- Email: [email protected]
Effects of epichlorohydrin–dimethylamine on coagulation and membrane performance of ferric chloride in coagulation–ultrafiltration hybrid process
Chemical Engineering Journal 280 (2015) 634–642
Contents lists available at ScienceDirect
Chemical Engineering Journal journal homepage: www.elsevier.com/locate/cej
Effects of epichlorohydrin–dimethylamine on coagulation and membrane performance of ferric chloride in coagulation–ultrafiltration hybrid process Shenglei Sun, Fan Bu, Xin Huang, Shuang Zhao, Hongyu Dong, Baoyu Gao ⇑, Qinyan Yue ⇑, Lijuan Feng, Yan Wang, Qian Li Shandong Key Laboratory of Water Pollution Control and Resource Reuse, School of Environmental Science and Engineering, Shandong University, Ji’ nan 250100, China
h i g h l i g h t s Coagulation efficiency of FC on HA–Kaolin could be enhanced apparently by DAM–ECH. Initial pH would remarkably influence the coagulation performance and floc properties. Application of DAM–ECH result in larger flocs with better recovery ability. DAM–ECH addition would generate flocs with loose and open structure. Membrane performance was significantly improved by DAM–ECH.
a r t i c l e
i n f o
Article history: Received 12 March 2015 Received in revised form 3 June 2015 Accepted 4 June 2015 Available online 18 June 2015 Keywords: Membrane fouling Epichlorohydrin–dimethylamine Ferric chloride Floc properties Coagulation–ultrafiltration
a b s t r a c t Cationic polymer epichlorohydrin–dimethylamine (DAM–ECH) was applied as coagulation aid of ferric chloride (FC) in treatment of humic acid–kaolin (HA–Kaolin) simulated water in this study. Impacts of DAM–ECH on coagulation performance, flocs characteristics and membrane fouling of FC in coagulation–ultrafiltration (C–UF) process were investigated under different pH conditions. Results showed that the removal rates of UV254 (ultraviolet adsorption at 254 nm) and DOC (dissolved organic carbon) of FC/DAM–ECH was much better than that of FC at same dosage. The optimum coagulation condition was FC at 10 mg/L with FC/DAM–ECH mass ratio (MR) of 3:1 based on the consideration of treatment cost and performance. Moreover, turbidity, UV254 and DOC removal rates increased when pH increased from 4 to 6 and then decreased as pH grew from 7 to 10. Thus, the optimum pH for coagulation was also determined to be 6 accordingly. Addition of DAM–ECH also increased sizes of produced flocs and made them looser in structure comparing with FC–HA flocs. Meanwhile, results obtained from ultrafiltration experiments showed that membrane fouling would be reduced by introducing DAM–ECH in coagulation treatment; the optimum membrane performance was given by FC/DAM–ECH under pH 6. In conclusion, proper dose of FC/DAM–ECH under pH 6 could ensure an ideal HA removal rate and meanwhile increase operating duration of C–UF system by restricting membrane fouling. Ó 2015 Elsevier B.V. All rights reserved.
1. Introduction It is now widely recognized that as a kind of natural organic matter (NOM) which commonly existing in most of drinking water resources, humic substances are jeopardizing human health [1,2]. Thus, removal of humic substances has drawn worldwide attention
⇑ Corresponding authors. Tel.: +86 531 88366771; fax: +86 531 88364513 (B. Gao). Tel.: +86 531 88365258; fax: +86 531 88364513 (Q. Yue). E-mail addresses: [email protected] (B. Gao), [email protected] (Q. Yue). http://dx.doi.org/10.1016/j.cej.2015.06.011 1385-8947/Ó 2015 Elsevier B.V. All rights reserved.
recent decades in modern water treatment. As an extensively used coagulant in water treatment processes, ferric chloride (FC) can remove humic acid molecules in water bodies by forming strong and large flocs [3–5]. Moreover, many researchers have proved that compounding of iron salt flocculants with organic polymer would be more effective in HA and turbidity removal than single use of iron salts [6]. Thus, a cationic polymer epichlorohydrin–dimethylamine (DAM–ECH) was applied as coagulation aid of ferric chloride (FC) in this study. Meanwhile as a membrane filtration process, ultrafiltration has been broadly applied as postprocessing of water treatment process to further improve the reduction of
S. Sun et al. / Chemical Engineering Journal 280 (2015) 634–642
particle concentration and NOM in portable water resources. This is because that UF processes could not remove most of dissolved organic matter from water bodies because of its large pore sizes [7]. Another main bottle-neck in ultrafiltration utility is the drop of permeate flux caused by membrane fouling, which would hugely increase the power demand to maintain an acceptable flow rate of the UF device [8,9]. It has been proved that pretreatment (such as adsorption, coagulation or peroxidation) can remarkably reduce improve membrane performance by achieving larger permeate flux in combined water treatment systems, in which coagulation has been proved to be more effective than the other pretreatments in reduction of membrane fouling [10]. Breakthroughs have been made by previous researches in all aspect of C–UF system, including influence of coagulants/flocculants by Xu et al. and Zhao et al. [11,12]; optimization of operating conditions in application of C–UF in drinking water treatment and pilot-scale hybrid municipal wastewater reclamation [13,14]. Therefore, coagulation–ultrafiltration hybrid process (C–UF) is becoming more and more popular in high quality water treatment processes for its high removal efficiency of NOM. Generally, C–UF is significantly affected by many elements such as feed water types, coagulation methods and hydraulic conditions (shearing force) [15]. As mentioned above, combination of coagulation and ultrafiltration process is mutually beneficial which can further improve the water quality comparing with single use of each process. Furthermore, combine use of these two processes should be optimize through designed experiment processes. Thus, proper addition method of coagulants was studied to achieve acceptable coagulation and membrane performance in this study.
635
mixture was transferred into a 1000 mL volumetric flask and diluted to volume (1 L). Kaolin stock solution was applied in this study to adjust turbidity of HA simulated water, which was prepared as follow: 5.0 g of kaolin (Tianjin Kermel Chemical Reagent Co., Ltd, China) was weighted and mixed with 1.0 L of deionized water, the suspension liquid were stirred by a magnetic stirring apparatus for at least 30 min; then the upper 500 mL liquid was withdrawn for later use. HA–Kaolin simulated water of this study was prepared by diluting 10 mL of the HA stock solution into 1.0 L of tap water. Initial turbidity of the raw water was adjusted to 15.0 ± 0.5 NTU using kaolin stock solution. Characteristics of raw water were as follow: UV254 (ultraviolet absorbance at 254 nm) = 0.320 ± 0.010.
DOCðdissolved organic matterðDOMÞ concentrationÞ ¼ 3:55 0:03 mg=L Raw water pH ¼ 8:32 0:02 Many researchers have found that UV254 can be used as indicator of humic substances and aromatic organic matter contain C@C and C@O in water bodies, that higher humic concentration in water always show high UV254 adsorption [13,16]. Furthermore, the relationship between UV254 and DOC have been studied by many researchers, results showed that 70% to 80% of DOC removal was aquatic humic fraction with high UV254 absorption value, only 10% of DOC was hydrophobic organic with low UV254 adsorption. In other words, UV254 removal should be positive correlated with DOC removal [16]. The initial pH of the raw water sample was adjusted by 0.1 M HCl and 0.1 M NaOH solution when necessary during the tests.
2. Materials and methods
2.3. Jar test
2.1. Preparation of coagulants
In this study, jar tests were conducted to determine the optimum coagulation method for HA–Kaolin water, which were carried out by a program-controlled flocculator with six 1.5 L cylindrical plexiglass beakers and six flat paddle impellers (ZR4-6, Zhongrun Water Industry Technology Development Co. Ltd., China). During jar tests, 6 L of prepared HA–Kaolin solution was filled into six breakers averagely (1 L of raw water in each breaker). And the operation program was preseted as follow: first, synthetic water sample in each breakers was stirred rapidly under 200 rpm (rpm) for 0.5 min; then certain dose of coagulant was added into each breakers, followed by another stirring period of 200 rpm for 0.5 min; following the third period, the mixture in each breakers would be stirred under 200 rpm for 1.5 min for dispersion of coagulant; in the fourth stage, floc would grow under slow stirring speed of 40 rpm for 10 min; during the final stage, the shearing force would be removed for flocs to settle for 20 min. After the sedimentation period, 200 ml of supernatant water was withdrawn from each jar to use as water samples in the following tests (Turbidity, DOC, UF, etc.). Turbidity of water sample would be measured by a 2100P turbidimeter (Hach, USA) and zeta potential was measured using a Zetasizer 3000Hsa (Malvern Instruments, UK) at room temperature (around 25 °C). Each water sample was filtered through a 0.45 lm fiber membrane before DOC analysis using a Shimadzu TOC-VCPH analyzer and UV254 measurement by a UV-754 UV/VIS spectrophotometer (Precision Scientific Instrument Co. Ltd., Shanghai, China).
FC solution with concentration of 2 g/L (as Fe3+) was prepared by an analytical reagent FeCl3 6H2O (CAS No: 10025-77-1). 0.9661 g of FeCl3 6H2O was completely dissolved in 100 ml of deionized water stirred by a magnetic stirring apparatus until the solution became clarified. DAM–ECH copolymer was polymerized by analytical reagent (A.R.) epichlorohydrin and dimethylamine (A.R.) with ethanediamine (A.R.) as cross-linker. The polycondensation reaction that producing DAM–ECH copolymer was conducted in a 250 mL of four-neck round bottom flask. In addition, several facilities were required during the synthesis process: a mechanical stirrer, a thermometer, a dropping funnel and corresponding glass spigots. Certain dosage of epichlorohydrin was poured into the four-neck round bottom flask and meanwhile the environmental temperature within the flask was kept at 10 °C by a thermostated water bath device. Then dimethylamine was constantly dripped into the flask using a dropping funnel, and the mixture was under steady stirring throughout the whole dripping process. At last, certain dose of ethanediamine was introduced into flask under continuous stirring as the cross-linker [5].
