Bioresource Technology 198 (2015) 789–796

Contents lists available at ScienceDirect

Bioresource Technology journal homepage: www.elsevier.com/locate/biortech

Harvesting Chlorella vulgaris by magnetic flocculation using Fe3O4 coating with polyaluminium chloride and polyacrylamide Yuan Zhao, Wenyan Liang ⇑, Lijun Liu, Feizhen Li, Qianlong Fan, Xiaoli Sun Beijing Key Lab for Source Control Technology of Water Pollution, College of Environmental Science and Engineering, Beijing Forestry University, No. 35 Qinghua East Road, Haidian District, Beijing 100083, China

h i g h l i g h t s
Polymers of PACl and PAM were combined with Fe3O4 to harvest Chlorella vulgaris.
The optimum dosing strategy was adding the composite PACl/Fe3O4 first and then PAM.
This strategy could harvest 99% of cells in less than 0.5 min.
This strategy could avoid the influence of pH and algal organic matter.
Charge neutralization of PACl/Fe3O4 and sweeping of PAM were the main mechanism.

a r t i c l e

i n f o

Article history: Received 12 July 2015 Received in revised form 24 September 2015 Accepted 25 September 2015

Keywords: Microalgae harvesting Magnetic flocculation Polyaluminium chloride Polyacrylamide Composite flocculant

a b s t r a c t The harvesting of Chlorella vulgaris was investigated by magnetic flocculation, where the natural magnetite was used as magnetic seeds and the polyaluminium chloride (PACl) and polyacrylamide (PAM) were used as the coating polymer on the Fe3O4 surface. The composite modes of PACl, PAM, and Fe3O4 and their effects on harvesting were studied. The results showed that adding the composite PACl/Fe3O4 first (at (0.625 mmol Al/L)/(10 g/L)) followed by the addition PAM (at 3 mg/L) was the optimum dosing strategy. Following this strategy, 99% of cells could be harvested in less than 0.5 min, and it could overcome negative impacts from pH and algal organic matter. Compared to PACl, f-potentials of PACl/Fe3O4 were found to be increased substantially from 4.9–8.5 mV to 1.5–19.5 mV at pH range 2.1–12.3. The charge neutralization of PACl/Fe3O4 and sweeping of PAM play an important role in magnetic harvesting of microalgal cells. Ó 2015 Elsevier Ltd. All rights reserved.

1. Introduction Increasing energy demands coupled with environmental concerns, regarding the generation of greenhouse gases during fossil fuel combustion, have drawn attention toward the development of renewable biofuels (Foley et al., 2011). Microalgae are considered to be one of the most promising feedstock option due to its low cost, high lipid content, and bulk biomass (Chisti, 2007). However, small size of microalgae cells and their colloidal stability in suspension result in low efficiency and expensive consumption of energy in harvesting and dewatering; this has always been a major obstacle for the algae-to-fuel approach (Slade and Bauen, 2013).

⇑ Corresponding author. Tel./fax: +86 10 62336615. E-mail address: [email protected] (W. Liang). http://dx.doi.org/10.1016/j.biortech.2015.09.087 0960-8524/Ó 2015 Elsevier Ltd. All rights reserved.

To obtain the microalgae paste for drying, a primary harvesting step is often used to concentrate the dilute microalgae suspension from 0.02% to 0.06% total suspended solids (TSS) into a slurry with 2–7% TSS, then followed by a secondary dewatering step that produces a paste of 15–25% TSS (Uduman et al., 2010). The desired microalgae concentration is important to the drying process for lipid extraction. The major methodologies currently used for harvesting and recovering microalgae include gravity sedimentation (Smith and Davis, 2013), filtration and screening (Bilad et al., 2013), flotation (Kurniawati et al., 2014), electrophoresis (Uduman et al., 2010), centrifugation (Dassey and Theegala, 2013), and flocculation (Vandamme et al., 2013). In sedimentation processes in a mass force field, the settling behavior of particles strongly depends on physico-chemical properties, concentration, and size distribution of the particles. Although it is considered to be a low-cost and simple technique, the technique is only suitable for microalgae larger than 70 lm (Gultom and Hu, 2013). For

790

Y. Zhao et al. / Bioresource Technology 198 (2015) 789–796

smaller microalgae (5–20 lm), flocculants must be used to achieve higher settling velocities (Smith and Davis, 2013). Flotation is a physico-chemical type of gravity separation process, where air or gas bubbles attach to algae cells and help in carrying them to surrounding liquid’s surface (Chen et al., 2011). It is suitable for harvesting small and unicellular algae and is reportedly more effective and beneficial than sedimentation (Chen et al., 2011), although is often limited by the energy requirements of bubble production. As for electrophoresis techniques, the microalgae can be separated from water-based solutions through movement using an electric field, including electrolytic coagulation, electrolytic flotation, and electrolytic flocculation (Uduman et al., 2010). In filtration process, algal culture runs through filters that allow water and small molecules to pass, while inhibiting passage of larger microalgae cells. Such filtration methods can achieve an almost complete retention of biomass, however, consumes considerable amount of energy due to the application of high pressure and liquid velocity (Bilad et al., 2013). Centrifugation is the predominant method for harvesting microalgae as it does not require addition of chemicals. It can often be used in the secondary dewatering process. However, algal broth often requires preconcentration to reduce energy demands for centrifugation and the associated costs. Due to relatively low cost and energy consumption, flocculation receives greater attention for primary harvesting than the other aforementioned techniques. During flocculation, single cells can be concentrated from microalgae suspensions to form large aggregates that can be easily separated from medium by simple gravity sedimentation (Vandamme et al., 2013). This can be achieved using several approaches including traditional chemical flocculation methods widely used in water treatment to novel ideas based on the biology of microalgae such as bioflocculation (González-Fernán dez and Ballesteros, 2012). Magnetic flocculation is a newly emerging technology, which applies modified magnetic particles that attach directly to microalgal cells and later separate them from medium using an external magnetic field (Wang et al., 2014). In the modified technique, magnetite acts as the core coated with a protective organic layer carrying specific functional groups ensuring selective separations and/or targeting, resulting in the formation of composite magnetic beads (Prochazkova et al., 2013). The coating organics can bind magnetic core with microalgae cells and make the cells to aggregate together. The binding organics in microalgae harvesting include poly(diallyldimethylammonium chloride) (Lim et al., 2012; Toh et al., 2012), chitosan (Lee et al., 2013; Toh et al., 2014), diethylaminoethyl (Prochazkova et al., 2013), polyethylenimine (Ge et al., 2015; Prochazkova et al., 2013), and cationic polyacrylamide (Wang et al., 2014). Such magnetic separation technique can achieve more than 90% of cell recovery in less than 5 min during the harvesting processes of microalgae. The magnetic harvesting combines flocculation and magnetic separation in a single process, offering quick, simple, energy-efficient, and cost effective advantages. In addition, external magnetic field enables to concentrate magnetically modified cells into compact slurry and remove large amounts of bulk liquid in a short time. In the magnetic flocculation technique, organic polymer used for coating and the dosage of magnetic composite are the main factors influencing the harvesting efficiency (Prochazkova et al., 2013). The magnetic particles, pH value, reaction time, and stirring speed can also affect cell recovery (Wang et al., 2014; Xu et al., 2011). Due to negative surface charges on most microalgae, cationic polymers are often used as coating chemicals. In previous studies, only one cationic polymer has been tested to make composite with Fe3O4. Using two organic polymers for dosing is an approach that not only improves separation effects, but also shortens settling time in the conventional flocculation (Ahmad et al.,

