- Email: [email protected]
A general model for biosorption of Cd2+, Cu2+ and Zn2+ by aerobic granules
Journal of Biotechnology 102 (2003) 233 /239 www.elsevier.com/locate/jbiotec
A general model for biosorption of Cd2, Cu2 and Zn2 by aerobic granules Yu Liu *, Hui Xu, Shu-Fang Yang, Joo-Hwa Tay Environmental Engineering Research Centre, School of Civil and Environmental Engineering, Nanyang Technological University, 50 Nanyang Avenue, Singapore 639798, Singapore Received 1 October 2002; received in revised form 15 November 2002; accepted 11 December 2002
Abstract Aerobic granules are microbial aggregates with a strong and compact structure. This study looked into the feasibility of aerobic granules as a novel type of biosorbent for the removal of individual Cd2 , Cu2 and Zn2 from aqueous solution. Based on the thermodynamics of biosorption reaction, a general model was developed to describe the equilibrium biosorption of individual Cd2 , Cu2 and Zn2 by aerobic granules. This model provides good insights into the thermodynamic mechanisms of biosorption of heavy metals. The model prediction was in good agreement with the experimental data obtained. It was further demonstrated that the Langmuir, Freundlich and Sips or Hill equations were particular cases of the proposed model. The biosorption capacity of individual Cd2 , Cu2 and Zn2 on aerobic granules was 172.7, 59.6 and 164.5 mg g 1, respectively. These values may imply that aerobic granules are effective biosorbent for the removal of Cd2 , Cu2 and Zn2 from industrial wastewater. # 2003 Elsevier Science B.V. All rights reserved. Keywords: Aerobic granule; Biosorption; Thermodynamics; Cadmium; Copper; Zinc; Equilibrium; Model
1. Introduction Heavy metals have been detected in a wide variety of industrial wastewater streams. Chemical precipitation is one of the most common methods used to remove heavy metals from aqueous solution. However, chemical slurry produced is a secondary waste that should be properly handled before discharging. Recently, research attention has been given to look for efficient and low cost
* Corresponding author. E-mail address: [email protected] (Y. Liu).
metal absorbents. For this purpose, many biomaterials had been tested as biosorbents for the biosorption of heavy metals, such as marine algae, fungal biomass, wasted activated sludge, digested sludge and so on (Eccles, 1999; Zhou, 1999; Hamdy, 2000; Wang et al., 2001). It should be realized that most biosorbents currently used are in the form of suspended biomass. The major problems-associated with this kind of biosorbents are post-separation of suspended biomass from the treated effluent, maintenance of biosorbent stability and regeneration of used biosorbents. These drawbacks of conventional biosorbents in the form of bioflocs seriously limit application of
0168-1656/03/$ - see front matter # 2003 Elsevier Science B.V. All rights reserved. doi:10.1016/S0168-1656(03)00030-0
234
Y. Liu et al. / Journal of Biotechnology 102 (2003) 233 /239
biological process in the removal of metals from wastewater. In the past few years, aerobic granulation technology had been successfully developed (Beun et al., 1999; Tay et al., 2001). Aerobic granules are microbial aggregates that have a strong and compact microbial structure. The settling velocity and density of aerobic granules are much higher than conventional bioflocs, while the granules have large surface area. In fact, when selecting appropriate biosorbents for the removal of heavy metals from industrial wastewater, the settleability of biosorbent is one of the most important criteria for effective separation of biosorbent from treated effluent. To date, no information is currently available on the biosorption kinetics of heavy metals by aerobic granules. Therefore, the main objective of the present study was to derive a general model for the biosorption of three individual model heavy metals, cadmium, copper and zinc by aerobic granules. It is expected that this work would be useful in the development of aerobic granules-based biosorbers for the removal of heavy metals from aqueous solution.
2. Materials and methods Aerobic granules used for biosorption tests were collected from a laboratory scale sequencing batch reactor (SBR) fed with acetate as sole carbon source. The operation conditions and setup of the SBR can be found elsewhere (Tay et al., 2001). The mean diameter of aerobic granule was around 1.0 mm (Fig. 1a). Fig. 1b shows the microbial structure of the aerobic granules, and it can be seen that a dominant species is present in the acetate-fed aerobic granules. These aerobic granules had a settling velocity of 71 m h1, and a sludge volumetric index of 49 ml g1. Individual cadmium, copper and zinc were used as the model heavy metals in biosorption tests. The respective Cd2, Cu2 and Zn2 solution was prepared by dissolving CdCl2, CuCl2 and ZnCl2 in deionic water. In this study, three series of batch experiments with the individual Cd2, Cu2 and Zn2 were conducted at a constant initial granule concentration of 100 mg dry weight l 1. The
initial Cd2, Cu2 and Zn2 concentration varied in the range of 5 /200 mg l1 and solution pH was adjusted to 4. The batch biosorption tests were performed in 1-l glass beakers with gentle agitation at a constant temperature of 26 8C. Each test lasted for 5 h. Soluble Cd2, Cu2 and Zn2 concentrations were determined in the course of batch experiments by using Inductively Coupled Plasma Emission Spectrometer (ICP) (PerkinElmer P400, Perkin-Elmer Corporation, Norwalk, USA).
