The filtration and fouling performance of membranes with different pore sizes in algae harvesting

The filtration and fouling performance of membranes with different pore sizes in algae harvesting

STOTEN-21972; No of Pages 7 Science of the Total Environment xxx (2017) xxx–xxx Contents lists available at ScienceDirect Science of the Total Envir...

2MB Sizes 0 Downloads 126 Views

STOTEN-21972; No of Pages 7 Science of the Total Environment xxx (2017) xxx–xxx

Contents lists available at ScienceDirect

Science of the Total Environment journal homepage: www.elsevier.com/locate/scitotenv

The filtration and fouling performance of membranes with different pore sizes in algae harvesting Fangchao Zhao, Huaqiang Chu, Zhenjiang Yu, Shuhong Jiang, Xinhua Zhao, Xuefei Zhou, Yalei Zhang ⁎ State Key Laboratory of Pollution Control and Resource Reuse, Tongji University, Shanghai 200092, China

H I G H L I G H T S

G R A P H I C A L

A B S T R A C T

• Effect of membrane pore size on critical flux was studied. • Continuous filtration was conducted to find the optimal pore size. • The drag force revealed how pore size affected membrane fouling. • Compared with 0.03 and 0.05 μm, 0.1 μm was suitable for harvesting C. pyrenoidosa.

a r t i c l e

i n f o

Article history: Received 11 November 2016 Received in revised form 4 February 2017 Accepted 5 February 2017 Available online xxxx Editor: Simon Pollard Keywords: Algae harvesting Filtration Pore size Permeate drag force Membrane fouling

a b s t r a c t In this study, ultrafiltration membranes with three different pore sizes were applied for algae harvesting to investigate filtration performance. The critical fluxes (JC) increased as the pore size increased, and the JC of 0.03-, 0.05and 0.1-μm membranes were 20.0, 25.0 and 42.0 L m−2 h−1, respectively. During continuous filtration, 0.7JC was selected as the operation flux and the 0.1-μm membrane had the highest initial flux and final flux. It also had the highest flux decline rate, and therefore, the 0.1-μm membrane was more appropriate for algae separation compared to the 0.03- and 0.05-μm membrane. The mechanism by which pore size influenced filtration performance and membrane fouling was presented from the viewpoint of permeate drag force (FD). A lower FD retarded the velocity of algae cells towards the membrane, which could decelerate the deposition of particles on the membrane and thus reduce the membrane fouling rate. As the pore size increased, the membrane hydraulic resistance (Rm) decreased, which led to a decrease of FD. © 2017 Elsevier B.V. All rights reserved.

1. Introduction The production of biofuel from algae is a promising new technology. Biofuel, as an alternative to fossil fuels, can reduce global warming caused by fossil fuels and has very promising application prospects because of the renewable and nontoxic properties of microalgae (Davis et ⁎ Corresponding author. E-mail address: [email protected] (Y. Zhang).

al., 2014; Zhao et al., 2015b). Biofuel produces approximately zero net carbon dioxide and releases fewer gaseous pollutants than fossil fuels (Uduman et al., 2010). Because microalgae can absorb carbon dioxide from the atmosphere for growth, the consumption of biofuel emits carbon dioxide into the atmosphere. Thus, the use of biofuel can play an important role in mitigating the emission of greenhouse gases. Although microalgae can be a good source of biofuel, there are substantial challenges in efficient harvesting and dewatering for commercial use. The challenges of harvesting and dewatering algae derive from the

http://dx.doi.org/10.1016/j.scitotenv.2017.02.035 0048-9697/© 2017 Elsevier B.V. All rights reserved.

Please cite this article as: Zhao, F., et al., The filtration and fouling performance of membranes with different pore sizes in algae harvesting, Sci Total Environ (2017), http://dx.doi.org/10.1016/j.scitotenv.2017.02.035

