Adsorption of copper and zinc on pseudomonas putida CZ1: Particle concentration effect and adsorption reversibility

Colloids and Surfaces B: Biointerfaces 54 (2007) 46–52

Adsorption of copper and zinc on pseudomonas putida CZ1: Particle concentration effect and adsorption reversibility XinCai Chen, WeiXiang Wu, JiYan Shi ∗ , XiangHua Xu, Hui Wang, YingXu Chen Department of Environmental Engineering, Zhejiang University, HangZhou 310029, China Received 28 April 2006; received in revised form 3 July 2006; accepted 18 July 2006 Available online 25 July 2006

Abstract The adsorption and desorption processes of Cu(II) and Zn(II) on the biomass of Pseudomonas putida CZ1 as a function of particle concentrations (Cp ) were studied. In a 0.01 M KNO3 solution, the Cu-biomass and Zn-biomass adsorption systems displayed a clear Cp effect. The overall adsorption 0.8971 isotherms under three Cp conditions could be described as a Freundlich-type Cp effect isotherm equation: Γ = 2.553Cp−0.7106 Ceq for Cu-biomass −0.8305 0.6504 Ceq for Zn-biomass system. The results of experiments, designed to eliminate several typical sources of experimental system, Γ = 2.412Cp artifact, agree with the prediction of the metastable-equilibrium adsorption theory. Results from laboratory equilibration studies also indicate that biomass-adsorbed Cu(II) or Zn(II) fractions may be comprised of both reversibly and strongly bound or resistant components. A computational method has been derived to allow prediction of the magnitude of the reversible and more strongly adsorbed Cu(II) or Zn(II) fractions from conventional isotherm data. This methodology provides an initial quantitative approximation of the strongly bound, resistant, biomass fractions while utilizing relatively simple experimental adsorption–desorption data. © 2006 Elsevier B.V. All rights reserved. Keywords: Adsorption; Desorption; Particle concentration; Reversibility; Hysteresis

1. Introduction The extent of heavy metal sorption is a critical component in the evaluation of fate in natural water and soil systems. In studies of adsorption in natural waters, an anomalous phenomenon, “the particle concentration (Cp ) effect”, i.e. the decline of the adsorption isotherm or the decrease of the partition coefficient with the increasing of Cp , has been observed widely over the last two decades [1,2]. This phenomenon cannot be explained by classical thermodynamic theory of adsorption, but many laboratory studies on heavy metal or organic chemical adsorption to soils, natural sediments, clays and hydrous metal oxides have observed the Cp effect [3–5]. Before 1998, most researchers attributed the Cp effect to a variety of experimental artifacts, such as the presence of the third phase, e.g. dissolved organic matter (DOM) phase or “non-settling” colloid phase; or particle–particle interactions, or implicit adsorbate competition [2,6–9]. However,

∗

Corresponding author. Tel.: +86 571 869 71975; fax: +86 571 869 71898. E-mail address: [email protected] (J. Shi).

0927-7765/$ – see front matter © 2006 Elsevier B.V. All rights reserved. doi:10.1016/j.colsurfb.2006.07.009

there are still other observations of the Cp effect that cannot be accounted for by experimental artifacts [10]. In 1998, Pan and Liss proposed that the Cp effect has fundamental reasons besides some experimental artifacts [11,12]. In metastableequilibrium adsorption (MEA) theory, it is proposed that a fundamental deficiency in the theoretical foundation of adsorption thermodynamics that adsorption density Γ had been incorrectly used as a thermodynamic state variable in the past. The MEA theory can give a reasonable explanation to the Cp effect [13]. At the same time, less well understood is the desorption reaction. The conventional desorption tests for heavy metals are extractions using concentrated solutions [14–16]. This is in contrast to the usual desorption tests for organic chemicals in which the same aqueous phase is used for both adsorption and desorption to directly test reversibility. Heavy metal desorption into various extraction solutions has been found to be incomplete, suggesting that the reaction is not completely reversible [17]. The usual finding is that substantial quantities of the heavy metals remain associated with the particle even at extractant concentrations (∼0.1 M) that should displace all the physically adsorbed metal [18]. The desorption literature contains almost

