Chemosphere 73 (2008) 593–599

Contents lists available at ScienceDirect

Chemosphere

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

Dissolved organic matter from treated effluent of a major wastewatertreatment plant: Characterization and influence on copper toxicity

Benoît Pernet-coudrier a, Ludiwine Clouzot a, Gilles Varrault a,*, Marie-Hélène Tusseau-vuillemin b,Alain Verger c, Jean-Marie Mouchel d

a Université Paris-Est, CEREVE UMR MA 102 – ENPC – ENGREF – Univ Paris 12, 61 av. du Gal de Gaulle, 94010 Créteil Cedex, Franceb CEMAGREF – Unité QHAN, Parc de Tourvoie, BP 44, 92163 Antony Cedex, Francec SIAAP – Direction de la Recherche et du Développement, 82 av. Kléber, 92700 Colombes, Franced SISYPHE, Université Pierre et Marie Curie – Paris 6, 4 place Jussieu, 75252 Paris Cedex 05, France

a r t i c l e i n f o

Article history:Received 22 January 2008Received in revised form 23 May 2008Accepted 26 May 2008Available online 15 July 2008

Keywords:Daphnia magnaComplexationHydrophilic/hydrophobic substancesIsolationTrace metalsToxicity

0045-6535/$ - see front matter � 2008 Elsevier Ltd. Adoi:10.1016/j.chemosphere.2008.05.064

* Corresponding author. Tel.: +33 (1) 45 17 16 31;E-mail address: varrault@univ-paris12.fr (G. Varra

a b s t r a c t

A combination of reverse osmosis (RO) concentration and DAX-8/XAD-4 resin adsorption techniques isused to isolate the various constituents of urban dissolved organic matter (DOM) from inorganic salts.Three fractions: hydrophobic (HPO), transphilic (TPI) and hydrophilic (HPI) accounting respectively for35%, 20% and 45% of extracted carbon, are isolated from effluents of a major French wastewater treatmentplant. This atypical DOC distribution, in comparison with natural water where the HPO fraction domi-nates, shows the significance of HPI fraction which often gets neglected because of extraction difficulties.A number of analytical techniques (elemental, spectroscopic: UV, FTIR) allow highlighting the weak aro-maticity of wastewater effluent DOM (EfOM) due to fewer degradation and condensation processes andthe strong presence of proteinaceous structures indicative of intense microbial activity. Copper toxicity inthe presence of DOM is estimated using an acute toxicity test on Daphnia Magna (Strauss). Results revealthe similar protective role of each EfOM fraction compared to reference Suwannee river fulvic aciddespite lower EfOM aromaticity (i.e. specific UV absorbance). The environmental implications of theseresults are discussed with respect to the development of site-specific water quality criteria.

� 2008 Elsevier Ltd. All rights reserved.

1. Introduction

In aquatic systems, dissolved organic matter (DOM) constitutesa key component of the carbon cycle controlling the speciation,bioavailability and toxicity of trace metals (Buffle, 1988; Tessierand Turner, 1995). The sources of dissolved organic carbon (DOC)may be categorized as natural autochthonous (1), being derivedfrom biota (e.g. algae, bacteria, macrophytes) growing in the waterbody, natural allochtonous (2a) entering the system from the ter-restrial watershed and anthropogenic allochtonous (2b), streamingwater (urban runoff and landfill leachate) and urban sewage (bothdomestic and industrial) whether treated or untreated.

Over the past few decades, many studies have been publishedregarding the capacity of DOM to complex copper. Most of thesestudies have found that binding capacities and binding constantsvary significantly depending on the experimental conditions em-ployed (conditional constants) or the range of binding constantsconsidered (analytical windows) (Town and Filella, 2000). Theyalso depend on the nature and origin of isolated organic matter

ll rights reserved.

fax: +33 (1) 45 17 16 27.ult).

(Mantoura et al., 1978; Buffle et al., 1980). It is interesting to notehowever that the published data pertain mainly to the so called‘‘humic substances” (HS) and demonstrate the ability of these sub-stances to complex metals. HS are heterogeneous polyelectrolyteorganic material and the most hydrophobic fraction of the DOMisolated from natural water, based on XAD-8 resin adsorption atacid pH. HS are derived from oxidative and hydrolytic biodegrada-tion of plants and animals (Stevenson, 1994) and they make up40–60% of DOC in natural surface water (Martin-Mousset et al.,1997). In urbanized water, the hydrophobic characteristic ofDOM is weaker as a result of various urban DOM discharges andthe strong primary productivity induced by these discharges (Maet al., 2001; Imai et al., 2002). Recent studies carried out on urbanwastewater (Sarathy and Allen, 2005; Buzier et al., 2006) haveemphasized its great potential to complex metals despite the ex-pected large amount of hydrophilic substances it contains. Becauseof the difficulty in isolating the hydrophilic fraction of DOM, verylittle information is available in the literature regarding hydro-philic DOM and their influence on metal complexation. MoreoverDe Schamphelaere et al. (2004), states that the variability of natu-ral DOM with respect to metal binding is important chemically,biologically and toxicologically because metal speciation severelyaffects the fate of trace metals in the water column.

