ww.sciencedirect.com

wat e r r e s e a r c h 4 6 ( 2 0 1 2 ) 9 4 3e9 5 4

Available online at w

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

Quantifying PPCP interaction with dissolved organic matterin aqueous solution: Combined use of fluorescence quenchingand tandem mass spectrometry

Selene Hernandez-Ruiz a, Leif Abrell b, Samanthi Wickramasekara b, Benny Chefetz c,Jon Chorover a,b,*aDepartment of Soil, Water and Environmental Science, University of Arizona, 1177 E 4th St, Tucson, AZ 85721, USAbArizona Laboratory for Emerging Contaminants, University of Arizona, 1040 East 4th St, Tucson, AZ 85721, USAcDepartment of Soil and Water Sciences, The Hebrew University of Jerusalem, P.O. Box 12, Rehovot 76100, Israel

a r t i c l e i n f o

Article history:

Received 17 August 2011

Received in revised form

11 November 2011

Accepted 20 November 2011

Available online 26 November 2011

Keywords:

DOM

IHSS

Wastewater

Pharmaceuticals

LC-MS/MS

Fluorescence

Interaction

* Corresponding author. Department of Soil,USA. Tel.: þ1 520 626 5635; fax: þ1 520 621 1

E-mail addresses: selene.hernandez27@arizona.edu (S. Wickramasekara), chefetz0043-1354/$ e see front matter ª 2011 Elsevdoi:10.1016/j.watres.2011.11.061

a b s t r a c t

The documented presence of pharmaceuticals and personal care products (PPCPs) in

water sources has prompted a global interest in understanding their environmental fate.

Dissolved organic matter (DOM) can potentially alter the fate of these contaminants in

aqueous systems by forming contaminant-DOM complexes. In-situ measurements were

made to assess the interactions between three common PPCP contaminants and two

distinct DOM sources: a wastewater treatment plant (WWOM) and the Suwannee River,

GA (SROM). Aqueous DOM solutions (8.0 mg L�1 C, pH 7.4) were spiked with a range of

concentrations of bisphenol-A, carbamazepine and ibuprofen to assess the DOM fluo-

rophores quenched by PPCP interaction in excitationeemission matrices (EEM). Interac-

tion effects on target analyte (PPCP) concentrations were also quantified using direct

aqueous injection ultra high performance liquid chromatography tandem mass spec-

trometry (LC-MS/MS). At low bisphenol-A concentration, WWOM fluorescence was

quenched in an EEM region attributed to microbial byproduct-like and humic acid-like

DOM components, whereas carbamazepine and ibuprofen quenched fulvic acid-like flu-

orophores. Fluorescence quenching of SROM by bisphenol-A and carbamazepine was

centered on humic acid-like components, whereas ibuprofen quenched the fulvic acid-

like fluorophores. Nearly complete LC-MS/MS recovery of all three contaminants was

obtained, irrespective of analyte structure and DOM source, indicating relatively weak

PPCP-DOM bonding interactions. The results suggest that presence of DOM at

environmentally-relevant concentration can give rise to PPCP interactions that could

potentially affect their environmental transport, but these DOM-contaminant interac-

tions do not suppress the accurate assessment of target analyte concentrations by

aqueous injection LC-MS/MSMS.

ª 2011 Elsevier Ltd. All rights reserved.

Water and Environmental Science, University of Arizona, 1177 E 4th St, Tucson, AZ 85721,647.gmail.com (S. Hernandez-Ruiz), abrell@u.arizona.edu (L. Abrell), samanthw@email.@agri.huji.ac.il (B. Chefetz), chorover@cals.arizona.edu (J. Chorover).ier Ltd. All rights reserved.

wat e r r e s e a r c h 4 6 ( 2 0 1 2 ) 9 4 3e9 5 4944

1. Introduction concentrations in wastewater effluents are in the mg L�1 range

The chemical quality of treated wastewater is receiving

increased attention, particularly as human population growth

makes it a progressively larger component of the near-surface

hydrologic cycle. Treated wastewater contains partially

degraded or non-degraded endocrine disrupting compounds,

pharmaceuticals and personal care products (PPCPs) that can

make their way into freshwater sources with treatment plant

discharges (Joss et al., 2008; Dickenson et al., 2011). Human

and ecosystemhealth concerns derive from the fact that some

PPCPs are known to cause cancer, mutations, and/or impede

the reproduction and hormone function of living organisms

(Liu et al., 2005; Zhou et al., 2007). Over the past decade,

technical improvements in chemical analytical methods have

enabled the detection and quantification of these contami-

nants in water at sub-parts-per-trillion levels in environ-

mental samples including wastewater effluents and receiving

surface and ground water sources (Capdeville and Budzinski,

2011). Despite the documented adverse effects of PPCPs,

their persistence, transport, and fate in thewater cycle are not

well understood.

Organic contaminants co-occur with dissolved

natural organic matter (DOM), which is ubiquitous and

compositionally diverse in natural waters (Leenheer and

Croue, 2003). It comprises a heterogeneous mixture of vari-

ously aggregated organic molecules (<0.45 mm) deriving from

decaying biomass, including biomolecules and their degra-

dation products. Interactions between DOM and PPCPs can

potentially alter not only contaminant bioavailability and

transport, but also accurate detection and quantification of

these compounds (Lajeunesse and Gagnon, 2007). Aqueous

solubility and hydrophobicity parameters such as the octa-

nolewater partitioning coefficient (Kow) are insufficient

predictors of PPCP-DOM interaction (Kwon and Armbrust,

2008; Yamamoto et al., 2003) because intermolecular mecha-

nisms of association other than the hydrophobic effect e e.g.,

hydrogen bonding, cation bridging e can affect the bonding to

DOM of these relatively polar compounds (Tolls, 2001; Pan

et al., 2009). We hypothesize that physico-chemical proper-

ties of both DOM and PPCPs, including charge and function-

ality, influence the types of bonds formed between them. To

begin testing this hypothesis, the present study employed (i)

fluorescence spectroscopy to probe DOM-PPCP molecular

interaction, and (ii) direct aqueous injection tandem mass

spectrometry to measure DOM impacts on analyte quantifi-

cation. The study sought to probe PPCP-DOM interactions in

the aqueous environment, and at environmentally relevant

pH, ionic strength and dissolved organic carbon (DOC)

concentration.

In the present study, three model compounds e ibuprofen,

bisphenol-A, and carbamazepine e were used to probe PPCP-

DOM interactions in circumneutral pH aqueous systems

(Table 1). Ibuprofen (IBU), which is anionic at circumneutral

pH, is the third most used analgesic worldwide, and is

consumed on average at a level of 1200e1600 mg per person

per day. Of this, ca. 29% is estimated to be metabolized,

whereas the remainder is introduced to the water cycle

(Meulenberg et al., 2005; Kagle et al., 2009). Reported IBU

(Pedrouzo et al., 2007), whereas surface freshwaters typically

show lower values. Bisphenol-A (BPA), which is uncharged at

circumneutral pH, is used worldwide in the production of

plastics (Stavrakakis et al., 2008) and is known to bind to

estrogen receptors and may also be carcinogenic (Vom Saal

et al., 1998; Soto and Sonnenschein, 2010). BPA has been re-

ported to occur at 10e20 mg L�1 in several freshwaters (Kolpin

et al., 2002; Jjemba, 2006) and at several-fold higher values in

wastewaters (Lee et al., 2000e02; Wickramasekara et al., in

press). Carbamazepine (CBZ), which is also uncharged at cir-

cumneutral pH, is a potent pharmaceutical commonly used

for epilepsy and bipolar disorder (Bai et al., 2008). It has been

reported at concentrations ranging from ca. 0.2 mg L�1 in

freshwaters (Focazio et al., 2008) to >5 mg L�1 in wastewater

effluents (Pedrouzo et al., 2007; Jjemba, 2006; Wickramasekara

et al., in press) and can accumulate in plant biomass (Shenker

et al., 2011).

