This article appeared in a journal published by Elsevier.

The attached copy is furnished to the author for internal non-commercial research and education use, including for instruction at the authors institution and sharing with colleagues. Other uses, including reproduction and distribution, or selling or licensing copies, or posting to personal, institutional or third party websites are prohibited. In most cases authors are permitted to post their version of the article (e.g. in Word or Tex form) to their personal website or institutional repository. Authors requiring further information regarding Elseviers archiving and manuscript policies are encouraged to visit: http://www.elsevier.com/copyright

Author's personal copy

Available online at www.sciencedirect.com

Bioresource Technology 99 (2008) 46844693

Biosorption of heavy metals by Fucus spiralis

lez *, A. Ballester, M.L. Bla zquez, J.A. Mun E. Romera, F. Gonza oz
a Metalu rgica, Facultad de C. Qu micas, Universidad Complutense, Dpto. Ciencia de los Materiales e Ingenier Ciudad Universitaria, 28040 Madrid, Spain Received 19 December 2006; received in revised form 24 September 2007; accepted 27 September 2007 Available online 5 November 2007

Abstract The sorption uptake of cadmium, nickel, zinc, copper and lead by marine brown alga Fucus spiralis was investigated in bimetallic, trimetallic and multimetallic solutions. The experimental data tted very well to Langmuir model. In bimetallic systems, the anity of biomass for lead and copper increased and the sorption uptake of these metals was not aected by increasing concentrations of cadmium, nickel or zinc. However, in solutions with both metals there was a signicant mutual decrease of their sorption levels at high concentrations of the other metal. The sorption uptake of cadmium, nickel and copper was investigated in trimetallic aqueous systems. Based on the kinetic parameter b, the anity of F. spiralis for copper was considerably higher than for cadmium or nickel: bCd = 6.39, bNi = 1.82 and bCu = 17.89. In all tests, the maximum sorption uptake remained practically constant around 1 mmol/g, indicating that the number of active sites on the biomass was limited. Tests with four and ve metals showed that copper was preferentially adsorbed. The dierences between the experimental sorption data and those given by the chemical speciation program PHREEQCI were negligible. In general, the software used provided satisfactory estimated data for each metal and hence can be a useful tool to predict or simulate the real process. 2007 Elsevier Ltd. All rights reserved.
Keywords: Biosorption; Heavy metals; Fucus spiralis; Langmuir; PHREEQCI

1. Introduction The recovery of heavy metals from industrial euents is normally achieved by physicochemical processes such as: chemical precipitation, coagulation, chemical/electrochemical reduction, membrane technologies, ion exchange and sorption (Seheng et al., 2004). The quest for readily-available low-cost adsorbents has led to research into biological materials. Microorganisms such as fungi, bacteria, yeast and algae can extract heavy metals from aqueous solutions through a biological phenomenon known as biosorption (Volesky and Holan, 1995). The term covers a number of processes independent of the metabolism (physical and chemical sorption, ion exchange, complexation, chelation and inorganic microprecipitation) which take place essentially on the cell wall (Veglio and Beolchini, 1997). The biorecovery of heavy metals can be aected by physico*

Corresponding author. Tel.: +34 91 394 43 35; fax: +34 91 394 43 57. lez). E-mail address: fgonzalezg@quim.ucm.es (F. Gonza

chemical parameters of the solution such as pH, ion strength and temperature, and by other characteristics of the biomass (concentration, presence of organic and inorn ganic functional groups and pre-treatments) (Rinco et al., 2005; Mehta and Gaur, 2005). The principle virtues of biosorption are the economy of the process that uses cheap adsorbents, and good kinetics of metal recovery. In addition, there is a great variety of biosorbents with metal uptakes comparable to that achieved with industrial sorbents such as ion exchange resins (Volesky, 1999). For that reason, biosorption is an innovative technology for the recovery of heavy metals from aqueous solutions. The process is of great interest to the scientic community in view of the good results obtained. Industrial euents, however, far from being single-component, are complex solutions containing several metals simultaneously. In the biosorption of complex solutions, dierent metal ions may compete for the active sites present on the cell wall of the biomass. Consequently, the preference of the biomass for some metals is an important issue,

0960-8524/$ - see front matter 2007 Elsevier Ltd. All rights reserved. doi:10.1016/j.biortech.2007.09.081

