Environmental Chemistry

BINDING INTERACTIONS OF 1-NAPHTHOL WITH DISSOLVED ORGANICMATTER OF LAKE BIWA AND TREATED SEWAGE WASTEWATER:

THE ROLE OF MICROBIAL FULVIC ACID

TANVEER AHMED, KEIICHI OHTA, OSAMU NAGAFUCHI, and MASAHIRO MARUO*Department of Ecosystem Studies, School of Environmental Science, The University of Shiga Prefecture, Shiga, Japan

(Submitted 15 December 2011; Returned for Revision 15 January 2012; Accepted 28 May 2012)

Abstract—The binding interactions of 1-naphthol with effluent and whole natural dissolved organic matter (DOM) samples wereanalyzed by using a fluorescence quenching technique. Nonfractionated DOM samples from Lake Biwa (Japan), creek water, treatedsewage effluents, and an extracted Lake Biwa fulvic acid (LBFA) standard were used as quenchers and compared at the same 1-naphtholwith DOM organic carbon ratios found for low natural dissolved organic carbon (DOC) levels (�4.5mg/L). Natural and effluentDOM (eDOM) samples were characterized by the DOC level, relative hydrophobicity (RH%), ultraviolet (UV)-visible absorbanceand fluorescence excitation emission spectroscopy. These parameters were compared with those of the reference LBFA standard.Concave-up Stern-Volmer plots accounted for both the partitioning and the adsorptive binding in the eDOM–polycyclic aromatichydrocarbons (PAH) system as compared with the nonspecific partitioning in the natural DOM-PAH system. Strong linear regressions(r2> 0.80) between the log KDOC values, the RH%, the UV absorbances, and the Fl340-435-UV340 indices for the structural compositionand molecular weights of the DOM samples were obtained. These results suggest that low molecular weight microbial fulvic acid(<800 Da) is dominant in the eDOM-PAH binding interactions, as well as in the distinct molecular structure of the eDOM samples,which resulted in fivefold to sixfold higher binding magnitude for 1-naphthol than for the other samples. Environ. Toxicol. Chem.# 2012 SETAC

Keywords—Fluorescence quenching Dissolved organic carbon 1-naphthol Treated wastewater Microbial fulvic acid

INTRODUCTION

The binding interactions of nonionic organic contaminantswith dissolved organic matter (DOM) and their magnitude(KDOC, often referred to as the partitioning coefficient) are wellstudied because of their influence on the environmental fate,mobility, toxicity, and distribution of nonionic organic con-taminants in the soil, water, and sediments. Recently, DOMfrom treated wastewater has drawn much interest because ofdifferences in structure and characteristics as compared withnatural DOM. The differences arise from the DOM sources andtreatment, and this could further influence the fate of polycyclicaromatic hydrocarbons (PAHs) in the soil and in aquaticsystems [1–5]. Many aspects of DOM-PAH binding interac-tions, such as the type of binding isotherm, the hydrophobicityof the PAHs, the solution chemistry, and the physicochemicalproperties of DOM, including its source, polarity, molecularweight, aromatic, aliphatic contents, and sorption domainshave been thoroughly addressed elsewhere [3,6–11]. However,several contentious points remain regarding the comparativeanalysis of different DOM-PAH binding coefficients (KDOC),because of their diversity in physiochemical characteristics,molecular structure, composition, and weight distribution inaquatic systems [9]. Mainly, two factors make resolving theseissues difficult. The DOM samples used in these studies origi-nated from various sources, and the different analytical and

fractionation procedures used to isolate DOM from soil–aquaticsystems may affect the structural and functional properties ofDOM, which in turn may alter the DOM-PAH binding inter-actions. Although the former results from natural formation anddistribution of DOM in aquatic systems, the latter effects couldbe minimized by using nonfractionated whole natural DOMsamples at low, environmentally referenced natural dissolvedorganic carbon (DOC) levels. This approach may also allow thecomparative and analytical assessment of DOM-PAH bindinginteractions in relation to different DOM functionalities.

In recent studies, effects of DOM aromaticity and molecularsize distribution have been investigated to elucidate the mech-anistic relationship between the physicochemical properties ofPAHs and those of aquatic and commercial humic substances,by using concentrated, fractionated aquatic humic substances[11,12]. Nevertheless, controversies are ongoing in light ofrecent findings [13]. Currently, humic substances are viewedas collections of diverse, relatively low molecular weight(LMW) components, which form dynamic associations thatare stabilized by hydrophobic interactions and hydrogen bonds.High-performance size-exclusion chromatography (HPSEC)has shown that the humic fragments are aggregates of organicsmall molecules, which contain both hydrophobic and hydro-philic segments [14]. In contrast, the apparent molecular sizedistribution of fulvic acid showed a minimal change overvarious solution conditions, compared with humic acid. Sol-utions of humic substances were also analyzed by using amultidimensional nuclear magnetic resonance spectroscopictechnique, and aggregation and disaggregation behavior wasobserved, which is consistent with other studies [14,15]. Whenthe effect of concentration was investigated, concentrated sol-utions of peat humic acid (100mg/ml) contained aggregateswith diffusivities corresponding to an average molecular weight

Environmental Toxicology and Chemistry# 2012 SETAC

Printed in the USADOI: 10.1002/etc.1934

This paper was presented at the International Conference on Environ-mental Pollution and Remediation (ICEPR) in August 2011 and was amonga group of papers the conference committee selected for submission toEnvironmental Toxicology and Chemistry.

* To whom correspondence may be addressed(maruo@ses.usp.ac.jp).

Published online 3 July 2012 in Wiley Online Library(wileyonlinelibrary.com).

1

greater than 66,000 Da, whereas a less concentrated solution(5mg/ml) contained aggregates with diffusivities correspond-ing to an average molecular weight between 2,500 and 6,100Da. The diffusivities of the spectra of fulvic acid solutions alsosuggested a typical average molecular weight of approximately1,000 Da and provided little evidence of aggregation [15],which is again consistent with previous studies [14]. Morerecently, fulvic acid isolate was separated into three sizefractions to determine the elemental composition by size-exclusion chromatography coupled to electrospray ionizationFourier transform ion cyclotron resonance mass spectrometry[16], and the results suggested that high-molecular-weight(HMW) fulvic acid is formed from LMW fulvic acid. In thestudy, the authors concluded that HMW fulvic acid is anaggregate held together by the electrostatic interaction of thecarboxylate groups through hydrogen bonds and polyvalentcations or by hydrophobic interactions of their carbon back-bone, or that HMW fulvic acid consists of LMW fulvic acidscovalently bound to each other or to aliphatic alcohols [16].Therefore, the use of nonfractionated in situ DOM samples atlow DOC levels for investigating DOM-PAH binding inter-actions may elucidate the cumulative effects of DOM micro-components on the supramolecular association of DOM-PAHcomplexes.

