Reduction of Nitrosobenzenes and N-Hydroxylanilines by Fe (II

Athens, Georgia 30605-2720, and Department of Chemistry, .... University of Georgia. Environ. ..... (11) Colón, D.; Weber, E. J.; Anderson, J. L. QSA...
0 downloads 0 Views 155KB Size
Download PDF
Environ. Sci. Technol. 2006, 40, 4449-4454

Reduction of Nitrosobenzenes and N-Hydroxylanilines by Fe(II) Species: Elucidation of the Reaction Mechanism DALIZZA COLO Ä N , * ,† E R I C J . W E B E R , † JAMES L. ANDERSON,‡ P A U L W I N G E T , † A N D L U I S A . S U AÄ R E Z † National Exposure Research Laboratory, U.S. Environmental Protection Agency, 960 College Station Road, Athens, Georgia 30605-2720, and Department of Chemistry, University of Georgia, Athens, Georgia 30602-2556

Although there has been a substantial effort toward understanding the reduction of nitroaromatics in Fe(II)treated ferric oxide systems, little has been done to gain insight into the factors controlling the transformation of their reaction intermediates, nitrosobenzenes and N-hydroxylanilines, in such systems. Nitrosobenzenes, the first intermediates, were reduced by Fe(II) solutions as well as by Fe(II)-treated goethite suspensions at pH 6.6. Experimental observations indicate a reactivity trend in which the presence of electron-withdrawing groups in the para position increased the rate of reduction of the nitrosobenzenes. N-Hydroxylanilines, the second intermediates, were reduced in Fe(II)-treated goethite suspensions but were not reduced by Fe(II)aq. Their reactivity trend indicates that electron-withdrawing groups in the para position decreased their rate of reduction. The bond dissociation enthalpy of the N-O linkage was the most useful molecular descriptor for predicting the rates of reduction of N-hydroxylanilines in Fe(II)-treated goethite suspensions, suggesting that the cleavage of the N-O bond is the rate-determining step for reduction. The rate of reduction of p-cyano-Nhydroxylaniline showed a linear relationship against the concentration of surface-associated Fe(II) in hematite, goethite, and lepidocrocite suspensions, while having a relatively low sensitivity toward changes in pH within the nearneutral range in hematite suspensions.

Introduction Nitroaromatic compounds are ubiquitous contaminants in the environment. Their primary transformation pathway in natural anoxic systems is reduction, which predominantly occurs by mineral-surface-associated Fe(II), Fe(II)surf (1-3). The higher reactivity of Fe(II)surf, versus that of Fe(II)aq, toward the reduction of nitroaromatics and other chemicals has been attributed to surface complexation of mineral surface hydroxyl groups, which act as ligands, stabilizing Fe(III) and significantly shifting the redox potential to lower values (4, 5). This reaction proceeds through the formation of two evenelectron-reduction intermediates, namely, the nitroso and * Corresponding author phone: (706) 355-8223; fax: (706) 3558202; e-mail: [email protected]. † U.S. Environmental Protection Agency. ‡ University of Georgia. 10.1021/es0600429 CCC: $33.50 Published on Web 06/10/2006

