Journal of Environmental Management 131 (2013) 222e227
Contents lists available at ScienceDirect
Journal of Environmental Management journal homepage: www.elsevier.com/locate/jenvman
Chitosan and alginate biopolymer membranes for remediation of contaminated water with herbicides Mariana Agostini de Moraes a, Daniela Sgarbi Cocenza b, Fernando da Cruz Vasconcellos a, Leonardo Fernandes Fraceto b, Marisa Masumi Beppu a, * a School of Chemical Engineering, University of Campinas e UNICAMP, Av. Albert Einstein, 500, Cidade Universitária Zeferino Vaz, 13083-852 Campinas, SP, Brazil b Department of Environmental Engineering, São Paulo State University e UNESP, Av. Três de Março, 511, Alto da Boa Vista, 18087-180 Sorocaba, SP, Brazil
a r t i c l e i n f o
a b s t r a c t
Article history: Received 13 March 2012 Received in revised form 11 September 2013 Accepted 23 September 2013 Available online 28 October 2013
This study investigated the adsorption behavior of the herbicides diquat, difenzoquat and clomazone on biopolymer membranes prepared with alginate and chitosan (pristine and multi-layer model) for contaminated water remediation applications. Herbicides, at concentrations ranging from 5 mM to 200 mM, were adsorbed in either pure alginate, pure chitosan or a bilayer membrane composed of chitosan/alginate. No adsorption of clomazone was observed on any of the membranes, probably due to lack of electrostatic interactions between the herbicide and the membranes. Diquat and difenzoquat were only adsorbed on the alginate and chitosan/alginate membranes, indicating that this adsorption takes place in the alginate layer. At a concentration of 50 mM, diquat adsorption reaches ca. 95% after 120 min on both the alginate and chitosan/alginate membranes. The adsorption of difenzoquat, at the same concentration, reaches ca. 62% after 120 min on pure alginate membranes and ca. 12% on chitosan/ alginate bilayer membranes. The adsorption isotherms for diquat and difenzoquat were further evaluated using the isotherm models proposed by Langmuir and by Freundlich, where the latter represented the best-fit model. Results indicate that adsorption occurs via coulombic interactions between the herbicides and alginate and is strongly related to the electrostatic charge, partition coefficients and dissociation constants of the herbicides. Biopolymer based membranes present novel systems for the removal of herbicides from contaminated water sources and hold great promise in the field of environmental science and engineering. Ó 2013 Elsevier Ltd. All rights reserved.
Keywords: Adsorption Alginate Chitosan Multilayer
1. Introduction Currently, there is a great concern for environmental preservation, where the contamination of water is one of the major problems of the degradation of the environment by human activity in the world. Environmental contamination by chemical products and herbicides is an eminent risk for nature preservation and human health in general. Herbicides become biologically active when applied to plantations, and are later transported to the soil and aquatic systems (Nunez et al., 2002). The excessive use of pesticides and herbicides has given rise to several short- and long-term adverse effects on human health such as in immune-suppression,
* Corresponding author. Tel.: þ55 19 3521 3893; fax: þ55 19 3521 3922. E-mail addresses: [email protected] (M. Agostini de Moraes), [email protected] (D.S. Cocenza), [email protected] (F. da Cruz Vasconcellos), [email protected] (L.F. Fraceto), [email protected] (M.M. Beppu). 0301-4797/$ e see front matter Ó 2013 Elsevier Ltd. All rights reserved. http://dx.doi.org/10.1016/j.jenvman.2013.09.028
hormone disruption, reproductive abnormalities and cancer diseases, among others (Singh et al., 2010). Three commonly used herbicides include Diquat (DQ), Difenzoquat (DF) and Clomazone (CLO). DQ is a non-selective contact herbicide, used in the control of weeds and grasses in plantation crops. DF is a selective herbicide used in cereal crops to control the post emergence of wild oats. CLO is used in rice crops to control the preemergence of mono and dicotyledonous plants (Gunasekara et al., 2009; Nunez et al., 2002). These compounds have different chemical structures (Fig. 1). In terms of polarity and number of charges, DQ is a very polar compound with two positive charges, while DF has just one positive charge and CLO does not have any charge, which will have an effect on the adsorption behavior of each compound. Table 1 shows values of water solubility, dissociation constant (pKa), octanolewater partition coefficient (log Kow) and molecular weight (Mw) for DQ, DF and CLO, taken from the pesticide properties database (PPDB-UK). The persistence of biological activity in moist and fertile soils in moderate temperature is one month or less for DQ, and 3e12
M. Agostini de Moraes et al. / Journal of Environmental Management 131 (2013) 222e227
223
Fig. 1. Chemical structures of the herbicides: (a) Diquat, (b) Difenzoquat and (c) Clomazone.
months for DF and CLO (Senseman, 2007). While active, DQ, DF and CLO are harmful to human health and contaminate the environment, which may lead to several problems in various ecosystems. The World Health Organization (WHO, 2009) has classified DQ, DF and CLO, based on their toxicity, as moderately hazardous. The maximum contaminant level for DQ in drinking water is 0.02 mg L1, while no maximum limits have been established for DF and CLO (EPA, 2011). The European Union has not yet established maximum levels of individual herbicides in water; however the total amount of contaminants, including herbicides, permitted is up to 0.5 mg L1 (EU, 2007). Herbicides are traditionally removed from water through adsorption, microfiltration or oxidation processes (Montgomery, 2007); however the development of novel, efficient and low cost methods to remediate contaminated areas are needed. In this context, the use of membranes composed of natural polymers, such as chitosan (CS) and alginate (AG), appears to be an eco-friendly alternative for the adsorption of herbicides and remediation of contaminated aqueous sources. Chitosan is a linear polysaccharide obtained from the deacetylation of chitin, a major component of crustacean shells. CS has been studied as a suitable biopolymer for removal of metal ions from wastewater, since its amino and hydroxyl groups can act as chelating sites (Beppu et al., 2004; Vieira and Beppu, 2006). Alginate is a polysaccharide derived from seaweed and possesses a backbone comprised of two repeating carboxylated monosaccharide units (manuronic acid and guluronic acid), the ratio of which influences the physical properties of the biopolymer. AG has proved to be an efficient adsorbent for dies (Jeon et al., 2008) and metal ions, such as copper and arsenic (V) (Lim and Chen, 2007). CS and AG are obtained from renewable sources that are biodegradable and non-toxic and are potential materials for the production of multilayer membranes, using the electrostatic forces between the polyanions of the carboxylate groups of AG and the polycations of the protonated amines of CS (Lawrie et al., 2007). These membranes present binding sites for removal of different herbicides displaying net positive or negative charges. Few studies in the literature report the use of biopolymers (as membranes, fibers or nanoparticles) for adsorption of herbicides (Sadasivam et al., 2011; Silva et al., 2011). Particularly, our previous study on CS and AG membranes for adsorption of Paraquat was the first one to use a bilayer membrane for herbicide removal and it
Table 1 Physicochemical properties of the herbicides DQ, DF and CLO. Herbicide
Water solubility (g L-1) pKa
Diquat 718 Difenzoquat 765 Clomazone 1.1
log Kow Mw (g mol-1)
Totally dissociated 4.60 7.0 0.61 No dissociation 2.54
*Values taken for the pesticide properties database, UK.