2.2. Simulated water sample Humic acid–kaolin (HA–Kaolin) water was used as raw water in this study. The HA–Kaolin stock solution was prepared as follow: 1.0 g of HA powder (Aladdin, Shanghai, China) and 0.40 g of NaOH (Tianjin Damao Co., Tianjin, China) were weighted and dissolved in deionized water under continuously stirring of a magnetic stirring apparatus till completely dissolved; then the
2.4. On-line monitoring of floc formation, breakage and re-growth The evolution of flocs during coagulation process was monitored by a laser diffraction instrument (Malvern Mastersizer 2000, Malvern, UK). Water samples were monitored through the
636
S. Sun et al. / Chemical Engineering Journal 280 (2015) 634–642
optical unit of the Malvern Mastersizer 2000 where a circulation was formed by drawing water from and back into the jar using a peristaltic pump on a 5 mm internal diameter return tube at a flow rate of 2.0 liter per hour (L/h). Malvern Mastersizer 2000 was connected to a computer to collect data every 30 s during the whole process including floc formation, breakage and regrowth. The coagulation program was set to be the same as jar test mentioned in Section 2.3, except after the slow stirring period for floc growth, the flocs were exposed to shearing forces (100 rpm, 200 rpm, and 400 rpm) for 5 min; at last there would be a 15 min of floc regrowth period. 2.4.1. Floc size In this study, the median volumetric diameter (d50) was used to denote the floc size. The floc size distribution was also investigated by Matersizer 2000. 2.4.2. Floc breakage factor and recovery factor Flocs breakage and recovery factors are normally used to reveal the breakage and re-growth of flocs, in which breakage factor indicates the ability of flocs to resist the shear power and recovery factor shows the capacity of flocs after breakage period. The breakage and recovery factor of flocs can be calculated by the following equations which have been used in other studies [17–20]:
Breakage factor ð%Þ ¼
d2 100 d1
ð1Þ
Recovery factor ð%Þ ¼
d3 d2 100 d1 d2
ð2Þ
2.4.3. Floc dimension Different coagulation conditions may produce different evolution of flocs structure. Fractal dimension is one commonly used parameter to characterize flocs structure. Fractal dimension of flocs was measured by Laser Light Scattering in this study. Smaller fractal dimension means the more open or irregular flocs structure and lower effective density. In a conventional coagulation process for water treatment, the flocs with high density are preferable because of their better settleability in the subsequent settling unit. However, it may not be the case in the coagulation–membrane processes. Fractal dimension can be calculated by the formula:
ð3Þ
where Df is the mass fractal dimension, which can be determined by the slope of a plot of I as a function of Q on a log–log scale, while I is the light density and Q is the difference between the incident and scattered wave vectors of the radiation beam in the medium, which is given by equation:
Q¼
4p n sinðh=2Þ k
The UF membranes used in this study were purchased from Shanghai Mosu Science Equipment Co., Ltd and the molecular weight cut-off (MWCO) of membrane was 100 K Dalton (kDa). UF experiment was conducted within a magnetically stirred cell (MSC050, Mosu, China), which had a total volume of 300 mL and an effective membrane area of 50.2 cm2. Pre-coagulated water samples were added directly into the magnetically stirred ultrafiltration batch cell of for ultrafiltration, and the coagulation program was set to be the same as that of the jar test but without sedimentation. The magnetically stirred cell was pressurized with nitrogen gas (N2) provide by a high pressure gas tank with constant pressure of 0.15 ± 0.05 MPa. An electronic balance (MSU5201S-000-D0, SARTORIUS AG GERMANY) was applied to measure permeate mass. Meanwhile, the balance was connected with a computer to record permeate mass value on the balance per 10 s. The detailed hybrid C–UF system is stated in Fig. 1, which revealed that coagulated water was drawn from the coagulation breakers and then the suspension liquid was filtrated through the ultrafiltration membrane under a constant pressure provided by N2. The permeate flux was monitored by a computer which connected with an electronic balance. Modified fouling index (MFI) was designed to measure the fouling potential of micro-filtration membrane and ultrafiltration membrane, and details can be found in previous study [21,22]:
t gRm gaC b ¼ þ V V DPA 2DPA2 |fflfflffl{zfflfflffl}
ð5Þ
MFI
d1, d2, d3 stand for flocs sizes in the steady phase before breakage, after the breakage period and after the re-growth to another steady phase, respectively. Larger breakage factors means the flocs are stronger and more difficult to be broken, meanwhile larger recovery factor expresses better regrowth ability of flocs.
I / Q Df
2.5. UF apparatus and C–UF hybrid process
ð4Þ
where n is the refractive index of the suspending medium, h is the scattering angle, and k is the wavelength of the radiation in a vacuum. In this study, five points (D0.5 value) in steady stage of formation, breakage and regrowth period were chosen and their average value were used for calculation of strength and recovery factors.
where V is the filtrate volume, t is the filtration time, DP is the applied trans-membrane pressure, Rm is the resistance of the membrane, A is available membrane area, g is the water viscosity, a is the specific resistance of the cake deposited and Cb is the concentration of particles in feed water. The MFI is determined from the gradient of the linear portion of t/V versus V. A higher MFI value means more serious membrane fouling [23]. 3. Results and discussion 3.1. Coagulation performance As mentioned in Section 2, jar tests were conducted to investigate the coagulation performance of FC and DAM–ECH with different coagulation methods. Turbidity, UV254, and DOC were used as main indicators to determine the coagulation performance and zeta potential was also monitored to help analysis the phenomena. 3.1.1. Effects of different FC dosage and FC/DAM–ECH MR on HA– Kaolin water treatment As shown in Fig. 2 A, it is obvious that at removal rate of turbidity grew dramatically as FC dosage increased from 2 mg/L to 4 mg/ L, no matter with or without DAM–ECH; then the growth became gentle when 6 mg/L of FC was added and finally almost remain steady as FC dosage increased to 12 mg/L. It is noticeable that turbidity removal rates were enhanced as DAM–ECH was introduced even at FC dosage of 2 mg/L. According to Fig. 2D, zeta potential did not changed much at that point and kept far below 10 mV, which means sweep and adsorption should take the main part of coagulation process [5]. This result revealed that DAM–ECH could effectively increase turbidity removal effect of FC under low dosage. Meanwhile UV254 and DOC removal by FC trended to be similar to each other, which can be concluded that the removal rates kept increase as FC dosage increased from 2 mg/L to 8 mg/L; then the growth slowdown in range of 8 mg/L to 10 mg/L and became negligible when FC dosage further increased to 12 mg/L. Furthermore,
S. Sun et al. / Chemical Engineering Journal 280 (2015) 634–642
Fig. 1. Schematic of coagulation–ultrafiltration hybrid experiment.
A
B
C
D
Fig. 2. Turbidity, UV254 and DOC removal rates and zeta potential as functions of FC/DAM–ECH dosages under different coagulation methods.
637
638
S. Sun et al. / Chemical Engineering Journal 280 (2015) 634–642
inconsideration of DAM–ECH dosage, it can be seen that increase of UV254 and DOC removal rate became insignificant when FC/DAM– ECH mass ratio was 3:1 at FC dosage of 10 mg/L. This can be explained as follow: Zeta potential almost remained unchanged when FC was used alone (Fig. 2D). But as high positively charged DAM–ECH was added, the charge neutralization mechanism of FC with HA molecule was dramatically enhanced, so that NOM (indicated by UV254 and DOC) removal rate also increased as zeta potential became even closer to 0 mV; then when zeta potential became higher than 0 mV, which means that as more DAM–ECH was added, colloids in raw water would become positively charged and inner repulsive force between colloids would interfere the destabilization of particles. Thus, the growth of NOM removal did not kept increase as growth of DAM–ECH dose, instead, it slowed down and became negligible eventually. Combining consider results in Fig. 2 B–D, a conclusion could be drawn that zeta potential was always around zero when of UV254 and DOC removal grew up to certain points and trended to be stopped, which is consist with the results of other researchers that the optimum coagulation performance always occur when zeta potential was around 0 mv [24]. So considering both the cost in wastewater treatment and coagulation performance, 10 mg/L of FC with FC and DAM–ECH mass ratio 3:1 were chosen to be the optimum dosage for the following tests in this study. 3.1.2. Effect of pH variation on coagulation performance The initial raw water pH was adjusted in range between 4.0 and 10.0 to determine the optimum condition for coagulation performance of FC/DAM–ECH [25,26]. Fig. 3 reveals turbidity, UV254 and DOC removal of chosen FC/DAM–ECH dose (FC 10 mg/L, FC/ DAM–ECH mass ratio 3:1) in selected pH range. It is obvious that turbidity, UV254 and DOC removal reached high points at pH 6. This could be explained as follow: pH determines the surface charge of HA molecule and significantly affect the hydrolysis of iron salts, which in this case was FC. Under alkaline conditions functional groups of HA such as carboxylic or phenolic acids would be deprotonated, which would increase negative charge of HA, so that more FC would be consumed to neutralize the increased negative charges. The hydrolysis process of FC was essentially determined by the concentration of H+ (solution pH) [27,28], which has been proved by many researchers and the process can be revealed conveniently by the oversimplified equation below:
Fe3þ ! FeðOHÞ2þ ! FeðOHÞþ2 ! FeðOHÞ3 ! FeðOHÞ4 :
Fig. 3. Turbidity, UV254 and DOC removal efficiencies as a function of pH under optimum coagulation method.
