- Email: [email protected]
PPy and the membrane antifouling property in filtrating yeast suspensions in EMBR
Journal of Membrane Science 437 (2013) 99–107
Contents lists available at SciVerse ScienceDirect
Journal of Membrane Science journal homepage: www.elsevier.com/locate/memsci
Preparation of highly conductive cathodic membrane with graphene (oxide)/PPy and the membrane antifouling property in filtrating yeast suspensions in EMBR Lifen Liu n, Feng Zhao, Jiadong Liu, Fenglin Yang MOE, Key lab of industrial Ecology and Environmental Engineering, School of Environmental Science and Technology, Dalian University of Technology, 116024 Dalian, China
a r t i c l e i n f o
abstract
Article history: Received 27 July 2012 Received in revised form 14 November 2012 Accepted 21 February 2013 Available online 4 March 2013
The EMBR (MBR, attached with electric field) is effective in suppressing membrane fouling. In this paper, modified polyester cathode membranes were prepared by coating graphene (Gr)/polypyrrole (PPy) or graphene oxide (GO)/PPy, which had higher conductivity and better antifouling property in EMBR than the membrane modified with PPy only. The membrane modification using a vapor phase polymerization method (VPM) was proved better than using a liquid phase polymerization method (LPM). The modified membranes were tested and compared for antifouling performances in EMBR, by short time filtration of a yeast suspension, under 1 V/cm electric field. The VPM Gr/PPy modified membrane suppressed membrane fouling and increased the flux as high as 20%. The VPM Gr/PPy coating was more uniform and firm, and had higher conductivity (resistance lowered by 67%) than PPy modified membrane. It had better antifouling performance, smaller slopes in its t/V–V curves (adopting the classical cake filtration model, a suitable description for the membrane fouling process). By applying 1 V/cm electric field, an increase of 20% in the short-term cumulative permeate volume was obtained for the Gr/PPy modified membrane and it was only 10% for PPy modified membrane. & 2013 Elsevier B.V. All rights reserved.
Keywords: Conductive polymer composites Cathodic membrane Membrane fouling suppression EMBR Graphene
1. Introduction Membrane bioreactors (MBRs) have a lot of advantages over other conventional methods for wastewater treatment [1]. It has become more and more popular for its advantages in higher quality effluent, smaller space occupation, higher degree of automation and lower sludge production. However, the decline of flux caused by membrane fouling is a big problem in wastewater treatment since it affects the performance significantly [2,3]. To efficiently improve the membrane anti-fouling performance and control membrane fouling in MBR, it has been proved effective by applying an electric field in MBR (the so-called EMBR) [4,5], in which negatively charged sludge and extra-cellular polymeric substances (EPS) are forced to move away from the membrane surface under applied voltage (Fig. 1). Also, the conductive layer on the membrane surface, especially around the pore entrance, may greatly suppress fouling in EMBR. The influence of the electric field and the conductivity of the cathode membrane on membrane flux are considered critical for membrane fouling suppression performance in EMBR. In general, membrane flux increases when the electric field strength
n
Corresponding author. Tel.: þ86 411 84706173; fax: þ 86 411 84708083. E-mail address: [email protected] (L. Liu).
0376-7388/$ - see front matter & 2013 Elsevier B.V. All rights reserved. http://dx.doi.org/10.1016/j.memsci.2013.02.045
increases within certain range, then beyond certain points, it is no longer economic, after reaching certain value (the critical potential) the flux remained almost constant [6]. Conducting at or above the critical potential can avoid the formation of filtration cake layer, but the high voltage requires highly stable electrodes and implies high energy consumption/cost. To reduce energy consumption, an intermittent electric field, applied only when the permeate flux has drastically declined, can also effectively suppress the membrane fouling [7]. Also, applying low voltage to the built-in cathode in the flat panel membrane module and better arranged anodes in configured MBR, can effectively utilize electrophoresis force in inhibiting bio-fouling of membrane [8]. Moreover, using the conductive membrane as cathode combines the membrane and electrode as one, so higher efficiency in material usage and better anti-fouling performance can be achieved. To make the conductive cathode membrane, depositing conductive polymers on the surface of membrane directly is effective and low cost, than arranging metal cathode in membrane modules. Polypyrrole (PPy) is a polymer with high conductivity, stable in MBR and safe for humans, and it has been widely used in electro-chemistry [9,10]. In our lab, conductive membrane made of PPy based on filter cloth have been successfully prepared [11] and showed really good property. However, the polymer conductivity on polyester cloth is not as good as conventional metal cathode (iron, aluminum, copper). And higher
100
L. Liu et al. / Journal of Membrane Science 437 (2013) 99–107
Negatively charged fouling particle
More Conductive layer Non conductive layer
Electrostatic rejection force Better fouling suppression
Less conductive layer
less fouling suppression
Cathode Membrane
Difference in membrane pore structure, conductivity and surface property affects fouling suppression performance
Fig. 1. The mechanism for membrane-fouling suppression in EMBR.
Fig. 2. (A) VPM process for GO(Gr)/PPy membrane (PPy membrane without the process in green dash line). (B) LPM process for GO(Gr)/PPy membrane (PPy membrane without the process in orange dash line). (For interpretation of the references to color in this figure caption, the reader is referred to the web version of this article.)
resistance means high energy consumption to produce the same effect in fouling reduction. Preparing high conductivity membrane from large pored, low price polyester filter cloth is critical for our research. Graphene and its oxides are cheap carbon materials with unique electronic properties [12]. Their composites with PPy have better conductivity than the pure PPy, widely used as supercapacitor, electrodes for microbial fuel cell, electrochemical oxygen reduction and pollutant removal [13–15]. In this paper, we prepared highly electronic-conductive membrane by modifying polyester filter cloth with Gr/PPy or GO/PPy via the VPM and/or LPM. The better conductive membrane was tested as cathode
membrane for anti-fouling performance in filtrating yeast suspension with 1 V in EMBR. Also for comparing membrane fouling performance, fouling models suitable for describing the flux decrease in our experimental system were studied and compared.
2. Experiment 2.1. Preparation and reduction of graphene oxide The graphene oxide (GO) was prepared using the Hummer’s method [16,17]. First, graphite (2.0 g) was mixed with 96 mL
L. Liu et al. / Journal of Membrane Science 437 (2013) 99–107
H2SO4 (98%) and NaNO3 (2 g) by stirring for 30 min in a beaker in an ice bath. Potassium permanganate (KMnO4, 12.0 g) was added very slowly in the suspension with vigorous stirring for 3.5 h while maintaining reaction temperature below 20 1C. Then the ice bath was removed, the reaction mixture was stirred at 35 1C for another 2 h. Next, 184 mL water was added to the pasty solution with constant agitation, and temperature was raised to 98 1C. There upon, the color of the solution changed to yellowish brown. After 2 h of vigorous stirring, 20 mL of 30% H2O2 was added to remove the excess KMnO4. The mixture was washed several times with 5% HCl and deionized (DI) water until the solution was free of sulfate and acid. Then ultrasonic treatment of the mixture was carried out until no more solid could be centrifuged. The concentration of GO solution was obtained by setting the volume of solution and weighing the formed GO, thus the stock GO solution was prepared, it is then diluted to different concentrations and used for immersing the membrane. Graphene was obtained from reducing its oxide by using hydrazine mono-hydrate, a strong reducing agent. The standard GO solution was diluted to form different concentration solution with DI water and the pH value was adjusted to 10 [18]. Then, hydrazine mono-hydrate (ratio of hydrazine: GO was 7:10 by weight) was added to the mixture and ultrasonically treated for 8 h. After the completion of the reaction, the reduced graphene oxide (Gr) was collected by filtration and dried. The black powder product thus obtained was washed with DI water several times to remove excess hydrazine, and the final product was dried in a vacuum oven at 60 1C for 24 h.
2.2. Vapor phase polymerization method and membrane modification The VPM membrane modifying process was shown in Fig. 2(A) [11]. First, blank filter cloth (39.4672.50 mm for the average and 66.51710.76 mm for the maximal pore size, 8 11 cm2) was immersed and ultrasonically treated for 10 min in GO(0.09 mg/mL) or Gr(0.02 mg/mL) solution. The filter cloth was taken out to dry in the air for 30 min, sufficient APS (ammonium persulfate) solution (50 mL, 20 mg/mL) was sprayed onto it. 0.5 mL pyrrole was first added to the bottom of a 1000 mL beaker, and the unmodified membrane was hanged in it above the pyrrole liquid without any contact with the liquid. The beaker was sealed with plastic film and it is filled with pyrrole (Py) vapor when its bottom was heated by 90 1C water bath. About 15 min, pyrrole molecules were completely adsorbed from the vapor phase onto the membrane and polymerized with GO (Gr) in situ to PPy with APS as initiator. The white filter cloth turned dark and was successfully modified by PPy and Gr/GO. Finally the membrane was rinsed with tap water.
