Polymerization of 2,4,6-Trinitrotoluene

Alkaline hydrolysis has been investigated as a nonbiological procedure for the destruction of 2,4,6-trinitrotoluene (TNT) in explosives contaminated s...
0 downloads 0 Views 282KB Size
Download PDF
Environ. Sci. Technol. 2004, 38, 2224-2231

Alkaline Hydrolysis/Polymerization of 2,4,6-Trinitrotoluene: Characterization of Products by 13C and 15N NMR K E V I N A . T H O R N , * ,† PHILIP G. THORNE,‡ AND LARRY G. COX† U.S. Geological Survey, P.O. Box 25046, Denver Federal Center, Denver, Colorado 80225-0046, and Applied Research Associates, 415 Waterman Road, South Royalton, Vermont 05068

Alkaline hydrolysis has been investigated as a nonbiological procedure for the destruction of 2,4,6-trinitrotoluene (TNT) in explosives contaminated soils and munitions scrap. Nucleophilic substitutions of the nitro and methyl groups of TNT by hydroxide ion are the initial steps in the alkaline degradation of TNT. Potential applications of the technique include both in situ surface liming and ex situ alkaline treatment of contaminated soils. A number of laboratory studies have reported the formation of an uncharacterized polymeric material upon prolonged treatment of TNT in base. As part of an overall assessment of alkaline hydrolysis as a remediation technique, and to gain a better understanding of the chemical reactions underlying the hydrolysis/polymerization process, the soluble and precipitate fractions of polymeric material produced from the calcium hydroxide hydrolysis of unlabeled and 15N-labeled TNT were analyzed by elemental analysis and 13C and 15N nuclear magnetic resonance spectroscopy. Spectra indicated that reactions leading to polymerization included nucleophilic displacement of nitro groups by hydroxide ion, formation of ketone, carboxyl, alcohol, ether, and other aliphatic carbons, conversion of methyl groups to diphenyl methylene carbons, and recondensation of aromatic amines and reduced forms of nitrite, including ammonia and possibly hydroxylamine, into the polymer. Compared to the distribution of carbons in TNT as 14% sp3- and 86% sp2-hybridized, the precipitate fraction from hydrolysis of unlabeled TNT contained 33% sp3- and 67% sp2-hybridized carbons. The concentration of nitrogen in the precipitate was 64% of that in TNT. The 15N NMR spectra showed that, in addition to residual nitro groups, forms of nitrogen present in the filtrate and precipitate fractions include aminohydroquinone, primary amide, indole, imine, and azoxy, among others. Unreacted nitrite was recovered in the filtrate fraction. The toxicities and susceptibilities to microbial or chemical degradation of the polymeric materials remain unknown.

Introduction Contamination of soils by the explosive 2,4,6-trinitrotoluene (TNT) on firing ranges, in open burning/open detonation * Corresponding author phone: (303) 236-3979; fax: (303) 2363934; e-mail: [email protected]. † U.S. Geological Survey. ‡ Applied Research Associates. 2224

