A New Method to Radiolabel Natural Organic Matter by Chemical

Sep 1, 2007 - In the first step, 10 mg of FA was added to the reaction vial (1.95 mL of 5.13 g L-1 FA stock solution) through the gastight addition fu...
0 downloads 0 Views 234KB Size
Environ. Sci. Technol. 2007, 41, 6776-6782

A New Method to Radiolabel Natural Organic Matter by Chemical Reduction with Tritiated Sodium Borohydride RUTH M. TINNACHER AND BRUCE D. HONEYMAN* Environmental Science and Engineering Division, Colorado School of Mines, Golden, Colorado 80401

In this paper, we describe a new method for labeling NOM with the radioisotope tritium (3H) using fulvic acid (FA) as the target NOM fraction. During labeling, FA ketone groups are chemically reduced with tritiated sodium borohydride (NaBH4), while the chemical functionality of the carboxyl and phenol groups is preserved. The labeling procedure was optimized in efficiency experiments that determined the excess concentration of tritiated NaBH4 required for optimum reduction conditions. The chemical characterization of the labeled FA product using FTIR and 1H NMR spectral analysis confirms the proposed reaction mechanism and rules out any significant amounts of impurities or undesirable side reactions. Results from size exclusion chromatography indicate that the tritium label is distributed uniformly over the whole molecular size range of FA and that it is stable over time and under various pH conditions. Potential differences in FA sorption behavior onto mineral surfaces due to labeling were excluded based on experimental data. This method produces NOM of high specific activity (e.g., 1.9 mCi mg-1 FA); this permits the tracing of FA at a detection limit of 0.3 µg L-1 FA.

1. Introduction Natural organic matter (NOM) presents a number of challenges in understanding its behavior in natural and engineered systems. NOM is a mixture of natural polyelectrolytic acids in soils and waters that cannot be further classified based on their chemical structure (1). It has the ability to associate with metal ions and organic compounds (2, 3) and exhibits a relatively high persistence in the environment due to a comparably low potential for microbial decomposition (2, 4). The use of tagged compounds as tracers for studying the physical, chemical, and biological behavior of a chemical species has a long and rich history in environmental science. Uses of tracers include isotope dilution for determining the mass and quantity of a substance, isotopic exchange to elucidate the labile fraction of a chemical species of interest, and the biodegradation of substrates during microbial activity. Characteristics of the ideal tracer are that (i) the tracer’s behavior is identical to the target compound and (ii) it can be distinguished from the material being traced. Another advantage of a labeled form of NOM is that it provides the ability to conduct laboratory experiments on * Corresponding author phone: (303)273 3420; fax: (303)273 3413; e-mail: [email protected]. 6776

9

ENVIRONMENTAL SCIENCE & TECHNOLOGY / VOL. 41, NO. 19, 2007

systems containing NOM at environmentally relevant concentrations. For example, typical concentration ranges of humic substances in groundwater have been reported between 0.04 and 8.6 mg L-1 (5) and 0.1 and 10 mg L-1 organic carbon (1). Conventional analytical techniques for NOM quantification, such as dissolved organic carbon (DOC) analysis, UV absorbance, and fluorescence, often lack sufficient sensitivity and/or exhibit a dependence on solution conditions and organic ligand speciation (6-9). For instance, the UV absorbance of NOM can be affected by pH and NOMmetal complexation reactions in solution and may not be evenly distributed over the whole NOM molecular size range. The radiolabeling of organic ligands can provide a sensitive tool for the analysis of organic ligand concentrations over a range of experimental conditions and concentrations. The requirements for an ideal labeling method for humic substances include the following: (i) high specific activities and stability of the labeled product over a wide range of system conditions (e.g., pH, Eh), including the chemical stability of the label as well as potential isotope exchange reactions; (ii) minimal alteration, addition, or blocking of relevant reactive NOM groups; (iii) preservation of the intrinsic, complex nature of humic substances; (iv) sound understanding of the chemical mechanisms involved in the labeling reaction and a detailed chemical characterization of the labeled product regarding its physicochemical differences relative to the original material; both ensure that potential method limitations are well-understood and can be evaluated prior to the use of the product in specific applications; and (v) a predictive model for the labeling efficiency and specific activity of the target product allowing for the tailoring of the method to new applications. Other organic matter fractions containing known concentrations of reactive groups suitable for the labeling reaction can then be labeled without extensive experimental testing. Currently available radiolabeling techniques for humics and NOM are hampered by various chemical problems (e.g., a limited chemical stability of the label and a lack of detailed understanding regarding the underlying mechanisms of the labeling procedure (Table S1, Supporting Information)). The objective of this research was to develop a method for the radiolabeling of NOM that meets as many of the previously stated criteria as possible. Some of the required characteristics of the radiolabeled FA are demonstrated experimentally; other properties can be assumed based on the current understanding of the chemical characteristics of reduction reactions with sodium borohydride and of its reaction products. Furthermore, the development of several types of reduction efficiency models for this radiolabeling reaction will be presented in detail in a companion publication.