594 B. Pernet-coudrier et al. / Chemosphere 73 (2008) 593–599

The primary objective of this study is to exhaustively extractDOM from wastewater effluent (EfOM) then to fractionate itaccording to polarity criteria (Croué, 2004) and to further charac-terize the obtained fractions by means of various analyses (ele-mental, spectroscopic UV and FTIR). The second goal is to studythe influence of every isolated fraction on copper toxicity. Sincethe speciation of dissolved metals strongly affects their biologicalavailability, i.e. their capacity to reach the biological target (Tessierand Turner, 1995), a biotest based on an acute toxicity test usingDaphnia Magna has been applied to assess the influence of eachEfOM fraction on metal toxicity. A Suwannee river fulvic acid ob-tained from International Humic Substances Society (IHSS) waspurchased to compare EfOM with a natural organic matter. DGT(diffusion gradient in thin films) were dipped in the solutions usedfor the bioassay in order to measure labile metal and to assesswhether or not the DOM influence recorded during the bioassaycould be attributed to trace metal complexation.

2. Materials and methods

2.1. Sample collection

Treated effluent from the Seine-Aval wastewater treatmentplant (WWTP) was sampled (250 L) in April 2006 during a dryweather period. This WWTP was chosen for the present study sinceit collects over 70% of the dry weather (combined sewer) flowsfrom Paris and the surrounding suburban metropolitan area (�8million inhabitants). These effluents are highly representative ofdischarged effluent from Paris and are treated at this WWTP by pri-mary settling and aerobic activated sludge. The sample was col-lected using a double membrane pump (IR ARO�) and filteredonsite through subsequent 10-lm and 0.45-lm polypropylene car-tridge filters (Predel�). One IHSS standard fulvic acid (SuwanneeRiver Fulvic Acid SRFA 1S101F) was used as a natural organic mat-ter for follow-up investigations.

2.2. Dissolved organic matter isolation: The RO/DAX resins protocol

The sample was then softened on sodium cation-exchange resinin order to eliminate calcium and magnesium ions that co-precip-itate with DOM and could clog the reverse osmosis (RO) mem-branes. The RO step enables reducing sample volume by oneorder of magnitude. During the concentration step, conductivitywas monitored through permeate to stop the concentration stepbefore any major DOM leakage could occur. The aliquots for DOCmeasurement and UV absorbance were also sampled in permeatein order to accurately calculate the slight DOM leakage. After theconcentration step, a final rinsing of membranes (Filmtech TW30) with a 0.05 M sodium hydroxide solution allowed recoveringadsorbed DOM in the eluate (Croué, 2004). Sample filtration, soft-ening and concentration were carried out in line and onsite so as tolimit process duration and potential DOM biodegradation. Theobtained sodic eluate and the RO concentrate, added and collectedin a stainless steel bottle (50 L), were then acidified (HCl,0.01 mol L�1) and filtered back at the laboratory on nonionic mac-roporous Amberlite� DAX-8 resins (an attractive substitute for thewell-known XAD-8, according to Peuravuori et al., 2001, 2002)(acrylate ester) and Supelite� XAD-4 (divynil benzene) combinedwith one another. According to Leenheer (1981) a column capacityfactor, k’ equal to 50 was applied to isolate HPO substances. DOCmeasurements were conducted in both influent and effluent ofDAX-8 and XAD-4 resins to assess the proportion of each fraction.Hydrophobic (HPO) and transphilic (TPI) fractions were derivedfollowing the procedure described in Fig. 1. The HPI fraction, i.e.the fraction of DOM not adsorbed on the DAX-8 and XAD-4 resins,

contains all the salts initially present in the sample and requiresfurther purification given that salts would interfere during the var-ious investigations. The separation of inorganic salts from the HPIfraction was performed according to a protocol described byLeenheer et al. (2000) which consists of a zeotrophic distillationand successive evaporation steps to precipitate the inorganic salts.All Teflon and stainless steel materials used for sampling and iso-lation were washed several times with ultrapure deionised water.The glass material was washed with a detergent (TFD4, 5%), rinsedthoroughly with ultrapure deionised water and then precombustedfor 5 h at 500 �C. Before application, the resins were washed bymeans of soxhlet extraction and successive acid and sodic rinsingin accordance with a protocol described by Leenheer (1981).Blanks experiments were carried out, by implementing this proto-col with ultrapure deionised water throughout the system (pipes,pump, and filters all prewashed using ultrapure deionised water).The DOC values for these blank tests never reached the DOC quan-tification limit of 0.5 mgC L�1.

2.3. Effluent dissolved organic matter characterization

DOC content were determined using the O.I. Analytical (quanti-fication limit = 0.5 mgC L�1) carbon analyzer. Elemental analyseswere carried out by the Service Central des Analyses Laboratory ofthe Centre National de la Recherche Scientifique (CNRS, Solaize)and enabled determining the C, H, O, N, S contents to a precisionof 0.3% and a standard deviation of 0.2%. Specific UV absorbance(SUVA) is defined as the UV absorbance of a given sample deter-mined at 254 nm and divided by the organic carbon content ofthe solution, according to U.S. EPA recommended method 415.3(United State Environmental Protection Agency, 2005). SUVA wasalso evaluated at 350 nm. UV absorbance at 254 nm was obtainedusing a Lambda Perkin Elmer spectrophotometer with 1-cm longquartz cells. Infrared spectra of 2–5 mg of DOM fraction isolatesin potassium bromide pellets were determined on a Fourrier Trans-form InfraRed spectrometer (Perkin Elmer Spectrum BX). All spec-tra were normalized after acquisition with a maximum absorbanceof 1.0 for comparative purpose.

