demonstrate the recovery abil-ity of the membrane fouled by proteins. From the columnsshown inFig. 8(b), it can be seen that the membranes modi-
Fig. 9. Permeation flux after static BSA adsorption (J3), static adsorptivefouling as well as the dynamic adsorptive fouling for the PNVP-modifiedPPMMs with immobilization degrees of (1) 0 wt.%; (2) 4.23 wt.%; (3)7.44 wt.%; (4) 10.12 wt.%; (5) 14.18 wt.%.
fied with PNVP are much more easily recovered with wash-ing, and the overall water flux recovery ((J2 − J1)/(J0 − J1))is larger than that of unmodified PPMM. For 10.12 wt.%PNVP-immobilized PPMM, about 69% of the pure waterflux loss was recovered after washing, indicating that a greatpart of BSA adsorptive fouling is reversible and the adhe-sion between the BSA and the PNVP-modified membranesis small. Nevertheless, for the PNVP-modified PPMMs, evencleaning the membrane with NaOH solution does not restorethe flux after BSA filtration. This might be reasonable thatPNVP chains are mainly immobilized on the membrane sur-face and BSA can be adsorbed in the pores during the filtrationand then block the pores. Furthermore, the hydrophilicity ofPNVP-modified PPMM is not strong enough to prevent theirreversible adsorption of proteins on the membrane surface.
The static adsorption of BSA and its effect on the per-meation property for the studied membranes were also mea-sured. The water fluxes for the membranes adsorbed withBSA (J3) as well as the static adsorptive fouling (1− J3/J0)are shown inFig. 9, the data of dynamic adsorptive fouling(1− J2/J0) are also listed here. Compared the static adsorp-tion of BSA with the dynamic ones, one can clearly see thedisparity of membrane properties performing at different con-ditions. For nascent PPMM, the dynamic adsorptive foulingis much higher than the static one. Upon the immobilizationo thed witht nceb ases.A mica ean-i form SAs
herp uced
f PNVP macromolecules on the PPMM surface, bothynamic and static adsorption of BSA decrease. While
he increase of PNVP immobilization degree, the differeetween dynamic protein fouling and static one decret high PNVP immobilization degree, the values of dynand static adsorptive fouling reach almost the same, m
ng there is nearly no difference in adsorptive foulingembranes permeating or just keeping in touch with B
olutions.In a summary, all PNVP-immobilized PPMMs have hig
ure water fluxes with enhanced flux recovery and red
30 Z.-M. Liu et al. / Journal of Membrane Science 249 (2005) 21–31
flux loss from BSA adsorptive fouling. Nevertheless, thefluxes for the PNVP-immobilized PPMM do not change inthe same order of immobilized degree. A 4.23 wt.% PNVP-modified PPMM has the highestJ0 andJ2, while J1 of the10.12 wt.% PNVP-modified PPMM is the highest among thefive membrane samples. From the SEM images we knowthat the immobilized PNVP chains may obstruct micropores,leading to the decrease in both pore size and porosity. Mean-while, the immobilization of PNVP macromolecules onto thePPMM surface produces a more hydrophilic surface. This in-dicates that at low immobilization degree the hydrophilicityhas a major effect on the permeability, while this may beovercompensated by pore blocking due to the immobilizedpolymer at high immobilization degree.
4. Conclusions
Polypropylene microfiltration membranes were modifiedby the immobilization of poly-(N-vinyl-2-pyrrolidone) on thesurface. The modification procedure differs from those pre-vious reports in that poly(N-vinyl-2-pyrrolidone) is boundon the membrane by plasma treatment instead of generat-ing from the corresponding monomer. FT-IR/ATR spectraand XPS results demonstrate the chemical changes occurringo anglem odi-fi ase ofi evealt r thei -t ad-s f them
A
nceF onalB d theH hina( ed.P othf e Ed-u ests
R
onse-
ul-ne
[3] G.B. Sigal, M. Mrksich, G.M. Whitesides, Effect of surface wetta-bility on the adsorption of proteins and detergents, J. Am. Chem.Soc. 120 (1998) 3464.
[4] F. Fang, I. Szleifer, Effect of molecular structure on the adsorptionof protein on surfaces with grafted polymers, Langmuir 18 (2002)5497.
[5] A. Sadana, Protein adsorption and inactivation on surfaces. Influenceof heterogeneities, Chem. Rev. 92 (1992) 1799.
