Environ. Sci. Technol. 2002, 36, 2443-2450
Sediment-Associated Reactions of Aromatic Amines. 2. QSAR Development DALIZZA COLO Ä N,† E R I C J . W E B E R , * ,† A N D GEORGE L. BAUGHMAN‡ National Exposure Research Laboratory, U.S. Environmental Protection Agency, Athens, Georgia, 30605, and Department of Textiles, Merchandising and Interiors, University of Georgia, Athens, Georgia 30613
The fate of aromatic amines in soils and sediments is dominated by irreversible binding through nucleophilic addition and oxidative radical coupling. Despite the common occurrence of the aromatic amine functional group in organic chemicals, the molecular properties useful for predicting reaction kinetics in natural systems have not been thoroughly investigated. Toward this goal, the sorption kinetics for a series of anilines with substituents in the ortho, meta, or para positions were measured in sediment slurries. The sorption kinetics of the substituted anilines were characterized by an initial, rapid sorption process followed by a much slower sorption process. The initial rates of sorption varied with the type and position of the substituent group. Rate constants for the initial sorption process were correlated with molecular descriptors, including dissociation constants (pKa’s), Hammett σ constants, polarographic half-wave potentials (E1/2), one-electron oxidation potentials (E1), highest occupied molecular orbital (HOMO) energies (EHOMO), and ionization energies (EIE). On the basis of the strength of linear correlations and the availability of data, dissociation constants and Hammett σ constants appear to be the most useful molecular descriptors for predicting reaction rates of substituted anilines in the sediment slurries. The slow rates of sorption were much less sensitive to substituents effects than the rate constants for the faster sorption process, suggesting that the slower process was not controlled by the rate of electron transfer (i.e., nucleophilic addition or radical formation) but was limited by the availability of covalent binding sites.
Introduction Aromatic amines, which are the building blocks for many textile dyes, agrochemicals, and other classes of synthetic chemicals, comprise an important class of environmental contaminants. Concern exists over their release to the environment during production processes or improper treatment of industrial waste streams. Aromatic amines are also the end products resulting from the reductive transformation of nitroaromatics (1, 2) and azo dyes (3-5). A thorough understanding of how these chemicals interact with * Corresponding author phone: 706-355-8224; fax: 706-355-8201; e-mail:
[email protected]. † National Exposure Research Laboratory. ‡ University of Georgia. 10.1021/es0113551 CCC: $22.00 Published on Web 05/01/2002
2002 American Chemical Society
FIGURE 1. Proposed pathways for the covalent binding and transformation of aromatic amines in soil and sediment systems. natural surfaces is important to successfully model their transport and transformation in aquatic and soil ecosystems. Sorption of aromatic amines is observed to occur initially by a rapid, reversible equilibrium, followed by a slower, irreversible sorption process. The rapid sorption step has generally been attributed to electrostatic interactions, hydrophobic partitioning, and the formation of labile aminecarbonyl adducts (e.g., imines; 6-10), though, recently, we reported that the initial rapid sorption can also be dominated by irreversible covalent binding (11). The slower process has been attributed to irreversible covalent binding through the nucleophilic addition of the amino functional group to electrophilic sites (i.e., carbonyl moieties) or oxidative mechanisms resulting in the formation of radical species that couple with sediment-bound radicals (Figure 1; 6-10). The anilino radical species can also form dimers through head-to-tail or head-to-head coupling reactions (Figure 1). Dimerization in soils has been reported for readily oxidizable aromatic amines (10). 15N NMR studies have provided direct spectroscopic evidence for the reaction pathway occurring through nucleophilic addition to carbonyl moieties in dissolved organic matter (12, 13). Although there has been a substantial effort toward understanding the processes controlling the sorption of aromatic amines to natural surfaces, few quantitative structure-activity relationships (QSARs) have been developed for the purpose of predicting sorption kinetics for aromatic amines in natural soils and sediments. The development of successful QSARs for describing the reactivity of substituted anilines in well-defined model systems suggests that such relationships for natural systems should be obtainable. The initial oxidation rates of substituted anilines in MnO2 suspensions were log-linearly correlated with polarographic half-wave potentials (E1/2) and Hammett σ+ constants (14, 15). Also, a log-linear correlation of the pseudo-first-order rate constant for the nucleophilic addition of substituted anilines with p-benzoquinone, a model compound for the quinone functional group, with Hammett σ+ constants has been reported (6). More recently, a log-linear relationship VOL. 36, NO. 11, 2002 / ENVIRONMENTAL SCIENCE & TECHNOLOGY
9
2443