Many trimeric and polynuclear hydrolysis products of Fe can would be formed during the actual process, but they could not greatly affect the overall metal speciation in diluted solution so that can be ignored [27,29]. Hydrolysis of Fe as shown above would proceed from left to right as the pH increase. To be more specific, under certain acidic condition (pH range 5–7), more H+ in the solution would ensure there are more cationic hydrolysis products (eg. Fe(OH)2+, FeðOHÞþ 2 ) in the solution instead of FeðOHÞ4 . In this way, under acidic condition, charge neutralization between HA molecules (with less negative charges) and FC hydrolysates (more positively charged) would be enhanced. Thus, the demand of FC for would be reduced to achieve the same coagulation effect [30– 32]. Accordingly, pH 6 was determined to be the optimum pH condition for FC/DAM–ECH coagulation in this study, and its membrane performance in the follow up UF process would also be investigated in comparing with other coagulation methods. 3.2. Floc properties Floc sizes, breakage/recovery factors and structure were investigated as parameters to further analyze the membrane fouling in UF process. All parameters were generated from Fig. 4, of which each picture can be divided into 3 stages, formation period (0–12.5 min), followed by 5 min of breakage stage and the regrowth period lasted for 10 min. 3.2.1. Floc sizes under various coagulation and hydraulic conditions As revealed in Fig. 4, in general, formation of flocs started right after the coagulants added into raw water, the size increased dramatically for in the first 5 min; then slowdown and trend to be steady in the following 7.5 min. Afterwards, when shearing force were introduced, floc sizes drop sharply especially when rotate speed was high up to 400 rpm. Finally, during the last 10 min of regrowth period, flocs restored to certain sizes but far below their original sizes in steady stage of formation period. Fig. 5 is introduced to show the floc sizes in all stages more accessible, in which is obvious floc sizes dropped more dramatically as higher stirring speed was applied. Also, the lower shearing forces, the larger flocs would be obtained in the regrowth period; when rotation speed was high up to 400 rpm, the regenerated flocs under original pH and pH 6 were almost with the same size but still a little higher than FC flocs, which means that shearing force would be the crucial impact that influences floc size. It can also be observed form Fig. 5 that addition of DAM–ECH led to larger floc size which high up around 600 lm; meanwhile when FC was applied as the only coagulant, floc size was only around 400 lm. Also sizes of regenerated floc by FC/DAM–ECH were much larger than that of FC, especially when low shearing forces (100 & 200 rpm) were introduced. Then when pH adjusted to 6, FC/ DAM–ECH floc grew even larger than under original pH which climbed over 680 lm in the formation period. Phenomena above about floc size could be explained as follow: according to previous researchers, flocs formed under acidic condition would be larger than that formed in alkaline region, which was because that more colloid particles and microflocs would be bridged by coagulant hydrolyzates under acidic condition and form larger flocs [33]. In conclusion, addition of DAM–ECH would apparently increase floc sizes, also larger flocs would be generated under acidic condition than under raw water pH. 3.2.2. Floc strength and recoverability As mentioned in Section 3.2.1, higher shearing force would produce smaller flocs in breakage period and meanwhile generated smaller flocs after regrowth period. Also larger flocs formed in growth period, the sharper drop of floc sizes would occur at the
S. Sun et al. / Chemical Engineering Journal 280 (2015) 634–642
A
B
c
Fig. 4. Flocs diameters on medium as a function of DAM–ECH dosage and pH in formation, breakage and regrowth stages.
Fig. 5. Flocs sizes in stable stage of formation and breakage periods with different coagulation conditions.
639
640
S. Sun et al. / Chemical Engineering Journal 280 (2015) 634–642
Table 1 Effect of shearing force on flocs strength factors under different coagulation conditions. Coagulation conditions
FC (raw water pH) FC/DAM–ECH (raw water pH) FC/DAM–ECH (pH 6)
Stirring speed (rpm) 100
200
400
57.24 49.92 48.71
25.38 24.28 23.60
15.05 15.01 13.82
Table 2 Effect of shearing force on flocs recovery factors under different coagulation conditions. Coagulation conditions
FC (raw water pH) FC/DAM–ECH (raw water pH) FC/DAM–ECH (pH 6)
Stirring speed (rpm) 100
200
400
28.24 31.52 34.81
26.17 31.72 37.61
20.21 21.56 22.03
beginning of breakage period, suggesting that larger flocs were easier to be broken [19]. Floc strength and recoverability were calculated by Eqs. (1) and (2), which were stated in Section 2.4.2. The strength factors were shown in Table 1, which is obvious that larger flocs formed by FC/DAM–ECH under pH 6 and original pH were more frangible than smaller flocs produced by FC alone. This result consist with the previous studies that flocs in larger sizes would be easier broken when shearing force were introduced [19]. Li et al. suggested that floc strength was of the following order: bridging > charge neutralization > sweep [34], which means the addition of DAM–ECH changed the main coagulation mechanism from bridging to charge neutralization.
A
Table 2 shows that addition of DAM–ECH would not only tremendously increase the floc sizes formed by FC but also increase their recovery ability against the shearing force, especially when pH was adjusted into weak acidic range (pH 6). It is noticeable that when the rotation speed was set to be 200 rpm, flocs formed by FC/ DAM–ECH had much better recovery ability under original or acidic condition. This was because that highly positively charged DAM–ECH enhanced the charge neutralization of FC with HA molecules, and flocs formed by charge neutralization have better recovery ability than the flocs formed by sweep enmeshment [35,36]. In conclusion, flocs formed by FC/DAM–ECH were much larger than FC–HA flocs, especially under acidic region. Though larger flocs were easier to be broken, the FC/DAM–ECH flocs had better recovery ability which would make sure to produce larger flocs even under high shearing force.
3.2.3. Floc structure Fractal dimension value (Df) of floc would be used in demonstration of floc structure in this study. As common knowledge, higher Df means denser structure of floc [37–39]. As shown in Fig. 6, it can be observed that FC flocs (A) had most compact structure and largest Df value and formed cake layer with smallest pore sizes; meanwhile FC/DAM–ECH flocs (C) with lowest Df value formed loosest cake layer on membrane surface with biggest pore sizes. Also for each coagulation method, Df reached the highest point during the breakage period which revealed that floc size dropped to the low point and their values were in the following order: FC (2.8567) > FC/DAM–ECH (2.836) > FC/DAM–ECH pH 6 (2.6923); then Df dropped to certain level but still higher than the formation stage, which also consist with the change of floc size that regenerated flocs would not reach the floc size of formation period.
B
C Fig. 6. SEM images of flocs formed by different coagulation methods on membrane surface at 30,000 times magnification and their fractal dimensions: (A) FC (2.481); (B) FC/DAM–ECH (2.352); (C) FC/DAM–ECH pH 6 (2.271).
S. Sun et al. / Chemical Engineering Journal 280 (2015) 634–642
641
(2) Initial pH of raw water was proved to be a crucial parameter which influenced the turbidity and HA removal rates. The optimum pH condition for coagulation was around weak acidic condition, which in this study was pH 6. (3) Introduce of DAM–ECH influenced the floc properties by enlarge the floc size and lower Df. Furthermore, flocs formed under pH 6 were much bigger than those at original raw water pH and more fragile. Meanwhile recovery ability of flocs formed by FC/DAM–ECH was much better than by FC, especially when neutralization was enhanced as raw water pH was adjusted to 6. (4) FC combined use with DAM–ECH would improve the membrane performance in the following UF process. Adjustment raw water pH to 6 further reduce membrane fouling of FC/ DAM–ECH coagulated water.
Acknowledgment Fig. 7. Normalized UF permeate flux profiles for pre-coagulated HA suspension by different coagulation methods.
The project was financially supported by the National Natural Science Foundation of China (21377071).
3.3. Assessment of membrane fouling in ultrafiltration References UF experiments were conducted as post-treatment of coagulation jar test to investigate the effect of FC/DAM–ECH dose and pH on membrane fouling. The MFI and trend of permeate flux decline caused by different coagulation methods and are shown in Fig. 7. Generally, sharp drop of permeate fluxes occurred right after the filtration started under all water conditions; then the decrease trend to be gentler as filtration proceed and finally became steady after certain filtration period. It is obvious that the permeate flux influenced remarkably by different coagulation conditions, and the final permeate fluxes with coagulation processes were all higher than that of raw water, which has been proved by other researchers [10]. The sequence of MFI for different coagulation methods was as follows: FC/DAM–ECH pH 6 (28.43) < FC/DAM– ECH (76.79) < FC (96.91) < Raw water (128.77). The permeate flux of FC/DAM–ECH coagulated water at pH 6 dropped much slower than raw water and the other coagulation methods, meanwhile its final flux also appears to be much higher which was consist with MFI values. Coagulated water by FC/DAM–ECH also performed lighter membrane fouling comparing with raw water and FC coagulated water. Previous studies suggested membrane fouling is markedly influenced by floc properties, especially cake layer which mainly causes the revisable fouling [15,40,41]; moreover, increase of floc size and decrease of Df would reduce the membrane specific membrane resistance [42]. According to Fig. 4C, the average D0.5 of regenerated FC/DAM–ECH–HA flocs formed under pH 6 after broken by 200 rpm of shearing force was 225.05 lm, which was much higher than that of FC–HA flocs (139.28 lm) and FC/DAM–ECH–HA flocs (182.86 lm). Also considering that stated in Section 3.2.3, FC/DAM–ECH flocs formed under pH 6 had the lowest Df, which is consist with the membrane performance that the optimum membrane performance was achieved by FC/DAM–ECH at pH 6. As for FC/DAM–ECH coagulated water under original raw water pH, flocs had medium size and medium Df, so it achieved medium final permeate flux among all three coagulated water. 4. Conclusions (1) Addition of DAM–ECH with proper dosage would improve the turbidity and HA removal efficiencies by enhancing neutralization mechanism of coagulation process.
