Environ. Sci. Technol. 2002, 36, 3797-3805
15
N NMR Investigation of the Reduction and Binding of TNT in an Aerobic Bench Scale Reactor Simulating Windrow Composting K. A. THORN* U.S. Geological Survey, P.O. Box 25046, MS 408, Denver Federal Center, Denver, Colorado 80225-0046 J. C. PENNINGTON Environmental Laboratory, U.S. Army Engineer Research and Development Center, 3909 Halls Ferry Road, Vicksburg, Mississippi 39180-6199 C. A. HAYES DynTel, 3530 Manor Dr., Suite 4, Vicksburg, Mississippi 39180
T15NT was added to a soil of low organic carbon content and composted for 20 days in an aerobic bench scale reactor. The finished whole compost and fulvic acid, humic acid, humin, and lignocellulose fractions extracted from the compost were analyzed by solid-state CP/MAS and DP/ MAS 15N NMR. 15N NMR spectra provided direct spectroscopic evidence for reduction of TNT followed by covalent binding of the reduced metabolites to organic matter of the composted soil, with the majority of metabolite found in the lignocellulose fraction, by mass also the major fraction of the compost. In general, the types of bonds formed between soil organic matter and reduced TNT amines in controlled laboratory reactions were observed in the spectra of the whole compost and fractions, confirming that during composting TNT is reduced to amines that form covalent bonds with organic matter through aminohydroquinone, aminoquinone, heterocyclic, and imine linkages, among others. Concentrations of imine nitrogens in the compost spectra suggest that covalent binding by the diamines 2,4DANT and 2,6DANT is a significant process in the transformation of TNT into bound residues. Liquid-phase 15N NMR spectra of the fulvic acid and humin fractions provided possible evidence for involvement of phenoloxidase enzymes in covalent bond formation.
Introduction In the companion paper to this study, the monoamine (2ADNT and 4ADNT), diamine (2,4DANT and 2,6DANT), and triamine (TAT) reductive degradation products of TNT, labeled with 15N in the amine positions, were shown to undergo covalent binding with soil humic acid, model lignin and quinone compounds, and lignocellulose (sawdust) in the presence and absence of horseradish peroxidase and birnessite (1). In this paper, we report 15N NMR analyses on a compost produced in a bench scale aerobic reactor from a low organic carbon soil spiked with T15NT. The aerobic conditions and compost amendments were designed to * Corresponding author phone: (303)236-3979; fax: (303)236-3934; e-mail:
[email protected]. 10.1021/es011382r CCC: $22.00 Published on Web 08/06/2002
2002 American Chemical Society
simulate large scale windrow composting. The objective of the study was to observe in situ both the reduction of the labeled nitro groups and their subsequent condensation with organic matter in the compost and confirm whether the same types of bonds formed between the individual amines and soil organic matter are reproduced in the compost. In contrast to laboratory experiments in which the individual amines are reacted with dissolved humic substances or model compounds under controlled conditions, the composting environment encompasses a potentially more complicated set of reactions. Specifically, the transitory nitroso and hydroxylamino intermediates, formed as the TNT is reduced to the amines, are potentially reactive. The nitroso group can act as an electrophile and the hydroxylamino group as a nucleophile. Nitroso groups may be subject to nucleophilic attack by naturally occurring amines and thiols (2). The trace quantities of azoxytoluenes routinely detected at contamination sites and in laboratory microcosm experiments result from condensation of the hydroxylamino intermediates with nitroso intermediates of TNT. Condensation reactions between the nitroso or hydroxylamino intermediates of TNT and functional groups present in the larger molecules comprising the naturally occurring organic matter of the compost are possible. Daun et al. (3) reported irreversible binding of 2- and 4-hydroxylaminodinitrotoluene to soil humic acid. Hydroxylamine itself readily condenses with carbonyl groups in humic substances (4); however, the hydroxylaminonitrotoluenes are secondary amines, and so the number of potential condensation reactions between hydroxylaminonitrotoluenes and carbonyl groups in soil organic matter is restricted. The analysis of a whole compost by 15N NMR poses analytical challenges. As evident from a comparison of liquidversus solid-state 15N NMR analyses of humic acid reacted with the amines (1), critical