was reported between reaction rate constants for irreversible binding for a limited number of p-substituted aromatic amines in soil slurries and both Hammett σ+ constants and E1/2 oxidation potentials (10). To further develop QSARs for aromatic amines in natural systems, the sorption kinetics for 21 anilines with substituents in the ortho, meta, or para positions were measured in sediment slurries. The substituent groups (-OCH3, -CH3, -Cl, -Br, -CF3, -C(O)CH3, and -CN) were selected to generate a wide range in reactivity. Sorption rate constants were correlated with seven molecular descriptors describing the reactivity of the substituted anilines. In addition to providing a predictive tool for sorption rate constants for aromatic amines in sediment slurries, the correlation analyses also provided insight into the rate-determining step for irreversible binding (i.e., chemical kinetics versus mass transfer).
Experimental Section Materials. The monosubstituted anilines were analytical grade (Aldrich Chemical, Milwaukee, WI) and were used without further purification. Sediment was collected from Cherokee Park (CP) pond in Athens, GA. The sediment samples were passed through a 1-mm sieve and then airdried by spreading to a depth of several millimeters in glass pans and being allowed to stand for 5 days. Further experimental details concerning sediment collection and preparation are described in Weber et al. (11). Sediment analysis was conducted using standard methods in the Crop and Sciences Department at the University of Georgia (Athens, GA). The percent organic carbon of the CP sediment was 3.3 ( 0.5. The percentages of sand, silt, and clay in the sediment were 36.5%, 32.1%, and 31.4%, respectively. The CEC, base saturation, and particle density for CP sediment were 11.8 meq/100 g, 3.98%, 2.41 g/cm3, respectively. The clay mineralogy of Cherokee Park sediment was dominated by kaolinite. Other minerals detected in smaller amounts included chlorite/vermiculite, hydrous mica, gibbsite, goethite, and quartz. Kinetic Studies. The air-dried pond sediment, doubledistilled water, and a 250-mL Erlenmeyer flask containing a magnetic stirring bar were heat-sterilized (20 min, 121 °C, 20 psi). Using sterile techniques, 95 mL of H2O was added to 5.0 g of sediment in the 250-mL Erlenmeyer flask. An aqueous stock solution (1.0 × 10-4 M) of the substituted aniline, which had been sterilized by filtering through a 0.2-µm Nylon 66, was introduced with a sterile, disposable 1-mL pipet. The pH of the sediment slurry treated in this manner was 5.5. The flask was placed on a temperature-controlled platform shaker at 20 °C. At preselected times, the flask was removed from the shaker and placed on a magnetic stirrer. With vigorous stirring, a 3-mL aliquot was removed using sterile techniques. The aliquot was transferred to two 1.5-mL polypropylene centrifuge tubes and centrifuged for 10 min at 14 000 rpm. The supernatant was removed and transferred to glass autosample vials for LC analysis. A control study conducted in duplicate demonstrated that sampling the sediment suspension in this manner resulted in an increase in the sediment-to-water ratio of 0.95. Rate constants with r2 values 20 h) for o-, m-, and p-chloroaniline (Figure 3). The first-order rate constants for both the fast and slow sorption processes for each of the 21 monosubstituted anilines are summarized in Table 1. Molecular Descriptors. The common feature of the reaction processes through which covalent binding is thought to occur, nucleophilic addition and radical formation and coupling, both involve electron transfer (Figure 1). Nucleophilic addition to carbonyl moieties (e.g., quinones) is a twoelectron process, whereas the formation of the amino radical is a one-electron process. Consequently, molecular descriptors that describe the “willingness” of the aromatic amine to donate electrons have the potential to correlate with reaction rate constants measured in the sediment slurries. The list of available molecular descriptors that meet this criterion include dissociation constants (pKa), Hammett σ and σ+ constants, polarographic half-wave potentials (E1/2), oneelectron oxidation potentials (E1), highest occupied molecular orbital (HOMO) energies (EHOMO), and ionization energies (EIE) (10, 15-17). The molecular descriptors for the ortho-, meta-, and para-substituted anilines are summarized in Table 1. The basicity of aromatic amines, as manifested by their pKa, reflects the tendency of an amine to donate a pair of electrons to a proton in a partly covalent/partly ionic bond with the development of positive charge on nitrogen:
Reaction processes involving aromatic amines with transition states that resemble the transition state for the protonation of the amine have the potential to correlate with pKa’s (16). The transition states for the nucleophilic addition of anilines VOL. 36, NO. 11, 2002 / ENVIRONMENTAL SCIENCE & TECHNOLOGY
9
2445
FIGURE 4. Scatter plot matrix of the indicated molecular descriptors for the ortho-, meta-, and para-substituted anilines. Coefficients of determination, r 2, are given in the boxes. to simple carbonyl compounds (I) and quinones (II) are expected to have similar characteristics (i.e., development of positive charge on nitrogen).
