No job name

Environ. Sci. Technol. 2003, 37, 3601-3608
Modeling of Lithium Interference in
describe the isotherm data. However, these models arecompletely insensitive to the environment in which bio- sorption takes place. Furthermore, the biosorption perfor-mance, depending on the pH, ionic impurities, competingions, and other environmental variables, cannot be reliably predicted using conventional empirical models. The binding of heavy metals to algae (10) and seaweed biomass (11-13) Division of Environmental and Chemical Engineering, and has recently been confirmed to be mainly due to an ion Research Institute of Industrial Technology, Chonbuk exchange process. Although the molecular structure of National University, Chonbuk 561-756, Korea, and seaweed biomass is extremely complex when compared with Department of Chemical Engineering, McGill University, that of synthetic resins, the key material for metal binding 3610 University Street, Montreal, Quebec H3A 2B2, Canada is known to be alginate in the cell wall (14). Therefore, theion exchange reactions on the binding sites can provide thebasis for modeling of biosorption.
In this study, the protonated biomass of brown seaweed Biosorption of Cd by the brown seaweed Sargassum Sargassum polycystum was used for biosorption of cadmium polycystum biomass was experimentally investigated and as a model heavy metal. The toxicity of Cd is extremely high mathematically modeled at different pH and ionic strength (15), and it has been classified in the USEPA's Group B of levels. From the potentiometric titration of the biomass, pollutants (probable human carcinogen) (16). A well-known three types of functional groups were identified, and the problem of the skeletal and renal systems associated with dissociation constant and the numbers of these groups were cadmium exposure is the major symptoms of "Itai-Itai determined. The carboxyl group (pK disease", first diagnosed in Japanese individuals who con- H 3.70 ( 0.09) was found to play a major role in binding protons and Cd. The sumed rice and water contaminated with high levels of Cd background ion, Li+, could interfere with the uptake of (17). A special interest has focused on biosorption of Cd (12-13, 18-19). In addition to health and environmental concerns, protons and Cd by competition for the carboxyl sites. Whereas a simple water chemistry of Cd was another reason to choose the binding mechanism on the carboxyl group was it as a model metal in this study. Up to pH 6.8 Cd does not established as an ion exchange process, the second precipitate and is present in solution as a single form (Cd2+) functional group, phosphonate (pKH 5.41 ( 0.31), most likely without any hydrolysis reactions (20). The modeling study bound metals by a complexation reaction. A biosorption can thus focus on the binding of Cd2+ without considering model was developed based upon the binding mechanisms the effect of hydrolysis reactions in the aquatic environment and was successfully used for predicting the isotherm (21-22). In this study, the pH-sensitive Cd binding to the and pH edge experiments. In addition, the speciation of functional groups in the presence of an alkali ion, Li, and the binding sites as a function of the pH was simulated using with ionic strength as an additional parameter, is math- the developed model in order to visualize the distribution ematically modeled reflecting the underlying biosorptionmechanisms.
of Cd on the binding sites.
Materials and Methods
Materials. The brown seaweed Sargassum polycystum was
obtained in Cebu, Philippines, courtesy of MCP, Inc. The
Biosorption of heavy metals has been considered as an sun-dried biomass was treated with a 0.5 N HNO alternative cost-effective means for the treatment of metal- for 24 h, replacing the natural mix of ionic species with bearing wastewaters (1). The low cost of biosorbents is a protons. The acid-treated biomass, designated as protonated tangible advantage of biosorption over other technologies, biomass in this article, was washed with deionized distilled such as ion exchange and reverse osmosis (2). Therefore, water several times and thereafter dried at 60 °C in an oven new types of biosorbent materials able to effectively sequester for 24 h. The resulting dried Sargassum biomass was used heavy metals have been explored (2-4). Among available as a biosorbent in the experiments.
biosorbent materials (algae, bacteria, fungi, yeast, etc.), the All chemicals used in this study were of analytical grade.
size of seaweed biomass is large enough to facilitate its The Cd stock solution was prepared using Cd(NO application without biomass immobilization and a cumber- some solid-liquid separation process (5, 6). Furthermore, 3 was used for changing the ionic strength. To adjust the pH, concentrated LiOH‚H various kinds of seaweed proliferate ubiquitously and 2O and HNO3 solutions were used. Although Na+ is generally present in the aquatic abundantly in the littoral zones of the world's oceans, making systems, it may cause the alginate leaching from the biomass the seaweed biomass readily available and inexpensive.