2. Experimental Procedures 2.1. Radiolabeling Technique. 2.1.1. Chemical Background. The use of tritium as an ancillary tracer for organic molecules has numerous advantages. Tritium-labeled compounds can be produced at a relatively high specific activity due to the short half-life of tritium (12.26 years as compared to 5730 years for 14C). Analysis of the label is straightforward through liquid scintillation counting (10, 11). However, the successful application of tritium-labeled compounds depends on sufficient knowledge of the integrity of the carbon-hydrogen bond under various chemical conditions. In this method, fulvic acid (FA) is radiolabeled with tritium (T) by its chemical reduction with tritiated sodium borohydride (NaBH4-xTx). FA carbonyl groups of aliphatic and aromatic ketones as well as quinones are selectively reduced 10.1021/es070563b CCC: $37.00

 2007 American Chemical Society Published on Web 09/01/2007

to yield secondary alcohols and hydroquinones/phenols, respectively. This reduction is a two-step process based on the following reaction (12):

of successful tritium labeling of other organic compounds with NaBH4 are reported in the literature (e.g., 23). Tritiated NaBH4, however, can decompose in neutral and acidic aquatic solutions, leading to the production of tritiated hydrogen gas (18).

NaBH4 + 4H2O f NaB(OH)4 + 4H2 The first step is a nucleophilic addition reaction, where NaBH4 acts as a donor of a hydride ion that is attracted to the partial positive charge of the carbonyl carbon. In the second step, water protonates the tetrahedral alkoxide intermediate and yields the alcohol product. In the case of a reaction with tritiated NaBH4, at most one tritium atom can be bound to one carbon per carbonyl group reduction. Since the proton forming the secondary alcohol originates from water, the stability of the tritium label is not affected by the dissociation of alcohols, and the labeled product can be assumed to be chemically stable over a wide pH range. Furthermore, the covalent carbon-tritium bond is expected to be stable under both oxidizing and reducing conditions, as the oxidation of alcohols by dissolved oxygen is a slow process, and reducing conditions should not have any significant effects on the products of a reduction reaction. The possibility of the loss of tritium from radiolabeled FA due to isotope exchange reactions in water was tested experimentally using size exclusion chromatography (SEC) as a characterization tool. While FA carbonyl groups of both ketones and quinones are selectively reduced, the reduction products of quinones are not expected to be stable in aerobic environments over time. Literature suggests that FA hydroquinones formed by NaBH4 reduction can be reoxidized in reactions with molecular oxygen or by intramolecular redox processes (13, 14). Since these reactions would lead to free T+ in solution, this could result in large analytical errors for the quantification of radiolabeled organic matter concentrations. Therefore, as part of the radiolabeling procedure, aeration and cleanup steps are performed after completion of the reduction reaction to reoxidize and remove unstable products under controlled experimental conditions. FA contains various types of functional groups. Carboxyl and phenol groups are assumed to play a major role in FA metal complexation reactions, and aromatic and aliphatic moieties contribute to the hydrophobic properties of FA, which affect its accumulation on the solid water interface (15). While ketones may also be involved in metal complexation reactions, they only represent weak complexing sites as compared to carboxyl and phenol groups. Therefore, the consumption of FA carbonyl groups during labeling minimally alters the chemical behavior of FA. Carbonyl groups may be of importance, however, for the investigation of the covalent binding of aniline to humic substances (13, 14), the photoproduction of carbonyl sulfide and carbon monoxide from dissolved organic matter (16), and in direct studies of humic ketone groups (17). Furthermore, the production of additional alcoholic groups during labeling will render the organic matter slightly more hydrophilic. In terms of the physical size characteristics of FA, an even distribution of the label over the whole FA size range is expected, as similar ketone concentrations can be assumed in all FA size fractions. Furthermore, this assumption was tested experimentally using SEC as a characterization tool. Sodium borohydride, used in this method, is the primary reductant for organic synthesis on the industrial scale (18). Considerable information is available on NaBH4 reaction mechanisms and products (18-22), as well as on examples of successful reductions of humic substances with NaBH4 (2, 13, 14, 16, 17). Tritiated NaBH4 exchanges no tritium upon dissolution in water or alkali (10), and numerous examples

