J. of Advanced Engineering and Technology Vol. 2, No. 3 (2009) pp.

201-205

Removal of Algae and Organic Compounds by Dissolved Air Flotation Process

Abstract : A new hybrid system consisting of powdered activated carbon adsorption and dissolved air flotation (DAF) processes were studied for simultaneous removal of algae and the organics produced from algae (anabaena and mycrocystis). Before studying the hybrid system, adsorption equilibrium and kinetics of organics on powdered activated carbon (PAC) were investigated. Three types of powdered activated carbons (wood-based, coal-based, coconut-based) were chosen as an adsorbent. The 2-methylisoboneol (MIB) and geosmin were used as a representative organic compound produced from algae. PACs were characterized using BET, SEM and particle size analyzer. It was found that MIB and geosmin can be effectively removed by PAC adsorption when a correct dose was applied. The homogeneous surface diffusion model was applied to predict PAC doses required to reduce MIB and geosmin concentrations to below 10 ng/l. On the other hand, the flotation efficiency of algae and organicloaded PAC particles in DAF process was evaluated with zeta potential measurements. Our experimental results revealed that simultaneous removal of algae and organics dissolved in water can be successfully achieved using the hybrid system of adsorption/DAF processes. Keywords : zeta potential, DAF, algae, MIB, geosmin, hybrid process.

1. Introduction
Blue-green algae in water often produces the musty-earthy taste and odor compound of 2-methylisoborneol (MIB) and geosmin. Therefore, removing taste and odor compounds from drinking water has been attracting much attention internationally because they pollute water sources significantly. It has been generally known that MIB and geosmin are detected by consumers as mustyearthy odors at levels as low as 10 ng/l [1]. Thus, the treatment of these compounds must be very effective. Among many water treatment strategies, powdered activated carbon (PAC) has been successfully applied because it is relatively inexpensive and can be applied only when required. However, removing the spent PAC from water is a drawback. On the other hand, it has been known that dissolved air flotation (DAF) is an effective solid/liquid separation process for low density particles such as algae, color, clay flocs produced from low turbidity water [2]. However, DAF is limited in removing the organic compounds dissolved in water. A DAF unit consists of four steps: (1) coagulation and flocculation prior to flotation, (2) bubble generation, (3) bubble-floc collision and attachment in a mixing zone, (4) rising of bubble-floc agglomerates in a flotation tank. If a bubble-PAC particles collision results in successful attachment, and if the resulting bubble-particle agglomerate
Corresponding Author Tel : +1-714-699-0900 E-mail : kwak124@hanamail.net

is positively buoyant, the agglomerate can rise to the top of the liquid column and collect in a foam layer which can subsequently be skimmed off. In other words, when the PAC can be floated by DAF, a combination of DAF and PAC adsorption seems to be successfully applied in water and wastewater treatment. Many studies have been conducted on individual process of PAC adsorption of MIB and geosmin, and DAF for algae removal [2]. However, little has been known about combining the two processes systematically. Therefore, this study focuses on the feasibility of the hybrid processes of adsorption and DAF for simultaneous removal of algae (anabaena and mycrocystis) and organic pollutants (MIB and geosmin) produced from algae in drinking water treatment. Prior to the studies on the hybrid system, adsorption equilibrium and kinetics of organics were investigated using three types of powdered activated carbons (wood-based, coal-based, coconut based). The homogeneous surface diffusion model was used to determine the internal mass transfer coefficient which contributes to the prediction of PAC doses required to reduce MIB and geosmin concentrations to below 10 ng/l. On the other hand, the flotation efficiency of algae and PAC adsorbing organics was evaluated in a DAF process.

2. Experimental
The adsorption experiments were carried out using two organic compounds (2-MIB and geosmin) purchased from Aldrich Co.
201