matrix at elevated concentrations (23). To avoid an additional To make better and more effective use of biosorption and complexity (i.e., the biomass leaching effect), Li+, which does its potential, biomass functional metal-binding chemical not induce such an effect, was used as a background and groups should be identified, and the binding mechanism competing light ion.
needs to be understood. Additionally, modeling of the Potentiometric Titration. The potentiometric titration
biosorption process is essential for the prediction of the was carried out with 5 g/L of biomass concentration at two process performance and for its design and optimization.
different ionic strengths (0.1 and 1 M). Plastic bottles (high- In general, empirical models such as those proposed by density polyethylene) of 40-mL capacity were used for the Langmuir (7), Freundlich (8), and Sips (9) have been used to titration experiments and for isotherm and pH edge experi-ments. First, the biomass and 20 mL of water (CO2-free) of * Corresponding author phone: +1-514-398-4276; fax: +1-514- the desired ionic strength were put into each bottle. CO2- 398-6678; e-mail: [email protected].
† Chonbuk National University.
free water was obtained by stripping water with nitrogen gas ‡ McGill University.
for 2 h with vigorous mixing. Different volumes of 1 N LiOH 10.1021/es011454e CCC: $25.00  2003 American Chemical Society VOL. 37, NO. 16, 2003 / ENVIRONMENTAL SCIENCE & TECHNOLOGY 9 3601
Published on Web 07/22/2003 or 1 N HNO3 were added into each bottle containing the Measurements of Cd Uptake. The dissolved Cd concen-
biomass suspension and the bottles were capped. The capped tration of samples was analyzed using an inductively coupled bottles were then agitated using a shaker (200 rpm) at room plasma atomic emission spectrophotometer (ICPAES, Trace temperature for 24 h. Preliminary tests showed that 24 h was Scan, Thermo Jarrel Ash). So that the change of working a sufficient time for achieving the proton sorption equilib- volume (up to 5%) by added Cd stock and LiOH solutions rium. Thereafter, the equilibrium pH was measured using might be considered, the metal uptake (qCd) in both isotherm an electrode (Ingold) as bubbling nitrogen gas was used to and pH edge experiment was calculated from the mass keep the CO2-free condition. The control titration experi- balance as follows: ments with water solution of the same ionic strength werecarried out without the biomass in order to compare them [Cd] - V [Cd] with titration data when the biomass was present. During the titration experiments, the CO2-free condition was alwaysmaintained to avoid the influence of inorganic carbon onthe solution pH.
where [Cd]o and [Cd]f are the Cd concentration of the From potentiometric titration data, the proton release Cd(NO3) stock solution and the final Cd concentration after the sorption experiment, respectively. Here, subscript o qH,f) was calculated according to eq 1. A detailed derivation procedure of eq 1 is given in the Supporting represents the original state, in which the reactants (e.g., Information with raw data of potentiometric titration results.
metal stock solution, electrolyte solution, and biomass) arenot mixed but separated. VCd,o and Vf are the volume of Cd stock solution and the final volume, respectively. M stands for the weight of biomass used. Here, Vf is the sum of the initial working volume (20 mL), the Cd stock solution volume(V Cd,o), and the amount of LiOH solution added for adjusting B and CA are the concentrations of added base and acid, respectively, and X is the biomass concentration.
Subscripts B and W represent the systems containing the Results and Discussion
biomass and only water without the biomass, respectively.
Isotherm Experiments. The isotherm represents the
Model Building and Parameter Estimation. Proton Bio-
equilibrium relationship between the metal uptake by the sorption. To describe the potentiometric titration data, the sorbent and the final metal concentration in the aqueous proton uptake (qH) should be modeled as a function of the phase, showing the sorption capacity of the biosorbent. In sorption system pH and of the background ions present (Li+ this study, the Cd biosorption isotherms were obtained at in this study). It can be considered that the biomass contains different solution pHs and ionic strengths. The isotherm some negative or neutral groups capable of binding the experiments were carried out with 0.1 g of the biomass in 40-mL plastic bottles with 20 mL of working solution volume.
