Available online at www.sciencedirect.

com

Bioresource Technology 99 (2008) 28642871

Biosorption of methylene blue from aqueous solution using free and polysulfone-immobilized Corynebacterium glutamicum: Batch and column studies
K. Vijayaraghavan, Juan Mao, Yeoung-Sang Yun
*
Division of Environmental and Chemical Engineering, Research Institute of Industrial Technology, Chonbuk National University, Chonbuk 561-756, South Korea Received 7 May 2007; received in revised form 11 June 2007; accepted 11 June 2007 Available online 30 July 2007

Abstract The amino acid fermentation industry waste, Corynebacterium glutamicum, has been found to possess excellent biosorption capacity towards methylene blue (MB). Due to practical diculties in solidliquid separation and biomass regeneration, C. glutamicum was immobilized in a polysulfone matrix. The pH edge experiments revealed that neutral or alkaline pH values favored MB biosorption. Isotherm experiments indicated that C. glutamicum, when in immobilized state, exhibited slightly inferior dye uptake compared to free biomass. Also considering the two forms, immobilized biomass took a long time to attain equilibrium. An attempt to identify the diusion limitations in immobilized beads was successful, with the WeberMorris model clearly indicating intraparticle as the rate controlling step. Regeneration of the free biomass was not possible as it tended to become damaged under strong acidic conditions. On the other hand, immobilized biomass performed well with 99% desorption of MB from the biosorbent with the aid of 0.1 mol/l HCl. The immobilized biomass was also successfully regenerated and reused for three cycles without signicant loss in sorption capacity. An up-ow packed column loaded with immobilized biomass was employed for the removal of MB. The column performed well in the biosorption of MB, exhibiting a delayed and favorable breakthrough curve with MB uptake and % removal of 124 mg/g biomass and 70.1%, respectively. 2007 Elsevier Ltd. All rights reserved.
Keywords: Wastewater treatment; Immobilization; Polysulfone; Packed column; Modeling

1. Introduction In recent years, several adsorbents have been identied as possessing good dye-binding capabilities (Rao and Rao, 2006; Batzias and Sidiras, 2007; Ozer et al., 2007; Pavan et al., 2007). In particular, biomaterials of microbial origin have been very eective because of their cell wall constituents. Important fungal biosorbents include Aspergillus (Fu and Viraraghavan, 2002), Penicillium (Iscen et al., 2007) and Rhizopus (OMahony et al., 2002; Aksu and Cagatay, 2006; Kumari and Abraham, 2007). Very few biosorbents under the class of bacteria have been
*

Corresponding author. Tel.: +82 63 270 2308; fax: +82 63 270 2306. E-mail address: ysyun@chonbuk.ac.kr (Y.-S. Yun).

reported for the removal of dyes; but of these, Corynebacterium (Won et al., 2004; Vijayaraghavan and Yun, 2007b) ` and Streptomyces (Nacera and Aicha, 2006) are noteworthy. Since these microorganisms are used widely in dierent food/pharmaceutical industries, they are generated as waste and can be acquired free or at low cost from these industries. Microbial biosorbents are basically small particles with low density, poor mechanical strength and little rigidity. Even though they have merits, such as high biosorption capacity, rapid equilibrium attainment, less process cost and good particle mass transfer, they often suer from several drawbacks. The important demerits include solid liquid separation problem, possible biomass swelling, impossibility of regeneration/reuse and the development

0960-8524/$ - see front matter 2007 Elsevier Ltd. All rights reserved. doi:10.1016/j.biortech.2007.06.008