2

F. Zhao et al. / Science of the Total Environment xxx (2017) xxx–xxx

nature of microalgae, such as the small size of algal cells (mostly below 30 μm), the similarities in the densities of microalgae (approximately 1.025 g/L) and water, the concentration of algae in culture solutions at approximately 0.5–3 g/L and the negative surface charge on microalgae that results in dispersed stable algal suspensions, especially during the growth phase (Bhave et al., 2012; Uduman et al., 2010). Due to the nature of microalgae, harvesting is considered a major obstruction to algal biofuel production and a main factor limiting the commercial use of microalgae (Hwang et al., 2015; Milledge and Heaven, 2012; Weschler et al., 2014). Recently, the application of membrane technology to algae harvesting has drawn growing attention, due to their advantages of high efficiency and stability (Drexler and Yeh, 2014; Zhang et al., 2014; Zhao et al., 2015a). Membrane separation processes can recover microorganisms and yield stable, clean effluent water (Pavez et al., 2015); moreover, membrane technology can remove viruses and protozoans from culture media but retaining the residual nutrients, and thus the cultivation medium can be recycled. As the manufacturing techniques for membranes improve and their applications increase, the cost of membranes steadily falls, making it possible to apply membranes to algae harvesting. Therefore, an increasing number of researchers have studied how to make better use of membranes to harvest microalgae (Ahmad et al., 2013; Rossignol et al., 1999). However, some obstacles must still be overcome in harvesting algae. The primary problem for this process is membrane fouling, which may result in a decline in flux or increase in transmembrane pressure (TMP) (Drews et al., 2006; Rickman et al., 2012; Zhao et al., 2015a). These changes lower harvesting efficiency and increase harvesting cost. Thus, reducing membrane fouling and improving membrane flux are very significant goals in algae harvesting. Membrane filtration in algae harvesting is a physical separation process whose separation performance depends on membrane pore size. Pores that are too large cannot retain all algae cells, but pores that are too small dramatically reduce permeate flux, which seriously affects harvesting efficiency (Batista et al., 2013). However, the literature shows that the pore size of membranes is not a crucial factor because during algae filtration, the fouling layer on the membrane, caused by the deposition of algae cells and extracellular organic matter, can act as a membrane-selective layer (Batista et al., 2013; Nguyen et al., 2012; Zhou et al., 2014). However, pore size is still an important parameter that influences membrane performance. Selecting an appropriate pore size and operation flux can reduce membrane fouling rate and flux decline rate, which can extend filtration period and reduce membrane cleaning frequency. In this study, membranes with three different pore sizes (0.03, 0.05 and 0.1 μm) were used to filter algae to find the appropriate membrane pore size for algae harvesting. Moreover, the mechanism by which membrane pore size influenced membrane fouling and flux was presented and investigated. First, critical flux experiments were conducted to measure the critical flux of the three membranes. Second, the permeate drag force was calculated to theoretically demonstrate how pore size affected membrane filtration. Finally, continuous filtration tests were conducted using membranes with three different pore sizes to obtain the appropriate membrane pore size for filtering algae. This research can provide valuable information for algae harvesting using ultrafiltration membrane technology.

2. Materials and methods

other cultivation conditions were the same: light intensity 127 μmol m−2 s, light/dark = 14 h/10 h. 2.2. Reactor set-up The filtration was performed in a lab-scale tank, as shown in Fig. 1. The effective volume of the filtration tank was 4 L. A micro-porous pipe was placed below the membrane to reduce membrane fouling and the aeration flow rate was controlled at 6 L min−1. Three different hydrophilic PVDF ultrafiltration membranes (pore sizes of 0.03, 0.05 and 0.1 μm, Minglie, Shanghai, China) were employed to determine the appropriate pore size for algae harvesting. The effective filtration area of the membrane was 0.02 m2. The filtrate was filtered using a peristaltic pump. The volume of filtrate was measured and recorded using an electric balance and a computer, respectively. The values of the transmembrane pressure (TMP) were recorded using a vacuum meter. 2.3. Filtration experiments 2.3.1. Critical flux tests In filtration, a high flux will result in severe membrane fouling and sharp flux decline, while a low flux adversely influences efficiency. Thus, selecting an appropriate sub-critical flux will prolong membrane service life and reduce membrane cleaning frequency (Bilad et al., 2012c; van der Marel et al., 2009; Wicaksana et al., 2012). Moreover, the value of critical flux can be utilized to compare the propensity towards membrane fouling. In the present study, an improved flux-step method (IFM) was used to determine the value of critical flux (van der Marel et al., 2009). The difference between IFM and the common flux-step method is that in IFM, the level of the successive membrane flux (JH) increases, including an intermediate flux decrease to an initial low flux (JL) after each JH step. In IFM, filtration at JL is considered a form of relaxation, although real relaxation filtration should be 0 L m−2 h−1. However, a flux larger than 0 L m−2 h−1 must be applied to measure a value for TMP before and after a JH (van der Marel et al., 2009). At low flux, the convective flow towards the membrane is reduced and, due to air scouring, the reversible fouling reduced. For the 0.03-, 0.05- and 0.1-μm membranes, the JL of each module was identical (10 L m−2 h−1), with JH starting from 10, 10 and 15 L m−2 h−1 and stepwise increasing by 2.5, 2.5, and 3 L m− 2 h− 1, respectively. The duration for all JL and JH processes was 15 min. In this research, an arbitrary minimum increase in the TMP of 20 Pa min−1 was used to determine JC. In theory, once the JC of all the modules was achieved, the test should stop; however, to better comprehend the effect of temperature on the performance of the membranes, the maximum JH of 0.03-, 0.05- and 0.1-μm membranes reached 22.5, 27.5 and 42 L m− 2 h− 1, respectively. In the critical flux tests, there were two phases, namely, the ascending and descending phase. In the ascending phase, with the increase in JH, critical flux is achieved; in the descending phase, JH decreases stepwise and can achieve fouling condition. The algae in the stationary phase was used for all IFM tests and continuous filtration experiments, and the concentrations were approximately 0.3 g L−1. The experiments were performed for several days to reduce the impact of the fast growth of algae on filtration, and we thus selected the stationary phase rather than the exponential phase. To reduce the influence of change in temperature on algae, all filtration experiments were conducted at 35 ± 0.5 °C.

