INFORMATION TO USERS

This manuscript has been reproduced from the microfilm master. UMI films the text directly from the original or copy submitted. Thus, some thesis and dissertation copies are in typewriter free, while others may be from any type of computer printer.

The quality of this reproduction is dependent upon the quality of the copy submitted. Broken or indistinct print, colored or poor quality illustrations and photographs, print bleedthrough, substandard margins, and improper alignment can adversely affect reproduction.

In the unlikely event that the author did not send UMI a complete manuscript and there are missing pages, these will be noted. Also, if unauthorized copyright material had to be removed, a note will indicate the deletion.

Oversize materials (e.g., maps, drawings, charts) are reproduced by sectioning the original, beginning at the upper left-hand comer and continuing from left to right in equal sections with small overlaps. Each original is also photographed in one exposure and is included in reduced form at the back of the book.

Photographs included in the original manuscript have been reproduced xerognq)hically in this copy. Ifigher quality 6” x 9” black and white photographic prints are available for any photognq>hs or illustrations tq>pearing in this copy fiar an additional charge. Contact UMI directly to order. UMI A Bell & Howell bifinnialion Conquny 300 North Zed> Road, Ann Arbor MI 48106-1346 USA 313/761-4700 800/521-0600

ABIOTIC REMEDIATION OF GROUND WATER CONTAMINATED BY CHLORINATED SOLVENTS

DISSERTATION

Presented in Partial Fulfillment of the Requirements for the Degree Doctor of Philosophy in the Graduate School of The Ohio State University

By

Yuejun Yan, B.S., M.S.

*****

The Ohio State University 1998

Dissertation Committee: Approved by Dr. Franklin W. Schwartz, advisor

Dr. E. Scott Bair A.

Dr. Yu-Ping Chin Advisor

Department of Geological Sciences UMI Number: 9822390

UMI Microform 9822390 Copyright 1998, by UMI Company. All rights reserved.

This microform edition is protected against unauthorized copying under Title 17, United States Code.

UMI 300 North Zeeb Road Ann Arbor, MI 48103 Dedicated to Beiwen and Melissa ABSTRACT

The oxidative treatment of chlorinated solvents in aqueous phase using

permanganate was investigated in the laboratory through a series of systematic batch

experiments. Kinetic experiments established the rate of breakdown of five chlorinated

ethylenes including tetrachloroethylene (PCE), trichloroethylene (TCE), and three

isomers of dichloroethylenes (DCEs). The degradation process was rapid and all pseudo

first-order rate constants are greater than 6.5 x 10^ s ' when using 1 mM Mn 0 4 ‘, except

for PCE with a relatively low rate constant (4.5 x 10'^ s '). Among the chlorinated

ethylenes, the degradation rate increases with a decreasing number of chlorine

substituents on the ethylene.

The reaction pathway and kinetics of oxidation of chlorinated ethylenes by

permanganate were studied using TCE as an example. The oxidation process mainly

involves three reaction steps: ( 1) the destruction of TCE to form a cyclic complex, (2) the decomposition of the complex to carboxylic acids, and (3) the oxidation of carboxylic acids to form the final product, CO 2 The initial reaction step in TCE oxidation is a rate- limiting step controlling the destruction rate of TCE. The kinetic study suggests that this reaction step is second order with a rate constant of 0.67 ± 0.03 M 's ' for TCE, which is

m independent of pH over a pH range of 4-8. A series of fast reactions are involved in the second step and determine the nature and distribution of intermediate products. Four carboxylic acids, including formic, oxalic, giyoxylic and glycolic acids were identified and quantified in kinetic measurements. The distribution of products is highly dependent upon experimental conditions, particularly pH. At low pH, the transformation of cyclic complex to formic acid is the dominant pathway. However, at high pH, decomposition favors the formation of oxalic and giyoxylic acids. During the decomposition process, chlorine substituents on ethylene are completely liberated to the solution. The rate constant for this step is at least 100 times greater than that for the first step. The final reaction step controls formation rate of the final product, COi. The rate constants for this step are the same as or one order less than that for the first step. With decreasing pH, the formation rate of CO 2 increases.

Based on the understanding of various processes involved with TCE oxidation, a kinetic model was developed and the model parameters (seven rate constants) were estimated. The kinetic model proposed in this study successfully simulates the observed experimental data.

IV ACKNOWLEDGMENTS

I would like to express my sincere gratitude to my advisor Dr. Franklin W. Schwartz for his guidance, encouragement, and support throughout this study. 1 would like to thank Drs. E. Scott Bair, Yu-Ping Chin, Samuel Traina for serving on my research committee and for their suggestions and comments. Special thanks are due to Dr. Samuel Traina for his advice and valuable discussion over the course of the study, to Dr. Yu-Ping Chin for providing unlimited access to his laboratory and assistance in chemical analysis, and to Dr. Olli Tuovinen for his generous help and use of his laboratory in conducting radiolabled product analysis. Thanks are also extended to many of my colleagues in the Hydrogeology Program at The Ohio State University for their help and support at various stages of the study. I am indebted to my parents for their patience and support. Finally, I would like to express my special appreciation to my dear wife Beiwen and my lovely daughter Melissa for their love, understanding, patience, and support. This work was funded by Browning-Ferris Industries (BFI) and Department of Energy. VTTA

1982 ...... B.S. Marine Geological Engineering Tongji University, Shanghai, China

1985 ...... M.S. Geology Tongji University, Shanghai, China

1993...... M.S. Geophysics Kent State University, Kent, Ohio

1994-present ...... Graduate Research and Teaching Associate The Ohio State University, Columbus, Ohio

FIELD OF STUDY

Major field: Geological Sciences

Specialization: Contaminant Hydrogeology

VI TABLE OF CONTENTS

DEDICATION...... ii

ACKNOLEDGEMENTS...... v

VITA ...... vi

LIST OF TABLES...... x

LIST OF FIGURES...... xi

CHAPTERS Page

1. INTRODUCTION...... 1

LI Bioremediation ...... 2

1.2 Abiotic Remediation ...... 5

1.3 Motivation of the Study ...... 7

1.4 Scope and Objectives ...... 8

1.5 References ...... 9

2. OXIDATIVE DEGRADATION OF CHLORINATED ETHYLENES BY

POTASSIUM PERMANGANATE...... 13

2.1 Introduction ...... 13

VII 2.2 Chemical Background ...... 15

2.3 Materials and Methods ...... 17 2.3.1 Materials...... 17 2.3.2 Kinetic experiments ...... 18 2.3.3 Chemical analyses ...... 19

2.4 Results and Discussion ...... 22 2.4.1 Reaction order ...... 22 2.4.2 Degradation of chlorinated ethylenes ...... 24 2.4.3 Dechlorination ...... 25 2.4.4 Change in acidity and pH dependence ...... 27 2.4.5 Permanganate decomposition and products ...... 28 2.4.6 Permanganate consumption in contaminated ground water ...... 30 2.5 Conclusions ...... 34

2.6 References ...... 36

3. KINETICS AND MECHANISM FOR TCE OXIDATION

BY PERMANGANATE...... 55

3.1 Introduction ...... 55

3.2 Experimental Methods ...... 57 3.2.1 Materials...... 57 3.2.2 Product studies and kinetic experiments ...... 58 3.2.3 Chemical analysis ...... 60 3.3 Results and Discussion ...... 6 1

3.3.1 Products ...... 61 3.3.2 Kinetic equations for TCE oxidation ...... 64

3.3.3 Estimation of rate constants ...... 6 6 3.3.4 pH dependence ...... 70 3.3.5 Effect of temperature ...... 71

VUl 3.4 Mechanism for TCE Oxidation ...... 73

3.5 Conclusions ...... 77

3.6 References ...... 79

4. SUMMARY...... 93

LIST OF REFERENCES...... 99

IX LIST OF TABLES

Table Page

2.1 Half-cell redox of manganese ions at various pH ranges ...... 41

2.2 Rate constants and half lives for the oxidative degradation of chlorinated ethylenes by permanganate (I mM)...... 41

3.1 Rate constants obtained for TCE transformation pathways using equations, 3.16- 3.19. Estimation based on experiments conducted with an initial concentration of

0.63 mM Mn 0 4 ...... 84 LIST OF FIGURES

Figure Page

Figure 2. I Reaction scheme. The oxidation of ethylene in a neutral to weak acidic condition ...... 42

Figure 2. 2 Schematic diagram of the bath reactor. In kinetic experiments, each sample aliquot was withdrawn from the sampling port by a syringe when the same volume of solution as the aliquot was injected via the injection port. Zero headspace was maintained at all times ...... 43

Figure 2. 3 Plot of initial rates versus initial concentration for eight kinetic experiments.

TCE ranging from 0.031 to 0.083 mM was oxidized by 1 mM Mn 0 4 at pH 7.1. A slope a = 1.01 ± 0.02 is first order with respect to TCE ...... 44

Figure 2.4 Plot of pseudo-order rate constant kobs versus initial concentration of

permanganate. TCE at 0.078 mM was oxidized by Mn 0 4 concentration varying from 0.37 to 1.2 mM at pH 7.1. A slope P = 1.05 ± 0.03 is first order with respect to M n04 ...... 45

Figure 2. 5 Degradation of chlorinated ethylenes by Mn 0 4 ( 1 mM), at pH 7.1. Lines represent best fits using a pseudo-first-order kinetic model ...... 46

Figure 2. 6 Liberation of CF ions in the TCE oxidative transformation by I mM Mn 0 4 at pH 7.1. The error bars for Cl ions are the standard deviation of triplicate samples.

XI The solid lines were calculated from Eq. 2.10 with kobs obtained from a best fit to TCE observations using Eq. 2.10(a) ...... 47

Figure 2. 7 TCE degradation rate constant (kobs) with 1 mM Mn 0 4 in the pH range 4-8. Error bars are standard deviations of kobs ...... 48

Figure 2. 8 Pseudo-first-order plot of TCE transformation in 1 mM MnO^ over the pH

range of 4-8. Rate constant kobs = 0.67 ± 0.03 x 10'^ s ' and half life T 1/2 = 17.3 ± 0.5 minutes ...... 49

Figure 2. 9 Overlay of UV-vis spectra at time intervals of 10 minutes during TCE (3.81

mM) oxidation by Mn 0 4 (0.23 mM) at pH 7.1. The solid bold line is the initial

Mn 0 4 spectrum and the dashed line is the final spectrum due to a product from the

decomposition of Mn 0 4 ...... 50

Figure 2. 10 Log A versus log X for the spectrum recorded at 80 minutes after the

reaction between MnÛ 4 (0.23 mM) and TCE (6.85 mM) at pH 8.2. The linear relationship with a slope =-4.22±0.05 obeys Rayleigh’s law ...... 51

Figure 2. 11 Comparison of TCE degradation by 0.13 mM (20 mg/L) Mn 0 4 ‘ in Milli-Q water, ground water, and contaminated ground water (2% leachate). Lines represent best fits using a pseudo-first-order kinetic equation ...... 52

Figure 2. 12 Plot of the maximum pseudo-first-order rate constant (kobs) versus TOC concentration ...... 53

Figure 2. 13 TCE degradation by Mn 0 4 ‘ at various concentrations in a synthesized contaminated ground water containing 101 mg/L TOC. The solid lines calculated using kjcE = 0.67 M'* s ' and kjoc = 0.60 M*' s '. The dashed lines calculated for TCE degradation in Milli-Q water without TOC ...... 54

XU Figure 3. 1 Carboxylic products of TCE oxidation as determined by HPLC. Samples were collected one hour after the reaction began in three eight-hour experiments at

pH 4, 6, and 8 , respectively, and monitored by absorbance at 210 nm ...... 85

Figure 3. 2 Proposed TCE oxidation pathways. Shadowed boxes are products identified

in this study ...... 8 6

Figure 3. 3 Product distribution with time in TCE oxidation at pH=4. Lines for TCE and carboxylic acids represent best fits using equations of 3.18-3.21, whereas the line for

CO2 is calculated by equation 3.23 using rate constants from the previous curve fitting ...... 87

Figure 3.4 Product distribution with time in TCE oxidation at pH= 6 . Lines for TCE and carboxylic acids represent best fits using equations of 3.18-3.21. whereas the line for

CO2 is calculated by equation 3.23 using rate constants from the previous curve

fitting ...... 8 8

Figure 3. 5 Product distribution with time in TCE oxidation at pH= 8 . Lines for TCE and carboxylic acids represent best fits using equations of 3.18-3.21, whereas the line for

CO2 is calculated by equation 3.23 using rate constants from the previous curve fitting ...... 89

Figure 3. 6 Plot of determination coefficients of curve fitting versus logarithm of ratio of

lc2 to k|. The determination coefficients were obtained by fitting equations of 3.19-

3.21 to data from carboxylic measurements using various ratios, lc 2/ki...... 90

Figure 3. 7 Dependence of the pseudo-first-order rate constants (ksa, kab, and ksc) on pH for oxidation of carboxylic acids. The reaction involved reacting 0.09 mM TCE with

0.63 mM Mn 0 4 in a phosphate buffered aqueous solution at ionic strength 0.05 M...... 91

X lll Figure 3.8 . Arrhenius plot for the oxidation of TCE (0.06 mM) by potassium permanganate (I mM) in a phosphate buffered solution at ionic strength 0.05M and pH 7.1. Error bars are standard deviation for each estimated second-order rate constant k[ ...... 92

Figure 4. I . Half lives of degradation of chlorinated ethylenes in permanganate oxidation and iron reduction. The initial concentration of permanganate is 158 mg/L, while initial concentration of iron powder is 250 g/L with high surface area. The half lives in iron reduction are from Gillham and O’Hannesin ( 1994) ...... 98

XIV CHAPTER I

INTRODUCTION

Chlorinated solvents are the most common industrial ground-water contaminants in the United State and Europe. They were first produced in Germany at the end of the last century and were imported into the United States in about 1906 (Wiseman, 1979). In the early to middle twentieth century, they gained widespread use as ingredients in industrial cleaning solutions, a "universal" degreasing agent for metals, and a detergent in dry cleaning applications. Patents issued for new organochloride formulations exploded during this period (Hutzinger and Veerkamp, 1981) and usage reached a peak in mid-

1980s.

Chlorinated solvents are the largest single group of compounds on the list of priority pollutants compiled by the United States Environmental Protection Agency

(EPA). They are either known or suspected carcinogens (Ram et al., 1990). When presenting as contaminants in ground-water systems, chlorinated solvents pose an extremely difficult clean-up challenge. First, even low aqueous concentrations are greatly in excess of public drinking-water standards, although many chlorinated solvents are sparingly soluble in water. Thus, a large volume of ground water can be contaminated

1 by a small amount of spilled solvent. Secondly, due to their resistance to natural

attenuation and persistence in the subsurface, pure-phase chlorinated solvents can remain over many decades, or even centuries. In place, these contaminants serve as a long-term source of continued ground-water contamination. Finally, by virtue of a density much greater than water, chlorinated solvents, as pure liquids, may move deep into the subsurface.

Over the past decade, the development of improved methods for remediation of ground water contaminated by chlorinated solvents has emerged as a significant environmental priority. With growing awareness of the limitation of many conventional technologies (e.g. pump-and-treatment), substantial research efforts are now focusing on in-situ remediation technologies using either biological process or chemical process.

1.1 Bioremediation

Biological degradation of chlorinated solvents mainly involves anaerobic reductive and aerobic oxidative dechlorination (Ensley, 1991; Wackett, 1995). In the early 1980s, the biodégradation of chlorinated solvents was demonstrated through an anaerobic dechlorination process (Bouwer et al., 1981). Barrio-Lage et al. (1986) proposed a reductive dechlorination mechanism for the anaerobic conversion of trichloroethylene (TCE) to dichloroethylene (DCE) isomers, and vinyl chloride (VC).

With studies of pure culture of methanogens, a clearer picture has emerged on how microorganisms and enzymes are involved in reductive dechlorination (Belay and

Daniels, 1987; Fathepure et al., 1987 and 1988). In detail, reductive dechlorination by anaerobes involves the transfer of electrons directly to chlorinated ethylenes via

biochemical processes in which the reduction of the chlorinated hydrocarbon is coupled

to the oxidation of electron transfer agents involved in methane biosynthesis. Wackett

(1995) pointed out that the reductive reactions could actually be envisioned as a short

circuit within an oxidative reaction cycle. In effect, electrons are transferred directly to

the chlorinated hydrocarbon rather than to molecular oxygen as in the oxygenative cycle,

which is a reaction cycle with aerobic oxidative dechlorination.

The greatest attraction of anaerobic dechlorination is that highly chlorinated molecules, which resist attack by aerobic organism (Ewers et al.. 1990; Abramowicz,

1990), could be reductively dechlorinated, such as PCE to TCE (Fathepure and Boyd,

1988) and highly chlorinated PCBs to less chlorinated PCBs (Quensen et al., 1988).

However, there are disadvantages. First, toxic and carcinogenic products (DCE isomers and VC) form from reductive dechlorination of PCE or TCE. These compounds tend to be increasingly persistent because the fewer chlorine atoms left on a molecule, the slower the rate of dechlorination in an anaerobic system (Vogel et al., 1987; Freedman and

Gossett, 1989). Thus, field studies indicate that there is the tendency for significant amounts of VC and cis-1,2-dichloroethylene (cDCE) to accumulate under anaerobic condition (Beeman et al., 1993; Smith and Ferguson, 1993). A second disadvantage is that reductive dechlorination will not occur normally in an aerobic aquifer because this process is strongly inhibited by oxygen (Logan et al., 1993).

An alternative approach has been developed to avoid the formation of vinyl chloride, a known human carcinogen (Maltoni and Lefemine, 1974). This approach involves the cooxidative metabolism of chlorinated solvents under aerobic conditions

(Wilson and Wilson, 1985). Chlorinated solvents such as TCE can be transformed by

organisms that use a primary substrate, such as methane, for metabolism. Many broadly-

specific bacterial oxygenases have been found to generate a highly reactive monoatomic

or diatomic active oxygen species which could bind to and electrophilically attack many

hydrophobic compounds including chlorinated solvents (Wackett et al., 1989). Li and

Wackett ( 1992) found that the major TCE oxidation products were formic acid and giyoxylic acid, which are nontoxic. Since 1985, much work has been devoted to this subject.