202

Dong-Heui Kwak

(USA). Chemical properties of MIB and geosmin are listed in Table 1. The concentration of organics was measured by GC/MS. Three different PACs, namely, wood based (WB), coconut-shell based (HA), and coal based (CB), were obtained from James Cumming & Sons PTY Ltd. (Australia). Adsorption experiments were conducted by adding different amounts of PAC into the flask containing 100 ml of synthetic wastewater. After shaking in a constant temperature incubator at constant temperature (298.15 K) for 3 days to give sufficient contacting time for equilibrium, samples were taken from the flask and filtered through 150 mm filter paper (ADVANTEC, Japan). The filtrate was then measured for organic concentrations. The results of adsorption experiments were used to determine the equilibrium isotherms for two organics (MIB and geosmin). The adsorption capacity of PAC was determined by material balance. On the other hand, adsorption kinetic experiments were conducted in a Carberry-type batch adsorber (1.0 ~ 2.0 10-3 m3) at 300 rpm to obtain the concentration decay curves as a function of time. Cells of algae, anabaena and mycrocystis, were obtained from Academia Sinica. Cultivation was conducted using a medium. 350 ml of medium was added to each 500 ml Erlenmeyer flask equipped with a guaze stopper and autocleave. The flasks were irradiated with UV light for 2 hr before cells were added into, and placed on a thermostat shaker. The temperature was controlled at 25oC, and the speed of shaker at 80 rpm. 350 ml of medium was added to each 500 ml Erlenmeyer flask. Cultivation flasks were shaken for 20 days with illumination of solar light for 15 hr each day. Cells were dried under vacuum state for further experiments. Figure 1 shows SEM photos of anabaena and mycrocystis. Prior to sedimentation and DAF operation, coagulation of algae and PAC was conducted on a Jar-Test apparatus using polyaluminium chloride as a coagulant. 10 ~ 50 mg of coagulant and 3 ml of NaOH (0.1 M) were added to the distilled water or raw water (1 L) and rapidly mixed (150 rpm) for 1 min followed by slow mixing
Table 1. Chemical Properties of MIB and Geosmin 2-MIB Name MW Chemical fomular 168 C7H5OCl Geosmin 182 C12H22O 2-methyl-isoborneol Trans-1,10-dimethyl-trans-9-decalol

Figure 1. SEM photos of anabaena and mycrocystis.

(50 rpm) for 20 min. The quality of raw water (Dong-Hwa Dam, Korea). The pH was adjusted by adding HCl or NaOH. The supernatant after the treatment was examined for residual algae and PAC concentration. Zeta potential (Photal Otsuka ELS-8000, Japan) was measured to examine the underlying surface charge to obtain further insight in the mechanism of removal. The schematic diagram of the DAF apparatus is shown in Figure 2. The diameter of the flotation column made of plexiglass was 10 cm, and the height was 30 cm. Algae particles both in the presence and absence of PAC were suspended initially in the column, then bubbles were introduced in the column from the bottom side of the column. The dissolved air solution was fed into the flotation column and the particles in the column were removed by the rising bubbles. The mean diameter of the bubbles fed into the column was 25 m. After all the bubbles in the cell reached the top of the column, the solution was sampled to obtain the flotation efficiency. The turbidity was measured using the turbidity meter (HACH 2100P).

3. Results and Discussion

Chemical structure

OTC (ng/L)a Ordor

30 Musty Camphorous

10 Earthy-musty Figure 2. The schematic diagram of the DAF apparatus.

ordor threshold concentration

2 3 (2009)

Removal of Algae and Organic Compounds by Dissolved Air Flotation Process

203

The surface area was calculated by using the BET method, and the pore size distribution was measured by BJH (Barrett, Joyner and Halenda) method using the nitrogen desorption data. The surface area was found to be 882 to 1200 m2/g, and the average pore diameter was 24 to 31 , which belongs to the mesopore range. Information on the particle size distribution of the PACs used is very important, especially when it is used in conjunction with DAF. The average particle size of three PACs was in the range of 11 to 34 m. It is meaningful to investigate how much concentration of PAC (i.e., dose) is required to remove dissolved organics by PAC adsorption in water treatment. Removal efficiency of suspended solids by DAF process is highly related with the concentration of solids (i.e., turbidity). Prior to the studies on the hybrid system of adsorption and DAF, adsorption equilibrium and kinetics of two organics (MIB and geosmin) were investigated. Adsorption isotherms of PACs used in this work were the most important information for estimating carbon dose for the removal of MIB and geosmin [1]. The adsorption capacity mainly depended on the properties of the adsorbate and adsorbent. From our previous experimental results, the adsorption capacity of geosmin was found to be much higher than that of MIB for the same PACs at constant temperature (298.15 K). In addition, MIB and geosmin were successfully removed using PACs at levels as low as 10 ng/l. The adsorption capacity for MIB and geosmin was the order of PAC-HA > PAC-WB > PAC-CB. The greater affinity of geosmin compared to MIB can be attributed to its chemical structure and solubility (Table 1). This is in agreement with the results reported by Cook et al. [1]. In general, it has been known that the adsorption capacity on activated carbon is enhanced by increasing the molecular size and aromaticity, and by decreasing solubility, polarity and carbon chain branching. Geosmin has a slightly lower solubility and molecular weight and has a flatter structure (Table 1), which may make it easier to adsorption in the slit-shaped pores of the activated carbon. The experimental adsorption kinetics data for MIB and geosmin were predicted by a homogeneous surface diffusion model [3-5]. Results showed that the model predictions were in very good agreement with the experimental data. The determined external mass transfer coefficient and internal surface diffusivities of MIB are 6.26 10-5 m/s and 2.59 10-14 m2/s, and those of geosmin are 2.84 10-5 m/s and 4.71 10-15 m2/s. Unlike the results of equilibrium, the kinetics of adsorption of MIB was faster than geosmin. On the basis of equilibrium and kinetics obtained in this work, PAC-HA seems to be applied to the hybrid system of PAC adsorption and DAF processes because it has higher adsorption capacity and kinetics. 3.2. DAF study Zeta potential is an important parameter of double layer repulsion for individual particles, and it can be used to interpret the trend of coagulation efficiency. It has been known that colloidal particles should have zero net surface charge (isoelectric point,

