Biosorption of copper and zinc by Cymodocea nodosa

Transcription

1 FEMS Microbiology Reviews 23 (1999) 527^536 Abstract Biosorption of copper and zinc by Cymodocea nodosa Antonio Sänchez, Antonio Ballester *, Mar a Luisa Bläzquez, Felisa Gonzälez, Jesus Mun oz, Adil Hammaini Departamento de Ciencia de los Materiales e Ingenier a Metalürgica, Facultad de Ciencias Qu micas, Universidad Complutense, Madrid, Spain The adsorption of the two metal ions Cu and Zn in a single-component system by Cymodocea nodosa, a brown alga, under different ph conditions was investigated. The solution ph significantly affected the exhibited uptake, being maximum at a ph value of 4.5. Multi-component mixture biosorption in aqueous solutions is also reported. A comparison was made between the single-component saturation uptake and the multi-component uptakes. To evaluate the two-metal sorption system performance, simple isotherm curves had to be replaced by three-dimensional sorption isotherm surfaces. In order to describe the isotherm surfaces mathematically, three Langmuir-type models were evaluated. The isotherms indicate a competitive uptake with Cu being preferentially adsorbed. In addition, different tests were carried out to compare the process efficiency working continuously in small columns. ß 1999 Federation of European Microbiological Societies. Published by Elsevier Science B.V. All rights reserved. Keywords: Biosorption; Cymodocea nodosa; Copper; Zinc; Wastewater treatment; Biomass Contents 1. Introduction Materials and methods Biomass preparation and chemicals In uence of ph Three-dimensional sorption isotherm surfaces Continuous experiment Analysis of heavy metal ions Results and discussion The in uence of ph in the biosorption of Cu and Zn Cu-Zn system Model Model Model Continuous adsorption * Corresponding author. Tel.: +34 (91) ; Fax: +34 (91) ; ambape@eucmax.sim.ucm.es / 99 / $20.00 ß 1999 Federation of European Microbiological Societies. Published by Elsevier Science B.V. All rights reserved. PII: S (99)

2 528 A. Sänchez et al. / FEMS Microbiology Reviews 23 (1999) 527^ Conclusion Acknowledgements References Introduction The increased use of metals and chemicals in the process industries has resulted in the generation of large quantities of aqueous e uents that contain high levels of heavy metals and pose environmental disposal problems. In addition, mining, mineral processing and extractive metallurgical operations generate huge volumes of toxic liquid waste [1]. Reducing the toxic metals to environmentally acceptable limits in a cost-e ective and environmentally friendly manner thus assumes great signi cance. The conventional physicochemical methods used for this purpose, such as chemical precipitation, electrowinning, membrane separation, evaporation or resin ionic exchange, can be very expensive, and sometimes not very e ective. Biological treatment, based on living micro-organisms or plants, could be an alternative method to clean up industrial wastewater containing heavy metals. However, these processes are very sensitive to the characteristics of the e uent, such as temperature, ph and chemical composition, and are not suitable for wastewaters with high metal concentrations. The use of dead biomass and biomass derivative (biosorbent materials) to remove heavy metals from aqueous solution has been widely studied in recent years. These systems are less expensive than the traditional physicochemical processes. They do not need nutrients and are resistant to the physicalchemical properties of heavy metal solutions [2]. Simple sorption isotherm curves are usually constructed as a result of studying equilibrium batch sorption behaviour of di erent biosorbent materials. These curves enable quantitative evaluation of biosorption performance of these materials for only one metal at a time. However, when more than one metal at a time is present in the sorption system, the evaluation, interpretation, and representation of biosorption results become much more complicated. With two metals in the solution, instead of two-dimensional biosorption isotherm curves the system evaluation results in a series of three-dimensional sorption isotherm surfaces. This novel approach is very useful particularly because there is no control over the values of the nal, equilibrium, residual concentration of the sorbate as it results from standard sorption experiments in which the residual liquid is analysed for the equilibrium concentration of the sorbate (C f ). Random experimental ndings of both C f values when a two-sorbate system is studied require such an approach [3,4]. The objective of the present work was to evaluate the in uence of ph on the uptake of Cu and Zn by Cymodocea nodosa and to compare the biosorption performance of a single-metal system in batch and continuous mode and that in batch mode between single- and two-metal systems. The choice of metals was made with regard to their industrial use and potential pollution impact. Copper is not acutely toxic to humans but its extensive use and increasing levels in the environment are causes for concern. Zinc is used very extensively by many industries, mainly in galvanisation and in manufacturing brass and other alloys. Zinc is toxic for humans at levels of 100^500 mg day 31 [5]. 2. Materials and methods 2.1. Biomass preparation and chemicals Particles (20^30 mm in length and 4^5 mm in width) of raw biomass of C. nodosa originally collected in the Mar Menor (a coastal lagoon on the Mediterranean coast of south-east Spain) were washed in distilled water for the release of common ions present in seawater. Biomass was dried at 60 ³C overnight and then used in the experiments. Analytical grades of H 2 SO 4, NaOH, CuSO 4 W5H 2 O and ZnSO 4 WH 2 O were used In uence of ph Dried algae (3 g dry weight l 31 ) were put into contact with known concentrations of metal-bearing