2008). However, less reports are available till date regarding two polymers dosing strategy to make composite with magnetite in harvesting microalgae. Polyaluminium chloride (PACl) is often applied coupled with polyacrylamide (PAM), and the combination has already exhibited better efficiency in removing algae and pollutant during water treatment (Ahmad et al., 2008; Lou et al., 2013). Hence, PACl and PAM were selected as the two polymers to combine with natural magnetite in this study. The present study examines the composite modes of PACl, PAM, and Fe3O4 and their effects on harvesting of oleaginous microalgae Chlorella vulgaris (C. vulgaris) from freshwater. Natural magnetite was selected as the magnetic material and characterized its surface properties using field emission scanning electron microscopy (FESEM), energy-dispersive spectroscopy (EDS), and X-ray diffraction (XRD) patterns. The harvesting efficiency of different dosages, dosing strategy of composite flocculants and the ratio of polymer to Fe3O4 were also investigated. Moreover, the f-potentials of the composite flocculants and the supernatant after flocculation were measured. The influence of pH of microalgal broth, settling time, and algal organic matter (AOM) on cell harvesting were also evaluated. 2. Methods 2.1. Algae culture C. vulgaris (FACHB-31) was obtained from the Institute of Hydrobiology, Chinese Academy of Sciences. It was cultured in a 10 L round bottom flask containing 7 L sterilized BG-11 media. Algae culture was incubated at room temperature (22–25 °C) under a fluorescent lamp with an illumination of 34 lmol/m2/s (light:night = 20:4) with constant air feeding at 6.5  102 m3/h. After 24 days, algae broth was used directly without any pretreatment. The pH of the final broth was 8.4. 2.2. Preparation of composite flocculants Magnetite powder was purchased from the Baotou Bally Ken Industrial Technology Co., Ltd., China. PACl was obtained from Adamas Reagent Co., Ltd., China, while PAM (molecular weight P 3,000,000) was purchased from Sinopharm Chemical Reagent Co., Ltd., China. Before preparing composite flocculants, magnetite powder was washed three times using distilled water and then dried at room temperature. Stock solutions of concentrations 125 mmol Al/L and 1.00 g/L were prepared for PACl and PAM, respectively. The composite flocculants of PACl/Fe3O4 and PAM/Fe3O4 were synthesized by mixing magnetite powder with PACl or PAM stock solution at a given ratio with continuous shaking for 5 min in a shaking incubator. To obtain the composite PAM/PACl/Fe3O4 and PACl/PAM/Fe3O4, PAM and PACl stock solutions were added to the PACl/Fe3O4 and PAM/Fe3O4 composite solution, respectively and mixed thoroughly. The dosages of composite flocculants are described using ratio x/y or x/y/z for expressing the doses of PACl, PAM or Fe3O4; the concentration units used for calculations are mmol Al/L, mg/L and g/L, respectively. 2.3. Magnetic flocculation The experiments were carried out in a jar test apparatus (ZR4-6, Zhongrun Water Industry Technology Development Co., Ltd., China). The flocculant was first added to 400 mL algae suspension (2–3  1010 cells/L), freshly taken from broth. After stirring for 1 min at 500 rpm, the beaker was placed inside a magnetic field created by a cubic NdFeB permanent magnet

Y. Zhao et al. / Bioresource Technology 198 (2015) 789–796

(50 mm L  50 mm W  25 mm H) with a magnetic induction intensity of 3800 G (rB < 40 T/m). During magnetic precipitation, 10 ml sample was collected from 3 cm below the water surface at different time intervals (0.5 min, 3 min, 5 min, and 20 min) to determine the cell density. The harvesting efficiency was calculated using the Eq. (1):

