Environ. Sci. Technol. 2000, 34, 2535-2541
Kinetics and Mechanisms for TCE Oxidation by Permanganate Y . E U G E N E Y A N * ,† A N D FRANKLIN W. SCHWARTZ Department of Geological Sciences, The Ohio State University, Columbus, Ohio 43210
The oxidation of trichloroethylene (TCE) by permanganate proceeds in three sequential reaction steps. In the initial step, a cyclic hypomanganate ester is formed via an activated organometallic complex. The activation parameters for this step were determined to be Ea ) 41.46 kJ/mol, ∆Hq ) 39 kJ/mol, and ∆Sq ) -14 J/mol. The initial reaction is a rate-limiting step (second-order rate constant k1p ) 0.65-0.68 M-1 s-1 at 21 °C) and independent of pH. In the second step, the decomposition of the cyclic ester with complete chlorine liberation proceeds quickly via various reaction pathways to form four carboxylic acids. Approximately 77% of the TCE was transformed to formic acid at pH 4, while 95-97% of the TCE was transformed to oxalic and glyoxylic acids at pH values of 6-8. Kinetic data suggest that the decomposition rate of the cyclic ester is at least 100 times higher than its formation rate. In the final step, all carboxylic acids are oxidized by permanganate to the final product, CO2. Second-order rate constants of k3ap ) 0.075-0.35 M-1 s-1, k3bp ) 0.130.37 M-1 s-1, and k3cp ) 0.073-0.11 M-1 s-1 over a pH range of 4-8 at 21 °C were estimated for oxidation of formic, glyoxylic, and oxalic acids, respectively. The oxidation rate of carboxylic acids and accumulation rate of CO2 increase with decreasing pH. The kinetic model that was developed, formulated, and solved analytically on the basis of the understanding of various processes is consistent with results obtained in the kinetic experiments.
Introduction There is considerable interest in the in-situ oxidative degradation of chlorinated ethenes with a metal-oxo reagent (1-5). In terms of remediating contaminated groundwater, this method is attractive due to the rapid degradation of chlorinated ethenes, the nonreactive nature of the reagent with carbonate and bicarbonate, the ease of field implementation, and the relatively low costs (1, 5, 6). Degradation kinetics for five chlorinated ethenes, including PCE, TCE, and three isomers of DCE were reported by ref 7. TCE degradation was postulated to involve the following sequential reactions: k1
k2
C2HCl3 + MnO4- 98 I 98 γCA + MnO2 + 3Clk3(MnO4-)
γCA 98 ηCO2
(1) (2)
where I is a cyclic complex, CA are various carboxylic acids, * Corresponding author telephone: (630)252-6322; fax: (630)252-5747; e-mail:
[email protected]. † Present address: Argonne National Laboratory, 9700 S. Cass Ave., Argonne, IL 60439. 10.1021/es991279q CCC: $19.00 Published on Web 05/18/2000
2000 American Chemical Society
γ and η are stoichiometric coefficients, and k1, k2, and k3 are rate constants. In the first reaction, TCE disappearance was found to be independent of pH. Complete chlorine liberation was thought to occur in the second step of eq 1. However, the second and third reactions have not been studied in detail. This study was undertaken to understand fully the kinetics and mechanisms for TCE oxidation by permanganate and more specifically (i) to identify reaction products, (ii) to elucidate reaction pathways and their pH dependence, (iii) to develop a kinetic model, and (iv) to determine the rate constants for all major reactions in the KMnO4-TCE-H2O system.