information is lost in going from liquid- to solid-state spectra because of the inherently inferior resolution in the solid state. Also, determination of the quantitative distribution of nitrogens incorporated into the organic matter of the whole compost becomes problematic, because, even with optimal contact times, the CP/MAS experiment can underestimate nitrogens not directly bonded to protons, notably heterocyclic and imine nitrogens. The workaround to the quantitation problem in CP/MAS, use of the DP/MAS (direct polarization; i.e., single pulse or Bloch decay) experiment, requires long acquisition times. The approach of extracting humic, fulvic, and other soluble fractions from the compost for liquid- or solid-state NMR analyses also creates experimental difficulties. The acid and base conditions employed in the extraction can potentially cause hydrolysis of some of the bonds formed between the reduced TNT amines and organic matter. Amines bonded to organic matter through imine linkages are susceptible to acid and base hydrolysis (5-7). The basic conditions of the extraction could also effect reactions that would not have occurred in the natural composting environment. For example, acid and base treatment of the compost could solubilize organic matter and make available for reaction functional groups that were previously inaccessible. Unless carried out under N2, extraction with base may cause oxidative and free radical coupling reactions (8, 9). Under strongly basic conditions, hydroxide ion can displace the nitro or methyl groups on mono- or diamines covalently bonded to organic matter, with the consequence of further rearrangement or recondensation reactions (10, 11). At this time there is no completely satisfactory solution between the problems of resolution and quantitation in solid-state NMR analyses VOL. 36, NO. 17, 2002 / ENVIRONMENTAL SCIENCE & TECHNOLOGY
9
3797
TABLE 1. Characteristics of Grange Hall Soil soil property
value
classification pH organic carbon, % cationic exchange capacity, mequiv/100 g sand, % silt, % clay, %
C 37.32
Tunica; clayey over loamy, montmorillonitic, nonacid, thermic Vertic Haplaquept 8.2 0.29 16.7 39 51 10
Elemental Analyses of Finished T15NT Compost (Moisture Free Basis; Percent of Total) H O N S P ash
total
4.77
98.0
30.3
1.75
0.21
0.28
23.37
and chemical alteration of covalent bonds between the reduced TNT amines and organic matter during extraction of fulvic and humic acids. In this study, both the whole compost and the fulvic acid (FA), humic acid (HA), humin (HN), and lignocellulose fractions extracted from the compost are analyzed by a combination of CP/MAS and DP/MAS solidstate NMR. An extraction procedure that should minimize chemical alteration reactions was used. The FA and humin fractions are also analyzed by quantitative liquid-phase 15N NMR. An interpretation of the spectra is provided based upon our knowledge of how the individual reduced TNT amines react with soil organic matter and model compounds and with the above limitations as caveat.
Materials and Methods Reagents. The 2,4,6-trinitrotoluene-15N3, 99 atom % 15N, was purchased from ISOTEC. (Use of trade names in this report is for identification purposes only and does not constitute endorsement by the U.S. Geological Survey or U.S. Army.) Characteristics of the Grange Hall soil, collected near Vicksburg, MS, are listed in Table 1. Compost Preparation and Incubation. Composting experiments were described in Pennington et al. (12). The Grange Hall soil was prepared by air-drying and sieving (Number 10 sieve, 1.65 mm) to remove coarse gravel. T15NT was added as a methanol solution to a 1:4 soil to water slurry at a concentration of 200 mg T15NT per gram of soil. The slurry was added to a compost mixture and comprised 10% of the compost wet weight. The compost mixture consisted of 33% green cow manure, 22% alfalfa, 6% chopped apples, 22% sawdust, and 17% chopped potatoes. All components of the compost mixture were of particle size 10 mm or less and added as percent of wet weight. The finer particles were added first, with successively larger particle-sized amendments added as mixing continued. The final mixture was moistened to 75 wt %. The combined soil slurry and compost mixtures were added to reactors that consisted of widemouthed glass canning jars (473 cm3) with modified lids (13). The chambers were incubated in a water bath maintained at 55 °C to match the temperatures reached in large scale composting that are conducive for the biotransformation of TNT by thermophilic microorganisms (14). Airflow was from the bottom at 10 mL/min. Temperature was monitored automatically by thermocouples in the center of the compost. Compost was incubated for 20 days. The composting experiments consisted of eight separate treatments incubated in eight separate reactors. In addition to two treatments containing the 15N-labeled TNT, four treatments employed a combination of 14C-labeled and unlabeled (cold) TNT, one employed unlabeled TNT only, and one control contained no TNT. 3798