Reaction mechanisms in oxidative enzymes are understood relatively well, as compared to reductive dechlorination. With the progress of these studies, however, some findings emerge as limitations in field applications:

(1) Metabolite toxicity: TCE intermediate metabolites generated during the oxidation

process could inactivate the enzyme system to cause severe cell death and to inhibit

the further oxidation (Wackett and Gibson, 1988; Wackett and Householder, 1989).

The toxic effect increases with increasing TCE concentration.

(2) Competitive cosubstrate inhibition: To induce the necessary oxygenase and to

regenerate reducing power in the cells lost during the oxidation of chlorinated

solvents, the cosubstrate, such as methane, propane, toluene or phenol, must be present

in relatively high concentrations. At the same time, however, the cosubstrate

concentrations must be low enough so as not to completely inhibit oxidation of

chlorinated solvents through competition for active sites on the enzymes. Apparently, the proper set of cosubstrate-concentration conditions are extremely rare in

the environment where chlorinated solvents persist (Ensley, 1991).

Because of limitations with anaerobic reductive and aerobic oxidative processes,

further research in biodégradation has focused on engineering sequential anaerobic and aerobic reactors (Phelps et al., 1991; Wackett, 1995). To date, this process presents

formidable challenges to practical use in subsurface application.

A logical and complementary research strand involves examining the opportunities for abiotic destruction of compounds in situ. The following section will review research related to abiotic processes for solvent destruction.

1.2 Abiotic Remediation

In general, the abiotic degradation of chlorinated solvents can involve reductive and oxidative processes. Reductive dechlorination was first demonstrated in experiments through carbon tetrachloride (CT) transformation to chloroform and CS? in the presence of pyrite, sulfide, biotite or vermiculite (Kriegman-King and Reinhard, 1992 and 1994).

The disappearance of CT was, however, very slow with half lives varying from 12 days in a pyrite system to 160 days in a vermiculite system. In 1994, Gillham and O'Hannesin demonstrated the degradation of 14 halogenated solvents using zero-valent iron.

Meanwhile, Matheson and Tratnyek (1994) proposed several possible reductive dechlorination mechanism involving either a direct electron transfer from Fe° to the absorbed chloroethylene at the iron surface or the reaction with dissolved Fe""^ or Hi, which are products of iron corrosion. Because of the relatively short degradation half lives for many halogenated solvents, this process has received considerable attention.

Iron reduction of PCE or TCE could produces toxic byproducts DCE isomers and vinyl chloride if the dechlorination process is incomplete. The degradation rate of solvents decreases as their degree of chlorination is reduced. For example, the half life for TCE degradation is 13.2 hours as compared to 432 hours for cDCE (Gillham and

O'Hannesin, 1994). Further efforts are underway in understanding the process of iron reduction and inhibition by blocking reactive sites on the iron surface due to corrosion products.

Oxidative dechlorination emerged as extensions of advanced oxidation processes

(AOPs), which have been extensively studied as a non-Cl? oxidation treatment for drinking water. AOPs have received considerable attention since 1979 when U.S. EPA regulated four trihalomethanes, the by-products of chlorine treatment. AOPs involve the use of reactive hydroxyl radicals, which are generated by ozone (O 3 ), hydrogen peroxide

(H2O2), or ultraviolet radiation (UV), to oxidize organic compounds. Glaze et al. (1987) reported that four processes, 0 3 /high pH (OH ), O 3/H 2O2 , O3/UV, and H 2O2/UV, could accelerate the generation of hydroxyl radicals. Among these processes, the O 3/H2O2 combination is the most effective. The degradation of TCE and PCE through the

O3/H2O2 process was successfully demonstrated by Glaze and Kang (1988) in bench- scale experiments and by Aieta et al. (1988) in pilot-scale studies. The half life for the degradation is short, about four minutes. However, the consumption of hydroxyl radicals by carbonate and bicarbonate (Hoigné and Bader, 1983) provides a significant limitation in subsurface applications.

Potassium permanganate has been widely used as a powerful oxidant for several decades in organic synthesis (Stewart, 1965; Haines, 1985). A useful study on oxidative dechlorination of TCE was undertaken by Vella and Veronda (1992). They compared potassium permanganate with HoO? in oxidation of TCE and found that potassium permanganate also can oxidize TCE. Apparently, the permanganate oxidation of chlorinated solvents does not follow the hydroxyl radical reaction pathway and, thus, effects of carbonate and bicarbonate may be limited. This feature would appear to be a significant benefit for in-situ remediation.

1.3 Motivation of the Study

In spite of significant progress in both in-situ bioremediation and abiotic remediation, there remain limitations and potential problems in field application.

Exploring new in-situ remediation methods and technologies is an environmental priority.

As a potential in-situ remediation method, permanganate oxidative dechlorination has several advantages. Permanganate is a very simple and common chemical and is less expensive than AOPs. Due to the nature of oxidation, the byproducts and products are much less harmful than those from reduction reaction. Degradation process are rapid

(Vella and Veronda, 1992). Permanganate as a metal-oxo reagent may have the similar function as symmetrical iron-bridged dioxygen species to electrophilically attack chlorinated solvents in aerobic dechlorination by dioxygenase. Thus, the reaction might not be affected by carbonate and bicarbonate as AOPs do. To date, the permanganate oxidation for the purpose of dechlorination has not been well investigated and the reaction mechanisms have not been well defined, although numerous studies and reviews of permanganate oxidation are reported (Amdt, 1981). The potential of using permanganate oxidative dechlorination in pollutant remediation provides the main motivation for this study.

1.4 Scope and Objectives

The major goal of this study is to understand oxidative destruction of chlorinated solvents using potassium permanganate. Of particular interests are the kinetics and mechanisms of oxidation and the factors that influence these reactions including the effects of subsurface environment and competition from the other compounds in contaminated groundwater. The study will provide a theoretic basis to evaluate the feasibility of in-situ applications, to couple kinetic reaction with transport models, and to develop an appropriate field test for further assessing the approach.

Among chlorinated solvents, chlorinated ethylenes are major toxic compounds found in ground-water systems. TCE and PCE are ranked number one and number three in the 25 most frequently detected ground-water contaminants at hazardous waste sites in the United States (National Research Council, 1994). Therefore, five chlorinated ethylenes including PCE, TCE, and three isomers will be the target compounds in this investigation. The study mainly involves a series of systematic batch experiments to

8 elucidate basic mechanism and kinetics of the reaction. The column tests are not included in this thesis, even though they were conducted for the project supported by

CECOS/BFI.

The thesis includes four chapters. The current chapter serves as a general introduction to the study, including an overview of remediation technologies, and the motivation and the objectives of the study. Chapter 2 investigates the oxidative degradation of five chlorinated ethylenes and measures their destruction rates. This section demonstrates the oxidative dechlorination process and examines the effect of experimental conditions on degradation rate and the competition for permanganate from other organic compounds in ground water or highly contaminated ground water. Chapter

3 focuses on the mechanism of oxidation using TCE as an example. Sequential reactions were postulated based on identified products with time in the TCE degradation process.

Based on an understanding of degradation process, a kinetic model is developed, which simulates the experimental data well. Chapter 4 summarizes the results from this investigation, outlines the major findings of the study, and discusses the direction of the future research on this subject.

1.5 References

Abramowicz, D. A., 1990. Aerobic and anaerobic biodégradation of PCBs; A review. Crit. Rev. Biotechnol., 10:241-251.

Aieta, E. M., K. M. Regan, J. S. Lang, L. McReynolds, J. W. Kang, and W. H. Glaze, 1988. Advanced oxidation processes for treating groundwater contaminated with TCE and PCE: Pilot-scale evaluations. Jour. AWWA, 80:64-72. Aradt, D., 1981. Manganese compounds as oxidizing agents in organic chemistry. Open Court Publishing Company, La Salle, Illinois, 344 pp.

Barrio-Lage, G. A., F. Z. Parsons, R. S. Nassar and P. A. Lorenzo, 1986. Sequential dehalogenation of chlorinated ethenes. Environmental Science and Technology, 20:96-99.

Beeman, R., S. Shoemaker, J. Howell, E. Salazar, and J. Buttram, 1993. A field evaluation of in situ microbial reductive dehalogenation by the biotransformation of chlorinated ethylenes, abstr. A-2. In In situ and on-site bioreclamation: 2nd International Symposium, San Diego, Calif., Battelle, Columbus. Ohio.

Belay, N. and L. Daniels, 1987. Production of ethane and ethylene from halogenated hydrocarbon by methanogenic bacteria. Applied Environ. Microbiol., 53:1604- 1610.

Bouwer, E.J., B.E. Rittman and P.L. McCarty. 1981. Anaerobic degradation of halogenated 1- and 2-carbon organic compounds. Environmental Science and Technology, 15:596-599.

Ensley, B. D., 1991. Biochemical diversity of trichloroethylene metabolism. Annu. Res. Microbiol., 45:283-299.

Ewers, J., D. Freier-Schroder and H-J Knackmuss, 1990. Selection of trichloroethene (TCE) degrading bacteria that resist inactivation by TCE. Arch. Microbiol.. 154:410-413.

Fathepure, B. Z., J. P. Nengu and S. A. Boyd, 1987. Anaerobic bacteria that dechlorinate perchloroethylene. Applied Environ. Microbiol., 53:2671-2674.

Fathepure, B. Z. and S. A. Boyd, 1988. Dependence of tetrachloroethylene dechlorination on methanogenic substrate consumption by Methanosarsina sp. strain DSM. Applied Environ. Microbiol., 54:2976-2980.

Freedman, D. L. and J. M. Gossett, 1989. Biological reductive dechlorination of tetrachloroethylene and trichloroethylene to ethylene under methanogenic conditions. Applied and Environmental Microbiology, 55:2144-2151.

GiIlham, R. W. and S. F. O'Hannesin, 1994. Enhanced degradation of halogenated aliphatics by zero-valent iron. Ground Water, 32:958-967.

Glaze, W. H., J. W. Kang, and D. H. Chapin, 1987. The chemistry of water treatment processes involving ozone, hydrogen peroxide, and ultraviolet radiation. Ozone Science and Engineering., 9:335.

1 0 Glaze, W. H. and J. K. Kang, 1988. Advanced oxidation processes for treating groundwater contaminated with TCE and PCE: laboratory studies. Jour. AWWA, 80:57-63.

Haines, A. H., 1985. Methods for the oxidation of organic compounds. Academic Press. London, 388 pp.

Hoigné, J. and H. Bader, 1983. Rate constants of reaction of ozone with organic and inorganic compounds in water, I - non-dissociating organic compounds. Water Research, 17:173.

Hutzinger, O. and Veerkamp W., 1981. Xenobiotic compounds with pollution potential. In Leisinger T., Cook A., Hutter R., and Nuesch J., eds: Microbial Degradation of Xenobiotic and Recalcitrant Compounds. Academic Press, London, p 3.

Kriegman-King, M. R. and M. Reinhard, 1992. Transformation of carbon tetrachloride in the presence of sulfide, biotite, and vermiculite. Environmental Science and Technology, 26:2198-2206.

Kriegman-King, M. R. and M. Reinhard, 1994. Transformation of carbon tetrachloride by pyrite in aqueous solution. Environmental Science and Technology, 28:692- 700.

Li, S. and L. P. Wackett, 1992. Trichloroethylene oxidation by toluene dioxygenase. Biochem. Biophys. Res. Commun., 185:443-451.

Logan, M. S. P . , L. M. Newman, C. A . Schanke, and L. P . Wackett, 1 9 9 3 . Co-substrate effects in reductive dehalogenation by Pseudomonas putida 0786 expressing cytochrome P 4 5 0 c a m - Biodégradation 4 : 3 9 - 5 0 .

Maltoni, C. and G. Lefemine, 1974. Carcinogenicity bioassays of vinylchloride. 1. Research plan and early results. Environ. Res., 7:387-396.

Matheson, L. J., and P. G. Tratnyek, 1994. Reductive dehalogenation of chlorinated methanes by iron metal. Environmental Science and Technology, 28:2045-2053.

National Research Council, 1994. Alternatives for ground water cleanup. National Academic Press, Washington, D. C., p 26.

Phelps, T. J., J. J. Niedzielski, K. J. Malachowsky, R. M. Schram, S. E. Herbes, and D. C. White, 1991. Biodégradation of mixed-organic wastes by microbial consortia in continuous-recycle expanded-bed bioreactors. Environmental Science and Technology, 25:1461-1465.

11 Quensen, J. F. m , J. M. Tiedje and S. A. Boyd, 1988. Reductive dechlorination of polychlorinated biphenyls by anaerobic microorganisms from sediment. Science, 242:752-754.

Ram, N. M., Christman, R. F., and Cantor, K. P., 1990. Significance and treatment of volatile organic compounds in water supplies. Lewis Publishers, Chelsa, MI.

Smith, G. and G. Ferguson, 1993. In situ remediation of groundwater contaminated with chlorinated solvents using anaerobic biotransformation, abstr. C-2. In In situ and on-site bioreclamation: 2nd International Symposium, San Diego. Calif.. Battelle. Columbus, Ohio.

Stewart, R., 1965. Oxidation by permanganate. In Oxidation in organic chemistry. K. B. Wiberg, ed.. Academic Press, New York, Part A Chapter 1, pp 1-68.

Vella, P. A. and B. Veronda, 1992. Oxidation of trichloroethylene: comparison of potassium permanganate and Fenton's reagent. The third international symposium on chemical oxidation technology for the nineties, Nashville, TN.

Vogel, T. M., C. S. Criddle and P. L. McCarty, 1987. Transformations of halogenated aliphatic compounds. Environmental Science and Technology, 21:722-736.

Wackett, L. P. and D. T. Gibson, 1988. Degradation of trichlorethylene by toluene dioxygenase in whole-cell studies with Pseudomonas putida FI. Appl. Environ. Microbiol., 54:1703-1708.

Wackett, L. P. and S. R. Householder, 1989. Toxicity of trichloroethylene to Pseudomonas putida FI is mediated by toluene dioxygenase. Appl. Environ. Microbiol., 55:2723-2725.

Wackett, L. P., 1995. Bacterial co-metabolism of halogenated organic compounds. In Microbial transformation and degradation of toxic organic chemicals, L. Y. Young and C. E. Cemiglia, ed., Wiley-Liss, Inc, New York, pp 217-241.

Wilson, J. T. and B. H. Wilson, 1985. Biotransformation of trichloroethylene in soil. Applied Environ. Microbiol., 29:242-243.

Wiseman, P., 1979. An Introduction to Industrial Organic Chemistry, 2nd edition. Elsevier Applied Science, New York, 366 pp.

1 2 CHAPTER 2

OXIDATIVE DEGRADATION OF CHLORINATED ETHYLENES BY

POTASSIUM PERMANGANATE

2.1 Introduction

Chemical approaches for the remediation of ground water contaminated by

chlorinated solvents commonly utilize reduction or oxidation scheme to transform

organic contaminants. To date, most work has concentrated on reductive dechlorination,

which occurs with hydrolysis (Jeffers et al., 1989) and surface reactions involving pyrite,

sulfide, biotite or vermiculite (Kriegman-King and Reinhard, 1992 and 1994) and zero-

valent iron (Gillham and O'Hannesin, 1994; Matheson and Tratnyek, 1994). In

particular, remedial schemes developed around zero-valent iron have shown considerable

promise due to the relatively short half lives of the reactions.

There has been much less recent work on oxidation processes, even though

experience has shown oxidative degradation to be fast with the half life of degradation for chlorinated organic compounds, of the order of several minutes in a O 3/H2O2 system

(Glaze and Kang, 1988). An apparent limitation with this reaction is that the key reactive

intermediate hydroxyl radical, generated in this advanced oxidation process (AGP),

13 strongly reacts with common inorganic species in ground water such as carbonate and

bicarbonate (Hoigné and Bader, 1983).

Not all oxidants suffer from this limitation. Permanganate, as a metal-oxo reagent

(Gardner and Mayer, 1995), does not apparently rely on generating a hydroxyl radical to oxidize halogenated ethylenes as AOPs do. Experience spanning more than a century in laboratory-scale organic synthesis indicates that metal-oxo reagents can attack a double carbon-carbon bond (Stewart, 1964) powerfully through direct oxygen transfer (Wiberg and Saegebarth, 1957). This feature of metal-oxo reagents facilitates the degradation of chlorinated ethylenes with little scavenging of carbonate or bicarbonate.

There have been both laboratory and field experiments that have demonstrated the ability of potassium permanganate to oxidize common chlorinated ethylenes like trichloroethylene (TCE) and tetrachloroethylene (PCE). In a series of batch experiments with both water (Vella and Veronda, 1992) and soil (Gates et al., 1995), both TCE in water and TCE and PCE in soil were oxidized by permanganate. Compared to Fenton’s reagent (a mixture of hydrogen peroxide and ferrous ion), permanganate was less dependent on pH and had a higher efficiency in water or soil treatment. Gonullu et al.

( 1997) observed in column tests that more than 90% of the TCE and PCE was degraded after flushing with several pore volumes of aqueous permanganate. Most of these experiments can be considered as a proof-of-concept that demonstrates the efficacy of the remedial concept with less emphasis on reaction pathways and kinetics.

Investigations are designed to provide a detailed process-level understanding of the oxidative destruction of chlorinated ethylenes by permanganate. In this paper, the

14 specific objectives are ( I ) to examine reaction order, degradation rate and kinetic behavior of chlorinated ethylenes in reactions with permanganate, ( 2 ) to demonstrate the extent of dechlorination, and (3) to assess the effects of pH and other organic compounds in subsurface environment on the TCE degradation rate. Further work is underway to elucidate the detailed reaction pathways based on product analysis.

2.2 Chemical Background

There has been little direct process-oriented work on the oxidation of chlorinated ethylenes by permanganate. However, the body of work on alkene (C=C) oxidation in chemical synthesis provides a general understanding of the oxidation of chlorinated ethylenes. Figure 2.1 depicts a reaction scheme for the oxidation of ethylene in a neutral to weak acidic condition. The oxidation reaction begins with the formation of a cyclic hypomanganate ester (1) (Wiberg and Saegebarth, 1957; Wiberg et al., 1973; Lee and

Brownridge, 1973). The cyclic ester then undergoes oxidative decomposition in a neutral to weak acidic medium through hydrolysis, with fission of the Mn-0 bonds to form a glycol aldehyde (Wiberg and Saegebarth, 1957). The glycol aldehyde could be further oxidized to glyoxylic acid and oxalic acid (2) (Arndt, 1981 ; Szammer and Jaky, 1992).