2.1. Cultivation of microalgae 2.3.2. Continuous filtration experiments Chlorella pyrenoidosa (C. pyrenoidosa, FACHB-9) was purchased from the Institute of Hydrobiology at the Chinese Academy of Sciences. C. pyrenoidosa was cultured in Basal medium prepared in sterilized distilled water. The algae was inoculated in 3-L glass flasks and placed in incubators. The cultivation temperature was kept at 35 ± 0.5 °C. The

2.3.2.1. Short-term filtration. Short-term filtration was performed for 1 h because after 1 h, filtration with the 0.03-μm membrane was slowed by serious fouling and filtration could not continue. The operational flux of these three membranes was set as 20 L m−2 h−1.

Please cite this article as: Zhao, F., et al., The filtration and fouling performance of membranes with different pore sizes in algae harvesting, Sci Total Environ (2017), http://dx.doi.org/10.1016/j.scitotenv.2017.02.035

F. Zhao et al. / Science of the Total Environment xxx (2017) xxx–xxx

3

Fig. 1. Schematic diagram of the experimental setup in the filtration experiment.

where JC is the critical flux using different pore size membranes (L m−2 h−1); ΔJ is the specific rate of flux change (s−1), J is the filtration flux (L m−2 h−1), JI is the initial flux (L m−2 h−1), and t is the filtration time (h). In the previous study, 0.7JC (0.7 times JC) was selected as JI (Chu et al., 2016), so in this study 0.7JC was also selected as JI. In continuous filtration, when flux was kept constant at the given initial flux, the membrane will be seriously fouled in several hours, and continuous filtration cannot be conducted for long periods. Thus, during 24-h filtration experiments, the peristaltic pump was kept at a constant speed (Li et al., 2014).

membrane had the lowest. The reason was that 0.1 μm membrane had the highest flux and the longest filtration time and it might generate more pollutants, so it naturally had the highest fouling when the flux reached JC (Lee et al., 2013). Although 0.05 μm membrane had a longer filtration time and larger flux compared to 0.03 μm membrane, it had a lower fouling when the flux reached JC. That might be because the pore size of 0.05 μm membrane was larger than 0.03 μm membrane, and 0.05 μm membrane only had a slight longer filtration time and slight higher critical flux than 0.03 μm membrane. In the descending phase, the value of TMP was higher compared with the corresponding fluxes in ascending stage, which indicated that the membranes have been influenced by their fouling history and fouled by the irreversible fouling in the ascending phase (van der Marel et al., 2009). The membranes were more seriously affected, especially at large flux with high TMP, resulting in a more cohesive fouling that could not be completely removed by bubbles in the descending phase. For the 0.1-μm membrane at a flux of 36 L m− 2 h−1 in the descending phase, the value of TMP was higher than that at 42 L m− 2 h− 1. The 0.1-μm membrane was more seriously fouled with irreversible fouling due to long filtration times and high flux.

