Photochemically Induced Bound Residue Formation of

May 5, 2017 - More than 400 new nitrogen containing products were detected upon experimental sunlight photolysis of the pharmaceutical carbamazepine ...
0 downloads 0 Views 1MB Size
Article pubs.acs.org/est

Photochemically Induced Bound Residue Formation of Carbamazepine with Dissolved Organic Matter Julia Raeke,† Oliver J. Lechtenfeld,†,‡ Bettina Seiwert,† Till Meier,† Christina Riemenschneider,† and Thorsten Reemtsma*,† †

Department of Analytical Chemistry and ‡ProVIS − Centre for Chemical Microscopy, Helmholtz Centre for Environmental Research - UFZ, Permoserstrasse 15, 04318 Leipzig, Germany S Supporting Information *

ABSTRACT: More than 400 new nitrogen containing products were detected upon experimental sunlight photolysis of the pharmaceutical carbamazepine (CBZ) in the presence of dissolved organic matter (DOM) by Fourier transform ion cyclotron resonance mass spectrometry (FTICR-MS). These products were presumably formed through covalent binding of CBZ phototransformation products with DOM molecules. About 50% of these newly formed bound residues contained one nitrogen atom and had a molecular mass between 375 and 525 Da, which was 150 to 200 Da higher than for an average DOM molecule. In addition, a previously unknown CBZ phototransformation product, 3-quinolinecarboxylic acid (3QCA), was identified by liquid chromatography high resolution tandem mass spectrometry (LC-HRMS/MS). 3-QCA was likely formed through oxidative ring cleavage and subsequent decarboxylation of acridine, a well-known phototransformation product of CBZ. Collision induced dissociation experiments and Kendrick mass defect analyses corroborated that about 160 of the new products were formed via covalent binding of 3-QCA with DOM molecules of above-average O/C and H/C ratios. Experiments at lower CBZ concentration suggested that the importance of bound residue formation increases with increasing DOM/CBZ ratios. Photochemically induced bound residue formation of polar contaminants with DOM in the aqueous phase is thus a disregarded pathway along which contaminants can be transformed in the environment. The method presented here offers a new possibility to study the formation of bound residues, which may be of relevance also for other transformation processes in natural waters where radical intermediates are generated.



INTRODUCTION

One possible process leading to bound residues in the aquatic environment is sunlight photolysis in surface waters as this can generate reactive radical intermediates. The coupling of polychlorobenzenes with benzoic acid or syringic acid, considered model monomers for humic substances, upon photolysis has been shown already 30 years ago.6 The need for further research on this topic was addressed in the late 1980s,7 but little new knowledge has been gained since then. Recently, the incorporation of benzotriazole, a widely used and distributed corrosion inhibitor, into DOM through photochemical processes has been shown.5 In that case the reaction of the benzotriazole phototransformation product aniline with quinone functionalities in DOM has been suggested as the main process responsible for the coupling to DOM.5 The present study was initiated to clarify if the formation of bound residues with DOM by sunlight photolysis might be a more general process, relevant also for other contaminants than

The term “bound residues” was originally defined by the International Union of Pure and Applied Chemistry (IUPAC) to describe the environmental fate of pesticides as “chemical species originating from pesticides [...] that are unextracted by methods which do not significantly change the chemical nature of these residues”.1 Bound residue formation in soils has been extensively studied, beyond pesticides, and can result from different types of interactions, with adsorption, entrapment, or covalent binding being most important.2,3 It has been discussed controversially whether bound residues are an environmentally acceptable way to immobilize potentially harmful chemicals or if they pose an unknown risk because their potential for future release and toxicity cannot be predicted.4 Up to now, the photochemically induced formation of bound residues from organic contaminants with dissolved organic matter (DOM) in the aqueous phase has rarely been studied.5,6 However, this process could describe a so far disregarded sink for anthropogenic contaminants in the aquatic environment. Contrary to the soil environment such bound residues in surface waters would remain as mobile as the DOM. © XXXX American Chemical Society

Received: February 14, 2017 Revised: April 21, 2017 Accepted: April 25, 2017

A

DOI: 10.1021/acs.est.7b00823 Environ. Sci. Technol. XXXX, XXX, XXX−XXX

Article

Environmental Science & Technology

were used to unambiguously distinguish newly formed products from naturally present molecules and to obtain mechanistic insight into how phototransformation products of CBZ form bound residues with DOM.

