Water Research 36 (2002) 30633073

Equilibrium modelling of individual and simultaneous biosorption of chromium(VI) and nickel(II) onto dried activated sludge
. Z. Aksu*, U. Ackel, E. Kabasakal, S. Tezer Department of Chemical Engineering, Hacettepe University, Beytepe, 06532 Ankara, Turkey Received 30 November 2000; accepted 16 November 2001

Abstract The biosorption of chromium(VI) and nickel(II) ions, both singly and in combination, by dried activated sludge was investigated in a batch system as a function of initial pH and single- and dual-metal ion concentrations. The working initial pH values for single chromium(VI) and nickel(II) biosorptions were determined as 1.0 and 4.5, respectively. It was observed that the co-ion effect on the equilibrium uptake became more pronounced as the co-ion concentration in solution increased and pH level increased for chromium(VI) and decreased for nickel(II). Adsorption isotherms were developed for both the single- and dual-metal ion systems at these two pH values and expressed by the mono- and multi-component Langmuir and Freundlich adsorption models and model parameters were estimated by the non-linear regression. It was seen that the mono-component adsorption equilibrium data tted very well to both the monocomponent adsorption models for both the components and the pH values studied while the multi-component Freundlich adsorption model adequately predicted the multi-component adsorption equilibrium data at moderate ranges of initial mixture concentrations for both the studied pH values. r 2002 Elsevier Science Ltd. All rights reserved.
Keywords: Simultaneous biosorption; Chromium(VI); Nickel(II); Dried activated sludge; Mono- and multi-component adsorption models

1. Introduction It is well recognized that the presence of heavy metals in the environment can be detrimental to a variety of living species, including man. The methods which have been used to remove heavy metal ions, such as chemical precipitation, adsorption, ion-exchange, solvent extraction have been found to be limited, since they often involve high capital and operational costs and may also be associated with the generation of secondary wastes which present treatment problems. The use of microorganisms as biosorbents for heavy metals offers a potential alternative to existing methods for detoxica*Corresponding author. Tel.: +90-312-2977434; fax: +90312-2992124. E-mail address: zaksu@hacettepe.edu.tr (Z. Aksu).

tion and recovery of these components from industrial wastewaters and is a subject of extensive studies. The special surface properties of microorganisms enable them to adsorb heavy metal ions from solutions. This passive bioaccumulation process (biosorption) has distinct advantages over the conventional methods: the process does not produce chemical sludges (i.e. nonpolluting), it could be highly selective, more efcient, easy to operate and hence cost effective for the treatment of large volumes of wastewaters-containing low pollutant concentrations. Industrial applications of biosorption often make use of dead biomass, which does not require nutrients and can be exposed to environments of high toxicity [118]. Much of the work on the biosorption of heavy metal ions by various kinds of microorganisms has focused on the uptake of single metals. In practice, wastewaters are

0043-1354/02/$ - see front matter r 2002 Elsevier Science Ltd. All rights reserved. PII: S 0 0 4 3 - 1 3 5 4 ( 0 1 ) 0 0 5 3 0 - 9

3064

Z. Aksu et al. / Water Research 36 (2002) 30633073

Nomenclature ni b mono-component (non-competitive) Langmuir adsorption constant of the single component (dm3 mg1) individual Langmuir adsorption constant of each component (mg g1) unadsorbed concentration of the single component at equilibrium (mg dm3) unadsorbed concentration of each component at equilibrium (mg dm3) initial concentration of each component (mg dm3) mono-component (non-competitive) Freundlich adsorption constant of the single component ((mg g1) (mg dm3)n) individual Freundlich adsorption constant of each component ((mg g1) (mg dm3)n) mono-component (non-competitive) Freundlich adsorption constant of the single xi ; yi ; zi qeq

bi Ceq Ceqi C0 KF

qeqi

Q0

Q0 i X Zi

KFi n

component individual Freundlich adsorption constant of each component multi-component (competitive) Freundlich adsorption constants of each component adsorbed quantity of the single component per gram of dried anaerobic activated sludge at equilibrium (mg g1) adsorbed quantity of each component per gram of dried anaerobic activated sludge at equilibrium (mg g1) mono-component (non-competitive) Langmuir adsorption constant of the single component (mg g1) individual Langmuir adsorption constant of each component (mg g1) biosorbent (dried activated sludge) concentration (g dm3) multi-component (competitive) Langmuir adsorption constant of each component

polluted with multiple metals. In addition, the equilibrium modelling of multi-metal biosorption, which is important in the design of treatment systems, has largely been neglected. The examination of the effects of binary metal ions in various combinations is more representative, of the actual environmental problems faced by organisms, than the single-metal studies. Bioremoval of single species of metal ions using microorganisms is affected by several factors. These factors include the specic surface properties of the microorganism and the physicochemical parameters of the solution such as temperature, pH, initial metal ion concentration and biomass concentration [1,3,5,6,8,9,1113,16]. The combined effects of two or more metal ions on microorganisms also depend on the number of metal ions competing for binding sites, metal ion combination, levels of metal ion concentration and order of metal ion addition (Ting and Prince, 1991) [4,10,12,14,15,18]. The equilibrium established between adsorbed component on the biosorbent and unadsorbed component in solution can be represented by adsorption isotherms. The most widely used isotherm equation for modelling equilibrium is the Langmuir equation which is valid for monolayer sorption onto a surface with a nite number of identical sites which are homogeneously distributed over the sorbent surface and is given by Eq. (1). qeq Q0 bCeq ; 1 bCeq 1