2.3. Liquid phase polymerization method and membrane modification The LPM membrane modifying process is shown in Fig. 2(B). First, blank filter cloth (8 11 cm2) was immersed in APS solution (20 mg/mL, 50 mL) for 10 min and drip-dried in the air for 30 min. Dissolve 1 mL pyrrole in ethanol and water solution, then ultrasonically mix it with certain concentration graphene (or its oxide) solutions to get the GO/PPy and Gr/PPy composites. Then dip the filter cloth into the composites solution. Very soon a large quantity of composite GO/PPy (Gr/PPy) was formed by polymerization on the surface of blank filter cloth and made the membrane really dark. After 5 min, the membrane was taken out, hanged in air and rinsed with water.
101
2.4. Membrane characterization FT-IR spectrometer (IR Prestige-21, Shimadzu, Japan) was used to investigate the functional groups of graphite, graphene oxide (GO) and reduced graphene oxide (RGO). Membrane samples were viewed by digital camera (canon ixus 105) and the surface microcosmic morphologies of the prepared membranes were visualized by scanning electron microcosmic (SEM, JEOL JSM5600LV). Pore size distributions were measured with the Bubble Point Method according to the China national standard, GB/T 24219–2009 (determination of aperture of woven filtering fabric). The electric resistance of the membrane was measured by digital multimeter (Victor VC 830L) and averaged from values at six different locations. The quantities polymerized on the base membrane were measured using four 8n11 cm2 samples by calculating the weight increase of the membrane after modification. The stability of VPM and LPM was studied by observing their flux changes in different filtration cycles. 2.5. Filtration and the anti-fouling properties of the membrane The experimental set-up was schematically shown in Fig. 3. The flat sheet membrane module has an effective permeate area of 80 cm2. The separation unit has a volume of 4 L and the aeration rate was kept at 0.2 m3/h. According to the experiment above, the modified membrane with better conductivity was obtained for both VPM and LPM. They were chosen for anti-fouling property test in yeast solution (Zeta potential¼ 12.9 mV, size ¼7 mm), which was to simulate the condition in MBR. The results were compared with unmodified and only PPy modified membrane, with and without electric field. Membrane module was set and fixed in MBR container (MLSS¼ 5 g/L). When electric field was applied, the voltage was supplied from a steady DC power source. The anodes were made of two pieces of stainless steel meshes which were set 1 cm from the cathode at either side of the membrane module. The vertical distance between the liquid level and the bottom of effluent tube was fixed to 102 cm in every test, ensuring a corresponding transmembrane pressure at 10.00 kPa, so the gravity flow always has the same water head drop. The test was conducted in three cycles and each lasted for 1.5 h, between the intervals of cycles, the membrane was washed with tap water. Each three-cycle filtration test was repeated for three times. Each test uses one new piece of membrane of the same batch/kind.
3. Results and discussion The doping ratio of GO (Gr) and PPy had an effect on the property of membrane. To a certain degree, higher conductivity would be obtained if more Gr (larger than the ratio mentioned above in this paper) could strongly and uniformly fixed with PPy on the base membrane, just as other papers reported [19,20]. However, in our experiment, too much GO or Gr was against uniform fixation of PPy on the base membrane, the PPy could not coat the graphene (or its oxide) totally and the polymerized composites were on the surface of the membrane instead of among the woven fibers, which is a big disadvantage for the conductivity improvements. So the low concentration (GO¼0.09 mg/mL, Gr¼0.02 mg/mL) was chosen to treat the base membrane and better comprehensive property was obtained. 3.1. FT-IR of the composite membrane The Fourier-transform infrared (FTIR) spectroscopic study analyzed the functional groups on GO, compared with natural
102
L. Liu et al. / Journal of Membrane Science 437 (2013) 99–107
Fig. 3. The set-up of yeast filtration experiment (1, feed tank; 2, balance tank; 3, EMBR; 4, membrane module; 5, stainless iron meshes anode; 6, air pump; 7, air manometer; 8, aerating apparatus; 9, DC electrical source; 10, outflow tube).
1562 PPy Transmittance(%)
Gra phene oxide 2976 1047 Reduced GO 3383
3500
2927
1739
1636
Transmittance(%)
1477
Natural graphite
1046
1401
1681 1286
1216 927
GO/PPy
Gr/PPy
1400
3000 2500 2000 1500 Wavenumbers (cm-1)
961
1000
2000
1800
1600
1400
1200
1000
800
Wavenumbers (cm-1)
Fig. 4. The FT-IR of natural graphite, GO and Gr (A) and the FT-IR of PPy, GO/PPy and Gr/PPy (B).
graphite and the reduced GO. Upon oxidation of graphite to GO, the observed representative peaks of GO confirm the presence of the oxygen-containing functional moieties in carbon frameworks, which include the bands at 3383 cm 1(O–H stretching), 1739 cm 1(C¼O stretching vibration of carboxyl), 1636 cm 1 (water H–O–H bending vibration), 1400 cm 1(O–H deformation vibration ) and 1047 cm 1(C–O vibration). The curves of graphite and Gr are much smoother than GO, the typical peak is around 1400 cm 1, confirming that by Hummer’s method, natural graphite is successfully transformed to GO, and reduction of GO did take place by chemical reaction with hydrazine (Fig. 4(A)). Fig.4(B) shows the bands of typical pyrrole ring stretching at 1401, 1477 and 1562 cm 1, identifying the presence of the conductive polymer PPy. For PPy, the peaks at 1681 cm 1 and 1562 cm 1 are assigned to the carbonyl group and the C ¼C/C–C
ring stretching of pyrrole. The bands at 1477 cm 1 and 1286 cm 1 are corresponding to the stretching vibration in the ring of C–N bond and C–H vibrations. The bands at 1046 cm 1 are caused by N–H bending vibration. In-plane and out-plane ¼C–H bending vibration is located at 961 cm 1 and 927 cm 1. This agrees well with other reports in literature [21,22] for PPy. The curves of GO/PPy and Gr/PPy are similar to the PPy, with some differences. PPy, GO and Gr curves all have an obvious peak at 1401 cm 1 for O–H deformation vibration, then in the composite membrane of GO/PPy and Gr/PPy, this representative peaks intensified (weakened) in strength by GO (Gr) doping. For the same at peak 1047 cm 1, where the transmittance peak strength for Gr/PPy weakened than pure PPy, whereas the GO/PPy peak intensified because of the existence of GO whose absorbance peak is strong here.
L. Liu et al. / Journal of Membrane Science 437 (2013) 99–107
3.2. Macroscopic view and the morphology of the conductive membrane surface 3.2.1. Macroscopic view of the membrane Membranes modified by VPM were shown in Fig. S1. The color was uniform and very dark compared with the white blank membrane. All the modified membrane had firm interfacial bonding between the composite coating and base membrane, and membrane colors were slightly different as different coating components were used. The color of the membrane prepared by LPM (Fig. S2) seemed uniform from macroscopic view. But only when more reagents (about two times Py quantity of VPM) were used, could the filter cloth be totally modified by conductive polymer and the membrane seemed much darker than the VPM membrane, in which less PPy composite was coated.
3.2.2. The morphology of the conductive membrane surface Micro-scale morphology of the membrane was shown by SEM (Fig. 5) and it further demonstrated that all the VPM was successful in modifying the membranes (images B for PPy, C for GO/PPy, D for Gr/PPy), the modification layer was more uniform than in the LPM modified membranes (image E). For the VPM modified GO/PPy or Gr/PPy membrane, the surface of the coated fibers had less wrinkles or ripples/ridges than in image B (PPy only), was generally more even and uniform, though typical tiny nodes of PPy can be seen. And the tiny sheet structures from PPy
103
coated GO (image C2) or Gr (image D2) can be clearly seen in the SEM images. 3.3. The pore size of the modified membrane In Fig. 6(A), the relative ratios of pore size of the VPM modified membrane to the average pore size of blank membrane were shown, the average pore size of the modified membrane did not shrink or narrow much than the blank membrane. Although, from the SEM images, we can see the thin layers of composite coating on the fiber, the pore size measured by bubble point method did not show much change because the pore size was much bigger and wider than the thin coating layer. As pyrrole in liquid phase tends to polymerize on the coarse surface of the base membrane instead of around the fibers, the modified membrane by LPM was not uniform and pore blocking was serious. Membrane prepared in this way had much smaller average pore size than the blank membrane (Fig. 6(B)). 3.4. Conductivity The conductivity comparison of the modified membranes prepared by VPM and LPM methods was listed in Table 1. For the membrane prepared by VPM, the electric resistance for the modified membrane can be as low as 710 O cm 1 (modified by GO/PPy) and 680 O cm 1 (modified by Gr/PPy), much lower than 2.03 k O cm 1 for the only PPy modified filter cloth. And the
Fig. 5. The SEM of the membranes (A, blank membrane; B, PPy membrane by VPM; C, GO/PPy membrane by VPM; D, Gr/PPy membrane by VPM; E, Gr/PPy membrane by LPM).
