ultrafiltration combined process

Chemical Engineering Journal 262 (2015) 1161–1167

Contents lists available at ScienceDirect

Chemical Engineering Journal journal homepage: www.elsevier.com/locate/cej

Coagulation performance and membrane fouling of different aluminum species during coagulation/ultrafiltration combined process Lijuan Feng a,b, Wenyu Wang b, Ruiqi Feng b, Shuang Zhao b, Hongyu Dong b, Shenglei Sun b, Baoyu Gao b, Qinyan Yue b,⇑ a b

Key Laboratory of Inorganic Chemistry in Universities of Shandong, Department of Chemistry and Chemical Engineering, Jining University, Qufu, Shandong 273155, China Shandong Provincial Key Laboratory of Water Pollution Control and Resource Reuse, School of Environmental Science and Engineering, Shandong University, Jinan 250100, China

h i g h l i g h t s  Coagulation performance of three high content Al species was first studied.  Membrane fouling of ultrafiltration would be different because of flocs performance.  The different floc properties and membrane fouling with Al species were discussed.  Alb has the best benefit to alleviate the membrane fouling.

a r t i c l e

i n f o

Article history: Received 25 August 2014 Received in revised form 17 October 2014 Accepted 19 October 2014 Available online 27 October 2014 Keywords: Al species Coagulation performance Coagulation/ultrafiltration Membrane fouling

a b s t r a c t Three species of aluminum coagulants mainly consist of mononuclear Al species (Ala), medium polymeric species (Alb) and colloidal species (Alc) were prepared, separated and purified by organic solvent precipitation method. The effect of different Al species on coagulation performance and membrane fouling were investigated through the coagulation–ultrafiltration hybrid process. The influence of flocs characteristics were also studied to build the relation model between Al species coagulation and membrane fouling mechanisms. The results showed different Al species combined with humic acid (HA) in diverse ways feature distinct floc properties and membrane fouling. Ala coagulant had better UV254 and dissolved organic carbon (DOC) removal efficiency. Furthermore, Ala and Alc were beneficial to form larger flocs with lower strength and looser structure, whereas Alb was prone to form small flocs with higher strength and compact structure. Ultrafiltration experiments indicated that Alb is the most effective Al species to alleviate membrane fouling. Ó 2014 Elsevier B.V. All rights reserved.

1. Introduction Ultrafiltration (UF) technology has been recognized as an attractive process for water treatment due to its ability of entrapping particles, colloids and microorganisms [1]. However, the main barriers of this technology are low removal efficiency of natural organic matters (NOM) and serious membrane fouling. Thus researchers focus on pretreatment prior to UF combined technology to improve the removal of NOM and reduce membrane fouling [2]. Nowadays, coagulation is used frequently before UF process, namely, coagulation/ultra-filtration (C-UF) combined technology. In this system, flocs were formed by instable colloids after addition of coagulants, and then absorb NOM. Meanwhile, the flocs were apt ⇑ Corresponding author. Tel.: +86 531 88365258; fax: +86 531 88364513. E-mail address: [email protected] (Q. Yue). http://dx.doi.org/10.1016/j.cej.2014.10.078 1385-8947/Ó 2014 Elsevier B.V. All rights reserved.

to be entrapped by membrane which would form cake layer and improve the filtration performance [3]. Wang et al. conducted CUF experiments for removing NOM in water and found the combined technology is able to enhance the removal of DOC and UV254. Moreover, it could reduce the formation of trihalomethanes (THMs) [4]. Hence, flocs formed by coagulation were the key to enhance the performance of membrane. Many researchers focus on the relationship between membrane fouling and floc characteristics, such as floc size, fractal dimension and floc strength etc. [4,5]. For example, it is reported that the floc size and fractal dimension formed by PACl had impact on the reduction of membrane fouling. Waite et al. reported the affinity between floc properties (size and structure) and the permeability of the cake that accumulated on the ultrafiltration membranes [6]. Guan et al. stated floc size and fractal dimension were related with specific cake resistance [7]. Cho et al. believed that flocs with