Another possible reaction pathway has the ester directly cleaved by permanganate to form two formic acids (3) (Wiberg and Saegebarth, 1957). All the carboxylic products could be further oxidized to carbon dioxide under certain conditions (Stewart, 1965).

The kinetics and mechanism of the reaction are affected when halogens substitute for hydrogen on ethylene. Surdon and Tatlow (1958) observed a faster oxidation

15 reaction (a few seconds) with fluorinated alkene, as compared to hydrocarbon alkene, using permanganate in acetone during the synthesis of carboxylic acids. They thought

that an electron-withdrawing group such as fluorine on the alkene facilitates the

nucleophilic attack by permanganate ions. Lee ( 1982) also pointed out that the oxidation of hydrocarbons can sometimes be facilitated by prior halogénation of the oxidation site.

However, Freeman ( 1975) considered the permanganate ion as an electrophile in the reaction. In his overview of activated complexes in addition reactions, he suggested that the attack of permanganate ion on carbon double bond, as an electrophilic addition, resulted in the formation of a five-member cyclic activated complex in the transition state. A chlorine as an electron-withdrawing substituent induces a deficiency of electrons at the carbon double bond of the substituted ethylene. Based on the concept of electrophilicity, a decrease of electron availability at the carbon double bond should decrease the rate of the electrophilic addition reaction.

Although only a few studies are available on the oxidation of halogenated ethylenes. The author postulates a possible reaction as shown in equation 2.1,

cc C 2 C l n H ^ .n + P MnO^ —> yCA + S Mn02 + C Cl (2 .1)

where a, P, y, ô, and C are stoichiometric coefficients, C 2ClnH 4.„ represents various chlorinated ethylenes, including PCE, TCE and dichloroethylenes (DCEs), and CA is a group of intermediate products. These intermediate products could be either chlorinated or hydrocarbon carboxylic acids, which might be further oxidized to carbon dioxide under certain conditions as follows.

1 6 CA CO, (2 .2 )

In oxidation by permanganate, pH is considered as a primary variable because it strongly influences the redox potential in a system. Table 2.1 lists the redox reactions and the corresponding potentials of various couples of manganese ions at different pH, based on Stewart’s overview (1965) on oxidation by permanganate. Table 2.1 suggests that the pH of the system determines the number of electrons involved in the over-all reaction. In general, the overall redox potential of the system increases with decreasing pH. With organic substrates, however, mechanistic factors are of major importance, whereas the over-all free energy change of the oxidant, determined by its potential, is of minor importance. For the chlorinated ethylenes, a lack of knowledge of the reaction mechanism makes it unclear whether pH will affect the degradation rates of chlorinated ethylenes.

2.3 Materials and Methods

2.3.1 Materials

The chlorinated ethylenes. PCE (C 2CI4 , 99+%), TCE (C2 HCI3, 99.5+%), cis-DCE

(C2H2CI2 97%), trans-DCE (C 2H2CI2 97%), and 1 , 1-DCE (C2H2CI2, 99%) were obtained from Aldrich Chemical Co.(Milwaukee, WI) and used as received. A high purity pentane of GC grade (Burdick & Jackson) from Baxter Diagnostics Inc. (McGaw Park, IL) was used as a liquid-liquid extraction solvent for analysis of chlorinated ethylenes. Potassium permanganate stock solution of 1-10 mg/ntiL was prepared by dissolving KMn 0 4 crystals in Milli-Q water or phosphate-buffered Milli-Q water. The stock solution was stored in

17 brown glass bottles and used freshly. The reducing agents, thiosuifate and hydrazine hydrate (95+%), are both research grade and were prepared as a stock solution ( 6 mg/mL) for quenching the reaction in some kinetic experiments. A total organic carbon (TOC) stock solution ( 1000 mg C/L) for use in the TOC analysis was prepared by dissolving

0.2125 g of dried, reagent grade potassium hydrogen phthalate in 100 mL of Milli-Q water.

Experiments were performed using solutions prepared by dissolving the various chlorinated ethylenes in Milli-Q water (Millipore Corp., Bedford, MA), phosphate- buffered Milli-Q water, ground water, and contaminated ground water, respectively.

Contaminated ground water was synthesized as a mixture of ground water and landfill leachate at various ratios. Ground water was collected from the uppermost perched aquifer below the CECOS/BFI landfill site in Cincinnati, Ohio, and landfill leachate was obtained from the same site. These waters were filtered by 0.2 pm filters to provide a homogeneous reaction solution but were not acidified. Apparently, most of metals in solution were oxidized when they were exposed to the atmosphere. For example, no ferrous iron was detected by phenanthroline colorimetric method. Thus, the effect of metal on oxidation of TCE by permanganate was not evaluated in this investigation.

2.3.2 Kinetic experiments

Most of kinetic experiments were conducted in a sealed and water-jacketed spherical glass reaction vessel (Figure 2.2) where a zero headspace was maintained at all times. To test impact of subsurface environment on the reaction, experiment conditions

1 8 were adjusted with pH ranging from four to eight. One experiment was mn without a pH

buffer to monitor the change in acidity of the solution. Other experiments were run with

aqueous solutions prepared from phosphate-buffered Milli-Q water, ground water or contaminated ground water, and with initial concentrations of TCE ranging from two

mg/L (1.52 X 10'^ M) to 10 mg/L (7.61 x 1 0 '^ M). The test solution of contaminated ground water was a mixture of ground water with 2 % or 2 0 % landfill leachate.

The rate at which the chlorinated-ethylene is degraded can be simply measured when the other reactant, permanganate, is held essentially constant through experiment.

To maintain a constant concentration of KMnO^, a more than ten-fold excess concentration of KMn 0 4 was employed in each experiment. The activity of Cl', pH, temperature and the concentration of chlorinated ethylenes were monitored with time.

In addition to the above experiments, a set of experiments was designed to examine the consumption rate of KMn 0 4 by organic compounds other than the chlorinated ethylenes. The organic compounds in ground water and contaminated ground water were characterized by TOC. In this set of experiments, KMnÜ 4 (20 mg/L) was isolated by using solutions containing various percentages of landfill leachate with excess concentration of TOC varying from 302 to 506 mg/L. The concentration of KMn 0 4 was monitored through the experiment.

2.3.3 Chemical analyses

(a) Analysis of chlorinated ethylenes. A sample ranging from 20-pL to 2-mL was collected at a fixed time interval from the reaction vessel using a syringe. The sample

19 was diluted as necessary with Milli-Q water in 10-mL to 1-L volumetric glass flasks,

depending on the concentration of the chlorinated ethylenes. A 10-mL sample after

dilution was transferred to an extraction vial containing 4 mL of pentane. The vial was

shaken for one minute and then equilibrated for 10 minutes. The extractant was analyzed

for the particular chlorinated ethylene using a Fisons Instruments 8060 gas

chromatograph equipped with a electron capture detector and a DB-5 capillary

column (J&W Scientific, Rancho Cordova, CA), 30 m x 0.32 mm l.D, with a film

thickness of 1.0 pm. Helium was used as the carrier gas and nitrogen as the make-up gas.

The gas chromatograph was calibrated daily with a minimum four calibration standards,

and duplicate measurements were made for each sample or standard. If the standard

deviation of measurements was greater than 1 0 %, another measurement would be made.

(b) Analysis ofKMn0 4 and its product. A 1.5-mL sample was collected and transferred

into a quartz cuvette with 1 cm pathline and analyzed using a Varian Cary 1 UV-visible

spectrophotometer at wavelengths ranging from 400 to 700 nm. In this range of

wavelength, permanganate (Mn 0 4 l and hypomanganate (Mn 0 4 ^ ) ions have a maximum

of absorbance at 525/546 nm and 667 nm, respectively, while manganate (Mn 0 4 " ) has

two absorbance peaks at 439 and 606 nm (Stewart, 1965). These species are almost transparent to radiation at 418 nm. The absorbance at 418 nm is a measure of either a cyclic hypomanganate ester (Wiberg et al., 1973; Lee and Brownridge, 1973) or a soluble form of colloidal MnOa (Simandi and Jaky, 1976; Mata-Pérez and Pérez-Benito, 1985).

These interpretations have been controversial for a long time.

2 0 (c) Measurement ofCU. Initially, the activity of Cl' in reaction vessel was measured using an Orion ion selective electrode (ISE) with glass body (Model 9617). The ISE was calibrated using a standard solution of NaCl in the range from 0 .1 to 10'^ M. Because of low sensitivity of the ISE to the trace amount of Cl', ranging from 10"^ - 10'^ M, additional samples were measured by a Buchler Digital Chloridometer, using a coulometric titration of chloride ions (Cotlove, 1958). Before the titration, a 0 .1-tnL aliquot of reducing reagent stock solution ( 6 mg/mL), hydrazine hydrate, was added to a

0.9-mL sample taken from the reaction vessel to quench the reaction in the sample solution. To prepare for the titration, the total 1-mL sample aliquot was added to a test vial containing 3-mL of an acid reagent comprised of 0.4 N HNO 3 and 40% glacial acetic acid. Finally, four drops of gelatin reagent were added to the test vial. The samples in the test vials were titrated at either the LOW or HIGH switch position in order to provide the proper concentration range from 3.3x10'^ to 33.3 mM.

(d) TOC analysis. TOC was analyzed using a Shimadzu TOC-5000 TOC Analyzer. The four standards for calibration were prepared by diluting the TOC stock solution to a concentration in the range of 10 - 60 mg C/L. All standards and samples were acidified with concentrated HCl and sparged with CO 2 free air. Samples of contaminated ground water and landfill leachate with high TOC concentration were diluted as needed in order for concentrations to fall within the calibrated range of the TOC analyzer. Standards and

2 1 samples were analyzed in triplicate and the values of TOC were accepted when the standard deviation was less than 2 %.

2.4 Results and Discussion

2.4.1 Reaction order

Based on the Eq. 2.1, the degradation of chlorinated ethylenes can be described with the following rate equation

r =------= k[CiClnH^-n r [ MnO'4 f (2.3) a dt

When the concentration of [Mn 0 4 '] is in excess, Eq. 2.3 can be simplified as Eqs. 2.4 and

2.5.

r = k„h,[C2ClnH4.n r (2.4)

k„bx = k [ MnO'4 (2.5) where r is a reaction rate, k is a rate constant, and kobs is a pseudo-order rate constant because [Mn 0 4 "] is effectively constant during the course of the experiment. By varying the values of [CzClnfL*.» ] and measuring reaction rate, the order a with respect to

[C2 ClnH 4 .n ] Can be simply determined by a log-log form of Eq. 2.4:

log r = log k„hx + oc log [C2Cl„H4-nJ (2.4)

2 2 In a similar way but varying initial concentration of [Mn 0 4 lo and measuring kobs, the order P with respect to [Mn 0 4 l can be obtained by a log-log form of Eq. 2.5:

log k„bs = log A: + jS log [ MnO'^ Jo (2S)

To avoid complications from subsequent reactions or catalysis, an initial rate method

(Casado, 1986) was used here and Eq. 2.6 can be expressed as:

log ro = log k„b, + a\o% [C2 CLH 4.nl0 (2 .6 )

The two sets of kinetic experiments were designed to estimate a and P values for

TCE oxidation. The first set of eight experiments was conducted with initial TCE concentrations varying from 0.031 to 0.083 mM. The initial permanganate concentration was fixed at 1 mM for all experiments. The initial reaction rates were estimated as the tangent to the TCE concentration-time curve. As shown in Figure 2.3, the slope a = 1 .01

Ï 0.02 was calculated (based on Eq. 2.8) through a linear regression of the logarithm of initial rates versus the logarithm of initial TCE concentration (r^=0.998). The reaction order with respect to TCE is unity and kobs is a pseudo-first-order rate constant.

In the second set of five duplicate experiments, the initial concentration of TCE was fixed at 0.078 mM and TCE was reacted with excess Mn 0 4 ranging from 0.37 to 1.2 mM. A slope of P = 1.05 ± 0.03 was determined from a plot of the logarithm of kobs versus the logarithm of Mn 0 4 (Figure 2.4). The reaction order with respect to Mn 0 4 is also unity.

23 Hence, the results from both Figures 2.3 and 2.4 demonstrate that the initial reaction between TCE and MnÛ 4 is a second-order reaction with a= l and P=I. The second-order rate constant k of 0.66 ± 0.02 M*‘s*‘ were estimated using Eq. 2.5 with P =

1 to fit the second set of experimental data.

2.4.2 Degradation of chlorinated ethylenes

Five chlorinated ethylenes were investigated through kinetic experiments using 1 mM MnO^. As before, to maintain MnO^ in excess, the concentration of chlorinated ethylenes is at least 10 times less than M nO/. Figure 2.5 depicts the degradation behavior of the chlorinated ethylenes. Apparently, the disappearance of chlorinated ethylenes can be simply characterized by a pseudo-first-order model. The pseudo-first- order rate constants were calculated from Figure 2.5. Values range from 0.45 to 300 x

10"^ s ' and are listed in Table 2.2. The degradation is rapid for most of compounds with half lives generally less than 20 minutes. PCE is the exception with a half life 257 minutes or about four hours.

Figure 2.5 illustrates that the degradation rate is inversely proportional to the number of chlorines as substituents on ethylenes. The fact that chlorine as a substituent slows down the reaction is consistent with the idea of electrophilic addition proposed by

Freeman ( 1975). The high deficiency of electrons in the carbon double bond, induced by four chlorine substituents in PCE, reduces the rate of electrophilic attack. Therefore,

PCE degradation is slow and its rate constant is small compared to the others.

24 Trans isomers are generally more stable than the corresponding cis isomers in alkenes. However, Figure 2.5 shows a higher reactivity of trans-DCE as compared to the degradation rate for cis-DCE. The ratio of kck to k^ans is only 0.03. It indicates that a significant steric effect was evident in the reaction. The steric interaction of cis substituents is caused by the change in bond angles in the addition reaction involving the large cyclic activated complexes, such as five- and six-membered cyclic complexes

(Freeman, 1975). Apparently, a significant steric effect involving cis-DCE is consistent with the formation of the five-membered cyclic hypomanganate ester, which was originally proposed by Wagner (1895) and later supported by Wiberg and Saegebarth’s

(1957) experimental data.

2.4.3 Dechlorination

The oxidation of chlorinated ethylenes likely starts with a Mn 0 4 attack on the

C=C double bond to form a cyclic complex, similar to the oxidation of ethylenes.

However, it is unclear whether dechlorination proceeds by hydrolysis or further oxidation during the decomposition of the cyclic complex. If chlorines remain in Intermediate products after the cleavage of C=C bond, the most likely compounds would be chlorinated organic acids such as formyl chloride, oxalyl chloride and phosgene.

Otherwise, all chloride ions would be released to the solution.

It is known that most of chlorinated organic compounds are much more toxic than the corresponding hydrocarbon compounds due to the existence of chlorine substituents.

In treating chlorinated ethylenes, the ideal by-products and final products in the reaction

25 are carboxylic acids and CO: without any chlorine substituent. To evaluate the extent of

dechlorination of chlorinated ethylenes in more detail, the degradation of TCE was

examined by monitoring Cl ions through a kinetic experiment. Stoichiometry indicates

that the dechlorination of I M TCE (CiHCb) releases 3 M Cl' ions, or

Cl HCl3 —> CA ■¥ 3 Cl (2.7)

where CA is a group of carboxylic acids without chlorine substituents. When excess

Mn 0 4 ' is used for the oxidation of TCE, the degradation of TCE and the formation of

chloride ions can be described by

lC‘Ha,i _ ici-j _ , = g (a) — —------—— = / - e (b) [C 2 HCI3 J 0 3[C2HCI,1„ (2.8) where kobs is the pseudo-first-order rate constant confirmed previously.

A kinetic experiment was conducted using 0.06 mM TCE with more than 15 fold excess of Mn 0 4 (1 mM). The measurement of chloride concentration in the experiment was triplicated and the standard deviation is shown by error bars in Figure 2.6. About three times the concentration of Cl' ions as compared to transformed TCE was observed over time (Figure 2.6). As shown in Figure 2.6, the total chlorine mass is accounted as both chlorine substituents in untransformed TCE and chlorides liberated from transformed TCE. This result indicates that the amount of chlorine substituents remaining in intermediate products is negligible.

2 6 The pseudo-first-order rate constant kobs=7-0±0.5 x lO"^ s ‘ was obtained from a best fit to the data points of TCE concentrations using Eq. 2.10(a). Based on a kobs of

7.ÜX ID"* s*‘, both the degradation of TCE and the formation of C! with time, as predicted from Eq. 2.10 (see lines in Figure 2.6), coincided very well with the observations. The kinetic results from this experiment suggest that the complete dechlorination be achieved rapidly during the decomposition of the cyclic complex after TCE is transformed.

2.4.4 Change in acidity and pH dependence

The formation of carboxylic acids has been proposed to accompany the rapid decomposition of the cyclic complex. To examine the change in acidity of the solution due to the production of carboxylic acids, an experiment with a high concentration of

TCE (0.76 mM) reacting with 3.8 mM Mn 0 4 ' was conducted without the pH buffer. A drop in pH to 2-3 was observed within several minutes. Stewart (1965) pointed out that pH of the solution would be close to the pK of an acid produced in the reaction using permanganate. The carboxylic acids that likely formed would be oxalic acids, formic acids and so on. Oxalic acid has a pK|=l.23 and a pKz=4.19, and formic acid has a pK=3.75. A more detailed analysis of these carboxylic acids is now underway.

The effect of pH on the TCE degradation rate was studied over the pH range of 4-

8 at constant concentration of Mn 0 4 ( 1 mM). As shown in Figure 2.7, the pseudo-first- order rate constant, kobs, did not show strong pH dependence in the pH range 4-8. The differences in the rate constants at various pHs are not significant as compared to the experimental errors which are indicated by the error bars. A TCE degradation rate.

