Environ. Sci. Technol. 2008, 42, 8352–8359
Variability of Nitrogen Isotope Fractionation during the Reduction of Nitroaromatic Compounds with Dissolved Reductants ´ E. HARTENBACH,† AKANE T H O M A S B . H O F S T E T T E R , * ,† MICHAEL AESCHBACHER,† MICHAEL SANDER,† DONGWOOK KIM,‡ TIMOTHY J. STRATHMANN,‡ WILLIAM A. ARNOLD,§ CHRISTOPHER J. CRAMER,| AND ´ P. SCHWARZENBACH† RENE Institute of Biogeochemistry and Pollutant Dynamics (IBP), ETH Zurich, Universita¨tstrasse 16, 8092 Zurich, Switzerland, Department of Civil and Environmental Engineering, University of Illinois at Urbana-Champaign, 205 North Mathews, Urbana, Illinois 61801, Department of Civil Engineering, University of Minnesota, 500 Pillsbury Dr. SE, Minneapolis, Minnesota, and Department of Chemistry and Supercomputing Institute, University of Minnesota, 207 Pleasant St. SE, Minneapolis, Minnesota 55455
Received April 17, 2008. Revised manuscript received August 13, 2008. Accepted August 15, 2008.
Compound-specific nitrogen isotope analysis was shown to be a promising tool for the quantitative assessment of abiotic reduction of nitroaromatic contaminants (NACs) under anoxic conditions. To assess the magnitude and variability of 15N fractionation for reactions with dissolved reductants, we investigated the reduction of a series of NACs with a model quinone (anthrahydroquinone-2,6-disulfonate monophenolate; AHQDS-) and a Fe(II)-catechol complex (1:2 Fe(II)-tiron complex; Fe(II)L26-) over the pH range from 3 to 12 and variable reductant concentrations. Apparent kinetic isotope effects, AKIEN, for the reduction of four mononitroaromatic compounds by AHQDS- ranged from 1.039 ( 0.003 to 1.045 ( 0.002 (average(1σ),consistentwithpreviousresultsforvariousmineralbound reductants. 15N fractionation for reduction of 1,2dinitrobenzene and 2,4,6-trinitrotoluene by AHQDS- and that of 4-chloronitrobenzene by Fe(II)L26-, however, showed substantial variability in AKIEN-values which decreased from 1.043 to 1.010 with increasing pH. We hypothesize that the isotopesensitive and rate-limiting step of the overall NAC reduction can shift from the dehydration of substituted N,N-dihydroxyanilines (large 15N fractionation upon N-O bond cleavage) to protonation or reduction of nitroaromatic radical anions (small 15N isotope effect upon electron transfer) consistent with calculations of semiclassical 15N isotope effects. Our results
* Corresponding author phone: +41 44 632 83 28; fax: +41 44 633 11 22; e-mail:
[email protected]. † ETH Zurich. ‡ University of Illinois at Urbana-Champaign. § Department of Civil Engineering, University of Minnesota. | Department of Chemistry and Supercomputing Institute, University of Minnesota. 8352
9
ENVIRONMENTAL SCIENCE & TECHNOLOGY / VOL. 42, NO. 22, 2008
imply that a quantitative assessment of NAC reduction using compound-specific isotope analysis (CSIA) might need to account for homogeneous and heterogeneous reactions separately.