IEP) for agglomeration. This can result from the adsorption of hydrogen ions or positive-charged ions (i.e., such as aluminum ions), on negatively-charged surfaces. When the zeta potential of particles is approaching zero, coagulation efficiency is generally improved. The coagulation mechanism of PAC is thought to be complex and involves several reaction routes, including charge neutralization (electrostatic interaction) and complex formation between PAC and alumino precipitates, adsorption, bridge formation and surface precipitation on alumino hydroxide solid precipitate. In addition, it has been addressed that pH is one of the most important parameters in the coagulation processes. Depending on pH values, the interfacial properties and reaction routes may be different. The results of zeta potential measurements of PACs are carried out under various experimental conditions, including the effects of water properties (distilled water and reservoir water), coagulant dose (5 ~ 20 mg/L), and organic adsorption on PAC dose (5, 10, 20 mg/L). The zeta potential values ranged approximately between +20 to -50 mV for distilled water and always negative values (3 ~ 20 mV) for reservoir water. The zeta potential decreased with an increasing PAC dose and pH. However, it was observed that the influence of the types and dose of PACs on the variation of zeta potential was quite low. IEP of PACs dissolved in reservoir water was not seen while that in distilled water was present in the pH range of 4 to 5. The zeta potential of organics (MIB and gesosmin) ranged approximately between +40 to 40 mV in terms of solution pH (4 ~ 9) with PACs dose (5, 10, 20 mg/ L). It was observed that IEP moved to the neutral range (pH 6 ~ 7) after adsorbing organics. This result implies that the organic adsorption affects the zeta potential of PACs. The optimal coagulant condition is the functions of the types of algae, coagulant dose, and other operating conditions such as turbidity, pH, and temperature. Polyaluminum chloride was used for altering the surface charge of PAC and algae. The variation of zeta potential of PACs depending on coagulant dose (10 ~ 50 mg/L) and pH (4 ~ 9) were investigated. As expected, the values increased with coagulant dose. Also, the surface charge of the PAC particles dissolved in water was easily controlled by the adjustment of coagulant doses. In general, the particles were negatively charged in water and bubbles were also negatively charged. In this case, the removal efficiency by DAF was very low without adjustment of the surface charge of the particles. Thus, the surface charge of PAC should be changed to neutral or positive to float in the DAF process because PAC is negatively charged without coagulant doses. The variation of zeta potential of two algae was investigated in terms of solution pH of 4 ~ 9 with a coagulant dose of 10 ~ 50 mg/L. It was observed that the zeta potential increased much with a coagulant dose at lower pH (4 ~ 6), while it moderately increased at higher pH (7 ~ 9). The values of zeta potential ranged between approximately 15 ~ +25 mV, and they were slightly different according to the types of algae. The values were 15 ~ +23 mV
J. of Adv. Eng. and Tech., Vol. 2, No. 3 (2009)