9
ENVIRONMENTAL SCIENCE & TECHNOLOGY / VOL. 36, NO. 17, 2002
Portions of the finished composts were analyzed for solvent extractable TNT metabolites prior to NMR analyses. The finished composts were extracted overnight in methanol with sonication. The methanol extracts were analyzed by HPLC according to EPA SW-846 Method 8330 (15). Analytes assayed included TNT, 1,3,5-trinitrobenzene (TNB), 1,3-dinitrobenzene (DNB), 4ADNT, 2ADNT, 2,6-dinitrotoluene (2,6DNT), 2,4-dinitrotoluene (2,4DNT), nitrobenzene (NB), 2-nitrotoluene (2NT), 3-nitrotoluene (3NT), 2,6DANT, 2,4DANT, 2,2′,6,6′-tetranitro-4,4′-azoxytoluene (4,4′AZOXY). The following compounds were not detected in the assays: DNB; 2,6DNT; 2,4DNT; NB; 2NT; and 3NT. Preparation of Finished Compost for NMR. The T15NT and control compost samples were dried in a desiccator and then ground in a mortar and pestle to achieve as homogeneous a particle size possible without performing an actual physical size or density fractionation. Solid-state 15N NMR spectra were recorded on the samples. The T15NT compost was then extracted with 100% acetonitrile. Approximately 0.360 g of the compost was sonicated in 100 mL of acetonitrile for 8 h and then filtered through a 40-60 micron fritted glass filter. The T15NT compost was air-dried and desiccated, and the solid-state 15N spectrum was re-recorded. Fractionation of Finished Compost. After solid-state NMR analysis, the two T15NT composts were combined, solvent extracted with acetonitrile, and fractionated into humic acid, humin, fulvic acid, and cellulose, as described by Pennington (13). The cellulose fraction is referred to as the lignocellulose fraction in this paper. The fractionation procedure is a modification of the MIBK extraction of Rice and MacCarthy (16). The operationally defined humin from this fractionation is soluble in DMSO. NMR Spectrometry. Solid-state NMR spectra were recorded on a Chemagnetics CMX-200 NMR spectrometer as described (1). Acquisition parameters for solid-state CP/MAS 15N NMR experiments included a 30 000 Hz spectral window, 17.051-ms acquisition time, 0.5-s pulse delay, 5.0-ms contact time (2.0-ms for unlabeled control compost), and spinning rate of 5 kHz. For DP/MAS experiments, a 90° pulse angle in conjunction with a 5-s pulse delay were employed. Baseline distortions in the DP/MAS spectra were alleviated using backward linear prediction of 5 to 6 points. Solid-state CP/MAS 13C NMR spectra of the compost fractions were acquired using a 30 000 Hz spectral window, 17.051-ms acquisition time, 5.0-ms contact time (2-ms for lignocellulose), 1.0-s pulse delay, and spinning rate of 5 kHz. Liquidphase ACOUSTIC 15N NMR spectra of the fulvic acid and humin fractions dissolved in DMSO-d6 were recorded as described previously, using a 0.5-s pulse delay (1).
Results HPLC Analyses of Composts. Concentrations of solvent extractable TNT and metabolites from the initial and finished 14C, cold, and control composts that were run concurrently with the 15N composts are listed in Table 2. For the 14C treatments, the decreases in solvent extractable TNT range from 90.64% to 95.38% after 20 days (Table 2). The predominance in concentrations of solvent extractable 4ADNT over 2ADNT and 2,4DANT over 2,6DANT after 20 days, reflecting the regioselective reduction of nitro groups, is consistent with literature reports on aerobic composting and reduction (17, 18). The recoveries of 14C radioactivity for the 20-day {14C}-TNT composted soil (combination of four treatments) after fractionation are listed in Table 3. Most of label (56.24%) resides in the lignocellulose fraction. The amount of 14CO2 measured was below background. The total recovery of 69.5% for the 14C label in the compost appears to be a low outlier for this particular run of experiments, as recoveries in previous runs have been g 95% (13).
TABLE 2. Concentrations of TNT and Its Transformation Productsa treatment
T0 TNTb
T20 TNT
T20 TNB
T20 4ADNT
T20 2ADNT
T20 2,6DANT
T20 2,4DANT
T20 4,4′AZOXY
cold + 14C cold + 14C cold + 14C cold + 14C cold control
21 300 32 900 34 900 33 300 32 000