X. Chen et al. / Colloids and Surfaces B: Biointerfaces 54 (2007) 46–52

no data that include adsorption and desorption isotherms using the same aqueous phase [16–26]. Although many examples of these isotherms with significant nonreversibility exist in the literature for various adsorbate–absorbent systems: aldicarb, terbufos-soil [27]; PCB (Aroclor 1254)-estuarine sediment [28]; atrazine-soil [29]; Fluometuron-fine sandy loam [30]; HCBP-sediment [31]; nickel (cobalt)-montmorillonite (quartz) [32]; Cu, Zn and Cd-goethite [1,12]; Zn-manganite [33], it was first shown for copper and zinc-biomass (Pseudomonas putida CZ1) systems to analyze adsorption–desorption isotherm data and examine reversible component sorption as a function of particle concentration. In this paper, copper and zinc P. putida CZ1 sorption data are presented below which demonstrate that reversibility is not complete and varies as a function of physical (particle concentration) variables. The MEA theory will be tested by verifying one of its inferences, namely the control of adsorption reversibility on the Cp effect. Experimentally, we used the change of adsorption hysteresis to approximate a change in adsorption reversibility. A method for analyzing adsorption–consecutive desorption isotherms, which yields the partition coefficients of the reversible and resistant components is applied in order to isolate the behavior of reversibly sorbed heavy metal as a function of particle concentration. 2. Materials and methods 2.1. Microorganism and its preparation for biosorption The bacterial strain used in the present study was P. putida CZ1, isolated from metal-contaminated soil of mining activities as a copper and zinc tolerant strain. The production of this strain has been described previously [34]. After cultivated, cells were harvested by means of centrifugation at 15,000 × g for 5 min and washed three times with phosphate buffer solution (PBS) (pH 7.0) and deionized water, respectively. The harvested cells were freeze-dried, autoclaved at 121 ◦ C for 30 min, crushed in a blender for further experiments. 2.2. Adsorption and desorption experiments An adsorption isotherm under a specified Cp condition was produced by preparing a series of initial concentrations of Cu(II) and Zn(II) in a series of centrifuge tubes with each tube containing the same concentration of biomass. The total volume of each tube was made to 10 ml by adding 0.01 M KNO3 solution. Adsorption and desorption suspensions were generated by using 0.01 M KNO3 to maintain constant ionic strength. 0.1 M NaOH or 0.1 M HNO3 were used to control a constant pH of 4.5 for Cu(II) and 5.0 for Zn(II) during the course of the experiments. After this, the tubes were capped and shaken at 30 ◦ C for 48 h. After 48 h of equilibration, solids were removed by centrifugation and dissolved metal concentrations were determined by flame atomic absorption spectrometry (AAS, Thermo Element MKII-M6). Desorption experiments were conducted at 30 ◦ C by removing the equilibrium supernatant solution after centrifugation,

47

replacing the solution with equal volume of 0.01 M KNO3 background solution, adjusting pH to 4.5 for Cu(II) and 5.0 for Zn(II) and then shaking the mixture for 48 h. The other analytical procedures were the same as described above. The desorption process was repeated in this way three times so that a consecutive desorption isotherm was obtained. 2.3. Theory The basis of the MEA theory is that adsorption density Γ (moles or mass of adsorbate per unit surface area) is not a thermodynamic state variable. Adsorption isotherms are therefore fundamentally dependent on the metastable-equilibrium state of the adsorbate. Based on this concept, the MEA inequality can be used to reformulate some of the existing equilibrium isotherm equations into metastable-equilibrium isotherm equations. One example is the Freundlich-type MEA isotherm equation, where α is a constant under isothermal conditions [11]: Γ = αKme Cβ

(1)

Since changes in particle concentration can affect the rate of adsorption, which, in turn, could affect the MEA state and adsorption reversibility, we assume [7,11,12,35]: Kme = γCp−n

(2)

where γ is a constant and n is an empirical parameter, n ≥ 0. A semi-empirical Freundilich-type Cp effect isotherm equation can be obtained [11]: β Γ = ksp Cp−n Ceq

(3)