3. Model development It had been proposed that the overall biosorption reaction could be regarded as a simple change in the state of metal ion (Morel and Hering, 1993), that is, C 0 Cads
DGo?
(1)
in which C and Cads is metal concentration in bulk solution, and that adsorbed at time t, while DGo? is the effective free energy change of biosorption, which changes with the proceeding of the biosorption reaction. If the metal concentration (C ) in bulk solution increases, its adsorption is more favorable, i.e. DGo? would decrease with the increase of the metal concentration (Metcalf and Eddy, 1991). According to Morel and Hering (1993), DGo? can be expressed as DG o? DG o aRT ln C
(2)
in which a is positive coefficient. Evidence shows that the real driving force of biosorption is the difference between the amount adsorbed by unit biosorbent (Q ) at a given metal concentration and the theoretical amount that could be adsorbed by unit biosorbent at that concentration (Qth), and this driving force is disappearing when the biosorption reaction gradually approaches its equilibrium state (Metcalf and Eddy, 1991). As biosorption proceeds, the driving force decreases and the adsorption resistance will increase. It is a reasonable consideration that the overall change of free energy of the biosorption process (DG ) would increase with the increase of adsorption resistance, and decrease with the increase of the
Y. Liu et al. / Journal of Biotechnology 102 (2003) 233 /239
235
Fig. 1. (a) Morphology of aerobic granules; (b) microbial structure of aerobic granules.
driving force of adsorption reaction. We assume that the overall change of free energy of the biosorption reaction should be formulated as the function of the driving force and resistance of adsorption such that DG DG o? bRT ln Resistance=(Driving force)
(3)
in which b is a positive coefficient. In a theoretical sense, Eq. (3) is indeed consistent with the expression for free energy change of an ideal gas and solution, while this equation is also similar to the Maxwell /Boltzmann distribution law (Morel and Hering, 1993). As pointed out earlier, the adsorption reaction becomes less favorable as the adsorption proceeds, i.e. DG must increase accordingly. These seem to imply that Q would reflect the magnitude of adsorption resistance. On the other hand, the difference between Qth and Q represents the actual driving force of the biosorption process, a larger difference leads to a smaller value of DG . Therefore, Eq. (3) can be written as follows: DG DG o? bRT ln Q=(Qth Q)
(4)
Eq. (4) shows that when Q /0.5Qth, DGo? is equal to DG . This in turn implies that DGo? can be defined as the overall free energy change at Q / 0.5Qth, i.e. the driving force of biosorption is equal to the resistance force. As Q approaches Qth, DG goes to infinity and further adsorption becomes energetically impossible, this is indeed in agree-
ment with that stated by Morel and Hering (1993). Substitution of Eq. (2) into Eq. (4) yields DG DGo aRT ln C bRT ln Q=(Qth Q) (5) When the adsorption reaches its equilibrium, DG is zero. Hence, 0DG o aRT ln Ce bRT ln Qe =(Qeth Qe ) (6) in which Ce, Qe and Qeth are the respective value of C , Q and Qth at equilibrium. Solving Eq. (6) for Qe gives Qe Qeth e
DG o RT
C a=b e 1=b
(7) a=b
Ce
Eq. (7) can be rearranged as Qe Qeth
Cen Kads Cen
in which DGo m Kads e RT
(8)
(9)
and n and m equal a /b and 1/b, respectively. Analogue to a chemical reaction, the thermodynamic equilibrium constant of adsorption reaction (Keq) can be defined as Keq e
DG o RT
(10)
236
Y. Liu et al. / Journal of Biotechnology 102 (2003) 233 /239
Comparison of Eq. (9) and Eq. (10) shows that m 1 Kads (11) Keq Eq. (11) reveals the real physical meaning of Kads. A simple least square-based computer program was developed to estimate the constants involved in Eq. (8). However, as soon as Qeth is known in Eq. (8), plotting ln Qe/(Qeth/Qe) versus ln Ce gives a straight line as follows: lnQe =(Qeth Qe ) n ln Ce ln Kads
(12)
Values of Kads and n can be easily determined from the slope and intercept of Eq. (12).
Fig. 3. Biosorption isotherm of Cu2 by aerobic granules. Eq. (8) prediction is shown by solid curve. Qeth /59.6 mg g 1; Kads /7.1; n/0.52 and correlation coefficient/0.90.