1162

L. Feng et al. / Chemical Engineering Journal 262 (2015) 1161–1167

lower fractal dimension have loose and porous structure which can enhance membrane permeability [8]. Latterly, research began to focus on the relation among the floc strength, the ability of regrowth and membrane fouling [9,10]. Results imply that floc with good characteristics could enhance coagulation efficiency and reduce membrane fouling. So improvement of floc characteristics is one of the most important affects in the combined technology. At present, as one of the most popular traditional coagulants, polyaluminium chloride (PACl) has been widely used in C–UF processes. PACl that composed of partially hydrolyzed products of Al(III) is superior to traditional Al salt coagulants in performance of particulate and higher organic matter removal efficiency [11]. In PACl, Al species were divided into three categories according to the size of their hydrolysis compound: mononuclear Al species (Ala), medium polymeric species (Alb) and colloidal species (Alc). Previous research showed the different Al species feature distinct characteristic, and distribution of Al species is the important factor of coagulation performance and mechanism. Furthermore, flocs formed by different Al species could be predicated and departed from water through different mechanisms of charge neutralization or bridge. So researchers paid attention to the structures of three Al species and their coagulation efficiency. Yan et al. concluded that Alb was the most effective species in coagulation process [12]. Xu et al. found that pre-hydrolyzed Al13 polycation ([Al13O4(OH)24(H2O)12]7+) had been confirmed to be correlated well with Alb species, generally resulting in smaller but more compact HA flocs than the commonly used PACl [13]. Luan et al. concluded that Keggin-type Al30 [(AlO4)2Al28(OH)56(H2O)26]18+ is the highest charged polycation in hydrolytic polymeric Al and belongs to a new optimum Al coagulation species [14]. However, study on floc performance formed by three species coagulates and membrane fouling was not clearly enough. Moreover, the most efficient specie of C-UF combined technology is remaining indistinct. Intrigued by these sealed effects, we undertook systematic studies in the effect of Al species on coagulation performance and membrane fouling of C-UF process. In this study, three species of Al salts were used in coagulation process to demonstrate the relationship between removal of HA-Kaolin raw water and floc properties. Then we chose the optimum project and discuss the relation between floc performance and membrane fouling.

2. Materials and methods 2.1. Preparation and purification of Ala, Alb and Alc All the reagents used in this experiment were of analytical grade. Ala was prepared by adding 1.7884 g AlCl36H2O in 200 mL deionized water directly. Alb: High Alb content of PACl was prepared before purification. Then Alb specie was purified with ethanol-acetone precipitation method. The method could be found in the paper [13,15]. Alc was prepared as follow: PACl was heated at 85 °C and stirred continually for 24 h. Then 48 h for curing, the PACl was purified by mixed methanol/acetone solvent (1:9).

2.2. Characterization and concentration calculation of Al species Total Al concentrations (AlT) were determined by EDTA complexometric method according to the National Standard of China Standard method (GB15892-1995). Ala and Alb species were analyzed with an UV-754 spectrophotometer (Precision Scientific Instrument Co. Ltd., Shanghai, China) by Ferron colorimetric

method [16,17]. The concentration of Alc was calculated by the equation:

Alc ¼ AlT  Ala  Alb

ð1Þ

The results of purification Alb were shown in Table 1, which showed that the sample of number 4 was optimal and used for the following experiments. The sample of number 4 was characterized with NMR spectra method to ensure the experiment accuracy [15]. The concentration of Alb was calculated according to the that of Al13 due to its high correlation to Alb [18]. NMR spectra of sample 4 (Fig. 1b) indicate there only Al13 specie peak and internal standard peak. The concentration of Al13 was 96.4% from formula (2), which is according to the result based on Ferron method. The more details can be found in literature [13].

Al13 ¼ ½ða2 =ar  C r  kÞ=C T   13  100%

ð2Þ

where a2 and ar are the peak areas at 62.5 ppm and 80.0 ppm of chemical shift, respectively, and Cr is the concentration of internal standard, and CT represents the sample concentration and k is the ratio of sectional area between capillary and sample cell. The results of purified Alc were shown in Table 2, Number 2 were used for the highest content of Alc. 2.3. Coagulation experiments 2.3.1. Synthetic test water HA stock solution was prepared as follows: 1.0 g of HA (Aladdin, Shanghai, China) and 0.4 g of NaOH (Tianjin Damao Co., Tianjin, China) was dissolved in 1000 mL deionized water. Kaolin stock solution was prepared as follows: 5.0 g Kaolin was dissolved into 800 mL deionized water with 30 min magnetic stirring, diluted to 1 L, then transfer into a 1 L measuring cylinder. After 30 min for standing, the upper 500 mL was drawn out for using. The synthetic test water was prepared by diluting 10.0 mL of HA stock solution with deionized water to 1 L. Kaolin (Kermel Co., China) was added to adjust turbidity 15.0 ± 0.5 NTU. The property of this synthetic test water was as following: UV254 = 0.29 ± 0.03, TOC = 5.24 ± 0.50 mg/L, pH = 8.3 ± 0.1, Turbidity = 15.0 ± 0.5 NTU, Zeta = 16 ± 2 mV. 2.3.2. Coagulation procedure Coagulation experiments were operated by jar test apparatus (ZR4-6, Zhongrun Water Industry Technology Development Co. Ltd., China). Coagulation procedure contains 4 steps: 1. Rapid mix 200 rpm lasting for 30 s to form sample uniformly; 2. Coagulant of certain dose was added, keep 200 rpm for 90 s; 3. Slow mixing speed of 40 rpm for 15 min to make particles gathering each other; 4. The solution settling for 20 min. After sedimentation, the supernatant was carefully withdrawn from about 2 cm below the surface for analysis. The turbidity was measured with turbidimeter (Hach 2100P Co., US). Zeta potential was tested by a Zetasizer (Malvern 3000Hsa, UK). Dissolved organic carbon (DOC) and UV254 absorbance were measured after the samples filtered through 0.45 lm fiber membrane. DOC values were measured by TOC analyzer (TOC-VCPH, Shimadzu, Japan) and