[6] L.J. Zeman, A. Zydney, Microfiltration and Ultrafiltration: Principlesand Applications, Marcel Dekker, New York, 1996.
[7] D.R. Lloyd (Ed.), Material Science of Synthetic Membrane, ACSSymposium Series 26, American Chemical Society, Washington, DC,1985.
[8] H. Chen, G. Belfort, Surface modification of poly(ether sulfone)ultrafiltration membranes by low-temperature plasma-induced graftpolymerization, J. Appl. Polym. Sci. 72 (1999) 1699.
[9] P. Wang, K.L. Tan, E.T. Kang, K.G. Neoh, Plasma-induced immobi-lization of poly(ethylene glycol) onto poly(vinylidene fluoride) mi-croporous membrane, J. Membr. Sci. 195 (2002) 103.
[10] Q. He, Z. Liu, P. Xiao, R. Liang, N. He, Z. Lu, Preparation of hy-drophilic poly(dimethylsiloxane) stamps by plasma-induced grafting,Langmuir 19 (2003) 6982.
[11] Y.-J. Li, T. Hanada, T. Nakaya, Surface modification of segmentedpolyurethanes by grafting methacrylates and phosphatidylcholine po-lar headgroups to improve hemocompatibility, Chem. Mater. 11(1999) 763.
[12] D.S. Wavhal, E.R. Fisher, Hydrophilic modification of polyethersul-fone membranes by low temperature plasma-induced graft polymer-ization, J. Membr. Sci. 209 (2002) 255.
n the membrane surface. Results from water contacteasurements indicate that the hydrophilicity of the m
ed membranes can be enhanced actually with the incremmobilization degree. Platelets adhesion experiments rhat the membranes are obviously hemocompatible aftemmobilization of poly(N-vinyl-2-pyrrolidone) chains. Waer and BSA solution permeation as well as static BSAorption results demonstrate the antifouling property oembrane is improved.
cknowledgements
Financial supports from the National Nature Scieoundation of China (Grant no. 20074033), the Natiasic Research Program of China (2003CB15705), anigh-Tech Research and Development Program of C
Grant no. 2002AA601230) are gratefully acknowledgrof. Zhi-Kang Xu also thanks the financial supports b
rom Deutsche Forschungsgemeinschaft (DFG) and thcation Ministry of China for the visit of a Chinese gucientist to Germany.
eferences
[1] J.L. Nilsson, Protein fouling of UF membranes: causes and cquences, J. Membr. Sci. 52 (1990) 121.
[2] D. Mockel, E. Staude, M.D. Giver, Static protein adsorption,trafiltration behavior and cleanability of hydrophilized polysulfomembranes, J. Membr. Sci. 158 (1999) 63.
plasma-induced polymerization of acrylamide for improvedface properties and reduced protein fouling, Langmuir 19 (279.
14] M.L. Steen, A.C. Jordan, E.R. Fisher, Hydrophilic modificationpolymeric membranes by low temperature H2O plasma treatment,Membr. Sci. 204 (2002) 341.
15] M.L. Steen, M.L. Steen, L. Hymas, N.E. Capps, E.D. Havey, DCastner, E.R. Fisher, Low temperature plasma treatment of ametric polysulfone membranes for permanent hydrophilic sumodification, J. Membr. Sci. 188 (2001) 97.
16] S. Piletsky, H. Matuschewski, U. Schedler, A. Wilpert, E. Pska, T.A. Thirle, M. Ulbricht, Surface functionalization of poropolyprolylene membranes with molecularly imprinted polymersphotograft copolymerization in wate, Macromolecules 33 (23092.
17] J. Pieracci, J.V. Crivello, G. Belfort, UV-assisted graft polymerizaof N-vinyl-2-pyrrolidinone onto poly (ether sulfone) ultrafiltratimembranes using selective UV wavelengths, Chem. Mater. 14 (256.
18] D.S. Wavhal, E.R. Fisher, Hydrophilic modification of polyetherfone membranes by low temperature plasma-induced graft polization, J. Membr. Sci. 209 (2002) 255.
19] F. Cardona, G.A. George, Copolymers obtained by the radiainduced grafting of styrene onto poly(tetrafluoroethyleneperfluoropropylvinyl ether) substrates. 1. Preparation and struinvestigation, Macromolecules 35 (2002) 355.