Given the variation in charge and structural chemistry of

these three model contaminants (Table 1), we anticipated

that their impact on DOM fluorescence might also depend on

DOM source and structure. Hence, we employed two distinct

bulk DOM sources, one from a wastewater treatment plant

and the other from a surface freshwater. Two complemen-

tary methods were employed. Fluorescence spectroscopy

was used to elucidate molecular components of DOM

interacting with the target PPCPs, and tandem mass spec-

trometry was used to measure the impact of associated bond

formation on analyte chromatographic separation and

quantification.

2. Materials and methods

2.1. Extraction and characterization of DOM

The wastewater DOM (WWOM) was obtained by pore water

extraction of primary municipal wastewater treatment plant

sludge (Tucson, AZ). Within an hour of collection, wet sludge

was centrifuged at 21,875 g for 20 min and the supernatant

solution filtered through 0.7 mm and 0.45 mm hydrophilic

polypropylene filters to obtain WWOM. The filtrate was

freeze-dried for characterization and experiments. The

freshwater DOM-Suwannee River natural organic matter

(SROM) e was purchased in freeze-dried form from the

International Humic Substances Society (IHSS). The SROMhad

been extracted from the Suwannee River in GA (USA) via

reverse osmosis with a final in-line 0.4 mm filter followed by

a desalting step mediated by cation exchange resin (Hþ form)

(Serkiz and Perdue, 1990).

Prior to characterization, freeze-dried DOM samples were

suspended in nanopure water and incubated at 7 rpm for 24 h

on an orbital shaker for analysis of total organic carbon (TOC)

using high temperature combustion and infrared detection of

CO2 on a Shimadzu TOC-V CSH TOC/TN analyzer (Columbia,

MD). Standard calibrations for organic carbonweremadewith

oven-dried potassium hydrogen phthalate.

Infrared spectra were collected in transmission mode by

drying four 1 mL aliquots of 20 mg L�1 DOM solution (C basis)

wat e r r e s e a r c h 4 6 ( 2 0 1 2 ) 9 4 3e9 5 4 945

onto germanium windows followed by collection of Fourier

Transform Infrared (FTIR) spectra with a Nicolet MagnaeIR

560 spectrometer equipped with a CsI beam splitter, DTGS-

detector and OMNIC software (Chorover et al., 2004;

Omoike and Chorover, 2006). A blank germanium window

was subtracted from sample data as background, and

a 20 min purge time was employed between data collection

at 400 scans over the spectral range of 500e3800 cm�1 at

4 cm�1 resolution.

Apparent molar mass distributions for each sample were

determined via high pressure size exclusion liquid chroma-

tography (HPLC-SEC) using a Waters 600 HPLC (Milford, MA)

unit equipped with multisolvent delivery 600 pump, 717 plus

auto sampler, and 996 photodiode array (PDA) detector oper-

ating at 280 nmwavelength. A guard columnand two stainless

steel (8� 300 mm) SEC columns (MCXGPC 1000 & 100,000 �A,

PSS Polymer Standard Service-USA, Inc Warnick RI) were

connected in series to give a linear relationship between log

molar mass and elution time for polystyrene sulfonate (PSS)

standards (Omoike and Chorover, 2006; Navon et al., 2011).

Circumneutral pH values (ca. 7.4) were maintained using

a phosphate buffer solution prepared by bringing 60 mL of

stock 400 mM sodium phosphate buffer solution to 1.0 L with

filtered (0.2 mm) nanopure water. DOM solutions were

prepared in this background electrolyte to give 30 mg L�1

Table 1 e Physico-chemical properties of PPCPs used in this st

PPCP Structure at pH 7.4 pKa Log Kow Log Dow

d

IBU 5.2d 3.72b 1.15c

0.36b

BPA 9.73b 3.32b 3.43b

CBZ 2.3c 2.93e 2.67b

a Collision Energy (CE), Capillary Voltage (CV), Cone Voltage (CNV), and D

b Nghiem and Hawkes (2009).

c Nghiem et al. (2005).

d Jjemba (2006).

e Bai et al. (2008).

f Stavrakakis et al. (2008).

g Gomez et al. (2006).

h Zhang and Zhou (2007).

i Cahill et al. (2004).

j Kuster et al. (2008).

organic C. PSS standards (w2.5 mgmL�1 from 910e48,600 Da)

and 4-ethylbensulfonic acid (186 Da standard) were used for

linear calibration of retention time against log molar mass

(Cabaniss et al., 2000). For each calibration or sample run,

100 mL of solution were injected onto the SEC columns in iso-

cratic mode at a flow rate of 1.0 mLmin�1. Number-average

and weighted-average molar mass (Mn and Mw, respectively),

and polydispersity (r) were calculated in the Empower Pro

software program (Waters 2002, Medford, MA) according to:

Mw ¼XN

i¼1hiðMiÞ=

XN

i¼1hi (1)

Mn ¼XN

i¼1hi=

XN

i¼1

�hi

Mi

�(2)

r ¼ Mw=Mn (3)

where hi andMi are the height andmolarmass, respectively of

the sample SEC-HPLC curve at elution volume i.

2.2. PPCP incubations

All solutions and incubations were prepared on a mass basis

in baked (550 �C, 4 h) amber glass bottles with Teflon caps.

Stock IBU, BPA and CBZ solutions were prepared in 250 mL

udy.

Parent toaughter (m/z)

CEa (eV) CVa (kV) CNVa (V) DTa (�C)

205/ 161g,i,j 8 2.8 20 300

2.8 170

227/ 212f 18 3.3 40 150

3.4 150

237/ 194g,h 16 2.8 50 1502.5 150

esolvation Temperature (DT).

wat e r r e s e a r c h 4 6 ( 2 0 1 2 ) 9 4 3e9 5 4946

nanopure water with 1.0 mL of MeOH added to enable

complete dissolution in 48 h. Experimental solutions for IBU

and CBZ were then prepared in an LC-MS/MS-compatible

(volatile) 24 mM NH4HCO3 buffer (1.4 g L�1 of NH4HCO3 in

nanopure water) back-titrated with 0.06 M HCl to pH 7.4.

Freeze-dried WWOM and SROM were likewise dissolved in

24 mM NH4HCO3 buffer to give 16 mg L�1 DOC stock concen-

tration, as verified with Shimadzu TOC-V CSH TOC/TN

analyzer (Columbia, MD). Samples were incubated for 24 h in

shaker at 100 rpm to attain sorption equilibrium (Yamamoto

et al., 2003; Polubesova et al., 2007; Ilani et al., 2005) in dark

amber bottles prior to LC-MS/MS and fluorescence spectros-

copy analyses.

