Environ. Sci. Technol. 2006, 40, 1055-1061
Kinetics of Contaminant Degradation by Permanganate RACHEL H. WALDEMER AND PAUL G. TRATNYEK* Department of Environmental and Biomolecular Systems, OGI School of Science and Engineering, Oregon Health and Science University, 20000 NW Walker Road, Portland, Oregon 97006
To provide a more complete understanding of the kinetics of in situ chemical oxidation (ISCO) with permanganate (MnO4-), we measured the kinetics of oxidation of 24 contaminantssmany for which data were not previously available. The new data reported here were determined using an efficient method based on continuous measurement of the MnO4- concentration by absorbance spectrometry. Under these conditions, the kinetics were found to be firstorder with respect to both contaminant and MnO4concentrations, from which second-order rate constants (k ′′) were readily obtained. Emerging contaminants for which k ′′ was determined (at 25 °C and pH 7) include 1,4dioxane (4.2 × 10-5 M-1 s-1), methyl t-butyl ether (MTBE) (1.0 × 10-4 M-1 s-1), and methyl ethyl ketone (MEK) (9.1 × 10-5 M-1 s-1). Contaminants such as 2,4,6-trinitrotoluene (TNT), the pesticides aldicarb and dichlorvos, and many substituted phenols are oxidized with rate constants comparable to tetrachloroethene (PCE) and trichloroethene (TCE) (i.e., 0.03-1 M-1 s-1) and therefore are good candidates for remediation with MnO4- in the field. There are severalssometimes competingsmechanisms by which MnO4- oxidizes contaminants, including addition to double bonds, abstraction of hydrogen or hydride, and electron transfer.
Introduction Of the various treatments used for in situ chemical oxidation (ISCO) (1), permanganate (MnO4-) is sometimes preferred because of its comparative stability (and therefore mobility) in the subsurface, effectiveness over a wide pH range, ease of handling, and relatively low cost (2). To support the use of MnO4- (or any other ISCO technology) for particular contaminants of concern (COCs), a detailed understanding of the reaction kinetics and mechanisms is desirable. A few studies are available that provide such background for COC oxidation by MnO4- (3-5), but these studies are focused only on the chlorinated ethenes, which means that the reactivity of most potential COCs with MnO4- is not fully characterized. The lack of detailed kinetic and mechanistic data on the wide range of COCs that might be treatable with MnO4- is an impediment to expanded application of this remediation technology. Of the data obtained from detailed kinetic and mechanistic studies, second-order rate constants (k ′′) are particularly useful because they characterize reaction rates at a level of generality that is appropriate for comparisons among systems and yet they can usually be used to calculate system-specific * Corresponding author phone: (503)748-1023; fax: (503)748-1273; e-mail:
[email protected]. 10.1021/es051330s CCC: $33.50 Published on Web 12/21/2005
2006 American Chemical Society
measures of reaction rates such as half-lives (6, 7). For oxidation by MnO4-, however, reliable values of k ′′ have only been reported for a few COCs, including the chlorinated ethenes (3-5, 8-12), methyl t-butyl ether (MTBE) (13), hexahydro-1,3,5-trinitro-1,3,5-triazine (RDX) (14), and some polycyclic aromatic hydrocarbons (PAHs) (15). Other COCss such as the pesticides aldrin, dieldin, and chlordane (16); phenol; and various substituted phenols (17)sare known to be labile to oxidation by MnO4-, but the available kinetic data on these COCs are not sufficient to determine values of k ′′. No quantitative data are available for the COCs that are known qualitatively as unreactive with MnO4-, such as 1,1,1-trichloroethane (2, 9). In light of the importance of good kinetic data, we have compiled the values of k ′′ from the literature, determined the priority data gaps, designed an efficient method for filling the data gaps, and applied the method to many priority COCs (focusing on emerging contaminants, contaminants that are prevalent at military sites, and contaminants that form nonaqueous phase liquids in the environment).