27 independent of pH, is in agreement with the suggestion that neither hydrogen nor hydroxyl ions significantly facilitate the attack by permanganate on C=C in TCE in an initial transformation step, where a cyclic complex forms. It is possible that pH affects the reaction afterwards. In a later step when hydrolysis and further oxidation proceed, it is well known that the decomposition of the complex is highly pH dependent and that both permanganate and hydroxyl ions compete for reaction with a common cyclic intermediate (Wiberg and Saegebarth, 1957). A typical example is diol formation in a basic solution and ketol formation in a neutral solution.

To obtain the TCE disappearance rate over the pH range 4 - 8 , log [TCE]/[TCE]o is plotted versus time for all data points from eight kinetic experiments at four pH levels with 1 mM Mn 0 4 *(Figure 2 . 8 ) . A linear regression with a determination coefficient of

0 . 9 8 8 gives a rate constant kobs=0.67±0.02 x 10'^ s '. A second-order rate constant, thus, can be simply estimated as k=0.67±0.02 M ' s ' by dividing 1 mM MnOT. TCE degradation by permanganate is rapid with a half life about 1 7 minutes.

2.4.5 Permanganate decomposition and products

Figure 2.9 shows the spectra of the reacting solution over time. The absorbance at both 525 and 546 nm at the beginning of the reaction (see a solid bold line in Figure 2.9) is representative of the initial concentration of Mn 0 4 (0.23 mM). The reduction in absorbance with time in a first-order fashion indicates a decreasing concentration of

Mn 0 4 '. A slight decrease in absorbances at 606 and 667 nm suggests that Mn 0 4 ~' and

Mn 0 4 ^' apparently do not form as reaction products. At a wavelength 418 nm, the

2 8 spectrum of the initial solution has the lowest absorbance and Mn 0 4 is almost

transparent to the source light. The increase of absorbance at 418 nm with time indicates

the formation of a decomposition product of Mn 0 4 *. Based on independent observations

when crotonic and cinnamic acid were oxidized by Mn 0 4 , respectively, both Wiberg et

al. ( 1973) and Lee and Brownridge (1973) interpreted a similar spectrum (a dashed line

in Figure 2.9) as indicative of cyclic hypomanganate ester. However, Simandi and Jaky

(1976) and Freeman et al. (1981) using an iodometric technique believed that it was a

manganate (IV) with the +4 oxidation state.

In our spectrophotometric examination, all final spectra representing a decomposition product of Mn 0 4 in TCE oxidation show a linear relationship between the

logarithm of the absorbance and the logarithm of the wavelength. A typical plot of log A

versus log X has been reproduced in Figure 2.10. The linear relationship with ihe slope of

-4.22±0.05 appears to be a reflection of Rayleigh’s law, which can be written as in Eq.

2. 11,

C A = (a) log A = log C - 4 log A (b) (2.9) A where A is the absorbance, A. the wavelength, and C a constant depending on the polarizability, cell path length, mass and concentration of the colloidal particles. As a consequence of Rayleigh’s law, the lost energy is due to light scattering by a product present in the form of colloidal particles. Figure 2.10, thus, confirms the suggestion that the product is actually soluble colloidal manganese dioxide.

29 2.4.6 Permanganate consumption in contaminated ground water

Permanganate is a powerful and reactive oxidant that would likely oxidize other

organic compounds existing in ground water or contaminated ground water. The

presence of these compounds would reduce the rate of TCE oxidation because of

competition for the permanganate. To compare TCE loss rate in various solutions

containing different quantities of other compounds, experiments were designed using 2

mg/L (0.0l5mM) TCE reacting with 20 mg/L (0.13 mM) MnO^ in the Milli-Q water,

ground water, and contaminated ground water, respectively. The contaminated ground

water was synthesized by mixing ground water with 2% landfill leachate. Figure 2.11

depicts the reduction of TCE degradation rate due to the consumption of permanganate

by other compounds in the solution. Based on the ratio of second-order rate constant in

ground water (kcw=0.61 M * s ') to that in Milli-Q water (kMQ=0.68 M ' s '), the TCE

degradation rate in ground water was slightly decreased by 1 0 % , compared to that in

Milli-Q water. In the ground water contaminated by 2% landfill leachate, the second-

order rate constant (kccw) is 0.44 M*' s '. The rate was reduced to 65 % of that observed

in Milli-Q water.

Although chlorinated ethylenes are the most frequently detected ground-water

contaminants at hazardous waste sites, they are often found with other organic compounds. To account for the effects of competition in multi-contaminant solutions is complicated by the fact that many different organic compounds could be involved and

many of which may not be identifiable. As a first-step in dealing with the issue of complex aqueous solutions, we provide results of several experiments involving ground

30 water and contaminated ground water, where the relative abundance of these other organic compounds is represented by total organic carbon (TOC).

Clearly, using TOC to represent other reactive compounds assumes that all compounds present in solution would react with permanganate is able to oxidize most organic contaminants characterized by carbon double bonds (e.g. most alcohol, ketones, organic acids, phenolic compounds, and humic substances). Apparently, the competition of TOC for permanganate varies from site to site. However, the simplified kinetic- reaction model, which accounts for the general effect of other organic compounds in the reaction, would be a quick and an appropriate approach to evaluate the site-specific consumption of permanganate by TOC for the field application.

A series of experiments were designed using 20 mg/L (0.13 mM) MnO^ reacting with excess TOC at five different concentrations from 302 to 506 mg/L. MnO+', as a monitoring species, was measured over time. The MnO^' disappearance rate is the greatest at the beginning of the 30-minute reaction period. The maximum rate constant was obtained by fitting the initial data (at the first 1 0 % of reaction time) using a simple pseudo-first-order kinetic equation. Figure 2.12 is a plot of the derived decomposition rate constant against TOC concentration. The rate constant (kobs) depends almost linearly on TOC concentration. The over-all kinetic behavior of TOC reacting with Mn 0 4 in this case is close to second-order. To simplify the evaluation of Mn 0 4 decomposition by

TOC, a second order kinetic equation was used and the second order rate constant k =

3.40x10^ L/mg s ' was obtained from the slope of a linear fitting line in Figure 2.12. If a composite molecular weight is effective in representing TOC in the reaction, the

31 maximum rate constant for Mn 0 4 decomposition by TOC is in the range from 0.17 to 1 .02 M"' s ' when the composite weight is between 50 and 300 g. This rate constant is similar to that observed for TCE.

To predict the TCE degradation in the system involving TOC, TCE and MnO^. the rate equations are simplified based on the over-all kinetic behavior of TOC in the contaminated ground water we tested,

dfTCE] = -kTCE[TCE] [M n O ,l [TCEJ(t = 0) = [TCE 1, dt d[TOC] ^-kToclTOCJ [MnO',1 [TOCJ(t = 0) = [TOC ]„ dt d[M nO ,J =-(kTca[TCE] + kToc[TOC])[MnQ-J, [M nO:j(t = 0) = [ M nO'Ja

(2. 10)

The equations can be generalized as follows dv- — = f i(t, y,. >2 , yj), yi(t = 0) = y.^, i= I, 2, and 3. (2 .1 1 )

and solved numerically using the Runge-Kutta method with adaptive step-size control. In this case, the molar concentration of TOC is unknown without a composite molecular weight for TOC. However, it can be find from experimental data by solving an inverse problem.

Three experiments were conducted in a highly contaminated ground water containing 101 mg/L TOC. TCE at a fixed initial concentration of 2 mg/L reacted with three different concentrations of MnO 4 (20, 80, and 120 mg/L). The data points in

Figure 2.13 illustrate the TCE degradation at various concentrations of Mn 0 4 . To model

32 the experimental data using Eq. 2.13, an initial composite weight of 50 g was assigned to compute TCE degradation with time and the residual between predicted TCE and observed TCE. The automatic adjustment of the composite weight with a fixed step was made until a minimum residual was calculated. A composite molecular weight, 176 g, for TOC was obtained based on the first experimental data (20 mg/L MnO^l. The simulated results with a composite weight of 176 g match the observed rate of TCE degradation well. In the second experiment, where 80 mg/L Mn 0 4 was used, the simulated TCE-time curve is consistent with that observed in the first hour but deviates at later time. The same behavior was evident in the third experiment in which a relatively high concentration of Mn 0 4 (120 mg/L) was used. However, the later-time deviation in the third experiment is much larger than that in the second. The observed TCE loss with time in the third experiment is close to the calculated result for Milli-Q water without

TOC (the dashed line in Figure 2.13). Apparently, reduction in the TCE degradation rate due to the presence of other organic compounds was much less than the expected when a high concentration of Mn 0 4 was used.

These results do not necessarily imply that some of the MnÛ 4 ‘ was not being used in competing reactions with TOC. We believe that there was an actual increase in the

TCE degradation rate. In effect, the increase in TCE degradation rate masked the effect of competition with TOC. The increased TCE loss rate could be caused by the large amount of colloidal MnOz produced in the reaction with the high concentrations of TOC and Mn 0 4 . In the third experiment, the calculations indicate that 91 mg/L (0.58 mM)

Mn 0 4 was lost and converted to MnOi. Meanwhile, 2.15 mg/L (0.016 mM) TCE and 99

33 mg/L of the other organic compounds represented by TOC were transformed. The MnO% has been reported to catalyze reactions between Mn 0 4 and many organic compounds

(Pérez-Benito et ai., 1987; Pérez-Benito and Arias, 1991). A further study is now underway to examine autocatalysis of MnOi on TCE degradation rates.

2.5 Conclusions

This study shows that chlorinated ethylenes can be rapidly degraded by permanganate in aqueous solution. The half lives of TCE, cis-1,2-DCE, trans-1,2-DCE and 1,1-DCE with 1 mM Mn 0 4 ' range from 0.4-18 minutes. The half life of PCE, however, is much longer, about four hours. In PCE degradation, the attack of permanganate ion, as an electrophile, is slowed by the deficiency of electrons in the carbon double bond induced by four chlorines in PCE. The ratio of kcis to ktnms (0.03) represents a much higher reactivity of trans-DCE than that of cis-DCE. The significant steric effect of cis substituents on reaction rate supports the postulate that a five- membered cyclic complex is formed during the transition state and leads to an intermediate product, the cyclic hypomanganate ester.

Extensive kinetic studies of TCE oxidation by permanganate suggest that TCE degradation is a second-order reaction. The reaction can be reasonably described as;

C.HCl^ + M nOi —^ I —^ CA + MnO, + 3C /' (2.12) where I is a cyclic complex, CA is carboxylic acids, k| is a second-order rate constant, and k% is an unknown-order rate constant. The permanganate attack, as an electrophilic

34 addition, on the C=C bond in TCE, leads to the formation of a cyclic complex in the first

reaction step. The fact that the rate constant ki is independent of pH over the range 4-8

indicates that this attack is not affected by either hydrogen or hydroxyl ions. However,

the decomposition of I in the second step may involve hydrogen or hydroxyl ions as

suggested by numerous studies on the oxidation of other organic compounds by

permanganate in chemical synthesis. The reaction in the second step proceeds rapidly. A

dechlorination of the complex (I) over time was observed in the kinetic experiments. The

results show that Cl' ions are completely liberated from I immediately after its formation.

Because the rate constant ki is much greater than ki, the liberation of Cl' ions in the

kinetic experiment is in excellent agreement with the calculation based on a second-order

reaction model using k|. Therefore, the reaction rate in Eq. 2.14 can be approximated by the rate-limiting step (the first step) using the second-order rate constant k|. Over the pH range 4-8, a k, value of 0.67 ± 0.03 M ' s'" was calculated based on eight experiments.

The essentially complete dechlorination and possible formation of organic acids suggest that the degradation products of chlorinated ethylenes are much less harmful than parent compounds and miscible with water. In an in situ scheme-based permanganate flushing, those products would be readily removed from ground water as flushing proceeds. Research is continuing to identify the carboxylic acids and to elucidate the pathways of formation and further oxidation of acids.

In TCE oxidation, spectrophotometric evidence shows that permanganate is reduced to form soluble colloidal manganese dioxide. The consumption of permanganate in ground water with low TOC is limited. Ground water contaminated by landfill

35 leachates, however, will consume permanganate depending on the TOC level. The second-order rate constant for permanganate consumption by TOC is about the same order as measured for TCE if a composite molecular weight is assumed in the range from

50 to 300 g.

TCE degradation by permanganate at concentrations of 20, 80, and 120 mg/L

(0.13,0.51, and 0.76 mM) in contaminated ground water containing 101 TOC mg/L was modeled with a system of ordinary deferential equations (Eq. 2.13) and solved using the

Runge-Kutta method. The simulated results are consistent with the observations at the early time but deviate from those at the later time, especially in the experiment with a relatively high concentration of permanganate. A greater TCE degradation rate than expected at later time might be caused by autocatalysis on the surface of colloidal MnO% produced in the reaction. Competition of TOC for permanganate would be offset by the presence of a large quantity of MnOj, which promotes the reaction between TCE and permanganate.

2.6 References

Arndt, D., 1981. Manganese compounds as oxidizing agents in organic chemistry. Open

Court Publishing Company, La Salle, Illinois.

Burdon, J., and Tatlow, J. C., 1958. The reactions of highly fluorinated organic

compounds. X. The oxidation of fluoro-olefins by potassium permanganate in

acetone. Journal of Applied Chemistry, 8:293-296.

36 Casado, J., Lôpez-Quintela, M. A., and Lorenzo-Barral, F. M., 1986. The initial rate

method in chemical kinetics. Journal of Chemical Education, 63:450-452.

Cotlove, E., 1958. An instrument and method for automatic, rapid, accurate and sensitive

titration of chloride in biologic samples. Journal of Laboratory and Clinical

Medicine, 51:461-468.

Gardner, K. A., and Mayer, J. M., 1995. Understanding C-H bond oxidations: H- and H

transfer in the oxidation of toluene by permanganate. Science, 269:1849-1851.

Gates, D. D, Siegrist, R. L., and Cline, S. R., 1995. Chemical oxidation of volatile and

semivolatile organic compounds in soil. Proceedings of 88th annual meeting and

exhibition, San Antonio, Texas.

Gillham, R. W., and O'Hannesin, S. P., 1994. Enhanced degradation of halogenated

aliphatics by zero-valent iron. Ground Water, 32:958-967.

Glaze, W. H., and Kang, J. K., 1988. Advanced oxidation processes for treating

groundwater contaminated with TCE and PCE: laboratory studies. Journal

AWWA, 80:57-63.

Gonullu, T., Farquhar, G. J., Truax, C., Schnarr, M. J., and Stickney, B., 1997. Studies

on the use of permanganate to oxidize chlorinated solvents in soil. Journal of

Contaminant Hydrology, in press.

Hoigné, J., and Bader, H., 1983. Rate constants of reaction of ozone with organic and

inorganic compounds in water, I. Non-dissociating organic compounds. Water

Research, 17:173-183.

37 Jeffers, P. M., Woytowitch, L. M„ and Wolfe, N. L., 1989. Homogeneous hydrolysis

rate constants for selected chlorinated methanes, ethanes and propanes.

Environmental Science and Technology, 23:965-969.

Freeman, P., 1975. Possible criteria for distinguishing between cyclic and acyclic

activated complexes and among cyclic activated complexes in addition reactions.

Chemical reviews, 75:439-491.

Freeman, P., Puselier, C.O., Armstead, C.R., Dalton, C.E., Davidson, P.A., Karchefski,

E.M., krochman, D.E., Johnson. M.N., and Jones, N.K., 1981. Permanganate ion

oxidation. 13. Soluble manganese (IV) species in the oxidation of 2,4( 1H.3H)-

pyrimidinediones (Uracils). Journal of American Chemical Society, 103:1154-

1159.

Kriegman-King, M. R., and Reinhard, M., 1992. Transformation of carbon tetrachloride

in the presence of sulfide, biotite, and vermiculite. Environmental Science and

Technology, 26:2198-2206.

Kriegman-King, M. R., and Reinhard, M., 1994. Transformation of carbon tetrachloride

by pyrite in aqueous solution. Environmental Science and Technology, 28:692-

700.

Lee, D. G., and Brownridge, J. R., 1973. The oxidation of cinnamic acid by

permanganate ion. Spectrophotometric detection of an intermediate. Journal of

American Chemical Society, 95: 3034-3035.

38 Lee, D. G-, 1982. Phase transfer assisted permanganate oxidations. In Oxidation in

organic chemistry, W. S. Trahanovsky, ed.. Academic Press. New York, Part D.

Charpter 2, pp 147-206.

Mata-Pérez, F., and Pérez-Benito, J. P., 1985. Identification of the product from the

reduction of permanganate ion by trimethylamine in aqueous phosphate buffers.

Canadian Journal of Chemistry, 63:988-992.

Matheson, L. J., and Tratnyek, P. G., 1994. Reductive dehalogenation of chlorinated

methanes by iron metal. Environmental Science and Technology, 28:2045-2053.

Pérez-Benito J. P., and Arias, C., 1991. A kinetic study of the permanganate oxidation of

triethyl amine. Catalysis by soluble colloids. International Journal of Chemical

Kinetics, 23:717-732.

Pérez-Benito, J. P., Mata-Pérez, P., and Brillas, E., 1987. Permanganate oxidation of

glycine: kinetics, catalytic effect, and mechanisms. Canadian Journal of

Chemistry, 65:2329-2337.

Simandi, L. I., and Jaky, M., 1976. Nature of the detectable intemediate in the

permanganate oxidation of trans-cinnamic acid. Journal of American Chemical

Society, 98: 1995-1997.

Stewart, R., 1964. Oxidation mechanisms. Benjamin, New York, pp 58-76.

Stewart, R., 1965. Oxidation by permanganate. In Oxidation in organic chemistry.

Wiberg, K. B.,ed., Academic Press, New York, Part A, Charpter 1, pp 1-68.

39 Szammer, J, and Jaky, M., 1992. Oxidation by permanganate in strong alkaline medium.

Oxidation of ethane-1,2-dioI, glycol aldehyde, glycollic acid, and glyoxylic acid.

International Journal of chemical kinetics, 24:145-154.

Vella, P. A., and Veronda, B., 1992. Oxidation of trichloroethylene: comparison of

potassium permanganate and Fenton’s reagent. In Chemical oxidation

technologies for the nineties, Eckenfelder, W. W., ed., Technomic publishing,

Lancaster, Basel.