after binding, respectively. Q0 is the maximum amount of metal ion per unit weight of sorbent to form a complete monolayer on the surface bound at high Ceq ; and b is a constant related to the afnity of the binding sites. Q0 and b can be determined from Ceq =qeq versus Ceq plot [1,3,5,6,813,1521]. The Freundlich expression is an empirical equation based on sorption on a heterogeneous surface suggesting (as expected) that binding sites are not equivalent and/or independent. The mono-component Freundlich equation is given below:
1=n qeq KF Ceq ;

where qeq is the amount of metal ion bound to per gram of dried biomass at equilibrium and Ceq is the residual (equilibrium) metal ion concentration left in solution

where KF and n are the mono-component Freundlich constants related to the sorption capacity and sorption intensity of the sorbent, respectively. Eq. (2) can be linearized in logarithmic form and Freundlich constants can be determined. Both models were developed for a single-layer adsorption. However, the Freundlich model physically provides a more realistic description of adsorption by organic matter because it accounts for different binding sites. But, in most cases, both equations t the data set reasonably well for the experimental data over moderate ranges of concentration [1,5,6, 8,10,11,13,16,18,20,21]. One of the difculties in describing the adsorption of pollutants from wastestreams is that wastewaters contain not one, but many kinds of pollutants. When several components are present, interference and competition phenomena for adsorption sites occur and lead to a more complex mathematical formulation of the equilibrium. Therefore, multi-component isotherms

Z. Aksu et al. / Water Research 36 (2002) 30633073

3065

attempt to express relationships between the adsorbed quantity of one component and the concentrations of all other components, either in solution or already adsorbed. Several isotherms have been proposed to describe equilibrium for such a system. These isotherms range from simple models related to the individual isotherm parameters only (non-modied adsorption models), to more complex models related to the individual isotherm parameters and to correction factors (modied adsorption models) [10,13,15,1822]. The long and tedious experimental method for obtaining competitive adsorption data has rapidly given rise to the idea of predicting multi-component isotherms from the single component ones only and competitive non-modied adsorption models were developed. The non-modied Langmuir model is written as: qeqi Q0 bi Ceqi i ; P 1 N bj Ceqj j1 3

where Ceqi and qeqi are the unadsorbed concentration of each component at equilibrium and the adsorbed quantity of each component per gram of dried alga at equilibrium, respectively. bi and Q0 are derived from the i corresponding individual Langmuir isotherm equations [10,13,15,1822]. Individual adsorption constants may not dene exactly the multi-component adsorption behaviour of metal ion mixtures. For that reason, better accuracy may be achieved by using modied isotherms related to the individual isotherm parameters and to correction factors. For instance, an interaction term Zi which is a characteristic of each species and depends on the concentrations of the other components has been dened in the modied Langmuir model. The modied Langmuir isotherm becomes qeqi Q0 bi Ceqi =Zi i ; P 1 N bj Ceqj =Zj j1 4

where KF1; KF2 and n1 and n2 are derived from the corresponding individual Freundlich isotherm equations and the other six parameters (x1 ; y1 ; z1 and x2 ; y2 ; z2 ) are the multi-component Freundlich adsorption constants of the rst and the second components [10,13,15,1820]. However, there is no general formula applicable to all sorbate/sorbent systems due to the wide range of systems leading to a wide variety of equilibrium behaviour. Probably, the most abundant source of mixed microbial biomass is the activated sludge wastewater treatment process. The ability of activated sludge biomass to remove and accumulate heavy metals has been recognized and activated sludge systems have been studied to a certain degree for their biosorption capabilities [3]. However, there has been no research on the simultaneous bioremoval and the expression of the adsorption isotherms of metal ion mixtures by the biomass to date. In this study, a process of simultaneous biosorption of chromium(VI) and nickel(II), which are frequently encountered together in wastewaters such as from stainless steel, metal plating, dye and painting industries [23,24] on dried activated sludge containing a mixed culture of microorganisms, was studied and compared to a single-component situation. Some of the adsorption models used to predict and/or correlate mono- and multi-component equilibrium data were presented and applied to all combinations of two metal ions in their aqueous solutions. Moreover, a statistical comparison of the methods was carried out.

2. Materials and methods 2.1. Activated sludge, growth conditions and preparation as the biosorbent Activated sludge, a complex consortium of microorganisms mainly containing bacteria obtained from Hacettepe University, Environmental Engineering Department was used in this study. The microorganisms were grown in agitated and aerated liquid media containing glucose (10 g dm3), urea (2.62 g dm3), diammoniumphosphate (0.853 g dm3) and magnesium sulphate (0.05 g dm3). The pH of the medium was adjusted to 6.56.8 with dilute and concentrated H2SO4 and NaOH solutions before sterilization. The experimental set-up basically consisted of a chemostat unit with 2.5 dm3 working volume and control and monitoring units for temperature, stirring rate, air ow rate (Biotech LP 100 model). Conditions in the reactor were maintained at an agitation rate of 150 rpm and a temperature of 251C. Air was continuously fed to the reactor. After the growth period, the cells were harvested and washed thoroughly with sterile distilled water and then