 2006 American Chemical Society

N-hydroxylamino compounds, with subsequent formation of the terminal aromatic amine product (Figure 1) (6). The nitroso derivatives are more reactive (7) and in many instances more toxic than the parent molecule (8). The polar and oxygen-sensitive N-hydroxylamino intermediates are usually more mutagenic than the nitroaromatic parent compound (9, 10). Our earlier research efforts, directed toward the study of the reduction of nitrobenzenes (NBs) in Fe(II)-treated ferric oxide suspensions (11), prompted an investigation of the reactivity of their reaction intermediates in such systems. Nitrosobenzenes (NOs) have not been found in reduction studies of nitrobenzenes in Fe(II)-treated ferric oxide systems and in natural systems under ferrogenic redox conditions (1, 11-17), but have been detected in studies of nitrobenzene reduction by zero-valent iron (18, 19). For this reason, there is still a lack of information regarding what species are responsible for the facile reduction of nitrosobenzenes in reducing systems. There is evidence that N-hydroxylaniline (HA) intermediates are formed from the reduction of NBs in Fe(II)-treated ferric oxide systems and are quite stable in such systems (12-14, 16, 17). Also, it has been reported that surface-bound hydroxylaminodinitrotoluenes, formed from the reduction of trinitrotoluene (TNT), are recalcitrant toward biological reduction (20). The complex environmental fate of HAs can make the design of remediation schemes a challenging task; the possible transformations they could undergo in the environment include reduction to anilines (ANs), binding to natural organic matter, condensation with nitroso functional groups, forming azoxy compounds, and rearrangement, forming aminophenols (see ref 9 and references therein). A few studies have measured the reduction rates of HA intermediates in Fe(II)/Fe(III) and sediment systems. Elsner et al. (17) studied the reduction of p-chloronitrobenzene, and monitored the formation and disappearance of the p-chloro-N-hydroxylaniline (ClHA) intermediate as well, in various Fe(II)-treated mineral suspensions. They reported the rate of reduction of ClHA to decrease in the order goethite > magnetite > lepidocrocite and that ClHA was consistently reduced at a slower rate than the parent nitro compound in such systems. In the same study, ClHA was not detected in an Fe(II)-treated hematite suspension. In another study, the reduction intermediate pentachloro-N-hydroxylaniline of the fungicide pentachloronitrobenzene was reduced about 3 times slower than its parent nitro compound in Fe(II)-treated goethite suspensions (16). Simon et al. (12) were able to detect the formation of p-cyano-N-hydroxylaniline (CNHA) when p-cyanonitrobenzene (CNNB) was fed into an anoxic sediment column. CNHA was reduced in the iron-reducing zone of the column. In the same study, it was assessed that CNHA was reduced about 20 times slower than CNNB in an Fe(II)treated goethite suspension. The purpose of this work was to (i) study the effect of electron-withdrawing and -donating functional groups on the reactivity of HA and NO structural moieties toward their reduction by Fe(II) species, (ii) determine the appropriate molecular descriptors for predicting the rate of reduction of HA compounds in Fe(II) systems, (iii) determine the reactivity of surface area loading of various ferric oxides that are abundant and commonly found in the environment (i.e., hematite, goethite, and lepidocrocite) toward the reduction of CNHA, and (iv) study the effect of pH on the reduction of CNHA and p-cyanonitrosobenzene (CNNO) in the presence of Fe(II) species. CNHA was selected as our model compound for the study of the effect of several environmentally relevant VOL. 40, NO. 14, 2006 / ENVIRONMENTAL SCIENCE & TECHNOLOGY

9

4449

FIGURE 1. Reaction scheme for the reduction of nitrobenzene, through the formation of the nitrosobenzene and hydroxylamine intermediates, followed by the formation of aniline. variables on HA reduction because its pKa is low enough, i.e., -0.7 to -0.85 (21) or -3.35 (22), to rule out changes in its speciation at the various experimental conditions used in this work.

Experimental Section Chemicals. The source, treatment, and analysis of the water used throughout this work are discussed in the Supporting Information. The following compounds were synthesized for this work (see the Supporting Information for details): p-Nhydroxylaminobenzonitrile (CNHA), p-N-hydroxylaminoacetophenone (AcHA), p-chloro-N-hydroxylaniline (ClHA), p-bromo-N-hydroxylaniline (BrHA), N-hydroxylaniline (HHA), p-N-hydroxyltoluidine (MeHA), p-cyanonitrosobenzene (CNNO), and p-nitrosotoluene (MeNO). The commercially available chemicals (Aldrich, unless otherwise noted) include nitrosobenzene (HNO), p-cyanonitrobenzene (CNNB), pcyanoaniline (CNNA), p-aminoacetophenone (AcAN), pchloroaniline (ClAN), p-bromoaniline (BrAN), aniline (HAN) (Fluka), and p-toluidine (MeAN). Mineral Phases. Goethite (R-FeOOH) was prepared as by Atkinson et al. (23) and Torrents and Stone (24). Hematite (R-Fe2O3), lepidocrocite (γ-FeOOH), and two-line ferrihydrite (∼Fe(OH)3) were prepared as by Schwertmann and Cornell (25). Additional details about the synthesis and characterization of oxides are offered in the Supporting Information. Molecular Modeling Calculations. The energy of the lowest unoccupied molecular orbital (ELUMO) and the bond energy (BE) were calculated using the GAMESSPLUS software. The LUMO values of the HAs were obtained by using the B3LYP/6-31G* theory/basis set. Solvation was introduced by SM5.42R calculations. Batch Studies. To study the reduction of a series of HA and NO compounds in the presence of ferrous ion species, individual batch kinetic experiments were performed in 59 mL serum bottles inside an anoxic glovebox operated as described elsewhere (11). The system contained ferric oxide and/or 375 µM ferrous ion (added as FeCl2‚4H2O (Fluka) unless otherwise noted) and 25 mM sulfonic acid buffer for pH control. Serum bottles containing the reaction mixture were crimped with Teflon-faced gray butyl septa and agitated in an end-over-end rotor at 49 rpm; the final volume of the suspensions was 50 mL. Fe(II) was equilibrated in the systems for 24 h except as otherwise noted. The reactions were started by spiking the organic chemical from a degassed stock solution prepared in methanol. The kinetic time was defined as the time elapsed between the spike of the reducible organic chemical and the end of the filtration of the sample aliquot of 1.3 mL using a 0.02 µm Anotop (unless otherwise noted) syringe filter. Removal of an aliquot and the filtering process usually took about 12 s. Some of the reactions were stopped by the addition of perchloric acid (60%, J. T. Baker). The progress of the reactions was monitored by quantifying the disappearance of the compound of interest, the formation and consumption of the intermediate, if applicable, and the concurrent appearance of the corresponding aniline compound in the filtrate. Organic Compounds and Fe(II) Analysis. See the Supporting Information. 4450