DOM-PPCP incubations were performed in triplicate by

adding 2 mL of each DOC stock (16 mg L�1) to 2 mL of various

concentration PPCP stock solutions for final reactor volume of

4.0 mL, DOC at 8 mg L�1, and BPA at 100, 200, or 1000 mg L�1, or

IBU or CBZ at 10, 20, 200, and 1000 mg L�1. Experiments with

BPA were limited to the higher concentrations because of

a higher limit of LC-MS/MS detection in the aqueous buffer

solution for BPA (50 mg L�1) relative to those for IBU or CBZ

(0.16 and 0.15 mg L�1, respectively). The 4 mL bottles were

incubated (no headspace) for 24 h in an end-over-end rotating

mixer, followed by separation into 1.0 and 3.0 mL aliquots for

LC-MS/MS and fluorescence excitation-emission analyses,

respectively. Controls included DOM alone (no PPCP) to

account for possible analyte recoveries in excess of experi-

mental spikes, PPCP at the respective concentrations alone

(no DOM) to measure recovery from aqueous buffer solutions

and to confirm that analyte fluorescence was insignificant

relative to DOM across the excitation-emission range

employed.

2.3. Fluorescence spectroscopy

To elucidate DOM-PPCP interactions, excitation-emission

matrices (EEMs) of DOM were collected in the absence and

presence of PPCP analytes. EEMs were obtained with a Jobin

Yvon Horiba/Spex Fluoromax-4 fluorometer (Edison, NJ)

equipped with a xenon lamp as the excitation source. The

3.0 mL subsample for each treatment was placed in a square

quartz cuvette cell (light path 10 mm� 10 mm) for EEM

collection. The signal to reference detector ratio was collected

and EEMs were produced using the FluorEssence software

with excitation and emission wavelength ranges of

200e450 nm (5 nm slit) and 250e650 nm (2 nm slit), respec-

tively, both at 5 nm increments. Quenching EEMs were

calculated by subtracting EEMs for the DOM-PPCP treatment

from those of corresponding DOM alone. The contribution of

PPCPs themselves to total fluorescence in DOM-PPCP systems

was found to be negligible relative to that of DOM, such that

quenching EEMs represent accurately the diminished fluo-

rescence of DOM in the presence of PPCPs.

2.4. Fluorescence quantification

Fluorescence intensity was integrated beneath each of five

EEM regions previously characterized as (I) “tyrosine-like”, (II)

“tryptophan-like”, (III) “fulvic acid-like”, (IV) “microbial byproduct-

like”, and (V) “humic acid-like” (Fig. 1C) (Chen et al., 2003). The

volume (Fi) beneath region “i” of the EEM surfacewas obtained

according to

Fi ¼XEx

XEm

IðlExlEmÞDlExDlEm (4)

where Dlex is the excitation wavelength interval, Dlem is the

emission wavelength interval and I (Dlex, Dlem) is the fluo-

rescence intensity >104 at each excitation-emission wave-

length pair. The intensity cut-off of 104 was applied to remove

background noise (fluorescence values at emission wave-

lengths smaller than excitation). The total number of data

points >104 (Ni) for each region were counted to produce the

fractional projected excitation-emission factor (Fi). The

normalized fluorescence intensity volume beneath region ‘i’

of the DOM sources (Fin) was obtained from Fin¼ Fi*Fi, and

from this the fluorescence percentage of reach region was

calculated (Pin¼Fin/FT,n*100%), where FTn is the total

normalized EEM fluorescence deriving from all regions i

(FTn¼ SFin).

2.5. PPCP spike recovery by LC-MS/MS

The recovery of PPCPs following incubation with and without

DOM solutions was accomplished using an Acquity Ultra

Performance Liquid Chromatograph and triple quadruple

Quattro Premier XE mass spectrometer, equipped with

a sample organizer (Waters Corp., Milford, MA). Calibration

standards for IBU, BPA, and CBZ, were evaluated in the same

matrix as experimental treatments (24 mM NH4HCO3 buffer,

pH 7.4). Analytes were separated following 5.0 mL direct

injections through an Acquity UPLC BEH C18 (1.7 um;

2.1 mm� 50 mm) column maintained at 40 �C with a gradient

mobile phase starting at 80:20% water:acetonitrile and ending

at 100% acetonitrile at flow rate 0.3 mlmin�1. Positive mode

electrospray ionization (ESIþ) was used for CBZ, whereas

negative mode (ESI�) was used for IBU and BPA. Recovery was

quantified using multiple reaction monitoring (MRM). Frag-

mentation transitions, collision energies (CE), capillary volt-

ages (CV), cone voltages (CNV), and desolvation temperatures

(DT) are provided in Table 1. Mass lynx software (Waters

Corp.) was used to for identification and quantification of

analytes. An ANOVA/Tukey’s statistical test (95% CI) was used

in Origin 8.5 (Northampton, MA) to assess variance in recov-

eries of all treatments.

3. Results

3.1. Characterization of WWOM and SROM

Chemical analyses indicate that the mass fraction of carbon

is higher for SROM than for WWOM (Table 2). The measured

values for SROM are consistent with those reported previ-

ously (Serkiz and Perdue, 1990). Although the pH and elec-

trical conductivity (EC) of freshly prepared DOM solutions

are shown to vary by source, both samples were normalized

to pH 7.4 and an EC 1200 mS cm�1 for subsequent

experiments.

Table 2 e Baseline characterization for DOM sources.

OC (%) pH EC (mS cm�1) ε (Lmol�1 C cm�1)

SROM 54� 4 5.5� 0.7 31.4 � 2.3 347� 15

WWOM 32� 6.5 7.4� 0.3 1178� 40.1 32� 7

Organic carbon percentage in w/v (OC), Electrical conductivity (EC),

and Molar absorptivity (ε).

wat e r r e s e a r c h 4 6 ( 2 0 1 2 ) 9 4 3e9 5 4 947

3.1.1. Fourier transform infrared spectra of WWOM andSROMTransmission FTIR spectra of WWOM and SROM are overlain

for direct comparison in Fig. 2. Both samples display broad

peaks associated with hydroxyl stretching in the

3000e3500 cm�1 range (Leenheer, 1981), and symmetric/

asymmetric CH stretching of eCH3 and eCH2 moieties at

2840e2930 cm�1 (Swift, 1996). SROM displays a major peak at

1731 cm�1 from ester and/or C]O carboxyl stretching

(Swift, 1996), and a peak at 1610 cm�1 attributed to the

asymmetric stretch of eCOO� Swift (1996). Additionally,

SROM shows a peak at 1429 cm�1 ascribed to CeH deforma-

tion of CH2 or CH3 (Chefetz et al., 1998), at 1094 cm�1 due to

CeC stretching of aliphatics (Swift, 1996), and a peak at

829 cm�1 due to CeH out-of-plane bend of condensed

aromatic compounds (Santamaria et al., 2006; Swift, 1996).

Conversely, the WWOM sample displays a peak at 1796 cm�1

that could be attributed to ester C]O stretching (Stuart and

Ando, 1996) and a peak at 1658 cm�1 from C]O vibration of

amides (amide I vibration) associatedwith proteins (Leenheer,

Fig. 1 e Fluorescence EEMs of bulk WWOM (A) and SROM (B). C)

et al., 2003): (I) Tyrosine-like, (II) Tryptophan-like, (III) Fulvic acid

D) Regional normalized fluorescence percentage in accordance

(24 mM NH4HCO3 of pH 7.4 buffer and 8.0 mg LL1 of DOC).

