Chemosphere 112 (2014) 177–184

Contents lists available at ScienceDirect

Chemosphere journal homepage: www.elsevier.com/locate/chemosphere

The effects of acrylamide polyelectrolytes on aquatic organisms: Relating toxicity to chain architecture R. Costa a,⇑, J.L. Pereira b, J. Gomes a, F. Gonçalves b, D. Hunkeler c, M.G. Rasteiro a a CIEPQPF – Research Centre for Chemical Process Engineering and Forest Products, Department of Chemical Engineering, University of Coimbra, Pólo II, Rua Sílvio Lima, 3030-790 Coimbra, Portugal b CESAM – Centre for Environmental and Marine Studies, Department of Biology, University of Aveiro, Campus de Santiago, 3810-193 Aveiro, Portugal c Aqua+Tech, Chemin du Chalet-du-Bac 4, CH-1283 La Plaine CP 28, Geneva, Switzerland

h i g h l i g h t s  The effects of polyelectrolytes on aquatic biota depend on chain architecture.  Branched polymers are less toxic due to lower coverage of the biological surfaces.  Chain architecture is a critical design variable to optimise environmental safety.  Improved biofouling bivalve control may rely on chain architecture manipulation.

a r t i c l e

i n f o

Article history: Received 22 October 2013 Received in revised form 20 March 2014 Accepted 27 March 2014

Handling Editor: A. Gies Keywords: Acrylamide copolymers Aquatic biota Biofouling bivalves control Ecotoxicity Polyelectrolytes Polymer chain architecture

a b s t r a c t Understanding the inherent toxicity of water-soluble synthetic polyelectrolytes is critical for adequate risk management as well as enhancing product design when biological activity is a key performance index (e.g. for application in biofouling bivalves’ control). The toxicity of two cationic acrylamide copolymers with different chain branching degree was evaluated. Standard ecotoxicity tests were conducted with microalgae and daphnids. The susceptibility of Corbicula fluminea, as a biofouling bivalve, was also evaluated. The effect of polyelectrolyte on the test media viscosity and the polymer chain size distributions under the experimental conditions were also examined. The susceptibility of the microalgae to both polymers was similar. As the complexity and size of the test organisms increased, differences in toxicity due to different chain architecture were noticeable. The more branched polymer was significantly less toxic to both daphnids and the bivalves, which could be linked to the distinctive features of its bimodal size chain distribution. This architecture resulted in both more compact globular molecules and the formation of aggregates, which reduce the polymer interaction with the biological surfaces. The results of this study promote the incorporation of environmental considerations in polyelectrolyte development and contribute to the design of improved solutions for controlling biofouling bivalves. Ó 2014 Published by Elsevier Ltd.

1. Introduction Water-soluble synthetic polyelectrolytes are routinely employed in a number of fields, including mining, drinking water treatment, papermaking, personal care products manufacturing (Cary et al., 1987; Rowland et al., 2000; de Rosemond and Liber, 2004; Cumming, 2008; Harford et al., 2010). Despite the common use of these chemicals, their ecotoxicity has not been extensively investigated (Liber et al., 2005; Harford et al., 2010). Often their pollutant potential is somewhat ⇑ Corresponding author. Tel.: +351 239798700. E-mail address: [email protected] (R. Costa). http://dx.doi.org/10.1016/j.chemosphere.2014.03.096 0045-6535/Ó 2014 Published by Elsevier Ltd.

disregarded because low controlled discharges into the environment are generally expected, and there is a tendency for natural waters to further mitigate polyelectrolyte toxicity through adsorption phenomena (Cary et al., 1987; Goodrich et al., 1991; Bolto and Gregory, 2007; Cumming, 2008). Although this might be true in most cases, under some circumstances polyelectrolytes may pose significant ecological concerns. For instance, de Rosemond and Liber (2004) identified a cationic polyquaternary ammonium compound as the main toxic component of a diamond mine effluent. In some specific applications, as mining, the rates of use and discharge are such that only a very small fraction of the polyelectrolyte released needs to be in unbounded state to potentially affect the biota in the receiving waters (Liber et al., 2005). Additionally,