80
80
Max Pore Size Average Pore Size
70
60 Pore size (μm)
60 Pore size (μm)
Max Pore Size Average Pore Size
70
50 40 30
50 40 30
20
20
10
10
0
0 Blank
PPy-M.
GO/PPy-M.
Membrane by VPM
Gr/PPy-M.
Blank
PPy-M.
GO/PPy-M.
Gr/PPy-M.
Membrane by LPM
Fig. 6. The average and maximal pore size for VPM membrane (A) and the average and maximal pore size for LPM membrane (B).
104
L. Liu et al. / Journal of Membrane Science 437 (2013) 99–107
membrane prepared by LPM had a resistance of 14.81 k O cm 1 for GO/PPy modified membrane and 2.7 k O cm 1 (modified by Gr/PPy), compared with 8.15 k O cm 1 (modified only by PPy). The low conductivity of LPM membrane may be caused by nonuniform polymerization of Py on the membrane surface. Just as we have observed from the SEM (Fig. 5), the PPy is lumpish and polymerized on the surface instead of among the fibers, which contributes to the poor conductive performance of LPM membrane. For both VPM and/or LPM, the Gr/PPy modified membrane had the lowest electric resistance. Moreover, the VPM membrane
Table 1 The resistance of membrane modified with PPy, GO/PPy and Gr/PPy by different polymerization method. The modification method
Sample Blank membrane
Resistance (kO/cm) þN
Quality (mg/cm2) –
VPM
PPy-M GO/PPy-M Gr/PPy-M PPy-M GO/PPy-M Gr/PPy-M
2.03 7 0.61 0.71 7 0.10 0.68 7 0.08 8.15 7 0.79 13.21 7 0.59 2.70 7 1.05
1.70 1.59 1.70 1.07 1.74 2.08
LPM
Blank membrane Gr/PPy-membrane
500
tends to have smaller deviation in the conductivity than LPM membrane when the same dopant was used. 3.5. Fouling reduction of VPM and LPM modified membrane The membrane prepared by LPM did not show any regularity in flux changes in the three cycles yeast filtration test (Fig. 7). When no electric field was applied, the flux of modified membrane was much smaller than the blank membrane in the first filtration cycle. This is caused by the macroscopic lumpish PPy and Gr/PPy coatings, blocking the space between membrane fibers, resulted in a much smaller pore size. Then in the second and third cycles, as the membrane wash may remove/wipe off some Gr/PPy, the flux for the modified membrane increased to approach the level of blank membrane. Applying 1 V electric field to the LPM Gr/PPy modified membrane, enabled an increase of flux than without applied voltage, in the first cycle. But in the second and third cycles, it did not show any improvement as the conductive materials may partly fall off during the membrane wash, and resulted in the decrease of conductivity. During the yeast filtration test without electric field applied, flux of all the modified membrane by VPM did not show obvious difference with blank membrane in the three cycles (Fig. S3)
PPy-membrane Gr/PPy-membrane with E=1V
Flux(L/m2h)
450 175 150 125 100 75 50 0
25
50
75
100 125 150 175 Permeate time (min)
200
225
250
275
Fig. 7. Permeate flux of blank, PPy, Gr/PPy LPM membrane with or without electric field applying in three filtration cycles (MLSS ¼ 5000 mg/L, TMP¼ 10.0 kPa).
PPy-M.
PPy-M. with E=1V
Gr/PPy-M. with E=1V
Flux(L/m2h)
500
180 160 140 120 100 80 0
25
50
75
100 125 150 175 Permeate time (min)
200
225
250
275
Fig. 8. Permeate flux of PPy, Gr/PPy VPM membrane with or without applying electric field in three filtration cycles (MLSS ¼ 5000 mg/L, TMP¼ 10.0 kPa).
L. Liu et al. / Journal of Membrane Science 437 (2013) 99–107
because of the better hydrophilic property of PPy [11]. However, as the modified membrane has smaller pore size than the blank membrane, it shows better turbidity removal property (Fig. S4), which stands for the membrane rejection property. According to the figure, more than 90% turbidity was removed in the first 10 min for the blank membrane, and that turbidity removal for the modified membrane was 100%, which was maintained to the end of the first filtration cycle. Further research was conducted to exhibit the conductive membrane property. The anti-fouling property of PPy and Gr/PPy modified membrane was shown in Fig. 8, when 1 V electric field is applied, comparing with the PPy modified membrane without electric field. According to the results, modified membrane had an increase in flux and the biggest improvement was obtained for membrane with better conductivity, the Gr/PPy modified membrane. The fouling on VPM modified membrane was reversible, since the flux did not present any obvious change
Table 2 The increasing percentage of cumulative volume for conductive membrane with E compared with PPy-M without applying E. Membrane sample tested
1st cycle
2nd cycle
3rd cycle
PPy-M with E Gr/PPy-M with E
8.9% 16.9%
10.8% 21.3%
10.8% 20.3%
in the three cycles after membrane washing with tap water, indicating a good performance. Applying electric field, increased total permeate volume from each cycle and the fluxes of all the modified membranes (Fig. S5). Comparing with the PPy modified membrane without applied electric field, the total permeate volume increase by applying electric field was 10% for PPy and the increase for Gr/PPy modified membrane was 20% (Table 2). 3.6. The membrane fouling mechanism In order to investigate the membrane fouling process, we studied two models according to the previous research [23], the standard filtration model and classical filtration model. First, the standard filtration model, suits to the situation that the particle size is much smaller than the pore size of the primary membrane and the membrane fouling is caused mainly by the internal pore clogging. t 1 k ¼ þ t V Q0 2 where t is the filtration time (min), V is the cumulative permeate volume for corresponding filtration time (L), Q0 is the initial flux rate (L/min) and k is the filtration constant (L 1). When the membrane pore size was initially narrowed down by the particles to be filtrated, the deposit particles started to form a cake layer on membrane surface and decreased the flux.
60
60 PPy-M. PPy-M. E=1v Gr/PPy-M.E=1v
50
50
PPy-M. PPy-M. E=1v Gr/PPy-M. E=1v
40 t/V (min/L)
40 t/V (min/L)
105
30 20
y=1.134x+a1 R2=0.9383
30 20 y=32.86x+a1 R2=0.9950
2
y=1.015x+a2 R =0.9455
10
10
y=28.97x+a2 R2=0.9971
y=0.994x+a3 R2=0.9344 0
y=23.39x+a3 R2=0.9943
0 0
20
40 60 Permeate time(min)
80
100
0.5
1.0 1.5 Cumulative volume(L)
2.0
Fig. 9. The t/V–t (A) curves (at the filtration beginning stage) and t/V–V (B) curves (all from second points started) for PPy membrane without electric field, PPy membrane and/or Gr/PPy membrane with 1 V/cm electric field (R2 means linear correlation coefficient).
PPy-M.
PPy-M. with E=1V
Linear Fit of PPy-M.
Gr/PPy-M. with E=1V Linear Fit of PPy-M. E=1v
Linear Fit of Gr/PPy-M. E=1v 50
t/V(min/L)
40
1st cycle
30 20 2nd cycle
10
3rd cycle
0 0.0
0.5
1.0
1.5
2.0
2.5 3.0 3.5 4.0 Permeate volume (L)
4.5
5.0
5.5
6.0
Fig. 10. The three filtration cycles t/V–V curves for PPy membrane with or without electric field, and Gr/PPy VPM membrane with electric field (all starts from the second points in each cycle and when a new cycle start, set t as 0 min).