Correlations of basicity with nucleophilicity have been observed over restricted ranges of basicity (18). Of special significance is the observation that log k (first-order rate constant) versus pKa for the nucleophilic addition of a series of meta- and para-substituted anilines with benzoquinone were linear (19). Quinone moieties are thought to be the dominant site for the nucleophilic addition of aromatic amines to natural organic matter (12, 13, 20). Rate constants for the yeast-mediated biotransformation of a series of meta2446
9
ENVIRONMENTAL SCIENCE & TECHNOLOGY / VOL. 36, NO. 11, 2002
and para-substituted anilines also were found to correlate with pKa (21). The biotransformation appeared to involve the enzyme-catalyzed nucleophilic attack of the substituted aniline on acetyl-Coenzyme A to form the corresponding acetanilide. As with most ionizable organic chemicals, an extensive database of measured pKa’s exists for organic bases (22). In addition, pKa’s can be obtained from computational tools such as SPARC (23). The use of Hammett σ constants is the most general means for estimating the electronic effect of substituents on reaction centers (24). Hammet σ constants are powerful molecular descriptors because they account for solution effects on substituents such as hydrogen bonding and dipole interactions. Furthermore, Hammett σ constants are additive for substituents in meta and para positions. Hammett σ+ constants are also available for para substituents that have the ability to directly conjugate with an electron-demanding reaction center in the transition state (24). Although Hammett σ and σ+ constants are available for most meta and para substituents, there is a very limited data set for ortho substituents (25, 26). Successful correlations based on Hammett σ and σ+ constants have been observed for the nucleophilic addition of substituted anilines to benzoquinone (6, 19) and the oxidation of substituted anilines by MnO2 (14, 15).