9

ENVIRONMENTAL SCIENCE & TECHNOLOGY / VOL. 40, NO. 14, 2006

Kinetic Studies. The batch experiments for the study of the reduction of HAs consisted of individual reactors containing 0.142 g/L goethite (51 m2/g), 25 mM MES (4morpholineethanesulfonic acid) for a pH of 6.6, and 375 µM Fe(II). Individual reactors were spiked with a stock solution of the HA compound (namely, CNHA, AcHA, ClHA, BrHA, HHA, or MeHA) to obtain an initial concentration of 15 µM. Each kinetic point aliquot was immediately analyzed after being filtered. Batch reactors were set to study the effect of the type and surface area loading of various oxides on the rate of reduction of CNHA, formed in situ from the transformation of CNNB. The oxides and their respective surface areas were as follows: hematite (34 m2/g), lepidocrocite (49 m2/g), and goethite (51 m2/g). The individual system components were hematite (0.55-60.2 m2/L), lepidocrocite (0.990-117 m2/L), or goethite (1.03-112 m2/L) and 25 mM MES (pH 6.6), 0.1 M NaCl, and 375 µM Fe(II). Fe(II) was equilibrated for 48 h in oxide suspensions. The sample aliquots were filtered and acidified as mentioned above. Individual batch experiments were performed with hematite (0.032 g/L) equilibrated with 375 µM Fe(II) and 0.1 M NaCl using sulfonic acid buffers (25 mM) to attain a range of pHs between 6.1 and 7.7 (i.e., MES for pH 6.1 and 6.6, MOPSO (β-hydroxy-4-morpholinepropanesulfonic acid) for pH 6.8, MOPS (4-morpholinepropanesulfonic acid) for pH 7.0, 7.28, and 7.5, and EPPS (N-(2-hydroxyethyl)piperazineN′-(3-propanesulfonic acid)) for pH 7.67). The reactions were started by the addition of 15 µM CNHA. Sample aliquots were taken, filtered (0.2 µM PTFE), and acidified to pH 1.4 with deoxygenated perchloric acid. The reduction of MeNO, HNO, and CNNO was studied in individual batch experiments containing goethite (0.0142 g/L) and/or 375 µM Fe(II) (added as FeCl2‚4H2O) at pH 6.6, 5.5, or 4.5 (MES). Mohr’s salt (Fe(NH4)2(SO4)2‚∼6H2O; Fluka) was evaluated as an alternative Fe(II) source. Sample aliquots were promptly removed from the anoxic chamber and immediately analyzed.

Results and Discussion Reduction of N-Hydroxylanilines. The reduction of a series of HA compounds, having electron-withdrawing or -donating functional groups in the para position, was studied in Fe(II)treated goethite suspensions. The HA compounds selected for this study were CNHA, AcHA, ClHA, BrHA, HHA, and MeHA. The observed order of elution by LC analysis of the species possibly present in samples was HA < AN < NB < NO. In reactors containing Fe(II)-treated goethite suspensions, the HAs were reduced with the concurrent formation of the respective aniline. In such systems, about 18 µM Fe(II) was associated with the surface of the goethite (7.24 m2/L) after a 24-h equilibrium period and Fe(II)aq was continuously consumed along the course of the reaction (Figure 2). The stoichiometry of the reaction between CNHA and Fe(II)aq was close to the expected 1:2 ratio on the basis of the consumed concentrations, 15 and 37 µM, respectively. Reduction rate constants of the HAs (kHA) were calculated using a pseudo-first-order approximation for this set of experiments (Supporting Information Table S1), ranging from 0.902 to 0.0897 h-1, and throughout the remaining portion of this work. The R2 values for calculated kHA values were 0.998 or higher, except for one AcHA reactor having a value of 0.990. The sorption (%) of HAs to system components, determined by spiking the HA into anoxic Fe(II)-free goethite suspensions, was 10% for CNHA, 9.0% for AcHA, 5.0% for ClHA, 7.8% for BrHA, 4.8 for HHA, and 2.0% for MeHA. In particle-free Fe(II)aq systems only two of the HAs, CNHA and MeHA, were reduced; the contribution of this side reaction to the overall rate, estimated as [kFe(II)aq/kgoethite+Fe(II)aq] × 100,