benzotriazole and, possibly, involving different reactive groups in DOM. This appears important as such a bound residue formation would offer another pathway beyond mineralization along which a contaminant and its transformation products apparently disappear from the environment. Carbamazepine (CBZ), an antiepileptic drug, which is not completely metabolized in the human body, hardly eliminated in wastewater treatment plants8−10 and widely determined in surface waters in concentrations in the μg L−1 range,8,9,11 was chosen for this study. DOM is a complex and heterogeneous mixture of organic molecules derived from primary production and subsequent biological and chemical transformation and plays an important role in biogeochemical processes in natural waters. Due to its large functional diversity, it can influence the phototransformation of anthropogenic contaminants in different ways. On the one hand, it can decrease the sunlight photolysis of chemicals in natural waters by scavenging reactive species, filtering photochemically active light, or reduction of intermediates.12−14 On the other hand, it can also enhance their degradation by acting as a photosensitizer, producing reactive intermediates such as singlet oxygen (1O2), hydroxyl radicals (OH•), or triplet state DOM (3DOM*).15,16 The indirect photolysis of CBZ has been studied extensively. While in some studies it was suggested that the inner filter effect of DOM hinders the photodegradation of CBZ,12,17 in other studies indirect photolysis was reported to promote its degradation.18−20 These contradictory data suggest that the effect of indirect photolysis likely depends on the composition and/or the concentration of the DOM. The transformation products formed upon photolysis of CBZ were also investigated in several studies.19,21,22 Acridine (ACD) and 10,11-dihydro-10,11-trans-dihydroxy-CBZ were observed as the main products for direct photolysis and for the reaction with hydroxyl radicals.17,21 Photochemical reactions can also lead to the transformation of DOM itself. Besides mineralization also low molecular weight compounds such as glyoxal, formaldehyde, or pyruvate can be formed, which serve as substrates for microbes.23−25 Electrospray ionization Fourier transform ion cyclotron resonance mass spectrometry (ESI-FTICR-MS) revealed that mainly unsaturated and/or aromatic molecules were photodegraded and converted to more saturated molecules with a higher oxygen content.26−28 One reason for the limited knowledge on covalent binding of organic contaminants to DOM induced by sunlight photolysis might be that their detection is hindered by the intrinsic complexity of DOM and the presumably large number of possible reaction sites. As a consequence, many possible reaction products can be expected, each formed at very low concentration. However, ultrahigh resolution ESI-FTICR-MS, which is an established method for DOM analysis,29,30 has the potential to overcome the analytical challenges associated with the detection of bound residues with DOM due to its very large resolving power and its high sensitivity.5 In the present study ESI-FTICR-MS was used to follow the incorporation of CBZ into Suwannee River Fulvic Acid (SRFA), a well-defined DOM standard with low nitrogen content, by sunlight photolysis. This technique was used to identify reactive phototransformation products of CBZ (in combination with LC-HRMS and LC-MS/MS), the reaction products with SRFA as well as reactive functional groups within the SRFA, which are involved in the bound residue formation. Collision induced fragmentation and isotopically labeled CBZ



EXPERIMENTAL SECTION Chemicals and Reagents. Carbamazepine (99%), acridine (97%), 3-quinolinecarboxylic acid (98%), and ammonium bicarbonate (≥99.5%) were purchased from Sigma-Aldrich (St. Louis, USA). [D215N]-carbamazepine was obtained from Toronto Research Chemicals (Toronto, CA). ULC/MS grade methanol, isopropanyl alcohol, and acetic acid were purchased from Biosolve (Valkenswaard, Netherlands). SRFA (1R101F) was obtained from the International Humic Substances Society (IHSS, St. Paul, MN, USA). All solutions were prepared with ultrapure water (Milli-Q, Merck KGaA, Darmstadt, Germany). Irradiation Experiments. Irradiation experiments were conducted using a sun simulator Q-SUN Xe-1 (Q-Lab, Westlake, USA) equipped with a xenon arc lamp (1800 W) and a Daylight-Q filter (noon summer sunlight). The continuous emission spectra of a xenon arc lamp can closely mimic natural light with a wavelength range of 290 to 800 nm. An internal temperature of 32 °C was adjusted with a chiller. All experiments were carried out under a radiation intensity of 784 W m−2 (integration of the spectral power distribution over the wavelength range of 290−800 nm).5 The specimen tray sizes of 251 × 457 mm enabled the simultaneous irradiation of eight quartz cuvettes (SUPRASIL, light path 10 mm, V = 3500 μL) at a time. An irradiation time of 2.7 h corresponded to one summer sunny day (SSD; with 12.3 sunlight hours, 15 July, 45°N latitude).5 Each experiment was conducted in duplicate with 500 μL sample volume per cuvette. CBZ (0.5 μmol L−1, 5 μmol L−1, and 50 μmol L−1), ACD (50 μmol L−1), and [D215N]-CBZ (50 μmol L−1) were irradiated alone and in the presence of SRFA (25 mg L−1) for 24 h. SRFA was also irradiated separately for 24 h. To examine the reaction of SRFA with ammonium, SRFA was irradiated with 20 μmol L−1 ammonium bicarbonate. The pH of the solutions was around 5.5 and remained stable during the irradiation experiments. Because of strong signal suppression in ESI-FTICR-MS no buffer was used in the experiments. For each sample a dark control sample was prepared by wrapping a vial containing 500 μL of the sample with aluminum foil and placing it in the sun simulator. All samples were stored at −20 °C until analysis. FTICR-MS Analysis. An aliquot of each sample was diluted 1:1 (v/v) with methanol. The samples were analyzed using an FTICR mass spectrometer equipped with a dynamically harmonized analyzer cell (solariX XR, Bruker Daltonics Inc., Billerica, MA, USA) and a 12 T refrigerated actively shielded superconducting magnet (Bruker Biospin, Wissembourg, France). Negative electrospray ionization (capillary voltage 4.5 kV) was used at an infusion rate of 240 μL h−1. For each spectrum 300 scans were coadded in the mass range m/z 147− 3000. The acquired spectra were externally calibrated with an arginine cluster. Mass spectra of the samples containing SRFA were internally calibrated with a list of peaks commonly present in natural organic matter (m/z 150−600, n > 150), the other mass spectra with a list of omnipresent fatty acids. The rootmean-square of the calibration masses after internal calibration was lower than 0.1 ppm. Peaks were considered if the signal-tonoise (S/N) ratio was greater than four. Only the peaks that were present in both duplicates but not in the respective dark B

DOI: 10.1021/acs.est.7b00823 Environ. Sci. Technol. XXXX, XXX, XXX−XXX

Article

Environmental Science & Technology

Figure 1. Structures of CBZ and seven of its phototransformation products and their relative contribution (percentage of CBZ removed) when CBZ was irradiated alone (black) or in the presence of SRFA (red). Each symbol represents a separate experiment.