Materials and Methods Experimental. The COCs were obtained in high purity from either Sigma-Aldrich (St. Louis, MO) or ChemService (West Chester, PA) and used without further purification. Potassium permanganate of 99% purity was obtained from SigmaAldrich (St. Louis, MO) and used as received. The KMnO4 crystals were dissolved in deionized water to make 5 mM (790 mg/L) stock solutions, which were stored in the dark until use (1-3 days). For the COCs that were dense nonaqueous phase liquids (DNAPLs), stock solutions were prepared as saturated solutions: 1 mL of a DNAPL was added to 20 mL of deionized water, shaken vigorously, and allowed to equilibrate overnight. Stock solutions for light nonaqueous phase liquids (LNAPLs) were also prepared as saturated solutions by adding 1 mL of LNAPL to 20 mL of deionized water; however, the solution was inverted overnight, and the aqueous phase was drawn with a syringe from the upturned vial to avoid contamination of the sample with pure-phase LNAPL. The stock solutions for the other contaminants were made by dissolving a measured quantity of each COC in deionized water or phosphate buffer. At least a 5:1 molar ratio of COC to MnO4- was used to ensure that conditions were pseudo-first-order in MnO4throughout each experiment. Although the kinetics of chlorinated ethene oxidation by MnO4- have been shown to be independent of pH over a range of pH 4-8 (3), a phosphate buffer was used to control pH at 7 for consistency. An additional purpose of the phosphate buffer was to prevent flocculation of the colloidal particles of manganese dioxide (MnO2), which is a product formed during the oxidations with MnO4- (18). The concentration of the phosphate buffer was 50 mM in most experiments (resulting in a total ionic strength of 0.2 M); however, for some contaminants, a 100 mM phosphate buffer was required to control flocculation of MnO2 (ionic strength ) 0.4 M). Detailed information about the concentration of reagents in individual experiments is provided as Supporting Information Table S1. A UV/vis spectrophotometer (Perkin-Elmer model Lambda 20) was used to measure the decrease in absorbance at 525 nm, an absorption maximum for MnO4-. Two experimental methods, designated stopped-flow and static, were used to study the kinetics of the reactions of COCs with MnO4-. The static method involved premixing the MnO4- and COC in a 1 cm path length, 5 mL quartz cuvette with a screw cap seal (Starna Cells, Atascadero, CA) or, for TNT only, a 5 cm path VOL. 40, NO. 3, 2006 / ENVIRONMENTAL SCIENCE & TECHNOLOGY
9
1055
length, 3.5 mL cuvette (Starna Cells). This method was used for slow reactions (reactions that took 60 min or longer to reach completion). The stopped-flow apparatus consisted of two 5 mL gastight Hamilton syringes (one each for the MnO4- and COC stock solutions) that were connected to a mixing cell with 2 µL of dead volume (Upchurch Scientific, Oak Harbor, WA), which in turn led to a quartz flow-through cuvette (Starna Cells) with a 1 cm path length. A syringe pump (KD Scientific, Holliston, MA) was used to obtain rapid delivery (∼1 s) from the mixing cell to the cuvette. We verified that the static and stopped-flow methods gave similar values of k ′′ using TCE (0.67 ( 0.05 M-1 s-1 for the stopped flow method and 0.46 ( 0.05 to 0.76 ( 0.03 M-1 s-1 for the static method) and cis-DCE (0.69 ( 0.1 M-1 s-1 for the stopped flow method and 0.71 ( 0.06 M-1 s-1 for the static method). Details on which experimental method was used for each reaction are in Table S1. Cuvettes used in both methods were cleaned daily with a solution of 6% hydrogen peroxide and 2% sulfuric acid to remove any MnO2 residue generated during experiments. The cell in which reactions were performed was maintained at 25.0 ( 0.1 °C. Literature Review. Previously reported values of k ′′ for oxidation reactions by MnO4- were gathered from all citable sources available as of December 2004. Values of k ′′ for roughly 115 compounds were found, of which ∼35 were deemed relevant as potential COCs or model compounds for COCs. Some of the literature data were obtained at extreme pH values (ca. 1 and 14) or in nonaqueous solvents, and these values were not used in this study. Literature values of k ′′ that were obtained at a pH between 4 and 8 were used without attempting any a priori correction for pH effects. In this pH range, it is known that k ′′ for MnO4- oxidation of nonionogenic COCs such as TCE is unaffected by pH (3). In general, when a literature source reported k ′′ values over a range of temperatures, only the data from within the range of 20-30 °C were included in our analysis. In a few cases, however, the only data available for an important COC was for temperatures outside this range. In these cases, we did not exclude the value a priori but rather included it with a flag to prevent its use in inappropriate comparisons. The pH and temperature conditions for the literature data we have used as well as our own experimental results are summarized as Supporting Information in Table S2.