204

Dong-Heui Kwak

Table 2. Removal Efficiency of PACs by Sedimentation and DAF (unit:%) PAC* PAC-CB PAC-WB PAC-HA
*

without coagulant Sedimentation 1.7 3.8 1.5 DAF 9.7 5.3 6.0

with coagulant Sedimentation 77.5 72.9 69.5 DAF 94.6 94.2 93.6

Conditions : PAC (10 mg/L), coagulant (polyaluminum chloride, 30 mg/L)

Table 3. Removal Efficiency of Algae by DAF with and without PAC (unit:%) without PAC Anabana Mycrocystis
*

with PAC* 94 95

96 96

Conditions : PAC-HA (10 mg/L), coagulant (polyaluminum chloride, 35 mg/L)

for anabaena and 3 ~ 25 mV for microcystis. Unlike the results of PACs, the zeta potential of algae was quite different depending on the solution pH and coagulant dose. The turbidity removal efficiencies of anabaena and microcystis by DAF and CGS in terms of coagulant dose (10 ~ 50 mg/L) at pH 6. The removal efficiency of two algae increased with coagulant dose although its tendency was changed when the coagulant dose was over 40 mg/L. We believe that DAF is an effective method of removing algae. On the other hand, the removal efficiency of three types of PACs was compared in the absence and presence of coagulant in order to verify the DAF methodology to float PACs (Table 2). CGS results were included for comparison. The removal efficiency without coagulation was very low (< 10%), while with increasing coagulant dose the efficiency increased up to 70% by CGS and up to 95% by DAF. We found that the electrostatic interactions between positively charged algal particles and negatively charged bubbles became mutually attractive when the coagulant was used. However, the interactions between algal particles and bubbles are

Figure 3. Variation of zeta potential of algae depending on coagulant dose.

Figure 4. Removal efficiency of algae depending on coagulant dose. 2 3 (2009)

Removal of Algae and Organic Compounds by Dissolved Air Flotation Process

205

repulsive without the use of coagulant. It was also found that the removal efficiency was almost independent of the types of PAC. On the other hand, the removal efficiency of algae in the absence and presence of PAC was examined (Table 3). Results showed that the removal efficiencies of anabana and mycrosystis were very high about 94 ~ 96% regardless of the presence of PAC when a coagulant of 35 mg/L was used at solution pH 6. This result implies that DAF is an effective process for the removal of algae even in the presence of PAC. On the basis of this finding, therefore, we believed that simultaneous removal of algae and PAC adsorbing organic pollutants produced from algae can be successfully achieved by employing the hybrid system of PAC adsorption and DAF processes.

the zeta potential values of algae and PAC increased with coagulation dose, and the flotation efficiencies were also enhanced. The removal efficiency of PACs was very low without coagulation, while the efficiency increased up to 70% by CGS and to 95% by DAF when coagulant (i.e., polyaluminium chloride) was used. In especial, it was found that the removal efficiency of PAC was almost independent of their types. Unlike the results of PAC, the removal efficiencies of algae by DAF or CGS were very high both in the absence and the presence of PAC. Our experimental results suggest that the hybrid system consisting of adsorption and DAF processes can be effectively applied for simultaneous removal of algae and PAC adsorbing organic pollutants produced from algae in water treatment..

4. Conclusions
Experimental and theoretical studies on adsorption equilibrium and kinetics were carried out to investigate the adsorption characteristics of MIB and geosmin. Adsorption capacity of geosmin on three PACs was much higher than that of 2-MIB. The equilibrium data were well fitted by Freundlich isotherm. In addition, the adsorption kinetic data of MIB and geosmin were predicted with a homogeneous surface diffusion model. The determined surface diffusivities of MIB and geosmin are 2.59 10-14 m2s-1 and 4.71 10-15 m2s-1, respectively. On the other hand,

References
(1) D. Cook, G. Newcombe and P. Sztajnbok, Wat. Res., 35, 1325 (2001). (2) Okada, K., Akagi, Y., Kogure, M. and Yoshioka, N., Can. J. Chem. Eng., 68, 614 (1990). (3) J.W. Lee, W.G. Shim, J.Y. Ko and H. Moon, Sep. Sci. Tech., 39, 2041 (2004). (4) J.W. Lee, T.W. Kwon and I.S. Moon, Carbon, 42, 371 (2004). (5) J.W. Lee, H.J. Jung, D.H. Kwak and P.G. Chung, Wat. Res., 39, 617 (2005).

J. of Adv. Eng. and Tech., Vol. 2, No. 3 (2009)