Harvesting efficiency ð%Þ ¼

x0  xt x0

ð1Þ

where x0 (cells/L) and xt (cells/L) denote initial cell density and cell density at time t, respectively. Microalgal harvesting was first tested at different dosage and ratio of composite flocculant PACl/Fe3O4. The dosage of PACl was fixed at either 0.625 mmol Al/L or 1.25 mmol Al/L, while the dosage of Fe3O4 was varied from 1 g/L to 30 g/L. There were three strategies in magnetic flocculation to test the effects of PAM on harvesting — ‘‘immobilized-on” strategy, ‘‘attached-to” strategy (Lim et al., 2012), and ‘‘hybrid” strategy. In the ‘‘immobilized-on” test, composite flocculant, such as PACl/ Fe3O4, PAM/Fe3O4, PAM/PACl/Fe3O4, PACl/PAM/Fe3O4, was dosed directly to microalgae broth. In the ‘‘attached-to” test, PACl, Fe3O4, and/or PAM were dosed separately and sequentially. The ‘‘hybrid” strategy combined the above two strategies; the composite PACl/Fe3O4 or PAM/Fe3O4 was dosed first followed by PAM or PACl (expressed as PACl/Fe3O4 + PAM and PAM/Fe3O4 + PACl, respectively). The final concentrations of PACl, PAM and Fe3O4 in all the three strategies were kept the same, i.e., 0.625 mmol Al/L, 3 mg/L, and 10 g/L, respectively. To investigate the effects of pH on harvesting, the pH of algal broth was adjusted in the range 6–10 using either 1 M HCl or 1 M NaOH. In the test of AOM, the C. vulgaris culture (2–3  1010 cells/L) was centrifuged to remove AOM, and the algal pallet was resuspended in the same volume of medium BG-11. All the experiments were performed induplicate. Mean values were presented as results. 2.4. Analytical methods The surface properties of Fe3O4 particles were investigated using FE-SEM (FEI QUANTA 200 FEG) and EDS (S-4800, HITACHI, Japan). XRD patterns were recorded on a Rigaku Dmax/2400 X-ray diffractometer equipped with graphite monochromatized Cu Ka (k = 0.1542 nm), operating at 40 kV and 80 mA; a scanning rate of 0.02°/s was used in 2h ranges from 20° to 90°. The cell density was enumerated in a counting chamber under an optical microscope (BX 51, Olympus, Japan), and the cell images were obtained with a charge-coupled-device (CCD) camera (DM 320, Chongqing Aote Optical Instrument Co., Ltd., China). The pH value was detected using a digital pH meter (PB-10, BSISL, China). Zeta potentials of C. vulgaris cells (1–2  109 cells/L), magnetite (10 g/L), PACl (0.625 mmol Al/L), PAM (3 mg/L), PACl/Fe3O4 (0.625/10), PAM/Fe3O4 (3/10), PAM/PACl/Fe3O4 (3/0.625/10), and PACl/PAM/Fe3O4 (0.625/3/10) were measured in model environments (distilled water, pH 2–12) at 25 °C using Zetasizer Nano (2000HSA, Malvern, United Kingdom). The supernatant after flocculation was detected directly. All the samples were measured five times to obtain average values. 3. Results and discussion 3.1. Characterization of the magnetic particles FE-SEM and EDS approach can provide detailed information regarding morphology and surface texture of individual particle along with elemental sample composition. The minimum particle size of Fe3O4 was approximately 0.25 lm, while the maximum size

791

was greater than 40 lm (Fig. S1). Because of the pulverization of the magnetic powder using natural giant magnet, magnetic particles were irregular and angular in shape with uneven size distribution. The EDS spectra indicated that the magnetic particles were mainly consist of Fe and O, as well as a small amount of Zn, Ti, Mg, Al, Ca, Si, S, and Cl (Table S1). The XRD pattern of the magnetic particles showed peaks that were mainly characteristic of Fe3O4 (Fig. S2). Most of the magnetic materials used for flocculation experiments are generally prepared by chemical reactions such as co-precipitation, thermal decomposition, metal reduction, hydrothermal synthesis, and solvothermal synthesis (He et al., 2014). These particles possess a near-spheroidal shape and are evenly distributed. Due to extremely small particle dimension that even reaches to nanoscale, lab-made magnetite particles are ideal for specific functions, such as protein immobilization, drug delivery, waste-water treatment, and metal removal (Li et al., 2011; Ma et al., 2003). However, these materials usually show instability during storage and usage as they are prone to oxidation, especially when particle diameter is less than 8 nm (Wang et al., 2005). For magnetic separation, it is important to generate magnetic force (using magnetic field) that is large enough to overcome other opposing forces such as Brownian motion and viscous drag (Yavuz et al., 2006). In this study, we also investigated finer magnetic particles (10–50 nm) from the same magnetite company. However, these fine particles obtained only 52% harvesting efficiency, lower than 95% of the above particles (0.25–40 lm) at the same dosage of PACl/Fe3O4 (0.625/10). The aggregation of nano materials in aqueous solution may influence the dispersion of nano-Fe3O4 particles and then the coating of PACl, resulting in the decrease of harvesting efficiency. 3.2. Zeta potentials of composite flocculants Fig. 1 depicts zeta potentials of C. vulgaris cells (1–2  109 cells/ L), Fe3O4, and flocculants. Measurements of f-potentials revealed that C. vulgaris cells were all negatively charged, ranging from 4.6 mV to 16.8 mV at pH 2.1–12.3. The cell membrane of Chlorella sp. contains ACOOH, ASH, and AOH functional groups and C@O and NAH bonds. Deprotonation of these groups generates a net negative charge on cell surface at natural pH of water (Toh et al., 2014). The negative charge of microalgae often need cationic flocculants to attach and separate them from aqueous solution (Lou et al., 2013). As also shown in Fig. 1a, the magnetite particles were negatively charged from 17.5 mV to 25.6 mV at 10 g/L (pH 2.1–12.3). However, f-potential for aqueous PACl (0.625 mmol Al/ L) were positive when the pH was below the isoelectric point (at pH 8.5). After mixing PACl with Fe3O4, f-potential of the composite PACl/Fe3O4 was increased drastically to 1.5–19.5 mV in the whole range of studied pH values, and the isoelectric point of PACl/ Fe3O4 shifted to approximately 13.1. Fourier transform infrared (FTIR) spectra of Fe3O4 also indicated the change of surface property of the coated magnetite particles and suggested the successful coating of magnetite surface by PACl (Fig. S3). Due to the presence of bare Fe and O atoms on the particle surface, it adsorbs OH or H+ ions from aqueous solution (Fig. S4). When Fe3O4 was dissolved in distilled water to obtain a solution of 10 g/L, the pH values of the resulting solution increased from 6.5 to 9.5, suggesting that Fe3O4 particles capture more H+ ions on their surfaces. Hydrogen ion (H+) adsorbed on Fe3O4 surface forms hydroxyl group (AOH), and thus particles become negatively charged (Ma et al., 2003). The hydroxyl group on Fe3O4 surface can react with PACl through ion exchange, e.g., [AAln(OH)m] group from PACl displaces H+ ions from hydroxyl to form FeAOAAln(OH)m. Such ion exchange reaction changes f-potential of Fe3O4 from negative to positive.

Y. Zhao et al. / Bioresource Technology 198 (2015) 789–796

30

30

20

20

10

10

0

2

4

6

8

10

12

-10

14 pH

-20 -30 -40

Chlorella vulgaris PACl Fe3O4 PACl/Fe3O4

(a)

Zeta potential (mV)

Zeta potential (mV)

792

0

2

4

6

8

-10 -20

PAM PAM/Fe3O4

-30

PAM/PACl/Fe3O4 PACl/PAM/Fe3O4

10

12

pH

14

-40

(b)

Fig. 1. (a) The zeta potentials of C. vulgaris cells (1–2  109 cells/L), Fe3O4, PACl, and composite flocculant PACl/Fe3O4. (b) The zeta potentials of PAM and the composite flocculants PAM/Fe3O4, PAM/PACl/Fe3O4, and PACl/PAM/Fe3O4. The final concentrations of PACl, PAM and Fe3O4 were 0.625 mmol Al/L, 3 mg/L and 10 g/L, respectively.