Results and Discussion Method Refinement and Validation. As we have discussed previously (19), a factor that complicates the use of absorbance spectrometry to follow oxidation reactions involving MnO4- is that the reduction product, colloidal MnO2, contributes to the absorbance measured at 525 nm (an absorbance maximum for MnO4-) (3, 18, 20, 21). Fortunately, the measured absorbance of colloidal MnO2 suspensions obeys Beer’s law until the particles begin to flocculate, resulting in increased scattering of light by the larger aggregates (18, 20). To determine where Beer’s law was obeyed in our system, full scans (350-700 nm) were taken periodically during the course of reaction between MnO4and the COCs that were studied using the static method. As illustrated in Figure 1A for 1,4-dioxane, a sharp isosbestic point was observed for most of the reaction, which suggests that MnO4- and MnO2 were the only reactants absorbing at this wavelength and that their molar absorptivities were constant (i.e., aggregation was negligible and Beer’s law is obeyed (22)). Only data from before the isosbestic point failed were used in subsequent analysis. For COCs studied using the stopped-flow method, sequential scans were not possible using our apparatus, so we checked for Beer’s law behavior by comparing the absorbance at 525 nm (where both MnO2 and MnO4- absorb) and 418 nm (where only MnO2 absorbs) measured over the course 1056
9
ENVIRONMENTAL SCIENCE & TECHNOLOGY / VOL. 40, NO. 3, 2006
FIGURE 1. Analysis used to identify the Beer’s Law region for (A) 1,4-dioxane, a compound analyzed with the static method and (B) 3-chlorophenol, a compound analyzed with the stopped-flow method. The data used in further analysis were those obtained before the isosbestic point degraded (in the case of 1,4-dioxane) or those obtained in the linear portion of the figure (for 3-chlorophenol). The data for 1,4-dioxane represent 114 h of reaction (12 h between displayed scans), and the data for 3-chlorophenol represent 3 min of reaction (∼4.5 s between data points). Experimental conditions: 0.1 mM KMnO4, pH 7, 50 mM phosphate buffer, 25 °C. of reaction in independent but identically prepared experiments. As illustrated in Figure 1B for 3-chlorophenol, a linear relationship was observed for most of each reaction. Only data from within the linear portion of these plots (between the arrows in Figure 1B) were used in further analysis. After excluding the data where scattering by colloidal MnO2 was significant, we found that most COCs gave data for absorbance at 525 nm versus time that fit pseudo-firstorder kinetics with a modest deviation from this kinetic model as the reaction approached completion (Figure 2A,B). We concluded that the latter effect was due to increasing light absorption by suspended MnO2, based on comparison with previously published spectra for colloidal MnO2 suspensions (20). To obtain pseudo-first-order rate constants (kobs) in a manner that systematically accounts for the absorbance of
FIGURE 2. Kinetics for the reactions of MnO4- with 3-chlorophenol and with 1,4-dioxane: panels A and B show data for the two reactions (note the difference in time scales) with the fit to a simple first-order kinetic model (curve in red) and to the MnO2-corrected model represented by eq 1 (curve in blue). Panels C and D are kobs vs concentration of COC plots; the slope of the regression lines shown was used to determine k′′. Experimental conditions: 0.1 mM KMnO4, pH 7, 50 mM phosphate buffer, 25 °C. colloidal MnO2, we fit all of our kinetic data to a two-term model 525 -kobst 525 ]εMnO2 A525 ) [Cie-kobst ]εMnO - + [Ci - Cie 4
(1)
where A525 is the absorbance at 525 nm, Ci is the initial 525 concentration of MnO4-, εMnO - is the absorptivity of MnO4 4 525 (2.46 cm-1 mM-1), and εMnO is the absorptivity of MnO . We 2 2 525 treated Ci and εMnO - as constants and simultaneously fit kobs 4 525 and εMnO by least squares regression. This treatment allows 2 525 for variability in εMnO , which is expected because the optical 2 properties of the particles will depend on their size distribution, which will be affected by the rate of MnO4- reduction, which will depend on the COC that is being oxidized. The 525 fitted values we obtained for εMnO have a roughly normal 2 distribution (raw data available in Table S1, statistical analysis given in ref 23). Overall, eq 1 fit the experimental data very well, as illustrated in Figure 2A for a fast reaction (3chlorophenol, with a half-life of seconds) and Figure 2B for a slow reaction (1,4-dioxane, with a half-life of days). Further details on the derivation and validation of eq 1 can be found in refs 19 and 23.