3 A. Sänchez et al. / FEMS Microbiology Reviews 23 (1999) 527^ solutions in the range of 25^300 mg l 31 in Erlenmeyer asks at initial ph values of 3, 4, 4.5 and 5.3 for Cu and 3.5, 4.5, 5.5 and 6.5 for Zn. The ph was adjusted by adding 10% H 2 SO 4. Maximum ph values were xed to avoid complications due to precipitation of the metals. The asks were magnetically agitated for 120 min, which is more than ample time for sorption equilibrium. One-millilitre samples of solution were taken at de nite intervals and analysed for their Cu or Zn content, ph was also measured Three-dimensional sorption isotherm surfaces The biomass was put into contact with a solution containing Cu and Zn at an initial ph value of 4.5 and at a concentration range of 0^300 mg l 31 of each of the metals. The three-dimensional (3D) sorption surfaces were obtained by plotting the experimentally nal (equilibrium) metal concentrations of both metals on the x and y coordinates respectively, against the Cu, Zn, or total metal uptakes, respectively, on the z coordinate. The computer program MATLAB (version 4) was used for this purpose Continuous experiment These experiments were carried out in three columns (10 cm in height by 2.5 cm in diameter) in cascade con guration as shown in Fig. 1. The metal-bearing solution at initial concentrations of 50, 100 and 300 mg l 31 was passed through the biomass (1 g in each column) by means of a peristaltic pump with a ow rate of 15 ml min 31. The level of solution was kept constant in each column at 50 ml Analysis of heavy metal ions The concentrations of unadsorbed Cu and Zn ions in the sample supernatant liquid were determined using an atomic absorption spectrophotometer (Perkin Elmer 1100) with an air-acetylene ame. Copper and zinc hollow cathode lamps were used. Metal uptake (q) was calculated using the general de nition: q mg or mmol g 31 ˆV C i 3C f =S 1 where C i and C f are the initial and nal (equilibrium) metal concentrations in the solution, respectively, V is the solution volume, and S is the mass of the biosorbent used. 3. Results and discussion 3.1. The in uence of ph in the biosorption of Cu and Zn A Langmuir-type adsorption isotherm model [6] was used to approximate biosorption of the two metals by C. nodosa according to the equation: q ˆ bc f q max = 1 bc f 2 where q max is the maximum sorbate (metal) uptake and b is the Langmuir constant, a ratio of the adsorption rate constant to the desorption rate constant. The apparent dissociation constant for the sorption system (K) is the ratio of the desorption to the adsorption rate constant and is the inverse of the Langmuir constant b. The ph of the solution and therefore the presence of competing protons is an important environmental condition that a ects the capacity of the biomass to remove cations [1]. The e ects of ph on the biosorption of Cu by C. nodosa are shown in Fig. 2. In general at ph 3 the uptake of Cu was very low, probably due to the e ect of proton competition. Fig. 1. Schematic outline of the experimental set-up used for continuous sorption mode studies.