344.05 360.43 239.70
indicated a strong dependency between herbicide adsorption and the biopolymer properties (Cocenza et al., 2012). However, no studies have compared the adsorption capacity of various herbicides, with different net charges and physicochemical properties, on biopolymer membranes. The present study investigated the adsorption behaviors of the Diquat, Difenzoquat and Clomazone herbicides in membranes prepared with natural polymers (chitosan and alginate), in order to evaluate the use of these membranes as potential materials for remediation of contaminated aquatic systems. 2. Material and methods 2.1. Materials The herbicides Diquat PESTANALÒ, Difenzoquat PESTANALÒ, Clomazone PESTANALÒ and chitosan (extracted from crab shells, minimum deacetylation degree of 85%) were purchased from Sigma Chem. Co. Sodium alginate was purchased from Vetec Química Fina (Brazil). 2.2. Membrane preparation and characterization CS and AG were used as polyelectrolytes in concentrations of 1% (w/v). CS was dissolved in 2% vol acetic acid solution and AG was dissolved in 0.1 M NaOH aqueous solution. CS solution pH was ca. 3 and AG solution pH was ca. 10. The membranes were prepared according to the procedure proposed by Moon et al. (1999), with modifications. First, the CS solution was cast onto polystyrene dishes and dried at 50 C, for ca. 2 h (70% dried). The AG solution was then cast onto the CS film and completely dried at 50 C, forming a bilayer membrane (CS/AG). When completely dried, the membranes were immersed in a 0.1 M sulfuric acid solution in 50% vol ethanol, for 1 h to neutralize the functional groups of CS and AG. Membranes of the pure components (pure AG and pure CS) were also prepared in order to compare their characteristics and properties with those of the bilayer membrane. All membranes were prepared using the same total mass of biopolymer, i.e. the CS/AG membranes were produced with 50% of the biopolymer mass utilized in either the pure AG or CS membranes giving a total mass contribution equal to that used for the production of the pure membranes. The morphology of the membranes was observed by scanning electron microscopy (SEM), using a LEO 440i scanning electron microscope, after being cryo-fractured and freeze-dried for 24 h, at 54 C under vacuum. The membrane’s thickness was measured using a digital micrometer (MDC-25S, Mitutoyo), prior to (dry state) and after (wet state) the swelling test. For the swelling test, the membranes were cut to 2.5 cm in diameter and stored for 48 h in an environment with 50% relative humidity. The samples were weighed in the
224
M. Agostini de Moraes et al. / Journal of Environmental Management 131 (2013) 222e227
dry state (mi), immersed in 100 mL of distilled water and weighed after several time intervals, until reaching equilibrium (mf). The swelling capacity (g of water/g of sample) was calculated as Swelling ¼ ðmf mi Þ=mi . Fourier transform infrared spectroscopy with attenuated total reflectance (FTIR-ATR) was performed using a Bomem MB 102 spectroscope, prior and after the herbicides adsorption tests, in order to evaluate possible interaction effects between the CS and AG functional groups with those of the herbicides. 2.3. Adsorption tests The CS, AG and CS/AG membranes (with 4 cm2 of area) were immersed in aqueous solutions with the herbicides in several concentrations (5 mM, 25 mM, 50 mM, 100 mM and 200 mM) under continuous shaking (100 rpm) for 4 h at 25 C. The pH of the DQ, DF and CLO solutions were 7.3, 6.1 and 6.0, respectively. The experiments were carried out in duplicate for each one of the herbicides in each concentration. Aliquots of the solution were taken at 0, 15, 30, 45, 60, 120, 180 and 240 min and analyzed in a UVeVIS spectrophotometer (Varian e Cary 50) in the absorption wavelength range of each herbicide (DQ ¼ 308 nm; DF ¼ 257 nm; CLO ¼ 210 nm). Adsorption isotherms were obtained, in triplicate (n ¼ 3), from the changes observed in the absorbance of the solutions as a function of time, and mass balance calculations were used to determine the adsorption capacity of the membranes. The same experimental procedure was performed in the absence of the membranes (blank) in order to verify whether the herbicides had undergone any degradation throughout the experiment. The adsorption of herbicides may be described using an adsorption isotherm model. The isotherm represents the relationship between the amount of adsorbed herbicide per unit of sorbent, and the equilibrium concentration. Langmuir (Equation (1)) and Freundlich (Equation (2)) are the most frequently used models to describe herbicide adsorption:
Ce 1 Ce þ ¼ ðQ0 $Kl Þ Q0 qe lnðqe Þ ¼ ln Kf þ N lnðCe Þ
(1)
(2)
where Ce (mmol L1) is the equilibrium adsorbate concentration in the aqueous phase; qe (mmol g1) is the equilibrium adsorbate concentration in the solid phase; Q0 (mmol g1) is the maximum adsorption capacity and Kl (L mmol1) is the Langmuir constant and shows the adsorption intensity; Kf (L g1) is the Freundlich equilibrium constant and is an indicative of adsorption capacity and N (dimensionless) is the adsorption constant whose reciprocal is indicative of adsorption intensity. 3. Results and discussion 3.1. Membrane characterization The typical morphology of the CS/AG membranes, with a homogeneous and uniform cross section, characteristics of dense polymeric materials were obtained by scanning electron microscopy (SEM) (Fig. 2). The membranes present a different thickness for each biopolymer layer, even though the same mass for both biopolymers was used in their preparation. The CS layer (w7 mm), at the bottom, is thicker than the AG layer (w4 mm), at the top. This occurs due to the difference between the solutions densities: rCS > rAG. The CS/AG bilayer is well defined, signifying that the biopolymers interact strongly once a cohesive membrane is
Fig. 2. SEM micrograph of the cross section of the CS/AG bilayer membrane.
obtained and no delamination layers are observed. During the membrane preparation, CS (pK ¼ 6.5) and AG (pK ¼ 3.0) solutions had pHs of approximately 3 and 10, respectively. At these pH values, CS amino functional groups are fully protonated (NH3 þ ) and can easily interact with the AG carboxyl groups that are fully deprotonated (COO), at the membrane interface. Table 2 presents the results for the membrane swelling and thickness ratio in the dry and wet states. Swelling of the membranes occurred mainly within the first 30 s after their immersion in water (data not shown). The AG swelling was approximately two times higher than the CS swelling and the same proportionality is observed for the ratio between the dry and wet thicknesses of each biopolymer. These results indicate that the AG layer retains a higher quantity of water in the AG molecular chains and this water retention capacity could also be an indication of better adsorption properties. The swelling is related to chain mobility and crosslinking degree (i.e. an increase in crosslinking reduces swelling capacity) (Furuyama Lima et al., 2007). The post-treatment of membranes with sulfuric acid diluted in ethanol leads to the stabilization of AG membrane in water, even though its chain mobility is not lost. Traditional AG crosslinking in CaCl2 results in a swelling capacity of approximately 0.80 g of water/g of sample (RemunanLopez and Bodmeier, 1997), significantly lower than the results obtained in this study. 3.2. Adsorption results The analysis of the UVeVIS spectra of the herbicide solutions (blank) showed that there were no changes in solution concentrations with time, indicating that the herbicides were not adsorbed onto the glassware, and did not undergo any decomposition during the period of the experiment. Adsorption isotherms comparing adsorption of DQ, DF and CLO, each at a concentration of 50 mM, in the AG, CS and CS/AG membranes are shown in Fig. 3. The average experimental error for the Table 2 Swelling and thickness of the AG, CS and CS/AG membranes.