Analyst Software (version 1.5.2, AB Sciex) was applied. Further details are described by Riemenschneider et al.33 UPLC-QTOF-MS. For the identification of unknown phototransformation products of CBZ Ultra-Performance Liquid Chromatography-Quadrupole-Time-of-Flight-Mass Spectrometry (UPLC-QTOF-MS) was used. The samples were analyzed on a Xevo G2-XS QTOF mass spectrometer (Waters) with positive mode electrospray ionization. A C18 reversed phase column (Waters, Acquity UPLC HSS T3, 2.1 mm × 100 mm, 1.8 mm) was used for separation. Solvent A was ultrapure water with 0.1% formic acid (ULC/MS grade, Biosolve); solvent B was methanol with 0.1% formic acid (ULC/MS grade, Biosolve). The gradient used is shown in Table S2. The flow rate was 0.45 mL min−1. The injection volume was 10 μL per sample.

control were considered for further analysis. For the samples containing SRFA the relative peak intensities were calculated relative to the highest SRFA peak in the spectrum. Highest sensitivity for the experiments at lower CBZ concentration (0.5 μmol L−1 and 5 μmol L−1) was achieved by applying the CASI (continuous accumulation of selected ions) mode to a mass window of 10 Da with increased ion accumulation time (1 s instead of 0.05 s). Product ion spectra were recorded at collision energies between 5 and 15 V. The calculation of mass differences was performed with a self-developed R script inspired by the total mass difference statistics (TMDS) algorithm of Kunenkov et al.31 After elimination of the 13C isotopes and doubly charged peaks a mass difference matrix composed of all pairwise mass differences between monoisotopic peaks was generated, and the appearance probability P(d) was calculated according to Kunenkov et al.31 The discrete mass difference spectrum was then filtered to remove combinations of mass differences d3 = d1 ± d2 which had a lower appearance probability than the individual mass difference d1 and d2. The Kendrick mass (KM) was calculated for each mass peak by eq 1 according to Kendrick et al.32 Here, the base mass 173.0477 (C10H7NO2) was used. KM = massdetected ×

173.0000 173.0477



RESULTS AND DISCUSSION Phototransformation Products of CBZ. In the presence of SRFA the photolysis of CBZ was more effective (95% of initial CBZ removed) than without SRFA (72% removed), which points at the importance of indirect photolysis. CBZ and seven of its phototransformation products (details in Table S1) were quantified after irradiation of CBZ (50 μmol L−1) alone and in the presence of SRFA (Figure 1, Table S3). The distribution of phototransformation products differed with an increasing proportion of carbamazepine-10,11-epoxide (EPCBZ), 10,11-dihydro-10,11-trans-dihydroxycarbamazepine, and 10,11-dihydro-10,11-cis-dihydroxycarbamazepine (trans- and cis-DiOH−CBZ) and a decreasing proportion of acridine (ACD) and 10-hydroxy-carbamazepine (10-OH−CBZ) in the presence of SRFA (Figure 1). Hydroxyl radicals are reported to be one of the main reactive species in photochemical reactions of SRFA34 and to play an important role as a photosensitizing agent during indirect photodegradation of CBZ in the presence of DOM.19,35 Indirect photolysis not only accelerates CBZ transformation but also leads to changes in the product distribution. EP-CBZ and trans-DiOH−CBZ are known to be preferentially formed through the reaction with hydroxyl radicals,19,21 which explains the higher portion of these products after indirect photolysis. However, the quantified CBZ phototransformation products only accounted for 6% of CBZ degraded in pure solution and 11% degraded in the presence of SRFA. This indicates that other transformation products were also formed through

(1)

The Kendrick mass defect (KMD) was then calculated by eq 2 as the difference between the closest integer mass and the Kendrick mass. KMD = KM integer − KM

(2)

Mass peaks with the same KMD values differ exactly by the base mass. LC-MS/MS. A high performance liquid chromatography (HPLC) system (Agilent1260 Infinity, Agilent Technologies, Böblingen, Germany) coupled to a triple quadrupole mass spectrometer (QTrap 5500, AB Sciex, Darmstadt, Germany) was used for quantification of CBZ and seven commercially available CBZ transformation products listed in Table S1. The analysis of all target compounds was performed by direct injection of 5 μL of each sample into the LC-ESI-MS/MS system. Quantification was carried out by external standard calibration in methanol/water (1:1, v/v). For data analysis the C

DOI: 10.1021/acs.est.7b00823 Environ. Sci. Technol. XXXX, XXX, XXX−XXX

Article

Environmental Science & Technology

Figure 2. a) Intensity-based mass distribution of newly formed mass peaks containing one nitrogen atom (CxHyN1Oz) after irradiation of 50 μmol L−1 CBZ (magenta) and ACD (yellow) in the presence of SRFA, and common mass peaks in CBZ and ACD irradiated with SRFA (blue). CxHyN0Oz mass peaks after irradiation of SRFA alone are shown for reference (grey). b) van Krevelen diagram with CxHyN1Oz molecular formulas originally present in SRFA and newly formed after irradiation of CBZ (50 μmol L−1) in the presence of SRFA, black triangles: CBZ, ACD, and 3QCA.

TiO2.22 In contrast, ammonium was only formed at minor concentration in irradiated SRFA. Nitrate was not detected in any experiment (Table S4). Formation of Bound Residues with DOM. FTICR-MS analyses revealed that more than 400 new molecular formulas containing nitrogen were formed upon irradiation of CBZ in the presence of SRFA. These are collectively defined here as bound residues. The corresponding mass peaks were not detected in SRFA or CBZ irradiated alone or in the dark control of CBZ and SRFA. Most of these molecular formulas contained one (about 50%) or two (about 40%) nitrogen atoms. In the following, we will focus on the molecular formulas containing one nitrogen atom (CxHyN1Oz) because of their higher abundance and higher average intensities in the FTICR mass spectra. The intensity maximum of the corresponding CxHyN1Oz mass peaks was between m/z 400 and 450 (Figure 2a, CBZ+SRFA). Assuming that the new nitrogen containing molecular formulas were formed by a reaction of an SRFA derived CxHyOz molecule with an N1 compound generated by phototransformation of CBZ, the intensity maximum was shifted by 150 to 200 Da (Figure 2a) as compared to SRFA irradiated alone (m/z 250 for CxHyN0Oz mass peaks). This suggests that the CBZ phototransformation products reacting with the SRFA molecules also fell into the same mass range as known CBZ phototransformation products (Table S1). The majority of the newly formed molecules had molecular formulas with O/C ratios between 0.4 and 0.6 and H/C ratios between 0.5 and 1.0 (Figure 2b, CBZ+SRFA). These ratios were lower as compared to the average O/C and H/C ratios after irradiation of SRFA alone (Figure 2b). This could be explained by the reaction of DOM molecules with CBZ phototransformation products, which have a comparatively low oxygen content and high unsaturation (Table S1). Photoincorporation of inorganic nitrogen was previously considered as a possible pathway for the formation of new dissolved organic nitrogen.38,39 Knowing that ammonium was formed upon photolysis of CBZ (see above), it was tested in a separate experiment with SRFA and ammonium whether the incorporation of ammonium was a significant process for the formation of new dissolved organic nitrogen (DON). The FTICR mass spectra of SRFA irradiated in the presence of ammonium (20 μmol L−1) revealed that less than 40 new mass peaks with molecular formulas containing nitrogen were