As sodium borohydride degradation can be minimized by the optimization of solution parameters, such as pH and temperature (18), the labeling reaction is performed at alkaline pH and at two temperature levels. However, additional safety measures are included in the labeling setup to avoid potential exposure to tritium gas. 2.1.2. Reagents. All solutions were prepared with autoclaved UV water (Barnstead EASYpure UV compact ultrapure water system). Suwannee River Reference fulvic acid (International Humic Substances Society, cat. no. 1R101F-1) was selected for radiolabeling as it represents a wellcharacterized, commonly available humic material. Tritiated sodium borohydride (100 mCi at 500 mCi mmol-1) was purchased from American Radiolabeled Chemicals, Inc. (ART 121), and non-radioactive NaBH4 powder from Fisher Scientific (>98% purity, S678-25). A cation exchange resin (Biorad AG MP-50, cation exchange capacity: 1.5 mequiv mL-1 (moist) and 3.5 mequiv g-1 (dry)) in the hydrogen form was used in the first NaBH4 cleanup step after preconditioning the resin in the following manner: resin (25 g) was packed as a slurry into two glass columns, fed with 100 mL of UV water for the removal of fines, and treated with 100 mL of ∼10% hydrochloric acid. Then, the resin beds were rinsed with UV water (at least 900 mL in 100 mL batches) until the pH of the effluent stabilized at the original value. The preconditioned resin was partially air-dried and stored airtight. This resin was selected due to its previous successful use with humic substances (24). 2.1.3. Experimental Setup for Labeling Reaction. All relevant steps of the labeling experiments were performed in hoods. Glassware in contact with FA was acid and base washed (10% (v/v) HCl and 1% (w/v) NaOH) and autoclaved. A gastight, round flask (50 mL, three-neck 19/22, Kemtech America: F439950) was set up as the reaction vial in a water bath at 60 °C on a Corning stirrer/hotplate (Figure S1, Supporting Information). The openings of the flask were used to insert a small pH probe (VWR Scientific: cat. no. 34105021), a thermocouple (Pasco Scientific CI-6505 temperature sensor), and a gastight glass addition funnel with a stopcock (14/20 joint, ACE glass: 7257/9498). They also provided the gas in- and outlets (Masterflex tubing through greased rubber stoppers). Measurements of pH and temperature were recorded electronically over the course of the experiment (Pasco Scientific-CI 6507 pH electrode amplifier, Pasco Scientific/Science Workshop 500 interface, Science Workshop 2.2.5 data acquisition program). Tritiated hydrogen gas, potentially produced by NaBH4 decomposition, was continuously removed from the gastight reaction vial by supplying a regulated flow of zero-grade air (CGA590) to the gas inlet (Masterflex tubing 06419-16, Matheson 600 tube cube flow regulator). The off-gas was lead from the reaction vessel through heat-resistant tubing (Masterflex platinum-cured silicone 3350 tubing, 96420-16) to a stainless steel column (1 ft long, 0.75 in. × 0.065 in. wall tubing, Denver Valve and Fitting Co.: SS-T12-S-065-20) for off-gas treatment. This column was packed with a platinum catalyst bed (platinum on 1/8 in. alumina pellets, 0.5% Pt, 25 g, ACROS Organics: 19530 0250) to promote the oxidation of tritiated H2 (g) to tritiated water vapor. To avoid condensation of tritiated water inside the tubing and catalyst bed, all relevant parts were held at an elevated temperature (g115 °C) by means of a flexible small diameter fiberglass VOL. 41, NO. 19, 2007 / ENVIRONMENTAL SCIENCE & TECHNOLOGY

9

6777

heating cord (Cole-Parmer: U-03122-6, Variac SE Powerstat Variable Transformer: 3PN116B). The tip of a type-K thermocouple (TE Omega) was packed inside the catalyst bed for manual monitoring and control of the bed temperature with a digital thermometer (Omega HH-25KC). The catalyst system was designed to handle the complete decomposition of NaBH4 into hydrogen gas over a short time frame (40 min). The air flow rate through the reaction vessel (58.8 mL min-1) was set accordingly to dilute hydrogen concentrations to sufficiently low levels for optimum oxidation conditions on the Pt catalyst (e2.5% H2 (g)). The removal of H2 (g) concentrations of up to 7.2% from the off-gas under the specified conditions was verified in a separate experiment prior to radiolabeling. For this purpose, various mixtures of zero-grade air and H2/N2 gas were injected into the Pt catalyst column at an elevated temperature (above 115 °C), while the off-gas was continuously monitored for H2 (g) with a TIF 8800 permissible gas detector (data not reported). Tritiated water vapor, produced in the catalyst column, was finally collected in a series of two ice-water cooled cooling traps filled with glass wool. The off-gas of the cooling trap was monitored for H2 (g) with a handheld detector (TIF 8800 permissible gas detector; detection limit: 500 ppm) and released into the radioisotope hood. 2.1.4. Stepwise Labeling Procedure and Sample Cleanup. In the following discussion, we describe the experimental procedure for the radiolabeling of 10 mg of FA. In addition to the preparation of hot FA, a batch of 110 mg of FA was treated with non-radioactive NaBH4 following the same procedure, with proportionally increased amounts of reagents. This produced cold-labeled FA for the detailed chemical product characterization under a minimum of safety restrictions. In the first step, 10 mg of FA was added to the reaction vial (1.95 mL of 5.13 g L-1 FA stock solution) through the gastight addition funnel. This funnel was carefully rinsed with known volumes of UV water after each addition of FA or NaBH4 to prevent early FA reduction and tritiated H2 (g) formation outside of the reaction vial. For optimum reduction efficiency, a mass of 10 mg (2.65 × 10-4 mol) of NaBH4 is needed (see Labeling Efficiency Experiments), which forced us to supplement the available 7.6 mg of radioactive NaBH4 with 2.4 mg of non-radioactive NaBH4. Both NaBH4 powders were dissolved in an alkaline solution (0.5 mL of 0.1 N NaOH) to prevent reagent decomposition and were transferred to the funnel. Earlier tests determined that this NaOH addition resulted in a final pH of approximately 9.6 in the reaction solution (data not reported). Additional rinsing of the NaBH4 vials with known amounts of UV water brought the total volume of the NaBH4 solution in the addition funnel to 4 mL. After a short mixing time, 5 µL of this NaBH4 solution was removed for later determination of the tritium counting efficiency of the liquid scintillation counter. The remaining NaBH4 solution was then transferred into the reaction vial. Final UV water rinses of the NaBH4 vials and the addition funnel (2.554 mL) brought the total volume of the reaction solution to 10 mL. After starting the monitoring of solution pH and temperature, the reaction was first allowed to proceed under stirring at room temperature for 30 min, then at an elevated temperature (reaction vessel in 60 °C water bath) for four additional hours. In the first NaBH4 cleanup step, a water slurry of about 920 mg of preconditioned cation exchange resin was added into the reaction vessel. The resin provides reactive sites for the uptake of Na+ originating from NaBH4 in exchange for H+ in approximately 10-fold molar excess. Sodium borohydride is transformed into HBH4, which further reacts to boric acid (B(OH)3) in the presence of water. Because of the proton release from the resin, a significant decrease of solution pH 6778