Author's personal copy

E. Romera et al. / Bioresource Technology 99 (2008) 46844693

4685

which is aected by the type of biomass, its physicochemical characteristics, its preparation or previous conditionn ing, and the nature of the metallic solution (Rinco et al., 2005). Many dierent biomasses can eectively uptake heavy metals, but recent research has shown that algae are more eective (Romera et al., 2006). An additional advantage is that algae are present in unlimited amounts in seas and oceans especially brown (Phaeophyta) and red (Rhodophyta) varieties that grow exclusively in salt water. Green (Chlorophyta) algae, on the other hand, can also grow in both fresh water and soils (Klimmek et al., 2001). Most research in this eld deals with brown algae (Chong and Volesky, 1995; Matheickal and Yu, 1999; Matheickal et al., 1999; Leusch et al., 1995); while there has been less interest in green (Aksu et al., 1997, 1999; Do nmez et al., 1999; Fraile et al., 2005) and red algae (Murphy et al., 2007), probably because of brown algae are more eective biosorbents (Davis et al., 2003). That was one reason to select Fucus spiralis for the present research. Its sorption uptake was evaluated using multimetallic solutions (containing combinations of Cd2+, Cu2+, Ni2+, Pb2+ and Zn2+) in order to quantify the eect of each metal on the sorption uptake of the remaining metals. The ultimate aim of this study was the treatment of complex and multimetallic solutions (e.g. industrial euents) in a sequential way, testing rstly bimetallic systems and later systems with three, four and up to ve metals in solution. In this study, experimental data were tted to sorption isotherms, which are a very valuable tool for quantifying the maximum sorption uptake of the biomass under equilibrium conditions of constant temperature. Langmuir models were chosen, since they provide the equilibrium constants necessary to dene the chemical equilibria between metal and biomass. These studies are particularly important in order to determine how dierent metals compete for active sites present in the cell walls of the biomass responsible for sorption, and to ascertain the degree of biomass selectivity with respect to dierent metals present in solution (Sag et al., 2001). Finally, the process was also simulated using the chemical speciation computer program PHREEQCI 6.2 in order to check the predictability of the biosorption process for the treatment of complex euents. 2. Methods h Samples of F. spiralis, a marine brown alga (Cabioc et al., 1995; Van den Hoek et al., 1995; Graham and Wilas Bajas (Galicia, cox, 1998), collected in autumn in Las R Spain) were used as biosorbent. The biomass was conditioned as follows: rstly, it was thoroughly cleaned until visually free of impurities and then washed with distilled water repeatedly. The overow from each wash was centrifuged at 5000 rpm for 1015 min and the clear liquid removed. The resulting pellet was mixed with the previously washed biomass and the whole system was dried in

a stove at 60 C till reach a constant weight. The samples were then ground to <0.5 mm, the particle size used in the biosorption tests. The experiments were performed with 1000 mg/L synthetic solutions of Cd2+, Cu2+, Ni2+, Pb2+ and Zn2+ prepared from the corresponding chemically pure nitrate by dilution in distilled water. pH was adjusted with diluted nitric acid 1% or 10% v/v and with two alkaline reagents: ammonia 1% or 10% v/v or chemically pure NaOH dissolved in distilled water at a concentration of 1 g/L. Tests were carried out in Erlenmeyer asks containing the solution with the heavy metal and the appropriate amount of biomass. The asks were magnetically stirred for 2 h on a multiple stirring plate and the process was controlled by removing samples from the mixture at dierent times. The pH was measured in the overow and the metal concentration determined by atomic absorption spectrophotometry previous centrifugation of the sample at 5000 rpm. In biosorption tests with two metals, the biomass was in contact with 100 mL of synthetic solutions. In this case, a series of tests was done at constant concentration of the rst metal (0, 10, 25, 50, 100 and 150 mg/L) and varying concentration of the second one (within the same range). Then, a total of ten dierent bimetallic systems were tested combining the ve metals studied. In all cases, the initial pH was set at 4.5 and the biomass concentration at 0.5 g/L. The main drawback for calculating the sorption isotherms in multimetallic systems is that the number of experiments necessary to dene completely each system grows exponentially with the number of metals. Therefore, only three of the ve metals used were selected to carry out the study in trimetallic systems: cadmium, copper and nickel. Dierent series of experiments were performed by setting the concentration of the rst metal and varying the concentration of the second and third metals in the full range between 0 and 150 mg/L. In tetrametallic systems, aqueous solutions of the three previous metals plus zinc were employed, and in pentametallic systems lead was also considered. In all cases, the metallic solutions tested contained the same initial concentration of each metal. Finally, the process was simulated using the chemical speciation program PHREEQCI 6.2. 3. Results and discussion 3.1. Bimetallic systems 3.1.1. Sorption isotherms The best way to study sorption equilibria with two metals is to adjust the experimental data to a mathematical model from which a number of parameters can be obtained for quantitative evaluation of sorption equilibrium uptake. Those parameters should also provide information about the inuence that the presence of a given metal has on the sorption uptake of another one. There is scarce information on multimetallic system modelling (Aksu et al., 1997; Chong and Volesky, 1996;

Author's personal copy

4686

E. Romera et al. / Bioresource Technology 99 (2008) 46844693

Yu and Kaewsarn, 1999; Singh et al., 2000; Silva et al., 2003). For the present research, the Langmuir model, dened by Eqs. (1) and (2) was selected for the sake of simplicity and good results. The model represents a binarytype equation which establishes the equilibrium between two metals, M1 and M2, and the biomass, resulting in two dierent sorption species, BM1 and BM2: B M 1 B M 1 B M 2 B M 2 B M 1 BM 1 B M 2 b2 BM 2 b1 1 2

The nal expressions are as follows: q M 1 q M 2 qmax b1 C e M 1 1 b1 C e M 1 b 2 C e M 2 qmax q b C M 2 K 2 max 2 e 1 b1 C e M 1 b 2 C e M 2 3 4

For each bimetallic system, the dierent parameters (b1, b2 and qmax) that quantify the process were obtained from the experimental equilibrium data (Ce (M1), Ce (M2), q (M1) and q (M2)) using Matlab 5.1 software. This program allowed to plot Eqs. (3) and (4) in a series of sorption isotherms in 3D. For that, the nal equilibrium concentrations of both metals, on the X and Y-axes, were plotted against the sorption uptake of one metal or the sum of both, on the Z-axis. Table 1 shows the values of Langmuir constants qmax and b for each metal (b1 and b2) in molar units and their corresponding linear regression coecients (R12 and R22); which provide the degree of adjustment between experimental data and model and hence the corresponding sorption surfaces. The maximum sorption uptake of the biomass remained practically constant around 1 mmol/g. This suggests that the total number of metal active sites available on the biomass is xed and independent of the solution used. The lowest value was registered in the NiZn system, which is consistent with the fact that recovery of individual metals in monometallic systems was lower (Romera et al., 2007)