Experimental Methods Materials. Radiolabeled [1,2-14C]TCE (3.1 mCi/mmol) was purchased from Sigma Chemical Co. (St. Louis, MO) and diluted with methanol in a 10-mL glass volumetric flask and stored at 4 °C in the dark. Carboxylic acids were obtained from Fluka (Buchs, Switzerland) with purity 99+% and used as received. All other chemicals were of the highest purity available and were described in the previous study (7). Product Studies and Kinetic Experiments. With TCE oxidation by permanganate, carboxylic acids and carbon dioxide would be expected as major intermediate and final products (7). Two types of kinetic tests were designed to identify the carboxylic acids and to trace CO2 formation. Each type consists of three experiments conducted at pH values of 4, 6, and 8 at 21 °C for about 8 h, and each experiment has one set of vials with two replicates and one control vial containing only TCE aqueous solution. For investigating carboxylic acid distributions over time, a first kinetic test was conducted in 50-mL glass vials with PTFE/silicone septum-lined screw-top caps. Each vial was filled with a 50-ml TCE solution of 0.1 mM. The experiment began by injecting 5 mL of 6.3 mM KMnO4 through the septum. A second needle enabled an equal volume of TCE solution to be displaced. Vials were placed on a vibratory shaker until sampled. During sampling for each experiment, two vials were taken. A 1-mL aliquot of reaction solution from each vial was transferred to a volumetric flask with appropriate dilution following immediate TCE analysis. To quench the reaction, 1 mL of thiosulfate from the stock solution was added to each vial. The quenched solution was centrifuged at 3000 rpm for 20 min and filtered by a 0.2-µm filter to separate the precipitated manganese dioxide. After filtration, the solutions were withdrawn for carboxylic acid analysis. To trace the CO2 production, a second kinetic test was undertaken with radiolabeled 14C TCE. To maintain the same initial conditions as the first kinetic test, experiments began with the addition of 0.1 mL of permanganate stock solution to a 3-mL test tube filled with 2.9 mL of 0.093 mM TCE solution. Each experiment consists of one set of test tubes with duplicates. At predetermined interval, the reaction was quenched by adding 0.1 mL of thiosulfate from the stock solution. The test solution was then analyzed for CO2. In addition, five sets of duplicated kinetic experiments were performed at various temperatures from 5 to 25 °C to investigate how temperature affects the rate of TCE loss and to determine activation parameters for the initial oxidation. This temperature range covers variation commonly observed for natural groundwaters. All experiments were conducted by reacting 0.06 mM TCE with an excess of MnO4- (1 mM) at pH 7 and ionic strength of 0.05 M. Chemical Analysis. The organic acids were analyzed on a Waters high-performance liquid chromatography (HPLC) VOL. 34, NO. 12, 2000 / ENVIRONMENTAL SCIENCE & TECHNOLOGY
9
2535
fitted with a Bio-Rad Aminex HPX-87H column (300 × 7.8 mm i.d.). The column was operated at room temperature using degassed dilute sulfuric acid (4 mM) as eluent at a flow rate of 0.6 mL/min. The column effluents were monitored by an UV spectrophotometric detector (Waters 486) at 210 nm. The products were identified and quantified by comparison with external standards. CO2 was determined by retaining CO2 under basic conditions and stripping CO2 under acidic conditions. This method was based on and adapted from ref 8. TCE concentrations were measured as described in ref 7.
Results and Discussion Products. In the experiments for identifying products, the reaction was initiated with 0.09 mM TCE reacting with 0.63 mM permanganate in a phosphate-buffered solution having an ionic strength of 0.05 M at pH 4, 6, and 8, respectively. Four carboxylic acids (formic, oxalic, glyoxylic, and glycolic acids) were identified in the system as intermediate products. Either formic acid or oxalic acid predominated, depending on pH. At 1 h after the reaction started, a maximum of 43% of the initial TCE was converted to either formic acid or oxalic acid, and up to 25% of the TCE was transformed to glyoxylic acid. Glycolic acid was measured in very small quantities ( 10. The large k2/k1 ratio supports the idea that the decomposition of cyclic ester is much faster than its formation, as suggested by refs 7 and 9. When k2 g 102 k1, r 2 does not show any significant improvement (the increase of r 2 is less than 0.001 from k2 ) 102 k1 to k2 ) 105 k1). The rate constant ratios (k2a/k2, k2b/k2, and k2c/k2) for the second step and rate constants (k3a, k3b, and k3c) for the third step remain constant with increasing k2/k1. Thus, the TCE oxidation process involving TCE decomposition and carboxylic acid formation and decomposition is significantly controlled by the initial step (k1), a rate-limiting step, regardless of the rate (k2) for the second step when k2 g 102 k1. Thus, k2 ) 102 k1 was assumed in all curve fittings. The ratios k2a/k2, k2b/k2, and k2c/k2 represent the fraction of the total decomposition rate of the cyclic ester along each pathway and quantify the contributions from the various pathways in Figure 1. Their values (Table 2) suggest that the pathway leading to formic acid or oxalic acid is dominant in the decomposition process. The pseudo-first-order rate constants, k3a, k3b, and k3c, describe the oxidation of formic, glyoxylic, and oxalic acids 2538