[1] X. Guo, Q. Li, W. Hu, W. Gao, D. Liu, Ultrafiltration of dissolved organic matter in surface water by a polyvinylchloride hollow fiber membrane, J. Membr. Sci. 327 (2009) 254–263. [2] D.R. Simpson, Biofilm processes in biologically active carbon water purification, Water Res. 42 (2008) 2839–2848. [3] M. Rossini, J.G. Garrido, M. Galluzzo, Optimization of the coagulation– flocculation treatment: influence of rapid mix parameters, Water Res. 33 (1999) 1817–1826. [4] W. Gan, Q. Liu, Coagulation of bitumen with kaolinite in aqueous solutions containing Ca2+, Mg2+ and Fe3+: effect of citric acid, J. Colloid Interface Sci. 324 (2008) 85–91. [5] Z. Yang, X. Liu, B. Gao, S. Zhao, Y. Wang, Q. Yue, Q. Li, Flocculation kinetics and floc characteristics of dye wastewater by polyferric chloride–polyepichlorohydrin–dimethylamine composite flocculant, Sep. Purif. Technol. 118 (2013) 583–590. [6] E.E. Chang, P.C. Chiang, S.H. Chao, C.H. Liang, Effects of polydiallyldimethyl ammonium chloride coagulant on formation of chlorinated by products in drinking water, Chemosphere 39 (1999) 1333–1346. [7] J.M. Laîné, D. Vial, P. Moulart, Status after 10 years of operation — Overview of UF technology today, Desalination 131 (2000) 17–25. [8] W. Yu, N.J.D. Graham, Application of Fe(II)/K2MnO4 as a pre-treatment for controlling UF membrane fouling in drinking water treatment, J. Membr. Sci. 473 (2015) 283–291. [9] W. Yuan, A.L. Zydney, Humic acid fouling during ultrafiltration, Environ. Sci. Technol. 34 (2000) 5043–5050. [10] J. Liu, B. Liu, T. Liu, Y. Bai, S. Yu, Coagulation–bubbling–ultrafiltration: effect of floc properties on the performance of the hybrid process, Desalination 333 (2014) 126–133. [11] W. Xu, Q. Yue, B. Gao, B. Du, Impacts of organic coagulant aid on purification performance and membrane fouling of coagulation/ultrafiltration hybrid process with different Al-based coagulants, Desalination 363 (2015) 126–133. [12] S. Zhao, B. Gao, Q. Yue, W. Song, R. Jia, P. Liu, Evaluation of floc properties and membrane fouling in coagulation–ultrafiltration system: the role of enteromorpha polysaccharides, Desalination 367 (2015) 126–133. [13] A.W. Zularisam, A.F. Ismail, M.R. Salim, M. Sakinah, T. Matsuura, Application of coagulation–ultrafiltration hybrid process for drinking water treatment: optimization of operating conditions using experimental design, Sep. Purif. Technol. 65 (2009) 193–210. [14] K. Chon, J. Cho, H.K. Shon, A pilot-scale hybrid municipal wastewater reclamation system using combined coagulation and disk filtration, ultrafiltration, and reverse osmosis: removal of nutrients and micropollutants, and characterization of membrane foulants, Bioresour. Technol. 141 (2013) 109–116. [15] W.-Z. Yu, N. Graham, H.-J. Liu, J.-H. Qu, Comparison of FeCl3 and alum pretreatment on UF membrane fouling, Chem. Eng. J. 234 (2013) 158–165. [16] J.K. Edzwald, Coagulation in drinking water treatment: particles, organics and coagulants, Water Sci. Technol. 27 (1993) 21–35. [17] P. Jarvis, B. Jefferson, J. Gregory, S.A. Parsons, A review of floc strength and breakage, Water Res. 39 (2005) 3121–3137. [18] M.A. Yukselen, J. Gregory, The reversibility of floc breakage, Int. J. Miner. Process. 73 (2004) 251–259. [19] J.C. Wei, B.Y. Gao, Q.Y. Yue, Y. Wang, Strength and regrowth properties of polyferric-polymer dual-coagulant flocs in surface water treatment, J. Hazard. Mater. 175 (2010) 949–954.
642
S. Sun et al. / Chemical Engineering Journal 280 (2015) 634–642
[20] R.J. François, Strength of aluminium hydroxide flocs, Water Res. 21 (1987) 1023–1030. [21] S.F.E. Boerlage, M.D. Kennedy, M.P. Aniye, E. Abogrean, Z.S. Tarawneh, J.C. Schippers, The MFI-UF as a water quality test and monitor, J. Membr. Sci. 211 (2003) 271–289. [22] S.F.E. Boerlage, M.D. Kennedy, M.R. Dickson, D.E.Y. El-Hodali, J.C. Schippers, The modified fouling index using ultrafiltration membranes (MFI-UF): characterisation, filtration mechanisms and proposed reference membrane, J. Membr. Sci. 197 (2002) 1–21. [23] R. Mao, Y. Wang, B. Zhang, W. Xu, M. Dong, B. Gao, Impact of enhanced coagulation ways on flocs properties and membrane fouling: increasing dosage and applying new composite coagulant, Desalination 314 (2013) 161–168. [24] B. Bolto, D. Dixon, R. Eldridge, S. King, Cationic polymer and clay or metal oxide combinations for natural organic matter removal, Water Res. 35 (2001) 2669– 2676. [25] Q. Wang, B. Gao, Y. Wang, Z. Yang, W. Xu, Q. Yue, Effect of pH on humic acid removal performance in coagulation–ultrafiltration process and the subsequent effects on chlorine decay, Sep. Purif. Technol. 80 (2011) 549–555. [26] Z.L. Yang, B.Y. Gao, Q.Y. Yue, Y. Wang, Effect of pH on the coagulation performance of Al-based coagulants and residual aluminum speciation during the treatment of humic acid–kaolin synthetic water, J. Hazard. Mater. 178 (2010) 596–603. [27] C.M. Flynn, Hydrolysis of inorganic iron(III) salts, Chem. Rev. 84 (1984) 31–41. [28] A.J. AB Lamb, The slow hydrolysis of ferric chloride in dilute solution, J. Am. Chem. Soc. (1938). [29] J. Duan, J. Gregory, Coagulation by hydrolysing metal salts, Adv. Colloid Interface Sci. 100–102 (2003) 475–502. [30] X. Bo, B. Gao, N. Peng, Y. Wang, Q. Yue, Y. Zhao, Effect of dosing sequence and solution pH on floc properties of the compound bioflocculant–aluminum sulfate dual-coagulant in kaolin–humic acid solution treatment, Bioresour. Technol. 113 (2012) 89–96. [31] Y.L. Wang, J. Feng, S.K. Dentel, J. Lu, B.Y. Shi, D.S. Wang, Effect of polyferric chloride (PFC) doses and pH on the fractal characteristics of PFC–HA flocs, Colloids Surf. A 379 (2011) 51–61.
[32] A. Adin, Y. Soffer, R. Ben Aim, Effluent pretreatment by iron coagulation applying various dose-pH combinations for optimum particle separation, Water Sci. Technol. 38 (1998) 27–34. [33] B. Cao, B. Gao, X. Liu, M. Wang, Z. Yang, Q. Yue, The impact of pH on floc structure characteristic of polyferric chloride in a low DOC and high alkalinity surface water treatment, Water Res. 45 (2011) 6181–6188. [34] T. Li, Z. Zhu, D. Wang, C. Yao, H. Tang, Characterization of floc size, strength and structure under various coagulation mechanisms, Powder Technol. 168 (2006) 104–110. [35] V. Chaignon, B.S. Lartiges, A. El Samrani, C. Mustin, Evolution of size distribution and transfer of mineral particles between flocs in activated sludges: an insight into floc exchange dynamics, Water Res. 36 (2002) 676– 684. [36] K. McCurdy, K. Carlson, D. Gregory, Floc morphology and cyclic shearing recovery: comparison of alum and polyaluminum chloride coagulants, Water Res. 38 (2004) 486–494. [37] T. Serra, J. Colomer, B.E. Logan, Efficiency of different shear devices on flocculation, Water Res. 42 (2008) 1113–1121. [38] X.-Y. Li, B.E. Logan, Permeability of fractal aggregates, Water Res. 35 (2001) 3373–3380. [39] J. Gregory, The density of particle aggregates, Water Sci. Technol. 36 (1997) 1– 13. [40] T.D. Waite, A.I. Schäfer, A.G. Fane, A. Heuer, Colloidal fouling of ultrafiltration membranes: impact of aggregate structure and size, J. Colloid Interface Sci. 212 (1999) 264–274. [41] K. Listiarini, D.D. Sun, J.O. Leckie, Organic fouling of nanofiltration membranes: Evaluating the effects of humic acid, calcium, alum coagulant and their combinations on the specific cake resistance, J. Membr. Sci. 332 (2009) 56–62. [42] S.A. Lee, A.G. Fane, T.D. Waite, Impact of natural organic matter on floc size and structure effects in membrane filtration, Environ. Sci. Technol. 39 (2005) 6477–6486.