K. Vijayaraghavan et al. / Bioresource Technology 99 (2008) 28642871

2865

of a high pressure drop when used in column mode. Many ` investigators have highlighted these limitations (Veglio and Beolchini, 1997; Hu and Reeves, 1997; Vijayaraghavan and Yun, 2007a), which often limit the application of microbial biosorbents to industrial applications. Several established techniques are available that make these biosorbents suitable for process applications. Of these, immobilization techniques such as entrapment and cross linking have been found practical for biosorption applications (Bai and Abraham, 2003; Beolchini et al., 2003). Immobilization of microorganisms in a polymeric matrix exhibits greater potential, especially in packed or uidized bed reactors, with benets including control of particle size, regeneration and reuse of biomass, easy separation of biomass and euent, high biomass loading and minimal clogging under continuous ow conditions (Hu and Reeves, 1997). Corynebacterium glutamicum, used in the present study, is widely employed for the biotechnological production of amino acids (Hermann, 2003); therefore, a large amount of biomass is produced as a byproduct. Due to practical diculties in desorption, C. glutamicum was immobilized in polysulfone matrix and used for the biosorption of methylene blue. Methylene blue was selected as the adsorbate, since it is one of the most important and widely used cationic dyes in the textile and paper industries. Thus, the euents emanating from these industries often colored due to methylene blue and require proper treatment prior to their discharge. Hence, this study employed polysulfone-immobilized C. glutamicum for the biosorption of methylene blue in both batch and packed column modes of operation. 2. Methods 2.1. Sorbate Methylene blue (C16H18ClN3S 3H2O) was purchased from SigmaAldrich Korea Ltd. (Yongin, Korea), with a purity of 82% and molecular weight of 373.9. 2.2. Preparation of biosorbent The fermentation wastes (C. glutamicum biomass) were obtained in the form of a ne powder from a lysine fermentation industry (Deasang, Gunsan, Korea). The biomass was dried at 60 C for 12 h and subsequently used for biosorption experiments. Through potentiometric titration (Won et al., 2005), pHPZC of C. glutamicum was determined as 2.1. For immobilization of the biomass, a 9% (w/v) solution of polysulfone was prepared in N,N-dimethyl formamide (DMF) solution. After stirring the mixture for 10 h, the C. glutamicum biomass (14%) was mixed with the polysulfone slurry, with the resultant slurry dripped in deionized water, where beads were formed by a phase inversion process. The beads were then washed with deionized water, and placed in a water bath for 18 h to remove any residual

DMF. The resultant beads (12 mm diameter) were then stored at 4 C, and are designated as PIC in this paper. The pattern of immobilization and other bead characterization has been reported elsewhere (Vijayaraghavan et al., 2007). 2.3. Batch studies 2.3.1. Biosorption Batch biosorption experiments were conducted by bringing into contact 0.1 g (dry weight) of raw biomass or 1 g (wet weight) of PIC with 40 ml dye solution, at the desired pH, in 50 ml plastic bottles (high-density polyethylene), which were maintained on a incubated rotary shaker at 160 rpm and 25 C. The pH of the solution was initially adjusted using either 0.1 mol/l HCl or NaOH, which were subsequently used to control the pH during the experiments. After 12 h of contact with the dye solution, the biosorbent was separated by centrifugation at 3000 rpm for 5 min. The dye (MB) concentration in the supernatant was determined using a spectrophotometer (UV-2450, Shimadzu, Kyoto, Japan) at 660 nm, after appropriate dilution. Kinetic experiments were conducted in the same manner described above, except the samples were collected at dierent time intervals to determine the time at which biosorption equilibrium had been attained. The amount of MB sorbed by the biosorbent was calculated from the dierences between the concentrations of MB added to those detected in the supernatants, using the following equation: Q V C 0 C f =M 1

where Q is the MB uptake (mg/g), C0 and Cf the initial and equilibrium MB concentrations in the solution (mg/l), respectively, V the solution volume (l) and M the mass of biosorbent (g). 2.3.2. Desorption The MB-loaded biomass, which was exposed to 250 mg/ l of MB at pH 7 and temperature 25 C, was separated from the biosorbent-water slurry by centrifugation. The biosorbent was then brought into contact with 20 ml of 0.1 mol/l HCl for 3 h on a rotary shaker at 160 rpm. The remaining procedure was the same as employed in the biosorption equilibrium experiments. After desorption, the biosorbent was washed several times with deionized water, with the regenerated biosorbent reused for the next cycle. These cycles of biosorption followed by elution were repeated three times, with the biosorbent capacity then evaluated. 2.4. Column studies A glass column (1.5 cm ID and 25 cm height) was packed with a known quantity of biosorbent to yield the desired bed height. The column was then fed with 50 mg/l of MB solution in an up-ow mode, at a ow rate of