where Zi is the Langmuir correction coefcient of the i component estimated from competitive adsorption data. For binary mixtures, Eq. (4) can be rewritten as for the rst and the second component, respectively and these two equations can be solved simultaneously to obtain the multi-component Langmuir adsorption constants of the rst and the second components, respectively [10,13,15,1820]. The empirical extended form of the Freundlich model restricted to binary mixtures can be given by Eqs. (5) and (6) for each component of binary system: qeq1 KF1 Ceq1 1
1=n x1 x1 z1 Ceq1 y1 Ceq2

qeq2

KF2 Ceq2 2

1=n x2

x2 z2 Ceq2 y2 Ceq1

3066

Z. Aksu et al. / Water Research 36 (2002) 30633073

dried at 601C for 24 h before use. For the biosorption studies, 5 g of dried biomass was suspended in 100 dm3 of double-distilled water and homogenized in a homogenizer (Janke and Kunkel, IKA-Labortechnick, Ultra Turrax T25) at 8000 rpm for 20 min and then stored in the refrigerator. 2.2. Chemicals The test solutions containing single chromium(VI) or nickel(II) ions were prepared by diluting 1 g dm3 of stock solutions of chromium(VI) and nickel(II) to the desired concentrations. Stock solutions of chromium (VI) and nickel(II) were obtained by dissolving the exact quantities of K2Cr2O7 (Merck) and NiSO4 6H2O (Merck), in 1 dm3 of double-distilled water, respectively. The ranges of concentrations of both components prepared from stock solutions varied between 25 and 500 mg dm3. The test solutions containing desired combinations of chromium(VI) and nickel(II) were prepared by diluting stock solutions of chromium(VI) and nickel(II) and mixing. Before mixing the dried activated sludge solution, the pH of each test solution was adjusted to the required value with dilute and concentrated H2SO4 and NaOH solutions. 2.3. Batch biosorption studies A microorganism suspension of 10 cm3 was mixed with 90 cm3 of solution containing a known concentration of chromium(VI), nickel(II) or binary mixture of these components in an Erlenmeyer ask at the desired temperature and pH. All the nal solutions contained 0.5 g dm3 of biosorbent. The asks were agitated on a shaker at 150 rpm constant shaking rate for 24 h to ensure equilibrium was reached. Samples of 5 cm3 were taken before mixing the biosorbent solution and chromium(VI), nickel(II) or chromium(VI)- and nickel(II)-bearing solution, at 5 min intervals at the beginning of adsorption and 1530 min intervals after reaching equilibrium, centrifuged at 5000 rpm for 3 min and then the supernatant liquid was used to analyse for each metal ion. The studies were performed at a constant temperature of 251C to be representative of environmentally relevant conditions. The biosorption experiments were done in duplicates. The data were the mean values of two replicate determinations. 2.4. Analysis of chromium(VI) and nickel(II) The concentration of residual nickel(II) in the biosorption medium was determined in Unicam 929 atomic absorption spectrophotometer with the detection limit of 0.063 ppm at the wavelength of 232 nm. The concentration of residual chromium(VI) in the biosorption medium was determined spectrophotometrically.

The absorbance of the violet coloured complex of chromium(VI) with diphenyl carbazide was read at 540 nm [25].

3. Results and discussion Simultaneous biosorption of chromium(VI) and nickel(II) to dried activated sludge from binary mixture was investigated and compared with single-chromium (VI) or single-nickel(II) situation in a batch stirred system in this study. The mono- and multi-component sorption phenomena were expressed by the mono- and multi-component Langmuir and Freundlich adsorption models and the adsorption isotherms were developed for both the single- and dual-component systems at pH 1.0 and 4.5 and model parameters were determined by nonlinear regression. The equilibrium results were given as the units of adsorbed chromium(VI) or nickel(II) quantity per gram of dried biomass (qeq ; mg g1) and unadsorbed chromium(VI) or nickel(II) concentration (Ceq ; mg dm3) at equilibrium. 3.1. Effect of initial pH on the biosorption of chromium(VI) and nickel(II) One of the most critical parameters in the treatment of this multi-component system by the dried activated sludge is the initial pH of biosorption medium. Different metal ions may have different pH optima, possibly due to the different solution chemistry of the species. The initial pH optimum for biosorption is also microorganism-dependent because of different adsorptive sites of different species of microorganisms. The uptake of chromium(VI) and nickel(II) by the biomass is a function of initial pH too. The effect of initial pH on the equilibrium uptake of nickel(II) ions was investigated between pH 1.06.0 since the precipitation by the formation of nickel hydroxide (Ni(OH)2) may occur above pH 7.7 [26]. The initial pH values for chromium (VI) also changed from 1.0 to 6.0 where chromium(VI) precipitation was not observed. As shown in Fig. 1, the uptake of free ionic nickel(II) increased by increasing initial pH and was the greatest at pH 4.5. The chromium(VI) ions were more effectively adsorbed by the biomass than nickel(II) at low pH values. The working adsorption pH was determined as 1.0 for chromium(VI). It is well-known that the dominant form of chromium(VI) at this pH value is the acid chromate ion species (HCrO) and increasing the pH will shift 4 the concentration of HCrO to other forms, CrO2 and 4 4 Cr2O2. It can be concluded that the active form of 7 chromium(VI) was adsorbed by the biomass [26,27]. The different pH binding proles for these components could be due to the nature of the chemical

Z. Aksu et al. / Water Research 36 (2002) 30633073

100 Chromium(VI) Nickel(II) 80

3067

8 Chromium(VI); pH=1.0 Chromium(VI); pH=4.5 6 Nickel(II); pH=1.0 Nickel(II); pH=4.5

60

C eq /q eq (g dm-3 )

q eq (mg g -1 )

40

20

0 0 1 2 3 pH 4 5 6 7

100

200

300

400

500

C eq (mg dm -3 )

Fig. 1. The effect of initial pH on the individual equilibrium uptake of chromium(VI) and nickel(II) (C0 : 100 mg dm3, T: 251C, X : 0.5 g dm3).