1981; Omoike and Chorover, 2006). A minor amide II peak is

also present (1540 cm�1) confirming a greater relative preva-

lence of protein acious material in WWOM than in SROM

(Chefetz et al., 1998). The WWOM sample also shows peaks at

1356 cm�1from symmetric COO� stretching and CeH bending

of aliphatics (Swift, 1996). The most intense peaks in WWOM

occur around 1000 cm�1 and correspond to the CeO and

CeOeC vibrations of polysaccharides (Chefetz et al., 1998;

Omoike and Chorover, 2006). Overall, the FTIR data suggest

stronger plant derivation of SROM,which is enriched in lignin-

derived products, whereas prevalence of amide and carbo-

hydrate functionalities in WWOM is consistent with domi-

nantly microbial derivation associated with biological waste

water treatment (Hudson et al., 2007).

3.1.2. High performance size exclusion chromatographyThe SROM exhibits a weight average molar mass (Mw) of

2466 Da and polydispersity of three (Fig. 3), consistent with

a prior report of 2190 and 2320 Da for OM in Suwannee River

water (Chin et al., 1994). Similarly, values ranging from 2200 to

2300 Da have been reported for Suwannee River fulvic acid

(Chin et al., 1994; Cabaniss et al., 2000; Zhou et al., 2000). The

Mw and Mn values for WWOM are lower, and the chromato-

gram reveals a subpopulation of small molecules ca. 192 Da

and a large polydispersity value of 12, indicating significant

heterogeneity. Mw values for both DOM types are within the

range reported for DOM from soils, freshwater sources, and

wastewater sources (Leenheer, 1981; Cabaniss et al., 2000;

Chin et al., 1994; Li et al., 2005).

Location of five 1 operationally-defined EEMs regions (Chen

-like, (IV) Microbial byproduct-like, and (V) Humic acid-like.

with regions specified in (C) for bulk SROM and WWOM

Fig. 2 e Transmission FTIR spectra of SROM and WWOM.

Fig. 3 e HPLC-SEC chromatograms for SROM and WWOM

(24 mM sodium phosphate buffer of pH 7.4 and 30 mg LL1

of DOC). Weight-average molar mass (Mw), molar mass of

peak location (MP), number-average molar mass (Mn), and

polydispersity Index (r[Mw/Mn) are indicated.

wat e r r e s e a r c h 4 6 ( 2 0 1 2 ) 9 4 3e9 5 4948

3.1.3. Fluorescence spectroscopy of WWOM and SROMMolecular fluorescence is the result of energy emitted during

electron return to the ground state following radiation-

induced excitation (Lakowicz, 1999). Since fluorescence

spectra vary with excitation wavelengths and DOM composi-

tion, excitation-emission matrices (EEMs) are often used to

provide information on the prevalence and structural

composition of DOM fluorophore mixtures (Hudson et al.,

2007; Her et al., 2003).

No correction for inner filtering effects was required as

DOC concentrations (8 mg L�1) were well below threshold

values (ca. 25 mg L�1) for which inner-filtering effects are

observed (Hudson et al., 2007). The selected concentration is

representative of aqueous environmental DOM concentra-

tions in surface water and wastewater solutions (Henderson

et al., 2009; Serkiz and Perdue, 1990), and it enabled collec-

tion of data in the low (200e250 nm) excitation range of EEM

that was often neglected in prior research, but that provides

important information about soluble proteins and lower

molar mass DOM components of potential high reactivity

such as fulvic-acid like molecules.

Fluorescence EEMs for the DOM samples are consistent

with sample origins and literature review (Fig. 1-A and B). The

SROM EEM shows three peaks: a minor intensity peak at Ex/

Em 325/450 ascribed to humic-acid-like molecules in region V,

a secondary peak at Ex/Em 250/450 nm and a high intensity

peak at Ex/Em 220/450 nm attributed to fulvic acid-like

molecules of region III (Chen et al., 2003; Wu et al., 2003; Her

et al., 2003; Henderson et al., 2009). The WWOM EEM shows

a distinct peak at 275/350 nm, which has been reported

previously for wastewater effluents and is attributed to

microbial byproducts (Chen et al., 2003) such as amino acids

including tryptophan and tyrosine (Hudson et al., 2007; Her

et al., 2003; Henderson et al., 2009)and polysaccharides

(Her et al., 2003; Chen et al., 2003).

The SROM sample is rich in aromatic, humic acid-like

components that account for 88% of total EEM fluorescence,

whereas 11% is attributable to fulvic acid-like fluorescence. In

contrast,WWOMdisplayed greater fluorescence in fulvic acid-

like (33%) relative to humic acid-like (49%) regions. Moreover,

the EEM region defined as “microbial byproduct-like”

accounted for up to 10% of the total WWOM fluorescence

(Fig. 1D).

3.2. Fluorescence quenching of WWOM and SROM byPPCPs

DOM fluorescence is diminished (quenched) if static or

dynamic (collisional) interaction occurs with PPCP molecules.

In the case of the dynamic process, the quencher

(PPCP compound) collides with the fluorophore(s) during the

lifetime of the excited state, whereas static quenching is the

result of ground statemolecular interactions (Lakowicz, 1999).

Although fluorescence quenching does not discern dynamic

from static interaction (doing so requires additional fluores-

cence lifetime measurements at different temperatures), the

moieties undergoing the greatest quenching effect are

assumed to be directly involved in contaminant association

(Sun et al., 2007; Yamamoto et al., 2003). Prior studies have

reported differential quenching of DOM fluorophores during

molecular interaction with polar contaminant molecules

(Yamamoto et al., 2003; Bai et al., 2008), indicating the

potential for quenching EEMs to provide a fingerprint of DOM-

PPCP interaction.

Regions of DOM fluorescence that were quenched by IBU,

BPA and CBZ in the present work are depicted as difference

EEMs (DOM alone minus DOM-PPCP) in Figs. 4 (IBU), 5 (BPA)

and 6 (CBZ). These results are quantified in terms of regional

normalized fluorescence quenching percentage (Pin), which is

plotted for all DOM-PPCP combinations (Fig. 7). Fluorescence

quenching of WWOM by addition of IBU was observed in

Fig. 4 e DOM quenching by ibuprofen represented by difference EEMs plotted with increasing PPCP concentration for

WWOM (A-10, B-20, C-200 and D-1000 mg LL1) and for SROM (E-10, F-20, G-200 and H-1000 mg LL1). All treatments conducted

in 24 mM NH4HCO3, pH 7.4, and 8 mg LL1 of DOC.

wat e r r e s e a r c h 4 6 ( 2 0 1 2 ) 9 4 3e9 5 4 949

region III, and to a lesser extent in regions I and II at all

concentrations (Figs. 4 and 7A). Similarly, IBU quenching of

SROMfluorescence occurred primarily in region III followed by

region I (Figs. 4 and 7D). Fluorescence quenching of WWOM in

the presence of BPAwasmanifested more broadly throughout

the EEM, and clearly within regions IV and V, attributed to

Fig. 5 e DOM quenching by bisphenol-A represented by differen

WWOM (A-100, B-200 and C-1000 mg LL1) and for SROM (D-100, E

NH4HCO3, pH 7.4, and 8 mg LL1 of DOC.

humic substances (Figs. 5 and 7B). Reaction of BPA with SROM

resulted in quenching of region III fluorophores with an

increasing quenching trend from low to high BPA concentra-

tion (Fig. 5). Although region V showed significant quenching

at low BPA concentration, the relative quenching effect in this

region decreased with increasing BPA concentration as region

ce EEMs plotted with increasing PPCP concentration for

-200 and F-1000 mg LL1). All treatments conducted in 24 mM

Fig. 6 e DOM quenching by carbamazepine represented by difference EEMs plotted with increasing PPCP concentration for

WWOM (A-10, B-20, C-200 and D-1000 mg LL1) and for SROM (E-10, F-20, G-200 and H-1000 mg LL1). All treatments conducted

in 24 mM NH4HCO3, pH 7.4, and 8 mg LL1 of DOC.

wat e r r e s e a r c h 4 6 ( 2 0 1 2 ) 9 4 3e9 5 4950

III became increasingly important (Fig. 7E). Quenching in

regions I, II and IV was negligible for all three BPA

concentrations.