9

ENVIRONMENTAL SCIENCE & TECHNOLOGY / VOL. 41, NO. 19, 2007

is observed; the pH quickly drops to a value around 2.3, followed by a slow decrease to a pH value of approximately 2 over the following 2-3 h. For the controlled destabilization of unstable reaction products and the removal of free tritium ions in solution, zero-grade air was sparged directly through the reaction solution until the water in the reaction vessel was completely evaporated (water bath temperature at around 45 °C). Previous experimental testing based on the UV-vis absorbance of FA solutions determined that air sparging over a minimum of 24 h was required for the complete reoxidation of unstable reduction products (Figures S2 and S3, Supporting Information). It was assumed that after this step, the tritium remaining in the reaction vial is associated with FA in a stable, covalent bond. After drying, the FA-resin material was redissolved in UV water and removed from the reaction vessel. Then, the resin was separated from solution in a series of decanting steps, subsequent centrifugation (Baxter Biofuge 22 R, 30 min at 10 000 rpm), and two filtration steps (0.7 µm Ahlstrom glass microfiber filter). The remaining FA solution was dried in a rotary evaporator flask. To clean up the remaining boric acid in solution, 5 mL of methanol (Optima Grade, Fisher Chemical: A454-4) was added to the dry sample to form the volatile trimethyl borate (B(OCH3)3), which was then removed by rotoevaporation (Bu ¨ chi Rotovapor R). Since methanol can potentially methylate and permanently block relevant reactive groups of FA over time, it is necessary to minimize the contact time between FA and methanol to approximately 5 min. Our improved rotoevaporation setup included the following additional features: cooling water circulation through the condenser at 1-2 °C from a temperature regulated water bath; wrapping of the condenser with a cool, wet cloth; heating of the distillation tube with a heating gun before and during rotoevaporation; rotation of the sample flask at a high speed in the hot water bath at 80 °C; and partial submersion of the collection flask into ice-water. After rotoevaporation of B(OCH3)3 and excess methanol, autoclaved UV water was added to the sample flask immediately, and the radiolabeled FA was completely redissolved under shaking. For sample preservation, cold-labeled FA was freeze-dried (Labconco Freeze Dryer 4.5), and the radiolabeled FA product was frozen prior to storage. Both samples were kept in the dark to prevent photocatalytic reactions. In addition, the removal of dissolved oxygen from FA solutions by sparging is recommended to minimize the potential for reoxidation of the radiolabel. 2.2. Labeling Efficiency Experiments. As part of the optimization of the labeling procedure, the efficiency of the reduction reaction during labeling was determined experimentally as a function of sodium borohydride concentrations in solution. Because of the instability of the formed hydroquinones/phenols, there is a need to clarify the term reduction efficiency. Maximum reduction efficiency is achieved when all FA reactive groups available for carbonyl reduction by NaBH4 (ketones as well as quinones) are completely reduced. The specific activity of the labeled product, however, is only based on the labeling of ketones due to the long-term stability of their products. The setup of efficiency experiments is similar to that for the labeling reaction and is summarized briefly in the following discussion. In an oxygen-free glove box environment, a series of 10 mL solutions containing aliquots of 10 mg of freeze-dried FA were reduced with various amounts of non-radioactive NaBH4 (from approximately 1.25 to 15 mg) at 60 °C and a pH of 9.6 ( 0.06 over 4 h. In addition, FA standard solutions containing no NaBH4 were prepared in the same manner. After reaction completion, all solutions were scanned for