178

R. Costa et al. / Chemosphere 112 (2014) 177–184

polyelectrolyte accidental release or uncontrolled overdosing imply increased hazard (Rowland et al., 2000; ARC, 2004; Cumming et al., 2010). Therefore, the potential environmental risk posed by these substances should always be acknowledged. Understanding their inherent ecotoxicity is crucial for adequate management of such risk. The functionality, for example flocculant action, of polyelectrolytes may be maximised by manipulating not only the chemistry but also parameters such as polymer charge density, molecular weight and chain architecture (Rasteiro et al., 2010). Comprehending how these variables also affect ecotoxicity enables environmental considerations to be incorporated into the first steps of the polyelectrolyte development process (Hamilton et al., 1994), instead of being postponed to later stages when the product is close to get into the market. Furthermore, this type of information may assist the optimisation of the overall product performance, related to the specific application context, through the maximisation of polymer functionality and the minimisation of environmental impacts. A particular situation where molecular parameters may be exploited for optimising the overall product performance is the multi-objective molecular design of polyelectrolytes for the industrial control of biofouling bivalves, such as the Asian clam Corbicula fluminea and the zebra mussel Dreissena polymorpha. Polyelectrolytes have proved effective to mitigate these pests, and many are particularly promising for application in the highly regulated drinking water treatment industry because they are licensed for dosing in potable water and may also be used in the plant for other treatment purposes (Costa et al., 2011b). In this context, optimal overall polyelectrolyte performance would involve maximal pest control activity and operational functionality (for example as a flocculant) and ideally minimal impacts on nontarget species. Some attention has been devoted to the effects of polymer chemistry, molecular weight and charge density on polyelectrolyte ecotoxicity (Goodrich et al., 1991; Hall and Mirenda, 1991; Jop et al., 1998). While polymer chain architecture is a useful design variable, able to significantly affect operational functionality (Rasteiro et al., 2010), little is known about the way this parameter influences the impacts on aquatic biota and the importance of further exploring this theme has been acknowledged (Cumming, 2008). In this study the ecotoxicity of two cationic acrylamide/dimethylaminoethylacrylate methyl chloride copolymers (AM-co-DMAEA) with different chain branching degree, synthesised by inverseemulsion polymerisation, was investigated. The green microalgae Pseudokirchneriella subcapitata and the cladoceran zooplankter Daphnia magna, both standard test species for ecotoxicological evaluation, as well as the Asian clam C. fluminea, representing biofouling bivalves, whose industrial control may be achieved through polyelectrolytes, were used as model organisms. For aquatic organisms, it is reasonable to envisage that polyelectrolyte effects result from both indirect mechanical impairment, related to changes in the medium viscosity, and physiological disturbance, due to the direct action of the polymer molecules on the organisms. For this reason, the viscosity of solutions containing the polyelectrolytes in the test media and the polymer chain size distributions under the experimental conditions were also measured in an attempt to provide a physiochemical basis for the interpretation of the toxicity data. The former assisted the assessment of the mechanical impairment contribution to polyelectrolyte toxicity while the latter provided an insight into the interaction of the polymers with the biological surfaces. The outcome of this work may be relevant not only for risk management as part of polyelectrolyte development process, but also for the design of improved solutions for controlling biofouling bivalves.

2. Material and methods 2.1. Polymer synthesis and characteristics The AM-co-DMAEA were synthesised through an inverseemulsion polymerisation process (Barajas et al., 2004; Rasteiro et al., 2010; Palomino et al., 2012). The two model flocculants had similar composition and charge density (approximately 42 wt% acrylamide) as well as molecular weight (approximately 1.3  106 g mol 1, measured by analytical ultracentrifugation in a Beckman Coulter OPTIMA XL-I analytical ultracentrifuge and calculated by the modified Flory–Mandelkern–Scheraga equation (Mandelkern and Flory, 1952)). One (hereinafter designated as E2+) was a slightly branched polymer, containing, on average, one branch per chain. The other (thereafter referred to as E2++++) was characterised by highly branched architecture, with the individual chains containing, on average, four branches. Reference values of the polymers’ intrinsic viscosity, determined in a 0.05 M NaCl aqueous solution (Bourdillon et al., 2006), reflected the different chain branching degree: 1164 mL g 1 for E2+ and 977 mL g 1 for E2++++. 2.2. Ecotoxicological evaluation 2.2.1. Bioassays with the green microalgae P. subcapitata P. subcapitata was maintained in the laboratory as a non-axenic batch culture in Woods Hole MBL medium, at 20 ± 2 °C, under permanent illumination. New cultures were started by inoculating microalgae harvested during the exponential growth phase in fresh medium. Microalgae growth inhibition tests were conducted by adapting the appropriate OECD guideline (OECD, 2011) to the use of 24-well microplates (Geis et al., 2000). The microalgae were exposed to 12 concentrations (defined as a geometric series of ratio 1.75 in the range 0.318–150 mg L 1) of E2+ and E2++++, tested in triplicate. Each treatment (1 mL of algal suspension in Woods Hole MBL medium per microplate well) consisted of 900 lL of test solution plus 100 lL of algal inoculum, the latter adjusted to provide each well with an initial cell density of 104 cells mL 1. The cell density of the original inoculum was determined microscopically using a Neubauer haemocytometer. Each microplate held the replicates for two test concentrations plus two replicates of control treatments, where no chemicals were dosed. The microplates were incubated for 72 h at 20 ± 2 °C, under continuous illumination. The algal suspension in each well was thoroughly mixed by repetitive pipetting twice a day to prevent cell clumping. At the end of the test, the microalgae density in the wells was quantified by measuring the absorbance at 440 nm using a Jenway 6505 UV/ Vis spectrophotometer, based on a previously determined calibration curve. The microalgae production in each individual treatment (yield) was calculated as the difference between the cell densities at the end and the beginning of the test. The inhibition in yield (IY) was then expressed as IY = (YC YT)/YC  100, where YC and YT represent the mean value of yield for the controls and the yield in each replicated treatment, respectively. As a complementary analysis to aid the interpretation of the bioassay results, the zeta potential of the microalgae, an indication of the cells’ charge surface, was determined in a Zetasizer Nano ZS, model number ZEN3600, using filtered (0.2 lm pore size) Woods Hole MBL medium as dispersing agent. 2.2.2. Bioassays with the cladoceran D. magna The neonates used in the study were obtained from monoclonal bulk cultures of D. magna (clone A sensu Baird et al., 1989a). The cultures were maintained in synthetic ASTM hard water (ASTM, 1980), supplemented with a standard organic additive (Baird