Fig. 2. The mono-component Langmuir adsorption isotherms of chromium(VI) and nickel(II) obtained at pH 1.0 and 4.5 in the single-component situation (T: 251C; X : 0.5 g dm3).

interactions of each component with the microbial cells. The low level of nickel(II) uptake at lower pH values could be attributed to the increased concentration of hydrogen (H+) and hydronium (H3O+) ions competing for nickel(II)-binding sites on the biomass. The increase in nickel(II) biosorption at higher pH values may be explained by the ionization of functional groups on the cell surface which serve as the binding sites related to the isoelectric point of the cells. Heavy metal ions such as nickel(II) have a strong afnity for proteins of the cell wall. At pH values above the isoelectric point, there is a net negative charge on the cell surface and the ionic state of ligands such as carboxyl, phosphoryl, sulphydryl, hydroxyl and amino groups will be such that so as to promote reaction with the metal cations. As the pH is lowered, however, the overall surface charge on the cells will become positive, which will inhibit the approach of positively charged metal cations. The chromium(VI) ions being in the anionic form of HCrO 4 in the aqua solution will interact with the cell surface in this case [13,21,24,2628]. It is also obvious that the proposed biosorption mechanisms due to the initial pH are not sufcient to explain the biosorptions of both components observed at all the pH values studied. It is thought that additional types of biosorption mechanisms such as complex formation, chelation and microprecipitation or membrane transport and physicochemical forces such as van der Waals, H-binding are also important for the bioremoval of chromium(VI) and nickel(II) ions by the biomass, irrespective of initial pH [2,13,28]. However, the initial pH of wastewater could provide selectivity for the removal of the desired component in the mixture of chromium(VI) and nickel(II) and this situation was observed in the simultaneous removal of these components studying at these two initial pH values.

3.2. Biosorption of chromium(VI) and nickel(II): application of mono-component adsorption models Equilibrium data, commonly known as adsorption isotherms, are basic requirements for the design of adsorption systems. To determine the mono-component isotherms, initial concentrations of chromium(VI) or nickel(II) were varied while the amount of biosorbent in each sample was kept constant for each component at these two pH values. The individual Langmuir and Freundlich adsorption isotherms of chromium(VI) and nickel(II) ions obtained at pH 1.0 and 4.5 are shown in Figs. 2 and 3. The individual adsorption constants for each component obtained by evaluating the isotherms are also listed in Table 1 with the linear regression coefcients. In view of the values of linear regression coefcients in the table, the Freundlich model exhibited a slightly better t to the adsorption data of chromium (VI) and nickel(II) than the Langmuir model in the studied concentration range for both the pH values studied. However, the Langmuir model also seemed to agree very well with the experimental data of two components considering that obtained linear regression coefcients are >0.987. Adsorption model constants, the values of which express the surface properties and afnity of the biosorbent, can be used to compare the adsorptive capacity of biosorbent for different components. The magnitude of KF and n; the mono-component Freundlich constants, showed a selective uptake of chromium(VI) at pH 1.0 and of nickel(II) at pH 4.5 from wastewater with higher adsorptive capacity of biomass. Table 1 also showed that n > 1; indicating that both the chromium(VI) and nickel(II) ions are favourably adsorbed by dried activated sludge at both the studied pH values.

3068

Z. Aksu et al. / Water Research 36 (2002) 30633073

While the Freundlich model does not describe the saturation behaviour of the biosorbent, Q0 ; the monocomponent Langmuir constant is the monolayer saturation at equilibrium. The other mono-component Langmuir constant, b; corresponds to the concentration at which a nickel(II) or a chromium(VI) ion amount of Q0 =2 is bound and indicates the afnity for the binding of nickel(II) or chromium(VI) ions. The magnitude of Q0 indicated that the amount of chromium(VI) ions per unit weight of biosorbent to form a complete monolayer on the surface was much higher than that of nickel(II) at pH 1.0. The value of Q0 also appeared to be signicantly higher for nickel(II) at pH 4.5 in comparison with the uptake of chromium(VI) at the same pH value. The higher value of b implied the strong bonding of chromium(VI) to the cells at pH 4.5 and nickel(II) ions at pH 1.0.