Introduction Nitrogen isotope analysis has great potential for assessing transformation processes of N-containing organic contaminants including nitroaromatic compounds (NACs 1-4), an important class of soil and groundwater contaminants (5). We have recently shown for the abiotic reduction of NACs by mineral-bound Fe(II) species that the transformation of the aromatic nitro group to the nitroso intermediate is accompanied by large 15N fractionation (1-3). In all systems investigated, the corresponding apparent kinetic isotope effects, AKIEN, were found to be quite consistent between 1.03 and 1.04, indicating that light isotopologues (containing 14NO2) react 3-4% faster than the heavy isotopologues (containing 15 NO2). In contrast to the overall rates of reduction, 15N isotope effects were insensitive to aromatic substituents in reactions of the NACs with Fe(II) adsorbed to Fe(III) oxides or structural Fe(II) in reduced clay minerals (6, 7). Therefore, we hypothesized that the corresponding bulk 15N enrichment factors, , are likely representative of the N abiotic reduction of NACs and can hence be used to quantify the extent of transformation for a wide variety of nitroaromatic contaminants in complex subsurface environments. Besides solid-bound reductants, however, NACs can also be transformed by a variety of homogeneous species including dissolved natural organic matter (8), naptho- and anthraquinones in the presence of H2S (9, 10), and Fe(II) complexes with catechol and organothiol ligands (11). While the overall rates of NAC reduction generally depend on the speciation of the reductant, reactions rates determined for homogeneous reductants are more sensitive to the electron-accepting properties of the NACs (8, 10, 12-15). With reduced naphthoquinone species, for example, changing pH, aromatic NAC substituent, or reductant concentration results in variation of reduction rate constants by orders of magnitude. The only available AKIEN-values for reduction of 2- and 4-chloronitrobenzene with homogeneous reductants (i.e., 8-hydroxy-1,4-naphthoquinone (juglone) in solutions of H2S (1)) were consistent with observations reported for heterogeneous reductants. This data set, however, is too limited for drawing general conclusions on the variability of 15N fractionation with homogeneous reductants. It remains unclear from our previous studies on 15N fractionation whether this strong variability of NAC transformation rates bears consequences for the variability of 15N isotope effects. The objective of this study was to provide AKIEN-values for NACs reduced by homogeneous reductants that allow for a more comprehensive evaluation of NAC reduction in anoxic environments with compound-specific 15N analysis. To assess the variability of 15N kinetic isotope effects we investigated the 15N fractionation over a wide range of experimental conditions. From our earlier studies, we proposed that the 15N fractionation occurs predominantly during the cleavage of the first N-O bond of substituted N,N-dihydroxyanilines after a series of e-- and H+-transfers (3). To explore whether these e-- and H+-transfers influence the overall 15N fractionation, in this study, we compared one- and two-electron transfer agents as homogeneous model reductants, that is Fe(II)-complexes with tiron (catechol-3,5-disulfonate) and 9,10-anthrahydroquinone10.1021/es801063u CCC: $40.75
2008 American Chemical Society
Published on Web 10/15/2008
FIGURE 1. 15N fractionation during reduction of 2-nitrotoluene in solutions of AH2QDS at pH 7 using two different experimental setups vs fraction of remaining reactant (c/c0, panel a). Circles represent reactors in which 2-nitrotoluene reduction was followed in the presence of excess reductant concentrations compared to squares, which represent reactors in which the reductant concentration was limited to achieve fractional conversion of 2-nitrotoluene. The linearized 15N enrichment calculated according to eq 1 for the two experimental setups is shown in panel b (uncertainty of the slope represents ( σ); for 15N enrichment factors see entries 4 and 15 of Table 1. 2,6-disulfonate (AH2QDS), respectively. As was shown previously in studies on the kinetics of NAC reduction by Fe(II)-tiron (11) and AH2QDS (9), rates strongly correlated with the concentration of the Fe(II)L26- species (L4-)fully deprotonated tiron) and monophenolate species AHQDS-, respectively. By varying either reductant concentration or pH, that is by restricting the amount of available electrons or protons, we investigated the relevance of e-- and H+transfer kinetics in modulating 15N fractionation during NAC reduction. Similarly, substituent effects on 15N fractionation were investigated because the e- and H+accepting properties of NACs and their reaction intermediates differ with aromatic substituents, thus providing evidence for the origins of variable 15N fractionation.
