Article pubs.acs.org/est
Impact of Halide Ions on Natural Organic Matter-Sensitized Photolysis of 17β-Estradiol in Saline Waters Janel E. Grebel,† Joseph J. Pignatello,†,‡ and William A. Mitch†,* †
Department of Chemical and Environmental Engineering, Yale University, Mason Lab 313b, 9 Hillhouse Ave., New Haven, Connecticut 06520, United States ‡ Department of Environmental Sciences, Connecticut Agricultural Experiment Station, 123 Huntington St., New Haven, Connecticut 06504, United States ABSTRACT: Indirect (sensitized) photolysis by natural organic matter (NOM), mainly from terrestrial sources, can be an important mechanism for attenuation of organic contaminants in estuarine waters, but the effect of salt gradients has been poorly investigated. We studied Suwannee River NOM-sensitized photolysis of 17β-estradiol (E2) in freshwater and saline media. Indirect photolysis by 4 mg-C/L SRNOM was much faster than direct photolysis, and quenching by sorbic acid verified the importance of triplet-excited NOM chromophores. Increasing halide concentrations up to seawater levels decreased the photolysis rate by 90%, with approximately 70% of this decrease associated with ionic strength effects, and the remainder due to halide-specific effects. Bromide (0.8 mM in seawater) accounted for 70% of the halide-specific effect. Halide promotion of NOM chromophore photobleaching was shown to play a major role in the halide-specific effect. Compared to chromophore bleaching, indirect photolysis of E2 was 230% faster in freshwater, but 63% slower in seawater. The involvement of hydroxyl radical (HO•) in indirect photolysis of E2 was ruled out by the lack of suppression by tert-butanol. Experiments in D2O−H2O demonstrated that 1O2 was unimportant in freshwater, but accounted for 42% of NOM-sensitized photolysis of E2 in seawater. We project that, as a parcel of water containing E2 moves through the gradient from freshwater to seawater, overall photolysis will decline due to ionic strength, indirect photolysis will decrease due to specific halide effects on NOM photobleaching, and indirect photolysis will decline relative to direct photolysis. Estuarine contaminant fate models may need to account for halide impacts on indirect photolysis of contaminants.
■
(1SENS) by the emission of a photon (fluorescence) or by interaction with quenching agents. Alternatively, the excitedstate electron 1SENS* undergoes a low-probability spin flip to form a triplet-excited sensitizer (3SENS*), known as intersystem crossing (ISC). 3SENS* can fragment (causing photobleaching), or revert to the ground state (1SENS) by emission of a photon (phosphorescence) or by interaction with quenching agents. As reversion to the ground state also requires a low-probability electron spin flip, 3SENS* tend to be longerlived than 1SENS*, and therefore are important photochemical intermediates. Moreover, 3SENS* are reactive species that may target organic contaminants (C) directly by the following: (1) energy transfer leading to contaminant fragmentation and reversion of 3SENS* to 1SENS (photosensitization) or (2) by oxidation (C → Cox in Scheme 1). Additionally, 3SENS* could
INTRODUCTION Estuaries have long been considered critical ecosystems, serving as nutrient-rich nurseries for a host of marine life. With increased development along coastlines, concerns are mounting regarding the impact of organic contaminants on estuarine ecosystems. Estuaries are particularly susceptible because they receive the cumulative inputs from upstream communities along with direct local inputs from industry, oil spills, aquaculture, and so forth. Indirect photolysis of chemical contaminants is an important natural attenuation mechanism protecting these ecosystems, because it leads to destruction of chemicals that do not absorb light, and would otherwise be resistant to photochemical destruction. Scheme 1 provides an overview of indirect photolytic processes.1 Absorption of sunlight by a groundstate singlet sensitizer [1SENS; e.g., light-absorbing groups within natural organic matter (NOM)], promotes an electron to a higher energy orbital, forming an excited singlet-state sensitizer (1SENS*). The 1SENS* may fragment (potentially resulting in photobleaching) or relax to the ground state © 2012 American Chemical Society
Received: Revised: Accepted: Published: 7128
July 29, 2011 June 4, 2012 June 8, 2012 June 8, 2012 dx.doi.org/10.1021/es3013613 | Environ. Sci. Technol. 2012, 46, 7128−7134
Environmental Science & Technology
Article
scavengers of HO•, a secondary photo-oxidant produced during irradiation of NOM.13,14 Because of the association of tripletexcited NOM chromophores and secondary photo-oxidants with indirect photolytic processes, these results suggested that halide ions affect indirect photolysis rates. The objective of this study was to isolate the impact of chloride and bromide from general ionic strength effects on the indirect photolysis of 17β-Estradiol (E2), selected as a model because its direct photolysis under environmentally relevant conditions is minimal,15,16 and because it has been detected in a variety of surface waters.17 Estrogens such as E2 can adversely affect aquatic organisms at extremely low concentrations.18−20 Even in freshwaters, the mechanism of E2 indirect photolysis is unclear. E2 was degraded in the presence of known 1O2 sensitizers (e.g., Rose Bengal).21 The phenolic group of E2 is likely to be involved in its photoreactions. However, previous research on phenols in freshwaters indicated that 1O2 reactions could account for anywhere from a negligible percentage to 33% of the indirect photolytic decay.22−24 In this work, experiments were conducted to quantify the impact on E2 indirect photolysis rates of ionic strength as well as specific halide effects. Further experiments were conducted to evaluate the pathways associated with these salinity effects, including assessing the importance of NOM chromophores, 1O2, and HO•.