4. Model verification Biosorption isotherms of Cd2, Cu2 and Zn2 at a constant granule concentration of 100 mg l 1 are shown in Figs. 2/4. It can be seen that the proposed model (Eq. (8)) can provide a satisfactory description for the experimental data obtained, indicated by a correlation coefficient greater than 0.90. For the Cd2 biosorption on aerobic granules, the value of n estimated is 1.2, 0.52 for Cu2 biosorption, and 1.02 for Zn2 biosorption. These n values seem to suggest that the Langmuir model could fit to the Zn2 biosorption data very well. The respective Kads value obtained is 13.6 for Cd2 biosorption, 7.1
Fig. 4. Biosorption isotherm of Zn2 by aerobic granules. Eq. (8) prediction is shown by solid curve. Qeth /164.5 mg g 1; Kads /51.3; n /1.02 and correlation coefficient/0.96.
for Cu2 biosorption and 51.3 for Zn2 biosorption on the aerobic granules. In order to further test the proposed model, the data published by Yang and Volesky (1999) were used. Fig. 5 shows that those data can be described by Eq. (8) very well. In fact, many literature data support Eq. (8) (Ruiz-Manriquez et al., 1998; Zhou, 1999; Aksu, 2001; Wang et al., 2001).
5. Discussion 2
Fig. 2. Biosorption isotherm of Cd by aerobic granules. Eq. (8) prediction is shown by solid curve. Qeth /172.7 mg g 1; Kads /13.6; n/1.2 and correlation coefficient/0.96.
Eq. (8) provides excellent fitting to the experimental data obtained in the biosorption tests of
Y. Liu et al. / Journal of Biotechnology 102 (2003) 233 /239
Fig. 5. Biosorption isotherm of Uranium on Sargassum biomass at pH of 3.2. Data from Yang and Volesky (1999). Eq. (8) prediction is shown by solid curve. Qeth /1.38 mmol g 1; Kads /0.24; n /0.73 and correlation coefficient/ 0.99.
individual Cd2, Cu2 and Zn2 by aerobic granules. The values of n estimated by Eq. (8) imply that the biosorption reaction of Cd2 by aerobic granules is subject to 1.2-order kinetics, 0.52-order for the Cu2 biosorption and 1.02order for Zn2 biosorption on the aerobic granules. In a sense of chemical reaction, the reaction order is directly related to the reaction mechanisms. Metal biosorption by microorganisms would be through either specific ion exchange mechanism on the surface of biosorbent or surface precipitation of metal hydroxide species. It seems certain that the overall reaction order of biosorption is dependent upon the characteristics of heavy metal as well as the nature of biosorbent used, and should be variable as found in this study. The second-order kinetic equation had been used to describe the biosorption of heavy metals (Ho and McKay, 1999; Aksu, 2001; Singh et al., 2001), while the biosorption process of Pb2 on fungal biomass followed the first-order reaction kinetics (Wang et al., 2001). As Eq. (11) shows, the value of 1/Kads is related to the magnitude of the thermodynamic equilibrium constant, Keq. When the value of Kads is small, i.e. the thermodynamic equilibrium constant, Keq is large. In this case, the biosorption reaction would proceed far towards
237
completion, and the position of equilibrium lies far toward the biosorption of metal. On the other hand, when the value of Kads is very large, i.e. Keq is very small, thus the position of equilibrium lies far towards the soluble metal. It is most likely that the value of Kads is determined by the natures of both biosorbent and the physicochemical characteristics of heavy metal. Consequently, the magnitude of Kads value may represent the equilibrium position of a biosorption process. The Langmuir and Freundlich sorption isotherms have been commonly used to describe the equilibrium behavior of biosorbents (Zhou, 1999; Aksu, 2001). When n equals 1, Eq. (8) is reduced to the following form, which is the same as the well known Langmuir adsorption isotherm: Qe Qeth
Ce Kads Ce
(13)
Eq. (13) shows that the Langmuir adsorption isotherm is only a particular case of Eq. (8) when n has a value of 1. This is confirmed by the data of Zn2 biosorption by aerobic granules (Fig. 4). When the value of Cne is much less than the value of Kads, Eq. (8) is simplified to the Freundlich adsorption isotherm such that Qe
Qeth n Ce Kads
(14)
Eq. (14) reveals that the Freundlich constant indeed equals Qeth/Kads. It appears from Eq. (13) and Eq. (14) that Eq. (8) can be regarded as a generalized form of the Langmuir and Freundlich models. It should be realized that if a /1, Eq. (7) thus reduces to the Sips or so-called Hill model (Roels, 1983; LeVan et al., 1998): Qe Qeth
DG
o
e RT
Cem m
(15) Cem
As Roels (1983) noted, the Hill model is a purely empirical equation. Eq. (15) seems to provide a theoretical interpretation for the empirical Hill model. The biosorption capacity of aerobic granules at equilibrium was 172.7 mg g1 for Cd2, 59.6 mg g1 for Cu2 and 164.5 mg g1 for Zn2.