Experimental Section For a complete list of chemicals, suppliers, and purities see the Supporting Information (SI). Hydroquinone Reductants. The reductions of NACs in solutions containing 9,10-anthrahydroquinonone-2,6-disulfonate acid (AH2QDS, reduced hydroquinone form) were conducted in an anoxic glovebox (N2 atmosphere at 25 ( 1 °C) over the pH range from 3 to 12 using 0.1 M buffer solutions of citrate, acetate, phosphate, borate, or carbonate in 0.1 M KCl. AH2QDS was generated via electrochemical reduction of the disodium salt of 9,10-anthraquinone-2,6-disulfonate (AQDS) following a procedure adapted from ref 16. Briefly, solid AQDS was transferred into the glovebox and suspended in the desired buffer (final AQDS concentrations were 10 to 28 mM). Dissolution was facilitated by immersion in an ultrasonic bath at temperature up to 70 °C outside the glovebox. Reduction of AQDS was carried out in a bulk electrolysis cell (Bioanalytical Systems Inc.) which consisted of a glass vessel closed with a Teflon cover including cathode (reticulated vitreous carbon electrode), anode (coiled platinum wire), and reference electrode (Ag/AgCl). The cathode compartment was separated from the anode compartment by a glass frit, thus preventing the reoxidation of AH2QDS. Aliquots of 90 mL AQDS solution were reduced by stepwise lowering the cathode potential using an Autolab PG 302 instrument (Eco Chemie B.V.). Final applied potentials ranged from -0.55 V (pH 3) to -0.8 V (pH 12). Total time of reduction was 2 and 6 h for solutions containing 10 mM and 30 mM AQDS, respectively. Complete reduction of AQDS to AH2QDS was verified by coulometry and absorption measurements at 328 nm.
Reduction of 2- and 4-nitrotoluene and 2- and 4-chloronitrobenzene in solutions of AH2QDS was performed in sealed 20 mL amber flasks, which were mixed with a magnetic stirrer. For the isotope ratio measurements, seven separate reactors were prepared containing 20 mL of AH2QDS at the desired pH and AH2QDS concentration. Reactions were initiated by addition of 200 µL of 40 mM methanolic NAC solution resulting in initial NAC concentrations of 400 µM. At given time intervals, the NAC reduction was quenched by addition of 0.5 M hexacyanoferrate. For each experiment, an AH2QDS-free buffer solution was spiked with the NAC and hexacyanoferrate. This blank reactor served as control and was used to determine the NAC initial concentration and the corresponding δ15N value. From each reactor, 0.5 mL was used for concentration measurements by HPLC, and the remaining solution was stored in the dark at 4 °C until δ15N analysis. Because the reduction of 1,2-dinitrobenzene (1,2-DNB) and 2,4,6-trinitrotoluene (2,4,6-TNT) by AH2QDS was complete within seconds, the degree of NAC conversion was controlled by limiting the amount of AH2QDS added to the reactor (Table 1, c(AH2QDS) is “variable” because AH2QDS is consumed during the reaction with the NAC and the AH2QDS concentration is, therefore, not constant). According to this procedure, seven reactors were prepared by diluting the required amount of AH2QDS with buffer to yield a final volume of 20 mL, and the reaction was initiated by addition of methanolic NAC stock solution. All other steps were identical to that of mononitroaromatic compounds. This approach did not lead to experimental artifacts on 15N fractionation. As shown in Figure 1, experiments conducted with 2-nitrotoluene at pH 7 according to this setup showed an almost identical 15N fractionation within one standard deviation (N ) -40 ( 1‰) as that measured from a kinetic experiment involving reaction with excess AH2QDS and quenching aliquots with 0.5 M hexacyanoferrate at different reaction times (N ) -38 ( 1‰, Figure 1). Catechol Ligand-Complexed FeII Reductant. 4-Chloronitrobenzene reduction by Fe(II)-tiron (catechol-3,5-disulfonate) complexes were conducted under strict anoxic conditions according to ref 11. In brief, reactions were performed in sealed 40 mL serum flasks with a total solution volume of 35 mL. Like 1,2-DNB and 2,4,6-TNT experiments with AH2QDS, the degree of 4-chloronitrobenzene conversion (initial concentration 150 µM) was controlled by varying the amount of Fe(II) (0-1 mM) added to a series of reactors VOL. 42, NO. 22, 2008 / ENVIRONMENTAL SCIENCE & TECHNOLOGY
9
8353
TABLE 1. Nitroaromatic Compounds (NACs), Pseudo-First-Order Reduction Rate Constants, kobs, Second-Order Rate Constant for NAC Reduction by 2,6-Disulfonate Anthrahydroquinone Monophenolate, kAHQDS, Bulk 15N Enrichment Factors, EN, and Apparent 15N Kinetic Isotope Effect, AKIEN AH2QDS entry no.