2.4. Acute toxicity tests on Daphnia Magna

The acute 24-h immobilization assay was performed in accor-dance with the ISO 6341 standard procedure (International Stan-dard Organization, 1993). The test organisms originating from ahealthy Daphnia Magna clone (K6) were cultivated in soft springwater (Mont-Dore spring) under controlled lighting (100 lux, witha 8:16 h light:dark period), temperature (22 ± 1 �C) and feedingconditions (daily feeding with Selanastrum capricornutum in anexponential growth phase). Five Daphnia magna juveniles (<24 hold) were exposed in triplicate in 15 mL of different media at pH7: Mont-Dore soft spring mineral water (reference inorganic med-ia), and four organic media. Mont Dore water was chosen for itsneutral pH and low mineralization. The four tested organic mediawere obtained by dissolving in Mont-Dore water 2 mgC L�1 of thethree previously-isolated fractions and SRFA. All solutions wereprepared the day before the experiment in order to allow for over-night equilibration (one night at 22 �C). The copper stock solutionused was an atomic absorption spectroscopy standard solution(Acros Organics, 1000 mg L�1 Cu in 2% HNO3). The number ofimmobilized daphnia was counted after 24 h of exposure. Thedose-effect curves and the toxic concentrations for 50% of popula-tion (EC50) were obtained by a nonlinear fit of the logistical model(REGTOX software, E. Vindimian, available at <http://perso.wana-doo.fr/eric.vindimian>) to experimental data (Isnard et al., 2001),with a 95% confidence interval. The control samples (medium withno copper spike) showed no toxicity of DOM fractions to daphnids.

Preparation

Concentration

Fractionation

Cationic exchange resin (Na+)

Effluent of WWTP Seine

Reverse Osmosis

Field filtration (10 µm and 0.45 µm)

Hydrophobic DOM (HPO)

DAX-8 XAD-4

Acidified concentrate (pH 2, HCl)

Transphilic DOM(TPI)

Hydrophilic DOM(HPI)

Elution withAcetonitrile/Water

(75 % / 25 %)

Zeotrophic distillation to remove inorganic salts

Freeze-dry

Vacuum evaporate acetonitrilePurification

Preparation

Concentration

Fractionation

+

Effluent of WWTP Seine-Aval

Reverse Osmosis

Hydrophobic DOM (HPO)

DAX-8 XAD-4

Acidified concentrate (pH 2, HCl)

Transphilic DOM(TPI)

Hydrophilic DOM(HPI)

Elution withAcetonitrile/Water

(75 % / 25 %)

Zeotrophic distillation to remove inorganic salts

Freeze-dry

Vacuum evaporate acetonitrilePurification

Fig. 1. Comprehensive isolation protocol of DOM: The RO/DAX protocol.

B. Pernet-coudrier et al. / Chemosphere 73 (2008) 593–599 595

2.5. Total and labile copper measurements

The DGT holders (piston type, 2-cm diameter window) werepurchased from DGT Research (Lancaster, UK), along with the dif-fusive gel disks (restricted gel type, 0.8 mm thickness) and the Che-lex binding resin. An additional cellulose ester filter membrane(Millipore, thickness 0.13 mm) was used to separate the gel fromthe solution. For each of the 30 treatments (5 media, 6 copperconcentrations), two DGT were deployed in a 500 mL polypropyl-ene-stirred beaker containing a well-stirred sample solution. Thebeakers were previously soaked at least 24 h in HNO3 10%, thenthoroughly rinsed with ultrapure deionised water. At both thebeginning and the end of each experiment, an aliquote of the solu-tion was collected and acidified (ultrapure HNO3, 10%) for totalcopper measurement. Once retrieved from the DGT, the Chelex re-sin was eluted overnight in 5 mL of 1 M ultrapure nitric acid.Copper was measured in each sample by graphic furnace-atomicabsorption spectroscopy (GFAAS) (Varian SpectrAA 220Z). Themass of metal sequestered on the resin was evaluated consideringan 80% elution yield (Zhang and Davison, 1995). The labile copperconcentration could then be calculated (Davison and Zhang, 1994)as

CDGT ¼M � Dgt � A � D

where M is the mass of copper accumulated on the resin after adeployment time t, Dg is the thickness of the diffusive layer (here0.8 mm hydrogel thickness plus 0.13 mm due to the membrane),A the window area and D the diffusion coefficient of copper (4.7610�6 cm2 s�1 at 22 �C in restricted gel, data from DGT Research,available at <http://www.dgtresearch.com>). The quantificationlimit of GFAAS for copper analyses is 0.5 lg L�1 and the deviationamong replicates was typically less than 15%.