238
Y. Liu et al. / Journal of Biotechnology 102 (2003) 233 /239
Aksu (2001) reported that the biosorption capacity of Cd2 on C. vulgaris was 85.3 mg g1, while Zhou (1999) found that the biosorption capacity of Zn2 by fungi was about 213 mmol g1, equivalent to 14 mg g1. The biosorption capacity of Cu2 by Thiobacillus ferrooxidans was found to be 40 mg g 1 (Ruiz-Manriquez et al., 1998). These seem to indicate that the biosorption capacities of individual cadmium, copper and zinc by aerobic granules are comparable with those conventional suspended biosorbents in the forms of microbial flocs or dispersed bacterial species. One serious operation problem-associated with those biosorbents in the forms of bioflocs or dispersed bacteria is the post-separation of used biosorbents from the treated effluent, and an extra settling tank is required. As compared to conventional floc-form biosorbents, aerobic granules have the advantages of regular, compact and strong microbial structure, and excellent settling ability, e.g. the settling velocity of the aerobic granules used in this study was about 71 m h1, while the settling velocity of conventional activated sludge was generally less than 10 m h1 (Campos et al., 1999). After the biosorption of heavy metals, the aerobic granules were settled down by gravity in 1 min, and were easily separated out from the treated effluent. It seems that the aerobic granule-based biosorption process is an efficient and cost-effective technology for the removal of heavy metals from industrial wastewaters.
6. Conclusions A general model (Eq. (8)) for the biosorption of heavy metals was derived, and was verified by the experimental results obtained in the biosorption tests of individual Cd2, Cu2 and Zn2. It was demonstrated that the proposed model was reduced to different well-known adsorption equations, such as the Langmuir, Freundlich and Hill models in some special cases. It is expected that one may gain good insights from Eq. (8) into the biosorption behaviors of heavy metals at equilibrium. The excellent settle ability of aerobic granules can ensure a rapid separation of biosolids from the treated effluent, which in turn leads to a
simple process design. This study probably for the first time shows that aerobic granules could be used as an effective biosorbent for efficient Cd2, Cu2 and Zn2 removal from industrial wastewater.
Acknowledgements We thank Kok-How Woon, Cher-Chong Ang and Su-Feng Tan for their skilled experimental works.
References Aksu, Z., 2001. Equilibrium and kinetic modeling of cadmium(II) biosorption by C. vulgaris in a batch system: effect of temperature. Sep. Purif. Technol. 21, 285 /294. Beun, J.J., Hendriks, A., van Loosdrecht, M.C.M., Morgenroth, E., Wilderer, P.A., Heijnen, J.J., 1999. Aerobic granulation in a sequencing batch reactor. Water Res. 33, 2283 /2290. Campos, J.L., Garrido-Fernandez, J.M., Mendez, R., Lema, J.M., 1999. Nitrification at high ammonia loading rates in an activated sludge unit. Bioresource Technol. 68, 141 / 148. Eccles, H., 1999. Treatment of metal-contaminated wastes: why select a biological process. Trend Biotechnol. 17, 462 / 465. Hamdy, A.A., 2000. Biosorption of heavy metals by marine algae. Current Microbiol. 41, 232 /238. Ho, Y.S., McKay, G., 1999. Pseudo-second order model for sorption processes. Proc. Biochem. 34, 451 /465. LeVan, M.D., Carta, G., Yon, C.M., 1998. Adsorption and ion exchange. In: Perry, R.H., Green, D.W., Maloney, J.O. (Eds.), Perry’s chemical engineer’s handbook. McGrawHill, New York. Metcalf, A., Eddy, S., 1991. Wastewater Engineering, 3rd ed. McGraw-Hill, Singapore. Morel, F.M.M., Hering, J.G., 1993. Principles and Applications of Aquatic Chemistry. Wiley, New York. Roels, J.A., 1983. Energetics and Kinetics in Biotechnology. Elsevier, New York. Ruiz-Manriquez, A., Magana, P.I., Lopez, V., Guzman, R., 1998. Biosorption of Cu by Thiobacillus ferrooxidans . Bioproc. Eng. 18, 113 /118. Singh, S., Rai, B.N., Rai, L.C., 2001. Ni (II) and Cr (VI) sorption kinetics by Microcystis in single and multimetallic system. Proc. Biochem. 36, 1205 /1213. Tay, J.H., Liu, Q.S., Liu, Y., 2001. The effects of shear force on the formation, structure and metabolism of aerobic granules. Appl. Microbiol. Biotechnol. 57, 227 /233.
Y. Liu et al. / Journal of Biotechnology 102 (2003) 233 /239 Wang, J.L., Zhan, X.M., Ding, D.C., Zhou, D., 2001. Bioadsorption of lead(II) from aqueous solution by fungal biomass of Aspergillus niger . J. Biotechnol. 87, 273 /277.
239
Yang, J.B., Volesky, B., 1999. Biosorption of Uranium on Sargassum biomass. Water Res. 33, 3357 /3363. Zhou, J.L., 1999. Zn biosorption by Rhizopus arrhizus and other fungi. Appl. Microbiol. Biotechnol. 51, 686 /693.