3.3. Biosorption of binary mixture of chromium(VI) and nickel(II): application of the multi-component adsorption isotherms To determine the effects of initial chromium(VI) and nickel(II) ion concentrations on the equilibrium uptake of chromium(VI), the initial chromium(VI) concentrations were varied between 25 and 500 mg dm3 while the initial nickel(II) concentration in each biosorption medium was held constant at 0, 25, 50, 100, 250 or 500 mg dm3 at pH 1.0 and 4.5. The non-linearized adsorption isotherms of chromium(VI) in the absence of nickel(II) ions and in the presence of increasing concentrations of nickel(II) ions obtained at the two pH values are shown in Figs. 4 and 5. Equilibrium chromium(VI) uptake increased by increasing the initial chromium(VI) concentration up to 500 mg dm3 at both the pH values. The curvilinear relationship between the amount of chromium(VI) adsorbed per unit weight of microorganism (qeq ) and the residual chromium(VI) concentration at equilibrium (Ceq ) suggests that saturation of constant number of cell-binding sites occurred at higher concentrations of this metal ion. As seen from the gures, higher chromium(VI) uptakes were obtained at pH 1.0 as expected since the pH value of 1.0 had a selective effect for the removal of chromium(VI) in the mixture of nickel(II) and chromium(VI). When the equilibrium uptakes of chromium(VI) ions by dried activated sludge biomass in the presence of nickel(II) ions are compared with the situation where the chromium(VI) ions are present singly, there appeared to be an inhibition in the equilibrium uptakes of chromium(VI) ions. The equilibrium uptake of chromium(VI) decreased regularly with increasing concentrations of nickel(II) ions at pH 4.5. The inhibitory effect of nickel(II) ions on the equilibrium chromium(VI) uptake was dominant at higher initial nickel(II) concentrations at both the pH values. At 100 mg dm3 of

6 Chromium(VI); pH=1.0 Chromium(VI); pH=4.5 5 Nickel(II); pH=1.0 Nickel(II); pH=4.5 lnq eq

2 2 3 4 lnC eq 5 6 7

Fig. 3. The mono-component Freundlich adsorption isotherms of chromium(VI) and nickel(II) obtained at pH 1.0 and 4.5 in the single-component situation (T: 251C; X : 0.5 g dm3).

Table 1 Comparison of the individual Langmuir and Freundlich adsorption constants obtained from the mono-component Langmuir and Freundlich adsorption models for chromium(VI) and nickel(II) at pH 1.0 and 4.5 pH 1.0 Component Q (mg g )
0
1

pH 4.5 b (dm mg ) 0.0063 0.0079

3 1

Q0 (mg g1) 95.2 238.1 pH 4.5

b (dm3 mg1) 0.0055 0.0048

R2 0.992 0.987

The mono-component Langmuir model Chromium(VI) 294.1 Nickel(II) 106.4 pH 1.0 KF ((mg g ) (mg dm ) ) The mono-component Freundlich model Chromium(VI) 4.99 Nickel(II) 3.21
1 3 n

0.986 0.991

n 1.55 1.82

R2 0.994 0.992

KF ((mg g1) (mg dm3)n) 1.60 3.80

n 1.62 1.56

R2 0.991 0.994

Z. Aksu et al. / Water Research 36 (2002) 30633073

250

3069

200 q eq,Cr(VI) (mg g -1 )

150

100

50

0 0 75 150 225 300 375 450

C eq,Cr(VI) (mg dm -3 ) Co,Ni(II) =0 mg dm-3 Co,Ni(II) =25 mg dm-3 Co,Ni(II) =50 mg dm-3 Co,Ni(II) =100 mg dm-3 Co,Ni(II) =250 mg dm-3 Co,Ni(II) =500 mg dm-3

Fig. 4. The comparison of the non-linearized adsorption isotherms of chromium(VI) adsorption to dried activated sludge with chromium(VI) present as the single component and in the presence of increasing concentrations of nickel(II) at pH 1.0 (T: 251C, X : 0.5 g dm3).

80

60 q eq,Cr(VI) (mg g -1 )

40

20

0 0 100 200 300 400 500 C eq,Cr(VI) (mg dm -3 ) Co,Ni(II) = 0 mg dm-3 Co,Ni(II) =25 mg dm-3 Co,Ni(II) =50 mg dm-3 Co,Ni(II) =100 mg dm-3 Co,Ni(II) =250 mg dm-3 Co,Ni(II) =500 mg dm-3

Fig. 5. The comparison of the non-linearized adsorption isotherms of chromium(VI) adsorption to dried activated sludge with chromium(VI) present as the single component and in the presence of increasing concentrations of nickel(II) at pH 4.5 (T: 251C, X : 0.5 g dm3).

initial chromium(VI) concentration, in the absence of nickel(II) ions and in the presence of 100 mg dm3 of nickel(II) concentration at pH 1.0 and 4.5, adsorbed chromium(VI) quantities at equilibrium were found as