Contents lists available at ScienceDirect
Chemical Engineering Journal journal homepage: www.elsevier.com/locate/cej
Effects of epichlorohydrin–dimethylamine on coagulation and membrane performance of ferric chloride in coagulation–ultrafiltration hybrid process Shenglei Sun, Fan Bu, Xin Huang, Shuang Zhao, Hongyu Dong, Baoyu Gao ⇑, Qinyan Yue ⇑, Lijuan Feng, Yan Wang, Qian Li Shandong Key Laboratory of Water Pollution Control and Resource Reuse, School of Environmental Science and Engineering, Shandong University, Ji’ nan 250100, China
h i g h l i g h t s Coagulation efficiency of FC on HA–Kaolin could be enhanced apparently by DAM–ECH. Initial pH would remarkably influence the coagulation performance and floc properties. Application of DAM–ECH result in larger flocs with better recovery ability. DAM–ECH addition would generate flocs with loose and open structure. Membrane performance was significantly improved by DAM–ECH.
a r t i c l e
i n f o
Article history: Received 12 March 2015 Received in revised form 3 June 2015 Accepted 4 June 2015 Available online 18 June 2015 Keywords: Membrane fouling Epichlorohydrin–dimethylamine Ferric chloride Floc properties Coagulation–ultrafiltration
a b s t r a c t Cationic polymer epichlorohydrin–dimethylamine (DAM–ECH) was applied as coagulation aid of ferric chloride (FC) in treatment of humic acid–kaolin (HA–Kaolin) simulated water in this study. Impacts of DAM–ECH on coagulation performance, flocs characteristics and membrane fouling of FC in coagulation–ultrafiltration (C–UF) process were investigated under different pH conditions. Results showed that the removal rates of UV254 (ultraviolet adsorption at 254 nm) and DOC (dissolved organic carbon) of FC/DAM–ECH was much better than that of FC at same dosage. The optimum coagulation condition was FC at 10 mg/L with FC/DAM–ECH mass ratio (MR) of 3:1 based on the consideration of treatment cost and performance. Moreover, turbidity, UV254 and DOC removal rates increased when pH increased from 4 to 6 and then decreased as pH grew from 7 to 10. Thus, the optimum pH for coagulation was also determined to be 6 accordingly. Addition of DAM–ECH also increased sizes of produced flocs and made them looser in structure comparing with FC–HA flocs. Meanwhile, results obtained from ultrafiltration experiments showed that membrane fouling would be reduced by introducing DAM–ECH in coagulation treatment; the optimum membrane performance was given by FC/DAM–ECH under pH 6. In conclusion, proper dose of FC/DAM–ECH under pH 6 could ensure an ideal HA removal rate and meanwhile increase operating duration of C–UF system by restricting membrane fouling. Ó 2015 Elsevier B.V. All rights reserved.
1. Introduction It is now widely recognized that as a kind of natural organic matter (NOM) which commonly existing in most of drinking water resources, humic substances are jeopardizing human health [1,2]. Thus, removal of humic substances has drawn worldwide attention
⇑ Corresponding authors. Tel.: +86 531 88366771; fax: +86 531 88364513 (B. Gao). Tel.: +86 531 88365258; fax: +86 531 88364513 (Q. Yue). E-mail addresses: [email protected] (B. Gao), [email protected] (Q. Yue). http://dx.doi.org/10.1016/j.cej.2015.06.011 1385-8947/Ó 2015 Elsevier B.V. All rights reserved.
recent decades in modern water treatment. As an extensively used coagulant in water treatment processes, ferric chloride (FC) can remove humic acid molecules in water bodies by forming strong and large flocs [3–5]. Moreover, many researchers have proved that compounding of iron salt flocculants with organic polymer would be more effective in HA and turbidity removal than single use of iron salts [6]. Thus, a cationic polymer epichlorohydrin–dimethylamine (DAM–ECH) was applied as coagulation aid of ferric chloride (FC) in this study. Meanwhile as a membrane filtration process, ultrafiltration has been broadly applied as postprocessing of water treatment process to further improve the reduction of
S. Sun et al. / Chemical Engineering Journal 280 (2015) 634–642
particle concentration and NOM in portable water resources. This is because that UF processes could not remove most of dissolved organic matter from water bodies because of its large pore sizes [7]. Another main bottle-neck in ultrafiltration utility is the drop of permeate flux caused by membrane fouling, which would hugely increase the power demand to maintain an acceptable flow rate of the UF device [8,9]. It has been proved that pretreatment (such as adsorption, coagulation or peroxidation) can remarkably reduce improve membrane performance by achieving larger permeate flux in combined water treatment systems, in which coagulation has been proved to be more effective than the other pretreatments in reduction of membrane fouling [10]. Breakthroughs have been made by previous researches in all aspect of C–UF system, including influence of coagulants/flocculants by Xu et al. and Zhao et al. [11,12]; optimization of operating conditions in application of C–UF in drinking water treatment and pilot-scale hybrid municipal wastewater reclamation [13,14]. Therefore, coagulation–ultrafiltration hybrid process (C–UF) is becoming more and more popular in high quality water treatment processes for its high removal efficiency of NOM. Generally, C–UF is significantly affected by many elements such as feed water types, coagulation methods and hydraulic conditions (shearing force) [15]. As mentioned above, combination of coagulation and ultrafiltration process is mutually beneficial which can further improve the water quality comparing with single use of each process. Furthermore, combine use of these two processes should be optimize through designed experiment processes. Thus, proper addition method of coagulants was studied to achieve acceptable coagulation and membrane performance in this study.
635
mixture was transferred into a 1000 mL volumetric flask and diluted to volume (1 L). Kaolin stock solution was applied in this study to adjust turbidity of HA simulated water, which was prepared as follow: 5.0 g of kaolin (Tianjin Kermel Chemical Reagent Co., Ltd, China) was weighted and mixed with 1.0 L of deionized water, the suspension liquid were stirred by a magnetic stirring apparatus for at least 30 min; then the upper 500 mL liquid was withdrawn for later use. HA–Kaolin simulated water of this study was prepared by diluting 10 mL of the HA stock solution into 1.0 L of tap water. Initial turbidity of the raw water was adjusted to 15.0 ± 0.5 NTU using kaolin stock solution. Characteristics of raw water were as follow: UV254 (ultraviolet absorbance at 254 nm) = 0.320 ± 0.010.
DOCðdissolved organic matterðDOMÞ concentrationÞ ¼ 3:55 0:03 mg=L Raw water pH ¼ 8:32 0:02 Many researchers have found that UV254 can be used as indicator of humic substances and aromatic organic matter contain C@C and C@O in water bodies, that higher humic concentration in water always show high UV254 adsorption [13,16]. Furthermore, the relationship between UV254 and DOC have been studied by many researchers, results showed that 70% to 80% of DOC removal was aquatic humic fraction with high UV254 absorption value, only 10% of DOC was hydrophobic organic with low UV254 adsorption. In other words, UV254 removal should be positive correlated with DOC removal [16]. The initial pH of the raw water sample was adjusted by 0.1 M HCl and 0.1 M NaOH solution when necessary during the tests.
2. Materials and methods
2.3. Jar test
2.1. Preparation of coagulants
In this study, jar tests were conducted to determine the optimum coagulation method for HA–Kaolin water, which were carried out by a program-controlled flocculator with six 1.5 L cylindrical plexiglass beakers and six flat paddle impellers (ZR4-6, Zhongrun Water Industry Technology Development Co. Ltd., China). During jar tests, 6 L of prepared HA–Kaolin solution was filled into six breakers averagely (1 L of raw water in each breaker). And the operation program was preseted as follow: first, synthetic water sample in each breakers was stirred rapidly under 200 rpm (rpm) for 0.5 min; then certain dose of coagulant was added into each breakers, followed by another stirring period of 200 rpm for 0.5 min; following the third period, the mixture in each breakers would be stirred under 200 rpm for 1.5 min for dispersion of coagulant; in the fourth stage, floc would grow under slow stirring speed of 40 rpm for 10 min; during the final stage, the shearing force would be removed for flocs to settle for 20 min. After the sedimentation period, 200 ml of supernatant water was withdrawn from each jar to use as water samples in the following tests (Turbidity, DOC, UF, etc.). Turbidity of water sample would be measured by a 2100P turbidimeter (Hach, USA) and zeta potential was measured using a Zetasizer 3000Hsa (Malvern Instruments, UK) at room temperature (around 25 °C). Each water sample was filtered through a 0.45 lm fiber membrane before DOC analysis using a Shimadzu TOC-VCPH analyzer and UV254 measurement by a UV-754 UV/VIS spectrophotometer (Precision Scientific Instrument Co. Ltd., Shanghai, China).
FC solution with concentration of 2 g/L (as Fe3+) was prepared by an analytical reagent FeCl3 6H2O (CAS No: 10025-77-1). 0.9661 g of FeCl3 6H2O was completely dissolved in 100 ml of deionized water stirred by a magnetic stirring apparatus until the solution became clarified. DAM–ECH copolymer was polymerized by analytical reagent (A.R.) epichlorohydrin and dimethylamine (A.R.) with ethanediamine (A.R.) as cross-linker. The polycondensation reaction that producing DAM–ECH copolymer was conducted in a 250 mL of four-neck round bottom flask. In addition, several facilities were required during the synthesis process: a mechanical stirrer, a thermometer, a dropping funnel and corresponding glass spigots. Certain dosage of epichlorohydrin was poured into the four-neck round bottom flask and meanwhile the environmental temperature within the flask was kept at 10 °C by a thermostated water bath device. Then dimethylamine was constantly dripped into the flask using a dropping funnel, and the mixture was under steady stirring throughout the whole dripping process. At last, certain dose of ethanediamine was introduced into flask under continuous stirring as the cross-linker [5].