In addition, the effects of autochthonous hydrophilic frac-tions derived from aquatic biota on the DOM-PAH bindinginteractions have rarely been studied because of the lack ofisolation techniques [9]. The use of effluent and whole naturalDOM samples for the DOM-PAH binding study should eluci-date the competitive influence of these DOM subcomponents onthe DOM-PAH binding relationship. More recently, it has beenshown that the variable PAH concentration affects the DOM-PAH binding mechanism, which resulted in nonlinear bindingat high PAH concentration. The eDOM-pyrene distributioncoefficients have been shown to vary between 3.5- and seven-fold for various pyrene concentrations [5]. Hence, a realisticcomparison of the different DOM-PAH binding magnitudesmight be difficult at varying PAH concentrations and molecularsizes. Thus, similar DOM:PAH organic carbon ratios shouldallow us to compare the functionality of different natural andeffluent DOM samples and help to quantify the limits ofcompetitive dual-mode binding interactions (partitioning/adsorptive mechanism) in terms of DOM:PAH effective carbonratios [17].

In the present study, structural and functional diversity ofdifferent DOM samples were compared to draw a mechanisticrelationship with their PAH binding coefficients (KDOC), usingsimilar PAH:DOM organic carbon ratios at a low in situ DOClevel for the natural, eDOM and extracted fulvic acid standard.Nonfractionated natural, eDOM, and extracted fulvic acidstandard were characterized by total DOC, HPSEC analysis,specific ultraviolet (UV) absorbance, and fluorescence spectro-scopy. Fluorescence quenching technique was used to deter-mine the binding coefficients (KDOC) of the DOM samples withan LMW PAH, namely, 1-naphthol, which is a U.S. Environ-mental Protection Agency (U.S. EPA) priority pollutant. It iswidely used as an agricultural chemical, and it is a commondegradation product of naphthalene and other PAHs. Therelationship of the optical characteristics and molecular sizedistributions of the DOM samples to the binding coefficient(KDOC) was also evaluated, to reveal the mechanistic relation-ships in the DOM-PAH interactions. Thus, the present studyaims to: (1) compare the different chemical and optical char-acteristics of whole natural, eDOM, and an extracted fulvic acid

standard, and examine how they affect different functionalitiesof DOM in aquatic systems, and (2) elucidate how the changesin DOM functionalities influence the magnitude andmechanismof PAH-DOM binding formation at the same PAH:DOMorganic carbon ratios.

MATERIAL AND METHODS

Reagents and chemicals

The DOM-PAH binding interactions were studied by using aweakly polar, LMW PAH, 1-naphthol. Solutions of 1-naphthol(1.0mM) were prepared by dissolving high purity 1-naphthol(>99% pure, Aldrich Chemical) in methanol (fluorometricgrade, �99%, Dojindo). The solution was transferred to glassbottles wrapped in aluminum foil and stored at 58C. The finalconcentration of methanol was kept at 0.01% (v/v) in the batchexperiments; this concentration does not alter the fluorescenceof 1-naphthol. A phosphate buffer stock solution (1M,KH2PO4/K2HPO4, Wako Pure Chemical) at pH 7.0 was used.

Dissolved organic matter sample collection and preparation

The Lake Biwa samples were collected at a depth of 10m(pelagic zone, maximum depth¼ 90m). Lake Biwa is thelargest freshwater lake in Japan and is located in Shiga Pre-fecture, northeast of the former capital city of Kyoto. The lake ischaracterized as mesotrophic and serves as the major source ofdrinking water and industrial water for 14 million people in theKinki region. Samples were collected from a creek fed by waterfrom the Shiga Prefectural Fisheries Experimental Station,which also mainly uses water from the Lake Biwa system.Water used to prepare the eDOM samples was collected fromthe end of the pipe from the Tohokubu wastewater treatmentplant reservoir, and from the final outfall near the Lake Biwasystem, which is 1.6 km away from the main reservoir. Inaddition, a commercially extracted Lake Biwa fulvic acidstandard (LBFA), supplied by the Japanese Humic SubstanceSociety, was used in the present study. The whole natural andeffluent DOM samples, which were defined as containingparticles less than 0.7mm in size, were prepared by filteringthe water samples on the day they were collected, using aprecombusted (4508C for 5 h) filter (0.7mm, GF/F glassfiber; Whatman). The whole natural DOM samples werealso preconcentrated using a freeze-drying technique (Taitec,VD-800F, freeze dryer) to achieve an optimum DOC level(�5mg/L). The DOC level of the samples was kept in the samerange (3–5.5mg/L). The LBFA standard (Japanese HumicSubstance Society) was used without further purification,and a stock solution (10mg/L) was prepared by diluting thestandard with aqueous buffered solution (pH 7.0, 1mMKH2PO4, and 1mMK2HPO4). All of the DOM samples werekept in the dark at 48C. The collected samples are termed freeze-dried Lake Biwa DOM (LBFD), freeze-dried creek waterDOM (CreekFD), LBFA, Tohokubu wastewater treatment plantat the discharge point (TWW), and Tohokubu wastewatertreatment plant reservoir at the end of pipe (TPWW).

Laboratory methods

The DOC was determined using a total organic carbonanalyzer (Shimadzu TOC-5000, catalyzed combustion) as pre-viously described [18]. The UV absorbances at 254, 340, and365 nm were measured using a double beam spectrophotometer(Shimadzu UV-1600), and the aromatic content of DOM sam-ples was interpreted as previously described [12]. A quartz cell(1 cm) was used to measure the UV absorbance and to correct

2 Environ. Toxicol. Chem. 31, 2012 T. Ahmed et al.

inner filter effects [10]. The fluorescence intensity was deter-mined on a spectrofluorophotometer (Shimadzu RF-1500 DR-15) equipped with a xenon arc lamp (150W), and a cuvettemagnetic stirring system containing Teflon micro stir bars(9.5mm� 7.5mm). The excitation and emission monochroma-tor slit width was 10 nm. The fluorescence intensities weremeasured at the excitation/emission wavelengths of 240/422 nm, 340/435 nm, and 380/437 nm for the natural, effluent,and reference DOM samples. The identification, application,and coherence of these distinctive fluorescence peaks for differ-ent compositions and functions of DOM have been reported[19,20]. For further clarification of the characteristic fluorescentpeaks, initial experiments on the collected samples were carriedout using fixed scan excitation spectra at 5-nm intervals over anemission wavelength range of 300 to 500 nm and an excitationwavelength of 200 to 400 nm. The fluorescent intensities werestandardized using a quinine sulfate standard to eliminateinstrumental drift; an intensity of 16.6 was equivalent to 1quinine sulfate unit (1mg/L in 0.1M H2SO4). The fluorescenceintensities were normalized using the DOC concentrations ofthe respective samples and were expressed as a specific fluo-rescence intensity for comparative analysis. The relative hydro-phobicity (RH%) of the DOM samples was measured using areversed-phase C18 column and a mass balance calculation ofthe bulk DOM DOC (%) contents as described previously [21].