Table 1 Purification of the distribution of the Alb species (AlT 0.25 mol/L, B = 2.4, 85 °C). Number

20% ethanol/acetone (mL)

Ala%

Alb%

Alc%

1 2 3 4 5

0 160 60 100 160

9.76 6.04 5.85 0 1.3

90 78.96 72.88 100 81.82

0.24 15 21.27 0 16.88

1163

L. Feng et al. / Chemical Engineering Journal 262 (2015) 1161–1167

a

indicates the abilities of regrowth. A high Rf value shows better regrowth abilities after high shear.

NaAlO2 (the internal standard)

Ala

2.4. coagulation/ultrafiltration combined hybrid experiment

Al13

b

Ultrafiltration experiments were conducted in a dead-end batch unit after coagulation (without sedimentation). The membranes were commercial polyethersulfone membrane with cut-off molecular weight of 100 KDa and diameter of 80 mm (Mosu Co. Ltd., China). The cylindrical apparatus with capacity of 300 mL was pressurized with nitrogen gas at 150 ± 5 KPa to maintain a constant pressure. The pre-coagulated water was added gently into the cylindrical apparatus. A magnetic stirring apparatus was used under the cylindrical apparatus to avoid floc setting. Data of permeate mass was measured by an electronic balance (MSU5201S000-D0, SARTORIUS AG GERMANY) which logged onto a computer. The fouling behavior may be assessed by calculating the above data. The fouling potential of ultrafiltration membrane is explained by using modified fouling index (MFI) according to the cake filtration theory [20,21]:

NaAlO2 (the internal standard)

Al13

t

v Fig. 1. The typical 27Al NMR chart of PACl (a) and Alb sample (b). (The signal near 0.0 ppm represents Ala species; the signal at 62.5 ppm represents the Al13 species; and the signal at 80.0 ppm indicates the formation of Al(OH) 4 (the internal standard)).

¼

gRm gaC b þ V DPA 2DP A2

where t is the filtration time, V is filtrate volume, g is water viscosity, Rm is resistance of the membrane, DP is applied trans-membrane pressure, A is available membrane area, a is specific resistance of the cake deposited and Cb is concentration of particles in a feed water. The slop

Table 2 Purification of the distribution of the Alc species (AlT 0.25 mol/L, B = 2.4, 85 °C). Number

10% methanol/ acetone (mL)

Ala%

Alb%

Alc%

1 2 3 4

0 180 100 150

7.72 4.88 2.87 6.92

11.46 8.75 17.2 21.97

80.82 86.37 79.92 71.11

UV254 absorbance at 254 nm by UV spectrophotometer (Jinghua Science and Technology instrument Co., Shanghai, China), respectively. 2.3.3. Floc characteristic A laser diffraction instrument Mastersizer 2000 (Malvern Instruments, UK) was used to measure dynamic floc size during the whole experiment. The mean size (d50) was used to denote floc size. As the coagulation/flocculation process proceeded, floc was monitored through optical unit of the instrument by a peristaltic pump circulated on a 5 mm inside diameter tubing. The water sample procedure was carried out as follows: rapid mixing speed of 200 rpm for 30 s, the chemical coagulants were added in the jar, keep 200 rpm for 90 s, followed by a growth phase at 40 rpm for 15 min. Then a breakage period of 5 min at a speed of 200 rpm, followed by a regrowth phase at 40 rpm for 15 min. The floc strength and recovery factors were used to demonstrate the floc characteristic [19]:

Strength factor ¼

d2  100 d1

ð3Þ

Recovery factor ¼

d3  d2  100 d1  d2

ð4Þ

d1 is the average floc size of the initial steady phase; d2 is the average floc size of the breakage period; d3 is the floc size after regrowth to the steady phase. Strength factor (Sf) indicates the abilities of resisting high shear. A higher Sf value means lower degree of breakage and higher intensity that flocs have. Recovery factor (Rf)

ð5Þ

gaC b 2DP A2

is defined MFI and the higher value

indicates the more serious membrane fouling. Generally, Rt (total resistance) is considered to be the sum of Rm (membrane hydraulic resistance), Ra (adsorption resistance), Rp (pore blocking resistance) and Rc (cake formation resistance). To further study the membrane pollution mechanism, experiments were done as follows: After the pre-coagulated water was ultra-filtrated about 4 h, the membrane was taken out and wiped carefully to remove the cake layer. The membrane was reused to ultra-filtrate deionized water for 4 h under the same condition as above mentioned. From these procedures, Rc proportion to the Rm can be calculated (6).

Rc ¼

J0  J  100% J0  J

ð6Þ

where J0 is the permeate flux of deionized water filtration, J is the eventual permeate flux of water sample filtration and J0 is the initial permeate flux of water sample filtration [22]. 3. Results and discussions 3.1. Effect of Al species on coagulation removal Initially, coagulation experiments were carried out to ascertain the optimal coagulant dosage. The variation tendencies of residual turbidity, zeta potential, UV254 and DOC removal efficiencies with the coagulant dosages (0–8 mg/L as Al) were shown in Fig. 2. As presented, the variation tendencies of different Al species were all similar. Residual turbidity decreased sharply when dosage increased from 1 mg/L to 4 mg/L and then slow down when the dosage range from 4 mg/L to 8 mg/L. The residual turbidity values were in the following order: Alb > Alc > Ala. In Fig. 2(b), it is manifestly found that zeta potentials increased continuously with the rise of coagulant concentration and remained negative within dosage of 8 mg/L. However, zeta potential variation tendencies of different Al species were different. The zeta potential of Ala coagulated water increased obviously as dosage increased. While that of Alc coagulated water increased gently. It indicated that

1164

L. Feng et al. / Chemical Engineering Journal 262 (2015) 1161–1167

a

16 14

b

-2 -4

12

Zeta potential (mV)

Residual turbidity (NTU)

0

Ala Alb Alc

10 8 6

-6 -8 -10 -12 -14

4

Ala Alb Alc

-16

2

-18

0 0

2

4

6

8

0

2

4

Dosage (mg/L)

Dosage

90

60

c

6

8

mg/L)

d

80

50

TOC removal (%)

UV254 removal (%)

70 60 50 40 30

Ala Alb Alc

20

40

30

20

Ala Alb Alc

10

10

0 0

1

2

3

4

5

6

7

8

9

0

1

Dosage (mg/L)

2

3

4

5

6

7

8

9

Dosage (mg/L)

Fig. 2. Coagulation performance of Al species as a function of dosage: (a) UV254 removal efficiency, (b) zeta potential, (c) residual turbidity and (d) DOC removal efficiency.

Ala own the maximum of free positive charge and could change the zeta potential trend sharply. Fig. 2(c) and (d) showed that the UV254 and DOC feature similar removal curve. Both of them increased sharply with the dosage ranging from 1 mg/L to 3 mg/L, and then reached a plateau when dosage was higher than 6 mg/L. Therefore, the optima dosage determined as 6 mg/L. Ala and Alb had the similar UV254 removal efficiencies, which were all around 90% at the dosage of 6 mg/L. Ala had the best removal efficiency of DOC at the same dosage. Meanwhile, zeta potential of Ala, Alb and Alc were 6 mV, 8.6 mV and 12 mV at optimum dosage, respectively. The main coagulation mechanisms of Al salt are adsorption and charge neutralization [23], meanwhile HA-Kaolin water is an alkaline solution which can provide hydroxyl. As Al salt coagulant added in HA-Kaolin water, significant precipitation of polymeric positive hydroxide could obtain immediately. Then the charge neutralization would be occurred by adsorbed precipitate. Ala had the best removal efficiency of UV254 and DOC when dose was low. This suggested hydrolytic precipitation may be the dominant mechanism for Ala. It was mainly because Ala had the most hydrolysis reactions and formed the most precipitate. This would also explain why efficient turbidity removal occurred at low zeta potentials which somewhat below zero. It seemed that Ala had dual function of both hydrolytic absorption and charge neutralization. This