NACa
pH
c(AH2QDS)tot (mM)b
kobs (s-1)
kAHQDS (M-1 s-1)c
EN (‰)d
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20
2-CH3-NB 2-CH3-NB 2-CH3-NB 2-CH3-NB 2-CH3-NB 2-CH3-NB 2-CH3-NB 2-CH3-NB 2-CH3-NB 2-CH3-NB 2-CH3-NB 4-CH3-NB 2-Cl-NB 4-Cl-NB 2-CH3-NB 1,2-DNB 1,2-DNB 1,2-DNB 2,4,6-TNT 2,4,6-TNT
3 5 6 7 8 9 10 11 12 5 5 5 5 5 7 5 6 7 5 7
3.0 3.0 3.0 3.0 3.0 3.0 3.0 3.0 3.0 10.1 28.3 3.0 3.0 3.0 variable variable variable variable variable variable
1.06 · 10-6 5.63 · 10-5 5.61 · 10-4 2.83 · 10-3 9.22 · 10-3 9.42 · 10-3 8.70 · 10-3 2.69 · 10-3 4.29 · 10-4 2.69 · 10-4 1.51 · 10-3 6.27 · 10-4 2.44 · 10-3 9.20 · 10-3 n.d.e n.d.e n.d.e n.d.e n.d.e n.d.e
1.11 · 10-1 5.95 6.10 3.93 4.05 3.34 3.82 3.73 4.67 8.46 1.69 6.63 · 101 2.58 · 102 9.72 · 102 n.d.e 2.55 · 105f n.d.e n.d.e 1.60 · 105f n.d.e
-43.3 ( 0.3 -38.9 ( 0.8 -37.7 ( 1.0 -38.1 ( 0.8 -40.5 ( 0.7 -40.2 ( 0.4 -41.2 ( 0.7 -41.3 ( 0.5 -43.1 ( 0.8 -41.9 ( 0.3 -39.1 ( 0.8 -39.5 ( 0.5 -40.4 ( 0.2 -37.1 ( 0.4 -40.3 ( 0.6 -17.1 ( 0.3 -13.6 ( 0.1 -5.3 ( 0.5 -8.6 ( 0.4 -3.4 ( 0.2
Fe(II)-tiron entry no.
NACa
pH
c(tiron)tot (mM)
kobs (s-1)I
c(FeL26-) (M)j
EN (‰)d
4-Cl-NB 4-Cl-NB 4-Cl-NB 4-Cl-NB 4-Cl-NB 4-Cl-NB 4-Cl-NB 4-Cl-NB 4-Cl-NB
6.0 6.5 7.0 7.5 8.0 8.5 9.0 7.0 7.0
8.15 · 10-4 4.72 · 10-2 1.24 · 100 1.02 · 101 2.56 · 101 3.39 · 101 3.68 · 101 1.09 · 10-1 6.86 · 100
2.14 · 10-8 1.24 · 10-6 3.25 · 10-5 2.67 · 10-4 6.71 · 10-4 8.90 · 10-4 9.65 · 10-4 2.85 · 10-6 1.80 · 10-4
-34.6 ( 1.3 -30.3 ( 1.3 -28.1 ( 0.4 -20.7 ( 1.6 -13.3 ( 3.0 -12.1 ( 0.5 -13.7 ( 1.2 -41.2 ( 4.4 -17.5 ( 1.3
21 22 23 24 25 26 27 28 29
10 10 10 10 10 10 10 2 50
AKIEN (-)e 1.0453 ( 0.0003 1.0405 ( 0.0009 1.0392 ( 0.0011 1.0396 ( 0.0008 1.0423 ( 0.0008 1.0419 ( 0.0004 1.0429 ( 0.0007 1.0431 ( 0.0006 1.0450 ( 0.0008 1.0438 ( 0.0003 1.0407 ( 0.0008 1.0412 ( 0.0005 1.0421 ( 0.0002 1.0385 ( 0.0004 1.0420 ( 0.0006 1.0354 ( 0.0007g 1.0281 ( 0.0003g 1.0107 ( 0.0010g 1.0265 ( 0.0011h 1.0100 ( 0.0005h AKIEN (-)e 1.0358 ( 0.0013 1.0312 ( 0.0013 1.0289 ( 0.0004 1.0212 ( 0.0016 1.0135 ( 0.0030 1.0122 ( 0.0005 1.0139 ( 0.0012 1.0429 ( 0.0044 1.0178 ( 0.0013
a Abbreviations: 2-CH3-NB (2-nitrotoluene), 4-CH3-NB (4-nitrotoluene), 2-Cl-NB (2-chloronitrobenzene), 4-Cl-NB (4-chloronitrobenzene), 1,2-dinitrobenzene (1,2-DNB), 2,4,6-trinitrotoluene (2,4,6-TNT). b Total AQDS concentration in solution. c kAHQDS ) kobs/[AHQDS-]; for 2-CH3-NB kAHQDS equals 3.62 M-1 s-1 from linear regression of kobs vs [AHQDS-] for all pH-values. d Uncertainty represents ( 1 standard deviation. e n.d. ) Not determined. f Estimated from correlation of kAHQDS vs E1h/0.059, see Figure S2. g Calculated from eq 2 with n/x · z ) 2. h Calculated from eq 2 with n/x · z ) 3. I kobs ) kFeL2 [FeL2], kFeL2 ) 3.81 · 104 M-1s-1 (11). j Calculated concentration of the 1:2 Fe-tiron complex assuming [Fe2+]total ) 1 mM.