9

ENVIRONMENTAL SCIENCE & TECHNOLOGY / VOL. 38, NO. 7, 2004

areas, and in washout lagoon soils on former manufacturing plants and military bases is a worldwide problem that has been documented in numerous reviews (1, 2). Concentrations of TNT in contaminated soils have been reported to range from 10 to 12 000 mg kg-1 (1). In response to the need for cleanup of TNT-contaminated soils, a variety of biological and nonbiological remediation techniques have been investigated. Biological technologies have included soil piles, bioslurries, and windrow composting, the latter having been studied in detail and implemented at numerous contaminated sites within the U.S. (3). During composting, TNT undergoes microbial reduction via nitroso and hydroxylamino intermediates to the monoamines 2ADNT (2-amino4,6-dinitrotoluene) and 4ADNT (4-amino-2,6-dinitrotoluene), the diamines 2,4DANT (2,4-diamino-6-nitrotoluene) and 2,6DANT (2,6-diamino-4-nitrotoluene), and, under strictly anaerobic conditions, the triamine TAT (2,4,6-triaminotoluene) (4, 5). The aromatic amines subsequently form covalent bonds with the organic matter of the composted soil and become nontoxic to organisms (4-7). Once sequestered into the organic matter via covalent bonding, the amines are prevented from leaching into groundwater and surface water. Alkaline hydrolysis is one nonbiological procedure that has been investigated in several laboratory studies for the direct chemical destruction of TNT in explosives-contaminated soils (8-14) and munitions scrap (10, 15). The susceptibility of the nitro group of nitroaromatic compounds to nucleophilic substitution by hydroxide ion has been well documented (16, 17) and is the basis for these investigations. The literature on reactions of nitroaromatic compounds and TNT in basic solution dates back to 1889 (15, 18, 19). Alkaline hydrolysis of RDX and HMX, high explosives typically occurring along with TNT in contaminated soils, has also been investigated (20). The disappearance kinetics and percent destruction of TNT in alkaline solution and of TNT in soil matrixes subjected to alkaline treatment have been measured over a range of concentrations, temperatures, and pH (8-11, 21). Emmrich (9), for example, reported that up to two nitro groups per molecule of TNT were released during alkaline hydrolysis of TNT at pH 10-12 and 20 °C. In soils contaminated at a concentration of 116 mg of TNT/kg, 99% destruction of the TNT was achieved upon alkaline treatment. Saupe et al. (8) reported essentially complete elimination of TNT (∼0.15 M solution) by 4 h of alkaline hydrolysis at 80 °C and an initial pH of 14. Greater than 99.8% reduction of TNT was achieved by alkaline hydrolysis at 80 °C of a soil contaminated with 580 mg of TNT/kg. The occurrence of polymerization reactions upon prolonged alkaline hydrolysis of TNT has also been documented (8, 10, 13). From ultrafiltration measurements of polymers formed from alkaline hydrolysis of TNT at 80 °C, Saupe (8) reported that 40% of the DOC fell within the molecular mass range of 3-30 kDa while 60% had molecular masses greater than 30 kDa. Felt et al. (13) reported that 98% of the products from potassium hydroxide hydrolysis of 100 ppm TNT at pH 13 and 25 °C were polar compounds that varied in molecular mass from 3000-6000 to 99.8%) and 100 mg of CaO were added to 200 mL of reagent-grade water. The flask was capped and stirred for three weeks. The suspended material was then allowed to settle out for 3 days and the solution decanted, and then the precipitates were washed four times with reagent-grade water and once with acetonitrile and air-dried. A portion of the dried precipitate was subsequently treated with 1 N HCl and rinsed with deinonized and distilled water to remove the calcium carbonate. Elemental analyses of the precipitates were performed by Huffman Laboratories, Inc., Golden, CO. Methylation of Unlabeled TNT Precipitate. A phasetransfer methylation procedure employing tetrabutylammonium hydroxide (TBAH) and iodomethane (25) was performed on the precipitate from the unlabeled TNT hydrolysis. In a 25 mL Erlenmeyer under nitrogen, 250 mg of the precipitate was dissolved into 1.5 mL of TBAH solution, and the resulting solution was charged with 3 mL of tetrahydrofuran and 200 µL of iodomethane-13C and stirred for 24 h. The THF was then evaporated off, the solution neutralized with 5 mL of 1 N HCl, and the precipitate filtered and washed with water. The precipitate was then dialyzed in a 100 MWCO tube and freeze-dried. A portion of the methylated precipitate was dissolved in dimethyl-12C2,d6 sulfoxide for 13C NMR analysis. Hydrolysis/Polymerization of TNT-15N3. In a 1 L Teflon centrifuge bottle, 0.3750 g of TNT-15N3 was added to a solution of 0.375 g of CaO dissolved in 600 mL of water (pH 11.6). The reaction solution was stirred for 13 days at room temperature and open to the atmosphere and then filtered through a 0.45 µM cellulose acetate filter. The precipitate was washed with water, air-dried, and then desiccated. The filtrate (final pH 6.59) was freeze-dried. A portion of the precipitate fraction was also de-ashed by treatment with 1 N HCl as above. NMR Spectrometry. Solid-state 13C and 15N NMR spectra were recorded on a Chemagnetics CMX-200 NMR spectrometer at carbon and nitrogen resonant frequencies of 50.3 and 20.3 MHz, respectively, using a 7.5 mm ceramic probe (zirconium pencil rotors). Acquisition parameters for the cross polarization/magic angle spinning (CP/MAS) 15N spectrum of the TNT-15N3 precipitate included a 23094.7 Hz spectral window, 17.051 ms acquisition time, 2.0 ms contact time, 0.5 s pulse delay, and spinning rate of 6 kHz. Parameters for the

filtrate spectrum were similar but with a 26666.7 Hz spectral window and 5 kHz spinning rate. Parameters for CP/MAS 13 C spectra included a 30 030 Hz spectral window, 17.051 ms acquisition time, 5.0 ms contact time, 1.0 s pulse delay, and 5 kHz spinning rate. The direct polarization/magic angle spinning (DP/MAS) 13C spectrum of the precipitate from unlabeled TNT was acquired with a 90° pulse angle and 7.0 s pulse delay. The 7.0 s pulse delay was determined to be adequate for complete relaxation of all nuclei, as no changes in signal intensities were observed at a longer pulse delay of 10.0 s. Nitrogen-15 chemical shifts were referenced to glycine, taken as 32.6 ppm, and 13C shifts to hexamethylbenzene. Liquid-phase 13C and 15N NMR spectra were recorded on a Varian 300 MHz NMR spectrometer at carbon and nitrogen resonant frequencies of 75.4 and 30.4 MHz, respectively, using a 10 mm broad-band probe. Acquisition parameters for the continuous decoupled 13C NMR spectra of the de-ashed and of the methylated precipitate from the unlabeled TNT (dissolved in dimethyl-12C2,d6 sulfoxide) included a 50 000 Hz spectral window, 45° pulse angle, 0.2 s acquisition time, and 1.0 s pulse delay. ACOUSTIC and DEPT 15N NMR spectra were recorded on the de-ashed TNT-15N3 precipitate fraction dissolved in dimethyl-12C2,d6 sulfoxide. Acquisition parameters for the ACOUSTIC sequence (26) included a 24325.1 Hz spectral window, 0.2 s acquisition time, 1.0 s pulse delay, and τ delay of 0.1 ms; the spectrum was acquired with the addition of Cr(Acac)3 as paramagnetic relaxation reagent. Acquisition parameters for the DEPT spectra included a 24 325.1 Hz spectral window, 0.2 s acquisition time, 0.5 s delay for proton relaxation, and 1JNH of 90 Hz. Neat formamide in a 5 mm NMR tube, assumed to be 112.4 ppm, was used as an external reference standard for all spectra. The 15N NMR chemical shifts are reported in parts per million downfield of ammonia, taken as 0.0 ppm. Quantitation in NMR. In solid-state CP/MAS experiments, peak areas can accurately represent the number of nuclei resonating, when the time constant for cross polarization is significantly less than the time constant for proton spinlattice relaxation in the rotating frame, TCH or TNH , Τ1FΗ. Since no analyses of the spin dynamics were performed, the solid-state CP/MAS 13C and 15N NMR spectra can be interpreted only semiquantitatively. The DP/MAS 13C NMR spectrum of the precipitate fraction represents the quantitative distribution of carbons in the sample. The liquid-state continuous decoupled 13C NMR spectra are also nonquantitative, as differential T1 and NOE effects have not been eliminated. The liquid-state ACOUSTIC and inverse gated decoupled 15N NMR spectra recorded with relaxation reagent represent the quantitative distribution of nitrogens in the samples.