2866

K. Vijayaraghavan et al. / Bioresource Technology 99 (2008) 28642871

1 ml/min using a peristaltic pump. The samples were collected at desired time intervals at the exit of the column, using a sample collector arrangement, and the MB concentrations then analyzed. The operation of the column was stopped when the euent MB concentration exceeded 97% of the inlet dye concentration. After exhaustion of the column, desorption was carried out by pumping 0.1 mol/l HCl upwards through the column at a ow rate of 2 ml/min. To determine the weight loss after sorption desorption cycle, the biosorbent was washed with deionized water and dried naturally. Both, batch and column data modeling were performed by non-linear regression using the Sigma Plot (version 4.0, SPSS, USA) software. The average percentage error between the experimental and predicted values is calculated using: PN Qexp;i Qcal;i =Qexp;i e % i1 100 2 N where Qexp and Qcal represents experimental and calculated dye uptake values, respectively, and N is the number of measurements. 3. Results and discussion 3.1. Eect of pH and mechanism In the rst series of batch experiments, the inuence of equilibrium pH on the biosorption of MB was examined (Fig. 1). The solution equilibrium pH found to severely aect the MB biosorption capacity of C. glutamicum, with pH values at and above neutral resulted in maximum uptake. The cell wall of Gram-positive bacterium mainly comprised of peptidoglycan layer connected by amino acid bridges (Mera et al., 1992). Imbedded in the Gram-positive cell wall are polyalcohols called teichoic acids, which give an overall negative charge to the bacterial cell wall due to the presence of phosphodiester bonds between teichoic acid monomers (Beveridge et al., 1982). The zero charge potential of the biomass was determined as 2.1 and thus the biomass will have a net negative charge above pH 2.1. On the other hand, basic dyes release colored positively charged
200
Free biomass PIC (based on beads) PIC (based on biomass)

dye ions when in solution, which will exhibit electrostatic attraction towards the negatively charged cell surface. In particular, the carboxyl groups present in C. glutamicum (Won et al., 2005) were mainly responsible for the biosorption of MB. The pKa value of carboxylic groups usually lies within the range of 3.85.0 (Roberts and Caserio, 1977). Therefore, the carboxyl groups have a negative charge at pHs approximately higher than 5; therefore, will electrostatically bind MB to the bacterial biomass.
O Biomass C ODye + Biomass C O+

O NH Dye

150

100

50

The above reaction (Pavan et al., 2007) clearly explains the pH eect on the biosorption of MB onto C. glutamicum. At low pH values, the carboxyl groups will be in their protonated form and; thus, the overall charge of the biomass will be positive. Under this situation, the occupation of MB onto carboxyl groups will be dicult and; hence, a low uptake was observed at acidic pH values (Fig. 1). The results suggest that C. glutamicum possesses an excellent binding capacity for MB. However, the C. glutamicum biomass has been known to cause problems during column operations due to its poor mechanical stability, high swelling and solidliquid separation problems (Vijayaraghavan and Yun, 2007a). Therefore in this research, C. glutamicum was immobilized in a polysulfone matrix (Vijayaraghavan et al., 2007) and subsequently employed for biosorption studies. The polysulfone-immobilized C. glutamicum (PIC) also exhibited a similar trend as that of the free biomass, with maximum uptakes at pH P 7. Comparing the uptakes, PIC exhibited less uptake compared to that of the free cells. However, it should also be noted that the uptake capacity of PIC was based on the dry weight of the beads. To gain a clear picture of the contribution of the biomass in the biosorption of MB, the amount of biomass per gram of dry beads was calculated (Vijayaraghavan and Yun, 2007a) and on the basis of mass balance it was determined as 0.54 g biomass/g dry beads. Subsequently, the MB uptake by PIC was determined, as shown in Fig. 1, which was comparable to that by the free biomass. Nevertheless, the slight decrease in the uptake may be attributed to some binding sites not being easily accessible or blocked due to the immobilization process. Several researchers (Bai and Abraham, 2003; Hu and Reeves, 1997) observed that immobilization usually decrease the binding capacity of the biomass; however, this decrease will be minimal considering the possibility of recycling of biosorbent for a number of cycles. Biosorption control experiments revealed that biomass-free polysulfone beads exhibited 11.7 mg/g dry beads at pH 7. 3.2. Biosorption isotherms and modeling The results of the previous section revealed that the biomass of C. glutamicum can be competent for the biosorp-

Uptake (mg/g)

10

Equilibrium pH

Fig. 1. Eect of pH on the uptake of MB by the free and polysulfoneimmobilized C. glutamicum biomasses (C0 = 500 mg/l; temperature = 25 1 C; agitation speed = 160 rpm).

K. Vijayaraghavan et al. / Bioresource Technology 99 (2008) 28642871

2867

tion of MB at pH P 7; therefore, isotherm experiments were conducted at pHs 7 and 8 to elucidate the complete biosorption potential of the biomass. Typical biosorption isotherms were observed for both the free and immobilized forms of C. glutamicum (Fig. 2). The biomass of C. glutamicum exhibited very high MB uptakes at pHs 7 and 8, with signicant variation in the uptakes observed only at high residual MB concentrations. The isotherms were steep, indicating the high anity of the sorbate towards the sorbent. Modeling of MB isotherm data was attempted using the Langmuir and Toth models, which can be represented as follows, Langmuir model Langmuir; 1918 : Toth model Toth; 1971 : Q Qmax bC f 1 bC f Qmax bT C f Q
1=nT nT