containing a constant total tiron concentration (2-50 mM), pH (6-9; 50 mM MES (2-(N-morpholino)ethanesulfonic acid), MOPS (3-(N-morpholino)ethane sulfonic acid), or TAPS (3-(tris(hydroxymethyl)methyl)-3-aminopropanesulfonicacid)), and ionic strength (0.25 M NaCl). Reactions were initiated by adding 350 µL of a 15 mM 4-chloronitrobenzene methanolic stock solution and were allowed to proceed to completion before analyzing for concentration and δ15N. For each isotopic experiment, a solution containing all constituents except Fe(II) was spiked with 4-chloronitrobenzene and served as a control for determining the initial concentration and δ15N value of the NAC. Analytical Procedures and Data Evaluation. Concentrations of NACs were determined by HPLC on a LC-18 reverse phase column with UV/vis detection as described previously (1). δ15N signatures of NACs were determined according to a method using solid-phase microextraction (SPME) coupled to a GC/C/IRMS (gas chromatography isotope-ratio mass spectrometry with combustion interface (17)). All δ15N values were derived from triplicate measurements with good precision (< (0.5‰, (1σ) and are reported relative to N2 in air as δ15NAir. To avoid uncertainty of δ15N due to instrument nonlinearity (18), aqueous samples were diluted to the level of the least concentrated NAC solution prior to analysis and measured at constant peak amplitudes between 0.8 and 2 V. 8354
9
ENVIRONMENTAL SCIENCE & TECHNOLOGY / VOL. 42, NO. 22, 2008
Bulk 15N enrichment factors, N, of the NACs were derived from linear regression analysis of eq 1 (19) ln(δ15N + 1000) )
εN 1000
(
· ln(c) + ln
(δ15N0 + 1000) c0εN ⁄ 1000
)
(1)
where δ15N0 and δ15N are the initial 15N signatures of the NACs and its value during the reaction, and c0 and c are the substrate’s initial concentration and concentration following different extents of NAC reduction, respectively. Apparent kinetic isotope effects, AKIEN, were calculated according to eq 2 AKIEN )
(
1 1 + n ⁄ x · z · εN ⁄ 1000
)
(2)
where n, x, and z are correction factors accounting for isotopic dilution (n number of isotopic N atoms), number of reactive sites (x), and of reactive positions in intramolecular isotopic competition (z) (20). The overall correction factor (n/x · z) was unity for 2-/4-nitrotoluenes and 2-/4-chloronitrobenzenes and equaled 2 and 3 for 1,2-DNB and 2,4,6-TNT, respectively. Computational Methods. Calculations to determine kinetic isotope effects for the second electron transfer were performed with the Gaussian 03 electronic structure
FIGURE 2. Panel a: Logarithms of pseudo-first-order rate constants of 2-nitrotoluene reduction in solutions of AH2QDS (log kobs, empty squares) and logarithm of AHQDS- concentration (log [AHQDS-], filled squares) as a function of pH. Panel b: Apparent kinetic isotope effects, AKIEN, of 2-nitrotoluene reduction by AHQDS- as a function of the pH (circles) in solutions containing 3 mM total AQDS and in solutions of 10 and 28 mM total AQDS at pH 5 (diamonds and squares). Panel c, Logarithms of calculated pseudo-first-order rate constants 4-chloronitrobenzene reduction by Fe(II)-tiron complexes (log