Table 1 Values of Langmuir constants for the ten bimetallic systems studied System CdNi CdZn CdCu CdPb NiZn NiCu NiPb ZnCu ZnPb CuPb qmax (mmol/g) 0.75 0.80 0.96 1.09 0.64 0.92 0.97 0.95 1.04 1.07 b1 (L/mmol) 21.79 21.50 7.28 3.03 9.89 1.98 1.03 1.98 1.36 5.48 b2 (L/mmol) 5.32 7.57 17.61 44.25 12.61 18.66 93.46 15.70 50.00 44.44 R12 0.99 0.99 0.97 0.81 0.94 0.85 0.69 0.71 0.70 0.92 R22 0.99 0.95 0.98 0.86 0.99 0.98 0.70 0.99 0.84 0.89

Note: For each system, metals 1 and 2 appear in rst and second place respectively.

A higher value of b for the rst metal (b1) than for the second one (b2) means a higher anity of the biosorbent towards the former. High values of b are associated with a high ratio of sorption/desorption of the metal by the biomass (Veglio and Beolchini, 1997; Lau et al., 1999). Therefore, the values of b are an indication of the anity of the biomass for one metal in the presence of another. A comparative analysis of b1 and b2 reveals that: the highest values of b were obtained for lead and copper whereas they were signicantly low for Ni and Zn (Table 1). The dierence between the corresponding values of b was greater in the NiPb system and smaller in the NiZn system. Cadmium presented an intermediate behaviour. The results indicate that metal sorption uptake by the biomass diers according to the following sequence: Pb > Cu > Cd > Zn > Ni. This is consistent with the hypothesis, proposed by various researchers, that the sorption process depends on the characteristics of the metal adsorbed (Chong and Volesky, 1996; Singh et al., 2000; Sag et al., 2001). In addition, the linear regression coecients were very high, indicating that the experimental data tted the Langmuir model very well. In most cases the coecient was higher than 90%. Fig. 1 shows the sorption isotherms of the ZnCu system, one of the 10 bimetallic systems studied. While the zinc sorption uptake was greatly aected by the presence of copper in solution (Fig. 1a), the presence of zinc in solution had no eect upon copper sorption (Fig. 1b). Thus, the eect of zinc on the sorption of Cu was negligible but copper highly aected the sorption of Zn, even at low concentrations of copper. Those dierences would be due to both the competence between metals for the same active sites on the biomass (Lau et al., 1999) and the higher anity of the biomass for copper than for zinc because of a higher value of b (Table 1). The evolution of the process in each bimetallic system can be followed more precisely through transversal cuts of the isotherms plotted in Fig. 1. In this way, the changes in the sorption process of a metal can be recorded as the concentration of the second metal increases. Therefore, for each system, the sorption uptake of each metal can be plotted versus the corresponding concentration at equilibrium, considering the concentration of the second metal as variable. For the ZnCu system, the presence of copper in solution decreased the sorption of zinc up to 80% while the sorption of copper stabilized around 1 mmol/g regardless the zinc concentration (Fig. 2). Another important result, as shown in Fig. 1c, is that the total amount of metal adsorbed by the biomass remains constant at equilibrium around 1 mmol/g (Table 1). That conrms the Langmuir model hypothesis that the total number of available active sites on the biomass also remains constant over the entire surface. Moreover, this is consistent with the idea that the metal uptake by the biomass is related not only to the chemical characteristics of the metal itself, but also to the type of biomass.

Author's personal copy

E. Romera et al. / Bioresource Technology 99 (2008) 46844693

0.80

4687

q Zn (mmol/g)

0.60 0.40 0.20 0.00 0.00

0.50

1.00

1.50

2.00

Ce Zn (mmol/L)
0 ppm Cu 50 ppm Cu 10 ppm Cu 100 ppm Cu 25 ppm Cu 150 ppm Cu

1.00

q Cu (mmol/g)

0.80 0.60 0.40 0.20 0.00 0.00 0.50 1.00 1.50 2.00

Ce Cu (mmol/L)
0 ppm Zn 50 ppm Zn 10 ppm Zn 100 ppm Zn 25 ppm Zn 150 ppm Zn

Fig. 2. Transversal sections of sorption isotherms in the ZnCu system (Fig. 1) at dierent concentrations of the second metal.

Fig. 1. Sorption isotherms surfaces in bimetallic system: ZnCu.

3.1.2. Simulation of the process The next step in this study was to simulate the corresponding equilibrium data with a computer program for chemical speciation, PHREEQCI 6.2 (Charlton et al., 1997) and compare them with the experimental data. The known data for each system are: b, the equilibrium constant or anity constant of Langmuir model; the maximum metal uptake of the biomass (qmax), and the total number of available active sites for the metal on the biomass surface.