Results Background. Initial reactions of TNT in base consist of nucleophilic substitution of nitro groups by hydroxide ion, nucleophilic addition at the C3 and C5 carbons by hydroxide ion, nucleophilic substitution of the methyl group (demethylation) by hydroxide ion to form picric acid, and abstraction of a proton from the methyl group to form the benzylic carbanion (8, 11, 15, 18, 19). Some of the resultant monoand dihydroxy-substituted transformation products, including catechol and hydroquinone derivatives, are shown in Figure 1. The formation of σ-Meisenheimer complexes as intermediates in the formation of the hydroxy-substituted transformation products, as well other general reactions of nitroarenes with bases, including formation of donoracceptor complexes and radical anions, has been reviewed by Buncel (27). Compounds detected during the hydrolysis of TNT include nitrite, nitrate, ammonia, carbonate, acetate, and formate (8-10). Trace quantities of 2ADNT and 4ADNT were reported by Karasch (11). Saupe (8) also reported trace VOL. 38, NO. 7, 2004 / ENVIRONMENTAL SCIENCE & TECHNOLOGY

9

2225

TABLE 1. Elemental Analysis of Precipitate from Alkaline Hydrolysis of Unlabeled TNTa TNT precipitate

FIGURE 1. Structures from initial attack of hydroxide ion on TNT. quantities of nitrophenols, dinitrobenzenes (e.g., 2-methyl1,3-dinitrobenzene), nitroanilines (e.g., 4-methyl-2-nitroaniline), and anilines. Thermal treatment of the hydrolysates resulted in the production of N2, N2O, and NH3 gases (8). Formation of nitrate and ammonia from the initially released nitrite ion indicates the occurrence of both oxidative and reductive transformation pathways. Presumably, hydroxylamine is an intermediate in the reduction of nitrite to ammonia. Once the initial products such as catechols, hydroquinones, ammonia, hydroxylamine, aminodinitrotoluenes, and other aromatic amines have formed, a complex series of recondensation and free radical coupling reactions are possible. Nitrite itself has the potential for reincorporation into the developing polymeric material via nitrosation reactions. Although nitrosation reactions are unlikely to be significant above a pH of about 8, as noted in the Experimental Section, the pH of the filtrate had dropped to 6.6 by the end of the TNT-15N3 hydrolysis experiment. The pH conditions may therefore be amenable to nitrosation reactions toward the later stages of the hydrolyses. Some of the reactions of ammonia, aromatic amines, hydroxylamine, and nitrite with functional groups are summarized in Figure 2. In our experiments, and as reported in other studies, as the TNT dissolved into the basic solution, a red-pink color developed, followed by gradual formation of a brown-colored precipitate. The precipitate formed in our experiments was essentially

Ca ash total C carbonate C organic C H N O H/C O/C C/N

18.3 38.45 29.15 3.75 25.4 2.63 6.47 34.92

TNT precipitate (HCl-treated)

TNT std (calcd)

NDb 0.68 49.03 0.20 48.83 3.61 11.92 35.78 0.0739 0.7327 4.0965

37.00 2.20 18.50 42.29 0.0595 1.1430 2.000

a Elemental analyses reported on moisture-free basis. determined.