R. Costa et al. / Chemosphere 112 (2014) 177–184

et al., 1989b) and vitamins (Elendt and Bias, 1990), at 20 ± 2 °C, on a 16-h light/8-h dark photoperiod cycle. The medium was renewed and the daphnids were fed P. subcapitata (3  105 cells mL 1) three times a week. The acute immobilisation tests were conducted in accordance with the OECD guideline 202 (OECD, 2004). Neonates aged less than 24 h and born within the 3rd to 5th culture broods were used in the experiment. For each treatment, twenty offspring were randomly distributed by four vials containing 25 mL of test solution. Following a static design, the daphnids were exposed to seven concentrations (3.0, 5.1, 8.7, 14.7, 25.1, 42.6 and 72.4 mg L 1) of each flocculant. Control treatments, where no polymers were dosed, were used in all trials. Immobilised organisms were counted after a 48-h exposure period. The test was performed using ASTM hard water (ASTM, 1980) enriched with vitamins (Elendt and Bias, 1990), under the same conditions as culture maintenance with the exception that the daphnids were not fed during the treatment period.

179

2.4. Polymer chain size distribution measurements Polymer chain size distributions, expressed in equivalent spherical diameters (hydrodynamic diameters), were determined by dynamic light scattering in a Zetasizer Nano ZS, model number ZEN3600, employing backscatter detection (173° angle). Stock solutions (1 g L 1) of each flocculant were prepared in Milli-Q water. The samples for size analysis were then obtained by diluting the stock solutions in either Woods Hole MBL medium, ASTM hard water or dechlorinated municipal water at a concentration that ensured an adequate attenuation level for the measurement (approximately 6 mL of stock solution in 40 mL of medium). The samples were passed through 1-lm syringe filters prior to analysis. The measurement temperature was set to 20 °C. As an aid for polymer chain size data interpretation, the conductivity of the single media was measured using a multi-parameter measuring instrument WTW Multi 3430 SET F. 2.5. Data analysis

2.2.3. Bioassays with the bivalve C. fluminea In early summer, adult Asian clams were collected from a freshwater canal in Mira (littoral centre of Portugal) by sieving sediment into a 1-mm mesh bag. Individuals falling in the shell length range 20–25 mm were selected and immediately transported in field water to the laboratory, where they were kept in aerated dechlorinated municipal water, at 20 ± 2 °C, on a 16-h light/8-h dark photoperiod cycle until use. The bioassays were initiated 1 week after collection and run under the maintenance conditions. Sets of 10 clams were placed into individual test beakers containing 500 mL of continuously aerated water. After a 24-h acclimation period in the test beakers, over which all bivalves were confirmed to be actively filtering, E2+ and E2++++ were dosed so that the clams were exposed to six concentrations of each polymer (75, 150, 300, 600, 1200 and 1500 mg L 1) in triplicate. Untreated sets of animals, also in triplicate, were used as control. Mortality was monitored every 24 h after polyelectrolyte dosing for 96 h. Open clams not showing evident siphoning activity and not responding to external tactile stimulus as well as closed animals that did not offer resistance as valve opening was carefully forced with a blunt dissection needle were considered dead. At each mortality assessment, dead organisms were discarded. The results of the test were analysed in terms of the mortality data obtained after 96 h of exposure to the polyelectrolytes. The mean shell length of the test organisms was 22.7 ± SD 1.2 mm.

2.3. Viscosity measurements Solutions of E2+ and E2++++ in Woods Hole MBL medium, ASTM hard water and dechlorinated municipal water were prepared for rheological studies. The concentration of the solutions fell in the ranges 0.25–256, 2.5–80 and 50–1600 mg L 1 respectively, covering those used in the corresponding bioassays. The single test media with no polyelectrolytes were also analysed. Each experimental condition was tested in triplicate. A Haake Model RS1 controlled stress rheometer, equipped with a coaxial cylinder system Z34 DIN with a gap of 7.2 mm and a Thermo Haake C35P refrigerated circulator Phoenix-line for temperature control, was used. Flow tests were conducted, subjecting the samples (50 mL) to an upward shear rate ramp for 120 s, followed by a stage of constant shear rate for 60 s and then a decreasing shear rate ramp for 120 s. Maximum values of shear rate up to 1500 s 1 were employed. The measurements were carried out at 20 ± 2 °C. No significant deviations from Newtonian behaviour were observed, and hence constant values of viscosity were recorded.