9
ENVIRONMENTAL SCIENCE & TECHNOLOGY / VOL. 34, NO. 12, 2000
in the final step. Compared to the other steps in TCE oxidation, the oxidation rates for the carboxylic acids are slower than any of the rates in the previous steps. The secondorder rate constants, k3ap, k3bp, and k3cp, can be calculated with eqs 12-14 (Table 1). In this final step, the oxidation rate of carboxylic acids (k3) might be underestimated by eqs 1214 because of relatively low initial concentration of MnO4used in the experiments. However, the estimated values for k3ap, k3bp, and k3cp are comparable with results from other studies. For example, k3ap for the oxidation of formic acid ranging from 0.075 to 0.35 M-1 s-1 is consistent with the range of 0.003-0.25 M-1 s-1 obtained in ref 21. The estimated k3cp in the range of 0.073-0.11 M-1 s-1 at pH values of 4-8 for oxalic acid is also reasonable if it is compared to k ) 0.47-0.58 M-1 s-1, which is obtained at relatively low pH (pH 1.9) without the initial presence of manganese(II) species (19). The rate constant k3bp for glyoxylic acid was estimated to be 0.13-0.37 M-1 s-1. On the basis of estimated rate constants, the oxidation rate of formic and glyoxylic acids is faster than oxalic acid over a pH range of 4-8. Using all the estimated rate constants, the calculated accumulation of CO2 is plotted in Figure 2. The results for
CO2 are in general agreement with the experimental data, except at pH 8 in which the predicted CO2 is higher than the observed at later time. This overestimation may result from the adsorption of carboxylic acids on MnO2, which is more readily precipitated at pH 8 than at other lower pH conditions. Because the precipitates were left in the filtrates in our experiments for acid analysis, acid adsorption on MnO2 could cause a mass deficiency, which occurs at late time (Figure 2c). Consequently, the rate constant k3 at pH 8 may be slightly overestimated. Effects of pH and Temperature. During the first step in TCE oxidation, the rate constant k1 does not show strong correlation with pH. Our previous study (7) indicated that (i) the variation of k1 at different pH values is small and within the experimental error and (ii) the second-order rate constant k1p is equal to 0.67 ( 0.03 M-1 s-1 over the pH range of 4-8 at 21 °C. Thus, the rate of TCE disappearance is independent of pH. However, the fate of cyclic hypomanganate ester and the product distribution are strongly dependent on pH. Kinetic data demonstrate two major decomposition pathways for cyclic hypomanganate ester. At pH 4, the transformation of cyclic ester to formic acid is overwhelmingly dominant (k2a/k2 ) 0.77). With this relatively acidic condition, the majority of the cyclic ester has been cleaved at the carboncarbon double bond to form two formic acids 5 via 4 (Figure 1). At pH 6-8, decomposition favors the formation of oxalic and glyoxylic acids ((k2b + k2c)/k2 ) 0.95-0.97). A different reaction pathway has been suggested for this higher pH range; most of cyclic ester transforms to acyclic ester 6 and 7 followed by hydrolysis to form oxalic and glyoxylic acids, 13 and 12, via 11. The rate constants (k3a, k3b, and k3c) for oxidation of carboxylic acids are strongly correlated with pH over a range of 4-8 (Figure 3). With decreasing pH, the oxidation rates (k3) increase for all three carboxylic acids. This pH effect on k3 results in the relatively rapid formation of CO2 at low pH values as compared to higher pH values. The temperature effect on TCE disappearance rate by MnO4 was investigated by five sets of duplicated kinetic experiments at temperatures varying from 5 to 25 °C. The estimated logarithm of the second-order rate constant k1p from the experiments is plotted against the reciprocal of temperature in Figure 4. The dependence of k1p on temperature, as expected, follows the Arrhenius equation. The activation energy (Ea) and preexponential (A) were calculated and are listed in Figure 4. These values can provide a rate constant, k1p, at any temperature relevant to groundwater. On the basis of Ea ) 41.46 kJ/mol and A ) 1.678 × 107 M-1 s-1, activation parameters ∆Hq and ∆Sq can be calculated. The low activation enthalpy (∆Hq ) 39 kJ/mol) obtained here is comparable to the results from many other studies in alkene oxidation by permanganate. This result indicates that initial step of TCE oxidation may proceed via a similar transition state as other alkene oxidation. The negative entropy (∆Sq ) -14 J/mol) is consistent with a bimolecular reaction in which the transient complex may be highly structured and possibly solvated.