The molecular weight distribution was obtained by using anHPSEC system (Shimadzu) equipped with both UV (254 nm)and fluorescence (340/435 nm) detectors, and separationcolumn (Superose 12 10/300 GL, GE Healthcare) designedfor high-performance size exclusion. Mobile phase (NaOH,0.01M) was passed at a flow rate of 0.4ml/min. The poly-dispersity (r), number (MN), and weight-averaged (MW) molec-ular weight were determined for the DOM samples aspreviously described [12]. A fluorescence-UV index basedon the excitation/emission peaks at 340/435 nm, normalizedto the absorbance at 340 nm, was applied to elucidate themolecular size distribution based on optical characteristics[19,22].

The DOM-PAH binding coefficients (KDOC) were measuredusing a fluorescence quenching technique based on Stern-Volmer plots, which describes the static quenching of a fluo-rophore [10]. This technique was adopted because it allows fastmeasurement of the KDOC, and it is very sensitive to structuraland functional changes in DOM [18,23]. In addition, thefluorescence quenching of 1-naphthol by DOM was attributedto static quenching by previous studies that explored theinfluence of temperature and diffusion and fluorescence effi-ciency calculations [24]. Stock solutions of 1-naphthol, DOMsamples and Milli-Q water were adjusted to pH 7.0 and 0.01Mionic strength with a KH2PO4/K2HPO4 buffer solution. Thequenching of 1-naphthol by the whole natural DOM, eDOM,and fulvic acid standard samples was quantified using threereplicates at four concentrations. An evaluation of the 1-naph-thol partitioning kinetics in the unfractionated DOM samplesand the fulvic acid standard was conducted to determine asuitable equilibrium time for the effective 1-naphthol-DOMorganic carbon ratio in the quenching experiment. The parti-tioning of 1-naphthol in the effluent and whole natural LakeBiwa DOM samples was complete in 12 h, as previouslyreported [24], whereas 20 h was required for the equilibrationof the creek DOM sample and the extracted fulvic acid standard.Therefore, an equilibration time of 24 h was chosen forthese reactions. Batch experiments were carried out in Pyrexglass vials (20ml) with Teflon screw caps. No adsorption of

1-naphthol on the Pyrex glass walls was observed. For thequenching experiments with the eDOM samples, differenteDOM concentrations were obtained by diluting the stock withMilli-Q water containing sodium azide (50mg/L). The sameconcentration of sodium azide was also used in the blanks. In allof the DOM batch experiments, pH values between 7.0 and 7.5and an effective 1-naphthol-DOM organic carbon ratio from0.01 to 0.04 were maintained. The fluorescence emission wasrecorded using the same spectrofluorometer as described pre-viously, at excitation/emission wavelengths of 294/468 nm,respectively, with slit widths of 10 nm. The fluorescence inten-sity was measured in a quartz cell (1.0 cm), prewashed with thesample, onto which no adsorption of 1-naphthol was observed.The fluorescence intensities were obtained by averaging every10 sample readings. The resulting fluorescence intensities werealso corrected for the inner filter effect, as previously described[10]. The correction factors for different DOM samples werenegligible and never exceeded 1.05. In addition, Cl� did notaffect the 1-naphthol fluorescence intensity at the chlorideconcentration (�70–80mg/L) in the eDOM samples.

The percentage of the bound 1-naphthol in different DOMsamples was calculated using the equation derived for thesolubility enhancement technique that has previously been usedto calculate KDOC [19]. The similarity of the apparent solutesolubility to the fluorescence quenching as well as the similarityof the solute liquid–solid partitioning coefficient to the Stern-Volmer constants in the presence of DOM samples allow theamount of 1-naphthol bound to DOM to be measured [11,25].The calculation of the bound 1-naphthol contents will thushelp to elucidate the role of different DOM samples on thequantitative distribution of 1-naphthol in aquatic systems.

RESULTS AND DISCUSSION

General and optical characterization of DOM samples

The physicochemical and optical characteristics of the DOMsamples are detailed in Table 1. The in situ pH of the wholenatural and eDOM solutions were between 6.5 and 7.8(Table 1). The LBFD and CreekFD samples represent theDOC level in the preconcentrated solution, which wererecorded before preconcentration as 1.0 and 1.3mg/L, respec-tively. The UV absorbance at 254 nm and the fluorescenceexcitation/emission peaks were expressed as the specific UVabsorbance at 254 nm (SUVA254) and the specific fluorescenceintensity as normalized to the DOC. The SUVA values at254 nm are often used to characterize the aromatic structuresin the fractionated DOM samples, which are independent ofthe DOC concentration [19,26,27]. The SUVA254 values forthe DOM samples were between 1.10 and 2.12 L/(mgC. m).The SUVA254 data suggested that the aromatic moieties werehydrophilic, with the highest value for the LBFA sample[28,29]. Based on previous studies, two distinct fluorescencepeaks were identified and used to characterize the DOM sam-ples [19,21]. The fluorescence peak with an excitation/emissionat 340/435 nm is known as the fulvic-like peak (Fig. 1A), andthe emission peak at 340 nm with an excitation at 220 nm iscalled the tryptophan-like peak (Fig. 1B). The tryptophan-likepeak relates to both the algal and microbial organic matter [19].For the whole natural and effluent DOM samples, the SUVA254

values correlated strongly (r¼þ0.95, p< 0.05) with the fulvic-like peaks, suggesting a very large contribution of the fulvicmolecules to the whole aromatic content of the DOM samples.In contrast, a high SUVA254 value was associated with a smallfulvic-like peak, which suggests a significant loss of fulvic