2.6. Biological determination of copper complex stability constants

If we consider the elementary complexation scheme of a freemetal ion M capable of forming a complex ML by association withthe ligand L, i.e.:

Mþ L ()Ka

Kd

ML ð1Þ

where charges have been omitted for the sake of simplicity, the rateconstants for association and dissociation, Ka and Kd, respectivelyrelate to the conditional complex stability constant K (Eq. (2)):

K ¼ Ka

Kd¼ ½ML�½M�½L� ¼

½M�tot � ½M�½M�½L� ð2Þ

Under assumptions of the free ion activity model (FIAM) (Morel,1983), the bioavailability of copper to Daphnia magna is proportionalto Cu2+. Since ionic copper at pH 7 is the predominant toxic speciesin short-term toxicity (Andrew et al., 1977; Meador, 1991), the pos-sible contribution to the toxicity of other inorganic copper specieshas been discarded. Based on FIAM, the free metal concentrationsinducing 50% effect on Daphnia magna (EC50free) were consideredto be identical (for the same effect) in each medium and taken asthe value in the inorganic reference medium calculated by the Visu-alMinteq software (v 2.50). According to the biotic ligand model(BLM), major cations mitigate the free metal ion toxicity by meansof the competition effect (De Schamphelaere and Janssen, 2002).This effect was assessed and then neglected herein due to the verylow major cations concentrations (soft spring water) as well as thelow effect of major cations concentrations beside the strong effectof DOM. In using the EC50free instead of [M], and [DOC] instead of[L] (as a result of the excess ligand condition) in (2), we can calculatea biologically determined constant Kbio expressed in L mg�1 of DOCfor a critical effect to Daphnia magna (Eq. (3)):

596 B. Pernet-coudrier et al. / Chemosphere 73 (2008) 593–599

Kbio ¼½M�tot � EC50free

EC50free½DOC� ð3Þ

In other words, Daphnia magna has been used as a biological speci-ation tool to estimate free copper concentration for the purpose ofcalculating the biologically determined constant for each DOMfractions.

3. Results and discussion

3.1. DOM Isolation and fractionation

The treated wastewater sample contains 38 mgC L�1. The loss ofDOM in the permeate amounted to less than 0.5 mgC l�1, which inthis case enabled calculating a DOC rejection efficiency for the ROmembrane higher than 98%. This rejection rate has typically rangedbetween 91% and 98% for a number of studied water samples(Croué, 2004), which demonstrates the efficiency of the RO mem-brane to concentrate DOM into a restricted volume. The additionalrinsing of the membrane surface (with sodium hydroxide solution)leads to overall DOC recovery rates of 100 ± 3% among concentrate,permeate and membrane surface. The last rinsing solution of ROmembranes (15 L vs. 25 L for the concentrate) contains 5% of thetotal DOM. Carbon balances before and after each purification stephave enabled to reconstituting the initial DOM sample compositionand estimating DOM loss. The DOM recovered as brownish powderrepresents 75% of the total DOC (in the raw sample). The majorityof DOC loss stems from the HPI fraction due to a certain level of co-precipitation of organic matter with inorganic salts during thepurification step. As a matter of fact, only 62% of this fractionwas recovered, thereby revealing the difficulty of HPI isolation.The others DOC losses originate from the lack of complete elutionof HPO and TPI fractions. Fractionation of the RO concentratedsample on DAX-8/XAD-4 resins led to three fractions, with themost hydrophilic one (HPI) representing 45% of the concentratedDOC, hydrophobic fraction (HPO) and transphilic fraction (TPI) rep-resent 35% and 20%, respectively, of the concentrated DOC (see Ta-ble 1). Such a distribution, atypical in natural water (Violleau,1999), has also been observed in another study on WWTP effluent,wherein the hydrophobic fraction represents less than 40% of DOCand the hydrophilic fraction varies between 60% and 75% (Imaiet al., 2002).

Table 1Elemental composition and SUVA of isolated fractions from WWTP Seine-Aval

DOM Fractions Distribution (%) Elemental c

C H

Treated effluent from WWTP This study HPO 35 53.6 6.5TPI 20 48.4 6.4HPI 45 43.8 7.1

IHSS SRFA SRFA – 52.3 4.4

Ribou water Croué (2004) HPO 68 49.2 5.2TPI 25 45.6 5.0HPI 8 35.2 4.3

Blavet river Violleau (1999) HPO 79 47.0 4.6TPI 10 45.2 5.6HPIa 11 – –

Gartempe river Violleau (1999) HPO 66 47.0 4.7TPI 19 43.2 4.7HPI a 15 – –

Loire river Violleau (1999) HPO 60 49.7 5.0TPI 28 45.7 5.0HPIb 12 26.0 3.5

a not isolated.b High content of inorganic salts.