106
L. Liu et al. / Journal of Membrane Science 437 (2013) 99–107
Then classical cake filtration model represents the decrease in permeate volume t mrm mcRc 1 K ¼ þ þ V UV ¼ V Q0 2 AUTMP 2A2 TMP
m is the dynamic viscosity (Pa s), rm is the membrane resistance (m 1), and A is the membrane area (m2). TMP is the transmembrane pressure (Pa). c is the concentration of particles (kg/L), Rc means the resistance of the cake layer (m/kg), and K is the cake filtration constant. For the models, the smaller the filtration constant (K) and the cake filtration constant (K), the lower filtration resistance and the higher the mass transfer efficiency is. This is what we only considered to study the membrane fouling in our experiment. According to the models, t/V–V and t/V–t curves were drawn and analyzed. Comparing Fig. 9(A) with (B), we know the fouling mechanism for our filtration experiment fits the classical model better on the whole, confirmed by the better coincidence with
Table 3 The slopes and linear correlation coefficients of the curves correspond to Fig. 10 (R2 means linear correlation coefficient). Filtration cycle
PPy-M R
1st cycle 2nd cycle 3rd cycle
2
0.9950 0.9951 0.9967
PPy-M, E¼ 1 V 2
Gr/PPy-M, E ¼1 V
Slope
R
Slope
R2
Slope
32.86 36.99 36.45
0.9971 0.9952 0.9939
28.97 30.48 30.55
0.9943 0.9904 0.9932
23.39 24.01 24.89
linear relation of t/V–V curves than the t/V–t curves. This means the internal pore clogging quickly occurred, then the follow-up flux decrease was mainly caused by the filtration cake layer and the modified membrane with applied electric field could effectively suppress the depositing of particles on the surface of membrane. Moreover, at the initial filtration period, slopes of t/V–t curves for both conductive membrane with applied voltage did not show significant difference, indicating that the low electric field has little effect on improving the membrane flux at the beginning stage of test. However, for the t/V–V curves, Gr/PPy modified membrane with electric field has the smallest slope, second is the PPy with electric field and then the PPy-membrane without applied electric field, which is in accordance with their resistances, indicating membrane with better conductivity has better anti-fouling property. When the tests were finished after three cycles (Fig. 10 and Table 3), the advantage of better conductive membrane in fouling suppression was shown well kept, during the tests. This observation justifies the explanation proposed in Fig. 1(B). PPy has cathodic catalytic oxygen reduction reactivity (ORR), doped by AQS (anthraquinone monosulfate), this ORR property may be increased significantly [22], thus, the fouling reduction/ suppression mechanism of PPy modified membrane cathodes in EMBR can be extended, the results are also expected to increase under applied electric field. This is being studied in our lab, and the results will be reported elsewhere. Also we have supplied the SEM of the membranes (Fig. 11) after being used in E-MBR. Comparison of the images showed the stable membrane coating for both the Gr/PPy and PPy membranes, and confirmed the better fouling reduction property for Gr/PPy membrane than PPy membrane.
Fig. 11. The SEM of the membranes (A, C: PPy membrane by VPM; B, D: Gr/PPy membrane by VPM) after applying in E-MBR.
L. Liu et al. / Journal of Membrane Science 437 (2013) 99–107
4. Conclusion The following conclusion can be drawn from this study: The Gr/PPy modified membrane prepared by VPM had better conductivity than that prepared by LPM. Both GO and Gr doping of PPy during modification improved the conductivity of the modified membranes. The Gr/PPy modified membrane had better property than the GO/PPy modified membrane when the Gr or GO doping concentration was at the same level, or with much lower concentration of Gr than GO. For the Gr/PPy membrane, its electric resistance was 1/3 of the PPy modified membrane. With 1 V/cm electric field applied, Gr/PPy modified membrane can suppress the fouling more effectively than PPy modified membrane, and the cumulative permeate increase after 1.5 h filtration was 20% while that for PPy modified membrane was only 10%.
Acknowledgment Authors acknowledge the financial support from China National Natural Science Foundation (Project No. 21177018).
Appendix A. Supporting information Supplementary data associated with this article can be found in the online version at http://dx.doi.org/10.1016/j.memsci.2013. 02.045.
Nomenclature A c k K Q0 Rc rm t TMP V
the membrane area (m2) the concentration of particles (kg/L) the filtration constant (L 1) the cake filtration constant the initial flux rate (L/min) the resistance of the cake layer (m/kg) the membrane resistance (m 1) the filtration time (min) the trans-membrane pressure (Pa) the cumulative permeate volume for corresponding filtration time (L)
References [1] F.G. Meng, S. Chae, A. Drews, M. Kraume, H. Shin, F.L. Yang, Recent advances in membrane bioreactors (MBRs): membrane fouling and membrane material, Water Res. 43 (2009) 1489–1512.
107
[2] P. Le-Clech, V. Chen, T.A.G. Fane, Fouling in membrane bioreactors used in wastewater treatment, J. Membr. Sci. 284 (2006) 17–53. [3] A. Drews, Membrane fouling in membrane bioreactors—characterisation, contradictions, cause and cures, J. Membr. Sci. 363 (2010) 1–28. [4] K. Bani-melhem, M. Elektorowicz, Development of a novel submerged membrane electro-bioreactor (SMEBR): performance for fouling reduction, Environ. Sci. Technol. 44 (9) (2010) 3298–3304. [5] K. Akamatsu, Y. Yoshida, T. Suzaki, Y. Sakai, H. Nagamoto, S. Nakao, Development of a membrane–carbon cloth assembly for submerged membrane bioreactors to apply an intermittent electric field for fouling suppression, Sep. Purif. Technol. 88 (2012) 202–207. [6] J.P. Chen, C.Z. Yang, J.H. Zhou, X.Y. Wang, Study of the influence of the electric field on membrane flux of a new type of membrane bioreactor, Chem. Eng. J. 128 (2007) 177–180. [7] K. Akamatsu, W. Lu, T. Sugawara, S. Nakao, Development of a novel fouling suppression system in membrane bioreactors using an intermittent electric field, Water Res. 44 (2010) 825–830. [8] L.F. Liu, J.D. Liu, B. Gao, F.L. Yang, Minute electric field reduced membrane fouling and improved performance of membrane bioreactor, Sep. Purif. Technol. 86 (2012) 106–112. [9] G. Kang, R.B. Borgens, Y. Cho, Well-ordered porous conductive polypyrrole as a new platform for neural interfaces, Langmuir 27 (2011) 6179–6184. [10] A. Laith, D. Catherine, B. Stephan, W. Wojtek, Electropolymerized polypyrrole nanowires for hydrogen gas sensing, J. Phys. Chem. C 116 (2012) 13388–13394. [11] L. Liu, J. Liu, B. Gao, F. Yang, S. Chellam, Fouling reductions in a membrane bioreactor using an intermittent electric field and cathodic membrane modified by vapor phase polymerized pyrrole, J. Membr. Sci. 394–395 (2012) 202–208. [12] J.A. Matthew, C.T. Vincent, B.K. Rechard, Honeycomb carbon: a review of graphene, Chem. Rev. 110 (2010) 132–145. [13] P. Si, S. Ding, X.W. (David) Lou, D.H. Kim, An electrochemically formed threedimensional structure of polypyrrole/graphene nanoplatelets for highperformance supercapacitors, RSC Adv. 1 (2011) 1271–1278. [14] S. Biswas, L.T. Drzal, Multilayered nanoarchitecture of graphene nanosheets and polypyrrole nanowires for high performance supercapacitor electrodes, Chem. Mater. 22 (2010) 5667–5671. [15] S. Bose, T. Kuila, E. Uddin, N.H. Kimb, A.K.T. Lau, J.H. Lee, In-situ synthesis and characterization of electrically conductive polypyrrole/graphene nanocomposites, Polymer 51 (2010) 5921–5928. [16] W.S. Hummers, R.C. Offeman, Preparation of graphitic oxide, J. Am. Chem. Soc. 80 (1958) 1339. [17] Y. Wang, Y.M. Li, L.H. Tang, J. Lu, J.H. Li, Application of graphene-modified electrode for selective detection of dopamine, Electrochem. Commun. 11 (2009) 889. [18] D. Li, M.B. Muller, S. Gilje, R.B. Kaner, G.G. Wallace, Processable aqueous dispersions of graphene nanosheets, Nat. Nanotechnol. 3 (451) (2008) 101–105. [19] A.R. Liu, C. Li, H. Bai, G.Q. Shi, Elecrochemical deposition of polypyrrole/ sulfonated graphene composite films, J. Phys. Chem. C 114 (2010) 22783–22789. [20] L.L. Zhang, S.Y. Zhao, X.N. Tian, X.S. Zhao, Layered graphene oxide nanostructures with sandwiched conducting polymers as supercapacitor electrodes, Langmuir 26 (22) (2010) 17624–17628. [21] Y.C. Zhao, L. Zhan, J.N. Tian, S.L. Nie, Z. Ning, Enhanced electrocatalytic oxidation of methanol on Pd/polypyrrole–graphene in alkaline medium, Electrochim. Acta 56 (2011) 1967–1972. [22] S. Zhao, G.Q. Zhang, L. Fu, L.F. Liu, X.H. Fang, F.L. Yang, Enhanced electrocatalytic performance of anthraquinonemonosulfonate-doped polypyrrole composite: electroanalysis for the specific roles of anthraquinone derivative and polypyrrole layer on oxygen reduction reaction, Electroanalysis 23 (2) (2011) 355–363. [23] M.H. Al-Malack, A.A. Bukhari, N.S. Abuzaid, Crossflow microfiltration of electrocoagulated kaolin suspension: fouling mechanism, J. Membr. Sci. 243 (2004) 143–153.
