Bioresource Technology 151 (2014) 383–387

Contents lists available at ScienceDirect

Bioresource Technology

journal homepage: www.elsevier .com/locate /bior tech

Short Communication

Floc characteristics of Chlorella vulgaris: Influence of flocculation modeand presence of organic matter

0960-8524/$ - see front matter � 2013 Elsevier Ltd. All rights reserved.http://dx.doi.org/10.1016/j.biortech.2013.09.112

⇑ Corresponding author. Tel.: +32 56 246257; fax: +32 56 246999.E-mail address: Dries.Vandamme@kuleuven-kulak.be (D. Vandamme).

Dries Vandamme ⇑, Koenraad Muylaert, Ilse Fraeye, Imogen FoubertKU Leuven Kulak, Laboratory Aquatic Biology, E. Sabbelaan 53, 8500 Kortrijk, Belgium

h i g h l i g h t s

� Floc characteristics are depending on flocculation mode.� AOM influences the floc characteristics, especially the concentration factor.� Floc characteristics seem to be most optimal for cationic starch flocculation.

a r t i c l e i n f o

Article history:Received 25 July 2013Received in revised form 22 September 2013Accepted 25 September 2013Available online 9 October 2013

Keywords:DewateringDOMNOMMicroalgaeBiomass

a b s t r a c t

Floc characteristics such as settling velocity, concentration factor and floc size were studied for five dif-ferent flocculation modes for Chlorella: aluminum sulphate, electro-coagulation–flocculation, chitosan,cationic starch and pH induced flocculation. These floc characteristics were influenced by the flocculationmode, which depends on the coagulation mechanism: adsorption – charge neutralization, sweeping orbridging. Secondly, the influence of the presence of AOM was evaluated. The presence of AOM led toan increase between 1.5 and 5-fold in needed flocculant dosage for all flocculation modes. This resultedparticularly in a comparable decrease of the concentration factor. The floc characteristics upon floccula-tion using cationic starch were least affected by the presence of AOM, while flocculation using chitosanwas most affected. The impact on floc characteristics is an important parameter next to flocculation effi-ciency to consider in the assessment of flocculation-based harvesting of microalgae.

� 2013 Elsevier Ltd. All rights reserved.

1. Introduction natively, cationic biopolymers, such as chitosan or the cheaper

Microalgae are rich in lipids and proteins and are assumed toachieve a much higher areal productivity than traditionalagricultural crops. Therefore, they have received much interestas a source of biomass and biofuels (Georgianna and Mayfield,2012). However, one of the major challenges for economicallyfeasible production of biomass on commodity scale is energyefficient harvesting (Molina Grima et al., 2003). Due to theirsmall size (5–50 lm) and low concentration (0.05–0.5%),harvesting using centrifugation is not efficient from an energeticpoint of view. If microalgae could however be concentratedabout 30–50 times by coagulation–flocculation and gravity sedi-mentation prior to centrifugation, the energy demand for overallharvesting could be significantly reduced (Jorquera et al., 2010).Flocculation based preconcentration can be induced by the addi-tion of metal salts such as aluminum sulphate (alum) or by elec-trochemical release of metal ions from a sacrificial anode inelectro-coagulation–flocculation (Vandamme et al., 2011). Alter-

alternative cationic starch can also be used (Vandamme et al.,2010). Flocculation induced by increasing pH, which leads toprecipitation of magnesium hydroxides, was also evaluated aspromising in recent studies (Vandamme et al., 2012a).

Flocs that are formed by various coagulation mechanisms canexhibit different floc characteristics such as floc size, structureand density and this will affect important parameters such assettling velocity and concentration factor (Li et al., 2006). Thesettling velocity is a key parameter in the design of large scalesedimentation units while the concentration factor is importantto evaluate the water content of the particulate phase. The finalwater content of the particulate phase in addition to the floccu-lation efficiency will determine the overall efficiency of a floccu-lation mode. Most studies however only focus on the evaluationof flocculation based on flocculation efficiency. Recently, a fewstudies investigated these additional parameters for pH inducedand chitosan flocculation (i.e. Smith and Davis, 2012).However, those results were not related to coagulation mecha-nism, floc size and structure. Therefore, in this study, the influ-ence of coagulation mechanism on the floc characteristics ofChlorella vulgaris using five different flocculation modes wasevaluated.

