Article pubs.acs.org/est

Sunlight-Driven Photochemical Halogenation of Dissolved Organic Matter in Seawater: A Natural Abiotic Source of Organobromine and Organoiodine José Diego Méndez-Díaz,†,‡ Kyle K. Shimabuku,† Jing Ma,† Zachary O. Enumah,§ Joseph J. Pignatello,*,§,∥ William A. Mitch,*,⊥ and Michael C. Dodd*,† †

Department of Civil and Environmental Engineering, University of Washington, Seattle, Washington 98195-2700, United States Department of Inorganic Chemistry, University of Granada, Granada. 18071, Spain § Department of Chemical and Environmental Engineering, Yale University, New Haven, Connecticut 06520−8267, United States ∥ Department of Environmental Sciences, The Connecticut Agricultural Experiment Station, 123 Huntington St, P.O. Box 1106, New Haven, Connecticut 06504-1106, United States ⊥ Department of Civil and Environmental Engineering, Stanford University, Stanford, California 94305, United States ‡

S Supporting Information *

ABSTRACT: Reactions of dissolved organic matter (DOM) with photochemically generated reactive halogen species (RHS) may represent an important natural source of organohalogens within surface seawaters. However, investigation of such processes has been limited by difficulties in quantifying low dissolved organohalogen concentrations in the presence of background inorganic halides. In this work, sequential solid phase extraction (SPE) and silver-form cation exchange filtration were utilized to desalt and preconcentrate seawater DOM prior to nonspecific organohalogen analysis by ICP-MS. Using this approach, native organobromine and organoiodine contents were found to range from 3.2−6.4 × 10−4 mol Br/mol C and 1.1−3.8 × 10−4 mol I/mol C (or 19−160 nmol Br L−1 and 6−36 nmol I L−1) within a wide variety of natural seawater samples, compared with 0.6−1.2 × 10−4 mol Br/mol C and 0.6−1.1 × 10−5 mol I/mol C in terrestrial natural organic matter (NOM) isolates. Together with a chemical probe method specific for RHS, the SPE+ICP-MS approach was also employed to demonstrate formation of nanomolar levels of organobromine and organoiodine during simulated and natural solar irradiation of DOM in artificial and natural seawaters. In a typical experiment, the organobromine content of 2.1 × 10−4 mol C L−1 (2.5 mg C L−1) of Suwannee River NOM in artificial seawater increased by 69% (from 5.9 × 10−5 to 1.0 × 10−4 mol Br/ mol C) during exposure to 24 h of simulated sunlight. Increasing I− concentrations (up to 2.0 × 10−7 mol L−1) promoted increases of up to 460% in organoiodine content (from 8.5 × 10−6 to 4.8 × 10−5 mol I/mol C) at the expense of organobromine formation under the same conditions. The results reported herein suggest that sunlight-driven reactions of RHS with DOM may play a significant role in marine bromine and iodine cycling.

■

seawater;7,11 and CH2ICl may form through the photolysis of CH2I2 in surface seawater.3 A much wider variety of organohalogens−including organobromines−may arise from the nonspecific halogenation of DOM by photochemically generated reactive halogen species (RHS) at the ocean surface. Radical and nonradical RHS (e.g., X•, X2•−, XY•−, X2, XY, and HOX; where X = Br or I, and Y = Cl) can be produced in sunlit surface seawater from the oxidation of halides by •OH,12,13 the DOM-photosensitized reduction of IO3−,14 and potentially also by photoexcited DOM or chlorophyll.15−17 Radical RHS may yield organohalogens via