photolysis of CBZ. Full scan mass spectra of FTICR-MS revealed more than 100 mass peaks likely originating from phototransformation products of CBZ in the m/z range of 150−500 (Figure S1a), the majority containing one or two nitrogen atoms. Due to the large number of new, unaccounted mass peaks in CBZ irradiated alone, we exemplarily examined the identity and interaction with SRFA of one prominent peak at m/z 172.0404 ([C10H6NO2]−). This mass peak showed a high intensity in the CBZ photolysis experiments, especially in the presence of SRFA and was also formed upon photolysis of ACD (see below). It was identified as 3-quinolinecarboxylic acid (3-QCA) by UPLC-QTOF-MS through comparison of its retention time and product ion spectrum with a standard (Figure S2, Figure S3). This phototransformation product of CBZ has not been described previously. Its concentration was estimated with an external calibration as 0.2 μmol L−1 (0.6% of CBZ degraded) in irradiated CBZ and 0.4 μmol L−1 in irradiated ACD. The concentration in irradiated CBZ was in the range of other phototransformation products such as ACD or acridine-9carboxylic acid (ACD-9-COOH) (Table S3). 3-QCA may be formed through oxidative ring cleavage of ACD and subsequent decarboxylation. The mass peak of the corresponding intermediate quinoline-2,3-dicarboxylic acid at m/z 216.0302 was also detected in the FTICR mass spectra after photolysis of CBZ and ACD (Figure S1). Photolytic cleavage of aromatic rings has been observed previously and was attributed to the reaction with hydroxyl radicals.36,37 As already discussed above, hydroxyl radicals are also the main reactive intermediate responsible for indirect photolysis of CBZ in the presence of DOM. This might explain why the relative intensity of the mass peak at m/z 172.0404 ([C10H6NO2]−) in the FTICR mass spectrum was much higher in the presence of SRFA than without (Figure S1a, b). Besides the organic nitrogen compounds, 23 ± 4 μmol L−1 ammonium were formed during photolysis of 50 μmol L−1 CBZ (n = 5, Table S4). Irradiation of CBZ in the presence of SRFA resulted in a higher ammonium concentration at the end of the experiment (38 ± 3 μmol L−1, n = 3), which agrees with the increased degradation of CBZ. Cleavage of the carbamoylgroup of CBZ seems to be the major process of ammonium formation. Ammonium was also observed as the major nitrogen species upon photocatalytic transformation of CBZ with D

DOI: 10.1021/acs.est.7b00823 Environ. Sci. Technol. XXXX, XXX, XXX−XXX

Article

Environmental Science & Technology

Figure 3. left: Extension of the FTICR mass spectra at m/z 428. Red circles indicate newly formed nitrogen mass peaks in a) CBZ (50 μmol L−1) irradiated with SRFA, b) CBZ (50 μmol L−1) irradiated alone, c) SRFA irradiated alone, and d) ACD (50 μmol L−1) irradiated with SRFA; right: product ion spectra of m/z 428 of e) CBZ (50 μmol L−1) irradiated with SRFA and f) ACD (50 μmol L−1) irradiated with SRFA.

Figure 4. Kendrick mass defect (based on C10H7NO2) vs m/z plot showing pairs of molecular formulas where the mass difference is 173.0477 Da for a) CBZ (50 μmol L−1) irradiated in the presence of SRFA, b) SRFA irradiated alone, and c) van Krevelen diagram with SRFA molecular formulas derived from the KMD analysis of a).

SRFA, the most commonly occurring exact mass differences between the molecular formulas originally present in SRFA and the newly formed nitrogen containing molecular formulas after irradiation of CBZ in the presence of SFRA were calculated. The ten most common mass differences ranged between m/z 149 and m/z 175, contained only one nitrogen atom, and had molecular compositions in the range of C7−10H3−13N1O1−3 (Table S5). These mass differences may point at the phototransformation products that reacted with DOM. Indeed, molecular ions corresponding to three of these calculated mass differences were detected in the FTICR-MS spectra of CBZ irradiated alone and in the presence of SRFA. One of these mass differences corresponds to the molecular formula C10H7NO2, which is the same for the transformation product of CBZ that was identified as 3-QCA (see above). To obtain structural information on the nitrogen containing products, product ion spectra were recorded from a series of newly formed nitrogen containing mass peaks at m/z 428 (Figure 3a). These mass peaks were not present after irradiation of CBZ or SRFA alone (Figure 3b, c) but were also formed after irradiation of ACD in the presence of SRFA (Figure 3d). The major fragment ions were m/z 172.0404 ([C10H6NO2]−) and m/z 192.0821 ([C14H10N]−) (Figure 3e). The first fragment ion corresponds to the molecular anion of 3QCA and was also observed in the case of ACD (Figure 3f). This supports the hypothesis that 3-QCA is one of the