The methods for general statistical analysis of the data were used as outlined by Zar (1999) and implemented in STATISTICA software (StatSoft, Inc., 2003, STATISTICA – data analysis software system, www.statsoft.com). The concentration–response data were modelled by non-linear regression also using STATISTICA software in the case of the algal growth inhibition test, and by Probit analysis using StatsDirect software (StatsDirect, Ltd., 1990, StatsDirect statistical software, www.statsdirect.com) in the case of both the daphnia immobilisation test and the Asian clam mortality test. A significance level of 0.05 was used in all statistical analyses. 3. Results 3.1. Ecotoxicological evaluation The control organisms performed satisfactorily in all bioassays. In the microalgae growth inhibition tests, the algal production in the control treatments was 6.1  106 ± SE 2.1  105 cells mL 1, which, by definition, corresponded to a null inhibition in yield. In the daphnid acute immobilisation tests, no immobilised organisms were observed in the control vials. In the Asian clam mortality tests, the survival amongst untreated bivalves was higher than 95%. The acute toxicity of both flocculants was found to decrease as the complexity of the test organisms increased, with the median lethal concentration for the Asian clam being around 30 times higher than the median immobilisation concentration for the daphnids and more than 500-fold the median effective concentration obtained in the algal growth inhibition tests (Table 1). The effect of chain architecture on polymer toxicity also varied amongst the model species (Fig. 1). As for P. subcapitata, E2+ and E2++++ produced similar responses (two-factor ANOVA following arcsine transformation of the yield inhibition percent data; F = 0.62; df = 1; p = 0.434; Fig. 1a). Conversely, the less branched E2+ was significantly more toxic to D. magna (two-factor ANOVA following arcsine transformation of the immobilisation percent data; F = 23.54; df = 1; p < 0.0001; Fig. 1b) and C. fluminea (two-factor ANOVA following arcsine transformation of the mortality percent data; F = 5.78; df = 1; p = 0.024; Fig. 1c) than the more branched E2++++ architecture. In the case of the Asian clam in particular, where polyelectrolyte toxicity may be exploited for industrial mitigation purposes, the increased activity of E2+ was especially evident for higher response levels, which are those of major interest from the pest control perspective (Fig. 1c). The concentration of E2+ producing 90% mortality almost halved the corresponding concentration

180

R. Costa et al. / Chemosphere 112 (2014) 177–184

Table 1 Median effect concentrations (EC50) for E2+ and E2++++ (p 6 0.002 for all model fits). EC50 (95% confidence interval limits) (mg L

72-h microalgae yield inhibition 48-h daphnid immobilisation 96-h bivalve mortality

1

)

E2+

E2++++

1.4 (1.2–1.6) 21.6 (14.1–28.4) 834.2 (687.7–955.8)

1.3 (1.1–1.5) 37.0 (29.2–48.5) 1052 (817.7–1248.0)

As complementary data, the zeta potential of the microalgae was 33.9 ± SE 0.2 mV. 3.2. Viscosity measurements In general, regardless of the medium used, when present at low concentrations (up to approximately 100 mg L 1), neither E2+ nor E2++++ seemed to alter the solution viscosity (Fig. 2). At a higher concentration range, the viscosity increased as the polymer content augmented, with E2+ leading to more viscous solutions compared to E2++++ (Fig. 2). Detailed statistical analysis shows that for the solutions in Woods Hole MBL medium both polyelectrolyte concentration (two-factor ANOVA; F = 100.8; df = 6; p < 0.0001) and chain architecture (two-factor ANOVA; F = 26.7; df = 1; p < 0.0001) had a significant effect on the viscosity (Fig. 2a). A multiple comparison Bonferroni test demonstrated that such statistical significance came from the behaviour observed at the highest flocculant concentration: the viscosities of both 256 mg L 1 solutions were significantly different (p 6 0.003) from the viscosities of all other solutions as well as from each other, the E2+ solution being slightly more viscous than the E2++++ solution. It should be noticed that this polymer concentration is more than two orders of magnitude higher than the median effect concentration estimated for P. subcapitata. For ASTM hard water, only solutions containing up to 80 mg L 1 of polymer were analysed in agreement with the concentrations used in the daphnid acute immobilisation tests. At such low polyelectrolyte content, neither polymer concentration (two-factor ANOVA; F = 1.3; df = 6; p = 0.302) nor chain architecture (two-factor ANOVA; F = 0.7; df = 1; p = 0.415) were observed to significantly affect the medium viscosity (Fig. 2b). In accordance with the treatments applied in the Asian clam mortality tests, the concentration of the dechlorinated municipal water solutions analysed ranged from 0 to 1600 mg L 1. Within such extended range, both polymer concentration (two-factor ANOVA; F = 2409.6; df = 6; p < 0.0001) and chain architecture (two-factor ANOVA; F = 1032.2; df = 1; p < 0.0001) significantly affected the medium viscosity. As the polymer content increased, the solutions became more viscous, this behaviour being more pronounced for E2+ than for E2++++ (Fig. 2c). A comparative analysis of the relation between solution viscosity and flocculant concentration for different media shows that, within the overlapping range of concentrations tested, for both E2+ (ANCOVA; F = 1.02; df = 2; p = 0.370; Fig. 2) and E2++++ (ANCOVA; F = 0.48; df = 2; p = 0.621; Fig. 2), such relation was not significantly affected by the type of medium. 3.3. Polymer chain size distribution measurements Fig. 1. Concentration–response data obtained for the exposure of (a) P. subcapitata, (b) D. magna and (c) C. fluminea exposed to E2+ and E2++++. The points refer to the experimental data and the lines represent concentration–response models (p 6 0.002). The experimental data is represented as mean values and standard error bars.