INTRODUCTION

More than 4700 natural organohalogen compounds have been identified, most of which are produced in marine environments.1,2 The vast majority of these are thought to originate from biological processes.1,2 Only a small number of volatile organoiodine and organochlorine compounds, including the important atmospheric ozone-depleting gases CH3I, CH2I2, CH2ICl, and CH3Cl, are presently understood to arise from abiotic processes in marine systems.3−9 For example, CH2I2, CHI3, and CHClI2 may form through reactions of dissolved organic matter (DOM) with HOI and/or I2 generated via I− oxidation by O3 at the ocean surface;6 CH3Cl is believed to arise from nucleophilic substitution of CH3I and CH3Br by Cl−,10 as well as from the photosubstitution of aromatic methoxy groups by Cl− during solar irradiation of DOM in © 2014 American Chemical Society

Received: Revised: Accepted: Published: 7418

April 4, 2014 May 20, 2014 May 27, 2014 June 16, 2014 dx.doi.org/10.1021/es5016668 | Environ. Sci. Technol. 2014, 48, 7418−7427

Environmental Science & Technology

Article

transparent to UV light at wavelengths greater than 250 nm, precluding direct photolysis during exposure to natural sunlight.44 Due to the presence of activating methyl groups at its 3 and 5 positions, DMPZ is susceptible to rapid, specific halogenation at the 4 position during exposure to HOX and X2 species, yielding the respective 4-halo-3,5-dimethyl-1H-pyrazoles (XDMPZs) as products (Scheme 1).45,46

addition to unsaturated C−C bonds and/or recombination with carbon-centered radicals,18−23 while nonradical RHS (e.g., X2, HOX) are widely known to halogenate a variety of organic functional groups via electrophilic substitution and/or addition.24−26 The formation of CH3I at the ocean surface is believed to proceed in part through recombination of I• and • CH3 arising from photochemical reactions of DOM in the presence of I−.8 Several recent investigations highlight the photoinitiated halogenation of specific organic probe compounds (e.g., bisphenol A, phenol, salicylic acid, and allyl alcohol) during sunlight illumination of natural and simulated seawater,27−31 also consistent with a role for RHS. To the authors’ knowledge, however, nonspecific photochemical halogenation of DOM under such conditions has not been reported, due in part to difficulties in quantifying the bulk organohalogen (org-X) content of DOM in seawater in the presence of high background levels of inorganic halides. Methods developed for quantification of org-X content in sediments32−34 and of total organic halogen, or TOX, levels in drinking water35,36 are unsuitable for quantification of org-X in seawater due to interference from high inorganic halide concentrations. Although dissolved organic iodine (DOI) can be quantified in seawater indirectly by measuring iodide concentrations before and after UV photomineralization of DOM,37 the much higher native Br− concentrations in seawater prevent the use of such an approach for organobromine quantification. A number of studies have demonstrated that inductively coupled plasma mass spectrometry (ICP-MS) can serve as a sensitive tool for quantification of organobromine and organoiodine in various natural waters (including seawater in the case of organoiodine), provided that separation of organic and inorganic halogen species is assured prior to detection.38−42 In these earlier investigations, separations were achieved by online coupling of ICP-MS instrumentation with front-end high performance liquid chromatography (HPLC) systems equipped with ion exchange,38,41,42 size-exclusion,39,40 or reversed-phase columns.38 In the present work, a novel sample pretreatment method has been developed to enable the offline purification and preconcentration of DOM from seawater prior to quantification of its organobromine and organoiodine content using ICP-MS. In this method, seawater samples are initially passed through solid-phase extraction (SPE) cartridges to achieve both halide removal and DOM preconcentration, followed by filtration of the reconstituted aqueous DOM extracts through a silver-form anion exchange resin to sequester any residual inorganic halides remaining after SPE. Levels of total Br and I in the recovered DOM are then quantified by ICP-MS and the values compared to corresponding measurements of recovered dissolved organic carbon (DOC) to determine the bulk organobromine (org-Br) and organoiodine (org-I) content of the DOM (in mol X/mol C). This org-X method has in turn been utilized here to quantify native org-Br and org-I levels in a wide variety of natural seawater samples, and to demonstrate that nanomolar levels of org-Br and org-I are generated through photochemical processes during the exposure of DOM in natural and artificial seawater to simulated and natural sunlight. As a complement to the bulk org-X analytical method, an additional technique has been employed that utilizes the synthetic organic compound 3,5-dimethyl-1H-pyrazole (DMPZ) as a selective probe for RHS halogenating activity in sunlight illuminated seawater. DMPZ is highly soluble in water (estimated solubility = 3.8 × 10−1 mol L−1),43 and is