light absorbance on a Hach DR/4000 U spectrophotometer over a wavelength range from 200 to 800 nm. The data of UV-vis spectral analysis at 465 nm were used to determine the FA reduction efficiency as a function of NaBH4 concentration in solution, as the reduction of quinones is known to result in a decrease in the light absorbance and visual color of FA. It is expected that increasing NaBH4 concentrations in solution will result in a decrease of FA absorbance until a constant absorbance value is reached at the maximum reduction efficiency and the optimum chemical yield of the reaction. Furthermore, the data interpretation was based on the assumption that the number of transformed quinone groups is proportional to the total number of reduced FA carbonyl groups. Using this absorbance data, chemical models describing the reduction efficiency of FA with sodium borohydride were developed, which will be presented elsewhere. 2.3. Determination of Specific Activity of Radiolabeled FA. The tritium counting efficiency was determined during the labeling experiment. Five microliters of the 4 mL NaBH4 solution in the addition funnel (100 mCi tritiated NaBH4 with cold NaBH4 in 0.0125 N NaOH) was diluted with 0.001 N NaOH to give 5 nCi mL-1 tritium in solution. The solution was used to count a series of NaBH4 volumes in an Ultima Gold XR liquid scintillation cocktail on a PerkinElmer TR2500 liquid scintillation counter (LSC). For the determination of the specific activity of the radiolabeled FA product, fractions of a 1:50 dilution of the original radiolabeled FA stock solution were analyzed for TOC (Shimadzu TOC 5000, high-sensitivity catalyst) and for tritium activity (PerkinElmer TR2500 LSC). 2.4. Comparison of FTIR Spectra. The comparison of FTIR spectra of original, untreated FA stock with cold-labeled FA allowed for the investigation of potential changes in FA functional group composition due to labeling. In particular, we needed to ensure that FA carboxyl groups had not been methylated to methyl esters during the rotoevaporation step with methanol and that FA was not contaminated with methanol or boric acid impurities. Samples of original and cold-labeled FA were dried over desiccant for approximately 48 h. From each sample, approximately 2 mg was mixed with 100 mg of potassium bromide (Sigma-Aldrich, FT-IR grade, 221864-25g) and pressed to pellets (Carver hydraulic unit, model #3925). FTIR spectra of FA samples and a KBr blank were collected over the mid-IR region (4000-400 cm-1) with 200 scans at 2 cm-1 resolution (Nexus 670 FT-IR ESP). The instrument performed an automatic blank subtraction. 2.5. Proton NMR Spectral Analysis. Proton NMR analysis of cold-labeled FA was used to confirm the reduction of FA aliphatic and aromatic ketones to secondary alcohols by NaBH4 and allowed for the determination of potential methanol impurities in the reduced FA sample. Approximately 17 mg of untreated and cold-labeled FA was dissolved in 0.9 mL of deuterium oxide (Cambridge Isotope Products, 99.9%, DLM-4). Proton NMR spectra were generated on a Chemomagnetics CMX Infinity 400 Solids/Liquids NMR spectrometer operating at 400 MHz for 1H, with a pulse width of 9 µs and a pulse repetition time of 4.277 s. 2.6. FA Characterization by SEC. FA characterization by SEC allowed us to investigate two aspects of the label: (i) the distribution of the tritium label over the molecular size range of FA and (ii) label stability during short-term experimental conditions and long-term storage. Following recommendations from literature (25), samples were injected into a Universal Fractionator (Model F-1000 FFFractionation, Inc.) with a home-built injection loop (approximately 200 µL) connected to a TSK-50S SEC column (8 mm × 300 mm, 30 µm particles). A Na2HPO4 buffer was

used as a carrier solution (0.004 M Na2HPO4 in 0.088 M NaCl, ITot ) 0.1 M, pH ∼8.3) at 0.7 mL min-1 (Acuflow Series II pump). A 254 nm UV measuring cell (ISCO Type-II Optical Unit) coupled with a UV-vis detector (ISCO Model 229) provided online recording of UV absorbance (Flow 2003 software). Additionally, eluent fractions were collected for highly concentrated and tritiated FA samples for TOC analysis (Shimadzu TOC 5000, regular sensitivity catalyst) or tritium counting (PerkinElmer TR2500 LSC). Prior to sample characterization, the molecular size separation of the column was calibrated based on the injection of a series of polyethylene glycol (polyethylene oxide) molecular size standards from 400 to 12 600 Da (Figure S4, Supporting Information). Initially, three sample solutions were prepared in the buffer matrix for the comparison of their SEC profiles: tritiated FA at a concentration of approximately 10 mg L-1 TOC and two regular FA samples of approximately 10 and 200 mg L-1 TOC. After SEC sample characterization over 4 days, the solutions were split up into three volume fractions each and individually adjusted to specific pH values of 4, 7, or 10. Then, solutions were allowed to age on a shaker table at room temperature for approximately 10.5 days (253 h). Finally, the solution pH values after shaking were recorded and readjusted to the original values prior to adjustment, followed by a second sample characterization by SEC. In addition to short-term stability, the long-term behavior of the label was also tested based on SEC characterization of a tritiated FA dilution (12.4 µCi (0.46 MBq) mL-1; 6.56 mg L-1 FA, pH 6.65) stored at 4 °C for approximately 21 months. 2.7. Comparison of FA Sorption Behavior. The goal of batch sorption experiments was to investigate potential changes in FA sorption behavior due to labeling. Indirectly, the experimental setup also allowed us to test the chemical stability of the label over a wide pH range and in the presence of a mineral surface. Following the procedure by Lenhart and Honeyman (26), a total concentration of 10 mg L-1 FA was sorbed to 1 g L-1 hematite in a series of sample vials, each representing a different solution pH in a total volume of 25 mL. FA samples contained either untreated FA or cold-labeled FA combined with small volumes of tritium-labeled FA (10 µL of 12.4 µCi mL-1, equal to 2.6 µg L-1 tritiated FA). Therefore, the radiolabeled FA was used as radioactive tracer at a ratio of 1/3810. The ionic strength (I) for samples of pH < 6 was 0.01 N NaClO4/NaHCO3; for solutions at higher pH values, it was slightly higher (pH 7, I ) 0.011 N; pH 7.5, I ) 0.012 N; pH 8, I ) 0.017 N; and pH 9, I ) 0.029 N). This is different from the published procedure, where all sample solutions were kept at the same ionic strength. After equilibration at room temperature over approximately 50 h, pH measurements, and sample centrifugation (Baxter Biofuge 22R, 10 000 rpm for 159 min), fractions of the supernatant solutions were analyzed for FA in terms of TOC (Shimadzu TOC 5000, high sensitivity catalyst) or tritium activity (PerkinElmer TR2500 LSC). The fraction of FA sorbed to the mineral phase was calculated based on concentration differences, by comparing remaining FA concentrations in sample solutions with FA concentrations in standard vials containing no solid phase.