Supposing a certain negatively charged group (iB-), its For each bottle, the initial Cd concentration was differently reaction with a proton, and its related equilibrium constant adjusted, ranging between 0 and 19 mM, which resulted in (iKH) are defined as follows: different final Cd concentrations and Cd uptakes after thesorption equilibrium was achieved. Following the addition [iB ]γ{H of biomass into Cd-containing solutions, the suspensions BH ) iB + H were agitated on a rotary shaker (200 rpm, room temperature) and the solution pH was maintained at the desired valueusing 1 N LiOH, as the pH had a decreasing tendency with where [ ] and { } indicate the concentration and activity, Cd binding. The isotherm experiments were carried out for respectively. The activity coefficient (γ) of the monovalent 24-36 h, which was enough contact time for achieving ion (here iB-) is expressed by Davies' equation (25): equilibrium. Thereafter, the supernatant samples were takenfrom the bottles and diluted properly with deionized waterfor analysis of the Cd concentration. A detailed standard 10γ ) - AZ2( xI procedure for determination of the sorption isotherm has been reported elsewhere (1).
pH Edge Experiments. In addition to the sorption
where the constants A and b used have values of 0.5 and 0.24, isotherm, the pH edge experiments were carried out: an respectively, as chosen from the software MINEQL+ (20). Z equilibrium relationship between the metal uptake and the is the ion valence, and I stands for the ionic strength defined final pH, which is helpful to understand the pH dependence as I ) 0.5 ∑ z 2 i Ci, where Ci is the concentration of any ionic of metal biosorption (21). Whereas the isotherm experiments species in the system.
were performed with different initial metal concentrations, Furthermore, the negatively charged group is assumed to the pH edge experiments were carried out with different base be associated with the other cation Li+ as seen in the earlier (or acid) additions. In the pH edge experiments, the initial studies (10, 26). This assumption is based upon the reports Cd concentration and ionic strength were 2 mM Cd and 0.1 that even light ions can be bound to the biomass, however very weakly, through the ion exchange mechanism (10, 27).
3, respectively. The pH was intentionally altered by means of adding 1 N LiOH or 1 N HNO The association of Li+ with a negatively charged group is 3 into the bottles. The biomass concentration was the same (3 g/L) in all bottles for the pH edge experiment. Other conditions for the pH edgeexperiments were identical with those in the isotherm [iBLi]{H } experiments. After the system reached the equilibrium state iBH + Li ) iBLi + H (24-36 h), the final pH was measured and the liquid samples were taken for analysis of the Cd concentration after properdilution.
Also, a certain positively charged functional group (jBH + Because the leaching of constituents of seaweed biomass in the neutral condition may be present in the biomass (for is significant at a high pH (23, 24), the sorption experiments 3 ), and its reaction with the protons can be (isotherm and pH edge) were carried out only below pH 6.
3602 9 ENVIRONMENTAL SCIENCE & TECHNOLOGY / VOL. 37, NO. 16, 2003
TABLE 1. Dissociation Constants and the Number of Three
Types of Functional Groups in the Sargassum polycystum
Biomassa
first group
second group
third group
a The coefficient of determination was 0.992. Standard errors of the estimated parameters are given in the parentheses. b The first functionalgroup indicates the carboxyl site; the second group indicates possiblythe phosphonate site; and the third group is the amine site. c The pKHvalues represent the dissociation constants of the functional groups.
d The b values are the numbers of the functional groups. e The KLi valueis the Li-binding constant to the first functional group.
FIGURE 1. Uptakes of proton and lithium depending on the
equilibrium pH and ionic strength. The biomass concentration was
5 g/L and the initial ionic strengths were 0.1 M (O: first trial, 4:
second trial) and 1 M (2). The lines were produced using the
developed model. The proton uptake: s 0.1 M and - - - - 1 M initial
ionic strength. The lithium uptake: -‚‚- 0.1 M and -‚- 1 M initial
ionic strength.
However, the positively charged group cannot be associatedwith the Li+ cation.
The total uptake of protons by the biomass is the sum of protons taken up by all kinds of negative and positive groups.
b{H }γ FIGURE 2. Prediction of the cadmium uptake depending on the pH
and initial ionic strength. Experimental data: (O) pH 3 and 0.1 M
i)1 K γ + iKLi initial ionic strength, (2) pH 5 and 0.1 M initial ionic strength, (4)
pH 5 and 0.01 M initial ionic strength. Performance of the two
The detailed procedure of model derivation is provided in models is compared: - - - - ion exchange/ion exchange model, s
the Supporting Information. The total uptake of Li by the ion exchange/complexation model.
biomass (only by negatively charged groups) can be obtainedin the same manner: residues) (31). Therefore, the first functional group wasbelieved to be the carboxyl group. The pKH value and the biK [Li ]γ2 number of the second binding sites (negatively charged) were estimated to be 5.41 ( 0.31 and 0.45 ( 0.07 mmol/g, i)1 K + {H } γ + iK [Li ]γ2 respectively. The second sites were likely to be phosphonate groups. The phosphonate groups of phospholipids present Because the proton uptake (eq 7) is modeled as a function in the plasma membrane of brown seaweed (32) have a similar range of the pK 10{H+}) and the concentration of background H value (30). The last group (positively charged) cations Li+, eq 1 can be simultaneously fitted to all the titration was considered to be the amine group with pKH values in curves (Figure 1) obtained at two ionic strengths (or Li various biomaterials ranging between 8 and 10 (30, 33).