Quenching of WWOM fluorescence by CBZ occurred

entirely in regions I, II and III at all tested concentrations (Figs.

6 and 7C). However the quenching of SROM fluorescence by

CBZwas negligible for regions I and II, low for regions III and IV

and consistently high for region V at all concentrations (Fig. 7C

and F). The trends discussed above were highly reproducible

as indicated by replicated experiments.

3.3. LC-MS/MS recovery

The mass balance recoveries of DOM-reacted PPCPs were

evaluated by direct LC-MS/MS injection of DOM-PPCP aqueous

solutions following incubation. Recovery statistics were

analyzed by ANOVA (a¼ 0.05) to determine the significance of

DOM effects relative to DOM-free positive controls (Table 3).

A negative control confirmed that no PPCP signal was detected

from the DOM material alone. The only case where DOM

treatment deviated from the positive control was in the case

of SROM-IBU, where enhancement (109% recovery) was

observed for the 10 mg L�1and for the WWOM-IBU at 1000

concentration. In all other cases, no statistically significant

difference was observed for analyte recovery in the presence

or absence of DOM (Table 3).

4. Discussion

4.1. Comparison of WWOM and SROM

Since traditional wastewater treatments are only partially

effective at removing PPCPs (Snyder et al., 2007), these

contaminants are introduced to environmental water

supplies following discharge of treated wastewater into

surface or ground water systems. Hence, the interactions of

these variably charged and often polar organic compounds

with both wastewater and fresh water DOM constituents are

relevant to their transport and fate (Maoz and Chefetz, 2010;

Chefetz et al., 2008). Wastewater and freshwater DOM sour-

ces are expected to vary significantly in their chemical prop-

erties. Wastewater DOM is dominantly sourced from

residential and industrial waste materials and microbial

biomass associated with biological treatment processes

(Chefetz et al., 2006). Conversely, freshwater DOM is derived

from decay of terrestrial and/or aquatic biomass, with the

relative predominance of plant or algal contributions being

dependent on location and season (McKnight et al., 2001).

These chemical differences between wastewater and

freshwater OM are apparent for the samples used in the

present study. Their distinct EEMs indicate lower contribu-

tions of humic acid-like e and greater contributions of

microbial byproduct-like and proteinaceous constituents e to

fluorescence ofWWOMrelative to SROM (Fig. 1D). TheWWOM

sample also exhibited low apparent molar mass and high

polydispersity (Fig. 3), low molar absorptivity (ε, Table 2) and

hence low aromaticity, and dominantly comprised function-

alities associated with microbial biomolecular fragments,

particularly those deriving from polysaccharides and proteins

observed in FTIR (Fig. 2). Conversely, the SROM exhibited

higher apparent molar mass and lower polydispersity (Fig. 3),

more than 10-fold higher molar absorptivity (ε, Table 2), and

prevalent polyphenolic aromatic moieties characteristic of

partially-degraded lignin originating from vegetation decay

(Fig. 2) (Leenheer and Croue, 2003; Chen et al., 2003). Hence,

WWOM and SROM are expected to exhibit different reactivity

toward PPCPs, with other experimental conditions (DOC

concentration, pH, background electrolyte concentration)

held constant.

10 20 200 10000

25

50

75

100

SROM

Ibuprofen ( g L-1)

Pin

A D

E

F

WWOM

10 20 200 10000

25

50

75

100

Ibuprofen ( g L-1)

Pin

100 200 10000

25

50

75

100

BPA ( g L-1)

Pin

B

100 200 10000

25

50

75

100

BPA ( g L-1)

Pin

10 20 200 10000

25

50

75

100

Carbamazepine ( g L-1)

Pin

Region I Region II Region III

Region IV Region V

C

10 20 200 10000

25

50

75

100

Carbamazepine ( g L-1)

Pin

Fig. 7 e Regional normalized fluorescence quenching

percentage (Pin) of WWOM (AeC) and SROM (DeF) in

24 mM NH4HCO3 of pH 7.4 buffer and 8 mg LL1 of DOC.

wat e r r e s e a r c h 4 6 ( 2 0 1 2 ) 9 4 3e9 5 4 951

4.2. DOM-PPCP interactions

The fluorescence quenching of DOM as reflected in an EEM is

evidently sensitive to DOM-PPCP interaction under

environmentally-relevant conditions (Figs. 4e6). Although the

method cannot provide a quantitative assessment of the DOM

bound fraction of PPCP, it does provide “fingerprints” of

reactive DOMfluorophores. A consistent result with respect to

the quenching EEMs is that quenching patterns are qualita-

tively consistent for a given DOM-PPCP pair, whereas signifi-

cant differences are observed between DOM types for a given

PPCP (compare top and bottom rows in Figs. 4e6), and

different PPCPs exhibit distinctly different quenching

patterns. Conversely, within a DOM-PPCP pair, the effects of

PPCP concentration appear to be relatively small.

This qualitative assessment is supported, in part, by

quantitative integration of the EEM “quenching matrix”,

which provides an index for direct comparison among

different PPCPs and DOM types (Fig. 7). These data further

demonstrate the strong dependence of quenched regions of

the EEM on both compound and DOM type, and quenching

trends with PPCP concentration also become evident when

viewed this way.The present experiments probed DOM fluorescence

quenching by three PPCP analytes that were anionic (IBU), and

neutral (BPA and CBZ) at the experimental pH of ca. 7.4.

Ibuprofen, bearing a negatively-charged carboxyl group

attached to a neutral backbone (Table 1), yielded reproducible

quenching of regions III (fulvic acid-like), II and I (aromatic

protein-like) in bothWWOM and SROM.We postulate that the

dominance of quenching of fulvic acid-like moieties may be

the result of cation bridging between deprotonated carboxyls

of fulvate and ibuprofen, irrespective of DOM type. Indeed,

density functional theory calculations indicate that cation

bridging between carboxylated contaminants and ionized,

carboxylated DOM is energetically favorable at circumneutral

pH (Aquino et al., 2008).

In contrast, the neutral BPA molecule showed distinctly

different quenching patterns between WWOM and SROM.