We call ksp the specific adsorption constant, which is independent of the Ceq and Cp conditions. n is called the Cp effect index, which is the measure of the degree of the Cp effect. 3. Results 3.1. The Cp effect adsorption isotherm The adsorption and desorption isotherms under three different Cp conditions are shown in Fig. 1. As Cp increase in the Cu-biomass and Zn-biomass systems, the adsorption isotherm (solid line only) clearly decline and the degree of hysteresis (the angle between the dotted and solid lines for the same Cp condition, see the inset in Fig. 1) increases as Cp increases. However, the changes in the degree of hysteresis are not very distinct for all adsorption–desorption systems. In the Freundlich-type Cp effect isotherm equation (Γ = β ksp Cp−n Ceq ), like β, ksp and n can be calculated from adsorption isotherm data. Based on the adsorption isotherm data of Fig. 1, the plots of log Γ versus log Ceq and the plots of log Γ versus log Cp for Cu-biomass and Zn-biomass systems are presented in Figs. 2 and 3, respectively. Good linear relationships were obtained. For the Cu-biomass system, ksp = 2.553, β = 0.8971, n = 0.7106. For the Zn-biomass system, ksp = 2.412, β = 0.6504, n = 0.8305. After the specific adsorption constant (ksp ), the Cp effect index (n), and β are calculated, the Cp effect

48

X. Chen et al. / Colloids and Surfaces B: Biointerfaces 54 (2007) 46–52

Fig. 1. Adsorption (solid lines, filled symbols) and desorption (dotted lines, open symbols) isotherms under different Cp conditions in the Cu-biomass and Zn-biomass systems.

adsorption isotherm equations for Cu-biomass and Zn-biomass 0.8971 and systems can be expressed as Γ = 2.553Cp−0.7106 Ceq −0.8305 0.6504 Γ = 2.412Cp Ceq , respectively. The comparisons between calculated and measured isotherms under different Cp conditions in the Cu-biomass and Zn-biomass systems were shown in Fig. 4. It can be seen that the calculated isotherms fitting the experimental data well for the Zn-biomass system under different Cp conditions, however, which only under low Cp and equilibrium concentration conditions for Cu-biomass system. In other words, the adsorption of Zn-biomass system can be described well by the Freundlich-type Cp effect isotherm equation but not for the Cu-biomass system. 3.2. The Cp –reversibility relationship Fig. 2. Plots of log qeq vs. log Ceq under the condition of Cp = 2.0 g/l for the Cu-biomass and Zn-biomass systems, using the data in Fig. 1.

Adsorption and desorption isotherms corresponding to the forward and backward reactions were not coincide, indicating that adsorptions of Cu(II) and Zn(II) onto the biomass of P. putida CZ1 were not reversible. The degree of hysteresis (the angle α between an adsorption and its corresponding desorption isotherm) was used to approximate a change in adsorption reversibility [12,36]. Fig. 5 shows the angle α between an adsorption isotherm (solid line) and its corresponding desorption isotherm (dotted line). We used a trigonometric formula [Eq. (4)] to describe quantitatively the degree of adsorption hysteresis. The detailed explanations were obtained as follows: cos(α) =

Fig. 3. Plots of log qeq vs. log Cp under the condition of Ceq = 4 mg/l for the Cu-biomass system and Ceq = 15 mg/l for the Zn-biomass system, using the data in Fig. 1.

a2 + c 2 − b 2 2ac

(4)

where a is the length of the linear desorption line, b is the intercept of the linear desorption line, and c is the length of the linear adsorption line. In order to study the influence of Cp on the degree of adsorption hysteresis, we selected samples 1–3 under three Cp isotherms which adsorption density (Γ av =10.2 mg Zn/g biomass) was similar while their equilibrium concentrations of Zn increased with the increasing of Cp (Fig. 6, samples 1–3

X. Chen et al. / Colloids and Surfaces B: Biointerfaces 54 (2007) 46–52

49

Fig. 4. Comparison between calculated and measured isotherms under different Cp conditions in the Cu-biomass and Zn-biomass systems.

biomass adsorption system, the degree of hysteresis increased from 0.36◦ to 10.21◦ (indicating high irreversibility in the system). 3.3. Relationship to consecutive desorption isotherm