Magnetite surface is capable of concentrating more functional groups from PACl, making magnetite surface more positive than that of PACl solution. As evident from Fig. 1b, f-potentials of PAM were more positive than those of Fe3O4, and exhibited an isoelectric point of 3.8. After mixing PAM with Fe3O4, the isoelectric point of PAM/Fe3O4 also shifted rightly to 6.5. FTIR spectra also showed the successful coating of magnetite surface by PAM (Fig. S3). The ion exchange character of Fe3O4 particles depends on the type of functional group present on their surface (Prochazkova et al., 2013). Particles with strong ion exchange groups, such as polyethylenimine, show a high isoelectric point to magnetic beads, while particles bearing weak ion exchange group, such as diethylaminoethyl, exhibit a lower isoelectric point (Prochazkova et al., 2013). Compared to [AAln(OH)m] group of PACl, PAM having ACH2ACHAC(O)ANH2 group exhibits relatively weak ion exchange ability for hydroxyl on Fe3O4 surface. Consequently, isoelectric point of PAM/Fe3O4 was lower than that of PACl/Fe3O4. The lower charge of PAM resulted in the f-potentials of the composite PAM/PACl/Fe3O4 and PACl/PAM/Fe3O4 descending obviously. Furthermore, the mixing order among Fe3O4, PACl, and PAM influenced the coating components on Fe3O4 surface obviously (Fig. S3). Due to the first mixing of Fe3O4 with PAM, the PAM occupied more adsorption sites on Fe3O4 surface than PACl, which resulted in the f-potential of PACl/PAM/Fe3O4 lower than that of PAM/PACl/Fe3O4. 3.3. PACl/Fe3O4 flocculation Before investigating the flocculating effects of PACl/Fe3O4, experiments about the exploring PACl dosage were conducted (Fig. S5). When cell density of C. vulgaris culture was 2–3  1010 cells/L, harvesting efficiency increased with PACl concentration, and the optimum dosage occurred at around 2.50 mmol Al/L with 96% harvesting efficiency. Following this, harvesting efficiency decreased as the dosage increased further and reached 52% at 6.25 mmol Al/L. The flocculation processes were complete within 5–10 min for dosages between 0.625 mmol Al/L and 3.75 mmol Al/L. In contrast, flocculation was not complete even after 20 min for the high 6.25 mmol Al/L dosage. Fig. 2 shows the harvesting efficiency of composite PACl/Fe3O4. At PACl dosage 0.625 mmol Al/L, harvesting efficiency increased with the addition of magnetite, obtaining the highest efficiency of 95% with 10 g/L Fe3O4; although both at 2 g/L and 5 g/L, the efficiency was only marginally less with near identical trends. When the concentration of Fe3O4 increased to 20 g/L, harvesting efficiency of PACl/Fe3O4 drastically decreased to 62%. In spite of such

descent, 20 g/L Fe3O4 enhanced the harvesting efficiency dramatically from 52% to 97% for 6.25 mmol Al/L PACl (Fig. S6). However, the similar magnetite dosage effects were not observed when PACl concentration was set higher at 1.25 mmol Al/L. Harvesting efficiency increased in presence of magnetite with very less variation in efficiency at different concentrations. The optimal dosage of Fe3O4 was the same as the suboptimal values identified for 0.625 mmol Al/L PACl, i.e., 2 g/L and 5 g/L of Fe3O4, achieving approximately 99% harvesting efficiency. Compared to the optimal dosage of 2.50 mmol Al/L PACl alone, PACl/Fe3O4 reduced PACl consumption largely, and it is obvious that there exists an optimum matching of PACl and Fe3O4 dosage in magnetic recovery of microalgae. It is well-known that flocculation can be achieved by four common mechanisms acting alone or in combination, namely double layer compression, charge neutralization, bridging and colloid entrapment or sweeping (Wang et al., 2014). The fpotential measurement (Fig. 1a) showed that the C. vulgaris cells were negatively charged, while PACl/Fe3O4 was positively charged at the tested pH values (2.1–12.3). Charge neutralization and double layer compression occur when two particles carry opposite charges (Wang et al., 2014), and this helps microalgae cells and PACl/Fe3O4 to form PACl/Fe3O4-cells floc due to electrostatic attraction. Compared to 0.625 mmol Al/L PACl alone, more positive PACl/Fe3O4 (0.625/10) achieved greater harvesting. The f-potentials of the supernatants (Table S2) after PACl/Fe3O4 flocculation showed that the dosage of 0.625/10 nearly achieved the ideal charge neutralization (f-potential = 0.48 mV), while the f-potential of the dosage 0.625/20 still remained in the negative point of 15.6 mV. Due to the negative charge on Fe3O4, the excess Fe3O4 can make the f-potential of PACl/Fe3O4 (0.625/20) negative. The more negative charge strengthened the repulsive force between composite flocculant and algal cells and resulted in the worse effects of composite flocculant. This suggests that charge neutralization plays a very important role among the four common mechanisms. In addition, the higher ratio of polymer to magnetite provides more adsorption sites to cells; therefore, a composite with a higher polymer to magnetite ratio exhibited higher harvesting efficiency with same dosage (Lee et al., 2013). That was another reason for achieving higher harvesting efficiency with the dosage 0.625/10 compared to that of 0.625/20. The kinetics of PACl/Fe3O4 harvesting was comparable to that of PACl harvesting. As soon as the magnet was brought under the beaker containing the mixture, it pulled PACl/Fe3O4-cells flocs toward the beaker bottom very rapidly. PACl/Fe3O4 of 0.625/10 completed

793

Y. Zhao et al. / Bioresource Technology 198 (2015) 789–796

90

90

80 70 60 50 40 0.625/0 0.625/2 0.625/10

30 20 0

5

10 15 Time (min)

0.625/1 0.625/5 0.625/20 20

(a)

Harvesting efficiency (%)

100

Harvesting efficiency (%)

100

80 70 60 50 1.25/0 1.25/2 1.25/10 1.25/30

40 30

1.25/1 1.25/5 1.25/20

20 0

5

10 15 Time (min)

20

(b)

Fig. 2. Effects of the composite flocculant PACl/Fe3O4 on harvesting C. vulgaris (2–3  1010 cells/L) from culture broth. (a) PACl dosage of 0.625 mmol Al/L. (b) PACl dosage of 1.25 mmol Al/L. The ratios are described as PACl dosage (mmol Al/L) to Fe3O4 dosage (g/L) of the composite PACl/Fe3O4.