Table 1 Biosorption isotherm constants for the biosorption of MB by the free and immobilized C. glutamicum biomasses Isotherm models Free biomass pH 7 Langmuir Qmax (mg/g) b (l/mg) R2 e (%) Toth Qmax (mg/g) bT (l/mg) nT R2 e (%) 336.2 0.023 0.972 9.6 328.1 0.020 0.849 0.973 8.1 pH 8 339.2 0.034 0.987 2.9 343.6 0.048 1.241 0.992 0.2 Polysulfone-immobilized biomass pH 7 158.0 (292.4) 0.045 0.988 3.1 168.9 (312.8) 0.070 1.423 0.993 1.4 pH 8 166.5 (308.1) 0.086 0.986 4.5 184.8 (342.1) 0.211 1.767 0.998 0.8

3 4

The values in parentheses indicate the uptake based on gram of biomass in the immobilized bead. R2 = correlation coecient.

1 bT C f

where Qmax is the maximum dye uptake (mg/g), b the Langmuir equilibrium constant (l/mg), bT the Toth model constant and nT the Toth model exponent. The Langmuir adsorption isotherm has traditionally been used to quantify and contrast the performance of dierent biosorbents. It also served to estimate the maximum dye uptake values where these could not be attained experimentally. The constant b represents anity between a sorbent and sorbate. The Langmuir constant values (Table 1) revealed that the free biomass obviously performed well. However, the recorded Langmuir anity constants were found to have maxima for the immobilized biomass. For favorable biosorption, high Qmax and a steep initial isotherm slope (i.e. high b) are desirable. The MB biosorption capacity observed in this study was superior compared to the results published in the literature (Table 2). To improve the tness of the biosorption isotherm data, a three parameter isotherm model, viz. the Toth model, was also used. The Toth isotherm, derived from potential theory, has proven to be useful in describing the sorption in heterogeneous systems, such as phenolic compounds onto carbon. It assumes an asymmetrical quasi-Gaussian

energy distribution with a widened left-hand side, i.e. most sites have sorption energy less than the mean value (Ho et al., 2002). As expected, the Toth model described the isotherm data well with high R2 and low % error vales (Table 1). The successful application of the Toth model to the present data supports the fact that the surfaces of the biosorbent were heterogeneous and contained dierent functional groups. 3.3. Biosorption kinetics and modeling Fig. 3 shows the plots of MB uptake against time for the dierent forms of C. glutamicum. In general, the results revealed that the removal of MB was fast during the initial stages of the contact period, but thereafter became slower near the equilibrium. This is obvious, in that a large number of vacant binding sites will be available for sorption during the initial stage; and after a lapse of time, the occupation of the remaining vacant sites will be dicult due to the repulsive forces between the solute molecules on the solid and bulk phases. On comparing the two forms of biomass, it was clearly visible, as shown in Fig. 3, that the time for equilibrium to be attained was relatively fast for the free C. glutamicum compared to the PIC. In the case of free biomass, the removal of MB was rapid for the initial 3 h, followed by the slow attainment of equilibrium at around 5 h. On the contrary, for immobilized biomass; the removal of MB was slow, only reaching equilibrium after 7 h. This was not altogether surprising, since when immobilized, the biomass was retained within the interior of the immobilized matrix; whereas the free biomass had its binding sites freely exposed to MB. For immobilization systems, mass transfer resistances play a signicant role in deciding the rate of biosorption. However, for successful immobilization systems, these mass transfer resistances should not inuence the overall biomass biosorption performance. Most importantly, the dye molecules should have access to all possible binding sites, even at a slower rate.

400

MB uptake (mg/g)

300

200

100

pH 7 (Free biomass) pH 8 (Free biomass) pH 7 (PIC) pH 8 (PIC)

200

400

600

800

1000

1200

1400

1600

1800

Final MB concentration (mg/l)

Fig. 2. Biosorption isotherms for the free and polysulfone-immobilized C. glutamicum biomasses (temperature = 25 1 C; agitation speed = 160 rpm). Curves predicted by the Toth model.