Fig. 5. Schematic representation of the degree of hysteresis between the adsorption and desorption isotherm.

were labeled with arrows). By affecting MEA state or adsorption reversibility, Cp can fundamentally influence adsorption isotherms. According to Eq. (4), we could calculate the degree of hysteresis of samples 1–3, respectively. Table 1 shows the degree of hysteresis as a function of Cp . As Cp increased in the Zn-

Fig. 6. Adsorption (closed symbols, solid curve) and desorption (open symbols, dotted curve) isotherm of Zn(II) on biomass.

The results of experimental consecutive desorption studies indicate that there exists a significant component of the adsorbed Cu and Zn that is extremely difficult to desorb. To idealize the situation as shown in Fig. 7, the adsorption isotherm is assumed to be linear as is the initial stages of the consecutive desorption isotherm, which is presumed to describe the behavior of the readily reversible fraction. This idealization is important since it limits the applicability of the analysis to those aqueous concentrations for which the consecutive desorption data conform to the linear assumption [31]. If this idealization is a useful characterization, then the reversible component should follow an isotherm that does not distinguish between either the adsorption or the desorption points but is related only to the aqueous concentration. That is

Fig. 7. Schematic illustration of the definitions of the resistant, r0 , and the reversible components of adsorption, rxa , and desorption, rxd , with the assumption of a linear consecutive desorption isotherm.

50

X. Chen et al. / Colloids and Surfaces B: Biointerfaces 54 (2007) 46–52

Table 1 Zn-biomass adsorption conditions and the degree of hysteresis for samples (pH 5.0, 0.01 M KNO3 medium) Sample

Cp (g/l)

[Zn] initial C0 (mg/l)

[Zn] final Ceq (mg/l)

Adsorption density, Γ (mg/g)

Degree of hysteresis, α (◦ )

1 2 3

0.5 1.0 2.0

9.75 18.90 41.68

4.42 8.98 21.60

10.65 9.91 10.04

0.36 7.86 10.21

the reversible component isotherm should not display any hysteresis. Fig. 8c, which contain the data in Fig. 1 (and replotted in Fig. 8a) analyzed under the idealized situation, show that indeed all the reversible adsorption and desorption points are described

by a single, in this case linear, isotherm. Fig. 8b illustrates that for these data the resistant component is a linear function of the equilibrium adsorption aqueous concentration, ca , leading to an isotherm for the resistant component.

Fig. 8. Cu and Zn adsorption–desorption isotherms, biomass of P. putida CZ1 (Cp = 2.0 g/l); (a) adsorption and single desorption data and linear isotherms; (b) resistant component estimates and linear isotherm; (c) reversible component linear isotherm, adsorption and desorption estimates of the reversible component.

X. Chen et al. / Colloids and Surfaces B: Biointerfaces 54 (2007) 46–52

51

Table 2 Isotherm parameters for Cu-biomass and Zn-biomass adsorption–desorptiona Particle concentration, Cp (g/l)

0.5 1.0 2.0 a

Adsorption and desorption time (h)

48 48 48

Partition coefficients (l/g) πa

πd

π0

πx

Cu

Zn

Cu

Zn

Cu

Zn

Cu

Zn

2.788 2.676 1.263

1.575 0.915 0.514

22.551 15.857 6.918

2.992 2.481 1.700

2.534 2.481 1.163

0.622 0.557 0.361

0.254 0.189 0.100

0.961 0.361 0.153

The aqueous phase is 0.01 M KNO3 solution for all experiments; temperature is 30 ◦ C.

Since the adsorption isotherm is also linear, the resistant component can also be expressed in terms of the adsorbed concentration. The experimental conditions and isotherm parameters for data are listed in Table 2. As Cp increases in the Cu-biomass and Zn-biomass adsorption systems, the partition coefficients decrease, which indicating that there exist significant particle concentration effect in the adsorptions of Cu(II) and Zn(II) onto the biomass of P. putida CZ1.