2.2. Simulated water sample Humic acid–kaolin (HA–Kaolin) water was used as raw water in this study. The HA–Kaolin stock solution was prepared as follow: 1.0 g of HA powder (Aladdin, Shanghai, China) and 0.40 g of NaOH (Tianjin Damao Co., Tianjin, China) were weighted and dissolved in deionized water under continuously stirring of a magnetic stirring apparatus till completely dissolved; then the
2.4. On-line monitoring of floc formation, breakage and re-growth The evolution of flocs during coagulation process was monitored by a laser diffraction instrument (Malvern Mastersizer 2000, Malvern, UK). Water samples were monitored through the
636
S. Sun et al. / Chemical Engineering Journal 280 (2015) 634–642
optical unit of the Malvern Mastersizer 2000 where a circulation was formed by drawing water from and back into the jar using a peristaltic pump on a 5 mm internal diameter return tube at a flow rate of 2.0 liter per hour (L/h). Malvern Mastersizer 2000 was connected to a computer to collect data every 30 s during the whole process including floc formation, breakage and regrowth. The coagulation program was set to be the same as jar test mentioned in Section 2.3, except after the slow stirring period for floc growth, the flocs were exposed to shearing forces (100 rpm, 200 rpm, and 400 rpm) for 5 min; at last there would be a 15 min of floc regrowth period. 2.4.1. Floc size In this study, the median volumetric diameter (d50) was used to denote the floc size. The floc size distribution was also investigated by Matersizer 2000. 2.4.2. Floc breakage factor and recovery factor Flocs breakage and recovery factors are normally used to reveal the breakage and re-growth of flocs, in which breakage factor indicates the ability of flocs to resist the shear power and recovery factor shows the capacity of flocs after breakage period. The breakage and recovery factor of flocs can be calculated by the following equations which have been used in other studies [17–20]:
Breakage factor ð%Þ ¼
d2 100 d1
ð1Þ
Recovery factor ð%Þ ¼
d3 d2 100 d1 d2
ð2Þ
2.4.3. Floc dimension Different coagulation conditions may produce different evolution of flocs structure. Fractal dimension is one commonly used parameter to characterize flocs structure. Fractal dimension of flocs was measured by Laser Light Scattering in this study. Smaller fractal dimension means the more open or irregular flocs structure and lower effective density. In a conventional coagulation process for water treatment, the flocs with high density are preferable because of their better settleability in the subsequent settling unit. However, it may not be the case in the coagulation–membrane processes. Fractal dimension can be calculated by the formula:
ð3Þ
where Df is the mass fractal dimension, which can be determined by the slope of a plot of I as a function of Q on a log–log scale, while I is the light density and Q is the difference between the incident and scattered wave vectors of the radiation beam in the medium, which is given by equation:
Q¼
4p n sinðh=2Þ k
The UF membranes used in this study were purchased from Shanghai Mosu Science Equipment Co., Ltd and the molecular weight cut-off (MWCO) of membrane was 100 K Dalton (kDa). UF experiment was conducted within a magnetically stirred cell (MSC050, Mosu, China), which had a total volume of 300 mL and an effective membrane area of 50.2 cm2. Pre-coagulated water samples were added directly into the magnetically stirred ultrafiltration batch cell of for ultrafiltration, and the coagulation program was set to be the same as that of the jar test but without sedimentation. The magnetically stirred cell was pressurized with nitrogen gas (N2) provide by a high pressure gas tank with constant pressure of 0.15 ± 0.05 MPa. An electronic balance (MSU5201S-000-D0, SARTORIUS AG GERMANY) was applied to measure permeate mass. Meanwhile, the balance was connected with a computer to record permeate mass value on the balance per 10 s. The detailed hybrid C–UF system is stated in Fig. 1, which revealed that coagulated water was drawn from the coagulation breakers and then the suspension liquid was filtrated through the ultrafiltration membrane under a constant pressure provided by N2. The permeate flux was monitored by a computer which connected with an electronic balance. Modified fouling index (MFI) was designed to measure the fouling potential of micro-filtration membrane and ultrafiltration membrane, and details can be found in previous study [21,22]:
t gRm gaC b ¼ þ V V DPA 2DPA2 |fflfflffl{zfflfflffl}
ð5Þ
MFI
d1, d2, d3 stand for flocs sizes in the steady phase before breakage, after the breakage period and after the re-growth to another steady phase, respectively. Larger breakage factors means the flocs are stronger and more difficult to be broken, meanwhile larger recovery factor expresses better regrowth ability of flocs.
I / Q Df
2.5. UF apparatus and C–UF hybrid process
ð4Þ
where n is the refractive index of the suspending medium, h is the scattering angle, and k is the wavelength of the radiation in a vacuum. In this study, five points (D0.5 value) in steady stage of formation, breakage and regrowth period were chosen and their average value were used for calculation of strength and recovery factors.
where V is the filtrate volume, t is the filtration time, DP is the applied trans-membrane pressure, Rm is the resistance of the membrane, A is available membrane area, g is the water viscosity, a is the specific resistance of the cake deposited and Cb is the concentration of particles in feed water. The MFI is determined from the gradient of the linear portion of t/V versus V. A higher MFI value means more serious membrane fouling [23]. 3. Results and discussion 3.1. Coagulation performance As mentioned in Section 2, jar tests were conducted to investigate the coagulation performance of FC and DAM–ECH with different coagulation methods. Turbidity, UV254, and DOC were used as main indicators to determine the coagulation performance and zeta potential was also monitored to help analysis the phenomena. 3.1.1. Effects of different FC dosage and FC/DAM–ECH MR on HA– Kaolin water treatment As shown in Fig. 2 A, it is obvious that at removal rate of turbidity grew dramatically as FC dosage increased from 2 mg/L to 4 mg/ L, no matter with or without DAM–ECH; then the growth became gentle when 6 mg/L of FC was added and finally almost remain steady as FC dosage increased to 12 mg/L. It is noticeable that turbidity removal rates were enhanced as DAM–ECH was introduced even at FC dosage of 2 mg/L. According to Fig. 2D, zeta potential did not changed much at that point and kept far below 10 mV, which means sweep and adsorption should take the main part of coagulation process [5]. This result revealed that DAM–ECH could effectively increase turbidity removal effect of FC under low dosage. Meanwhile UV254 and DOC removal by FC trended to be similar to each other, which can be concluded that the removal rates kept increase as FC dosage increased from 2 mg/L to 8 mg/L; then the growth slowdown in range of 8 mg/L to 10 mg/L and became negligible when FC dosage further increased to 12 mg/L. Furthermore,
S. Sun et al. / Chemical Engineering Journal 280 (2015) 634–642
Fig. 1. Schematic of coagulation–ultrafiltration hybrid experiment.
A
B
C
D
Fig. 2. Turbidity, UV254 and DOC removal rates and zeta potential as functions of FC/DAM–ECH dosages under different coagulation methods.
637
638
S. Sun et al. / Chemical Engineering Journal 280 (2015) 634–642
inconsideration of DAM–ECH dosage, it can be seen that increase of UV254 and DOC removal rate became insignificant when FC/DAM– ECH mass ratio was 3:1 at FC dosage of 10 mg/L. This can be explained as follow: Zeta potential almost remained unchanged when FC was used alone (Fig. 2D). But as high positively charged DAM–ECH was added, the charge neutralization mechanism of FC with HA molecule was dramatically enhanced, so that NOM (indicated by UV254 and DOC) removal rate also increased as zeta potential became even closer to 0 mV; then when zeta potential became higher than 0 mV, which means that as more DAM–ECH was added, colloids in raw water would become positively charged and inner repulsive force between colloids would interfere the destabilization of particles. Thus, the growth of NOM removal did not kept increase as growth of DAM–ECH dose, instead, it slowed down and became negligible eventually. Combining consider results in Fig. 2 B–D, a conclusion could be drawn that zeta potential was always around zero when of UV254 and DOC removal grew up to certain points and trended to be stopped, which is consist with the results of other researchers that the optimum coagulation performance always occur when zeta potential was around 0 mv [24]. So considering both the cost in wastewater treatment and coagulation performance, 10 mg/L of FC with FC and DAM–ECH mass ratio 3:1 were chosen to be the optimum dosage for the following tests in this study. 3.1.2. Effect of pH variation on coagulation performance The initial raw water pH was adjusted in range between 4.0 and 10.0 to determine the optimum condition for coagulation performance of FC/DAM–ECH [25,26]. Fig. 3 reveals turbidity, UV254 and DOC removal of chosen FC/DAM–ECH dose (FC 10 mg/L, FC/ DAM–ECH mass ratio 3:1) in selected pH range. It is obvious that turbidity, UV254 and DOC removal reached high points at pH 6. This could be explained as follow: pH determines the surface charge of HA molecule and significantly affect the hydrolysis of iron salts, which in this case was FC. Under alkaline conditions functional groups of HA such as carboxylic or phenolic acids would be deprotonated, which would increase negative charge of HA, so that more FC would be consumed to neutralize the increased negative charges. The hydrolysis process of FC was essentially determined by the concentration of H+ (solution pH) [27,28], which has been proved by many researchers and the process can be revealed conveniently by the oversimplified equation below:
Fe3þ ! FeðOHÞ2þ ! FeðOHÞþ2 ! FeðOHÞ3 ! FeðOHÞ4 :
Fig. 3. Turbidity, UV254 and DOC removal efficiencies as a function of pH under optimum coagulation method.