AG CS CS/AG a b
Swellinga (g of water/g of sample)
Initial thicknessb (ti) (mm)
Final thicknessb (tf) (mm)
Ratioa (tf/ti)
1.58 0.12 0.72 0.16 1.75 0.14
0.027 0.001 0.038 0.001 0.026 0.003
0.075 0.002 0.049 0.001 0.056 0.001
2.778 0.289 1.289 0.066 2.221 0.342
Average average deviation (n ¼ 3). Average average deviation (n ¼ 5).
M. Agostini de Moraes et al. / Journal of Environmental Management 131 (2013) 222e227
225
Fig. 3. Adsorption of 50 mM DQ, DF and CLO in the AG membrane (a), CS membrane (b) and CS/AG bilayer membrane (c).
isotherms is 2%. For the AG membranes (Fig. 3a), the adsorption equilibrium was reached within 120 min for DQ and DF, although with different adsorption percentages, where the adsorption was approximately 54% higher for the herbicide that had the most positive charges in its structure (DQ). No adsorption was observed in the experiment with CLO. In fact, CLO was not adsorbed on either the AG or CS membranes. This is probably due to the absence of charges in the CLO chains. This indicates that electrostatic interactions between the herbicides and the biopolymer chains are the main forces responsible for the adsorption. Due to the low CLO adsorption results observed on the membranes, further adsorption and characterization studies on this herbicide were not conducted. Additionally, pure CS membranes did not adsorb the herbicides studied (Fig. 3b), probably due to the fact that the herbicides DQ and DF present positive charges, as do the amino groups of the CS. Thus, no electrostatic attraction is achieved between the herbicides and the CS polymer chains, as reflected by the absence of adsorption. The CS/AG membranes (Fig. 3c) exhibited similar behaviors to the pure AG membrane with regard to DQ herbicide adsorption; however the kinetics of incorporation were slower in the former case. The incorporation of DF in CS/AG membranes was much lower than that observed for the pure AG membrane, indicating that saturation of the AG sites available for the adsorption had occurred. Fig. 4 shows the adsorption profiles of DQ at several concentrations of the herbicide in AG and CS/AG membranes (average error ¼ 3%). These results demonstrate that the DQ herbicide has a better adsorption on the AG membrane than on the CS/AG membrane and that, even at high concentrations, such as 200 mM,
the saturation of the membrane had not yet been completely achieved. The CS/AG membrane was also capable of adsorbing a high quantity of the DQ herbicide. Fig. 3b shows that CS does not adsorb DQ, signifying that all the DQ adsorption on the CS/AG membrane takes place in the AG layer. The CS/AG membrane has 50% of the AG mass present in the pure AG membrane, and is still capable of providing a fast adsorption of DQ, since within 120 min ca. 90% of the DQ herbicide, at concentrations in the range of 50 mMe200 mM, were already adsorbed. The good incorporation of DQ into the AG or CS/AG membranes is probably due to the negative charges of the carboxyl groups present in the AG chains, which favors their electrostatic attraction to the positive charges of the DQ molecules. For DF (data not shown), saturation of the AG membrane was observed even at small concentrations of herbicide. The adsorption capacity of the AG membrane for DF decreased with increasing DF concentrations. The adsorption capacity of the herbicides in the membranes can also be related to the physicochemical properties of the herbicides. For example, for the AG membrane, the adsorption capacity follows the order: DQ > DF > CLO (Fig. 3). As displayed by Table 1, the partition coefficient (log Kow) and the dissociation constant (pKa) parameters influence the herbicide adsorption capacity on AG membranes. An herbicide with a higher dissociation constant will have a structure with more ionisable groups that may interact with the AG membrane. In solution, DQ is totally dissociated, while DF displays pKa of approximately 7 and CLO does not undergo dissociation. Regarding the partition coefficient, DQ and DF can be considered hydrophilic, while CLO is hydrophobic. The lower the
Fig. 4. Adsorption profile at several DQ concentrations in the AG membrane (a) and in the CS/AG membrane (b).
226
M. Agostini de Moraes et al. / Journal of Environmental Management 131 (2013) 222e227
Table 3 Adjustment parameters of the Langmuir and Freundlich models for DQ and DF adsorption in AG and CS/AG membranes. Langmuir
Freundlich
Q0 (mmol g1) Kl (L mmol1) R
N (dimensionless) Kf (L g1) R
Diquat AG e CS/AG e Difenzoquat AG 34.19 CS/AG 7.80
e e
e e
1.12 1.01
4.49 2.69
0.99 0.88
35.76 8.23
0.89 0.22 0.91 0.64
3.34 0.81
0.99 0.90
partition coefficient, the more hydrophilic the herbicide is, leading to better adsorption on the AG biopolymer membrane. The solubility of the herbicides in water does not seem to be a key factor for the adsorption capacity. Although DQ and DF have similar solubility, DQ adsorption capacity was much higher that DF. In terms of molecular weight, all herbicides display similar values, and this parameter is not expected to significantly influence adsorption capacities. The Langmuir and linear Freundlich adsorption isotherm models were applied to the experimental data of DQ and DF adsorption in the membranes, as shown in Table 3 and Fig. 5 (AG membrane). The correlation coefficient obtained by the models showed that the Freundlich model fitted the data best (Table 3). These results can be explained by the fact that the Langmuir model normally does not fit data well for heterogeneous surfaces. The isotherm shape provides information on the strength by which the sorbate is held to the membrane, and also allows the determination of the maximum adsorption capacity of a sorbent (Echeverria et al., 1998). Experimental adsorption data for DQ fitted best in the linear Freundlich model and is a C-shaped type or a linear isotherm (N ¼ 1). This type of isotherm suggests that the solution has a high affinity with the sorbent. Interestingly, for DF, the Freundlich isotherm model fits with a good correlation coefficient; however, the isotherm is L shaped (N < 1), which indicates that DF has a moderate affinity for the membrane during the initial stage of adsorption, but there is decrease in the interaction when the adsorption sites begin to be filled (Echeverria et al., 1998). Additionally, according to the classification given by Giles and coworkers (Giles et al., 1960), the isotherms for DQ and DF have the L2 shape, which usually indicates that the molecules adsorb flat on the surface, or that vertically oriented ions with strong intermolecular attraction are adsorbed.
Fig. 6. FTIR-ATR spectra of (a) pure AG membrane and AG membrane with (b) DF and (c) DQ incorporated.