FIGURE 2. Reduction of CNHA with the concurrent consumption of Fe(II)aq and formation of CNAN in an Fe(II)-treated goethite suspension as a function of time. Initial conditions: goethite concentration 0.142 g/L, [Fe(II)] ) 375 µmol/L, and pH 6.6 (MES). was minimal, 0.17% and 0.73%, respectively. Good mass recoveries (%) were found for the reduction of HAs in Fe(II)treated goethite suspensions (i.e., 96% for CNHA, 95% for AcHA, 93% for ClHA, 90% for BrHA, 84% for HHA, and 100% for MeHA). The reactivity trend observed was opposite to what we originally expected (i.e., HAs with electronwithdrawing groups (CNHA, AcHA, BrHA, and ClHA) reacted more slowly than the one with an electron-donating group (MeHA)). QSAR Analysis. Several molecular descriptors were evaluated as independent descriptor variables for the rate of reduction through log-linear quantitative structure-activity relationships (QSARs). Because the reduction of HAs and formation of the corresponding aniline must involve the capture of two electrons, and the cleavage of the N-O bond, we evaluated molecular descriptors that would describe (i) the ability of the HAs to accept electrons and (ii) the strength of the N-O bond. The list of molecular descriptors that meet these criteria, and either are available in the literature or can be calculated using commercially available software, include the lowest unoccupied molecular orbital (LUMO) energy (ELUMO), the negative logarithm of the acidity constant of the hydrolylammonium form (pKa), the Hammett σ constant (σp), the electron affinity (EA), and the bond dissociation enthalpy (BDE) for the N-O linkage. Values for the molecular descriptors used for the training set can be found in Supporting Information Table S1. ELUMO was selected as a descriptor because the LUMO is the frontier molecular orbital into which electron donation takes place regardless of the mechanism for electron transfer (26). We initially tested the usefulness of the ELUMO (B3LYP, 6-31G*, water) as a predictor for kHA values by Fe(II) species (Figure 3a) motivated by our past experience in which we built successful QSAR relationships using ELUMO for the reduction of NBs in Fe(II)-treated goethite suspensions (11), and the fact that ELUMO values are easily calculated with commercially available molecular modeling packages. ELUMO provided a good quality correlation with log kHA (R2 ) 0.917). The reactivity trend observed offers insight about the ratedetermining step for reduction of HAs. That is, transfer of the first electron is most likely not the rate-determining step since there is an inverse relationship between the reactivity of the HAs and their LUMO values. Consequently, we proposed a mechanism in which the cleavage of the N-O bond is the rate-determining step (Figure 4). To evaluate this mechanism, other molecular descriptors were evaluated (i.e., pKa, σp, EA, and BDE), aiming to obtain more insight into whether the electron transfer or the cleavage of the N-O bond was the rate-determining step. We regarded pKa as a measure of the electron density on nitrogen. Assuming our proposed mechanism is valid, the higher the electron density on nitrogen (higher pKa), the faster