concentrations). To estimate the model parameters, the Because the third group has a positive charge, the Li nonlinear regression was performed by means of the Mar- binding to this group was not expected. Furthermore, quardt-Levenberg algorithm (28) using the technical soft- although the phosphonate group had a potential to bind Li, ware Sigma Plot (Version 5.0, SPSS).
the binding constant (2KLi) was estimated to be nearly zero As shown in Figure 1, the three-site model (two types of (8.67 × 10-13) with a relatively large standard error (2.83 × negative groups and one type of a positive group) was able 10-5), indicating that the Li binding to the phosphonate group to describe the entire proton uptake data whereas two- or would be negligible indeed.
one-site functional group models could not describe the Cd Biosorption. The Cd uptake increased with increasing titration data, especially at a high pH (data not shown). The pH or decreasing ionic strength (Figure 2). At pH 5, the estimated parameters are summarized in Table 1.
maximum uptake was estimated to be 1.44 ( 0.16 mmol/g The first group was found to be most abundant (2.57 ( by fitting the Langmuir equation to the data. The metal 0.06 mmol/g) and its pKH value was estimated at 3.70 ( 0.09.
binding to carboxyl groups of alginate in seaweed biomass Carboxyl groups in biological polymers have pKH values has been well-known to be an ion exchange process (10- ranging 3.5 to 5.0 (30). Furthermore, the pKH values of carboxyl 13). Therefore, the maximum uptake of divalent Cd2+ groups of alginate, the material mainly responsible for metal corresponds to 2.88 mequiv/g of binding sites, which is sequestering by seaweed (14), are known to range between more than the total number of the carboxyl sites (2.57 ( 0.06 3.38 (for mannuronic residues) and 3.65 (for guluronic mmol/g). Considering that a part of the carboxyl sites is VOL. 37, NO. 16, 2003 / ENVIRONMENTAL SCIENCE & TECHNOLOGY 9 3603
TABLE 2. Cadmium Binding Constants of Biosorption Models Based on Two Types of Binding Mechanisms
first group
second group
a The 1KCd and 2KCd values are the Cd binding constants to the first and second groups, respectively. b The values R2 are the coefficients of occupied with Li, the available sites should be far less than Finally, the Cd uptake by complexation with the phos- the total number of carboxyl sites. Therefore, it can be noted phonate group can be expressed as follows: that the phosphonate group takes part in binding Cd.
However, it should be verified whether the binding mech- anism to this group is also based on ion exchange.
The Cd-H ion exchange on the carboxyl group (1BH) and γ + 2KCd the corresponding equilibrium constant can be defined as Although the Cd uptake is separately derived according to the types of binding sites, only the total uptake (qCd) is 21BH + Cd2+ ) 21BCd measured and compared with experimental data.
) 1q + 2q where γ is the activity coefficient of monovalent ions; γ4 In Figure 2, two kinds of models were shown and compared corresponds to that of divalent Cd2+ according to Davies with experimental isotherm data. Estimated parameters are equation (eq 4). Here, the formulation of metal-bound form given in Table 2. When both functional groups were assumed can be chosen as 2BCd0.5 instead of B2Cd, to emphasize that to bind Cd via the ion exchange mechanism, the model output two bonds have to be broken in competitive binding or upon had a bias: overestimation at a lower pH and underestimation desorption of metal (33).