Proposed Mechanism for TCE Oxidation
FIGURE 2. Product distribution with time in TCE oxidation at pH 4 (a), pH 6 (b), and pH 8 (c). Lines for TCE and carboxylic acids represent best fits using eqs 16 and17, whereas the line for CO2 is calculated by eq 18 using rate constants from the previous curve fitting.
The fact that the loss of TCE is independent of pH but that the nature and distribution of products are highly dependent on pH strongly supports the existence of a short-lived intermediate in the initial stage of oxidation. There has been more than a century of study on alkene oxidation by permanganate. Cyclic hypomanganate ester is generally thought to form as an intermediate, on the basis of cisaddition products, cis-diol/ketol (22, 13), and oxygen transfer from MnO4- to diol observed in 18O-labeling experiments (9). In recent studies, however, a more comprehensive reaction mechanism has been postulated with the introducVOL. 34, NO. 12, 2000 / ENVIRONMENTAL SCIENCE & TECHNOLOGY
9
2539
TABLE 2. Pseudo-First-Order Rate Constants Obtained for TCE Transformation Pathways Using Eqs 16 and 17a pH
k1 (10-4 s-1)
k2a/k2
k3a (10-4 s-1)
r2
k2b/k2
k3b (10-4 s-1)
r2
k2c/k2
k3c (10-4 s-1)
r2
4 6 8
4.30 4.11 4.11
0.77 0.02 0.04
2.19 1.53 0.47
0.98 0.91 0.97
0.03 0.32 0.42
2.34 1.05 0.80
0.98 0.99 0.99
0.20 0.63 0.55
0.70 0.64 0.46
0.95 0.95 0.95
a
Estimation was based on experiments conducted with an initial concentration of 0.63 mM MnO4-; k2 ) 102 k1.
