Citation
Behavior of partially miscible organic compounds in simulated ground water systems

Material Information

Title:
Behavior of partially miscible organic compounds in simulated ground water systems
Creator:
Cline, Patricia V. ( Dissertant )
Delfino, Joseph J. ( Thesis advisor )
Chadik, Paul A. ( Reviewer )
Dorsey, John ( Reviewer )
Rao, P. S. C. ( Reviewer )
Yost, Richard A. ( Reviewer )
Place of Publication:
Gainesville, Fla.
Publisher:
University of Florida
Publication Date:
Copyright Date:
1988
Language:
English
Physical Description:
vii, 194 leaves : ill. ; 28 cm.

Subjects

Subjects / Keywords:
Chemicals ( jstor )
Eggshells ( jstor )
Gasoline ( jstor )
Groundwater ( jstor )
Halides ( jstor )
Hydrocarbons ( jstor )
Hydrolysis ( jstor )
pH ( jstor )
Solubility ( jstor )
Solvents ( jstor )
Dissertations, Academic -- Environmental Engineering Sciences -- UF
Environmental Engineering Sciences thesis Ph.D
Gasoline -- Solubility ( lcsh )
Groundwater -- Pollution ( lcsh )
Trichloroethane -- Solubility ( lcsh )
Genre:
bibliography ( marcgt )
theses ( marcgt )
non-fiction ( marcgt )

Notes

Abstract:
Serious ground water contamination problems result from leaks or spills of organic liquids which are partially miscible in water. Two important categories of these liquids include low molecular weight chlorinated solvents and gasoline. 1,1,1,-Trichloroethane (TCA) abiotically degrades in water forming approximately 17-25%, 1,1-dichloroethen (1,1-DCE) via an elimination reaction. The substituion product is acetic acid. The arrhenius activation energy is 11+/- 3kj/mol with an Arrhenius factor of 2 x 10^13 s^-1, which results in an estimated half-life for the degradation at 25*C of 10.2 months. Brominate analogs of TCA hydrolyze 11-13 times faster than TCA. As the number of bromines increase, the percent of elimination products increases. These geminal trihalides degrade by unimolecular mechanism (E1/SN1). The rate coefficient for TCA degradation in buffered water at elevated temperature is approximately six times greater than hydrolysis of 1-chloropropane (SN2 mechanism) and more than 100 times greater than 1,1-dichloroethane rate by 22 and TCA degradation by approximately two. Halogenated ethenes are stable at various temperatures and reaction conditions. Trichloroethen degrades in alkaline solution at elevated temperature. 1,1,1,-Tricholoroethane and 1,1,-DCE form a near ideal solution in the solvent phase. The solubility of 1,1-DCE at 24*C is 3200 mg/l and the solubility of TCA is approximately 1580mg/l. The range of concentrations for major components of gasoline which partition into water was determined for 65 gasoline samples. Benzene concentrations in the water extracts ranged from 12.3-130 mg/l and toluene concentrations ranged from 23-185 mg/l. Fuel/water partition coefficients of seven major aromatic constituents were measured for 31 gasoline types and showed a standard deviation of 10-30%. Those coefficients were highly correlated with the pure component solubilities. Chemometric techniques were applied to 20 peaks measured in the aqueous extracts of the 65 gasolines. Bivariate plots and principal component analyses show selected brands have distinguishing equilibrium concentrations, but complete separation of brands was not observed.
Thesis:
Thesis (Ph. D.)--University of Florida, 1988.
Bibliography:
Includes bibliographical references.
General Note:
Typescript.
General Note:
Vita.
Statement of Responsibility:
by Patricia V. Cline.

Record Information

Source Institution:
University of Florida
Holding Location:
University of Florida
Rights Management:
Copyright [name of dissertation author]. Permission granted to the University of Florida to digitize, archive and distribute this item for non-profit research and educational purposes. Any reuse of this item in excess of fair use or other copyright exemptions requires permission of the copyright holder.
Resource Identifier:
024589839 ( ALEPH )
19924511 ( OCLC )
AFL3692 ( NOTIS )

Downloads

This item has the following downloads:

UF00086020_00001.pdf

00006.txt

UF00086020_00001_0152.txt

00199.txt

UF00086020_00001_0010.txt

UF00086020_00001_0140.txt

UF00086020_00001_0170.txt

00026.txt

UF00086020_00001_0086.txt

00047.txt

00080.txt

UF00086020_00001_0164.txt

UF00086020_00001_0155.txt

00058.txt

UF00086020_00001_0118.txt

UF00086020_00001_0081.txt

UF00086020_00001_0092.txt

UF00086020_00001_0014.txt

UF00086020_00001_0137.txt

00105.txt

00060.txt

00054.txt

UF00086020_00001_0110.txt

00092.txt

UF00086020_00001_0034.txt

UF00086020_00001_0114.txt

UF00086020_00001_0179.txt

UF00086020_00001_0073.txt

00051.txt

UF00086020_00001_0028.txt

UF00086020_00001_0139.txt

00177.txt

00055.txt

00061.txt

UF00086020_00001_0204.txt

UF00086020_00001_pdf.txt

00153.txt

00162.txt

00137.txt

UF00086020_00001_0079.txt

UF00086020_00001_0176.txt

UF00086020_00001_0011.txt

UF00086020_00001_0067.txt

00183.txt

00067.txt

00142.txt

00181.txt

00037.txt

UF00086020_00001_0127.txt

00033.txt

UF00086020_00001_0190.txt

UF00086020_00001_0186.txt

00100.txt

00096.txt

UF00086020_00001_0083.txt

UF00086020_00001_0141.txt

00145.txt

UF00086020_00001_0037.txt

UF00086020_00001_0120.txt

00108.txt

UF00086020_00001_0104.txt

UF00086020_00001_0192.txt

UF00086020_00001_0013.txt

00174.txt

UF00086020_00001_0175.txt

UF00086020_00001_0122.txt

UF00086020_00001_0005.txt

00062.txt

00002.txt

UF00086020_00001_0113.txt

00112.txt

00146.txt

UF00086020_00001_0090.txt

00076.txt

00057.txt

UF00086020_00001_0074.txt

UF00086020_00001_0165.txt

00148.txt

UF00086020_00001_0084.txt

00182.txt

UF00086020_00001_0052.txt

UF00086020_00001_0180.txt

UF00086020_00001_0072.txt

00158.txt

UF00086020_00001_0151.txt

00087.txt

00066.txt

UF00086020_00001_0055.txt

00186.txt

UF00086020_00001_0108.txt

00073.txt

00075.txt

UF00086020_00001_0195.txt

UF00086020_00001_0001.txt

UF00086020_00001_0058.txt

UF00086020_00001_0123.txt

UF00086020_00001_0032.txt

UF00086020_00001_0111.txt

00194.txt

UF00086020_00001_0154.txt

00007.txt

UF00086020_00001_0082.txt

00127.txt

UF00086020_00001_0045.txt

00027.txt

00063.txt

UF00086020_00001_0061.txt

UF00086020_00001_0178.txt

00114.txt

00091.txt

00071.txt

00120.txt

00059.txt

00136.txt

UF00086020_00001_0161.txt

00150.txt

UF00086020_00001_0106.txt

UF00086020_00001_0148.txt

UF00086020_00001_0103.txt

00042.txt

UF00086020_00001_0133.txt

00012.txt

00201.txt

UF00086020_00001_0042.txt

00156.txt

UF00086020_00001_0033.txt

00125.txt

00023.txt

00167.txt

00039.txt

UF00086020_00001_0018.txt

00122.txt

UF00086020_00001_0043.txt

00163.txt

UF00086020_00001_0095.txt

00133.txt

UF00086020_00001_0078.txt

UF00086020_00001_0194.txt

00072.txt

UF00086020_00001_0060.txt

00081.txt

00020.txt

UF00086020_00001_0091.txt

UF00086020_00001_0098.txt

UF00086020_00001_0160.txt

UF00086020_00001_0163.txt

00038.txt

UF00086020_00001_0132.txt

00188.txt

00179.txt

00193.txt

UF00086020_00001_0107.txt

UF00086020_00001_0017.txt

00151.txt

00101.txt

00011.txt

00190.txt

00160.txt

UF00086020_00001_0054.txt

UF00086020_00001_0187.txt

UF00086020_00001_0075.txt

00034.txt

00010.txt

00083.txt

UF00086020_00001_0142.txt

00157.txt

UF00086020_00001_0136.txt

00143.txt

00024.txt

UF00086020_00001_0097.txt

UF00086020_00001_0169.txt

00110.txt

UF00086020_00001_0046.txt

00093.txt

UF00086020_00001_0198.txt

UF00086020_00001_0035.txt

00117.txt

UF00086020_00001_0094.txt

UF00086020_00001_0093.txt

UF00086020_00001_0006.txt

00152.txt

UF00086020_00001_0126.txt

00184.txt

00022.txt

UF00086020_00001_0024.txt

00119.txt

00189.txt

00168.txt

00111.txt

00154.txt

UF00086020_00001_0016.txt

UF00086020_00001_0157.txt

UF00086020_00001_0156.txt

00019.txt

00203.txt

00126.txt

00135.txt

UF00086020_00001_0048.txt

UF00086020_00001_0069.txt

00172.txt

00191.txt

00170.txt

00169.txt

00070.txt

00032.txt

UF00086020_00001_0105.txt

UF00086020_00001_0044.txt

00138.txt

00068.txt

UF00086020_00001_0153.txt

UF00086020_00001_0117.txt

UF00086020_00001_0041.txt

UF00086020_00001_0177.txt

00107.txt

UF00086020_00001_0026.txt

UF00086020_00001_0159.txt

00128.txt

UF00086020_00001_0185.txt

UF00086020_00001_0138.txt

00140.txt

00064.txt

UF00086020_00001_0174.txt

00008.txt

UF00086020_00001_0076.txt

00035.txt

UF00086020_00001_0031.txt

00095.txt

UF00086020_00001_0057.txt

00200.txt

UF00086020_00001_0077.txt

UF00086020_00001_0068.txt

UF00086020_00001_0202.txt

00090.txt

UF00086020_00001_0089.txt

UF00086020_00001_0040.txt

00196.txt

UF00086020_00001_0135.txt

00016.txt

UF00086020_00001_0182.txt

00116.txt

UF00086020_00001_0203.txt

00118.txt

00005.txt

00103.txt

UF00086020_00001_0188.txt

UF00086020_00001_0128.txt

UF00086020_00001_0087.txt

UF00086020_00001_0099.txt

00166.txt

00197.txt

00017.txt

00139.txt

00178.txt

UF00086020_00001_0191.txt

00097.txt

UF00086020_00001_0036.txt

UF00086020_00001_0158.txt

UF00086020_00001_0056.txt

00050.txt

00121.txt

UF00086020_00001_0166.txt

UF00086020_00001_0167.txt

00085.txt

UF00086020_00001_0050.txt

00195.txt

UF00086020_00001_0168.txt

00018.txt

00098.txt

copyright.txt

00113.txt

00052.txt

UF00086020_00001_0197.txt

00144.txt

UF00086020_00001_0131.txt

00084.txt

00069.txt

00134.txt

00004.txt

UF00086020_00001_0183.txt

UF00086020_00001_0162.txt

UF00086020_00001_0012.txt

UF00086020_00001_0003.txt

00088.txt

00187.txt

UF00086020_00001_0020.txt

UF00086020_00001_0129.txt

00029.txt

UF00086020_00001_0015.txt

00175.txt

UF00086020_00001_0112.txt

UF00086020_00001_0100.txt

UF00086020_00001_0065.txt

00074.txt

UF00086020_00001_0146.txt

UF00086020_00001_0143.txt

UF00086020_00001_0009.txt

UF00086020_00001_0080.txt

UF00086020_00001_0101.txt

UF00086020_00001_0125.txt

00132.txt

UF00086020_00001_0134.txt

UF00086020_00001_0002.txt

UF00086020_00001_0200.txt

UF00086020_00001_0173.txt

UF00086020_00001_0144.txt

00077.txt

00041.txt

00053.txt

UF00086020_00001_0007.txt

UF00086020_00001_0038.txt

00164.txt

00198.txt

UF00086020_00001_0196.txt

UF00086020_00001_0066.txt

00104.txt

00185.txt

00115.txt

00078.txt

UF00086020_00001_0184.txt

00149.txt

00141.txt

00131.txt

UF00086020_00001_0071.txt

00021.txt

00028.txt

UF00086020_00001_0181.txt

UF00086020_00001_0189.txt

00031.txt

00009.txt

UF00086020_00001_0039.txt

UF00086020_00001_0008.txt

UF00086020_00001_0085.txt

00046.txt

UF00086020_00001_0021.txt

UF00086020_00001_0004.txt

UF00086020_00001_0029.txt

UF00086020_00001_0147.txt

00147.txt

UF00086020_00001_0116.txt

UF00086020_00001_0030.txt

00044.txt

00013.txt

UF00086020_00001_0119.txt

00001.txt

00109.txt

00099.txt

00102.txt

00180.txt

UF00086020_00001_0022.txt

UF00086020_00001_0063.txt

00040.txt

UF00086020_00001_0070.txt

00129.txt

UF00086020_00001_0201.txt

00094.txt

UF00086020_00001_0199.txt

00159.txt

UF00086020_00001_0096.txt

00014.txt

UF00086020_00001_0047.txt

00086.txt

UF00086020_00001_0124.txt

00130.txt

00049.txt

00079.txt

UF00086020_00001_0150.txt

00048.txt

00165.txt

UF00086020_00001_0115.txt

00123.txt

UF00086020_00001_0172.txt

00065.txt

00106.txt

UF00086020_00001_0062.txt

UF00086020_00001_0171.txt

00015.txt

UF00086020_00001_0064.txt

UF00086020_00001_0019.txt

00056.txt

00192.txt

00045.txt

00161.txt

00171.txt

00176.txt

00173.txt

00202.txt

UF00086020_00001_0109.txt

UF00086020_00001_0049.txt

UF00086020_00001_0059.txt

UF00086020_00001_0102.txt

00030.txt

UF00086020_00001_0193.txt

UF00086020_00001_0053.txt

UF00086020_00001_0027.txt

00089.txt

UF00086020_00001_0025.txt

00082.txt

UF00086020_00001_0121.txt

UF00086020_00001_0149.txt

00155.txt

UF00086020_00001_0130.txt

00036.txt

00124.txt

UF00086020_00001_0145.txt

00043.txt

00025.txt

UF00086020_00001_0088.txt

00003.txt

UF00086020_00001_0023.txt

UF00086020_00001_0051.txt


Full Text












BEHAVIOR OF PARTIALLY MISCIBLE ORGANIC COMPOUNDS
IN SIMULATED GROUND WATER SYSTEMS

















BY

PATRICIA V. CLINE


A DISSERTATION PRESENTED TO THE GRADUATE SCHOOL
OF THE UNIVERSITY OF FLORIDA IN
PARTIAL FULFILLMENT OF THE REQUIREMENTS
FOR THE DEGREE OF DOCTOR OF PHILOSOPHY

UNIVERSITY OF FLORIDA

1988


i l Sy I OF FORIDA LIBRARIES















ACKNOWLEDGMENTS

I sincerely appreciate the technical and editorial

assistance provided by my research director, Dr. J. Delfino.

I also thank Dr. P. S. C. Rao and Dr. P. Chadik for their

advice and for providing opportunities for challenging

discussions, and Dr. J. Dorsey and Dr. R. Yost for serving

on my committee and reviewing this dissertation. Each

member of my committee has contributed to my graduate career

through excellent teaching and creating a positive

intellectual environment.

This work was funded by grants from the Florida

Department of Environmental Regulations. Special thanks are

extended to Dr. Geoffrey Watts for his role in securing

funds and providing technical support and comments.

I am grateful to Dr. M. Battiste for discussions of

reaction mechanisms, and for providing the use of his

laboratory for the synthesis of brominated ethanes.

Special thanks go to Angle Harder for her hard work,

Linda Lee for her generosity with analyses and information,

Tom Potter for unselfish computer and mathematical

assistance, and Bill Davis for technical support.

I extend warmest and deepest thanks to my husband Ken

for technical assistance and emotional support, and my son

Brendan for giving me joy.















TABLE OF CONTENTS


ACKNOWLEDGEMENTS .


ABSTRACT .

INTRODUCTION


Chemistry of Alkyl Halides
Gasoline Partitioning .


MATERIALS AND METHODS .. ..


Alkyl Halides
Gasoline .


DEGRADATION OF ALKYL HALIDES


Introduction . . .
Degradation of 1,1,1-Trichloroethane .
Degradation of Brominated Ethanes .
Halogenated Ethenes . .
Structure/Rate Relationships of Alkyl Halides .
Simple SN1/E1 Reactions . .
Comparisons of Geminal Trihalides .
Effect of Additional Halogens on the Alpha
Carbon . . .
Sediment Matrix Affects . .


SOLUBILIZATION AND DEGRADATION OF RESIDUAL TCA .

Behavior of Residual Solvent .
Aqueous Phase Concentrations .
Advection . . .
Degradation Rate . .
Model Parameters and Procedures .
Limitations of the Model Assumptions ..


iii


. . ii


. . 1


S. 3
. 8


. . 1 2
. . 18


. 22











GASOLINE IN GROUND WATER


Background . . 95
Composition of Gasoline .. 95
Multicomponent Liquid-Liquid Equilibria 97
Statistics and Pattern Recognition
Applications . 102
Partitioning of Gasoline Components into Water 105
Fuel/Water Partition Coefficients .. 105
Water Soluble Blending Agents .. ... 113
Prediction of Kfw for Other Components 120
Changes in Concentrations with Time .. 122
Differences in Water Extracts of Gasolines 129
Equilibrium Concentrations of Major
Constituents . .. 131
Visual Comparison of Water Extracts of
Gasoline . .. 131
Preparation of the Data Base for Statistical
Analysis . . 135
Basic Descriptive Statistics .. 138
Bivariate Plots . 141
Stepwise Discriminant Analysis .. 148
Principal Component Analysis .. 155
Summary . . 164

SUMMARY AND CONCLUSIONS . 166

APPENDIX A. SOLUBILITY MEASUREMENTS BY LINDA LEE 171

APPENDIX B. FORTRAN PROGRAM FOR MODELING LOSS OF
RESIDUAL TCA . .. 173

APPENDIX C. AREA COUNT DATA SET FOR STATISTICAL ANALYSIS
OF WATER EXTRACTS OF GASOLINE .. 179

REFERENCES . . 187

BIOGRAPHICAL SKETCH . . .. 194















Abstract of Dissertation Presented to the Graduate School
of the University of Florida in Partial Fulfillment of the
Requirements for the Degree of Doctor of Philosophy


BEHAVIOR OF PARTIALLY MISCIBLE ORGANIC COMPOUNDS
IN SIMULATED GROUND WATER SYSTEMS



By

Patricia V. Cline

August 1988

Chairman: Joseph J. Delfino
Major Department: Environmental Engineering Sciences

Serious ground water contamination problems result from

leaks or spills of organic liquids which are partially

miscible in water. Two important categories of these

liquids include low molecular weight chlorinated solvents

and gasoline.

l,l,l-Trichloroethane (TCA) abiotically degrades in

water forming approximately 17-25% l,l-dichloroethene (1,1-

DCE) via an elimination reaction. The substitution product

is acetic acid. The Arrhenius activation energy is 119 +/-

3 kj/mol with an Arrhenius factor of 2 X 1013 s-1, which

results in an estimated half-life for the degradation at

250C of 10.2 months.

Brominated analogs of TCA hydrolyze 11-13 times faster








than TCA. As the number of bromines increase, the percent

of elimination products increases.

These geminal trihalides degrade by a unimolecular

mechanism (E1/SN1). The rate coefficient for TCA

degradation in buffered water at elevated temperature is

approximately six times greater than hydrolysis of 1-

chloropropane (SN2 mechanism) and more than 100 times

greater than l,l-dichloroethane. In the presence of sodium

thiosulfate, the l-chloropropane degradation rate increased

by more than a factor of 100, l,l-dichloroethane rate by 22

and TCA degradation by approximately two.

Halogenated ethenes are stable at various temperatures

and reaction conditions. Trichloroethene degrades in

alkaline solution at elevated temperature.

l,l,l-Trichloroethane and 1,1-DCE form a near ideal

solution in the solvent phase. The solubility of 1,1-DCE at

24C is 3200 mg/l and the solubility of TCA is approximately

1580 mg/l.

The range of concentrations for major components of

gasoline which partition into water was determined for 65

gasoline samples. Benzene concentrations in the water

extracts ranged from 12.3-130 mg/l and toluene

concentrations ranged from 23-185 mg/l.

Fuel/water partition coefficients of seven major

aromatic constituents were measured for 31 gasoline types

and showed a standard deviation of 10-30%. These









coefficients were highly correlated with the pure component

solubilities.

Chemometric techniques were applied to 20 peaks

measured in the aqueous extracts of the 65 gasolines.

Bivariate plots and principal component analyses show

selected brands have distinguishing equilibrium

concentrations, but complete separation of brands was not

observed.


vii















INTRODUCTION


Liquids organic compounds ar frequent causes of ground

water contamination. Nonaqueous-phase liquids (NAPL) fall

into two broad categories based on their migration patterns

upon reaching ground water. Mineral oils, including crude

oils as well as various refined products like gasoline, are

less dense than water and move vertically through the

unsaturated (vadose) zone and tend to spread laterally upon

reaching the water table. The majority of spills involving

organic fluids which contaminate ground water result from

this group of compounds (Schwille, 1984).

In many industrialized countries, serious threats to

ground water supplies result from low molecular weight

chlorinated solvents. These anthropogenic substances are

more dense than water and vertical rather than lateral

movement dominates upon reaching the water table. The more

common solvents detected in ground water include 1,1,1-

trichloroethane (TCA), trichloroethene (TCE),

tetrachloroethene or perchloroethene (PCE), and various

dichloroethene isomers. In addition to common usage, the

high frequency of detection is attributed to the compounds'

high mobility and relatively high resistance to degradation.








2

Decreases in the concentration of contaminants measured

in environmental samples can occur as a result of various

attenuation mechanisms. These include biodegradation,

volatilization, photooxidation, and dispersion. In the

subsurface, losses from pathways like photoxidation are not

important. Other pathways like volatilization occur at

rates slower than those measured from exposed surfaces.

Aerobic biodegradation can occur in the subsurface providing

adequate oxygen and nutrients are available and that the

contaminants are not present in concentrations which are

toxic for microorganisms.

The major objectives of this research include

determining fuel/water partitioning patterns and measuring

chemical degradation rates to aid in the interpretation of

data from contaminated ground water sites. Field

investigations of sites contaminated by gasoline or

chlorinated solvents typically analyze and report the

presence of constituents which are regulated by the state or

federal government (e.g. priority pollutants). These

components are emphasized in my research.

Many chlorinated organic compounds will degrade in

water by hydrolysis or elimination mechanisms. Due to the

extended residence times of organic pollutants in ground

water, this typically slow abiotic degradation within months

or years can be a significant attenuation mechanism. The

focus of my research on the halogenated solvents is on the











transformation processes, and the factors which affect

reaction pathway and the rates of degradation.

Gasoline is a complex mixture of hydrocarbons. Ground

water contamination by gasoline is characterized by elevated

concentrations of the more water-soluble constituents. The

focus of my research on gasoline hydrocarbons is on the

distribution or partitioning of various components of the

gasoline mixtures into ground water and the variability in

the equilibrium concentrations of major constituents.



Chemistry of Alkyl Halides

Abiotic transformation has been reported for TCA, TCE

and PCE, with less work reported on the dichloroethene

isomers. My research reevaluates previous studies and

further examines the chemistry of these compounds.

Mechanisms are evaluated to aid in predicting behavior of

alkyl halides in complex subsurface environments which can

catalyze reactions, lead to the formation of complexes, or

provide localized microenvironments of variable pH or redox

potential.

Evidence of the importance of the abiotic

transformation of l,l,l-trichloroethane (TCA) has been

presented (Cline et al., 1986). l,l-Dichloroethene or

vinylidene chloride (1,1-DCE) was one of the five most

frequently detected volatile organic compounds in finished

drinking water supplies, other than trihalomethanes,










according to a survey by the US Environmental Protection

Agency (Westrick et al., 1984). Vinylidene chloride (1,1-

DCE) is a highly reactive, flammable liquid which is

primarily used in the production of copolymers with vinyl

chloride or acrylonitrile. Emissions occur during

manufacturing, shipping and production; however, these

emissions represent less than 1% of the total 1,1-DCE

produced (Environmental Protection Agency, 1985). The

common occurrence of this compound as a ground water

contaminant cannot be entirely explained by its production

and usage patterns.

One source of 1,1-DCE develops during the abiotic

degradation of 1,1,1-trichloroethane (TCA). The production

of TCA is more than three times the production of 1,1-DCE,

and unlike 1,1-DCE, it is an end-use product indicating that

emission to the environment is essentially equivalent to the

production (Environmental Protection Agency, 1985). The

presence of 1,1-DCE is typically associated with the

presence of other alkyl halides. Since 1,1-DCE is more

toxic than TCA, the conversion to 1,1-DCE in ground water

can increase the toxicity of the water supply.

The association of 1,1-DCE with TCA can be seen more

dramatically in field data from sites which show high levels

of chlorinated solvents in ground water. A summary of

volatile organic compounds (VOC's) in Arizona's ground water

(Graf, 1986) states that, of the six most commonly detected











VOC's, only three (1,1,1-trichloroethane (TCA),

trichloroethene (TCE), and tetrachloroethene (PCE)) are used

in large quantities at the industrial facilities. The

presence of 1,2-dichloroethene isomers and 1,1-

dichloroethane, particularly with frequent detections of

vinyl chloride, suggest anaerobic biodegradation (Parsons

and Lage, 1985; Bouwer and McCarty, 1983). Selected

locations show very high levels of 1,1-DCE in association

with TCA, and frequently little evidence of biodegradation

(Table 1). The primary source of 1,1-DCE at these locations

appears to be the chemical degradation of TCA, prompting

questions as to the rate of formation of the 1,1-DCE and its

stability in ground water.



Table 1. Maximum Concentrations (jg/L) of VOC's Detected
at Selected Sites in Arizona (Graf, 1986)

Site TCA 1,1-DCE TCE 1,2-DCE

1 630 3320 13000 20
2 490 1320 9
3 9800 10400 410 933
4 98 206 139 106


Two products are formed during the abiotic degradation

of TCA. The elimination product is 1,1-DCE, while the

substitution or hydrolysis product is acetic acid (Figure

1). Previous research (Cline, 1987) described the rate of

degradation of TCA and formation of 1,1-DCE in dilute buffer

solutions (pH 4-10) at temperatures from 27 to 700C.








|CI
C= C
I CI


pl
H3C-CQ'+ + CI
Cl
C1


1,1 -Dichloroethene


ELIMINATION PATHWAY


1,1,,1 -Trichloroethane


OH
H3C-C-CI
CI


H3C-C-CI H3C-C-OH

Acetic Acid


SUBSTITUTION PATHWAY


Figure 1. Abiotic degradation pathways for 1,1,1-trichloroethane.


Cl
CI

H3C-C-CI
CI
Cl








7

Transformation processes are most evident in field data

when the degradation products accumulate and are analyzed

and reported, as shown for TCA. The slow degradation of

priority pollutants to products which are not analyzed or

reported (alcohols, aldehydes, or acids which are not

regulated substances) is not as easily characterized in

field investigations. This may occur during degradation of

chlorinated ethenes. The common occurance of TCE and PCE,

as well as the formation of dichloroethene isomers during

degradation, suggest additional study of pathways of the

chlorinated ethenes.

The relative importance of anaerobic biodegradation

versus chemical degradation on a site (Table 1) may be

inferred by observations of the amount of the biodegradation

product of TCE (cis-1,2-DCE) or of TCA (l,l-dichloroethane)

as compared to the chemical degradation product of TCA, 1,1-

DCE. Specific site conditions can affect the relative rates

of these attenuation mechanisms. Abiotic degradation rates

increase as the ground water temperature increases.

Biodegradation rates may be influenced by many factors

including presence of other organic, redox potential,

oxygen concentration, and nutrients.

The specific objectives of my research are to examine

the degradation rate and pathways for halogenated ethanes

and ethenes and determine factors which may affect these

processes.










Gasoline Partitioning

Gasoline contamination of ground water has become a

major environmental concern. Documented cases of

contamination from underground storage tanks (Florida

Department of Environmental Regulation, 1985) have prompted

enactment of additional legislation, the "State Underground

Petroleum Environmental Response Act of 1986" (SUPER Act),

to protect the ground water and surface waters of the state

of Florida. The SUPER act was designed to maximize ground

water protection, encourage early detection, reporting, and

clean-up of leaking underground storage tanks.

Issues relating to the behavior of gasoline components

in ground water are diverse and complex. Gasoline itself is

a complex mixture of hydrocarbons and some of the factors

which affect the concentration of these constituents in the

subsurface environment (vadose zone and ground water)

include solubility, biodegradability, volatility, soil

sorptive capacity, and dilution.

Components of gasoline may undergo abiotic chemical or

photochemical oxidations through free radical formation.

Thermal degradation is negligible at environmental

temperatures below 800C. Since free radicals are limited in

the subsurface environment, chemical degradation is not

expected to play a significant role there (Bossert and

Bartha, 1984).








9

Aerobic biodegradation will be an important attenuation

mechanism provided that sufficient oxygen and nutrients are

present, and these components typically become limiting

after a spill or leak. Attempts to stimulate aerobic

biodegradation of underground petroleum need to remedy both

nutrient and oxygen deficiencies. In addition, hydrocarbons

in the C5-C9 range (which are typical of gasoline) have

relatively high solvent-type membrane toxicity which will

reduce the number of microorganisms and therefore, decrease

the amount of biodegradation following a gasoline spill

(Bossert and Bartha, 1984).

Sites which have been contaminated by gasoline spills

occasionally report results of the analysis of the "floating

layer." Recovery wells to remove the residual organic

liquid are typically installed as an early remediation

measure. Ground water is typically analyzed for benzene,

toluene, and the xylenes (BTX) and more recently for the

oxygenated gasoline additive methyl tertiary butyl ether

(MTBE).

Concentrations of the BTX or oxygenated constituents

will vary spatially and temporally. At the source, changes

in relative concentrations of hydrocarbon components occur

through weathering, primarily volatilization and

solubilization of the liquid residual organic constituents,

resulting in increasing concentrations of the least mobile

constituents. Compounds detected in ground water











downgradient from the spill occur as a result of transport

from the source, and therefore show higher concentrations of

the more mobile constituents.

The downgradient aqueous concentrations are dependent

on the initial partitioning of the gasoline components into

water at the source. The presence of the residual

hydrocarbon will dominate the partitioning process, with

soils playing an increasing role as the residual hydrocarbon

is depleted. Field data are complex to interpret. This is

due to many factors, including site heterogeneities, well

construction and sampling variables, and lack of detailed

information which can provide estimates of the rates of

partitioning and transport. However, patterns resulting

from physical processes, i.e. partitioning and transport,

may be observed. In Table 2 are summarized the highest

concentrations of BTX components measured in monitoring

wells at various gasoline spill/leak sites in Florida.


Table 2. Maximum concentrations (mg/L) of BTX components
in monitoring wells at selected gasoline
contamination sites in Florida.

County Benzene Toluene Xylenes

Hillsborough 24 64 16
11 46 15
Volusia 10 28 11
8 46 9
Desoto 0.8 60 9



These concentrations are similar to those measured in

laboratory gasoline-water partitioning experiments in this








11

study in spite of differences which exist in the age of the

spills and various physical and biological factors. The

time component for the weathering of gasoline at the source

is dependent on many site-specific factors. Even the

relative contributions of volatilization and solubilization

will depend on conditions like the depth of the water table

at the time of the spill and subsequent water table

fluctuations.

A simplification of the complex problem of determining

patterns of gasoline constituent concentrations following a

spill is to initially focus on the partitioning of gasoline

components from the fuel to water. This allows estimations

of equilibrium concentrations of different components from a

fresh.spill in contact with water. Different brands and

grades of gasolines may then be evaluated to determine if

differences among the source types are distinguishable, and

how differences in composition affect the partitioning

behavior.

The major objectives of the gasoline study include

determination of the variability in the fuel/water partition

coefficients for aromatic constituents. Factors which may

affect the partitioning (concentration, cosolvents) will be

evaluated. Chemometric analyses on hydrocarbon components

present in the aqueous solution in equilibrium with gasoline

will be performed to evaluate similarities and differences

in various brands and grades of gasolines.















MATERIALS AND METHODS

Alkvl Halides

Reagent grade chemicals (Fisher Scientific) were used

to prepare buffers and standard solutions. Phosphate

solutions (0.05 M) were prepared at pH 4.5, 7.0 and 8.5 by

mixing stock solutions and monitoring the pH with a Fisher

Accumet model 230A pH meter. Solutions of 0.05 M potassium

dihydrogen phosphate and 0.05 M potassium hydrogen phosphate

were prepared using distilled deionized water. Equal molar

volumes were used for the pH 7.0 buffer. The phosphate

solutions at pH 4.5 (potassium dihydrogen phosphate) and pH

8.5 (potassium hydrogen phosphate) required minor pH

adjustment using 0.05 M phosphoric acid or potassium

hydroxide solutions.

Stock standard solutions of TCA and 1,1-DCE were

prepared in methanol at concentrations of approximately 1

mg/mL. Working standards were prepared by spiking

approximately 5 pL of the stock standard solution into 10 mL

of distilled deionized water. Aliquots of 100-500 pL of.the

working standards were used to prepare standard curves for

the response of the gas chromatograph (GC) to the

concentration of analyte.










Seawater samples were obtained from the coastal

Atlantic Ocean near Ormond Beach, Florida. Samples were

filtered and subsequently handled similar to the phosphate

solutions.

Ground water samples from two monitoring wells were

obtained from a site in Orlando, Florida, which had been

contaminated by chlorinated solvents. These samples were

purged to remove existing solvents and interfering

substances, then filter (10 pm) sterilized.

Approximately 6.6 mL of the phosphate solutions,

seawater or distilled deionized water were added to 5 mL

(nominal volume) glass ampules (Wheaton Scientific). The

ampules were plugged with cotton and autoclaved for 15

minutes at 1210C.

These ampules were then aseptically spiked with 10 pL

of the stock solution of TCA in methanol and flame sealed

using a Model 524PS sealing unit manufactured by O.I.

Corporation. Final concentrations were approximately 1-3

mg/l. Approximately 0.5 to 1 mL of air space was present in

the ampules after sealing.

Ampules were incubated at 280C (Precision Scientific

Model 6) and at 370C (Precision Scientific Model 4).

Experiments at higher temperatures were performed in a

Magna-Whirl constant temperature water bath (Blue M).

Samples were analyzed using a purge and trap device

(Tekmar LSC-2), interfaced with a Perkin Elmer Model 8410 GC










with flame ionization detector (FID) which employed a 30 m

J&W DB-I, 0.53 mm i.d. wide bore capillary column with a 3

pm stationary film thickness. The temperature program

included a 10 minute hold time at 300C and temperature

ramping of 5C/min to 800C. The helium flow was 2.5 mL/min.

Selected analyses were performed by gas chromatography/mass

spectrometry (GC/MS) for quantification and confirmation of

the formation of 1,1-DCE.

The brominated analogs of TCA were not commercially

available. These compounds were synthesized according to

the protocol described by Stengle and Taylor (1970). Two

hundred and fifty milliliters of carbon disulfide (CS2) were

added to a 500 mL, 3-neck flask that was saturated with HBr

vapors at OOC. Excess vapors were trapped over aqueous KOH.

Five milliliters of TCA were added. Five grams of aluminum

bromide (AlBr3) were added to 100 mL anhydrous CS2, placed

in a dropping funnel, and gradually added to the TCA/CS2/HBr

solution over a period of one hour.