at a higher pH. The model based on complexation with the Therefore, the Cd uptake by the carboxyl group via ion second group was successfully able to describe the isotherm exchange mechanism can be derived as follows: data and the determination coefficient was R2 ) 0.945. Thisimplies that complexation is more likely than ion exchange for Cd binding on the phosphonate group. Therefore, the second model was used for the following model-based The proton uptakes can be also derived on the basis of the ion exchange/complexation mechanism γ + 1KLi γ2 + x1KCd q ) 1q + 2q + 3q The detailed model derivation procedure is available in the where the proton uptakes by three types of functional groups The Cd binding to the phosphonate group should also be considered, as this binding group is apparently functional asdiscussed in the previous section. If the mechanism for Cd b{H } γ binding to the phosphonate group is also based on ion exchange, the Cd binding to this group can simply be derived γ + 1KLi γ2 + x1KCd according to the same procedure as applied for the carboxylgroup. However, as shown previously, Li is not bound to this b{H } γ group. Therefore, based upon the ion exchange mechanism, the Cd uptake by the second group can be expressed as γ + 2KCd γ + x2KCd 1 + 3KHγ/{H Another possible mechanism for Cd binding to the Also, the Li uptake by only the carboxyl group can be obtained phosphonate group is likely to be complexation, as postulated in the previous study (34). In the present study, the following complexation reaction is presumed: b1K [Li ] q ) 1q ) γ + 1KLi γ2 + x1KCd [2BHCdOH ]γ{H Discussion on Assumptions Made during Model Derivation. In this study, the biosorption system comprising biomass As can be seen in eq 12, this reaction may occur easily without and solution was treated as one solution phase rather than deprotonation of the functional group, especially at pH < two phases (solid biomass plus solution). In other words, the biomass functional groups, as well as ions, were assumed to 3604 9 ENVIRONMENTAL SCIENCE & TECHNOLOGY / VOL. 37, NO. 16, 2003
be homogeneously distributed in the aqueous phase. Thisassumption was made from the gellike nature of seaweedbiomass, where all ions were known to penetrate freely inand out. With this postulation, therefore, concentrations ofall species related to biomass functional groups (B-, BH, BLi,BH + 2 ) were expressed as mole per working volume. Fur- thermore, to keep consistency with one phase approximation,the activity coefficients of these species were not taken as 1but estimated using Davies equation. This was not likelycommon in studies on interactions between solution andimpermeable oxide surfaces.
In general, the Donnan model is widely used for the analysis of gel systems. However, since the Donnan typebiosorption models (35, 36) are more complicated than themodels presented in this study, it is likely difficult to applyor extend the Donnan model to complex multi-ion systemssuch as actual wastewaters.
In this study, Li+ was used instead of Na+ as a model alkali ion interfering heavy metal biosorption, because Na+results in the leaching of organic components from theseaweed biomass (23). Therefore, the biosorption perfor-mance in actual aquatic systems containing Na+ may beaffected by the leaching property. To predict the leachingeffect, the present models should be further modified.
Nitrate was used as a model anion in this study. This ion can form complexes with Cd2+ such as CdNO + previously (21), these metal complexes are possibly able toaffect the biosorption performance. This aspect was neglectedin the present work and more studies will be needed tounderstand (or model) the effect of Cd-NO3 complexes.
Li Interaction with Proton Binding. The proton uptake
depending on the pH and Li concentration is shown in Figure1 together with the model output. The higher the ionicstrength (or Li concentration), the more protons were FIGURE 3. Prediction of equilibrium state from the initial cadmium
concentration. Conditions: 5 g/L of biomass concentration, 0.1 M
required for the maximum uptake, mainly due to the initial ionic strength, and pH 3. Symbols: (
interference of Li with the proton binding. The Li uptake b) final cadmium
predicted by eq 8 increased with increasing pH. However, 4) cadmium uptake, (O) initial lithium concentration,
at an elevated pH (>4.5), the Li uptake was independent of 2) concentration of added base (LiOH). The lines were predicted
using the developed biosorption model.
the pH, because the Li binding was considered to occur onlyon the carboxyl group.
C ) ([H ] - [H ] ) - ([OH ] - [OH ] ) + X(q The equilibrium constant (1K ) 2.64 ( 0.46 × 10-3) for the Li binding to the carboxyl group corresponds to theselectivity coefficient of Li-H ion exchange as shown in eq [Li ] ) [Li ] - Xq 6. Therefore, the affinity of Li for the biomass can be considered as low, approximately 0.26% of that displayed by protons. Nevertheless, the reason the Li uptake was con- siderable at pH > 4.5 ([H+] < 3.2 × 10-5 M) was because theLi concentration was fairly high ([Li+] ) 0.1 or 1 M).
where the proton balance was established on the basis of the From Figure 1 it was noted that when a high concentration original state that was composed of the (dried) protonated of Li (probably also of other light alkali ions) was present, the biomass (q 3.66 mmol/g), the Cd stock solution, and the Li competition with protons for the carboxyl group could solution containing LiNO3.
not be neglected, although the Li affinity was far lower than The previously developed Cd binding model (eq 14) could that of the protons. Especially at an elevated pH (up to pH be used for the Cd uptake term (qCd) in eq 22. The uptakes 6), Li was considerably competitive because of the lack of of proton and lithium (eqs 15 and 19) were substituted into protons. In addition to competition for the binding sites, the eqs 20 and 21, respectively. To obtain four unknowns (CB, background ions (Li+ and NO - in this study) can disturb the [Li+]f, [Cd2+]f, γ), four equations (eqs 20-22 and Davies proton uptake via reducing the activities of binding sites and equation (eq 4)) should be simultaneously solved. In this protons as well.
study, a damped Newton's method (37) was applied using Prediction of Cd Isotherms Depending on the pH and
the technical software Mathematica 3.0 (38).