3.2. Chemical characterization of DOM

Table 1 lists the results from an elemental analysis along withSUVA of the HPO, TPI and HPI fractions. The sums of elementalmasses exceed 94% in each of the fractions, which indicates thatthe isolation protocol has efficiently removed inorganic constitu-ents from the sample. The carbon contents of EfOM fractions areclose to those observed for SRFA and in the literature for surfacewater (Violleau, 1999; Ma et al., 2001; Peuravuori et al., 2001;Croué, 2004). The hydrogen, nitrogen and sulphur contents how-ever are especially high and on the opposite the oxygen contentis lower for EfOM fractions (Croué et al., 2003). As previouslyobserved (Violleau, 1999; Croué, 2004), H/C, N/C and O/C ratiosincrease with the hydrophilic nature of the samples (HPO < TPI <HPI), thus indicating a smaller degree of unsaturation and anenrichment in nitrogenized structures and oxygenated functionalgroups. The likely presence of more aromatic structure in HPOand, to a lesser extent TPI than in HPI has once again been con-firmed again by SUVA, with adsorption at 254 nm being an indica-tor of aromatic structures (Leenheer and Croué, 2003). SUVAobtained for these fractions are ranging between 2.0 and0.9 m�1 L mgC�1 and close to those obtained from fractions of trea-ted effluent of WWTP by Imai et al. (2002). Nevertheless, EfOMSUVA remains weaker than that of SRFA (4.3 m�1 L mgC�1) andDOM in natural water, since humic substances reach values ofSUVA of 5 m�1 l mgC�1 (Violleau, 1999; Imai et al., 2001; Croué,2004). This finding indicates a low aromatic characteristic of EfOM,most likely due to fewer degradation and condensation processes.

Fig. 2 presents the FTIR spectra of the three EfOM fractions andthe SRFA. The first observation is that the absence of bond vibra-tion for inorganic salts (sodium, carbonate or silicate) shows thepurity of EfOM fractions and the isolation protocol efficiency to re-move inorganic salts. Ten bands can be attributed to the variouschemical bond vibrations. The first band, very broad, is attributedto the O–H bond in alcohols and carboxylic acids. Bands 2, 6 and7 are all correlated with amides and amines, which are character-istic of proteins and amino sugars (Leenheer and Rostad, 2004),thus explaining the high nitrogen content previously determinedin EfOM. Proteinaceous structures are also observed for the HPOfraction of wastewater treatment effluents, and this is interpretedas an indicator of intense microbial activity (Drewes and Croué,2002). Because of the low nitrogen content in SRFA and the slight

omposition (%) Atomic ratios SUVA (m�1 L mgC�1)

O N S Total H/C O/C N/C 254 nm 350 nm

27.9 5.7 2.6 96.3 1.45 0.39 0.09 2.0 0.6128.8 8.4 2.1 94.1 1.56 0.45 0.15 1.4 0.3229.4 12.3 2.1 94.7 1.92 0.50 0.24 0.9 0.25

43.0 0.7 0.5 100.9 0.99 0.62 0.01 4.3 1.38

38.7 2.7 – 95.8 1.25 0.59 0.05 3.4 –41.5 4.6 – 96.7 1.31 0.68 0.09 2.1 –44.4 4.8 – 88.7 1.44 0.95 0.12 1.2 –

38.8 2.0 – 92.4 1.17 0.62 0.04 4.3 –41.0 2.3 – 94.1 1.49 0.68 0.04 2.3 –– – – – – – – – –

38.1 1.9 – 91.7 1.20 0.61 0.03 4.0 –44.4 2.6 – 94.9 1.31 0.77 0.05 2.9 –– – – – – – – – –

33.5 2.1 – 90.3 1.21 0.51 0.04 2.9 –39.8 4.2 – 94.7 1.31 0.65 0.08 1.9 –36.2 4.3 – 70.0 1.61 1.04 0.14 1.2 –

4000 3600 3200 20002400 18002800 1400 1200 1000 800 6001600

Abs

Wavenumber (cm-1)

HPO

TPI

HPI

27 10

SRFA

4000 3600 3200 20002400 18002800 1400 1200 1000 800 6001600

Abs

Wavenumber (cm-1)

HPO

TPI

HPI

27 10

SRFA

61 54 983

Fig. 2. FTIR spectra of SRFA and HPO, TPI, HPI fractions isolated from WWTP Seine-Aval effluents (band 1: O–H bonds; bands 2, 6, 7: nitrogen groups; bands 3, 8: aliphaticchains; bands 4, 5: carboxylic groups; band 9: C–O bonds; band 10: sulphonic groups).

0

20

40

60

80

100

0 10 20 30 40 50 60 70 80[Cu] (µg L-1)

Effect (%)

Total dissolved Cu

Labile Cu

EC50

Labile Cu

EC50

Fig. 3a. Dose-effect curves obtained for the HPO media.

5

10

15

20

25

30

inorganic media HPO TPI HPI

40EC50 (µg l -1) EC50 (total Cu)

EC50 (labile Cu)

SRFA

a a ab b

c cc c

a

µg -1)

Fig. 3b. Copper EC50 obtained for each media. (groups a, b and c are statisticallydifferent).