4 530 A. Sänchez et al. / FEMS Microbiology Reviews 23 (1999) 527^536 Fig. 2. E ect of ph on the biosorption of Cu by C. nodosa. High ph values gave similar adsorption behaviour, with ph 4.5 being the optimum. Similar results were observed for Zn (Fig. 3). Tables 1 and 2 give the q max and K values derived from Eq. 2 at di erent initial ph values for Cu and Zn, respectively. As expected, the values of both constants were similar when ph was 4.5 or above, although, in the case of Cu, when ph was 5.3 a decrease in q max was observed. At this ph value, Cu is present as two forms in solution: Cu 2 and (CuOH). Since in this latter state copper in solution would present a larger size it would be adsorbed less easily and therefore a diminution in the biosorption capacity would be expected. Fig. 4 presents the variation of the concentration of Cu when its initial concentration was 200 mg l 31 and that of ph when the initial concentration of Cu was 200, 300 or 500 mg l 31 with respect to time. From this gure, it can be seen that there was a considerable adsorption of protons during the rst few minutes, then, when the rate of Cu adsorption increased, protons began to be desorbed in a similar manner as occurs in the ion exchange process. These results indicate a clear competition for the Table 1 Langmuir isotherm constants for Cu at di erent initial ph values ph Q max (mg g 31 ) K (mg l 31 ) Fig. 3. E ect of ph on the biosorption of Zn by C. nodosa. biomass adsorption sites between the Cu and protons. At the beginning of the biosorption both cations are adsorbed, then, when the Cu concentration is still high, a partial desorption of protons occurs allowing Cu to be adsorbed in the biomass sites left by protons. In the case of Zn, as shown in Fig. 5, the equilibrium ph decreased as the Zn concentration increased. It can also be seen that at an initial concentration of 200 mg l 31 the increase in ph (di erence between the nal and the initial ph values) was greater at ph 3.5 than at ph 4.5. This clearly indicates that protons displace more Zn from the adsorption sites at low ph. A kinetic study of the adsorption process was made using the following expression: r ˆ K c C3C f n 3 where r is the adsorption rate (mg min 31 ), K c is the kinetic constant (h mg 313n min 31 ) and n is the reaction order. As can be seen from Tables 3 and 4, the kinetic constant values were very similar for both metals, indicating that the adsorption rate is similar Table 2 Langmuir isotherm constants for Zn at di erent initial ph values ph Q max (mg g 31 ) K (mg l 31 )

5 A. Sänchez et al. / FEMS Microbiology Reviews 23 (1999) 527^ Fig. 4. Evolution of ph versus time for copper sorption. regardless of the working ph. In Table 5 the K c values for both metals are presented in units of moles. The results indicated a similar order of the reaction for both metals, although the kinetic constant was greater for copper, con rming a greater a nity of the biomass for this metal Cu-Zn system Table 3 Kinetic constants for Cu adsorption at di erent initial ph values ph K c (l n31 mg 13n min 31 )U10 2 n R Fig. 5. Evolution of ph versus time for zinc sorption. ph 4.5 was used in the study of the bicomponent system also to avoid complications due to Cu-hydroxide precipitation at higher ph. Molar concentration units were used which are useful for stoichiometric comparison of the sorption capacity for each of the two metals. The values obtained for the q max and b constants at ph 4.5 are summarised in Table 6. These values indicate a decidedly better biomass a nity for, and higher sorption of, Cu ions versus Zn ions. To propose the most suitable equation to represent the sorption data in 3D space, three models were investigated. The rst model produced an equation with three parameters, while the second and the third models had four and ve parameters, respectively. These parameters were evaluated using the MATLAB program. The following are descriptions of each of the models [5] Model 1 When equilibrium is established: B M 1 HB3M 1 K 1 ˆ k 31 =k 1 4 B M 2 HB3M 2 K 2 ˆ k 32 =k 2 5 Assuming that the sorption system is in equilibrium (there are no net changes of [B3M 1 ] and [B3M 2 ] with respect to time), therefore: d B3M 1 Š=dt ˆ 0; d B3M 2 Š=dt ˆ 0 6 B 0 Šˆ BŠ B3M 1 Š B3M 2 Š 7 Combining Eqs. 5 and 6: B3M 1 Šˆ B 0 Š M 1 Š=fK 1 M 1 Š K 1 =K 2 M 2 Šg 8 Table 4 Kinetic constants for Zn adsorption at di erent initial ph values ph K c (l n31 mg 13n min 31 )U10 2 n R