Wagner, G., 1895. History of oxidation reaction of unsaturated compounds. Journal of

Russian Physical-Chemical Society, 27:219-236.

Wiberg, K. B., Deutsch, C. J., and Rocek, J., 1973. Permanganate oxidation of crotonic

acid. Spectrometric detection of an intermediate. Journal of American Chemical

Society, 95: 3034-3035.

Wiberg, K. B., and Saegebarth, K. A., 1957. The mechanisms of permanganate

oxidation. IV. Hydroxylation of olefins and related reactions. Journal of

American Chemical Society, 79: 2822-2824.

40 pH range Half-cell reactions (Mn) E" (volts) >12 Mn04+e=Mn04-' 4-0.56

3.5-12 Mn 0 4 4- 3e" -i- 2HzO = MnOi -t- 40H 40.59

Mn 0 4 4- 3e" 4- 4H^ = MnO? 4- 2 H2O 4-1.70

<3.5 Mn 0 4 4- 5e' 4- 8H+ = Mn"^ 4- 4 H2O 4-1.51

Table 2.1 Half-cell redox of manganese ions at various pH ranges.

Chlorinated kobs ( 10"* s ' ) Determination coef. T |/2 (min) Ethylenes of regression (r^) PCE 0.45±0.03 0.924 256.7 TCE 6.5±0.1 0.997 17.8 cis-DCE 9.2±0.5 0.976 12.6 trans-DCE 300±20 0.991 0.4 I, I-DCE 23.8±1.3 0.980 4.9

Table 2.2 Rate constants and half lives for the oxidative degradation of chlorinated ethylenes by permanganate ( 1 mM).

41 Reaction Scheme (in neutral and weakly acidic conditions) OH OH O OH O H MnO/ I II MnO- : I c - c C - C / \ H.0 '* I I i H.0 H H o o O H 6 OH \ / Glycol Aldehyde Glyoxylic Add Oxalic Add Mn / \ H H H 9 9 H MnO/ ) c - c (

CycOcHyponrangnale O o Ethylene EMer 2 H li H-C-H H.0 H-C-OH Formaldehyde FoimicAcid

3

Figure 2.1 Reaction scheme. The oxidation of ethylene in a neutral to weak acidic condition.

42 Injection Port Cl Ion Analyzer

Septum

Sampling Port W Silicon pipe

Septum Teflon Clamp to Seal the System

Glass Reactor Water Bath

Water Bath Stir-Bar

Water Jacket

Figure 2. 2 Schematic diagram of the bath reactor. In kinetic experiments, each sample aliquot was withdrawn from the sampling port by a syringe when the same volume of solution as the aliquot was injected via the injection port. Zero headspace was maintained at all times.

43 - 7.2

-7.4

CDO

-7.6

Log r,= -3.1843 + 1.0085 Log [TCE], (R'=0.998)

-7.8

-4.6 -4.5 -4.4 -4.3 -4.2 -4.1 -4.0 Log TCE, (M)

Figure 2.3 Plot of initial rates versus initial concentration for eight kinetic experiments. TCE ranging from 0.031 to 0.083 mM was oxidized by 1 mM Mn 0 4 at pH 7.1. A slope a = 1.01 ± 0.02 is first order with respect to TCE.

44 - 3.0

-3.2

-3.4 — O) o i

-3.6

Slope = 1.05 t 0.03 (R' = 0.993)

-3.8

-3.5 -3.4 -3.3 -3.2 -3.1 -3.0 -2.9 Log [MnOJo (M)

Figure 2.4 Plot of pseudo-order rate constant kobs versus initial concentration of permanganate. TCE at 0.078 mM was oxidized by MnO^ concentration varying from 0.37 to 1.2 mM at pH 7.1. A slope p = 1.05 ± 0.03 is first order with respect to MnOj'.

45 0.5

0.0

-0.5

o 1 : -1.0 + -f Chlorinated ethylenes k«.(10 's ') R: OJ JI l \! + 1,1-DGE 23.8 0.980 -1.5 □ trans-DCE 300 0.991 1 A cis-DCE 9.2 0.976 -2.0 J O TCE 6.5 0.997 o PCE 0.45 0.924

-2.5

100 200 300 400 Time (min)

Figure 2. 5 Degradation of chlorinated ethylenes by Mn 0 4 ( 1 mM), at pH 7.1 . Lines represent best Fits using a pseudo-first-order kinetic model.

46 1.2

+ 1.0 + # LU ï * O ? 0.8 : l . ü

- 0.6 —

X ü ♦ TCE 1 Cl ! 0.4 # e + Cl Mass Balance ! 1 ü 1

20 40 60 80 100 Time (min)

Figure 2.6 Liberation of CT ions in the TCE oxidative transformation by 1 mM Mn 0 4 at pH 7.1. The error bars for CT ions are the standard deviation of triplicate samples. The solid lines were calculated from Eq. 2.10 with kohs obtained from a best fit to TCE observations using Eq. 2.10(a).

47 1.0

0.8 —I

# 0.6 — ! CO o

0.4 1

0.2

0.0 — | -

4 pH

Figure 2. 7 TCE degradation rate constant (kobs) with I mM MnÛ 4 in the pH range 4-8. Error bars are standard deviations of kobs-

48 0.0

= 0.67 ± 0.03 x1 O ' s ' (r"= 0.988 T,,j= 17.3 ± 0.5 min.

-0.5

Ü Ü LU Ü I— O) o pH 4.2 pH 6.3 pH 7.1 pH 8.0

- 2.0

20 40 60 80 100 120 Time (min)

Figure 2.8 Pseudo-first-order plot of TCE transformation in 1 mM MnO^ over the pH range of 4-8. Rate constant kobs = 0.67 ± 0.03 x 10'^ s'' and half life T,/^ = 17.3 ± 0.5 minutes.

49 0.6

0.4

O(D C CO JD O CO <

0.2

0.0

400 450500 550 600 650 700 Wavelength (nm)

Figure 2. 9 Overlay of UV-vis spectra at time intervals of 10 minutes during TCE (3.81 mM) oxidation by Mn 0 4 * (0.23 mM) at pH 7.1. The solid bold line is the initial MnO^ spectrum and the dashed line is the final spectrum due to a product from the decomposition of MnO^.

50 0.00

-0.40 —

- 1.20 —

log A = 10.72 - 4.22 log À (r" = 0.996)

-1.60 2.60 2.65 2.70 2.75 2.80 2.85 log A. (nm)

Figure 2.10 Log A versus log X for the spectrum recorded at 80 minutes after the reaction between M nO/ (0.23 mM) and TCE (6.85 mM) at pH 8.2. The linear relationship with a slope =-4.22±0.05 obeys Rayleigh’s law.

51 0 .2

Ground water with 2% leachate Ground water 0.0

- 0.2

O O LU -0.4 O h- O) o r t - 0.6

- 0.8

- 1.0

Time (hour)

Figure 2.11 Comparison of TCE degradation by 0.13 mM (20 mg/L) Mn 0 4 in Milli-Q water, ground water, and contaminated ground water (2% leachate). Lines represent best fits using a pseudo-first-order kinetic equation.

52 4.0

Slope = 3.4 X 10® (r"= 0.90)

3.0 -

to o

2.0 H

1.0

200 300 400 500 600 TOC (mg/L)

Figure 2.12 Plot of the maximum pseudo-first-order rate constant (kobs) versus TOC concentration.

53 10 ^

_l cn Ê

LU O H-

0.01

0 80 mg/L MnO, 1 I 120 mg/L MnO,

0.001

0 1 2 3 4 5 6 7 Time (hour)

Figure 2. 13 TCE degradation by Mn 0 4 ' at various concentrations in a synthesized contaminated ground water containing 101 mg/L TOC. The solid lines were calculated using kxcE = 0.67 s ' and kjoc = 0.60 M*‘ s '. The dashed lines were calculated for TCE degradation in Milli-Q water without TOC.

54 CHAPTER 3

KINETICS AND MECHANISM FOR TCE OXIDATION BY PERMANGANATE

3.1 Introduction

In the past few years, there has been considerable interest in the in-situ oxidative degradation of chlorinated ethenes with a metal-oxo reagent. In terms of a remedial strategy, the most important advantages of the method are the rapid degradation of chlorinated ethenes, the non-reactive nature of the reagent with carbonate and bicarbonate, the ease of field implementation, and the relatively low costs (Vella and

Veronda, 1992; Valla, 1996; Gonullu et al., 1997; Yan and Schwartz, 1995; and Gates et al., 1995). An increasing number of research groups are working to assess the utility of these reactions in the treatment of ground water and soil contaminated by chlorinated ethenes like tetrachloroethene (PCE), trichloroethene (TCE) and dichloroethene (DCE).

Use of potassium permanganate in treatment schemes started in early 1980’s and showed promise in the destruction and/or detoxification of a wide range of environmentally harmful chemicals in industrial waste water (Walton et al., 1991 ). Since early I990’s, specific studies have focused on the oxidative destruction of chlorinated ethenes in ground water and soil with permanganate. Preliminary batch and column

55 experiments showed that degradation is rapid and effective across a wide pH range (Vella and Veronda, 1992). More than 90% of the TCE and PCE was degraded after flushing column with several pore volumes of permanganate solution (Gonullu et al., 1997). On the basis of these results, a pilot-scale field study was initiated using aqueous permanganate flushing to treat an simulated spill of eight-liters of a TCE/ PCE mixture at

Canadian Forces Base Borden. Both TCE and PCE were effectively oxidized as evidenced by a roughly stoichiometric increase in dissolved chloride (Truax, 1993).

About 90% of PCE and TCE was transformed in 10 months.

A process-oriented study was conducted by Yan and Schwartz (1998), to provide a more detailed understanding of the reactions. Their study reported degradation kinetics for five chlorinated ethenes including PCE, TCE, and three isomers of DCE. TCE degradation was postulated by Yan and Schwartz (1997) to involve the following three sequential reactions:

C,//C /, + MnO^' —^ I — ^ CA + MnO. + 3C r (3.1)

CA + MnO; —^ CO. + MnO. (3.2) where I is a cyclic complex, CA are various carboxylic acids, and k[, ka, and kg are rate constants. In the first reaction, TCE disappearance was found to be independent of pH.

Complete chlorine liberation was observed and believed to occur as a consequence of the second reaction. However, the second and third reactions were not studied in detail. It is

56 well known that oxidation of alkene and its products are highly pH dependent and that

both permanganate and hydroxyl ions may compete for a common cyclic intermediate in

the reaction (Wiberg and Saegebarth, 1957). To fully understand kinetics and mechanism

of the process of TCE oxidation by permanganate, the author investigated the products

and kinetics of the second and third reactions. The specific objectives of this study were

(I) to identify and quantify reaction products, ( 2 ) to elucidate reaction pathways and their

pH dependence, and (3) to determine the rate constants for reactions in KMnO^-TCE-HiO

system.

3.2 Experimental Methods

3.2.1 Materials

TCE (C2HCI3, 99.5+%) was obtained from Aldrich Chemical Co. (Milwaukee,

WI) and used as received. Radiolabeled [l,2-'“^C] TCE was purchased from Sigma

Chemical Co. (St. Louis, MO) and its specific activity was 3.1 mCi/mmol with

radiochemical purity 98+%. '“^C-labled TCE at 31.5 |imol was diluted with methanol in a

lO-ml glass volumetric flask and stored at 4“C in the dark. High purity pentane of GC grade (Burdick & Jackson) from Baxter Diagnostics Inc. (McGaw Park, IL) was used as a

liquid-liquid extraction solvent for the TCE analysis. Potassium permanganate stock solutions at 3 mg/mL were prepared by dissolving KMn 0 4 crystals in phosphate-buffered

Milli-Q water at pH 4 , 6 and 8 . The stock solutions were stored in brown glass bottles and used freshly. The standards for four organic acids were prepared by the dilution of

57 glycolic acid (99%), sodium glyoxylate (99%), sodium oxalate (99.5%), and sodium formate (99.5%) from Huke (Buchs, Switzerland). The eluent for the analysis of organic acids was 4 mM sulfuric acid, prepared by making appropriate dilutions of analytical- reagent grade sulfuric acid. The eluent was degassed when it was used. The reducing agent, thiosulfate, is research grade and was prepared as a stock solution (6 mg/mL) for quenching the reaction in kinetic experiments.

3.2.2 Product studies and kinetic experiments

In a TCE oxidation reaction with permanganate, carboxylic acids and carbon dioxide would be expected as major intermediate and final products (Yan and Schwartz,

1998). In this study, two types of experiments were performed to identify the carboxylic acids and to trace the formation of CO 2 .

Kinetic experiments with analyses of TCE, carboxylic acids, and carbon dioxide were used to identify and quantify the various products over time in a TCE oxidation reaction. Experiments were conducted by mixing the oxidant with each TCE test solution. The reaction was allowed to proceed for some specific time, and then was quenched by adding a reducing agent. Solutions of TCE, permanganate, and the quenching agent were adjusted to the desired pH prior to the start of each experiment.

Drift in pH was limited to a maximum of 0.2 during the experiments.

Two kinetic tests were employed. In each test, three experiments were conducted at three different pHs, 4,6, and 8. Each experiment has one set of vials with two or three

58 replicates and one control vial containing only TCE aqueous solution. For investigating carboxylic acid distributions over time, a 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 was initiated by injecting 5-ml of permanganate of

6.3 mM through the septum. A second needle was used to allow an equal volume of TCE solution to be displaced, in order to keep the headspace free. All the vials were then placed in a vibratory shaker until sampling was required. At each sampling time, one vial was taken. A I-ml aliquot of reaction solution from the vial was transferred to a volumetric flask with appropriate dilution following immediate TCE analysis. To quench the reaction, 1-ml thiosulfate from the stock solution was added to the vial. The quenched solution was centrifuged at 3000 rpm for 20 minutes and filtered by 0.2-p.m filter to separate the precipitated manganese dioxide. After filtration, the solutions were withdrawn for carboxylic acid analysis using the method described below.

To trace the CO? production in the reaction, a second kinetic test was undertaken with radiolabeled ^‘^C TCE. To maintain the same initial conditions as the first kinetic test, experiments began with the addition of 0.1-ml permanganate stock solution to a 3-ml test tube filled with a 2.9-ml TCE solution of 0.093 mM. Each experiment involved one set of test tubes with duplicates. At predetermined interval, the reaction was quenched by adding 0 .1-ml thiosulfate from the stock solution. Once the reaction was quenched, the test solution was taken for CO? analysis.

59 3.2.3 Chemical analysis

(a) TCE Analysis

A 1-ml sample taken from the test solution was diluted appropriately by 10-100.

depending on the concentration of TCE. A 10-mL sample after dilution was immediately

transferred to an extraction vial containing 4 mL of pentane. The vial was shaken for one

minute and equilibrated for 10 minutes. The extractant was analyzed for TCE using a

Fisons Instruments 8060 gas chromatograph equipped with a Ni^^ electron capture detector and a DB-5 capillary column (J&W Scientific, Rancho Cordova, CA), 30 m x

0.32 mm I.D, with a film thickness of 1.0 pm. Helium was used as the carrier gas and nitrogen as the make-up gas. The gas chromatograph was calibrated daily with a minimum four calibration standards, and duplicate measurements were made for each sample or standard. If the standard deviation of measurements was greater than 10 %. another measurement would be made.

(h) Organic Acid Analysis

Acidic products were analyzed on a Waters high performance liquid chromatography (HPLC) fitted with a Bio-Rad Aminex HPX-87H strong cation-exchange resin column (300 x 7.8 mm I.D.). Samples were injected through a 20-pl sample loop valve into the liquid chromatographic system. The column was operated at room temperature using degassed dilute sulfuric acid (4 mM) as eluent at a flow rate of

0.6ml/min. The column effluents were monitored by an UV spectrophotometric detector

60 (Waters 486) at 210 nm. The products were identified and quantified by comparison with external standards.

{c) CO2 Analysis

CO2 was determined by retaining CO 2 under basic conditions and stripping CO 2 under acidic conditions. This method was based on and adapted from Kriegman-lCing and Reinhard ( 1992). Three 0.8-ml aliquots were withdrawn from each test tube.

Aliquot A was acidified with 0.3 ml of 1 N H 2SO4 and purged with N 2 for 20 minutes to strip all volatiles and CO 2 from the solution. Aliquot B was treated with 0.3 ml of 1 N

NaOH and purged for 20 minutes to retain CO 2 and to strip only non- CO 2 volatiles. After purging both aliquots were added to 10 ml of liquid scintillation cocktail and then counted by liquid scintillation counter. Measured counts per minute were converted to disintegrations per minute using the external standard method.

3.3 Results and Discussion

3.3.1 Products

In the kinetic experiments for identifying product distribution over time, 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. The liquid chromatograph shown in Figure 3.1 came from the samples collected one hour after the reaction started in 8-hour kinetic experiments. Four carboxylic acids, formic, oxalic, glyoxylic, and glycolic acids, were identified in the

6 1 system as intermediate products. Either formic or oxalic acid predominated, depending on pH. In general, a maximum of 43% initial TCE was converted to either formic or oxalic acid and up to 25% of TCE was transformed to glyoxylic acid. Glycolic acid was measured in very small quantities (<2%) in TCE transformation.

The radiolabeled product analysis indicates that the majority of TCE was transformed to CO?, the final product. Assuming TCE is stoichiometrically converted to

CO2, 57-88% of the initial TCE, varying depending on pH, was converted to CO 2 at the time when the experiments were terminated at 8 hours.

With an overall understanding of alkene oxidation by permanganate from organic synthesis chemistry (Wiberg and Saegebarth, 1957; Stewart, 1964; Freeman, 1976; and

Lee and Chen, 1989), it is possible to propose chemical transformation pathways for TCE oxidation (Figure 3.2). The products and intermediates in the shadowed boxes were identified in the experiments. TCE oxidation is initiated by the attack of permanganate ion, as an electrophile, on the carbon-carbon double bond (Yan and Schwartz, 1998). An organometallic compound, cyclic hypomanganate ester (2) in Figure 2, is formed via an activation complex in transition state, which has been postulated by many authors

(Freeman, 1976; Littler, 1971; Sharpless et al., 1977; and Lee and Brown, 1982). The rapid decomposition of cyclic ester (2) can follow several different pathways in an aqueous system such as oxidative hydrolysis and hydrolysis.