The kobs values obtained by fitting eq 1 to experimental data were used to determine second-order rate constants, k ′′. For COCs with low solubility ( 1.5 days), k ′′ was calculated with eq 2
k ′′ ) kobs/[COC]
(2)
For these COCs, the treatment effectively assumes that the order of reaction with respect to the COC is one. The COCs for which k ′′ was determined this way were methyl t-butyl ether (MTBE), toluene, perchloroethene (PCE), picric acid, and 2,4-dinitrophenol. MTBE, PCE, and toluene have been shown by others to be first-order with respect to the COC (3, 5, 13, 21). Picric acid and 2,4-dinitrophenol were assumed to be first-order with respect to COC based on their structural similarity to the mononitrophenols. For COCs with sufficiently high solubilities that experiments could be set up over a significant range of initial COC concentrations ([COC]0), we plotted kobs versus [COC]0, determined that the plots were linear (and therefore approximately first-order in [COC]), then fit the data by linear regression, and assigned the fitted slope to k ′′. Figures 2C and 2D illustrates this analysis for 3-chlorophenol and 1,4-dioxane, respectively. VOL. 40, NO. 3, 2006 / ENVIRONMENTAL SCIENCE & TECHNOLOGY
9
1057
FIGURE 3. Correlation between k ′′ obtained from pseudo-firstorder conditions by either analyzing decreasing concentrations of MnO4- in the presence of an excess amount of COC (this paper) or analyzing decreasing concentrations of COC in the presence of an excess amount of MnO4- (k ′′ data obtained from the literature). Perfect agreement between the two sets of data would follow the dashed line. For the chlorinated ethenes, the literature values of k ′′ used in the figure are the values Huang et al. measured at 25 °C (5) because their experimental conditions most closely matched ours. Multiple literature values of k ′′ for toluene are shown because the only literature value obtained at 25 °C was also obtained with excess COC conditions. For MTBE and 2,4-dichlorophenol, no literature data from 25 °C were available for this comparison, so data for 23 and 16 °C, respectively, were used. All literature values used in this figure are from Table S2. In addition to absorption by MnO2, another factor that can complicate the determination of k ′′ by measuring the disappearance of MnO4- rather than the COC is that MnO4can be consumed by processes other than the primary reaction of interest: for example, the autodecomposition of MnO4- (even in pure water (24)) and reaction with daughter products derived from the initial oxidation of the COC. Under the conditions used in this study, however, we found that autodecomposition of MnO4- was negligible (Supporting Information, Figure S1). We also concluded that the reaction of MnO4- with any daughter products of the COC oxidation was not likely to be significant because the COC concentration was highsat least 5-fold (and often >10-fold) greater than MnO4-sand therefore would be the dominant reactant. Further evidence that these (or other) factors did not significantly bias our measurement of k ′′ is provided by the strong correlation between values of k ′′ obtained from this study and previously published values of k ′′ that were obtained with the opposite experimental design (i.e., analyzing decreasing concentrations of COC in the presence of excess MnO4- (Figure 3)). In Figure 3, perfect agreement between our data and literature data would fall on a line with slope ) 1 and intercept ) 0, which is shown in the figure for comparison. Second-Order Rate Constants. Figure 4 summarizes all values of k ′′ that were measured in this study along with all values of k ′′ that we found in the literature for MnO4oxidation of potential COCs under relevant conditions. Details on the data shown in Figure 4sincluding the conditions used to determine each value of k ′′sare given as Supporting Information (Table S2). 1058
9
ENVIRONMENTAL SCIENCE & TECHNOLOGY / VOL. 40, NO. 3, 2006
Inspection of the data in Figure 4 shows that variability in k ′′ for any particular COC is generally less than 10-fold, while variability among closely related COCs is 2-3 orders of magnitude, and variability among families of unrelated COCs is almost 7 orders of magnitude. The variability in k ′′ for any particular COC is satisfactory considering that the data often are from different laboratories obtained by different methods under slightly different conditions. With respect to the variability in k ′′ within and across families of COCs, a general observation that can be made from Figure 4 is that a number of COCs react with MnO4nearly as fast or faster than the chlorinated ethenes. This comparison is potentially significant because the chlorinated ethenes have been remediated successfully with MnO4- in field-scale applications of ISCO (2, 25); therefore, it follows that the other fast-reacting COCs shown in Figure 4 (e.g., TNT, aldicarb, dichlorvos, many phenols, and some PAHs) are possible candidates for ISCO with MnO4-. Of course, having established that k ′′ is suitably large for these COCs is only the first step. To ensure successful treatment of a COC with MnO4- in field-scale applications, other factors, such as the tendency of the COC to adsorb to soil or its potential to form toxic daughter products, need to be evaluated. Another useful generalization that can be made from Figure 4 concerns COCs that do not react rapidly with MnO4-. The families of COCs with values of k ′′ that are significantly