the reaction will occur since the N-O bond would be broken more easily. Since pKa values for HAs are scarce in the literature, we used the on-line SPARC calculator (27) to predict the pKa values for the training set (Supporting Information Table S1) for QSAR analysis. Reported experimentally determined pKa values for HAs (see ref 21 and references therein) are consistently several units higher than those predicted by SPARC. The SPARC-calculated pKa values gave a moderately good prediction of log kHA (R2 ) 0.883, Figure 3b), and account for the reactivity trend that is in agreement with the proposed rate-determining step: faster reaction rates were observed for HAs having higher pKa values. Another method for estimating the electronic effects of substituents on reaction centers is the use of Hammett σ constants (28), which account for solution effects on substituents such as hydrogen bonding and dipole interactions. kHA values showed a dependency on the electron-donating (negative σ value) and electron-withdrawing (positive σ value) capabilities of the substituents (Figure 3c). A moderately good prediction of log kHA was achieved using σ (R2 ) 0.888) along with the reactivity trend that is consistent with the proposed mechanism. The EA is defined as the negative of the enthalpy change at 0 K accompanying the gain of an electron (29), which is a measure of the ease of reduction of a compound. The SPARC-calculated EA (30) provided a good correlation with log kHA (R2 ) 0.916, Figure 3d). The correlation of log kHA with EA is another indication that the gain of the first electron is not the rate-determining step for electron transfer since less negative EAs did not lead to faster reductions. Knowing that the N-O bond must be broken to form the corresponding aniline, we proceeded to calculate its BDE. We hypothesized that if the cleavage of this bond is involved in the rate-determining step for reduction of HAs in Fe(II)treated goethite suspensions, then BDE should provide a robust predictor for log kHA. In fact, the use of BDE (B3LYP/ 6-31G*, water) allowed us to explain 97.6% of the variability in log kHA data (Figure 3e), and provided a correlation in which electron donors decreased the BDE and increased the rate of reactions (i.e., promoting N-O bond cleavage). While the correlation of BDE with log k cannot prove that the N-O bond cleavage is the rate-determining step, it is consistent with a thermodynamic picture where the endothermicity of the reaction would determine the transition state. Effect of the Type and Surface Area Loading of Ferric Oxides. It has been previously demonstrated that surface area loading (SAL; area of oxide/volume of suspension) of oxides can have a significant impact on the reduction of NBs in Fe(II)-treated oxide suspensions (11, 14, 31). In this work we studied the effect of a variation in the SAL of oxides on the reduction of HA intermediates. In general, an increase in the rate of reduction of CNHA (kCNHA) (Figure 5a), as well as an increase in [Fe(II)surf,o], was observed for all suspensions as a function of the SAL (Supporting Information Table S2). These findings are consistent with Fe(II) forming a more reactive species once it is associated with the surface (32). We found that kCNHA had a linear dependency, to a good approximation, on [Fe(II)surf,o] (Figure 5b). The reactivity trend of the mineral surfaces was hematite > goethite > lepidocrocite. The effect of an increase in SAL on kCNHA was not as pronounced as that on kCNNB, with the effect being 100, 40, and 2.5 times greater on kCNNB (11) than on kCNHA in Fe(II)-treated hematite, goethite, and lepidocrocite suspensions, respectively. A plausible explanation for the observed difference in the response of CNHA versus CNNB could be found in the differences of their rate-determining steps for reduction. It is accepted that the rate-determining step for the nitroaromatics is the trapping of the first electron (33), a process that should be surface-mediated under the conditions of this work VOL. 40, NO. 14, 2006 / ENVIRONMENTAL SCIENCE & TECHNOLOGY

9

4451

FIGURE 3. Best fit for the reduction rate constants of para-substituted N-hydroxylanilines, derived from their reaction in Fe(II)-treated goethite suspensions, versus the (a) ELUMO(HF/6-31G*,water), (b) pKa, (c) σp, (d) EA, and (e) BDE (B3LYP/6-31G*, water). Effect of pH. Since it has been demonstrated that pH has a tremendous impact on the reduction rate of NBs in aqueous Fe(II)/Fe(III) systems, in work performed in our laboratory and by other researchers (11, 13, 14, 31), we decided to evaluate the extent of this effect on the reduction kinetics of CNHA in Fe(II)-treated hematite suspensions.

FIGURE 4. Proposed reaction scheme for the reduction of Nhydroxylaniline in an Fe(II)-treated ferric oxide suspension. (1, 14). On the other hand, our results suggest that the cleavage of the N-O bond is the rate-determining step for reduction of the HAs, a process that takes place subsequently to the trapping of the first electron. The effect the electron-transfer kinetics has on kCNHA could be masked by the kinetics of the N-O bond cleavage. 4452

9

ENVIRONMENTAL SCIENCE & TECHNOLOGY / VOL. 40, NO. 14, 2006

kCNHA in hematite suspensions as a function of pH (i.e., pH 6.1-7.67) varied by a factor of 25, the reaction being faster and [Fe(II)surf,o] being greater as the pH increased (Figure 6 and Supporting Information Table S3). This trend is consistent with those observed for the reduction of nitroaromatic compounds (11, 13-16, 31, 34) and pentachloro-N-hydroxylaniline (16) in the presence of Fe(II)treated mineral suspensions, where reduction took place at a faster rate since Fe(II) sorption to oxides occurred to a larger extent as the pH was increased. The slope and R2 value of the plot of log kCNHA against pH was 0.921 and 0.977, respectively. CNHA did not show reactivity in the Fe(II)-only control studies with the exception of the system at pH 7.67, in which it was reduced at a rate of 0.00422 h-1. The lower

FIGURE 5. (a) Effect of changes in surface area loadings of ferric oxides on the rate of reduction of CNHA in Fe(II)-treated hematite, goethite, and lepidocrocite suspensions. (b) Dependence of kCNHA on the initial concentration of surface-associated Fe(II) in Fe(II)treated hematite, goethite, and lepidocrocite suspensions as a function of the surface area loading of oxides.