FIGURE 3. Dependence of the pseudo-first-order rate constants (k3a, k3b, and k3c) on pH for oxidation of carboxylic acids. Kinetic data were obtained by reacting 0.09 mM TCE with 0.63 mM MnO4in a phosphate-buffered aqueous solution at ionic strength 0.05 M. tion of an organometallic complex between the reactants and the intermediate (14). In the classical view, as summarized in ref 10, the formation of cyclic ester involves a 3+2 electrocyclic addition of permanganate ion to the carbon-carbon double bond
(19)
where the superscript q indicates an activated complex at transition state. The modern ideas on alkene oxidation by permanganate have been advanced by refs 14 and 23. The initial reaction proceeds with 2+2 insertion of an alkene π-bond directly into a metal-oxo bond of manganese, forming a transient organometallic complex (14) followed by rearrangement to the cyclic ester (12, 15). This modern theory was confirmed by showing that relatively lower activation energy was required for the 2+2 reaction pathway as compared to 3+2 (23). With the modern theory, TCE is more likely attacked by electropositive metal (manganese) via a transition state:
(20)
This reaction mechanism more reasonably explains why permanganate ion with electron-rich oxygen termini could attack the carbon-carbon double bond electrophilically. The attack is controlled by orbital overlap via delocalized charge2540
9
ENVIRONMENTAL SCIENCE & TECHNOLOGY / VOL. 34, NO. 12, 2000
FIGURE 4. Arrhenius plot for the oxidation of TCE (0.06 mM) by potassium permanganate (1 mM) in a phosphate-buffered solution at ionic strength 0.05 M and pH 7.1. Error bars are standard deviation for each estimated second-order rate constant k1p. transfer interaction rather than net charge (24, 12). The interaction is initiated by a “slipping” of the metal toward one end of the alkene double bond leaving the other end more electropositive to reaction with the oxygen in permanganate. The proposed slipping motion along the carboncarbon double bond in the transition state apparently favors trans among dichloroethene (DCE) isomers but is relatively difficult to be induced when ethene is fully substituted by four chlorines (a strong electron-withdrawing group), such as tetrachloroethene (PCE). This mechanism is supported by our previous results that showed trans-DCE with much greater reactivity than cis-DCE and PCE with a reaction rate more than 1 order of magnitude smaller than TCE, cis-DCE, and 1,1-DCE (7). The small enthalpy ∆Hq (39 kJ/mol) and negative entropy ∆Sq (-14 J/mol) are consistent with the observations of the oxidation of other alkenic substrates by permanganate under various conditions, even though their products are different. The results suggest that TCE oxidation proceeds via similar transition states but that these transformation products are subjected to various post-transition-state reactions, which are highly dependent on reaction conditions. The existence of various reaction pathways is evidenced by the formation of different intermediate products as a function of pH. At pH 4, the decomposition of cyclic ester 2 involves both oxidative hydrolysis (followed by carboncarbon double bond cleavage) to form formic acid and hydrolysis (with subsequent reactions) to form oxalic acid (Figure 1). The calculated ratios of rate constants (k2a/k2 and k2c/k2) indicate that the former reaction pathway is predominant (77% of total production) while the latter one is much less important (20%). With increasing pH, almost all the cyclic esters are hydrolyzed via 6. Products are further decomposed via oxidative hydrolysis, hydrolysis, and electron
transfer to form acyl halides, which are very reactive and easily hydrolyzed to glyoxylic and oxalic acids (95-97%). Decomposition of acyclic ester 7 via 8 to form glycolic acid is insignificant. The estimated k2/k1 ratio from our kinetic data also supports the postulate that the decomposition of cyclic ester is much faster than its formation. The series of reactions, during the fast ester decomposition, liberate almost all the chlorine substitutes on TCE (7). In the final step, the oxidation of carboxylic acids is relatively slow as compared to the previous steps. The oxidation may involve hydride abstraction by permanganate ion (25) and carbon-carbon bond cleavage. In the aqueous solution, carboxylic acids can be dissociated:
HCOOH S HCOO- + H+
pKa ) 3.74
OHC-COOH S HOC-COO- + H+ pKb ) 3.34
(21) (22)
HOOC-COOH S HOOC-COO- + H+ pKc1 ) 1.27 (23) HOOC-COO- S -OOC-COO- + H+ pKc2 ) 4.28 (24) where Ka, Kb, Kc1, and Kc2 are ionization constants for formic, glyoxylic, and oxalic acids, respectively. If all species are considered to react independently with permanganate, eqs 7-9 can be further developed to eqs 25-27 (Table 1). As compared to eqs 25-27 with eqs 7-9, the second-order rate constants for oxidation of formic, glyoxylic, and oxalic acids can be simplified as
k3cp )
k3ap )
k3ap1 + [H ] + k3ap2 Ka
(28)
k3bp )
k3bp1 + [H ] + k3bp2 Kb
(29)
k3cp1[H+]2 + k3cp2K[H+] + k3cp3KclKc2 KclKc2 + Kc1[H+]
(30)
when pH is much higher than pKa, pKb, and pKc1, resulting in [H+] , Ka, [H+] , Kb, and [H+]2 , Kc1Kc2 and Kc1[H]+. The relationship between logarithms of second-order rate constants with pH shown in Figure 3 is consistent with the mechanism described by eqs 28-30. Because of the relatively
high oxidation rate of carboxylic acids at low pH, as illustrated in Figure 2, CO2 is accumulated more rapidly with decreasing pH.