This solution was extracted with ice water made basic

with ammonium hydroxide. The solvent was then removed by

distillation and the residue was filtered. An aliquot of

the mixture was added to methanol and spiked into ampules

containing water. Analysis by purge and trap GC showed two

primary peaks and a secondary peak. The major peaks were

determined by GC/MS to be tribromoethane and

dibromochloroethane. A smaller peak was shown to be










bromodichloroethane. Trichloroethane was below detection

levels ( <30 pg/L )in these analyses.

Some of the spiked ampules were heated for a few hours

to determine if halogenated ethenes would be formed, and if

so, to subsequently determine their corresponding retention

times. Two major peaks were identified by GC/MS to be 1,1-

dibromoethene and l-bromo-l-chloroethene. A sample GC

chromatogram containing reactants and products is shown in

Figure 2, with mass spectra of TBA and DBE in Figure 3.

The same analytical conditions were used for the

brominated compounds as were used for TCA, although the

final temperature was slightly increased.

Pure standards of these compounds were not available

for quantification. The degradation rate was determined

directly from the decrease in area counts, since the

response of the external standard remained consistent during

the time of the experiments. However, determination of

molar concentrations was required to determine the percent

transformation to the elimination product.

The response on an FID is generally related to the

number of carbons and can be affected by functional groups.

To determine if the molar response on the FID was affected

by the type of halogen on the molecule (bromine or

chlorine), I examined the response of the trihalomethane

series for which standards were available (Table 3). The

molar response on the FID was the same for this series of


















FID Response


7.88
10. 1 .
II .15

14.?3


4.87



12.30


17...I


.f-9.47
21.3!

5. 04
J- ,28.11


Figure 2. Sample chromatogram of partially degraded geminal
trihalides.


Compound

1,1,1-Tribromoethane
l,l-Dibromo-l-chloroethane
l-Bromo-l,l-dichloroethane

1,1-Dibromoethene
1-Bromo-l-chloroethene
1,1-Dichloroethene


Retention Time

25.93
23.24
19.47

17.33
12.30
7.80


a t-: r PGN

W .


3. 93









17
73S RCT. TIHC: 18.55 TOT AIJV4D- 132141. BASE PK/MrUtli: 136.9 34530.
U
C 100 26.1
a TEA
S TBA
c 80


los
S40

S105
4 0
"j 20
( 93 10 172 251 266
( D) O 1 i- r ,,
32 80 100 120 140 60 18 200 22 240 260
p m/z *

1MLPVYC/TBA/KCL*e0i1nt4T ,45-450,2:0~OP.24HOV8 7,unD ~ 13S66,~ I 31
DiCS,30M,IUn,8030eS-2SO,S.2D,3SR,SCRYO 874 SCaNS ( 874 SCANS, 15.88 MIMS)
S1.0 MAftS RACCE: 44.0, 269.8 TOTAL 3SUNDs 323S689.

DEE

TBA


SS 110 16S 220 276 331 384 439 494 550 65S 660 715 77 82S


2* 29 RET. TIME: 10.2 TOT GEUII[- 199?422. EFSE PK/ACiUUI': 10S.O/ 46640.

S100 23.4
Sov DBE
*0


4-


S1-,,
I *" o I- -

m/z











Figure 3. Total ion chromatogram (center) for partially
degraded geminal trihalide mixture, with mass spectra for
1,1,1-tribromoethane and 1,1-dibromoethene.








18

compounds. Therefore, the molar response factor for TCA was

used to quantify the ethanes containing bromine and the

molar response factor for 1,1-DCE used to quantify the

brominated ethenes.


Table 3. Relative response


Trihalomethane

1 Chloroform
2 Bromoform
3 Chloroform
4 Bromoform
5 Bromodichloromethane
6 Dibromochloromethane
7 Bromodichloromethane
8 Dibromochloromethane


e of trihalomethanes on

Area
ng nmoles Counts

616 5.15 24.18
924 3.65 15.19
924 7.73 43.98
1386 5.48 22.02
502 3.06 18.47
386 1.85 9.89
1255 7.65 38.79
965 4.63 21.56

Average
Std. Dev.
Rel. Dev.


Response Factor, nmoles/area counts.




Gasoline

Analyses for gasoline constituents were also performed

by GC/FID, using a Perkin-Elmer Model 8410 gas chromatograph

with a 30 m wide bore capillary column (J&W, DB-1) having a

3 pm film thickness. The neat gasoline samples were

analyzed by direct injection of 0.05 pL of the fuel.

Gasoline components dissolved in water were determined by

sparging volatiles from water using a Tekmar LSC-2 Purge and

Trap instrument interfaced to the Perkin-Elmer GC. The

temperature program for both neat gasolines and water


GC/FID


rf*

0.21
0.24
0.18
0.25
0.17
0.19
0.20
0.21

0.21
0.03
13.5%








19

extracts included a 13-minute hold time at 350C, temperature

ramping of 30C/min to 90C, then 5C/min to 2000C. The

helium flow rate was 3.0 mL/min.

Between August and December 1986, subsamples of

gasolines were obtained from the Department of Agriculture

and Consumer Services (DACS) Petroleum Laboratory in

Tallahassee, Florida. These samples were originally

collected by field inspectors and shipped for analysis to

assess compliance with ASTM guidelines and represent various

terminals in northern and central Florida. These samples

represented both summer and winter blends. Subsamples were

collected into 40 mL VOA screw cap vials with Teflon lined

septa and stored on ice prior to analysis.

Local samples were also collected from selected gas

stations in Gainesville. Samples were obtained from the

pump in gasoline safety containers, then a subsample was

transferred to a VOA vial and cooled.

Procedures for evaluating the partitioning of gasoline

into the aqueous phase were reported by Coleman (1984) and

Brookman et al. (1985a). Brookman et al. (1985a) measured

concentrations of aromatic compounds in the aqueous phase

with varying rotation contact times and found a maximum

concentration after two hours. Samples were then

centrifuged to separate the two phases. Coleman et al.

(1984) determined that a rotation contact time of 30 minutes

and an equilibration period of approximately 1 hour produced








20

consistent results and that longer periods had little effect

on the final concentrations.

Saturated, equilibrated solutions of neat gasolines in

contact with distilled, deionized, organic-free water were

prepared. Two mL of gasoline were added to 40 mL water in

VOA vials having Teflon septa. Samples were mixed on a

rotating disk apparatus for 30 minutes at room temperature

(generally 21-230C). The vials then sat undisturbed for one

hour, in an inverted position. Each separated water phase

was removed through the septum at the bottom of the VOA

bottle using a 5 mL syringe. A separate needle was inserted

to allow air to enter the vial so that a vacuum did not form

preventing withdrawal of the water.

Triplicate samples of each water phase were then sealed

in 2 mL crimp-seal vials and refrigerated until the GC

analysis was performed, typically within 2 days. Replicate

extractions, and replicate analyses of extracts were

performed for quality control.

Some overlap or incomplete peak resolution occurred in

the early eluting compounds for both the neat gasoline

samples and the water extracts. Enhancement of the more

water soluble components occurred following aqueous

extraction, making it easier to identify compounds like

benzene and MTBE in the water extract. Toluene was easily

identified in both the neat and water fractions.










When the objective of comparing gasoline samples

involved identification and quantitation of MTBE, analysis

of the water extract provided the most straightforward

interpretation. Although MTBE may be present in gasoline in

quantities approaching 11%, it was more commonly present at

about 5%. MTBE has a lower FID response than the

hydrocarbons, and eluted early in the chromatogram where

several other components also eluted. In samples that did

not contain MTBE, hydrocarbon peaks were present at lower

concentrations at MTBE's retention time. Since MTBE has a

much greater water solubility than these other constituents,

the relative proportion of MTBE to hydrocarbons was

increased in the water extract.















DEGRADATION OF ALKYL HALIDES


Introduction

In this section the degradation kinetics for 1,1,1-

trichloroethane (TCA) and other 1,1,1-trihaloethanes will be

presented and discussed. These compounds degrade in water

forming both elimination and substitution products.

Specific experiments were performed to determine the

mechanism of this reaction and to describe factors which may

effect the rate or pathway of the degradation.

Mechanisms of hydrolysis/elimination have been studied

for many years and numerous reviews, textbook chapters and

empirical concepts have been developed to describe the

chemical degradation of alkyl halides in water. The

following review provides the framework for subsequent

discussions of alkyl halide structure and reaction

mechanisms where specific examples will be presented. The

information was synthesized from several sources (March,

1985; Carey and Sundberg, 1984; Mabey and Mill, 1978;

Bentley and Schleyer, 1977).

Classical SN1, SN2, El and E2 mechanisms have been

defined as early as 1933 (Figure 4). The distinction

between SN1 and SN2 is whether or not the nucleophilic









Unimolecular Mechanisms


I I



-C-C+ + OH-
I I


I I
-C-C+ + X
I I


I I
-C-C-OH
I I


ESh:I


Step 1.



Step 2.



Step 2.


I I-'
-C-C+

H


Bimolecular Mechanisms



OH- + -C-X D- HO C X-- HO-C- + XSN2
i I I


-C-C-
OHI
SH
OH-


C=C
/


+ H20 +


x- E2


Figure 4. Classical substitution and elimination reaction
mechanisms for degradation of alkyl halides in water.


C C
/C = C










attack at the alpha carbon (carbon containing the halogen

leaving group) occurs before the transition state in the

rate determining step, not the extent to which the bond to

the leaving group is broken. Clear cut differences in

substitution reaction mechanisms are apparent in many

reactions. In practice, there is a spectrum of SN2

mechanisms involving varying amounts of nucleophilic attack,

with SN1 being the limiting case where nucleophilic attack

does not occur before the transition state of the rate

determining step.

Unimolecular (SN1 or El) processes are favored by

systems that form stable carbocationsI. A classic example

would be the hydrolysis of t-butyl bromide. The more polar

the solvent, the faster the reaction. An increase in ionic

strength will typically increase the reaction rate, unless

the anion is the leaving group ion (common ion effect). The

reaction is independent of the concentration of nucleophile.

The classic SN2 case occurs in molecules with low

steric hindrance and low carbocation stability. Simple

primary halides react by the SN2 mechanism, while secondary

halides react by an SN2 or intermediate mechanism. Solvent



1 For years these were called "carbonium ions".
Recently, it was determined that the term "carbonium ions"
more accurately refers to pentacoordinated positive ions
(e.g. CH5+) and the more typical positive ion intermediates
(R3C ) are "carbenium ions". The term "carbocation"
includes either type and is generally used to describe any
of these intermediates (March, 1985, p. 141-142).










polarity has less effect on the reaction rate than is

observed for SN1 reactions, but the rate is more sensitive

to changes in concentration or strength of nucleophiles.

The E2 reaction occurs when base attacks the hydrogen

at the carbon adjacent to the carbon containing the leaving

group (beta carbon). This reaction occurs at higher pH and

is more rapid for molecules containing a more acidic

hydrogen.


Degradation of 1.1.1-Trichloroethane (TCA)

The abiotic degradation of TCA was the subject of my

master's thesis (Cline, 1987) which included a detailed

discussion of related degradation studies and illustrations

of the first order decay of TCA in aqueous solution.

Additional data were collected subsequent to those studies.

This included additional concentration measurements in long

term degradation studies and measurements of rate

coefficients in additional matrices. In this section, a

concise comprehensive summary of these data are presented.

A brief synopsis of previous degradation studies of TCA

which have been reported in the literature is summarized

here. Dilling et al. (1975) performed reactivity studies on

selected chlorinated solvents, including TCA. Estimated

rate coefficients were based on four measurements over a

period of one year for each of two sets of reaction ampules;

one set was maintained in the laboratory and a second set

kept outdoors in Midland, Michigan. The same estimated rate











was reported for each experiment, with half-lives of

approximately six months. Reaction products were not

measured.

The hydrolysis of TCA in seawater was reported by

Pearson and McConnell (1975). A half-life of 39 weeks (9

months) was estimated for TCA at 100C with the predominant

reaction being dehydrochlorination to 1,1-DCE. Walraevens

et al. (1974) examined the degradation of TCA in 0.5, 1.0

and 2.0 M sodium hydroxide solutions. The elimination

reaction was not observed, and sodium acetate was shown by

infrared analysis to be the sole reaction product. The

elimination product, 1,1-DCE, was assumed to be stable under

all experimental conditions.

Vogel and McCarty (1987) monitored the degradation of

TCA and formation of 1,1-DCE in water at pH 7 and a

temperature of 20C. The TCA half-life at 200C was

estimated to be between 2.8 and 19 years. Haag and Mill

(1988) report approximately 22% conversion of TCA to the

elimination product, with an extrapolated half-life of 350

days (11.5 months) at 250C.

Degradation experiments were performed at various

temperatures and in different sample matrices. The results

of these experiments are summarized in Table 4. First order

degradation kinetics were observed (Figure 5) in the

data as verified by plotting In [TCA] versus time. Linear

regression analyses were performed on each data set. All

















Table 4. Summary of TCA Degradation Rates
and Product Formation


Temp.
oC


Matrix


108 k
s-1


ke/k
%


70 pH 4
pH 5
pH 7
pH 10
GW1
GW2


62 pH 13

53 pH 4.5
pH 7.0
pH 7.0
pH 8.5
Seawater
DW

39 pH 4.5
pH 7.0
pH 8.5

28 pH 4.5
pH 7.0
pH 8.5


1390 +/- 85
1530 +/- 90
1410 +/-100
1400 +/- 95
1480 +/- 90
1400 +/- 80


565 +/- 35

140 +/- 12
140 +/- 15
144 +/- 20
145 +/- 16
155 +/- 18
133 +/- 14

25 +/- 1.2
24 +/- 1.1
24 +/- 1.2

4.4 +/- 0.2
3.9 +/- 0.2
4.2 +/- 0.2


26 +/- 1

25 +/- 1
26 +/- 2




38 +/- 1

25 +/- 2
24 +/- 2
24 +/- 2
25 +/- 2
"25
23 +/- 3

19 +/- 1
22 +/- 1
17 +/- 1

23 +/- 2
19 +/- 2
21 +/- 2


DW, Distilled organic free water
GW, Ground water matrix


#obs







8.0
7.8
7.6-
.7.4 i
7.2 -
S7.0
6.8 "" --
6.6-
6.4
6.2
6.0 ,
0 100 200 300
Time (days)

350

300

250-

S200

L 150-

100-

50
50-
0 i i -O- -- -- --
0 0.2 0.4 0.6 0.8 1
1 e-kt


Figure 5. First order kinetic data for the degradation of
1,1,1-trichloroethane at 280C and pH 4.5, with the
corresponding data for the formation of the elimination
product, 1,1-dichloroethene.









29

rate constants were based on reactions showing a minimum of

75% degradation of the initial concentration of TCA.

Statistical analyses were performed to assess if the

slopes measured at any given temperature were significantly

different, thus determining the extent to which the sample

matrix, or pH affected the rate constant. The reaction

rates in the buffer solutions (pH 4.5, 7 and 8.5) were not

significantly affected by pH (p < 0.01). In addition, the

rates measured in ground water matrices at 70C (GW1, GW2)

were not significantly different from rates measured in the

buffer solutions at the same temperature.

The spiking solutions typically were prepared with

methanol, which resulted in approximately 0.1% methanol in

the final solution. Separate experiments were conducted

without the use of methanol with no apparent affect on the

rates. The use of methanol decreased the variability in

concentrations observed among ampules, apparently due to the

decreased volatility of TCA in the methanol spiking

solution.

Reaction rates at 530C in seawater, distilled deionized

water and 0.05 M phosphate buffer solutions showed that the

ionic matrix affected the rate of reaction. The fastest

rate was observed for seawater, while the rate in distilled

deionized water (DW) was 14% lower and those in the buffer

solutions were approximately 10% lower. The rates measured

in the distilled deionized water and the buffer solutions











were not significantly different; however, the rate in the

seawater matrix was higher than these at the p<0.01 level.

The 10-14% increase in reaction rate observed in the

seawater matrix at this temperature may be due to the

catalytic influence of some component of that matrix, or to

the increase of ionized species concentration in the

solution.

The relationship between the rate coefficient, k, and

temperature is expressed by the Arrhenius equation,

In k In A EA/RT, where EA is the Arrhenius activation

energy, R is the gas constant, T is the temperature and A is

the Arrhenius pre-exponential factor. The plot of the data

from this and other studies is shown in Figure 6. The plot

includes rates for a variety of matrices including seawater

and sodium hydroxide solutions. Since two products were

formed, the degradation process was complex, but the overall

linearity of the Arrhenius plot implies that a single rate-

determining step is involved in the degradation. Based on

these results, an activation energy of 119+/-3 kJ/mol and an

Arrhenius (A) factor of 2.0x1013 s'1 were calculated.

Extrapolated rate constants and estimated half-lives are

shown in Table 5.

Table 5. Extrapolated Half-Lives for the Degradation of TCA

Temperature (C) Half-life (years)

15 4.5 +/- 0.8
20 2.0 +/- 0.3
25 0.85 +/- 0.13












-10 This dissertation.
v Pearson and McConnell, 1975.
x Vogel and McCarty, 1987.

-12 o Haag and Mill, 1988.
O Dilling et al., 1975.
+ Wolroevens et al., 1974.
-14
Io
O0

S-16-
C-
-J v
-18-



-20



-22 ,
2.8 3 3.2 3.4 3.6

1000/T (OK)

Figure 6. Arrhenius plot for the abiotic degradation of 1,1,1-trichloroethane.









32

Included in the Arrhenius plot are the degradation rate

coefficient for TCA in a pH 13 buffer and also the rate

coefficients calculated by Walraevens et al. (1974) for the

sodium hydroxide solutions. The rates for these high pH

solutions were within the confidence interval for the

regression line, indicating the reaction rate was not

significantly accelerated in alkaline media. The lack of

change in the rate in the presence of a high concentration

of a strong nucleophile (i.e. OH') suggested that the

reaction with the nucleophile occurs after the rate

determining step, characteristic of SN1 reactions.

Similarly, the increase in base strength did not shift the

elimination to an E2 mechanism through a large rate increase

and/or increase in formation of the elimination product.

The rate data which exceeded the confidence interval of

the regression line (Figure 6) were from studies (Vogel and

McCarty, 1987; Pearson and McConnell, 1975) which estimated

the rates of the slow reactions with less than 50%

degradation of the parent compound occurring. Rate

constants calculated for low conversion are more variable

than rates established based on higher amounts of conversion

(Levenspiel, 1972, p. 85). The strong linear Arrhenius

relationship between temperature and rate observed between

25 and 800C, regardless of sample matrix, suggests that

reaction rates at temperatures below 250C can be estimated

by extrapolation.










The elimination product, 1,1-DCE, was measured to

establish the factors which influenced the reaction pathway

(substitution versus elimination). Degradation of 1,1-DCE

was observed only at very high pH and even under those

conditions the rate was slow compared with the degradation

of TCA. Therefore, the ratio of the rate for elimination

(ke) to the total rate of degradation (k) was estimated by

plotting the concentration of 1,1-DCE versus (l-e-kt) where

t is time. The slope of the line equals ([TCA]o (ke/k)),

where (TCA]o is the concentration of TCA at time zero.

This calculation required an estimate for the starting

concentration of TCA. For most experiments, multiple

analyses were performed for the estimate of the initial

concentration. Other authors (eg. Vogel and McCarty, 1987)

have used the intercept in the regression analysis of the

degradation, and this value was used as the estimate of

initial concentrations in this study.

Increases in pH and/or temperature theoretically favor

elimination over substitution. The elimination pathway

(Table 4) ranged from 17 to 38% of the total degradation

rate of TCA. Higher temperatures showed slightly more

transformation to 1,1-DCE over the temperature range

evaluated in these experiments. The percent of TCA

degradation due to elimination was not affected by matrix in

the pH range of 4.5 to 8.5. Seawater had no apparent effect

on the relative proportion of products. The highest percent












elimination pathway was measured in the strongest sodium

hydroxide (pH 13) solution.

Qualitative observations (GC and GC/MS) of TCA

degradation at approximately 600C in 0.5, 1.0 and 2.0 molar

sodium hydroxide solutions, showed the presence of 1,1-DCE,

and separate experiments indicated that 1,1-DCE also slowly

degraded under those conditions. These findings contradict

the results reported by Walraevens et al. (1974) in which

1,1-DCE was not detected in TCA degradation experiments at

high pH. This may be due to differences in analytical

methods, or the slow degradation of 1,1-DCE under their

reaction conditions.


Degradation of Brominated Ethanes

The degradation rates of brominated versus chlorinated

1,1,1-trihaloethanes were compared to provide insight into

the mechanisms and overall behavior of these compounds.

Since bromine is a better "leaving group" than chlorine,

brominated compounds typically degrade faster than their

chlorinated counterparts. In reviewing hydrolysis

degradation processes, Mabey and Mill (1978) concluded that

Br is more reactive than Cl by a factor of 5 to 10.

In a search of Chemical Abstracts, fewer than 20

references were reported for the brominated analog of TCA,

1,1,1-tribromoethane (TBA). Most of the papers addressed

spectra and bond energy studies, while no information on the

hydrolysis of this compound was reported.











Brominated analogs of TCA were not commercially

available. Therefore, TBA was synthesized according to the

methods reported by Stengle and Taylor (1970). The

procedure for the synthesis of 1,1,1-tribromoethane (TBA)

produced a mixture of brominated analogs of TCA. The

primary components were TBA and l,l-dibromo-l-chloroethane

(DBCA), while smaller quantities of l-bromo-l,l-

dichloroethane (BDCA) were present. Kinetic data for

abiotic degradation of TBA and DBCA were measured for

several temperatures while data for BDCA were obtained in

only selected experiments conducted at higher overall

concentrations. Compound structures are illustrated in

Figure 7. The elimination pathway involved loss of HBr to

form the corresponding alkene, the dominant elimination

product was the ethene formed by loss of a bromine. The

substitution pathway forms acetic acid.

Initial degradation experiments involving the

synthesized brominated mixture were conducted in reagent

grade (Milli-Q) water to obtain preliminary data on the

transformation process. Subsequent experiments were

conducted in buffer solutions at pH 4, 7, and 10. The

results of these experiments are summarized in Table 6.

First-order kinetics of degradation were observed, as

were also seen for TCA. Rate constants were calculated

from the linear regression analysis of the plots of the













H Br
HC-CBr
H Br


1,1,1-Tribromoethane (TBA)


H Br
HC-CBr
H CI


1,1-Dibromo-1 -chloroethane (DBCA)


H Br
HC-CCI
H CI


1 -Bromo-1.1 -dichloroethone (BDCA)


/Br
C=C
/ \Br


1,1-Dibromoethene (DBE)


S /Br
C=C


1-Bromo-1 -chloroethene (BCE)


C=C
/ \CI


1,1-Dichloroethene (DCE)


Figure 7. Brominated analogs of 1,1,1-trichloroethane and
corresponding elimination products. Since bromine is a
better leaving group than chlorine, the predominant pathway
is elimination of HBr.














Table 6. Summary of Brominated Compound Degradation
Rate Coefficients and Product Formation


Tem


p.


Matrix Compound


20 DW
20 pH 4
20 pH 7
20 pH 10
28 DW
30 pH 4
30 pH 7
30 pH 10
37 1 M KC1
65 DW
65 Na2S203

20 DW
20 pH 4
20 pH 7
20 pH 10
28 DW
30 pH 4
30 pH 7
30 pH 10
37 1 M KC1
65 DW
65 Na2S203


65 DW


TBA
TBA
TBA
TBA
TBA
TBA
TBA
TBA
TBA
TBA
TBA

DBCA
DBCA
DBCA
DBCA
DBCA
DBCA
DBCA
DBCA
DBCA
DBCA
DBCA


BDCA


108 k
s-1

14
10
9
11
42
65
64
71
492
7700
13000

17
14
11
15
51
81
69
73
492
7860
14000


5350


ke/k


50.9
60.7
58.5
61.8
64.1
61.8
56.3
63.1
38.0
51.6
60.1

35.6
33.9
32.5
33.4
45.6
39.8
31.9
38.5
24.4
33.8
40.6


29.6


DW, Distilled organic free water
Na2S203, 1 M Sodium thiosulfate


Extrapolated Half-Lives for Degradation of TBA and DBCA


TBA


Temp.


15
20
25


108 k
s-1

5.54
12.83
28.92


T 1/2
Days

145
62
28


108 k
s-1

6.90
15.56
34.14


DBCA


T 1/2
Days

116
52
23











natural log of the concentrations versus time. All rate

constants were based on reactions showing a minimum of 75%

degradation.

The results of the degradation of TBA, DBCA and BDCA at

65C are illustrated in Figure 8. The differences in

slopes for the degradation of these compounds were not

statistically significant indicating that the rate

determining step was similar for each compound.

The formation of products (Figure 9) was calculated as

discussed previously for the formation of 1,1-DCE. The

percent elimination (ke/k) was the slope of the regression

line divided by the initial concentration of the parent

product. The smaller slope for 1,1-DCE, and its lower

maximum concentration, was a function of both lower initial

concentration of reactant (BDCA) and lower percent of BDCA

degradation which occurred through the elimination pathway.

The Arrhenius plot for TCA as determined in this study

is compared in Figure 10 with that of the brominated

compounds, TBA and DBCA. The Arrhenius plot for the two

brominated compounds was represented by a single regression

line. The regression line for TCA was essentially parallel

to that of the brominated compounds. The Arrhenius

activation energy (EA) for all of these compounds was almost

identical, since EA is a function of the slope of this line.

The rate of degradation of TCA at 25C was

approximately a factor of 11 to 13 times slower than for the


























0 2 4 6
Time (Hours)
Figure 8. First order kinetic data for the abiotic
degradation of TBA, DBCA and BDCA in water at 650C.


45-
40-
35-
30-
25-
20-
15-
10-


5

0


0.2


0.4


0.6


0.8


-kt
1-e-

Figure 9. Formation of the elimination products (BCE, DBE,
DCE) in water at 650C from the abiotic degradation of the
corresponding 1,1,1-trihaloethanes.


+ BCE :
* DBE ,
* DCE ++


+
+ +
+



S0 0











+ TBA
DBCA
0 TCA
-10




Y -15




-20




2.8 3 3.2 3.4 3.6

1000/T (OK)

Figure 10. Arrhenius plot for the abiotic degradation of 1,1,1-trihaloethanes.










brominated analogs. The observed rate constants for the

various pH values were not significantly different, as was

also observed for the degradation of TCA.

Experiments were performed to determine the reaction

mechanism for these 1,1,1-trihaloethanes. First, the

degradation experiment was conducted in a 1 molar sodium

thiosulfate solution. Sodium thiosulfate is a much stronger

nucleophile than water or hydroxide (Swain and Scott, 1953)

and a dramatic increase in degradation rate in this solution

is indicative of an SN2 reaction in which the nucleophile is

directly involved in the rate determining step. The

degradation rates measured for the brominated 1,1,1-

trihaloethanes increased less than a factor of 2 in the

thiosulfate solution, which may be attributed to the

increased ionic strength of the solution. The percent

elimination was also unaffected by this sample matrix.

To further characterize the mechanism for these

degradation reactions, the brominated geminal trihalides

were placed in a 1 M KC1 solution at 370C. High ionic

strength solutions generally increase the rate of SN1 or El

reactions. When a common ion is present, the rate of the

reverse reaction is enhanced. In the presence of high

concentrations of chloride, chloride may be exchanged for

bromide when an ion pair forms. If BDCA forms an ion pair

and chloride is exchanged, TCA will be formed (Figure 11)

providing evidence of a carbocation intermediate. Even










though TCA will degrade, it is more stable than the

brominated compounds and it may accumulate to detectable

levels.

The BDCA compound had the lowest concentration in the

mixture of the three geminal trihalides in reaction

solution, and TCA concentration was less than 40 ug/l.

After three days of incubation at 370C, the concentration of

TCA rose to approximately 200 ug/l. This was a minor

pathway (less than 5% of the BDCA degraded forming

detectable TCA) in the overall degradation process. 1,1,1-

Trichloroethane was not detected in other sample matrices

during the degradation experiments of the brominated

compounds, indicating that its presence in this solution was

a result of the reverse reaction of carbocation with the

chloride in the solution.

Increasing the extent of bromination increased the

percent of the degradation resulting in the elimination

product (Table 6). The proportion of the total degradation

which resulted in elimination for BDCA at 650C was within

the error estimate for the percent elimination of TCA at

elevated temperatures, and both of these parent compounds

produced 1,1-DCE. The highest percent elimination was

observed for TBA which formed approximately 60% 1,1-

dibromoethene (Figure 12). This may be due to an increase

in steric hindrance in carbocations containing bromine

rather than chlorine, slowing the substitution pathway.









H Cl
HC-CCI
H CI


Exchange
Reaction


H CI
HC-CBr
H Cl

BDCA


H CI
HC-C+
H CI


Ion Pair


Br-


/ El
----v


SN1


HC-C
H \OH
OH


Acetic Acid


Figure 11. Reaction pathways for BDCA in 1 M KC1 solution. The exchange reaction of Cl
with the ion pair forms TCA, which degrades more slowly than the brominated compound.


H
\
HC
H








100

90-

80-

70-

60-

S50-

2 40-

30-

20-

10-

0


Figure 12. Comparison of the percent of the elimination pathway for 1,1,1-trihaloethanes.


TBA


DBCA


BDCA


I~1


TCA










These experiments provided evidence that the abiotic

degradation of 1,1,1-trihaloethanes occurred by SN1/El

rather than SN2 and E2 mechanisms. The trihaloethanes

containing one or more bromine atoms degraded at similar

rates, approximately a factor of 11-13 faster than TCA,

reflecting that bromine was a better leaving group. As the

number of bromines present on the trihaloethanes increased,

the percent of the degradation occurring through the

elimination pathway increased.


Degradation of Halogenated Ethenes

One of the primary objectives of examining the behavior

of halogenated ethenes was to provide an accurate evaluation

of their formation and stability during degradation of the

corresponding ethanes. The literature provided some

evidence that slow degradation of these ethenes may occur at

a rate of interest for ground water studies.

Supporting the possibility of degradation, Dilling et

al. (1975) reported half-lives for the abiotic degradation

of trichloroethene (TCE) of 10.7 months (0.002 day-1) and

for tetrachloroethene (PCE) of 9.9 months at 250C.

Molecular oxygen was present and the degradation rates were

suggested to result from oxidation as well as hydrolysis.

In this often referenced work, it was suggested that

mechanisms of degradation at lower temperatures may differ

from rates extrapolated from studies at higher temperatures.










In a study of hydrolytic decomposition by Pearson and

McConnell (1975), volatilization was extrapolated to zero

and a degradation half-life for TCE of 30 months was

estimated.

Roberts (1985) examined field evidence for the

degradation of various chlorinated organic and estimated

rate constants for both TCE and PCE of approximately 0.003

day-1, which may be due to a variety of factors including

sorption and dilution.

Wilson et al. (1985) studied the aerobic degradation of

TCE, PCE and other compounds in actual aquifer materials

from two sites in Oklahoma and Louisiana. No detectable

biodegradation of these compounds was observed under the

experimental conditions. Since degradation was noted in

autoclaved samples, the authors postulated that TCE and PCE

degradation was likely due to abiotic processes with rates

similar to those reported by Dilling et al. (1975).

The dehydrochlorination reaction of TCE occurs under

basic conditions and generates dichloroacetylene and

hydrogen chloride. This reaction of TCE with base is

spontaneous at room temperature and was responsible for

dichloroacetylene intoxication observed in patients inhaling

TCE-containing air in closed systems equipped with alkali

absorbers (Environmental Protection Agency, 1979).

Dichloroacetylene was detected in the gas phase above

aqueous alkaline solutions with pH 11 to 13 and upon











incubation with moderately alkaline material such as

concrete (Greim et al., 1984). They concluded

dehydrohalogenation can occur under these relatively mild

conditions resulting in toxicity from exposure to the

dichloroacetylene.

Many substitution and addition reactions of TCE have

been carried out in the presence of base. What initially

appeared to be a direct substitution reaction may in fact

have been multistep processes involving intermediates like

carbanions, chloroacetylenes, or carbenes. Rappaport (1969)

reviewed the mechanisms for nucleophilic vinylic

substitution processes in alkaline solutions at elevated

temperatures.

Mechanisms may differ for chemical studies performed

under extreme conditions of temperature and high pH compared

to reactions occurring under more typical environmental

conditions. The possibility of slow nucleophilic attack in

aqueous solution was considered because March (1985) reports

that although vinyl halides are generally considered

resistant to nucleophilic attack, the presence of electron-

withdrawing groups like halogen lower the electron density

of the double bond enhancing nucleophilic substitution or

addition reactions.

In ground water, even very slow degradation may be an

important attenuation mechanism. Since environmental

studies report slow degradation of TCE or PCE in water and











the chemical studies show presence of electron withdrawing

groups like chlorine increases the susceptibility of an

olefin to nucleophilic attack, experiments to evaluate

possible reactions were performed.

References to possible hydrolysis reactions of 1,1-DCE

or its brominated analogs were not found upon review of the

literature. The 1,1-dihaloethenes would be less susceptible

to nucleophilic attack than TCE since fewer electron

withdrawing groups are present. The pure compounds however,

are very reactive and polymerize readily. Their

reactivities in dilute aqueous solution have not been

examined.

The focus of my research with halogenated ethenes was

to examine the stability of these compounds in relatively

dilute aqueous solutions and to determine their

susceptibility to nucleophilic attack. Autooxidation or

other reactions of the pure liquid compounds which may be

present in the vadose zone following a spill could occur,

but these reactions are not addressed here.

There were two major purposes for the examination of

the degradation behavior of halogenated ethenes. First, the

stability of the ethene products formed during the

transformation of the geminal trihalides needed to be

determined to accurately describe the kinetics of the

appearance of these elimination products. Secondly,

previous studies which indicated that halogenated ethenes












like trichloroethene (TCE) and tetrachloroethene (PCE) may

undergo slow abiotic degradation in water at room

temperature with a half-life of less than one year were

reevaluated. The question of possible nucleophilic attack

by water, hydroxide ion, or other nucleophiles must be

addressed to understand the stability of these commonly

detected ground water contaminants.

The stability of 1,1-DCE was evaluated in experiments

that were performed concurrently with the evaluation of TCA

degradation. In the buffer solutions, seawater, and

distilled deionized water, no significant degradation of

1,1-DCE occurred during the course of the evaluation of the

degradation of TCA.

The formation of ethenes containing bromine was

monitored during the degradation studies of the brominated

ethanes, and their concentrations were continually monitored

for some time after the ethane degradation was completed.

Trichloroethene was studied in separate experiments

performed at various temperatures selected to repeat the

experiments conducted by Dilling et al. (1975). In addition

to buffer solutions, one set of ampules was prepared with a

nutrient solution which was not autoclaved, and to which

ground water known to show biological activity was added.

This was done to determine if any degradation which might

have occurred during the long term studies could have been

due to biological activity.










A summary of the results of these experiments is

presented in Table 7. No significant degradation of these

compounds was found in the experimental matrices during the

indicated reaction times, as evidenced by the slopes of In C

vs time which were not significantly different from zero.

The overall coefficient of variation for the observations is

similar to values obtained for simple replicate analyses.

Experiments were also conducted to evaluate the overall

behavior of these compounds under more rigorous conditions.

The literature indicated that halogenated ethenes such as

TCE can undergo elimination to form chloroacetylenes at

elevated pH (Rappaport, 1969). This reaction was verified

by using GC/MS to confirm the formation of dichloroacetylene

from TCE and also chloroacetylene from 1,1-DCE by analysis

of the headspace vapor above an alkaline (1 M NaOH) aqueous

solution of the halogenated ethene which was warmed to

approximately 600C.

The rate of degradation of components in a mixture of

1,1-DCE, TCE and PCE in sodium hydroxide solutions was

examined at 600C (Table 8). These were the matrices used by

Walraevens et al. (1974) in their examination of the

degradation of TCA, wherein they did not observe formation

of 1,1-DCE. One objective was to establish if the

elimination product was stable under their reaction

conditions.











Summary of Experimental Conditions for which
Halogenated Ethenes were Stable.


Time No.
Cmpd. Days Obs.