observation can also be explained by another reason of charge neutralization. Ala had the greatest capacity of charge neutralization which could furthest absorb hydroxyl in HA (Fig. 2(b)). Results of this table are consistent with the results of Hu et al. [24]. They claimed that Ala could not only hydrolysis to be Alb species but also function as pH control agent in the coagulation process. It was also in good agreement with the literature [25]. Further investigation was done to confirm the correctness of this experiment result. Ala was determined by Ferron method within coagulation process with detail results shown in Table 3. During the coagulation process, Ala quickly hydrolyzed to be Alb aggregations. The degree of hydrolysis grew as the dosage increased. Alb with highly positive charge had great capacity for charge neutralization. However, the removal efficiency of Alb was lower than that of Ala mainly due to the less free positive of Alb Table 3 The species changes of Ala during coagulation procedure at 1 min. Ala dosage (mg/L)

Ala%

Alb%

Alc%

1 4 6 8

18 6 3.5 2.7

82 94 96.5 97.3

0 0 0 0

1165

L. Feng et al. / Chemical Engineering Journal 262 (2015) 1161–1167

([Al13O4(OH)24(H2O)12]7+) compared with Ala. Alc had the lowest efficiency among three Al species. However, its UV254 removal rate was 78%, turbidity removal was 87% at the dosage of 8 mg/L. Fig. 2(b) showed that the capacity for charge neutralization of Alc was low. Although Al30 in Alc was highly positive charged, there were also inactive components. That’s why Zeta potential of Alc coagulated water was lower than the other. Thus, coagulation mechanisms of Alc were mainly absorption and sweep. The mechanisms of Ala and Alb were charge neutralization and absorption. Different mechanisms led to different coagulation performance, which could explain coagulation performance of Ala and Alb were better than Alc. 3.2. Floc characteristic 3.2.1. Effect of Al species on floc size Floc growth of different Al species coagulants during coagulation process was shown in Fig. 3. In this study, median diameter of floc (d50) was used to represent the floc size. During the coagulation procedure, positive charged coagulants neutralized quickly with negative charged colloidal particles and then particles formed into primary flocs which were in small sizes. Subsequently, the floc size began to grow rapidly with the increasing number of primary flocs. While the flocs grow, large flocs would joined together to form larger flocs via aggregating. Meanwhile, larger flocs would break by collision during aggregating. When the velocity of breakage and aggregation was in a balance condition,

250

flocs size reaches a steady plateau [26]. When high shear forces were introduced, the floc sizes decreased sharply. But floc sizes could come to a plateau again as the shear turn into low shear. It can be seen from Fig. 3 that floc sizes were significantly affected by coagulants dosage. Floc size at steady plateau first increased and then decreased as dosage increased. While the largest flocs were not obtained at the largest dosage were added in raw water. The floc sizes at dosage 4 mg/L with Ala and Alc had the largest sizes, while the floc sizes at dosage 6 mg/L with Alb were the largest. It was mainly attributed to that NOMs would charge positive charge after they were captured by coagulant. When more coagulant was added, redundant positive charge attached to NOM surface, which made NOMs excluded each other. Then floc sizes would not grow anymore. Results mentioned above indicated that the highest organics removal was not achieved in coagulation system with the largest floc size. This was due to the loose structure of larger flocs, which were not easy to settle down. Floc sizes at steady plateau were following the order of Ala > Alc > Alb at the same dosage of 6 mg/L and/or at the maximum floc sizes, respectively. That was attributing to different coagulation mechanisms with Al species. During coagulation process, Ala was complexes with HA and also hydrolyzed into Alb at the same time. It was reported that the primary aggregates which formed from two different reactions would be larger by charge neutralization [27]. Alc contained large polymer and colloidal species, which had large size that could easily precipitate. It also had a better capacity of bridging flocculation because of the large 180

2mg/L 4mg/L 6mg/L 8mg/L

Ala

200

2mg/L 4mg/L 6mg/L 8mg/L

Alb

160 140

Floc size d50(µm)

Floc size d50(µm)

120

150

100

100 80 60 40

50 20 0

0

-20

0

5

10

15

20

25

30

35

40

0

5

10

Time (min)

250

2mg/L 4mg/L 6mg/L 8mg/L

Floc size d50(µm)

200

150

100

50

0 5

10

15

20

20

Time (min)

Alc

0

15

25

30

35

40

Time (min) Fig. 3. Floc sizes during coagulation with Ala, Alb and Alc species.