9

ENVIRONMENTAL SCIENCE & TECHNOLOGY / VOL. 38, NO. 7, 2004

ND ) not

insoluble in organic solvents, and could not be redissolved in acidic solution. The precipitate could be redissolved in tetrabutylammonium hydroxide solution, as noted above. Upon treatment with 1 N hydrochloric acid to remove calcium carbonate and bound calcium, the precipitate became soluble in DMSO. In contrast to TNT, the dinitrotoluenes 2,4DNT and 2,6DNT did not form polymerization products in our laboratory experiments using similar reaction conditions of CaO hydrolysis. This is in agreement with results of Bunte et al. (28), who noted only limited release of nitrite from 2,4DNT under pressurized alkaline hydrolysis. Elemental Analysis. Elemental analysis was performed on the precipitate fraction from hydrolysis of the unlabeled TNT, before and after treatment with dilute HCl. The data are compared to the calculated values for pure TNT (Table 1). The de-ashed precipitate contains 11.9% nitrogen, compared to 18.5% in the pure standard. Thus, the concentration of nitrogen in the precipitate is approximately 64% of that in the TNT standard. The O/C ratio decreases

FIGURE 2. Reactions of aromatic amines, ammonia, hydroxylamine, and nitrite with organic functional groups. 2226

b

complex is formed from reaction of the TNT carbanion with a second molecule of TNT (27):

FIGURE 3. Solid-state CP/MAS 13C NMR spectra of unlabeled TNT and precipitate fraction from calcium hydroxide hydrolysis/ polymerization of unlabeled TNT. ct ) contact time. LB ) line broadening. from 1.1430 in the standard to 0.7327 in the precipitate, consistent with the release of nitro groups from the TNT. The H/C ratio increases from 0.0595 in the standard to 0.0739 in the precipitate, indicating formation of aliphatic material during the hydrolysis/polymerization process, confirmed in the 13C NMR spectra. 13C NMR Spectra. The solid-state CP/MAS 13C NMR spectrum of the precipitate from the unlabeled TNT is compared to that of the TNT standard in Figure 3. Peaks in the spectrum of the TNT standard are assigned as C1, 135 ppm; C2&6, 153 ppm; C4 , 148 ppm; C3&5, 123 ppm; and methyl, 17 ppm. The broad bands in the spectrum of the precipitate suggest a multicomponent mixture of significant complexity. The spectrum indicates that formation of aliphatic and carbonyl carbons is a significant process during the hydrolysis/polymerization of TNT. Distinct peaks in the aliphatic region occur at 13, 30, and 42 ppm. The peak at 13 ppm corresponds to the residual methyl carbons in the polymerized TNT molecules. The peak at 30 ppm would correspond to polymethylene carbons; the formation pathway leading to these carbons is unclear. The peak at 42 ppm likely corresponds to diphenylmethylene carbons:

These structures may derive by elimination of a substituent from a Janovsky-type σ complex. The Janovsky

Nitrodibenzyl derivatives (e.g., 2,2′,4,4′,6,6′-hexanitrobibenzyl) have been postulated to result from reaction of TNT with base (18). The methylene carbons of these structures would occur at about 38 ppm. The broad peak from approximately 60 to 90 ppm corresponds to alcohol and ether carbons. The aromatic region exhibits distinct peaks at 122 and 148 ppm, derived from the C3&5 and C4 carbons of the TNT, respectively. The broad region from approximately 135 to 165 ppm also includes phenolic and arylamine carbons. The peak at 169.0 ppm corresponds to carbonate. It overlaps with other carbonyl carbons from 160 to 190 ppm, which include carboxyl and possibly quinone carbons. The peak at 213 ppm corresponds to ketone carbons. Peak areas as percent of total carbon for the quantitative DP/MAS spectrum of the precipitate (Figure A, Supporting Information) are listed in Table 2. Considering the DP/MAS spectrum from 0 to 90 ppm and from 90 to 230 ppm to represent sp3and sp2-hybridized carbons, respectively, sp3-hybridized carbons comprise 33% of the total carbons in the precipitate, compared to 14% in the TNT standard. The liquid-phase 13C NMR spectrum (nonquantitative continuous decoupled) of the acid-treated (de-ashed) precipitate affords additional details of structure, including a more clearly resolved view of the carbonyl region (Figure 4). Approximate chemical shift ranges in DMSO are 168-175 ppm for carboxylic acids, 178-188 ppm for quinones, and 176-210 ppm for ketones. The broad peak of low intensity from approximately 180 to 220 ppm in the precipitate spectrum encompasses diaryl, alkyl aryl, and dialkyl ketones, in addition to quinones. The distinct peak at 206.3 ppm is likely attributable to a dialkyl ketone structure. Other resolved peaks occur at 172.0, 166.7, 157.6, 147.1, 131.4, 113.6, and 109.9 ppm. The peak maximum of the alcohol and ether region occurs at 85.0 ppm. DEPT 13C NMR spectra (not shown) indicated that carbons downfield of 138.7 ppm are nonprotonated. Methylation analysis confirms the presence of several functional groups in the precipitate material. Peaks resulting from oxygen, nitrogen, and carbon methylation are present in the liquid-phase 13C NMR spectrum of the 13C-methylated precipitate in Figure 5. Most significant are the methyl ester peaks at 52.3 ppm, resulting from methylation of carboxylic acids, and the methyl ether peaks at 56.9, 62.2, and 67.0 ppm, resulting from methylation of phenolic and alcoholic hydroxyl groups (29). Methyl ethers of phenolic hydroxyls adjacent to two substituents on an aromatic ring give rise to the peak at 62.2 ppm; methyl ethers of phenolic hydroxyls adjacent to

TABLE 2. Peak Areas as Percent of Total Carbon for Solid State DP/MAS 13C NMR Spectrum of Precipitate Fraction from Alkaline Hydrolysis of Unlabeled TNTa

precipitate fraction a

230-190 ppm, ketone

190-160 ppm, carboxyl, quinone, carbonate

160-90 ppm, aromatic

90-60 ppm, alcohol, ether

60-0 ppm, aliphatic

0.2

16.8

50.1

5.8

27.2

Electronic integration.