Contents lists available at SciVerse ScienceDirect
Journal of Membrane Science journal homepage: www.elsevier.com/locate/memsci
Preparation of highly conductive cathodic membrane with graphene (oxide)/PPy and the membrane antifouling property in filtrating yeast suspensions in EMBR Lifen Liu n, Feng Zhao, Jiadong Liu, Fenglin Yang MOE, Key lab of industrial Ecology and Environmental Engineering, School of Environmental Science and Technology, Dalian University of Technology, 116024 Dalian, China
a r t i c l e i n f o
abstract
Article history: Received 27 July 2012 Received in revised form 14 November 2012 Accepted 21 February 2013 Available online 4 March 2013
The EMBR (MBR, attached with electric field) is effective in suppressing membrane fouling. In this paper, modified polyester cathode membranes were prepared by coating graphene (Gr)/polypyrrole (PPy) or graphene oxide (GO)/PPy, which had higher conductivity and better antifouling property in EMBR than the membrane modified with PPy only. The membrane modification using a vapor phase polymerization method (VPM) was proved better than using a liquid phase polymerization method (LPM). The modified membranes were tested and compared for antifouling performances in EMBR, by short time filtration of a yeast suspension, under 1 V/cm electric field. The VPM Gr/PPy modified membrane suppressed membrane fouling and increased the flux as high as 20%. The VPM Gr/PPy coating was more uniform and firm, and had higher conductivity (resistance lowered by 67%) than PPy modified membrane. It had better antifouling performance, smaller slopes in its t/V–V curves (adopting the classical cake filtration model, a suitable description for the membrane fouling process). By applying 1 V/cm electric field, an increase of 20% in the short-term cumulative permeate volume was obtained for the Gr/PPy modified membrane and it was only 10% for PPy modified membrane. & 2013 Elsevier B.V. All rights reserved.
Keywords: Conductive polymer composites Cathodic membrane Membrane fouling suppression EMBR Graphene
1. Introduction Membrane bioreactors (MBRs) have a lot of advantages over other conventional methods for wastewater treatment [1]. It has become more and more popular for its advantages in higher quality effluent, smaller space occupation, higher degree of automation and lower sludge production. However, the decline of flux caused by membrane fouling is a big problem in wastewater treatment since it affects the performance significantly [2,3]. To efficiently improve the membrane anti-fouling performance and control membrane fouling in MBR, it has been proved effective by applying an electric field in MBR (the so-called EMBR) [4,5], in which negatively charged sludge and extra-cellular polymeric substances (EPS) are forced to move away from the membrane surface under applied voltage (Fig. 1). Also, the conductive layer on the membrane surface, especially around the pore entrance, may greatly suppress fouling in EMBR. The influence of the electric field and the conductivity of the cathode membrane on membrane flux are considered critical for membrane fouling suppression performance in EMBR. In general, membrane flux increases when the electric field strength
n
Corresponding author. Tel.: þ86 411 84706173; fax: þ 86 411 84708083. E-mail address: [email protected] (L. Liu).
0376-7388/$ - see front matter & 2013 Elsevier B.V. All rights reserved. http://dx.doi.org/10.1016/j.memsci.2013.02.045
increases within certain range, then beyond certain points, it is no longer economic, after reaching certain value (the critical potential) the flux remained almost constant [6]. Conducting at or above the critical potential can avoid the formation of filtration cake layer, but the high voltage requires highly stable electrodes and implies high energy consumption/cost. To reduce energy consumption, an intermittent electric field, applied only when the permeate flux has drastically declined, can also effectively suppress the membrane fouling [7]. Also, applying low voltage to the built-in cathode in the flat panel membrane module and better arranged anodes in configured MBR, can effectively utilize electrophoresis force in inhibiting bio-fouling of membrane [8]. Moreover, using the conductive membrane as cathode combines the membrane and electrode as one, so higher efficiency in material usage and better anti-fouling performance can be achieved. To make the conductive cathode membrane, depositing conductive polymers on the surface of membrane directly is effective and low cost, than arranging metal cathode in membrane modules. Polypyrrole (PPy) is a polymer with high conductivity, stable in MBR and safe for humans, and it has been widely used in electro-chemistry [9,10]. In our lab, conductive membrane made of PPy based on filter cloth have been successfully prepared [11] and showed really good property. However, the polymer conductivity on polyester cloth is not as good as conventional metal cathode (iron, aluminum, copper). And higher
100
L. Liu et al. / Journal of Membrane Science 437 (2013) 99–107
Negatively charged fouling particle
More Conductive layer Non conductive layer
Electrostatic rejection force Better fouling suppression
Less conductive layer
less fouling suppression
Cathode Membrane
Difference in membrane pore structure, conductivity and surface property affects fouling suppression performance
Fig. 1. The mechanism for membrane-fouling suppression in EMBR.
Fig. 2. (A) VPM process for GO(Gr)/PPy membrane (PPy membrane without the process in green dash line). (B) LPM process for GO(Gr)/PPy membrane (PPy membrane without the process in orange dash line). (For interpretation of the references to color in this figure caption, the reader is referred to the web version of this article.)
resistance means high energy consumption to produce the same effect in fouling reduction. Preparing high conductivity membrane from large pored, low price polyester filter cloth is critical for our research. Graphene and its oxides are cheap carbon materials with unique electronic properties [12]. Their composites with PPy have better conductivity than the pure PPy, widely used as supercapacitor, electrodes for microbial fuel cell, electrochemical oxygen reduction and pollutant removal [13–15]. In this paper, we prepared highly electronic-conductive membrane by modifying polyester filter cloth with Gr/PPy or GO/PPy via the VPM and/or LPM. The better conductive membrane was tested as cathode
membrane for anti-fouling performance in filtrating yeast suspension with 1 V in EMBR. Also for comparing membrane fouling performance, fouling models suitable for describing the flux decrease in our experimental system were studied and compared.
2. Experiment 2.1. Preparation and reduction of graphene oxide The graphene oxide (GO) was prepared using the Hummer’s method [16,17]. First, graphite (2.0 g) was mixed with 96 mL
L. Liu et al. / Journal of Membrane Science 437 (2013) 99–107
H2SO4 (98%) and NaNO3 (2 g) by stirring for 30 min in a beaker in an ice bath. Potassium permanganate (KMnO4, 12.0 g) was added very slowly in the suspension with vigorous stirring for 3.5 h while maintaining reaction temperature below 20 1C. Then the ice bath was removed, the reaction mixture was stirred at 35 1C for another 2 h. Next, 184 mL water was added to the pasty solution with constant agitation, and temperature was raised to 98 1C. There upon, the color of the solution changed to yellowish brown. After 2 h of vigorous stirring, 20 mL of 30% H2O2 was added to remove the excess KMnO4. The mixture was washed several times with 5% HCl and deionized (DI) water until the solution was free of sulfate and acid. Then ultrasonic treatment of the mixture was carried out until no more solid could be centrifuged. The concentration of GO solution was obtained by setting the volume of solution and weighing the formed GO, thus the stock GO solution was prepared, it is then diluted to different concentrations and used for immersing the membrane. Graphene was obtained from reducing its oxide by using hydrazine mono-hydrate, a strong reducing agent. The standard GO solution was diluted to form different concentration solution with DI water and the pH value was adjusted to 10 [18]. Then, hydrazine mono-hydrate (ratio of hydrazine: GO was 7:10 by weight) was added to the mixture and ultrasonically treated for 8 h. After the completion of the reaction, the reduced graphene oxide (Gr) was collected by filtration and dried. The black powder product thus obtained was washed with DI water several times to remove excess hydrazine, and the final product was dried in a vacuum oven at 60 1C for 24 h.
2.2. Vapor phase polymerization method and membrane modification The VPM membrane modifying process was shown in Fig. 2(A) [11]. First, blank filter cloth (39.4672.50 mm for the average and 66.51710.76 mm for the maximal pore size, 8 11 cm2) was immersed and ultrasonically treated for 10 min in GO(0.09 mg/mL) or Gr(0.02 mg/mL) solution. The filter cloth was taken out to dry in the air for 30 min, sufficient APS (ammonium persulfate) solution (50 mL, 20 mg/mL) was sprayed onto it. 0.5 mL pyrrole was first added to the bottom of a 1000 mL beaker, and the unmodified membrane was hanged in it above the pyrrole liquid without any contact with the liquid. The beaker was sealed with plastic film and it is filled with pyrrole (Py) vapor when its bottom was heated by 90 1C water bath. About 15 min, pyrrole molecules were completely adsorbed from the vapor phase onto the membrane and polymerized with GO (Gr) in situ to PPy with APS as initiator. The white filter cloth turned dark and was successfully modified by PPy and Gr/GO. Finally the membrane was rinsed with tap water.