384 D. Vandamme et al. / Bioresource Technology 151 (2014) 383–387

Microalgae are known to release significant amounts of organicmatter (AOM). In microalgae cultivation systems, AOM can amountto 60–80 mg C L�1 (Hulatt and Thomas, 2010). The major fractionof this AOM consists of neutral or charged polysaccharides, butother compounds such as proteins, nucleic acids, lipids and othersmall molecules can be present as well (Myklestad, 1995; Hender-son et al., 2010). In a recent study, the effect of AOM on flocculationefficiency of C. vulgaris, induced by five different flocculationmodes, was evaluated (Vandamme et al., 2012b). To the best ofour knowledge, however, no publications have focused on theeffect of the presence of AOM on the floc characteristics such assettling velocity, concentration factor and floc size. Therefore, inthis study, also the influence of the presence of AOM on those floccharacteristics of C. vulgaris using five different flocculation modeswas evaluated.

2. Methods

2.1. Cultivation of C. vulgaris

C. vulgaris (211-11b SAG, Germany) was cultivated in dechlori-nated tap water enriched with inorganic nutrients according to theconcentration of the Wrights cryptophyte medium (Vandammeet al., 2012a). Bubble column photobioreactors (30 L) were usedto cultivate the microalgae. The system was mixed by spargingwith 0.2 lm filtered air (5 L min�1) and pH was controlled at 8.5through 2–3% CO2 addition using a pH-stat system. Growth ofthe microalgae was monitored by measuring the absorbance at550 nm. Microalgal dry weight was determined gravimetricallyby filtration using Whatman glass fibre filters (Sigma–Aldrich)and drying until constant weight at 105 �C. Flocculation experi-ments were performed in the early stationary phase at a biomassconcentration of 0.3 g L�1.

2.2. Flocculation protocol

Five different flocculation modes were assessed: addition ofalum (AL), electro-coagulation-flocculation using aluminum an-odes (ECF), addition of chitosan (CH), addition of cationic starch(CS) and flocculation induced by high pH (pH). In addition, theinfluence of AOM on the flocculation process was studied for eachflocculation mode. To do so, flocculation of Chlorella was comparedin medium with and without AOM. To remove AOM, Chlorella wasseparated from the medium using centrifugation and resuspendedin fresh medium. This approach has already been used successfullyin previous studies and it has also been demonstrated that centri-fugation and resuspension of Chlorella, as a treatment as such, hasno influence on the flocculation efficiency (Vandamme et al.,2012b).

In a preliminary study, the flocculation parameters (pH anddosage) resulting in a flocculation efficiency higher than 85% weredetermined in jar tests on 100 ml scale. The results of this preli-minary study are reported in Supplementary Table 1. For each floc-culation mode, this was based on a protocol used in a previousstudy (Vandamme et al., 2012a). These parameters were then usedfor the remainder of the study to induce coagulation on 1 L scalefor floc characterization based on sedimentation and particle sizeanalysis.

For alum (Al2(SO4)3�18H2O; Sigma–Aldrich), the pH was ad-justed to 5.5 prior to and immediately after addition of the coagu-lant. For ECF, the setup described in a previous study was used(Vandamme et al., 2011). In short, this setup consisted of a 1-Lrectangular PVC reactor with an aluminum anode and an inert tita-nium oxide cathode and a power supply controller (EHQ PowerPS3010 DC). Current density in the experiments was set at

1.5 mA cm�2. For flocculation using chitosan (from crab shells,practical grade, Sigma–Aldrich), the pH of the suspension wasadjusted to 7.5 prior to and immediately after addition. For cationicstarch flocculation, a stock solution of 10 g L�1 Greenfloc 120(Hydra 2002 Research, Hungary) was prepared. The pH was notadjusted, as it has been shown that flocculation efficiency wasnot pH-dependent (Vandamme et al., 2010). For pH-induced floc-culation, 0.5 N NaOH was used to increase the pH.