VOL. 38, NO. 7, 2004 / ENVIRONMENTAL SCIENCE & TECHNOLOGY

9

2227

FIGURE 4. Liquid-state continuous decoupled 13C NMR spectrum of the de-ashed precipitate fraction from calcium hydroxide hydrolysis/polymerization of unlabeled TNT.

FIGURE 6. Solid-state CP/MAS 13C NMR spectrum of the filtrate fraction from calcium hydroxide hydrolysis/polymerization of 15Nlabeled TNT. ct )contact time. LB ) line broadening.

FIGURE 5. Liquid-state continuous decoupled 13C NMR spectrum of the precipitate fraction from calcium hydroxide hydrolysis/ polymerization of unlabeled TNT, methylated with 13C-labeled iodomethane and tetrabutylammonium hydroxide. Peaks designated A (13.4 ppm), B (19.3 ppm), C (23.2 ppm), and D (57.5 ppm) correspond to residual tetrabutylammonium salts. one or no substituents on an aromatic ring give rise to the peak at 56.9 ppm. The region in the spectrum from approximately 29 to 47 ppm is comprised of N-methyl carbons. Distinct peaks occur at 30.0, 32.0, and 42.7 ppm. These peaks are likely attributable to methyl- and dimethylamine groups. The structural forms of nitrogen in the precipitate are discussed in further detail in the next section. The approximate chemical shift range for C-methyl groups is 20-30 ppm. A minor but distinct peak due to C-methyl occurs at 26.2 ppm in the spectrum. The diphenylmethylene carbons observed in the solid-state spectrum of the precipitate fraction (Figure 3) are potential sites for carbon methylation. Additional C-methyl carbon may contribute to the area of broad resonances underlying the sharp peaks due to the residual TBAH salts at 23.2 ppm (C) and 19.3 ppm (B). Assignments are summarized in Table 3. The broad bands in the solid-state CP/MAS 13C NMR spectrum of the filtrate fraction from hydrolysis of the TNT15N (Figure 6) indicate that the soluble polymeric material 3 is also a multicomponent mixture of significant complexity. The qualitative and quantitative distribution of carbons in the filtrate fraction is similar to the precipitate fraction from the unlabeled TNT, including significant concentrations of 2228

9

ENVIRONMENTAL SCIENCE & TECHNOLOGY / VOL. 38, NO. 7, 2004

aliphatic and carbonyl carbons. Noteworthy is the occurrence of methylene, diphenylmethane, and alcohol/ether carbons. 15N NMR Spectra. Solid-state CP/MAS 15N NMR spectra were recorded on the precipitate and filtrate fractions from the alkaline hydrolysis of TNT-15N3 (Figure 7). The residual nitro groups (368 ppm) are the peaks of major intensity in both spectra. (The solid-state δ(15N) for the TNT-15N3 standard is 366 ppm.) The broad distribution of other forms of nitrogen in the spectra confirms the complexity of the materials. In the spectrum of the precipitate, peak maxima occur at 72, 133, 186, 271, and 304 ppm. In the spectrum of the filtrate, peak maxima occur at 76, 116, 134, 201, 256, 303, and 618 ppm. The peak at 618 ppm in the filtrate spectrum corresponds to nitrite. Subsequent analyses confirmed that the nitrite could be removed from the filtrate fraction by dialysis through a 100 MWCO membrane. A number of peaks in the 15N spectra can be interpreted as the condensation products of reduced forms of nitrogen, i.e., ammonia, aminodinitrotoluenes, and other aromatic amines, and possibly hydroxylamine or nitrite, with catechol, hydroquinone, and other carbonyl groups. Whereas ammonia and hydroxylamine may undergo nucleophilic addition reactions with catechols and hydroquinones upon oxidation to their corresponding quinones, the aminodinitrotoluenes and other aromatic amines potentially may undergo both nucleophilic addition and free radical coupling reactions to the catechols or quinones. Assignment of the spectra is complicated by the fact that ammonia, aminodinitrotoluenes, hydroxylamine, and nitrite react with the hydroquinones, catechols, and other carbonyls to form products whose chemical shifts overlap one another. Possible assignments for the spectrum are listed in Table 4. These are derived from the chemical shift literature (Figure B, Supporting Information) and our previous studies on the reactions of ammonia, hydroxylamine, nitrite, aniline, aminonitrotoluenes, and reduced TNT amines with humic substances and model compounds (5, 30-35). In Figure 7, peaks at 72 (precipitate) and 76 (filtrate) ppm would encompass primary aminoquinone, secondary aminohydroquinone, phenoxazine, diphenylamine, arylamine, and hydrazine nitrogens. Peaks at 133 (precipitate) and 134 (filtrate) ppm would encompass unsubstituted indole and anilide nitrogens; the peak at 186

FIGURE 7. Solid-state CP/MAS 15N NMR spectra of filtrate and precipitate fractions from calcium hydroxide hydrolysis/polymerization of 15N-labeled TNT. ct ) contact time. LB ) line broadening. ss ) spinning speed. Asterisks denote spinning sidebands. In the filtrate spectrum, the upfield spinning sideband of the nitrite peak at 618 ppm occurs from approximately 330 to 400 ppm, underlying the nitro peak at 368 ppm. The solid-state δ(15N) for the TNT-15N3 standard is 366 ppm.