for E2++++: 1218.1 mg L 1 (95% confidence interval limits for the estimate: 1064.6–451–1477.1 mg L 1) and 1995.5 mg L 1 (95% confidence interval limits for the estimate: 1587.5–3690.6 mg L 1), respectively.

Fig. 3 illustrates typical equivalent spherical diameter (hydrodynamic diameter) distribution curves for E2+ and E2++++ in the three test media. The temporal dimension, and hence the potential polyelectrolyte agglomeration dynamics, were not considered in the measurements. For this reason, these curves may not be an absolute representation of the flocculants’ behaviour throughout the duration of the bioassays. Nevertheless, they provide a common basis for the comparative analysis of E2+ and

R. Costa et al. / Chemosphere 112 (2014) 177–184

Fig. 2. Mean viscosity of solutions of E2+ and E2++++ in (a) Woods Hole MBL medium, (b) ASTM hard water and (c) dechlorinated municipal water, at varying concentrations covering the concentration range used in the corresponding bioassays. Error bars represent the standard error.

E2++++ structural attributes, which affect their interaction with biological surfaces. The degree of branching consistently affected the polymer chain size distribution in all three test media (Fig. 3). E2+ always exhibited unimodal chain size distributions, with modes in the range 360–480 nm. In contrast, E2++++ was characterised by bimodal chain size distributions, an indication that significant aggregation phenomena occurred. The first peak of the bimodal curves was located in the range 210–290 nm, meaning that the free E2++++ chains tended to have smaller hydrodynamic diameter (more coiled conformation) than the E2+ molecules. Aggregated E2++++ chains produced a second mode towards the highest size range (above 1000 nm).

181

Fig. 3. Typical polymer chain size distributions (intensity percent vs equivalent spherical diameter) for E2+ and E2++++ in (a) Woods Hole MBL medium, (b) ASTM hard water and (c) dechlorinated municipal water. The dashed lines represent no adjusted model and were used for clarity purposes.

The comparison of the size data for different media (Fig. 3a–c) shows that, on average, the E2+ molecules seem to have adopted a slightly larger, more extended form in dechlorinated municipal water, whose conductivity was 123 ± SE 12 lS cm 1, as compared to the structure assumed in Woods Hole MBL medium and ASTM hard water, where the conductivity was higher (517 ± SE 10 and 502 ± SE 6 lS cm 1, respectively). Conversely, a similar effect of the medium conductivity on the E2++++ chain sizes was not evident (Fig. 3a–c). 4. Discussion 4.1. The effect of chain architecture on polyelectrolyte toxicity While the susceptibility of the green microalgae to the polyelectrolytes did not seem to be affected by the chain architecture, for