2.3. Sedimentation analysis

After flocculation, using each of the flocculation modes, sedi-mentation was followed in order to calculate the settling velocity.This analysis could not be performed for ECF because flotationsimultaneously occurred with flocculation. For each flocculationtreatment, coagulation was induced in 1 L cylindrical vesselsaccording to the dosage obtained in the preliminary flocculationexperiment. The suspension was stirred for 15 min at 300 rpmusing an overhead stirrer. Then the suspension was transferredto 1 L imhoff cones, allowing sedimentation for 15 min. Images(Supplementary Fig. 1A) were automatically taken at fixed timeintervals using a webcam. Grey values were analysed as functionof height using ImageJ (NIH, USA) allowing to determine the heightof the interface between the suspended and particular phase ateach time step (Supplementary Fig. 1A; red line + SupplementaryFig. 1B). The corresponding height was then plotted as functionof sedimentation time (Supplementary Fig. 1C). The settling veloc-ity is defined as the velocity in cm s�1 to achieve complete biomasssettling without further observed increase of settled floc volume.

After 15 min of sedimentation, the suspension was allowed tosettle an additional 15 min to determine the concentration factor(CF) and the aggregated volume index (AVI). Both parameters arerelated to each other and provide information about the residualwater content of the particulate phase. The CF was determinedby dividing the total volume of 1000 ml by the volume of the par-ticulate phase after 30 min of sedimentation. The AVI was calcu-lated according to the method of (Javaheri and Dick, 1969). It isdefined as the volume in milliliters occupied by 1 g of algal suspen-sion in the particulate phase after 30 min of settling and is calcu-lated as:

AVI ðmL g�1Þ ¼ volume of settled biomass ðmL L�1Þmicroalgal biomass dry weight ðmg L�1Þ� 1000ðmg g�1Þ ð1Þ

2.4. Particle size analysis

After sedimentation, a subsample of the particulate phase wastaken and diluted 10 times. The flocs were then analysed using astereo zoom microscope (Olympus SZX10) and images were takenusing a camera (Lumenera Infinity 2). Particle size analysis wasconducted using ImageJ (NIH, USA). The original images weretransformed to 8 bit, the background was subtracted and particlessmaller than 100 were thresholded (Supplementary Fig. 2). Aftertransformation of the image, the average Feret’s diameter was cal-culated based on the determination of the Feret’s diameter of eachfloc (Supplementary Table 2).

3. Results and discussion

After inducing flocculation, the sedimentation of the Chlorellamicroalgae flocs was monitored. Fig. 1 presents floc front heightas function of time for four different flocculation modes containingAOM (indicated as AOM+; Fig. 1A) and resuspended in fresh med-ium without AOM (indicated as AOM�; Fig. 1B). This analysis could

Fig. 1. Sedimentation analysis for four flocculation modes: alum flocculation (AL), chitosan flocculation (CH), cationic starch flocculation (CS) and pH induced flocculation(pH) for Chlorella vulgaris (A) with (AOM+) and (B) without the presence of AOM (AOM�).

D. Vandamme et al. / Bioresource Technology 151 (2014) 383–387 385

not be performed for ECF because flotation occurred simulta-neously with flocculation. For all flocculation modes with andwithout the presence of AOM, the biomass settled within 15 minof sedimentation. Alum and chitosan flocculation resulted in afaster sedimentation than pH induced flocculation. In both treat-ments, cationic starch flocculation resulted in the slowestsedimentation. The settling velocities were calculated for each floc-culation mode and ranged between 0.06 and 0.6 cm s�1 with amaximal standard deviation of 12%. They confirmed the first obser-vations made on the basis of the graphs. When AOM was present,the settling velocity was the highest for chitosan flocculation(0.4 cm s�1), followed by alum flocculation (0.2 cm s�1) and thelowest for pH induced flocculation (0.09 cm s�1) and cationicstarch flocculation (0.06 cm s�1). Without AOM, settling velocitiesmostly increased, but the order as function of flocculation modein general remained the same. In the fast settling cases of chitosanand alum flocculation, the settling velocity increased to 0.6 cm s�1.For pH induced flocculation, it increased to 0.2 cm s�1. For cationicstarch however, the settling velocity decreased to 0.04 cm s�1.