Ionic Strength. To obtain the isotherm data, the initial Cd
Figure 3 shows a biosorption performance prediction concentration ([Cd] oVCd,o/Vf) is changed at a given pH and example (pH 3, Ii 0.1 M). The final Cd concentration and the initial ionic strength. While the only operable variable in the Cd uptake increased with the increasing initial Cd isotherm experiments is the initial Cd concentration, the concentration (Figure 3A). The plot of the resulting final Cd final Cd concentration and the Cd uptake are dependent on concentration and uptake became the isotherm curve at pH the initial Cd concentration. In this study, the isotherms were 3 and 0.1 M initial ionic strength as shown in Figure 2. The "predicted" from known initial conditions (initial Cd con- Li concentration was also predicted as a function of the initial centration, initial ionic strength (or Li concentration)) and Cd concentration and compared with the initial Li concen- the fixed pH. For the batch equilibrium experiments, the tration (Figure 3B). Within a low range of initial Cd following mass balances were valid for protons (eq 20), Li concentrations, a small difference between the initial and (eq 21), and Cd (eq 22): final Li concentrations was observed that was due to a higher VOL. 37, NO. 16, 2003 / ENVIRONMENTAL SCIENCE & TECHNOLOGY 9 3605
FIGURE 4. Effect of pH on the cadmium biosorption. Experimental
FIGURE 5. Effect of initial ionic strength on the cadmium biosorption.
data: (O) pH 3, (2) pH 5. The biomass concentration and the initial
Experimental data: (O) 0.1 M, (b) 0.01 M, and (4) 0 M initial ionic
ionic strength were 5 g/L and 0.1 M, respectively. The lines were
strength. The biomass concentration and the solution pH were 5
predicted using the developed biosorption model.
g/L and 5.0, respectively. The lines were predicted using the
developed biosorption model.
Li uptake in this range. Although the proton concentration(∼10-3 M) was much lower than the Li concentration (∼10-1M), the proton uptake was larger than the Li uptake, becausethe Li affinity toward the biomass was relatively low. Thehigher the initial Cd concentration, the more LiOH wasconsumed to maintain the constant pH and the more Cdions were present at the equilibrium state. Therefore, theionic strength increased along with the initial Cd concentra-tion as predicted in Figure 3C. However, the activitycoefficient remained nearly constant (∼0.78).
From the practical point of view, it is of great importance to estimate the base (or acid) requirement in a pH-stat processoperation and, in turn, to predict the final pH according tothe known amount of base (or acid) addition. As can be seenin Figure 3C, the base consumption to maintain a constantpH was very well predicted, considering that, unlike the CO2-free titration experiments, these isotherm experiments wereperformed in an open-air system.
The above method was used to simulate the effect of pH on the Cd uptake (Figure 4). The pH significantly affectedthe Cd uptake between pH 2 and 4, where the p1KH value ofthe carboxyl group is dominant. However, at pH < 2 andpH > 5 (cf. p2K ) 5.4), the Cd uptake was almost independent of the pH. This reflects the fact that most of the binding sitesare occupied by protons at pH < 2, and by Cd, and also byLi in part, at pH > 5.
The effect of initial ionic strength is shown in Figure 5.
At an extremely high ionic strength (I ) 1 M), the Cd uptake was reduced by approximately 50% of the maximum value(I ) 0 M). For this case, it should be noted that even if no LiNO3 was used, the ionic strength for each data point wasnot zero but varied depending on the concentrations of Cd FIGURE 6. Speciation of the binding sites as a function of the
and added LiOH. For example, at the final Cd concentration solution pH. The lines were produced using the developed
of 10 mM and I ) biosorption model. Experimental conditions: 3 g/L of biomass
0 M, the ionic strength was as high as 53 mM, which could suppress the Cd uptake. Therefore, data concentration, 0.1 M initial ionic strength, and 2 mM Cd initial
plots of metal uptake versus final metal concentration (generally called an "isotherm") would not present equilib-rium relationships at a constant ionic strength medium concentration of added base (CB), whereas in the isotherm although such a plot may be still useful for evaluating the prediction it was the initial Cd concentration.