Temp.
oC


27
27
27
27
37
65

27
27
27
27
37
65


BCE
BCE
BCE
BCE
BCE
BCE

DBE
DBE
DBE
DBE
DBE
DBE

DCE
DCE
DCE
DCE
DCE
DCE
DCE
DCE
DCE
DCE
DCE
DCE
DCE
DCE

TCE
TCE
TCE
TCE
TCE
TCE
TCE
TCE
TCE
TCE


Average
Concentration
mg/L


4.2
4.2
4.2
88
80
392


32
32
28
120
110
690


160
160
160
54
126
6

160
160
160
54
126
6

274
386
386
386
386
386
386
15
15
15
14
14
14
6

274
386
386
386
386
386
386
14
14
14


C.V. Matrix


7%
9%
6%
4%
4%
11%

4%
7%
10%
2%
11%
12%

7%
13%
14%
12%
9%
8%
10%
16%
6%
9%
12%
9%
14%
12%

4%
9%
10%
8%
12%
9%
16%
8%
9%
9%


pH 10
pH 7
pH 4
DW
DW
Thio

pH 10
pH 4
pH 7
DW
DW
Thio


Nutrient
pH 4
pH 7
pH 8.5
pH 4
pH 7
pH 8.5
DW
pH 7
Seawater
pH 4
pH 7
pH 8.5
Thio

Nutrient
pH 4
pH 7
pH 8.5
pH 4
pH 7
pH 8.5
pH 4
pH 7
pH 8.5


DW Distilled organic free water
Thio 1 M Sodium thiosulfate solution


Table 7.


2.7
1.1
1.1
1.1
2.4
2.4
2.4
2.1
2.2
2.3
2.3
2.3
2.3
.4

3.0
1.2
1.2
1.2
1.7
1.7
1.7
1.7
1.7
1.7













Table 8. Second Order Degradation Rates (1 mole-1 hr-1)
of Halogenated Ethenes at 600C
in Sodium Hydroxide Solutions

NaOH TCE 1,1-DCE PCE
Concentration
0.1 M 0.6 0.02 nd
0.5 M 0.28 0.01 nd
1.0 M 0.17 0.01 nd
2.0 M 0.12 0.004 nd

nd no significant degradation occurred after 260 hours.



The rate of degradation of TCE was the greatest among

the tested compounds due to the presence of an acidic

hydrogen (a hydrogen present on a carbon containing a

halogen). The elimination reaction was also an available

pathway for the degradation of 1,1-DCE, although the rate of

degradation was approximately 30 times slower than for TCE

in all solutions except for the 1.0 M NaOH.

Tetrachloroethene (PCE) did not degrade since the

dehydrohalogenation reaction could not occur, and apparently

conditions were not favorable for an addition process.

For environmental applications, there are concerns with

the mildest conditions (temperature and pH) which may still

result in degradation of these compounds. The pathway for

the degradation of ethenes at elevated temperature could

differ from reactions at lower temperatures where the

elimination reaction would be less favorable and a possible

addition reaction could occur instead. Therefore, TCE was








53

incubated at 200C in a solution at pH 12.5. No degradation

was observed during four months of incubation (Table 7).

The experiments demonstrate the resistance of the

halogenated ethenes to degradation in dilute aqueous

solution. Reports of the degradation of these compounds

with half-lives of less than 1 year appear to represent a

process other than abiotic degradation in water. In the

same way as Dilling et al. (1975), my experiments were

conducted in sealed ampules containing a headspace, however

degradation was not observed as reported in their study. I

believe their results may be a result of analytical error.

The half-lives for each experiment were based on four

measurements. The results showed a chemically diverse group

of compounds had similar decreases in concentration and

temperature had little effect on these decreases. A

possible explanation for these results would be a decrease

in instrument response over the year of the study.

The halogenated ethenes generally showed very little

degradation, with the exception of the rapid degradation of

TCE at high pH and temperature. It appears that any

degradation of these compounds in aqueous solution which

occurs, does so under rather extreme conditions and is not

expected to be a dominant process.


Structure/Rate Relationships of Alkyl Halides

In the previous sections degradation patterns and

kinetics were evaluated for various 1,1,1-trihaloethanes in












aqueous solution. A broader perspective on hydrolysis /

elimination reactions can be obtained by comparisons with

other haloalkanes reported in the literature. The

objectives are

1. To compare degradation rates measured for

trihaloethanes of other simple alkyl halides which react by

an SN1/E1 mechanism.

2. To compare degradation rates of trihaloethanes with

other geminal trihalides reported in the literature to

determine structure/activity relationships with changes in

the substituents on the beta carbon, and describe shifts in

mechanisms which may occur for these trihalides.

3. To compare degradation rates and pathways of 1-

chloropropane and l,l-dichloroethane with TCA to show the

effects of increasing number of chlorines on the alpha

carbon.

The classic reaction mechanisms for substitution and

elimination reactions are SN1, SN2, El and E2, as previously

discussed. The presence of various functional groups can

effect the rate and pathway of degradation of an alkyl

halide. For example, rates of hydrolysis are greater for

alkyl halides containing Br rather than for Cl by a factor

of 5 to 10. The rates also increase as the alkyl group goes

from primary to secondary to tertiary in the ratio of

1:10:1000 for chloride. Allyl groups enhance the rate of

hydrolysis of a primary halide by a factor of 5 to 100,










while benzyl groups enhance the rate by a factor of 50.

(Mabey and Mill, 1978)

The formation of stabilized carbocations by electron

donation from the non-bonded electron pairs of halogens

adjacent to the cationic carbon center have been reported

(Olah, 1974). The stabilizing effect was enhanced when two

or even three electron-donating heteroatoms coordinate with

the electron-deficient carbon atom as illustrated in Figure

13. Specific examples, designated as "chlorocarbenium

ions" by Olah (1974), have been identified and are

illustrated in Figure 14.

Simple SN1/E1 Reactions

My data suggested 1,1,1-trihaloethanes form carbocation

intermediates. The intermediate would contain two halogens

and one methyl group. The observed rates and pathways are

compared (Table 9) to compounds containing two methyl groups

and one halogen (2,2-dihalopropanes) and three methyl groups

(t-butyl chloride).

Degradation of tertiary halides like t-butyl chloride

occurs with a carbocation intermediate and these compounds

are resistant to bimolecular nucleophilic displacement. The

half-life for the aqueous degradation of t-butyl chloride is

approximately 23 seconds at 25C with about 19% of the

degradation occurring through the elimination pathway. The

carbocation intermediate is stabilized by the three methyl












+
RC
2


+
x-C c


- x


+
X =


R C
2


c -x


Figure 13. Stabilization of carbocations by halogen (Olah, 1974).


x


zX


X c
I







H

C -CC12
H3
;-< ^ cl


HC


-I
I


- CH3


CH3

-CCI2


-~cI2


CH3


Figure 14. Examples of "chlorocarbenium ions" (Olah, 1974)


3


HC-
3C


~Ilr?


















Table 9. Summary of Degradation Rate Coefficients and
Pathways for Tertiary and Secondary Halides


Compound


t-Butyl Bromide
t-Butyl Chloride

2-Bromo-2-chloropropane
2,2-Dibromopropane
2,2-Dichloropropane

1,1,1-Tribromoethane
l,l-Dibromo-l-chloroethane
1,1,1-Trichloroethane

2-Bromopropane
2-Chloropropane


k (sec-1)
2500


2.98x10-2

1.78x10-4
4.62x10-5
9.09x10-6

2.89x10-7
3.41x10-7
2.62x10-8

3.82x10-6
2.11x10-7


ke/kt
%


Reference


100
100
100


References:
1. March, 1985.
2. Queen and Robertson, 1966.
3. This dissertation.










groups. The rate coefficient at 25C is approximately 106

faster than for TCA.

Queen and Robertson (1966) examined the hydrolysis of

2,2-dihalopropanes. These compounds form carbocation

intermediates with two methyl and one halogen group. The

rate coefficients for the degradation of 2,2-dihalopropanes

are intermediate between t-butyl chloride and the 1,1,1-

trihaloethanes. The mechanism was reported to be SN1/El

based on results of experiments with deuterated gem-

dihalides. The degradation rates of these compounds were

10-50 times higher than of the corresponding secondary

halides (e.g., 2-chloropropane).

The degradation rates were affected by the leaving

group, bromine or chlorine. Also, the structure and

stability of the resulting carbocation affected the rate and

pathway (elimination and/or substitution) of the reaction.

Since bromine was a better leaving group than chlorine,

there was a rate increase when bromine was present as

compared to the corresponding chlorinated compound. 2,2-

Dibromopropane degraded 19 times faster than 2,2-

dichloropropane, while 2-bromo-2-chloropropane degraded 5

times faster than the dichloro compound (Queen and

Robertson, 1966). The 1,1,1-trihaloethanes containing

bromine degraded 11-13 times faster than TCA.

Rates were also increased as the number of methyl

groups present on the carbocation increased. The t-butyl










chloride degraded approximately 3000 times faster than 2,2-

dichloropropane and 106 faster than TCA.

There were two major differences between my results and

those reported by Queen and Robertson (1966). First, they

reported a rate nearly four times higher for 2-bromo-2-

chloropropane than for 2,2-dibromopropane, while the rate

coefficients I measured for the trihaloethanes containing at

least one bromine were approximately equal (within 20%).

Secondly, they report only formation of the elimination

product for all 2,2-dihaloethanes, while the percent

elimination in my experiments was a function of the number

of bromines and was always less than 60%. The percent

elimination for t-butyl chloride was less than the value

obtained for the trihaloethanes.

The effect of alpha halogen is complex, "combining a

negative inductive effect and an electron-releasing

resonance effect" (Queen and Robertson, 1966, p. 1364).

Based on my results and the results for t-butyl chloride,

elimination was not expected as the primary pathway nor the

large difference in rates observed for the two

dihalopropanes which contained a bromine. The rate data

were determined for the dihalopropanes by a conductance

method. Extraction of the products of solvolysis of 2,2-

dibromoethane with CC14 and analysis by vapor phase

chromatography (GC) and nmr showed 2-bromopropene was the

only product in other than trace amounts. It may be that










the substitution product, acetone, would not have

partitioned and been measured using that analytical

protocol.

Comparisons of Geminal Trihalides

A number of compounds in the literature contain a

geminal trihalide group (R-CX3), and many of these compounds

have environmental implications. My experiments on 1,1,1-

trihaloethanes indicated that the -CX3 group was sterically

hindered and resistant to attack by an SN2 mechanism, and

that the halogens could help to stabilize the formation of a

carbocation. The overall rate of degradation of other

geminal trihalides will increase if R also stabilizes the

carbocation. If the beta carbon contains an acidic hydrogen

the mechanism may shift to E2 at elevated pH.

A summary of degradation rates (expressed as reaction

half-lives) of various geminal trihalides is presented in

Table 10. The simplest compounds, trihalomethanes, were

very resistant to hydrolysis. The R- consists only of

hydrogen, which was inadequate to stabilize a carbocation.

The mechanism for this degradation has been determined to be

a base catalyzed process with a carbanion intermediate

(Hine, 1950). The extremely low reactivity also suggests

that steric hindrance may prevent SN2 attack.

By contrast alpha,alpha,alpha-trichlorotoluene has a

half-life of 19 seconds at 250C, which corresponds to a rate

of a factor of 106 greater than for TCA. Therefore, the












Table 10. Half-lives for Abiotic Degradation of
Geminal Trihalides


COMPOUND


STRUCTURE


HALF-LIFE
250C


REFERENCE


Chloroform


Bromoform


CHC13


CHBr3


1,1,1-Trichloroethane CH3CC13


1,1,1-Tribromoethane CH3CBr3


DDT


cI C H CC[3

2


Methoxychlor


3500 yr


690 yr



10.2 mo


1 mo


Mabey and Mill,
1978

Mabey and Mill,
1978


This dissertation


This dissertation


12 yr Wolfe et al., 1977
(pH5, 270C)


1 yr Wolfe et al., 1977
(pH5, 270C)


(CH 0 CH CCI3

2


a,a,a-Trichlorotoluene / CC a
Cd/ 3


19 s Lyman et al., 1982










rate increase was much greater than the factor of 50

reported by Mabey and Mill (1978).

Quemeneur et al. (1971) determined that tri-chloro

compounds of the type p-RC6H4-CC13 (R is OMe, Me, H, Cl, or

NO2) were hydrolyzed in neutral or acidic medium via a

cationic transition state for all types of R substituents.

The hydrolysis of the p-substituted alpha,alpha-

dichlorotoluenes reacted via a cationic mechanism when R is

an electron-donor, and a bimolecular mechanism when R is an

electron-attracting group. These results also supported the

observation that halogens contributed to the stability of

the carbocation. Monochlorotoluene reacts nearly 3000 times

more slowly by an SN2 mechanism than the trichlorotoluene

reacts by the SN1.

Methoxychlor and DDT are two environmentally important

pesticides which contain a geminal trihalide functional

group. Wolfe et al. (1977) provided an in depth examination

of the degradation of these compounds. There is a beta

hydrogen on each of these compounds. At elevated pH the

degradation rate increased as a function of pH and the

elimination products were dominant, suggesting these

structures were more susceptible to degradation by the E2

mechanism than is TCA. While the elimination product, DDE,

was the major product of DDT hydrolysis even at lower pH,

the major product of methoxychlor at pH 7 was the hydrolysis

product, with minor amounts of the elimination product,










DMDE. The hydrolysis products formed were anisoin and

anisil, which were explained by phenyl group rearrangement

after the formation of the carbocation.

Mochida et al. (1967) showed 1,1,1,2-tetrachloroethane

and pentachloroethanes reacted more slowly than TCA under

lower pH conditions, which indicated that chlorines on the

beta carbon decrease the stability of the carbocation. The

presence of these chlorines on the beta carbon however,

increased the acidity of the hydrogens, with enhanced

degradation rates for the tetra and pentachloroethanes by an

E2 mechanism at elevated pH.

There is considerable evidence that geminal trihalides

can form carbocations in the presence of an appropriate

neighboring group. Subsequent reaction pathways may vary

according to the structure of the carbocation resulting in

elimination, substitution, or rearrangements. An E2

reaction may also occur for compounds containing an acidic

hydrogen on the beta carbon.

Effect of Additional Halogens on the Alpha Carbon

The hydrolysis of a simple primary halide, 1-

chloropropane, was compared with the reactivity of 1,1-

dichloroethane and TCA in experiments I performed at

elevated temperature. As the number of hydrogens on the

alpha carbon decrease, steric hindrance can increase and

result in a shift in reaction mechanism. The experiments

were designed to demonstrate the relative rates of










hydrolysis in aqueous solution, and the response to an

increase in concentration of a strong nucleophile whose

effect would be a function of the mechanism.

Based on the literature, simple primary alkyl halides

like l-chloropropane are expected to degrade by an SN2

mechanism. Therefore, l-chloropropane should show an

increase in degradation rate in the presence of a strong

nucleophile, since the nucleophile is involved in the rate

determining step.

Predicting the degradation rate of l,l-dichloroethane

is more difficult. Secondary chlorides, like isopropyl

chloride, have been shown to degrade more quickly than the

primary alkyl halides, possibly by an intermediate

mechanism. Chloride can contribute somewhat to the

stability of a carbocation, however, it is not as effective

as a methyl group as discussed previously. In addition, the

presence of a halogen can increase the steric hindrance at

the alpha carbon.

Comparisons of the degradation rates of these compounds

were made at elevated temperature (650C) in pH 7 buffer

solution, and in a 1 M thiosulfate solution. In the buffer

solution the degradation of TCA was approximately 6 times

faster than the hydrolysis of 1-chloropropane. Degradation

of 1,1-dichloroethane was less than 6% of the rate of 1-

chloropropane degradation. This rate comparison is

illustrated in Figure 15.



















O

o 2



1 +






0 200 400 600
Time (hours)
Figure 15. Pseudo-first-order kinetic data plots for hydrolytic degradation of TCA,
1-chloropropane, and 1,1-dichloroethane in pH 7 buffer solution at 650C.








67

The degradation of 1-chloropropane was enhanced by more

than a factor of 100 in the thiosulfate solution, 1,1-

dichloroethane degraded approximately 22 times faster, and

TCA degradation rate increased less than a factor of 2. The

differences in rate enhancement among these compounds is

attributed to differences in mechanism. Part of the

increase in rate of degradation of TCA in thiosulfate is

attributed to the increasing ionic strength, and TCA

degradation rate was clearly less affected by the presence

of thiosulfate than the other compounds. The rate

enhancement for l,l-dichloroethane was similar to the type

of rate increase which would be observed for secondary

halides which react by an intermediate mechanism.

The thiosulfate solution was used as a matter of

convenience as a strong nucleophile to assist in

demonstrating how knowledge of mechanism may be necessary in

estimating degradation rates as matrices change. Greatest

changes in rates in the presence of sulfur nucleophiles may

be expected for simple primary alkyl halides, and the least

effect occur with compounds which react via an SN1 or El

mechanism.


Sediment Matrix Effects

There is considerable interest in possible effects of

solid surfaces on rates of hydrolysis. Most hydrolysis

experiments are performed in simple buffered aqueous

solution. Contaminants in the vadose zone or ground water










have considerable contact with a variety of aquifer

materials which could potentially affect degradation rate.

Hydrolysis reactions may be affected by factors like acid or

base catalysis, sorption and ionic strength. Since

compounds which react by different mechanisms may be

impacted differently by these solid surfaces, both 1-

chloropropane and TCA were used in degradation experiments

performed in various matrices.

Catalysis of hydrolysis or elimination reactions of

alkyl halides by saturated aquifer materials has not been

demonstrated. Because high concentrations of 1,1-DCE have

been observed in Florida and Arizona at solvent spill sites

contaminated with TCA, the role of sand or other materials

which may influence the degradation of TCA was evaluated.

The nonbiological degradation of pesticides in the

unsaturated zone was shown to play an important role for a

few groups of pesticides, mainly organophosphates and s-

triazines. Clay mineral surfaces have shown catalytic

activity, correlated to their acid strength. This catalytic

process is most important at low moisture content, and

therefore is more important in the vadose zone than beneath

the water table (Saltzman and Mingelgrin, 1984).

Haag and Mill (1988) did not observe significant

differences in the kinetics or products of TCA in contact

with sediment pore water. Epoxide hydrolysis was










accelerated by a factor of four in sediment as compared to

rates in buffered water.

Mabey and Mill (1978) indicated that acid promotion of

the aqueous hydrolysis of halogenated aliphatic hydrocarbons

has not been observed. March (1985) stated that gem-

dihalides can be hydrolyzed in water with either acid or

basic catalysis to give aldehydes or ketones, although the

strength of acid was not addressed.

In a review of elimination reactions in the presence of

polar catalysts, Noller and Kladnig (1976) stated that

"interaction of X with an acid is probably as indispensable

as the reaction of H with base in liquid-phase elimination

reactions, but this function is probably taken over by the

solvent and is less pronounced than the base promoted

process."

Clarification of interactions with polar surfaces may

provide insight into possible effects of sediments or soil

on reaction rates. Clays, for example, contain polar

surfaces which have been shown to catalyze degradation of

some pesticides (Saltzman and Mingelgrin, 1984).

Noller and Kladnig (1976) illustrated elimination

reaction products were a function of the specific catalyst

with 1,1,2-trichloroethane


Cl H
Cl C1 C2 Cl
H H










as reactant. Basic catalysts (e.g., KOH-SiO2) attack the

most acidic H, that at C1, forming more 1,1- than 1,2-

dichloroethene. Acidic catalysts (e.g., silica-alumina)

attack C1 on Cl because the formation of the carbocation is

facilitated by the other Cl on that carbon resulting in the

formation of much more of the 1,2-dichloroethene isomer.

The choice of catalyst will determine the predominant

product giving selectivity to the reaction.

Mochida et al. (1967) reported that the reactivity of

TCA on solid acids was greater than that for other

chlorinated ethanes (mono-, di-, tri- and tetra- chloro

compounds). On solid bases it was less reactive than

penta-, tetra-, and 1,1,2-tri- chloroethanes. The shift in

reactivity of the ethanes with change in catalyst showed

enhanced ability of TCA to form a carbocation by

accelerating the reaction on an acid surface as compared to

the other chlorinated ethanes. There was also the lack of

an acidic beta hydrogen to permit catalysis by base.

Possible catalysis would be compound- and mechanism-

specific. Degradation experiments were performed on 1-

chloropropane (SN2) and TCA (SN1,E1) at 650C in 5 ml

distilled deionized water, with a final concentration of

approximately 2 mg/l. Separate ampules were prepared with

the addition of 0.4 g bentonite clay, 1 g limestone, 1 g

sand, and 0.2 g silica gel.








71

Similar trends were observed for both compounds (Table

11). The slowest rates relative to water were obtained for

both compounds in the sample containing clay, while the

fastest rates were observed in the sand.

The data generally showed greater variability in the

samples containing the solids as compared to the DW system

(Figures 16 and 17) as evidenced by correlation coefficients

less than 0.99. However, the rates of l-chloropropane

degradation in ampules containing solids differed by less

than 10% of the rate obtained for Milli-Q water.

The relative degradation rates for TCA differed more as

a function of matrix than observed for chloropropane,

however, there was also greater variability as evidenced by

the correlation coefficients. In the case of TCA, the

formation of 1,1-DCE was similar in all matrices suggesting

the ratio of products was not affected by the presence of

these solids.

The relatively small differences in rates measured in

these matrices may be due to a variety of factors including

sorption, however significant surface catalysis was not

observed. For this type of saturated system, the amount of

alkyl halide in contact with the surface would be small.

Differences may be attributed to normal variability and

differences in ionic strength or composition of the aqueous

phase in contact with the solids.













Table 11. Matrix Effects for Degradation Rates of
1-Chloropropane and 1,1,1-Trichloroethane at 700C.
Linear Regression Output for the Plot of in C (ug/1)
vs. Time (hours).

Chloropropane


Regression Output: MQ


Clay Limestone Sand


Silica Gel


Constant (Ci)
Std Err of Y Est
R Squared
No. of Observation
Degrees of Freedom


7.20
0.12
0.99
12
10


7.59
0.26
0.94
9
7


7.44
0.07
0.99
7
5


7.14
0.13
0.99
11
9


X Coeff. (Rate) -0.0102 -0.0094 -0.0098 -0.0113

Std Err of Coef. 0.0004 0.0009 0.0004 0.0004


Relative rate
(MQ 1)


1.00


0.92


0.96


1.11


1.1.1-Trichloroethane


Regression Output:

Constant (Ci)
Std Err of Y Est
R Squared
No. of Observations
Degrees of Freedom


MQ Clay Limestone Silica Gel Sand


5.86
0.11
0.99
11
9


6.26
0.12
0.99
7
5


5.23
0.07
1.00
6
4


5.91
0.18
0.98
7
5


6.01
0.28
0.96
9
7


X Coeff. (Rate)
Std Err of Coef.

Relative rate
(MQ 1)


-0.027 -0.020 -0.038
0.0008 0.0009 0.0010


1.00


0.74


1.38


-0.034 -0.040
0.0020 0.0030


1.25


1.45


7.11
0.23
0.96
12
10

-0.0107

0.0007

1.06

























0 --a
0


0 20


Figure 16. Effect
rate of hydrolytic


40 60 80 100 120 140
Time (hours)


of the presence of
degradation of TCA


solid material
at 650C.


on the


Clay
+ Silica Gel
0 Limestone
x
4 Milli-Q Water
Sand

3- A


2-
0

1 + 0
-_ _


100


200


300


400


Time (hours)

Figure 17. Effect of the presence of solid material on the
rate of hydrolytic degradation of l-chloropropane at 650C.








74

These experiments do suggest that TCA in sand aquifers

may show a slightly increased rate as compared to low ionic

strength buffered water experiments. The rate coefficient,

however, will fall within the error limits for the rate

estimate for the degradation of TCA based on the experiments

in buffered distilled water.














SOLUBILIZATION AND DEGRADATION OF RESIDUAL TCA

A computational model was constructed to describe the

attentuation of TCA beneath the water table in the presence

of multiple phases. This simplified scenario for a TCA

spill considered the chemical transformation of TCA to 1,1-

DCE along with advective transport resulting from ground

water flow, of TCA and 1,1-DCE out of this zone containing

the residual solvent. Biodegradation of TCA in this highly

contaminated zone was considered negligible.

The major objective in developing this model was to

describe the relative concentrations of the major

constituents and how their concentrations may change with

time. These trends are illustrated for various ground water

flow rates, change in initial concentrations, and initial

composition.

Behavior of Residual Solvent

The migration pattern of chlorinated hydrocarbons

following a spill is illustrated in Figure 18. These dense

nonaqueous phase liquids (NAPL) will infiltrate the porous

media, with some of the NAPL retained in residual

concentration. The retention capacity for these NAPL in the

unsaturated zone may range from 5 L m-3 (approximately 12

mL/L of pore space) in highly permeable media to 30-50 L m-3








76







Vadose Zone


.. ..... ................................................................


G r o u n d ... ..... ............................................
W a te r .........................
Flow
----...::.---- II ....i....iKE.....uo4 S

S.of.. e:.Pore.sp e...:.....
::::::::::::::::: ::: :: ::::: :::::::::: :::::::::::::::::::::::T C A
:- : --" "," 1 D' '
S.............................................................. ..C




W at.................................................................a t e r
..................:::::..........:::::. Saturated
-..:with Solvent






Chlorinated Solvent Pool
at impermeable layer




Figure 18. Equilibrium model for the attenuation of
residual TCA present beneath the water table.








77

in media of lower permeability (Schwille, 1984). Additional

factors which influence whether the NAPL will reach the

water table include the spilled volume and infiltration

process.

If sufficient volume of dense NAPL reach the water

table, it will sink into the saturated zone and continue to

migrate downward as long as the retention capacity of the

zone is exceeded. Wilson and Conrad (1985) reported

residual hydrocarbon occupying 15-40% of the pore space in

the saturated zone.

Water continues to flow laterally through the water

saturated zone containing residual NAPL. The globules of

NAPL provide a large interface with the water providing a

solution zone, where the initial concentration of a given

component is proportional to its aqueous solubility as

determined by the NAPL composition. These globules are

generally trapped in the larger pore spaces and are being

prevented from entering the smaller pores due to the high

capillary entrance pressure. There is a reduction in

permeability to water where the residual NAPL is present, as

the largest channels become blocked at several places by

discontinuous solvent ganglia. This forces water to flow

around the solvent in fairly thin films and/or be diverted

into the smaller channels whose carrying capacity

(conductivity) is low (Jones, 1985).








78

In laboratory experiments, the initial concentration of

chlorinated solvent was at saturation concentration even

when the layer of sand with residual solvent was thin

(Schwille, 1988). The concentration gradually decreased

until the levels in the water were sufficiently low that

further removal of solvent was slow. At this point

approximately 86% of the residual had been removed.

My model was developed assuming that equilibrium

saturation was maintained, the dissolution of residual

solvent being faster than the degradation or advective

transport of components. Diffusion or hydrodynamic

dispersion was not considered to be a limiting factor in

maintaining equilibrium. The solvent-contaminated zone was

then treated similar to a well-mixed flow reactor.

Interactions of the solutes in the water with the solid

matrix of the saturated zone were considered minimal

providing residual solvent was present; the porous medium

was assumed to provide a matrix in which the residual

solvent was retained.

Once the flow of the NAPL stopped, the subsequent

losses were assumed to occur through degradation or

advection of the compound in the aqueous phase.

Hydrolysis/elimination of TCA occurs much faster in dilute

aqueous solution than would occur for water dissolved in the

TCA solvent phase (Walraevens et al., 1974). Ground water

continues to flow through this zone, although at somewhat










reduced velocities, carrying dissolved components out of

this zone.

Aqueous Phase Concentrations

The quantity of solvent lost each day by advection or

degradation is a function of the concentration of each

component (TCA and 1,1-DCE) in the aqueous phase, which in

turn depends on the composition of the residual NAPL. The

solvent phase may contain TCA and/or 1,1-DCE, or another

solvent which may have been spilled with the TCA.

The distribution of a component between the two liquid

phases can be expressed in terms of fugacity. For ideal

mixtures, the solubility of the solute at any composition is

estimated by multiplying the unit solubility by the mole

fraction of the component in the solvent phase at

equilibrium. Nonideal mixtures form deviations from

linearity. Estimates of aqueous concentrations resulting

from a nonideal solvent mixture requires knowledge of the

activity coefficients at the various mole fraction

compositions. For the simpler ideal case,

[TCA]w x STCA

[DCE]w (l-x) SDCE

x TCAs / (TCAs + DCEs)

where TCAs and DCEs are the number of moles of that compound

in the solvent phase at equilibrium, x is the mole fraction

of TCA in the solvent phase, STCA and SDCE are the pure

component solubilities, and [TCA]w and [DCE]w are the










aqueous phase concentrations at equilibrium. The total

number of moles of TCA in a unit volume of porous media is

the sum of the moles present in the aqueous and solvent

phases.

The model describes changes for TCA spilled on a high

permeability material like sand. As TCA degrades and forms

1,1-DCE, the degradation product partitions into the NAPL

affecting the aqueous phase concentration of TCA (and DCE).

Both the individual solubilities and the solubility of

a mixture of TCA and 1,1-DCE are required in the model and

it was also necessary to assess if mixtures of TCA and 1,1-

DCE deviate significantly from ideality. Literature values

for the solubilities of these constituents vary widely

(Table 12). The solubility data for 1,1-DCE reported by

Lyman (1981), showed as much as a 700% error from a

predicted concentration based on regression relationships.

That estimated concentration is much closer to the

concentrations reported by Verschueren (1977).

Table 12. Solubilities of TCA and 1,1-DCE (mg/L)

Temp (0C TCA 1.1-DCE Source
20 480 400 Pearson and McConnell (1975)
20 4400 2640 Verschueren (1977)
30 1088 3675 Verschueren (1977)
25 273 Lyman (1982)

4 1700 4200 This study.
24 1580 3200 This study.



Measurements (Figure 19) were made on the solubility of

the individual components (TCA and 1,1-DCE) and on the













3000
3000- + + DCE







0
0 +

o i







0 0.2 0.4 0.6 0.8 1.0
Equilibrium Mole Fraction of TCA
(Solvent Phase)
Figure 19. Aqueous solubilities of a binary mixture of TCA and 1,1-DCE as a function of
mole fraction composition in the solvent phase (24 C).










solubility of each with varying compositions of the binary

mixture. Mixtures were at room temperature, approximately

24C.

The pure component solubility of TCA (1580 mg/L or 11.8

mmoles/L) and the solubility of 1,1-DCE in the aqueous phase

(3200 mg/L or 33 mmoles/L) measured at 24C were within the

concentration range listed by Verscheuren (1977) who

reported solubilities at 20 and 300C. This is significantly

higher than solubilities reported by Lyman (1981) and

Pearson and McConnell (1975). The solubility for 1,1-DCE

reported in this dissertation was verified independently by

solubility measurements performed using high performance

liquid chromatography (HPLC) (Linda Lee, University of

Florida, Personal communication, 1988). She measured an

average for the solubility of 1,1-DCE at 24C as 2990 mg/L.

Her report is included in Appendix B.

Verscheuren (1977) reported that the solubility of TCA

at 200C was four times greater than at 3000, a value

approximately three times greater than our result at 240C.

Since the mass lost per unit time from degradation is a

function of aqueous concentration and the first-order

degradation rate coefficient, higher aqueous concentrations

at lower temperatures could compensate for the lower

degradation rate. The solubility of TCA at 40C was measured

to verify this trend. As shown in Table 12, a significant








83

increase in solubility of TCA at lower temperatures was not

observed.

The linearity of the change in solubility with

increasing mole fraction for these two compounds suggested

that 1,1-DCE and TCA form a near-ideal solution in the

solvent phase. Based on these measured data, I assumed that

mole fraction in the solvent phase multiplied by the aqueous

solubility of the pure compound provided a reasonable

estimate of aqueous phase concentration of TCA and 1,1-DCE.

Advection

Loss of TCA from this hypothetical contaminated zone

occurs via advection and degradation, both of which are a

function of the aqueous phase concentration. The relative

importance of these two mechanisms is a function of the flow

velocity advectionn) and the temperature (solubility and

degradation rate). Observations of selected field data

suggest higher concentrations of 1,1-DCE appear in southern

state aquifers where the ground water temperatures are

higher. The model therefore, assigns a temperature of 250C.

The volume of water exchanged through the contaminated

zone is a function of the ground water flow velocity and the

length of the contaminated zone. Fresh water upgradient of

the spill enters the contaminated zone while an equal volume

of water at equilibrium saturation of the contaminants is

displaced. Velocities for the model are expressed as the

per cent of the volume of contaminated water exchanged per







84

day. These values include the "no flow" or "low flow" (0.1%

per day) cases, in which the dominant loss occurs through

degradation. At 0.25% per day, the rate of advection is

comparable to the rate of degradation. Finally, a flow rate

of 0.5% per day represented the case in which the loss of

TCA is primarily due to advection. At flows greater than

0.5% per day the losses would be dominated by the advective

term. These volume exchange rates represent slow flows

and/or very large spill areas. An exchange of 0.5% per day

represents an approximate flow through 5 meters of

contaminated porous media at a rate of 2.5 cm/day.

Degradation Rate

The solubility of TCA affects not only its rate of

advection from the contaminated zone, but also the total

mass of TCA degraded per unit time. The first-order rate

constant at 250C is approximately 0.00226 day-1 as measured

in this study. In a contaminant plume, the half-life for

the degradation of TCA is approximately 10.2 months.

Although the first-order rate coefficient remains constant,

the mass of TCA converted per unit time decreases as the

concentration of TCA in the aqueous phase decreases.

In the model, it was assumed that the TCA concentration

remained at saturation within the zone containing residual

solvent since the TCA that degraded was replaced by

dissolution of the residual solvent. The amount of TCA

degraded per unit time follows zero-order kinetics. The










zero-order rate equals the mass converted per unit time in

the first-order equation as the time increment approaches

zero. This becomes 0.00226 day-1 multiplied by the aqueous

concentration of TCA. A 50% decrease in the solubility

would therefore result in a corresponding 50% decrease in

the mass of TCA degraded per unit time.

Model Parameters and Procedures

Initial conditions for the model include a unit volume

of water (1 liter) in contact with 100 mmoles of TCA. After

equilibrium 11.8 mmoles of TCA will be in the aqueous phase

leaving 88.2 mmoles (approximately 11.8 grams or 8.5 mL) in

the residual solvent phase. The changes in concentration of

TCA or 1,1-DCE in this unit volume are displayed graphically

illustrating the effects of different flow rates, higher

initial mass of TCA, and effect of the presence of an inert

solvent mixed in the residual phase.

Iterative calculations (Appendix C) are made in the

model for advection and degradation in relatively small time

increments, with subsequent reequilibration of the solvent

remaining in the zone of residual contamination. The

residual solvent mass will continue to decrease until at

some point a separate solvent phase does not exist.

Calculations become more difficult (smaller time increments

must be used to attain convergence of the iterative

mathematical solution) and other factors would become more

important as the NAPL is depleted. Therefore, the









86

calculations are stopped when amounts of TCA in the residual

NAPL are less than 10 mmoles. At lower levels of residual

NAPL, the process may become diffusion limited as the NAPL

is trapped in regions of the soil matrix removed from the

aqueous flow. The results of the model are shown in Figures

20-26.

The total mass of TCA in the NAPL showed zero-order

decay with flows from 0.1-0.5% per day (Figure 20). As the

flow rate decreases, slight nonlinearity is observed. This

reflects the slow accumulation of 1,1-DCE in the solvent

phase which begins to decrease the aqueous concentration of

TCA.

The decrease in aqueous concentration of TCA (Figure

21) as the total mass of TCA in the system goes from 100

mmoles to approximately 15 mmoles (slightly in excess of the

solubility) is dependant on the flow. The larger decrease

is observed for the case of no-flow, which results in a 45%

decrease in the aqueous phase concentration after 10 years.