formed. Only four of those matched with mass peaks formed upon irradiation of CBZ and SRFA. We conclude that the photoincorporation of ammonium was not relevant under the selected experimental conditions. Reactive CBZ Phototransformation Products. ACD has been described as one of the major phototransformation products of CBZ17,21 and was also determined in our experiments. Its concentration decreased slightly in the presence of SRFA (Table S3). In contrast to other known CBZ phototransformation products ACD carries only one nitrogen atom and was hence assumed to be a potential precursor for the formation of the new CxHyN1Oz molecules. Therefore, an experiment with ACD irradiated in the presence of SRFA was also conducted. Accordingly, almost all new molecular formulas contained only one nitrogen atom, precluding extensive dimer formation in the presence of SRFA. The intensity-based maximum of these mass peaks was between m/z 425 and 475 (Figure 2a, ACD+SRFA), which is about 25 Da higher than observed for CBZ (Figure 2a, CBZ +SRFA). Most of the molecular formulas with one nitrogen atom (CxHyN1Oz) formed after irradiation of CBZ in the presence of SRFA were also formed from ACD (78%). This indicates that ACD or one of its phototransformation products was involved in the formation of these new nitrogen containing mass peaks from CBZ. As an alternative approach to identify possible CBZ phototransformation products that were reacting with the E

DOI: 10.1021/acs.est.7b00823 Environ. Sci. Technol. XXXX, XXX, XXX−XXX

Article

Environmental Science & Technology

Figure 5. One possible reaction pathway of 3-QCA with hydroxyl or carboxyl groups of SRFA.

was detected (Table S6). C−C bond formation is also possible with carboxylic acids, as indicated by experiments with 3-QCA and propionic acid (Table S6). The mass peak of the 3-QCA dimer, likely resulting from the reaction of two 3-QCA radicals without the loss of any small unit (like H2 or H2O), was also detected in the FTICR mass spectra of irradiated CBZ at m/z 345.0881 ([C20H13N2O4]−). The mechanism proposed here for the formation of bound residues from CBZ differs from the one suggested previously for benzotriazole. There, aniline was assumed to be the major reactive intermediate and thought to preferentially react with quinoid groups in DOM.5 The above data all point to the fact that the CBZ phototransformation product 3-QCA is an important reactant forming bound residues with DOM. The proposed reaction pathways would lead to a C−C bond between the previous contaminant and the DOM, implying that these bound residue molecules would be comparatively stable and not prone to hydrolytic liberation of the contaminant. However, not all of the newly formed nitrogen containing mass peaks can be explained by these reactions. Other reactive phototransformation products of CBZ and/or other functional groups in DOM likely contribute to the multitude (>400) of bound residues in our experiments. Bound Residue Formation at Lower CBZ Concentrations. The photolysis experiments described above were performed with a CBZ concentration of 50 μmol L−1 (∼12 mg L−1) to allow for the detection of larger numbers of bound residue mass peaks in the complex DOM pool by FTICR-MS analysis. Typical concentrations of CBZ in surface waters are in the range of 0.3−1.7 μg L−1.8,11 The DOC concentration in the experiments was approximately 12 mg L−1 (25 mg L−1 SRFA) which is at the upper end of concentrations typically observed in lakes and rivers (1−10 mg L−1).43 To assess whether bound residue formation that was observed under these experimental conditions may also be of relevance in the environment, similar experiments were performed with CBZ concentrations that were one and two orders of magnitude lower (5 μmol L−1 and 0.5 μmol L−1). Due to sensitivity issues the number of detected newly formed nitrogen containing mass peaks decreased drastically, but some of the signals found with 50 μmol L−1 CBZ could be also detected with these lower concentrations at the same instrument settings (Table S6), including m/z 172.0404 ([C10H6NO2]−). Using high sensitivity settings for FTICR-MS, the previously discussed products at m/z 428 (Figure 3) were also detected at 5 μmol L−1 and even at 0.5 μmol L−1 CBZ irradiated with SRFA (25 mg L−1). The relative intensities of these three signals ([C21H18NO9]−, [C22H22NO8]−, and [C23H26NO7]−) normalized to the neighboring 13C-isotope of [C20H27O10]−) decreased with decreasing initial CBZ concentration (Figure 6). Interestingly, the decrease of the relative signal intensity with decreasing CBZ concentration was not linear (Figure 6): while CBZ concentrations decreased by two orders of magnitude from 50 to 0.5 μmol L−1, the relative signal intensity of these