20] N.I. Shtanko, V.Y. Kabanov, P.Y. Apel, M. Yoshida, A.I. VilenskPreparation of permeability-controlled track membranes on theof ‘smart’ polymers, J. Membr. Sci. 179 (2000) 155.
21] M. Nakamura, S. Kiyohara, K. Saito, K. Sugita, T. Sugo, Higholution of dl-tryptophan at high flow rates using a bovine sealbumin-multilayered porous hollow-fiber membrane, Anal. Ch71 (1999) 1323.
22] D. Meinhold, R. Schweiss, S. Zschoche, A. Janke, A. Baier, Fmon, H. Dorschner, C. Werner, Hydrogel characteristics of elecbeam-immobilized poly (vinyl pyrrolidone) films on poly(ethyleterephthalate) supports, Langmuir 20 (2004) 396.
[25] S. Robinson, P.A. Williams, Inhibition of protein adsorption ontosilica by polyvinylpyrrolidone, Langmuir 18 (2002) 8743.
[26] A. Higuchi, K. Shirano, M. Harashima, B.O. Yoon, M. Hara, M.Hattori, K. Imamura, Chemically modified polysulfone hollow fiberswith vinylpyrrolidone having improved blood compatibility, Bioma-terials 23 (2002) 2659.
[27] J.S. Kang, K.Y. Kim, Y.M. Lee, Preparation of PVP immobilizedmicroporous chlorinated polyvinyl chloride membranes on fabric andtheir hydraulic permeation behavior, J. Membr. Sci. 214 (2003) 311.
[28] Z.-M. Liu, Z.-K. Xu, Q. Yang, J.-Q. Wang, J. Wu, J.-J. Fu, Surfacemodification of polypropylene microfiltration membranes by graftpolymerization ofN-vinyl-2-pyrrolidone, Eur. Polym. J. 40 (2004)2077.
[29] Z.-K. Xu, J.-L. Wang, L.-Q. Shen, D.-F. Men, Y.-Y. Xu, Microporouspolypropylene hollow fiber membrane. Part I. Surface modificationby the graft polymerization of acrylic acid, J. Membr. Sci. 196 (2002)221.
[30] R.-Q. Kou, Z.-K. Xu, H.-T. Deng, Z.-M. Liu, P. Seta, Surface modifi-cation of microporous polypropylene membranes by plasma-inducedgraft polymerization of �-allyl glucoside, Langmuir 19 (2003)6869.
[31] Z.-K. Xu, Q.-W. Dai, J. Wu, X.-J. Huang, Q. Yang, Covalent attach-ment of phospholipid analogous polymers to modify a polymericmembrane surface: a novel approach, Langmuir 20 (2004) 1481.
[32] J.G.A. Terlingen, J. Feijen, A.S. Hoffman, Immobilization of sur-face active compounds on polymer supports using glow dischargeprocesses: 1. Sodium dodecyl sulfate on poly(propylene), J. ColloidInterface Sci. 155 (1993) 55.
[33] H.D.W. Roesink, M.A.M. Beerlage, W. Potman, Th. Van den Boom-gaard, M.H.V. Mulder, C.A. Smolders, Characterization of new mem-brane materials by means of fouling experiments adsorption of BSAon polyetherimide-polyvinylpyrrolidone membranes, Colloids Surf.55 (1991) 231.
[34] M. Hudis, Techniques and Applications of Plasma Chemistry, Wi-ley/Interscience, New York, 1974.
[35] A.R. Roudman, F.A. DiGiano, Surface energy of experimental andcommercial nanofiltration membranes: effects of wetting and naturalorganic matter fouling, J. Membr. Sci. 175 (2000) 61.
[36] W. Chen, A.Y. Fadeev, M.C. Hsieh, D.Oner, J. Youngblood, T.J.McCarthy, Ultrahydrophobic and ultralyophobic surfaces: some com-ments and examples, Langmuir 15 (1999) 3395.
[37] B.C. Menno, M.B. Olde Riekerink, J.D. de Bruijn, D.W. Grijpma,G.H.M. Engbers, J. Feijen, Enhanced bone marrow stromal celladhesion and growth on segmented poly(ether ester)s based onpoly(ethylene oxide) and poly(butylenes terephthalate), Biomacro-molecules 4 (2003) 57.
[38] J.B. Park, Biomaterials Science and Engineering, Plenum Press, NewYork, 1984.