The program can predict the behaviour of the biomass when the nal metal concentration in solution to be reached is provided. This assumption is of special interest since the environmental legislation on reduction of the level of metals in contaminated euents is becoming stricter. Hence, when this value of nal metal concentration is entered, the program can predict the amount of metal retained by the biomass (q or number of available active sites for that metal occupied under those conditions) and the free active sites after reaching equilibrium (qmaxq). In this way, the reactions between the biomass and each metal, with their corresponding equilibrium constants given by the b constant of the Langmuir model, are loaded into a data base in the program. In addition, the program requires that the experimental conditions of the solution (pH, temperature, equilibrium metal concentration, ionic species present in solution, etc.) and the characteristics of the biomass (qmax, specic area and weight) be entered. Table 2 shows the dierent parameters obtained in the bimetallic system ZnCu: the initial and equilibrium metal concentrations (C0 and Ce, respectively, for Cu and Zn), the experimental amount of metal retained by the biomass (qe) and the amount calculated by the program (qp). The table also shows the dierence between the percentage of active sites on the biomass occupied by the metal [qe = (qe/qmax).100] obtained experimentally and the percentage calculated by the simulation program [qp = (qp/ qmax).100]. Based on the minimum dierences obtained between experimental and calculated data, the program satisfactorily estimated the experimental data for each

Author's personal copy

4688

E. Romera et al. / Bioresource Technology 99 (2008) 46844693

Table 2 Comparison between experimental data and those estimated by the computer program PHREEQCI in bimetallic system: ZnCu CoZn (mg/L) 0 0 0 0 0 0 10 10 10 10 10 10 25 25 25 25 25 50 50 50 50 50 50 100 100 100 100 100 100 150 150 150 150 150 150 CoCu (mg/L) 0 10 25 50 100 150 0 10 25 50 100 150 10 25 50 100 150 0 10 25 50 100 150 0 10 25 50 100 150 0 10 25 50 100 150 Experimentation CeZn (mol/L) 0.00E+00 0.00E+00 0.00E+00 0.00E+00 0.00E+00 0.00E+00 4.00E05 5.00E05 8.00E05 1.10E04 1.20E04 1.20E04 1.80E04 2.40E04 2.80E04 3.10E04 3.10E04 4.10E04 4.60E04 5.00E04 5.70E04 6.10E04 6.30E04 1.22E03 1.22E03 1.22E03 1.22E03 1.22E03 1.22E03 1.68E03 1.68E03 1.68E03 1.68E03 1.68E03 1.68E03 CeCu (mol/L) 0.00E+00 2.00E 1.00E04 2.90E04 9.90E04 1.65E03 0.00E+00 2.00E05 1.20E04 3.20E04 1.02E03 1.68E03 3.00E05 1.10E04 3.50E04 1.02E03 1.72E03 0.00E+00 4.00E05 1.20E04 3.60E04 1.04E03 1.68E03 0.00E+00 5.00E05 1.50E04 4.10E04 1.07E03 1.73E03 0.00E+00 5.00E05 1.70E04 4.30E04 1.09E03 1.75E03 qeZn (mol/g) 0.00E+00 0.00E+00 0.00E+00 0.00E+00 0.00E+00 0.00E+00 2.00E04 1.70E04 9.00E05 5.00E05 2.00E05 2.00E05 3.10E04 2.00E04 1.10E04 6.00E05 6.00E05 5.80E04 3.10E04 2.80E04 1.50E04 6.00E05 3.00E05 6.10E04 3.10E04 3.10E04 3.10E04 3.10E04 3.10E04 6.10E04 3.10E04 3.10E04 3.10E04 3.10E04 3.10E04 qeCu (mol/g) 0.00E+00 2.50E 5.60E04 8.50E04 9.40E04 1.01E03 0.00E+00 2.20E04 4.90E04 8.00E04 8.80E04 9.80E04 2.10E04 5.10E04 7.60E04 8.80E04 8.50E04 0.00E+00 1.90E04 4.80E04 7.10E04 8.30E04 9.10E04 0.00E+00 1.70E04 4.00E04 6.30E04 7.70E04 8.50E04 0.00E+00 1.80E04 3.80E04 5.80E04 7.60E04 8.50E04 Simulation qpZn (mol/g) 0.00E+00 0.00E+00 0.00E+00 0.00E+00 0.00E+00 0.00E+00 6.61E05 6.43E05 4.99E05 3.42E05 1.37E05 8.60E06 1.78E04 1.40E04 7.69E05 3.46E05 2.14E05 3.93E04 3.26E04 2.40E04 1.41E04 6.45E05 4.34E05 6.27E04 5.22E04 3.91E04 2.36E04 1.18E04 7.82E05 6.86E04 5.91E04 4.44E04 2.88E04 1.52E04 1.03E04 qpCu (mol/g) 0.00E+00 2.07E04 5.48E04 7.54E04 8.80E04 9.06E04 0.00E+00 1.90E04 5.53E04 7.38E04 8.69E04 8.99E04 2.19E04 4.76E04 7.13E04 8.48E04 8.87E04 0.00E+00 2.10E04 4.27E04 6.61E04 8.21E04 8.64E04 0.00E+00 1.59E04 3.58E04 5.92E04 7.71E04 8.31E04 0.00E+00 1.31E04 3.35E04 5.51E04 7.39E04 8.07E04 qeqp Zn (%) 0.0 0.0 0.0 0.0 0.0 0.0 14.1 11.1 4.2 1.7 0.7 1.2 13.9 6.3 3.5 2.7 4.1 19.7 1.6 4.2 1.0 0.5 1.4 1.8 22.3 8.5 7.8 20.3 24.4 8.0 29.6 14.1 2.3 16.6 21.8 Cu (%) 0.0 4.5 1.3 10.2 6.3 10.9 0.0 3.2 6.7 6.5 1.2 8.6 1.0 3.6 5.0 3.3 3.9 0.0 2.1 5.6 5.2 1.0 4.8 0.0 1.1 4.5 4.1 0.1 2.0 0.0 5.1 4.7 3.1 2.2 4.5