The major reason 1,1-DCE fails to accumulate

significantly in the solvent phase is its higher water

solubility. Having a solubility twice that of TCA, 1,1-DCE

is advected from the zone containing residual solvent more

readily. In the special case of no flow through the system,

1,1-DCE is not advected and begins to accumulate in the

solvent phase affecting the aqueous phase concentration of

TCA. However, since only approximately 20% of the TCA is








100


80


60


40


20


0 2 4 6 8 10


YEARS


Figure 20. Model results: Decrease in total TCA mass in
the residual zone as a function of flow.


C)
Q)
0
F-
E


Q)





H---


YEARS


Figure 21. Model results: Change in aqueous concentration
of TCA as a function of flow.








88

converted to 1,1-DCE, the effect of the accumulation is not

observed until substantial degradation has occurred. If all

the TCA degraded in this closed system, 20 mmoles of 1,1-DCE

would be produced, which is 60% of the pure component

aqueous solubility of 1,1-DCE. Therefore, for the initial

conditions of the model, a residual NAPL will exist only

when excess TCA is present.

A comparison of different initial conditions for a

constant flow (0.25%) is shown in Figure 22. With an

increase in amount of residual TCA, the same zero-order

decay rate is observed, indicating that doubling the amount

of TCA in the solvent phase doubles the time needed for

removal of the residual.

In addition, Figure 22 illustrates the rate of loss of

TCA when the initial 100 mmoles is mixed with another

solvent, a hypothetical mixture in which the mole fraction

of the "inert" compound remains at 0.5 in the solvent phase.

This represents a case where a compound with solubility

similar to TCA (like TCE) is present in the residual. The

presence of this other compound causes a 50% reduction in

the aqueous phase concentration of TCA, and therefore the

rate of loss of TCA, doubling the time to remove the TCA

from the residual phase.

The patterns of change in mass of 1,1-DCE in the

solvent or aqueous phase over time are more complex when

there is advection from the system Figure 23. The aqueous







200
180
160
140
120
100
80
60
40
20
0


YEARS (Flow, 0.25%/Day)


Figure 22. Model results: Change in total mass of TCA as a
function of initial mass of TCA and composition of the
solvent phase.


YEARS (Flow, 0.25%/Day)

Figure 23. Model results: Pattern of 1,1-DCE formation and
advection as 100 mmoles of TCA in the residual zone
degrades.








90

concentration of 1,1-DCE continues to increase for some time

as the mass of 1,1-DCE in the solvent phase begins to

decrease because its mole fraction continues to increase in

the solvent phase.

The total mass of 1,1-DCE in the zone of residual

contamination increased over time, reaching a maximum as the

TCA mass in the solvent phase approached zero. Increasing

the flow rate not only shortened the time in which 1,1-DCE

was accumulating, but decreased the maximum amount of 1,1-

DCE present in that zone. This is true for the aqueous

phase concentrations (Figure 24) and amount in the solvent

phase (Figure 25). The maximum concentration of 1,1-DCE in

the aqueous phase for a flow of 0.5% per day is

approximately 1 mmole/L (100 mg/L) at the point where some

residual phase is still present. The concentration of TCA

at that time is nearly at saturation (approximately 1500

mg/L).

The changes in aqueous concentration of 1,1-DCE for

larger amounts of TCA originally present or in the presence

of an inert solvent as previously discussed, are shown in

Figure 26. The changes in the amount of 1,1-DCE in the

solvent phase is shown in Figure 27. The inert solvent

increases partitioning into the organic phase, keeping the

aqueous concentration low.

The model illustrates factors which affect the time for

removal of a residual phase under varying conditions, and
























2-


0 2 4 6 8
YEARS

Figure 24. Model results: Increase in aqueous
concentration of 1,1-DCE forming from degradation
a function of flow.

7 .


10


of TCA as


YEARS


Figure 25. Model results: Pattern of accumulation of 1,1-
DCE in the solvent phase as TCA degrades.








2.8

2.4

2.0

1.6


1.2

0.8

0.4


0 2 4 6
YEARS (Flow, 0.25%/Day)
Figure 26. Model results: Change in aqueous concentration
of 1,1-DCE as a function of initial mass of TCA and
composition of the solvent phase.
7
---- 100 mmole
-- 200 mmole and
6
100 mmole TCA +
inert organic solvent
5
LU

0 4-


2 3

2-

1

0 ,
0 2 4 6 8 10
YEARS
Figure 27. Model results: Change in total mass of 1,1-DCE
in the residual zone as a function of initial mass of TCA
and composition of the solvent phase.










the different concentrations of 1,1-DCE which would result.

Given a constant initial mass of TCA, the maximum

concentration of 1,1-DCE in the aqueous phase occurs at the

lowest flow rates. For flow rates higher than the 0.5%

volume exchange per day the advective term is dominant and

concentrations of 1,1-DCE in the residual zone remain

negligible.

As long as a residual NAPL is present, aqueous

concentrations are dominated by TCA. Equal concentrations

of TCA and 1,1-DCE in the water from monitoring well data

from various sites would occur according to the model

primarily in the plume of dissolved constituents

downgradient from the residual zone, or in the original

spill area after all residual solvent was dissolved or

degraded. The presence of a low solubility compound in the

solvent phase with the TCA will considerably slow TCA rate

of advection and degradation.

First-order degradation will continue in the ground

water plume downgradient from the source and this process

could be modeled (Kinzelbach, 1985). Evidence of the

formation of 1,1-DCE would support the assignment of a

degradation rate. Assuming similar retardation factors for

TCA and 1,1-DCE, equal concentrations of TCA and 1,1-DCE

would occur after approximately 3 half-lives, approximately

2.5 years at 250C.




Full Text
Toluene
Figure 39. Bivariate plot of area counts of toluene and benzene for the aqueous
extractions of 65 gasoline samples.
145


3
transformation processes, and the factors which affect
reaction pathway and the rates of degradation.
Gasoline is a complex mixture of hydrocarbons. Ground
water contamination by gasoline is characterized by elevated
concentrations of the more water-soluble constituents. The
focus of my research on gasoline hydrocarbons is on the
distribution or partitioning of various components of the
gasoline mixtures into ground water and the variability in
the equilibrium concentrations of major constituents.
Chemistry of Alkvl Halides
Abiotic transformation has been reported for TCA, TCE
and PCE, with less work reported on the dichloroethene
isomers. My research reevaluates previous studies and
further examines the chemistry of these compounds.
Mechanisms are evaluated to aid in predicting behavior of
alkyl halides in complex subsurface environments which can
catalyze reactions, lead to the formation of complexes, or
provide localized microenvironments of variable pH or redox
potential.
Evidence of the importance of the abiotic
transformation of 1,1,1-trichloroethane (TCA) has been
presented (Cline et al., 1986). 1,1-Dichloroethene or
vinylidene chloride (1,1-DCE) was one of the five most
frequently detected volatile organic compounds in finished
drinking water supplies, other than trihalomethanes,


133
consistant dilution and attenuation settings were used.
This approach" would not necessarily be a viable alternative
for routine ground water analyses since the overall
concentrations in the field samples would be quite variable.
The most obvious visual difference (Figure 36) among
the chromatograms resulted from the presence of MTBE. The
peak was very distinct in the water extracts of the
gasolines in which it was detected. Methyl tertiary butyl
ether was detected in Amoco regular unleaded and "Silver"
grades (what is called here "super regular", an intermediate
grade of unleaded between regular and premium) and detected
only randomly in other brands. Because of its irregular
usage, and probable increase in future use, it did not
provide a unique marker.
For samples which did not contain MTBE, the visual
differences became more subtle, since most gasolines contain
similar major constituents. Even when these patterns
appeared distinctive, they were difficult to describe
quantitatively. Premium gasolines can sometimes be
distinguished from regular samples by GC of the neat
gasolines because it usually has fewer peaks corresponding
to aliphatic compounds (Senn and Johnson, (1987). Since the
aliphatic compounds do not partition readily into the
aqueous phase, the ratios of aromatic to aliphatic
constituents were not as easily observed in the gas
chromatograms of the water extracts. However, the higher


Principal Component 2
4
3
2
1
0
-1
-2
-3
-4
-6 -4 -2 0 2 4 6
Principal Component 1
Figure 48. Principal component plot of all grades of Gulf and Phillips gasolines.
G Gulf
P Phillips
(All Grades)
G
G
P
P
P G
p P G
P
P
G
i 1 1 1 1 1 r
163


79
reduced velocities, carrying dissolved components out of
this zone.
Aqueous Phase Concentrations
The quantity of solvent lost each day by advection or
degradation is a function of the concentration of each
component (TCA and 1,1-DCE) in the aqueous phase, which in
turn depends on the composition of the residual NAPL. The
solvent phase may contain TCA and/or 1,1-DCE, or another
solvent which may have been spilled with the TCA.
The distribution of a component between the two liquid
phases can be expressed in terms of fugacity. For ideal
mixtures, the solubility of the solute at any composition is
estimated by multiplying the unit solubility by the mole
fraction of the component in the solvent phase at
equilibrium. Nonideal mixtures form deviations from
linearity. Estimates of aqueous concentrations resulting
from a nonideal solvent mixture requires knowledge of the
activity coefficients at the various mole fraction
compositions. For the simpler ideal case,
[TCA]w = x STCA
[DCE]w = (1-x) SDCE
x = TCAs / (TCAs + DCEs)
where TCAs and DCEs are the number of moles of that compound
in the solvent phase at equilibrium, x is the mole fraction
of TCA in the solvent phase, S>pcA and are the pure
component solubilities, and [TCA]w and [DCE]w are the


Principal Component 2
4
3
2
1
0
-1
-2
-3
-4
-4 -2 0 2 4
A
Amoco
A
C
Chevron
G
Gulf
P
Phillips
S
Shell
U
Union
A
>
O
A
C
A
S
c
sS
G
S
S
su
J
Up
P
P
s
p
U
S
Principal Component 1
Regular Unleaded Gasolines
Figure 43. Plot of principal component scores for 6 brands of regular unleaded gasoline.
157


148
distinguishing characteristic of various types of gasolines.
However, the absolute concentrations in the water extracts
varied among brands. Many Shell samples contained the
higher concentrations of both of these compounds, while the
concentrations in Chevron samples were frequently low.
The plot of toluene versus benzene (Figure 39)
highlighted four Amoco gasoline samples which contained very
high concentrations of toluene, distinguishing them from all
other samples. These samples were all Amoco Superpremium
grade, and only these four samples of this brand and grade
were tested during this study.
Stepwise Discriminant Analysis
The technique of stepwise discriminant analysis was
used to reduce the number of variables to those which are
most useful for discriminating among the several brands and
grades. The SAS procedure "STEPDISC" (SAS Institute Inc.,
1985) was used to select the appropriate subset of
variables. This statistical technique selects variables
based on a minimum level of significance to explain the
variation that exists among types of gasolines. The
variable with the most discriminating power was added first,
then the remaining variables were reevaluated for their
ability to provide additional information. Therefore, a
peak which varies for different types of gasoline may not be
selected if another previously selected peak explains the
same variation.


Ill
800
700
600
500
i 400
300
200
100
0
10 30 50 70 90 110
Benzene (mg/I) in aqueous phase
2.0
1.8
1.6
-o 1.4
O
V)
8 1.2
1.0
0.8
0.6
Toluene (mg/I) in aqueous solution
0 20 40 60 80 100 120 140
Figure 29. Fuel/Water partition coefficient (Kfw) as a
function of concentration of benzene and toluene.


74
These experiments do suggest that TCA in sand aquifers
may show a slightly increased rate as compared to low ionic
strength buffered water experiments. The rate coefficient,
however, will fall within the error limits for the rate
estimate for the degradation of TCA based on the experiments
in buffered distilled water.


85
zero-order rate equals the mass converted per unit time in
the first-order equation as the time increment approaches
zero. This becomes 0.00226 day'-*- multiplied by the aqueous
concentration of TCA. A 50% decrease in the solubility
would therefore result in a corresponding 50% decrease in
the mass of TCA degraded per unit time.
Model Parameters and Procedures
Initial conditions for the model include a unit volume
of water (1 liter) in contact with 100 mmoles of TCA. After
equilibrium 11.8 mmoles of TCA will be in the aqueous phase
leaving 88.2 mmoles (approximately 11.8 grams or 8.5 mL) in
the residual solvent phase. The changes in concentration of
TCA or 1,1-DCE in this unit volume are displayed graphically
illustrating the effects of different flow rates, higher
initial mass of TCA, and effect of the presence of an inert
solvent mixed in the residual phase.
Iterative calculations (Appendix C) are made in the
model for advection and degradation in relatively small time
increments, with subsequent reequilibration of the solvent
remaining in the zone of residual contamination. The
residual solvent mass will continue to decrease until at
some point a separate solvent phase does not exist.
Calculations become more difficult (smaller time increments
must be used to attain convergance of the iterative
mathematical solution) and other factors would become more
important as the NAPL is depleted. Therefore, the


7
Transformation processes are most evident in field data
when the degradation products accumulate and are analyzed
and reported, as shown for TCA. The slow degradation of
priority pollutants to products which are not analyzed or
reported (alcohols, aldehydes, or acids which are not
regulated substances) is not as easily characterized in
field investigations. This may occur during degradation of
chlorinated ethenes. The common occurance of TCE and PCE,
as well as the formation of dichloroethene isomers during
degradation, suggest additional study of pathways of the
chlorinated ethenes.
The relative importance of anaerobic biodegradation
versus chemical degradation on a site (Table 1) may be
inferred by observations of the amount of the biodegradation
product of TCE (cis- 1,2-DCE) or of TCA (1,1-dichloroethane)
as compared to the chemical degradation product of TCA, 1,1-
DCE. Specific site conditions can affect the relative rates
of these attenuation mechanisms. Abiotic degradation rates
increase as the ground water temperature increases.
Biodegradation rates may be influenced by many factors
including presence of other organics, redox potential,
oxygen concentration, and nutrients.
The specific objectives of my research are to examine
the degradation rate and pathways for halogenated ethanes
and ethenes and determine factors which may affect these
processes .


130
Emphasis was placed on a selected number of brands of
gasoline. Sixty-five samples were collected at different
times, and from different sources, to establish if samples
of a single brand provided a unique pattern of constituent
concentrations to distinguish it from a second brand of
gasoline. Chromatograms of the actual neat gasoline samples
were presented and compared among themselves (Harder et al.,
1987). The focus here is on the water-soluble extracts
since the dissolved gasoline constituents may be the only
(or at least the major) components identified during an
investigation of a subsurface contamination site. The
predominant peaks (area counts and frequency of detection)
in the gas chromatograms were selected for statistical
analysis and evaluation.
To be able to generalize about particular brands or
grades of gasoline, samples were collected on at least two
occasions each from different sources (terminals or gasoline
stations). This was done to evaluate the variability in
constituent concentrations for a single gasoline brand and
grade. In addition to the samples analyzed in the partition
experiment described earlier, samples were also collected
locally from gasoline stations located in Gainesville. A
complete list of the data used to establish constituent
concentration patterns for various gasoline brands and
grades is presented in Appendix D.


Table 14. Solubilities of Gasoline Components
in Distilled Water
Component
AROMATICS1
benzene
e thylbenzene
o- xylene
m-xylene
p-xylene
isopropylbenzene
n-propylbenzene
3- 4-ethy1toluene
1.2.4-trimethylbenzene
1,2,3-trimethylbenzene
1.3.5-trimethylbenzene
n-butylbenzene
s-butylbenzene
t-butylbenzene
toluene
naphthalene
PARAFFINS AND OLEFINS1
pentane
me thylcyclopentane
n-butane
1- butene
1 -pentene
dodecane
OXYGENATED BLENDING AGENTS2
MTBE
t BA
E thano1
Me thano1
n- butano 1
s-butanol
1Brookman et al., 1985a
O 7
^Verschueren, 1977
Solubility (mg/L 0 250
1740
161
170
146
156
65.3
55
40
59.0
75 2
48.2
11.8
17.6
29.5
532
31.3
39.5
41.8
61.4
222
148
0.0037
48,000
Miscible
Miscible
Miscible
77,000
125-250,000 @20C


27
Table 4. Summary of TCA Degradation Rates
and Product Formation
Temp .
Matrix
108 k
ke/k
#obs
C
s' 1
%
70
pH 4
1390 +/- 85
26 +/-
1
9
pH 5
1530 +/- 90
6
pH 7
1410 +/-100
25 +/-
1
8
pH 10
1400 +/- 95
26 +/-
2
6
GW1
1480 +/- 90
8
GW2
1400 +/- 80
8
62
PH
13
565
+/-
35
38
+/-
1
15
53
PH
4.5
140
+/-
12
25
+/-
2
21
PH
7.0
140
+/-
15
24
+/-
2
29
PH
7.0
144
+/-
20
24
+/-
2
25
PH
8 5
145
+/-
16
25
+/-
2
21
Seawater
155
+/-
18
"25
24
DW
133
+ /-
14
23
+/-
3
20
39
PH
4.5
25
+/-
1. 2
19
+/-
1
20
PH
7.0
24
+ /-
1.1
22
+/-
1
14
PH
8 5
24
+ /-
1.2
17
+/-
1
18
28
PH
4.5
4 .
4 +/-
0.2
23
+/-
2
22
PH
7.0
3 .
9 +/-
0.2
19
+/-
2
25
PH
8 5
4 .
2 +/-
0.2
21
+/-
2
23
DW, Distilled organic free
GW, Ground water matrix
water


107
si
Figure 28. Comparison of
(A) with the chromatogram
the chromatogram of neat
of its water extract (B)
gasoline


03/18/88
LEE
Solubility Determination of 1,1 DCE
RPLC Analysis: Vaters Radial Compression Column C-18
Mobile Phase 8/70/22 Acetonitrile/Methanol/Water
Flow Rate- 1.5 ml/min
Vaters 490 UV Detector Wavelength-240 nm
AUF-0.02 Time Constant-1.Osee
Retention Time of 1,1 DCE 3.85min
(No impurities detected at operating conditions)
Equilibration Time 24 hours on platform shaker (Low speed)
Centrifuged at 2400rpm (6000 RCF) for 20 minutes
Inmiscible phase still present in all systems at end of equilibration
1,1 DCE received from Pat Cline with no further purification
1,1 DCE Standard prepared in MeOH according to EPA 601 for VOC
Standard Curve:
InJ. Vol.
Cone.
Mass
(uL)
(ug/ml)
ugx!0A3
Area
25
2067.2
51680
780800
15
2067.2
31008
490670
40
2067.2
82688
1176400
25
2067.2
51680
770230
25
2067.2
51680
796700
Samples:
HPLC
Water
1,1 DCE
InJ. Vol.
Cone.
#
Vol.(ml)
Vt. (g)
(uL)
Area
(ug/ml)
1
4.04
0.54
25
1102800
2982.9
2
4.02
0.49
25
1072800
2901.8
3
4.02
0.52
25
1139100
3081.1
Average
Std. Dev.
% Dev.
2988.6
73.3
2.45
Standard Regression Output:
Constant 0
Std Err of Y Est 33716.90
R Squared 0.980940
No. of Observations 5
Degrees of Freedom 4
X Coefficient(s) 14.78815
Std Err of Coef. 0.268141
172


4
o
o
\
o
c.
I
3 -
2 -
1 -
O
+ 1-Chloropropane
600
Time (hours)
Figure 15. Pseudo-first-order kinetic data plots for hydrolytic degradation of TCA,
1-chloropropane, and 1,1-dichloroethane in pH 7 buffer solution at 65C.


21
When the objective of comparing gasoline samples
involved identification and quantitation of MTBE, analysis
of the water extract provided the most straightforward
interpretation. Although MTBE may be present in gasoline in
quantities approaching 11%, it was more commonly present at
about 5%. MTBE has a lower FID response than the
hydrocarbons, and eluted early in the chromatogram where
several other components also eluted. In samples that did
not contain MTBE, hydrocarbon peaks were present at lower
concentrations at MTBE's retention time. Since MTBE has a
much greater water solubility than these other constituents,
the relative proportion of MTBE to hydrocarbons was
increased in the water extract.


132
Table 20. Concentrations of Gasoline Components in
31 Water Extracts (mg/L).
Average
Std. Dev.
Min
Max
Benzene
42.6
18.9
12.3
130
Toluene
69.4
25.4
23.0
185
Ethylbenzene
3.2
0.8
1.3
5 7
m,p- Xylene
11.4
3 8
2.6
22.9
o-Xylene
5.6
1.8
2.6
9 7
n-Propylbenzene
0.4
0.1
0.1
3.0
3-, 4-Ethyltoluene
1.7
0.3
0.8
3.8
1,3,5-Trimethylbenzene
1.0
0.2
0.5
2.8
2-Ethyltoluene
0.7
0.1
0.4
1.6
1,2,3-Trimethylbenzene
0.7
0.2
0.2
2.0


* VV"V UN IVI RSITY of
yr FLORIDA
Internet Distribution Consent Agreement
In reference to the following dissertation:
AUTHOR: Cline, Patricia
TITLE: Behavior of partially miscible organic compounds in simulated ground
water systems / (record number: 1116907)
PUBLICATION DATE: 1988
I, Patricia Cline a as copyright holder for the aforementioned dissertation, hereby grant
specific and limited archive and distribution rights to the Board of Trustees of the University of
Florida and its agents. I authorize the University of Florida to digitize and distribute the
dissertation described above for nonprofit, educational purposes via the Internet or successive
technologies.
This is a non-exclusive grant of permissions for specific off-line and on-line uses for an
indefinite term. Off-line uses shall be limited to those specifically allowed by "Fair Use" as
prescribed by the terms of United States copyright legislation (cf, Title 17, U.S. Code) as well as
to the maintenance and preservation of a digital archive copy. Digitization allows the University
of Florida to generate image- and text-based versions as appropriate and to provide and enhance
access using search software.
This grant of permissions prohibits use of the digitized versions for commercial use or profit.
Signature of Copyright Holder
Patricia Cline
Printed or Typed Name of Copyright Holder/Licensee
Personal Information Blurred
Printed or Typed Phone Number and Email Address of Copyright Holder/Licensee
/A, y
T)ate of Signature


BEHAVIOR OF PARTIALLY MISCIBLE ORGANIC COMPOUNDS
IN SIMULATED GROUND WATER SYSTEMS
BY
PATRICIA V. CLINE
A DISSERTATION PRESENTED TO THE GRADUATE SCHOOL
OF THE UNIVERSITY OF FLORIDA IN
PARTIAL FULFILLMENT OF THE REQUIREMENTS
FOR THE DEGREE OF DOCTOR OF PHILOSOPHY
UNIVERSITY OF FLORIDA
fflWSITY OF FLORIDA LIBRARIES
1988

ACKNOWLEDGMENTS
I sincerely appreciate the technical and editorial
assistance provided by my research director, Dr. J. Delfino.
I also thank Dr. P. S. C. Rao and Dr. P. Chadik for their
advice and for providing opportunities for challenging
discussions, and Dr. J. Dorsey and Dr. R. Yost for serving
on my committee and reviewing this dissertation. Each
member of my committee has contributed to my graduate career
through excellent teaching and creating a positive
intellectual environment.
This work was funded by grants from the Florida
Department of Environmental Regulations. Special thanks are
extended to Dr. Geoffrey Watts for his role in securing
funds and providing technical support and comments.
I am grateful to Dr. M. Battiste for discussions of
reaction mechanisms, and for providing the use of his
laboratory for the synthesis of brominated ethanes.
Special thanks go to Angie Harder for her hard work,
Linda Lee for her generosity with analyses and information,
Tom Potter for unselfish computer and mathematical
assistance, and Bill Davis for technical support.
I extend warmest and deepest thanks to my husband Ken
for technical assistance and emotional support, and my son
Brendan for giving me joy.

TABLE OF CONTENTS
ACKNOWLEDGEMENTS ii
ABSTRACT v
INTRODUCTION ....... 1
Chemistry of Alkyl Halides 3
Gasoline Partitioning 8
MATERIALS AND METHODS 12
Alkyl Halides 12
Gasoline 18
DEGRADATION OF ALKYL HALIDES 22
Introduction 22
Degradation of 1,1,1-Trichloroethane 25
Degradation of Brominated Ethanes 34
Halogenated Ethenes 45
Structure/Rate Relationships of Alkyl Halides . 53
Simple SN1/E1 Reactions 55
Comparisons of Geminal Trihalides 61
Effect of Additional Halogens on the Alpha
Carbon 64
Sediment Matrix Affects 67
SOLUBILIZATION AND DEGRADATION OF RESIDUAL TCA .... 75
Behavior of Residual Solvent 75
Aqueous Phase Concentrations 79
Advection 83
Degradation Rate 84
Model Parameters and Procedures 85
Limitations of the Model Assumptions .... 94
i i i

GASOLINE IN GROUND WATER 95
Background 95
Composition of Gasoline 95
Multicomponent Liquid-Liquid Equilibria ... 97
Statistics and Pattern Recognition
Applications 102
Partitioning of Gasoline Components into Water . 105
Fuel/Water Partition Coefficients 105
Water Soluble Blending Agents 113
Prediction of Kfw for Other Components . 120
Changes in Concentrations with Time 122
Differences in Water Extracts of Gasolines .... 129
Equilibrium Concentrations of Major
Constituents 131
Visual Comparison of Water Extracts of
Gasoline 131
Preparation of the Data Base for Statistical
Analysis 135
Basic Descriptive Statistics 138
Bivariate Plots 141
Stepwise Discriminant Analysis 148
Principal Component Analysis 155
Summary 164
SUMMARY AND CONCLUSIONS 166
APPENDIX A. SOLUBILITY MEASUREMENTS BY LINDA LEE . 171
APPENDIX B. FORTRAN PROGRAM FOR MODELING LOSS OF
RESIDUAL TCA 173
APPENDIX C. AREA COUNT DATA SET FOR STATISTICAL ANALYSIS
OF WATER EXTRACTS OF GASOLINE 179
REFERENCES 187
BIOGRAPHICAL SKETCH 194
iv

Abstract of Dissertation Presented to the Graduate School
of the University of Florida in Partial Fulfillment of the
Requirements for the Degree of Doctor of Philosophy
BEHAVIOR OF PARTIALLY MISCIBLE ORGANIC COMPOUNDS
IN SIMULATED GROUND WATER SYSTEMS
By
Patricia V. Cline
August 1988
Chairman: Joseph J. Delfino
Major Department: Environmental Engineering Sciences
Serious ground water contamination problems result from
leaks or spills of organic liquids which are partially
miscible in water. Two important categories of these
liquids include low molecular weight chlorinated solvents
and gasoline.
1,1,1-Trichloroethane (TCA) abiotically degrades in
water forming approximately 17-25% 1,1-dichloroethene (1,1-
DCE) via an elimination reaction. The substitution product
is acetic acid. The Arrhenius activation energy is 119 +/-
3 kj/mol with an Arrhenius factor of 2 X 10^ s"^, which
results in an estimated half-life for the degradation at
25C of 10.2 months.
Brominated analogs of TCA hydrolyze 11-13 times faster
v

than TCA. As the number of bromines increase, the percent
of elimination products increases.
These geminal trihalides degrade by a unimolecular
mechanism (E1/SN1). The rate coefficient for TCA
degradation in buffered water at elevated temperature is
approximately six times greater than hydrolysis of 1-
chloropropane (SN2 mechanism) and more than 100 times
greater than 1,1 -dichloroethane. In the presence of sodium
thiosulfate, the 1-chloropropane degradation rate increased
by more than a factor of 100, 1,1-dichloroethane rate by 22
and TCA degradation by approximately two.
Halogenated ethenes are stable at various temperatures
and reaction conditions. Trichloroethene degrades in
alkaline solution at elevated temperature.
1,1,1-Trichloroe thane and 1,1-DCE form a near ideal
solution in the solvent phase. The solubility of 1,1-DCE at
24C is 3200 mg/1 and the solubility of TCA is approximately
1580 mg/1.
The range of concentrations for major components of
gasoline which partition into water was determined for 65
gasoline samples. Benzene concentrations in the water
extracts ranged from 12.3-130 mg/1 and toluene
concentrations ranged from 23-185 mg/1.
Fuel/water partition coefficients of seven major
aromatic constituents were measured for 31 gasoline types
and showed a standard deviation of 10-30%. These
vi

coefficients were highly correlated with the pure component
solubilities .
Chemometric techniques were applied to 20 peaks
measured in the aqueous extracts of the 65 gasolines.
Bivariate plots and principal component analyses show
selected brands have distinguishing equilibrium
concentrations, but complete separation of brands was not
observed.
vi i

INTRODUCTION
Liquids organic compounds ar frequent causes of ground
water contamination. Nonaqueous-phase liquids (NAPL) fall
into two broad categories based on their migration patterns
upon reaching ground water. Mineral oils, including crude
oils as well as various refined products like gasoline, are
less dense than water and move vertically through the
unsaturated (vadose) zone and tend to spread laterally upon
reaching the water table. The majority of spills involving
organic fluids which contaminate ground water result from
this group of compounds (Schwille, 1984).
In many industrialized countries, serious threats to
ground water supplies result from low molecular weight
chlorinated solvents. These anthropogenic substances are
more dense than water and vertical rather than lateral
movement dominates upon reaching the water table. The more
common solvents detected in ground water include 1,1,1-
trichloroethane (TCA), trichloroethene (TCE),
tetrachloroethene or perchloroethene (PCE), and various
dichloroethene isomers. In addition to common usage, the
high frequency of detection is attributed to the compounds'
high mobility and relatively high resistance to degradation.
1

2
Decreases in the concentration of contaminants measured
in environmental samples can occur as a result of various
attenuation mechanisms. These include biodegradation,
volatilization, photooxidation, and dispersion. In the
subsurface, losses from pathways like photoxidation are not
important. Other pathways like volatilization occur at
rates slower than those measured from exposed surfaces.
Aerobic biodegradation can occur in the subsurface providing
adequate oxygen and nutrients are available and that the
contaminants are not present in concentrations which are
toxic for microorganisms.
The major objectives of this research include
determining fuel/water partitioning patterns and measuring
chemical degradation rates to aid in the interpretation of
data from contaminated ground water sites. Field
investigations of sites contaminated by gasoline or
chlorinated solvents typically analyze and report the
presence of constituents which are regulated by the state or
federal government (e.g. priority pollutants). These
components are emphasized in my research.
Many chlorinated organic compounds will degrade in
water by hydrolysis or elimination mechanisms. Due to the
extended residence times of organic pollutants in ground
water, this typically slow abiotic degradation within months
or years can be a significant attenuation mechanism. The
focus of my research on the halogenated solvents is on the

3
transformation processes, and the factors which affect
reaction pathway and the rates of degradation.
Gasoline is a complex mixture of hydrocarbons. Ground
water contamination by gasoline is characterized by elevated
concentrations of the more water-soluble constituents. The
focus of my research on gasoline hydrocarbons is on the
distribution or partitioning of various components of the
gasoline mixtures into ground water and the variability in
the equilibrium concentrations of major constituents.
Chemistry of Alkvl Halides
Abiotic transformation has been reported for TCA, TCE
and PCE, with less work reported on the dichloroethene
isomers. My research reevaluates previous studies and
further examines the chemistry of these compounds.
Mechanisms are evaluated to aid in predicting behavior of
alkyl halides in complex subsurface environments which can
catalyze reactions, lead to the formation of complexes, or
provide localized microenvironments of variable pH or redox
potential.
Evidence of the importance of the abiotic
transformation of 1,1,1-trichloroethane (TCA) has been
presented (Cline et al., 1986). 1,1-Dichloroethene or
vinylidene chloride (1,1-DCE) was one of the five most
frequently detected volatile organic compounds in finished
drinking water supplies, other than trihalomethanes,

4
according to a survey by the US Environmental Protection
Agency (Westrick et al., 1984). Vinylidene chloride (1,1-
DCE) is a highly reactive, flammable liquid which is
primarily used in the production of copolymers with vinyl
chloride or acrylonitrile. Emissions occur during
manufacturing, shipping and production; however, these
emissions represent less than 1% of the total 1,1-DCE
produced (Environmental Protection Agency, 1985). The
common occurrence of this compound as a ground water
contaminant cannot be entirely explained by its production
and usage patterns.
One source of 1,1-DCE develops during the abiotic
degradation of 1,1,1-trichloroethane (TCA). The production
of TCA is more than three times the production of 1,1-DCE,
and unlike 1,1-DCE, it is an end-use product indicating that
emission to the environment is essentially equivalent to the
production (Environmental Protection Agency, 1985). The
presence of 1,1-DCE is typically associated with the
presence of other alkyl halides. Since 1,1-DCE is more
toxic than TCA, the conversion to 1,1-DCE in ground water
can increase the toxicity of the water supply.
The association of 1,1-DCE with TCA can be seen more
dramatically in field data from sites which show high levels
of chlorinated solvents in ground water. A summary of
volatile organic compounds (VOC's) in Arizona's ground water
(Graf, 1986) states that, of the six most commonly detected

5
VOC's, only three (1,1,1-trichloroethane (TCA),
trichloroethene (TCE), and tetrachloroethene (PCE)) are used
in large quantities at the industrial facilities. The
presence of 1,2-dichloroethene isomers and 1,1-
dichloroethane, particularly with frequent detections of
vinyl chloride, suggest anaerobic biodegradation (Parsons
and Lage 1985; Bouwer and McCarty, 1983 ). Selected
locations show very high levels of 1,1-DCE in association
with TCA, and frequently little evidence of biodegradation
(Table 1). The primary source of 1,1-DCE at these locations
appears to be the chemical degradation of TCA, prompting
questions as to the rate of formation of the 1,1-DCE and its
stability in ground water.
Table 1. Maximum Concentrations (/g/L) of VOC's Detected
at Selected Sites in Arizona (Graf, 1986)
Site
TCA
1,1-DCE
TCE
1,2-DCE
1
630
3320
13000
20
2
490
1320
9
-
3
9800
10400
410
933
4
98
206
139
106
Two products are formed during the abiotic degradation
of TCA. The elimination product is 1,1-DCE, while the
substitution or hydrolysis product is acetic acid (Figure
1). Previous research (Cline, 1987) described the rate of
degradation of TCA and formation of 1,1-DCE in dilute buffer
solutions (pH 4-10) at temperatures from 27 to 70C.

Cl
I
HvC-C-CI
J I
Cl
1,1,1 -Trichloroethane
F
h3c-q>
ci
+ ci
\
H
Cl
\
/
c = c
/
\
H Cl
1,1 -Dichloroethene
ELIMINATION PATHWAY
r ~
0
i
'/
h3c-c-ci
1

h3c-c-ci
Cl __J
h3c-c-oh
Acetic Acid
SUBSTITUTION PATHWAY
Figure 1. Abiotic degradation pathways for 1,1,1-trichloroethane.

7
Transformation processes are most evident in field data
when the degradation products accumulate and are analyzed
and reported, as shown for TCA. The slow degradation of
priority pollutants to products which are not analyzed or
reported (alcohols, aldehydes, or acids which are not
regulated substances) is not as easily characterized in
field investigations. This may occur during degradation of
chlorinated ethenes. The common occurance of TCE and PCE,
as well as the formation of dichloroethene isomers during
degradation, suggest additional study of pathways of the
chlorinated ethenes.
The relative importance of anaerobic biodegradation
versus chemical degradation on a site (Table 1) may be
inferred by observations of the amount of the biodegradation
product of TCE (cis- 1,2-DCE) or of TCA (1,1-dichloroethane)
as compared to the chemical degradation product of TCA, 1,1-
DCE. Specific site conditions can affect the relative rates
of these attenuation mechanisms. Abiotic degradation rates
increase as the ground water temperature increases.
Biodegradation rates may be influenced by many factors
including presence of other organics, redox potential,
oxygen concentration, and nutrients.
The specific objectives of my research are to examine
the degradation rate and pathways for halogenated ethanes
and ethenes and determine factors which may affect these
processes .