FIGURE 6. Effect of pH on the rate of reduction of CNHA in Fe(II)treated hematite suspensions. dependency of kCNHA (factor of 25 increase) compared to that of kCNNB (4 orders of magnitude increase) (11) on pH in Fe(II)-treated hematite suspensions is consistent with our proposed mechanism for N-hydroxylaniline reduction (Figure 4) in which a H+ is consumed prior to the rate-determining step (i.e., cleavage of the N-O bond). The wide difference in the reactivity of CNNB and CNHA as a function of pH could, in part, be rationalized with the same argument offered to account for the difference in the responses of these two compounds with respect to SAL variation (i.e., masking effect). Reduction of Nitrosobenzenes. This work studied the reduction of MeNO, HNO, and CNNO by Fe(II) species. The reduction of MeNO was faster in Fe(II)-treated goethite suspensions than in Fe(II) controls, 5.6% and 10% of MeNO was left after 1.1 and 1.2 min of reaction, respectively. The Fe(II) control turned yellow immediately after MeNO was spiked into it, indicating that Fe(II) had been oxidized. A statistically robust calculation of a rate constant for MeNO in the Fe(II) solution was not possible since only two kinetic points containing MeNO were collected before it was totally consumed. Recoveries of MeNO added to Fe(II)-treated

goethite suspension, particle-free Fe(II), and goethite sorption control experiments were 89% as MeAN, 85% as MeAN, and 102% as MeNO, respectively. HNO and CNNO were fully reduced to the corresponding HA intermediate when spiked into Fe(II)-treated goethite suspensions and particle-free Fe(II) solutions buffered to pH 6.6. Fe(II) particle-free experiments turned yellow immediately after the system was spiked with HNO and within 14 s after treatment with CNNO. The reduction of the HAs in this system was, most likely, facilitated by the presence of a freshly formed iron oxide or hydroxides produced from the oxidation of Fe(II). The recoveries for the reduction products of HNO and CNNO were 94% and 102%, respectively. The observed reactivity trend of CNNO > HNO > MeNO (Hammett σ constant values of 0.66, 0, and -0.17, respectively) (28) and speed of change in the color of Fe(II) solutions upon addition of NOs, though limited in its quantitative value, is aligned with a trend observed by Lutz and Lytton (35), which illustrates the tendency of electron-withdrawing substituents, particularly in the para position, to increase the reduction potential of nitrosobenzenes. For instance, they reported the following order of oxidation-reduction potential for various nitrosobenzenes: o-nitro (σ ) 0.78) > p-carbomethoxyl (σ ) 0.50) > p-chloro (σ ) 0.23) > p-methyl (σ ) -0.17) > p-methoxyl (σ ) -0.27) (28, 35). This work evaluated the influence of the vessel material, Fe(II) source, and pH on the reduction of CNNO. The reaction of CNNO and Fe(II) was conducted in polypropylene bottles. The color development took place instantaneously upon addition of CNNO (recovery of CNHA, 96%), ruling out the possibility that the reaction was being catalyzed by the silica surface. Mohr’s salt was reacted with CNNO at pH 6.6. The Fe(II) solution instantaneously turned yellow upon the addition of CNNO (recovery of CNHA, 101%). Also, CNNO was reacted with Fe(II) at pH 5.5 and 4.5. Under these conditions, CNNO was fully reduced to CNHA before the aliquots were injected into the LC instrument (∼2 min). It took approximately 45 min for a faint yellow hue to appear in the reactors at pH 5.5 (recovery of CNHA, 97%), and no color development occurred at pH 4.5 (recovery of CNHA, 99%). Under these conditions, oxidation of Fe(II) should slow to a significant extent (36) and Fe(II) sorption to any ferric oxide in the system should be negligible. Our results strongly suggest that the nitrosobenzenes studied in this work are reduced by both Fe(II)aq and Fe(II)surf, which is likely mass-transfer-limited. These observations account for the fact that nitrosobenzenes have not been detected in reduction studies of nitrobenzenes in Fe(II)treated oxide suspensions or natural sediments and aquifer systems under ferrogenic redox conditions (1, 11-17).

Environmental Significance Environmental Significance. This work reconciles previously published work on the transformation of reduction intermediates of nitroaromatics, namely, nitrosobenzenes and N-hydroxylanilines, by Fe(II)-treated ferric oxide aqueous systems. According to our findings, the reduction of nitrosobenzenes by Fe(II)aq and mineral-surface-associated Fe(II) is a facile reaction, which accounts for the lack of detection of nitrosobenzenes in Fe(II)-treated ferric oxide suspensions, natural sediments, or aquifer systems under ferrogenic conditions. In addition, this work offers an explanation to the observed slower reduction of N-hydroxylanilines containing an electron-withdrawing group, compared to their parent nitro compound. Thus, there exists a possibility that HAs formed in natural anoxic environments from nitro compounds bearing one or more electron-withdrawing groups could be more persistent than the parent compounds toward abiotic reduction in Fe(II)/Fe(III) systems. Of particular environVOL. 40, NO. 14, 2006 / ENVIRONMENTAL SCIENCE & TECHNOLOGY

9

4453

mental relevance in this respect is the case of TNT, which has been reported to produce 2-hydroxylaminodinitrotoluene and 4-hydroxylaminodinitrotoluene as reduction intermediates, which in turn were recalcitrant toward biological reduction (20).