In order to verify whether any chemical interaction occurred between the AG layer and the herbicides, FTIR analysis was performed. Fig. 6 shows the FTIR spectra of pure AG membranes before and after DQ and DF adsorptions. For AG membranes, the main absorption bands observed at 1635 and 1409 cm1 are the C¼O antisymmetric and a symmetric stretch of the carboxylate group. The band at 1240 cm1 is the skeletal vibration of alginate and a sharp band at 1080-1020 cm1 is related to the antisymmetric stretch Ce OeC (Lawrie et al., 2007). No change was observed in the spectrum of the AG membrane with DF, when compared to that of the pure AG membrane. This is probably due to the weak interaction between DF and the carboxylate groups of AG. However, two new bands are observed at 1500 and 1284 cm1 in the AG membrane after DQ adsorption. These bands are related to characteristic bands of DQ, associated with ring stretching vibration and inter-ring stretching mode (Brienne et al., 1995). Future studies may further investigate the mechanisms for DQ and DF adsorption on AG and CS/AG membranes and relate them to the characteristic group vibrations. 4. Conclusions The results of this study provide new perspectives for using biopolymeric membranes (particularly AG and CS/AG) for remediation of areas contaminated with herbicides. AG proved to be an efficient biopolymer for the adsorption of DQ and DF, mainly due to the possible coulombic interaction between the carboxyl groups of AG and the positive charges of these herbicides. The dissociation constants and the partition coefficients of the herbicides also seem to be a key factor determining the adsorption capacity in the membrane, since higher dissociation constants and lower partition coefficients resulted in higher adsorption. The CS/AG membranes did not show the best result; however, this type of membrane may be of interest for the adsorption of different herbicides in each layer of the membrane, e.g., a positively charged herbicide can be adsorbed on the AG layer and a negatively charged herbicide can simultaneously be adsorbed in the CS layer, while also being a very effective adsorbent for DQ. In addition, the pH can be an important parameter on determining the adsorption behavior of the studied herbicides. The investigation of adsorption mechanisms at different pH values is subject of future work. Acknowledgments
Fig. 5. Linear Freundlich isotherm for adsorption of DQ and DF in AG membranes.
Financial support was provided by São Paulo Research Foundation (FAPESP), CNPq and Fundunesp.
M. Agostini de Moraes et al. / Journal of Environmental Management 131 (2013) 222e227
References Beppu, M.M., Arruda, E.J., Vieira, R.S., Santos, N.N., 2004. Adsorption of Cu(II) on porous chitosan membranes functionalized with histidine. J. Membr. Sci. 240 (1e2), 227e235. Brienne, S.H.R., Boyd, P.D.W., Schwerdtfeger, P., Bowmaker, G.A., Cooney, R.P., 1995. Infrared-spectra and theoretical-studies of bridged diquaternary 2,2’-dipyridinium compounds. J. Mol. Struct. 356 (2), 81e94. Cocenza, D.S., de Moraes, M.A., Beppu, M.M., Fraceto, L.F., 2012. Use of biopolymeric membranes for adsorption of Paraquat herbicide from water. Water Air Soil Pollut. 223 (6), 3093e3104. Echeverria, J.C., Morera, M.T., Mazkiaran, C., Garrido, J.J., 1998. Competitive sorption of heavy metal by soils. Isotherms and fractional factorial experiments. Environ. Pollut. 101 (2), 275e284. EPA, United States Environmental Protection Agency, 2011. Drinking Water Standards and Health Advisories. Washington. EU, 2007. European Communities Drinking Water Regulations (No. 2). SI 278, Dublin. Furuyama Lima, A.M., Andreani, L., Soldi, V., Borsali, R., 2007. Influence of plasticizer addition and crosslinking process on morphology, water absorption and mechanical properties of sodium alginate films. Quimica Nova 30 (4), 832e837. Giles, C.H., Macewan, T.H., Nakhwa, S.N., Smith, D., 1960. Studies in adsorption .11. A system of classification of solution adsorption isotherms, and its use in diagnosis of adsorption mechanisms and in measurement of specific surface areas of solids. J. Chem. Soc. (OCT), 3973e3993. Gunasekara, A.S., dela Cruz, I.D.P., Curtis, M.J., Claassen, V.P., Tjeerdema, R.S., 2009. The behavior of clomazone in the soil environment. Pest Manage. Sci. 65 (6), 711e716. Jeon, Y.S., Lei, J., Kim, J.H., 2008. Dye adsorption characteristics of alginate/polyaspartate hydrogels. J. Ind. Eng. Chem. 14 (6), 726e731.
227
Lawrie, G., Keen, I., Drew, B., Chandler-Temple, A., Rintoul, L., Fredericks, P., et al., 2007. Interactions between alginate and chitosan biopolymers characterized using FTIR and XPS. Biomacromolecules 8 (8), 2533e2541. Lim, S.F., Chen, J.P., 2007. Synthesis of an innovative calcium-alginate magnetic sorbent for removal of multiple contaminants. Appl. Surf. Sci. 253 (13), 5772e5775. Montgomery, J., 2007. Groundwater Chemicals Desk Reference, fourth ed. CRC Press. Moon, G.Y., Pal, R., Huang, R.Y.M., 1999. Novel two-ply composite membranes of chitosan and sodium alginate for the pervaporation dehydration of isopropanol and ethanol. J. Membr. Sci. 156 (1), 17e27. Nunez, O., Kim, J.B., Moyano, E., Galceran, M.T., Terabe, S., 2002. Analysis of the herbicides paraquat, diquat and difenzoquat in drinking water by micellar electrokinetic chromatography using sweeping and cation selective exhaustive injection. J. Chromatogr. A 961 (1), 65e75. RemunanLopez, C., Bodmeier, R., 1997. Mechanical, water uptake and permeability properties of crosslinked chitosan glutamate and alginate films. J. Controlled Release 44 (2e3), 215e225. Sadasivam, S., Kandasamy, K.S., Ponnusamy, K., Nagarajan, G.S., Kang, T.W., 2011. Film diffusion mechanism of methyl parathion in biopolymers: kinetics and thermodynamic equilibrium. J. Chem. Eng. Data 56 (11), 4024e4030. Senseman, S.A., 2007. Herbicide Handbook, ninth ed. Weed Science Society of America. Silva, M. d. S., Cocenza, D.S., Grillo, R., Silva de Melo, N.F., Tonello, P.S., de Oliveira, L.C., et al., 2011. Paraquat-loaded alginate/chitosan nanoparticles: preparation, characterization and soil sorption studies. J. Hazard. Mater. 190 (1e3), 366e374. Singh, B., Sharma, D.K., Kumar, R., Gupta, A., 2010. Development of a new controlled pesticide delivery system based on neem leaf powder. J. Hazard. Mater. 177 (1e 3), 290e299. Vieira, R.S., Beppu, M.M., 2006. Dynamic and static adsorption and desorption of Hg(II) ions on chitosan membranes and spheres. Water Res. 40 (8), 1726e1734. WHO, World Health Organization, 2009. The WHO Recommended Classification of Pesticides by Hazard and Guidelines to Classification 2009. Geneva.