6 532 A. Sänchez et al. / FEMS Microbiology Reviews 23 (1999) 527^536 K c Table 5 The average of kinetic constants for Cu and Zn adsorption Metal K c n (l n31 mg 13n min 31 )U10 2 (l n31 mg 13n min 31 ) Cu Zn Table 6 Langmuir isotherm constants for Cu and Zn at initial ph 4.5 Metal q max (mmol g 31 ) b (l mmol 31 ) Cu Zn K (mmol l 31 ) We de ne [B3M 1 ] as the number of binding sites occupied by M 1 per gram of biosorbent, and [B 0 ] as the total number of binding sites per gram of biosorbent. Then, multiplying both sides by a value having the units of `mmol M 1 per number of binding sites', the following expression can be obtained: q M 1 ˆ q max =K 1 C f M 1 Š= f1 1=K 1 C f M 1 Š 1=K 2 C f M 2 Šg 9 Table 7 gives the three parameters derived from model Model 2 When equilibrium is established: B M 1 HB3M 1 K 1 B M 2 HB3M 2 K 2 B3M 1 M 2 HB3M 1 3M 2 K 1;2 Table 9 Langmuir constants derived from model B3M 2 M 1 HB3M 2 3M 1 K 2;1 B M 1 M 2 HB3M 2 3M 1 K ˆ K 1 K 1;2 K ˆ K 2 K 2; Assuming that the sorption system is in equilibrium (there are no net changes of [B3M 1 ], [B3M 2 ] and [B3M 1 3M 2 ] with respect to time), therefore: d B3M 1 Š=dt ˆ 0; d B3M 2 Š=dt ˆ 0 And and d B3M 1 3M 2 Š=dt ˆ 0 15 B 0 Šˆ BŠ B3M 1 Š B3M 2 Š B3M 1 3M 2 Š 16 Combining Eqs. 14 and 15: B3M 1 Šˆ B 0 Š M 1 Šf1 K 1 =K 2 M 2 Šg= fk 1 M 1 Š K 1 =K 2 M 2 Š 2 K 1 =K M 1 Š M 2 Šg 17 Table 7 Langmuir constants derived from model 1 Metal system K 1 (mmol l 31 ) K 2 (mmol l 31 ) q max (mmol g 31 ) Cu-Zn Cu: 0.02 Zn: Table 8 Langmuir constants derived from model 2 Metal system K 1 (mmol l 31 ) K 2 (mmol l 31 ) K 1;2 (mmol l 31 ) K 2;1 (mmol l 31 ) K (mmol l 31 ) 2 q max (mmol g 31 ) Metal system K 1 (mmol l 31 ) K 2 (mmol l 31 ) K 3 (dimensionless) K 4 (dimensionless) q max (mmol g 31 ) Cu-Zn Cu: 0.02 Zn:

7 A. Sänchez et al. / FEMS Microbiology Reviews 23 (1999) 527^ Following similar procedures as for model 1, the next expression can be obtained: q M 1 ˆq max C f M 1 Šf1 K 1 =K 2 C f M 2 Šg= fk 1 C f M 1 Š K 1 =K 2 C f M 2 Š 2 K 1 =K C f M 1 ŠC f M 2 Šg 18 Table 8 gives the four independent parameters derived from model Model 3 The following equation is based on the multi-component sorption isotherm model: q M 1 ˆ q max =K 1 C f M 1 Š= f1 1=K 1 C f M 1 ŠK 3 1=K 2 C f M 2 ŠK 4 g 19 Table 9 gives the ve independent parameters derived from model 3. Model 1 presented here is a binary Langmuir-type equation. For the Cu-Zn system, a higher value of the K parameter for Zn than for Cu implies that the biosorbent has a higher a nity for Cu than for Zn. Higher values of K are associated with a higher ratio of the desorption rate constant to the adsorption constant. Similar to the rst model, a Langmuir-type equation resulted from applying the second model but there are extra terms in the numerator and the denominator. It can be seen that the values of K 1;2 and K 2;1 are generally severalfold higher than those of K 1 and K 2. This implies that formation of the B3M 1 3M 2 complex is not as favourable when compared to the B3M 1 and B3M 2 complexes. As in Table 10 Percentage of calculated data which deviated (0^10%) from the experimental data Model 1 (%) Model 2 (%) Model 3 (%) Cu (31.1) (31.1) (32.9) Zn (73.7) (65.9) (73.1) Cu+Zn (25.56) (26.9) (20.35) Figures in parentheses are the maximum percentage deviation of calculated data from experimental data. Table 11 Residual sum of squares Degrees of freedom Sum of square residuals Model Model Model Mean squares model 1, the K values indicate the preference of the biosorbent according to the order Cu s Zn. The third equation, derived from the multi-component isotherm model, is similar to the rst equation except that there is a new parameter incorporated as an exponent to each of the residual concentrations in the denominator. The values of parameters K 1 and K 2 obtained from this model also lead to similar conclusions as those drawn from models 1 and 2. The values of K 3 and K 4 are quite close to unity. From these data, it can be concluded that model 3 is very similar to model 1. Tables 10 and 11 show the percentage of calculated data which deviated less than 10% from the experimental metal uptakes and the sum of the squared residuals (SSR) which resulted from applying the three models, respectively. Generally, the selection criterion for the best model is usually based on the minimum variance. Table 11 shows that the three models can make a good prediction of the metal uptake for the system studied. However, because the three models represent the data in a very similar manner, the choice of the best model is restricted by looking for the one with the lowest number of parameters. Since model 1 has only three parameters, it can be judged to be the simplest and most practical in this case and it will be further applied in describing the behaviour of the two-metal system. The equation from model 1 can be represented by 3D biosorption isotherm surfaces as shown in Figs. 6^8. While these 3D surfaces represent the overall of the two-metal equilibrium results, the selected slides (cuts) through the 3D diagrams reveal the quantitative trends observed in the two-metal sorption system better. The e ect of di erent levels of Cu on the biosorbent uptake of Zn is quantitatively much better demonstrated in Fig. 6, showing how the biosorption uptake of Zn decreases in the presence of Cu. The