metal. The linear regression coecients for each bimetallic system were, in all cases, better than 90%. Similar results were obtained for the other nine bimetallic systems studied. The programs prediction function is therefore potentially a useful tool for simulating the real process. 3.2. CdNiCu system 3.2.1. Sorption isotherms The ternary-type Langmuir equation used denes the equilibrium between the three metals (M1, M2 and M3) present in solution and the species resulting from sorption by the biomass B (BM1, BM2 and BM3). The mathematical expression is totally equivalent to the binary Eqs. (3) and (4), but with an additional term corresponding to the third metal present in solution. The parameters dened, in this case, by the model (qmax = 0.93 mmol/g, bCd = 6.39 L/mmol, bNi = 1.82 L/ mmol and bCu = 17.89 L/mmol) were again determined

using the Matlab 5.1 software from the experimental data (values of Ce (Cd), Ce (Cu), Ce (Ni), q (Cd), q (Cu) and q (Ni)) in all the 216 tests performed on this trimetallic system (combinations of six concentrations taken three by three). In addition, the adjustment of experimental data to the Langmuir model was very good, with linear regression coecients of the three metals close to unity: RCd2 = 0.94, RNi2 = 0.82 and RCu2 = 0.99. The maximum sorption uptake achieved is around 1 mmol/g, comparable to the values in bimetallic systems. In addition, the b values indicate that the biomass preferentially takes up copper. Thus, the sorption of the rest of metals is aected by the presence of copper in solution. This issue is addressed more fully later on. In systems with more than two metals, the graphic plot of the sorption isotherms and of the mathematical model is more complicated since representation in more than three dimensions is not possible and four dimensions would be necessary, one for the sorption uptake of the biomass for

Author's personal copy

E. Romera et al. / Bioresource Technology 99 (2008) 46844693

4689

each metal or the sum of all of them, plus three more for the equilibrium concentrations of the three metals. Therefore, for a xed concentration of one metal the corresponding isotherms have been drawn for the other two. This implies that several surfaces are represented in the same plot, one for each concentration of the rst metal. Fig. 3 shows the dierent possible ways representations of the sorption isotherms. The CdCu isotherm at increasing concentrations of nickel is shown in Fig. 3a, the NiCu isotherm at increasing concentrations of cadmium in Fig. 3b and the CdNi isotherm at increasing concentrations of copper in Fig. 3c. In general, surfaces decrease with increasing concentration of the third metal. For instance, sorption of cadmium is very noticeable in the presence of copper, especially at low concentrations of nickel (Fig. 3a). If the amounts of the second and the third metals in solution are increased, competition for the active sites of the biomass increases,

while the sorption of cadmium appreciably diminishes. Sorption of copper, on the other hand, was practically unaected by the presence of either cadmium or nickel in solution. Sorption of nickel decreased drastically in the presence of copper and less so with cadmium (Fig. 3b). The sorption of nickel was practically zero, even at relatively low copper concentrations. Also, at low copper concentrations, the presence of cadmium was another key factor in the sorption of nickel since its value practically halved. Again, sorption of copper was unaected by the presence of cadmium in solution and little aected by nickel. Copper controlled the sorption uptake of cadmium and nickel by the biomass (Fig. 3c). In fact, biosorption of both metals was practically zero at the highest copper concentration tested (150 ppm). However, sorption of cadmium was unaected by the presence of nickel in solution. Conversely, the presence of cadmium aected sorption of

Fig. 3. Sorption isotherms surfaces in the CdNiCu trimetallic system: (a) CdCu system at dierent initial concentrations of Ni; (b) NiCu system at dierent initial concentrations of Cd and (c) CdNi system at dierent initial concentrations of Cu.

Author's personal copy

4690

E. Romera et al. / Bioresource Technology 99 (2008) 46844693

nickel, but only if the copper concentration in solution was not very high. Fig. 3 also indicates that the number of active sites on the biomass was constant and most of them were occupied by copper, especially at high concentrations of that metal. 3.2.2. Simulation of the process The chemical speciation program PHREEQCI 6.2 (Charlton et al., 1997) also was used to simulate the experimental data for equilibrium conditions in the trimetallic system. The program predicted the behaviour of the biomass after providing the nal metal equilibrium concentration in solution to be reached. For that purpose the chemical reactions between the biomass and each metal, with their corresponding equilibrium constants as determined by the b constant of the Langmuir model, were loaded into the program database. The treated data were used to compile a table with the values of the active sites of the biomass occupied by the metal [qe = (qe/qmax).100], both ascertained experimentally and calculated by the simulation program [qp = (qp/ qmax).100]. The number of experiments done to complete the study of this system was too high (216 tests), given that table with the same number of rows and so unmanageable. The linear regression coecients giving the degree of t between experimental data and those calculated by the program were: RCd2 = 0.94, RNi2 = 0.83 and RCu2 = 0.98. For cadmium and copper the dierence between calculated and experimental data was smaller than 10%. For nickel that dierence was slightly higher (17%) but still acceptable. In short, the program PHREEQCI 6.2 is a useful tool for simulating the biosorption process in trimetallic systems with the biomass and under the experimental conditions described here. 3.3. Systems with four and ve metals For a good approximation of the experimental data to the real situation, euents with more metals needed to be treated. Therefore, tests were performed with solutions containing four and ve metals. The purpose of these tests was to determine whether: (1) in tetra- and pentametallic systems the biomass still exhibited a preference for a given metal irrespective of the type of solution, as in the case of copper in the trimetallic system; and (2) the PHREEQCI program could predict the behaviour of the biomass without the need of all the experimental tests given that according to the trimetallic system, 1290 experiments would be necessary to complete the study in the system with four metals (combinations of six concentrations taken four by four) and 7740 experiments in the system with ve metals (combinations of six concentrations taken ve by ve). Tests were carried out using equal concentrations, that is, solutions with each metal at the same initial concentration. In this way, even with an incomplete study of each