2868

K. Vijayaraghavan et al. / Bioresource Technology 99 (2008) 28642871

Table 2 Comparison of methylene blue biosorption capacity of various adsorbents Adsorbent Corynebacterium glutamicum PIC Sargassum muticum Fomes fomentarius Peanut hull Ulva lactuca Streptomyces rimosus Kaolin Sawdust Coir pith carbon Posidonia oceanica (L.) bres Flyash Experimental condition pH 8; T = 25 1 C pH 8; T = 25 1 C T = 25 C T = 20 C pH 3.5; T = 50 C pH 10; T = 25 2 C T = 20 C pH 8; T = 23 C pH 6.9; T = 35 C pH 6; T = 30 C T = 30 C Uptake (mg/g) 339.2 (L) 308.1 (L) 279.2 (L) 232.7 (L) 161.3 (L) 40.2 (L) 34.3 (L) 14.0 (L) 9.78 (L) 5.87 (L) 5.56 (L) 3.1 (L) References This work This work Rubin et al. (2004) Maurya et al. (2006) Ozer et al. (2007) El Sikaily et al. (2006) ` Nacera and Aicha (2006) Ghosh and Bhattacharyya (2002) Batzias and Sidiras (2007) Kavitha and Namasivayam (2007) Ncibi et al. (2007) Rao and Rao (2006)

200
Free biomass PIC

qt k i t1=2

150
200

(b)
qt (mg/g)

100

150 100

50

50 0

10

0 0 100 200 300

t0.5 (min0.5)

15

20

25

400

500

600

Time (min)

Fig. 3. Biosorption kinetics of the uptake of MB for dierent forms of C. glutamicum (C0 = 500 mg/l; pH 7; temperature = 25 1 C; agitation speed = 160 rpm). Curves predicted by the non-linear pseudo-second order model. (b) WeberMorris intraparticle diusion model plot.

The experimental biosorption kinetic data were modeled using pseudo-second order kinetics, which can be represented in a non-linear form, as follows: qt q2 kt e 1 qe kt 5

where qe is the amount of dye sorbed at equilibrium (mg/g); qt the amount of dye sorbed at time t (mg/g); and k the equilibrium rate constant (g/mg min). The pseudo-second order model is based on the sorption capacity onto the solid phase, and predicts the sorption behavior over the entire study range (Ho and McKay, 1998; Ku et al., 2007). The correlation coecients were always greater than 0.994, and the predicted curves showed excellent agreement with the experimental data (Fig. 3). The equilibrium rate constant values were recorded as 3.2 104 and 0.8 104 g/ mg min for free biomass and PIC, respectively. The predicted equilibrium uptake (qe) values were 185.8 and 178.2 mg/g biomass for free biomass and PIC, respectively; which were in good agreement with those found experimentally. In an attempt to visualize the inuence of mass transfer resistance on the binding of MB to the free biomass and PIC, the kinetic data were analyzed using the equation proposed by Weber and Morris (1963), as follows:

Thus, the intraparticle diusion constant ki (mg/g min0.5), can be obtained from the slope of the plot of qt (uptake at any time, mg/g) versus the square root of time. If this plot passes through the origin, then intraparticle diusion is the rate controlling step. Fig. 3 shows the plots of qt versus t1/2, with multilinearity clearly observed in both the case of the free and immobilized biomasses, which implies the process involves more than one kinetic stage (or sorption rates) (Guo et al., 2003). For instance, the PIC exhibited three stages, including an initial linear portion (up to p p t < 4.5), a second portion (4.5 < t < 13.4) and a nal p third portion (13.4 < t). The rst was attributed to the sorption of MB over the surface of the biomass, and; hence, was the fastest sorption stage. The second, ascribed to intraparticle diusion, was a delayed process. The third stage may be regarded as the diusion through smaller pores, which is followed by the establishment of equilibrium. Thus, the inuence of intraparticle diusion was clearly seen in the case of the polysulfone-immobilized beads. Conversely, from the free cell kinetic data, intraparticle diusion is clearly not the fully operative mechanism, as the slope of the plot (second linear portion) did not pass through the origin. The calculated intraparticle diusion constants (ki) were 7.6 and 10.3 mg/g biomass min0.5 for the free biomass and PIC, respectively. 3.4. Desorption and regeneration of biomass To investigate the possibility of recycling the biomass over multiple cycles, batch desorption experiments were conducted. The sorption experiments revealed that neutral or basic conditions were necessary for the binding of MB to the biomass. Thus, it would be logical that the sorbed dye molecules can be recovered only under acidic conditions. Therefore, 0.1 mol/l HCl was used as an elutant, the performance of which was very satisfactory with both the free and immobilized biomasses. Under strong acidic (low pH) conditions, the number of positively charged sites increases. These positively charged sites on the sorbent surface favor desorption of the dye cations due to electrostatic

Uptake at any time (mg/g)

