Bromination of Marine Dissolved Organic Matter ... - ACS Publications

Jul 13, 2015 - Bromination of Marine Dissolved Organic Matter following Full Scale ... Research Services, Baltimore, Maryland 21231, United States...
0 downloads 0 Views 2MB Size
Article pubs.acs.org/est

Bromination of Marine Dissolved Organic Matter following Full Scale Electrochemical Ballast Water Disinfection Michael Gonsior,*,† Carys Mitchelmore,† Andrew Heyes,† Mourad Harir,‡ Susan D. Richardson,§ William Tyler Petty,§ David A. Wright,†,⊥ and Philippe Schmitt-Kopplin‡,∥ †

Chesapeake Biological Laboratory, University of Maryland Center for Environmental Science, Solomons, Maryland 20688, United States ‡ Helmholtz Zentrum München, Analytical BioGeoChemistry, D-85764 Neuherberg, Germany § Department of Chemistry and Biochemistry, University of South Carolina, Columbia, South Carolina 29208, United States ∥ Technische Universität München, Analytical Food Chemistry, D-85354 Freising-Weihenstephan, Germany ⊥ Environmental Research Services, Baltimore, Maryland 21231, United States S Supporting Information *

ABSTRACT: An extensively diverse array of brominated disinfection byproducts (DBPs) were generated following electrochemical disinfection of natural coastal/estuarine water, which is one of the main treatment methods currently under consideration for ballast water treatment. Ultra-high-resolution mass spectrometry revealed 462 distinct brominated DBPs at a relative abundance in the mass spectra of more than 1%. A brominated DBP with a relative abundance of almost 22% was identified as 2,2,4-tribromo-5-hydroxy-4cyclopentene-1,3-dione, which is an analogue to several previously described 2,2,4-trihalo-5-hydroxy-4-cyclopentene-1,3-diones in drinking water. Several other brominated molecular formulas matched those of other known brominated DBPs, such as dibromomethane, which could be generated by decarboxylation of dibromoacetic acid during ionization, dibromophenol, dibromopropanoic acid, dibromobutanoic acid, bromohydroxybenzoic acid, bromophenylacetic acid, bromooxopentenoic acid, and dibromopentenedioic acid. Via comparison to previously described chlorine-containing analogues, bromophenylacetic acid, dibromooxopentenoic acid, and dibromopentenedioic acid were also identified. A novel compound at a 4% relative abundance was identified as tribromoethenesulfonate. This compound has not been previously described as a DBP, and its core structure of tribromoethene has been demonstrated to show toxicological implications. Here we show that electrochemical disinfection, suggested as a candidate for successful ballast water treatment, caused considerable production of some previously characterized DBPs in addition to novel brominated DBPs, although several hundred compounds remain structurally uncharacterized. Our results clearly demonstrate that electrochemical and potentially direct chlorination of ballast water in estuarine and marine systems should be approached with caution and the concentrations, fate, and toxicity of DBP need to be further characterized.



2004).2 Forty-four countries so far have acceded to the convention, representing 32.86% of the required 35% of gross commercial tonnage required for full ratification. The convention is scheduled to come into force exactly 12 months from the date of full ratification, which is expected sometime in 2015. After the convention enters into force, ballast water exchange will no longer be accepted as a management option, and all qualifying commercial vessels must operate an effective ballast water treatment system capable of complying with the International Maritime Organization standards and a set of very similar standards adopted by the U.S. Federal Government

INTRODUCTION With the ever-increasing movement of goods around the world, the number and size of ocean-going vessels have increased.1 These vessels require the uptake of ballast water of varying amounts to stabilize the ship in transit, especially because cargo flow between ports is not always equal. Water is taken in and discharged from large holding tanks in response to the differing cargo and fuel loads. With many cases of invasive species introduction globally linked to ballast water movement, current reduction policies include open ocean water exchange to minimize introductions of invasive species. Open ocean exchange varies in efficacy, is not suitable in all transport situations, and therefore represents a compromise solution to ballast water management pending ratification of the 2004 International Convention for the Control and Management of Ships’ Ballast Water and Sediments (BWM © XXXX American Chemical Society