TABLE 3. Assignments for the 13C NMR Spectrum of the 13C-Methylated Precipitate Fraction from Calcium Hydroxide Hydrolysis of Unlabeled TNT peak, ppm

assignment

67.0, 62.0, 56.0

R-O-*CH3, methyl ether of phenol or alcohol R-CO2-*CH3, methyl ester of carboxylic acid N-*CH3, methylamine C-*CH3, C-methyl

52.3 42.7, 32.1, 30.0 26.2

ppm in the precipitate would correspond to N-substituted indole, pyrrole, and other heterocyclic nitrogens. Imidazole, oxazole, pyrazole, or nitrile nitrogens may contribute to the

peaks at 271 (precipitate) and 256 (filtrate) ppm. Imine, pyridine, quinoline, and azoxybenzene nitrogens are possible assignments for the peaks at 304 (precipitate) and 303 (filtrate) ppm. The liquid-phase 15N NMR spectra of the de-ashed TNT15N precipitate fraction (Figure 8) provide additional resolu3 tion over the solid-state spectrum. A weak shoulder from approximately 390 to 430 ppm is visible downfield from the nitro peak at 371 ppm in the ACOUSTIC spectrum. This shoulder corresponds to the tautomeric equilibrium position between nitrosophenol and quinone monoxime nitrogens (31, 34, 36). It constitutes evidence either for the oximation of quinone groups in the polymer by hydroxylamine or for the recondensation of nitrite into the polymer by nitrosation of phenolic moieties. The nitrosophenol/quinone monoxime peak is present at a much greater signal-to-noise ratio in the spectrum of the soil humic acid reacted with TNT-15N3 under conditions of alkaline degradation (Figure C, Supporting Information). The evidence for reincorporation of hydroxylamine or nitrite into the precipitate fraction suggests the possible occurrence of two other reactions of these compoundssoximation of ketones and nitrosation of activated methylene carbons. Both reactions would result in the formation of ketoximes, which would overlap with the nitro groups in the 15N NMR. Secondary Beckmann reactions of both nitrosophenols and ketoximes are also possible. The ACOUSTIC spectrum represents the quantitative distribution of nitrogens in the precipitate fraction. The DEPT spectra show only nitrogens directly bonded to protons. These occur downfield from the ammonia peak to approximately 170 ppm. The peak at 109 ppm is confirmed as primary amide nitrogen. Nitrogens from approximately 120 to 170 ppm are singly protonated. The filtrate molecules may be considered either as intermediates in the progression from TNT to precipitate material or as the products of a different degradation pathway. If the latter case is true, then differences between the solidstate 15N NMR spectra of the filtrate and precipitate suggest that chemical changes, not just increases in molecular weight, are associated with the transformation of filtrate into precipitate molecules. Two differences between the filtrate and precipitate materials are the lower ratio of nitro to other nitrogens in the precipitate (Table 4), and the greater intensity of the peak at 304 ppm in the precipitate compared to the peak at 303 ppm in the filtrate. Possible assignments for the peaks at 304 and 303 ppm include azoxy, imine, and pyridine nitrogens.

Discussion Previous studies have clearly demonstrated the destruction of TNT alone or as a constituent in soil matrixes upon alkaline hydrolysis. This study provides information on the structural characteristics of the polymeric material produced from

TABLE 4. Assignments and Peak Areas (%) for Solid-State CP/MAS 15N NMR Spectra of Filtrate and Precipitate Fractions from Alkaline Hydrolysis of TNT-15N3a chem shift range, ppm

assignment

filtrateb

precipitate

652-566 413-332 332-283 283-229 229-178 178-98 98-0

nitrite nitro, oxime, nitrosophenol azoxybenzene, imine, pyridine, pyrazine, quinoxaline, quinoline, phenoxazinone oxazole, imidazole, pyrazole, nitrile imidazole, various heterocyclic primary amide, secondary amide, N-phenylpyrrole, pyrrole, indole, aminoquinone aminohydroquinone, phenoxazine, diphenylamine, hydrazine

23.6 25.4 (33.3) 7.8 (10.2) 3.0 (3.9) 2.1 (2.7) 25.8 (33.8) 12.3 (16.1)

0 19.7 17.8 6.3 7.3 30.5 18.5

a The CP/MAS experiment does not provide absolute quantitation, and peak areas therefore represent a semiquantitative distribution of nitrogens in the precipitate and filtrate fractions. b Peak areas in the filtrate spectrum were corrected for contributions from the spinning sidebands to the nitrite peak at 618 ppm. Numbers in parentheses represent peak areas as percent of organic nitrogen.