75.0 and 26.1 mg g1 and as 58.0 and 21.5 mg g1, respectively. Although chromium(VI) was preferentially adsorbed at pH 1.0, nickel(II) ions were also adsorbed by the cells at this pH value. This situation shows the partial competition between the chromium(VI) and nickel(II) ions although the initial pH has a signicant effect in the selective biosorption. The presence of second metal ion develops a competition for the adsorption sites on the surface and some sites are occupied by this component. As a consequence, the rst metal ion has a smaller parking space and its uptake is decreased. It is also obvious that not only one but different kinds of biosorption mechanisms are important in the simultaneous biosorption of chromium(VI) and nickel(II) ions on the dried activated sludge. There are possible interactions between different species in solution and potential interactions with the surface, particularly depending on the biosorption mechanism. Factors that affect the adsorption capacity of a biosorbent for different kinds of adsorbates may be related to the characteristics of the binding sites (e.g. functional groups, structure, surface properties, etc.), the properties of the adsorbates (e.g. concentration, ionic size, ionic weight, molecular structure, ionic nature or standard reduction potential, etc.) and the solution chemistry (e.g. pH, ionic strength, etc.). It is difcult and complicated to nd a common rule from the physical and chemical properties of chromium(VI) and nickel(II) to identify how these properties affect the sorption capacity of biosorbent. This is because the observed behaviour may result from a combination of all the factors above [13,15,21]. At the second part of biosorption studies, this time, the uptake of nickel(II) in the presence of increasing concentrations of chromium(VI) was investigated at the initial pH values of 1.0 and 4.5. While the initial nickel(II) concentration was changed from 25 to 500 mg dm3, initial chromium(VI) concentration was held constant between 0 and 500 mg dm3 for each experiment set. Figs. 6 and 7 depicted the variations of nickel(II) uptakes at equilibrium with increasing the initial chromium(VI) concentrations at these pH values. Similar biosorption patterns are obtained both in the single-nickel(II) and nickel(II)chromium(VI) systems; nickel(II) equilibrium uptakes increased with increasing initial nickel(II) concentrations up to 500 mg dm3, but increase of chromium(VI) concentration also reduced the nickel(II) quantities at equilibrium. The selectivity of the biomass for nickel(II) increased while the concentration of chromium(VI) in the binary mixture was decreased. At 100 mg dm3 of initial nickel(II) concentration, in the absence of chromium(VI) ions and in the presence of 100 mg dm3 of chromium(VI) concentration at pH 1.0 and 4.5, adsorbed nickel(II) quantities at equilibrium were found as 36.0 and 60.2 mg g1 and as 29.1 and 53.0 mg g1, respectively.

3070
125

Z. Aksu et al. / Water Research 36 (2002) 30633073

100 q eq,Ni(II) (mg g -1 )

75

50

25

0 0 100 200 C eq,Ni(II) Co,Cr(VI) =0 mg dm-3 Co,Cr(VI) =25 mg dm-3 Co,Cr(VI) =50 mg dm-3 300 (mg dm -3 ) Co,Cr(VI) =100 mg dm-3 Co,Cr(VI) =250 mg dm-3 Co,Cr(VI )=500 mg dm-3 400 500

Fig. 6. The comparison of the non-linearized adsorption isotherms of nickel(II) adsorption to dried activated sludge with nickel(II) present as the single component and in the presence of increasing concentrations of chromium(VI) at pH 1.0 (T: 251C, X : 0.5 g dm3).

200

150 q eq,Ni(II) (mg g -1 )

100

50

0 0 75 150 C eq,Ni(II) Co,Cr(VI) =0 mg dm-3 Co,Cr(VI) =25 mg dm-3 Co,Cr(VI) =50 mg dm-3 225 300 375 450

(mg dm -3 ) Co,Cr(VI) =100 mg dm-3 Co,Cr(VI) =250 mg dm-3 Co,Cr(VI) =500 mg dm-3

Fig. 7. The comparison of the non-linearized adsorption isotherms of nickel(II) adsorption to dried activated sludge with nickel(II) present as the single component and in the presence of increasing concentrations of chromium(VI) at pH 4.5 (T: 251C, X : 0.5 g dm3).

The data obtained in the binary system indicate the manner in which the two metal ions affected each others biosorption equilibrium due to solution pH as compared to the results from single-metal adsorption situation. There was a weaker competition in adsorptive capacity of nickel(II) in the presence of chromium(VI) at pH 4.5,

the optimum value of nickel(II), whereas the uptake of chromium(VI) was much more reduced by the addition of nickel(II) at pH 1.0, the optimum value of chromium(VI). The adsorption data for one component approached the single-ion situation at lower concentrations of the other component while the equilibrium uptake of one component decreased with increasing concentration of the other ion; hence, the combined effects of two metal ions seemed to be antagonistic. The prediction of multi-component equilibrium data has always been complicated due to the interactive and competitive effects involved. The behaviour of each species in a multi-component system depends strongly on the physical and chemical properties of both sorbent and sorbate. This determines the sorbatesorbent chemical relation which affects the equilibrium behaviour hereafter. In addition, the number and kind of species present, concentration of each component, the pH of solution decide the shape and equilibrium constants of the isotherm (Ting, 1991) [4,14]. Nevertheless, attempts are carried out to predict and correlate multi-component data from single-component data. The simultaneous biosorption phenomena of chromium(VI) and nickel(II) on the dried activated sludge were expressed by the multi-component modied Langmuir and Freundlich adsorption models. The Langmuir and Freundlich correlation coefcients for each component evaluated from the modied adsorption models at pH 1.0 and 4.5 are given in Table 2. All the Langmuir correlation coefcients (Z) estimated close to 1.0 showed that non-modied multi-component Langmuir model related to the individual isotherm parameters only could also be used to predict the two-component data. Using the individual and modied multi-component Langmuir and Freundlich adsorption constants, qeq values were predicted from the related multi-component Langmuir and Freundlich adsorption formulas at both the pH values studied. According to the theoretical base of both multi-component Langmuir and Freundlich adsorption models, the adsorbed quantity of the rst component decreased with increasing the concentration of the second component, depending on the values of the individual Langmuir and Freundlich constants of both the two components. The comparisons of the experimental and calculated qeq values of chromium(VI) and nickel(II) in mixtures were presented in Figs. 8 and 9 for both the studied pH values. Basically, if most of the data are distributed around the 451 line this indicates that the models represent well the experimental data of the system so as shown in Figs. 8 and 9. In comparison with the multi-component Langmuir model, the multi-component Freundlich model tted reasonably well the binary uptake data of chromium(VI) and nickel(II) in the studied concentration range and for both of the studied pH values, although slight deviations were observed between the experimental and calculated