Many trimeric and polynuclear hydrolysis products of Fe can would be formed during the actual process, but they could not greatly affect the overall metal speciation in diluted solution so that can be ignored [27,29]. Hydrolysis of Fe as shown above would proceed from left to right as the pH increase. To be more specific, under certain acidic condition (pH range 5–7), more H+ in the solution would ensure there are more cationic hydrolysis products (eg. Fe(OH)2+, FeðOHÞþ 2 ) in the solution instead of FeðOHÞ4 . In this way, under acidic condition, charge neutralization between HA molecules (with less negative charges) and FC hydrolysates (more positively charged) would be enhanced. Thus, the demand of FC for would be reduced to achieve the same coagulation effect [30– 32]. Accordingly, pH 6 was determined to be the optimum pH condition for FC/DAM–ECH coagulation in this study, and its membrane performance in the follow up UF process would also be investigated in comparing with other coagulation methods. 3.2. Floc properties Floc sizes, breakage/recovery factors and structure were investigated as parameters to further analyze the membrane fouling in UF process. All parameters were generated from Fig. 4, of which each picture can be divided into 3 stages, formation period (0–12.5 min), followed by 5 min of breakage stage and the regrowth period lasted for 10 min. 3.2.1. Floc sizes under various coagulation and hydraulic conditions As revealed in Fig. 4, in general, formation of flocs started right after the coagulants added into raw water, the size increased dramatically for in the first 5 min; then slowdown and trend to be steady in the following 7.5 min. Afterwards, when shearing force were introduced, floc sizes drop sharply especially when rotate speed was high up to 400 rpm. Finally, during the last 10 min of regrowth period, flocs restored to certain sizes but far below their original sizes in steady stage of formation period. Fig. 5 is introduced to show the floc sizes in all stages more accessible, in which is obvious floc sizes dropped more dramatically as higher stirring speed was applied. Also, the lower shearing forces, the larger flocs would be obtained in the regrowth period; when rotation speed was high up to 400 rpm, the regenerated flocs under original pH and pH 6 were almost with the same size but still a little higher than FC flocs, which means that shearing force would be the crucial impact that influences floc size. It can also be observed form Fig. 5 that addition of DAM–ECH led to larger floc size which high up around 600 lm; meanwhile when FC was applied as the only coagulant, floc size was only around 400 lm. Also sizes of regenerated floc by FC/DAM–ECH were much larger than that of FC, especially when low shearing forces (100 & 200 rpm) were introduced. Then when pH adjusted to 6, FC/ DAM–ECH floc grew even larger than under original pH which climbed over 680 lm in the formation period. Phenomena above about floc size could be explained as follow: according to previous researchers, flocs formed under acidic condition would be larger than that formed in alkaline region, which was because that more colloid particles and microflocs would be bridged by coagulant hydrolyzates under acidic condition and form larger flocs [33]. In conclusion, addition of DAM–ECH would apparently increase floc sizes, also larger flocs would be generated under acidic condition than under raw water pH. 3.2.2. Floc strength and recoverability As mentioned in Section 3.2.1, higher shearing force would produce smaller flocs in breakage period and meanwhile generated smaller flocs after regrowth period. Also larger flocs formed in growth period, the sharper drop of floc sizes would occur at the
S. Sun et al. / Chemical Engineering Journal 280 (2015) 634–642
A
B
c
Fig. 4. Flocs diameters on medium as a function of DAM–ECH dosage and pH in formation, breakage and regrowth stages.
Fig. 5. Flocs sizes in stable stage of formation and breakage periods with different coagulation conditions.
639
640
S. Sun et al. / Chemical Engineering Journal 280 (2015) 634–642
Table 1 Effect of shearing force on flocs strength factors under different coagulation conditions. Coagulation conditions
FC (raw water pH) FC/DAM–ECH (raw water pH) FC/DAM–ECH (pH 6)
Stirring speed (rpm) 100
200
400
57.24 49.92 48.71
25.38 24.28 23.60
15.05 15.01 13.82
Table 2 Effect of shearing force on flocs recovery factors under different coagulation conditions. Coagulation conditions
FC (raw water pH) FC/DAM–ECH (raw water pH) FC/DAM–ECH (pH 6)
Stirring speed (rpm) 100
200
400
28.24 31.52 34.81
26.17 31.72 37.61
20.21 21.56 22.03
beginning of breakage period, suggesting that larger flocs were easier to be broken [19]. Floc strength and recoverability were calculated by Eqs. (1) and (2), which were stated in Section 2.4.2. The strength factors were shown in Table 1, which is obvious that larger flocs formed by FC/DAM–ECH under pH 6 and original pH were more frangible than smaller flocs produced by FC alone. This result consist with the previous studies that flocs in larger sizes would be easier broken when shearing force were introduced [19]. Li et al. suggested that floc strength was of the following order: bridging > charge neutralization > sweep [34], which means the addition of DAM–ECH changed the main coagulation mechanism from bridging to charge neutralization.
A
Table 2 shows that addition of DAM–ECH would not only tremendously increase the floc sizes formed by FC but also increase their recovery ability against the shearing force, especially when pH was adjusted into weak acidic range (pH 6). It is noticeable that when the rotation speed was set to be 200 rpm, flocs formed by FC/ DAM–ECH had much better recovery ability under original or acidic condition. This was because that highly positively charged DAM–ECH enhanced the charge neutralization of FC with HA molecules, and flocs formed by charge neutralization have better recovery ability than the flocs formed by sweep enmeshment [35,36]. In conclusion, flocs formed by FC/DAM–ECH were much larger than FC–HA flocs, especially under acidic region. Though larger flocs were easier to be broken, the FC/DAM–ECH flocs had better recovery ability which would make sure to produce larger flocs even under high shearing force.
3.2.3. Floc structure Fractal dimension value (Df) of floc would be used in demonstration of floc structure in this study. As common knowledge, higher Df means denser structure of floc [37–39]. As shown in Fig. 6, it can be observed that FC flocs (A) had most compact structure and largest Df value and formed cake layer with smallest pore sizes; meanwhile FC/DAM–ECH flocs (C) with lowest Df value formed loosest cake layer on membrane surface with biggest pore sizes. Also for each coagulation method, Df reached the highest point during the breakage period which revealed that floc size dropped to the low point and their values were in the following order: FC (2.8567) > FC/DAM–ECH (2.836) > FC/DAM–ECH pH 6 (2.6923); then Df dropped to certain level but still higher than the formation stage, which also consist with the change of floc size that regenerated flocs would not reach the floc size of formation period.
B
C Fig. 6. SEM images of flocs formed by different coagulation methods on membrane surface at 30,000 times magnification and their fractal dimensions: (A) FC (2.481); (B) FC/DAM–ECH (2.352); (C) FC/DAM–ECH pH 6 (2.271).
S. Sun et al. / Chemical Engineering Journal 280 (2015) 634–642
641
(2) Initial pH of raw water was proved to be a crucial parameter which influenced the turbidity and HA removal rates. The optimum pH condition for coagulation was around weak acidic condition, which in this study was pH 6. (3) Introduce of DAM–ECH influenced the floc properties by enlarge the floc size and lower Df. Furthermore, flocs formed under pH 6 were much bigger than those at original raw water pH and more fragile. Meanwhile recovery ability of flocs formed by FC/DAM–ECH was much better than by FC, especially when neutralization was enhanced as raw water pH was adjusted to 6. (4) FC combined use with DAM–ECH would improve the membrane performance in the following UF process. Adjustment raw water pH to 6 further reduce membrane fouling of FC/ DAM–ECH coagulated water.
Acknowledgment Fig. 7. Normalized UF permeate flux profiles for pre-coagulated HA suspension by different coagulation methods.
The project was financially supported by the National Natural Science Foundation of China (21377071).
3.3. Assessment of membrane fouling in ultrafiltration References UF experiments were conducted as post-treatment of coagulation jar test to investigate the effect of FC/DAM–ECH dose and pH on membrane fouling. The MFI and trend of permeate flux decline caused by different coagulation methods and are shown in Fig. 7. Generally, sharp drop of permeate fluxes occurred right after the filtration started under all water conditions; then the decrease trend to be gentler as filtration proceed and finally became steady after certain filtration period. It is obvious that the permeate flux influenced remarkably by different coagulation conditions, and the final permeate fluxes with coagulation processes were all higher than that of raw water, which has been proved by other researchers [10]. The sequence of MFI for different coagulation methods was as follows: FC/DAM–ECH pH 6 (28.43) < FC/DAM– ECH (76.79) < FC (96.91) < Raw water (128.77). The permeate flux of FC/DAM–ECH coagulated water at pH 6 dropped much slower than raw water and the other coagulation methods, meanwhile its final flux also appears to be much higher which was consist with MFI values. Coagulated water by FC/DAM–ECH also performed lighter membrane fouling comparing with raw water and FC coagulated water. Previous studies suggested membrane fouling is markedly influenced by floc properties, especially cake layer which mainly causes the revisable fouling [15,40,41]; moreover, increase of floc size and decrease of Df would reduce the membrane specific membrane resistance [42]. According to Fig. 4C, the average D0.5 of regenerated FC/DAM–ECH–HA flocs formed under pH 6 after broken by 200 rpm of shearing force was 225.05 lm, which was much higher than that of FC–HA flocs (139.28 lm) and FC/DAM–ECH–HA flocs (182.86 lm). Also considering that stated in Section 3.2.3, FC/DAM–ECH flocs formed under pH 6 had the lowest Df, which is consist with the membrane performance that the optimum membrane performance was achieved by FC/DAM–ECH at pH 6. As for FC/DAM–ECH coagulated water under original raw water pH, flocs had medium size and medium Df, so it achieved medium final permeate flux among all three coagulated water. 4. Conclusions (1) Addition of DAM–ECH with proper dosage would improve the turbidity and HA removal efficiencies by enhancing neutralization mechanism of coagulation process.