A general model for biosorption of Cd2, Cu2 and Zn2 by aerobic granules Yu Liu *, Hui Xu, Shu-Fang Yang, Joo-Hwa Tay Environmental Engineering Research Centre, School of Civil and Environmental Engineering, Nanyang Technological University, 50 Nanyang Avenue, Singapore 639798, Singapore Received 1 October 2002; received in revised form 15 November 2002; accepted 11 December 2002
Abstract Aerobic granules are microbial aggregates with a strong and compact structure. This study looked into the feasibility of aerobic granules as a novel type of biosorbent for the removal of individual Cd2 , Cu2 and Zn2 from aqueous solution. Based on the thermodynamics of biosorption reaction, a general model was developed to describe the equilibrium biosorption of individual Cd2 , Cu2 and Zn2 by aerobic granules. This model provides good insights into the thermodynamic mechanisms of biosorption of heavy metals. The model prediction was in good agreement with the experimental data obtained. It was further demonstrated that the Langmuir, Freundlich and Sips or Hill equations were particular cases of the proposed model. The biosorption capacity of individual Cd2 , Cu2 and Zn2 on aerobic granules was 172.7, 59.6 and 164.5 mg g 1, respectively. These values may imply that aerobic granules are effective biosorbent for the removal of Cd2 , Cu2 and Zn2 from industrial wastewater. # 2003 Elsevier Science B.V. All rights reserved. Keywords: Aerobic granule; Biosorption; Thermodynamics; Cadmium; Copper; Zinc; Equilibrium; Model
1. Introduction Heavy metals have been detected in a wide variety of industrial wastewater streams. Chemical precipitation is one of the most common methods used to remove heavy metals from aqueous solution. However, chemical slurry produced is a secondary waste that should be properly handled before discharging. Recently, research attention has been given to look for efficient and low cost
* Corresponding author. E-mail address: [email protected] (Y. Liu).
metal absorbents. For this purpose, many biomaterials had been tested as biosorbents for the biosorption of heavy metals, such as marine algae, fungal biomass, wasted activated sludge, digested sludge and so on (Eccles, 1999; Zhou, 1999; Hamdy, 2000; Wang et al., 2001). It should be realized that most biosorbents currently used are in the form of suspended biomass. The major problems-associated with this kind of biosorbents are post-separation of suspended biomass from the treated effluent, maintenance of biosorbent stability and regeneration of used biosorbents. These drawbacks of conventional biosorbents in the form of bioflocs seriously limit application of
0168-1656/03/$ - see front matter # 2003 Elsevier Science B.V. All rights reserved. doi:10.1016/S0168-1656(03)00030-0
234
Y. Liu et al. / Journal of Biotechnology 102 (2003) 233 /239
biological process in the removal of metals from wastewater. In the past few years, aerobic granulation technology had been successfully developed (Beun et al., 1999; Tay et al., 2001). Aerobic granules are microbial aggregates that have a strong and compact microbial structure. The settling velocity and density of aerobic granules are much higher than conventional bioflocs, while the granules have large surface area. In fact, when selecting appropriate biosorbents for the removal of heavy metals from industrial wastewater, the settleability of biosorbent is one of the most important criteria for effective separation of biosorbent from treated effluent. To date, no information is currently available on the biosorption kinetics of heavy metals by aerobic granules. Therefore, the main objective of the present study was to derive a general model for the biosorption of three individual model heavy metals, cadmium, copper and zinc by aerobic granules. It is expected that this work would be useful in the development of aerobic granules-based biosorbers for the removal of heavy metals from aqueous solution.
2. Materials and methods Aerobic granules used for biosorption tests were collected from a laboratory scale sequencing batch reactor (SBR) fed with acetate as sole carbon source. The operation conditions and setup of the SBR can be found elsewhere (Tay et al., 2001). The mean diameter of aerobic granule was around 1.0 mm (Fig. 1a). Fig. 1b shows the microbial structure of the aerobic granules, and it can be seen that a dominant species is present in the acetate-fed aerobic granules. These aerobic granules had a settling velocity of 71 m h1, and a sludge volumetric index of 49 ml g1. Individual cadmium, copper and zinc were used as the model heavy metals in biosorption tests. The respective Cd2, Cu2 and Zn2 solution was prepared by dissolving CdCl2, CuCl2 and ZnCl2 in deionic water. In this study, three series of batch experiments with the individual Cd2, Cu2 and Zn2 were conducted at a constant initial granule concentration of 100 mg dry weight l 1. The
initial Cd2, Cu2 and Zn2 concentration varied in the range of 5 /200 mg l1 and solution pH was adjusted to 4. The batch biosorption tests were performed in 1-l glass beakers with gentle agitation at a constant temperature of 26 8C. Each test lasted for 5 h. Soluble Cd2, Cu2 and Zn2 concentrations were determined in the course of batch experiments by using Inductively Coupled Plasma Emission Spectrometer (ICP) (PerkinElmer P400, Perkin-Elmer Corporation, Norwalk, USA).