Napthol-Binding Complexes Environ. Toxicol. Chem. 31, 2012 3

components from the LBFA standard samples, which has alsobeen previously observed [30,31]. The LBFA samples also hadthe lower McKnight Index values because of their higherfluorescence intensity at a longer wavelength compared withthe LBFD or CreekFD samples, which was contrary to theexpectation. McKnight proposed this index to distinguish themicrobially derived fulvic acids from the terrestrially derivedfulvic acids. McKnight index is defined as the ratio of fluo-rescence emission intensity at wavelength 450 nm to that at500 nm (or E450/E500) with an excitation at 370 nm. Becausethe microbially derived organic materials generally contain lowfluorophores and a low aromaticity, they give low fluorescenceintensity at longer a wavelength compared with the terrestriallyderived organic materials [32]. Typically McKnight indicesranging from 1.9 to 2.5 suggested the dominance of microbialfulvic acids in the LBFD, CreekFD, TPWW, TWW, and DOMsamples, respectively (Fig. 2) [32,33]. On the contrary, aMcKnight index value of 1.59 coupled with a high SUVA254

value of 2.12 and a small fulvic-like peak suggested thedominance of terrestrially derived fulvic acids in the LBFAsample [33]. However, the characteristic McKnight Index andSUVA254 value for the LBFA sample could also be influencedby the highly synthesized refractory organic matter of bio-logical origin. The high SUVA254 value of the LBFA samplealso may reflect the enrichment of other aromatic or polyphe-nolic components from allochthonous sources [28,30] in thetotal aromatic pool. To compare the ratio of microbial fulvicmolecules to the terrestrial fulvic component of the DOMsamples, the fluorescence emission ratio (E450/E500) of 450/500 nm at excitations of 340 nm and 370 nm were recordedand compared with previous values [32,33]. The values ofthe E450/E500 ratio for excitation at 340 nm ranged from 1.94to 2.87, compared with the range of 1.59 to 2.48 for theMcKnight Index (Fig. 2). However, both of the indices were

positively correlated (r¼þ0.95, p< 0.005) for all of the DOMsamples except the LBFA samples, which suggested the dom-inance of autochthonous (microbially derived) fulvic compo-nents in the whole effluent and natural DOM samples. Similarhigh SUVA254 values and the absorbance ratio at 250/365 nm(E2/E3) for the fulvic acid samples have been previouslyreported [19]. A strong positive correlation of the trypto-phan-like peak with the E450/E500 (r¼ 0.95, p< 0.01) and theMcKnight Indices (r¼ 0.90, p< 0.05) again confirmed thatthe algal or microbial fulvic molecules were most dominantin the TWW and TPWW samples, followed by the LBFD,CreekFD, and LBFA samples (Table 2).

In addition to these functional optical parameters, the RH%of the DOM samples were measured by reverse-phase chro-matography, using C18 reversed-phase columns [21]. A goodstatistical correlation of the RH% parameter with polydisper-sity, E450/E500, and the McKnight index of the DOM sampleswas also obtained (Table 2). The RH% data showed theextracted LBFA standard had the greatest hydrophobicity,followed by the CreekFD, TWW, LBFD, and TPWW samples(Fig. 1). The RH% data are in good agreement with the opticalcharacteristics of the DOM samples.

Molecular weight distribution

The HPSEC chromatograms of the whole natural Lake Biwawater, eDOM, and the commercial LBFA standard are shown inFigure 3A and B. The chromatograms for the LBFD, CreekFD,LBFA, TWW, and TPWW samples showed two or three broad,multimodal, well-resolved peaks.

As shown in Figure 3A and B, three poorly resolved majorpeaks were observed in the chromatogram of LBFA sample ascompared with that of LBFD sample. In the case of an LBFAsample, differences in retention time and molecular weightof the first peak (retention time¼ 33.9min, molecular wt¼

Table 1. General and optical properties of the dissolved organic matter samples

DOM source pH DOC (mg/L) SUVA254 (L/[mg carbon m]) SFI for FA component

LBFD 7.62 4.88 1.37 2.16TWW 6.58 3.55 1.69 18.38TPWW 7.23 3.21 1.90 22.65CreekFD 7.80 4.46 1.10 1.61LBFA 6.55 5.32 2.12 5.33

DOM¼ dissolved organic matter; DOC¼ dissolved organic carbon; SUVA254¼ specific UV absorbance at 254 nm; SFI¼ specific fluorescence intensities;FA¼ fulvic acid; LBFD¼ freeze-dried Lake Biwa DOM; TWW¼Tohokubu wastewater treatment plant at the discharge point; TPWW¼Tohokubuwastewater treatment plant reservoir at the end of the pipe; CreekFD¼ freeze-dried creek water DOM; LBFA¼ extracted Lake Biwa fulvic acid standard.

Fig. 1. Fluorescence emission spectrum of different dissolved organic matter (DOM) samples showing (A) ‘‘fulvic-like’’ and (B) ‘‘tryptophan-like’’ peaks.Specific fluorescence intensities (SFI) were shown as fluorescence intensities normalized to dissolved organic carbon (DOC) levels of the DOM sample.LBFD¼ freeze-dried Lake Biwa DOM; CreekFD¼ freeze-dried creek water DOM; LBFA¼Lake Biwa fulvic acid; TTW¼Tohokubu wastewater treatmentplant at the discharge point; TPWW¼Tohokubu wastewater treatment plant reservoir at the end of the pipe.

4 Environ. Toxicol. Chem. 31, 2012 T. Ahmed et al.

1,370 Da) and third peak (37.4min, molecular wt¼ 760 Da)were approximately 3.5min and 610 Da. Conversely, thesedifferences in LBFD sample were approximately 5.5min and1,240 Da (first peak retention time¼ 31.0min, molecularwt¼ 2,160 Da, and third peak retention time¼ 36.5min, molec-ular wt¼ 920 Da). These facts suggested a narrower molecularweight distribution and a more homogenous structural compo-sition of LBFA. Calibration with the polystyrene sodium sul-fonate standards curve allowed the molecular size distributionsof the DOM samples to be classified into three series: large(molecular wt> 2,000 Da), medium (2,000 Da>molecularwt> 1,000 Da), and small (molecular wt< 1,000 Da). TheLBFD samples showed the largest molecular weight distribu-tion; 20% / 33% / 35% (large/medium/small), and the creekFDsamples had a similar molecular weight distribution. However,for the LBFA sample, more than 70% of the molecular weightdistribution was found in the mediummolecular size series. TheTWW sample showed the lowest molecular weight distribution,with 15 and 65% in the medium and small series, respectively.The TPWW samples had a higher molecular weight distributioncompared with TWW; more than 20% of the molecules were in

