Article pubs.acs.org/est
Dark Formation of Hydroxyl Radical in Arctic Soil and Surface Waters Sarah E. Page,† George W. Kling,‡ Michael Sander,† Katherine H. Harrold,§ J. Robert Logan,† Kristopher McNeill,*,† and Rose M. Cory*,§ †
Institute of Biogeochemistry and Pollutant Dynamics, Swiss Federal Institute of Technology (ETH), Zurich, Switzerland Department of Ecology and Evolutionary Biology, University of Michigan, Ann Arbor, Michigan 48109, United States § Department of Earth and Environmental Sciences, University of Michigan, Ann Arbor, Michigan 48109, United States ‡
S Supporting Information *
ABSTRACT: Hydroxyl radical (•OH) is a highly reactive and unselective oxidant in atmospheric and aquatic systems. Current understanding limits the role of DOMproduced •OH as an oxidant in carbon cycling mainly to sunlit environments where •OH is produced photochemically, but a recent laboratory study proposed a sunlight− independent pathway in which •OH forms during oxidation of reduced aquatic dissolved organic matter (DOM) and iron. Here we demonstrate this non− photochemical pathway for •OH formation in natural aquatic environments. Across a gradient from dry upland to wet lowland habitats, •OH formation rates increase with increasing concentrations of DOM and reduced iron, with highest •OH formation predicted at oxic−anoxic boundaries in soil and surface waters. Comparison of measured vs expected electron release from reduced moieties suggests that both DOM and iron contribute to •OH formation. At landscape scales, abiotic DOM oxidation by this dark •OH pathway may be as important to carbon cycling as bacterial oxidation of DOM in arctic surface waters.
■
INTRODUCTION Hydroxyl radical is one of the strongest oxidants in the environment1,2 and affects carbon oxidation rates and the degradation of primary and formation of secondary pollutants in the atmosphere.3 In aquatic environments, •OH is produced by photochemical reactions of dissolved organic matter (DOM) and may be important for transforming recalcitrant molecules within the DOM pool that are otherwise difficult to photo- or biodegrade.4−6 The commonly held assumption that sunlight is necessary for •OH production from DOM has been recently questioned,7,8 especially because electrochemically reduced isolates of aquatic DOM containing redox−active hydroquinone moieties9,10 were shown to transfer three electrons to O2 to form •OH.8 In addition, oxidation of Fe(II) upon exposure of iron−rich waters to oxygen can result in the formation of •OH by Fenton−type chemistry.11−17 These findings suggest that dark •OH formation occurs in environments where reduced DOM and Fe(II) produced by anaerobic microbial respiration18,19 are exposed to O2, such as at oxic− anoxic interfaces (Figure 1). Such environments include wetlands, soils, sediments, and surface waters containing oxic−anoxic boundaries where oxygen concentrations vary intermittently due to soil flushing, hyporheic exchange, or mixing events in the overlying water.20,21 We measured the formation of •OH in natural waters from 50 sampling sites across gradients in redox status, pH, and iron and DOM concentrations in the arctic tundra near Toolik Lake, Alaska during summer 2012 (Table 1). We sampled along vegetation toposequences from hilltop heath to mid−slope tussock tundra to lowland wet sedge that are drained by water © 2013 American Chemical Society
Figure 1. Dark formation of hydroxyl radical from reduced dissolved organic matter and iron. Reduced dissolved organic matter (DOM) and ferrous iron are produced during anaerobic microbial respiration and are stable under anoxic conditions. When these reduced species come into contact with O2, the reduced DOM and ferrous iron will reduce O2 by three electrons to form •OH. The •OH formed is expected to irreversibly oxidize DOM, forming CO2 and low molecular weight, bio−available products in the process. After the reduced DOM and iron are oxidized by oxygen, they may again be reduced if anaerobic conditions are re-established.