Additional information about the residual water content of theparticulate phase is provided by calculating the concentration fac-tor (CF; Fig. 2A) and the aggregated volume index (AVI; Fig. 2B). Ahigh concentration factor corresponds to a low AVI. Using cationicstarch, the algal biomass could be concentrated more than 100times, in the presence of AOM. After removal of AOM, this even in-creased to 180 times. This corresponds with an AVI lower than25 mg g�1.When AOM was present, the other flocculation modes

Fig. 2. Concentration factor (A), AVI (B) and Feret’s diameter (C) for five flocculationflocculation (CH), cationic starch flocculation (CS) and pH induced flocculation (pH) for

resulted in a clearly lower CF of between 15 and 35, correspondingto a clearly higher AVI of between 70 and 140 mg g�1. In the ab-sence of AOM, the CF increased and the AVI decreased for all floc-culation modes. The Feret’s diameter of the settled flocs for the fivedifferent flocculation modes was also determined (Fig. 2C). It var-ied between 130 and 2300 lm. When AOM was present, the flocswere larger, except for flocculation by cationic starch. For chitosanflocculation in the presence of AOM, flocs with a diameter higherthan 2 mm were observed.

The present results show that the type of flocculation mode af-fects the microalgal floc characteristics such as settling velocity,concentration factor and floc size. Flocculation of microalgaeshould not only be effective in terms of flocculation efficiency,but also in terms of settling rate and concentration of the biomass.Those parameters are important in the design of a harvesting pro-cess including a secondary harvesting step using for example cen-trifugation. As a consequence of this, the delivery of a fast settledand high concentrated biomass is desirable before centrifugationin order to improve overall energy efficiency. Three mechanisms,i.e. charge neutralization, sweeping by precipitation enmeshmentand bridging have been demonstrated in coagulation processesand differences in these mechanisms may explain the influenceof flocculation mode on floc characteristics (Bache and Gregory,2007).

For flocculation modes using inorganic metal salts, the coagula-tion mechanism is absorption–charge neutralization or sweepingflocculation caused by precipitate enmeshment or a combination

modes: alum flocculation (AL), electro-coagulation–flocculation (ECF), chitosanChlorella vulgaris with (AOM+) and without the presence of AOM (AOM�).

386 D. Vandamme et al. / Bioresource Technology 151 (2014) 383–387

of both (Bache and Gregory, 2007). The usage of for example alumas coagulant introduces water-binding amorphous precipitates,which are present in large amounts especially when coagulationoccurs by sweeping flocculation. During absorption–chargeneutralization those precipitates are present in limited amountsand this has consequently less impact on the floc size (Knockeet al., 1987; Bache and Gregory, 2007). Absorption–charge neutral-ization may thus result in smaller flocs compared to flocculationbased on sweeping. For alum flocculation, operating conditionssuch as biomass density, coagulation pH and coagulant dosagedetermine the coagulation mechanism (Duan and Gregory, 2003).In this study, pH was adjusted and controlled at 5 for alum floccu-lation, which is known to facilitate coagulation dominated byabsorption–charge neutralization (Knocke et al., 1987; Garzon-Sanabria et al., 2012). In contrast, pH was not controlled duringEC flocculation using aluminum anodes. The initial pH was 8.5(Supplementary Table 1) and is known to rise as function of oper-ation time because of the hydroxide ions released at the cathode(Vandamme et al., 2011). In those conditions, sweeping floccula-tion is the dominant flocculation mechanism (Duan and Gregory,2003). This could explain the overall bigger flocs for the EC floccu-lation mode when compared to alum flocculation (Fig. 2C).