182

R. Costa et al. / Chemosphere 112 (2014) 177–184

larger organisms (daphnids and bivalves) the flocculant toxicity decreased with increasing chain branching degree. These results may be rationalised based on the chemicals’ hypothesised mechanism of toxicity. In the aquatic environment, polyelectrolyte ecotoxicity may result from both indirect mechanical impairment and physiological disturbance. When the dissolved polymer significantly increases water viscosity, mechanical impairment is expected to occur, including movement inhibition, decreased filtering rates, disruption of food capture mechanisms and probably an increase of the energy expenditure on movement and filtering processes (Rowland et al., 2000; Liber et al., 2005; Bolto and Gregory, 2007; Acharya et al., 2010; Harford et al., 2010). More generally, cationic polyelectrolytes should exert toxicity by directly acting on the organisms, producing physiological disturbance. Since large polymer chains do not readily pass across the cell membranes, the most likely primary route of physiological disturbance is adsorption onto the biological surfaces in contact with water (e.g. gill tissues). As a result, the cell membranes’ structure is altered and transport mechanisms between the cells and the surrounding medium are disrupted, with impacts on respiratory and ion regulation processes, for example (Rowland et al., 2000; Liber et al., 2005; Bolto and Gregory, 2007). Depending on the species, further specific complications may result from the adsorption of the flocculant molecules onto biological surfaces. These include, for instance, microalgae flocculation and the disturbance of antennas’ movement with consequences to the filtration processes in daphnids. Both E2+ and E2++++ were comparatively very toxic to the green microalgae, producing particularly steep concentration– response curves. As changes in the test medium viscosity occurred only at the highest treatment dosages, these severe effects must have resulted mostly from physiological damage due to the especially strong affinity of the cationic polymer chains to the algal cells’ negatively charged surface, the latter confirmed by the negative zeta potential value obtained experimentally. Another reason for the pronounced susceptibility of the microalgae to the flocculants is the fact that they have a greater surface area-tovolume ratio compared to the larger test organisms, which favours toxicity due to adsorption. Since both E2+ and E2++++ produced sudden overwhelming toxicity on P. subcapitata, no differences could be detected between the responses elicited by each of them, and thus polymer chain architecture did not affect this species’ susceptibility to the polyelectrolytes. Conversely, when the complexity and size of the test organisms increased and the polymers’ acute toxicity decreased, differences in the polymers’ toxic potential due to different degree of branching became apparent. At similar, fairly high concentrations, E2+ originated more viscous solutions, which is consistent with its higher intrinsic viscosity reflecting a more extended conformation of this polymer molecule. Such higher solution viscosity may have contributed to the polymer’s increased toxicity due to indirect mechanical effects, in particular in the case of the bivalves’ mortality tests. However, over a wide range of treatments the increased susceptibilities of D. magna and C. fluminea were unrelated to changes in the medium viscosity. Therefore, again, it is likely that the observed toxicity resulted primarily from the direct interaction between the flocculants and the biological surfaces. Such interaction must have depended on the arrangement adopted by the polymer chains in solution, which explains the chemicals’ relative toxicity. The less branched E2+ molecules tended to be larger than the E2++++ chains, which is consistent with a more expanded, eventually more flexible, structure of the former that is likely to facilitate the accommodation of the molecules to the animals’ surfaces and thus promote toxicity. In contrast, even though the flocculants had similar molecular weight and charge density, the

E2++++ architecture resulted in a more compact globular configuration as expected (Bourdillon et al., 2006), translated by both lower polymer intrinsic viscosity and lower equivalent spherical diameters of the free molecules. Such a stiffer configuration must have limited the binding to the surfaces. Furthermore, the aggregation observed amongst E2++++ molecules, a phenomenon also reported by other authors (Bourdillon et al., 2006), reduced the number of free chains to originate larger structures with smaller surface area-to-volume ratio, decreasing the total polyelectrolyte surface available for interaction with the biological surfaces. Overall, these two effects must have resulted in a less effective coverage of the biological surfaces by the more branched flocculant and hence in reduced toxicity. As the polymer chain size distribution, linked to the arrangement of the molecules in solution, seems to be an important determinant of flocculants biological activity, it is worth noting that E2+ chains expanded less as the test medium conductivity increased. Ions in solution tend to screen the polymer charged segments, decreasing the electrostatic repulsion in the chains, which may thus assume a more compact configuration (Bolto and Gregory, 2007). Such a screening effect is less evident for more branched polymers (as E2++++) because their charged segments are somewhat protected from the dissolved ions by the chains’ more curled structure. Other authors have also shown that branched polyelectrolytes are more resistant to changes in the ionic content of solutions (Bourdillon et al., 2006). 4.2. Practical implications of the study While this study contributes to the understanding of the inherent toxicity of cationic polyelectrolytes, its immediate practical relevance may be somewhat questioned because the model compounds were tested at considerably high concentrations and in the absence of dissolved organic matter and suspended solids. Under normal process operating conditions, where optimal polyelectrolyte dosages are used and stringent wastewater practices are in place, it is unlikely that significant quantities of residual polyelectrolytes are discharged into the environment. Additionally, the tendency for dissolved and suspended materials to further decrease the concentration of unattached polyelectrolyte molecules, mitigating toxicity in natural waters is well-documented (Cary et al., 1987; Goodrich et al., 1991; Bolto and Gregory, 2007; Cumming, 2008). Nonetheless, under some circumstances, such as overdosing, accidental release or applications requiring particularly high polymer dosages, significant amounts of unbounded polyelectrolytes may reach the receiving environment, posing serious ecological concerns (Rowland et al., 2000; ARC, 2004; Liber et al., 2005; Cumming et al., 2010). For this reason, informed risk management requires a comprehensive view of polyelectrolyte inherent toxicity, which entails a step by step fundamental study approach as initiated in the present work. This study has also practical implications for the industrial control of biofouling bivalves. Such species may be deliberately exposed to toxic free polyelectrolytes by overcoming the mitigative effects of natural dissolved and particulate materials. This may be achieved by either using adequate dosages or specifically targeting the delivery of the chemical at the bivalves, for example through encapsulation (Costa et al., 2011a). Both E2+ and E2++++ were shown to exhibit molluscicidal activity towards the Asian clam, and thus seem promising for pest control, in particular in industries where they can be employed with dual function (for example, as flocculant and biocide). The potential of the polyelectrolytes as pest mitigation agents has to be further examined, namely by conducting long-term bioassays, in which low realistic polymer dosages are applied for prolonged exposure periods, as well as by assessing the polymers’ performance in the context of