Received: March 23, 2015 Revised: June 15, 2015 Accepted: July 13, 2015

A

DOI: 10.1021/acs.est.5b01474 Environ. Sci. Technol. XXXX, XXX, XXX−XXX

Article

Environmental Science & Technology

flushing of the system, and the question of whether discharges are single, pulsed, or continual releases must be taken into consideration together with the inherent abiotic and biotic characteristics of the receiving waters. To date, more than 600 DBPs have been identified and grouped.4g Recent studies have suggested that there are many more compounds5b,10 and that chlorination of drinking water potentially accounts for only 50% of the total organic halogen (TOX) concentration. 11 The relative paucity of such information stems from the limitations of the methods employed. The primary method of identifying and quantifying regulated compounds has been gas chromatography (GC)/ mass spectrometry (MS), which, while effective in detecting the submicrogram per liter concentrations, is not well-suited for the broad spectrum analysis required to identify many of the brominated compounds that appear to form. For example, ultraperformance liquid chromatography (UPLC) interfaced with electrospray ionization (ESI) MS/MS was applied to measure brominated DBPs in chlorinated sewage effluent water.10d Ultra-high-resolution mass spectrometry was used to confirm the high abundances of 2,2,4-trihalo-5-hydroxy-4cyclopentene-1,3-diones (trihaloHCDs),10aand the same technique was applied to reveal the presence of previously unknown brominated DBPs in drinking water.5c Although much concern has been placed on the products resulting from freshwater chlorination for drinking water, the chlorination of estuarine or saltwater introduces much greater complexity because of the very fast formation of reactive bromine. For example, chlorination of ballast water has already been shown to create DBPs such as haloacetic acids, halogenated phenols, halogenated acetonitriles, and halogenated hydrocarbons.12 Furthermore, the chlorination of saltwater at power plants has resulted in various compounds, such as dibromoacetonitrile, dibromochloromethane, bromodichloromethane, and 2,4,6-trichlorophenol.13 It has also been reported that brominated phenols are responsible for taste and odor problems in desalinated drinking water.14 It was also previously observed that the chlorination of sewage discharged into saltwater resulted in the formation of numerous other brominated DBPs such as bromomaleic acid, 5-bromosalicylic acid, 3,5-dibromo-4-hydroxybenzaldehyde, 3,5-dibromo-4-hydroxybenzoic acid, 2,6-dibromo-4-nitrophenol, and 2,4,6tribromophenol.10d The bromination of compounds following chlorination is extremely fast, occurring at a rate that is difficult to quantify.8 It can be interpreted that the addition of chlorine to ballast water will primarily form brominated DBPs that can be, in general, considered to be more toxic compared to their chlorinated analogues.15 The impacts of brominated compounds have often been expressed through genotoxicity.5d The assumed stability of at least some of these compounds suggests they are present in the water following release, but the persistence and fate of brominated DBPs discharged into coastal or estuarine ecosystems are not well-known. At present, it is not known if differences exist between the direct addition of hypochlorous acid and the electrochemical disinfection method. However, electrolytic disinfection or electrochlorination is an effective way to generate HOCl in situ for brackish water and seawater. The electrolysis of seawater is achieved within a Ti/IrO2 electrolyzer composed of anodes and cathodes. Several different electrode materials have been used in the past, but dimensionally stable anodes (DSA) with metal oxide coatings revolutionized electrocatalytic processes.16 Typical metal oxides on titanium anodes include