phototransformation products of CBZ (and ACD) that reacted with numerous SRFA molecules. The reaction of SRFA molecules with 3-QCA was further corroborated by an experiment using [D215N]-CBZ (Figure S4). In this isotopically labeled standard each aromatic ring of CBZ carries one deuterium (position 4 and 6), and the carbamoyl group carries the 15N-isotope (Figure S4). After irradiation the D1-isotopologues of [C21H18NO9]− and [C22H22NO8]− (Figure 3a) could be detected at m/z 429 (Figure S4a), indicating that one aromatic ring of CBZ was cleaved prior to reaction with the DOM molecules. Moreover the 15N-isotopologues of these mass peaks could not be detected, which confirms that the nitrogen in the product molecules originates from the azepine ring in CBZ and not from the 15N-labeled carbamoyl group. This is further supported by the fact that no 15N-isotopologue was detected for any other nitrogen containing mass peak formed after irradiation of the labeled CBZ with SRFA. Possible Reaction Partners in DOM. KMD analysis was used to further identify those molecular formula in SRFA that reacted with 3-QCA, by normalization to its exact mass (173.0477). For SRFA irradiated together with CBZ 161 mass pairs were found (Figure 4a), whereas in SRFA irradiated alone only 18 pairs were detected (Figure 4b). The tentative reaction partners in SRFA (low mass member of the respective mass pair) had a molecular mass between 160 and 370 Da. Most of them had O/C ratios between 0.5 and 1.0 and H/C ratios between 1.0 and 2.0, which is higher than for the average molecular formulas present in SRFA (Figure 4c). This suggests that molecules with this composition were most reactive toward photochemically generated 3-QCA. No mass peaks corresponding to the twofold addition of 3QCA (C10H7NO2) to SRFA were found in the data. This may, however, be attributed to the selected instrument settings resulting in decreasing sensitivity for m/z > 500. For the reaction products with m/z 428 (Figure 3) the calculated SRFA educt mass peaks at m/z 255.0146 ([C10H7O8]−), m/z 255.0510 ([C11H11O7]−), and m/z 255.0874 ([C12H15O6]−) show high relative intensities in SRFA. Product ion spectra were recorded for these three educt mass peaks, and the fragment ions [M−H−CO2]−, [M−H− CO2−H2O]−, [M−H-2×CO2]−, and [M−H-2×CO2−H2O]− were detected for all of them (Figure S5). Neutral losses of CO2 and H2O are mainly observed for carboxyl and hydroxyl groups in DOM.40,41 These findings indicate that at least one hydroxyl group and two carboxyl groups were present in these three SRFA educts. One possible reaction pathway involving 3-QCA and an alpha-standing carbon atom next to hydroxyl or carboxyl groups in an SRFA molecule is shown in Figure 5. This photoalkylation mechanism was previously proposed for the reaction of quinolines with ethanol.42 In a separate photochemical experiment 3-QCA and ethanol were irradiated, and a product corresponding to the coupling of 3-QCA with ethanol F

DOI: 10.1021/acs.est.7b00823 Environ. Sci. Technol. XXXX, XXX, XXX−XXX

Article

Environmental Science & Technology

residues in soil environment, those bound residues formed with DOM in the aqueous phase remain mobile.



ASSOCIATED CONTENT

S Supporting Information *

The Supporting Information is available free of charge on the ACS Publications website at DOI: 10.1021/acs.est.7b00823. Additional experimental details and results, Figures S1− S6, Tables S1−S7 (PDF)



AUTHOR INFORMATION

Corresponding Author

*Phone: +49 341 235 1261. E-mail: [email protected].

Figure 6. New nitrogen containing mass peaks [C21H18NO9]−, [C22H22NO8]−, and [C23H26NO7]− at m/z 428. Relative peak intensities (based on highest 13C mass peak at m/z 428) are displayed as a function of the CBZ concentrations (0.5, 5, and 50 μmol L−1) used in photolysis experiments with SRFA (25 mg L−1). Intensities of respective peaks present in SRFA irradiated alone (CBZ = 0 μmol L−1) were subtracted.

ORCID

Thorsten Reemtsma: 0000-0003-1606-0764 Notes

The authors declare no competing financial interest.



ACKNOWLEDGMENTS This work was supported by the German Federal Ministry of Education and Research (BMBF) in the framework of the project “Loading of drinking water reservoirs with dissolved organic matter−prediction and prevention” (TALKO, FKz 02WT1290A). The authors are grateful for using the analytical facilities of the Centre for Chemical Microscopy (ProVIS) at the Helmholtz Centre for Environmental Research which is supported by European regional Development Funds (EFRE − Europe funds Saxony) and the Helmholtz Association. We gratefully acknowledge Boris Koch (Alfred Wegener Institute, Helmholtz Centre for Polar and Marine Research) for developing and providing the FTICR-MS molecular formula calculation software and data evaluation pipeline used in this study. We thank three anonymous reviewers for their helpful comments and suggestions.

nitrogen containing molecules decreased by only one order. This suggests that the relative importance of bound residue formation is higher at lower initial concentrations of CBZ. The ratio of DOM to CBZ increases with decreasing CBZ concentration (in these experiments from 1 mg DOC/mg CBZ to 100 mg DOC/mg CBZ), resulting in more DOM molecules that are available to react with one molecule of a phototransformation product of CBZ. In surface water, with the concentrations mentioned above, the DOM/CBZ ratio would be even higher (approximately 1000 mg DOC/mg CBZ). Environmental Implications. Bound residues can be formed through photolytic processes when environmental contaminants are transformed and reactive intermediates are generated. The formation of bound residues from CBZ was studied with the well characterized DOM standard SRFA, and one possible pathway, the reaction of the CBZ transformation product 3-QCA, was examined in more detail. Due to the multitude of CBZ phototransformation products and structural elements in DOM, certainly other reactions also occur in parallel. The proposed mechanism involves hydroxyl and carboxyl groups, which are omnipresent in DOM. It can therefore be assumed that bound residue formation will also take place with DOM of different origin, although its extent and the product spectrum formed may differ. Although hundreds of new reaction products were detected by ultrahigh resolution mass spectrometry, the quantitative importance of bound residue formation upon photolysis of CBZ in the environment could not yet be determined. The data presented here suggest that this process is more important at lower CBZ concentrations (and increasing DOC/CBZ ratios). However, the detection of the respective molecules gets increasingly difficult. The formation of bound residues with DOM appears to be a novel pathway along which photochemically active contaminants disappear from sunlit surface waters, without being mineralized or forming detectable transformation products. Bound residue formation of contaminants with DOM via reactive transformation products may also be initiated by other transformation processes that generate radical intermediates, including biotic processes such as oxidation by laccases of fungi.44,45 Contrary to bound