3. Results and Discussion 3.1. Labeling Efficiency Experiments. The light absorbance data at 465 nm show an exponential-type decrease in absorbance with increasing NaBH4 excess concentrations in solution. This leads to the prediction of an exponential-type increase in reduction efficiencies with increasing NaBH4 excess concentrations (Table S2 and Figure S5, Supporting Information). Assuming a linear relationship (eff ) -k × abs + d) between reduction efficiency (eff) and solution absorbance (abs) as well as efficiency values of zero and 100% for VOL. 41, NO. 19, 2007 / ENVIRONMENTAL SCIENCE & TECHNOLOGY

9

6779

FIGURE 1. Comparison of size exclusion chromatograms of various FA samples based on TOC analysis (original, regular FA) or tritium counting (all tritiated samples): FA concentrations of injected solutions were approximately 200 (original, regular FA) or 10 mg L-1 TOC (tritiated FA samples). Retention times were normalized to the time of measurement at the online UV measuring cell by subtracting a delay time of 1.35 min (based on tubing length and flow rate) for fractions collected at the sample collection point. Each chromatogram represents the average of three replicate injections.

FIGURE 2. Results for 10 mg L-1 FA sorption to 1 g L-1 hematite at I ) 0.01 N NaClO4/NaHCO3 from this study and by Lenhart (29). Experimental results are based on TOC measurements (untreated FA) or on liquid scintillation counting (tritium-labeled FA tracer). For this study, error bars represent approximate 95% confidence intervals based on the variability of two (counting) or four (TOC) solution samples from a single experiment. Analytical errors for individual samples were negligibly small in comparison. Lenhart (29) reported estimated analytical errors of e7 % for TOC analysis based on multiple sample injections. zero and the maximum NaBH4 concentration in solution, we determined that eff ) -285.75abs + 221.23. Furthermore, we can conclude that at least a 10-fold molar NaBH4 excess concentration, equivalent to approximately equal masses of FA and NaBH4, is required to ensure maximum reduction efficiency of FA-ketone groups. 3.2. Determination of Specific Activity of Radiolabeled FA. The calculated tritium counting efficiency under the specified conditions is 0.47 with an estimated total error of ( 0.02. The tritium activity of the radiolabeled FA stock solution is computed as 0.62 mCi mL-1 (23 MBq mL-1) ( 3.5%. On the basis of a fulvic acid TOC content of 53.49% (27), this gives a specific activity of 1.9 mCi mg-1 FA (70 MBq mg-1 FA) ( 4.9%. The resulting detection limit for FA 6780

9

ENVIRONMENTAL SCIENCE & TECHNOLOGY / VOL. 41, NO. 19, 2007

concentrations in solution is 0.31 µg L-1 FA (4.4 × 10-10 mol L-1 FA or 0.16 µg L-1 TOC). In the determination of FA specific activity, FA chemiluminescence effects during counting were neglected since only small volumes of low FA concentrations were added to the LSC cocktail and the number of total counts was very high. As luminescence effects may be dependent on specific solution conditions (e.g., pH, FA concentration), an individual characterization for specific applications is recommended. 3.3. Comparison of FTIR Spectra. The FTIR transmittance spectrum of untreated FA agrees well with literature data (28), while the spectrum of the cold-labeled FA exhibits a few additional features that can be attributed to the creation of aliphatic hydroxyl groups from the reduction of aliphatic