the medium-sized series and 60% in the small molecular sizeseries. A distinct peak (�800 Da) was observed for the wholeeffluent DOM samples (TWW), which could be attributed tothe high microbial fulvic acid content of the samples [32]. Thecalculation of the peak height in this size range suggested amore than 60% loss of these fulvic acid moieties in the LBFAstandard, compared with the nonfractionated Lake Biwa DOMsamples. The estimatedMW,MN, and r of the DOM samples areshown in Table 3. The DOM samples had a MW moleculardistribution in the range of 830 to 1,260 Da, whereas the MN

molecular distribution was in the range of 669 to 982 Da. The rranged from 1.24 to 1.47 for the DOM samples, which waslower than previously reported values (1.5–1.9) [12]. These datasuggest a relatively narrow molecular weight distribution forthe DOM samples. For further clarification, the SUVA280 and afluorescence-UV (FlEx.340/Em.435/UV340) index was also corre-lated with the molecular weight distribution data of the samples.Both the parameters appeared to be effective for the molecularweight distribution assay of the nonfractionated DOM samples.However, the FlEx.340/Em.435/UV340 indices had a stronger cor-relation (Table 2) than the SUVA280; the HPSEC data showedthe lowest index value for the LBFD sample, which also had thehighest molecular weight distribution. The analysis of thesefunctional assay parameters for molecular weight distributionwas in good agreement with previous findings [19,24].

Determination of 1-naphthol-DOM binding coefficients (KDOC)

The 1-naphthol binding coefficient (KDOC) for the TWW,TPWW, LBFA, CKFD, and LBFD samples were between0.20� 105 and 1.10� 105 L/kg (Table 4). The KDOC valuefor the LBFD sample was the same as has previously beenreported for 1-naphthol/soil humic acid interactions [24]. How-ever, the TWWDOM sample exhibited a sixfold higher bindingaffinity for 1-naphthol compared with the LBFD sample. Thebinding coefficients for 1-naphthol were much lower for thewhole natural and extracted DOM samples than for the eDOMsamples. The correction factors for the primary and secondaryinner filter effects of the DOM samples were negligible, with avalue below 1.10. Linear Stern-Volmer plots were obtained formost of the DOM-1-naphthol binding interactions (r2> 0.93;lack of fit, F> 0.05), and concave-up isotherms for theeDOM:1-naphthol system, which indicates a distinctive bindinginteraction for the eDOM samples. Initially, the observation of anonlinear isotherm suggested the dynamic quenching of 1-naphthol by the DOM samples; however, the influence of thismechanism has already been shown not to be significant [24]. Inthe present study, the bimolecular quenching rate constants forthe eDOM system were calculated as 1.05� 1013/M/s, whichwas much higher than the maximum value expected for adiffusion-controlled process in an aqueous system [10].Recently, an adsorptive interaction has been shown to be mainlyresponsible for eDOM nonlinear binding isotherms at varyingconcentrations of pyrene (0.02–0.10mg/L) using a fluorescencequenching technique [5]. In that study, the role of the non-specific partitioning of pyrene between the aqueous phase andthe bulky DOM phase at higher concentrations was neglected.However, the sensitivity of fluorescence quenching techniquesis low at high PAH concentrations, which affects the reprodu-cibility as well as the proportionality of the fluorescenceintensities in response to the fluorophore (pyrene), while pro-ducing a higher inner filter effect (correction factor 1.50). Toconfirm these findings, a cross-comparative study is required,using solubility enhancement methods for measuring the pyr-ene-DOM binding coefficients, and detailed thermodynamic

Fig. 2. Distribution of McKnight indices, E450/E500, E2/E3, and relativehydrophobicity (RH%)of thedissolvedorganicmatter (DOM)samples.BothMcKnight indices andE450/E500 have the sameemission intensity ratio of 450to 500 nm, with excitation wavelengths of 370 and 340 nm, respectively.LBFD¼ freeze-dried Lake Biwa DOM; Creek Nat¼ freeze-dried creekwater DOM; LBFA¼Lake Biwa fulvic acid; TTW¼Tohokubu wastewatertreatment plant at the discharge point; TPWW¼Tohokubu wastewatertreatment plant reservoir at the end of the pipe.

Table 2. Stepwise linear regression model results from whole naturalmeasurable dissolved organic matter physiochemical and

optical characteristics

AssayID Linear regression model r p

1 SUVA254¼ 0.03 SFIEx.340/Em.435þ 1.17 0.95 0.052 SFIEx.220/Em.340 ¼ 66.18 E450/E500–128.37 0.96 0.013 SFIEx.220/Em.340 ¼ 70.03 McKnight

index – 114.240.89 0.04

4 RH%¼�36.14 E450/E500þ 151 0.83 0.085 McKnight index¼�0.016 Rh (%)þ 3.13 0.82 0.096 RH (%)¼ 149.5 rþ 136.4 0.77 0.127 log Mw¼�4.63� 10�5 Fl340/435–UV340

indexþ 3.090.89 0.04

8 log Mw¼ 0.98 SUVA280þ 1.68 0.94 0.069 log KDOC¼�4.23 log Mwþ 17.38 0.99 >0.00110 log KDOC¼ 0.0002 Fl340/435–UV340 indexþ 4.3 0.89 0.04

SUVA254¼ specific UV absorbance at 254 nm; SFI¼ specific fluorescenceintensities; RH%¼ relative hydrophobicity; UV¼ ultraviolet; MW¼weight-averaged molecular weight.