REFERENCES

(1) Roberts, T. Non-extractable pesticide residues in soils and plants. Pure Appl. Chem. 1984, 56 (7), 945−956. (2) Kästner, M.; Nowak, K. M.; Miltner, A.; Trapp, S.; Schäffer, A. Classification and Modelling of Nonextractable Residue (NER) Formation of Xenobiotics in Soil − A Synthesis. Crit. Rev. Environ. Sci. Technol. 2014, 44 (19), 2107−2171. (3) Gevao, B.; Semple, K. T.; Jones, K. C. Bound pesticide residues in soils: a review. Environ. Pollut. 2000, 108 (1), 3−14. (4) Barraclough, D.; Kearney, T.; Croxford, A. Bound residues: environmental solution or future problem? Environ. Pollut. 2005, 133 (1), 85−90. (5) Weidauer, C.; Davis, C.; Raeke, J.; Seiwert, B.; Reemtsma, T. Sunlight photolysis of benzotriazoles − Identification of transformation products and pathways. Chemosphere 2016, 154, 416−424. (6) Choudhry, G. G.; van den Broecke, J. A.; Webster, G. R. B.; Hutzinger, O. Environmental photoincorporations of polychlorobenzenes into several humic model monomers. Chemosphere 1987, 16 (2), 495−504. (7) Cooper, W. J.; Zika, R. G.; Petasne, R. G.; Fischer, A. M. Sunlight-Induced Photochemistry of Humic Substances in Natural Waters: Major Reactive Species. Adv. Chem. Ser. 1988, 219, 333−362. (8) Ternes, T. A. Occurrence of drugs in German sewage treatment plants and rivers. Water Res. 1998, 32 (11), 3245−3260. (9) Metcalfe, C. D.; Koenig, B. G.; Bennie, D. T.; Servos, M.; Ternes, T. A.; Hirsch, R. Occurrence of neutral and acidic drugs in the effluents of Canadian sewage treatment plants. Environ. Toxicol. Chem. 2003, 22 (12), 2872−2880.

G

DOI: 10.1021/acs.est.7b00823 Environ. Sci. Technol. XXXX, XXX, XXX−XXX

Article

Environmental Science & Technology (10) Castiglioni, S.; Bagnati, R.; Fanelli, R.; Pomati, F.; Calamari, D.; Zuccato, E. Removal of Pharmaceuticals in Sewage Treatment Plants in Italy. Environ. Sci. Technol. 2006, 40 (1), 357−363. (11) Feitosa-Felizzola, J.; Chiron, S. Occurrence and distribution of selected antibiotics in a small Mediterranean stream (Arc River, Southern France). J. Hydrol. 2009, 364 (1−2), 50−57. (12) Andreozzi, R.; Raffaele, M.; Nicklas, P. Pharmaceuticals in STP effluents and their solar photodegradation in aquatic environment. Chemosphere 2003, 50 (10), 1319−1330. (13) Canonica, S.; Laubscher, H.-U. Inhibitory effect of dissolved organic matter on triplet-induced oxidation of aquatic contaminants. Photochem. Photobiol. Sci. 2008, 7 (5), 547−551. (14) Brezonik, P. L.; Fulkerson-Brekken, J. Nitrate-induced photolysis in natural waters: Controls on concentrations of hydroxyl radical photo-intermediates by natural scavenging agents. Environ. Sci. Technol. 1998, 32 (19), 3004−3010. (15) al Housari, F.; Vione, D.; Chiron, S.; Barbati, S. Reactive photoinduced species in estuarine waters. Characterization of hydroxyl radical, singlet oxygen and dissolved organic matter triplet state in natural oxidation processes. Photochem. Photobiol. Sci. 2010, 9 (1), 78− 86. (16) Remucal, C. K. The role of indirect photochemical degradation in the environmental fate of pesticides: a review. Environ. Sci.-Process Impacts 2014, 16 (4), 628−653. (17) Chiron, S.; Minero, C.; Vione, D. Photodegradation processes of the antiepileptic drug carbamazepine, relevant to estuarine waters. Environ. Sci. Technol. 2006, 40 (19), 5977−5983. (18) Matamoros, V.; Duhec, A.; Albaiges, J.; Bayona, J. M. Photodegradation of Carbamazepine, Ibuprofen, Ketoprofen and 17 alpha-Ethinylestradiol in Fresh and Seawater. Water, Air, Soil Pollut. 2009, 196 (1−4), 161−168. (19) Lam, M. W.; Mabury, S. A. Photodegradation of the pharmaceuticals atorvastatin, carbamazepine, levofloxacin, and sulfamethoxazole in natural waters. Aquat. Sci. 2005, 67 (2), 177−188. (20) Doll, T. E.; Frimmel, F. H. Fate of pharmaceuticals−− photodegradation by simulated solar UV-light. Chemosphere 2003, 52 (10), 1757−1769. (21) De Laurentiis, E.; Chiron, S.; Kouras-Hadef, S.; Richard, C.; Minella, M.; Maurino, V.; Minero, C.; Vione, D. Photochemical Fate of Carbamazepine in Surface Freshwaters: Laboratory Measures and Modeling. Environ. Sci. Technol. 2012, 46 (15), 8164−8173. (22) Calza, P.; Medana, C.; Padovano, E.; Giancotti, V.; Baiocchi, C. Identification of the unknown transformation products derived from clarithromycin and carbamazepine using liquid chromatography/highresolution mass spectrometry. Rapid Commun. Mass Spectrom. 2012, 26 (15), 1687−1704. (23) Moran, M. A.; Zepp, R. G. Role of photoreactions in the formation of biologically labile compounds from dissolved organic matter. Limnol. Oceanogr. 1997, 42 (6), 1307−1316. (24) Mopper, K.; Stahovec, W. L. Sources and sinks of lowmolecular-weight organic carbonyl compounds in seawater. Mar. Chem. 1986, 19 (4), 305−321. (25) Cabaniss, S. E.; Madey, G.; Leff, L.; Maurice, P. A.; Wetzel, R. A stochastic model for the synthesis and degradation of natural organic matter. Part I. Data structures and reaction kinetics. Biogeochemistry 2005, 76 (2), 319−347. (26) Kujawinski, E. B.; Del Vecchio, R.; Blough, N. V.; Klein, G. C.; Marshall, A. G. Probing molecular-level transformations of dissolved organic matter: insights on photochemical degradation and protozoan modification of DOM from electrospray ionization Fourier transform ion cyclotron resonance mass spectrometry. Mar. Chem. 2004, 92 (1− 4), 23−37. (27) Gonsior, M.; Peake, B. M.; Cooper, W. T.; Podgorski, D.; D’Andrilli, J.; Cooper, W. J. Photochemically Induced Changes in Dissolved Organic Matter Identified by Ultrahigh Resolution Fourier Transform Ion Cyclotron Resonance Mass Spectrometry. Environ. Sci. Technol. 2009, 43 (3), 698−703. (28) Stubbins, A.; Spencer, R. G. M.; Chen, H. M.; Hatcher, P. G.; Mopper, K.; Hernes, P. J.; Mwamba, V. L.; Mangangu, A. M.;