K. Vijayaraghavan et al. / Bioresource Technology 99 (2008) 28642871

2869

repulsion. Desorption eciencies were always greater than 99%, but considerable weight loss (approximately 20%) was observed in the case of free biomass due to the extreme conditions. Also, the separation of the biomass from the nal elution solution was dicult; therefore, centrifugation and ltration are of utmost importance. Conversely, the polysulfone beads were stable under the acidic conditions examined, with no weight loss observed. Considering the good stability and ease of dye desorption, the PIC beads were employed in regeneration experiments. The regeneration experiments were aimed to explore the biosorption potential of PIC for MB over three sorption desorption cycles. The polysulfone beads approximately retained their rst cycle MB uptake (83.6 mg/g beads or 154.7 mg/g biomass) throughout the three cycles examined, aided with consistently high elution eciencies around 99% by 0.1 mol/l HCl. Most importantly, the weight loss was insignicant (<4.5%) at the end of three consecutive cycles. 3.5. Biosorption in an up-ow packed column The performance of biomaterial in continuous dye biosorption is often an important factor in accessing the feasibility of a biosorbent in real applications. Therefore, a PIC loaded up-ow packed column was devised and employed in this study for the biosorption of MB. The column was initially packed with 19.7 g (wet weight)/4.4 g (dry weight) of PIC, yielding an initial bed height and volume of 20 cm and 35.3 ml, respectively. Fig. 4 shows the breakthrough curve generated during the biosorption of MB. The PIC column bed performed well in biosorption of MB, with the breakthrough (1 mg/l in the euent) appeared only after 58 h of column operation. Thereafter, the PIC bed still performed well, and resulted in a smooth breakthrough curve, which nally became exhausted (49 mg/l in the euent) after around 144 h of column operation. The sorption zone (dierence between the breakthrough and exhaustion times) and the slope of the breakthrough curve (from the breakthrough and exhaustion time) were recorded as 86 h and 0.66 mg/ l.h, respectively. The column uptake (Vijayaraghavan et al., 2004) and percentage removal were calculated to be 67.1 mg/g beads (124 mg/g biomass) and 70.1%, respectively. Modeling of column data using the Thomas model enabled the predictions of breakthrough curve and column uptake. The non-linear form of the Thomas model can be expressed as follows:   C0 k TH 1 exp Q0 M C 0 V eff 7 C F where Q0 is the dye uptake (mg/g); kTH the Thomas model rate constant (l/mg h); M the PIC weight (g) and Ve the euent volume (l). The dye uptake predicted by the Thomas model (Q0) was 121.5 mg/g biomass, which agreed well with that obtained experimentally. The rate constant (kTH), which characterize the rate of solute transfer from the uid

Time (h) Final MB concentration (mg/l)

1 0.8 0.6 0.4 0.2 0
Sorption Elution

10

2000

1500

C/C0

1000

500

20

40

60

80

100

120

140

0 160

Time (h)

Fig. 4. Sorption and elution curves during the biosorption and desorption of MB, respectively, in the up-ow packed column (bed height = 20 cm; solute ow rate = 1 ml/min; C0 = 50 mg/l; inlet solution pH 7; elutant = 0.1 mol/l HCl; elutant ow rate = 2 ml/min). Breakthrough curve predicted by the Thomas model.

to the solid phase, was 0.002 l/mg h. The breakthrough curve simulated by the Thomas model, as presented in Fig. 4, coincided well with the experimental data. In the next stage of column experiment, the column loaded with MB was eluted using 0.1 mol/l HCl. The elution curve (Fig. 4) exhibited a general trend; that is, a sharp increase at the beginning, followed by a gradual decrease. The elution process was conducted for 9.5 h, with an elution eciency of 98.1%. At the end of the sorptiondesorption cycle, no decrease was observed in the bed height, with a biomass weight loss (based on dry weight) of less than 3%. Since the batch regeneration experiments revealed no signicant loss in the biosorption capacity over repeated cycles, the same would be expected in column regeneration experiments. The total volume of MB solution treated during the sorption process was 8.64 l; whereas 1.14 l of 0.1 mol/l HCl was used to elute the MB during the desorption process. This resulted in a very high concentrated MB solution, with a concentration factor (Vijayaraghavan et al., 2005) of 7.6. A high concentration factor is always preferable as the eventual recovery of dye will became more feasible. 4. Conclusions From the present study on the biosorption of methylene blue from aqueous solution, using free and polysulfoneimmobilized C. glutamicum, the following conclusions are suggested: The biomass of C. glutamicum, both in the free and immobilized forms, was found to be an excellent biosorbent for methylene blue. Biosorption isotherms were modeled using the Langmuir and Toth models. PIC exhibited MB uptake of 292.4 mg/g biomass compared to the free biomass uptake capacity of 336.2 mg/g biomass at pH 7, based on the Langmuir model. Based on the correlation coef-