the settling processes in approximately 5 min (Fig. 2a). The flocculation could be finished within 3 min with the ratio increasing to 1.25/10 (Fig. 2b). For the dosage of 6.25/20 (Fig. S6), the magnetic flocculation process was completed in less than 0.5 min. Irrespective of the separation effects, the flocculation process reaches the maximum efficiency within the first 0.5–5 min before reaching a plateau, less than 5–10 min in conventional flocculation using PACl. As the magnetically tagged microalgae cells approaches the magnetic field source, the microalgae cells collide, leading to cells chaining that further enhanced the magnetic removal rate (Toh et al., 2012). Therefore, formation of long chaining and larger floc during settling course could also be increased by increasing magnetic particle concentration and coated PACl adsorption sites. This would in turn favor the increase of settling velocity at the dosage of 6.25/20. 3.4. The dosing strategy for PAM Common tagging strategies involved in magnetic flocculation can be implemented either through ‘‘immobilized-on” or ‘‘attached-to” methods. In the ‘‘immobilized-on” approach, naked-magnetite surface is first functionalized with polyelectrolyte binder and then attached on cell surface. On the other hand, the ‘‘attached-to approach” involves coating of cells by polymer binder followed by the attachment of naked magnetite (Toh et al., 2014). In the ‘‘immobilized-on” processes, composite PAM/PACl/Fe3O4 achieved 99% of harvesting efficiency after 20 min, as shown in (Fig. 3a). However, the harvesting efficiency drops from 99% to 89% for another composite PACl/PAM/Fe3O4 with similar composition (PACl, PAM and Fe3O4) as that of PAM/PACl/Fe3O4. In the ‘‘attached-to” approach, the dosing of PACl first followed by Fe3O4 achieved only 72% harvesting efficiency, much less than that of PACl/Fe3O4 (95%), as evident from Fig. 3b. However, adding PAM after PACl and Fe3O4, the harvesting efficiency increased to 91%. Moreover, Fig. 3b shows that the hybrid strategy PACl/Fe3O4 + PAM, where composite PACl/Fe3O4 was dosed first followed by PAM, achieved a harvesting efficiency of 99%, higher than that of PACl/Fe3O4 (95%). The harvesting efficiency of another hybrid strategy PAM/Fe3O4 + PACl, where composite PAM/Fe3O4 was dosed first followed by PACl, achieved 86% harvesting, much higher than that of PAM/Fe3O4 (52%). The hybrid strategies PACl/Fe3O4 + PAM and PAM/Fe3O4 + PACl had the similar recovery effects of PAM/ PACl/Fe3O4 and PACl/PAM/Fe3O4 as obtained for ‘‘immobilizedon” technique. Therefore, application of two polymers in magnetic

flocculation can enhance microalgae harvesting greatly, irrespective of dosing strategy. Although the PACl/Fe3O4 + PAM and PAM/ PACl/Fe3O4 exhibited the nearly same effects of harvesting (99%), the settling process of PACl/Fe3O4 + PAM could be fulfilled in less than 0.5 min, much less than that of PAM/PACl/Fe3O4 (20 min). It can be concluded that the dosing strategies of magnetic flocculants not only influence harvesting efficiency but also influence settling velocity. The optimum method would be the addition of composite PACl/Fe3O4 to microalgae suspension first, followed by the addition PAM. Among the two strategies employed to harvest microalgae, the ‘‘attached-to” technique usually suffers from lower efficiency compared to the ‘‘immobilized-on” technique (Lim et al., 2012). In the ‘‘immobilized-on” approach, naked-magnetite particles agglomerates to form large particle clusters induced by van der Waals force and magnetostatic attraction prior to their attachment onto microalgae cells (Lim et al., 2012; Toh et al., 2014). The aggregation of magnetite particles inhibits to achieve good dispersibility and thus significantly lowers particles-to-cells ratio, which in turn substantially suppresses separation efficiency (Toh et al., 2014). The ‘‘immobilized-on” approach does not associated with such problem; hence, this is the preferred technique as polymer coating on particle’s surface forms an electrosteric barrier that enhances its colloidal stability (Toh et al., 2014). Moreover, when PACl is first added to culture, it adheres onto microalgal cells by the electrostatic attraction. The Fe3O4 particles cannot be embedded in PACl-cells flocs, inhibiting magnetite to fulfill its role as the floc core (Eroglu et al., 2013). In addition to the dosing strategy, the f-potentials of the PACl/Fe3O4 were more positive than those of PACl (Fig. 1a). The attachment of magnetic particles onto microalgae, either through ‘‘immobilized-on” technique or by ‘‘attached-to” strategy, is achieved via electrostatic interaction. The more is the intensity of electrostatic force, easier would be the attachment of particles to microalgae and the aggregation to form the floc. All these factors are responsible for the better performance of PACl/Fe3O4 compared to PACl and Fe3O4 in ‘‘attached-to” approach. The dosing strategy becomes more complicated than the conventional one, when two polymers are used for magnetic flocculation. PAM is an amorphous, water-soluble polymer formed from acrylamide. It is usually used as a flocculating agent in water treatment, because PAM can readily cross-link and bind with suspended solids. It can also be used as a flocculant aid to increase the viscosity of water (creating a thicker solution) or to encourage flocculation of particles present in water (Li et al., 2006). In this study, when PAM was used as the flocculant to be mixed with Fe3O4,

794

Y. Zhao et al. / Bioresource Technology 198 (2015) 789–796

90

90

80 70 60 50 40 PACl/Fe3O4 PAM/PACl/Fe3O4 PACl/PAM/Fe3O4 PAM/Fe3O4

30 20

0

5

10 15 Time (min)

20

Harvesting efficiency (%)

100

Harvesting efficiency (%)

100

80 70 60 dosing PACl first, then Fe3O4

50

dosing Fe3O4 first, then PACl dosing PACl, Fe3O4, PAM in order

40

dosing PACl/Fe3O4 first, then PAM

30 20

dosing PAM/Fe3O4 first, then PACl

0

5

10 Time (min)

(a)

15

20

(b)

Fig. 3. The harvesting efficiency of different composite flocculants in ‘‘immobilized-on strategy” (a) and ‘‘attached-to” and ‘‘hybrid” strategy (b). The final concentrations of PACl, PAM and Fe3O4 were 0.625 mmol Al/L, 3 mg/L and 10 g/L, respectively.