2.4. Analytical methods

3.2. Effect of pore size on membrane cleaning

The algae concentration was measured using the OD680 method. SEM analysis was applied to observing the fouling condition of fouled and clean membranes (XL30FEG, PHILIPS, Holland). The dynamic viscosity of the algae solution was measured using a viscometer (NDJ-8S, Shanghai, China). The size distribution of the microalgae was determined by a laser particle size analyzer (Mastersizer 3000, Malvern, England)

After IFM tests, the membranes were first cleaned by flushing with tap water, and then washed with an ultrasonic cleaner for 20 min and subsequently soaked in a 1000 ppm sodium hypochlorite (NaClO) solution followed by a 1000 ppm citric acid (C6H8O7) solution for 2 h each (Bilad et al., 2012a). In this way, the overwhelming majority of algae cells, inorganic particles, and extracellular organic matter (EOM) was removed from the surface and pores of the membrane (Bilad et al., 2012b). The fouled membrane and clean membranes were used for scanning electron microscope (SEM) analysis (shown in Fig. 3). The

2.3.2.2. Long-term filtration. The filtration time was set as 24 h. During long-term filtration, an initial flux (JI) and a specific rate of flux change (ΔJ) were applied as calculated using Eq. (1) and (2): JI ¼ 0:7 JC

ð1Þ

dJ JI dt

ð2Þ

ΔJ ¼

3. Results and discussion 3.1. Effect of pore size on critical flux As shown in Table 1, JC increases with increasing membrane pore size. Bilad et al. (2012c) also reported that JC can increase with increasing membrane pore size. As shown in Fig. 2, in the ascending phase, when the three pore size membranes were at the same flux, the 0.1μm membrane had the lowest TMP and the 0.03-μm membrane had the highest. However, when the fluxes reached the respective JC, 0.1 μm membrane had the highest TMP value, whereas 0.05 μm

Table 1 Critical flux and membrane hydraulic resistance (Rm) of membranes with different pore sizes and TMP when critical flux achieved. Pore size (μm)

0.03

0.05

0.1

Critical flux (L m−2 h−1) TMP (kPa) Rm (m−1) Water flux recovery (%)

20.0 0.5 1.5 × 1012 75.6 ± 3.9

25.0 0.6 8.5 × 1011 80.4 ± 4.1

42.0 2.5 2.9 × 1011 93.5 ± 5.1

Please cite this article as: Zhao, F., et al., The filtration and fouling performance of membranes with different pore sizes in algae harvesting, Sci Total Environ (2017), http://dx.doi.org/10.1016/j.scitotenv.2017.02.035

F. Zhao et al. / Science of the Total Environment xxx (2017) xxx–xxx

Ascending phase

(a) 25

Flux (L.m-2 h-1)

20

Descending phase 5 4.5 4 3.5 3 2.5 2 1.5 1 0.5 0

Flux 25 TMP TMP

15 10 5 0 10

(b)

30 25

12.5

15

17.5

20 22.5 20 Flux-step (L.m-2h-1)

17.5

15

12.5

TMP (kPa)

4

10

3

Flux

2.5

25 TMP TMP Peristaltic

Filtration tank

15

1.5

10

1

5

0.5 Membrane

0 10 12.5

15 17.5

20 22.5

0

25 27.5 25 22.5 Flux-step (L.m-2 h-1)

20 17.5

15 12.5

10

(c)

45 40

Flux (L.m-2h-1)

35

TMP(kPa)

2

pump

9

Flux

8

TMPTMP 25

7

30

6

25

5

20

4

15

3

10

2

5

1

0

TMP (kPa)

Flux (L.m-2 h-1)

20

0 15 18 21 24 27 30 33 36 39 42 39 36 33 30 27 24 21 18 15 Flux-step (L.m-2 h-1)

Fig. 2. Flux–TMP profile of the improved flux-step method using different PVDF membranes: pore size of (a) 0.03, (b) 0.05, and (c) 0.1 μm.