below those of the chlorinated ethenes are the chlorinated methanes and ethanes, simple aromatics such as benzene and toluene, most fuel oxygenates and solvent ethers, and some explosives (most notably RDX). As part of this study, we have filled many of the priority data gaps that existed previously, including those for TNT, 1,4-dioxane, some pesticides, and many phenols. However, because our method requires that COCs have at least moderate solubility (∼0.5 mM), we were not able to fill some data gaps for significant, low-solubility compoundssmost notably pentachlorophenol, dibenzodioxins, polychlorinated biphenyls, and some of the more insoluble pesticides, such as 1,1,1-trichloro-2,2-bis(p-chlorophenyl)ethane (DDT), and dieldrin. Most of these COCs are included in Figure 4seven though no k ′′ data for them are available yetsso that the figure provides a balanced perspective on current coverage of the whole range of potential COCs. Products and Mechanisms. The generally accepted mechanism of oxidation of alkenes with MnO4- is electrophilic attack by MnO4- on the carbon-carbon double bond, with formation of a cyclic hypomanganate diester as a reaction intermediate (2-4, 24, 26). For TCE, kinetic modeling has shown that formation of the cyclic hypomanganate diester is the rate-limiting step (3). Several lines of evidence suggest that this may be true for all the chlorinated ethenes. First, k ′′ decreases as the number of chlorine substituents increases (and correspondingly, the electron density in the double bond decreases due to the withdrawing effect of the chlorine atoms). Qualitatively, Yan and Schwartz made this observation from their data (3), and we have shown that combining data on the chlorinated ethenes from all sources gives a good quantitative correlation between log k ′′ and the total number of chlorines (n ) 46, R2 ) 0.77) (Supporting Information, Figure S2A). Most of the scatter in Figure S2 is due to separation between the trans- and cis-isomers of 1,2dichloroethene. The trans-isomer reacts with MnO4- significantly more rapidly than the cis-isomer, which is typical of steric effects on reactions with cyclic intermediates (3, 26) and therefore is a second line of evidence in support of this mechanism. A third and new line of evidence for the mechanism involving cyclic hypomanganate diester intermediates is provided by Figure S2B. These data show a strong correlation between log k ′′ for COC oxidations by MnO4and log k ′′ for oxidation of the corresponding COCs with
FIGURE 4. Summary of second-order rate constants (k ′′) for compounds in all classes. The size of the data points represents the temperature at which the data were collected, with smaller data points indicating lower temperatures. Most of the temperatures for the literature values vary between 16 to 35 °C, the exception being that the k ′′ reported for benzene was determined at 70 °C. The pH for the literature values of k ′′ ranges from 4.6 to 8.0. Compounds labeled in red text are too insoluble for the method we used. Experimental conditions for the data points from this study: 25 °C, pH 7, and phosphate buffer concentrations between 50 and 100 mM. See Table S2 of the Supporting Information for the data on which this figure is based. ozone (n ) 45, R2 ) 0.85). It is well-established that ozone reacts with unsaturated bonds to form a cyclic intermediate (ozonide) (26), much in the same way that is proposed for MnO4-. The initial products of the reaction of TCE and MnO4have been shown to be carboxylic acids, and they can react further with MnO4- eventually to be completely mineralized to CO2 (4). With respect to the aromatic COCs included in Figure 4, it can be seen that many of the substituted phenols are among the fastest reacting compounds with MnO4-, whereas the BTEX compounds (benzene, toluene, ethylbenzene, and xylene) are among the slowest. This is a notable difference,
as the unsaturated nature of these compounds might make both groups susceptible to oxidation by the cycloaddition mechanism described previously for the chlorinated ethenes. However, it has been suggested that instead of cycloaddition to the aromatic ring, the primary site of MnO4- attack on toluene, ethylbenzene, and the xylenes is at benzylic C-H bonds (15, 21, 27, 28). Evidence to support the hypothesis that MnO4- attacks the benzylic C-H bond during oxidation of benzylic compounds in aqueous solution includes significant kinetic isotope effects (21, 28, 29) and strong correlations to bond dissociation energies of the benzylic C-H bonds (15). Furthermore, the observed products of these VOL. 40, NO. 3, 2006 / ENVIRONMENTAL SCIENCE & TECHNOLOGY
9
1059