Scheme 1. Formation of XDMPZs in the Reactions of HOX and/or X2 with DMPZ

Neither DMPZ nor the XDMPZs are likely to be initially present in natural seawater.47,48 Consequently, the formation of XDMPZs during solar illumination of natural seawater samples amended with DMPZ can be utilized as a specific indicator of halogenating activity arising from photogenerated RHS, and hence confirmation of RHS formation. This technique has in turn been used to monitor photochemical production of XDMPZs (using HPLC-MS/MS for quantification) during solar illumination of several DMPZ-amended natural estuarine seawater samples, providing additional evidence for the role of RHS in driving bulk org-X formation during solar irradiation of DOM in seawater.

■

MATERIALS AND METHODS Chemicals and Materials. All solutions were prepared using Milli-Q grade (≥18.2 MΩ·cm) deionized water. Artificial seawater (ASW) was prepared according to the formula of Dickson and Goyet49 (Supporting Information (SI) Table S1), and amended with additional reagents (e.g., NaI) as needed. Natural estuarine, coastal, and oceanic seawater samples were collected from a variety of geographic locations (SI Figures S2− S7) and stored under refrigeration at 4 °C until processing. Additional details on chemicals and materials are provided in SI Text S1. Bulk Organohalogen Analytical Method. Prior to extractions, natural seawater samples were filtered through 400 °C baked, 0.7 μm nominal pore size glass fiber filters (Whatman GF/F, GE Healthcare Life Sciences) to remove suspended solids, and treated with 0.1 units/mL of catalase (Sigma-Aldrich C30, from bovine liver) for >30 min to destroy any hydrogen peroxide that could oxidize halides under the acidic extraction conditions.50 Artificial seawater samples subjected to irradiation with natural or simulated sunlight were also treated with catalase as above, but only after irradiations were completed and samples had been removed from the light sources. Control extractions conducted with and without catalase confirmed that it had no influence on organohalogen or DOC concentration measurements in recovered DOM extracts. After acidification to pH ∼ 2 by dropwise HCl addition, a typical 2-L seawater sample was split into 1000, 700, and 300 mL fractions. The DOM of each fraction was then concentrated and desalted according to the general approach depicted in Figure 1a, with essential aspects of SPE procedures adapted from Dittmar et al.51 A portion of the recovered Ba/Ag/H cartridge filtrate was analyzed for DOC while the rest was brought to 5% NH4OH and analyzed for 7419

dx.doi.org/10.1021/es5016668 | Environ. Sci. Technol. 2014, 48, 7418−7427

Environmental Science & Technology

Article

Figure 1. Bulk organohalogen analytical method. (a) Organic matter isolation, preconcentration, desalting, and analytical procedures. Notes: aAs necessary (for natural samples), bOptional−used only for extractions of samples collected in 2009 (see SI Text S6), cSPE manifold and casing rinsed thoroughly with DI H2O prior to elution to remove halides from surfaces in contact with samples, dSeveral samples also analyzed by ion chromatography + ICP-MS, to enable selective quantification of Br− and I−. (b) Example analysis for org-Br content of 2.1 × 10−4 mol C L−1 (2.5 mg C L−1) SRNOM in artificial seawater during duplicate exposure to natural sunlight over the course of 4 days (for an in situ fluence of 543 J cm−2 between 290 and 400 nm) at ambient temperature (293 (±4) K) on a rooftop at the University of Washington in Seattle, WA. (c) Example analysis for org-I content under the same conditions. Dark controls were subject to the same conditions as irradiated samples, except for coverage of glass bottles with aluminum foil. Uncertainties are reported as ±1 SE.