SEM images do not indicate any difference. The surfaces of the three fouled membranes were completely covered by algae cells and EOM, but no foulants were visible on the surface of the clean membrane. However, as shown in Table 1, the flux recoveries of the 0.03-μm, 0.05-μm and 0.1-μm membranes were 75.6%, 80.4% and 93.5%, respectively when filtering the water using clean membranes. Smaller membrane pore sizes lead to more difficult membrane cleaning and lower water flux recovery. However, the flux recoveries of the membranes in the algae solution were not measured, and thus, this does not mean that the flux recovery had large discrepancies from the algae solution. Old membranes were not reapplied in other filtration experiments because of the relatively large discrepancies for flux recovery of clean water. 3.3. Effect of pore size on short-term filtration Continuous filtration experiments were performed to compare the filtration and fouling performance of the three membranes at an operational flux of 20 L m−2 h− 1. As shown in Fig. 4, for the 0.03-μm

membrane, TMP rapidly increases from 0 to 67.0 kPa in 1 h, at which point the membrane has been seriously fouled, and filtration cannot continue. The filtration experiments for 0.05- and 0.1-μm membranes were also conducted for 1 h. For the 0.05- and 0.1-μm membranes, TMP increased to 54.0 and 38.5 kPa in 1 h, respectively. During the filtration, the smaller-pore membrane always had a higher TMP than the larger-pore membrane, indicating the increase in pore size from 0.03 to 0.1 μm, led to obviously reduced membrane fouling. 3.4. Mechanism of pore size on membrane fouling Lewis acid-base force, electrostatic double layer force, and Lifshitzvan der Waals force (i.e., “extended Derjaguin, Landau, Verwey, Overbeek” (XDLVO) forces), permeate drag force and inertial lift force are the main forces acting on algae cells, which can determine if algae are deposited on membranes in algae filtration (Kang et al., 2004; Zhao et al., 2016). According to the interfacial forces equations (Kang et al., 2004), Lewis acid-base force, electrostatic double layer force, Lifshitz-van der Waals force and inertial lift force vary only slightly

Please cite this article as: Zhao, F., et al., The filtration and fouling performance of membranes with different pore sizes in algae harvesting, Sci Total Environ (2017), http://dx.doi.org/10.1016/j.scitotenv.2017.02.035

F. Zhao et al. / Science of the Total Environment xxx (2017) xxx–xxx

5

Fig. 3. SEM images of fouled and clean membranes.

with membrane pore size, and changes in pore size only influence the permeate drag force. Therefore, in the present study, only permeate drag force was investigated. During membrane filtration, increasing FD can improve the speed of algae moving towards the membrane, which leads to the acceleration of deposition onto the membrane (Kang et al., 2004; Subramani and Hoek, 2008). The permeate drag force can be shown as: F D ¼ −6πμ w ra vw φH φH ¼

 0:5 2Rm r a þ 1:0722 3

ð3Þ ð4Þ

where FD is the permeate drag force at the membrane surface (nN), μw v is the solution viscosity (Pa s), w is the permeate water velocity (m s−1), ra is algae cell radius (m), φH is a hydrodynamic correction

Fig. 4. The changes in TMP during the short-term filtration.

factor (Kang et al., 2004; Subramani and Hoek, 2008) and Rm is the membrane hydraulic resistance based on the Darcy law (m−1). The minus sign in Eq. (3) demonstrates that the permeate drag force is an attractive force and the direction is towards the membrane surface. As observed from Eqs. (3) and (4), the permeate drag force is a function of viscosity and Rm. Table 1 shows the membrane hydraulic resistance of the three different membranes. At 35 °C, the viscosity of the algae solution was approximately 1.3 mPa s. As shown in Fig. 5(a), the size distribution of the microalgae mainly ranged from 2 to 10 μm. The average algae cell size is approximately 4 μm. From Fig. 5(b), the diameter of algae cells is mostly distributed from 3 to 5 μm. Thus, in this study the average diameter of algae cells was considered to be 4 μm (r a = 2 μm). At 20 L m − 2 h − 1 , with increasing separation between algae cell and membrane, the variations of FD for membranes with three different pore sizes are shown in Fig. 6. As the increase of membrane pore size R m decreased, F D acting on algae declined. As shown in Fig. 6, the value of F D increased with the decrease in separation distance between algae cell and membrane, and the F D of large-pore membranes was always lower than that of small-pore membranes. The decrease in F D resulted in less algae deposited on the membrane, leading to less membrane fouling. This is why at the same flux, membranes with larger pores had a low TMP. The calculation of permeate drag force could theoretically explain why smaller pores are more easily contaminated. EOM is another important factor in membrane fouling. Membranes with smaller pores could retain more EOM on the membrane surface and pore interiors, resulting in more serious membrane fouling compared to membranes with larger pores. However, the combined effect of algae cells and EOM could result in more serious membrane fouling compared with the independent effect of algae cells or EOM (Qu et al., 2012). In this study, for membranes with smaller pores there were more algae cells deposited on the membrane surface, and thus the combined effect of algae cells and EOM on membrane fouling was more obvious. The size of EOM is very small compared with algae cells, and the drag force acting on EOM can be negligible; thus, the drag force on EOM was not analyzed.