the PAM/Fe3O4 achieved only 52% of harvesting. However, when it was dosed after PACl/Fe3O4 as the flocculant aid, the harvesting efficiency rose to 99%. The charge neutralization between PACl/Fe3O4 and cells made them form the PACl/Fe3O4-cells floc. High molecular weight of PAM enabled it to produce threads or fibers that attached to PACl/Fe3O4-cells floc, capturing and binding them together to form PAM–PACl/Fe3O4-cells floc. The magnetic core of PAM–PACl/Fe3O4-cells floc and the surrounding magnetic field strengthen the bridging and entrapment effects of microalgae, resulting in enhancing floc settling separation. However, it was observed that while mixing PAM with PACl/Fe3O4 to obtain PAM/ PACl/Fe3O4, its f-potential decreased than those of PACl/Fe3O4 (Fig. 1b). This suggests that PAM might partially neutralize PACl/ Fe3O4 positive charge, and microalgal cells could not assembled onto Fe3O4 surface efficiently via electrostatic attraction. It took 20 min to achieve 99% of harvesting for PAM/PACl/Fe3O4 because of its lower f-potentials. When Fe3O4 was first mixed with PAM, f-potentials of another composite PACl/PAM/Fe3O4 prepared using immobilized-on strategy kept on decreasing; the isoelectric point of PACl/PAM/Fe3O4 was obtained as 3.7. The composite PACl/PAM/Fe3O4 finally achieved only 86% harvesting efficiency. The results indicate that the PAM should be dosed separately from PACl/Fe3O4. The microalgal aggregates were formed when the magnetically tagged microalgal cells went through a magnetic separation process induced by an externally applied magnetic field; clusters of micrometer-size were obtained. The morphology of aggregates depended on the composite flocculant and its dosing strategy (Fig. S7). On visual inspection, flocs produced via hybrid strategy of PACl/Fe3O4 + PAM were larger than those in PACl/Fe3O4 flocculation (Fig. S7). PACl/Fe3O4 was first attached to microalgal cell mainly through electrostatic interaction, while PAM attached to PACl/Fe3O4-cells floc mainly by bridging and sweeping. The number of freely PACl/Fe3O4-cells floc reduced with time as they continuously attached onto PAM and were caught-up in the sweeping flow of PAM–PACl/Fe3O4-cells along their magnetic settling pathway. The PAM–PACl/Fe3O4-cells floc became obviously larger and contained more magnetite particles in the cluster matrix. Since the magnetic force is directly proportional to the magnetic volume of a particle (Toh et al., 2012), the larger floc matrix experienced much larger force and could be collected more rapidly and easily compared to smaller particles. After magnetic treatment of PACl/Fe3O4 + PAM, the previously greenish C. vulgaris culture turned to be crystal clear, suggesting effective recovery of microalgae (99% harvesting). The supernatant after magnetic

flocculation using composite PACl/Fe3O4 still showed light green color, although 95% recovery was achieved. The crystal supernatant in PACl/Fe3O4 + PAM flocculation was collected, and the BG-11 was added to make a kind of recycling medium. The normal growth of C. vulgaris in the recycling medium indicated that the flocculants in hybrid strategy did not influence the supernatant reuse in further cultivation of microalgae (Fig. S8). 3.5. The effects of pH The pH of a solution is an important factor that often influences conventional flocculation intensively. Fig. 4 depicts the effects of pH on magnetic flocculation. Alkaline environment was more beneficial for microalgal harvesting using either composite PACl/Fe3O4 or PACl/Fe3O4 + PAM approach in the studied range of pH (6.04–10.1). Moreover, the pH values of the microalgae broth had nearly no influence on the magnetic floc movement to the bottom. The addition of PAM not only improved the harvesting efficiency to 99%, but also shortened the settling time from 5 min in PACl/Fe3O4 process to less than 0.5 min in the PACl/Fe3O4 + PAM process under alkaline environment. After magnetic flocculation, the final pH values of the supernatant dropped approximately 0.2–0.4, suggesting that the magnetic flocculation did not change the pH value of culture solution too much. The effects of pH on magnetic flocculation observed in this study were not inaccordance with the literature regarding microalgae harvesting. In case of naked Fe3O4, a low pH of 4 has been found to be the optimum for harvesting Botryococcus braunii cells, while the highest recovery efficiency of Chlorella ellipsoidea is achieved at pH 7.0 (Xu et al., 2011). The higher efficiencies have also been observed in harvesting C. vulgaris at pH 4 and pH 2 for Fe3O4 coated with polyethylenimine and Fe3O4 coated with diethylaminoethyl, respectively (Prochazkova et al., 2013). When pH was increased to 10, the efficiency of the diethylaminoethyl did not exceed 65%. In this study, PACl/Fe3O4 or PACl/Fe3O4 + PAM could achieve 99% harvesting efficiency in the studied pH range. This could be result from the highly positive f-potentials of the composite PACl/Fe3O4 and even positively charge at the strong alkaline pH. Under acidic condition, major hydrolyzed Al species of PACl are positively charged monomer ions such as Al3+, Al (OH)2+, and Al(OH)2+. As pH increases to neutral and alkaline, Al species transform into Al(OH)3 and Al(OH) 4 in the presence of OH, resulting in decreasing f-potentials (Li et al., 2014). However, the exchange of Al(OH)3 or Al(OH) 4 to the AOH on Fe3O4 surface results in the neutralization of OH from Al(OH)3 or Al(OH) 4 to

795

Y. Zhao et al. / Bioresource Technology 198 (2015) 789–796

100

95

95

90

Harvesting efficiency (%)

Harvesting efficiency (%)

100

85 80 75 70

6.04 9.01

65 60

0

5

7.05 10.1

8.45

10 15 Time (min)

20

90 85 80 75 6.04 9.01

70

7.05 10.1

8.45

65 60

0

5

10 15 Time (min)

20

(b)

(a)

Fig. 4. C. vulgaris harvesting efficiency from culture broth (2–3  1010 cells/L) at different pH using composite PACl/Fe3O4 (a) and composite PACl/Fe3O4 + PAM (b). The final concentrations of PACl, PAM and Fe3O4 were 0.625 mmol Al/L, 3 mg/L and 10 g/L, respectively.