VOL. 38, NO. 7, 2004 / ENVIRONMENTAL SCIENCE & TECHNOLOGY

9

2229

The precipitate and filtrate materials formed from the alkaline hydrolysis of TNT should be representative of products formed from alkaline hydrolysis of munitions scrap, and from surface liming or ex situ alkaline treatment of low organic carbon sandy soils contaminated with particles of TNT. Experiments were also carried out in which TNT-15N3 was subjected to alkaline hydrolysis in the presence of soil humic acid. These experiments demonstrated that the alkaline degradation products of TNT-15N3 condensed with the humic acid (Figure C, Supporting Information), and that the presence of the humic acid limited the extent of selfpolymerization of the TNT alkaline degradation products. Formation of precipitate material was not observed during the course of the hydrolysis. The implication of these results is that, in the case of soils with moderate to large concentrations of soil organic matter, the final products from alkaline treatment of TNT may differ significantly from the precipitate and filtrate materials described in this paper. The alkaline hydrolysis products of TNT would likely condense with the soil organic matter rather than polymerize with one another. The toxicities to organisms of the polymeric materials produced from alkaline hydrolysis of TNT need to be assessed. The susceptibilities of the polymeric materials to chemical and microbial degradation also need to be determined. One unresolved question regarding the environmental safety of composts of TNT-contaminated soils is the long-term potential for chemical or microbial rerelease of the reduced TNT amines covalently bonded to the organic matter (4). The same question applies to aromatic amines that may be incorporated into the polymeric materials produced from the alkaline hydrolysis of TNT.

Acknowledgments This work was supported in part by the Department of Defense SERDP Project, Explosives Conjugation Products in Remediation Matrices, Dr. Judith C. Pennington, Project Chief.

Supporting Information Available FIGURE 8. Liquid-state 15N NMR spectra of the de-ashed precipitate fraction from calcium hydroxide hydrolysis/polymerization of 15Nlabeled TNT. Solvent ) dimethyl-12C2,d6 sulfoxide. Chemical shifts for the TNT-15N3 standard in DMSO: 2,6-NO2, 367 ppm; 4-NO2, 361 ppm. LB ) line broadening. alkaline hydrolysis of pure TNT, as well as insight into some of the mechanisms of the hydrolysis and polymerization reactions. Carbon functional groups confirmed in the precipitate material include methyl, methylene, alcohol/ ether, phenolic, carboxylic acid, ketone, and possibly quinone. Nitrogen occurs in the precipitate material as aminohydroquinone, aminoquinone, heterocyclic, imine or azoxy, nitro, and nitrosophenol, among other possibilities. In general, the filtrate is structurally similar to the precipitate material. The NMR analyses confirm that release of nitrite ion by nucleophilic substitution of nitro groups by hydroxide ion, followed by recondensation of reduced forms of nitrite into the alkaline degradation products, is a primary process during the hydrolysis and polymerization of the TNT. The observation of diphenylmethane carbons in the polymeric material is an indication of the formation and subsequent reaction of the benzylic carbanion of TNT. The structural complexity of the TNT polymers offers an opportunity for further investigation by NMR, including homonuclear and heteronuclear shift correlation experiments, as well as other techniques for polymer characterization, such as ESR and MALDI-TOF. A more detailed understanding of the polymerization mechanisms may derive from such additional studies. 2230

9

ENVIRONMENTAL SCIENCE & TECHNOLOGY / VOL. 38, NO. 7, 2004

Solid-state DP/MAS 13C NMR spectrum of the precipitate fraction from calcium hydroxide hydrolysis of unlabeled TNT (Figure A), 15N NMR chemical shifts of model compounds relevant to this study (Figure B), and liquid-phase inverse gated decoupled 15N NMR spectrum of IHSS Elliot soil humic acid reacted with TNT-15N3 in base (Figure C). This material is available free of charge via the Internet at http:// pubs.acs.org.

Literature Cited (1) ) Spain, J. C. Biodegradation of Nitroaromatic Compounds; Plenum Press: New York, 1995. (2) Spain, J. C.; Hughes, J. B.; Knackmuss, H.-J. Biodegradation of nitroaromatic compounds and explosives; Lewis Publishers: Boca Raton, FL, 2000. (3) Jerger, D. E.; Woodhull, P. M. In Biodegradation of Nitroaromatic Compounds and Explosives; Spain, J. C., Hughes, J. B., Knackmuss, H. J., Eds.; Lewis Publishers: Boca Raton, FL, 2000; pp 395-423. (4) Thorn, K. A.; Pennington, J. C.; Hayes, C. A. Environ. Sci. Technol. 2002, 36, 3797-3805. (5) Thorn, K. A.; Kennedy, K. R. Environ. Sci. Technol. 2002, 36, 3787-3796. (6) Achtnich, C.; Fernandes, E.; Bollag, J.-M. Environ. Sci. Technol. 1999, 33, 4448-4456. (7) Bruns-Nagel, D.; Knicker, H.; Drzyzga, O. Environ. Sci. Technol. 2000, 34, 1549-1556. (8) Saupe, A.; Garvens, H. J.; Heinze, L. Chemosphere 1998, 36, 1725-1744. (9) Emmrich, M. Environ. Sci. Technol. 1999, 33, 3802-3805. (10) Bishop, R. L.; Flesner, R. L.; Larson, S. A.; Bell, D. A. J. Energ. Mater. 2000, 18, 275. (11) Karasch, C.; Popovic, M.; Qasim, M.; Bajpai, R. K. Appl. Biochem. Biotechnol. 2002, 98, 1173-1185.