3. Model development It had been proposed that the overall biosorption reaction could be regarded as a simple change in the state of metal ion (Morel and Hering, 1993), that is, C 0 Cads
DGo?
(1)
in which C and Cads is metal concentration in bulk solution, and that adsorbed at time t, while DGo? is the effective free energy change of biosorption, which changes with the proceeding of the biosorption reaction. If the metal concentration (C ) in bulk solution increases, its adsorption is more favorable, i.e. DGo? would decrease with the increase of the metal concentration (Metcalf and Eddy, 1991). According to Morel and Hering (1993), DGo? can be expressed as DG o? DG o aRT ln C
(2)
in which a is positive coefficient. Evidence shows that the real driving force of biosorption is the difference between the amount adsorbed by unit biosorbent (Q ) at a given metal concentration and the theoretical amount that could be adsorbed by unit biosorbent at that concentration (Qth), and this driving force is disappearing when the biosorption reaction gradually approaches its equilibrium state (Metcalf and Eddy, 1991). As biosorption proceeds, the driving force decreases and the adsorption resistance will increase. It is a reasonable consideration that the overall change of free energy of the biosorption process (DG ) would increase with the increase of adsorption resistance, and decrease with the increase of the
Y. Liu et al. / Journal of Biotechnology 102 (2003) 233 /239
235
Fig. 1. (a) Morphology of aerobic granules; (b) microbial structure of aerobic granules.
driving force of adsorption reaction. We assume that the overall change of free energy of the biosorption reaction should be formulated as the function of the driving force and resistance of adsorption such that DG DG o? bRT ln Resistance=(Driving force)
(3)
in which b is a positive coefficient. In a theoretical sense, Eq. (3) is indeed consistent with the expression for free energy change of an ideal gas and solution, while this equation is also similar to the Maxwell /Boltzmann distribution law (Morel and Hering, 1993). As pointed out earlier, the adsorption reaction becomes less favorable as the adsorption proceeds, i.e. DG must increase accordingly. These seem to imply that Q would reflect the magnitude of adsorption resistance. On the other hand, the difference between Qth and Q represents the actual driving force of the biosorption process, a larger difference leads to a smaller value of DG . Therefore, Eq. (3) can be written as follows: DG DG o? bRT ln Q=(Qth Q)
(4)
Eq. (4) shows that when Q /0.5Qth, DGo? is equal to DG . This in turn implies that DGo? can be defined as the overall free energy change at Q / 0.5Qth, i.e. the driving force of biosorption is equal to the resistance force. As Q approaches Qth, DG goes to infinity and further adsorption becomes energetically impossible, this is indeed in agree-
ment with that stated by Morel and Hering (1993). Substitution of Eq. (2) into Eq. (4) yields DG DGo aRT ln C bRT ln Q=(Qth Q) (5) When the adsorption reaches its equilibrium, DG is zero. Hence, 0DG o aRT ln Ce bRT ln Qe =(Qeth Qe ) (6) in which Ce, Qe and Qeth are the respective value of C , Q and Qth at equilibrium. Solving Eq. (6) for Qe gives Qe Qeth e
DG o RT
C a=b e 1=b
(7) a=b
Ce
Eq. (7) can be rearranged as Qe Qeth
Cen Kads Cen
in which DGo m Kads e RT
(8)
(9)
and n and m equal a /b and 1/b, respectively. Analogue to a chemical reaction, the thermodynamic equilibrium constant of adsorption reaction (Keq) can be defined as Keq e
DG o RT
(10)
236
Y. Liu et al. / Journal of Biotechnology 102 (2003) 233 /239
Comparison of Eq. (9) and Eq. (10) shows that m 1 Kads (11) Keq Eq. (11) reveals the real physical meaning of Kads. A simple least square-based computer program was developed to estimate the constants involved in Eq. (8). However, as soon as Qeth is known in Eq. (8), plotting ln Qe/(Qeth/Qe) versus ln Ce gives a straight line as follows: lnQe =(Qeth Qe ) n ln Ce ln Kads
(12)
Values of Kads and n can be easily determined from the slope and intercept of Eq. (12).
Fig. 3. Biosorption isotherm of Cu2 by aerobic granules. Eq. (8) prediction is shown by solid curve. Qeth /59.6 mg g 1; Kads /7.1; n/0.52 and correlation coefficient/0.90.