system the preferential sorption uptake of each metal by the biomass could be analyzed. Fig. 4a represents the same, but for the trimetallic system, which served as reference for the system with four metals (CdCuNiZn) shown in Fig. 4b. Nickel uptake was strongly aected by the presence of other metal ions in solution, especially with increasing concentrations (Fig. 4a). Cadmium, however, was practically unaected; while copper was preferentially adsorbed by the biomass. Identical results have been reported in earlier studies. In solutions with four metals the same trend was observed, but the sorption uptake of each metal diminished (Fig. 4b). Since the number of active sites on the biomass is xed, if the number of metals in solution is increased, the sorption level of each metal does not remain constant unless other metals are not adsorbed.

0.8 0.7 0.6

Cd Ni Cu

q (mmol/g)

0.5 0.4 0.3 0.2 0.1 0

10 ppm

25 ppm

50 ppm

100 ppm

150 ppm

0.6 Cd 0.5 0.4 0.3 0.2 0.1 0 10 ppm 25 ppm 50 ppm 100 ppm 150 ppm Ni Zn Cu

q (mmol/g)

0.35

Cd
0.3 0.25

Ni Zn Cu

q (mmol/g)

0.2 0.15 0.1 0.05 0 10 ppm 25 ppm 50 ppm 100 ppm 150 ppm

Fig. 4. Sorption uptake values using metal equal concentrations: (a) Cd CuNi system, (b) CdCuNiZn system and (c) CdCuNiPbZn system.

Author's personal copy

E. Romera et al. / Bioresource Technology 99 (2008) 46844693

4691

Copper was still the metal with the best recovery, followed by cadmium, while nickel and zinc presented very low sorption uptake. In pentametallic solutions (CdCuNiPbZn), the pH values reached were theoretically lower than that of lead precipitation. However, several researchers have drawn

attention to the formation of microprecipitates of lead at pH higher than 5 (Tsezos, 2001), which value was surpassed during the sorption process. For that reason, the results for lead are not plotted in Fig. 4c, which shows all the other metals while taking into account that the solution also contained lead.

Table 3 Comparison between experimental data and those estimated by the PHREEQCI program for the CdNiZnCu system Ce (mol/L) Cadmium 10 25 50 100 150 10 25 50 100 150 10 25 50 100 150 10 25 50 100 150 9.79E02 1.42E01 1.25E01 1.07E01 1.07E01 1.02E01 1.02E01 6.81E02 6.81E02 6.81E02 1.07E01 9.18E02 9.18E02 9.18E02 9.18E02 1.89E01 4.09E01 5.04E01 5.66E01 5.66E01 qe (mol/g) 9.79E05 1.42E04 1.25E04 1.07E04 1.07E04 1.02E04 1.02E04 6.81E05 6.81E05 6.81E05 1.07E04 9.18E05 9.18E05 9.18E05 9.18E05 1.89E04 4.09E04 5.04E04 5.66E04 5.66E04 qp (mol/g) 9.91E05 1.92E04 2.02E04 2.01E04 2.34E04 8.88E05 1.15E04 1.06E04 1.03E04 1.14E04 7.23E05 1.03E04 9.44E05 9.12E05 1.03E04 2.50E04 3.72E04 4.95E04 5.52E04 7.06E04 qe (%) 9.79 14.23 12.46 10.68 10.68 10.22 10.22 6.81 6.81 6.81 10.71 9.18 9.18 9.18 9.18 18.88 40.91 50.35 56.65 56.65 qp (%) 9.91 19.16 20.24 20.14 23.43 8.88 11.54 10.59 10.30 11.35 7.23 10.26 9.44 9.12 10.27 25.01 37.16 49.54 55.23 70.63 qeqp (%) 0.12 4.93 7.78 9.46 12.75 1.34 1.32 3.78 3.49 4.54 3.48 1.08 0.26 0.06 1.09 6.13 3.75 0.81 1.42 13.98