FIGURE 5. Correlations between log kfast and selected molecular descriptors for the ortho-, meta-, and para-substituted anilines. Measured E1/2 values are available for 12 of the 21 monosubstituted anilines used in this study (27). It was concluded that these half-wave oxidation potentials, which were measured by anodic voltammetry at a wax-impregnated graphite electrode, reflect the potential for the first oneelectron oxidation step (ArNH2 f ArNH•+). Strong correlations based on E1/2 values have been observed for the oxidation of substituted anilines by MnO2 (14, 15) and reaction rates for the irreversible binding of anilines in soil slurries (10). Because the complete data set of measured E1/2 values were not available, calculated one-electron oxidation potentials (E1) were also used for QSAR development. Cyclic voltammetry measurements have demonstrated that the first oneelectron transfer is the potential rate-determining step in the oxidation of anilines (28). Semiemperical molecular orbital theory and density functional theory were used to compute the E1 values for the set of 21 substituted anilines used in this study (29). In this same study, EHOMO values for
the set of 21 substituted anilines were also calculated. This quantum-chemical descriptor is the frontier molecular orbital from which electron transfer takes place (30). Experimentally determined EIE values, which are a measure of the energy for the first one-electron oxidation step in the gas phase (ArNH2 f ArNH2•+), were found for 13 of the substituted anilines (Table 1). Previous to this paper, correlations of reaction rate constants measured in soil or model systems for aromatic amines with one-electron oxidation potentials, HOMO energies, and ionization potentials have not been reported. A scatter plot matrix of the molecular descriptors illustrates their interdependence (Figure 4). These data suggest that there is a good correlation (r2 > 0.9) between several of the molecular descriptors. The strong correlation between pKa values and Hammett σ values was expected because Hammett σ values are obtained from differences in pKa values between substituted benzoic acids and the parent compound, benzoic acid (24). Likewise, the strong correlation observed for E1/2 VOL. 36, NO. 11, 2002 / ENVIRONMENTAL SCIENCE & TECHNOLOGY
9
2447
TABLE 2. Equations of Linear Regressions of log (kfast) and log (kslow) versus the Indicated Molecular Descriptors eq 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 a
regression equationa log (kfast) ) 0.411 ((0.040)pKa - 3.14 ((0.13) log (kfast) ) 0.463 ((0.039)pKa - 3.26 ((0.14) log (kfast) ) 0.269 ((0.050)pKa - 2.91 ((0.14) log (kfast) ) -1.429 ((0.200)σ - 1.25 ((0.08) log (kfast) ) -1.593 ((0.213)σ - 1.15 ((0.08) log (kfast) ) -0.849 ((0.192)σ - 1.65 ((0.09) log (kfast) ) -2.716 ((0.696)E1/2 + 0.221 ((0.465) log (kfast) ) -2.801 ((0.496)E1/2 + 0.457 ((0.328) log (kfast) ) -2.086 ((0.063)E1/2 - 0.564 ((0.043) log (kfast) ) -2.787 ((0.644)E1 + 0.211 ((0.472) log (kfast) ) -2.952 ((0.750)E1 + 0.486 ((0.537) log (kfast) ) -1.940 ((0.711)E1 - 0.739 ((0.545) log (kfast) ) 1.299 ((0.454)EHOMO + 9.341 ((3.891) log (kfast) ) 1.298 ((0.510)EHOMO + 9.511 ((4.367) log (kfast) ) 0.991 ((0.582)EHOMO + 6.333 ((5.012) log (kfast) ) -0.893 ((0.199)EIE + 5.497 ((1.594) log (kfast) ) -1.058 ((0.200)EIE + 6.933 ((1.601) log (kfast) ) -0.558 ((0.159)EIE + 2.494 ((1.277) log (kfast) ) -0.762 ((0.194)EIE + 4.697 ((1.578) log (kfast) ) -1.522 ((0.114)EIE + 10.449 ((0.899) log (kslow) ) 0.273 ((0.065)pKa - 3.458 ((0.218) log (kslow) ) -0.909 ((0.326)σ - 2.196 ((0.133) log (kslow) ) -1.864 ((1.198)E1/2 - 1.266 ((0.835) log (kslow) ) -1.327 ((0.867)E1 - 1.629 ((0.646) log (kslow) ) 1.039 ((0.501)EHOMO + 6.327 ((4.311) log (kslow) ) -0.612 ((0.238)EIE + 2.424 ((1.917)
substituents o, m, p m, p o o, m, p m, p o o, m, p m, p o o, m, p m, p o o, m, p m, p o o, m, p m, p o m p o, m, p o, m, p o, m, p o, m, p o, m, p o, m, p
n 21 14 7 18 14 4 12 8 4 21 14 7 21 14 7 15 11 4 5 6 17 14 10 17 17 12
r2 0.850 0.923 0.852 0.762 0.823 0.907 0.604 0.842 0.998 0.496 0.563 0.598 0.301 0.350 0.367 0.608 0.757 0.860 0.838 0.978 0.538 0.393 0.232 0.135 0.223 0.398
Numbers in parentheses are (95% CI.