Please cite this article as: Zhao, F., et al., The filtration and fouling performance of membranes with different pore sizes in algae harvesting, Sci Total Environ (2017), http://dx.doi.org/10.1016/j.scitotenv.2017.02.035

6

F. Zhao et al. / Science of the Total Environment xxx (2017) xxx–xxx

Fig. 5. (a) The size distribution of the microalgae and (b) SEM of fouled membrane.

3.5. Effect of pore size on long-term filtration To further verify the effect of pore size on performance during prolonged filtration, long-term filtration experiments were carried out. Fig. 7 shows the flux and TMP variation of three different membranes with time. The specific rate of change of the flux can be divided into three stages: the rapidly declining stage (ΔJ N 0.1), the slowly declining stage (0.01 b ΔJ b 0.1) and the stable stage (ΔJ b 0.01). The 0.7JC value was selected as JI, and the JI of the 0.03-, 0.05- and 0.1-μm membrane were 14.1, 17.4 and 29.4 L m−2 h−1, respectively. This is in contrast to some membrane filtrations, in which flux showed a rapid reduction during the initial 60 min, followed by a relatively slow reduction period (Tan et al., 2014; Zhang et al., 2014). For 0.1- and 0.03-μm membranes the flux and TMP were at a stable state from 0 to 60 min; from 60 to 120 min, the flux was at a sharply decreasing stage. As time went on, the flux/TMP experienced a slow decrease/increase stage followed by a relatively stable period. However, a strange phenomenon was observed for the 0.05-μm membrane: when the filtrations finished the value of TMP was very high (57.5 kPa) compared with the 0.03-μm membrane (8.8 kPa) and 0.1-μm membrane (16.5 kPa); whereas during almost the entire filtration process, flux remained in a relatively slow decline, and the average specific rate of change of flux was very low (0.013 h−1) compared with the 0.03-μm membrane (0.02 h−1) and the 0.1-μm membrane (0.024 h−1). When the continuous filtration experiments finished, the final flux did not obviously increase as pore size increased from 0.05 to 0.1 μm (shown in Fig. 7(b) and (c)). The 0.1-μm membrane had the highest

Fig. 6. The variations of drag force with the separation distance between algae cell and membrane surface at the same flux of 20 L m−2 h−1 (the average algae cells radius is 2 μm, the viscosity of algae solution is 1.3 mPa s, the permeate water velocity is 4.16 × 10−6 m s−1).

average specific rate of change of flux. This may be because the 0.1-μm membrane had a higher run flux and thus generated more contaminants, leading to an increase in the reduction rate of the flux (Lee et al., 2013). However, it remains unclear why the 0.05-μm membrane had a very high TMP but a low flux decline rate. Although the 0.1 μm membrane had the highest flux decline rate, after 24 h of filtration the flux was at a stable state, after which the flux decline became very slow; moreover, the final flux of the 0.1-μm membrane

Fig. 7. Variations in flux and TMP over time during continuous filtration using three different pore size ultrafiltration membranes: (a) 0.03, (b) 0.05, and (c) 0.1 μm.