4. Model verification Biosorption isotherms of Cd2, Cu2 and Zn2 at a constant granule concentration of 100 mg l 1 are shown in Figs. 2/4. It can be seen that the proposed model (Eq. (8)) can provide a satisfactory description for the experimental data obtained, indicated by a correlation coefficient greater than 0.90. For the Cd2 biosorption on aerobic granules, the value of n estimated is 1.2, 0.52 for Cu2 biosorption, and 1.02 for Zn2 biosorption. These n values seem to suggest that the Langmuir model could fit to the Zn2 biosorption data very well. The respective Kads value obtained is 13.6 for Cd2 biosorption, 7.1
Fig. 4. Biosorption isotherm of Zn2 by aerobic granules. Eq. (8) prediction is shown by solid curve. Qeth /164.5 mg g 1; Kads /51.3; n /1.02 and correlation coefficient/0.96.
for Cu2 biosorption and 51.3 for Zn2 biosorption on the aerobic granules. In order to further test the proposed model, the data published by Yang and Volesky (1999) were used. Fig. 5 shows that those data can be described by Eq. (8) very well. In fact, many literature data support Eq. (8) (Ruiz-Manriquez et al., 1998; Zhou, 1999; Aksu, 2001; Wang et al., 2001).
5. Discussion 2
Fig. 2. Biosorption isotherm of Cd by aerobic granules. Eq. (8) prediction is shown by solid curve. Qeth /172.7 mg g 1; Kads /13.6; n/1.2 and correlation coefficient/0.96.
Eq. (8) provides excellent fitting to the experimental data obtained in the biosorption tests of
Y. Liu et al. / Journal of Biotechnology 102 (2003) 233 /239
Fig. 5. Biosorption isotherm of Uranium on Sargassum biomass at pH of 3.2. Data from Yang and Volesky (1999). Eq. (8) prediction is shown by solid curve. Qeth /1.38 mmol g 1; Kads /0.24; n /0.73 and correlation coefficient/ 0.99.
individual Cd2, Cu2 and Zn2 by aerobic granules. The values of n estimated by Eq. (8) imply that the biosorption reaction of Cd2 by aerobic granules is subject to 1.2-order kinetics, 0.52-order for the Cu2 biosorption and 1.02order for Zn2 biosorption on the aerobic granules. In a sense of chemical reaction, the reaction order is directly related to the reaction mechanisms. Metal biosorption by microorganisms would be through either specific ion exchange mechanism on the surface of biosorbent or surface precipitation of metal hydroxide species. It seems certain that the overall reaction order of biosorption is dependent upon the characteristics of heavy metal as well as the nature of biosorbent used, and should be variable as found in this study. The second-order kinetic equation had been used to describe the biosorption of heavy metals (Ho and McKay, 1999; Aksu, 2001; Singh et al., 2001), while the biosorption process of Pb2 on fungal biomass followed the first-order reaction kinetics (Wang et al., 2001). As Eq. (11) shows, the value of 1/Kads is related to the magnitude of the thermodynamic equilibrium constant, Keq. When the value of Kads is small, i.e. the thermodynamic equilibrium constant, Keq is large. In this case, the biosorption reaction would proceed far towards
237
completion, and the position of equilibrium lies far toward the biosorption of metal. On the other hand, when the value of Kads is very large, i.e. Keq is very small, thus the position of equilibrium lies far towards the soluble metal. It is most likely that the value of Kads is determined by the natures of both biosorbent and the physicochemical characteristics of heavy metal. Consequently, the magnitude of Kads value may represent the equilibrium position of a biosorption process. The Langmuir and Freundlich sorption isotherms have been commonly used to describe the equilibrium behavior of biosorbents (Zhou, 1999; Aksu, 2001). When n equals 1, Eq. (8) is reduced to the following form, which is the same as the well known Langmuir adsorption isotherm: Qe Qeth
Ce Kads Ce
(13)
Eq. (13) shows that the Langmuir adsorption isotherm is only a particular case of Eq. (8) when n has a value of 1. This is confirmed by the data of Zn2 biosorption by aerobic granules (Fig. 4). When the value of Cne is much less than the value of Kads, Eq. (8) is simplified to the Freundlich adsorption isotherm such that Qe
Qeth n Ce Kads
(14)
Eq. (14) reveals that the Freundlich constant indeed equals Qeth/Kads. It appears from Eq. (13) and Eq. (14) that Eq. (8) can be regarded as a generalized form of the Langmuir and Freundlich models. It should be realized that if a /1, Eq. (7) thus reduces to the Sips or so-called Hill model (Roels, 1983; LeVan et al., 1998): Qe Qeth
DG
o
e RT
Cem m
(15) Cem
As Roels (1983) noted, the Hill model is a purely empirical equation. Eq. (15) seems to provide a theoretical interpretation for the empirical Hill model. The biosorption capacity of aerobic granules at equilibrium was 172.7 mg g1 for Cd2, 59.6 mg g1 for Cu2 and 164.5 mg g1 for Zn2.