Acknowledgments This paper has been reviewed in accordance with the U.S. Environmental Protection Agency’s peer and administrative review policies and approved for publication. Mention of trade names or commercial products does not constitute endorsement or recommendation for use by the U.S. EPA.

Supporting Information Available Experimental details, kinetic data, and molecular descriptors for para-substituted N-hydroxylanilines, effect of oxide surface area loading on the rate of reduction of p-cyanoN-hydroxylaniline, and effect of pH on the rate of reduction of p-cyano-N-hydroxylaniline in Fe(II)-treated hematite suspensions. This information is available free of charge via the Internet at http://pubs.acs.org.

Literature Cited (1) Hofstetter, T. B.; Heijman, C. G.; Haderlein, S. B.; Holliger, C.; Schwarzenbach, R. P. Complete reduction of TNT and other (poly)nitroaromatic compounds under iron-reducing subsurface conditions. Environ. Sci. Technol. 1999, 33, 1479-1487. (2) Ru ¨gge, K.; Hofstetter, T. B.; Haderlein, S. B.; Bjerg, P. L.; Knudsen, S.; Zraunig, C.; Mosbaek, H.; Christensen, T. H. Characterization of predominant reductants in an anaerobic leachate-contaminated aquifer by nitroaromatic probe compounds. Environ. Sci. Technol. 1998, 32, 23-31. (3) Heijman, C. G.; Grieder, E.; Hollinger, C.; Schwarzenbach, R. P. Reduction of nitroaromatic compounds coupled to microbial iron reduction in laboratory aquifer columns. Environ. Sci. Technol. 1995, 29, 775-783. (4) Stumm, W. In Aquatic Chemistry: Interfacial and Interspecies Processes; Huang, C. P., O’Melia, C. R., Morgan, J. J., Eds.; American Chemical Society: Washington, DC, 1995; pp 1-32. (5) Stumm, W. Catalysis of redox processes by hydrous oxide surfaces. Croat. Chem. Acta 1997, 70, 71-93. (6) Larson, R. A.; Weber, E. J. Reaction Mechanisms in Environmental Organic Chemistry; Lewis Publishers: Boca Raton, FL, 1994. (7) Corbett, M. D.; Corbett, B. R. In Biodegradation of Nitroaromatic Compounds; Spain, J. C., Ed.; Plenum Press: New York, 1995; Vol. 49, pp 151-182. (8) Spain, J. C. Biodegradation of nitroaromatic compounds. Annu. Rev. Microbiol. 1995, 49, 523-555. (9) Hughes, J. B.; Wang, C. Y.; Zhang, C. Anaerobic biotransformation of 2,4-dinitrotoluene and 2,6-dinitrotoluene by Clostridium acetobutylicum: A pathway through dihydroxylamino. Environ. Sci. Technol. 1999, 33, 1065-1070. (10) Hughes, J. B.; Wang, C.; Yesland, K.; Richardson, A.; Bhadra, R.; Bennet, G.; Rudolph, F. Bamberger rearrangement during TNT metabolism by Clostridium acetobutylicum. Environ. Sci. Technol. 1998, 32, 494-500. (11) Colo´n, D.; Weber, E. J.; Anderson, J. L. QSAR study of the reduction of nitroaromatics by Fe(II) species. Accepted for publication in Environ. Sci. Technol. (12) Simon, R.; Colo´n, D.; Tebes-Stevens, C. L.; Weber, E. J. Effect of redox zonation on the reductive transformation of pCyanonitrobenzene in a laboratory sediment column. Environ. Sci. Technol. 2000, 34, 3617-3622. (13) Schultz, C. A.; Grundl, T. J. pH dependence on reduction rate of 4-Cl-nitrobenzene by Fe(II)/montmorillonite systems. Environ. Sci. Technol. 2000, 34, 3641-3648. (14) Klausen, J.; Tro¨ber, S. P.; Haderlein, S. B.; Schwarzenbach, R. P. Reduction of substituted nitrobenzenes by Fe (II) in aqueous mineral suspensions. Environ. Sci. Technol. 1995, 29, 23962404. (15) Klupinski, T. P.; Chin, Y. P. Abiotic degradation of trifluralin by Fe(II): Kinetics and transformation pathways. Environ. Sci. Technol. 2003, 37, 1311-1318.