Contents lists available at ScienceDirect
Journal of Environmental Management journal homepage: www.elsevier.com/locate/jenvman
Chitosan and alginate biopolymer membranes for remediation of contaminated water with herbicides Mariana Agostini de Moraes a, Daniela Sgarbi Cocenza b, Fernando da Cruz Vasconcellos a, Leonardo Fernandes Fraceto b, Marisa Masumi Beppu a, * a School of Chemical Engineering, University of Campinas e UNICAMP, Av. Albert Einstein, 500, Cidade Universitária Zeferino Vaz, 13083-852 Campinas, SP, Brazil b Department of Environmental Engineering, São Paulo State University e UNESP, Av. Três de Março, 511, Alto da Boa Vista, 18087-180 Sorocaba, SP, Brazil
a r t i c l e i n f o
a b s t r a c t
Article history: Received 13 March 2012 Received in revised form 11 September 2013 Accepted 23 September 2013 Available online 28 October 2013
This study investigated the adsorption behavior of the herbicides diquat, difenzoquat and clomazone on biopolymer membranes prepared with alginate and chitosan (pristine and multi-layer model) for contaminated water remediation applications. Herbicides, at concentrations ranging from 5 mM to 200 mM, were adsorbed in either pure alginate, pure chitosan or a bilayer membrane composed of chitosan/alginate. No adsorption of clomazone was observed on any of the membranes, probably due to lack of electrostatic interactions between the herbicide and the membranes. Diquat and difenzoquat were only adsorbed on the alginate and chitosan/alginate membranes, indicating that this adsorption takes place in the alginate layer. At a concentration of 50 mM, diquat adsorption reaches ca. 95% after 120 min on both the alginate and chitosan/alginate membranes. The adsorption of difenzoquat, at the same concentration, reaches ca. 62% after 120 min on pure alginate membranes and ca. 12% on chitosan/ alginate bilayer membranes. The adsorption isotherms for diquat and difenzoquat were further evaluated using the isotherm models proposed by Langmuir and by Freundlich, where the latter represented the best-fit model. Results indicate that adsorption occurs via coulombic interactions between the herbicides and alginate and is strongly related to the electrostatic charge, partition coefficients and dissociation constants of the herbicides. Biopolymer based membranes present novel systems for the removal of herbicides from contaminated water sources and hold great promise in the field of environmental science and engineering. Ó 2013 Elsevier Ltd. All rights reserved.
Keywords: Adsorption Alginate Chitosan Multilayer
1. Introduction Currently, there is a great concern for environmental preservation, where the contamination of water is one of the major problems of the degradation of the environment by human activity in the world. Environmental contamination by chemical products and herbicides is an eminent risk for nature preservation and human health in general. Herbicides become biologically active when applied to plantations, and are later transported to the soil and aquatic systems (Nunez et al., 2002). The excessive use of pesticides and herbicides has given rise to several short- and long-term adverse effects on human health such as in immune-suppression,
* Corresponding author. Tel.: þ55 19 3521 3893; fax: þ55 19 3521 3922. E-mail addresses: [email protected] (M. Agostini de Moraes), [email protected] (D.S. Cocenza), [email protected] (F. da Cruz Vasconcellos), [email protected] (L.F. Fraceto), [email protected] (M.M. Beppu). 0301-4797/$ e see front matter Ó 2013 Elsevier Ltd. All rights reserved. http://dx.doi.org/10.1016/j.jenvman.2013.09.028
hormone disruption, reproductive abnormalities and cancer diseases, among others (Singh et al., 2010). Three commonly used herbicides include Diquat (DQ), Difenzoquat (DF) and Clomazone (CLO). DQ is a non-selective contact herbicide, used in the control of weeds and grasses in plantation crops. DF is a selective herbicide used in cereal crops to control the post emergence of wild oats. CLO is used in rice crops to control the preemergence of mono and dicotyledonous plants (Gunasekara et al., 2009; Nunez et al., 2002). These compounds have different chemical structures (Fig. 1). In terms of polarity and number of charges, DQ is a very polar compound with two positive charges, while DF has just one positive charge and CLO does not have any charge, which will have an effect on the adsorption behavior of each compound. Table 1 shows values of water solubility, dissociation constant (pKa), octanolewater partition coefficient (log Kow) and molecular weight (Mw) for DQ, DF and CLO, taken from the pesticide properties database (PPDB-UK). The persistence of biological activity in moist and fertile soils in moderate temperature is one month or less for DQ, and 3e12
M. Agostini de Moraes et al. / Journal of Environmental Management 131 (2013) 222e227
223
Fig. 1. Chemical structures of the herbicides: (a) Diquat, (b) Difenzoquat and (c) Clomazone.
months for DF and CLO (Senseman, 2007). While active, DQ, DF and CLO are harmful to human health and contaminate the environment, which may lead to several problems in various ecosystems. The World Health Organization (WHO, 2009) has classified DQ, DF and CLO, based on their toxicity, as moderately hazardous. The maximum contaminant level for DQ in drinking water is 0.02 mg L1, while no maximum limits have been established for DF and CLO (EPA, 2011). The European Union has not yet established maximum levels of individual herbicides in water; however the total amount of contaminants, including herbicides, permitted is up to 0.5 mg L1 (EU, 2007). Herbicides are traditionally removed from water through adsorption, microfiltration or oxidation processes (Montgomery, 2007); however the development of novel, efficient and low cost methods to remediate contaminated areas are needed. In this context, the use of membranes composed of natural polymers, such as chitosan (CS) and alginate (AG), appears to be an eco-friendly alternative for the adsorption of herbicides and remediation of contaminated aqueous sources. Chitosan is a linear polysaccharide obtained from the deacetylation of chitin, a major component of crustacean shells. CS has been studied as a suitable biopolymer for removal of metal ions from wastewater, since its amino and hydroxyl groups can act as chelating sites (Beppu et al., 2004; Vieira and Beppu, 2006). Alginate is a polysaccharide derived from seaweed and possesses a backbone comprised of two repeating carboxylated monosaccharide units (manuronic acid and guluronic acid), the ratio of which influences the physical properties of the biopolymer. AG has proved to be an efficient adsorbent for dies (Jeon et al., 2008) and metal ions, such as copper and arsenic (V) (Lim and Chen, 2007). CS and AG are obtained from renewable sources that are biodegradable and non-toxic and are potential materials for the production of multilayer membranes, using the electrostatic forces between the polyanions of the carboxylate groups of AG and the polycations of the protonated amines of CS (Lawrie et al., 2007). These membranes present binding sites for removal of different herbicides displaying net positive or negative charges. Few studies in the literature report the use of biopolymers (as membranes, fibers or nanoparticles) for adsorption of herbicides (Sadasivam et al., 2011; Silva et al., 2011). Particularly, our previous study on CS and AG membranes for adsorption of Paraquat was the first one to use a bilayer membrane for herbicide removal and it
Table 1 Physicochemical properties of the herbicides DQ, DF and CLO. Herbicide
Water solubility (g L-1) pKa
Diquat 718 Difenzoquat 765 Clomazone 1.1
log Kow Mw (g mol-1)
Totally dissociated 4.60 7.0 0.61 No dissociation 2.54
*Values taken for the pesticide properties database, UK.