B. Pernet-coudrier et al. / Chemosphere 73 (2008) 593–599 597

shift, the band 6 for SRFA is attributed to aromatic carboxylic acid.The presence of aliphatic chains is revealed by the bands 3 and 8:at 2960 cm�1 with a small shoulder (C–H asymmetric stretching inCH2 and CH3) and at 2890 cm�1 (C–H symmetric stretching in CH2)(Dignac et al., 2000). The variation in the intensity of the aliphaticpeaks confirms the differences in alkyl chain abundance betweenthese fractions. Bands 4 and 5 are associated with the carboxylicacid bonds. Peaks 5, at approximately 1700 cm�1 are very intenseand this highlights the acid character of organic matter (Croué,2004). Band 9 is attributed to the C–O of alcohols and sugars.The high intensity of this peak for the HPI fraction displays theabundance of sugars in HPI fraction (Leenheer and Rostad, 2004).Band 10 may be correlated to the presence of sulphonic groupassociated to carboxylic aromatic cycles. These compounds arebreakdown products (metabolites) of a surfactant (LAS: linear alkylsulphonates) as previously observed in municipal wastewater(Barber et al., 1997). In conclusion, EfOM fractions differ from SRFAin nitrogenized function and aromaticity due to the high content ofproteinaceous structures.

3.3. Influence of DOM on copper toxicity

Figs. 3a and 3b show the dose-effect curves obtained for theHPO media and EC50 expressed in total (EC50tot) and labile(EC50lab) copper (measured by DGT), obtained for all media. TheEC50tot and EC50lab are similar for the inorganic media (Mont-Doremineral water without DOM) given that copper inorganic com-plexes are totally labile with DGT techniques. The EC50tot for or-ganic media are definitely higher than the EC50tot of inorganicmedia, thereby highlighting the protective role of DOM. TheEC50tot obtained for EfOM are not statistically different from the

598 B. Pernet-coudrier et al. / Chemosphere 73 (2008) 593–599

EC50tot for SRFA, which indicates a similar and real protective ef-fect of EfOM fractions, as compared to natural hydrophobic DOM.The labile toxic concentration (EC50lab) for HPO and SRFA mediaare not statistically different from the toxic labile concentration(EC50) obtained in inorganic media. For HPI and TPI media, the la-bile toxic copper concentration is slightly higher statistically. In thepresent case, this feature reflects the relative effectiveness of DGTmeasurements in approximating copper toxicity within such med-ia. Buzier et al. (2006) also observed toxic concentrations in labilecopper (EC50lab) in the both raw and treated effluent of the sameWWTP higher than those obtained in inorganic media. Other stud-ies have also shown that, depending on the nature of DOM, DGTmeasurements could overestimate copper toxicity (Luider et al.,2004; Tusseau-Vuillemin et al., 2004; Divis et al., 2007).

Biologically determined constant values of 9.9, 11.3 and9.8 L mgC�1 are derived for EfOM fractions HPO, TPI and HPI,respectively. The biologically determined constants determinedfor isolated EfOM fractions are quite similar to the SRFA biologi-cally determined constant (9.5 L mgC�1), which underscores thereal protective effect of EfOM. A similar constant (8.4 L mgC�1)was found for the raw treated effluent from the same WWTP withdata from Buzier et al. (2006). The biologically determined con-stants of isolated fractions are slightly higher than that obtainedfor raw complete treated effluent of the same WWTP. This findingcan be explained by a competition effect relative to the toxicityand/or to the copper complexation with DOM since data of Buzieret al. (2006) were produced at higher major cations concentrationslevel. This finding could also underline that the DOM isolation pro-cedure does not modified EfOM and moreover presents the advan-tage of separating DOM from inorganic salts and inorganicpollutants. De Schamphelaere et al. (2004) and Kramer et al.(2004) demonstrated that SUVA at 350 nm was a good proxy ofbiologically (Daphnia magna) relevant differences in copper com-plexing properties of natural DOM. In this study, regardless thewavelength being considered (254 or 350 nm), SUVA can not beused to explain the observed copper toxicity on Daphnia magnain presence of EfOM because the protective effect of DOM appearsto be similar at the same DOC concentration whereas SUVA de-creases strongly (two-to-fivefold) between natural hydrophobicDOM (SRFA) and EfOM depending on the polarity criteria (seeTable 1). According to characterization results of EfOM, proteinsand polysaccharides, e.g. substances with low SUVA, are suspectedto shown metal binding properties. Lamelas et al. (2005) also havedemonstrated that some fraction of DOM (polysaccharide, espe-cially alginates) may exhibit complexing properties and a protec-tive role, with respect to the organism, that are comparable tothose of fulvic acids at similar concentrations. In this context, thedevelopment of site-specific water quality criteria should take intoaccount not only the fulvic and humic acids, but others substancesas well (proteins, polysaccharide, etc.), which exhibit a low SUVAbut yet still could play an important role in metal complexation.In this respect, the introduction of correction factors based onNOM aromaticity measurement (specific absorbance coefficientor SUVA values), to account for NOM quality, may only be appro-priate when the NOM is dominated by humic substances. Futureresearch is needed to determine the extent and conditions withwhich EfOM can exert a major impact in attenuating metal toxicity.

4. Conclusion

According to a RO/DAX protocol, the DOM of treated wastewa-ter was extracted and fractionated into three groups depending onpolarity property (HPO, TPI and HPI). Hydrophilic DOM appeared tobe the most significant fraction in WWTP discharge, as opposed tothe natural aquatic ecosystem where hydrophobic DOM prevails.