kobs, empty squares) at 10 mM total ligand concentration and logarithm of calculated FeL26- concentration (log [FeL26-], filled squares). Panel d, AKIEN of 4-chloronitrobenzene reduction by Fe(II)-tiron as function of the pH (diamonds); at pH 7, total tiron concentrations of 2 and 50 mM (circles and squares) were also considered. All uncertainties represent ( one standard deviation. program suite (21) as described in detail in the SI. Kinetic isotope effects were calculated following the formulation of Kavner et al. (22). Because the reaction is an electron transfer, activation free energies were determined using Marcus theory (23). The activation free energy ∆G* in Marcus theory is
(
l ∆G ∆G 2 + ∆G ) + 4 2 4λ /
)
(3)
where ∆G is the reaction driving force, and λ is the reorganization energy. The driving force (and thus the activation barrier and the KIE) is dependent on the reducing agent, so a series of ∆G values ranging from -0.6 to to 0.3 eV were surveyed. In addition, three different values of λ were considered, 100, 200, and 300 kJ/mol. Using eq 3, ∆G* values were determined for the 14N-containing species as a function of ∆G and λ. The activation free energy is sensitive to isotopic substitution because the 298 K driving force is sensitive to isotopic substitution. In principle, the difference in driving force for the two isotopes may be computed from isotopically sensitive molecular free energies determined by Gaussian 03 using standard statistical mechanical formulas associated with the ideal-gas, harmonic-oscillator, rigid-rotator approximation. Because of the output formatting employed by the program, however, an additional significant figure is available if one
instead employs the computed partition functions directly. In the latter case, the difference in driving forces may be computed as ∆G15N - ∆G14N ) -RT ln R
(4)
where R is the universal gas constant, T is temperature in K, and R is R)
Q15N
Q14N
Q14N
Q15N
dianion
dianion
radical anion
(5)
radical anion
and the Q values are the partition functions for the isotopologues of the product (dianion, Scheme 1 compound 3) and reactant (radical anion, Scheme 1, compound 1) as determined from the Gaussian 03 frequency calculations. Using eqs 3 and 4, activation energies were calculated for the 15N-containing species as a function of ∆G and λ. KIE values are then computed as 14N
KIE ) 15
k
N
k
)
MW14N MW15N
exp
(
∆G /15N - ∆G /14N RT
)
(6)
where the second term accounts for the change in activation free energy, while the first corrects for the change in collision frequency due to the change in molecular weight (MW). VOL. 42, NO. 22, 2008 / ENVIRONMENTAL SCIENCE & TECHNOLOGY
9
8355
FIGURE 3. (a) Linearized 15N enrichment behavior during the homogeneous reduction of 1,2-dinitrobezene (1,2-DNB) by AHQDS- at pH 5 (circles), pH 6 (squares), and pH 7 (triangles), lines represent the linear regression analysis used for the calculation of 15N enrichment factors, EN, according to eq 1. (b) Linearized 15N enrichment for 2,4,6-trinitrotoluene (2,4,6-TNT) reduction by AHQDS- at pH 5 (circles) and pH 7 (triangles) and corresponding linear regression analysis.