344.05 360.43 239.70
indicated a strong dependency between herbicide adsorption and the biopolymer properties (Cocenza et al., 2012). However, no studies have compared the adsorption capacity of various herbicides, with different net charges and physicochemical properties, on biopolymer membranes. The present study investigated the adsorption behaviors of the Diquat, Difenzoquat and Clomazone herbicides in membranes prepared with natural polymers (chitosan and alginate), in order to evaluate the use of these membranes as potential materials for remediation of contaminated aquatic systems. 2. Material and methods 2.1. Materials The herbicides Diquat PESTANALÒ, Difenzoquat PESTANALÒ, Clomazone PESTANALÒ and chitosan (extracted from crab shells, minimum deacetylation degree of 85%) were purchased from Sigma Chem. Co. Sodium alginate was purchased from Vetec Química Fina (Brazil). 2.2. Membrane preparation and characterization CS and AG were used as polyelectrolytes in concentrations of 1% (w/v). CS was dissolved in 2% vol acetic acid solution and AG was dissolved in 0.1 M NaOH aqueous solution. CS solution pH was ca. 3 and AG solution pH was ca. 10. The membranes were prepared according to the procedure proposed by Moon et al. (1999), with modifications. First, the CS solution was cast onto polystyrene dishes and dried at 50 C, for ca. 2 h (70% dried). The AG solution was then cast onto the CS film and completely dried at 50 C, forming a bilayer membrane (CS/AG). When completely dried, the membranes were immersed in a 0.1 M sulfuric acid solution in 50% vol ethanol, for 1 h to neutralize the functional groups of CS and AG. Membranes of the pure components (pure AG and pure CS) were also prepared in order to compare their characteristics and properties with those of the bilayer membrane. All membranes were prepared using the same total mass of biopolymer, i.e. the CS/AG membranes were produced with 50% of the biopolymer mass utilized in either the pure AG or CS membranes giving a total mass contribution equal to that used for the production of the pure membranes. The morphology of the membranes was observed by scanning electron microscopy (SEM), using a LEO 440i scanning electron microscope, after being cryo-fractured and freeze-dried for 24 h, at 54 C under vacuum. The membrane’s thickness was measured using a digital micrometer (MDC-25S, Mitutoyo), prior to (dry state) and after (wet state) the swelling test. For the swelling test, the membranes were cut to 2.5 cm in diameter and stored for 48 h in an environment with 50% relative humidity. The samples were weighed in the
224
M. Agostini de Moraes et al. / Journal of Environmental Management 131 (2013) 222e227
dry state (mi), immersed in 100 mL of distilled water and weighed after several time intervals, until reaching equilibrium (mf). The swelling capacity (g of water/g of sample) was calculated as Swelling ¼ ðmf mi Þ=mi . Fourier transform infrared spectroscopy with attenuated total reflectance (FTIR-ATR) was performed using a Bomem MB 102 spectroscope, prior and after the herbicides adsorption tests, in order to evaluate possible interaction effects between the CS and AG functional groups with those of the herbicides. 2.3. Adsorption tests The CS, AG and CS/AG membranes (with 4 cm2 of area) were immersed in aqueous solutions with the herbicides in several concentrations (5 mM, 25 mM, 50 mM, 100 mM and 200 mM) under continuous shaking (100 rpm) for 4 h at 25 C. The pH of the DQ, DF and CLO solutions were 7.3, 6.1 and 6.0, respectively. The experiments were carried out in duplicate for each one of the herbicides in each concentration. Aliquots of the solution were taken at 0, 15, 30, 45, 60, 120, 180 and 240 min and analyzed in a UVeVIS spectrophotometer (Varian e Cary 50) in the absorption wavelength range of each herbicide (DQ ¼ 308 nm; DF ¼ 257 nm; CLO ¼ 210 nm). Adsorption isotherms were obtained, in triplicate (n ¼ 3), from the changes observed in the absorbance of the solutions as a function of time, and mass balance calculations were used to determine the adsorption capacity of the membranes. The same experimental procedure was performed in the absence of the membranes (blank) in order to verify whether the herbicides had undergone any degradation throughout the experiment. The adsorption of herbicides may be described using an adsorption isotherm model. The isotherm represents the relationship between the amount of adsorbed herbicide per unit of sorbent, and the equilibrium concentration. Langmuir (Equation (1)) and Freundlich (Equation (2)) are the most frequently used models to describe herbicide adsorption:
Ce 1 Ce þ ¼ ðQ0 $Kl Þ Q0 qe lnðqe Þ ¼ ln Kf þ N lnðCe Þ
(1)
(2)
where Ce (mmol L1) is the equilibrium adsorbate concentration in the aqueous phase; qe (mmol g1) is the equilibrium adsorbate concentration in the solid phase; Q0 (mmol g1) is the maximum adsorption capacity and Kl (L mmol1) is the Langmuir constant and shows the adsorption intensity; Kf (L g1) is the Freundlich equilibrium constant and is an indicative of adsorption capacity and N (dimensionless) is the adsorption constant whose reciprocal is indicative of adsorption intensity. 3. Results and discussion 3.1. Membrane characterization The typical morphology of the CS/AG membranes, with a homogeneous and uniform cross section, characteristics of dense polymeric materials were obtained by scanning electron microscopy (SEM) (Fig. 2). The membranes present a different thickness for each biopolymer layer, even though the same mass for both biopolymers was used in their preparation. The CS layer (w7 mm), at the bottom, is thicker than the AG layer (w4 mm), at the top. This occurs due to the difference between the solutions densities: rCS > rAG. The CS/AG bilayer is well defined, signifying that the biopolymers interact strongly once a cohesive membrane is
Fig. 2. SEM micrograph of the cross section of the CS/AG bilayer membrane.
obtained and no delamination layers are observed. During the membrane preparation, CS (pK ¼ 6.5) and AG (pK ¼ 3.0) solutions had pHs of approximately 3 and 10, respectively. At these pH values, CS amino functional groups are fully protonated (NH3 þ ) and can easily interact with the AG carboxyl groups that are fully deprotonated (COO), at the membrane interface. Table 2 presents the results for the membrane swelling and thickness ratio in the dry and wet states. Swelling of the membranes occurred mainly within the first 30 s after their immersion in water (data not shown). The AG swelling was approximately two times higher than the CS swelling and the same proportionality is observed for the ratio between the dry and wet thicknesses of each biopolymer. These results indicate that the AG layer retains a higher quantity of water in the AG molecular chains and this water retention capacity could also be an indication of better adsorption properties. The swelling is related to chain mobility and crosslinking degree (i.e. an increase in crosslinking reduces swelling capacity) (Furuyama Lima et al., 2007). The post-treatment of membranes with sulfuric acid diluted in ethanol leads to the stabilization of AG membrane in water, even though its chain mobility is not lost. Traditional AG crosslinking in CaCl2 results in a swelling capacity of approximately 0.80 g of water/g of sample (RemunanLopez and Bodmeier, 1997), significantly lower than the results obtained in this study. 3.2. Adsorption results The analysis of the UVeVIS spectra of the herbicide solutions (blank) showed that there were no changes in solution concentrations with time, indicating that the herbicides were not adsorbed onto the glassware, and did not undergo any decomposition during the period of the experiment. Adsorption isotherms comparing adsorption of DQ, DF and CLO, each at a concentration of 50 mM, in the AG, CS and CS/AG membranes are shown in Fig. 3. The average experimental error for the Table 2 Swelling and thickness of the AG, CS and CS/AG membranes.