Oxidative hydrolysis transforms cyclic hypomanganate ester (2) rapidly to cyclic manganate (VI) ester (3) following electron transfer, subsequent fragmentation, and

62 hydrolysis to form formic acid (5) via 4. Two other possible pathways involve the

hydrolysis of cyclic ester (2) to acyclic hypomanganate (V) ester (6). The acyclic

manganate (VI) ester (7), oxidatively hydrolyzed from 6, may undergo rapid electron

transfer and then either hydrolysis to form glycolic acid (9) via 8 or release of one

hydrogen chloride and hydrolysis to form glyoxylic acid (12) or oxalic acid (13) via 11

In weak alkalinic to alkalinic solutions, acyclic ester (6) may simply hydrolyze to

trichloroglycol (10), which rapidly releases two hydrogen chlorides. This compound is

subsequently hydrolyzed to either glyoxylic acid (12) or oxalic acid (13). Based on the

large quantity of COi produced over time in kinetic experiments, carboxylic acids are not

final products that are observed in oxidation of many other organic substrates. All

carboxylic acids may be further oxidized to carbon dioxide at a relatively slow rate.

The oxidative hydrolysis of hypomanganate (V) ester to manganate (VT) ester

proceeds rapidly in aqueous media and involves both oxidation by permanganate ion and

disproportionation of hypomanganate, as suggested by Wolfe et al. ( 1981 )

H2O S '° Æ - o - ' " r ° l = o (3.3) fast  ' ° ' .  ' ° ' .  ' ° ' . IV VI where the intermediate Mn (IV) ester can be oxidized to Mn (VT) ester in the presence of excess permanganate (Wiberg et al, 1973).

63 MnQi S '° Æ = o (3.4) fast  ' ° ' . . 4 ' ° . IV VI

The liberation of chlorine substituents is hypothesized as a series of reactions, including the transformation of hypomanganate ester to manganate ester, the electron transfer to release hydrogen chloride, and hydrolysis. A stoichiometrically consistent release of chloride was observed along with disappearance of TCE (Yan and Schwartz,

1998). Thus, the decomposition of the ester with chloride release is a very rapid process, which is much faster than its formation.

3.3.2 Kinetic equations for TCE oxidation

TCE oxidation by permanganate ion generally involves the formation and decomposition of cyclic hypomanganate ester (CHME), and the oxidation of carboxylic acid, based on the TCE transformation pathways in Figure 3.2. A second order reaction for the formation of CHME was suggested by Yan and Schwartz (1998):

CzH Ch + MnÛ4 -> CHME (3.5) where CHME is cyclic hypomanganate ester and kip is a second-order rate constant. The decomposition of the cyclic ester via various pathways to form major carboxylic acids, and the release of chlorides can be described by equations, 3.6-3.8:

aCHME + P H 2 O > 2aHCOOH+aMnO^ + 3aCr+ yH ^ (3.6)

aCHME +t]H,0 —^^aOHC-COOH+aMnO,+3aCr-\-vH* (3.7)

64 aCHME > aHOOC-COOH+a MnO,+3aCr -ti;H* (3.8) where a, |3, y, r|, v, and Ç are stoichiometric coefficients and kia, kab, and k^c are rate constants for the formation of formic, glyoxylic , and oxalic acids, respectively.

The oxidation of formic, glyoxylic, and oxalic acids by permanganate involves a series of reactions. Most studies, however, have shown that the reaction at a rate limiting step can be described by second-order reaction for formic and oxalic acids (Taylor and

Halpem, 1959; Berka and Koreckova, 1973; Rodriguez and Sanchez Burgos, 1975; and

Mahmood and Begum, 1975). By analogy with the oxidation of formic and oxalic acids, it is reasonable to assume a second-order reaction for glyoxylic acid because its structure is similar to the other two acids. The rate law for these oxidation reactions is

d[HCOOH], -=-k,^[HCOOfn,[MnO,~] (3.9) dt

d[OHC-COOH], =-k^^ [OHC-COOH], [MnO,-] (3.10) dt

d[HOOC-COOH], [HOOC-COOH], [MnO^-] (3.11) dt where ksap. ksbp, and kscp are second-order rate constants and the subscript t for each acid denotes the total concentration of all species pertaining to that acid.

Because all of our kinetic experiments were conducted with an excess of Mn 0 4 in aqueous media with a fixed pH, the order of kinetic reaction can be reduced. The rates of disappearance/appearance for TCE, major intermediates and final product can be written as

65 k,=k,^[MnOi\ (3.12) dt

d[CHME\ =k, [C,HCL,\-k, [CHME\ k, =k,^ +A:,., (3.13) dt

d[HCOOH], = 2k,JCHME]-k,^ [HCOOHl k,, =k,^^ [MnO,'] (3.14) dt

d[OHC-COOHV —------^-^=k,_,[CHME]-k,, [OHC-COOHl k,, [MnO,~] (3.15)

d[HOOC-COOH], —------=k.^[CHME]-k,^ [HOOC-COOHl it,, =it„^ [MnO^] (3.16)

=k,^[COOHl +2k,, [OHC-COOH], +lk,„ [HOOC-COOH], (3.17) where ki, ksa, kab, and kac are pseudo-first-order rate constants and kaa, k%b, and k?c are first-order rate constants.

3.3.3 Estimation of rate constants

To find the rate constants for transformation and/or formation of TCE, three carboxylic acids, and CO 2 , equations of 3.12-3.17 were solved analytically using Laplace transformation. The solutions for TCE, formic, glyoxylic, and oxalic acids in terms of normalized concentration are:

[t c e i

66 [HCOOHl g-*:' [t c e i

(3.19)

[OHC-COOHl =A:,fe.26 [TCEl (^2 ~^|)(^36 ~ ^|) (^1 - ^ 2 X ^ 3 6 - ^ 2 ) (^1 -^3*)(^2 ~^3*) /

(3.20)

[HOOC- COOHl -w -kx^t = ^l^2c [TCEl ^(^2 ~ ^ l ) ( ^ 3 c ~ ^ l) (^I ~ ^ 2 X ^ 3 c “ ^ 2 ) (^1 ~ ^ 3 c -)(^ 2 ~^3c) ; (3.21)

The general form for equations, 3.19-3.21, is

[Cf]( , , Y _ i,j=[,2,and3l; 2 /=a I L j ^=( , . , : (3.22) [rc£],tri(t,-i,) l=a,b,andc; \ I l=b,c /(j*n where [C,]i is the concentration of the carboxylic acids and the subscript 1 =a, b. and c denotes formic, glyoxylic, and oxalic acids, respectively. The formation of CO? can be described by equation, 3.23.

[CO.] =2 \-ky % i,j= 1,2, and 31 (3.23) [TCEl /=Ü

To establish a kinetic model, the seven rate constants, ki, kza, k 2b, k 2c, ksa, ksb, and k]c need to be estimated using kinetic data from the TCE oxidation experiments.

Accordingly, three duplicated kinetic experiments were conducted in an aqueous media with H3PO4-K2HPO4- KH2 P0 4 -NaOH buffer at pHs 4,6, and 8. All experiments reacted

67 0.09 mM TCE with 0.63 mM KMn 0 4 at a fixed ionic strength of 0.05M. The reactant

TCE, the intermediates (including formic, glyoxylic and oxalic acids), and the final

product, CO? were analyzed over time in kinetic experiments. Mass balances in these

experiments ranged from 0.84 to 1.09. The results indicate that the experiments were

well controlled, even though various chemical analyses and three different types of

instruments were used.

To estimate rate constants, equations 3.18-3.21 were used to fit the measurements,

which are plotted as discrete data points in Figures 3.3-3.5. The fitted curves are shown

as solid lines. The results show good agreement across the experiments at three different

pH levels in terms of TCE destruction and cyclic-hypomanganate-ester decomposition

along various pathways. Thus, our kinetic model for TCE oxidation is consistent with

experimental results at pH ranging from 4 to 8.

The estimated rate constants are listed in Table 3.1. With the pseudo-first-order

rate constant k, ranging from 4.11 x 10“* to 4.30 x 10“* s ', the second-order rate constant

k|p, defined in equation 3.10, can be calculated. The calculated kip range of 0.65 - 0.68

M'* s ' is in agreement with the result (k=0.67+0.03 M"' s ') from the previous study (Yan and Schwartz, 1998).

To choose an appropriate k? for the second step in TCE oxidation, various ratios of k? to ki were tried in fitting data for the carboxylic acids observed in experiments with equations 3.19-3.21. The determination coefficient (r^) for each curve fitting was plotted against a logarithm of ratio of kz to ki in Figure 3.6. Curves do not fit the data well for all

68 carboxylic acids until kz is 10 times higher than k,. This large ratio, kz/ki, supports the

idea that the decomposition of cyclic ester is much faster than its formation, which was postulated by many authors (Wiberg and Saegebarth, 1957; and Freeman et al, 1981 ; Yan and Schwartz, 1998). As exhibited in Figure 3.6, when k? > 10" k,, r^ does not show any obvious Improvement and all other rate constants (k^., kzb, kzc, k.^a, ksb. and k]c) also remain constant with increasing kz/k,. Thus, k? = 10“ ki was assumed in all curve fittings.

The ratios, kza/kz, kzb/kz, and kzc/kz, represent the fraction of the total decomposition rate along each pathway and quantify the contribution from the various pathways shown in Figure 3.2. Their values (Table 3.1) suggest that the pathway leading to formic or oxalic acid is dominant in the decomposition process.

The pseudo-first-order rate constants, ksa, k]b. and ksc, describe the oxidation of formic, glyoxylic, and oxalic acids in the final reaction step. Compared to the rate constants for the other steps in TCE oxidation, the oxidation rate for the carboxylic acids are slower than any of the rate constants in the previous steps. The second-order rate constants, ksap, ksbp. and kscp, can be calculated based on equations, 3.14-3.16. ksap for the oxidation of formic acid ranges from 0.075 to 0.16 M 's ', which is comparable to the range of 0.003-0.25 M ''s ' obtained by Perez-Benito et al. (1990). The observed range in kjc values, 0.073-0.11 M*'s ', for oxalic acid is also consistent with k = 0.005-0.02 M 's ' in Mahmood and Begum (1975) study. Based on estimated rate constants, the oxidation rate of formic and glyoxylic acids is faster than oxalic acid over a pH range of 4-8.

69 From all estimated rate constants, the accumulation of CO 2 can be calculated using equation 3.23. Results are plotted as the solid lines in Figures 3.3-3.5. The calculated concentrations of CO 2 are in general agreement with measured value from the experiments, except at pH 8 (Figure 3.5). In Figure 3.5, the calculated concentration of

CO2 is higher than observed CO 2 at later time. This overestimation may result from the adsorption of carboxylic acids on Mn02, which is more readily precipitated at pH 8 than other pH conditions. Because the precipitates were left in the filtrates in the experiments for acid analysis, acid adsorption on Mn 0 2 could cause a mass deficiency in acid measurements, which occurs at late time (Figure 3.5). Consequently, the rate constant ks at pH 8 may be slightly overestimated due to increase in the concentration difference of acids between early time and later time.

3.3.4 pH dependence

The effect of pH on TCE oxidation was studied over a pH range of 4-8 at fixed initial concentration of Mn 0 4 and ionic strength. During the first step in TCE oxidation, the rate constant k; does not show strong correlation with pH. The previous extensive study on the effect of pH on k, (Yan and Schwartz, 1998) indicated that (1) the variation of k| at different pH values is small and within the experimental error and (2) the second- order rate constant k,p is equal to 0.67±0.03 M‘* s'*. Thus, the rate of TCE disappearance is independent of pH.

70 In the following step, however, the fate of hypomanganate ester and the product

distribution are strongly dependent on pH. At pH = 4, the transformation of cyclic ester

to formic acid is overwhelmingly dominant and its rate is 77% of the total decomposition

rate (kiVk? = 0.77). At high pH, 6-8, the decomposition pathway switches to the

formation of oxalic and glyoxylic acids. The sum of both rates comprises more than 95%

of total decomposition rate ((lc 2b+ kic)/k 2 = 0.95-96). Most of cyclic ester reacts to form

oxalic and glyoxylic acids in this higher pH range.

The rate constants (ksa, k 3b, and ksc) for oxidation of carboxylic acids are strongly correlated with pH over a pH range of 4-8 (Figure 3.7). With decreasing pH, the oxidation rates (ks) increase for all three carboxylic acids. This pH effect on k] results in the relatively rapid formation of COt at low pHs, as compared to higher pHs.

3.3.5 Effect of temperature

Five sets of duplicated kinetic experiments were performed at various temperatures from 5 to 25 “C, in order to investigate temperature effect on TCE disappearance rate by permanganate. This temperature range covers any variation in ground-water temperature. All experiments were conducted by reacting 0.06 mM TCE with excess of Mn 0 4 * (ImM) at pH 7 and an ionic strength of 0.05 M. The estimated logarithm of the second-order rate constant k,p from the experiments is plotted against the reciprocal of temperature in Figure 3.8. The dependence of kip on temperature, as expected (Figure 3.8), follows the Arrhenius equation. The activation energy, Ea, and

71 preexponential. A, were calculated and are listed in Figure 3.8. These values can provide

a rate constant, k[p, at any temperature relevant to ground water.

Based on the E a = 4 1 . 4 6 kJ/mol and A= 1 . 6 7 8 x 1 0 ^ M"' s ’, activation parameters,

AH* and AS* can be calculated by

AH* = A E ^-R T

AS* = R (In A - In (— ) - In T -1) h

where K and h are Boltzmann (1.3805 x 10'^ J K ‘) and Planck (6.6256 x 10'^ J s)

constants and T is temperature. AH*, the enthalpy of activation, is the difference in bond

energies between the reactants and the transition state. The low activation enthalpy (AH*

= 39 kJ/mol) obtained here is comparable to the results from many other studies in alkene

oxidation by permanganate. It suggests that the way to break down carbon-carbon double

bond and to form new bond may be similar and that initial step of TCE oxidation may proceed via a similar transition state as other alkene oxidation

Entropy of activation, AS*, is the difference in entropy between the reactants and the transition state. A considerable entropy loss (negative entropy) will be involve when two reacting molecules must approach each other in a specific orientation in order for the reaction to take place. The observed large negative entropy (AS* = -115 J/mol) indicates that the initial step of TCE oxidation is not favored by the entropy effect.

72 3.4 Mechanism for T ΠOxidation

A thorough mechanistic study does not appear to exist on the oxidation of chlorinated ethenes by permanganate in aqueous phase over the pH range of 4-8.

However, our work, together with the results from previous work on a variety of related systems in chemical synthesis, provides a consistent conceptual model of the reaction processes.

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. This intermediate is typically postulated for permanganate oxidation of many organic substrates with a carbon-carbon double bond. The formation of cyclic hypomanganate ester, as an intermediate, was first suggested by Wagner (1895), and later supported by Wiberg and Saegebarth (1957),

Wiberg et al. (1973), and Lee and Brownridge ( 1973). In the classical view, as summarized by Freeman ( 1976), the formation of cyclic ester involves a 3+2 electrocyclic addition of permanganate ion to the carbon-carbon double bond.

Mn (V) slow Mn O O (3.24) a>< O 'b where the superscript indicates that the intermediate is activated at transition state. The modem ideas on alkene oxidation by permanganate have been advanced by Sharpless et al. ( 1977), and Rappe and Goddard (1982). The initial formation of transient

73 organometallic intermediate proceeds with 2+2 insertion of an alkene k bond directly into

a metal-oxo bond of manganese.

* 0 ^ 9

This reaction mechanism more reasonably explains why permanganate ion with electron-

rich oxygen termini could attack carbon-carbon double bond electrophilically. The attack

is controlled by orbital-overlap via delocalized charge-transfer interaction rather than net-

charge (Toyoshimaet al., 1980; Lee and Brown, 1982). In the initial TCE oxidation,

TCE with electron-withdrawing group (Cl) Is more likely attacked by permanganate ion

via transition state in equation 3.25. This reaction mechanism may also favor trans

isomers with a charge-transfer interaction in the transition state (Toyoshima et al., 1980).

This view is supported by the early result that trans-DCE showed much greater reactivity

than cis-DCE (Yan and Schwartz, 1998).

The small activation enthalpy AH' (39 kJ/mol) is reasonable because bond

formation between n bond of TCE and a metal-oxo bond of manganese accompanies tz bond cleavage. The value of activation enthalpy is consistent with those observed from

the oxidation of other alkenic substrate by permanganate. The similar change in bond energies may indicate that TCE oxidation proceeds via similar transition states. The unfavorable entropy of activation (AS'=-115 J/mol) is to be expected for a transition state

74 that occurs during bimolecular reaction between two compounds filled with electron-rich

termini, oxygen and chlorine, respectively.

With the decomposition of the cyclic hypomanganate ester, the opportunity to

form different products as a function of pH indicates that various pathways exist. At low

pH (=4), the decomposition involves both oxidative hydrolysis (following carbon bond

cleavage) to form formic acid and hydrolysis (with subsequent reactions) to form oxalic

acid (Figure 3.2). As shown by the calculated ratios of rate constants, kiVk? and kic/kz

(Table 3.1), the former reaction pathway is predominant (77% of total production), while

the latter one is much less importance (20%). With increasing pH, almost all the cyclic

esters are hydrolyzed via 6 (Figure 3.2). Products are further decomposed via oxidative

hydrolysis, hydrolysis and electron transfer to form either glyoxylic or oxalic acid.

Decomposition of acyclic ester (7) via 8 to form glycolic acid is insignificant. The

estimated ki/k, ratio from the kinetic data also supports the postulate that the

decomposition of cyclic ester is much faster than its formation. The series of reactions,

such as hydrolysis and electron transfer during the fast ester decomposition, liberate

almost all chlorine substituents on TCE (Yan and Schwartz, 1998).