ketones (Figure S6, Supporting Information). The more pronounced absorption in the 3400 cm-1 region is caused by the OH groups of created aliphatic alcohols; the sharp bands from 1100 to 1050 cm-1 are the C-O stretches of these created alcohols. This confirms the reaction mechanism described in detail earlier. There is no evidence of the formation of methyl esters (peaks at 1740, 1439, and 1170 cm-1) or boric acid impurities (3200, 1194, and 548 cm-1), although a slight methanol contamination may be present at 1033 cm-1. 3.4. Proton NMR Spectral Analysis. A comparison of the 1 H NMR spectra of cold-labeled and untreated FA shows various features that confirm the reaction mechanism and the formation of secondary alcohols described earlier (Figure S7, Supporting Information). First, a significant increase in the peak region around 3.6 ppm for cold-labeled FA can be contributed to the formation of secondary alcohols. Second, a substantial decrease in the broad peak from 2.1 to 2.7 ppm (methyl groups of aliphatic and aromatic ketones) and an increase in the 1.25 ppm peak (saturated secondary hydrogen) can be observed for cold-labeled FA in comparison to untreated FA. Both characteristics have been described for the reduction of FA with NaBH4 previously (17). Further, the valley at 1.7 ppm in the reduced FA sample is less pronounced than for the original FA. The latter indicates that methylene hydrogen, adjacent to aromatic ketones, has been converted to methylene hydrogen, adjacent to aromatic carbinols (17). However, the spectrum of cold-labeled FA also shows an additional sharp peak at 3.3 ppm, which was identified as methanol contamination (NMR solvent data chart, Cambridge Isotope Laboratories, Inc.), probably due to the contact of FA with methanol during rotoevaporation. Based on the integration over the spectrum (Spinsight Software, Chemomagnetics), the methanol impurity contributes about 8% of the total number of protons in the cold-labeled FA sample. It is reasonable to assume that a similar level of methanol contamination can be expected in tritium-labeled FA. However, additional contributions to TOC due to methanol impurities can be neglected, as radiolabeled FA is expected to be used at trace level concentrations in future experiments. Furthermore, competing side reactions of methanol in FA metal complexation or sorption reactions can be ruled out. 3.5. FA Characterization by SEC. According to SEC results, the size distribution of radiolabeled FA determined by tritium counting follows FA chromatograms based on TOC and UV absorbance data sufficiently well (Figure 1, and Figures S8 and S9, Supporting Information). This indicates an even distribution of the tritium label over the FA molecular size range as well as the conservation of the original FA size distribution during the labeling process. However, an additional small peak at 19.1 min retention time was detected for tritiated FA, indicating a tritium-based impurity of approximately 3% (m/m) in the MW size range of 100 Da. This impurity concentration can be assumed to be negligible for most tracer applications of radiolabeled fulvic acid. During sample aging, the solution pH was fairly stable ((0.3) except for the pH 10 samples, which showed a pH decrease to approximately 9.4. SEC chromatograms of aged tritiated FA samples show very good agreement with the size distribution of original FA samples. A significant dissociation or exchange of the tritium label has not been detected, besides a potential small increase in the tritium-based impurity (maximum contribution of 5%). Furthermore, the size distribution of the 21 month old tritiated FA dilution follows the chromatogram of the tritiated FA stock very well, indicating long-term stability of the label during storage (Figure S10, Supporting Information). 3.6. Comparison of FA Sorption Behavior. The results for FA sorption to hematite based on tritium counting of radiolabeled FA are in good agreement with sorption data of untreated FA determined by TOC measurements (Figure

2). Possible small methanol impurities in the tritium-labeled FA tracer do not seem to affect FA sorption behavior. On the basis of this result, it can be further concluded that tritiumlabeled FA is chemically stable in solution and in the presence of an iron mineral phase over a pH range from about 3 to 9.

Acknowledgments We are indebted to Jerry Leenheer of the U.S. Geological Survey (Denver, CO) for invaluable discussions regarding the chemistry of natural organic matter. Further, we gratefully acknowledge the technical advice of Scott Cowley, Steven Dec, Andrew Herring, James Horan, Fangin Meng, and James Ranville (all of the Colorado School of Mines) as well as of Sungyun Lee (Gwangju Institute of Science and Technology). This research was supported by a grant from the U.S. Department of Energy’s Office of Science, Office of Biological and Environmental Research, Natural and Accelerated Bioremediation Research (NABIR) Program. Patent pending.

Supporting Information Available Overview of current NOM radiolabeling methods (Table S1), schematic of experimental setup for labeling procedure (Figure S1), results for stabilization of solution absorbance during aeration (Figures S2 and S3), SEC calibration curve (Figure S4), summary of results for labeling efficiency experiments (Table S2 and Figure S5), FTIR and 1H NMR spectral data (Figures S6 and S7), and additional SEC chromatograms based on FA TOC analysis and UV absorbance, as well as SEC data for FA long-term storage (Figures S8-S10). This material is available free of charge via the Internet at http://pubs.acs.org.

Literature Cited (1) Gaffney, J. S.; Marley, N. A.; Clark, S. B. Humic and fulvic acids and organic colloidal materials in the environment. In Humic and Fulvic Acids: Isolation, Structure, and Environmental Role; Gaffney, J. S., Marley, N. A., Clark, S. B., Eds.; American Chemical Society: Washington, DC, 1996; pp 2-16. (2) Stevenson, F. J. Reactive functional groups of humic substances. In Humus Chemistry: Genesis, Composition, Reactions; Wiley Interscience Publishers: New York, 1982; pp 221-243. (3) Leenheer, J. A.; Croue, J. P. Characterizing aquatic dissolved organic matter. Environ. Sci. Technol. 2003, 37, 18-26. (4) MacCarthy, P. The principles of humic substances. Soil Sci. 2001, 166, 738-751. (5) Thurman, E. M. Humic substances in groundwater. In Humic Substances in Soil, Sediment, and Water; Aiken, G. R., McKnight, D. M., Wershaw, R. L., MacCarthy, P., Eds.; John Wiley and Sons: New York, 1985; pp 87-103. (6) Northcott, G. L.; Jones, K. C. Experimental approaches and analytical techniques for determining organic compound bound residues in soil and sediment. Environ. Pollut. 2000, 108, 1943. (7) Novak, J. M.; Mills, G. L.; Bertsch, P. M. Estimating the percent aromatic carbon in soil and aquatic humic substances using ultraviolet absorbance spectrometry. J. Environ. Qual. 1992, 21, 144-147. (8) Dierckx, A.; Hall, A.; De Cannie`re, P.; Warwick, P.; Put, M. Stability of 125I and 14C labeled boom clay organic matter. Radiochim. Acta 1998, 82, 379-384. (9) Senesi, N. Molecular and quantitative aspects of the chemistry of fulvic acid and its interactions with metal ions and organic chemicals. Part II. The fluorescence spectroscopy approach. Anal. Chim. Acta 1990, 232, 77-106. (10) Evans, A. E. Tritium and Its Compounds; Van Nostrand Reinhold: New York, 1966. (11) The Radiochemical Center. The Radiochemical Manual, 2nd ed.; Amersham Radiochemical Centre: Amersham, U.K., 1966; pp 42-46. (12) McMurry, J. Organic Chemistry, 5th ed.; Brooks/Cole: Pacific Grove, CA, 2000; pp 768-769. (13) Thorn, K. A.; Pettigrew, P. J.; Goldenberg, W. S.; Weber, E. J. Covalent binding of aniline to humic substances. Part 2. 15N VOL. 41, NO. 19, 2007 / ENVIRONMENTAL SCIENCE & TECHNOLOGY