8 534 A. Sänchez et al. / FEMS Microbiology Reviews 23 (1999) 527^536 Fig. 6. Two-metal sorption isotherm surface corresponding to model 1. The uptake capacity of Cu is plotted as a function of the equilibrium concentrations of Cu and Zn. curves in Fig. 9 represent series of two Cu `iso-concentration cuts' (at C f [Cu] = 0.375, 1.5 mm) of the Zn sorption surface in Fig. 6. For example, whereas in the one-metal Zn system the Zn uptake was 0.61 mmol g 31 of biomass at equilibrium C f [Zn] = 1 mm, when mm and 1.5 mm Cu was present in the system ( nal equilibrium Cu concentration), the Zn uptake decreased to 0.15 mmol g 31 and to 0.05 mmol g 31, respectively (25.5% and 7.86% of the original value, respectively). Similarly and conversely, the e ect of Zn on the biosorption of Cu is seen in Fig. 10 showing the two `iso-concentration cuts' (at C f [Zn] = 0.5, 2 mm) of Fig. 7. Two-metal sorption isotherm surface corresponding to model 1. The uptake capacity of Zn is plotted as a function of the equilibrium concentrations of Cu and Zn. Fig. 8. Two-metal sorption isotherm surface corresponding to model 1. The total uptake capacity (Cu+Zn) is plotted as a function of the equilibrium concentrations of Cu and Zn. the Cu sorption uptake surface from Fig. 7. The interference of Zn with Cu uptake was much less pronounced and was observed at much higher ratios of Zn:Cu concentration. For instance, 2 mm Zn only caused a 19.8% reduction of the Cu uptake [q (Cu) = 0.75 mmol g 31 ] at the selected nal concentration, C f [Cu] = 1 mm. The total metal uptake sorption surface is the product of adding the two individual metal uptake surfaces (Cu and Zn uptake). Fig. 8 shows that, with high levels of overall metal concentration present in Fig. 9. `Iso-concentration cuts' of the two-metal sorption isotherm surfaces (a 1 and a 2 derive from Fig. 6, b 1 and b 2 derive from Fig. 8): sorption uptakes from the two-metal solution vs. the equilibrium concentration of Cu, while the equilibrium concentration of Zn is held as a constant parameter.