R. Costa et al. / Chemosphere 112 (2014) 177–184

proactive control programmes targeted at veligers rather than adult bivalves. The results of the study further contribute to the design of improved solutions for biofouling bivalves control by highlighting the effect of architecture on polyelectrolyte toxicity. Attention has been devoted to the possibility of exploiting such structural parameter as a design variable to achieve optimal performance in technological applications, such as flocculation (Rasteiro et al., 2010). The increased susceptibility of C. fluminea to the less branched polymer indicates that a similar reasoning may be eventually applied when molluscicidal activity is the performance index of interest. Taking one step further, this result may be the seed of multi-objective polyelectrolyte design where molecular variables are manipulated to optimise molluscicidal activity together with other functional indices. Multi-objective polyelectrolyte design as defined here will be evaluated in a next step, with feedback from this study used to design an ideal molecular architecture. 5. Concluding remarks This study showed that the toxicity of polyelectrolytes strongly depends on chain architecture, which is thus an important design variable not only from the operational functionality point of view, but also from the environmental safety and biological activity (for example as a molluscicide) perspectives. The toxicity data provided here were supported by a comprehensive physical characterisation of the polymers’ behaviour in the test media, and hence a contribution to the general understanding of polyelectrolyte inherent toxicity rather than a simple particularised toxicological evaluation was given. Such a general understanding ultimately aims at more effective polymer development by both promoting early risk assessment considerations and the optimisation of product performance in situations where biological activity is a concern. This work paves the way for future, more detailed studies. Further analysis is still required for a fuller account of the influence of chain architecture on the polymers’ ecotoxicity. Such analysis should, for example, involve the determination of trigger values as integrated risk indicators and contemplate environmental fate issues and the toxicity mitigative effects of dissolved and suspended materials. Whilst polymer coil size is important for polyelectrolyte performance in several technological applications, the results reported here indicate that this parameter also strongly influences ecotoxicity. It depends on several factors, including molecular weight, chemical structure, chain architecture and solvent quality. Therefore, the possibility of using intrinsic viscosity, directly related to the chains’ hydrodynamic volume, as a holistic toxicity predictor able to integrate the effect of multiple variables comes as a preliminary conclusion of this study. Further investigation, involving a wide range of intrinsic viscosity values, to test this hypothesis is also worth pursuing in the future. Being able to use such a holistic predictor would be relevant in the context of polyelectrolyte development, in particular when the environmental or biological activity of alternative polymer designs, different at several levels, must be compared. Ultimately, the intrinsic viscosity may constitute a path for structure–activity relationships. Acknowledgements Financial support from the Portuguese Foundation for Science and Technology, FCT is gratefully acknowledged. The study was supported by FEDER funds through the programme COMPETE and by national funds through FCT under the scope of the