under the auspices of the U.S. Coast Guard and U.S. Environmental Protection Agency. As such, ballast water treatment systems (BWTS) must eliminate a diverse array of invasive organisms, including zooplankton, phytoplankton, and bacteria, before discharge. BWTS under consideration include mechanical methods such as filtration, ultraviolet light (UV) irradiation and/or chemical treatments such as the use of chlorine, ozone, or advanced oxidation processes (AOP), and reduction of dissolved oxygen levels using exhaust fumes or a combination of the above, although it is clear that no one treatment will be suitable for all classes of vessels and various options will ultimately be employed globally. Besides effectiveness, safety (personnel and ship integrity), and cost considerations, there is also a requirement for the treated discharge to be environmentally safe. The rationale for using direct chlorination or electrochemical disinfection over the others is based on effectiveness against the majority of ballast water organisms, cost, scalability, and ease of use. Chlorination of water is not a new concept and has been routinely employed for decades to sterilize drinking water but also to prevent the fouling of aquatic equipment. Modern electrochlorination units are easily scalable and utilize the already present chloride in seawater to produce hypochlorite; hence, these systems have a big advantage over simple addition of hypochlorite, because they can generate hypochlorite as needed and the maintenance of a specific concentration of disinfectant can be easily achieved. It would be also more hazardous to store large quantities of a highly concentrated hypochlorite solution aboard ships than to electrolytically produce hypochlorite. In the study of the chlorination of freshwater for drinking water, a number of undesirable issues have been uncovered. One issue is the formation of disinfection byproducts (DBPs), which were recognized as early as the 1970s3 and have the potential for adverse health effects (e.g., cytotoxicity, carcinogenicity, and mutagenicity).4 These DBPs are mainly chlorinated organic chemicals, but they also include some brominated compounds, even though bromide is not a large component of freshwater.5 The toxicity of some of these compounds led to the first U.S. Environmental Protection Agency (EPA) regulation in 1979 with expansions in 1998 and 2006.6 It has also been recognized that natural dissolved organic matter (DOM) exacerbates the problem.5a,7 The impact of dissolved organic carbon (DOC) or DOM on chlorination, including the formation of brominated compounds, has been recognized as being complex; however, assessing the quantity and concentration of the reaction products has been limited, and proxies of change, such as measuring changes in the specific UV absorbance at 254 nm (SUVA) and in fluorescence, have been utilized.8 However, little is known about the toxicity of the great majority of DBP compounds. It has been recognized that brominated compounds are more toxic than chlorinated compounds,5d,9 but this is a general statement, as most DBPs have not yet been assessed for toxicity. Furthermore, because multiple chemicals can be produced, the toxicity of these chemical mixtures regarding the additive effect (or superposition), synergy, and potentiation has also not been established. Of obvious importance is whether the concentration of the chemicals released will be of concern in the aquatic environment in which they are discharged. The persistence, transport, and fate of these chemicals also need to be established. As with any point source pollutant, the volume of discharge water relative to the volume of receiving water, the B

DOI: 10.1021/acs.est.5b01474 Environ. Sci. Technol. XXXX, XXX, XXX−XXX

Article

Environmental Science & Technology

Figure 1. Ultra-high-resolution mass spectra of intake water DOM, after disinfection by electrolysis, and the methanol blank, including the molecular formula assignments of major brominated molecular ions at a nominal value of m/z 345.

are capable of assigning unambiguous and exact molecular formulas, although the structure cannot be determined for highmolecular weight compounds because of the large numbers of isomers for any given molecular formula. In comparing the compositions of waters before and after treatment19,21 or over time,22 we can determine shifts in the molecular composition of complex DOM in waters. The primary goal of this research was to utilize ultra-highresolution mass spectrometry to provide a proof of concept regarding the potential of a full operational electrochemical disinfection test system proposed for use in ballast water treatment in forming a variety of known and novel brominated DBPs. The investigated system was already in use and was tested during this study and provided samples in situ.

iridium or rubidium oxides, whereas Ti/RiO2 anodes showed the highest level of production of free chlorine in laboratorybased experiments.17 The electrolytic reaction takes place after a direct current is applied between the electrodes according to the following overall reaction: 2Cl− + 2H 2O → 2HOCl + H 2

(1)