As a result, the pH edge data were successfully predicted Speciation of Binding Sites. The pH edge experiment
as shown in Figure 6A. At the same time, the speciation of was carried out at the initial Cd concentration of 2 mM, 0.1 binding sites as a function of the equilibrium pH could be M of LiNO3, and 3 g/L of biomass (Figure 6). With these simulated using the developed model. The Cd binding to the initial conditions, the same set of model equations was carboxyl group increased with the pH up to 4.1, but slightly simultaneously solved for the four unknowns (final pH, [Li+]f, decreased over the optimum pH, because of competitive [Cd2+]f, γ). In this case, the only operable variable was the binding of Li. The contribution of the Cd binding to the 3606 9 ENVIRONMENTAL SCIENCE & TECHNOLOGY / VOL. 37, NO. 16, 2003
phosphonate functional group became significant at a higher dry weight of the biomass (g) pH (>4). At a low pH, most of the binding sites were occupied with protons (Figure 6B), and the Cd binding could not be total uptake of Cd by the biomass (mol/g) expected. With increasing pH, the free sites, especially the uptake of Cd by the carboxyl group (mol/g) carboxyl group, increased up to ∼39% of the total carboxyl sites. This was because most of the Cd (total 2 mM) existed uptake of Cd by the phosphonate group (mol/ in the biomass phase, and only a low level of Cd was presentin the aquatic phase (e.g., [Cd2+] ) 0.3 mM at pH ) 5). The total uptake of proton by the biomass (mol/g) free form of the phosphonate group was rather rare even at a high pH. As expected, the third functional group was present uptake of proton in the final state (mol/g) as a positively charged form in the pH range studied.
uptake of proton in the initial state (mol/g) proton uptake by the protonated biomass (mol/ This work was supported in part by the postdoctoralfellowships program of the Korea Science & Engineering uptake of proton by the carboxyl group (mol/ Foundation (KOSEF) and by research funds of Chonbuk National University. We thank Prof. J. M. Park (POSTECH), Prof. M. E. Carballo (University of Havana), and Dr. H. Niu uptake of proton by the phosphonate group and Mr. T. A. Davies (McGill University) for their help with experiments and manuscript preparation.
uptake of proton by the amine group (mol/g) Supporting Information Available
total uptake of Li by the biomass (mol/g) Handling method of potentiometric titration data and uptake of Li by the ith negatively charged derivation procedures of models for proton and cadmium functional group (mol/g) biosorption. This material is available free of charge via the Internet at http://pubs.acs.org.
uptake of Li by the carboxyl group (mol/g) total proton concentration in the system (mol/ representative of concentration for the brack- total proton concentration in the final state eted species (mol/L) representative of activity for the bracketed total proton concentration in the initial state ith negatively charged functional group solution volume (L) jth positively charged functional group concentration of the biomass (g/L) weight-specific number of ith negatively valence of ion (-) charged functional group (mol/g) activity coefficient of monovalent ion (-) weight-specific number of jth positively charged functional group (mol/g) representative of the functional group in the (1) Volesky, B., Ed. Biosorption of Heavy Metals; CRC Press: Boca Raton, FL, 1990; pp 7-43.
concentrations of added base and acid, re- (2) Bailey, S. E.; Olin, T. J.; Bricka, R. M.; Adrian, D. D. Water Res. spectively (mol/L) 1999, 33, 2469-2479.
(3) Volesky, B; Holan, Z. R. Biotechnol. Prog. 1995, 11, 235-250.
proton concentration in the final state where (4) McKay, G.; Ho, Y. S.; Hg, J. C. Y. Sep. Purif. Methods 1999, 28,
the system equilibrates (mol/L) (5) Davis, T. A.; Volesky, B.; Vieira, R. H. S. F. Water Res. 2000, 34,
proton concentration in the initial state (i.e., in the stock solution) where the biomass and (6) Zhao, Y.; Hao, Y.; Ramelow, G. J. Environ. Monit. Assess. 1994,
the LiOH solution are mixed together and 33, 61-70.
equilibrate (mol/L) (7) Langmuir, I. J. Am. Chem. Soc. 1918, 40, 1361-1403.
(8) Freundlich, H. Z. Phys. Chem. 1907, 57, 385-470.
proton concentration in the original state where (9) Sips, R. J. Chem. Phys. 1948, 16, 490-495.
the reactants are not mixed but separated (10) Crist, R. H.; Martin, J. R.; Carr D.; Watson, J. R.; Clarke, J.; Crist, D. R. Environ. Sci. Technol. 1994, 28, 1859-1866.