183

projects PEst-C/EQB/UI0102/2013, PTDC/AAC-AMB/113515/2009 and PEst-C/MAR/LA0017/2013. J.L. Pereira is recipient of an individual scholarship by FCT (SFRH/BPD/44733/2008). References Acharya, K., Schulman, C., Young, M.H., 2010. Physiological response of Daphnia magna to linear anionic polyacrylamide: ecological implications for receiving waters. Water Air Soil Pollut. 212, 309–317. ARC (Auckland Regional Council), 2004. Overview of the Effects of Residual Flocculants on Aquatic Receiving Environments. Technical Publication TP226. Auckland. ASTM (American Society for Testing and Materials), 1980. Standard Practice for Conducting Acute Toxicity Tests with Fishes, Macroinvertebrates and Amphibians. E 729-80. Philadelphia. Baird, D.J., Barber, I., Bradley, M., Calow, P., Soares, A.M.V.M., 1989a. The Daphnia bioassay – a critique. Hydrobiologia 188, 403–406. Baird, D.J., Soares, A.M.V.M., Girling, A., Barber, I., Bradley, M.C., Calow, P., 1989b. The long-term maintenance of Daphnia magna Straus for use in ecotoxicity tests: problems and prospects. In: Lokke, H., Tyle, H., Bro-Rasmussen, F. (Eds.), Proceedings of the First European Conference on Ecotoxicology. DIS Congress Service, Copenhagen, pp. 144–148. Barajas, J.H., Hunkeler, D., Wandrey, C., 2004. Polyacrylamide copolymeric flocculants with homogeneous branching: hetereophase synthesis and characterization. Polym. News 29, 239–246. Bolto, B., Gregory, J., 2007. Organic polyelectrolytes in water treatment. Water Res. 41, 2301–2324. Bourdillon, L., Hunkeler, D., Wandrey, C., 2006. The analytical ultracentrifuge for the characterization of polydisperse polyelectrolytes. Prog. Colloid Polym. Sci. 131, 141–149. Cary, G.A., Mcmahon, J.A., Kuc, W.J., 1987. The effect of suspended-solids and naturally-occurring dissolved organics in reducing the acute toxicities of cationic polyelectrolytes to aquatic organisms. Environ. Toxicol. Chem. 6, 469–474. Costa, R., Aldridge, D.C., Moggridge, G.D., 2011a. Preparation and evaluation of biocide-loaded particles to control the biofouling zebra mussel, Dreissena polymorpha. Chem. Eng. Res. Des. 89, 2322–2329. Costa, R., Elliott, P., Aldridge, D.C., Moggridge, G.D., 2011b. Enhanced mortality of the biofouling zebra mussel, Dreissena polymorpha, through the application of combined control agents. J. Gt. Lakes Res. 37, 272–278. Cumming, J.L., 2008. Environmental Fate, Aquatic Toxicology and Risk Assessment of Polymeric Quaternary Ammonium Salts from Cosmetic Uses. PhD Thesis. Griffith University. Cumming, J.L., Hawker, D.W., Matthews, C., Chapman, H.F., Nugent, K., 2010. Analysis of polymeric quaternary ammonium salts as found in cosmetics by metachromatic polyelectrolyte titration. Toxicol. Environ. Chem. 92, 1595– 1608. de Rosemond, S.J.C., Liber, K., 2004. Wastewater treatment polymers identified as the toxic component of a diamond mine effluent. Environ. Toxicol. Chem. 23, 2234–2242. Elendt, B.P., Bias, W.R., 1990. Trace nutrient deficiency in Daphnia magna cultured in standard medium for toxicity testing – effects of the optimization of culture conditions on life history parameters of Daphnia magna. Water Res. 24, 1157– 1167. Geis, S.W., Fleming, K.L., Korthals, E.T., Searle, G., Reynolds, L., Karner, D.A., 2000. Modifications to the algal growth inhibition test for use as a regulatory assay. Environ. Toxicol. Chem. 19, 36–41. Goodrich, M.S., Dulak, L.H., Friedman, M.A., Lech, J.J., 1991. Acute and long-term toxicity of water-soluble cationic polymers to rainbow-trout (Oncorhynchus mykiss) and the modification of toxicity by humic-acid. Environ. Toxicol. Chem. 10, 509–515. Hall, W.S., Mirenda, R.J., 1991. Acute toxicity of wastewater treatment polymers to Daphnia pulex and the fathead minnow (Pimephales promelas) and the effects of humic acid on polymer toxicity. Res. J. Water Pollut. Control Fed. 63, 895–899. Hamilton, J.D., Reinert, K.H., Freeman, M.B., 1994. Aquatic risk assessment of polymers. Environ. Sci. Technol. 28, A187–A192. Harford, A.J., Hogan, A.C., van Dam, R.A., 2010. Ecotoxicological Assessment of a Polyelectrolyte Flocculant. Internal Report 575. Australian Government, Department of Environment, Water, Heritage and the Arts, Supervising Scientist Division, Darwin. Jop, K.M., Allaway, J.R., Sampath, P.R., Guiney, P.D., 1998. Ecotoxicological hazard assessment of two polymers of distinctively different molecular weights. Arch. Environ. Contam. Toxicol. 34, 145–151. Liber, K., Weber, L., Levesque, C., 2005. Sublethal toxicity of two wastewater treatment polymers to lake trout fry (Salvelinus namaycush). Chemosphere 61, 1123–1133. Mandelkern, L., Flory, P.J., 1952. The frictional coefficient for flexible chain molecules in dilute solution. J. Chem. Phys. 20, 212–214. OECD (Organisation for Economic Co-operation and Development), 2004. Daphnia sp. – Acute Immobilisation Test. Revisal Proposed for Updating Guideline 202. Paris. OECD (Organisation for Economic Co-operation and Development), 2011. Freshwater Alga and Cyanobacteria – Growth Inhibition Test. Guideline 201. Paris.

184

R. Costa et al. / Chemosphere 112 (2014) 177–184

Palomino, D., Hunkeler, D., Stoll, S., 2012. Salt concentration influence on the efficiency of two cationic polymeric flocculants. Colloid Polym. Sci. 219, 1301–1308. Rasteiro, M.G., Garcia, F.A.P., Ferreira, P.J., Antunes, E., Hunkeler, D., Wandrey, C., 2010. Flocculation by cationic polyelectrolytes: relating efficiency with polyelectrolyte characteristics. J. Appl. Polym. Sci. 116, 3603–3612.

Rowland, C.D., Burton, G.A., Morrison, S.M., 2000. Implication of polymer toxicity in a municipal wastewater effluent. Environ. Toxicol. Chem. 19, 2136–2139. Zar, J.H., 1999. Biostatistical Analysis. Prentice Hall, Upper Saddle River.