8
Gasoline Partitioning
Gasoline contamination of ground water has become a
major environmental concern. Documented cases of
contamination from underground storage tanks (Florida
Department of Environmental Regulation, 1985) have prompted
enactment of additional legislation, the "State Underground
Petroleum Environmental Response Act of 1986" (SUPER Act),
to protect the ground water and surface waters of the state
of Florida. The SUPER act was designed to maximize ground
water protection, encourage early detection, reporting, and
clean-up of leaking underground storage tanks.
Issues relating to the behavior of gasoline components
in ground water are diverse and complex. Gasoline itself is
a complex mixture of hydrocarbons and some of the factors
which affect the concentration of these constituents in the
subsurface environment (vadose zone and ground water)
include solubility, biodegradability, volatility, soil
sorptive capacity, and dilution.
Components of gasoline may undergo abiotic chemical or
photochemical oxidations through free radical formation.
Thermal degradation is negligible at environmental
temperatures below 80C. Since free radicals are limited in
the subsurface environment, chemical degradation is not
expected to play a significant role there (Bossert and
Bartha, 1984).

9
Aerobic biodegradation will be an important attenuation
mechanism provided that sufficient oxygen and nutrients are
present, and these components typically become limiting
after a spill or leak. Attempts to stimulate aerobic
biodegradation of underground petroleum need to remedy both
nutrient and oxygen deficiencies. In addition, hydrocarbons
in the C5-C9 range (which are typical of gasoline) have
relatively high solvent-type membrane toxicity which will
reduce the number of microorganisms and therefore, decrease
the amount of biodegradation following a gasoline spill
(Bossert and Bartha, 1984).
Sites which have been contaminated by gasoline spills
occasionally report results of the analysis of the "floating
layer." Recovery wells to remove the residual organic
liquid are typically installed as an early remediation
measure. Ground water is typically analyzed for benzene,
toluene, and the xylenes (BTX) and more recently for the
oxygenated gasoline additive methyl tertiary butyl ether
(MTBE).
Concentrations of the BTX or oxygenated constituents
will vary spatially and temporally. At the source, changes
in relative concentrations of hydrocarbon components occur
through weathering, primarily volatilization and
solubilization of the liquid residual organic constituents,
resulting in increasing concentrations of the least mobile
constituents. Compounds detected in ground water

10
downgradient from the spill occur as a result of transport
from the source, and therefore show higher concentrations of
the more mobile constituents.
The downgradient aqueous concentrations are dependent
on the initial partitioning of the gasoline components into
water at the source. The presence of the residual
hydrocarbon will dominate the partitioning process, with
soils playing an increasing role as the residual hydrocarbon
is depleted. Field data are complex to interpret. This is
due to many factors, including site heterogeneities, well
construction and sampling variables, and lack of detailed
information which can provide estimates of the rates of
partitioning and transport. However, patterns resulting
from physical processes, i.e. partitioning and transport,
may be observed. In Table 2 are summarized the highest
concentrations of BTX components measured in monitoring
wells at various gasoline spill/leak sites in Florida.
Table 2. Maximum concentrations (mg/L) of BTX components
in monitoring wells at selected gasoline
contamination sites in Florida.
County
Benzene
Toluene
Xylenes
Hi11sbo rough
24
64
16
11
46
15
Volusia
10
28
11
8
46
9
Desoto
0.8
60
9
These concentrations are similar to those measured in
laboratory gasoline-water partitioning experiments in this

11
study in spite of differences which exist in the age of the
spills and various physical and biological factors. The
time component for the weathering of gasoline at the source
is dependent on many site-specific factors. Even the
relative contributions of volatilization and solubilization
will depend on conditions like the depth of the water table
at the time of the spill and subsequent water table
fluctuations.
A simplification of the complex problem of determining
patterns of gasoline constituent concentrations following a
spill is to initially focus on the partitioning of gasoline
components from the fuel to water. This allows estimations
of equilibrium concentrations of different components from a
fresh spill in contact with water. Different brands and
grades of gasolines may then be evaluated to determine if
differences among the source types are distinguishable, and
how differences in composition affect the partitioning
behavior.
The major objectives of the gasoline study include
determination of the variability in the fuel/water partition
coefficients for aromatic constituents. Factors which may
affect the partitioning (concentration, cosolvents) will be
evaluated. Chemometric analyses on hydrocarbon components
present in the aqueous solution in equilibrium with gasoline
will be performed to evaluate similarities and differences
in various brands and grades of gasolines.

MATERIALS AND METHODS
Alkvl Halides
Reagent grade chemicals (Fisher Scientific) were used
to prepare buffers and standard solutions. Phosphate
solutions (0.05 M) were prepared at pH 4.5, 7.0 and 8.5 by
mixing stock solutions and monitoring the pH with a Fisher
Accuiet model 230A pH meter. Solutions of 0.05 M potassium
dihydrogen phosphate and 0.05 M potassium hydrogen phosphate
were prepared using distilled deionized water. Equal molar
volumes were used for the pH 7.0 buffer. The phosphate
solutions at pH 4.5 (potassium dihydrogen phosphate) and pH
8.5 (potassium hydrogen phosphate) required minor pH
adjustment using 0.05 M phosphoric acid or potassium
hydroxide solutions.
Stock standard solutions of TCA and 1,1-DCE were
prepared in methanol at concentrations of approximately 1
mg/mL. Working standards were prepared by spiking
approximately 5 iL of the stock standard solution into 10 mL
of distilled deionized water. Aliquots of 100-500 pL of.the
working standards were used to prepare standard curves for
the response of the gas chromatograph (GC) to the
concentration of analyte.
12

13
Seawater samples were obtained from the coastal
Atlantic Ocean near Ormond Beach, Florida. Samples were
filtered and subsequently handled similar to the phosphate
solutions.
Ground water samples from two monitoring wells were
obtained from a site in Orlando, Florida, which had been
contaminated by chlorinated solvents. These samples were
purged to remove existing solvents and interfering
substances, then filter (10 pm) sterilized.
Approximately 6.6 mL of the phosphate solutions,
seawater or distilled deionized water were added to 5 mL
(nominal volume) glass ampules (Wheaton Scientific). The
ampules were plugged with cotton and autoclaved for 15
minutes at 121C.
These ampules were then aseptically spiked with 10 pL
of the stock solution of TCA in methanol and flame sealed
using a Model 524PS sealing unit manufactured by O.I.
Corporation. Final concentrations were approximately 1-3
mg/1. Approximately 0.5 to 1 mL of air space was present in
the ampules after sealing.
Ampules were incubated at 28C (Precision Scientific
Model 6) and at 37C (Precision Scientific Model 4).
Experiments at higher temperatures were performed in a
Magna-Whirl constant temperature water bath (Blue M).
Samples were analyzed using a purge and trap device
(Tekmar LSC-2), interfaced with a Perkin Elmer Model 8410 GC

14
with flame ionization detector (FID) which employed a 30 m
J&W DB-1, 0.53 mm i.d. wide bore capillary column with a 3
un stationary film thickness. The temperature program
included a 10 minute hold time at 30C and temperature
ramping of 5C/min to 80G. The helium flow was 2.5 mL/min.
Selected analyses were performed by gas chromatography/mass
spectrometry (GC/MS) for quantification and confirmation of
the formation of 1,1-DCE.
The brominated analogs of TCA were not commercially
available. These compounds were synthesized according to
the protocol described by Stengle and Taylor (1970). Two
hundred and fifty milliliters of carbon disulfide (CS2) were
added to a 500 mL, 3-neck flask that was saturated with HBr
vapors at 0C. Excess vapors were trapped over aqueous KOH.
Five milliliters of TCA were added. Five grams of aluminum
bromide (AlBr3) were added to 100 mL anhydrous CS2, placed
in a dropping funnel, and gradually added to the TCA/CS2/HBr
solution over a period of one hour.
This solution was extracted with ice water made basic
with ammonium hydroxide. The solvent was then removed by
distillation and the residue was filtered. An aliquot of
the mixture was added to methanol and spiked into ampules
containing water. Analysis by purge and trap GC showed two
primary peaks and a secondary peak. The major peaks were
determined by GC/MS to be tribromoethane and
dibromochloroethane. A smaller peak was shown to be

15
bromodichloroethane. Trichloroethane was below detection
levels ( <30 ng/L )in these analyses.
Some of the spiked ampules were heated for a few hours
to determine if halogenated ethenes would be formed, and if
so, to subsequently determine their corresponding retention
times. Two major peaks were identified by GC/MS to be 1,1-
dibromoethene and 1-bromo-1 -chloroethene. A sample GC
chromatogram containing reactants and products is shown in
Figure 2, with mass spectra of TBA and DBE in Figure 3.
The same analytical conditions were used for the
brominated compounds as were used for TCA, although the
final temperature was slightly increased.
Pure standards of these compounds were not available
for quantification. The degradation rate was determined
directly from the decrease in area counts, since the
response of the external standard remained consistent during
the time of the experiments. However, determination of
molar concentrations was required to determine the percent
transformation to the elimination product.
The response on an FID is generally related to the
number of carbons and can be affected by functional groups.
To determine if the molar response on the FID was affected
by the type of halogen on the molecule (bromine or
chlorine), I examined the response of the trihalomethane
series for which standards were available (Table 3). The
molar response on the FID was the same for this series of

Retention Time
16
u
16 .
20
J.B0N
FID Response
17.33
23.2*
25. ?3
Figure 2. Sample chromatogram of partially degraded geminal
trihalides.
Compound Retention Time
1.1.1-Tribromoethane
1.1-Dibromo-l-chioroethane
l-Bromo-1,1-dichloroethane
25.93
23.24
19.47
1.1-Dibromoethene
1-Bromo-l-chloroethene
1.1-Dichloroethene
17.33
12.30
7.80

ni.PVC-'TBP.'KCt.Me'i IHTI .4S-4S8,2 :88P .24M0V87 ,Ut10 HrHI 13S66, 31
SBS, 36M, 1UN,8#306S-2SG,5.2D,3Sft,SCRYO 874 SCOMS < 874 SC8HS, IS.88 MIHS1
1.8 | Mars RftMCCs 44.8, 269.8 TOTOL 8SUMB 3235689.
I I
*!!
ES
110
DEE
I JL
i TBA
16S 220 276 331 384
I
439
I
494
SSO60S
1
660
r"
7lS
778 825
OJC.)
u
ra/z
Figure 3. Total ion chromatogram (center) for
degraded geminal trihalide mixture, with mass
1,1,1-tribromoethane and 1,1-dibromoethene.
partially
spectra for

18
compounds. Therefore, the molar response factor for TCA was
used to quantify the ethanes containing bromine and the
molar response factor for 1,1-DCE used to quantify the
brominated ethenes.
Table 3. Relative response of trihalomethanes on GC/FID
Trihalomethane
ng
nmoles
Area
Counts
rf*
1
Chloroform
616
5.15
24.18
0.21
2
Bromoform
924
3.65
15 19
0.24
3
Chloroform
924
7 73
43.98
0.18
4
Bromoform
1386
5.48
22.02
0.25
5
Bromodichlorome thane
502
3.06
18.47
0.17
6
Dibromochloromethane
386
1 .85
9 .89
0.19
7
Bromodichloromethane
1255
7 .65
38 79
0.20
8
Dibromochloromethane
965
4.63
21.56
0.21
Average 0.21
S td. Dev. 0.03
Re 1. Dev. 13.5%
Response Factor, nmoles/area counts.
Gasoline
Analyses for gasoline consituents were also performed
by GC/FID, using a Perkin-Elmer Model 8410 gas chromatograph
with a 30 m wide bore capillary column (J&W, DB-1) having a
3 pm film thickness. The neat gasoline samples were
analyzed by direct injection of 0.05 pL of the fuel.
Gasoline components dissolved in water were determined by
sparging volatiles from water using a Tekmar LSC-2 Purge and
Trap instrument interfaced to the Perkin-Elmer GC. The
temperature program for both neat gasolines and water

19
extracts included a 13-minute hold time at 35C, temperature
ramping of 3C/min to 90C, then 5C/min to 200C. The
helium flow rate was 3.0 mL/min.
Between August and December 1986, subsamples of
gasolines were obtained from the Department of Agriculture
and Consumer Services (DACS) Petroleum Laboratory in
Tallahassee, Florida. These samples were originally
collected by field inspectors and shipped for analysis to
assess compliance with ASTM guidelines and represent various
terminals in northern and central Florida. These samples
represented both summer and winter blends. Subsamples were
collected into 40 mL VOA screw cap vials with Teflon lined
septa and stored on ice prior to analysis.
Local samples were also collected from selected gas
stations in Gainesville. Samples were obtained from the
pump in gasoline safety containers, then a subsample was
transferred to a VOA vial and cooled.
Procedures for evaluating the partitioning of gasoline
into the aqueous phase were reported by Coleman (1984) and
Brookman et al. (1985a). Brookman et al. (1985a) measured
concentrations of aromatic compounds in the aqueous phase
with varying rotation contact times and found a maximum
concentration after two hours. Samples were then
centrifuged to separate the two phases. Coleman et al.
(1984) determined that a rotation contact time of 30 minutes
and an equilibration period of approximately 1 hour produced

20
consistant results and that longer periods had little effect
on the final concentrations.
Saturated, equilibrated solutions of neat gasolines in
contact with distilled, deionized, organic-free water were
prepared. Two mL of gasoline were added to 40 mL water in
VOA vials having Teflon septa. Samples were mixed on a
rotating disk apparatus for 30 minutes at room temperature
(generally 21-23C). The vials then sat undisturbed for one
hour, in an inverted position. Each separated water phase
was removed through the septum at the bottom of the VOA
bottle using a 5 mL syringe. A separate needle was inserted
to allow air to enter the vial so that a vacuum did not form
preventing withdrawal of the water.
Triplicate samples of each water phase were then sealed
in 2 mL crimp-seal vials and refrigerated until the GC
analysis was performed, typically within 2 days. Replicate
extractions, and replicate analyses of extracts were
performed for quality control.
Some overlap or incomplete peak resolution occurred in
the early eluting compounds for both the neat gasoline
samples and the water extracts. Enhancement of the more
water soluble components occurred following aqueous
extraction, making it easier to identify compounds like
benzene and MTBE in the water extract. Toluene was easily
identified in both the neat and water fractions.

21
When the objective of comparing gasoline samples
involved identification and quantitation of MTBE, analysis
of the water extract provided the most straightforward
interpretation. Although MTBE may be present in gasoline in
quantities approaching 11%, it was more commonly present at
about 5%. MTBE has a lower FID response than the
hydrocarbons, and eluted early in the chromatogram where
several other components also eluted. In samples that did
not contain MTBE, hydrocarbon peaks were present at lower
concentrations at MTBE's retention time. Since MTBE has a
much greater water solubility than these other constituents,
the relative proportion of MTBE to hydrocarbons was
increased in the water extract.

DEGRADATION OF ALKYL HALIDES
Introduction
In this section the degradation kinetics for 1,1,1-
trichloroethane (TCA) and other 1,1,1 -trihaloethanes will be
presented and discussed. These compounds degrade in water
forming both elimination and substitution products.
Specific experiments were performed to determine the
mechanism of this reaction and to describe factors which may
effect the rate or pathway of the degradation.
Mechanisms of hydrolysis/elimination have been studied
for many years and numerous reviews, textbook chapters and
empirical concepts have been developed to describe the
chemical degradation of alkyl halides in water. The
following review provides the framework for subsequent
discussions of alkyl halide structure and reaction
mechanisms where specific examples will be presented. The
information was synthesized from several sources (March,
1985; Carey and Sundberg, 1984; Mabey and Mill, 1978;
Bentley and Schleyer, 1977).
Classical SN1, SN2, El and E2 mechanisms have been
defined as early as 1933 (Figure 4). The distinction
between SN1 and SN2 is whether or not the nucleophilic
22

23
Step 1.
Step 2.
Step 2.
OH- +
OH
Unimolecular Mechanisms
-C-C-X
i i
i i
-C-C +
I I
+ X
C C+ + OH
i i
i i
-C-C-OH
i i
SN1
i i
-C-C +
I
H
\
/
/
c = c
\
El
Bimolecular Mechanisms
C-X
\ /
HO C X
HO-C- + X
SN2
Figure 4. Classical substitution and elimination reaction
mechanisms for degradation of alkyl halides in water.

24
attack at the alpha carbon (carbon containing the halogen
leaving group) occurs before the transition state in the
rate determining step, not the extent to which the bond to
the leaving group is broken. Clear cut differences in
substitution reaction mechanisms are apparent in many
reactions. In practice, there is a spectrum of SN2
mechanisms involving varying amounts of nucleophilic attack,
with SN1 being the limiting case where nucleophilic attack
does not occur before the transition state of the rate
determining step.
Unimolecular (SN1 or El) processes are favored by
systems that form stable carbocations^. A classic example
would be the hydrolysis of t-butyl bromide. The more polar
the solvent, the faster the reaction. An increase in ionic
strength will typically increase the reaction rate, unless
the anion is the leaving group ion (common ion effect). The
reaction is independent of the concentration of nucleophile.
The classic SN2 case occurs in molecules with low
steric hindrance and low carbocation stability. Simple
primary halides react by the SN2 mechanism, while secondary
halides react by an SN2 or intermediate mechanism. Solvent
1 For years these were called "carbonium ions".
Recently, it was determined that the term "carbonium ions"
more accurately refers to pentacoordinated positive ions
(e.g. CH5+) and the more typical positive ion intermediates
(R3C+) are "carbenium ions". The term "carbocation"
includes either type and is generally used to describe any
of these intermediates (March, 1985, p. 141-142).

25
polarity has less effect on the reaction rate than is
observed for SN1 reactions, but the rate is more sensitive
to changes in concentration or strength of nucleophiles.
The E2 reaction occurs when base attacks the hydrogen
at the carbon adjacent to the carbon containing the leaving
group (beta carbon). This reaction occurs at higher pH and
is more rapid for molecules containing a more acidic
hydrogen.
Degradation of 1 1 1-Trichloroethane (TCA)
The abiotic degradation of TCA was the subject of my
master's thesis (Cline, 1987) which included a detailed
discussion of related degradation studies and illustrations
of the first order decay of TCA in aqueous solution.
Additional data were collected subsequent to those studies.
This included additional concentration measurements in long
term degradation studies and measurements of rate
coefficients in additional matrices. In this section, a
concise comprehensive summary of these data are presented.
A brief synopsis of previous degradation studies of TCA
which have been reported in the literature is summarized
here. Dilling et al. (1975) performed reactivity studies on
selected chlorinated solvents, including TCA. Estimated
rate coefficients were based on four measurements over a
period of one year for each of two sets of reaction ampules;
one set was maintained in the laboratory and a second set
kept outdoors in Midland, Michigan. The same estimated rate

26
was reported for each experiment, with half-lives of
approximately six months. Reaction products were not
measured.
The hydrolysis of TCA in seawater was reported by
Pearson and McConnell (1975). A half-life of 39 weeks (9
months) was estimated for TCA at 10C with the predominant
reaction being dehydrochlorination to 1,1-DCE. Walraevens
et al. (1974) examined the degradation of TCA in 0.5, 1.0
and 2.0 M sodium hydroxide solutions. The elimination
reaction was not observed, and sodium acetate was shown by
infrared analysis to be the sole reaction product. The
elimination product, 1,1-DCE, was assumed to be stable under
all experimental conditions.
Vogel and McCarty (1987) monitored the degradation of
TCA and formation of 1,1-DCE in water at pH 7 and a
temperature of 20C. The TCA half-life at 20C was
estimated to be between 2.8 and 19 years. Haag and Mill
(1988) report approximately 22% conversion of TCA to the
elimination product, with an extrapolated half-life of 350
days (11.5 months) at 25C.
Degradation experiments were performed at various
temperatures and in different sample matrices. The results
of these experiments are summarized in Table 4. First order
degradation kinetics were observed (Figure 5) in the
data as verified by plotting In [TCA] versus time. Linear
regression analyses were performed on each data set. All

27
Table 4. Summary of TCA Degradation Rates
and Product Formation
Temp .
Matrix
108 k
ke/k
#obs
C
s' 1
%
70
pH 4
1390 +/- 85
26 +/-
1
9
pH 5
1530 +/- 90
6
pH 7
1410 +/-100
25 +/-
1
8
pH 10
1400 +/- 95
26 +/-
2
6
GW1
1480 +/- 90
8
GW2
1400 +/- 80
8
62
PH
13
565
+/-
35
38
+/-
1
15
53
PH
4.5
140
+/-
12
25
+/-
2
21
PH
7.0
140
+/-
15
24
+/-
2
29
PH
7.0
144
+/-
20
24
+/-
2
25
PH
8 5
145
+/-
16
25
+/-
2
21
Seawater
155
+/-
18
"25
24
DW
133
+ /-
14
23
+/-
3
20
39
PH
4.5
25
+/-
1. 2
19
+/-
1
20
PH
7.0
24
+ /-
1.1
22
+/-
1
14
PH
8 5
24
+ /-
1.2
17
+/-
1
18
28
PH
4.5
4 .
4 +/-
0.2
23
+/-
2
22
PH
7.0
3 .
9 +/-
0.2
19
+/-
2
25
PH
8 5
4 .
2 +/-
0.2
21
+/-
2
23
DW, Distilled organic free
GW, Ground water matrix
water

DCE (ug/l) In [JCA] ug/l
1 e kt
Figure 5. First order kinetic data for the degradation of
1,1,1-trichloroethane at 28C and pH 4.5, with the
corresponding data for the formation of the elimination
product, 1,1-dichloroethene.
28

29
rate constants were based on reactions showing a minimum of
75% degradation of the initial concentration of TCA.
Statistical analyses were performed to assess if the
slopes measured at any given temperature were significantly
different, thus determining the extent to which the sample
matrix, or pH affected the rate constant. The reaction
rates in the buffer solutions (pH 4.5, 7 and 8.5) were not
significantly affected by pH (p < 0.01). In addition, the
rates measured in ground water matrices at 70C (GW1, GW2)
were not significantly different from rates measured in the
buffer solutions at the same temperature.
The spiking solutions typically were prepared with
methanol, which resulted in approximately 0.1% methanol in
the final solution. Separate experiments were conducted
without the use of methanol with no apparent affect on the
rates. The use of methanol decreased the variability in
concentrations observed among ampules, apparently due to the
decreased volatility of TCA in the methanol spiking
solution.
Reaction rates at 53C in seawater, distilled deionized
water and 0.05 M phosphate buffer solutions showed that the
ionic matrix affected the rate of reaction. The fastest
rate was observed for seawater, while the rate in distilled
deionized water (DW) was 14% lower and those in the buffer
solutions were approximately 10% lower. The rates measured
in the distilled deionized water and the buffer solutions

30
were not significantly different; however, the rate in the
seawater matrix was higher than these at the p<0.01 level.
The 10-14% increase in reaction rate observed in the
seawater matrix at this temperature may be due to the
catalytic influence of some component of that matrix, or to
the increase of ionized species concentration in the
solution.
The relationship between the rate coefficient, k, and
temperature is expressed by the Arrhenius equation,
In k In A E^/RT, where is the Arrhenius activation
energy, R is the gas constant, T is the temperature and A is
the Arrhenius pre-exponential factor. The plot of the data
from this and other studies is shown in Figure 6. The plot
includes rates for a variety of matrices including seawater
and sodium hydroxide solutions. Since two products were
formed, the degradation process was complex, but the overall
linearity of the Arrhenius plot implies that a single rate
determining step is involved in the degradation. Based on
these results, an activation energy of 119+/-3 kJ/mol and an
Arrhenius (A) factor of 2.0x10^ s" ^ were calculated.
Extrapolated rate constants and estimated half-lives are
shown in Table 5.
Table 5. Extrapolated Half-Lives for the Degradation of TCA
Temperature (C) Half-life (years)
15
4.5
+/-
0.8
20
2.0
+/-
0.3
25
0.85
+/-
0.13

Ln k (sec
1000/T (K)
Figure 6. Arrhenius plot for the abiotic degradation of 1,1,1-trichloroethane.

32
Included In the Arrhenius plot are the degradation rate
coefficient for TCA in a pH 13 buffer and also the rate
coefficients calculated by Walraevens et al. (1974) for the
sodium hydroxide solutions. The rates for these high pH
solutions were within the confidence interval for the
regression line, indicating the reaction rate was not
significantly accelerated in alkaline media. The lack of
change in the rate in the presence of a high concentration
of a strong nucleophile (i.e. OH") suggested that the
reaction with the nucleophile occurs after the rate
determining step, characteristic of SN1 reactions.
Similarly, the increase in base strength did not shift the
elimination to an E2 mechanism through a large rate increase
and/or increase in formation of the elimination product.
The rate data which exceeded the confidence interval of
the regression line (Figure 6) were from studies (Vogel and
McCarty, 1987; Pearson and McConnell, 1975) which estimated
the rates of the slow reactions with less than 50%
degradation of the parent compound occurring. Rate
constants calculated for low conversion are more variable
than rates established based on higher amounts of conversion
(Levenspiel, 1972, p. 85). The strong linear Arrhenius
relationship between temperature and rate observed between
25 and 80C, regardless of sample matrix, suggests that
reaction rates at temperatures below 25C can be estimated
by extrapolation.

33
The elimination product, 1,1-DCE, was measured to
establish the factors which influenced the reaction pathway
(substitution versus elimination). Degradation of 1,1-DCE
was observed only at very high pH and even under those
conditions the rate was slow compared with the degradation
of TCA. Therefore, the ratio of the rate for elimination
(ke) to the total rate of degradation (k) was estimated by
plotting the concentration of 1,1-DCE versus (l-e'^t) where
t is time. The slope of the line equals ([TCA]c (ke/k)),
where [TCA]0 is the concentration of TCA at time zero.
This calculation required an estimate for the starting
concentration of TCA. For most experiments, multiple
analyses were performed for the estimate of the initial
concentration. Other authors (eg. Vogel and McCarty, 1987)
have used the intercept in the regression analysis of the
degradation, and this value was used as the estimate of
initial concentrations in this study.
Increases in pH and/or temperature theoretically favor
elimination over substitution. The elimination pathway
(Table 4) ranged from 17 to 38% of the total degradation
rate of TCA. Higher temperatures showed slightly more
transformation to 1,1-DCE over the temperature range
evaluated in these experiments. The percent of TCA
degradation due to elimination was not affected by matrix in
the pH range of 4.5 to 8.5. Seawater had no apparent effect
on the relative proportion of products. The highest percent

34
elimination pathway was measured in the strongest sodium
hydroxide (pH 13) solution.
Qualitative observations (GC and GC/MS) of TCA
degradation at approximately 60C in 0.5, 1.0 and 2.0 molar
sodium hydroxide solutions, showed the presence of 1,1-DCE,
and separate experiments indicated that 1,1-DCE also slowly
degraded under those conditions. These findings contradict
the results reported by Walraevens et al. (1974) in which
1.1-DCE was not detected in TCA degradation experiments at
high pH. This may be due to differences in analytical
methods, or the slow degradation of 1,1-DCE under their
reaction conditions.
Degradation of Brominated Ethanes
The degradation rates of brominated versus chlorinated
1.1.1-trihaloethanes were compared to provide insight into
the mechanisms and overall behavior of these compounds.
Since bromine is a better "leaving group" than chlorine,
brominated compounds typically degrade faster than their
chlorinated counterparts. In reviewing hydrolysis
degradation processes, Mabey and Mill (1978) concluded that
Br is more reactive than Cl by a factor of 5 to 10.
In a search of Chemical Abstracts, fewer than 20
references were reported for the brominated analog of TCA,
1.1.1-tribromoethane (TBA). Most of the papers addressed
spectra and bond energy studies, while no information on the
hydrolysis of this compound was reported.

35
Brominated analogs of TCA were not commercially
available. Therefore, TBA was synthesized according to the
methods reported by Stengle and Taylor (1970). The
procedure for the synthesis of 1,1,1-tribromoethane (TBA)
produced a mixture of brominated analogs of TCA. The
primary components were TBA and 1,1-dibromo -1-chloroethane
(DBCA), while smaller quantities of l-bromo-1,1-
dichloroethane (BDCA) were present. Kinetic data for
abiotic degradation of TBA and DBCA were measured for
several temperatures while data for BDCA were obtained in
only selected experiments conducted at higher overall
concentrations. Compound structures are illustrated in
Figure 7. The elimination pathway involved loss of HBr to
form the corresponding alkene, the dominant elimination
product was the ethene formed by loss of a bromine. The
substitution pathway forms acetic acid.
Initial degradation experiments involving the
synthesized brominated mixture were conducted in reagent
grade (Milli-Q) water to obtain preliminary data on the
transformation process. Subsequent experiments were
conducted in buffer solutions at pH 4, 7, and 10. The
results of these experiments are summarized in Table 6.
First-order kinetics of degradation were observed, as
were also seen for TCA. Rate constants were calculated
from the linear regression analysis of the plots of the

36
H Br
HC-CBr
H Br
Hx /Br
C = C
xBr
1,1,1 Tribromoethane (TBA)
1,1 Dibromoethene (DBE)
H Br
HC-CBr
H Cl
Hx /Br
C = C
XCI
1,1 Dibromo-1 chloroethane (DBCA) 1 Bromo-1 -chloroethene (BCE)
H Br
HC-CCI
H Cl
Hx /Cl
C = C
H7 XCI
1 -Bromo1,1 -dichloroethane (BDCA)
1,1-Dichloroethene (DCE)
Figure 7. Brominated analogs of 1,1,1 -trichloroethane and
corresponding elimination products. Since bromine is a
better leaving group than chlorine, the predominant pathway
is elimination of HBr.

37
Table 6. Summary of Brominated Compound Degradation
Rate Coefficients and Product Formation
Temp .
Matrix
Compound
108 k
s" 1
ke/k
%
20
DW
TBA
14
50.9
20
pH 4
TBA
10
60.7
20
pH 7
TBA
9
58.5
20
pH 10
TBA
11
61.8
28
DW
TBA
42
64.1
30
pH 4
TBA
65
61.8
30
pH 7
TBA
64
56.3
30
pH 10
TBA
71
63.1
37
1 M KC1
TBA
492
38.0
65
DW
TBA
7700
51.6
65
Na2S203
TBA
13000
60.1
20
DW
DBCA
17
35.6
20
pH 4
DBCA
14
33.9
20
pH 7
DBCA
11
32.5
20
pH 10
DBCA
15
33.4
28
DW
DBCA
51
45.6
30
pH 4
DBCA
81
39.8
30
pH 7
DBCA
69
31.9
30
pH 10
DBCA
73
38.5
37
1 M KC1
DBCA
492
24.4
65
DW
DBCA
7860
33.8
65
Na2 S 2*7 3
DBCA
14000
40.6
65
DW
BDCA
5350
29.6
DW, Distilled organic free water
Na2S2C>3, 1 M Sodium thiosulfate
Extrapolated Half-Lives for Degradation of TBA and DBCA
Temp .
108 k
e-1
T 1/2
Days
108 k
e-1
T 1/2
Days
15 5.54
20 12.83
25 28.92
145 6.90
62 15.56
28 34.14
116
52
23
TBA
DBCA

38
natural log of the concentrations versus time. All rate
constants were based on reactions showing a minimum of 75%
degradation.
The results of the degradation of TBA, DBCA and BDCA at
65C are illustrated in Figure 8. The differences in
slopes for the degradation of these compounds were not
statistically significant indicating that the rate
determining step was similar for each compound.
The formation of products (Figure 9) was calculated as
discussed previously for the formation of 1,1-DCE. The
percent elimination (ke/k) was the slope of the regression
line divided by the initial concentration of the parent
product. The smaller slope for 1,1-DCE, and its lower
maximum concentration, was a function of both lower initial
concentration of reactant (BDCA) and lower percent of BDCA
degradation which occurred through the elimination pathway.
The Arrhenius plot for TCA as determined in this study
is compared in Figure 10 with that of the brominated
compounds, TBA and DBCA. The Arrhenius plot for the two
brominated compounds was represented by a single regression
line. The regression line for TCA was essentially parallel
to that of the brominated compounds. The Arrhenius
activation energy (E^) for all of these compounds was almost
identical, since E^ is a function of the slope of this line.
The rate of degradation of TCA at 25C was
approximately a factor of 11 to 13 times slower than for the

39
n 1 1 i 1 1
0 2 4 6
Time (Hours)
Figure 8. First order kinetic data for the abiotic
degradation of TBA, DBCA and BDCA in water at 65C.
Figure 9. Formation of the elimination products (BCE, DBE,
DCE) in water at 65C from the abiotic degradation of the
corresponding 1,1,1-trihaloethanes.

1000/T (K)
Figure 10, Arrhenius plot for the abiotic degradation of 1,1,1-trihaloethanes.
p'
o

41
brominated analogs. The observed rate constants for the
various pH values were not significantly different, as was
also observed for the degradation of TCA.
Experiments were performed to determine the reaction
mechanism for these 1,1,1-trihaloethanes. First, the
degradation experiment was conducted in a 1 molar sodium
thiosulfate solution. Sodium thiosulfate is a much stronger
nucleophile than water or hydroxide (Swain and Scott, 1953)
and a dramatic increase in degradation rate in this solution
is indicative of an SN2 reaction in which the nucleophile is
directly involved in the rate determining step. The
degradation rates measured for the brominated 1,1,1-
trihaloethanes increased less than a factor of 2 in the
thiosulfate solution, which may be attributed to the
increased ionic strength of the solution. The percent
elimination was also unaffected by this sample matrix.
To further characterize the mechanism for these
degradation reactions, the brominated geminal trihalides
were placed in a 1 M KCl solution at 37C. High ionic
strength solutions generally increase the rate of SN1 or El
reactions. When a common ion is present, the rate of the
reverse reaction is enhanced. In the presence of high
concentrations of chloride, chloride may be exchanged for
bromide when an ion pair forms. If BDCA forms an ion pair
and chloride is exchanged, TCA will be formed (Figure 11)
providing evidence of a carbocation intermediate. Even

42
though TCA will degrade, it is more stable than the
brominated compounds and it may accumulate to detectable
levels.
The BDCA compound had the lowest concentration in the
mixture of the three geminal trihalides in reaction
solution, and TCA concentration was less than 40 ug/1.
After three days of incubation at 37C, the concentration of
TCA rose to approximately 200 ug/1. This was a minor
pathway (less than 5% of the BDCA degraded forming
detectable TCA) in the overall degradation process. 1,1,1-
Trichloroethane was not detected in other sample matrices
during the degradation experiments of the brominated
compounds, indicating that its presence in this solution was
a result of the reverse reaction of carbocation with the
chloride in the solution.
Increasing the extent of bromination increased the
percent of the degradation resulting in the elimination
product (Table 6). The proportion of the total degradation
which resulted in elimination for BDCA at 65C was within
the error estimate for the percent elimination of TCA at
elevated temperatures, and both of these parent compounds
produced 1,1-DCE. The highest percent elimination was
observed for TBA which formed approximately 60% 1,1-
dibromoethene (Figure 12). This may be due to an increase
in steric hindrance in carbocations containing bromine
rather than chlorine, slowing the substitution pathway.

H Cl
HC-CCI
H Cl
H Cl
HC-CBr
H Cl
BDCA
TCA
^C-C
h' XCI
DCE
H
HC-C
H
Acetic Acid
Figure 11. Reaction pathways for BDCA in 1 M KC1 solution. The exchange reaction of Cl
with the ion pair forms TCA, which degrades more slowly than the brominated compound.
P-
co

100
90 -
80 -
70 -
TBA DBCA BDCA
1 r
TCA
Figure 12. Comparison of the percent of the elimination pathway for 1,1,1-trihaloethanes.
p-

45
These experiments provided evidence that the abiotic
degradation of 1,1,1-trihaloethanes occurred by SN1/E1
rather than SN2 and E2 mechanisms. The trihaloethanes
containing one or more bromine atoms degraded at similar
rates, approximately a factor of 11-13 faster than TCA,
reflecting that bromine was a better leaving group. As the
number of bromines present on the trihaloethanes increased,
the percent of the degradation occurring through the
elimination pathway increased.
Degradation of Haloeenated Ethenes
One of the primary objectives of examining the behavior
of halogenated ethenes was to provide an accurate evaluation
of their formation and stability during degradation of the
corresponding ethanes. The literature provided some
evidence that slow degradation of these ethenes may occur at
a rate of interest for ground water studies.
Supporting the possiblity of degradation, Billing et
al ( 1975) reported half-lives for the abiotic degradation
of trichloroethene (TCE) of 10.7 months (0.002 day~^) and
for tetrachloroethene (PCE) of 9.9 months at 25C.
Molecular oxygen was present and the degradation rates were
suggested to result from oxidation as well as hydrolysis.
In this often referenced work, it was suggested that
mechanisms of degradation at lower temperatures may differ
from rates extrapolated from studies at higher temperatures.