H+ from AOH. The magnetite surface concentrated more molecules of Al(OH)2+ or Al(OH)2+ and behaved more positively than Al(OH)3 or Al(OH) 4 . The intensive positive charge on PACl/Fe3O4 helped it overcoming the influence of pH and obtaining the effective recovery of cells. Algal growth often causes changes in pH of the culture media, and pH reaches a range of 8.5–11.0 (Dubinsky and Rotem, 1974), suggesting that composite PACl/Fe3O4 and PACl/Fe3O4 + PAM can achieve perfect flocculation for these cultures. 3.6. The effects of algal organic matter Algal organic matter (AOM), secreted during microalgal metabolism, are released into aqueous solution, and this causes disturbance in conventional flocculation and increase consumption of flocculants (Wang et al., 2013). Fig. 5 shows the effects of AOM on magnetic flocculation. In the conventional flocculation of PACl alone, AOM decreased the harvesting efficiency from 93% to 86%. Even in the magnetic flocculation using PACl/Fe3O4, AOM decreased the harvesting efficiency from 99% to 95%. However, AOM did not interfere in cell harvesting in the hybrid magnetic flocculation using PACl/Fe3O4 + PAM. The cells recovery was achieved 99% in less than 0.5 min for both the samples with and without AOM. Although the microalgal broth contained high cell

100

Harvesting efficiency (%)

90 80 70 60

PACl (without AOM) PACl (with AOM) PACl/Fe3O4 (without AOM)

50 40 30 20 10

0

5

density (2–3  1010 cells/L), the magnetic strategy of PACl/Fe3O4 + PAM could overcome the negative effects of AOM. In the cultivation system, microalgae excrete a wide range of compounds such as proteins, neutral and charged polysaccharides, nucleic acids, lipids, and small molecules (Vandamme et al., 2012). Polysaccharides comprise the major fraction of AOM and they possess negatively charged carboxyl groups that interact with the positively charged metal or polymer flocculants, making them unavailable for microalgae flocculation and thus resulting in a higher coagulant demand (Mikulec et al., 2015). Beside polysaccharides, protein fraction can react with metal flocculant forming complexes and thus also becomes unavailable for flocculation of microalgae (Vandamme et al., 2012). The presence of AOM leads to increase in dosage to 2–10-fold for obtaining the similar harvesting efficiency (Prochazkova et al., 2015; Vandamme et al., 2012). When AOM is removed, the flocculant consumption is drastically reduced and sedimentation of large flocs occurs very fast. AOM concentration increases with increase in the concentration of algal biomass and often reaches a maximum value at the stationary phase (Mikulec et al., 2015). After 24 days of cultivation, C. vulgaris in this study entered the stationary phase. The AOM from C. vulgaris cells brought about adverse impact on the magnetic flocculation of PACl/Fe3O4. However, the addition of PAM made PAM–PACl/Fe3O4-cells floc larger than PACl-cells and PACl/ Fe3O4-cells floc. The larger floc aggregated more magnetite particles inside it. Since the magnetophoretic force is directly proportional to the magnetic volume of a particle, larger particles not only experience a much larger force but can also fight against all other competing forces that introduce hindrance at varying degrees. Therefore, the addition of PAM in PACl/Fe3O4 + PAM strategy is meaningful. It not only results in larger floc, faster sedimentation rate, and higher harvesting efficiency, but also overcomes the influence of pH and AOM in magnetic flocculation.

PACl/Fe3O4 (with AOM) PACl/Fe3O4+PAM (without AOM)

4. Conclusions

PACl/Fe3O4+PAM (with AOM)

Microalgae harvesting efficiency can be enhanced by employing PACl and PAM in magnetic flocculation. The composite mode and dosing strategy have an intensive impact on cell recovery. The optimum strategy to achieve 99% harvesting of C. vulgaris in less than 0.5 min is to dose the composite PACl/Fe3O4 first, followed by dosing PAM. The increase in f-potentials of PACl/Fe3O4 is the main reason for the improvement in harvesting. The main mechanism in the PACl/Fe3O4 + PAM process involves charge neutralization of

10 Time (min)

15

20

Fig. 5. The effects of algal organic matter (AOM) on magnetic flocculation of C. vulgaris culture (2–3  1010 cells/L). The final concentrations of PACl, PAM and Fe3O4 were 0.625 mmol Al/L, 3 mg/L and 10 g/L, respectively.

796

Y. Zhao et al. / Bioresource Technology 198 (2015) 789–796

PACl/Fe3O4 and bridging and entrapment of PAM. Such strategy can overcome the negative influences of pH and AOM. Acknowledgements This work is financially supported by the Scientific Research Foundation for the Returned Overseas Chinese Scholars, State Education Ministry. Appendix A. Supplementary data Supplementary data associated with this article can be found, in the online version, at http://dx.doi.org/10.1016/j.biortech.2015.09. 087. References Ahmad, A.L., Wong, S.S., Teng, T.T., Zuhairi, A., 2008. Improvement of alum and PACl coagulation by polyacrylamides (PAMs) for the treatment of pulp and paper mill wastewater. Chem. Eng. J. 137, 510–517. Bilad, M.R., Discart, V., Vandamme, D., Foubert, I., Muylaert, K., Vankelecom, I.F., 2013. Harvesting microalgal biomass using a magnetically induced membrane vibration (MMV) system: filtration performance and energy consumption. Bioresour. Technol. 138, 329–338. Chen, C.Y., Yeh, K.L., Aisyah, R., Lee, D.J., Chang, J.S., 2011. Cultivation, photobioreactor design and harvesting of microalgae for biodiesel production: a critical review. Bioresour. Technol. 102, 71–81. Chisti, Y., 2007. Biodiesel from microalgae. Biotechnol. Adv. 25, 294–306. Dassey, A.J., Theegala, C.S., 2013. Harvesting economics and strategies using centrifugation for cost effective separation of microalgae cells for biodiesel applications. Bioresour. Technol. 128, 241–245. Dubinsky, Z., Rotem, J., 1974. Relations between algal populations and the pH of their media. Oecologia 16, 53–60. Eroglu, E., D’Alonzo, N.J., Smith, S.M., Raston, C.L., 2013. Vortex fluidic entrapment of functional microalgal cells in a magnetic polymer matrix. Nanoscale 5, 2627– 2631. Foley, J.A., Ramankutty, N., Brauman, K.A., Cassidy, E.S., Gerber, J.S., Johnston, M., Mueller, N.D., O’Connell, C., Ray, D.K., West, P.C., Balzer, C., Bennett, E.M., Carpenter, S.R., Hill, J., Monfreda, C., Polasky, S., Rockstrom, J., Sheehan, J., Siebert, S., Tilman, D., Zaks, D.P., 2011. Solutions for a cultivated planet. Nature 478, 337–342. Ge, S.J., Agbakpe, M., Wu, Z.Y., Kuang, L.Y., Zhang, W., Wang, X.Q., 2015. Influences of surface coating, UV irradiation and magnetic field on the algae removal using magnetite nanoparticles. Environ. Sci. Technol. 49, 1190–1196. González-Fernández, C., Ballesteros, M., 2012. Microalgae autoflocculation: an alternative to high-energy consuming harvesting methods. J. Appl. Phycol. 25, 991–999. Gultom, S.O., Hu, B., 2013. Review of microalgae harvesting via co-pelletization with filamentous fungus. Energies 6, 5921–5939. He, J.C., Huang, M.Y., Wang, D.M., Zhang, Z.M., Li, G.K., 2014. Magnetic separation techniques in sample preparation for biological analysis: a review. J. Pharm. Biomed. Anal. 101, 84–101. Kurniawati, H.A., Ismadji, S., Liu, J.C., 2014. Microalgae harvesting by flotation using natural saponin and chitosan. Bioresour. Technol. 166, 429–434. Lee, K., Lee, S.Y., Na, J.G., Jeon, S.G., Praveenkumar, R., Kim, D.M., Chang, W.S., Oh, Y. K., 2013. Magnetophoretic harvesting of oleaginous Chlorella sp. by using biocompatible chitosan/magnetic nanoparticle composites. Bioresour. Technol. 149, 575–578.