2.3. Liquid phase polymerization method and membrane modification The LPM membrane modifying process is shown in Fig. 2(B). First, blank filter cloth (8 11 cm2) was immersed in APS solution (20 mg/mL, 50 mL) for 10 min and drip-dried in the air for 30 min. Dissolve 1 mL pyrrole in ethanol and water solution, then ultrasonically mix it with certain concentration graphene (or its oxide) solutions to get the GO/PPy and Gr/PPy composites. Then dip the filter cloth into the composites solution. Very soon a large quantity of composite GO/PPy (Gr/PPy) was formed by polymerization on the surface of blank filter cloth and made the membrane really dark. After 5 min, the membrane was taken out, hanged in air and rinsed with water.
101
2.4. Membrane characterization FT-IR spectrometer (IR Prestige-21, Shimadzu, Japan) was used to investigate the functional groups of graphite, graphene oxide (GO) and reduced graphene oxide (RGO). Membrane samples were viewed by digital camera (canon ixus 105) and the surface microcosmic morphologies of the prepared membranes were visualized by scanning electron microcosmic (SEM, JEOL JSM5600LV). Pore size distributions were measured with the Bubble Point Method according to the China national standard, GB/T 24219–2009 (determination of aperture of woven filtering fabric). The electric resistance of the membrane was measured by digital multimeter (Victor VC 830L) and averaged from values at six different locations. The quantities polymerized on the base membrane were measured using four 8n11 cm2 samples by calculating the weight increase of the membrane after modification. The stability of VPM and LPM was studied by observing their flux changes in different filtration cycles. 2.5. Filtration and the anti-fouling properties of the membrane The experimental set-up was schematically shown in Fig. 3. The flat sheet membrane module has an effective permeate area of 80 cm2. The separation unit has a volume of 4 L and the aeration rate was kept at 0.2 m3/h. According to the experiment above, the modified membrane with better conductivity was obtained for both VPM and LPM. They were chosen for anti-fouling property test in yeast solution (Zeta potential¼ 12.9 mV, size ¼7 mm), which was to simulate the condition in MBR. The results were compared with unmodified and only PPy modified membrane, with and without electric field. Membrane module was set and fixed in MBR container (MLSS¼ 5 g/L). When electric field was applied, the voltage was supplied from a steady DC power source. The anodes were made of two pieces of stainless steel meshes which were set 1 cm from the cathode at either side of the membrane module. The vertical distance between the liquid level and the bottom of effluent tube was fixed to 102 cm in every test, ensuring a corresponding transmembrane pressure at 10.00 kPa, so the gravity flow always has the same water head drop. The test was conducted in three cycles and each lasted for 1.5 h, between the intervals of cycles, the membrane was washed with tap water. Each three-cycle filtration test was repeated for three times. Each test uses one new piece of membrane of the same batch/kind.
3. Results and discussion The doping ratio of GO (Gr) and PPy had an effect on the property of membrane. To a certain degree, higher conductivity would be obtained if more Gr (larger than the ratio mentioned above in this paper) could strongly and uniformly fixed with PPy on the base membrane, just as other papers reported [19,20]. However, in our experiment, too much GO or Gr was against uniform fixation of PPy on the base membrane, the PPy could not coat the graphene (or its oxide) totally and the polymerized composites were on the surface of the membrane instead of among the woven fibers, which is a big disadvantage for the conductivity improvements. So the low concentration (GO¼0.09 mg/mL, Gr¼0.02 mg/mL) was chosen to treat the base membrane and better comprehensive property was obtained. 3.1. FT-IR of the composite membrane The Fourier-transform infrared (FTIR) spectroscopic study analyzed the functional groups on GO, compared with natural
102
L. Liu et al. / Journal of Membrane Science 437 (2013) 99–107
Fig. 3. The set-up of yeast filtration experiment (1, feed tank; 2, balance tank; 3, EMBR; 4, membrane module; 5, stainless iron meshes anode; 6, air pump; 7, air manometer; 8, aerating apparatus; 9, DC electrical source; 10, outflow tube).
1562 PPy Transmittance(%)
Gra phene oxide 2976 1047 Reduced GO 3383
3500
2927
1739
1636
Transmittance(%)
1477
Natural graphite
1046
1401
1681 1286
1216 927
GO/PPy
Gr/PPy
1400
3000 2500 2000 1500 Wavenumbers (cm-1)
961
1000
2000
1800
1600
1400
1200
1000
800
Wavenumbers (cm-1)
Fig. 4. The FT-IR of natural graphite, GO and Gr (A) and the FT-IR of PPy, GO/PPy and Gr/PPy (B).
graphite and the reduced GO. Upon oxidation of graphite to GO, the observed representative peaks of GO confirm the presence of the oxygen-containing functional moieties in carbon frameworks, which include the bands at 3383 cm 1(O–H stretching), 1739 cm 1(C¼O stretching vibration of carboxyl), 1636 cm 1 (water H–O–H bending vibration), 1400 cm 1(O–H deformation vibration ) and 1047 cm 1(C–O vibration). The curves of graphite and Gr are much smoother than GO, the typical peak is around 1400 cm 1, confirming that by Hummer’s method, natural graphite is successfully transformed to GO, and reduction of GO did take place by chemical reaction with hydrazine (Fig. 4(A)). Fig.4(B) shows the bands of typical pyrrole ring stretching at 1401, 1477 and 1562 cm 1, identifying the presence of the conductive polymer PPy. For PPy, the peaks at 1681 cm 1 and 1562 cm 1 are assigned to the carbonyl group and the C ¼C/C–C
ring stretching of pyrrole. The bands at 1477 cm 1 and 1286 cm 1 are corresponding to the stretching vibration in the ring of C–N bond and C–H vibrations. The bands at 1046 cm 1 are caused by N–H bending vibration. In-plane and out-plane ¼C–H bending vibration is located at 961 cm 1 and 927 cm 1. This agrees well with other reports in literature [21,22] for PPy. The curves of GO/PPy and Gr/PPy are similar to the PPy, with some differences. PPy, GO and Gr curves all have an obvious peak at 1401 cm 1 for O–H deformation vibration, then in the composite membrane of GO/PPy and Gr/PPy, this representative peaks intensified (weakened) in strength by GO (Gr) doping. For the same at peak 1047 cm 1, where the transmittance peak strength for Gr/PPy weakened than pure PPy, whereas the GO/PPy peak intensified because of the existence of GO whose absorbance peak is strong here.
L. Liu et al. / Journal of Membrane Science 437 (2013) 99–107
3.2. Macroscopic view and the morphology of the conductive membrane surface 3.2.1. Macroscopic view of the membrane Membranes modified by VPM were shown in Fig. S1. The color was uniform and very dark compared with the white blank membrane. All the modified membrane had firm interfacial bonding between the composite coating and base membrane, and membrane colors were slightly different as different coating components were used. The color of the membrane prepared by LPM (Fig. S2) seemed uniform from macroscopic view. But only when more reagents (about two times Py quantity of VPM) were used, could the filter cloth be totally modified by conductive polymer and the membrane seemed much darker than the VPM membrane, in which less PPy composite was coated.
3.2.2. The morphology of the conductive membrane surface Micro-scale morphology of the membrane was shown by SEM (Fig. 5) and it further demonstrated that all the VPM was successful in modifying the membranes (images B for PPy, C for GO/PPy, D for Gr/PPy), the modification layer was more uniform than in the LPM modified membranes (image E). For the VPM modified GO/PPy or Gr/PPy membrane, the surface of the coated fibers had less wrinkles or ripples/ridges than in image B (PPy only), was generally more even and uniform, though typical tiny nodes of PPy can be seen. And the tiny sheet structures from PPy
103
coated GO (image C2) or Gr (image D2) can be clearly seen in the SEM images. 3.3. The pore size of the modified membrane In Fig. 6(A), the relative ratios of pore size of the VPM modified membrane to the average pore size of blank membrane were shown, the average pore size of the modified membrane did not shrink or narrow much than the blank membrane. Although, from the SEM images, we can see the thin layers of composite coating on the fiber, the pore size measured by bubble point method did not show much change because the pore size was much bigger and wider than the thin coating layer. As pyrrole in liquid phase tends to polymerize on the coarse surface of the base membrane instead of around the fibers, the modified membrane by LPM was not uniform and pore blocking was serious. Membrane prepared in this way had much smaller average pore size than the blank membrane (Fig. 6(B)). 3.4. Conductivity The conductivity comparison of the modified membranes prepared by VPM and LPM methods was listed in Table 1. For the membrane prepared by VPM, the electric resistance for the modified membrane can be as low as 710 O cm 1 (modified by GO/PPy) and 680 O cm 1 (modified by Gr/PPy), much lower than 2.03 k O cm 1 for the only PPy modified filter cloth. And the
Fig. 5. The SEM of the membranes (A, blank membrane; B, PPy membrane by VPM; C, GO/PPy membrane by VPM; D, Gr/PPy membrane by VPM; E, Gr/PPy membrane by LPM).