238
Y. Liu et al. / Journal of Biotechnology 102 (2003) 233 /239
Aksu (2001) reported that the biosorption capacity of Cd2 on C. vulgaris was 85.3 mg g1, while Zhou (1999) found that the biosorption capacity of Zn2 by fungi was about 213 mmol g1, equivalent to 14 mg g1. The biosorption capacity of Cu2 by Thiobacillus ferrooxidans was found to be 40 mg g 1 (Ruiz-Manriquez et al., 1998). These seem to indicate that the biosorption capacities of individual cadmium, copper and zinc by aerobic granules are comparable with those conventional suspended biosorbents in the forms of microbial flocs or dispersed bacterial species. One serious operation problem-associated with those biosorbents in the forms of bioflocs or dispersed bacteria is the post-separation of used biosorbents from the treated effluent, and an extra settling tank is required. As compared to conventional floc-form biosorbents, aerobic granules have the advantages of regular, compact and strong microbial structure, and excellent settling ability, e.g. the settling velocity of the aerobic granules used in this study was about 71 m h1, while the settling velocity of conventional activated sludge was generally less than 10 m h1 (Campos et al., 1999). After the biosorption of heavy metals, the aerobic granules were settled down by gravity in 1 min, and were easily separated out from the treated effluent. It seems that the aerobic granule-based biosorption process is an efficient and cost-effective technology for the removal of heavy metals from industrial wastewaters.
6. Conclusions A general model (Eq. (8)) for the biosorption of heavy metals was derived, and was verified by the experimental results obtained in the biosorption tests of individual Cd2, Cu2 and Zn2. It was demonstrated that the proposed model was reduced to different well-known adsorption equations, such as the Langmuir, Freundlich and Hill models in some special cases. It is expected that one may gain good insights from Eq. (8) into the biosorption behaviors of heavy metals at equilibrium. The excellent settle ability of aerobic granules can ensure a rapid separation of biosolids from the treated effluent, which in turn leads to a
simple process design. This study probably for the first time shows that aerobic granules could be used as an effective biosorbent for efficient Cd2, Cu2 and Zn2 removal from industrial wastewater.
Acknowledgements We thank Kok-How Woon, Cher-Chong Ang and Su-Feng Tan for their skilled experimental works.
References Aksu, Z., 2001. Equilibrium and kinetic modeling of cadmium(II) biosorption by C. vulgaris in a batch system: effect of temperature. Sep. Purif. Technol. 21, 285 /294. Beun, J.J., Hendriks, A., van Loosdrecht, M.C.M., Morgenroth, E., Wilderer, P.A., Heijnen, J.J., 1999. Aerobic granulation in a sequencing batch reactor. Water Res. 33, 2283 /2290. Campos, J.L., Garrido-Fernandez, J.M., Mendez, R., Lema, J.M., 1999. Nitrification at high ammonia loading rates in an activated sludge unit. Bioresource Technol. 68, 141 / 148. Eccles, H., 1999. Treatment of metal-contaminated wastes: why select a biological process. Trend Biotechnol. 17, 462 / 465. Hamdy, A.A., 2000. Biosorption of heavy metals by marine algae. Current Microbiol. 41, 232 /238. Ho, Y.S., McKay, G., 1999. Pseudo-second order model for sorption processes. Proc. Biochem. 34, 451 /465. LeVan, M.D., Carta, G., Yon, C.M., 1998. Adsorption and ion exchange. In: Perry, R.H., Green, D.W., Maloney, J.O. (Eds.), Perry’s chemical engineer’s handbook. McGrawHill, New York. Metcalf, A., Eddy, S., 1991. Wastewater Engineering, 3rd ed. McGraw-Hill, Singapore. Morel, F.M.M., Hering, J.G., 1993. Principles and Applications of Aquatic Chemistry. Wiley, New York. Roels, J.A., 1983. Energetics and Kinetics in Biotechnology. Elsevier, New York. Ruiz-Manriquez, A., Magana, P.I., Lopez, V., Guzman, R., 1998. Biosorption of Cu by Thiobacillus ferrooxidans . Bioproc. Eng. 18, 113 /118. Singh, S., Rai, B.N., Rai, L.C., 2001. Ni (II) and Cr (VI) sorption kinetics by Microcystis in single and multimetallic system. Proc. Biochem. 36, 1205 /1213. Tay, J.H., Liu, Q.S., Liu, Y., 2001. The effects of shear force on the formation, structure and metabolism of aerobic granules. Appl. Microbiol. Biotechnol. 57, 227 /233.
Y. Liu et al. / Journal of Biotechnology 102 (2003) 233 /239 Wang, J.L., Zhan, X.M., Ding, D.C., Zhou, D., 2001. Bioadsorption of lead(II) from aqueous solution by fungal biomass of Aspergillus niger . J. Biotechnol. 87, 273 /277.
239
Yang, J.B., Volesky, B., 1999. Biosorption of Uranium on Sargassum biomass. Water Res. 33, 3357 /3363. Zhou, J.L., 1999. Zn biosorption by Rhizopus arrhizus and other fungi. Appl. Microbiol. Biotechnol. 51, 686 /693.