compounds are those consistent with cleavage of the benzylic C-H bond, such as the corresponding benzoic acids, aldehydes, alcohols, or ketones (24, 28, 29). With respect to benzene, which has no benzylic C-H bonds, Figure 4 shows that k ′′ is 2-3 orders of magnitude smaller than k ′′ values for the other BTEX compounds (even though the benzene data are for higher temperatures). Presumably, MnO4oxidizes benzene by attack on the aromatic ring since there are no benzylic C-H bonds available (28). Although the pathway of reaction between MnO4- and BTEX compounds is fairly well-characterized, the exact mechanism by which these oxidations occur is still a subject of research. Rudakov (28, 29) suggested that the mechanism in aqueous solution involves hydrogen atom abstraction based on product analysis and the strong correlation of the activation energies for these reactions to those for reactions with another radical, •CCl3. Mayer and Gardner (21, 27) also proposed the mechanism of hydrogen atom transfer for the oxidation of toluene, but only when the reaction occurs in organic solvents. For the same reaction in aqueous solution, they suggested that the mechanism involves hydride transfer to one of the oxygen atoms of MnO4-, in part because the rate of oxidation of toluene was found to be very different in the two types of solvents. In the case of the phenols, there have not been any detailed mechanistic studies under environmentally relevant conditions. Other than the cresols, none of the substituted phenols studied here have a benzylic C-H bond, indicating that some of these phenols must be oxidized by a different mechanism than that which applies to toluene, ethylbenzene, and the xylenes. One mechanism that has been suggested for oxidation of phenols by MnO4- is single electron transfer (30, 31). Both quinones (31, 32) and products suggesting ring cleavage (33) have been seen in reactions of MnO4- with different substituted phenols. Correlation analysis using our data might aid in elucidating the mechanism of oxidation of phenols by MnO4-, much as the correlation of reactivity to bond dissociation energies of the benzylic C-H bonds provided support for the benzylic C-H bond as the site of oxidation of BTEX compounds. However, because the deprotonated form of phenols generally reacts with oxidants more rapidly than the protonated form (34, 35), the effect of pH on k ′′ must be taken into account during the process of correlation analysis (36, 37). To this end, we have measured k ′′ versus pH for MnO4oxidation of eight substituted phenols (4-nitrophenol, 2-chlorophenol, 3-chlorophenol, 4-chlorophenol, 2,3-dichlorophenol, 2,4-dichlorophenol, 2,6-dichlorophenol, and 2,4,6trichlorophenol). The results (not shown) do not appear to conform to the simple mixing model that has been used to describe the effect of pH on the kinetics of phenols reacting with other oxidants (e.g., ozone, chlorine dioxide, singlet oxygen, and manganese dioxide) (34-37). These results require further study before a full correlation analysis on the data for MnO4- oxidation of phenols would be appropriate. The mechanism by which MnO4- oxidizes alcohols is believed to be abstraction of the hydrogen atom or hydride from the R-carbon of the anion of the alcohol (24). Thus, although oxidation of primary and secondary alcohols by MnO4- is favored at high pH, oxidation occurs at appreciable rates near neutral pH, and the products formed are aldehydes from primary alcohols and ketones from the secondary alcohols. Tertiary alcohols are relatively unreactive with MnO4- at all pH values (24, 33). The hydrogen/hydride abstraction mechanism might also apply to ethers (38), and our data show the pattern of reactivity that would be expected if this were the case: TAME and MTBE, both tertiary ethers, react more slowly with MnO4- than diisopropyl ether, a secondary ether. Perhaps due to its cyclic nature, 1,4-dioxane is the most stable to oxidation by MnO4-. 1060
9
ENVIRONMENTAL SCIENCE & TECHNOLOGY / VOL. 40, NO. 3, 2006
There does not appear to have been any detailed mechanistic study of the oxidation of ethanes and methanes by MnO4-, undoubtedly because these reactions are too slow to be of much use in any application. However, our data suggest that chloroform is somewhat exceptional, as it reacts with MnO4- at rates that are easily observed (Figure 4). Such distinctive reactivity of chloroform, relative to the other chlorinated methanes, is fairly typical and arises due to the relatively facile deprotonation of chloroform and intermediacy of the trichloromethyl anion in a number of mechanisms for further reaction (39). There are three classes of COCs (PAHs, explosives, and pesticides) for which kinetic data are included in Figure 4, but a discussion of reaction products and mechanisms is not provided here. With respect to the PAHs, a discussion is provided in ref 15 to which the results reported here add little. With respect to the explosives and pesticides, it is tempting to speculate on their relative reactivities in terms of competition among the major oxidation mechanisms discussed previously (electrophilic attack on double bonds, abstraction of benzylic hydrogens, etc.), but the prevalence of moieties whose reactivity with permanganate have not been thoroughly studied (e.g., nitro aromatics, nitroamines, carbamoyloximes, and organophosphate esters) would make further discussion of this highly uncertain. Certainly, further study of the oxidation of these COCs by MnO4- is warranted since Figure 4 shows that some of their rates of reaction are sufficiently fast to be of potential value in groundwater remediation.
Acknowledgments This work was supported by the Strategic Environmental Research and Development Program (SERDP, Project CU1289). This report has not been subject to review by the agency and therefore does not necessarily reflect their views, and no official endorsement should be inferred. We thank Eric Hood (GeoSyntec) and Neil Thomson (U. Waterloo) for their thoughtful review of this manuscript.
Supporting Information Available Two tables (giving details on experimental conditions and results) and two figures (giving further validation of our results). This material is available free of charge via the Internet at http://pubs.acs.org.