9 A. Sänchez et al. / FEMS Microbiology Reviews 23 (1999) 527^ Fig. 10. `Iso-concentration cuts' of the two-metal sorption isotherm surfaces (c 1 and c 2 derive from Fig. 7, d 1 and d 2 derive from Fig. 8): sorption uptakes from the two-metal solution vs. the equilibrium concentration of Zn, while the equilibrium concentration of Cu is held as a constant parameter. Fig. 11. Breakdown curves for the copper sorption in continuous mode corresponding to the three columns. Fig. 12. Breakdown curves for the zinc sorption in continuous mode corresponding to the three columns. the solution, the biosorbent easily reaches the saturation level demonstrated by a wide plateau of the surface. When the residuals of Cu and Zn were the same (e.g., 0.5 mm each or 1.5 mm each), about 90.5% of the total metal uptake was due to Cu uptake. At C f [Cu] = 0.21 mm and C f [Zn] = 2 mm, the uptake of each metal was 0.35 mmol g 31. Thus, the equilibrium concentration of Zn would have to be 9.5 times greater than that of Cu to obtain the same proportion of uptake for both metals. In general, the biosorbent exhibited a net preference for the Cu ion over Zn. Similar competition was observed in the two-metal biosorption performance of another biosorbent [5,7] Continuous adsorption The way in which an experiment is carried out is another parameter that can a ect the capacity of a particular type of biomass to sequester metals. Since in batch experiments the metal concentration and the solution ph vary only with time, the same parameters also vary with the length of the column in continuous experiments. Figs. 11 and 12 present the metal concentration values in each column versus the total volume of the solution treated (denominated breakdown curve) at an initial concentration of 100 mg l 31 for copper and zinc, respectively. The maximum adsorption capacity of the biomass in each column is obtained by dividing the area below its corresponding curve (the amount of metal adsorbed) by the total amount of biomass used. Table 12 Maximum adsorption capacity in continuous mode for copper Initial concentration q max (mg l 31 ) 1st column 2nd column 3rd column ^ ^

10 536 A. Sänchez et al. / FEMS Microbiology Reviews 23 (1999) 527^536 Table 13 Maximum adsorption capacity in continuous mode for zinc Initial concentration q max (mg l 31 ) 1st column 2nd column 3rd column Tables 12 and 13 show the adsorption capacities of C. nodosa in each column for initial metal concentrations of 50, 100 and 300 mg l 31 for copper and zinc, respectively. The values obtained for Cu varied at around mg g 31 (0.85 mmol g 31 ), a value only slightly above that obtained in batch operation, suggesting that for this metal in particular the maximum adsorption capacity is independent of the mode of operation. In the case of zinc, the mean charge capacity in continuous mode (52.73 mg g 31 or 0.80 mmol g 31 ) was higher than that obtained in batch mode (44.96 mg g 31 or 0.69 mmol g 31 ), and similar to that obtained for copper. This would suggest that the operation mode has a strong in uence on the zinc adsorption process. 4. Conclusion In this study, the abilities of the algal biomass C. nodosa to bind copper and zinc were investigated. The ph of the solution had a strong in uence on the adsorption process, ph 4.5 being the optimum for the uptake of both metals. Ion exchange has been revealed as the predominant sequestering mechanism in the biosorption of both metals. The three Langmuir models tested represent the data corresponding to the two-metal sorption system in a very similar manner. Model 1 was used to describe the behaviour of the two-metal system revealing that the biomass had a higher a nity for Cu than for Zn. The maximum adsorption capacity of copper was not modi ed by the operating mode (batch or continuous) while it was greater for zinc in continuous mode. Acknowledgements The authors would like to thank the Commission Interministerial de Ciencia y Tecnolog a (CICYT) for supporting this research. References [1] Modak, J.M. and Natarajan, K.A. (1995) Biosorption of metals using nonliving biomass ^ A review. Miner. Metall. Process. 12 (4), 189^196. [2] Araüjo, M.M. and Teixeira, J.A. (1997) Trivalent chromium sorption on alginate beads. Int. Biodeterior. Biodegrad. 40, 63^74. [3] Figueira, M.M., Volesky, B. and Ciminelli, V.S.T. (1997) Assesment of interference in biosorption of a heavy metal. Biotechnol. Bioeng. 54, 344^350. [4] Figueira, M.M., Yang, J., Volesky, B. and Camargos, E.R.S. (1995) Interference of Fe in the Cd uptake by Sargassum biomass. Biohydrometall. Process. Vol. 2, Proc. Int. Biohydromet. Symp. Santiago de Chile, pp. 187^194. ISBN: [5] Chong, K.H. and Volesky, B. (1995) Description of two-metal biosorption equilibria by Langmuir-type models. Biotechnol. Bioeng. 47, 451^460. [6] Volesky, B. (1990) Removal and recovery of heavy metals by biosorption. In: Biosorption of Heavy Metals, Chapter 1.2. CRC Press, Boca Raton, FL. [7] Sag, Y., Kaya, A. and Kutsal, T. (1998) The simultaneous biosorption of Cu(II) and Zn on Rhizopus arrhizus: application of the adsorption models. Hydrometallurgy 50, 297^ 331.