Napthol-Binding Complexes Environ. Toxicol. Chem. 31, 2012 5

studies of the binding interactions at the same pyrene concen-tration. In contrast, the role of hydrophobic partitioning hasbeen highlighted in other studies for pyrene-DOM bindinginteractions [11], and some other studies have led to thedevelopment of the micellar model of DOM formation [13].In later studies, an initial increase in the fluorescence intensitywas observed because of the cation-enhanced formation ofmicelles, with hydrophobic domains created by charge neutral-ization and cation bridge formation [13,34–36]. A drop in thefluorescence was then observed because the cation bridges werebroken and hydrophobic regions, which participated in thepyrene binding interactions, were disaggregated. More micellarstructures were shown to form at neutral and acidic pH than atbasic pH, and fulvic acid formed micelles faster than humicacid. However, if the adsorptive mechanism, which relies onhydrogen bonding and electrostatic interactions, played a majorrole in pyrene–DOM interactions, then we could expect a rapidinitial decrease in the fluorescence intensity because of thesorptive interaction of pyrene with the hydrophilic boundarylayers of the micelles. An increase in the fluorescence wouldthen follow, arising from the dispersion of the hydrophobicregions of the DOM. However, such nonlinear concave-upisotherms for the binding interaction of a weakly polarLMW PAH such as 1-naphthol with an LMW hydrophiliceDOM are rarely reported. These distinctive binding isothermscan be attributed to the combined effects of partitioning andadsorptive mechanisms, according to the dual-mode adsorptionmodel [17]. The sorptive interactions, via electron donoracceptor interactions or hydrogen bonding, take place in a

limited number of high-energy polar binding sites in the eDOMat a low 1-naphthol/eDOM organic carbon ratio (0.04), andthis resulted in initial fluorescence quenching. Changes in thedielectric environment of 1-naphthol were also observed, andthis resulted in a 5-nm Stokes shift (294/460–294/465) in thecorrected excitation/emission fluorescence spectroscopy of the1-naphthol–eDOM complex system (Fig. 4). A further rapidincrease in the quenching because of an increase in the eDOMconcentration suggests a shift toward the nonpolar partitioningof 1-naphthol in the bulky eDOM phase caused by the saturationof the polar binding sites for the eDOM samples. A 35%decrease in the fluorescence intensity was observed for thefinal 1-naphthol–TWW organic carbon ratio compared with a15% decrease in the intensity for the 1-naphthol- LBFD bindinginteractions (Fig. 5A, B). The influence of this combinedbinding mechanism may not preclude colloidal formation oraggregation from LMWmicrobial fulvic acid in eDOM samplesat a neutral pH and an ionic strength of 0.01M (K2HPO4/KH2PO4 buffer solution) [13]. A detailed investigation of thesame binding interactions under the influence of cation-bridgingor charge neutralization of the DOM functional groups isnecessary. In contrast, nonspecific partitioning played a majorrole in the binding interactions for 1-naphthol with the LBFAand CreekFD samples, which resulted in more linear partition-ing isotherms (Fig. 6, Table 4) because their hydrophobicity waslarger than that of the other samples (Fig. 2).

Environmental significance

On the basis of the resulting DOM-PAH binding coefficientsand the respective DOC levels of the collected samples, wecalculated the percentage of 1-naphthol bound to the differentDOM samples using the equation derived from the Stern-Volmer plots (Table 4). The LMW microbial fulvic acid inthe eDOM samples (TWW, TPWW) exhibited twofold tofivefold higher uptake of 1-naphthol compared with the wholenatural LBFD, CreekFD, and extracted LBFA standard DOMsamples. The adsorptive (hydrophilic) mechanism appearedto play a major role in the dual-mode interactions between1-naphthol and the eDOM system. Although the data showed asignificant loss of these LMW DOM components from theLBFA standard, it showed a 1.5- to twofold higher uptake of1-naphthol than that of whole natural LBFD and CreekFDDOMsamples. Enrichment of the highly aromatic and hydrophobicfractions in the LBFD standard sample appeared to promotenonspecific partitioning of the 1-naphthol between the aquaticand bulky DOM phase of the sample. Therefore, the data for thecommercial fulvic acid standard appeared to overestimate thewhole natural DOM functionality and their binding interactions

Fig. 3. High-performance size exclusion chromatography (HPSEC) spectra for thedissolvedorganicmatter (DOM)samples: (A) Tohokubuwastewater treatmentplant at the discharge point (TWW), Tohokubu wastewater treatment plant reservoir at the end of the pipe (TPWW), freeze-dried Lake Biwa DOM (LBFD);(B) Lake Biwa fulvic acid (LBFA), and freeze-dried creek water DOM (CreekFD).

Table 3. Molecular weight, polydispersity and fluorescence-UV indicescharacterizing the molecular size distribution of the dissolved organic

matter samplesTable 3. Molecular weight, polydispersity andfluorescence-UV indices characterizing the molecular size distribution

of the dissolved organic matter samples

DOMsource MW [Da] MN [Da]

Polydispersityr

Fl-UV indexfor MW

LBFA 1,039 706 1.47 934CreekFD 1,173 842 1.39 556LBFD 1,266 982 1.29 534TPWW 929 723 1.28 3462TWW 830 669 1.24 2999

DOM¼ dissolved organic matter; LBFD¼ freeze-dried Lake Biwa DOM;TWW¼Tohokubu wastewater treatment plant at the discharge point;TPWW¼Tohokubu wastewater treatment plant reservoir at the end ofthe pipe; CreekFD¼ freeze-dried creek water DOM; LBFA¼ extractedLake Biwa fulvic acid standard; MW¼weight-averaged molecular weight;MN¼molecular weight number.

6 Environ. Toxicol. Chem. 31, 2012 T. Ahmed et al.

with 1-naphthol. The large discrepancies between the spectro-scopic and fluorescence characteristics that defined the standardLake Biwa DOM functionalities further narrowed its use as asuitable material for studying the behavior of the nonionicorganic contaminant speciation in aquatic systems. However,the data for eDOM samples suggest that 1-naphthol was highlymobile and easily transported in the watershed area that receivessignificant amounts of treated effluent water. This may posedifficulties for the on-site removal of LMW PAHs associatedwith eDOM in the tertiary treatment stages of wastewaterplants. In addition, the LMW DOM fractions pose problemsparticularly in drinking water treatment plants [37]. AlthoughLMW fulvic acid dominated the major portions of the DOMsample collected from the receiving watershed, they appeared tohave a negligible influence on the DOM samples collected fromthe pelagic zone of the Lake Biwa system. This phenomenoncould be attributed to the following factors: the effect of

Table 4. 1-Naphthol–dissolved organic matter binding coefficients (KDOC) with their statistics, correction factor for inner filter effects, and percentages of1-naphthol bound to different DOM samples

DOM sampleKDOC

(L/kg carbon)SD

(L/kg carbon) r2Lack of fit,p value

Correctionfactor

DOC(mg/L)

1-Naphthol boundto DOM [%]

LBFD 1.95� 104 � 0.10 0.96 > 0.05 1.05 4.88 2.00TWW 1.07� 105 � 0.07 0.93 < 0.05 1.05 3.55 10.00TPWW 7.56� 104 � 1.80 0.93 > 0.05 1.05 3.21 7.00CreekFD 2.42� 104 � 0.33 0.93 > 0.05 1.03 4.46 2.40LBFA 4.21� 104 � 0.07 0.99 > 0.05 1.08 5.32 4.00

DOM¼ dissolved organic matter; DOC¼ dissolved organic carbon; LBFD¼ freeze-dried Lake Biwa DOM; TWW¼Tohokubu wastewater treatment plantat the discharge point; TPWW¼Tohokubu wastewater treatment plant reservoir at the end of the pipe; CreekFD¼ freeze-dried creek water DOM;LBFA¼ extracted Lake Biwa fulvic acid standard.