HOCl and hydrogen gas are formed in the electrolyzer at high pH, which is beneficial for the stabilization of HOCl. Hydrogen is removed prior to adding the HOCl solution to the ballast water. The extent of formation of HOCl is directly proportional to the chloride concentration and the applied specific charge18 and makes it highly effective in brackish water and seawater, where less current can be applied to achieve sufficient levels of oxidant. Ultra-high-resolution mass spectrometry has been utilized to describe complex organic matrices at the molecular level and has successfully demonstrated the complexity of effluent organic matter and its changes during wastewater treatment processes,19 as well as the molecular diversity of marine DOM,20 and previously unknown DBPs formed during drinking water disinfection.5b,10a−c All of these studies have shown that highly precise measurements of m/z values combined with soft ionization



METHODS Sampling and Solid Phase Extraction of DOM. Shipboard trials consisted of the use of two ballast water tanks with a total capacity of ∼2500 m3. The uplift/discharge process took 1 h, and sampling ports consisted of a 2 in. hose that was connected to a sampling assembly. Natural estuarine water (salinity of ∼30, pH 7.7, ambient temperature of 16 °C, and ∼4.0 mg L−1 DOC) was collected upstream from the BWTS and immediately extracted to serve as the control or “untreated” ballast uptake water. Water was C

DOI: 10.1021/acs.est.5b01474 Environ. Sci. Technol. XXXX, XXX, XXX−XXX

Article

Environmental Science & Technology

spectrum in the ion cyclotron resonance (ICR) cell. Typically, masses of up to 1000 Da can be found in DOM, but in this study, intensities of ions greater than m/z 600 were already very low in abundance; hence, we decided to consider only ions smaller than or equal to m/z 600. A washing step using 600 μL (80% methanol and 20% water) between each sample was implemented to minimize any sample carryover. Blank methanol samples were frequently recorded and did not show any signs of sample carryover (see also Figure 1), but several m/z peaks are always present in blanks possibly caused by the SPE resin or remaining contaminants in methanol and the instrument itself. The mass accuracy of the used mass spectrometer was much better than the mass of an electron, and it would have been easy to distinguish between multiply charged molecular ions; however, only singly charged molecules were found. The very high mass accuracy in combination with ultrahigh resolution and soft ionization, such as ESI, allowed us to determine exact molecular formulas for the singly charged ions measured by using FT-MS. Further, simple isotope simulations can be used to cross validate assigned molecular formulas, and this technique is particularly useful when more than one abundant isotope of an element is present. Brominated compounds are therefore easily depicted because of the equal abundance of isotopes 79Br and 81 Br. Software-based automated mass assignments were used to quickly assign molecular formulas to a majority of the ions recorded. However, isotope simulation is much more difficult to automate and was done manually using the Bruker Daltonics data analysis software. The highest m/z peak in each individual mass spectrum was used as a reference of 100% relative abundance that corresponded to m/z ions always found in DOM, and all other masses of m/z peaks, including all DBPrelated m/z ions, were scaled accordingly. Ultra-high-resolution mass spectrometric data were visualized by using elemental plots or van Krevelen diagrams,26 where the hydrogen to carbon ratio (H/C) was plotted against the oxygen to carbon ratio (O/C) of all assigned molecular formulas. Kendrick plots27 were also useful to visualize homologous series of molecular formulas that were spaced by only CH2 groups. In this study, we used the modified Kendrick plot, in which unambiguous homologous series could be more easily identified.28