AG CS CS/AG a b
Swellinga (g of water/g of sample)
Initial thicknessb (ti) (mm)
Final thicknessb (tf) (mm)
Ratioa (tf/ti)
1.58 0.12 0.72 0.16 1.75 0.14
0.027 0.001 0.038 0.001 0.026 0.003
0.075 0.002 0.049 0.001 0.056 0.001
2.778 0.289 1.289 0.066 2.221 0.342
Average average deviation (n ¼ 3). Average average deviation (n ¼ 5).
M. Agostini de Moraes et al. / Journal of Environmental Management 131 (2013) 222e227
225
Fig. 3. Adsorption of 50 mM DQ, DF and CLO in the AG membrane (a), CS membrane (b) and CS/AG bilayer membrane (c).
isotherms is 2%. For the AG membranes (Fig. 3a), the adsorption equilibrium was reached within 120 min for DQ and DF, although with different adsorption percentages, where the adsorption was approximately 54% higher for the herbicide that had the most positive charges in its structure (DQ). No adsorption was observed in the experiment with CLO. In fact, CLO was not adsorbed on either the AG or CS membranes. This is probably due to the absence of charges in the CLO chains. This indicates that electrostatic interactions between the herbicides and the biopolymer chains are the main forces responsible for the adsorption. Due to the low CLO adsorption results observed on the membranes, further adsorption and characterization studies on this herbicide were not conducted. Additionally, pure CS membranes did not adsorb the herbicides studied (Fig. 3b), probably due to the fact that the herbicides DQ and DF present positive charges, as do the amino groups of the CS. Thus, no electrostatic attraction is achieved between the herbicides and the CS polymer chains, as reflected by the absence of adsorption. The CS/AG membranes (Fig. 3c) exhibited similar behaviors to the pure AG membrane with regard to DQ herbicide adsorption; however the kinetics of incorporation were slower in the former case. The incorporation of DF in CS/AG membranes was much lower than that observed for the pure AG membrane, indicating that saturation of the AG sites available for the adsorption had occurred. Fig. 4 shows the adsorption profiles of DQ at several concentrations of the herbicide in AG and CS/AG membranes (average error ¼ 3%). These results demonstrate that the DQ herbicide has a better adsorption on the AG membrane than on the CS/AG membrane and that, even at high concentrations, such as 200 mM,
the saturation of the membrane had not yet been completely achieved. The CS/AG membrane was also capable of adsorbing a high quantity of the DQ herbicide. Fig. 3b shows that CS does not adsorb DQ, signifying that all the DQ adsorption on the CS/AG membrane takes place in the AG layer. The CS/AG membrane has 50% of the AG mass present in the pure AG membrane, and is still capable of providing a fast adsorption of DQ, since within 120 min ca. 90% of the DQ herbicide, at concentrations in the range of 50 mMe200 mM, were already adsorbed. The good incorporation of DQ into the AG or CS/AG membranes is probably due to the negative charges of the carboxyl groups present in the AG chains, which favors their electrostatic attraction to the positive charges of the DQ molecules. For DF (data not shown), saturation of the AG membrane was observed even at small concentrations of herbicide. The adsorption capacity of the AG membrane for DF decreased with increasing DF concentrations. The adsorption capacity of the herbicides in the membranes can also be related to the physicochemical properties of the herbicides. For example, for the AG membrane, the adsorption capacity follows the order: DQ > DF > CLO (Fig. 3). As displayed by Table 1, the partition coefficient (log Kow) and the dissociation constant (pKa) parameters influence the herbicide adsorption capacity on AG membranes. An herbicide with a higher dissociation constant will have a structure with more ionisable groups that may interact with the AG membrane. In solution, DQ is totally dissociated, while DF displays pKa of approximately 7 and CLO does not undergo dissociation. Regarding the partition coefficient, DQ and DF can be considered hydrophilic, while CLO is hydrophobic. The lower the
Fig. 4. Adsorption profile at several DQ concentrations in the AG membrane (a) and in the CS/AG membrane (b).
226
M. Agostini de Moraes et al. / Journal of Environmental Management 131 (2013) 222e227
Table 3 Adjustment parameters of the Langmuir and Freundlich models for DQ and DF adsorption in AG and CS/AG membranes. Langmuir
Freundlich
Q0 (mmol g1) Kl (L mmol1) R
N (dimensionless) Kf (L g1) R
Diquat AG e CS/AG e Difenzoquat AG 34.19 CS/AG 7.80
e e
e e
1.12 1.01
4.49 2.69
0.99 0.88
35.76 8.23
0.89 0.22 0.91 0.64
3.34 0.81
0.99 0.90
partition coefficient, the more hydrophilic the herbicide is, leading to better adsorption on the AG biopolymer membrane. The solubility of the herbicides in water does not seem to be a key factor for the adsorption capacity. Although DQ and DF have similar solubility, DQ adsorption capacity was much higher that DF. In terms of molecular weight, all herbicides display similar values, and this parameter is not expected to significantly influence adsorption capacities. The Langmuir and linear Freundlich adsorption isotherm models were applied to the experimental data of DQ and DF adsorption in the membranes, as shown in Table 3 and Fig. 5 (AG membrane). The correlation coefficient obtained by the models showed that the Freundlich model fitted the data best (Table 3). These results can be explained by the fact that the Langmuir model normally does not fit data well for heterogeneous surfaces. The isotherm shape provides information on the strength by which the sorbate is held to the membrane, and also allows the determination of the maximum adsorption capacity of a sorbent (Echeverria et al., 1998). Experimental adsorption data for DQ fitted best in the linear Freundlich model and is a C-shaped type or a linear isotherm (N ¼ 1). This type of isotherm suggests that the solution has a high affinity with the sorbent. Interestingly, for DF, the Freundlich isotherm model fits with a good correlation coefficient; however, the isotherm is L shaped (N < 1), which indicates that DF has a moderate affinity for the membrane during the initial stage of adsorption, but there is decrease in the interaction when the adsorption sites begin to be filled (Echeverria et al., 1998). Additionally, according to the classification given by Giles and coworkers (Giles et al., 1960), the isotherms for DQ and DF have the L2 shape, which usually indicates that the molecules adsorb flat on the surface, or that vertically oriented ions with strong intermolecular attraction are adsorbed.
Fig. 6. FTIR-ATR spectra of (a) pure AG membrane and AG membrane with (b) DF and (c) DQ incorporated.