Literature Cited (1) Interstate Technology and Regulatory Council (ITRC). Technical and Regulatory Guidance for In Situ Chemical Oxidation of Contaminated Soil and Groundwater, 2nd ed.; 2005. http:// www.itrcweb.org. (2) Siegrist, R. L.; Urynowicz, M. A.; West, O. A.; Crimi, M. L.; Lowe, K. S. Principles and Practices of In Situ Chemical Oxidation Using Permanganate; Battelle Press: Columbus, OH, 2001. (3) Yan, Y. E.; Schwartz, F. W. Oxidative degradation and kinetics of chlorinated ethylenes by potassium permanganate. J. Contam. Hydrol. 1999, 37, 343-365. (4) Yan, Y. E.; Schwartz, F. W. Kinetics and mechanisms for TCE oxidation by permanganate. Environ. Sci. Technol. 2000, 34, 2535-2541. (5) Huang, K. C.; Hoag, G. E.; Chheda, P.; Woody, B. A.; Dobbs, G. M. Oxidation of chlorinated ethenes by potassium permanganate: A kinetics study. J. Hazard. Mater. 2001, 87, 155-169. (6) Hoigne´, J. Formulation and calibration of environmental reaction kinetics: Oxidations by aqueous photooxidants as an example. In Aquatic Chemical Kinetics: Reaction Rates of Processes in Natural Waters; Stumm, W., Ed.; Wiley-Interscience: New York, 1990; pp 43-70. (7) Tratnyek, P. G.; Macalady, D. L. Oxidation-reduction reactions in the aquatic environment. In Handbook of Property Estimation Methods for Chemicals: Environmental and Health Sciences; Mackay, D., Boethling, R. S., Eds.; Lewis: Boca Raton, FL, 2000; pp 383-415.
(8) Vella, P. A.; Veronda, B. Oxidation of trichloroethylene: A comparison of potassium permanganate and Fenton’s reagent. In 3rd International Symposium, Chemical Oxidation: Technology for the Nineties, Vanderbilt University: Nashville, TN, February 17-19, 1993; Technomic: Lancaster, UK, 1993; Vol. 3, pp 62-73. (9) Tratnyek, P. G.; Johnson, T. L.; Warner, S. D.; Clarke, H. S.; Baker, J. A. In situ treatment of organics by sequential reduction and oxidation. In Proceedings of the First International Conference on Remediation of Chlorinated and Recalcitrant Compounds, May 18-21, 1998, Monterey, CA; Battelle Press: Columbus, OH, 1998; Vol. 1 (5), pp 371-376. (10) Hood, E. D.; Thomson, N. R.; Grossi, D.; Farquhar, G. J. Experimental determination of the kinetic rate law for the oxidation of perchloroethylene by potassium permanganate. Chemosphere 2000, 40, 1383-1388. (11) Poulson, S. R.; Naraoka, H. Carbon isotope fractionation during permanganate oxidation of chlorinated ethylenes (cDCE, TCE, PCE). Environ. Sci. Technol. 2002, 36, 3270-3274. (12) Hunkeler, D.; Aravena, R.; Parker, B. L.; Cherry, J. A.; Diao, X. Monitoring oxidation of chlorinated ethenes by permanganate in groundwater using stable isotopes: Laboratory and field studies. Environ. Sci. Technol. 2003, 37, 798-804. (13) Damm, J. H.; Hardacre, C.; Kalin, R. M.; Walsh, K. P. Kinetics of the oxidation of methyl t-butyl ether (MTBE) by potassium permanganate. Water Res. 2002, 36, 3638-3646. (14) Adam, M. L.; Comfort, S. D.; Morley, M. C.; Snow, D. D. Remediating RDX-contaminated groundwater with permanganate: Laboratory investigations for the Pantex perched aquifer. J. Environ. Qual. 2004, 33, 2165-2173. (15) Forsey, S. P. In Situ Chemical Oxidation of Creosote/Coal Tar Residuals: Experimental and Numerical Investigation. Ph.D. Thesis, University of Waterloo, 2004. (16) Tollefsrud, E.; Schreier, C. G. Effectiveness of chemical oxidation to remove organochlorine pesticides from soil. In Proceedings of the Third International Conference on Remediation of Chlorinated and Recalcitrant Compounds, Monterey, CA; Battelle Press: Columbus, OH, 2002; Paper 2C-16. (17) Vella, P. A.; Munder, J. A. Toxic Pollutant Destruction. In Emerging Technologies in Hazardous Waste Management III; American Chemical Society: Washington, DC, 1993; ACS Symposium Series 518, pp 85-105. (18) Mata-Perez, F.; Perez-Benito, J. F. Identification of the product from the reduction of permanganate ion by trimethylamine in aqueous phosphate buffers. Can. J. Chem. 1985, 63, 988-992. (19) Waldemer, R. H.; Tratnyek, P. G. The