Acknowledgments We thank Drs. O. H. Tuovinen, Y. P. Chin, and S. Traina for their advice and assistance. This material is based upon work supported by the U.S. Department of Energy under Grant DE-FG07-96ER14735.
Literature Cited (1) Vella, P. A.; Veronda, B. Proceedings of the 2nd International Symposium on Chemical Oxidation: Technologies for the Nineties, Nashville, TN; Technomic: Lancaster, PA, 1992; pp 62-73. (2) Truax, C. M.A.Sc. Thesis, University of Waterloo, Canada, 1993. (3) Yan, Y. E.; Schwartz, F. W. EOS, Trans. Am. Geophys. Union 1995, 76 (46), F247. (4) Gates, D. D.; Siegrist, R. L.; Cline, S. R. J. Environ. Eng. 1995, 121, 582-588. (5) Schnarr, M.; Truax, C.; Farquhar, G.; Hood E.; Gonullu, T.; Stickney, B. J. Contam. Hydrol. 1998, 29, 205-224. (6) Vella, P. A. Proceedings of 3rd International Conference on Advanced Oxidation Technologies for Water and Air Remediation, Cincinnati, OH, October, 1996; pp 119-120. (7) Yan, Y. E.; Schwartz, F. W. J. Contam. Hydrol. 1999, 37, 343365. (8) Kriegman-King, M. R.; Reinhard, M. Environ. Sci. Technol. 1992, 26, 2198-2206. (9) Wiberg, K. B.; Saegebarth, K. A. J. Am. Chem. Soc. 1957, 79, 2822-2824. (10) Freeman, F. Rev. React. Species Chem. React. 1976, 1, 179-226. (11) Lee, D. G.; Chen, T. J. Am. Chem. Soc. 1989, 111, 7534-7538. (12) Lee, D. G.; Brown, K. C. J. Am. Chem. Soc. 1982, 104, 5076-5081. (13) Wolfe, S.; Ingold, C. F. J. Am. Chem. Soc. 1981, 103, 938-939. (14) Sharpless, K. B.; Teranishi, A. Y.; Backvall, J. E. J. Am. Chem. Soc. 1977, 99, 3120-3128. (15) Freeman, F.; Kappos, J. C. J. Am. Chem. Soc. 1985, 107, 66286633. (16) Kivinen, A. In The Chemistry of Acyl Halides; Patai, S., Ed.; Wiley: New York, 1972; pp 177-230. (17) Berka, A.; Koreckova, J. Anal. Lett. 1973, 6, 1113-1123. (18) Rodriguez, J.; Sanchez Burgos, F. Ion (Madrid) 1975, 35, 241245. (19) Mahmood, A. J.; Begum, M. Dacca Univ. Stud., Part B 1975, 23, 51-64. (20) Huang, K. C.; Hoag, G. E.; Chheda, P.; Woody, B. A.; Dobbs, G. M. Environ. Eng. Sci. 1999, 16, 265-274. (21) Perez-Benito, J. F.; Arias, C.; Brillas, E. Int. J. Chem. Kinet. 1990, 22, 261-287. (22) Wagner, G. J. Russ. Phys. Chem. Soc. 1895, 27, 219-236. (23) Rappe, A. K.; Goddard, W. A., III. J. Am. Chem. Soc. 1982, 104, 448-456. (24) Toyoshima, K.; Okuyama, T.; Fueno, T. J. Org. Chem. 1980, 45, 1600-1604. (25) Gardner, K. A.; Mayer, J. M. Science 1995, 269, 1849-1851.
Received for review November 15, 1999. Revised manuscript received March 20, 2000. Accepted March 27, 2000. ES991279Q
VOL. 34, NO. 12, 2000 / ENVIRONMENTAL SCIENCE & TECHNOLOGY
9
2541