For flocculation induced by high pH, similar floc sizes as foralum flocculation were observed, but the floc compaction was infe-rior compared to alum flocculation (Fig. 2A and B). Positivelycharged magnesium hydroxides are involved in initiating coagula-tion by absorption–charge neutralization and partially also bysweeping flocculation, depending on pH and magnesium concen-tration (Vandamme et al., 2012b). With an increasing pH, largequantities of amorphous hydroxides are allowed to precipitate.These are known to have a high affinity for water arising frommechanical trapping and hydrogen bonding and thus affect thewater content of the particulate phase (Bache and Gregory, 2007).

Cationic biopolymers have been used in coagulation/floccula-tion processes as flocculation aid for water purification and areknown to lower flocculation dosage requirements, increase settle-ment and decrease sludge volume (Bolto and Gregory, 2007).Generally, flocculation using biopolymers is induced by a bridgingmechanism. This type of flocculation occurs when a polymer servesas a bridge after formation of a lot of particle-polymer–particleaggregates (Li et al., 2006). In general, the usage of polymers resultsin a higher effective density and thereby improved settling and ahigher floc compaction (Bache and Gregory, 2007). In this study,chitosan and cationic starch flocculation indeed resulted in a high-er compaction of the settled biomass, in absence of AOM (Fig. 2Aand B). Furthermore chitosan flocculation indeed had the highestsettling velocity, although cationic starch flocculation resulted inthe lowest settling velocity (Fig. 1). It must be noted that in thisstudy both biopolymers were used as primary coagulant, while set-tling improvement is mostly achieved when biopolymers acts asflocculant aid in combination with a primary coagulant such asalum (Bache and Gregory, 2007).

From previous studies (Bernhardt et al., 1989; Vandamme et al.,2012a; Zhang et al., 2012), it is known that the flocculant dosageneeded to achieve efficient flocculation is increased by the pres-ence of AOM. In this study, similar results were obtained (Supple-mentary Table 1). The present results however show that thepresence of AOM also affects the microalgal floc characteristicssuch as settling velocity, concentration factor and floc size. It wasfound that in the presence of AOM the settled biomass has a great-er water content, bigger floc size and that the settling velocity islower (Figs. 1 and 2). Moreover, the importance of the influenceof AOM on the floc characteristics clearly depended on the flocmode and therefore on the used flocculant. Besides interactingwith flocculant, AOM will also interact with the formed flocs. Inboth cases, the composition of AOM will play an important role

to understand those interactions. Henderson et al. (2010) charac-terized in detail the AOM produced by C. vulgaris. In the stationarygrowth phase, the AOM had a protein:carbohydrate ratio of 0.4 anda hydrophilicity of 71%. These hydrophilic attributes can explainthe increase in water content of the settled biomass when AOMwas present (Fig. 2A). Further research is however needed intodynamics of AOM production and composition and the distributionof AOM after sedimentation in order to fully understand the inter-actions between flocculant, AOM and microalgal flocs.

4. Conclusion

Floc characteristics such as settling velocity, concentrationfactor, aggregated volume index and floc size were studied for fivedifferent flocculation modes for C. vulgaris. This study showed thatcoagulant dose and type determine coagulation mechanism and bythis affect the floc characteristics of the settled microalgal biomass.The presence of AOM resulted in a lower concentration of the set-tled biomass. Our study thus showed that in addition to floccula-tion efficiency, the impact on the characteristics of the formedflocs needs to be assessed as well in the overall assessment of floc-culation based harvesting methods for microalgae.

Acknowledgements

The research presented in this paper was financially supportedby the Institute for the Promotion of Innovation by Science andTechnology–Strategic Basic Research (IWT–SBO) project Sunlightand the KU Leuven Research Coordination Office-Industrial Re-search Fund (DOC–IOF) project Algae-Tech. I.Fr. is a PostdoctoralResearcher funded by the Research Foundation – Flanders (FWO).We thank Akshay Mhetras, Isabel Vanoverberghe and Jan Lievensfor their practical support.

Appendix A. Supplementary data

Supplementary data associated with this article can be found, inthe online version, at http://dx.doi.org/10.1016/j.biortech.2013.09.112.