Please cite this article as: Zhao, F., et al., The filtration and fouling performance of membranes with different pore sizes in algae harvesting, Sci Total Environ (2017), http://dx.doi.org/10.1016/j.scitotenv.2017.02.035

F. Zhao et al. / Science of the Total Environment xxx (2017) xxx–xxx

(12.8 L m−2 h−1) was still higher compared with the 0.03- and 0.05-μm membranes (7.4 and 12.0 L m−2 h−1, respectively). Thus, the 0.1 μm membrane was suitable for algae harvesting. 4. Conclusions In this study, critical fluxes improved from 20.0 to 42.0 L m−2 h−1 as the pore size increased from 0.03 to 0.1 μm. In continuous filtration experiments, although the 0.1-μm membrane had the highest flux decline rate, it was still suitable for algae harvesting compared with the 0.03and 0.05-μm membranes. Thus, the 0.1-μm membrane was suitable for harvesting C. pyrenoidosa. With increases in pore size, the membrane hydraulic resistance decreased, which caused the permeate drag force to decrease, resulting in less algae adsorption onto the membrane and less membrane fouling. Calculation of the permeate drag force demonstrated why large pore sizes are beneficial for membrane filtration. Acknowledgements This research was financially supported by the National Natural Science Foundation of China (No. 51478324, 51625804) and the Program of Shanghai Subject Chief Scientist (No. 14XD1403700). References Ahmad, A.L., Mat Yasin, N.H., Derek, C.J.C., Lim, J.K., 2013. Harvesting of microalgal biomass using MF membrane: kinetic model, CDE model and extended DLVO theory. J. Membr. Sci. 446, 341–349. Batista, A.P., Gouveia, L., Bandarra, N.M., Franco, J.M., Raymundo, A., 2013. Comparison of microalgal biomass profiles as novel functional ingredient for food products. Algal Res. 2 (2), 164–173. Bhave, R., Kuritz, T., Powell, L., Adcock, D., 2012. Membrane-based energy efficient dewatering of microalgae in biofuels production and recovery of value added coproducts. Environ. Sci. Technol. 46 (10), 5599–5606. Bilad, M.R., Vandamme, D., Foubert, I., Muylaert, K., Vankelecom, I.F., 2012a. Harvesting microalgal biomass using submerged microfiltration membranes. Bioresour. Technol. 111, 343–352. Bilad, M.R., Mezohegyi, G., Declerck, P., Vankelecom, I.F., 2012b. Novel magnetically induced membrane vibration (MMV) for fouling control in membrane bioreactors. Water Res. 46 (1), 63–72. Bilad, M.R., Vandamme, D., Foubert, I., Muylaert, K., Vankelecom, I.F.J., 2012c. Harvesting microalgal biomass using submerged microfiltration membranes. Bioresour. Technol. 111, 343–352. Chu, H., Zhao, F., Tan, X., Yang, L., Zhou, X., Zhao, J., Zhang, Y., 2016. The impact of temperature on membrane fouling in algae harvesting. Algal Res. 16, 458–464. Davis, R.E., Fishman, D.B., Frank, E.D., Johnson, M.C., Jones, S.B., Kinchin, C.M., Skaggs, R.L., Venteris, E.R., Wigmosta, M.S., 2014. Integrated evaluation of cost, emissions, and resource potential for algal biofuels at the national scale. Environ. Sci. Technol. 48 (10), 6035–6042. Drews, A., Lee, C.-H., Kraume, M., 2006. Membrane fouling - a review on the role of EPS. Desalination 200 (1–3), 186–188. Drexler, I.L.C., Yeh, D.H., 2014. Membrane applications for microalgae cultivation and harvesting: a review. Rev. Environ. Sci. Biotechnol. 13 (4), 487–504.