46
In a study of hydrolytic decomposition by Pearson and
McConnell (1975), volatilization was extrapolated to zero
and a degradation half-life for TCE of 30 months was
e s timate d.
Roberts (1985) examined field evidence for the
degradation of various chlorinated organics and estimated
rate constants for both TCE and PCE of approximately 0.003
day-1, which may be due to a variety of factors including
sorption and dilution.
Wilson et al. (1985) studied the aerobic degradation of
TCE, PCE and other compounds in actual aquifer materials
from two sites in Oklahoma and Louisiana. No detectable
biodegradation of these compounds was observed under the
experimental conditions. Since degradation was noted in
autoclaved samples, the authors postulated that TCE and PCE
degradation was likely due to abiotic processes with rates
similar to those reported by Dilling et al. (1975).
The dehydrochlorination reaction of TCE occurs under
basic conditions and generates dichloroacetylene and
hydrogen chloride. This reaction of TCE with base is
spontaneous at room temperature and was responsible for
dichloroacetylene intoxication observed in patients inhaling
TCE-containing air in closed systems equipped with alkali
absorbers (Environmental Protection Agency, 1979).
Dichloroacetylene was detected in the gas phase above
aqueous alkaline solutions with pH 11 to 13 and upon

47
incubation with moderately alkaline material such as
concrete (Greim et al., 1984). They concluded
dehydrohalogenation can occur under these relatively mild
conditions resulting in toxicity from exposure to the
dichloroacetylene.
Many substitution and addition reactions of TCE have
been carried out in the presence of base. What initially
appeared to be a direct substitution reaction may in fact
have been multistep processes involving intermediates like
carbanions, chloroacetylenes, or carbenes. Rappaport (1969)
reviewed the mechanisms for nucleophilic vinylic
substitution processes in alkaline solutions at elevated
temperatures.
Mechanisms may differ for chemical studies performed
under extreme conditions of temperature and high pH compared
to reactions occurring under more typical environmental
conditions. The possibility of slow nucleophilic attack in
aqueous solution was considered because March (1985) reports
that although vinyl halides are generally considered
resistant to nucleophilic attack, the presence of electron-
withdrawing groups like halogen lower the electron density
of the double bond enhancing nucleophilic substitution or
addition reactions.
In ground water, even very slow degradation may be an
important attenuation mechanism. Since environmental
studies report slow degradation of TCE or PCE in water and

48
the chemical studies show presence of electron withdrawing
groups like chlorine increases the susceptibility of an
olefin to nucleophilic attack, experiments to evaluate
possible reactions were performed.
References to possible hydrolysis reactions of 1,1-DCE
or its brominated analogs were not found upon review of the
literature. The 1,1-dihaloethenes would be less susceptible
to nucleophilic attack than TCE since fewer electron
withdrawing groups are present. The pure compounds however,
are very reactive and polymerize readily. Their
reactivities in dilute aqueous solution have not been
examined.
The focus of my research with halogenated ethenes was
to examine the stability of these compounds in relatively
dilute aqueous solutions and to determine their
susceptibility to nucleophilic attack. Autooxidation or
other reactions of the pure liquid compounds which may be
present in the vadose zone following a spill could occur,
but these reactions are not addressed here.
There were two major purposes for the examination of
the degradation behavior of halogenated ethenes. First, the
stability of the ethene products formed during the
transformation of the geminal trihalides needed to be
determined to accurately describe the kinetics of the
appearance of these elimination products. Secondly,
previous studies which indicated that halogenated ethenes

49
like trichloroethene (TCE) and tetrachloroethene (PCE) may
undergo slow abiotic degradation in water at room
temperature with a half-life of less than one year were
reevaluated. The question of possible nucleophilic attack
by water, hydroxide ion, or other nucleophiles must be
addressed to understand the stability of these commonly
detected ground water contaminants.
The stability of 1,1-DCE was evaluated in experiments
that were performed concurrently with the evaluation of TCA
degradation. In the buffer solutions, seawater, and
distilled deionized water, no significant degradation of
1,1-DCE occurred during the course of the evaluation of the
degradation of TCA.
The formation of ethenes containing bromine was
monitored during the degradation studies of the brominated
ethanes, and their concentrations were continually monitored
for some time after the ethane degradation was completed.
Trichloroethene was studied in separate experiments
performed at various temperatures selected to repeat the
experiments conducted by Dilling et al. (1975). In addition
to buffer solutions, one set of ampules was prepared with a
nutrient solution which was not autoclaved, and to which
ground water known to show biological activity was added.
This was done to determine if any degradation which might
have occurred during the long term studies could have been
due to biological activity.

50
A summary of the results of these experiments is
presented in Table 7. No significant degradation of these
compounds was found in the experimental matrices during the
indicated reaction times, as evidenced by the slopes of In C
vs time which were not significantly different from zero.
The overall coefficient of variation for the observations is
similar to values obtained for simple replicate analyses.
Experiments were also conducted to evaluate the overall
behavior of these compounds under more rigorous conditions.
The literature indicated that halogenated ethenes such as
TCE can undergo elimination to form chloroacetylenes at
elevated pH (Rappaport, 1969). This reaction was verified
by using GC/MS to confirm the formation of dichloroacetylene
from TCE and also chloroacetylene from 1,1-DCE by analysis
of the headspace vapor above an alkaline (1 M NaOH) aqueous
solution of the halogenated ethene which was warmed to
approximately 60C.
The rate of degradation of components in a mixture of
1,1-DCE, TCE and PCE in sodium hydroxide solutions was
examined at 60C (Table 8). These were the matrices used by
Walraevens et al. (1974) in their examination of the
degradation of TCA, wherein they did not observe formation
of 1,1-DCE. One objective was to establish if the
elimination product was stable under their reaction
conditions.

51
Table 7. Summary of Experimental Conditions for which
Halogenated Ethenes were Stable.
Temp .
C
Cmpd.
Time
Days
No .
Obs .
Average
Concentration
C V.
Matrix
27
BCE
160
12
mg/L
4.2
7%
pH 10
27
BCE
160
10
4.2
9%
pH 7
27
BCE
160
10
4.2
6%
pH 4
27
BCE
54
12
88
4%
DW
37
BCE
126
18
80
4%
DW
65
BCE
6
13
392
11%
Thio
27
DBE
160
12
32
4%
pH 10
27
DBE
160
10
32
7%
pH 4
27
DBE
160
10
28
10%
pH 7
27
DBE
54
12
120
2%
DW
37
DBE
126
18
110
11%
DW
65
DBE
6
13
690
12%
Thio
27
DCE
274
15
2.7
7%
Nutrient
27
DCE
386
26
1 1
13%
pH 4
27
DCE
386
27
1.1
14%
pH 7
27
DCE
386
25
1 1
12%
pH 8.5
37
DCE
386
16
2.4
9%
pH 4
37
DCE
386
17
2.4
8%
pH 7
37
DCE
386
26
2.4
10%
pH 8.5
55
DCE
15
8
2.1
16%
DW
55
DCE
15
15
2 2
6%
pH 7
55
DCE
15
12
2 3
9%
Seawater
80
DCE
14
8
2.3
12%
pH 4
80
DCE
14
8
2.3
9%
pH 7
80
DCE
14
8
2 3
14%
pH 8.5
65
DCE
6
13
.4
12%
Thio
27
TCE
274
14
3.0
4%
Nutrient
27
TCE
386
10
1 2
9%
pH 4
27
TCE
386
11
1.2
10%
pH 7
27
TCE
386
10
1.2
8%
pH 8.5
37
TCE
386
16
1.7
12%
pH 4
37
TCE
386
16
1.7
9%
pH 7
37
TCE
386
16
1 7
16%
pH 8 5
80
TCE
14
9
1.7
8%
pH 4
80
TCE
14
8
1.7
9%
pH 7
80
TCE
14
10
1.7
9%
pH 8.5
DW Distilled organic free water
Thio 1 M Sodium thiosulfate solution

52
Table 8. Second Order Degradation Rates (1 mole'^ hr"^)
of Halogenated Ethenes at 60C
in Sodium Hydroxide Solutions
NaOH
ncentration
TCE
1,1-DCE
PCE
0.1 M
0.6
0.02
nd
0.5 M
0.28
0.01
nd
1.0 H
0.17
0.01
nd
2.0 M
0.12
0.004
nd
nd no significant degradation occurred after 260 hours.
The rate of degradation of TCE was the greatest among
the tested compounds due to the presence of an acidic
hydrogen (a hydrogen present on a carbon containing a
halogen). The elimination reaction was also an available
pathway for the degradation of 1,1-DCE, although the rate of
degradation was approximately 30 times slower than for TCE
in all solutions except for the 1.0 M NaOH.
Tetrach1oroethene (PCE) did not degrade since the
dehydrohalogenation reaction could not occur, and apparently
conditions were not favorable for an addition process.
For environmental applications, there are concerns with
the mildest conditions (temperature and pH) which may still
result in degradation of these compounds. The pathway for
the degradation of ethenes at elevated temperature could
differ from reactions at lower temperatures where the
elimination reaction would be less favorable and a possible
addition reaction could occur instead.
Therefore, TCE was

53
incubated at 20C in a solution at pH 12.5. No degradation
was observed during four months of incubation (Table 7).
The experiments demonstrate the resistance of the
halogenated ethenes to degradation in dilute aqueous
solution. Reports of the degradation of these compounds
with half-lives of less than 1 year appear to represent a
process other than abiotic degradation in water. In the
same way as billing et al. (1975), my experiments were
conducted in sealed ampules containing a headspace, however
degradation was not observed as reported in their study. I
believe their results may be a result of analytical error.
The half-lives for each experiment were based on four
measurements. The results showed a chemically diverse group
of compounds had similar decreases in concentration and
temperature had little effect on these decreases. A
possible explaination for these results would be a decrease
in instrument response over the year of the study.
The halogenated ethenes generally showed very little
degradation, with the exception of the rapid degradation of
TCE at high pH and temperature. It appears that any
degradation of these compounds in aqueous solution which
occurs, does so under rather extreme conditions and is not
expected to be a dominant process.
Struc ture/Rate Relationships of Alkyl Halides
In the previous sections degradation patterns and
kinetics were evaluated for various 1,1,1-trihaloethanes in

54
aqueous solution. A broader perspective on hydrolysis /
elimination reactions can be obtained by comparisons with
other haloalkanes reported in the literature. The
obj ectives are
1. To compare degradation rates measured for
trihaloethanes of other simple alkyl halides which react by
an SN1/E1 mechanism.
2. To compare degradation rates of trihaloethanes with
other geminal trihalides reported in the literature to
determine structure/activity relationships with changes in
the substituents on the beta carbon, and describe shifts in
mechanisms which may occur for these trihalides.
3. To compare degradation rates and pathways of 1-
chloropropane and 1,1-dichloroethane with TCA to show the
effects of increasing number of chlorines on the alpha
carbon.
The classic reaction mechanisms for substitution and
elimination reactions are SN1, SN2, El and E2, as previously
discussed. The presence of various functional groups can
effect the rate and pathway of degradation of an alkyl
halide. For example, rates of hydrolysis are greater for
alkyl halides containing Br rather than for Cl by a factor
of 5 to 10. The rates also increase as the alkyl group goes
from primary to secondary to tertiary in the ratio of
1:10:1000 for chloride. Allyl groups enhance the rate of
hydrolysis of a primary halide by a factor of 5 to 100,

55
while benzyl groups enhance the rate by a factor of 50.
(Mabey and Mill, 1978)
The formation of stabilized carbocations by electron
donation from the non-bonded electron pairs of halogens
adjacent to the cationic carbon center have been reported
(Olah, 1974). The stabilizing effect was enhanced when two
or even three electron-donating heteroatoms coordinate with
the eleetron-deficient carbon atom as illustrated in Figure
13. Specific examples, designated as "chlorocarbenium
ions" by Olah (1974), have been identified and are
illustrated in Figure 14.
Simple SN1/E1 Reactions
My data suggested 1,1,1-trihaloethanes form carbocation
intermediates. The intermediate would contain two halogens
and one methyl group. The observed rates and pathways are
compared (Table 9) to compounds containing two methyl groups
and one halogen (2,2-dihalopropanes) and three methyl groups
(t-butyl chloride).
Degradation of tertiary halides like t-butyl chloride
occurs with a carbocation intermediate and these compounds
are resistant to bimolecular nucleophilic displacement. The
half-life for the aqueous degradation of t-butyl chloride is
approximately 23 seconds at 25C with about 19% of the
degradation occurring through the elimination pathway. The
carbocation intermediate is stabilized by the three methyl

o+
X
+
X = C X *+ x c
I 1
R R
Figure 13.
Stabilization of carbocations by halogen (Olah, 1974).
+x

Figure 14. Examples of Mchlorocarbenium ions" (Olah, 1974)
Ul

58
Table 9. Summary of Degradation Rate Coefficients and
Pathways for Tertiary
and Secondary
Halides
Compound
k (sec"^)
ke/^t
Reference
2 5C
%
t-Butyl Bromide
5
1
t-Butyl Chloride
2.98xl02
2
2-Bromo-2-chloropropane
1.78xl0'4
100
2
2,2 Dibromopropane
4.6 2x105
100
2
2,2-Dichloropropane
9.09xl0'6
100
2
1,1,1-Tribromoe thane
2.89x10 7
58
3
1,1-Dibromo-l-chloroethane
3.41xl0'7
37
3
1,1 1-Trichloroe thane
2.62xl08
23
3
2-Bromopropane
3.82xl06
0
2
2-Chloropropane
2.11x10'7
0
2
References:
1. March, 1985 .
2. Queen and Robertson, 1966.
3. This dissertation.

59
groups. The rate coefficient at 25C is approximately 10^
faster than for TCA.
Queen and Robertson (1966) examined the hydrolysis of
2,2-dihalopropanes. These compounds form carbocation
intermediates with two methyl and one halogen group. The
rate coefficients for the degradation of 2,2-dihalopropanes
are intermediate between t-butyl chloride and the 1,1,1-
trihaloethanes. The mechanism was reported to be SN1/E1
based on results of experiments with deuterated gem-
dihalides. The degradation rates of these compounds were
10-50 times higher than of the corresponding secondary
halides (e.g., 2-chloropropane).
The degradation rates were affected by the leaving
group, bromine or chlorine. Also, the structure and
stability of the resulting carbocation affected the rate and
pathway (elimination and/or substitution) of the reaction.
Since bromine was a better leaving group than chlorine,
there was a rate increase when bromine was present as
compared to the corresponding chlorinated compound. 2,2-
Dibromopropane degraded 19 times faster than 2,2-
dichloropropane, while 2 bromo 2-chloropropane degraded 5
times faster than the dichloro compound (Queen and
Robertson, 1966). The 1,1,1-trihaloethanes containing
bromine degraded 11-13 times faster than TCA.
Rates were also increased as the number of methyl
groups present on the carbocation increased. The t-butyl

60
chloride degraded approximately 3000 times faster than 2,2-
dichloropropane and 10^ faster than TCA.
There were two major differences between my results and
those reported by Queen and Robertson (1966). First, they
reported a rate nearly four times higher for 2-bromo-2-
chloropropane than for 2,2-dibromopropane, while the rate
coefficients I measured for the trihaloethanes containing at
least one bromine were approximately equal (within 20%).
Secondly, they report only formation of the elimination
product for all 2,2-dihaloethanes, while the percent
elimination in my experiments was a function of the number
of bromines and was always less than 60%. The percent
elimination for t-butyl chloride was less than the value
obtained for the trihaloethanes.
The effect of alpha halogen is complex, "combining a
negative inductive effect and an electron-releasing
resonance effect" (Queen and Robertson, 1966, p. 1364).
Based on my results and the results for t-butyl chloride,
elimination was not expected as the primary pathway nor the
large difference in rates observed for the two
dihalopropanes which contained a bromine. The rate data
were determined for the dihalopropanes by a conductance
method. Extraction of the products of solvolysis of 2,2-
dibromoethane with CCI4 and analysis by vapor phase
chromatography (GC) and nmr showed 2-bromopropene was the
only product in other than trace amounts. It may be that

61
the substitution product, acetone, would not have
partitioned and been measured using that analytical
protocol.
Comparisons of Geminal Trihalides
A number of compounds in the literature contain a
geminal trihalide group (R-CX3), and many of these compounds
have environmental implications. My experiments on 1,1,1-
trihaloethanes indicated that the -CX3 group was sterically
hindered and resistant to attack by an SN2 mechanism, and
that the halogens could help to stabilize the formation of a
carbocation. The overall rate of degradation of other
geminal trihalides will increase if R also stabilizes the
carbocation. If the beta carbon contains an acidic hydrogen
the mechanism may shift to E2 at elevated pH.
A summary of degradation rates (expressed as reaction
half-lives) of various geminal trihalides is presented in
Table 10. The simplest compounds, trihalomethanes, were
very resistant to hydrolysis. The R- consists only of
hydrogen, which was inadequate to stabilize a carbocation.
The mechanism for this degradation has been determined to be
a base catalyzed process with a carbanion intermediate
(Hine, 1950). The extremely low reactivity also suggests
that steric hindrance may prevent SN2 attack.
By contrast alpha,alpha,alpha-trichlorotoluene has a
half-life of 19 seconds at 25C, which corresponds to a rate
of a factor of 10^ greater than for TCA. Therefore, the

62
Table 10. Half-lives for Abiotic Degradation of
Geminal Trihalides
COMPOUND
STRUCTURE HALF-LIFE REFERENCE
25C
Chloroform
Bromo form
CHCI3 3500 yr Mabey and Mill,
1978
CHBr3 690 yr Mabey and Mill,
1978
1,1,1-Trichloroethane CH3CCI3 10.2 mo This dissertation
1,1,1-Tribromoethane CH3CBr3
1 mo This dissertation
DDT
12 yr Wolfe et al., 1977
(PH5, 27C)
Me thoxychlo r
CH CCI-
1 yr Wolfe et al., 1977
(PH5, 2 7 C)
a a a T r ichlo r o to luene // \
CCL
19 s Lyman et al., 1982

63
rate increase was much greater than the factor of 50
reported by Mabey and Hill (1978).
Quemeneur et al. (1971) determined that tri-chloro
compounds of the type P-RC6H4-CCI3 (R is OMe, Me, H, Cl, or
NO2) were hydrolyzed in neutral or acidic medium via a
cationic transition state for all types of R substituents.
The hydrolysis of the p-substituted alpha,alpha-
dichlorotoluenes reacted via a cationic mechanism when R is
an electron-donor, and a bimolecular mechanism when R is an
electron-attracting group. These results also supported the
observation that halogens contributed to the stability of
the carbocation. Monochlorotoluene reacts nearly 3000 times
more slowly by an SN2 mechanism than the trichlorotoluene
reacts by the SN1.
Methoxychlor and DDT are two environmentally important
pesticides which contain a geminal trihalide functional
group. Wolfe et al. (1977) provided an in depth examination
of the degradation of these compounds. There is a beta
hydrogen on each of these compounds. At elevated pH the
degradation rate increased as a function of pH and the
elimination products were dominant, suggesting these
structures were more susceptible to degradation by the E2
mechanism than is TCA. While the elimination product, DDE,
was the major product of DDT hydrolysis even at lower pH,
the major product of methoxychlor at pH 7 was the hydrolysis
product, with minor amounts of the elimination product,

64
DMDE. The hydrolysis products formed were anisoin and
anisil, which were explained by phenyl group rearrangement
after the formation of the carbocation.
Mochida et al. (1967) showed 1,1,1,2-tetrachloroethane
and pentachloroethanes reacted more slowly than TCA under
lower pH conditions, which indicated that chlorines on the
beta carbon decrease the stability of the carbocation. The
presence of these chlorines on the beta carbon however,
increased the acidity of the hydrogens, with enhanced
degradation rates for the tetra and pentachloroethanes by an
E2 mechanism at elevated pH.
There is considerable evidence that geminal trihalides
can form carbocations in the presence of an appropriate
neighboring group. Subsequent reaction pathways may vary
according to the structure of the carbocation resulting in
elimination, substitution, or rearrangements. An E2
reaction may also occur for compounds containing an acidic
hydrogen on the beta carbon.
Effect of Additional Halogens on the Alpha Carbon
The hydrolysis of a simple primary halide, 1-
chloropropane, was compared with the reactivity of 1,1-
dichloroethane and TCA in experiments I performed at
elevated temperature. As the number of hydrogens on the
alpha carbon decrease, steric hindrance can increase and
result in a shift in reaction mechanism. The experiments
were designed to demonstrate the relative rates of

65
hydrolysis in aqueous solution, and the response to an
increase in concentration of a strong nucleophile whose
effect would be a function of the mechanism.
Based on the literature, simple primary alkyl halides
like 1-chloropropane are expected to degrade by an SN2
mechanism. Therefore, 1-chloropropane should show an
increase in degradation rate in the presence of a strong
nucleophile, since the nucleophile is involved in the rate
determining step.
Predicting the degradation rate of 1,1-dichloroethane
is more difficult. Secondary chlorides, like isopropyl
chloride, have been shown to degrade more quickly than the
primary alkyl halides, possibly by an intermediate
mechanism. Chloride can contribute somewhat to the
stability of a carbocation, however, it is not as effective
as a methyl group as discussed previously. In addition, the
presence of a halogen can increase the steric hindrance at
the alpha carbon.
Comparisons of the degradation rates of these compounds
were made at elevated temperature (65C) in pH 7 buffer
solution, and in a 1 M thiosulfate solution. In the buffer
solution the degradation of TCA was approximately 6 times
faster than the hydrolysis of 1-chloropropane. Degradation
of 1,1-dichloroethane was less than 6% of the rate of 1-
chloropropane degradation. This rate comparison is
illustrated in Figure 15.

4
o
o
\
o
c.
I
3 -
2 -
1 -
O
+ 1-Chloropropane
600
Time (hours)
Figure 15. Pseudo-first-order kinetic data plots for hydrolytic degradation of TCA,
1-chloropropane, and 1,1-dichloroethane in pH 7 buffer solution at 65C.

67
The degradation of 1-chloropropane was enhanced by more
than a factor of 100 in the thiosulfate solution, 1,1-
dichloroethane degraded approximately 22 times faster, and
TCA degradation rate increased less than a factor of 2. The
differences in rate enhancement among these compounds is
attributed to differences in mechanism. Part of the
increase in rate of degradation of TCA in thiosulfate is
attributed to the increasing ionic strength, and TCA
degradation rate was clearly less affected by the presence
of thiosulfate than the other compounds. The rate
enhancement for 1,1-dichloroethane was similar to the type
of rate increase which would be observed for secondary
halides which react by an intermediate mechanism.
The thiosulfate solution was used as a matter of
convenience as a strong nucleophile to assist in
demonstrating how knowledge of mechanism may be necessary in
estimating degradation rates as matrices change. Greatest
changes in rates in the presence of sulfur nucleophiles may
be expected for simple primary alkyl halides, and the least
effect occur with compounds which react via an SNl or El
mechanism.
Sediment Matrix Effects
There is considerable interest in possible effects of
solid surfaces on rates of hydrolysis. Most hydrolysis
experiments are performed in simple buffered aqueous
solution. Contaminants in the vadose zone or ground water

68
have considerable contact with a variety of aquifer
materials which could potentially affect degradation rate.
Hydrolysis reactions may be affected by factors like acid or
base catalysis, sorption and ionic strength. Since
compounds which react by different mechanisms may be
impacted differently by these solid surfaces, both 1-
chloropropane and TCA were used in degradation experiments
performed in various matrices.
Catalysis of hydrolysis or elimination reactions of
alkyl halides by saturated aquifer materials has not been
demonstrated. Because high concentrations of 1,1-DCE have
been observed in Florida and Arizona at solvent spill sites
contaminated with TCA, the role of sand or other materials
which may influence the degradation of TCA was evaluated.
The nonbiological degradation of pesticides in the
unsaturated zone was shown to play an important role for a
few groups of pesticides, mainly organophosphates and s-
triazines. Clay mineral surfaces have shown catalytic
activity, correlated to their acid strength. This catalytic
process is most important at low moisture content, and
therefore is more important in the vadose zone than beneath
the water table (Saltzman and Mingelgrin, 1984).
Haag and Mill (1988) did not observe significant
differences in the kinetics or products of TCA in contact
with sediment pore water. Epoxide hydrolysis was

69
accelerated by a factor of four in sediment as compared to
rates in buffered water.
Mabey and Mill (1978) indicated that acid promotion of
the aqueous hydrolysis of halogenated aliphatic hydrocarbons
has not been observed. March (1985) stated that gem-
dihalides can be hydrolyzed in water with either acid or
basic catalysis to give aldehydes or ketones, although the
strength of acid was not addressed.
In a review of elimination reactions in the presence of
polar catalysts, Noller and Kladnig (1976) stated that
"interaction of X with an acid is probably as indispensable
as the reaction of H with base in liquid-phase elimination
reactions, but this function is probably taken over by the
solvent and is less pronounced than the base promoted
process."
Clarification of interactions with polar surfaces may
provide insight into possible effects of sediments or soil
on reaction rates. Clays, for example, contain polar
surfaces which have been shown to catalyze degradation of
some pesticides (Saltzman and Mingelgrin, 1984).
Noller and Kladnig (1976) illustrated elimination
reaction products were a function of the specific catalyst
with 1,1,2-trichloroethane
Cl H
Cl Ci C2 Cl
H H

70
as reactant. Basic catalysts (e.g., KOH-SO2) attack the
most acidic H, that at C^ forming more 1,1- than 1,2-
dichloroethene. Acidic catalysts (e.g., si1ica-alumina)
attack Cl on because the formation of the carbocation is
facilitated by the other Cl on that carbon resulting in the
formation of much more of the 1,2-dichloroethene isomer.
The choice of catalyst will determine the predominant
product giving selectivity to the reaction.
Mochida et al. (1967) reported that the reactivity of
TCA on solid acids was greater than that for other
chlorinated ethanes (mono-, di-, tri- and tetra- chloro
compounds). On solid bases it was less reactive than
penta-, tetra-, and 1,1,2-tri- chloroethanes. The shift in
reactivity of the ethanes with change in catalyst showed
enhanced ability of TCA to form a carbocation by
accelerating the reaction on an acid surface as compared to
the other chlorinated ethanes. There was also the lack of
an acidic beta hydrogen to permit catalysis by base.
Possible catalysis would be compound- and mechanism-
specific. Degradation experiments were performed on 1-
chloropropane (SN2) and TCA (SN1,E1) at 65C in 5 ml
distilled deionized water, with a final concentration of
approximately 2 mg/1. Separate ampules were prepared with
the addition of 0.4 g bentonite clay, 1 g limestone, 1 g
sand, and 0.2 g silica gel.

71
Similar trends were observed for both compounds (Table
11). The slowest rates relative to water were obtained for
both compounds in the sample containing clay, while the
fastest rates were observed in the sand.
The data generally showed greater variability in the
samples containing the solids as compared to the DW system
(Figures 16 and 17) as evidenced by correlation coefficients
less than 0.99. However, the rates of 1-chloropropane
degradation in ampules containing solids differed by less
than 10% of the rate obtained for Milli-Q water.
The relative degradation rates for TCA differed more as
a function of matrix than observed for chloropropane,
however, there was also greater variability as evidenced by
the correlation coefficients. In the case of TCA, the
formation of 1,1-DCE was similar in all matrices suggesting
the ratio of products was not affected by the presence of
the s e solids.
The relatively small differences in rates measured in
these matrices may be due to a variety of factors including
sorption, however significant surface catalysis was not
observed. For this type of saturated system, the amount of
alkyl halide in contact with the surface would be small.
Differences may be attributed to normal variability and
differences in ionic strength or composition of the aqueous
phase in contact with the solids.

72
Table 11. Matrix Effects for Degradation Rates of
1-Chloropropane and 1,1,1-Trichloroethane at 70C.
Linear Regression Output for the Plot of In C (ug/1)
vs. Time (hours).
Chloropropane
Regression Output: MQ
Constant (C) 7.20
Std Err of Y Est 0.12
R Squared 0.99
No. of Observation 12
Degrees of Freedom 10
X Coeff. (Rate) -0.0102
Std Err of Coef. 0.0004
Relative rate 1.00
(MQ 1)
Clay
Limes tone
S and
Silica
7 59
7.44
7.14
7.11
0.26
0.07
0.13
0.23
0.94
0.99
0.99
0.96
9
7
11
12
7
5
9
10
-0.0094
-0.0098
-0.0113
-0.0107
0.0009
0.0004
0.0004
0.0007
0.92
0.96
1.11
1.06
1.1.1-Trichloroethane
Regression Output:
MQ
Clay Limestone
Silica Gel
S and
Constant (C^)
5 .86
6.26
5.23
5.91
6.01
Std Err of Y Est
0.11
0.12
0.07
0.18
0.28
R Squared
0.99
0.99
1.00
0.98
0.96
No. of Observations
11
7
6
7
9
Degrees of Freedom
9
5
4
5
7
X Coeff. (Rate)
0.027
-0.020
-0.038
-0.034
-0.040
Std Err of Coef.
0.0008
0.0009
0.0010
0.0020
0.0030
Relative rate
1.00
0.74
1.38
1 25
1.45
(MQ = 1)

5
73
<
O
I
O
o
\
o
c

Clay
&
+
Silica Gel
4
4 -
0
Milli Q Water

3-
A
Sand
+
6
l 1
2-
s

A

1 -
o
* .* .
o
0 <
1

, 1 1
i 1
~1 1 1 1 1 1
O 20 40 60 80 100 120 140
Time (hours)
Figure 16. Effect of the presence of solid material on the
rate of hydrolytic degradation of TCA at 65C.
6
5
4
3
2
1
0
0 100 200 300 400
Time (hours)
Clay
+ Silica Gel
Limestone
A MilliQ Water
X

* Sand

0
Â¥
t h 1
J.:
Â¥
A

i 1 1 1 1 1 r
Figure 17. Effect of the presence of solid material on the
rate of hydrolytic degradation of 1-chloropropane at 65C.

74
These experiments do suggest that TCA in sand aquifers
may show a slightly increased rate as compared to low ionic
strength buffered water experiments. The rate coefficient,
however, will fall within the error limits for the rate
estimate for the degradation of TCA based on the experiments
in buffered distilled water.

SOLUBILIZATION AND DEGRADATION OF RESIDUAL TCA
A computational model was constructed to describe the
attentuation of TCA beneath the water table in the presence
of multiple phases. This simplified scenario for a TCA
spill considered the chemical transformation of TCA to 1,1-
DCE along with advective transport resulting from ground
water flow, of TCA and 1,1-DCE out of this zone containing
the residual solvent. Biodegradation of TCA in this highly
contaminated zone was considered negligible.
The major objective in developing this model was to
describe the relative concentrations of the major
constituents and how their concentrations may change with
time. These trends are illustrated for various ground water
flow rates, change in initial concentrations, and initial
composition.
Behavior of Residual Solvent
The migration pattern of chlorinated hydrocarbons
following a spill is illustrated in Figure 18. These dense
nonaqueous phase liquids (NAPL) will infiltrate the porous
media, with some of the NAPL retained in residual
concentration. The retention capacity for these NAPL in the
unsaturated zone may range from 5 L m"^ (approximately 12
mL/L of pore space) in highly permeable media to 30-50 L mJ
75

76
Vadose Zone
T
Ground y
Water
FlOW
tFFFFFFFFXtF TCA
1,1-DCE
Water
;; Saturated
with Solvent
Chlorinated Solvent Pool
at impermeable layer
Figure 18. Equilibrium model for the attenuation of
residual TCA present beneath the water table.

77
in media of lower permeability (Schwille, 1984). Additional
factors which influence whether the NAPL will reach the
water table include the spilled volume and infiltration
process.
If sufficient volume of dense NAPL reach the water
table, it will sink into the saturated zone and continue to
migrate downward as long as the retention capacity of the
zone is exceeded. Wilson and Conrad (1985) reported
residual hydrocarbon occupying 15-40% of the pore space in
the saturated zone.
Water continues to flow laterally through the water
saturated zone containing residual NAPL. The globules of
NAPL provide a large interface with the water providing a
solution zone, where the initial concentration of a given
component is proportional to its aqueous solubility as
determined by the NAPL composition. These globules are
generally trapped in the larger pore spaces and are being
prevented from entering the smaller pores due to the high
capillary entrance pressure. There is a reduction in
permeability to water where the residual NAPL is present, as
the largest channels become blocked at several places by
discontinuous solvent ganglia. This forces water to flow
around the solvent in fairly thin films and/or be diverted
into the smaller channels whose carrying capacity
(conductivity) is low (Jones, 1985).

78
In laboratory experiments, the initial concentration of
chlorinated solvent was at saturation concentration even
when the layer of sand with residual solvent was thin
(Schwille, 1988). The concentration gradually decreased
until the levels in the water were sufficiently low that
further removal of solvent was slow. At this point
approximately 86% of the residual had been removed.
My model was developed assuming that equilibrium
saturation was maintained, the dissolution of residual
solvent being faster than the degradation or advective
transport of components. Diffusion or hydrodynamic
dispersion was not considered to be a limiting factor in
maintaining equilibrium. The solvent contaminated zone was
then treated similar to a well-mixed flow reactor.
Interactions of the solutes in the water with the solid
matrix of the saturated zone were considered minimal
providing residual solvent was present; the porous medium
was assumed to provide a matrix in which the residual
solvent was retained.
Once the flow of the NAPL stopped, the subsequent
losses were assumed to occur through degradation or
advection of the compound in the aqueous phase.
Hydrolysis/elimination of TCA occurs much faster in dilute
aqueous solution than would occur for water dissolved in the
TCA solvent phase (Walraevens et al., 1974). Ground water
continues to flow through this zone, although at somewhat

79
reduced velocities, carrying dissolved components out of
this zone.
Aqueous Phase Concentrations
The quantity of solvent lost each day by advection or
degradation is a function of the concentration of each
component (TCA and 1,1-DCE) in the aqueous phase, which in
turn depends on the composition of the residual NAPL. The
solvent phase may contain TCA and/or 1,1-DCE, or another
solvent which may have been spilled with the TCA.
The distribution of a component between the two liquid
phases can be expressed in terms of fugacity. For ideal
mixtures, the solubility of the solute at any composition is
estimated by multiplying the unit solubility by the mole
fraction of the component in the solvent phase at
equilibrium. Nonideal mixtures form deviations from
linearity. Estimates of aqueous concentrations resulting
from a nonideal solvent mixture requires knowledge of the
activity coefficients at the various mole fraction
compositions. For the simpler ideal case,
[TCA]w = x STCA
[DCE]w = (1-x) SDCE
x = TCAs / (TCAs + DCEs)
where TCAs and DCEs are the number of moles of that compound
in the solvent phase at equilibrium, x is the mole fraction
of TCA in the solvent phase, S>pcA and are the pure
component solubilities, and [TCA]w and [DCE]w are the

aqueous phase concentrations at equilibrium. The total
number of moles of TCA in a unit volume of porous media is
the sum of the moles present in the aqueous and solvent
80
phase s.
The model describes changes for TCA spilled on a high
permeability material like sand. As TCA degrades and forms
1,1-DCE, the degradation product partitions into the NAPL
affecting the aqueous phase concentration of TCA (and DCE).
Both the individual solubilities and the solubility of
a mixture of TCA and 1,1-DCE are required in the model and
it was also necessary to assess if mixtures of TCA and 1,1-
DCE deviate significantly from ideality. Literature values
for the solubilities of these constituents vary widely
(Table 12). The solubility data for 1,1-DCE reported by
Lyman (1981), showed as much as a 700% error from a
predicted concentration based on regression relationships.
That estimated concentration is much closer to the
concentrations reported by Verschueren (1977).
Table 12. Solubilities of TCA and 1,1-DCE (mg/L)
TeniD (C)
TCA
1 1
20
480
400
20
4400
2640
30
1088
3675
25
273
4
1700
4200
24
1580
3200
Source
Pearson and McConnell (1975)
Verschueren (1977)
Verschueren (1977)
Lyman (1982)
This study.
This study.
Measurements (Figure 19) were made on the solubility of
the individual components (TCA and 1,1-DCE) and on the

Aqueous Solubility (mg/l)
Equilibrium Mole Fraction of TCA
(Solvent Phase)
Figure 19. Aqueous solubilities of a binary mixture of TCA and 1,1-DCE as a function of
mole fraction composition in the solvent phase (24 C).