It quenched fluorescence of humicacid-like like constituents

in both cases, but also quenched significantly microbial

byproduct-like fluorophores in WWOM (where such compo-

nents are more prevalent). Also, trends with BPA concentra-

tion were different between the two OM types: increasing BPA

concentration in WWOM solutions gave rise to a relative

increase in quenching of humic acid-like components,

whereas in SROM solutions, quenching increasingly affected

fulvic acid-like fluorophores. We speculate that these trends

signal a change in dominant interaction mechanism as well;

hydrophobic association is expected to be more important for

aromatic moieties characteristic of region V fluorophores,

whereas hydrophilic interaction (i.e., water bridging and/or

hydrogen bonding)would bemore likely to occurwith oxygen-

containing, polar functionalities characteristic of region III

fluorophores. The shift from one fluorophore group to another

not only provides an indication of favorable interactions, but

also offers insight to failed attempts to apply log Kow to

explain DOM-PPCP interactions. Predictions based on log Kow

assume that hydrophobic interactions are paramount,

whereas polar functional groups may in fact mediate

association between ionized chemicals and oxygen-

containing DOM moieties (Sangster, 1997; Tolls, 2001; Pan

et al., 2009).

Carbamazepine quenched distinctly different fluorophores

in the two DOM types: fulvic acid-like constituents in WWOM

and humic acid-like constituents in SROM. In both cases,

hydrogen bonding, pep and van der Waals interactions likely

prevailed (Navon et al., 2011), the difference in quenching

region being largely due to the difference in prevalence of

polar functionalities of fulvate versus humate fluorophores

between the two DOM sources (Fig. 1D). Carbamazepine

interactions with humic acid-like regions of DOM have been

reported in a prior fluorescence quenching experiment that

used a landfill leachate fulvic acid (Bai et al., 2008). However

that study did not measure the low excitation portion

(200e250 nm) of the EEM (regions I, II and III), which precluded

assessment of quenching in these regions.

Much prior research has suggested that the partitioning of

PPCP contaminants to dissolved organic matter may be

based largely on hydrophobic interactions (Cabaniss et al.,

2000; Pan et al., 2009). However, our limited understanding

is based on the prevalent use of terrestrial DOM sources in

such studies. As shown in this study, terrestrial DOM is

wat e r r e s e a r c h 4 6 ( 2 0 1 2 ) 9 4 3e9 5 4952

indeed more aromatic as a result of lignin-based precursors

than is DOM formed as a result of microbial processing that

occurs during wastewater treatment (Chen et al., 2003;

Leenheer and Croue, 2003; Leenheer, 2004). However, even

for such terrestrial sources, assessment of DOM-PPCP affinity

on the basis of Kow value does not account for polar inter-

actions that may be important to intermolecular association

(Tolls, 2001).

4.3. LC-MS/MS recovery

Tandemmass spectrometry has been previously used to aid in

understanding of complex formation in biological molecules,

and we assert that such an approach is profitably transferable

to PPCPs as well. Prior work has focused on non-covalent

bonding between known proteins and ligands wherein

complexation reduces the signal intensity detected by the LC-

MS/MS compared to that of a control investigated under the

same solution chemistry (Bolbach, 2005; Loo, 2000; Daniel

et al., 2002). Such a signal decrease can occur due to interac-

tion forces that prevent chromatographic or gas phase sepa-

ration of the complex. The strength and stability of

interactions can vary. For example, as shown by Loo (2000),

protein subunits can be associated via hydrophobic interac-

tion or hydrogen bonding, both of which are more fragile and

thus labile in the gas phase relative to electrostatic attractions

such as those between a cationic and anionic organic species.

Thus a reduction in LC-MS/MS signal compared to that of the

control signals potential strong ionic and/or covalent

interactions.

Table 3 e LC-MSMS Mean percent recovery and standarddeviations of spiked IBU, BPA and CBZ in WWOM andSROM solutions (DOC[ 8 mg LL1) and positive controls(DC) from 10e1000 mg LL1 in 24 mM NH4HCO3 electrolyteat pH 7.4.

PPCP (mg L�1) WWOM mean% recovery

SROM mean% recovery

IBU 10 95� 6 109� 8

þC 10 97� 3 97� 3

20 102� 2 107� 3

þC 20 99� 3 99� 3

200 98� 0 97� 1

þC 200 100� 2 100� 2

1000 93� 2 96� 2

þC 1000 100� 3 104� 6

BPA 200 101� 11 107� 5

þ C 200 96� 14 96� 14

1000 93� 9 86� 11

þC 1000 95� 5 95� 10

CBZ 10 103� 22 103� 3

þC 10 100� 5 100� 5

20 105� 18 100� 2

þC 20 100� 1 100� 1

200 106� 5 110� 2

þC 200 96� 4 96� 4

1000 108� 13 96� 3

þC 1000 101� 3 106� 10

The use of aqueous injection LC-MS/MS to measure PPCP

recovery in the presence of the two DOM sources with

different physico-chemical properties indicates that the

interactions giving rise to fluorescence quenching were not

strong enough to prevent chromatographic separation,

ionization, fragmentation and detection of the target analy-

tes. This suggests that the presence of DOM, at

environmentally-relevant DOC concentrations (8 mg L�1)

should not interfere with the detection and quantification of

these trace contaminant concentrations by direct aqueous

injection.

5. Conclusion

At low PPCP concentrations, as commonly occurs in natural

environments, BPA interacts with soluble wastewater DOM

components including microbially-derived biomolecular

fragments, while IBU and CBZ associate with fulvic acid-like

fluorophores. Interactions with low molar mass, soluble

DOM components likely facilitate the transport of PPCPs

from waste to freshwater environments (Cabaniss et al.,

2000). Convergence of wastewater effluents with freshwater

sources could likely result in transfer of the same contami-

nants to humic acid-like DOM components that are typically

of higher molar mass and greater aromaticity. For both

WWOM and SROM, interactions with the three PPCPs were

sufficiently weak to permit ca. 100% recovery with aqueous

injection LC-MS/MS, irrespective of analyte structure and

DOM source. Additionally, direct injection LC-MS/MS studies

indicate that the presence of DOM at concentrations

employed does not suppress the accurate assessment of

these target analytes.

Acknowledgments

Research support was provided by the Binational Agricultural

Research and Development (BARD) fund, Grant # IS-3822-06,

and Water Research Foundation (Award #4269) and Univer-

sity of Arizona Water Sustainability Program. The comments

and views detailed herein may be necessarily reflecting the

views of theWater Research Foundation, its officers, directors,

affiliates, or agents. Analyses in the Arizona Laboratory for

Emerging Contaminants were supported by NSF CBET

0722579.

Author contributions

Selene Hernandez Ruiz: Experimental design, execution, data

processing and interpretation, manuscript writing.

Leif Abrell: Technical: LC-MS/MS support in method

development.

Samanthi Wickarasemara: Fine tuning of LC-MS/MS

instrument.

Benny Chefetz: Experimental design planning, data interpre-

tation support, manuscript writing.

Jon Chorover: Experimental design planning, data interpre-

tation support, manuscript writing.

wat e r r e s e a r c h 4 6 ( 2 0 1 2 ) 9 4 3e9 5 4 953

r e f e r e n c e s

Aquino, A.J.A., Tunega, D., Pa�sali�c, H., Haberhauer, G.,Gerzabek, M.H., Lischka, H., 2008. The thermodynamicstability of hydrogen bonded and cation bridged complexes ofhumic acid modelsda theoretical study. Chem. Phys. 349(1e3), 69e76.

Bai, Y., Wu, F., Liu, C., Guo, J., Fu, P., Li, W., Xing, B., 2008.Interaction between carbamazepine and humic substances:a fluorescence spectroscopy study. Environ. Toxicol. Chem. 27(1), 95e102.