80
80
Max Pore Size Average Pore Size
70
60 Pore size (μm)
60 Pore size (μm)
Max Pore Size Average Pore Size
70
50 40 30
50 40 30
20
20
10
10
0
0 Blank
PPy-M.
GO/PPy-M.
Membrane by VPM
Gr/PPy-M.
Blank
PPy-M.
GO/PPy-M.
Gr/PPy-M.
Membrane by LPM
Fig. 6. The average and maximal pore size for VPM membrane (A) and the average and maximal pore size for LPM membrane (B).
104
L. Liu et al. / Journal of Membrane Science 437 (2013) 99–107
membrane prepared by LPM had a resistance of 14.81 k O cm 1 for GO/PPy modified membrane and 2.7 k O cm 1 (modified by Gr/PPy), compared with 8.15 k O cm 1 (modified only by PPy). The low conductivity of LPM membrane may be caused by nonuniform polymerization of Py on the membrane surface. Just as we have observed from the SEM (Fig. 5), the PPy is lumpish and polymerized on the surface instead of among the fibers, which contributes to the poor conductive performance of LPM membrane. For both VPM and/or LPM, the Gr/PPy modified membrane had the lowest electric resistance. Moreover, the VPM membrane
Table 1 The resistance of membrane modified with PPy, GO/PPy and Gr/PPy by different polymerization method. The modification method
Sample Blank membrane
Resistance (kO/cm) þN
Quality (mg/cm2) –
VPM
PPy-M GO/PPy-M Gr/PPy-M PPy-M GO/PPy-M Gr/PPy-M
2.03 7 0.61 0.71 7 0.10 0.68 7 0.08 8.15 7 0.79 13.21 7 0.59 2.70 7 1.05
1.70 1.59 1.70 1.07 1.74 2.08
LPM
Blank membrane Gr/PPy-membrane
500
tends to have smaller deviation in the conductivity than LPM membrane when the same dopant was used. 3.5. Fouling reduction of VPM and LPM modified membrane The membrane prepared by LPM did not show any regularity in flux changes in the three cycles yeast filtration test (Fig. 7). When no electric field was applied, the flux of modified membrane was much smaller than the blank membrane in the first filtration cycle. This is caused by the macroscopic lumpish PPy and Gr/PPy coatings, blocking the space between membrane fibers, resulted in a much smaller pore size. Then in the second and third cycles, as the membrane wash may remove/wipe off some Gr/PPy, the flux for the modified membrane increased to approach the level of blank membrane. Applying 1 V electric field to the LPM Gr/PPy modified membrane, enabled an increase of flux than without applied voltage, in the first cycle. But in the second and third cycles, it did not show any improvement as the conductive materials may partly fall off during the membrane wash, and resulted in the decrease of conductivity. During the yeast filtration test without electric field applied, flux of all the modified membrane by VPM did not show obvious difference with blank membrane in the three cycles (Fig. S3)
PPy-membrane Gr/PPy-membrane with E=1V
Flux(L/m2h)
450 175 150 125 100 75 50 0
25
50
75
100 125 150 175 Permeate time (min)
200
225
250
275
Fig. 7. Permeate flux of blank, PPy, Gr/PPy LPM membrane with or without electric field applying in three filtration cycles (MLSS ¼ 5000 mg/L, TMP¼ 10.0 kPa).
PPy-M.
PPy-M. with E=1V
Gr/PPy-M. with E=1V
Flux(L/m2h)
500
180 160 140 120 100 80 0
25
50
75
100 125 150 175 Permeate time (min)
200
225
250
275
Fig. 8. Permeate flux of PPy, Gr/PPy VPM membrane with or without applying electric field in three filtration cycles (MLSS ¼ 5000 mg/L, TMP¼ 10.0 kPa).
L. Liu et al. / Journal of Membrane Science 437 (2013) 99–107
because of the better hydrophilic property of PPy [11]. However, as the modified membrane has smaller pore size than the blank membrane, it shows better turbidity removal property (Fig. S4), which stands for the membrane rejection property. According to the figure, more than 90% turbidity was removed in the first 10 min for the blank membrane, and that turbidity removal for the modified membrane was 100%, which was maintained to the end of the first filtration cycle. Further research was conducted to exhibit the conductive membrane property. The anti-fouling property of PPy and Gr/PPy modified membrane was shown in Fig. 8, when 1 V electric field is applied, comparing with the PPy modified membrane without electric field. According to the results, modified membrane had an increase in flux and the biggest improvement was obtained for membrane with better conductivity, the Gr/PPy modified membrane. The fouling on VPM modified membrane was reversible, since the flux did not present any obvious change
Table 2 The increasing percentage of cumulative volume for conductive membrane with E compared with PPy-M without applying E. Membrane sample tested
1st cycle
2nd cycle
3rd cycle
PPy-M with E Gr/PPy-M with E
8.9% 16.9%
10.8% 21.3%
10.8% 20.3%
in the three cycles after membrane washing with tap water, indicating a good performance. Applying electric field, increased total permeate volume from each cycle and the fluxes of all the modified membranes (Fig. S5). Comparing with the PPy modified membrane without applied electric field, the total permeate volume increase by applying electric field was 10% for PPy and the increase for Gr/PPy modified membrane was 20% (Table 2). 3.6. The membrane fouling mechanism In order to investigate the membrane fouling process, we studied two models according to the previous research [23], the standard filtration model and classical filtration model. First, the standard filtration model, suits to the situation that the particle size is much smaller than the pore size of the primary membrane and the membrane fouling is caused mainly by the internal pore clogging. t 1 k ¼ þ t V Q0 2 where t is the filtration time (min), V is the cumulative permeate volume for corresponding filtration time (L), Q0 is the initial flux rate (L/min) and k is the filtration constant (L 1). When the membrane pore size was initially narrowed down by the particles to be filtrated, the deposit particles started to form a cake layer on membrane surface and decreased the flux.
60
60 PPy-M. PPy-M. E=1v Gr/PPy-M.E=1v
50
50
PPy-M. PPy-M. E=1v Gr/PPy-M. E=1v
40 t/V (min/L)
40 t/V (min/L)
105
30 20
y=1.134x+a1 R2=0.9383
30 20 y=32.86x+a1 R2=0.9950
2
y=1.015x+a2 R =0.9455
10
10
y=28.97x+a2 R2=0.9971
y=0.994x+a3 R2=0.9344 0
y=23.39x+a3 R2=0.9943
0 0
20
40 60 Permeate time(min)
80
100
0.5
1.0 1.5 Cumulative volume(L)
2.0
Fig. 9. The t/V–t (A) curves (at the filtration beginning stage) and t/V–V (B) curves (all from second points started) for PPy membrane without electric field, PPy membrane and/or Gr/PPy membrane with 1 V/cm electric field (R2 means linear correlation coefficient).
PPy-M.
PPy-M. with E=1V
Linear Fit of PPy-M.
Gr/PPy-M. with E=1V Linear Fit of PPy-M. E=1v
Linear Fit of Gr/PPy-M. E=1v 50
t/V(min/L)
40
1st cycle
30 20 2nd cycle
10
3rd cycle
0 0.0
0.5
1.0
1.5
2.0
2.5 3.0 3.5 4.0 Permeate volume (L)
4.5
5.0
5.5
6.0
Fig. 10. The three filtration cycles t/V–V curves for PPy membrane with or without electric field, and Gr/PPy VPM membrane with electric field (all starts from the second points in each cycle and when a new cycle start, set t as 0 min).