(12) Felt, D. R.; Larson, S. L.; Valente, E. J. Chemosphere 2002, 49, 287-295. (13) Felt, D. R.; Larsen, S. L.; Hansen, L. D. Molecular Weight Distribution of the Final Products of TNT-Hydroxide Reaction; Report No. ERDC/EL TR-01-16; U.S. Army Corps of Engineers Engineer Research and Development Center: Vicksburg, MS, 2001. (14) Emmrich, M. Int. J. Environ. Anal. Chem. 2003, 83, 769-776. (15) Review and Evaluation of Alternative Technologies for Demilitarization of Assembled Chemical Weapons; The National Academies Press: Washington, DC, 1999; pp 213-229, http:// books.nap.edu/books/0309066395/html/. (16) Effenberger, F.; Koch, M.; Streicher, W. Chem. Ber. 1991, 124, 163-173. (17) Beck, J. R. Tetrahedron 1978, 34, 2057-2068. (18) Newman, K. E. A Review of Alkaline Hydrolysis of Energetic Materials: Is it Applicable to Demilitarization of Ordnance?; Report No. IHTR 2167; Naval Surface Warfare Center Indian Head Division: Indian Head, MD, 1999. (19) Hammersley, V. L. Historical and Experimental Studies of Alkali and Trinitrotoluene Reaction; Report No. WQEC/C 75-192; Naval Weapons Support Center: Crane, IN, 1975. (20) Heilmann, H. M.; Wiesmann, U.; Stenstrom, M. K. Environ. Sci. Technol. 1996, 30, 1485-1492. (21) Emmrich, M. Environ. Sci. Technol. 2001, 35, 874-877. (22) Thorne, P. G.; Jenkins, T. F.; Brown, M. Continuous Treatment of Low Levels of TNT and RDX in Range Soils Using Surface Liming; Special Report; U.S. Army Corps of Engineers Cold Regions Research and Engineering Laboratory: Hanover, New Hampshire, in press. (23) Davis, J. Topical Lime Treatment for Containment of Source Zone Energetics Contamination; Strategic Environmental Research and Development Program, 2001; www.serdp.org/ research/cu/cu-1230.pdf (accessed October 2003). (24) Review and Evaluation of Alternative Technologies for Demilitarization of Assembled Chemical Weapons; The National

(25) (26) (27)

(28) (29) (30) (31) (32) (33)

(34) (35)

(36)

Academies Press: Washington, DC, 1999; pp 58-70; http:// books.nap.edu/books/0309066395/html/. Rose, K. D.; Francisco, M. A. Energy Fuels 1987, 1, 233. Patt, S. L. J. Magn. Reson. 1982, 49, 161-163. Buncel, E. In Supplement F The chemistry of amino, nitroso and nitro compounds and their derivatives Part 2; Patai, S., Ed.; John Wiley and Sons: New York, 1982; Vol. 2, pp 1225-1260. Bunte, G.; Krause, H.; Hirth, T. J. Propellants, Explosives, Pyrotechnics 1997, 22, 160-164. Thorn, K. A.; Steelink, C.; Wershaw, R. L. Org. Geochem. 1987, 11, 123-137. Thorn, K. A.; Mikita, M. A. Sci. Total Environ. 1992, 113, 67-87. Thorn, K. A.; Arterburn, J. B.; Mikita, M. A. Environ. Sci. Technol. 1992, 26, 107-116. Thorn, K. A.; Pettigrew, P. J.; Goldenberg, W. S.; Weber, E. J. Environ. Sci. Technol. 1996, 30, 2764-2775. Thorn, K. A.; Goldenberg, W. S.; Younger, S. J.; Weber, E. J. 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 299-326. Thorn, K. A.; Mikita, M. A. Soil Sci. Soc. Am. J. 2000, 64, 568582. Pennington, J. C.; Thorn, K. A.; Hayes, C. A.; Porter, B. E.; Kennedy, K. R. Immobilization of 2,4- and 2,6-Dinitrotoluenes in Soils and Compost; Report No. ERDC/EL TR-03-2; U.S. Army Engineer Research and Development Center: Vicksburg, MS, 2003; http://www.wes.army.mil/el/elpubs/pdf/trel03-2.pdf. Witanowski, M.; Stefaniak, L.; Januszewski, H.; Szymanski, S.; Webb, G. A. Tetrahedron 1973, 29, 2833-2836.

Received for review October 13, 2003. Revised manuscript received January 20, 2004. Accepted January 21, 2004. ES030655A

VOL. 38, NO. 7, 2004 / ENVIRONMENTAL SCIENCE & TECHNOLOGY

9

2231