25

30

35

40

L. Feng et al. / Chemical Engineering Journal 262 (2015) 1161–1167

3.3. Effect of Al species on membrane fouling Considering coagulation performances and floc properties, the dosage of Al was selected as 6 mg/L to conduct subsequent UF processes. The variation of permeate flux as a function of filtration time were shown in Fig. 4. Table 4 Floc strength (Sf) and recovery factors (Rf) of flocs with Al species. Dosage (mg/L)

2 4 6 8

Ala

Alb

Alc

Sf

Rf

Sf

Rf

Sf

Rf

44.14 40.82 40.09 41.73

43.73 23.65 14.28 17.70

30.73 36.14 45.21 43.91

47.47 36.58 31.26 26.36

40.39 37.94 44.58 40.43

27.93 22.09 12.13 22.01

Ala Alb Alc

Strength factor (Sf)

Recovery factor (Rf)

Fractal dimension (Df)

R2

40.9 45.21 44.58

14.28 31.26 12.13

2.16 ± 0.01 2.45 ± 0.02 1.82 ± 0.02

0.9988 0.9978 0.9984

16

1.1

Ala MFI=84.07 Alb MFI=44.04 Alc MFI=87.03

1.0 0.9 0.8 0.7

%

3.2.2. Effect of Al species on floc strength and compaction Al species also had effect on floc strength and compaction. Sf and Rf were calculated to evaluate the floc properties of breakage and regrowth. Table 4 showed the results of different Al species in dosage range of 2–6 mg/L. Usually, flocs with larger Sf values are stronger than those with lower Sf values. For Ala, as the dosage increased, the changes of Sf values were not obvious and the largest value is 44.14 at the dosage 2 mg/L (Table 5) which means the floc sizes small but strong. This was because flocs concentrations were low and hard to form big ones in the process of collisions. Small flocs would have strong ability to resist shear. It was also because Ala flocs grown slowly and not achieved plateau process at the setting time. There was too little coagulation and they could not completely combine with HA. Small flocs settled slowly and part of HA remained in the effluent at low dosage. When dosage increased to 6 mg/L, sufficient Ala served as primary aggregates to combine with flocs tightly. So, better flocs performance could be shown more accurately under this dosage condition. For Alb and Alc, they have the largest Sf values at dosage 6 mg/L. Table 5 was made to compare with three Al species in the same dosage of 6 mg/L. Under the same shear and dosage condition (6 mg/L, Table 5), Sf values were in following order: Alb > Alc > Ala. Rf values of Ala, Alb and Alc were all high under small dosages, but decreased gradually as dosage increased. Since different recoverability of flocs were related to different coagulation mechanism. Chaignon et al. found that flocs generated by charge neutralization were the strongest ones [28]. Meanwhile Cohen et al. believed flocs generated by sweep have poor regrowth after breakage [29]. So the values become lower as dosage higher. The results were consistent with the trend of UV254 and DOC removal, which indicated that flocs with better capability of floc strength and floc recovery had higher removal efficiency. Compared with Ala and Alc, Alb had higher Sf and Rf values, implying that Alb-HA flocs were stronger and had better ability to resist the shear forces and meanwhile the fractured flocs can be regrow more easily. Boller proposed smaller flocs (<200 lm) were more likely to withstand shearing forces [30]. Xu also believed that the Alb aggregates had much branched structure and could form bridges between particles [20]. Table 5 also summarized Df values of Al species. Alb-HA flocs have the largest Df value. That was attributed to the higher positive charge of Alb polymer which could weaken the repulsive forces between particles in the agglomerates. Therefore it led to a higher degree of compaction.

Table 5 Floc strength and recovery factors, fractal dimensions (Df) of Al species at the same dosage 6 mg/L.

14

Percentage of cake resistance

size. For Alb, the precipitate formed from Alb-HA complexions mainly occurred by adsorption charge neutralization. Considering coagulation efficiency, it is clearly that larger flocs can precipitate easier. The reason was larger flocs had a better capacity of bridging which were quite easy to settlement.

Normalized permeat flux, J/J 0 (-)

1166

12 10 8 6 4 2 0 Ala

Alb

Alc

Al species

0.6 0.5 0.4 0.3 0

2000

4000

6000

8000

10000

12000

Time (s) Fig. 4. Normalized flux declines of effluents and percentage of cake resistance (Rc) for Al species with Al species.