Scheme 1. Photochemical Framework
promote contaminant destruction by generating secondary photo-oxidants, including hydroxyl radical (HO•) and singlet oxygen (1O2). For example, energy transfer from triplets to dissolved oxygen forms 1O2. Indirect photolysis has been studied extensively in freshwater systems.2,3 However, the influence of seawater salts on photolysis has been investigated in relatively few works. Previous research noted that degradation of triclosan4 and carbofuran5 was faster in seawater than in freshwater upon illumination with sunlight-relevant wavebands, while no difference was observed for degradation of ibuprofen, ketoprofen, 17α-ethinylestradiol,6 florfenicol or thiamphenicol.7 As the freshwater and seawater samples differed in many ways (e.g., pH, dissolved organic matter type and concentration), the impact of salts could not be isolated. A few studies attempted to isolate a role for halides, the dominant salts in seawater, but most focused on chloride and neglected to include ionic strength controls. Addition of 0.54− 0.75 M chloride quadrupled the rate of 2,4-dinitrotoluene degradation in deionized water,8 and increased degradation of fipronil by 20% in the presence of 5 mg/L dissolved organic matter.9 However, the lack of ionic strength controls precluded isolation of halide effects. Addition of 0.5 M chloride and bromide decreased direct photodegradation of the pesticide fenarimol by 58% and >81%, respectively.10 Fluorescence quenching studies suggested that halides scavenged the excitedstate singlet intermediates. Seawater-relevant concentrations of chloride, bromide, and iodide decreased the photodegradation of chrysene adsorbed to laponite clay by 70%, 24%, and 25%, respectively.11 The halide effects were attributed to halide quenching of singlet oxygen formed by triplet-excited chyrsene. Only one study has specifically addressed an impact of halides on indirect photolysis of contaminants in the context of marine systems. Addition of 0.42 M chloride doubled the rate of indirect photodegradation of carbamazepine, forming some chlorinated products.12 The impact of chloride was attributed to chlorine radicals (e.g., Cl2•−) formed by oxidation of chloride by iron colloids (e.g., hematite). In previous research, we demonstrated that increasing chloride and bromide ion concentrations within seawater-relevant ranges enhanced photobleaching of NOM chromophores.13 The effect was specific to halides, rather than a general ionic strength effect. Bromide was particularly potent, achieving a similar enhancement in photobleaching at approximately 50-fold lower concentration than chloride. Seawater halides are also efficient
■
MATERIALS AND METHODS Sigma Aldrich furfuryl alcohol, deuterium oxide (D2O), sorbic acid, tert-butanol, 17β-estradiol (E2), and EMD sodium perchlorate were used as received. Suwannee River NOM (SRNOM) from the International Humic Substances Society (St. Paul, MN) had a specific UV absorbance at 254 nm (SUVA254) of 4.1 L mg−1 m−1. Halide salts used were Acros analytical grade NaBr (99.5%) and Fluka ultra grade NaCl containing 3.8 × 10−3 mol% Br−.25 Solutions generally were buffered at pH 8.1 with 10 mM phosphate. For NOM-sensitized reactions, 4 mg-C/L of NOM was included. Except where noted, solutions of varying halide ion concentrations were adjusted to constant seawater ionic strength by addition of the inert salt sodium perchlorate. (Observed initial NOM-sensitized degradation rates of 220 nM E2 were similar in the presence of 0.54 M perchlorate (4.5 (±0.8) nM/h) or 0.18 M sodium sulfate (4.6 (±1.8) nM/h), solutions with equivalent ionic strength of 0.54 M.) Some solutions were deaerated with N2 prior to their irradiation in airtight vials, while others were not deaerated and were left open to the atmosphere during irradiation. All samples were stirred during irradiation. Irradiation was conducted in a merrygo-round photoreactor equipped with UVB-UVA lamps (see Grebel et al.13 for emission spectra). The reactor path length was 0.57 cm, and an irradiance of 1.1 × 10−5 Einsteins L−1 s−1 was determined using ferrioxalate actinometry.26,27 Sample absorbance for photobleaching experiments was measured with a Cary 50 Bio UV−vis spectrophotometer. SRNOM was quantified by its absorbance at 360 nm. E2 was analyzed on a Rainin Dynamax HPLC with a Dynamax Fl-1 fluorescence detector. Samples (100 μL) were injected onto an Inertsil ODS-3 C18 column (250 × 4.6 mm, 5 μm particle size). Samples were eluted with 67%/33% methanol/deionized water at a flow rate of 1 mL/min for 24.5 min, a 2 min ramp to 90%/10% methanol/deionized water held for 8.5 min, a 2 min ramp back to 67%/33% methanol/deionized water, and held for 3 min. Excitation and emission wavelengths were 270 and 7129