0

50

100

150

200

250

300

530480430380

Emission (nm)

Fluo

resc

ence

inte

nsity

LBFD1-NaphtholI-Naphthol+LBFD

0

50

100

150

200

250

530480430380Emission (nm)

Fluo

resc

ence

inte

nsity

TWW

1-Naphthol+TWWComplex1-Naphthol

A B

Fig. 4. Fluorescence emission spectra of 1-naphthol, 1-naphthol plus dissolved organic matter (DOM), and DOM alone at the excitation of 294 nm.(A) 1-napthol¼ 0.30mM; Tohokubu wastewater treatment plant at the discharge point (TWW) eDOM¼ 79.16mMC; (B) 1-napthol¼ 0.40mM, freeze-driedLake Biwa DOM (LBFD) DOM¼ 108mMC.

Fig. 5. Fluorescence quenching of 1-naphthol. (A) Concentration (1-naphthol)¼ 0.30mmol/L with the addition of Tohokubu wastewater treatment plant atthe discharge point (TWW) effluent dissolved organicmatter (eDOM) (�102mmol/L dissolved organic carbon [DOC]), d-h: 0, 0.79, 1.18,1.76, 2.95, respectively.(B) Concentration (1-naphthol)¼ 0.41mmol/L with the addition of LBFD DOM (�102mmol/L DOC), d–h: 0, 1.08, 1.62, 2.44, 4.06, respectively.

Fig. 6. Stern-Volmer plots for the fluorescence quenching of 1-naphtholby different DOM samples. TWW¼Tohokubu wastewater treatment plantat the discharge point; TPWW¼Tohokubu wastewater treatment plantreservoir at the end of the pipe; LBFA¼Lake Biwa fulvic acid;CreekFD¼ freeze-dried creek water DOM; LBFD¼ freeze-dried LakeBiwa DOM.

Napthol-Binding Complexes Environ. Toxicol. Chem. 31, 2012 7

dilution, the natural photo degradation of DOM, or the signifi-cant addition of DOM to the lake system from terrestrial sources[38].

CONCLUSION

The distinct binding mechanisms for eDOM:1-naphtholcomplexes were largely attributed to the microbially derivedfulvic acids having LMW distribution in the whole effluentsamples. Satisfactory regression of the two distinct ‘‘fulvic’’ and‘‘tryptophan’’-like fluorescence peaks with the RH% andMcKnight indices also confirmed a greater hydrophilicity ofthe eDOM samples than others. Such microbially synthesizedhydrophilic fulvic acid moieties influenced the sorptive non-linear interactions of eDOM with 1-naphthol, with severalfoldhigher binding magnitudes, which might cause a greater trans-port of such LMW PAHs in the watershed. However, fewdetailed studies have focused on the role of such LMW micro-bial fulvic acids in the transport, distribution, and speciation ofnonionic organic contaminants, because of the lack of conven-ient extraction and analytical techniques. We have used non-fractionated whole natural and eDOM samples to probe the insitu physicochemical characteristics and DOM functionalities atnatural DOC levels. In addition, 1-naphthol was used as a probeto mimic the binding mechanisms of PAHs, which wereaffected by different DOM functionalities in the receiving waterbody at the same PAH:DOM organic carbon ratio. However,1-naphthol is a weakly polar aromatic molecule with a two-ringplanar configuration; it does not represent other classes ofnonionic PAHs containing more aromatic, HMW, or fused ringstructures. In addition, the collected DOM samples had a narrowmolecular weight distribution (LMW) and a homogenous com-position. Conversely, binding interactions of the HMW in situDOM samples with other higher-class PAHs may broadlyelucidated the mechanistic role of DOM functionalities onPAH speciation in the aquatic system. In that context, com-parative analysis of these interactions with their fractionatedcounterparts obtained by advanced separation techniques alsorequires further investigation. Thus, these combined methodsmay aid a holistic view on PAH speciation in natural, organiccarbon-rich waters. Most of the assay parameters in the presentstudy showed sufficiently good statistical agreement to correlatethe different DOM functionalities to the binding interactions for1-naphthol in the receiving water body. Thus, our findings maysignificantly contribute to the laboratory analysis, online mon-itoring, and prediction of whole natural DOM functionalitiesand their PAH binding affinities.

REFERENCES

1. Gauthier TD, Seitz WR, Grant CL. 1987. Effects of structural andcompositional variations of dissolved humic materials on pyrene Koc

values. Environ Sci Technol 21:243–248.2. Salloum MJ, Chefetz B, Hatcher PG. 2002. Phenanthrene sorption by

aliphatic rich natural organic matter. Environ Sci Technol 36:1953–1958.

3. Schlauman MA, Morgan JJ. 1993. Effects of aqueous chemistry on thebinding of polycyclic aromatic hydrocarbons by dissolved humicmaterials. Environ Sci Technol 27:961–969.

4. Rav-Acha C, Rebhun M. 1992. Binding of organic solutes to dissolvedhumic substances and its effect on adsorption and transport in the aquaticenvironment. Water Res 26:1645–1654.

5. Borisover M, Laor Y, Bukhanovsky N, Saadi I. 2006. Fluorescence-based evidence for adsorptive binding of pyrene to effluent dissolvedorganic matter. Chemosphere 65:1925–1934.

6. McCarthy JF, Jimenez BD. 1985. Interactions between polycyclicaromatic hydrocarbons and dissolved humic material: Binding anddissociation. Environ Sci Technol 19:1072–1076.

7. Wang XL, Cook R, Tao S, Xing BS. 2007. Sorption of organiccontaminants by biopolymers: Role of polarity, structures and domainspatial arrangements. Chemosphere 66:1476–1484.

8. Gunasekara A, Simpson MJ, Xing B. 2003. Identification andcharacterization of sorption domains in soil organic matter usingstructurally modified humic acids. Environ Sci Technol 37:852–858.