Wabakanghanzi, J. N.; Six, J. Illuminated darkness: Molecular signatures of Congo River dissolved organic matter and its photochemical alteration as revealed by ultrahigh precision mass spectrometry. Limnol. Oceanogr. 2010, 55 (4), 1467−1477. (29) Reemtsma, T. Determination of molecular formulas of natural organic matter molecules by (ultra-) high-resolution mass spectrometry Status and needs. J. Chromatogr. A 2009, 1216 (18), 3687−3701. (30) Minor, E. C.; Swenson, M. M.; Mattson, B. M.; Oyler, A. R. Structural characterization of dissolved organic matter: a review of current techniques for isolation and analysis. Environ. Sci.-Process Impacts 2014, 16 (9), 2064−2079. (31) Kunenkov, E. V.; Kononikhin, A. S.; Perminova, I. V.; Hertkorn, N.; Gaspar, A.; Schmitt-Kopplin, P.; Popov, I. A.; Garmash, A. V.; Nikolaev, E. N. Total Mass Difference Statistics Algorithm: A New Approach to Identification of High-Mass Building Blocks in Electrospray Ionization Fourier Transform Ion Cyclotron Mass Spectrometry Data of Natural Organic Matter. Anal. Chem. 2009, 81 (24), 10106− 10115. (32) Kendrick, E. A mass scale based on CH2 = 14.0000 for high resolution mass spectrometry of organic compounds. Anal. Chem. 1963, 35 (13), 2146−2154. (33) Riemenschneider, C.; Seiwert, B.; Goldstein, M.; Al-Raggad, M.; Salameh, E.; Chefetz, B.; Reemtsma, T. An LC-MS/MS method for the determination of 28 polar environmental contaminants and metabolites in vegetables irrigated with treated municipal wastewater. Anal. Methods 2017, 9, 1273−1281. (34) Janssen, E. M. L.; Marron, E.; McNeill, K. Aquatic photochemical kinetics of benzotriazole and structurally related compounds. Environ. Sci.-Process Impacts 2015, 17 (5), 939−946. (35) Challis, J. K.; Hanson, M. L.; Friesen, K. J.; Wong, C. S. A critical assessment of the photodegradation of pharmaceuticals in aquatic environments: defining our current understanding and identifying knowledge gaps. Environ. Sci.-Process Impacts 2014, 16 (4), 672−696. (36) McConville, M. B.; Hubert, T. D.; Remucal, C. K. Direct Photolysis Rates and Transformation Pathways of the Lampricides TFM and Niclosamide in Simulated Sunlight. Environ. Sci. Technol. 2016, 50 (18), 9998−10006. (37) Malouki, M. A.; Lavédrine, B.; Richard, C. Phototransformation of methabenthiazuron in the presence of nitrate and nitrite ions. Chemosphere 2005, 60 (11), 1523−1529. (38) Mesfioui, R.; Abdulla, H. A. N.; Hatcher, P. G. Photochemical Alterations of Natural and Anthropogenic Dissolved Organic Nitrogen in the York River. Environ. Sci. Technol. 2015, 49 (1), 159−167. (39) Helms, J. R.; Mao, J.; Schmidt-Rohr, K.; Abdulla, H.; Mopper, K. Photochemical flocculation of terrestrial dissolved organic matter and iron. Geochim. Cosmochim. Acta 2013, 121 (0), 398−413. (40) Witt, M.; Fuchser, J.; Koch, B. P. Fragmentation Studies of Fulvic Acids Using Collision Induced Dissociation Fourier Transform Ion Cyclotron Resonance Mass Spectrometry. Anal. Chem. 2009, 81 (7), 2688−2694. (41) Reemtsma, T.; These, A. Comparative investigation of lowmolecular-weight fulvic acids of different origin by SEC-Q-TOF-MS: New insights into structure and formation. Environ. Sci. Technol. 2005, 39 (10), 3507−3512. (42) Stermitz, F. R.; Wei, C. C.; O’Donnell, C. M. Photochemistry of n-heterocycles. V. Photochemistry of quinoline and some substituted quinoline derivatives. J. Am. Chem. Soc. 1970, 92 (9), 2745−2752. (43) Sigg, L.; Stumm, W. Aquatische Chemie: Einführung in die Chemie natürlicher Gewässer; UTB: Stuttgart, Germany, Vol. 5, completely revised ed., 2011. (44) Jahangiri, E.; Seiwert, B.; Reemtsma, T.; Schlosser, D. Laccaseand electrochemically mediated conversion of triclosan: Metabolite formation and influence on antibacterial activity. Chemosphere 2017, 168, 549−558. (45) Gulkowska, A.; Krauss, M.; Rentsch, D.; Hollender, J. Reactions of a Sulfonamide Antimicrobial with Model Humic Constituents: Assessing Pathways and Stability of Covalent Bonding. Environ. Sci. Technol. 2012, 46 (4), 2102−2111. H

DOI: 10.1021/acs.est.7b00823 Environ. Sci. Technol. XXXX, XXX, XXX−XXX