106
L. Liu et al. / Journal of Membrane Science 437 (2013) 99–107
Then classical cake filtration model represents the decrease in permeate volume t mrm mcRc 1 K ¼ þ þ V UV ¼ V Q0 2 AUTMP 2A2 TMP
m is the dynamic viscosity (Pa s), rm is the membrane resistance (m 1), and A is the membrane area (m2). TMP is the transmembrane pressure (Pa). c is the concentration of particles (kg/L), Rc means the resistance of the cake layer (m/kg), and K is the cake filtration constant. For the models, the smaller the filtration constant (K) and the cake filtration constant (K), the lower filtration resistance and the higher the mass transfer efficiency is. This is what we only considered to study the membrane fouling in our experiment. According to the models, t/V–V and t/V–t curves were drawn and analyzed. Comparing Fig. 9(A) with (B), we know the fouling mechanism for our filtration experiment fits the classical model better on the whole, confirmed by the better coincidence with
Table 3 The slopes and linear correlation coefficients of the curves correspond to Fig. 10 (R2 means linear correlation coefficient). Filtration cycle
PPy-M R
1st cycle 2nd cycle 3rd cycle
2
0.9950 0.9951 0.9967
PPy-M, E¼ 1 V 2
Gr/PPy-M, E ¼1 V
Slope
R
Slope
R2
Slope
32.86 36.99 36.45
0.9971 0.9952 0.9939
28.97 30.48 30.55
0.9943 0.9904 0.9932
23.39 24.01 24.89
linear relation of t/V–V curves than the t/V–t curves. This means the internal pore clogging quickly occurred, then the follow-up flux decrease was mainly caused by the filtration cake layer and the modified membrane with applied electric field could effectively suppress the depositing of particles on the surface of membrane. Moreover, at the initial filtration period, slopes of t/V–t curves for both conductive membrane with applied voltage did not show significant difference, indicating that the low electric field has little effect on improving the membrane flux at the beginning stage of test. However, for the t/V–V curves, Gr/PPy modified membrane with electric field has the smallest slope, second is the PPy with electric field and then the PPy-membrane without applied electric field, which is in accordance with their resistances, indicating membrane with better conductivity has better anti-fouling property. When the tests were finished after three cycles (Fig. 10 and Table 3), the advantage of better conductive membrane in fouling suppression was shown well kept, during the tests. This observation justifies the explanation proposed in Fig. 1(B). PPy has cathodic catalytic oxygen reduction reactivity (ORR), doped by AQS (anthraquinone monosulfate), this ORR property may be increased significantly [22], thus, the fouling reduction/ suppression mechanism of PPy modified membrane cathodes in EMBR can be extended, the results are also expected to increase under applied electric field. This is being studied in our lab, and the results will be reported elsewhere. Also we have supplied the SEM of the membranes (Fig. 11) after being used in E-MBR. Comparison of the images showed the stable membrane coating for both the Gr/PPy and PPy membranes, and confirmed the better fouling reduction property for Gr/PPy membrane than PPy membrane.
Fig. 11. The SEM of the membranes (A, C: PPy membrane by VPM; B, D: Gr/PPy membrane by VPM) after applying in E-MBR.
L. Liu et al. / Journal of Membrane Science 437 (2013) 99–107
4. Conclusion The following conclusion can be drawn from this study: The Gr/PPy modified membrane prepared by VPM had better conductivity than that prepared by LPM. Both GO and Gr doping of PPy during modification improved the conductivity of the modified membranes. The Gr/PPy modified membrane had better property than the GO/PPy modified membrane when the Gr or GO doping concentration was at the same level, or with much lower concentration of Gr than GO. For the Gr/PPy membrane, its electric resistance was 1/3 of the PPy modified membrane. With 1 V/cm electric field applied, Gr/PPy modified membrane can suppress the fouling more effectively than PPy modified membrane, and the cumulative permeate increase after 1.5 h filtration was 20% while that for PPy modified membrane was only 10%.
Acknowledgment Authors acknowledge the financial support from China National Natural Science Foundation (Project No. 21177018).
Appendix A. Supporting information Supplementary data associated with this article can be found in the online version at http://dx.doi.org/10.1016/j.memsci.2013. 02.045.
Nomenclature A c k K Q0 Rc rm t TMP V
the membrane area (m2) the concentration of particles (kg/L) the filtration constant (L 1) the cake filtration constant the initial flux rate (L/min) the resistance of the cake layer (m/kg) the membrane resistance (m 1) the filtration time (min) the trans-membrane pressure (Pa) the cumulative permeate volume for corresponding filtration time (L)
References [1] F.G. Meng, S. Chae, A. Drews, M. Kraume, H. Shin, F.L. Yang, Recent advances in membrane bioreactors (MBRs): membrane fouling and membrane material, Water Res. 43 (2009) 1489–1512.
107
[2] P. Le-Clech, V. Chen, T.A.G. Fane, Fouling in membrane bioreactors used in wastewater treatment, J. Membr. Sci. 284 (2006) 17–53. [3] A. Drews, Membrane fouling in membrane bioreactors—characterisation, contradictions, cause and cures, J. Membr. Sci. 363 (2010) 1–28. [4] K. Bani-melhem, M. Elektorowicz, Development of a novel submerged membrane electro-bioreactor (SMEBR): performance for fouling reduction, Environ. Sci. Technol. 44 (9) (2010) 3298–3304. [5] K. Akamatsu, Y. Yoshida, T. Suzaki, Y. Sakai, H. Nagamoto, S. Nakao, Development of a membrane–carbon cloth assembly for submerged membrane bioreactors to apply an intermittent electric field for fouling suppression, Sep. Purif. Technol. 88 (2012) 202–207. [6] J.P. Chen, C.Z. Yang, J.H. Zhou, X.Y. Wang, Study of the influence of the electric field on membrane flux of a new type of membrane bioreactor, Chem. Eng. J. 128 (2007) 177–180. [7] K. Akamatsu, W. Lu, T. Sugawara, S. Nakao, Development of a novel fouling suppression system in membrane bioreactors using an intermittent electric field, Water Res. 44 (2010) 825–830. [8] L.F. Liu, J.D. Liu, B. Gao, F.L. Yang, Minute electric field reduced membrane fouling and improved performance of membrane bioreactor, Sep. Purif. Technol. 86 (2012) 106–112. [9] G. Kang, R.B. Borgens, Y. Cho, Well-ordered porous conductive polypyrrole as a new platform for neural interfaces, Langmuir 27 (2011) 6179–6184. [10] A. Laith, D. Catherine, B. Stephan, W. Wojtek, Electropolymerized polypyrrole nanowires for hydrogen gas sensing, J. Phys. Chem. C 116 (2012) 13388–13394. [11] L. Liu, J. Liu, B. Gao, F. Yang, S. Chellam, Fouling reductions in a membrane bioreactor using an intermittent electric field and cathodic membrane modified by vapor phase polymerized pyrrole, J. Membr. Sci. 394–395 (2012) 202–208. [12] J.A. Matthew, C.T. Vincent, B.K. Rechard, Honeycomb carbon: a review of graphene, Chem. Rev. 110 (2010) 132–145. [13] P. Si, S. Ding, X.W. (David) Lou, D.H. Kim, An electrochemically formed threedimensional structure of polypyrrole/graphene nanoplatelets for highperformance supercapacitors, RSC Adv. 1 (2011) 1271–1278. [14] S. Biswas, L.T. Drzal, Multilayered nanoarchitecture of graphene nanosheets and polypyrrole nanowires for high performance supercapacitor electrodes, Chem. Mater. 22 (2010) 5667–5671. [15] S. Bose, T. Kuila, E. Uddin, N.H. Kimb, A.K.T. Lau, J.H. Lee, In-situ synthesis and characterization of electrically conductive polypyrrole/graphene nanocomposites, Polymer 51 (2010) 5921–5928. [16] W.S. Hummers, R.C. Offeman, Preparation of graphitic oxide, J. Am. Chem. Soc. 80 (1958) 1339. [17] Y. Wang, Y.M. Li, L.H. Tang, J. Lu, J.H. Li, Application of graphene-modified electrode for selective detection of dopamine, Electrochem. Commun. 11 (2009) 889. [18] D. Li, M.B. Muller, S. Gilje, R.B. Kaner, G.G. Wallace, Processable aqueous dispersions of graphene nanosheets, Nat. Nanotechnol. 3 (451) (2008) 101–105. [19] A.R. Liu, C. Li, H. Bai, G.Q. Shi, Elecrochemical deposition of polypyrrole/ sulfonated graphene composite films, J. Phys. Chem. C 114 (2010) 22783–22789. [20] L.L. Zhang, S.Y. Zhao, X.N. Tian, X.S. Zhao, Layered graphene oxide nanostructures with sandwiched conducting polymers as supercapacitor electrodes, Langmuir 26 (22) (2010) 17624–17628. [21] Y.C. Zhao, L. Zhan, J.N. Tian, S.L. Nie, Z. Ning, Enhanced electrocatalytic oxidation of methanol on Pd/polypyrrole–graphene in alkaline medium, Electrochim. Acta 56 (2011) 1967–1972. [22] S. Zhao, G.Q. Zhang, L. Fu, L.F. Liu, X.H. Fang, F.L. Yang, Enhanced electrocatalytic performance of anthraquinonemonosulfonate-doped polypyrrole composite: electroanalysis for the specific roles of anthraquinone derivative and polypyrrole layer on oxygen reduction reaction, Electroanalysis 23 (2) (2011) 355–363. [23] M.H. Al-Malack, A.A. Bukhari, N.S. Abuzaid, Crossflow microfiltration of electrocoagulated kaolin suspension: fouling mechanism, J. Membr. Sci. 243 (2004) 143–153.