9

6781

(14) (15) (16) (17) (18) (19) (20)

(21) (22)

(23)

NMR studies of nucleophilic addition reactions. Environ. Sci. Technol. 1996, 30, 2764-2775. Weber, E. J.; Spidle, D. L.; Thorn, K. A. Covalent binding of aniline to humic substances. Part 1. Kinetic studies. Environ. Sci. Technol. 1996, 30, 2755-2763. Stumm, W.; Morgan, J. J. Aquatic Chemistry: Chemical Equilibria and Rates in Natural Waters, 3rd ed.; John Wiley and Sons, Inc.: New York, 1996. Pos, W. H.; Riemer, D. D.; Zika, R. G. Carbonyl sulfide (OCS) and carbon monoxide (CO) in natural waters: Evidence of a coupled production pathway. Mar. Chem. 1998, 62, 89-101. Leenheer, J. A.; Wilson, M. A.; Malcolm, R. L. Presence and potential significance of aromatic-ketone groups in aquatic humic substances. Org. Geochem. 1987, 11, 273-280. Rohm and Haas. The Sodium Borohydride Digest; http:// www.rohmhaas.com/hydride/technical.html#; accessed March 14, 2007. Gaylord, N. G. Reduction with Complex Hydrides; Wiley Interscience Publishers, Inc.: New York, 1956; Chs. 2, 3, 6, and 7. Wigfield, D. C.; Gowland, F. W. Limited alkoxy group exchange in tetraalkoxyborohydrides: Evidence against the four-center transition state in the borohydride reduction of ketones. Tetrahedron Lett. 1976, 38, 3373-3376. McMurry, J.; Castellion, M. E. Fundamentals of Organic and Biological Chemistry, 2nd ed.; Prentice Hall, Inc.: Upper Saddle River, NJ, 1999. Periasamy, M.; Thirumalaikumar, M. Methods of enhancement of reactivity and selectivity of sodium borohydride for applications in organic synthesis. J. Organomet. Chem. 2000, 609, 137151. Lenz, A.-G.; Costabel, U.; Shaltiel, S.; Levine, R. L. Determination of carbonyl groups in oxidatively modified proteins by

6782

9

ENVIRONMENTAL SCIENCE & TECHNOLOGY / VOL. 41, NO. 19, 2007

(24)

(25)

(26)

(27)

(28)

(29)

reduction with tritiated sodium borohydride. Anal. Biochem. 1989, 177, 419-425. Aiken, G. R.; McKnight, D. M.; Thorn, K. A.; Thurman, E. M. Isolation of hydrophilic organic acids from water using nonionic macroporous resins. Org. Geochem. 1992, 18, 567-573. Her, N.; Amy, G.; Foss, D.; Cho, J.; Yoon, Y.; Kosenka, P. Optimization of method for detecting and characterizing NOM by HPLC-SEC with UV and on-line DOC detection. Environ. Sci. Technol. 2002, 36, 1069-1076. Lenhart, J. J.; Honeyman, B. D. Uranium(VI) sorption to hematite in the presence of humic acid. Geochim. Cosmochim. Acta 1999, 63, 2891-2901. Averett, R. C.; Leenheer, J. A.; McKnight, D. M.; Thorn, K. A. Humic Substances in the Suwannee River, Georgia: Interactions, Properties, and Proposed Structures; Water-Supply Paper 2373; U.S. Geological Survey: Denver, CO, 1994. MacCarthy, P.; Rice, J. A. Spectroscopic methods (other than NMR) for determining functionality in humic substances. In Humic Substances in Soil, Sediment, and Water; Aiken, G. R., McKnight, D. M., Wershaw, R. L., MacCarthy, P., Eds.; John Wiley and Sons: New York, 1985; pp 527-559. Lenhart, J. J. The application of surface complexation modeling to the adsorption of uranium(VI) onto hematite in the presence of humic and fulvic acids. Ph.D. Thesis, Colorado School of Mines, 1997.

Received for review March 6, 2007. Revised manuscript received June 19, 2007. Accepted July 12, 2007. ES070563B