Z. Aksu et al. / Water Research 36 (2002) 30633073

3071

Table 2 Comparison of the multi-component Langmuir and Freundlich adsorption constants for each component evaluated from the competitive modied Langmuir and Freundlich adsorption models for simultaneous biosorption of chromium(VI) and nickel(II) to dried activated sludge at pH 1.0 and 4.5 pH 1.0 Component Z1 Z2 F 1.000 pH 4.5 Z1 0.995 F Z2 F 0.986 pH 4.5 yi zi 0.380 0.665 xi 0.300 0.503 yi 0.119 0.483 zi 0.536 0.383

The multi-component modied Langmuir model Chromium(VI) 0.925 Nickel(II) F pH 1.0 xi

The multi-component modied Freundlich model Chromium(VI) 0.513 1.000 Nickel(II) 0.683 0.359

250 Chromium(VI)-Modified Langmuir Model Nickel(II)-Modified Langmuir Model 200 Chromium(VI-Modified Freundlich Model) Nickel(II)-Modified Freundlich Model q eq,cal (mg g -1) 150

100

50

0 0 50 100 150 200 250 q eq,exp (mg g -1)

results from the model. These results can be attributed to the insensitivity of both models to competitive and interactive effects existing in multi-component systems and the characteristics of Langmuir model which is not valid for high concentrations assuming limited number of identical sites for sorption. The average percentage errors between the experimental and predicted values (e%) were calculated using the following equation, where the subscripts exp and calc indicate the experimental and calculated values and N the number of measurements: PN jqeq;i;exp qeq;i;calc =qeq;i;exp j e% i1 100: 7 N The average percentage errors between the experimental and predicted qeq values at pH 1.0 and 4.5 for the entire data set of chromium(VI) and nickel(II) were 17.67%, and 14.44%, respectively, for multi-component Langmuir model and 12.05% and 9.80%, respectively, for multi-component Freundlich model. It was concluded that the competitive, multi-component Freundlich model provided a more realistic description of the biosorption process. As reported in the literature, published models describing multi-component adsorption equilibria have restricted validity for the specic experimental conditions for which they were derived, e.g. species of microorganism, metalmetal combination and levels of metal concentration. Moreover, there are few experimental adsorption data published in the literature to test the accuracy of the proposed models.

Fig. 8. The comparison of the experimental and calculated individual qeq values of Ni(II) and Cr(VI) ions in binary mixtures at pH 1.0.

200 Chromium(VI)-Modified Langmuir Model Nickel(II)-Modified Langmuir Model 150 q eq,cal (mg g -1 ) Chromium(VI)-Modified Freundlich Model Nickel(II)-Modified Freundlich Model

100

50

0 0 50 100 q eq,exp (mg g -1 ) 150 200

4. Conclusion The presence of other metal species in the wastewater during single-metal biosorption can have signicant

Fig. 9. The comparison of the experimental and calculated individual qeq values of Ni(II) and Cr(VI) ions in binary mixtures at pH 4.5.

3072

Z. Aksu et al. / Water Research 36 (2002) 30633073 [2] Volesky B. Removal and recovery of heavy metals by biosorption. In: Volesky B, editor. Biosorption of heavy metals. Boca Raton, FL: CRC Press, 1990. p. 743. [3] Aksu Z, Kutsal T, Gun S, Hacosmano& lu N, Gholami. g nejad M. Investigation of biosorption of Cu(II), Ni(II) and Cr(VI) ions to activated sludge bacteria. Environ Technol 1991;12:91521. [4] Ting YP, Lavson F, Prince IG. Uptake of cadmium and zinc by the algae C. vulgaris: Part II: Multi-ion species. Biotech Bioeng 1991;37:44555. [5] Mattuschka B, Straube G. Biosorption of metals by a waste biomass. J Chem Technol Biotechnol 1993;58: 5765. [6] Wilde EW, Benemann JR. Bioremoval of heavy metals by the use of microalgae. Biotech Adv 1993;11: 781812. [7] Brady D, Stoll AD, Starke L, Duncan JR. Bioaccumulation of metal cations by Saccharomyces cerevisiae. Appl Microbiol Biotechnol 1994;41:14954. [8] Corder SL, Reeves M. Biosorption of nickel in complex aqueous waste streams by cyanobacteria. Appl Biochem Biotechnol 1994;45/46:84759. [9] Volesky B, May-Philips HA. Biosorption of heavy metals by Saccharomyces cerevisiae. Appl Microbiol Biotechnol 1995;42:797806. . [10] Aksu Z, Ackel U, Kutsal T. Application of multicomponent adsorption isotherms to simultaneous biosorption of iron(III) and chromium(VI) on C. vulgaris. J Chem Technol Biotechnol 1997;70:36878. [11] Veglio F, Beolchini F. Removal of metals by biosorption: a review. Hydrometallurgy 1997;44:30116. [12] Veglio F, Beolchini F, Gasbarro A. Biosorption of toxic heavy metals: an equilibrium study using free cells of Arthrobacter sp. Process Biochem 1997;32: 99105. [13] Aksu Z. Biosorption of heavy metals by microalgae in batch and continuous systems. In: Wong Y-S, Tam NFY, editors. Algae for waste water treatment. Germany: Springer, 1998. p. 3753. [14] Williams CJ, Aderhold D, Edyvean RGJ. Comparison between biosorbents for the removal of metal ions from aqueous solutions. Water Res 1998;32:21624. . [15] Aksu Z, Ackel U, Kutsal T. Investigation of simultaneous biosorption of copper(II) and chromium(VI) on C. vulgaris from binary metal mixtures: application of multicomponent adsorption isotherms. Sep Sci Technol 1999;34: 50124. . . . [16] Donmez G, Aksu Z, Ozturk A, Kutsal T. A comparative study on heavy metal biosorption characteristics of some algae. Process Biochem 1999;34:88592. [17] Mamery N, Boudries N, Addour L, Belhocine D, Lounici H, Grib H, Pauss A. Batch zinc biosorption by a bacterial nonliving Streptomyces rimosus biomass. Water Res 1999;33:134754. . [18] Aksu Z, Ackel U. Modelling of a single-staged bioseparation process for simultaneous removal of iron(III) and chromium(VI) by using Chlorella vulgaris. Biochem Eng J 2000;4:22938. [19] Mckay G, Al Duri B. Prediction of multicomponent adsorption equilibrium data using empirical correlations. Chem Eng J 1989;41:923.