tracks of birch−willow vegetation and represent the dominant low−arctic landscapes.22 Shallow depths of thaw to permafrost Received: Revised: Accepted: Published: 12860
July 26, 2013 October 10, 2013 October 10, 2013 October 10, 2013 dx.doi.org/10.1021/es4033265 | Environ. Sci. Technol. 2013, 47, 12860−12867
Environmental Science & Technology
Article
Table 1. Chemistry of Surface and Soil Waters of Different Ecosystemsa variable sample number electron donating capacity (μmol L−1)c electrons released (μmol L−1) DOC (μmol L−1) reduced DOC (μmol L−1)d Fe(II) (μmol L−1) Fetotal (μmol L−1) Mntotal (μmol L−1) pH T (°C) conductivity (μS cm−1) DO (μmol L−1) NH4+ (μmol L−1) alkalinity (μmol L−1) Cl− sample number = Cl− (μmol L−1) SO42− sample number = SO42− (μmol L−1)
all water 69 167 ± 115 106 ± 84 1383 ± 503 22 ± 8 64 ± 94 69 ± 96 2.3 ± 2.6 5.2 ± 0.5 10.3 ± 2.8 57 ± 98 75 ± 102
surface water 8 66 ± 64 36 ± 44 915 ± 31 15 ± 5 1.8 ± 0.3 2.6 ± 0.5 0.8 ± 0.4 6.0 ± 1.2 10.4 ± 4 44 ± 38 282 ± 23 0.69 ± 0.56 357 ± 380 105a 5.2 ± 4.1a 132a 4.4 ± 3.4a
soil water all soil water
wet sedge
birch−willow
tussock
heath
61 178 ± 114 115 ± 83 1441 ± 496 23 ± 8 72 ± 97 78 ± 99 2.4 ± 2.6 5.2 ± 0.3 10.3 ± 2.6 70 ± 102 36 ± 43 0.88 ± 0.14 34 ± 3 1120b 15 ± 68b 1148b 1.3 ± 1.5b
33 237 ± 116 146 ± 87 1716 ± 478 28 ± 8 116 ± 108 120 ± 109 3.1 ± 3.2 5.2 ± 0.3 10.3 ± 2.6 52 ± 35 16 ± 8
12 125 ± 75 86 ± 76 1230 ± 158 20 ± 5 37 ± 61 45 ± 68 1.8 ± 1.9 5.1 ± 0.5
14 95 ± 48 68 ± 43 1027 ± 217 17 ± 4 7 ± 18 11 ± 25 1.1 ± 0.4 5.3 ± 0.3
2 111 ± 20 72 ± 61 1067 ± 949 17 ± 15 4±6 13 0.01 5.3
109 ± 175 33 ± 1
83 ± 135 85 ± 66
22
a
All sites were located near the National Science Foundation Arctic Long-Term Ecological Research (LTER) site at Toolik Lake (68° 38′ N, 149° 36′ W; elevation 720 m). Mean values and standard deviations are reported for all measurements when possible. Alkalinity and NH4+ values were measured on similar, nearby soil and surface waters collected within one week of our sampling. Data for SO42− and Cl− concentrations for surface waters were taken from ref 24, and concentrations for soil waters were taken from the Arctic LTER database (http://ecosystems.mbl.edu/arc/). The molar ratio of DOC to POC in surface waters was previously found to be 16.24 aRef 24. bhttp://ecosystems.mbl.edu/arc/ c680 ± 10 mV (vs SHE). d Calculated as a constant fraction of DOC as described in the Supporting Information, assuming all quinone C in the DOC pool was reduced.
solution, certified), and hydrochloric acid (trace metal grade) were obtained from Fisher Scientific. Nitrogen gas (99.0%) was obtained from Airgas. Water for stock solutions was obtained from a Barnstead E-Pure and B-Pure deionization (DI) system. Instrumentation. Fluorescence was measured on a Horiba Aqualog Fluorometer. UV−visible absorbance measurements were conducted either on the Fluorometer or on an USB4000 spectrophotometer (OceanOptics). Field Sampling. Soil water samples were collected 10−20 cm below the surface using stainless steel needles attached to plastic syringes with Tygon tubing and 3-way valves. The needle and syringe were rinsed with sample three times before collection of bubble-free soil water. Surface water samples were collected directly into triply rinsed amber HDPE bottles, and subsamples were then taken directly from the bottle. To test for possible biotic contributions to the observed •OH formation, bubble−free samples were poisoned in a syringe with anoxic mercuric chloride (1% v/v) before further analysis. No difference in •OH formation between untreated and poisoned samples was observed (Supporting Information, Figure S1). Hydroxyl Radical. Quantification of •OH formation began immediately after sample collection in the field. Subsamples were transferred from the sampling syringe to a smaller, gas− tight syringe using a 3-way valve. After rinsing the valve and smaller syringe with sample, 200 μL of soil water was injected into 3 mL aliquots of either oxic DI water (blank samples, duplicate) or oxic 1.5 mM TPA (triplicate). Samples were oxidized and incubated in the dark for 24 h to allow for oxidation of samples (note that lab studies indicated nearly complete •OH formation within 12 h of air exposure),8 after which 3 mL aliquots of either 1.5 mM TPA (blank samples) or DI water were added to the vials. Sample fluorescence was immediately measured after the addition of TPA or DI water (λex = 310 nm, λem = 425 nm). Standard addition of 0, 40, and
(30−50 cm in summer) characterized all habitats and commonly lead to water−logged, anoxic soils. Downslope patterns in vegetation and soil−water residence times resulted in gradients in concentrations of DOM, reduced iron, and dissolved oxygen (DO),20,23 the major variables proposed to control dark •OH formation (Figure 1).8 We exposed soil and surface water to air (O2) and quantified •OH formation by reaction with the added probe compound terephthalate. The extent of •OH formation was related to the number of electrons released from the sample during the oxidation event, as quantified by reductive de−colorization of the one electron oxidant ABTS+• (2,2′-azino-bis(3-ethylbenzothiazoline-6-sulfonic acid)) before and after 24 h of sample aeration. We analyzed each sample also for DO, dissolved organic carbon (DOC), ferrous iron, total iron, and manganese concentrations, along with pH, conductivity, and temperature (Table 1) to investigate the system properties that control •OH formation in natural environments.