Results and Discussion NAC Reduction by Anthrahydroquinones. The pseudo-firstorder rate constant kobs of 2-nitrotoluene reduction in solutions of AH2QDS for the pH range 3 to 12 are shown in Figure 2a. Overall rates of NAC reduction varied over four orders of magnitude with pH, consistent with previous observations using other dissolved naphthoquinone reductants (10). Figure 2a shows that log kobs and the log of the anthrahydroquinone-2,6-monophenolate concentration (AHQDS-, Figure 2a (filled squares)) exhibited an almost identical dependence on solution pH. In fact, the rate constant, kobs, was strongly correlated with the AHQDSconcentration over the entire pH range (Figure S1), showing that the monophenolate is the predominant reductant of NACs in our experimental systems. Our data are consistent with previous observations in ref 9, albeit with a higher accuracy and over a wider pH range. Changes of pH modulate the concentration of the reactive species and thus indirectly affect the overall rate of reduction. Figure 2b shows the effect of the solution pH on the 15N fractionation. In contrast to the overall rates of NAC reduction, 15N fractionation of 2-nitrotoluene was constant over the entire pH range and insensitive of the total reduced anthrahydroquinone concentration (between 3 and 28 mM of AH2QDS). Measured bulk 15N enrichment factors, N, of 2-nitrotoluene were between -37.7 ( 1.0‰ and -43.3 ( 0.3‰ resulting in apparent kinetic isotope effects (AKIEN) between 1.0392 ( 0.0011 and 1.0453 ( 0.0003 (mean ( standard deviation). These values are very similar to what was found for 2-nitrotoluene reduction by mineral-bound Fe(II) species (1.0392 ( 0.0024 (3)). Note that for 2-nitrotoluene, 15N fractionation remained constant despite a change in H+ concentration of 9 orders of magnitude (Figure 2b) and variable total reductant concentration, and, thus, regardless of the corresponding change in the reduction potential of the system (24, 25). The constant AKIEN-values therefore suggest that for all chosen experimental conditions, NAC reduction proceeded by the same mechanism, that is via e- and H+ transfers from AHQDS- (see discussion below). Substituent Effects on Rates of NAC Reduction and 15N Fractionation. In AQDS solutions at pH 5, we measured the reduction rates and AKIEN-values of 2- and 4-nitrotoluene and 2- and 4-chloronitrobenzene. As found earlier for other dissolved reductants (NOM, juglone/H2S, Fe(II)-tiron (8, 10, 11, 15)), reduction rate constants (Table 1) were very sensitive to changes of the electron-accepting properties of the NACs (Figure S2). Note that reduction rates of 1,2-dinitrobenzene (1,2-DNB) and 2,4,6-trinitrotoluene (2,4,6-TNT) by AHQDS8356
9
ENVIRONMENTAL SCIENCE & TECHNOLOGY / VOL. 42, NO. 22, 2008
were too fast to be experimentally accessible. While reduction rate constants of nitrotoluenes and chloronitrobenzenes varied by two orders of magnitude (Table 1, entries 2, 12-14), N-values were constant within a narrow range of a few per mil (i.e., -37.1 ( 0.4‰ to -40.4 ( 0.2‰). Corresponding AKIEN-values between 1.0385 ( 0.0004 and 1.0421 ( 0.0002 showed no substituent effects. This is in agreement with recent work in Fe(II)-mineral suspensions, where average AKIEN-values of 1.038 ( 0.001 for mono- and polynitroaromatic compounds were measured (2, 3). From the previously observed absence of substituent effects on 15N fractionation, we concluded that all NO2-groups in 1,2-DNB and 2,4,6-TNT are also equivalent with regard to 15N fractionation in homogeneous solution. Thus AKIEN-values of polynitroaromatic compounds (Table 1) were calculated by correcting the observed N-values (Figure 3) for isotopic dilution, the number of reactive sites, and intramolecular isotopic competition (eq 2). In contrast to mononitroaromatic compounds, AKIEN-values decreased for 1,2-DNB from 1.0354 ( 0.0007 at pH 5 to 1.0107 ( 0.0010 at pH 7 and for 2,4,6-TNT from 1.0265 ( 0.0011 (pH 5) to 1.0100 ( 0.0005 (pH 7). To evaluate whether decreasing 15N fractionation of 1,2DNB and 2,4,6-TNT with pH were caused by diffusion-limited reaction rates, second-order rate constants can be estimated from the correlation of log kAHQDS- vs the free energy changes of the first electron transfer to the NAC (E1h/0.059 V, Figure S2). This type of correlation has been used previously to quantify the sensitivity of reduction rate constants of a series of NACs toward different reductants (26). Resulting bimolecular rate constants of 1,2-DNB and TNT were 2.6 · 105 M-1s-1 and 1.6 · 105 M-1 s-1, respectively, that is clearly below diffusion-control (second-order rate constants (