Permeate flux ratio (J/J0) was used to normalize membrane fouling rate. As can be seen, the permeate fluxes declined sharply at first and then varied inconspicuous with filtration time increasing. The filtration experiments were carried on untill fluxes relatively reached a constant. It also can be observed that the permeate flux of Alb coagulation effluent was obviously higher than these of Ala and Alc, which meant that the membrane fouling of Alb was slighter than Ala and Alc. It was mainly due to the tougher flocs formed by Alb than by others which could leave less aggregates fractions in effluent water. Barbot et al. reported that the effect of flocs on the permeate flux depended on the ability resistant to shear stress [31]. This may account for the pressure in the cylinder, which broke large flocs formed by Ala and Alc. Alb flocs were stronger than the other two, so the flocs were avoided being broken. The membrane fouling of Ala is similar with that of Alc. This could be ascribed to different floc properties formed by three Al species. Ala flocs had biggest size, but had loose structure which could lead to the floc broken when the press on the cake increased. But for Alc, it had big floc size and compact structure which could reduce the membrane fouling, but it also had high strength which would increase membrane fouling. Ala flocs were larger than had Alc flocs, but meanwhile, Alc flocs also had low Df value and more compact in structure, which could reduce membrane fouling. Thus, Ala and Alc had similar membrane flux. Fig 4 also showed the percentage of cake resistance in total resistance, which was one of the factors affecting membrane pollution. The order was as follows: Ala > Alb > Alc. This order could be evaluated by floc properties. Park et al. concluded a relationship among Rc, floc size and fractal dimension [32]. They predicted that Rc decreased when floc size increased and Df values decreased. The outcome from Alc was consistent with this prediction. Alc flocs had the largest floc sizes and the smallest Df values, so it had the smallest Rc percentage. It should be pointed out that Rc percentage with

L. Feng et al. / Chemical Engineering Journal 262 (2015) 1161–1167

Ala slightly higher than that with Alb even though the flocs formed by Ala had large floc size and low Df. This was due to flocs with low Sf of Ala partly compensating the advantage of large size flocs and low Df values. Xu et al. reported the permeate flux was in agreement with floc strength [13]. Zhao et al. also believed that floc strength is a very important factor influencing membrane fouling [27]. It indicated that cake formed by flocs with improvement strength was not easy to be compressed and less porous. Thus, the flux profile exhibited similar by the treatment of Ala and Alb and higher than that of Alc. 4. Conclusion The following conclusions can be listed: 1. Different Al species in PACl combined with HA in diverse ways, which led to different floc properties and membrane fouling. 2. Ala has the highest removal efficiency of UV254 and TOC within coagulation process. Alc had the lowest efficiency among all Al species. 3. Ala and Alc were beneficial to form large flocs with low strength and loose structure. Alb was beneficial to form small flocs, with high strength and compact structure. 4. Alc contained large polymer and colloidal species which could easily precipitate. Alc also has a better capacity of bridging flocculation. Alc contains Al30 which has high positive charge and could do well with complexion HA. 5. Alb is the most effective Al species to alleviate membrane fouling; meanwhile Ala and Alb had similar impacts on membrane fouling.

Acknowledgements This work was financially supported by the National Natural Science Foundation of China (No. 21377071). References [1] J.M. Laine, D. Vial, P. Moula, Status after 10 years of operation – Overview of UF technology today, Desalination 131 (2000) 17–25. [2] S.J. Judd, P. Hillis, Optimisation of combined coagulation and microfiltration for water treatment, Water Res. 35 (2001) 2895–2904. [3] M.R. Wiesner, M.M. Clark, J. Mallevialle, Membrane filtration of coagulated suspension, J. Environ. Eng. 115 (1989) 20–40. [4] J. Wang, J. Guan, S.R. Santiwong, Characterization of floc size and structure under different monomer and polymer coagulants on microfiltration membrane fouling, J. Membr. Sci. 321 (2008) 132–138. [5] M. Stoller, On the effect of flocculation as pretreatment process and particle size distribution for membrane fouling reduction, Desalination 240 (2009) 209–217. [6] T.D. Waite, A.I. Schäfer, A.G. Fane, Colloidal fouling of ultrafiltration membranes: impact of aggregate structure and size, J. Colloid Interface Sci. 212 (1999) 264–274. [7] J. Guan, R. Amal, T. Waite, Effect of aggregate size and structure on specific resistance of biosolids filter cakes, Water Sci. Technol. 44 (2001) 215–220.