(11) Kuyucak, N.; Volesky, B. Biotechnol. Bioeng. 1989, 33, 823-831.
ionic strength (mol/L) (12) Da Costa, A. C. A.; De Franca, F. P. Sep. Sci. Technol. 1996, 31,
initial ionic strength (mol/L) (13) Romero-Gonzalez, M. E.; Williams, C. J.; Gardiner, P. H. E.
Environ. Sci. Technol. 2001, 35, 3025-3030.
Cd binding constant on the ith negatively (14) Fourest, E.; Volesky, B. Environ. Sci. Technol. 1995, 30, 277-
charged functional group (15) Zuiderveen, J. A. Ph.D. Dissertation, University of Kentucky, proton binding constant on the ith negatively Lexington, KY, 1994.
charged functional group (16) Evangelou, V. P. Environmental Soil and Water Chemistry: Principles and Applications; John Wiley & Sons: New York, 1998; proton binding constant on the jth positively charged functional group (17) Kjellstroem, T. In Cadmium and Health: A Toxicological and Epidemiological Appraisal; Friberg, L., Elinder, C. G., Kjellstroem, Li binding constant on the ith negatively T., Nordberg, G. F., Eds.; CRC Press: Boca Raton, FL, 1986; pp charged functional group VOL. 37, NO. 16, 2003 / ENVIRONMENTAL SCIENCE & TECHNOLOGY 9 3607
(18) Puranik P. R.; Chabukswar, N. S.; Paknikar, K. M. Appl. Microbiol. (30) Hunt, S. In Immobilization of Ions by Biosorption; Eccles, H., Biotechnol. 1995, 43, 1118-1121.
Hunt, S., Eds.; Ellis Horwood: Chichester, U.K., 1986; pp 15- (19) Valdman, E.; Leite, S. G. F. Bioprocess Eng. 2000, 22, 171-173.
(31) Haug, A. Acta Chem. Scand. 1961, 15, 1794-1795.
(20) Schecher, W. D. MINEQL+: A Chemical Equilibrium Model for (32) Schiewer, S.; Volesky, B. In Environmental Microbe-Metal Personal Computers, Users Manual Version 2.22; Environmental Interactions; Lovley, D. R., Ed.; ASM Press: Washington, DC, Research Software, Inc.: Hallowell, ME, 1991.
2000; pp 329-362.
(21) Yun, Y.-S.; Park, D.; Park, J. M.; Volesky, B. Environ. Sci. Technol. (33) Buffle, J. Complexation Reactions in Aquatic Systems: An 2001, 35, 4353-4358.
Analytical Approach; Ellis Horwood: Chichester, 1988; pp 156- (22) Yang, J. B.; Volesky, B. Environ. Sci. Technol. 1999, 33, 4079-
157, 280-283, 323.
(34) Pagnanelli, F.; Petrangeli-Papini, M.; Toro, L.; Trifoni, M.; Veglio, (23) Fourest, E.; Volesky, B. Appl. Biochem. Biotechnol. 1997, 67,
F. Environ. Sci. Technol. 2000, 34, 2773-2778.
(35) Schiewer, S.; Volesky, B. Environ. Sci. Technol. 1997, 31, 1863-
(24) Mesa Perez, J. M.; Valle Matos, M.; Brossard Perez, L.; Guerrero Haber, J. R. Technol. Quim. 1998, 18, 90-96.
(36) Schiewer, S.; Volesky, B. Environ. Sci. Technol. 1997, 31, 2478-
(25) Stumm, W.; Morgan, J. J. Aquatic Chemistry: An Introduction (37) Carnahan, B.; Luther, H. A.; Wilkes, J. O. Applied Numerical Emphasizing Chemical Equilibria in Natural Waters; John Wiley Methods; John Wiley & Sons: New York, 1969.
& Sons: New York, 1981.
(38) Wolfram, S. The Mathematica Book, 3rd ed.; Wolfram Research: (26) Westall, J. C.; Jones, J. D.; Turner, G. D.; Zachara, J. M. Environ. Champaign, IL, 1996.
Sci. Technol. 1995, 29, 951-959.
(27) Lee, H. S.; Volesky, B. Water Res. 1997, 31, 3082-3088.
Received for review November 30, 2001. Revised manuscript (28) Mardquardt, D. W. J. Soc. Ind. Appl. Math. 1963, 11, 431-441.
received April 22, 2003. Accepted April 29, 2003. (29) Garnham, G. W.; Codd, G. A.; Gadd, G. M. Biol. Met. 1991, 4,
3608 9 ENVIRONMENTAL SCIENCE & TECHNOLOGY / VOL. 37, NO. 16, 2003

Source: http://che.chonbuk.ac.kr/~ysyun/publications/pdf/ES&T(2003)modeling.pdf