2870

K. Vijayaraghavan et al. / Bioresource Technology 99 (2008) 28642871 Fu, Y., Viraraghavan, T., 2002. Dye biosorption sites in Aspergillus niger. Bioresourc. Technol. 82, 139145. Ghosh, D., Bhattacharyya, K.G., 2002. Adsorption of methylene blue on kaolinite. Appl. Clay Sci. 20, 295300. Guo, B., Hong, L., Jiang, H.X., 2003. Macroporous poly(calcium acrylate-divinylbenzene) bead a selective orthophosphate sorbent. Ind. Eng. Chem. Res. 42, 55595567. Hermann, T., 2003. Industrial production of amino acids by coryneform bacteria. J. Biotechnol. 104, 155172. Ho, Y.S., McKay, G., 1998. Sorption of dye from aqueous solution by peat. Chem. Eng. J. 70, 115124. Ho, Y.S., Porter, J.F., McKay, G., 2002. Equilibrium isotherm studies for the sorption of divalent metal ions onto peat: copper, nickel and lead single component systems. Water Air Soil Pollut. 141, 133. Hu, Z.Z.-C., Reeves, M., 1997. Biosorption of uranium by Pseudomonas aeruginosa strain CSU immobilized in a novel matrix. Biotechnol. Prog. 13, 6070. Iscen, C.F., Kiran, I., Ilhan, S., 2007. Biosorption of Reactive black 5 dye by Penicillium restrictum. J. Hazard. Mater. 143, 335340. Kavitha, D., Namasivayam, C., 2007. Experimental and kinetic studies on methylene blue adsorption by coir pith carbon. Bioresourc. Technol. 98, 1421. Ku, C.S., Sathishkumar, M., Mun, S.P., 2007. Binding anity of proanthocyanidin from waste Pinus radiata bark onto praline-rich bovine achilles tendon collagen type I. Chemosphere 67, 1618 1627. Kumari, K., Abraham, T.E., 2007. Biosorption of anionic textile dyes by nonviable biomass of fungi and yeast. Bioresourc. Technol. 98, 1704 1710. Langmuir, I., 1918. The adsorption of gases on plane surfaces of glass, mica and platinum. J. Am. Chem. Soc. 40, 13611403. Maurya, N.S., Mittal, A.K., Cornel, P., Rother, E., 2006. Biosorption of dyes using dead macro fungi: eect of dye structure, ionic strength and pH. Bioresourc. Technol. 97, 512521. Mera, M.U., Kemper, M., Doyle, R., Beveridge, T.J., 1992. The membrane-induced proton motive forces inuences the metal binding ability of Bacillus subtilis cell walls. Appl. Environ. Microbiol. 58, 38373844. ` Nacera, Y., Aicha, B., 2006. Equilibrium and kinetic modeling of methylene blue biosorption by pretreated dead Streptomyces rimosus: eect of temperature. Chem. Eng. J. 119, 121125. Ncibi, M.C., Mahjoub, B., Seen, M., 2007. Kinetic and equilibrium studies of methylene blue biosorption by Posidonia oceanica (L.) bres. J. Hazard. Mater. B139, 280285. OMahony, T., Guibal, E., Tobin, J.M., 2002. Reactive dye biosorption by Rhizopus arrhizus biomass. Enzyme Microb. Technol. 31, 456463. Ozer, D., Dursan, G., Ozer, A., 2007. Methylene blue adsorption from aqueous solution by dehydrated peanut hull. J. Hazard. Mater. 144, 171179. Pavan, F.A., Lima, E.C., Dias, S.L.P., Mazzocato, A.C., 2007. Methylene blue biosorption from aqueous solutions by yellow passion fruit waste. J. Hazard. Mater. doi:10.1016/j.jhazmat.2007.05.023. Rao, V.V.B., Rao, S.R.M., 2006. Adsorption studies on treatment of textile dyeing industrial euent by yash. Chem. Eng. J. 116, 77 84. Roberts, J.D., Caserio, M.C., 1977. Basic principles of organic chemistry. In: Benjamin, W.A. (Ed.), second ed., Menlo Park, CA. Rubin, E., Rodriguez, P., Herrero, R., Cremades, J., Barbara, I., de Vicente, M.E.S., 2004. Removal of methylene blue from aqueous solutions using as biosorbent Sargassum muticum: an invasive macroalga in Europe. J. Chem. Technol. Biotechnol. 80, 291298. Toth, J., 1971. State equations of the solidgas interface layer. Acta Chem. Acad. Hung. 69, 311317. ` Veglio, F., Beolchini, F., 1997. Removal of metals by biosorption: a review. Hydrometallurgy 44, 301316. Vijayaraghavan, K., Yun, Y.-S., 2007a. Chemical modication and immobilization of Corynebacterium glutamicum for biosorption of