In the final step, the oxidation of carboxylic acids is relatively slow compared to

the previous steps. The oxidation may involve hydride abstraction by permanganate ion

(Stewart, 1964 and Gardner and Mayer, 1995) and carbon-carbon bond cleavage, which

are different mechanisms from the first step. In the aqueous solution, carboxylic acids can be dissociated to the following anionic species;

75 HCOOH ^ HCOO-+ H* =3.74 (3^6 )

OHC - COOH 4=> OHC - COO~ + pK, = 334 (3.27)

HOOC - COOH <=> //OOC - COO" + H^ = 127 (3.28)

HOOC - COO" <=> "OOC - COO" + //^ pK^^ = 428 (3.29) where Ka, Kb, Kci, and Kc 2, are ionization constant for formic, glyoxylic and oxalic acids, respectively. If all species are considered to react independently with permanganate ion

(Taylor and Halpem, 1959; and Perez-Benito et al., 1990),

[HCOOHl =[HC00H'[+[HC00-\ [OHC - COOHl =[OHC - COOH]+[OHC - COO" ] (3.30) [HOOC - COOHl =[HOOC - COOH]+[HOOC - COO" ]+[" OOC - COO" ] equations of 3.9-3.11 can be further developed to 3.31-3.33:

d[HCOOH], [HCOOH]+k,^, [HCOO-])[MnO,-] dt

L'-* J "T «vjup [HCOOHl [M nO /]

d[OHC-COOHl ------^ [OHC-COOH]+k,,^, [OHC-COO~])[MnO,-]

3 tp l L J Jbp2 [OHC-COOH],mnO~] [H n+K, d[HOOC-COOHl — ------^ ------^=-(^3cpi [HOOC- COOH]+k,„^^ [HOOC - COO" ]

+ ^ 3cp3 ['OOC - COO" ] )[MnO,-]

fe3cp. k,.p2KAH^ + fc3cp3 1 [HOOC - COOHl [MnO^-]

(3.33)

76 where lc 3api, kgapz, ksbpu ksbpz, kscpi, kgcpz, and k3cp3 are rate constants for each species

listed in equation 3.30. Because the pH range of 4-8 is higher than pfQ pKb, pKci, [H^

« Ka, [H *l« Kb and « KciKc? and Kci [H]\ As compared to equations 3.31-3.33

with 3.9-3. II, the second-order rate constants for oxidation of formic, glyoxylic, and oxalic acids are

(3J4)

= (3.35) *■6

, JCP'"- J jipj Cl Ci ^3cp = ------r. c. ■ c. rr,+-.------(3.36)

The relationship between logarithm of second-order rate constants with pH shown in

Figure 3.7 is consistent with the mechanism described by equations 3.34-3.36. Because of the relatively high oxidation rate of carboxylic acids at low pH, as illustrated in Figure

3.3-3.5, CO2 is accumulated more rapidly with decreasing pH.

3.5 Conclusions

The kinetic investigation shows that TCE oxidation by permanganate involves three main reaction steps. In the initial step, an organometallic compound (cyclic hypomanganate ester) is formed via an insertion of a 7t bond in the carbon-carbon double bond into metal-oxo bond of permanganate. The initial reaction is a rate limiting step and independent of experimental conditions (such as pH). The estimated activation

77 parameters have the same character of low enthalpy and negative entropy as other alkene oxidation reactions. This result indicates that the initial reaction in TCE oxidation also proceeds via metal-oxo reacting with carbon-carbon double bonds, as understood in the study of chemical synthesis.

In the second step, the decomposition of cyclic ester (Mn-V) proceeds via different reaction pathways to form various carboxylic acids. The particular acids that form are highly dependent upon the experiment conditions. Kinetic data show that the pathway leading to formic acid via carbon-carbon bond cleavage predominates at low pH and that the pathways forming glyoxylic acid and oxalic acids via oxidative hydrolysis predominate at high pH. The decomposition of cyclic ester is a very rapid process. The kinetic data obtained in this study suggest that the decomposition rate is at least 100 times higher than the formation rate.

All carboxylic acids formed in the second reaction step will be eventually oxidized to the final product. CO 2. This feature of the reaction makes the method appropriate for in-situ ground-water remediation. In the pH range 4-8, permanganate ion mainly reacts with anionic species of the carboxylic acids. Their oxidation rate increases with increasing concentration of hydrogen ions, which is in good agreement with kinetic data observed in this study. Therefore, pH conditions will affect the accumulation rate of

CO2.

The kinetic model that was formulated and solved analytically in this paper is consistent with the results of the kinetic experiments. The parameterization of TCE

78 oxidation kinetics provides a basis for coupling kinetic reaction models with transport models for further study of in-situ remediation schemes.

3.6 References

Berka, A., and Koreckova, J., 1973. The determination of organic compounds by their

oxidation with permanganate. Analytical Letters, 6:1113-1123.

Freeman, F., 1976. Postulated intermediates and activated complexes in the

permanganate ion oxidation of organic compounds. Reviews on Reactive Species

in Chemical Reactions, 1:179-226.

Gardner, K. A., and Mayer, J. M., 1995. Understanding C-H bond oxidations: H and H

transfer in the oxidation of toluene by permanganate. Science, 269:1849-1851.

Gates, D. D, Siegrist, R. L., and Cline, S. R., 1995. Chemical oxidation of volatile and

semivolatile organic compounds in soil. Proceedings of 88th Annual Meeting and

Exhibition, San Antonio, Texas.

Gonullu, T., Farquhar, G. J., Truax, C., Schnarr, M. J., and Stickney, B., 1997. Studies

on the use of permanganate to oxidize chlorinated solvents in soil. Journal of

Contaminant Hydrology, in press.

Kriegman-King, M. R., and Reinhard, M., 1992. Transformation of carbon tetrachloride

in the presence of sulfide, biotite, and vermiculite. Environmental Science and

Technology, 26:2198-2206.

79 Lee, D. G., and Brown, K. C., 1982. Oxidation of hydrocarbons. 11. Kinetics and

mechanism of the reaction between methyl (E)-cinnamate and quaternary

ammonium permanganates. Joumal of American Chemical Society, 104:5076-

5081.

Lee, D. G., and Brownridge, J. R., 1973. The oxidation of cinnamic acid by

permanganate ion. Spectrophotometric detection of an intermediate. Joumal of

American Chemical Society, 95: 3034-3035.

Lee, D. G., and Chen, T., 1989. Oxidation of hydrocarbon. 18. Mechanism of the

reaction between permanganate and carbon-carbon double bonds. Joumal of

American Chemical Society, 111:7534-7538.

Littler, J. S., 1971. Oxidations of olefins, alcohols, glycols and other organic compounds,

by inorganic oxidants such as chromium (VI), manganese (VU), iodine (VU), lead

(IV), vanadium (V) and halogens. Considered in the light of the selection mles for

electrocyclic reactions. Tetrahedron, 27:81-91.

Mahmood, A. J., and Begum, M., 1975. Kinetics of initial processes in the oxidation by

potassium permanganate. I. Oxidation of oxalic acid. Dacca University Studies,

Pt. B., 23:51-64.

Rappe, A. K., and Goddard, W. A. HI, 1982. Olefin Metathesis. A mechanistic study of

high-valent group 6 catalysts. Joumal of American Chemical Society, 104:448-

456.

80 Rodriguez, J., and Sanchez Burgos, F., 1975. Study of the salt effects on the kinetics of

oxidations. EL Salt effects on the kinetics of the oxidation of formic acid by

permanganate. Ion (Madrid), 35:241-245.

Sharpless, K. B., Teranishi, A. Y., and Backvall, J. E., 1977. Chromyl chloride

oxidations of olefins. Possible role of organometallic intermediates in the

oxidations of olefins by oxo transition metal species. Joumal of American

Chemical Society, 99:3120-3128.

Stewart, R., 1964. Oxidation mechanisms. Benjamin, New York.

Taylor, S. M., and Halpem, J., 1959. Kinetics of the permanganate oxidation of formic

acid and formate ion in aqueous solution. Joumal of American Chemical Society,

81:2933-2937.

Toyoshima, K., Okuyama, T., and Fueno, T., 1980. Stmcture and reactivity of a, P-

Unsaturated ethers. 17. Oxidations by permanganate and osmium tetraoxide.

Joumal of Organic Chemistry, 45:1600-1604.

Tmax, C., 1993. Investigation of the in situ KMn 0 4 .oxidation of residual DNAFLs

located below the groundwater table. M. A. Sc. Thesis, University of Waterloo,

Waterloo, Ontario, Canada.

Vella, P. A., and Veronda, B., 1992. Oxidation of trichloroethylene: comparison of

potassium permanganate and Fenton's reagent. In Chemical oxidation

technologies for the nineties, Eckenfelder, W. W., ed., Technomic publishing,

Lancaster, Basel.

81 Vella, P. A., 1996. Potassium permanganate of industrial wasterwater. Proceedings of

the Third International Conference on Advanced Oxidation Technologies for

Water and Air Remediation, Cincinnati, Ohio.

Wagner, G., 1895. History of oxidation reaction of unsaturated compounds. Joumal of

Russian Physical-Chemical Society. 27:219-236.

Walton, J., Labine, P., and Reidies, A., 1991. The chemistry of permanganate in

degradative oxidations. In Chemical Oxidation Technologies for the Nineties,

Eckenfelder, W. W., Bowers, A. R., and Roth, J. A., ed., Technomic publishing,

Lancaster, Basel, pp 205-221.

Wiberg, K. B., Deutsch, C. J., and Rocek, J., 1973. Permanganate oxidation of crotonic

acid. Spectrometric detection of an intermediate. Joumal of American Chemical

Society, 95: 3034-3035.

Wiberg, K. B., and Saegebarth, K. A., 1957. The mechanisms of permanganate

oxidation. IV. Hydroxylation of olefins and related reactions. Journal of

American Chemical Society, 79: 2822-2824.

Wolfe, S., and Ingold, C. F., 1981. Oxidation of olefins by potassium permanganate.

Mechanism of a-ketol formation. Joumal of American Chemical Society,

103:938-939.

Yan, Y., and F. W. Schwartz, 1995. The Removal of Trichloroethylene from

Contaminated Groundwater Using Potassium Permanganate. EOS, 76(46), p.

F247.

82 Yan, Y. E., and Schwartz, F. W., 1998. Oxidative degradation of chlorinated ethylenes

by potassium permanganate. Joumal of Contaminant Hydrology, in press.

83 pH k,(ior»s^) Kt(ior*s^) Ka/Ki i^(ig^s^) ^ i^do^s^) ^ m <2 i^do^s^) ^ 4 4.30 0.77 2.19 0.98 0.03 234 0.98 020 0.70 0.95 6 4.11 >1(fki 0.02 1.53 0.91 0.32 1.05 0.99 0.63 0.64 0.95 8 4.11 0.04 0.47 0.97 0.42 0.80 0.99 0.55 0.46 0.95

Table 3.1 Rate constants obtained for TCE transformation pathways using equations. 16-19. Estimation was based on experiments conducted with an initial concentration of 0.63 mM Mn 0 4

84 2 Oxalic Acid

pH 8 0.0135

1.5 0.0125 pH 8

pH 6 Formic Acid ■e 0.0115 At pH G pH 4 1 Giycoilc Acid 0.0105 pH 8

pH 4 pH 6 0.0095

0.5 Tima (min)

pH 4

0 0 5 10 15 20 Time (min)

Figure 3.1 Carboxylic products of T Πoxidation as determined by HPLC. Samples were collected one hour after the reaction began in three eight-hour experiments at pH 4, 6, and 8, respectively, and monitored by absorbance at 210 nm.

85 1. Trichloroethylene M11O4 slow L(vm

' /O' (V) Mn (V) 6st OH OMnOg^' O O a ^ l I/H " Q \ l l/H H-,0 Mn 0 4 +H (vm 6. Acyclic hypoimnganate ester 2. Cyclic Q- hypoimnganate ester MnOj+HiO {vm Mn(Vn>(-Mn(IV)

(VI) I Mnv (VD 0=My=S OH OH H^l I^CI OH O ) H ^l |> l /

_CI C l. 10. Tnchloiogly CO I 3. Cyclic manganate ester 7. Acyclic imnganate ester

fasti (IV) MnO. (IV) X 1 HMnOj- HMnOj+Ha A X 0 1 OH 0 0 0 II H ^ l II II II 0 — a

4 '•S'' ^ 11 2HiO

OH O 0 0 II II c / —c.\ HO OH 5. Formic Acid 9. Qycolic Acid 12. QyojQrlic Acid 13. Oxalic Acid

%

0 = 0=0

14. Gitfoon Dioxide pH

Figure 3.2 Proposed TCE oxidation pathways. Shadowed boxes are products identified in this study. 86 1 . 2 pH =4

1.0

0.8

o ü 0.6 CO: ü COOH CHO-COOH 0.4 HOOC-COOH TCE 0.2 Mass Balance

0.0

4 5 Time (hr)

Figure 3.3 Product distribution with time in TCE oxidation at pH=4. Lines for TCE and carboxylic acids represent best fits using equations of 3.18-3.21, whereas the line for CO? is calculated by equation 3.23 using rate constants from the previous curve fitting.

87 1.2 pH = 6 D COz HOOC-COOH

• COOH o TCE 1.0 ♦ CHO-COOH + Mase Balance

0.8

o u 0.6

0.4

0.2

0.0 —r 3 4 5 Time (hr)

Figure 3 .4 Product distribution with time in TCE oxidation at pH=6. Lines for TCE and carboxylic acids represent best fits using equations of 3.18-3.21, whereas the line for CO? is calculated by equation 3.23 using rate constants from the previous curve fitting.

88 1 .2 pH = 8 ▲ HOOC-COOH

• COOH C TCE 1.0 ♦ CHO-COOH + Mass Balance

0.8

o 0.6 sa o

0 .4

0.2

0.0

4 5 Time (hr)

Figure 3 .5 Product distribution with time in TCE oxidation at pH=8. Lines for TCE and carboxylic acids represent best fits using equations of 3.18-3.21. whereas the line for CO? is calculated by equation 3.23 using rate constants from the previous curve fitting.

89 1 0.8 m. » ‘ ‘ ‘

Z. 0.6

0.4 gFormic Acid • Giyoxylic Acid 0.2 • ♦ Oxalic Acid •

log (kzAi)

Figure 3.6 Plot of determination coefficients of curve fitting versus logarithm of ratio of ki to k[. The determination coefficients were obtained by fitting equations of 3.19-3.21 to data from carboxylic measurements using various ratios, ki/k,.

90 -3.0

• Formic Acid H ■ Giyoxylic Acid A Oxalic Acid , i

-3.5 —

O) -4.0 — o ▲

-4.5

-5.0 1 i r

4 6 pH

Figure 3.7 Dependence of the pseudo-first-order rate constants (ksa, kab, and kjc) on pH for oxidation of carboxylic acids. The reaction involved reacting 0.09 mM TCE with 0.63 mM MnÛ 4 in a phosphate buffered aqueous solution at ionic strength 0.05 M.

91 0.4

k,p= 1.678 X 10' exp(-4987/T) r"=0.99

0.0

CO -0.4 -

-0.8 H

•1.2 -1

A=1.678 X 1 0 'N T s Ea=41.462 kJ moT

1.6 4 - 3.3 3.4 3.5 3.6 3.7 1/T (IQ:)

Figure 3 .8 . Arrhenius plot for the oxidation of TCE (0.06 mM) by potassium permanganate ( 1 mM) in a phosphate buffered solution at ionic strength 0.05M and pH 7.1. Error bars are standard deviation for each estimated second-order rate constant k,.

92 CHAPTER 4

SUMMARY

A detailed process-oriented study was conducted on permanganate oxidation of chlorinated ethylenes. The results show the significant potential of the method as a new in-situ remediation technology and provide a theoretical basis to assess the utility of the method for in-situ treatment of ground water and soil contaminated by chlorinated ethylenes. The following are major findings drawn in this work:

Destruction rate of chlorinated ethvlenes

The study shows that all chlorinated ethylenes can be rapidly degraded by permanganate in aqueous solution. The calculated second-order rate constants range from 0.045 to 30.0 M ''s'\ Compared to half lives of degradation of chlorinated ethylenes in iron reduction, which is called “magic sand” due to its reputation for fast degradation, this method apparently provides an even faster way to destroy chlorinated ethylenes

(Figure 4.1). Among chlorinated ethylenes, the degradation rate increases with a decreasing number of chlorine substituents on ethylene. The deficiency of electrons in the carbon-carbon double bond induced by chlorines in ethylene mainly slows down the

93 attack of an electrophile (permanganate ion) and explains the reason why the number of chlorines is inversely related to increase in degradation rate.

Oxidative dechlorination

Rapid and complete dechlorination was observed in this investigation. The dechlorination in this oxidation process is apparently different from the sequential dechlorination that occurs in the reduction of chlorinated ethylenes. This latter reaction process slows down as the number of chlorines on ethylenes decreases. This behavior could result in an accumulation of undesirable products such as DCE and vinyl chloride.

The former dechlorination proceeds rapidly in the second reaction step where cyclic ester decomposes at the rate at least 100 times greater than the disappearance rate of chlorinated ethylenes.

Mechanism and products of reactions

The oxidation pathways for chlorinated ethylenes by permanganate were investigated using TCE as an example. The oxidation process involves three different reaction steps: (1) the destruction of TCE to form an organometallic compound

(hypomanganate ester), (2) the decomposition of the ester to various carboxylic acids, and (3) eventually the oxidation to the final product, CO?. The initial and final reactions in TCE oxidation are subject to rate-limiting steps, which control the destruction rate of

TCE and formation rate of the final product. The fast reactions involved in the second step determine the nature of products.

94 The main intermediate products are four carboxylic acids, including formic, oxalic, giyoxylic and glycolic acids. The products were identified and quantified with time in a series of kinetic experiments. They were formed in the second reaction step and eventually oxidized to the final product, CO 2 , in the last reaction step. Complete mineralization occurred in the reaction, making the method suitable for in-situ ground­ water remediation. Even under conditions where permanganate is not present in excess amounts to enable the reaction to completion, the intermediate products (i.e. carboxylic acids, which are miscible with water), will still be removed readily with ground water as flushing proceeds.

Effects of pH. temperature, and background contaminants

pH conditions do not affect TCE destruction rate over the range of 4-8 but they control the nature and distribution of products and the accumulation rate of the final product, CO 2 . Among the various possible reaction pathways, the one leading to formic acid via carbon-carbon bond cleavage predominates at low pH. At high pH, the reaction pathway is overwhelmed by those forming giyoxylic acid and oxalic acids via oxidative hydrolysis. With a decrease in temperature in the range of 5-25 “C, TCE destruction rate decreases.