Li, H.B., Bi, S.D., Liu, L., Dong, W.F., Wang, X., 2011. Separation and accumulation of Cu(II), Zn(II) and Cr(VI) from aqueous solution by magnetic chitosan modified with diethylenetriamine. Desalination 278, 397–404. Li, R.H., Gao, B.Y., Huang, X., Dong, H.Y., Li, X.C., Yue, Q.Y., Wang, Y., Li, Q., 2014. Compound bioflocculant and polyaluminum chloride in kaolin-humic acid coagulation: factors influencing coagulation performance and floc characteristics. Bioresour. Technol. 172, 8–15. Li, T., Zhu, Z., Wang, D.S., Yao, C.H., Tang, H.X., 2006. Characterization of floc size, strength and structure under various coagulation mechanisms. Powder Technol. 168, 104–110. Lim, J.K., Chieh, D.C., Jalak, S.A., Toh, P.Y., Yasin, N.H., Ng, B.W., Ahmad, A.L., 2012. Rapid magnetophoretic separation of microalgae. Small 8, 1683–1692. Lou, I., Gong, S.Y., Huang, X.J., Liu, Y.J., Mok, K.M., 2013. Coagulation optimization using ferric and aluminum salts for treating high algae and high alkalinity source water in a typical North-China plant. Destin. Water Treat. 51, 3361– 3370. Ma, M., Zhang, Y., Yu, W., Shen, H.Y., Zhang, H.Q., Gu, N., 2003. Preparation and characterization of magnetite nanoparticles coated by amino silane. Colloid Surf. A 212, 219–226. Mikulec, J., Polakovicˇová, G., Cvengroš, J., 2015. Flocculation using polyacrylamide polymers for fresh microalgae. Chem. Eng. Technol. 38, 595–601. Prochazkova, G., Podolova, N., Safarik, I., Zachleder, V., Branyik, T., 2013. Physicochemical approach to freshwater microalgae harvesting with magnetic particles. Colloid Surf. B 112, 213–218. Prochazkova, G., Kastanek, P., Branyik, T., 2015. Harvesting freshwater Chlorella vulgaris with flocculant derived from spent brewer’s yeast. Bioresour. Technol. 177, 28–33. Slade, R., Bauen, A., 2013. Micro-algae cultivation for biofuels: cost, energy balance, environmental impacts and future prospects. Biomass Bioenergy 53, 29–38. Smith, B.T., Davis, R.H., 2013. Particle concentration using inclined sedimentation via sludge accumulation and removal for algae harvesting. Chem. Eng. Sci. 91, 79–85. Toh, P.Y., Yeap, S.P., Kong, L.P., Ng, B.W., Chan, D.J.C., Ahmad, A.L., Lim, J.K., 2012. Magnetophoretic removal of microalgae from fishpond water: feasibility of high gradient and low gradient magnetic separation. Chem. Eng. J. 211–212, 22–30. Toh, P.Y., Ng, B.W., Chong, C.H., Ahmad, A.L., Yang, J.W., Derek, C.J.C., Lim, J.K., 2014. Magnetophoretic separation of microalgae: the role of nanoparticles and polymer binder in harvesting biofuel. RSC Adv. 4, 4114–4121. 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. Renew. Sustain. Energy 2, 1–15. Vandamme, D., Foubert, I., Fraeye, I., Muylaert, K., 2012. Influence of organic matter generated by Chlorella vulgaris on five different modes of flocculation. Bioresour. Technol. 124, 508–511. Vandamme, D., Foubert, I., Muylaert, K., 2013. Flocculation as a low-cost method for harvesting microalgae for bulk biomass production. Trends Biotechnol. 31, 233– 239. Wang, C.M., Baer, D.R., Thomas, L.E., Amonette, J.E., Antony, J., Qiang, Y., Duscher, G., 2005. Void formation during early stages of passivation: initial oxidation of iron nanoparticles at room temperature. J. Appl. Phys. 98, 1–8. Wang, L., Liang, W.Y., Yu, J., Liang, Z.X., Ruan, L.L., Zhang, Y.C., 2013. Flocculation of Microcystis aeruginosa using modified larch tannin. Environ. Sci. Technol. 47, 5771–5777. Wang, S.K., Wang, F., Hu, Y.R., Stiles, A.R., Guo, C., Liu, C.Z., 2014. Magnetic flocculant for high efficiency harvesting of microalgal cells. ACS Appl. Mater. Interfaces 6, 109–115. Xu, L., Guo, C., Wang, F., Zheng, S., Liu, C.Z., 2011. A simple and rapid harvesting method for microalgae by in situ magnetic separation. Bioresour. Technol. 102, 10047–10051. Yavuz, C.T., Mayo, J.T., Yu, W.W., Prakash, A., Falkner, J.C., Yean, S.J., Cong, L.L., Shipley, H.J., Kan, A., Tomson, M., Natelson, D., Colvin, V.L., 2006. Low-field magnetic separation of monodisperse Fe3O4 nanocrystals. Science 314, 964– 967.