then passed through a full scale proprietary electrochemical test system that was installed aboard the vessel to mimic and represent ballast water treatment. The concentration of total residual oxidant (TRO) or active halogens in this case was sustained at ∼3 mg L−1 for 26 h and then neutralized with sodium bisulfite prior to sampling to represent the ballast water “discharge” sample (water quality conditions mentioned above for the control sample were maintained). The bromide concentration was not directly measured, but given the conservative behavior of bromide in seawater and the average concentration of bromide being 65 mg L−1 in 35 PSU seawater,23 the bromide concentration was estimated to be around 56 mg L−1 at 30 PSU; hence, the measured free halogen concentration of 3 mg L−1 represented all HOBr, because of the extremely fast reaction of HOCl with bromide.8 Seawater DOM cannot be directly measured by mass spectrometry, and desalting is always required; therefore, 1 L water samples of the uptake and discharge were extracted using an established solid phase extraction (SPE) procedure published previously.24 Briefly, the water sample was acidified to pH 2 using formic acid instead of hydrochloric acid to prevent any adduct formation with chloride ions. The sample was then gravity-fed (flow rate of ∼20 mL min−1) through an Agilent Bond Elut PPL SPE cartridge filled with 1 g of a proprietary nonpolar surface and highly functionalized styrenedivinylbenzene (SDVB) polymer. After all the sample had passed through the cartridge, it was rinsed with acidified MilliQ water, dried under vacuum, and eluted with 10 mL of methanol (Chromasolv LC/MS grade methanol, SigmaAldrich). The extraction efficiency for the DOC typically ranges between 50 and 60% for marine and estuarine systems. Previous studies have shown that this SPE method is suitable for extracting complex mixtures of DBPs,10b although it is expected that volatile DBPs are lost during the drying procedure of the SPE resin. However, the volatility of brominated DBPs is lower compared to that of chlorinated analogues, and more highly brominated, small molecules should still be retained by the resin. The pH, dissolved oxygen, salinity, and temperature were measured, and the total residual oxidant levels were determined using a Hach probe. Dissolved Organic Carbon (DOC). The Shimadzu TOC5000 instrument uses a high-temperature catalytic combustion method to analyze aqueous samples for total inorganic carbon (TIC), total organic carbon (TOC), and nonpurgeable organic carbon. Samples for DOC analysis were treated with hydrochloric acid and sparged with ultrapure carrier grade air to degas inorganic carbon. High-temperature combustion (680 °C) on a catalyst bed of platinum-coated alumina beats breaks down organic carbon into carbon dioxide (CO2). The CO2 is carried by ultrapure air to a nondispersive infrared detector (NDIR) where CO2 is detected.25 Ultra-High-Resolution Mass Spectrometry. A Bruker Apex QE 12 Tesla Fourier transform ion cyclotron resonance mass spectrometer (FT-MS) located at the Helmholtz Center for Environmental Health (Munich, Germany) was used to analyze duplicate samples after negative mode ESI. The ESI voltage was set to −3.6 kV, and the sample was injected using a flow rate of 3 μL min−1. Five hundred spectra were averaged to achieve a mass accuracy of 22%. This compound has been suggested in previous studies to be formed by disinfection in the presence of bromide,5b10c and a similar compound, 2,2,4trichloro-5-methoxycyclopent-4-ene-1,3-dione, was also previously reported.4h The molecular ions indicative of dibromo-5hydroxy-4-cyclopentene-1,3-dione as well as the bromo-5hydroxy-4-cyclopentene-1,3-dione were also found at relative abundances of 10 and 1.7%, respectively (Table S1 of the Supporting Information). The chlorinated analogues of these HCDs have been recently confirmed in DBPs formed in a Swedish drinking water treatment plant.10a We also found a small halogenated sulfonate, and the only plausible structure for C2HBr3O3S− is tribromoethenesulfonate (Figure 2). This is an interesting compound with an O/C of 1.5 and hence is often not considered in mass assignments of ultrahigh-resolution mass spectrometry. This is certainly a new DBP, and there are currently no data about this specific compound. However, the core structure of this novel compound is tribromoethene, which is a well-known compound that has been shown to have acute toxicity (LC50 values) at concentrations of 2.4−18 mg L−1.30 Although we did not determine the concentration of this novel compound in this E

DOI: 10.1021/acs.est.5b01474 Environ. Sci. Technol. XXXX, XXX, XXX−XXX

Article

Environmental Science & Technology

the coastal ocean. As stated in a previous review,32 disinfection of ballast water should not cause greater harm than it prevents. The electrolytic generation of chlorine from seawater places this treatment under IMO guideline G9 governing ballast water treatment systems using or generating “active substances”. This guideline essentially makes no differentiation between sodium hypochlorite generated electrolytically and its addition as a chemical solution with respect to discharge requirements. Criteria for safe discharge of treated ballast water include no significant toxic effects as judged by standard whole effluent toxicity (WET) tests, the demonstration of an environmental half-life for residual byproducts of