In order to verify whether any chemical interaction occurred between the AG layer and the herbicides, FTIR analysis was performed. Fig. 6 shows the FTIR spectra of pure AG membranes before and after DQ and DF adsorptions. For AG membranes, the main absorption bands observed at 1635 and 1409 cm1 are the C¼O antisymmetric and a symmetric stretch of the carboxylate group. The band at 1240 cm1 is the skeletal vibration of alginate and a sharp band at 1080-1020 cm1 is related to the antisymmetric stretch Ce OeC (Lawrie et al., 2007). No change was observed in the spectrum of the AG membrane with DF, when compared to that of the pure AG membrane. This is probably due to the weak interaction between DF and the carboxylate groups of AG. However, two new bands are observed at 1500 and 1284 cm1 in the AG membrane after DQ adsorption. These bands are related to characteristic bands of DQ, associated with ring stretching vibration and inter-ring stretching mode (Brienne et al., 1995). Future studies may further investigate the mechanisms for DQ and DF adsorption on AG and CS/AG membranes and relate them to the characteristic group vibrations. 4. Conclusions The results of this study provide new perspectives for using biopolymeric membranes (particularly AG and CS/AG) for remediation of areas contaminated with herbicides. AG proved to be an efficient biopolymer for the adsorption of DQ and DF, mainly due to the possible coulombic interaction between the carboxyl groups of AG and the positive charges of these herbicides. The dissociation constants and the partition coefficients of the herbicides also seem to be a key factor determining the adsorption capacity in the membrane, since higher dissociation constants and lower partition coefficients resulted in higher adsorption. The CS/AG membranes did not show the best result; however, this type of membrane may be of interest for the adsorption of different herbicides in each layer of the membrane, e.g., a positively charged herbicide can be adsorbed on the AG layer and a negatively charged herbicide can simultaneously be adsorbed in the CS layer, while also being a very effective adsorbent for DQ. In addition, the pH can be an important parameter on determining the adsorption behavior of the studied herbicides. The investigation of adsorption mechanisms at different pH values is subject of future work. Acknowledgments
Fig. 5. Linear Freundlich isotherm for adsorption of DQ and DF in AG membranes.
Financial support was provided by São Paulo Research Foundation (FAPESP), CNPq and Fundunesp.
M. Agostini de Moraes et al. / Journal of Environmental Management 131 (2013) 222e227
References Beppu, M.M., Arruda, E.J., Vieira, R.S., Santos, N.N., 2004. Adsorption of Cu(II) on porous chitosan membranes functionalized with histidine. J. Membr. Sci. 240 (1e2), 227e235. Brienne, S.H.R., Boyd, P.D.W., Schwerdtfeger, P., Bowmaker, G.A., Cooney, R.P., 1995. Infrared-spectra and theoretical-studies of bridged diquaternary 2,2’-dipyridinium compounds. J. Mol. Struct. 356 (2), 81e94. Cocenza, D.S., de Moraes, M.A., Beppu, M.M., Fraceto, L.F., 2012. Use of biopolymeric membranes for adsorption of Paraquat herbicide from water. Water Air Soil Pollut. 223 (6), 3093e3104. Echeverria, J.C., Morera, M.T., Mazkiaran, C., Garrido, J.J., 1998. Competitive sorption of heavy metal by soils. Isotherms and fractional factorial experiments. Environ. Pollut. 101 (2), 275e284. EPA, United States Environmental Protection Agency, 2011. Drinking Water Standards and Health Advisories. Washington. EU, 2007. European Communities Drinking Water Regulations (No. 2). SI 278, Dublin. Furuyama Lima, A.M., Andreani, L., Soldi, V., Borsali, R., 2007. Influence of plasticizer addition and crosslinking process on morphology, water absorption and mechanical properties of sodium alginate films. Quimica Nova 30 (4), 832e837. Giles, C.H., Macewan, T.H., Nakhwa, S.N., Smith, D., 1960. Studies in adsorption .11. A system of classification of solution adsorption isotherms, and its use in diagnosis of adsorption mechanisms and in measurement of specific surface areas of solids. J. Chem. Soc. (OCT), 3973e3993. Gunasekara, A.S., dela Cruz, I.D.P., Curtis, M.J., Claassen, V.P., Tjeerdema, R.S., 2009. The behavior of clomazone in the soil environment. Pest Manage. Sci. 65 (6), 711e716. Jeon, Y.S., Lei, J., Kim, J.H., 2008. Dye adsorption characteristics of alginate/polyaspartate hydrogels. J. Ind. Eng. Chem. 14 (6), 726e731.
227
Lawrie, G., Keen, I., Drew, B., Chandler-Temple, A., Rintoul, L., Fredericks, P., et al., 2007. Interactions between alginate and chitosan biopolymers characterized using FTIR and XPS. Biomacromolecules 8 (8), 2533e2541. Lim, S.F., Chen, J.P., 2007. Synthesis of an innovative calcium-alginate magnetic sorbent for removal of multiple contaminants. Appl. Surf. Sci. 253 (13), 5772e5775. Montgomery, J., 2007. Groundwater Chemicals Desk Reference, fourth ed. CRC Press. Moon, G.Y., Pal, R., Huang, R.Y.M., 1999. Novel two-ply composite membranes of chitosan and sodium alginate for the pervaporation dehydration of isopropanol and ethanol. J. Membr. Sci. 156 (1), 17e27. Nunez, O., Kim, J.B., Moyano, E., Galceran, M.T., Terabe, S., 2002. Analysis of the herbicides paraquat, diquat and difenzoquat in drinking water by micellar electrokinetic chromatography using sweeping and cation selective exhaustive injection. J. Chromatogr. A 961 (1), 65e75. RemunanLopez, C., Bodmeier, R., 1997. Mechanical, water uptake and permeability properties of crosslinked chitosan glutamate and alginate films. J. Controlled Release 44 (2e3), 215e225. Sadasivam, S., Kandasamy, K.S., Ponnusamy, K., Nagarajan, G.S., Kang, T.W., 2011. Film diffusion mechanism of methyl parathion in biopolymers: kinetics and thermodynamic equilibrium. J. Chem. Eng. Data 56 (11), 4024e4030. Senseman, S.A., 2007. Herbicide Handbook, ninth ed. Weed Science Society of America. Silva, M. d. S., Cocenza, D.S., Grillo, R., Silva de Melo, N.F., Tonello, P.S., de Oliveira, L.C., et al., 2011. Paraquat-loaded alginate/chitosan nanoparticles: preparation, characterization and soil sorption studies. J. Hazard. Mater. 190 (1e3), 366e374. Singh, B., Sharma, D.K., Kumar, R., Gupta, A., 2010. Development of a new controlled pesticide delivery system based on neem leaf powder. J. Hazard. Mater. 177 (1e 3), 290e299. Vieira, R.S., Beppu, M.M., 2006. Dynamic and static adsorption and desorption of Hg(II) ions on chitosan membranes and spheres. Water Res. 40 (8), 1726e1734. WHO, World Health Organization, 2009. The WHO Recommended Classification of Pesticides by Hazard and Guidelines to Classification 2009. Geneva.