9. Mei Y, Wu F, Wang L, Bai Y, Liao H, Liao H. 2009. Bindingcharacteristics of perylene, phenanthrene and anthracene to differentDOM fractions from lake water. J Environ Sci 21:414–423.

10. Gauthier TD, Shane EC, Guerin WF, Seitz WR, Grant CL. 1986.Fluorescence quenching method for determining equilibrium constantsfor polycyclic aromatic hydrocarbons binding to dissolved humicmaterials. Environ Sci Technol 20:1162–1166.

11. Chin YP, Aiken GR, Danielsen KM. 1997. Binding of pyrene to aquaticand commercial humic substances: The role of molecular weight andaromaticity. Environ Sci Technol 31:1630–1635.

12. Chin YP, Aiken GR, OLoughlin E. 1994. Molecular weight, poly-dispersity, and spectroscopic properties of aquatic humic substances.Environ Sci Technol 28:1853–1858.

13. SuttonR, SpositoG. 2005.Molecular structure in soil humic substances:The new view. Environ Sci Technol 39:9009–9015.

14. Piccolo A. 2001. The supramolecular structure of humic substances.Soil Sci 166:810–832.

15. Simpson AJ. 2002. Determining the molecular weight, aggregation,structures and interactions of natural organic matter using diffusionordered spectroscopy. Magn Reson Chem 40:S72–S82.

16. ReemtsmaT,TheseA,SpringerA,LinscheldM.2008.Differences in themolecular composition of fulvic acid size fractions detected by size-exclusion chromatography-on line Fourier transform ion cyclotronresonance (FTICR-) mass spectrometry. Water Res 42:63–72.

17. Xing BS, Pignatello JJ. 1997. Dual-mode sorption of low polaritycompounds in glassy poly (venyl chloride) and soil organic matter.Environ Sci Technol 31:792–799.

18. Kopinke FD, Georgi A, Mackenzie K. 2000. Sorption and chemicalreactionsofPAHswith dissolvedhumic substances andmodelpolymers.Acta Hydrochim Hydrobiol 28:385–399.

19. Baker A, Tipping E, Thacker SA, Gondar D. 2008. Relating dissolvedorganic matter fluorescence and functional properties. Chemosphere73:1765–1772.

20. GuoW,Xu J,Wang J,WenY, ZhuoY,YanY. 2010.Characterization ofdissolved organic matter in urban and sewage using excitation emissionmatrix fluorescence spectroscopy and parallel factor analysis. J EnvironSci 22:1728–1734.

21. Herbert BE, Bertech PM, Novak JM. 1993. Pyrene sorption by water-soluble organic carbon. Environ Sci Technol 27:398–402.

22. Stewart AJ, Wetzel RG. 1980. Asymmetrical relationships betweenabsorbance, fluorescence, and dissolved organic carbon. LimnolOceanogr 26:590–597.

23. Burdige DJ, Homstead J. 1994. Fluxes of dissolved organic carbon fromChesapeake Bay sediments. Geochim Cosmochim Acta 58:3407–3424.

24. Chen S, Inskeep WP, Williams SA, Callis PR. 1994. Fluorescencelifetime measurements of Fluoranthene, 1-naphthol, and Napropamide inthe presence of dissolved humic acid.Environ Sci Technol 28:1582–1588.

25. Mott HV. 2002. Association of hydrophobic organic contaminants withsoluble organic matter: Evaluation of the database of KDOC values. AdvEnviron Res 6:577–593.

26. Weishaar JL,AikenGR,BergamaschiBA, FramMS,Fujii R,MopperK.2003. Evaluation of specific ultraviolet absorbance as an indicator ofthe chemical composition and reactivity of dissolved organic carbon.Environ Sci Technol 37:4702–4708.

27. Wang L, Wu F, Zhang R, Li W, Liao H. 2009. Characterization ofdissolved organic matter fractions from Lake Hongfeng, SouthwesternChina Plateau. J Environ Sci 21:581–588.

28. Peuravuori J, Pihlaja K. 1997. Isolation and characterization of naturalorganic matter from lake water; comparison of isolation with solidadsorption and tangential membrane filtration. Environ Int 23:441–451.

29. Croue JP, Benfdetti MF, Violleauand D, Leenheer JA. 2003. Character-ization and copper binding of humic and nonhumic organic matterisolated from the South Platte river, evidence for the presence ofnitrogenous binding site. Environ Sci Technol 37:328–336.

30. FujitakeN,TsudaK,AsoS,KodamaH,MaruoM,YonebayashiK. 2012.Seasonal characteristics of surface water fulvic acids from Lake Biwaand Lake Tankai in Japan. Limnology 13:45–53.

31. Fujitake N, Kodama H, Nagao S, Tsuda K, Yonebayashi K. 2009.Chemical properties of aquatic fulvic acids isolated from Lake Biwa, aclear water system in Japan. Humic Substances Res 5/6:45–53.

32. McKnight DM, Boyer EW, Westerhoff PK, Doran PT, Kulbe T,Anderson DT. 2001. Spectrofluorometric characterization of dissolved

8 Environ. Toxicol. Chem. 31, 2012 T. Ahmed et al.

organic matter for indication of precursor organic materials andaromaticity. Limnol Oceanogr 46:38–48.

33. Chen J, LeBoeuf EJ, Dai S, Gu B. 2003. Fluorescence spectroscopicstudies of natural organic matter fractions. Chemosphere 50:639–647.

34. Engebretson RR, Von Wandruszka R. 1997. The effects of molecularsize on humic acid associations. Org Geochem 26:759–767.

35. Engebretson RR, Von Wandruszka R. 1998. Kinetic aspects of cation-enhanced aggregation in aqueous humic acids. Environ Sci Technol32:488–493.

36. Von Wandruszka R, Engebretson RR. 2001. Kinetics of humic acidassociations. In Clapp CE, Hayes MHB, Senesi N, Bloom PR, JardinePM, eds. Humic Substances and Chemical Contaminants. Soil ScienceSociety of America. Madison, WI, 119–126.

37. ChengWP,Chi FH. 2003. Influenceof eutrophication on the coagulationefficiency in reservoir water. Chemosphere 53:773–778.

38. Coble PG. 1996. Characterization of marine and terrestrial DOM inseawater using exciatation-emission matrix spectroscopy. Mar Chem51:325–346.

Napthol-Binding Complexes Environ. Toxicol. Chem. 31, 2012 9