Bolbach, G., 2005. Matrix-assisted laser desorption/ionizationanalysis of non-covalent complexes: fundamentals andapplications. Curr. Pharm. Des. 11 (20), 2535e2557.

Cabaniss, S.E., Zhou, Q.H., Maurice, P.A., Chin, Y.P., Aiken, G.R.,2000. A log-normal distribution model for the molecularweight of aquatic fulvic acids. Environ. Sci. Technol. 34 (6),1103e1109.

Cahill, J.D., Furlong, E.T., Burkhardt, M.R., Kolpin, D.,Anderson, L.G., 2004. Determination of pharmaceuticalcompounds in surface- and ground-water samples by solid-phase extraction and high-performance liquidchromatography-electrospray ionization mass spectrometry.J. Chromatogr. A 1041 (1e2), 171e180.

Capdeville, M.J., Budzinski, H., 2011. Trace-level analysis oforganic contaminants in drinking waters and groundwaters.Trac-Trends Analy. Chem. 30 (4), 586e606.

Chefetz, B., Illani, T., Schulz, E., Chorover, J., 2006. Wastewaterdissolved organic matter: characteristics and sorptivecapabilities. Water Sci. Technol. 53 (7), 51e57.

Chefetz, B., Hadar, Y., Chen, Y., 1998. Dissolved organic carbonfractions formed during composting of municipal solid waste:properties and significance. Acta Hydrochim. Hydrobiol. 26 (3),172e179.

Chefetz, B., Mualem, T., Ben-Ari, J., 2008. Sorption and mobility ofpharmaceutical compounds in soil irrigated with reclaimedwastewater. Chemosphere 73 (8), 1335e1343.

Chen, W., Westerhoff, P., Leenheer, J.A., Booksh, K., 2003.Fluorescence excitation e emission matrix regionalintegration to quantify spectra for dissolved organic matter.Environ. Sci. Technol. 37 (24), 5701e5710.

Chin, Y.P., Aiken, G., Oloughlin, E., 1994. Molecular-weight,polydispersity, and spectroscopic properties of aquatic humicsubstances. Environ. Sci. Technol. 28 (11), 1853e1858.

Chorover, J., Amistadi, M.K., Chadwick, O.A., 2004. Surface chargeevolution of mineral-organic complexes during pedogenesisin Hawaiian basalt. Geochim. Cosmochim. Acta 68 (23),4859e4876.

Daniel, J.M., Friess, S.D., Rajagopalan, S., Wendt, S., Zenobi, R.,2002. Quantitative determination of noncovalent bindinginteractions using soft ionization mass spectrometry. Int. J.Mass Spectrom. 216 (1), 1e27.

Dickenson, E.R.V., Snyder, S.A., Sedlak, D.L., Drewes, J.E., 2011.Indicator compounds for assessment of wastewater effluentcontributions to flow and water quality. Water Res. 45 (3),1199e1212.

Focazio, M.J., Kolpin, D.W., Barnes, K.K., Furlong, E.T.,Meyer, M.T., Zaugg, S.D., Barber, L.B., Thurman, M.E., 2008.A national reconnaissance for pharmaceuticals and otherorganic wastewater contaminants in the United States e II)Untreated drinking water sources. Sci. Total Environ. 402(2e3), 201e216.

Gomez, M.J., Petrovic, M., Fernandez-Alba, A.R., Barcelo, D., 2006.Determination of pharmaceuticals of various therapeuticclasses by solid-phase extraction and liquid chromatography-tandem mass spectrometry analysis in hospital effluentwastewaters. J. Chromatogr. A 1114 (2), 224e233.

Henderson, R.K., Baker, A., Murphy, K.R., Hamblya, A.,Stuetz, R.M., Khan, S.J., 2009. Fluorescence as a potentialmonitoring tool for recycled water systems: a review. WaterRes. 43 (4), 863e881.

Her, N., Amy, G., McKnight, D., Sohn, J., Yoon, Y.M., 2003.Characterization of DOM as a function of MW by fluorescenceEEM and HPLC-SEC using UVA, DOC, and fluorescencedetection. Water Res. 37 (17), 4295e4303.

Hudson, N., Baker, A., Reynolds, D., 2007. Fluorescence analysis ofdissolved organic matter in natural, waste and pollutedwaters e a review. River Res. Appl. 23 (6), 631e649.

Ilani, T., Schulz, E., Chefetz, B., 2005. Interactions of organiccompounds with wastewater dissolved organic matter: role ofhydrophohic fractions. J. Environ. Qual. 34 (2), 552e562.

Jjemba, P.K., 2006. Excretion and ecotoxicity of pharmaceuticaland personal care products in the environment. Ecotoxicol.Environ. Saf. 63 (1), 113e130.

Joss, A., Siegrist, H., Ternes, T.A., 2008. Are we about to upgradewastewater treatment for removing organic micropollutants?Water Sci. Technol. 57 (2), 251e255.

Kagle, J., Porter, A.W., Murdoch, R.W., Rivera-Cancel, G., Hay, A.G.,2009. Biodegradation of pharmaceutical and personal careproducts. Adv. Appl. Microbiol. 67, 65e108.

Kolpin, D.W., Furlong, E.T., Meyer, M.T., Thurman, E.M.,Zaugg, S.D., Barber, L.B., Buxton, H.T., 2002. Pharmaceuticals,hormones, and other organic wastewater contaminants in USstreams, 1999e2000: a national reconnaissance. Environ. Sci.Technol. 36 (6), 1202e1211.

Kuster, M., Jose Lopez de Alda, M., Dolores Hernando, M.,Petrovic, M., Martin-Alonso, J., Barcelo, D., 2008. Analysis andoccurrence of pharmaceuticals, estrogens, progestogens andpolar pesticides in sewage treatment plant effluents, riverwater and drinking water in the Llobregat river basin(Barcelona, Spain). J. Hydrol. 358 (1e2), 112e123.

Kwon, J., Armbrust, K.L., 2008. Aqueous solubility, n-octanol-water partition coefficient, and sorption of five selectiveserotonin reuptake inhibitors to sediments and soils. Bull.Environ. Contam. Toxicol. 81 (2), 128e135.

Lajeunesse, A., Gagnon, C., 2007. Determination of acidicpharmaceutical products and carbamazepine in roughlyprimary-treated wastewater by solid-phase extraction and gaschromatography-tandem mass spectrometry. Int. J. Environ.Anal. Chem. 87 (8), 565e578.

Lakowicz, J.R., 1999. Principles of Fluorescence Spectroscopy.Kluwer, New York.

Lee, K.E., Barber, L.B., Furlong, E.T., Cahill, J.D., Kolpin, D.W.,Meyer, M.T., Zaugg, S.D., 2000e02. Presence and distribution oforganic wastewater compounds in wastewater, surface,ground, and drinking waters, Minnesota, 2000e02. U.S.Geological Survey Scientific Investigations Report 2004-5138.

Leenheer, J.A., 1981. Comprehensive approach to preparativeisolation and fractionation of dissolved organic-carbon fromnatural-waters and wastewaters. Environ. Sci. Technol. 15 (5),578e587.

Leenheer, J.A., 2004. Comprehensive assessment of precursors,diagenesis, and reactivity to water treatment of dissolved andcolloidal organic matter. Water Sci. Technol. Water Supply 4(4), 1e9.

Leenheer, J.A., Croue, J.P., 2003. Characterizing aquatic dissolvedorganic matter. Environ. Sci. Technol. 37 (1), 18Ae26A.