References

Bache, D.H., Gregory, R., 2007. Flocs in Water Treatment. IWA Publishing.Bernhardt, H., Schell, H., Hoyer, O., Lüsse, B., 1989. Influence of algogenic organic

substances on flocculation and filtration. WISA 1, 41–57.Bolto, B., Gregory, J., 2007. Organic polyelectrolytes in water treatment. Water Res.

41, 2301–2324.Duan, J., Gregory, J., 2003. Coagulation by hydrolysing metal salts. Adv. Colloid

Interface Sci. 100–102, 475–502.Garzon-Sanabria, A.J., Davis, R.T., Nikolov, Z.L., 2012. Harvesting Nannochloris

oculata by inorganic electrolyte flocculation: effect of initial cell density, ionicstrength, coagulant dosage, and media pH. Bioresour. Technol. 118, 418–424.

Georgianna, D.R., Mayfield, S.P., 2012. Exploiting diversity and synthetic biology forthe production of algal biofuels. Nature 488, 329–335.

Henderson, R.K., Parsons, S.A., Jefferson, B., 2010. The impact of differing cell andalgogenic organic matter (AOM) characteristics on the coagulation and flotationof algae. Water Res. 44, 3617–3624.

Hulatt, C.J., Thomas, D.N., 2010. Dissolved organic matter (DOM) in microalgalphotobioreactors: a potential loss in solar energy conversion? Bioresour.Technol. 101, 8690–8697.

Javaheri, A., Dick, R., 1969. Aggregate size variations during thickening of activatedsludge. J. Water Pollut. Control Fed., 197–214.

Jorquera, O., Kiperstok, A., Sales, E.A., Embiruçu, M., Ghirardi, M.L., 2010.Comparative energy life-cycle analyses of microalgal biomass production inopen ponds and photobioreactors. Bioresour. Technol. 101, 1406–1413.

Knocke, W.R., Hamon, J.R., Dulin, B.E., 1987. Effects of coagulation on sludgethickening and dewatering. J. Am. Water W. Assoc. 79, 89–98.

Li, T., Zhu, Z., Wang, D., Yao, C., Tang, H., 2006. Characterization of floc size, strengthand structure under various coagulation mechanisms. Powder Technol. 168,104–110.

Molina Grima, E., Belarbi, E.H., Acién Fernández, F.G., Robles Medina, A., Chisti, Y.,2003. Recovery of microalgal biomass and metabolites: process options andeconomics. Biotechnol. Adv. 20, 491–515.

D. Vandamme et al. / Bioresource Technology 151 (2014) 383–387 387

Myklestad, S.M., 1995. Release of extracellular products by phytoplankton withspecial emphasis on polysaccharides. Sci. Total Environ. 165, 155–164.

Smith, B.T., Davis, R.H., 2012. Sedimentation of algae flocculated using naturally-available, magnesium-based flocculants. Algal Res. 1, 32–39.

Vandamme, D., Foubert, I., Fraeye, I., Meesschaert, B., Muylaert, K., 2012a.Flocculation of Chlorella vulgaris induced by high pH: role of magnesium andcalcium and practical implications. Bioresour. Technol. 105, 114–119.

Vandamme, D., Foubert, I., Fraeye, I., Muylaert, K., 2012b. Influence of organicmatter generated by Chlorella vulgaris on five different modes of flocculation.Bioresour. Technol. 124, 508–511.

Vandamme, D., Foubert, I., Meesschaert, B., Muylaert, K., 2010. Flocculation ofmicroalgae using cationic starch. J. Appl. Phycol. 22, 525–530.

Vandamme, D., Pontes, S.C.V., Goiris, K., Foubert, I., Pinoy, L.J.J., Muylaert, K., 2011.Evaluation of electro-coagulation–flocculation for harvesting marine andfreshwater microalgae. Biotechnol. Bioeng. 108, 2320–2329.

Zhang, X., Amendola, P., Hewson, J.C., Sommerfeld, M., Hu, Q., 2012. Influence ofgrowth phase on harvesting of Chlorella zofingiensis by dissolved air flotation.Bioresour. Technol. 116, 477–484.