and E1 values was anticipated because the redox reaction at the graphite electrode is thought to be a one-electron transfer resulting in the formation of the free radical species (27). Correlation Analysis of the Fast Sorption Kinetics. The rate constants for both the fast (kfast) and slow (kslow) rates of sorption were correlated with the molecular parameters describing the electron-donating (e.g., substituents with negative Hammett σ constants) and withdrawing (e.g., substituents with positive Hammett σ constants) properties of the substituents. Plots of log kfast versus the molecular descriptors are illustrated in Figure 5. In general, the rate of initial sorption was dependent on the electron-donating and electron-withdrawing capabilities of the substituents. The rate of sorption increased with the electron-donating capability of the substituents (e.g., with decreasing Hammett σ constants). This general trend is consistent with irreversible sorption, occurring through either nucleophilic addition or oxidative radical coupling (Figure 1) and reversible sorption occurring through cation exchange. The equations for the linear regressions of log kfast versus the selected molecular descriptors are summarized in Table 2. On the basis of the strength of linear correlations for the complete set of ortho-, meta-, and para-substituted anilines and the availability of data, dissociation constants, and Hammett σ constants appear to be the most useful molecular descriptor for predicting initial rates of sorption of substituted anilines in the pond sediment. Use of the Hammett σ+ constants for the para substituents did not significantly improve the correlation observed with the Hammett σ constants (data not shown). In a similar study, a log-linear relationship was reported between reaction rate constants for contact times 20 h) with Hammett σ+ constants and E1/2 values for the irreversible binding of a series of parasubstituted anilines in surface soils. The total number of reaction sites in their surface soils was determined experimentally to be much greater than the number of reacted sites (10). Thus, the strong correlation observed with the molecular descriptors describing the reactivity of the substituted anilines indicates that the rate of irreversible binding in these surface soils was controlled by the rate of electron transfer. Clearly, this work and the work of Li et al. (10) have provided valuable information concerning the molecular properties of substituted anilines that are most useful for predicting reaction kinetics in sediment and soil slurries. The successful development of predictive models for aromatic amines in natural systems, however, requires knowledge of soil properties that can be correlated with aromatic amine reactivity. To date, this objective remains elusive and will continue to be the focus of future studies. VOL. 36, NO. 11, 2002 / ENVIRONMENTAL SCIENCE & TECHNOLOGY
9
2449
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.
Literature Cited (1) Haderlein, S. B.; Schwarzenbach, R. P. In Biodegradation of Nitroaromatic Compounds; Spain, J. C., Ed.; Plenum Press: New York, 1995; pp 199-225. (2) Elovitz, M. S.; Weber, E. J. Environ. Sci. Technol. 1999, 33, 26172625. (3) Weber, E. J.; Wolfe, N. L. Environ. Toxicol. Chem. 1987, 6, 911919. (4) Baughman, G. L.; Weber, E. J. Environ. Sci. Technol. 1994, 28, 267-276. (5) Weber, E. J.; Adams, R. L. Environ. Sci. Technol. 1995, 29, 11631170. (6) Parris, G. E. Environ. Sci. Technol. 1980, 14, 1099-1106. (7) Graveel, J. G.; Sommers, L. E.; Nelson, D. W. Environ. Toxicol. Chem. 1985, 4, 607-613. (8) Lee, L. S.; Nyman, A. K.; Li, H.; Nyman, M. C.; Jafvert, C. Environ. Toxicol. Chem. 1997, 16, 1575-1582. (9) Li, H.; Lee, L. S. Environ. Sci. Technol. 1999, 33, 1864-1870. (10) Li, H.; Lee, L. S.; Jafvert, C. T.; Graveel, J. G. Environ. Sci. Technol. 2000, 34, 3674-3680. (11) Weber, E. J.; Colo´n, D.; Baughman, G. Environ. Sci. Technol. 2001, 35, 2470-2475. (12) Thorn, K. A.; Pettigrew, P. J.; Goldenberg, W. S.; Weber, E. J. Environ. Sci. Technol. 1996, 30, 2764-2775. (13) Thorn, K. A.; Goldenberg, W. S.; Younger, S. J.; Weber, E. J. In Humic/Fulvic Acids and Organic Colloidal Materials in the Environment; Gaffney, J. S., Marley, N. A., Clark, S. B., Eds.: ACS Symposium Series 651; American Chemical Society: Washington, DC, 1996; pp 299-326. (14) Laha, S.; Luthy, R. G. Environ. Sci. Technol. 1990, 24, 363-373. (15) Klausen, J.; Haderlein, S. B.; Schwarzenbach, R. P. Environ. Sci. Technol. 1997, 31, 2642-2649. (16) Jencks, W. P. Catalysis in Chemistry and Enzymology; Dover Publishers: New York, 1987; pp 79-84. (17) Luthy, R. G.; Aiken, G. R.; Brusseau, M. L.; Cunningham, S. D.; Gschwend, P. M.; Pignatello, J. J.; Reinhard, M.; Traina, S. J.; Weber, W. J.; Westall, J. C. Environ. Sci. Technol. 1997, 31, 33413347. (18) Bruice, T. C.; Lapinski, R. J. Am. Chem. Soc. 1958, 80, 22652267.