efficient determination of rate constants for oxidations by permanganate. In Proceedings of the Fourth International Conference on Remediation of Chlorinated and Recalcitrant Compounds, May 24-27, 2004, Monterey, CA; Battelle Press: Columbus, OH, 2004; Paper 2A09. (20) Perez-Benito, J. F.; Arias, C. Occurrence of colloidal manganese dioxide in permanganate reactions. J. Colloid Interface Sci. 1992, 152, 70-84. (21) Gardner, K. A. Permanganate oxidations of aromatic hydrocarbons in aqueous and organic solution. Ph.D. Thesis, University of Washington, 1996. (22) Morrey, J. R. Isosbestic points in absorbance spectra. J. Phys. Chem. 1963, 57, 1569. (23) Waldemer, R. H. Determination of the Rate of Contaminant Oxidations by Permanganate: Implications for In Situ Chemical
(24)
(25)
(26)
(27)
(28)
(29)
(30) (31)
(32)
(33)
(34)
(35)
(36)
(37)
(38) (39)
Oxidation (ISCO). M.S. Thesis, OGI School of Science and Engineering, Oregon Health and Science University, 2004. Stewart, R. Oxidation by permanganate. In Oxidation in Organic Chemistry; Wiberg, K. B., Ed.; Academic Press: New York, 1965; Vol. 5A, pp 1-68. Schnarr, M.; Truax, C.; Farquhar, G.; Hood, E.; Gonullu, T.; Stickney, B. Laboratory and controlled field experiments using potassium permanganate to remediate trichloroethylene and perchloroethylene DNAPLs in porous media. J. Contam. Hydrol. 1998, 29, 205-224. Freeman, F. Possible criteria for distinguishing between cyclic and acyclic activated complexes and among cyclic activated complexes in addition reactions. Chem. Rev. 1975, 75, 439490. Gardner, K. A.; Mayer, J. M. Understanding C-H bond oxidations: H• and H- transfer in the oxidation of toluene by permanganate. Science 1995, 269, 1849-1851. Rudakov, E. S.; Lobachev, V. L. The kinetics, kinetic isotope effects, and substrate selectivity of alkylbenzene oxidation in aqueous permanganate solutions. 2. Reaction with HMnO4. Kinet. Catal. 1994, 35, 180-187. Rudakov, E. S.; Lobachev, V. L. The first step of oxidation of alkylbenzenes by permanganates in acidic aqueous solutions. Russ. Chem. Bull. 2000, 49, 761-777. Nonhebel, D. C.; Tedder, J. M.; Walton, J. C. Radicals; Cambridge University Press: Cambridge, 1979. Lee, D. G.; Sebastian, C. F. The oxidation of phenol and chlorophenols by alkaline permanganate. Can. J. Chem. 1981, 59, 2776-2779. Radhakrishnamurti, P. S.; Sahu, S. N. Kinetics and mechanism of oxidation of nitrophenols and nitroanilines by potassium permanganate. J. Indian Chem. Soc. 1976, 53, 1154-1155. Walton, J.; Labine, P.; Reidies, A. The chemistry of permanganate in degradative oxidations. In 1st International Symposium, Chemical Oxidation: Technology for the Nineties, Vanderbilt University, Nashville, TN, February 20-22, 1991, Nashville, TN; Technomic: Lancaster, UK, 1991; Vol. 1, pp 205-221. Tratnyek, P. G.; Hoigne´, J. Kinetics of reactions of chlorine dioxide (OClO) in water. II. Quantitative structure-activity relationships for phenolic compounds. Water Res. 1994, 28, 57-66. Tratnyek, P. G.; Hoigne´, J. Oxidation of substituted phenols in the environment: a QSAR analysis of rate constants for reaction with singlet oxygen. Environ. Sci. Technol. 1991, 25, 1596-1604. Canonica, S.; Tratnyek, P. G. Quantitative structure-activity relationships (QSARs) for oxidation reactions of organic chemicals in water. Environ. Toxicol. Chem. 2003, 22, 1743-1754. Tratnyek, P. G. Correlation analysis of the environmental reactivity of organic substances. In Perspectives in Environmental Chemistry; Macalady, D. L., Ed.; Oxford: New York, 1998; pp 167-194. Barter, R. M.; Littler, J. S. Hydride ion transfer in oxidations of alcohols and ethers. J. Chem. Soc. B 1967, 205-210. Larson, R. A.; Weber, E. J. Reaction Mechanisms in Environmental Organic Chemistry; Lewis: Chelsea, MI, 1994.
Received for review July 8, 2005. Revised manuscript received November 3, 2005. Accepted November 9, 2005. ES051330S
VOL. 40, NO. 3, 2006 / ENVIRONMENTAL SCIENCE & TECHNOLOGY
9
1061