consequences. Since real wastewaters will contain all kinds of pollutants, adsorption systems design must be based on multi-component efuents, making multicomponent equilibrium data a necessity. The biosorption of chromium(VI), nickel(II) and chromium(VI) nickel(II) binary mixtures on the dried activated sludge was investigated in this study and the mono- and multicomponent Langmuir and Freundlich isotherm models were used to predict the equilibrium uptake of components, both singly and in combination. The obtained results showed that the dried activated sludge biomass selectively adsorbed nickel(II) ions at pH 4.5 while chromium(VI) ions were preferentially adsorbed by the biomass at pH 1.0. Although dried activated sludge had a higher adsorption capacity for chromium(VI) and nickel(II) at single-component situation due to the initial pH of solution, the equilibrium uptake of chromium(VI) and nickel(II) in the binary mixture were found to be decreasing due to the levels of chromium(VI) and nickel(II) concentrations because of the antagonistic interaction between the components. The applicability of mono-component Langmuir and Freundlich models at both the studied pH values indicated that the individual biosorption of chromium(VI) and nickel(II) ions is favourable and could be characterized as a monolayer, single-site-type phenomenon with no interaction between sorbed components and the microbial surface. The individual Langmuir and Freundlich constants evaluated from the mono-component isotherms were used to compare the biosorptive capacity of the dried activated sludge for both components and to describe the multi-component adsorption equilibrium. It was concluded that multicomponent Freundlich model agreed well with the results found experimentally in the studied initial mixture concentration range at both the studied pH values. This work could enable to extrapolate the prediction of adsorption equilibria of the binary system if experimental data are not available for a certain level of bisolute concentrations.

Acknowledgements . We thank TUBITAK (the Scientic and Technical Research Council of Turkey), Project No. YDABCAG198Y097, for the partial nancial support.

References
[1] Aksu Z, Kutsal T. A comparative study for biosorption characteristics of heavy metal ions with C. vulgaris. Environ Technol 1990;11:97987.

Z. Aksu et al. / Water Research 36 (2002) 30633073 [20] Bellot JC, Condoret JS. Modelling of liquid chromatography equilibria. Process Biochem 1993;28:36576. [21] Al-Asheh S, Banat F, Al-Omari R, Duvnjak Z. Prediction of binary sorption isotherms for the sorption of heavy metals by pine bark using single isotherm data. Chemosphere 2000;41:65965. [22] Khan AR, Al-Wheab IR, Al-Haddad A. A generalized equation for adsorption isotherms for multi-component organic pollutants in dilute aqueous solution. Environ Technol 1996;17:1323. [23] Patterson JW. Wastewater treatment technology. USA: Ann Arbor Science Publishers Inc., 1977. [24] Ajmal M, Rao RAK, Ahmad R, Ahmad J. Adsorption studies on Citrus reticulata (fruit peel of orange): removal

3073

[25]

[26]

[27]

[28]

and recovery of Ni(II) from electroplating wastewater. J Hazard Mater B 2000;79:11731. Snell FD, Snell CT. Colorimetric methods of analysis, vol. 2, 3rd ed. Canada: D. Van Nostrand Company, 1959. Ricou P, Lecuyer I, Cloirec P. Inuence of pH on removal of heavy metallic cations by y ash in aqueous solution. Environ Technol 1998;19:100516. Cimino G, Passerini A, Toscano G. Removal of toxic cation and Cr(VI) from aqueous solution by hazelnut shell. Water Res 2000;34:295562. Par P, Kazy SK, Asthana RK, Singh SP. Metal adsorption and desorption by lyophilized Pseudomonas aeruginosa. Int Biodeterior Biodegrad 1999;44:10110.