■
MATERIALS AND METHODS Chemicals. All chemicals were used as received unless otherwise noted. Terephthalic acid (98%) and 2-hydroxyterephthalic acid (97%) were obtained from Aldrich. The disodium salt of terephthalic acid (TPA) was prepared as previously described.25 2,2′-azino-bis(3-ethylbenzothiazoline-6sulfonic acid) (ABTS; 98%) was obtained from Biotang, Inc. Potassium persulfate (≥99%) was obtained from Acros Organics. Ferrous ammonium sulfate (ACS grade), 3-(2pyridyl)-5,6-diphenyl-1,2,4-triazine-p,p′-disulfonic acid disodium salt hydrate (ferrozine; ≥98%), 4-(2-hydroxyethyl)piperazine-1-ethanesulfonic acid (HEPES; ≥99.5%), hydroxylamine hydrochloride (≥99%), potassium phosphate monobasic (reagent grade), sodium phosphate dibasic anhydrous (reagent grade), mercuric chloride (>99%), sodium hydroxide (1.0 N 12861
dx.doi.org/10.1021/es4033265 | Environ. Sci. Technol. 2013, 47, 12860−12867
Environmental Science & Technology
Article
These samples were transferred to triplicate gas−tight exetainer vials (Labco, Inc.) in the field. Mercuric chloride (1% v/v) was added immediately to kill any microorganisms and the vials were capped and stored in the dark at 4 °C until analysis. Dissolved oxygen (DO) was measured from the exetainer vials using a membrane inlet mass spectrometer (MIMS; Bay Instruments).29
80 nM hTPA was used to calibrate the hTPA response in each sample, with a linear response observed for all samples. The moles of •OH formed per L of undiluted sample were calculated following eq 1.25 mol•OH =
mol•OH molhTPA × × Vsample V molhTPA
(1)
■
Electrons Released. ABTS was previously shown to quantify the number of electron donating moieties of DOM.26 Solutions containing dissolved ABTS+• were prepared by combining equal volumes of 3.6 mM ABTS and 1.4 mM potassium persulfate, followed by equilibration for 12 h. During sample oxidation measurements, there was minimal change in the redox potential of the solution, because the initial conditions were 0.56 < [ABTS]/[ABTS+•] < 1.0, and the final conditions were [ABTS]/[ABTS+•] < 1.78. On average, the potential of all solutions measured with ABTS oxidation, whether pre- or post-O2 exposure, was 680 ± 10 mV (versus Standard Hydrogen Electrode). Serum vials were prepared with 1.2 mL 320 μM total ABTS, crimped, and sparged with watersaturated nitrogen to ensure that there was no oxygen in the vials to interfere with the subsequent quantification of the number of electrons released. Analyses for number of electrons detectable by ABTS+• in the initial sample began immediately after the analyses for •OH started, and 100 μL subsamples from the original sample syringe were injected into the anoxic ABTS−containing vials as described above. The injection was made below the surface of the ABTS solution. Samples were kept in the dark for at least 15 min before addition of 2 mL DI water and measurement of absorbance of ABTS+• at 734 nm. Additionally, 100 μL samples were added to 1.2 mL airequilibrated DI water to initiate oxidation. These samples were kept in the dark for 24 h for oxidation, after which 2 mL 175 μM ABTS/ABTS+• was added. Samples were kept in the dark for at least 15 min before measurement of absorbance at 734 nm. The concentration of ABTS+• in each sample was calculated using the molar absorptivity (ε = 1.5 × 104 M−1 cm−1 at 734 nm).27 Because ABTS+• is a one−electron oxidant, the moles of electrons released can be calculated following eq 2.
RESULTS AND DISCUSSION Higher Initial Dissolved Oxygen Concentrations Decreased Hydroxyl Radical Formation. We detected formation of •OH in the dark in all samples when exposed to O2 (N = 69). Introduction of O2 initiated •OH formation, and the greatest extent of formation was in anoxic or suboxic waters (50 μM DO was on average ∼23% of the mean dark •OH formation in waters with