82
solubility of each with varying compositions of the binary
mixture. Mixtures were at room temperature, approximately
24C .
The pure component solubility of TCA (1580 mg/L or 11.8
mmoles/L) and the solubility of 1,1-DCE in the aqueous phase
(3200 mg/L or 33 mmoles/L) measured at 24C were within the
concentration range listed by Verscheuren (1977) who
reported solubilities at 20 and 30C. This is significantly
higher than solubilities reported by Lyman (1981) and
Pearson and McConnell (1975). The solubility for 1,1-DCE
reported in this dissertation was verified independently by
solubility measurements performed using high performance
liquid chromatography (HPLC) (Linda Lee, University of
Florida, Personal communication, 1988). She measured an
average for the solubility of 1,1-DCE at 24C as 2990 mg/L.
Her report is included in Appendix B.
Verscheuren (1977) reported that the solubility of TCA
at 20C was four times greater than at 30C, a value
approximately three times greater than our result at 24C.
Since the mass lost per unit time from degradation is a
function of aqueous concentration and the first-order
degradation rate coefficient, higher aqueous concentrations
at lower temperatures could compensate for the lower
degradation rate. The solubility of TCA at 4C was measured
to verify this trend. As shown in Table 12, a significant

83
increase in solubility of TCA at lower temperatures was not
observed.
The linearity of the change in solubility with
increasing mole fraction for these two compounds suggested
that 1,1-DCE and TCA form a near-ideal solution in the
solvent phase. Based on these measured data, I assumed that
mole fraction in the solvent phase multiplied by the aqueous
solubility of the pure compound provided a reasonable
estimate of aqueous phase concentration of TCA and 1,1-DCE.
Advec tion
Loss of TCA from this hypothetical contaminated zone
occurs via advection and degradation, both of which are a
function of the aqueous phase concentration. The relative
importance of these two mechanisms is a function of the flow
velocity (advection) and the temperature (solubility and
degradation rate). Observations of selected field data
suggest higher concentrations of 1,1-DCE appear in southern
state aquifers where the ground water temperatures are
higher. The model therefore, assigns a temperature of 25C.
The volume of water exchanged through the contaminated
zone is a function of the ground water flow velocity and the
length of the contaminated zone. Fresh water upgradient of
the spill enters the contaminated zone while an equal volume
of water at equilibrium saturation of the contaminants is
displaced. Velocities for the model are expressed as the
per cent of the volume of contaminated water exchanged per

84
day. These values include the "no flow" or "low flow" (0.1%
per day) cases, in which the dominant loss occurs through
degradation. At 0.25% per day, the rate of advection is
comparable to the rate of degradation. Finally, a flow rate
of 0.5% per day represented the case in which the loss of
TCA is primarily due to advection. At flows greater than
0.5% per day the losses would be dominated by the advective
term. These volume exchange rates represent slow flows
and/or very large spill areas. An exchange of 0.5% per day
represents an approximate flow through 5 meters of
contaminated porous media at a rate of 2.5 cm/day.
Degradation Rate
The solubility of TCA affects not only its rate of
advection from the contaminated zone, but also the total
mass of TCA degraded per unit time. The first-order rate
constant at 25C is approximately 0.00226 day'^ as measured
in this study. In a contaminant plume, the half-life for
the degradation of TCA is approximately 10.2 months.
Although the first-order rate coefficient remains constant,
the mass of TCA converted per unit time decreases as the
concentration of TCA in the aqueous phase decreases.
In the model, it was assumed that the TCA concentration
remained at saturation within the zone containing residual
solvent since the TCA that degraded was replaced by
dissolution of the residual solvent. The amount of TCA
degraded per unit time follows zero-order kinetics. The

85
zero-order rate equals the mass converted per unit time in
the first-order equation as the time increment approaches
zero. This becomes 0.00226 day'-*- multiplied by the aqueous
concentration of TCA. A 50% decrease in the solubility
would therefore result in a corresponding 50% decrease in
the mass of TCA degraded per unit time.
Model Parameters and Procedures
Initial conditions for the model include a unit volume
of water (1 liter) in contact with 100 mmoles of TCA. After
equilibrium 11.8 mmoles of TCA will be in the aqueous phase
leaving 88.2 mmoles (approximately 11.8 grams or 8.5 mL) in
the residual solvent phase. The changes in concentration of
TCA or 1,1-DCE in this unit volume are displayed graphically
illustrating the effects of different flow rates, higher
initial mass of TCA, and effect of the presence of an inert
solvent mixed in the residual phase.
Iterative calculations (Appendix C) are made in the
model for advection and degradation in relatively small time
increments, with subsequent reequilibration of the solvent
remaining in the zone of residual contamination. The
residual solvent mass will continue to decrease until at
some point a separate solvent phase does not exist.
Calculations become more difficult (smaller time increments
must be used to attain convergance of the iterative
mathematical solution) and other factors would become more
important as the NAPL is depleted. Therefore, the

86
calculations are stopped when amounts of TCA in the residual
NAPL are less than 10 mmoles. At lower levels of residual
NAPL, the process may become diffusion limited as the NAPL
is trapped in regions of the soil matrix removed from the
aqueous flow. The results of the model are shown in Figures
20-26.
The total mass of TCA in the NAPL showed zero-order
decay with flows from 0.1-0.5% per day (Figure 20). As the
flow rate decreases, slight nonlinearity is observed. This
reflects the slow accumulation of 1,1-DCE in the solvent
phase which begins to decrease the aqueous concentration of
TCA .
The decrease in aqueous concentration of TCA (Figure
21) as the total mass of TCA in the system goes from 100
mmoles to approximately 15 mmoles (slightly in excess of the
solubility) is dependant on the flow. The larger decrease
is observed for the case of no-flow, which results in a 45%
decrease in the aqueous phase concentration after 10 years.
The major reason 1,1-DCE fails to accumulate
significantly in the solvent phase is its higher water
solubility. Having a solubility twice that of TCA, 1,1-DCE
is advected from the zone containing residual solvent more
readily. In the special case of no flow through the system,
1,1-DCE is not advected and begins to accumulate in the
solvent phase affecting the aqueous phase concentration of
TCA. However, since only approximately 20% of the TCA is

[TCA] in water (mmoles/l) Total TCA (mmoles)
87
YEARS
20. Model results: Decrease in total TCA mass in
idual zone as a function of flow.
YEARS
Figure 21. Model results: Change in aqueous concentration
of TCA as a function of flow.

88
converted to 1,1-DCE, the effect of the accumulation is not
observed until substantial degradation has occurred. If all
the TCA degraded in this closed system, 20 mmoles of 1,1-DCE
would be produced, which is 60% of the pure component
aqueous solubility of 1,1-DCE. Therefore, for the initial
conditions of the model, a residual NAPL will exist only
when excess TCA is present.
A comparison of different initial conditions for a
constant flow (0.25%) is shown in Figure 22. With an
increase in amount of residual TCA, the same zero-order
decay rate is observed, indicating that doubling the amount
of TCA in the solvent phase doubles the time needed for
removal of the residual.
In addition, Figure 22 illustrates the rate of loss of
TCA when the initial 100 mmoles is mixed with another
solvent, a hypothetical mixture in which the mole fraction
of the "inert" compound remains at 0.5 in the solvent phase.
This represents a case where a compound with solubility
similar to TCA (like TCE) is present in the residual. The
presence of this other compound causes a 50% reduction in
the aqueous phase concentration of TCA, and therefore the
rate of loss of TCA, doubling the time to remove the TCA
from the residual phase.
The patterns of change in mass of 1,1-DCE in the
solvent or aqueous phase over time are more complex when
there is advection from the system Figure 23. The aqueous

89
YEARS (Flow, 0.25%/Day)
Figure 22. Model results: Change in total mass of TCA as a
function of initial mass of TCA and composition of the
solvent phase.
YEARS (Flow, 0.25%/Doy)
Figure 23. Model results: Pattern of 1,1-DCE formation and
advection as 100 mmoles of TCA in the residual zone
degrades.

90
concentration of 1,1-DCE continues to increase for some time
as the mass of 1,1-DCE in the solvent phase begins to
decrease because its mole fraction continues to increase in
the solvent phase.
The total mass of 1,1-DCE in the zone of residual
contamination increased over time, reaching a maximum as the
TCA mass in the solvent phase approached zero. Increasing
the flow rate not only shortened the time in which 1,1-DCE
was accumulating, but decreased the maximum amount of 1,1-
DCE present in that zone. This is true for the aqueous
phase concentrations (Figure 24) and amount in the solvent
phase (Figure 25). The maximum concentration of 1,1-DCE in
the aqueous phase for a flow of 0.5% per day is
approximately 1 mmole/L (100 mg/L) at the point where some
residual phase is still present. The concentration of TCA
at that time is nearly at saturation (approximately 1500
mg/L).
The changes in aqueous concentration of 1,1-DCE for
larger amounts of TCA originally present or in the presence
of an inert solvent as previously discussed, are shown in
Figure 26. The changes in the amount of 1,1-DCE in the
solvent phase is shown in Figure 27. The inert solvent
increases partitioning into the organic phase, keeping the
aqueous concentration low.
The model illustrates factors which affect the time for
removal of a residual phase under varying conditions, and

91
Figure 24. Model results: Increase in aqueous
concentration of 1,1-DCE forming from degradation of TCA as
a function of flow.
a>
C/5
o
-C
a.
c
<1>
>
O
to
c/5
_cu
o
E
E
CJ
Q
Figure 25. Model results: Pattern of accumulation of 1,1-
DCE in the solvent phase as TCA degrades.

YEARS (Flow, 0.25%/Day)
Figure 26. Model results: Change in aqueous concentration
of 1,1-DCE as a function of initial mass of TCA and
composition of the solvent phase.
YEARS
Figure 27. Model results: Change in total mass of 1,1-DCE
in the residual zone as a function of initial mass of TCA
and composition of the solvent phase.

93
the different concentrations of 1,1-DCE which would result.
Given a constant initial mass of TCA, the maximum
concentration of 1,1-DCE in the aqueous phase occurs at the
lowest flow rates. For flow rates higher than the 0.5%
volume exchange per day the advective term is dominant and
concentrations of 1,1-DCE in the residual zone remain
negligible.
As long as a residual NAPL is present, aqueous
concentrations are dominated by TCA. Equal concentrations
of TCA and 1,1-DCE in the water from monitoring well data
from various sites would occur according to the model
primarily in the plume of dissolved constituents
downgradient from the residual zone, or in the original
spill area after all residual solvent was dissolved or
degraded. The presence of a low solubility compound in the
solvent phase with the TCA will considerably slow TCA rate
of advection and degradation.
First-order degradation will continue in the ground
water plume downgradient from the source and this process
could be modeled (Kinzelbach, 1985). Evidence of the
formation of 1,1-DCE would support the assignment of a
degradation rate. Assuming similar retardation factors for
TCA and 1,1-DCE, equal concentrations of TCA and 1,1-DCE
would occur after approximately 3 half-lives, approximately
2.5 years at 25C.

94
Limitations of the Model Assumptions
This type of model does not address the slower rate of
removal of residual NAPL which would occur from a pool of
excess solvent reaching bedrock. In that case, the surface
area is much smaller and the process is limited by the rate
of diffusion from the surface of the pool. This is expected
to considerably lengthen the time that residual solvent is
present at the source (Schwille, 1988).
In an actual site, complete mixing and equilibrium
conditions will not be maintained over time. The NAPL may
be trapped in pores which have minimal contact with the
water. Although much of the NAPL may be removed through
dissolution and advection, some of the residual NAPL will
remain out of the major flow pathways for the water. Losses
would be limited by diffusion out of these regions and
trends in this zone may represent the "no flow" conditions.
This model addresses relatively small percentage of the
pore space occupied by residual NAPL while theoretically, up
to 40% of the pore space may be occupied by a NAPL. In that
case, the slower degradation of TCA in the NAPL may
contribute to the overall attenuation.

GASOLINE IN GROUND WATER
The overall goal of this part of the dissertation
research was to describe the patterns and the variability in
the partitioning of gasoline components into water.
Specific objectives were to
1. Evaluate concentration ranges of major components
in water extracts of various gasolines.
2. Measure fuel/water partition coefficients.
3. Assess the behavior of oxygenated additives and
their effect on partitioning of hydrocarbons.
4. Describe the changes in gasoline which would occur
through weathering (changes in composition resulting from
environmental exposure).
5. Determine if specific fuel sources could be
identified by the concentrations of components measured in
water extracts of gasoline.
Background
Composition of Gasoline
Gasoline is a complex mixture of volatile hydrocarbons.
The major components are branched-chain paraffins,
cycloparaffins and aromatics. The specific composition will
vary depending on the source of the petroleum as well as the
production method (e.g. distillation or fractionation,
95

96
thermal and catalytic cracking, reforming, isomerization).
Gasoline may also include a number of additives (dyes,
antiknock agents, lead scavengers, anti-oxidants, metal
deactivators, corrosion inhibitors, and volatility/octane
enhancers) (Lane, 1977; Youngless et al 1985).
The approximate volume percent composition is given in
Table 13. Unleaded gasoline generally has a higher fraction
of aromatic hydrocarbons than leaded brands. These "lead-
free brands contain no more than 0.05 gram of lead per
gallon. The use of tetraethyl lead as an antiknock agent
has been phased-out for environmental and health reasons.
Table 13. Volume Percent Composition of Gasoline
(Watts, 1986)
Compound
Unleaded
Leaded
Normal/iso hydrocarbons
55%
59%
isopentane
9-11%
9-11%
n-butane
4-5%
4-5%
n-pentane
2.6-2.7%
2.6-2 7%
Aromatic Hydrocarbons
34%
26%
Xylenes
6-7%
6-7%
Toluene
6-7%
6-7%
Ethylbenzene
5%
5%
Benzene
2-5%
2 5%
Naphthalene
0.2-0.5%
0.2 -0.5%
Benzo(b)fluoranthene
3.9 mg/1
3.9 mg/1
Anthracene
1.8 mg/1
1.8 mg/1
Olefins
5%
10%
Cyclic Hydrocarbons
5%
5%
Additives
Tetraethyl lead
-
600 mg/1
Tetramethyl lead
-
5 mg/1
Dichloroe thane
-
210 mg/1
Dibromoe thane
-
190 mg/1

97
The use of oxygenated blending agents in gasoline, e.g.
methyl tertiary-butyl ether (MTBE) is increasing. The
maximum permitted volume percent of oxygenated compounds is
10% in the state of Florida. This level is regulated under
Chapter 525 of the Florida Statutes (Department of
Agriculture and Consumer Services, Bureau of Petroleum
Inspection, personal communication, 1988). The levels may
vary among states, although EPA sets a limit of 2.0 weight
percent of oxygen in unleaded gasoline. Methanol is
specifically restricted and not sold in Florida.
Gasoline composition will vary in different parts of
the country and at different times of the year. Product
composition may change daily depending on refinery
operations and continues to change in response to changes in
regulations, for example the phase-out of lead. In
addition, independent service stations may obtain product
from different suppliers depending on market conditions.
All of these factors contribute to variability in gasoline
composition and changes will continue to occur in the future
(Coleman et al., 1984)
Multicomponent Liquid-Liquid Equilibria
Fluid-phase equilibria have been extensively studied
and numerous texts, reviews and data compilations have been
published (Prausnitz et al., 1980; Brookman et al. 1985b;
Novak et al., 1987). Basically, mixtures of structurally
related hydrophobic liquids have activity coefficients

98
approximately equal to one in the solvent phase. The
concentration of solute in the aqueous phase follows
Raoult's ideal solution law, and is expected to be
proportional to the mole fraction of solute in the mixture
(Banerjee, 1984).
Solute concentrations in water resulting from contact
with an immiscible mixture containing components which
interact in the solvent phase show deviations from ideal
behavior. The activity coefficients to explain these
deviations could be estimated by the UNIFAC model (Banerjee,
1984). Specific deviations from ideal behavior were
reported for mixtures of aromatic hydrocarbons and saturated
paraffins (Leinonen and Mackay, 1973; Sanemasa et al.,
1987). These deviations resulted in higher aqueous
concentrations (10-20%) than predicted from the ideal
behavior.
The addition of polar organic compounds which are
miscible or highly soluble in water (e.g. ethanol, tert-
butyl alcohol, MTBE) to a mixture of hydrocarbons and water
has a potential cosolvent effect, resulting in an increased
aqueous concentration of hydrocarbon (Groves, 1988; Rubino
and Yalkowsky, 1987; Fu and Luthy, 1986). Among the
methods for estimating the solubility of a solute in a
water cosolvent mixture is one based on a log linear
relationship. Rubino and Yalkowsky (1987) report the

99
expected solubility of a component in a water cosolvent
mixture to be
log Sm = f log Sc + (1-f) log Sw
where Sm is the solubility in the water cosolvent mixture, f
is the volume fraction of cosolvent, Sc is the solubility in
neat cosolvent, and Sw is the solubility in water.
Groves (1988) reported there was no cosolvent effect
for MTBE concentrations up to mole fractions of nearly 0.2
in the hydrocarbon phase, while methanol and ethanol had an
effect at sufficiently high concentrations of alcohol.
Prausnitz et al. (1980) stated that liquid-liquid
equilibria were much more sensitive than vapor-liquid
equilibria to small changes in the effect of composition on
activity coefficients. Therefore, calculations for liquid-
liquid equilibria should be based, whenever possible, at
least in part, on experimental liquid-liquid equilibrium
data. Calculations become increasingly difficult for larger
numbers of components.
The difficulties in applying the above partitioning
calculations to commercial gasoline mixtures include the
large number of components involved and the problem of
determining the moles of each in theoretically infinite
combinations. One approach has been the estimation of
fuel/water partition coefficients which are based on weight
percent or concentration in the fuel rather than mole
fraction.

100
The concentration of a component present in the aqueous
phase has been estimated based on the solubility and weight
percentage in the gasoline. This approach was examined in a
recent laboratory study (Brookman et al., 1985b) on the
solubility of petroleum hydrocarbons in groundwater, where a
reference regular unleaded gasoline (API PS-6) was
equilibrated with organic free, deionized water.
The partitioning of components into water is affected
by the solubility of each compound in pure water and the
gasoline composition. The partitioning of fuel oil
components can be described using a partitioning coefficient
based on the following equation (Brookman et al. 1985b):
Kfw Cf/Cw
where Kfw fuel/water partition coefficient
Cf concentration of component in fuel (g/1)
Cw concentration of component in water (g/1)
The concentration of a particular component in the fuel
was based on the area percent determined in the analysis
which was assumed to approximate the weight percent of that
compound. This was converted to concentration (g/1) based
on an average density for gasoline of 0.74 g/ml (Brookman et
al. 1985a).
There are greater sources of variation in values of
fuel/water partition coefficients to describe gasoline
component partitioning than found for measurements of
octanol/water partition coefficients (Kow). Kow
i s the

101
ratio of a chemicals concentration in a two-phase
octanol/water system. Kow is measured using low solute
concentrations and is a weak function of solute
concentration (Lyman, 1982). The concentrations of
particular gasoline components are variable and may be as
high as 20% of the fuel layer. The second major factor
affecting the fuel/water partition coefficient is the
overall composition of the gasoline. Gasolines can differ
in molar volume (number of moles/liter), which affects
partitioning because the solubility is a function of the
mole fraction of the component in the solvent phase. Also,
certain gasoline components may be cosolvents, or change the
activity coefficients in the solvent phase.
The hydrocarbons which partition into the aqueous phase
are predominantly aromatics, including benzene, toluene and
xylenes. Methyl-tert-butyl ether (MTBE) and other
oxygenated additives are highly water soluble additives
which can be identified in the aqueous phase if they are
present in a given brand of gasoline.
A literature survey on hydrocarbon solubilities
summarized several factors which have been found to affect
solubility (Brookman et al., 1985a). These include
temperature, salinity, and dissolved organic matter. Minor
increases in solubility were noted at 0C as compared to
25C, while hydrocarbon concentration decreased with an
increase in salt concentration. Dissolved organic matter

102
enhanced the solubility of higher molecular weight
hydrocarbons in seawater, however the application of this to
gasoline components in lower ionic strength solutions was
not reported. The solubilities of selected gasoline
constituents in water are summarized in Table 14.
Statistics and Pattern Recognition Applications
A large number of GC analyses of water extracts of
gasolines provided a large data base which became difficult
to characterize by graphical means alone. Various
statistical procedures were employed to describe
similarities and differences among gasoline brands
(commercial or trade name, e.g., Shell, Chevron, etc.) and
grades (e.g., regular, unleaded, super unleaded, etc.) and
to generally improve understanding of the data base.
The simplest solution to gasoline source identification
would be to find a single unique compound which repeatedly
and consistently identified a specific source. "Active
tagging", where a known chemical or physical label is added
to a gasoline, would be a superior method for making
absolute identification, but such procedures have not yet
been implemented in the gasoline market, although crude oils
are often tagged for identification purposes. "Passive
tagging" is a procedure that attempts to identify the fuel
source based on the natural composition of the product.
Fingerprinting, or the identification of fuel type
based on chromatographic patterns, has been successfully

Table 14. Solubilities of Gasoline Components
in Distilled Water
Component
AROMATICS1
benzene
e thylbenzene
o- xylene
m-xylene
p-xylene
isopropylbenzene
n-propylbenzene
3- 4-ethy1toluene
1.2.4-trimethylbenzene
1,2,3-trimethylbenzene
1.3.5-trimethylbenzene
n-butylbenzene
s-butylbenzene
t-butylbenzene
toluene
naphthalene
PARAFFINS AND OLEFINS1
pentane
me thylcyclopentane
n-butane
1- butene
1 -pentene
dodecane
OXYGENATED BLENDING AGENTS2
MTBE
t BA
E thano1
Me thano1
n- butano 1
s-butanol
1Brookman et al., 1985a
O 7
^Verschueren, 1977
Solubility (mg/L 0 250
1740
161
170
146
156
65.3
55
40
59.0
75 2
48.2
11.8
17.6
29.5
532
31.3
39.5
41.8
61.4
222
148
0.0037
48,000
Miscible
Miscible
Miscible
77,000
125-250,000 @20C

104
applied to various petroleum problems. Jet fuels (JP-4 or
Jet A fuel) were distiguishable by GC/FID, even if the fuels
were partially weathered (Roberts and Thomas, 1986).
Characterization of oil spills (Flanigan and Frame, 1977)
was successful using a nitrogen-sensitive detector rather
than with the FID detector.
Jones et al. (1983) reported that aromatic constituents
in crude oil aerobically degrade more quickly than the
normal alkanes, resulting in formation of an unresolved
complex mixture in the GC/MS scan of the aromatic fractions.
The latter could act as a marker for environmental
contamination caused by crude oil leakage or spills.
Dynes and Burns (1987) showed it was possible to detect
and identify petrol burned on cotton wool and weathered for
12 days using by GC with a Hall electrolytic conductivity
detector to obtain the sulfur chromatograms. The GC-FID
interpretations of these samples were inconclusive in
distinguishing the type of fuel product.
Gasoline, kerosene, and heavier oils represent
different boiling ranges of petroleum fuels. For example,
gasoline generally contains constituents having carbon
numbers less than C9, while diesel fuel ranges from Cn to
C20* These differences in composition are easily
distinguishable by GC-FID. The gas chromatogram of fuel for
different grades of gasolines can also show distinguishing
characteristics since higher grades frequently contain

105
higher ratios of aromatic constituents (Senn and Johnson,
1987) .
Background information on statistical procedures is
available from a variety of sources (Pielou, 1984; Wolff and
Parsons, 1983; SAS Institute, Inc., 1985; Gordon, 1981).
The basic procedures used in my analysis and described in
these sources included simple statistics, correlations,
stepwise discriminant analyses, and principal component
analyses.
The interpretation of complex chromatograms using
pattern recognition techniques has been demonstrated
(Hosenfeld and Bauer, 1985). Statistical procedures were
applied without prior knowledge of chromatogram peak
identity for either compound class or type. Peaks were
identified by relative retention time, with principal
component analysis used to define the patterns in the
complex data set. This approach was used to interpret the
gasoline chromatograms in my study.
Partitioning of Gasoline Components into Water
Fuel/Water Partition Coefficients
To examine the partitioning behavior of individual
gasoline components in various grades and brands of
gasolines, neat gasoline and water extracts of gasoline were
analyzed for 31 gasoline samples. Commercial brands
included Amoco, Gulf, Shell, Phillips and Union.

106
Analyses of the water extracts produced simplified
chromatograms compared to the analysis of the neat gasoline
as shown in an example in Figure 28. The chromatogram of a
neat gasoline may show as many as 180 peaks, whereas a water
extract typically showed 40 to 80 peaks under the analytical
conditions employed. The water extracts usually contained
about 10 peaks that had area percentages greater than 1.0.
Four of the gasoline samples contained MTBE, but these
four did not represent either a single brand or grade. One
of the four Shell premium gasolines and one of the two Union
premium gasolines contained MTBE. These samples illustrated
the changes in composition that may occur for a single brand
and grade of gasoline, and also indicated that MTBE was not
a clear "marker compound" for any particular brand or grade.
Methyl t-butyl ether is used as an octane enhancer.
Other octane enhancers include ethanol, methanol,
tertiary-butyl alcohol (TBA), "reformate", "alkylate" or
extra amounts of toluene and/or xylenes. The concentrations
of toluene or the xylene compounds may be lower where a
gasoline contains MTBE than in gasolines without MTBE. The
presence of this additive was also postulated to increase
the water solubility of other components due to its high
pure compound aqueous solubility of 48000 mg/1 (Csikos et
al., 1976). Concentrations of aromatics in the water
extracts for samples containing MTBE, however, were not

107
si
Figure 28. Comparison of
(A) with the chromatogram
the chromatogram of neat
of its water extract (B)
gasoline

108
significantly different (either higher or lower) than
samples without this additive.
Fuel/water partition coefficients were expected to show
greater variation than parameters like octanol/water
partition coefficients because of differences in the
gasoline compositions as previously discussed. The
usefulness of the measured Kfw to estimate the partitioning
of specific gasoline components into water was evaluated.
The 31 measurements of Kfw for the gasoline samples
were examined to evaluate the variability in the coefficient
(Table 15). The variability would determine the general
usefulness for estimating partitioning behavior. The
reported measurements were based on my extraction protocol,
room temperature (ca.22C) and a 1:20 fuel to water ratio.
The relative deviation in Kfw for the 31 samples varied
between 11.5 and 30.0%, while the Kfw 's for the 10
components varied over two orders of magnitude. This is
rather consistent considering the wide variations in the
compositions of the gasolines. To put this in perspective,
Log Kow has been more precisely defined with only a three-
component system and fixed low solute concentration and
Lyman et al. (1982) state "it is frequently possible to
estimate log Kow with an uncertainty of no more than plus or
minus 0.1-0.2 log Kow units."
There were analytical factors which affected the
results. The complex nature of the gasoline mixture created

Table 15. Fuel Water Partition Coefficients as
measured for 31
Average
gasolines.
Standard
Deviation
S olubi1ity
(mg/L)
Benzene
350
75
1740
Toluene
1250
180
530
Ethylbenzene
4500
600
160
m,p-Xylene
4350
530
146,156
o-Xylene
3630
420
170
n-Propylbenzene
18500
5600
55
3-, 4-Ethy1 toluene
12500
2350
40
1,3,5-Trime thylbenzene
10200
2350
48
2 Ethyl toluene
10300
2100
40
1,2,3-Trimethylbenzene
13800
3980
75
*
Brookman et al., 1985a

110
one source of Kfw measurement variability due to the gas
chromatographic coelution of some of the constituents. This
was an example of another problem which would not occur with
systems where composition could be closely controlled. This
complexity factor primarily caused difficulty in obtaining
precise area percentages from gas chromatograms for those
neat gasolines where baseline resolution was not
accomplished. For example, 1,2,4-trimethylbenzene coeluted
with a low solubility hydrocarbon in the analysis of the
neat gasoline, which resulted in a high value for the
partition coefficient.
Among the factors that could affect Kfw was the solute
concentration. The relationship between Kfw and constituent
concentrations in the fuel or aqueous phase was examined and
is illustrated for benzene and toluene in Figure 29 and m,p-
xylene and 1,3,5-trimethylbenzene in Figure 30. Generally,
the Kfw did not show any significant trends with
concentration. Benzene showed a somewhat higher Kfw for low
concentrations which may be due to the coelution of
nonaromatic hydrocarbon compounds during the gas
chromatographic quantification of benzene in the neat
gasolines. The effect of the presence of lower solubility
coeluting hydrocarbons on the estimate of Kfw for benzene
would become more pronounced as the overall area percent of
the benzene decreased. Toluene was generally in much higher
concentration in the neat gasoline than most other

Ill
800
700
600
500
i 400
300
200
100
0
10 30 50 70 90 110
Benzene (mg/I) in aqueous phase
2.0
1.8
1.6
-o 1.4
O
V)
8 1.2
1.0
0.8
0.6
Toluene (mg/I) in aqueous solution
0 20 40 60 80 100 120 140
Figure 29. Fuel/Water partition coefficient (Kfw) as a
function of concentration of benzene and toluene.

8
112
7-
6-
5-
4-
3-
2-
1 -
O
i 1 1 1 1 1 1 1 1 1 1 1
4 8 12 16 20 24
m.p-Xylene (mg/I) in aqueous solution
80
70
60
if)
> 1 50
§
^ I 40
30
20
10
10 14 18 22 26 30 34 38
1,3,5-Trimethylbenzene (g/l) neat gas
Figure 30. Fuel/Water partition coefficient (Kfw) as a
function of concentration of 1,3,5-trimethylbenzene and
m,p-xylene.

113
components and eluted in a region of the chromatogram where
less peak overlap was observed. The partition coefficient
for toluene was not a function of concentration for these
samples.
Water Soluble Blending Agents
The most water-soluble constituent detected in some of
the gasoline samples, MTBE, had no apparent effect on the
aqueous phase concentrations of aromatic compounds measured
for these gasoline-water mixtures. However, the
partitioning of MTBE could not be estimated from the data
generated for the four samples containing this additive
because we could not accurately quantitate the area percent
in the neat gasoline due to peak overlap and decreased FID
sens itivity.
Separate experiments were performed to estimate Kfw for
MTBE, tert-butyl alcohol (t-BA) and ethanol and to evaluate
the effect of these compounds on the partitioning of other
constituents. The aqueous solubility of MTBE is 48,000
mg/1 while ethanol and t-BA are completely miscible in
water.
A Shell regular unleaded gasoline which did not contain
oxygenated compounds was spiked with MTBE at concentrations
as high as 11% by weight, which is slightly in excess of the
maximum allowable concentration in the state of Florida.
Water extracts of the gas o1ine-blend mixtures were obtained

114
as described earlier. This procedure was repeated for
ethanol and t-BA.
The Kfw for MTBE was found to be 15.7 +/- 4.3 (Table
16). The standard deviation for the measured Kfw's is 0.05
log units, while the Kfw was about a factor of 20 less than
benzene. The partition coefficients for the alcohols were
all found to be less than one. Accurate calculations of the
partition coefficients for the highly water soluble
components, ethanol and t-BA, were difficult using the
analytical protocols and fuels described in this
dissertation. Most of these alcohols were detected in the
aqueous phase, and coelution with other fuel components made
quantitation of the small amount of alcohol in the fuel
layer difficult to accurately perform.
These Kfw results can be compared with those reported
by Groves (1988) by calculating a solvent water partitioning
coefficient based on concentrations, similar to Kfw, instead
of ratios of mole fractions as reported. The solvent/water
coefficients as recalculated from Groves (1988) were
Compound Solvent system K
MTBE
MTBE
E thano1
Me thanol
Benzene/water
Hexane/water
Benzene/water
Hexane/water
23.3
14.9-15.5
0.05
0.002-0.004
These solvent/water partition coefficents correlate with the
Kfw experimental results. A partially miscible cosolvent
like MTBE concentrated in the solvent or fuel phase, while
the completely miscible alcohols predominantly partitioned

115
Table 16.
Partitioning
of Oxygenated
Blending Agents
Compound
Original
Equi1ib rium
Kf W
Fuel
Fue 1
water
%
g/L
mg/L
MTBE
2 2
7.5
430
17
MTBE
4 3
17.8
730
24
MTBE
6.5
19.8
1450
14
MTBE
8.6
27.2
1910
14
MTBE
10.8
33.0
2460
13
MTBE
10.8
29.4
2640
11
t-BA
4
<1
1410
<1
t-BA
4
<1
1540
<1
t-BA
6
<1
2310
<1
t-BA
6
<1
2300
<1
t-BA
8
<1
2440
<1
t-BA
10
<1
3160
<1
t-BA
10
<1
3750
<1
Ethanol
4
<1
1540
<1
Ethanol
8
<1
3050
<1
Ethano1
8
<1
3050
<1
E thanol
10
<1
3880
<1

116
into the aqueous phase. The Kfw measured for MTBE of 15.7
was closer to the solvent/water partition coefficient for
the hexane/water (14.9-15.5) system than for the
benzene/water (23.3) system which suggests the fuel behaved
more like hexane than benzene in describing the partitioning
of MTBE.
Since the water solubility of the oxygenated additives
was so high relative to other fuel constituents, the
possibility of a cosolvent effect existed. This effect
would increase the solubility or partitioning of other fuel
constituents into the aqueous phase. The concentrations of
benzene and toluene measured in the aqueous phase in the
presence of these alcohol additives showed typical
analytical variations in concentration and were not enhanced
at higher percentages of additives (Figure 31).
The solubility of a solute in an aqueous cosolvent
mixture is a function of the mole fraction of the cosolvent
(oxygenated additive) in the aqueous phase. The mole
fraction of oxygenated compound in my extraction experiments
was a function of the volume percent of the additive in the
fuel, and the fuel to water ratio used in the extraction.
The fuel to water extraction ratio was less critical for the
equilibrium concentrations of the less soluble hydrocarbon
components since negligible amounts partition into the water
and a large excess of fuel was present. The addition of
oxygenated compounds to gasoline at a regulated maximum

117
cn
E
o
120
110-
100O
90-
80 -
70-
60-
50-
40 Jl
30-
20-
10
0
0
Benzene
Toluene
mpXylene
+
+
800 1600 2400 3200
Ethanol in aqueous phase (mg/l)
4000
140
120
100
\
I 80
60
40
20
0
Figure 31. Equilibrium concentrations of major gasoline
constituents in water as a function of the concentration of
oxygenated additive (ethanol or t-BA) in the aqueous phase.