total Br and I by ICP-MS,52 or in certain cases for Br− and I− by ICP-MS coupled with ion chromatography (IC+ICP-MS).53 Additional details on SPE and Ba/Ag/H filtration procedures, ICP-MS and DOC instrumentation and methods, and validation of DOC recovery and halide removal are provided in SI Texts S2, S3, S6, and S7. Due to the extensive processing of water samples in this method, it is unlikely that the org-Br and org-I thus quantified consist of significant proportions of volatile halogenated compounds. Quantification of organochlorine (org-Cl) was not pursued due to the low sensitivity of ICP-MS to Cl. Photochemical Experiments. General Irradiation Conditions. With the exception of some samples irradiated under direct natural sunlight, most irradiation experiments were carried out in an Atlas Suntest XLS+ solar simulator equipped with a 1700 W air-cooled Xe lamp. Lamp output was passed through a daylight filter for attenuation of lamp emission wavelengths below 290 nm and an infrared filter for attenuation of lamp emission wavelengths above 800 nm, resulting in a nominal irradiance of 765 W m−2 between these wavelengths. Simulated and natural solar spectral irradiance measurements were recorded with an Ocean Optics USB2000+XR spectro-

radiometer. Simulated and natural solar fluence rates were measured from 290 to 400 nm (under the same conditions as for sample irradiations), using chemical actinometry with aqueous solutions containing 10−5 M para-nitroanisole and 10−4 M pyridine.54 Additional details on actinometric measurements are included in SI Text S2. Bulk Organohalogen Formation Experiments. For experiments evaluating variations in org-X levels during irradiation under simulated sunlight, samples were loaded into 2 L borosilicate glass bottles containing approximately 100 mL of headspace, sealed with threaded caps (to preclude interference from ambient gas species such as O3), and placed on top of a multiposition magnetic stir plate submerged in a 20 L Plexiglas water bath within the test chamber of the Atlas solar simulator for irradiation. Each experiment involved duplicate irradiated samples and duplicate dark controls (2 L bottles wrapped in aluminum foil) treated in parallel, and each sample was continuously mixed with Teflon-coated magnetic stir bars. Temperature was maintained at 298 (±1) K using an external recirculating constant temperature bath connected to the water bath located within the Atlas test chamber. Control experiments in which the headspace in each bottle was continuously 7420

dx.doi.org/10.1021/es5016668 | Environ. Sci. Technol. 2014, 48, 7418−7427

47° 47° 48° 48° 37° 28° 50° 50° N, 145° W 50° N, 145° W 50° N, 145° W

05/09 06/09 06/09 07/09 07/09 07/09

12/12 12/12 12/12 12/12 06/09 07/09 10/12

10/12

10/12

10/12

Magerholm (Norway) Spurwink River (ME) Higgins Beach (ME) Great Kills Harbor (NY) Hood Canal - Union (WA) Hood Canal - Hansville (WA) Golden Gardens (WA) Golden Gardens (WA) Deception Pass (WA) Deception Pass (WA) Ocean Beach (CA) Sargasso Sea − Surface Pacific Ocean, Station Papa − Surface Pacific Ocean, Station Papa −1250 m Pacific Ocean, Station Papa −2500 m Pacific Ocean, Station Papa −3750 m

7421

O2

O2

O2

E7 E7 E8 E8 C1 O1 O2

E1 E2 E3 E4 E5 E6

NA

NA

map codeb

Oceanic

Oceanic

Oceanic

Estuarine, Estuarine, Estuarine, Estuarine, Coastal Oceanic Oceanic

high tide low tide high tide low tide

IHSS NOM isolate in ASW IHSS NOM isolate in ASW Estuarine Estuarine, low tide Estuarine, low tide Estuarine Estuarine Estuarine

character

34.8

35.3

35.2

21.5 23.6 29.2 28.7 29.8 37.1 33.1

29.9 13.8 27.9 25.0 22.5 28.7

35.0

35.0

salinity (ppt)