dx.doi.org/10.1021/es3013613 | Environ. Sci. Technol. 2012, 46, 7128−7134
Environmental Science & Technology
Article
order rate constants for direct E2 photolysis and NOM chromophore bleaching, the initial rates of NOM-sensitized E2 decay were converted to first order rate constants by dividing by the initial E2 concentrations. For example, for NOMsensitized E2 degradation in the presence seawater halides, the rate constant (1.1 (±0.6) × 10−2 h−1) is 63% slower than the chromophore bleaching rate constant. This result has practical implications for the photochemical fate of E2. Since chromophores initiate indirect photolytic processes, alterations in chromophore destruction rates (e.g., by photobleaching) should affect the rate of indirect photodegradation of contaminants. The overall photolysis rate is the sum of direct and indirect photolysis rates. The maximum contribution of direct photolysis in seawater, ignoring competition for photon absorption, would be only 20% of the observed rate. However, although NOM-sensitized photolysis dominates initially, the relative importance of direct photolysis increases as chromophores bleach, both because bleaching exposes E2 to more light, and because bleaching reduces indirect photolysis of E2 by destroying NOM chromophores that initiate indirect photolysis. The decrease in the contribution of indirect photolysis with increasing irradiation is illustrated in Figure 2 by experiments conducted
310 nm, respectively, with 20 nm bandwidths, 100 Hz lamp flash rate and a 600 V PMT voltage.
■
RESULTS AND DISCUSSION Importance of NOM-Sensitized vs Direct Photolysis and Effect of Ionic Strength. As rivers convey E2 into estuaries, E2 passes through a salinity gradient characterized by increasing ionic strength, as well as individual ion concentrations. Initial experiments were conducted to compare the rates of direct E2 photolysis, SRNOM-sensitized E2 photolysis, and SRNOM chromophore photobleaching in deionized water buffered at pH 8.1. Degradation of 50 nM E2 in the absence of SRNOM and added salts could be fit to a first order model giving a decay constant of 1.1 (±0.3) × 10−2 h−1. Photobleaching of 4 mg-C/L SRNOM chromophores in the presence of 220 nM E2 could be fit, just as in our previous study in the absence of E2,13 to a first order model giving a bleaching constant of 2.2 (±0.4) × 10−2 h−1 (Figure 1). Due to its low
Figure 1. Degradation of 220 nM E2 and bleaching at 360 nm of NOM chromophores (4 mg-C/L) over time in mixtures of the two in synthetic seawater or freshwater. Solutions were at equilibrium with atmospheric oxygen at pH 8.1 with 10 mM phosphate buffer. Merrygo-round UVB-UVA photoreactor. Freshwater = no further salt additions. Seawater = 0.54 M chloride and 0.8 mM bromide.
Figure 2. Degradation of 220 nM E2 and bleaching of NOM chromophores (4 mg-C/L) over time in mixtures of the two. Perchlorate = 0.54 M perchlorate. Prebleached perchlorate = 0.54 M perchlorate and NOM photobleached for 24 h prior to E2 addition. Seawater = 0.54 M chloride and 0.8 mM bromide. Solutions were at equilibrium with atmospheric oxygen at pH 8.1 with 10 mM phosphate buffer. Merry-go-round UVB−UVA photoreactor.
concentration, E2 was not expected to affect the rate of chromophore bleaching. In the presence of 4 mg-C/L Suwannee River NOM, degradation of 220 nM E2 was initially forced to fit a pseudo-first order decay model, which is valid if E2 degradation were significantly faster than the photobleaching of chromophores believed to initiate the SRNOMsensitized photolysis of E2. As can be seen (Figure 1), E2 degradation is significantly faster (by 230%) than chromophore bleaching. The resulting pseudo-first order sensitized E2 decay constant, 5.0 (±0.1) × 10−2 h−1, is 460% higher than the first order rate constant for direct photolysis. This result concurs with previous research suggesting the importance of indirect photolysis for the fate of E2 in freshwater systems.15,16 When these experiments were repeated in the presence of 0.54 M chloride and 0.8 mM bromide, representative of seawater conditions, the direct photodegradation rate constant of E2 declined 80% to 2.2 (±0.2) × 10−3 h. At the same time, chromophore photobleaching increased 36% to 3.0 (±0.4) × 10−2 h−1. However, in this case, E2 degradation was slower than chromophore bleaching (Figure 1), suggesting that pseudo-first order decay conditions do not pertain. Accordingly, initial rates of NOM-sensitized E2 degradation were monitored over times