7

Hwang, T., Kotte, M.R., Han, J.I., Oh, Y.K., Diallo, M.S., 2015. Microalgae recovery by ultrafiltration using novel fouling-resistant PVDF membranes with in situ PEGylated polyethyleneimine particles. Water Res. 73, 181–192. Kang, S., Subramani, A., Hoek, E., Deshusses, M., Matsumoto, M., 2004. Direct observation of biofouling in cross-flow microfiltration: mechanisms of deposition and release. J. Membr. Sci. 244 (1–2), 151–165. Lee, H., Kim, S.G., Choi, J.S., Kim, S.K., Oh, H.J., Lee, W.T., 2013. Effects of water temperature on fouling and flux of ceramic membranes for wastewater reuse. Desalin. Water Treat. 51 (25–27), 5222–5230. Li, Y., Bilad, M.R., Vankelecom, I.F.J., 2014. Application of a magnetically induced membrane vibration (MMV) system for lignocelluloses hydrolysate filtration. J. Membr. Sci. 452, 165–170. Milledge, J.J., Heaven, S., 2012. A review of the harvesting of micro-algae for biofuel production. Rev. Environ. Sci. Biotechnol. 12 (2), 165–178. Nguyen, T., Roddick, F.A., Fan, L., 2012. Biofouling of water treatment membranes: a review of the underlying causes, monitoring techniques and control measures. Membranes 2 (4), 804–840. Pavez, J., Cabrera, F., Azócar, L., Torres, A., Jeison, D., 2015. Ultrafiltration of non-axenic microalgae cultures: energetic requirements and filtration performance. Algal Res. 10, 121–127. Qu, F., Liang, H., Tian, J., Yu, H., Chen, Z., Li, G., 2012. Ultrafiltration (UF) membrane fouling caused by cyanobacteria: fouling effects of cells and extracellular organics matter (EOM). Desalination 293, 30–37. Rickman, M., Pellegrino, J., Davis, R., 2012. Fouling phenomena during membrane filtration of microalgae. J. Membr. Sci. 423-424, 33–42. Rossignol, N., Vandanjon, L., Jaouen, P., Quemeneur, F., 1999. Membrane technology for the continuous separation microalgae/culture medium: compared performances of cross-flow microfiltration and ultra-filtration. Aquac. Eng. 20 (3), 191–208. Subramani, A., Hoek, E., 2008. Direct observation of initial microbial deposition onto reverse osmosis and nanofiltration membranes. J. Membr. Sci. 319 (1–2), 111–125. Tan, X., Chu, H., Zhang, Y., Yang, L., Zhao, F., Zhou, X., 2014. Chlorella pyrenoidosa cultivation using anaerobic digested starch processing wastewater in an airlift circulation photobioreactor. Bioresour. Technol. 170, 538–548. Uduman, N., Qi, Y., Danquah, M.K., Forde, G.M., Hoadley, A., 2010. Dewatering of microalgal cultures: a major bottleneck to algae-based fuels. J. Renewable Sustainable Energy 2 (1), 012701. van der Marel, P., Zwijnenburg, A., Kemperman, A., Wessling, M., Temmink, H., van der Meer, W., 2009. An improved flux-step method to determine the critical flux and the critical flux for irreversibility in a membrane bioreactor. J. Membr. Sci. 332 (1– 2), 24–29. Weschler, M.K., Barr, W.J., Harper, W.F., Landis, A.E., 2014. Process energy comparison for the production and harvesting of algal biomass as a biofuel feedstock. Bioresour. Technol. 153, 108–115. Wicaksana, F., Fane, A.G., Pongpairoj, P., Field, R., 2012. Microfiltration of algae (Chlorella sorokiniana): critical flux, fouling and transmission. J. Membr. Sci. 387–388, 83–92. Zhang, Y., Zhao, Y., Chu, H., Zhou, X., Dong, B., 2014. Dewatering of Chlorella pyrenoidosa using diatomite dynamic membrane: filtration performance, membrane fouling and cake behavior. Colloids Surf. B: Biointerfaces 113, 458–466. Zhao, F., Su, Y., Tan, X., Chu, H., Zhang, Y., Yang, L., Zhou, X., 2015a. Effect of temperature on extracellular organic matter (EOM) of Chlorella pyrenoidosa and effect of EOM on irreversible membrane fouling. Colloids Surf. B: Biointerfaces 136, 431–439. Zhao, F., Tan, X., Zhang, Y., Chu, H., Yang, L., Zhou, X., 2015b. Effect of temperature on the conversion ratio of glucose to Chlorella pyrenoidosa cells: reducing the cost of cultivation. Algal Res. 12, 431–435. Zhao, F., Chu, H., Su, Y., Tan, X., Zhang, Y., Yang, L., Zhou, X., 2016. Microalgae harvesting by an axial vibration membrane: the mechanism of mitigating membrane fouling. J. Membr. Sci. 508, 127–135. Zhou, S., Shao, Y., Gao, N., Li, L., Deng, J., Tan, C., Zhu, M., 2014. Influence of hydrophobic/ hydrophilic fractions of extracellular organic matters of Microcystis aeruginosa on ultrafiltration membrane fouling. Sci. Total Environ. 470-471, 201–207.

Please cite this article as: Zhao, F., et al., The filtration and fouling performance of membranes with different pore sizes in algae harvesting, Sci Total Environ (2017), http://dx.doi.org/10.1016/j.scitotenv.2017.02.035