Nickel

Zinc

Copper

Table 4 Comparison between experimental data and those estimated by the PHREEQCI program for the CdNiZnCuPb system Ce (mol/L) Cadmium 10 25 50 100 150 10 25 50 100 150 10 25 50 100 150 10 25 50 100 150 10 25 50 100 150 3.56E02 1.65E01 3.83E01 8.19E01 1.16E+00 1.11E01 3.41E01 7.16E01 1.43E+00 2.18E+00 9.18E02 3.06E01 6.73E01 1.35E+00 1.96E+00 3.93E02 1.89E01 5.35E01 1.23E+00 2.01E+00 5.79E03 6.27E03 4.20E02 2.03E01 4.33E04 qe (mol/g) 8.90E05 9.79E05 8.90E05 7.12E05 7.12E05 8.52E05 6.81E05 6.81E05 6.81E05 6.81E05 9.18E05 9.18E05 6.12E05 6.12E05 6.12E05 1.89E04 3.15E04 3.15E04 3.15E04 3.15E04 8.11E05 2.30E04 3.89E04 4.63E04 4.85E04 qp (mol/g) 2.65E05 4.90E05 6.32E05 9.00E05 9.87E05 2.61E05 3.21E05 3.75E05 4.99E05 5.89E05 3.05E05 4.08E05 4.99E05 6.67E05 7.55E05 5.40E05 1.04E04 1.66E04 2.56E04 3.29E04 5.67E04 6.44E04 6.04E04 4.79E04 3.89E04 qe (%) 8.90 9.79 8.90 7.12 7.12 8.52 6.81 6.81 6.81 6.81 9.18 9.18 6.12 6.12 6.12 18.88 31.47 31.47 31.47 31.47 8.11 22.97 38.90 46.33 48.48 qp (%) 2.65 4.90 6.32 9.00 9.87 2.61 3.21 3.75 4.99 5.89 3.05 4.08 4.99 6.67 7.55 5.40 10.44 16.56 25.58 32.86 56.70 64.35 60.43 47.86 38.90 qeqp (%) 6.25 4.89 2.58 1.88 2.75 5.91 3.6 3.06 1.82 0.92 6.13 5.1 1.13 0.55 1.43 13.48 21.03 14.91 5.89 1.39 48.59 41.38 21.53 1.53 9.58

Nickel

Zinc

Copper

Lead

Author's personal copy

4692

E. Romera et al. / Bioresource Technology 99 (2008) 46844693

Tables 3 and 4 show the results produced by the chemical speciation program PHREEQCI. As regards interpretation of these tables, the rst row of each metal shows the results obtained with an initial metal concentration of 10 ppm; the second row shows the results at 25 ppm, and so on. The results were very satisfactory since the programs prediction approximated well to the real level of uptake of each metal by the biomass. The last column represents the dierence, in percentage terms, between the experimental results and the program predictions for active sites occupied on the biomass. The estimation of active sites occupied by each metal was also very good in both cases, especially considering that the study of the corresponding sorption isotherms was incomplete. In quaternary mixtures of metals, the dierence between data as estimated by the program and as determined experimentally was around 10%. In the pentametallic system (Table 4) the calculated and the experimental data for cadmium, nickel and zinc were very close. In the case of copper, although the dierences were larger, they were still indicative of the behaviour of the biomass under equilibrium conditions. However, in the case of lead the deviations were greater, especially for systems with low metal concentrations and higher pH values. The program indicated that this was the case, since it detected the presence of a precipitate of lead. Erroneous determination of the active sites for one metal in this case lead would aect calculations for all the other metals, since the total number of active sites remains constant. Such an eect would be all the greater in the case of copper since it is the metal preferentially adsorbed by the biomass. 4. Conclusions The brown alga F. spiralis showed high levels of sorption to Cd, Cu, Ni, Pb and Zn and experimental data tted very well of to the Langmuir model. The biosorption process of heavy metals depended on the type of biomass used, since the number of active sites, and hence the sorption uptake, is limited. In the case of Fucus spiralis, the uptake value was around 1 mmol/g. In bimetallic systems the biomass showed a more pronounced anity for lead and copper than for the other metals. For both metals, the sorption uptake was practically unaected by increasing concentrations of cadmium, nickel or zinc. The values of b or constant of interaction between the biomass and the metal cations for the trimetallic system (CdCuNi), indicate a marked anity of F. spiralis for copper, with much higher b values for this metal than for cadmium or nickel. For instance, the presence of cadmium or nickel in the same solution hardly aected the sorption uptake of the other two metals. Conversely, the presence of copper negatively aected the sorption uptake of cadmium and nickel. Tests with four and ve metals showed that copper was preferentially adsorbed. The dierences between experimental sorption data and those given by the chemical speciation