4. Discussion According to MEA theory, the specific adsorption constant (ksp ) is 2.553 for Cu and 2.412 for Zn, the Cp effect index (n) is 0.7106 for Cu and 0.8305 for Zn, which is a measure of the degree of the solids concentration effect. The Cp effect isotherm equation introduced a new variable—Cp (solids concentration) into the traditional isotherm equation, and it could describe the relationship between three variables: Γ , Ceq and Cp . From this Freundlich-type Cp effect isotherm equation, we could infer that adsorption density Γ would decrease with increasing the solid adsorbent concentrations (Cp ). Though many potential artifacts can cause the so-called solids concentration effect, there are two measures to avoid these artifacts in our experiments. First we use a simple model adsorption system-biomass and Cu2+ or Zn2+ ions in 0.01 M KNO3 mediums, which will preclude the dissolved organic matter (DOM). Second, the adsorption and desorption samples were shaken completely to minimize the particle–particle interactions. According to the hard-soft acid-base principle (HSAB), the hard Lewis acids prefer to complex or react with hard Lewis bases and soft acids prefer to complex or react with soft bases. ˚ Cu2+ and Zn2+ are Lewis acids, and the size of Cu2+ (0.73 A) ˚ so the chemical bond is approximate to that of Zn2+ (0.74 A), between Cu and biomass of P. putida CZ1 should be as much as that between Zn and CZ1 biomass. However, from Figs. 1 and 4 we can see that the degree of irreversibility and the fitting between calculated and measured isotherms in the Cu-biomass and Zn-biomass systems are different. In other words, although copper and zinc have the similar chemical properties, their behavior of adsorption and desorption on biomass of CZ1 is different. Pan et al. also indicate that the dependence of the adsorption isotherms on the particle concentration is most obvious in the Zn-goethite system, less obvious in the Cu-goethite system [1].

The computational method allow us to predict the magnitude of the reversible and more strongly adsorbed Cu or Zn fractions from conventional isotherm data. The predictions agree with the results of MEA theory, indicating that the decline of the adsorption isotherm or the decrease of the partition coefficient with the increasing of Cp . The partition coefficient for adsorption including two parts: the reversible partition coefficient and the resistant partition coefficient. From Table 2 we can see that about 91% resistant components could not be desorbed for Cu-biomass system, much higher than that of Zn-biomass system (60–70%), especially only about 40% under the condition of Cp = 0.5 g/l. It was also indicated in Fig. 1 that the degree of hysteresis of Cu-biomass system is distinctly higher than that of Zn-biomass system. The proposed linear approximation describing the binding of metal to cells is clearly an oversimplification of the actual process. In the case of Zn, experimental evidence suggests that under certain chemical conditions the binding to cells during consecutive desorption may be described by a curvilinear isotherm that may or may not ultimately demonstrate complete desorbability. It is also quite possible that the actual adsorption process may involve binding to sites of a gradation of energies rather than the two arbitrarily defined fractions (reversible and resistant). Nevertheless, there exist distinct advantages for data analysis that result from treating metal and other organic adsorption study results in terms of this approximation [31]. From the standpoint of environmental modeling the linear approximation offers a means of refining currently used assumptions such as complete reversibility for metal adsorption to biomass. From the environmental chemical perspective the approximation can be utilized to separate reaction variables and provide an aid in interpreting metal binding mechanisms. These applications may be demonstrated by a consideration of the results for both multiple cycle adsorption–desorption and kinetic studies. Acknowledgements This work was funded by the National Key Natural Science Foundation of China (40432004), the Natural Science Foundation of Zhejiang Province (Y504109). References [1] G. Pan, P.S. Liss, M.D. Krom, Colloid Surf. A-Physicochem. Eng. Asp. 151 (1999) 127.