4454

9

ENVIRONMENTAL SCIENCE & TECHNOLOGY / VOL. 40, NO. 14, 2006

(16) Klupinski, T. P.; Chin, Y. P.; Traina, S. J. Abiotic degradation of pentachloronitrobenzene by Fe(II): Reactions on goethite and iron oxide nanoparticles. Environ. Sci. Technol. 2004, 38, 43534360. (17) Elsner, M.; Schwarzenbach, R. P.; Haderlein, S. B. Reactivity of Fe(II)-bearing minerals toward reductive transformation of organic contaminants. Environ. Sci. Technol. 2004, 38, 799807. (18) Agrawal, A.; Tratnyek, P. G. Reduction of nitro aromatic compounds by zero-valent iron metal. Environ. Sci. Technol. 1996, 30, 153-160. (19) Hung, H. M.; Ling, F. H.; Hoffman, M. R. Kinetics and mechanism of the enhanced reductive degradation of nitrobenzene by elemental iron in the presence of ultrasound. Environ. Sci. Technol. 2000, 34, 1758-1763. (20) Daun, G.; Lenke, H.; Reuss, M.; Knackmuss, H.-J. Biological treatment of TNT-contaminated soil. 1. Anaerobic cometabolic reduction and interaction of TNT and metabolites with soil components. Environ. Sci. Technol. 1998, 32, 1956-1963. (21) Darchen, A.; Boudeville, P. Determination polarographique des pKa des phenylhydroxylamines substituees et morphologie des polarogrammes des nitrobenzenes correspondants. Bull. Soc. Chim. Fr. 1971, 9, 3809-3815. (22) Karickhoff, S. W.; McDaniel, V. K.; Melton, C. M.; Vellino, A. N.; Nute, D. E.; Carreira, L. A. Predicting Chemical Reactivity by Computer. Environ. Toxicol. Chem. 1991, 10, 1405-1416. (23) Atkinson, R. J.; Posner, A. M.; Quirk, J. P. Adsorption of potentialdetermining ions at the ferric oxide-aqueous electrolyte interface. J. Phys. Chem. 1967, 71, 550-558. (24) Torrents, A.; Stone, A. T. Hydrolysis of phenyl picolinate at the mineral/water interface. Environ. Sci. Technol. 1991, 25, 143149. (25) Schwertmann, U.; Cornell, R. M. Iron Oxides in the Laboratory: Preparation and Characterization; VCH: New York, 1991. (26) Fleming, I. Frontier Orbitals and Organic Chemical Reactions; John Wiley & Sons, Ltd.: London, 1976. (27) Hilal, S. H.; Karickhoff, S. W.; Carreira, L. A. A rigorous test for SPARC’s chemical reactivity models: Estimation of more than 4300 ionization pKas. Quant. Struct.-Act. Relat. 1995, 14, 348355. (28) Hansch, C.; Leo, A. Exploring QSAR. Fundamentals and Applications in Chemistry and Biology; American Chemical Society: Washington, DC, 1995. (29) Lias, S. G.; Bartmess, J. E.; Liebman, J. F.; Holmes, J. L.; Levin, R. D.; Mallard, W. G. Gas-phase ion and neutral thermochemistry. J. Phys. Chem. Ref. Data 1988, 17. (30) Hilal, S. H.; Carreira, L. A.; Karickhoff, S. W.; Melton, C. M. Estimation of electron affinity based on structure activity relationship. Quant. Struct.-Act. Relat. 1993, 12, 389-396. (31) Amonette, J. E.; Workman, D. J.; Kennedy, D. W.; Fruchter, J. S.; Gorby, Y. A. Dechlorination of Carbon Tetrachloride by Fe(II) Associated with Goethite. Environ. Sci. Technol. 2000, 34, 4606-4613. (32) Scherer, M. M.; Balko, B. A.; Gallagher, D. A.; Tratnyek, P. G. Correlation analysis of rate constants for dechlorination by zerovalent Iron. Environ. Sci. Technol. 1998, 32, 3026-3033. (33) Riefler, R. G.; Smets, B. F. Enzymatic Reduction of 2,4,6Trinitrotoluene and Related Nitroarenes: Kinetics Linked to One-Electron Redox Potentials. Environ. Sci. Technol. 2000, 34, 3900-3906. (34) Pecher, K.; Haderlein, S. B.; Schwarzenbach, R. P. Reduction of polyhalogenated methanes by surface-bound Fe(II) in aqueous suspensions of iron oxides. Environ. Sci. Technol. 2002, 36, 17341741. (35) Lutz, R. E.; Lytton, M. R. Oxidation-Reduction potentials of a series of nitrosobenzene-phenylhydroxylamine systems. J. Org. Chem. 1937, 2, 68-75. (36) Stumm, W.; Lee, G. F. Oxygenation of Ferrous Ion. Ind. Eng. Chem. 1961, 53, 143-146.

Received for review January 9, 2006. Revised manuscript received April 13, 2006. Accepted April 14, 2006. ES0600429