program PHREEQCI were minimal. The software provided satisfactory estimated data for each metal and can be used as a potential tool for simulating the real process. Acknowledgements The authors wish to express their gratitude to Spanish Ministry of Science and Technology for the funding of this work. References
., Kutsal, T., 1997. Application of multicomponent Aksu, Z., Acikel, U adsorption isotherms to simultaneous biosorption of iron(III) and chromium(VI) on C. vulgaris. J. Chem. Tech. Biotechnol. 70, 368 378. ., Kutsal, T., 1999. Investigation of simultaneous Aksu, Z., Acikel, U biosorption of copper(II) and chromium(VI) on dried Chlorella vulgaris from binary metal mixtures: application of multicomponent adsorption isotherms. Separ. Sci. Technol. 34, 501524. h, J., Floc h, J.Y., Le Toquin, A., Boudouresque, C.F., Meinesz, A., Cabioc a de las algas de los mares de Europa: Verlaque, M., 1995. Gu ntico y Mediterra neo. Ed. Omega. Atla Charlton, S.R., Macklin, C.L., Parkhurst, D.L., 1997. PHREEQCI- A graphical user interface for the geochemical computer program PHREEQC. US Geological Survey. Water-Resources Investigations Report 97-4222, Lakewood, CO. Chong, K.H., Volesky, B., 1995. Description of two-metal biosorption equilibria by Langmuir-type models. Biotechnol. Bioeng. 47, 451 460. Chong, K.H., Volesky, B., 1996. Metal biosorption equilibria in a ternary system. Biotechnol. Bioeng. 49, 629638. Davis, T.A., Volesky, B., Micci, A., 2003. A review of the biochemistry of heavy metal biosorption by brown algae. Water Res. 37, 43114330. ztu Do nmez, G., Aksu, Z., O rk, A., Kutsal, T., 1999. A comparative study on heavy metal biosorption characteristics of some algae. Proc. Biochem. 34, 885892. lez, F., Bla zquez, M.L., Mun Fraile, A., Penche, S., Gonza oz, J.A., Ballester, A., 2005. Biosorption of copper, zinc, cadmium and nickel by Chlorella vulgaris. Chem. Ecol. 21, 6175. Graham, Linda E., Wilcox, Lee W., 1998. Algae. Prentice Hall. Klimmek, S., Stan, H.J., Wilke, A., Bunke, G., Buchholz, R., 2001. Comparative Analysis of the biosorption of cadmium, lead, nickel and zinc by algae. Environ. Sci. Technol. 35, 42834288. Lau, P.S., Lee, H.Y., Tsang, C.C.K., Tam, N.F.Y., Wong, Y.S., 1999. Eect of metal interference, pH and temperature on Cu and Ni biosorption by Chlorella vulgaris and Chlorella miniata. Environ. Technol. 20, 953961. Leusch, A., Holan, Z.R., Volesky, B., 1995. Biosorption of heavy metals (Cd, Cu, Ni, Pb, Zn) by chemically reinforced biomass of marine algae. J. Chem. Technol. Biotechnol. 62, 279288. Matheickal, J.T., Yu, Q., 1999. Biosorption of lead(II) and copper(II) from aqueous solutions by pre-treated biomass of Australian marine algae. Bioresour. Technol. 69, 223229. Matheickal, J.T., Yu, Q., Woodburn, G.M., 1999. Biosorption of cadmium(II) from aqueous solutions by pre-treated biomass of marine alga Durvillaea potatorum. Water Res. 33, 335342. Mehta, S.K., Gaur, J.P., 2005. Use of algae for removing heavy metal ions from wastewater: progress and prospects. Crit. Rev. Biotechnol. 25, 113152. Murphy, V., Hughes, H., McLoughlin, P., 2007. Cu(II) binding by dried biomass of red, green and brown macroalgae. Water Res. 41, 731740. n, J., Gonza lez, F., Ballester, A., Bla zquez, M.L., Mun Rinco oz, J.A., 2005. Biosorption of heavy metals by chemically activated algae Fucus vesiculosus. J. Chem. Technol. Biotechnol. 80, 14031407.

Author's personal copy

E. Romera et al. / Bioresource Technology 99 (2008) 46844693 lez, F., Ballester, A., Bla zquez, M.L., Mun Romera, E., Gonza oz, J.A., 2006. Biosorption with algae: A statistical review. Crit. Rev. Biotechnol. 26, 223235. lez, F., Ballester, A., Bla zquez, M.L., Mun Romera, E., Gonza oz, J.A., 2007. Comparative study of biosorption of heavy metals using dierent types of algae. Bioresour. Technol. 98, 33443353. Sag, Y., Akc ael, B., Kutsal, T., 2001. Evaluation, interpretation and representation of three-metal biosorption equilibria using a fungal biosorbent. Proc. Biochem. 37, 3550. Seheng, P.X., Ting, Y.P., Chem, J.P., Hong, L., 2004. Sorption of lead, copper, cadmium, zinc and nickel by marine algal biomass: characterization of biosorptive capacity and investigation of mechanisms. J. Coll. Interf. Sci. 275, 131141. Silva, E.A., Cossich, E.S., Tavares, C.G., Cardozo Filho, L., Guirardello, R., 2003. Biosorption of binary mixtures of Cr(III) and Cu(II) ions by Sargassum sp. Braz. J. Chem. Eng. 20, 122. Singh, S., Pradhan, S., Rai, L.C., 2000. Metal removal from single and multimetallic systems by dierent biosorbent materials as evaluated by

4693

dierential pulse anodic stripping voltammetry. Proc. Biochem. 36, 175182. Tsezos, M., 2001. Biosorption of metals, The experience accumulated and the outlook for technology development. Hydrometallurgy 59, 241243. Van den Hoek, C., Mann, D.G., Jahns, H.M., 1995. Algae. An Introduction to Phycology. Cambridge University Press, ISBN 0521-31687-1. Veglio, F., Beolchini, F., 1997. Removal of metals by biosorption: a review. Hydrometallurgy 44, 301316. Volesky, B., Holan, Z.R., 1995. Biosorption of heavy metals. Biotechnol. Prog. 11, 235250. Volesky, B., 1999. Biosorption for the next century. In: Amils, R., Ballester, A. (Eds.), Biohydrometallurgy and Environment Toward the Mining of the 21st Century; Part B. Elsevier, Amsterdam, pp. 161170. Yu, Q., Kaewsarn, P., 1999. Binary adsorption of copper(II) and cadmium(II) from aqueous solutions by biomass of marine alga Durvillaea potatorum. Separation science and technology 34, 1595 1605.