52 [2] [3] [4] [5] [6] [7] [8] [9] [10] [11] [12] [13] [14] [15]

[16] [17] [18] [19]

X. Chen et al. / Colloids and Surfaces B: Biointerfaces 54 (2007) 46–52 T.C. Voice, W.J. Weber, Envrion. Sci. Technol. 19 (1985) 789. D.J. O’ Connor, J.P. Connolly, Water Res. 14 (1980) 1517. T.C. Voice, W.J. Weber, Water Res. 17 (1983) 1433. L. Cox, M.C. Hermosin, R. Celis, J. Cornejo, Water Res. 31 (1997) 1309. G. Benoit, Geochim. Cosmochim. Acta. 59 (1995) 2677. U.P. Nyffeler, Y.H. Li, P.H. Santschi, Geochim. Cosmochim. Acta. 48 (1984) 1513. R.L. Curl, G.A. Keioleian, Environ. Sci. Technol. 18 (1984) 916. J.J.W. Higgo, L.V.C. Rees, Environ. Sci. Technol. 20 (1986) 483. J.P. McKinley, E.A. Jenne, Environ. Sci. Technol. 25 (1991) 2082. G. Pan, P.S. Liss, J. Colloid Interface Sci. 201 (1998) 71. G. Pan, P.S. Liss, J. Colloid Interface Sci. 201 (1998) 77. G. Pan, M.D. Krom, B. Herut, Environ. Sci. Technol. 36 (2002) 3519. H. Farrah, W.F. Pickering, Water Air Soil Pollut. 9 (1978) 491. R.M. Engler, J.M. Brannon, J. Rose, G. Bigham, in: T.F. Yen (Ed.), Chemistry of Marine Sediments, Ann Arbor Science, Ann Arbor, MI, 1977. C.H. Van der Weijden, M.J.H.L. Arnoldus, C.J. Meurs, Neth J Res. 11 (1977) 130. W.F. Pickering, in: J.O. Nriagu (Ed.), Zinc in the Environment., Wiley–Interscience, New York, 1980, Part I. U. Forstner, G.T.W. Wittmann, Metal Pollution in the Aquatic Environment, Springer-Verlag, Berlin, New York, 1979. C.N. Murray, L. Murray, Radioactive Contamination of the Marine Envrionment, International Atomic Energy Agency, Vienna, 1973.

[20] D.P. Kharkar, K.K. Turekian, K.K. Bertine, Geochim. Cosmochim. Acta. 32 (1968) 285. [21] D.W. Evans, N.H. Cutshall, Radioactive Contamination of the Marine Envrionment, International Atomic Energy Agency, Vienna, 1973. [22] M.H. Bothner, R. Carpenter, Radioactive contamination of the Marine Envrionment, International Atomic Energy Agency, Vienna, 1973. [23] E.K. Duursma, in: J.D. Burton, P.S. Liss (Eds.), Estuarine Chemistry, Academic Press, New York, 1976. [24] N. Karimian, F.R. Cox, Soil Sci. Soc. Am. J. 42 (1978) 757. [25] B. Patel, S. Patel, S. Pawar, Estuarine Coastal Mar. Sci. 7 (1978) 49. [26] S.S. Singh, Can. J. Soil Sci. 59 (1979) 119. [27] A. Felsot, P.A. Dahm, J. Agri. Food Chem. 27 (1979) 557. [28] D.J. Wildish, C.D. Metcalfe, H.M. Akagi, D.W. McLeese, Bull. Environ. Contam. Toxicol. 24 (1980) 20. [29] R.A. Swanson, G.R. Dutt, Soil Sci. Soc. Am. Proc. 37 (1973) 872. [30] K.E. Savage, R.D. Wauchope, Weed Sci. 22 (1974) 106. [31] D.M. Di Toro, L.M. Horzempa, Envriron. Sci. Technol. 16 (1982) 594. [32] D.M. Di Toro, J.D. Mahony, P.R. Kirchgraber, A.L. O’Byrne, L.R. Pasquale, D.C. Picclrilll, Environ. Sci. Technol. 20 (1986) 55. [33] Y.W. Qin, G. Pan, M.M. Zhang, X.L. Li, J. Environ. Sci. (China) 16 (2004) 627. [34] X.C. Chen, Y.P. Wang, Q. Lin, J.Y. Shi, W.X. Wu, Y.X. Chen, Colloid Surf. B-Biointerfaces 46 (2005) 101. [35] D. Dzombak, F.M.N. Morel, J. Colloid Interface Sci. 112 (1986) 588. [36] K. Verburg, P. Baveye, Clay Clay Min. 42 (1994) 207.