Li, F.S., Yuasa, A., Matsui, Y., Cheong, E.J., 2005. Polydisperseadsorption characteristics of aqueous organic matrices inwater and wastewater sources. Adsorp. J. Int. Adsorp. Soc. 11,691e696.

Liu, R.X., Wilding, A., Hibberd, A., Zhou, J.L., 2005. Partition ofendocrine-disrupting chemicals between colloids anddissolved phase as determined by cross-flow ultrafiltration.Environ. Sci. Technol. 39 (8), 2753e2761.

wat e r r e s e a r c h 4 6 ( 2 0 1 2 ) 9 4 3e9 5 4954

Loo, J.A., 2000. Electrospray ionization mass spectrometry:a technology for studying noncovalent macromolecularcomplexes. Int. J. Mass Spectrom. 200 (1e3), 175e186.

Maoz, A., Chefetz, B., 2010. Sorption of the pharmaceuticalscarbamazepine and naproxen to dissolved organic matter:role of structural fractions. Water Res. 44 (3), 981e989.

McKnight, D., Boyer, E., Westerhoff, P., Doran, P., Kulbe, T.,Andersen, D., 2001. Spectrofluorometric characterization ofdissolved organic matter for indication of precursor organicmaterial and aromaticity. Limnol. Oceanogr. 46 (1), 38e48.

Meulenberg, E.P., Peelen, G.O.H., Lukkien, E., Koopal, K., 2005.Immunochemical detection methods for bioactive pollutants.Int. J. Environ. Anal. Chem. 85 (12e13), 861e870.

Navon, R., Hernandez-Ruiz, S., Chorover, J., Chefetz, B., 2011.Interactions of carbamazepine in soil: effects of dissolvedorganic matter. J. Environ. Qual. 40 (3), 942e948.

Nghiem, L., Schafer, A., Elimelech, M., 2005. Pharmaceuticalretention mechanisms by nanofiltration membranes. Environ.Sci. Technol. 39 (19), 7698e7705.

Nghiem, L.D., Hawkes, S., 2009. Effects of membrane fouling onthe nanofiltration of trace organic contaminants. Desalination236 (1e3), 273e281.

Omoike, A., Chorover, J., 2006. Adsorption to goethite ofextracellular polymeric substances from Bacillus subtilis.Geochim. Cosmochim. Acta 70 (4), 827e838.

Pan, B., Ning, P., Xing, B., 2009. Part V-sorption of pharmaceuticalsand personal care products. Environ. Sci. Pollut. Res. 16 (1),106e116.

Pedrouzo, M., Reverte, S., Borrull, F., Pocurull, E., Marce, R.M.,2007. Pharmaceutical determination in surface andwastewaters using high-performance liquid chromatography-(electrospray)-mass spectrometry. J. Separat. Sci. 30 (3),297e303.

Polubesova, T., Sherman-Nakache, M., Chefetz, B., 2007. Bindingof pyrene to hydrophobic fractions of dissolved organicmatter: effect of polyvalent metal complexation. Environ. Sci.Technol. 41 (15), 5389e5394.

Sangster, J., 1997. OctanoleWater Partition Coefficients:Fundamentals and Physical Chemistry. Wiley, Chichester;New York.

Santamaria, A., Mondragon, F., Molina, A., Marsh, N.D.,Eddings, E.G., Sarofim, A.F., 2006. FT-IR and H-1 NMRcharacterization of the products of an ethylene inversediffusion flame. Combust. Flame 146 (1e2), 52e62.

Serkiz, S.M., Perdue, E.M., 1990. Isolation of dissolved organicmatter from the suwannee river using reverse osmosis. WaterRes. 24 (7), 911e916.

Shenker, M., Harush, D., Ben-Ari, J., Chefetz, B., 2011. Uptake ofcarbamazepine by cucumber plants e A case study related toirrigation with reclaimed wastewater. Chemosphere 82 (6),905e910.

Snyder, S.A., Adham, S., Redding, A.M., Cannon, F.S., DeCarolis, J.,Oppenheimer, J., Wert, E.C., Yoon, Y., 2007. Role of

membranes and activated carbon in the removal of endocrinedisruptors and pharmaceuticals. Desalination 202 (1e3),156e181.

Soto, A.M., Sonnenschein, C., 2010. Environmental causes ofcancer: endocrine disruptors as carcinogens. Nat. Rev.Endocrinol. 6 (7), 364e371.

Stavrakakis, C., Colin, R., Hequet, V., Faur, C., Le Cloirec, P., 2008.Analysis of endocrine disrupting compounds in wastewaterand drinking water treatment plants at the nanogram per litrelevel. Environ. Technol. 29 (3), 279e286.

Stuart, B., Ando, D.J., 1996. Modern Infrared Spectroscopy.Published on behalf of ACOL (University of Greenwich) byWiley, New York,.

Sun, W.L., Ni, J.R., Xu, N., Sun, L.Y., 2007. Fluorescence ofsediment humic substance and its effect on the sorption ofselected endocrine disruptors. Chemosphere 66 (4), 700e707.

Swift, R.S., 1996. Organic matter characterization. In: Sparks, D.L.(Ed.), Methods of Soil Analysis. Part 3, Chemical Methods. SoilScience Society of America: American Society of Agronomy,Madison, Wis, pp. 1011e1070.

Tolls, J., 2001. Sorption of veterinary pharmaceuticals in soils:a review. Environ. Sci. Technol. 35 (17), 3397e3406.

Vom Saal, F.S., Cooke, P.S., Buchanan, D.L., Palanza, P.,Thayer, K.A., Nagel, S.C., Parmigiani, S., Welshons, W.V., 1998.A Physiologically based Approach to the Study of Bisphenol Aand Other Estrogenic Chemicals on the Size of ReproductiveOrgans, Daily Sperm Production, and Behavior. Toxicologyand Industrial Health, pp. 239e260.

Wickramasekara, S., Hernandez Ruiz, S., Abrell, L., Arnold, R.,Chorover, J. Natural dissolved organic matter affectselectrospray ionization during analysis of emergingcontaminants by mass spectrometry. Anal. Chim. Acta., inpress.

Wu, F.C., Evans, R.D., Dillon, P.J., 2003. Separation andcharacterization of NOM by high-performance liquidchromatography and on-line three-dimensional excitationemission matrix fluorescence detection. Environ. Sci. Technol.37 (16), 3687e3693.

Yamamoto, H., Liljestrand, H.M., Shimizu, Y., Morita, M., 2003.Effects of physicalechemical characteristics on the sorption ofselected endocrine disruptors by dissolved organic mattersurrogates. Environ. Sci. Technol. 37 (12), 2646e2657.

Zhang, Z.L., Zhou, J.L., 2007. Simultaneous determination ofvarious pharmaceutical compounds in water by solid-phaseextraction-liquid chromatography-tandem massspectrometry. J. Chromatogr. A 1154 (1e2), 205e213.

Zhou, J.L., Liu, R., Wilding, A., Hibberd, A., 2007. Sorption ofselected endocrine disrupting chemicals to different aquaticcolloids. Environ. Sci. Technol. 41 (1), 206e213.

Zhou, Q.H., Cabaniss, S.E., Maurice, P.A., 2000. Considerations inthe use of high-pressure size exclusion chromatography(HPSEC) for determining molecular weights of aquatic humicsubstances. Water Res. 34 (14), 3505e3514.