cients and % error values, the Toth model better simulated the MB biosorption isotherms under all examined conditions. Desorption was only possible with PIC, as the free biomass tended to become damaged under strongly acidic conditions. The 0.1 mol/l HCl solution was able to elute 99% of MB from the dye-loaded PIC, and allowed the PIC to be regenerated and reused for three sorption desorption cycles, without any signicant loss in MB uptake. In column studies, PIC performed well in the biosorption of MB, with both a delayed breakthrough time and high MB uptake of 58 h and 124 mg/g biomass, respectively. The column bed loaded with MB was successfully eluted using 0.1 mol/l HCl with elution eciency and overall process concentration factor of 98.1% and 7.6%, respectively. Polysulfone showed competitive properties for the immobilization of C. glutamicum biomass, and enabled the biomass to be regenerated and reused for biosorption of MB. Even in the presence of intraparticle diusion, PIC exhibited MB uptakes close to that of free biomass. Because of their good chemical and physical stabilities, polysulfone beads may be easier to use in industrial wastewater systems compared with the bacterial biomass powder.

Acknowledgements This research was nancially supported by the Ministry of Commerce, Industry and Energy (MOCIE) and Korea Industrial Technology Foundation (KOTEF) through the Human Resource Training Project for Regional Innovation and in part, by KOSEF through AEBRC at POSTECH. K. Vijayaraghavan was supported by a grant from the Post-Doc program, Chonbuk National University (the second half term of 2006). References
Aksu, Z., Cagatay, S.S., 2006. Investigation of bisoorption of Gemazaol Turquise blue-G reactive dye by dried Rhizopus arrhizus in batch and continuous systems. Sep. Purif. Technol. 48, 2435. Bai, R.S., Abraham, T.E., 2003. Studies on chromium(VI) adsorption desorption using immobilized fungal biomass. Bioresourc. Technol. 87, 1726. Batzias, F.A., Sidiras, D.K., 2007. Dye adsorption by prehydrolysed beech sawdust in batch and xed-bed systems. Bioresourc. Technol. 98, 12081217. ` Beolchini, F., Pagnanelli, F., Toro, L., Veglio, F., 2003. Biosorption of copper by Sphaerotilus natans immobilized in polysulfone matrix: equilibrium and kinetic analysis. Hydrometallurgy 70, 101112. Beveridge, T.J., Forsberg, C.W., Doyle, R.J., 1982. Major sites of meal binding in Bacillus licheniformis walls. J. Bacteriol. 150, 14381448. El Sikaily, A., Khaled, A., El Nemr, A., Abdelwahab, O., 2006. Removal of methylene blue from aqueous solution by marine green alga Ulva lactuca. Chem. Ecol. 22, 149157.

K. Vijayaraghavan et al. / Bioresource Technology 99 (2008) 28642871 Reactive black 5 from aqueous solution. Ind. Eng. Chem. Res. 46, 608617. Vijayaraghavan, K., Yun, Y.-S., 2007b. Utilization of fermentation waste (Corynebacterium glutamicum) for biosorption of Reactive black 5 from aqueous solution. J. Hazard. Mater. 141, 4552. Vijayaraghavan, K., Jegan, J., Palanivelu, K., Velan, M., 2004. Removal of nickel(II) ions from aqueous solution using crab shell particles in a packed bed up-ow column. J. Hazard. Mater. 113, 223230. Vijayaraghavan, K., Palanivelu, K., Velan, M., 2005. Crab shell-based biosorption technology for the treatment of nickel-bearing electroplating industrial euents. J. Hazard. Mater. 119, 251254. Vijayaraghavan, K., Han, M.H., Choi, S.B., Yun, Y.-S., 2007. Biosorption of Reactive black 5 by Corynebacterium glutamicum biomass

2871

immobilized in alginate and polysulfone matrices. Chemosphere. doi:10.1016/j.chemosphere.2007.03.030. Weber, W.J., Morris, J.C., 1963. Kinetics of adsorption on carbon solution. J. Sanit. Eng. Div. Am. Soc. Civil Eng. 89, 3159. Won, S.W., Choi, S.B., Chung, B.W., Park, D., Park, J.M., Yun, Y.-S., 2004. Biosorptive decolorization of Reactive orange 16 using the waste biomass of Corynebacterium glutamicum. Ind. Eng. Chem. Res. 43, 78657869. Won, S.W., Choi, S.B., Yun, Y.-S., 2005. Interaction between protonated waste biomass of Corynebacterium glutamicum and anionic dye Reactive red 4. Colloid Surf. A: Physicochem. Eng. Aspects 262, 175180.