2450
9
ENVIRONMENTAL SCIENCE & TECHNOLOGY / VOL. 36, NO. 11, 2002
(19) Iskander, M. L.; Medien, H. A. A.; Nashed, S. Z. Phys. Chem. 1988, 6, 1183-1193. (20) Weber, E. J.; Spidle, D. L.; Thorn, K. A. Environ. Sci. Technol. 1996, 30, 2755-2763. (21) Moore, S. A.; Pope, J. T.; Barnett, J. T.; Suarez, L. A. StructureActivity Relationships and Estimation Techniques for Biodegradation of Xenobiotics; U.S. Environmental Protection Agency: U.S. Government Printing Office: Washington, DC, 1990; pp 1-100. (22) Perrin, D. D. Dissociation Constants of Organic Bases in Aqueous Solution; Butterworth: London, U.K., 1965. (23) Hilal, S.; Karickhoff, S. W.; Carreira, L. A. Quantum Struct. Act. Relat. 1995, 14, 348-355. (24) Hansch, C.; Leo, A. Exloring QSAR: Fundamentals and Applications in Chemistry and Biology; American Chemical Society: Washington, DC, 1995; pp 1-68. (25) Clark, J.; Perrin, D. D. Q. Rev., Chem. Soc. 1964, 18, 295-320. (26) Norrington, F. E.; Hyde, R. M.; Williams, S. G.; Wootton, R. J. Med. Chem. 1975, 18, 604-607. (27) Suatoni, J. C.; Snyder, R. E.; Clark, R. O. Anal. Chem. 1961, 33, 4-1897. (28) Sharma, L. R.; Maanchanda, A. K.; Singh, G.; Verma, R. S. Electrochim. Acta 1982, 27, 223-233. (29) Winget, P.; Weber, E. J.; Cramer, C. J.; Truhlar, D. G. Phys. Chem. Chem. Phys. 2000, 2, 1231-1239. (30) Fleming, I. Frontier Orbitals and Organic Chemical Reactions; Wiley: London, U.K., 1976. (31) Lowry, T. H.; Richardson, K. S. Mechanism and Theory in Organic Chemistry, 3rd ed.; Harper and Row: New York, 1987; p 141. (32) University of Georgia. On-line SPARC Calculator. http:// ibmlc2.chem.uga.edu/sparc (accessed May 2001). (33) Kvita, V.; Sauter, H.; Rihs, G. Helv. Chim. Acta 1985, 68, 15691576. (34) Farrell, P. G.; Newton, J. Tetrahedron Lett. 1966, 45, 5517-5523. (35) Baldwin, M. A.; Loudon, A. G.; Maccoll, A.; Webb, K. S. Org. Mass Spectrom. 1976, 11, 1181-1193. (36) Brown, P. Org. Mass Spectrom. 1970, 4, 519-532. (37) Brown, P. Org. Mass Spectrom. 1970, 3, 639-646. (38) Buchs, A. Helv. Chim. Acta 1970, 53, 2026-2035. (39) Modelli, A.; Distefano, G. Z. Naturforsch. 1981, 36, 1344-1351.
Received for review October 11, 2001. Revised manuscript received March 5, 2002. Accepted March 5, 2002. ES0113551