The quantitative isolation and purification of DOM (without salts)proved difficult to achieve due to the high proportion of HPI frac-tion. Various analytical techniques (elemental, spectroscopic: UVand FTIR) made it possible to highlight the weak aromaticity andhigh rate of proteinaceous structures in the whole EfOM comparedto natural hydrophobic DOM. Concerning copper toxicity, resultsshow a similar protective role of treated wastewater and naturalDOM to the Daphnia magna organism. Substances with low SUVA,e.g. protein or polysaccharides, are strongly suspected to play a siz-able role in the copper complexation by EfOM. Further character-ization of DOM and an evaluation of complex stability constantwill be soon performed in order to better understand the effectof DOM (both urban and natural) on trace metal speciation andbioavailability in within an urbanized aquatic system.

Acknowledgement

The authors would like to thank the SIAAP (Syndicat Interdé-partemental pour l’Assainissement de l’Agglomération Parisienne)for making the sampling site accessible, as well as Mohamed Saad,Isabelle Chardon and Mathieu Cladière for their technical assis-tance, Daniel Stadtmüller and Emmanuelle Kuhn for the chemicalanalyses during toxicity testing, Jean-Philippe Croué and DavidViolleau for their help in DOM fractionation, France’s Ministry ofResearch and Higher Education for its financial support in the formof a Ph. D. Grant awarded to Benoı̂t Pernet-Coudrier. This work hasalso been supported by the French National Research Agency (aspart of project BIOMET JC05_59809).

Appendix. List of acronyms

BLM biotic ligand modelDGT diffusive gradient filmDOC dissolved organic carbonDOM dissolved organic matterEC50 lethal concentration for 50% of the populationEfOM wastewater effluent dissolved organic matterFIAM free ion activity modelFTIR fourier transform infraredGFAAS graphite-furnace atomic absorption spectroscopyHPI hydrophilicHPO hydrophobicHS humic substancesICP-AES inductively coupled plasma atomic emission spectroscopyRO reverse osmosisSRFA Suwannee river fulvic acidSUVA specific ultra violet absorbanceTPI transphilicWWTP wastewater treatment Plant

References

Andrew, R.W., Biesinger, K.E., Glass, G.E., 1977. Effects of inorganic complexing onthe toxicity of copper to Daphnia magna. Water Res. 11, 309–315.

Barber, L.B., Brown, G.K., Kennedy, K.R., Leenheer, J.A., Noyes, T.I., Rostad, C.E., Thorn,K.A., 1997. Organic constituents that persist during aquifer storage and recoveryof reclaimed water in Los Angeles County, California. Conjunctive use of waterresources: aquifer storage and recovery, In: Proceedings of the American WaterResources Association Symposium, October 19–23, Long Beach, California, pp.261–271.

Buffle, J., 1988. Complexation Reactions in Aquatic Systems: An AnalyticalApproach. Ellis Horwood, New York.

Buffle, J., Deladoey, P., Greter, F.L., Haerdi, W., 1980. Study of the complex formationof copper (II) by humic and fulvic substances. Anal. Chim. Acta 116, 255–274.

Buzier, R., Tusseau-Vuillemin, M.-H., Mouchel, J.-M., 2006. Evaluation of DGT as ametal speciation tool in wastewater. Sci. Total. Environ. 358, 277–285.

Croué, J.F., 2004. Isolation of humic and non-humic NOM fractions: structuralcharacterization. Environ. Monit. Asses. 92, 193–207.

Croué, J.P., Benedetti, M.F., Violleau, D., Leenheer, J.A., 2003. Characterization andcopper binding of humic and nonhumic organic matter isolated from the South

B. Pernet-coudrier et al. / Chemosphere 73 (2008) 593–599 599

Platte river: evidence for the presence of nitrogenous binding site. Environ. Sci.Technol. 37, 328–336.

Davison, W., Zhang, H., 1994. In situ speciation measurements of trace componentsin natural waters using thin-film gels. Nature 367, 546–548.

De Schamphelaere, K.A.C., Janssen, C.R., 2002. A biotic ligand model predicting acutecopper toxicity for Daphnia magna: the effects of calcium, magnesium, sodium,potassium and pH. Environ. Sci. Technol. 36, 48–54.

De Schamphelaere, K.A.C., Vasconcelos, F.M., Tack, F.M.G., Allen, H.E., Janssen, C.R.,2004. Effect of dissolved organic matter source on acute copper toxicity toDaphnia magna. Environ. Toxicol. Chem. 23, 1248–1255.

Dignac, M.-F., Derenne, S., Ginestet, P., Bruchet, A., Knicker, H., Largeau, C., 2000.Determination of structure and origin of refractory organic matter in bio-epurated wastewater via spectroscopic methods. comparison of conventionaland ozonation treatments. Environ. Sci. Technol. 34, 3389–3394.

Divis, P., Docekalova, H., Brulak, L., Pavlis, M., Hekera, P., 2007. Use of the diffusivegradients in thin films technique to evaluate (bio)available trace metalconcentrations in river water. Anal. Bioanal. Chem. 387, 2239–2244.

Drewes, J.E., Croué, J.P., 2002. New approaches for structural characterization oforganic matter in drinking water and wastewater effluents. In: 2nd WorldWater Congress: Drinking Water Treatment, 1–10.