ND

ND

ND

ND ND ND ND ND 7.8 8.1

7.9 ND ND ND 7.8 ND

8.1

8.1

pH 58 (±3) 61 (±1) 53 42 48 43 49 48 58 54 43 38 28 48 21 25 (±1) 30 (±1) 29 (±3)

2.1 × 10−4 2.1 × 10−4 0.92 × 10−4 3.9 × 10−4 1.1 × 10−4 2.2 × 10−4 0.83 × 10−4 0.58 × 10−4 0.75 × 10−4 0.83 × 10−4 0.75 × 10−4 0.83 × 10−4 1.0 × 10−4 0.92 × 10−4 1.0 × 10−4 0.58 × 10−4 0.58 × 10−4 0.58 × 10−4

(±3) (±3) (±1) (±2) (±1) (±5) (±1)

(±2) (±10) (±9) (±5) (±2) (±4)

DOC recovery (%)

[DOC] (mol C L−1)

(±0.1) (±0.6) (±0.4) (±0.2) (±0.2) (±0.1) (±0.8)

(±0.5) (±1.0) (±0.4) (±0.3) (±0.2) (±0.3)

3.3 (±0.4)

4.4 (±0.3)

6.0 (±0.1)

4.7 5.6 5.3 5.8 3.2 3.4 6.4

5.8 4.0 4.4 6.1 4.4 5.0

1.2 (±0.1)

0.59 (±0.03)

[org-Br]/[C] ( × 10−4 mol Br/mol C)

11 (±1)

14 (±1)

16 (±1)

25 (±1) 32 (±2) 37 (±2) 38 (±2) NDd NDd 36 (±3)

NDd NDd NDd NDd NDd NDd

1.1 (±0.1)

0.57 (±0.01)

[org-I]/[C] ( × 10−5 mol I/mol C)

isolated DOM propertiesc

a

NA, not applicable, ND, not determined. bMaps of sample locations are provided in SI Figures S2−S7. cUncertainties reported as ±1 SD for carbon recoveries, and ±1 SE for [org-X]/[C] values. Each extraction series comprised three sample volumes, ranging from 100 to 1000 mL, and was performed at least in duplicate. DOC recoveries represent average values across all extracted sample volumes. [org-X]/[C] values were calculated from ICP-MS measurements as depicted in Figure 1. dQuantification of org-I was only undertaken in samples collected after 2009, following expansion of the investigation’s scope to encompass both org-Br and org-I.

41′ 38″ N, 122° 24′ 21″ W 41′ 38″ N, 122° 24′ 21″ W 24′ 5″ N, 122° 39′ 22″ W 24′ 5″ N, 122° 39′ 22″ W 45′ 34″ N, 122° 30′ 58″ W N, 66° W N, 145° W

Vallsjøen Reservoir, (Skarnes, Norway) 62° 25′ 56″ N, 6° 29′ 46″ E 43° 34′ 55″ N, 70° 15′ 29″ W 43° 33′ 40″ N, 70° 16′ 29″ W 40° 32′ 7″ N, 74° 8′ 28″ W 47° 22′ 27″ N, 123° 6′ 36″ W 47° 55′ 34″ N, 122° 36′ 60″ W

NA

NRNOM in ASW

Okefenokee Swamp (GA)

sample location

NA

sample month

SRNOM in ASW

name

sample information

Table 1. Characteristics of Dissolved Organic Matter Isolated from Natural and Artificial Seawater Matrixesa

Environmental Science & Technology Article

dx.doi.org/10.1021/es5016668 | Environ. Sci. Technol. 2014, 48, 7418−7427

Environmental Science & Technology

Article

NOM isolates and (b) natural seawater samples collected from a variety of geographic locations (Table 1, SI Figures S2−S7). Carbon recoveries ranged from ∼40−60% for estuarine samples and NOM isolates spiked into ASW, and from ∼20− 50% for open ocean samples. Recoveries were highly repeatable, featuring relative standard deviations of