1167

[8] M.H. Cho, C.H. Lee, S. Lee, Influence of floc structure on membrane permeability in the coagulation–MF process, Water Sci. Technol. 51 (2005) 143–150. [9] Y.H. Choi, H.S. Kim, J.H. Kweon, Role of hydrophobic natural organic matter flocs on the fouling in coagulation–membrane processes, Sep. Purif. Technol. 62 (2008) 529–534. [10] E. Barbot et al., Coagulation and ultrafiltration: understanding of the key parameters of the hybrid process, Membr. Sci. 325 (2008) 520–527. [11] Z.S. Qian et al., Theoretical investigation of water exchange on the nanometersized polyoxocation AlO4Al12(OH)24(H2O)7+ 12(Keggin-Al13) in aqueous solution, J. Am. Chem. Soc. 130 (2008) 14402–14403. [12] M. Yan et al., Relative importance of hydrolyzed Al(III) species (Ala, Alb, and Alc) during coagulation with polyaluminum chloride: a case study with the typical micro-polluted source waters, J. Colloid Interface Sci. 316 (2007) 482– 489. [13] W.Y. Xu, B.Y. Gao, R.R. Mao, Influence of floc size and structure on membrane fouling in coagulation–ultrafiltration hybrid process – The role of Al13 species, J. Hazard. Mater. 193 (2011) 249–256. [14] Z.K. Luan, Z.Y. Chen, Y.Z. Li, Progress in the species of Al30 in hydrolytic polymeric aluminum solutions, Prog. Chem. 17 (2005) 1034–1039. [15] H.Z. Zhao et al., The formation mechanism of Al13 and its purification with an ethanol–acetone fractional precipitation method, Sep. Purif. Technol. 81 (2011) 466–471. [16] D.R. Parker, P.M. Bertsch, Identification and quantification of the ‘‘Al13’’ tridecameric polycation using ferron, Environ. Sci. Technol. 26 (1992) 908– 914. [17] D. Wang, W. Sun, Y. Xu, Speciation stability of inorganic polymer flocculantPACl, Colloids Surf. A 243 (2004) 1–10. [18] H.X. Tang, Z.K. Luan, The differences of coagulating behavior and mechanism between preproduced inorganic polymeric and traditional coagulants, Environ. Chem. 16 (1997) 497–505 (in Chinese). [19] P. Jarvis, B. Jefferson, Breakage, regrowth, and fractal nature of natural organic matter flocs, Environ. Sci. Technol. 39 (2005) 2307–2314. [20] S.F.E. Boerlage, M.D. Kennedy, M.R. Dickson, The modified fouling index using ultrafiltration membranes (MFI-UF): characterisation, filtration mechanisms and proposed reference membrane, J. Membr. Sci. 197 (2002) 1–21. [21] S.F.E. Boerlage, M.D. Kennedy, M.P. Aniye, The MFI-UF as a water quality test and monitor, J. Membr. Sci. 211 (2003) 271–289. [22] H. Dong et al., Effect of Fe (III) species in polyferric chloride on floc properties and membrane fouling in coagulation–ultrafiltration process, Desalination 335 (2014) 102–107. [23] J. Gregory, J.M. Duan, Hydrolyzing metal salts as coagulants, Pure Appl. Chem. 73 (2001) 2017–2026. [24] C.Z. Hu, H.J. Li, J. Qu, Coagulation behavior of Al13 species, Environ. Sci. 27 (2006) 2467–2471 (in Chinese). [25] W.Y. Xu, Influences of polysilicic acid in Al13 species on floc properties and membrane fouling in coagulation/ultrafiltration hybrid process, Chem. Eng. J. 181 (2012) 407–415. [26] J. Yu et al., Flocculation of kaolin particles by two typical polyelectrolytes: a comparative study on the kinetics and floc structures, Colloids Surf. A 290 (2006) 288–294. [27] B. Zhao et al., Influence of floc structure on coagulation–microfiltration performance. Effect of Al speciation characteristics of PACls, Sep. Purif. Technol. 72 (2010) 22–27. [28] V. Chaignon, B.S. Lartiges, 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 (3) (2002) 676–684. [29] R.A. Cohen, P. Fong, Nitrogen uptake and assimilation in Enteromorpha intestinalis (L.) Link (Chlorophyta): using 15N to determine preference during simultaneous pulses of nitrate and ammonium, J. Exp. Mar. Biol. Ecol. 309 (2004) 67–77. [30] M. Boller, S. Blaser, Particles under stress, Water Sci. Technol. 37 (1998) 9–29. [31] Y.H. Choi, H.S. Kim, J.H. Kweon, Role of hydrophobic natural organic matter flocs on the fouling in coagulation–membrane processes, Sep. Purif. Technol. 62 (2008) 529–534. [32] P.K. Park, C.H. Lee, S. Lee, Variation of specific cake resistance according to size and fractal dimension of chemical flocs in a coagulation–microfiltration process, Desalination 199 (2006) 213–215.