[1] X. Guo, Q. Li, W. Hu, W. Gao, D. Liu, Ultrafiltration of dissolved organic matter in surface water by a polyvinylchloride hollow fiber membrane, J. Membr. Sci. 327 (2009) 254–263. [2] D.R. Simpson, Biofilm processes in biologically active carbon water purification, Water Res. 42 (2008) 2839–2848. [3] M. Rossini, J.G. Garrido, M. Galluzzo, Optimization of the coagulation– flocculation treatment: influence of rapid mix parameters, Water Res. 33 (1999) 1817–1826. [4] W. Gan, Q. Liu, Coagulation of bitumen with kaolinite in aqueous solutions containing Ca2+, Mg2+ and Fe3+: effect of citric acid, J. Colloid Interface Sci. 324 (2008) 85–91. [5] Z. Yang, X. Liu, B. Gao, S. Zhao, Y. Wang, Q. Yue, Q. Li, Flocculation kinetics and floc characteristics of dye wastewater by polyferric chloride–polyepichlorohydrin–dimethylamine composite flocculant, Sep. Purif. Technol. 118 (2013) 583–590. [6] E.E. Chang, P.C. Chiang, S.H. Chao, C.H. Liang, Effects of polydiallyldimethyl ammonium chloride coagulant on formation of chlorinated by products in drinking water, Chemosphere 39 (1999) 1333–1346. [7] J.M. Laîné, D. Vial, P. Moulart, Status after 10 years of operation — Overview of UF technology today, Desalination 131 (2000) 17–25. [8] W. Yu, N.J.D. Graham, Application of Fe(II)/K2MnO4 as a pre-treatment for controlling UF membrane fouling in drinking water treatment, J. Membr. Sci. 473 (2015) 283–291. [9] W. Yuan, A.L. Zydney, Humic acid fouling during ultrafiltration, Environ. Sci. Technol. 34 (2000) 5043–5050. [10] J. Liu, B. Liu, T. Liu, Y. Bai, S. Yu, Coagulation–bubbling–ultrafiltration: effect of floc properties on the performance of the hybrid process, Desalination 333 (2014) 126–133. [11] W. Xu, Q. Yue, B. Gao, B. Du, Impacts of organic coagulant aid on purification performance and membrane fouling of coagulation/ultrafiltration hybrid process with different Al-based coagulants, Desalination 363 (2015) 126–133. [12] S. Zhao, B. Gao, Q. Yue, W. Song, R. Jia, P. Liu, Evaluation of floc properties and membrane fouling in coagulation–ultrafiltration system: the role of enteromorpha polysaccharides, Desalination 367 (2015) 126–133. [13] A.W. Zularisam, A.F. Ismail, M.R. Salim, M. Sakinah, T. Matsuura, Application of coagulation–ultrafiltration hybrid process for drinking water treatment: optimization of operating conditions using experimental design, Sep. Purif. Technol. 65 (2009) 193–210. [14] K. Chon, J. Cho, H.K. Shon, A pilot-scale hybrid municipal wastewater reclamation system using combined coagulation and disk filtration, ultrafiltration, and reverse osmosis: removal of nutrients and micropollutants, and characterization of membrane foulants, Bioresour. Technol. 141 (2013) 109–116. [15] W.-Z. Yu, N. Graham, H.-J. Liu, J.-H. Qu, Comparison of FeCl3 and alum pretreatment on UF membrane fouling, Chem. Eng. J. 234 (2013) 158–165. [16] J.K. Edzwald, Coagulation in drinking water treatment: particles, organics and coagulants, Water Sci. Technol. 27 (1993) 21–35. [17] P. Jarvis, B. Jefferson, J. Gregory, S.A. Parsons, A review of floc strength and breakage, Water Res. 39 (2005) 3121–3137. [18] M.A. Yukselen, J. Gregory, The reversibility of floc breakage, Int. J. Miner. Process. 73 (2004) 251–259. [19] J.C. Wei, B.Y. Gao, Q.Y. Yue, Y. Wang, Strength and regrowth properties of polyferric-polymer dual-coagulant flocs in surface water treatment, J. Hazard. Mater. 175 (2010) 949–954.
642
S. Sun et al. / Chemical Engineering Journal 280 (2015) 634–642
[20] R.J. François, Strength of aluminium hydroxide flocs, Water Res. 21 (1987) 1023–1030. [21] S.F.E. Boerlage, M.D. Kennedy, M.P. Aniye, E. Abogrean, Z.S. Tarawneh, J.C. Schippers, The MFI-UF as a water quality test and monitor, J. Membr. Sci. 211 (2003) 271–289. [22] S.F.E. Boerlage, M.D. Kennedy, M.R. Dickson, D.E.Y. El-Hodali, J.C. Schippers, The modified fouling index using ultrafiltration membranes (MFI-UF): characterisation, filtration mechanisms and proposed reference membrane, J. Membr. Sci. 197 (2002) 1–21. [23] R. Mao, Y. Wang, B. Zhang, W. Xu, M. Dong, B. Gao, Impact of enhanced coagulation ways on flocs properties and membrane fouling: increasing dosage and applying new composite coagulant, Desalination 314 (2013) 161–168. [24] B. Bolto, D. Dixon, R. Eldridge, S. King, Cationic polymer and clay or metal oxide combinations for natural organic matter removal, Water Res. 35 (2001) 2669– 2676. [25] Q. Wang, B. Gao, Y. Wang, Z. Yang, W. Xu, Q. Yue, Effect of pH on humic acid removal performance in coagulation–ultrafiltration process and the subsequent effects on chlorine decay, Sep. Purif. Technol. 80 (2011) 549–555. [26] Z.L. Yang, B.Y. Gao, Q.Y. Yue, Y. Wang, Effect of pH on the coagulation performance of Al-based coagulants and residual aluminum speciation during the treatment of humic acid–kaolin synthetic water, J. Hazard. Mater. 178 (2010) 596–603. [27] C.M. Flynn, Hydrolysis of inorganic iron(III) salts, Chem. Rev. 84 (1984) 31–41. [28] A.J. AB Lamb, The slow hydrolysis of ferric chloride in dilute solution, J. Am. Chem. Soc. (1938). [29] J. Duan, J. Gregory, Coagulation by hydrolysing metal salts, Adv. Colloid Interface Sci. 100–102 (2003) 475–502. [30] X. Bo, B. Gao, N. Peng, Y. Wang, Q. Yue, Y. Zhao, Effect of dosing sequence and solution pH on floc properties of the compound bioflocculant–aluminum sulfate dual-coagulant in kaolin–humic acid solution treatment, Bioresour. Technol. 113 (2012) 89–96. [31] Y.L. Wang, J. Feng, S.K. Dentel, J. Lu, B.Y. Shi, D.S. Wang, Effect of polyferric chloride (PFC) doses and pH on the fractal characteristics of PFC–HA flocs, Colloids Surf. A 379 (2011) 51–61.
[32] A. Adin, Y. Soffer, R. Ben Aim, Effluent pretreatment by iron coagulation applying various dose-pH combinations for optimum particle separation, Water Sci. Technol. 38 (1998) 27–34. [33] B. Cao, B. Gao, X. Liu, M. Wang, Z. Yang, Q. Yue, The impact of pH on floc structure characteristic of polyferric chloride in a low DOC and high alkalinity surface water treatment, Water Res. 45 (2011) 6181–6188. [34] T. Li, Z. Zhu, D. Wang, C. Yao, H. Tang, Characterization of floc size, strength and structure under various coagulation mechanisms, Powder Technol. 168 (2006) 104–110. [35] V. Chaignon, B.S. Lartiges, A. El Samrani, C. Mustin, Evolution of size distribution and transfer of mineral particles between flocs in activated sludges: an insight into floc exchange dynamics, Water Res. 36 (2002) 676– 684. [36] K. McCurdy, K. Carlson, D. Gregory, Floc morphology and cyclic shearing recovery: comparison of alum and polyaluminum chloride coagulants, Water Res. 38 (2004) 486–494. [37] T. Serra, J. Colomer, B.E. Logan, Efficiency of different shear devices on flocculation, Water Res. 42 (2008) 1113–1121. [38] X.-Y. Li, B.E. Logan, Permeability of fractal aggregates, Water Res. 35 (2001) 3373–3380. [39] J. Gregory, The density of particle aggregates, Water Sci. Technol. 36 (1997) 1– 13. [40] T.D. Waite, A.I. Schäfer, A.G. Fane, A. Heuer, Colloidal fouling of ultrafiltration membranes: impact of aggregate structure and size, J. Colloid Interface Sci. 212 (1999) 264–274. [41] K. Listiarini, D.D. Sun, J.O. Leckie, Organic fouling of nanofiltration membranes: Evaluating the effects of humic acid, calcium, alum coagulant and their combinations on the specific cake resistance, J. Membr. Sci. 332 (2009) 56–62. [42] S.A. Lee, A.G. Fane, T.D. Waite, Impact of natural organic matter on floc size and structure effects in membrane filtration, Environ. Sci. Technol. 39 (2005) 6477–6486.