The consumption of permanganate in ground water with low TOC is limited.

Ground water contaminated by landfill leachates, however, will consume permanganate depending on the TOC level. In the test solution containing landfill leachates used in this study, a simple second-order reaction rate is able to provide a rough estimate of the

95 consumption rate of permanganate by TOC. This approach may be a useful basis for estimating the site-specific consumption rate of permanganate in the field. Apparently, the competition of TOC for permanganate slows down the degradation rate of chlorinated ethylenes. However, the effect of TOC is partially countered by autocatalysis on the surface of colloidal MnO? (Yan and Schwartz, 1996) at late time, when a large quantity of MnOi is produced in the reaction.

Kinetic model

The kinetic model has been developed based on understanding of oxidation process involving various reaction pathways. The established model successfully simulates the independent experimental data in accumulation of C0%. The parameterization of oxidation kinetics provides a basis for coupling kinetic reaction model with transport model for further study on optimization of remediation scheme.

Closing comments

In general, the results of this study continue to support efforts to develop permanganate oxidation as the basis for the treatment of solvent spills. Issues, however, remain to be addressed. Foremost among them are the extent to which plugging by

MnOi may interfere with flushing schemes. The work here on dissolved contaminants has an obvious extension to problems of pure phase liquids, which is how solvents occur at many sites. The most important problems to consider in this respect are the rate limiting steps in the dissolution of free products and the effects of phase saturation on

96 relative permeability and displacement efficiencies. Finally, large-scale experiments are needed to assess issues of flooding efficiencies as a function of well placement and schemes for controlling dissolved contaminants in reactive barrier systems. Thus, a number of very fertile research areas remain to be followed in this work.

97 432 ■ Permanganate 158 mg/L o Iron powder 250 g/L_____

SO 00

0.30

PCE TCE cDCE tDGE 1,1-DCE

Figure 4.1 Half lives of degradation of chlorinated ethylenes in permanganate oxidation and iron reduction. The initial concentration of permanganate is 158 mg/L, while initial concentration of iron powder with high surface area is 250 g/L. The half lives in iron reduction are from Gillham and O’Hannesin (1994). UST OF REFERENCES

Abramowicz, D. A., 1990. Aerobic and anaerobic biodégradation of PCBs: A review. Grit. Rev. Biotechnoi., 10:241-251.

Aieta, E. M., K. M. Regan, J. S. Lang, L. McReynolds, J. W. Kang, and W. H. Glaze, 1988. Advanced oxidation processes for treating groundwater contaminated with TCE and PCE: Pilot-scale evaluations. Jour. AWWA, 80:64-72.

Arndt, D., 1981. Manganese compounds as oxidizing agents in organic chemistry. Open Court Publishing Company, La Salle, Illinois, 344 pp.

Barrio-Lage, G. A., F. Z. Parsons, R. S. Nassar and P. A. Lorenzo, 1986. Sequential dehalogenation of chlorinated ethenes. Environmental Science and Technology, 20:96-99.

Beeman, R., S. Shoemaker, J. Howell, E. Salazar, and J. Buttram, 1993. A field evaluation of in situ microbial reductive dehalogenation by the biotransformation of chlorinated ethylenes, abstr. A-2. In In situ and on-site bioreclamation: 2nd International Symposium, San Diego, Calif., Battelle, Columbus, Ohio.

Belay, N., and L. Daniels, 1987. Production of ethane and ethylene from halogenated hydrocarbon by methanogenic bacteria. Applied Environ. Microbiol., 53:1604- 1610.

Berka, A., and Koreckova, J., 1973. The determination of organic compounds by their oxidation with permanganate. Analytical Letters, 6:1113-1123.

Bouwer, E.J., B.E. Rittman and P.L. McCarty. 1981. Anaerobic degradation of halogenated 1- and 2-carbon organic compounds. Environmental Science and Technology, 15:596-599.

Surdon, J., and Tatlow, J. C., 1958. The reactions of highly fluorinated organic compounds. X. The oxidation of fluoro-olefins by potassium permanganate in acetone. Journal of Applied Chemistry, 8:293-296.

Casado, J., Lopez-Quintela, M. A., and Lorenzo-Barral, F. M., 1986. The initial rate method in chemical kinetics. Journal of Chemical Education, 63:450-452.

99 Cotlove, E., 1958. An instrument and method for automatic, rapid, accurate and sensitive titration of chloride in biologic samples. Journal of Laboratory and Clinical Medicine, 51:461-468.

Ensley, B. D., 1991. Biochemical diversity of trichloroethylene metabolism. Annu. Res. Microbiol., 45:283-299.

Ewers, J., D. Freier-Schroder, and H-J Knackmuss, 1990. Selection of trichloroethene (TCE) degrading bacteria that resist inactivation by TCE. Arch. Microbiol., 154:410-413.

Fathepure, B. Z., and S. A. Boyd, 1988. Dependence of tetrachloroethylene dechlorination on methanogenic substrate consumption by Methanosarsina sp. strain DSM. Applied Environ. Microbiol., 54:2976-2980.

Fathepure, B. Z., J. P. Nengu, and S. A. Boyd, 1987. Anaerobic bacteria that dechlorinate perchloroethylene. Applied Environ. Microbiol., 53:2671-2674.

Freedman, D. L., and J. M. Gossett, 1989. Biological reductive dechlorination of tetrachloroethylene and trichloroethylene to ethylene under methanogenic conditions. Applied and Environmental Microbiology, 55:2144-2151.

Freeman, F., 1975. Possible criteria for distinguishing between cyclic and acyclic activated complexes and among cyclic activated complexes in addition reactions. Chemical reviews, 75:439-491.

Freeman, F., 1976. Postulated intermediates and activated complexes in the permanganate ion oxidation of organic compounds. Reviews on Reactive Species in Chemical Reactions, 1:179-226.

Freeman, F., Fuselier, C.O., Armstead, C.R., Dalton, C.E., Davidson, P.A., Karchefski. E.M., krochman, D.E., Johnson, M.N., and Jones, N.K., 1981. Permanganate ion oxidation. 13. Soluble manganese (IV) species in the oxidation of 2,4( 1H,3H)- pyrimidinediones (Uracils). Journal of American Chemical Society, 103:1154- 1159.

Gardner, K. A., and Mayer, J. M., 1995. Understanding C-H bond oxidations: H» and H- transfer in the oxidation of toluene by permanganate. Science, 269:1849-1851.

Gates, D. D, Siegrist, R. L., and Cline, S. R., 1995. Chemical oxidation of volatile and semivolatile organic compounds in soil. Proceedings of 88th Annual Meeting and Exhibition, San Antonio, Texas.

Gillham, R. W., and O'Hannesin, S. F., 1994. Enhanced degradation of halogenated aliphatics by zero-valent iron. Ground Water, 32:958-967.

100 Glaze, W. H., and Kang, J. K., 1988. Advanced oxidation processes for treating groundwater contaminated with TCE and PCE: laboratory studies. Journal AWWA, 80:57-63.

Glaze. W. H., J. W. Kang, and D. H. Chapin, 1987. The chemistry of water treatment processes involving ozone, hydrogen peroxide, and ultraviolet radiation. Ozone Science and Engineering., 9:335.

Gonullu, T., Farquhar, G. J., Truax, C., Schnarr, M. J., and Stickney, B., 1997. Studies on the use of permanganate to oxidize chlorinated solvents in soil. Journal of Contaminant Hydrology, in press.

Haines, A. H., 1985. Methods for the oxidation of organic compounds. Academic Press, London, 388 pp.

Hoigné, J., and Bader, H., 1983. Rate constants of reaction of ozone with organic and inorganic compounds in water, I. Non-dissociating organic compounds. Water Research, 17:173-183.

Hutzinger, O., and Veerkamp W., 1981. Xenobiotic compounds with pollution potential. In Leisinger T., Cook A., Hutter R., and Nuesch J., eds: Microbial Degradation of Xenobiotic and Recalcitrant Compounds. Academic Press, London, p3.

Jeffers, P. M., Woytowitch, L. M., and Wolfe, N. L., 1989. Homogeneous hydrolysis rate constants for selected chlorinated methanes, ethanes and propanes. Environmental Science and Technology, 23:965-969.

Kriegman-King, M. R., and M. Reinhard, 1994. Transformation of carbon tetrachloride by pyrite in aqueous solution. Environmental Science and Technology, 28:692- 700.

Kriegman-King, M. R„ and Reinhard, M., 1992. Transformation of carbon tetrachloride in the presence of sulfide, biotite, and vermiculite. Environmental Science and Technology, 26:2198-2206.

Lee, D. G., and Brown, K. C., 1982. Oxidation of hydrocarbons. 11. Kinetics and mechanism of the reaction between methyl (E)-cinnamate and quaternary ammonium permanganates. Journal of American Chemical Society, 104:5076- 5081.

Lee. D. G., and Brownridge, J. R., 1973. The oxidation of cinnamic acid by permanganate ion. Spectrophotometric detection of an intermediate. Journal of American Chemical Society, 95: 3034-3035.

101 Lee, D. G., and Chen, T., 1989. Oxidation of hydrocarbon. 18. Mechanism of the reaction between permanganate and carbon-carbon double bonds. Journal of American Chemical Society, 111:7534-7538.

Lee, D. G., 1982. Phase transfer assisted permanganate oxidations. In Oxidation in organic chemistry, W. S. Trahanovsky, ed.. Academic Press, New York, Part D, Charpter 2, pp 147-206.

Li, S., and L. P. Wackett, 1992. Trichloroethylene oxidation by toluene dioxygenase. Biochem. Biophys. Res. Commun., 185:443-451.

Littler. J. S., 1971. Oxidations of olefins, alcohols, glycols and other organic compounds, by inorganic oxidants such as chromium (VI), manganese (VU), iodine (VU), lead (IV), vanadium (V) and halogens. Considered in the light of the selection rules for electrocyclic reactions. Tetrahedron, 27:81-91.

Logan, M. S. P., L. M. Newman, C. A. Schanke, and L. P. Wackett, 1993. Co-substrate effects in reductive dehalogenation by Pseudomonas putida G786 expressing cytochrome P450CAM. Biodégradation 4:39-50.

Mahmood, A. J., and Begum, M., 1975. Kinetics of initial processes in the oxidation by potassium permanganate. I. oxidation of oxalic acid. Dacca University Studies, Pt. B., 23:51-64.

Maltoni, C., and G. Lefemine, 1974. Carcinogenicity bioassays of vinylchloride. 1. Research plan and early results. Environ. Res., 7:387-396.

Mata-Pérez, F., and Pérez-Benito, J. P., 1985. Identification of the product from the reduction of permanganate ion by trimethylamine in aqueous phosphate buffers. Canadian Journal of Chemistry, 63:988-992.

Matheson, L. J., and P. G. Tratnyek, 1994. Reductive dehalogenation of chlorinated methanes by iron metal. Environmental Science and Technology, 28:2045-2053.

National Research Council, 1994. Altematives for ground water cleanup. National Academic Press, Washington, D. C., p 26.

Phelps, T. J., J. J. Niedzielski, K. J. Malachowsky, R. M. Schram, S. E. Herbes, and D. C. White, 1991. Biodégradation of mixed-organic wastes by microbial consortia in continuous-recycle expanded-bed bioreactors. Environmental Science and Technology, 25:1461-1465.

Pérez-Benito J. P., and Arias, C., 1991. A kinetic study of the permanganate oxidation of triethylamine. Catalysis by soluble colloids. International Journal of Chemical Kinetics, 23:717-732.

102 Pérez-Benito, J. F., Mata-Pérez, P., and Brillas, E., 1987. Permanganate oxidation of glycine: kinetics, catalytic effect, and mechanisms. Canadian Journal of Chemistry, 65:2329-2337.

Quensen, J. F. HI, J. M. Tiedje, and S. A. Boyd, 1988. Reductive dechlorination of polychlorinated biphenyls by anaerobic microorganisms from sediment. Science, 242:752-754.

Ram, N. M., Christman, R. F., and Cantor, K. P., 1990. Significance and treatment of volatile organic compounds in water supplies. Lewis Publishers, Chelsa, MI.

Rappé, A. K., and Goddard, W. A. m , 1982. Olefin Metathesis. A mechanistic study of high-valent group 6 catalysts. Journal of American Chemical Society, 104:448- 456.

Rodriguez, J., and Sanchez Burgos, P., 1975. Study of the salt effects on the kinetics of oxidations, m . Salt effects on the kinetics of the oxidation of formic acid by permanganate. Ion (Madrid), 35:241-245.

Sharpless, K. B., Teranishi, A. Y., and Backvall, J. E., 1977. Chromyl chloride oxidations of olefins. Possible role of organometallic intermediates in the oxidations of olefins by oxo transition metal species. Journal of American Chemical Society, 99:3120-3128.

Simdndi, L. I., and Jdky, M., 1976. Nature of the detectable intermediate in the permanganate oxidation of trans-cinnamic acid. Journal of American Chemical Society, 98: 1995-1997.

Smith, G., and G. Ferguson, 1993. In situ remediation of groundwater contaminated with chlorinated solvents using anaerobic biotransformation, abstr. C-2. In In situ and on-site bioreclamation: 2nd International Symposium, San Diego, Calif., Battelle, Columbus, Ohio.

Stewart, R., 1964. Oxidation mechanisms. Benjamin, New York, pp 58-76.

Stewart, R., 1965. Oxidation by permanganate. In Oxidation in organic chemistry, K. B. Wiberg, ed.. Academic Press, New York, Part A Chapter 1, pp 1-68.

Szammer, J., and J« ky, M., 1992. Oxidation by permanganate in strong alkaline medium. Oxidation of ethane-1,2-diol, glycol aldehyde, glycollic acid, and giyoxylic acid. International Journal of chemical kinetics, 24:145-154.

Taylor, S. M., and Halpem, J., 1959. Kinetics of the permanganate oxidation of formic acid and formate ion in aqueous solution. Journal of American Chemical Society, 81:2933-2937.

103 Toyoshima, K., Okuyama, T., and Fueno, T., 1980. Structure and reactivity of oc, P- Unsaturated ethers. 17. Oxidations by permanganate and osmium tetraoxide. Journal of Organic Chemistry, 45:1600-1604.

Truax, C., 1993. Investigation of the in situ KMn04.oxidation of residual DNAPLs located below the groundwater table. M. A. Sc. Thesis, University of Waterloo, Waterloo, Ontario, Canada.

Vella, P. A., and Veronda, B., 1992. Oxidation of trichloroethylene: comparison of potassium permanganate and Fenton’s reagent. In Chemical oxidation technologies for the nineties, Eckenfelder, W. W., ed., Technomic publishing, Lancaster, Basel.

Vella. P. A., 1996. Potassium permanganate of industrial wasterwater. Proceedings of the Third International Conference on Advanced Oxidation Technologies for Water and Air Remediation, Cincinnati, Ohio.

Vogel. T. M., C. S. Criddle, and P. L. McCarty, 1987. Transformations of halogenated aliphatic compounds. Environmental Science and Technology, 21:722-736.

Wackett, L. P., and D. T. Gibson, 1988. Degradation of trichlorethylene by toluene dioxygenase in whole-cell studies with Pseudomonas putida FI. Appl. Environ. Microbiol., 54:1703-1708.

Wackett, L. P., and S. R. Househoder, 1989. Toxicity of trichloroethylene to Pseudomonas putida FI is mediated by toluene dioxygenase. Appl. Environ. Microbiol., 55:2723-2725.

Wackett, L. P., 1995. Bacterial co-metabolism of halogenated organic compounds. In Microbial transformation and degradation of toxic organic chemicals, L. Y. Young and C. E. Cemiglia, ed., Wiley-Liss, Inc, New York, pp 217-241.

Wagner, G., 1895. History of oxidation reaction of unsaturated compounds. Journal of Russian Physical-Chemical Society. 27:219-236.

Walton, J., Labine, P., and Reidies, A., 1991. The chemistry of permanganate in degradative oxidations. In Chemical Oxidation Technologies for the Nineties, Eckenfelder, W. W., Bowers, A. R., and Roth, J. A., ed., Technomic publishing, Lancaster, Basel, pp205-221.

Wiberg, K. B., and Saegebarth, K. A., 1957. The mechanisms of permanganate oxidation. IV. Hydroxylation of olefins and related reactions. Journal of American Chemical Society, 79:2822-2824.

104 Wiberg, K. B., Deutsch, C. J., and Rocek, J., 1973. Permanganate oxidation of crotonic acid. Spectrometric detection of an intermediate. Journal of American Chemical Society, 95: 3034-3035.

Wilson, J. T., and B. H. Wilson, 1985. Biotransformation of trichloroethylene in soil. Applied Environ. Microbiol., 29:242-243.

Wiseman, P., 1979. An Introduction to Industrial Organic Chemistry, 2dn edition. Elsevier Applied Science, New York, 366 pp.

Wolfe, S., and Ingold, C. P., 1981. Oxidation of olefins by potassium permanganate. Mechanism of a-ketol formation. Journal of American Chemical Society, 103:938-939.

Yan, Y., and F. W. Schwartz, 1995. The removal of trichloroethylene from contaminated groundwater using potassium permanganate. EOS, 76(46), p. F247.

Yan, Y. E., and F. W. Schwartz, 1996. Kinetic studies on permanganate oxidation of trichloroethylene in ground water. EOS, 77(46), p. F216.

Yan, Y. E., and Schwartz, F. W., 1998. Oxidative degradation of chlorinated ethylenes by potassium permanganate. Joumal of Contaminant Hydrology, in press.

Yan, Y. E., and Schwartz, F. W., 1998. Kinetics and mechanism for TCE oxidation by permanganate. To be submitted for publication.

105 IMAGE EVALUATION TEST TARGET (Q A-3) A

- O.

A

1.0 y . |Z 5 U£ U b 122 |r L6 : Ls 1.1 1.8 1.25 II u

150mm

/APPLIED ^ IIVHGE. Inc 1653 East Main street Rochester, NY 14609 USA Phone: 716/4820300 . ^ 5 ^ ^ = Fax: 716/288-5989

0 1993. Applied Image. Inc.. AH Rights Reserved

4^ . > ' ?