Imai, A., Fukushima, T., Matsushige, K., Hwan Kim, Y., 2001. Fractionation andcharacterization of dissolved organic matter in a shallow eutrophic lake, itsinflowing rivers and other organic matter sources. Water Res. 35, 4019–4028.

Imai, A., Fukushima, T., Matsushige, K., Kim, Y.-H., Choi, K., 2002. Characterization ofdissolved organic matter in effluents from wastewater treatment plants. WaterRes. 36, 859–870.

Isnard, P., Flammarion, P., Roman, G., Babut, M., Bastien, P., Bintein, S., Essermeant,L., Ferard, J.F., Gallotti-Schmitt, S., Saouter, E., 2001. Statistical analysis ofregulatory ecotoxicity tests. Chemosphere 45, 659–669.

Kramer, K.J.M., Jak, R.G., van Hattum, B., Hooftman, R.N., Zwolsman, J.J.G., 2004.Copper toxicity in relation to surface water-dissolved organic matter: Biologicaleffects to Daphnia magna. Environ. Toxicol. Chem. 23, 2971–2980.

Lamelas, C., Wilkinson, K.J., Slaveykova, V.I., 2005. Influence of the composition ofnatural organic matter on Pb bioavailability to microalgae. Environ. Sci. Technol.39, 6109–6116.

Leenheer, J.A., 1981. Comprehensive approach to preparative isolation andfractionation of dissolved organic carbon from natural waters andwastewaters. Environ. Sci. Technol. 15, 578–587.

Leenheer, J.A., Croué, J.-P., 2003. Characterized aquatic dissolved organic matter.Environ. Sci. Technol. 37, 19–26.

Leenheer, J.A., Croué, J.F., Benjamin, M., Korshin, G.V., Hwang, C.J., Bruchet, A., Aiken,G.R., 2000. Comprehensive isolation of natural organic matter from water forspectral characterizations and reactivity testing. ACS Symposium Series 76,Washington, DC, pp. 68–83.

Leenheer, J.A., Rostad, C.E., 2004. Fractionation and characterization of organicmatter in wastewater from a Swine waste-retention basin. ScientificInvestigations Report 2004–5217. US Department of the Interior and USGeological Survey.

Luider, C.D., Crusius, J., Playle, R.C., Curtis, P.J., 2004. Influence of natural organicmatter source on copper speciation as demonstrated by Cu binding to fish gills,by ion selective electrode and by DGT gel sampler. Environ. Sci. Technol. 38,2865–2872.

Ma, H., Allen, H.E., Yin, Y., 2001. Characterization of isolated fractions of dissolvedorganic matter from natural waters and a wastewater effluent. Water Res. 35,985–996.

Mantoura, R.F.C., Dickson, A., Riley, J.P., 1978. The complexation of metals withhumic materials in natural waters. Estuar. Coast. Mar. Sci. 6, 387–408.

Martin-Mousset, B., Croué, J.P., Lefebvre, E., Legube, B., 1997. Distribution etcaractérisation de la matière organique dissoute d’eaux naturelles de surface.Water Res. 31, 541–553.

Meador, J.P., 1991. The interaction of pH, dissolved organic carbon and total copperin the determination of ionic copper and toxicity. Aquat. Toxicol. 19, 13–32.

Morel, F.M., 1983. Complexation: trace metals and microorganisms. In: Principles ofAquatic Chemistry. Wiley Interscience, New York.

Peuravuori, J., Ingman, P., Pihlaja, K., Koivikko, R., 2001. Comparisons of sorption ofaquatic humic matter by DAX-8 and XAD-8 resins from solid-state 13C NMRspectroscopy’s point of view. Talanta 55, 733–742.

Peuravuori, J., Lehtonen, T., Pihlaja, K., 2002. Sorption of aquatic humic matter byDAX-8 and XAD-8 resins: comparative study using pyrolysis gaschromatography. Anal. Chim. Acta 471, 219–226.

Sarathy, V., Allen, H.E., 2005. Copper complexation by dissolved organic matter fromsurface water and wastewater effluent. Ecotox. Environ. Safe. 61, 337–344.

Stevenson, F.J., 1994. Humus Chemistry – Genesis, Composition, Reactions. JohnWiley & Sons, New York.

Tessier, A., Turner, D.R., 1995. Metal Speciation and Bioavailability in AquaticSystems. John Wiley & Sons, Chischester.

Town, R.M., Filella, M., 2000. A comprehensive systematic compilation ofcomplexation parameters reported for trace metals in natural waters. Aquat.Sci. 62, 252–295.

Tusseau-Vuillemin, M.-H., Gilbin, R., Bakkaus, E., Garric, J., 2004. Performance ofdiffusion gradient in thin films to evaluate the toxic fraction of copper toDaphnia magna. Environ. Toxicol. Chem. 23, 2154–2161.

Violleau, D., 1999. Intérêt du fractionnement et de l’extraction des matièresorganiques naturelles d’eaux de surface pour l’étude de leur propriétésstructurales et de leur pouvoir complexant vis-à-vis du cuivre, Thesis report.Laboratory of water chemistry, University of Poitiers, France.

Zhang, H., Davison, W., 1995. Performance-characteristics of diffusion gradients inThin-films for the in situ measurement of trace-metals in aqueous-solution.Anal. Chem. 67, 3391–3400.