Benzene
+
Toluene
o
mp-Xylene
-
+ +
+

+ + +
-
+
Jl

-

o
o
o
o
o
o
o
0
lili
1000
I I 1 1 1 1
2000 3000
4000
TBA in aqueous phase (mq/l)

118
level of 10% for ethanol and t-BA would result in a volume
fraction cosolvent in the aqueous phase of 0.005 (aqueous
mole fraction of less than 0.002) for the 1:20 extraction
procedure used in these experiments. Groves (1988) measured
an increase in benzene solubility of about 18% when the
ethanol mole fraction in water was approximately 0.025
compared to the benzene solubility in water. Therefore,
using my experimental conditions a cosolvent effect would
not be significant. Assuming the only source of alcohol was
from the gasoline, an alternate water/gasoline extraction
ratio of less than 1:1 would be required to show a
measurable effect of cosolvency for the alcohols.
Additives which are only partially miscible in water,
like MTBE, will have lower aqueous concentrations. The
solubility of MTBE in water is 48 g/1 or a maximum volume
fraction cosolvent less than 0.065. In addition, it will
partition into the fuel or solvent layer reducing the
aqueous phase concentration and decreasing the effectiveness
as a cosolvent.
The fuel/water partition coefficients developed during
this study provided a reasonable estimate of the
partitioning behavior of specific constituents regardless of
the gasoline composition. The selection of the 1:20 ratio
for the water extraction procedure followed the work of
Coleman et al. (1984); however, other ratios were reported
in the literature. Brookman et al. (1985a) used a 1:10 fuel

119
to water ratio and an equilibration time of 2 hours for
three samples and 96 hours for two samples in a study of API
PS-6 Reference gasoline. A comparison of Kfw 's for
selected constituents obtained by Brookman et al. (1985a)
with results obtained in this study are presented in Table
17 .
Table 17.
Comparison of Kfw
Kf w
Compound
This Study Brookman et al., 1985a
Benzene
350
Toluene
1250
Ethylbenzene
4480
m,p- Xylene
4360
0-Xylene
3630
245
1050
3440
3550
2430
The Kfw values reported by Brookman et al. (1985a) were
lower than my average result by 18 to 30%, but within two
standard deviations of my average value. Their results,
which indicated more partitioning into the water phase, may
be due to differences in analytical and extraction protocols
or gasoline composition.
The average fuel/water partition coefficients for
benzene, toluene, and ethylbenzene were approximately twice
the Kow's for those compounds as shown in Table 18.

120
Table 18. Comparison of Kfw and Kow.
Kfw
Kow
Benzene
350
135
Toluene
1250
620
Ethylbenzene
4480
2190
(Lyman et al.
, 1982)
Prediction of Kfy for Other Gasoline Components
An average value for Kfw could be used to describe the
partitioning into water of gasoline components which have
been quantified in this research (e.g. benzene, toluene,
etc.). For other gasoline constituents, the Kfw's may be
estimated from a regression equation. A least squares
regression analysis of the log of the fuel/water partition
coefficient (Log Kfw )and log of the solubility (Log S) was
examined to determine the degree of correlation of those
parameters. If they were highly correlated, the equation
could be used to predict the Kfw for various constituents
which were not measured in this study. Therefore, the
aqueous solution concentration of other components could be
estimated from their solubility in pure water and their
weight percentage in the fuel (Cw = Cf/Kfw).
Brookman et al. (1985b) applied this regression model
to data generated in a study of partitioning for a well
characterized reference gasoline supplied by API. The
following equation was derived based on concentrations of

121
six aromatics: benzene, toluene, ethylbenzene, o-, m-, and
p- xylenes.
Log Kfw -0.884 Log S + 0.975 r2 0.99
where S is the solubility of the pure hydrocarbon in water
(moles/1) and r2 is the correlation coefficient. This
equation would apply to the specific gasoline, water ratio
(1:10) and temperature of 13C used to generate the values
of Kfw. The regression analysis provided a conservative
equation which would be most accurately applied only to
aromatic compounds.
Brookman et al. (1985b) then applied this type of
regression model to several components including 2-butene,
2-pentene, n-butane, 1,2,4-trimethylbenzene, 2 me thylbutane
and n-pentane in addition to the BTX components and ethyl
benzene that were originally included in the first equation.
The log plot of the twelve components' solubilities and
partition coefficients resulted in the linear equation:
Log Kfw -1.018 log S + 0.706 r2 = 0.87
The correlation coefficient was lower when paraffins
and olefins were included with aromatic compounds in the
regression analysis, but suggested that Kfw was correlated
to solubility for a variety of components in addition to the
aromatic constituents. Therefore, solubility may be used to
provide an estimate for the partitioning behavior of a
broader spectrum of gasoline constituents. Brookman et al.
(1985b) stated one limitation of this model was the caveat

122
that it was applicable to only the specific gasoline for
which it was derived.
The results of my research indicate that the Kfw may be
more widely applied to estimate fuel/water partitioning for
a range of fuel compositions i.e. different brands and
grades. The relationship between log of the average
fuel/water partition coefficient and the log of the
solubility was determined for MTBE and eight aromatic
compounds; benzene, toluene, ethylbenzene, m,p-xylene,
o-xylene, n-propylbenzene and 1,2,3-trimethylbenzene. The
plot of the log S versus log Kfw is show on Figure 32, and
the regression equation was
Log Kfw -0.84 log S + 1.29 r2 0.98
The plot showed a strong linear relationship for log
Kfw and log S. The variability was due, as stated earlier,
to a variety of factors including differences in composition
of the fuels and peak overlap.
Changes in Concentrations with Time
An ageing effect occurs when complex fuel mixtures
remain in contact with air or water. The more soluble and
volatile components will be differentially removed from the
fuel, producing a time-dependent composition for the
hydrocarbon layers, and also the waters in contact with
them.
To evaluate the patterns of the changes in composition
with time due to solubilization, multiple extractions were

LOG K
6
5 -
4 -
3 -
2 -
1 -
0
-4 -3 -2 -1 0
LOG Solubility
Figure 32. Correlation between fuel/water partition coefficient and pure component solubility.
1,2,3-Trimethylbenzene
MTBE

124
performed for a gasoline sample (Phillips regular unleaded)
which contained approximately 5% MTBE. After the first
extraction was performed using the methods described
previously, the water fraction was sampled and the remaining
water layer was removed. Forty ml of distilled deionized
water were added to the vial to reextract the fuel layer.
The procedure was repeated a third time. The experiment was
performed in triplicate, and average concentrations of
selected constituents are summarized in Table 19.
Concentrations of MTBE showed a dramatic decrease over the
three extractions, while the other components remained
essentially constant. The sum of the area counts of all
components also showed substantial decreases over the three
extractions, reflecting primarily the losses of benzene and
MTBE .
The changes in composition with sequential extractions,
which represent changes with time, can be calculated from
the fuel/water partition coefficients. For each extraction,
the concentration of the component in the aqueous phase was
estimated using the Kfw. The amount of the compound
partitioning into the water was then subtracted from the
fuel layer. The new concentration of the component in the
fuel layer was calculated with a correction made for the
change in volume (based on the sum of the major
constituents). The calculation was then repeated using the

125
Table 19. Multiple Aqueous Extractions
of Gasoline (mg/L)
Compound
1
Extract Number
2
3
MTBE
69
36
18
Benzene
43
40
35
Toluene
57
55
51
Ethylbenzene
3.4
3.4
3
m,p,-Xylene
12
12
12
o-Xylene
6.4
6 3
6
Total Hydrocarbons
198
168
150

126
new concentration of the component in the fuel layer. The
equations were
cw = Cf(initial) / Kfw
Cf(after extraction) = (Cf(initial)*Vf Cw*Vw)/Vf2
where Vf2 was the new fuel volume adjusted for loss of all
major components. Assuming the volume of the fuel was not
decreased significantly after contacting with water the
relative decrease in concentration after n extractions is
Cw,n/Cw,o Cf>n/Cf/o [1/(1 4- Vw/(Vf*Kfw))]n
The results of these calculations are illustrated in
graphic form in Figure 33. The shape of the plot of the
theoretical decrease in concentration in fuel or water was a
function of the partition coefficient. For highly soluble
components like MTBE or benzene, an exponential decay curve
was projected. Compounds with a high Kfw partition into
water at an essentially steady rate over the model range of
the theoretical 100 extractions because the concentration in
the fuel remains essentially constant.
For most of the gasolines evaluated in this study, the
initial aqueous extract concentration was lower for benzene
than toluene as illustrated for a sample of Shell SU2000
gasoline (Figure 34). The benzene concentration will
decrease more rapidly than the toluene concentration from
solubilization because of its lower Kfw, reaching
concentrations lower than for xylenes.

127
Figure 33. Theoretical change in relative aqueous
concentration of selected gasoline constituents with
repeated aqueous extractions of the fuel.

Aqueous Concentration (mg/I) Sc Concentration (mg/I)
128
e 34. Theoretical concentration changes from multiple
us extractions of a Shell super unleaded gasoline.
EXTRACTION NUMBER
Figure 35. Theoretical concentration changes from multiple
aqueous extractions of a Union super unleaded gasoline.

129
A few Union gasolines had higher initial concentrations
of benzene as compared to toluene in the water extracts
(Figure 35). As the sample ages, the benzene concentration
would decrease more rapidly than the concentration of other
aromatic components. Since high equilibrium concentrations
of benzene in water were less commonly measured, and benzene
concentration would decrease rapidly, high concentrations of
benzene measured in water in equilibrium with residual
gasoline in field sites would represent relatively
unweathered gasoline and may suggest particular sources.
The Kfw was measured for many constituents like toluene
which were in high concentrations in the fuel and water.
The fuel/water partition coefficient should be reevaluated
for low solute concentrations which will develop as ageing
occurs. The values presented in this study provide a
reasonable estimate for many of these compounds over a
fairly wide range of concentrations (e.g. toluene
concentrations in water ranging from 23 to 133 mg/1). The
slight increase in Kfw as concentration decreased, was
primarily noted for the more soluble constituents, benzene
and MTBE, but may occur for other constituents.
Differences in Water Extracts of Gasolines
The objective of this section is to describe and
evaluate similarities and differences in water extracts of
gasolines in part by using selected routine statistical
procedures.

130
Emphasis was placed on a selected number of brands of
gasoline. Sixty-five samples were collected at different
times, and from different sources, to establish if samples
of a single brand provided a unique pattern of constituent
concentrations to distinguish it from a second brand of
gasoline. Chromatograms of the actual neat gasoline samples
were presented and compared among themselves (Harder et al.,
1987). The focus here is on the water-soluble extracts
since the dissolved gasoline constituents may be the only
(or at least the major) components identified during an
investigation of a subsurface contamination site. The
predominant peaks (area counts and frequency of detection)
in the gas chromatograms were selected for statistical
analysis and evaluation.
To be able to generalize about particular brands or
grades of gasoline, samples were collected on at least two
occasions each from different sources (terminals or gasoline
stations). This was done to evaluate the variability in
constituent concentrations for a single gasoline brand and
grade. In addition to the samples analyzed in the partition
experiment described earlier, samples were also collected
locally from gasoline stations located in Gainesville. A
complete list of the data used to establish constituent
concentration patterns for various gasoline brands and
grades is presented in Appendix D.

131
Eaul1ibrium Concentrations of Major Constituents
The equilibrium concentrations of selected major
components in the water soluble fractions were quantitated
(Table 20). Water in equilibrium with a relatively fresh
spill of gasoline would be expected to show similar
concentrations. Benzene and toluene typically represented
approximately 55-65% by weight of the hydrocarbons detected
in the aqueous extracts.
Visual Comparison of Water Extracts of Gasoline
Appropriate dilutions for the purge and trap analysis
of the water extracts were, to a large extent, determined by
the presence of the major peaks, benzene, toluene and
xylenes. For this reason, during a typical ground water
investigations, these BTX constituents are the primary
reported analytes.
As discussed in the previous section, the aqueous
equilibrium concentration of a constituent was based on the
solubility of the individual components and the gasoline
composition. Given these critera, the other major peaks
detected in the water extracts were lower molecular weight
hydrocarbons which eluted early in the chromatogram, and
various aromatic constituents which eluted after toluene.
The attenuation settings selected during the printing
of chromatograms determined the visual appearance of the
chromatograms. My analytical protocol generated
chromatograms which could be visually compared, because

132
Table 20. Concentrations of Gasoline Components in
31 Water Extracts (mg/L).
Average
Std. Dev.
Min
Max
Benzene
42.6
18.9
12.3
130
Toluene
69.4
25.4
23.0
185
Ethylbenzene
3.2
0.8
1.3
5 7
m,p- Xylene
11.4
3 8
2.6
22.9
o-Xylene
5.6
1.8
2.6
9 7
n-Propylbenzene
0.4
0.1
0.1
3.0
3-, 4-Ethyltoluene
1.7
0.3
0.8
3.8
1,3,5-Trimethylbenzene
1.0
0.2
0.5
2.8
2-Ethyltoluene
0.7
0.1
0.4
1.6
1,2,3-Trimethylbenzene
0.7
0.2
0.2
2.0

133
consistant dilution and attenuation settings were used.
This approach" would not necessarily be a viable alternative
for routine ground water analyses since the overall
concentrations in the field samples would be quite variable.
The most obvious visual difference (Figure 36) among
the chromatograms resulted from the presence of MTBE. The
peak was very distinct in the water extracts of the
gasolines in which it was detected. Methyl tertiary butyl
ether was detected in Amoco regular unleaded and "Silver"
grades (what is called here "super regular", an intermediate
grade of unleaded between regular and premium) and detected
only randomly in other brands. Because of its irregular
usage, and probable increase in future use, it did not
provide a unique marker.
For samples which did not contain MTBE, the visual
differences became more subtle, since most gasolines contain
similar major constituents. Even when these patterns
appeared distinctive, they were difficult to describe
quantitatively. Premium gasolines can sometimes be
distinguished from regular samples by GC of the neat
gasolines because it usually has fewer peaks corresponding
to aliphatic compounds (Senn and Johnson, (1987). Since the
aliphatic compounds do not partition readily into the
aqueous phase, the ratios of aromatic to aliphatic
constituents were not as easily observed in the gas
chromatograms of the water extracts. However, the higher

134
Figure 36.
gasoline.
Retention Time
Comparison of chromatograms of water extracts of

135
area percentages of aromatic constituents in the fuel would
result in higher aqueous phase concentrations of those
components. Qualitative descriptions of similarities and
differences were not able to simultaneously consider
fluctuations for any single brand and grade.
Preparation of the Data Base for Statistical Analysis
The pattern recognition approach to data analysis
involved a series of steps that attempted to identify
patterns in the experimental data. First, specific peaks
in the chromatograms of the water extracts were selected for
inclusion in the data base. These data were then evaluated
using various statistical procedures.
Twenty-three peaks with the highest area counts and
most frequent appearance in the water extracts of the 65
gasoline samples, were selected from the chromatograms. The
peaks were identified by a peak number, which represented a
particular retention time for the chromatographic elution.
A summary of the peaks, retention times and identification
where available, is shown in Table 21.
The statistical analysis did not require complete
separation nor identification of the peak. As stated
earlier, this type of data interpretation has been done
without prior knowledge of chromatogram peak identity for
either compound class or type (Hosenfeld and Bauer, 1985).
However, most of the peaks which are identified in Table 21
consisted predominantly of the component named. This was

136
Table 21. Peaks Used to Statistically Characterize
Water Extracts of Gasoline
Peak
Approximate
Retention Time (rain)
Compound
P3
5.85
C5 Hydrocarbon
P4
6.26
C5 Hydrocarbon
P5
8.12
MTBE, C5 Hydrocarbon
P6
14.06
Benzene
P7
23.08
Toluene
P 8
30.09
Ethylbenzene
P9
30.69
m p- Xylene
P10
32.19
o-Xylene
Pll
32.78
unidentified
P12
35.69
n-Propylbenzene
P13
36.03
3- and 4- Ethyltoluene
P14
36.39
1,3,5-Trimethylbenzene
P15
36.89
2-Ethyltoluene
P16
37.58
unidentified
P17
38.27
unidentifie d
P18
38.81
1,2,4-Trimethylbenzene
P19
39.31
1,2,3-Trimethylbenzene
P20
44.06
unidentified
P21
44.66
Napthalene
P22
49.21
unidentified
P23
53.21
unidentified

137
supported by comparison of mass spectra and concentrations
measured by GC/MS analysis. Area counts were obviously
related to concentrations and were used without conversion
since not all response factors were available. The data
evaluation presented in this section is, therefore,
dependent on the protocol used for the analysis. Focus
could be placed on only the peaks which have been identified
and which show very little evidence of coeluting peaks;
however, this may eliminate from the analysis peaks which
may distinguish the gasoline types.
After further examination the data from the 23 peaks,
the two earliest eluting peaks were deleted from the
statistical analysis. These earliest eluting peaks showed
more variation than some of the later peaks, and the
concentrations measured were not as reliable because
(1.) there was considerable peak overlap in this area of the
chromatogram; (2.) methanol and acetone, common laboratory
solvents which were used for rinsing sampling equipment,
elute in this region; (3.) these early eluting components
represent the most volatile constituents, making
reproducible extraction and quantification more difficult;
and (4.) the area counts of these peaks were random and
highly variable for each brand and grade. The final data
base consisted of 21 peaks, designated P3 to P23, which were
evaluated for the 65 gasoline samples.

138
Basic Descriptive Statistics
Basic statistics on the peak area counts (Table 22)
summarize the range of values (as area counts) obtained
during the analyses using identical protocols. The results
of the application of these procedures can be used to
establish the consistency or variability in the data for
particular components and may be used in quality control by
highlighting anomolous data points. The concentrations in
the water extracts were approximately 0.3 times the area
count based on the sample volume (0.1 ml) and average
response factor determined by analysis of standards.
The large standard deviation for P5 suggests highly
variable area counts. The peak at this retention time most
commonly consists of an alkene with area counts typically
less than 30. Eight samples with MTBE had area counts at
this retention time ranging from 580-1320.
For peaks P11-P23, the maximum values were much higher
than the average values. The data were examined to
determine if only one sample was responsible for the high
values for each of these components. These maximum values
represented several different samples, although many of the
samples with a high value for one of these components showed
elevated concentrations of a number of these constituents.
Correlations between constituent concentrations were
calculated using the SAS procedure CORR, so linear
relationships between various constituents would be

139
Table 22. Basic Statistics on Parameter Area Count Data
for 21 Gas Chromatography Peaks Identified in the
Water Extracts of Gasolines.
VARIABLE
N
MEAN
STD DEV
SUM
MINIMUM
MAXIM
P 3
65
21.0
9.0
1365
3.05
39 5
P4
65
21.2
9.9
1378
1.83
45.0
P 5
65
139
305
9091
2.81
1315
P6
65
133
65.0
8634
38.5
406
P7
65
217
95.4
14100
80.5
578
P8
65
10.0
3.07
652
4.78
17 9
P9
65
35.6
13.7
2316
8.08
71.7
P10
65
17.6
5.85
1146
8 51
30.4
Pll
65
1 54
1.04
100
0.56
6.88
P12
65
1 30
1.39
84 .
6
0.49
9 .25
P13
65
5.18
1.61
337
2.44
11.9
P14
65
3 14
1.11
204
1.63
8.81
Pi 5
65
2 14
0.67
139
1.25
4.91
P16
65
7.58
2.91
493
3.25
19 9
P17
65
1.48
1.63
96 .
, 2
0.00
5 .95
P18
65
2 .38
0.97
155
1.10
7.40
PI 9
65
2 .11
1.05
137
0.79
6 14
P20
65
2.29
2 .65
149
0.00
11.3
P21
65
2.45
1.09
159
0.73
8.93
P 2 2
65
3 .31
3.42
215
0.00
15.6
P 2 3
65
2 .79
2.42
181
0.00
13.0

140
Identified. The Pearson product-moment correlation is
calculated by the formula
rxy S (x x)(y-y)/ J S(x-x)2 S(y-y)2
The correlation matrix is shown in Table 23. This type of
intervariable correlation indicated the strength of linear
relationships between any two variables. Values of 1 or -1
indicated complete correlation, and as values approached
zero, more scatter in a plot of the two variables was
expected. If two variables were highly correlated then one
variable could be expressed as a linear function of the
other variable, although this does not necessarily imply a
cause and effect. This approach did not address nonlinear
relationships that may exist. Values midrange between
complete correlation or total scatter were difficult to
interpret and would require further evaluation. Higher
values for the correlations (greater than 0.75) are
presented in bold type.
This analysis revealed fairly high correlation (0.78 to
0.90) between peaks P8, P9, and P10 which represent
ethylbenzene, m,p-xylene, and o-xylene respectively.
Similar correlations were not observed for benzene and
toluene (P6 and P7). Although the ratios of benzene and
toluene may be fixed for a particular type of gasoline,
differences in processing may result in varying composition
for different types of gasolines.

141
Table 23. Pearson Correlation Coefficients
N -
P 3
P4
P5
P3
1.00
P4
0.739
1.00
P 5
0.360
0.324
1.00
P6
-0.327
-0.397
-0.227
P7
-0.120
-0.120
-0.007
P8
-0.080
-0.299
0.133
P9
-0.181
-0.448
-0.025
P10
-0.219
-0.467
-0.011
Pll
0.077
-0.031
0.242
P12
0.281
0.157
0.118
PI 3
0.235
0.023
0.166
P14
0.241
0.057
0.229
P15
0.145
-0.004
0.173
P16
0.239
-0.000
0.117
P17
0.081
-0.007
0.319
PI 8
0.206
0.032
0.293
P19
0.408
0.262
0.478
P20
0.239
0.123
0.327
P21
0.179
0.125
0.026
P22
0.103
-0.002
0.166
P23
0.069
-0.007
-0.007
P10
Pll
P12
P10
1.00
Pll
0.147
1.00
P12
0.012
0.503
1.00
PI 3
0.522
0.303
0.503
P14
0.199
0.706
0.681
P15
0.288
0.557
0.769
PI 6
0.459
0.498
0.693
P17
0.225
0.423
0.086
Pi 8
0.369
0.683
0.671
P1 9
0.127
0.630
0.408
P 2 0
0.148
0.625
0.362
P21
-0.087
0.348
0.078
P22
0.056
0.462
0.069
P23
-0.226
0.218
0.091
PI 7
PI 8
P19
P17
1.00
P18
0.490
1.00
P19
0.554
0.650
1.00
P20
0.757
0.735
0.718
P21
0.280
0.267
0.229
P22
0 700
0.446
0.439
P 2 3
0.089
0.028
-0.073
65
P6
P7
P8
P9
1.00
-0.011
1.00
0.258
0.266
1.00
0.292
0.250
0.825
1.00
0.493
0.198
0.783
0.899
-0.140
0.056
0.256
0.205
0.094
-0.037
0.013
-0.052
0.256
0.233
0.310
0.340
0.038
0.146
0.154
0.149
0.257
0.178
0.191
0.150
0.191
0.104
0.317
0.346
-0.068
0.063
0.318
0.237
0.085
0.047
0.329
0.268
-0.212
-0.100
0.324
0.142
-0.147
0.037
0.280
0.155
-0.156
0.209
-0.059
-0.051
-0.178
0.214
0.136
0.079
-0.126
0.198
-0.181
-0.174
P13
P14
PI 5
P16
1.00
0.586
1.00
0.727
0.782
1.00
0.870
0.670
0.751
1.00
0.285
0.125
0.233
0.386
0.709
0.709
0.757
0.817
0.397
0.413
0.415
0.560
0.387
0.460
0.404
0.558
0.339
0.376
0.331
0.255
0.332
0.341
0.267
0.377
0.081
0.313
0.299
0.047
P20
P21
P 2 2
P23
1.00
0.262
1.00
0.673
0.487
1.00
0.058
0.487
0.541 1.00

142
The early eluting peaks, P3 and P4, showed some
correlation (0.739) while various combinations of peaks P12
to P20 showed some linear relationship. A high correlation
between peaks indicated some redundancy in the information
provided by those peaks in distinguishing among brands or
grades of gasolines.
Bivariate Plots
Bivariate plots of selected components provided a
better visualization of the relationship between variables
than did a single Pearson correlation coefficient. Figures
37 through 41 illustrate patterns in the data, with the
various brands identified by letter. Variables with higher
correlation coefficients, e.g., mp-xylene and ethylbenzene
(Figure 37), showed a linear relationship, while a plot of
P3 (C5 hydrocarbon) and m,p-xylene (Figure 38) demonstrated
considerable scatter. A bivariate plot of variables which
were not highly correlated may show random scatter or
perhaps a nonlinear relationship. The data in plot of P5
versus m,p-xylene (correlation coefficient of -0.025) showed
random scatter. Some clustering of the brands of gasoline
can be observed in this plot, particularly with some of the
Shell gasoline samples.
A high linear correlation between two variables implies
that the ratio of the two concentrations is fairly constant
for all the grades and brands evaluated in this study. The
ratio of m,p-xylene to ethylbenzene will not be a

-Xylene
Ethylbenzene
Figure 37. Bivariate plot of area counts of m,p-xylene and ethylbenzene for aqueous
extractions of 65 gasoline samples.
143

50
40 -
30 -
ro
CL
20 -
10 -
0 -
2 6 10 14 18 22 26 30
m,p-Xy!ene
Figure 38. Bivariate plot of area counts of m,p-xylene and peak P3, a C5 hydrocarbon,
for aqueous extractions of 65 gasoline samples.
r=-0.181
G
S
G
C
C
A
c
A
A
P A
p A U
C
G
AJ A
A
S
C G
G
C
e
U
u
s
s
A
U
S S
Pf§S
SS
u
s
u
A Amoco
C Chevron
G Gulf
P Phillips
S Shell
U Union
S
S
S S
S
S
S
d
S
u
144

Toluene
Figure 39. Bivariate plot of area counts of toluene and benzene for the aqueous
extractions of 65 gasoline samples.
145

28
Figure 40. Bivariate plot of area counts of peak P16 and 3- or 4-ethyltoluene for
aqueous extractions of 65 gasoline samples.
146

P3
Figure 41. Bivariate plot of area counts of peaks P3 and P4 for aqueous extractions
of 65 gasoline samples.
147

148
distinguishing characteristic of various types of gasolines.
However, the absolute concentrations in the water extracts
varied among brands. Many Shell samples contained the
higher concentrations of both of these compounds, while the
concentrations in Chevron samples were frequently low.
The plot of toluene versus benzene (Figure 39)
highlighted four Amoco gasoline samples which contained very
high concentrations of toluene, distinguishing them from all
other samples. These samples were all Amoco Superpremium
grade, and only these four samples of this brand and grade
were tested during this study.
Stepwise Discriminant Analysis
The technique of stepwise discriminant analysis was
used to reduce the number of variables to those which are
most useful for discriminating among the several brands and
grades. The SAS procedure "STEPDISC" (SAS Institute Inc.,
1985) was used to select the appropriate subset of
variables. This statistical technique selects variables
based on a minimum level of significance to explain the
variation that exists among types of gasolines. The
variable with the most discriminating power was added first,
then the remaining variables were reevaluated for their
ability to provide additional information. Therefore, a
peak which varies for different types of gasoline may not be
selected if another previously selected peak explains the
same variation.

149
This analysis was performed on the entire data set (21
variables, 65 observations), as well as various subsets of
the data. An example of the summary table (Table 24) shows
the information provided during the analysis of the entire
data set when the procedure was separating data by "type,"
which was a specific brand and grade. Seventeen types
(class levels) of gasoline were included in the data set.
The peaks included by this procedure were the earliest
eluting, P3 through P10, and P13. Therefore, these peaks
provided the most information on the differences between all
the various types of gasolines. The average squared
canonical correlation (ASCC) is close to one if all groups
are well separated and if all or most directions in the
discriminant space show good separation for at least two
groups. For this analysis the ASCC was 0.32 after addition
of the final variable, which indicates relatively poor
separation of the groups.
Typically, the smaller the number of types of gasolines
to be distinguished, the fewer number of peaks were included
by the procedure and the greater the resolution obtained
among the gasoline types. While this approach cannot
provide definitive ways to distinguish among all brands and
grades of gasoline, it is possible that selections among
only a limited number of gasoline types may show clear
relationships.

150
Table 24. Selection of Major Discriminating Peaks
to Detect Differences in Water Extracts of Gasoline
SAS STEPWISE DISCRIMINANT ANALYSIS
65 OBSERVATIONS
17 CLASS LEVELS
SIGNIFICANCE
SIGNIFICANCE
21 VARIABLE(S) IN THE ANALYSIS
0 VARIABLE(S) WILL BE INCLUDED
LEVEL TO ENTER 0.1500
LEVEL TO STAY 0.1500
STEPWISE SELECTION: SUMMARY
VARIABLE
NUMBER
PARTIAL
F
PROB
STEP
ENTERED
IN
R**2
STATISTIC
F
1
P6
1
0.8359
15.284
0.0001
2
P 7
2
0.8336
14.715
0.0001
3
P8
3
0.7741
9.849
0.0001
4
P4
4
0.7442
8.181
0.0001
5
P5
5
0.6671
5.511
0.0001
6
P 3
6
0.5751
3.637
0.0004
7
P9
7
0.5053
2.682
0.0053
8
PI 3
8
0.4567
2.154
0.0244
9
P10
9
0.4829
2.335
0.0151
VARIABLE
NUMBER
WILKS'
PROB <
STEP
ENTERED
IN
LAMBDA
LAMBDA
1
P6
1
0.16407433
0.0001
2
P 7
2
0.02730376
0.0001
3
P8
3
0.00616909
0.0001
4
P4
4
0.00157832
0.0001
5
P5
5
0.00052543
0.0001
6
P3
6
0.00022327
0.0001
7
P9
7
0.00011044
0.0001
8
PI 3
8
0.00006000
0.0001
9
P10
9
0.00003103
0.0001
AVERAGE
SQUARED
VARIABLE
NUMBER
CANONICAL
PROB >
STEP
ENTERED
IN
CORRELATION
AS C C
1
P6
1
0.0522
0.0001
2
P7
2
0.1033
0.0001
3
P8
3
0.1480
0.0001
4
P4
4
0.1878
0.0001
5
P5
5
0.2286
0.0001
6
P3
6
0.2538
0.0001
7
P9
7
0.2710
0.0001
8
PI 3
8
0.2973
0.0001
9
P10
9
0.3217
0.0001

151
The stepwise discriminant analysis was therefore
applied to various subsets of the data. In the first case,
each brand was evaluated separately to determine the peaks
that would distinguish among its various grades. Some
brands had only two classes (grades) to be distinguished,
while others had as many as four for my data set. The peaks
selected by this statistical procedure varied in number,
type and order of selection for the different brands (Table
25). The greatest number of peaks included by this
procedure were selected to distinguish four grades of Shell
gasolines, and clear separation was not obtained (ASCG of
0.92).
For a second subset, each grade was evaluated to select
peaks which would best distinguish between brands (Table
26). Similar to the previous subset, each grade did not
have the same number of classes (brands) or observations.
This approach could be further applied in a variety of
ways, for example, selecting only two brands of interest, or
possibly two specific types. The more limited the number of
classes to be distinguished, the better the possibility of
separating the samples.
The obvious advantage of this statistical procedure was
the focus it provided for subsequent evaluation of the
information. The procedure was applied here in a general
sense, however, its real utility would be in response to a
more specific question; for example, how does one determine

Table 25. Stepwise Discriminant Analysis
Peaks to distinguish grades for each brand
152
Step
Amoco
Chevron
Gulf
Phillips
Shell
Union
1
P7
P 3
P4
P3
P10
P 3
2
P5
P6
P6
PI 2
P13
PI 5
3
P13
P7
PI 3
P 7
4
P14
P18
P21
P14
5
P3
P20
P14
P4
6
P17
P22
Pll
7
P9
P21
8
P6
9
P17
10
P23
11
P15
ASCC*
0.98
1.00
1.00
0.98
0.72
0.90
# Ob s
12
8
6
6
23
10
#Grades
3
3
2
2
4
3
Average squared canonical correlation, following the
final step.
Table 26. Stepwise Discriminant Analysis
Peaks to distinguish brands for each grade
Re guiar
Super
Super
Step
Regular
Unleaded
Regular
Premium
1
P4
P 5
P4
P7
2
P9
P6
PI 2
P6
3
P20
P8
P22
P 8
4
P12
P13
P7
P4
5
P23
P23
P14
6
P7
P10
P23
7
P4
8
P10
9
P3
10
P22
11
P9
ASCC*
0.64
0.83
1.00
0.81
# Ob s
11
26
8
19
# BRANDS
4
6
2
5

153
which of two possible sources was more likely responsible
for a petroleum fuel spill.
There were two ways in which the output from this
procedure could be used to clarify the differences seen
among various samples. The first and simplest option exists
when only two or three peaks were selected by the SAS
procedure, or if those first two or three peaks explain most
of the variation between the classes. Simple plots of the
two or three peaks should provide visual assistance in
seeing the differences represented in the samples.
To illustrate, only two peaks were selected to
distinguish between two grades of Phillips gasoline (P3 and
P12) or among three grades of Union gasoline (P3 and P15).
Simple bivariate plots of these peaks reveal the separation
of grades (Figures 42).
Although graphic representations of two or three peaks
are possible, it is not easy to graphically represent the
difference that exist when a larger number of peaks need to
be included in the analysis. Therefore, one additional
statistical tool, principal component analysis, was used to
provide insight into the differences in the peaks of the
GC/FID chromatograms used in this analysis. The information
from the stepwise discriminant analysis determines the peaks
to include in the principal component analysis and the ASCC
following the selection of the final peak indicates how
successful the separation may be.

5
4-
3-
CM
CL
2-
1 -
0-
14 16 18 20 22 24 26 28 30
P3
Phillips Gasoline
Regular
+ Regular Unleaded
-i 1 1 1 1 1 1 1 1 1 1 1 1 1 r
in
CL
4 i
3
2-
1 -
0
Union Gasolines
* Regular
+ Super Premium
Regular Unleaded
i 1 1 1 1 1 1 1 1 1 1 1 1
4 8 12 16 20 24 28 32
P3
Figure 42. Bivariate plots based on peaks selected using
stepwise discriminant procedure to distinguish grades of
Phillips or Union gasolines.

155
Principal Component Analysis
Principal component analysis is a simple ordination
procedure for projecting a multidimensional data set into
two or three dimensions to reveal intrinsic patterns.
In essence, the data are projected, without any
differential weighting, onto a differently oriented s-space.
The axes of the original coordinate frame in which the data
points are plotted are rotated rigidly around their origin.
This is done in such a way that the pattern of the data is
simplified. Principal component scores are then determined,
being weighted sums of the quantities after they have been
centered by the species means. The first principal axis is
oriented to make the variance of the first principal
component scores as great as possible, and the second is
oriented to make the spread in the data as great as possible
with the restriction that the second axis must be
perpendicular to the first axis. This approach is continued
to the final axis which is equal to the number of variables.
The effort is considered successful when a large proportion
of the total dispersion of the data is parallel with the
first two or three principal axes; then this large
proportion of the information contained in the original,
unvisualizable, s-dimensional data swarm can be plotted in
two-space or three-space and examined (Pielou, 1984).
The stepwise discriminant analysis of the total data
set had an ASCC of 0.32 after the final step. This

156
indicates that the peaks selected will explain only about
32% of the variation in the data. Principal component
analysis will not improve the separation of gasoline types,
but simply provides a visual presentation of the
differences.
Principal component analyses were performed on various
subsets of the data. Each grade (regular unleaded, super
regular and super premium) was evaluated separately (Figures
43 to 45). The ASCC values for the discriminant analyses
for regular and super premium grades were about 0.8 and the
plots reveal incomplete separation. Amoco samples,
particularly for the super premium grades, were frequently
separated from other brands. Shell super premium gasolines
showed more clustering (similarities) than other brands.
Since there were only two super regular gasoline
brands, Amoco and Shell, and the discriminant ASCC value was
one, these two sample types were separated in the principal
component plot.
These principal component plots do not reveal adequate
separation of all brands included in the analysis. Specific
pairs of brands were examined to determine if they could be
separated in a principal component plot. The data subsets
included all grades of each the two gasolines brands. This
type of analysis, if successful, could be used in a specific
spill situation where two possible sources are present.

Principal Component 2
4
3
2
1
0
-1
-2
-3
-4
-4 -2 0 2 4
A
Amoco
A
C
Chevron
G
Gulf
P
Phillips
S
Shell
U
Union
A
>
O
A
C
A
S
c
sS
G
S
S
su
J
Up
P
P
s
p
U
S
Principal Component 1
Regular Unleaded Gasolines
Figure 43. Plot of principal component scores for 6 brands of regular unleaded gasoline.
157

Principal Component 2
2.5
2
1.5
1
0.5
0
-0.5
-1
-1.5
-2
-2.5
-5-3-11 35
Principal Component 1
Super Premium Gasolines
Figure 44. Plot of principal component scores for 6 brands of super premium gasolines.
A Amoco
C Chevron
G Gulf
P Phillips
S Shell
U Union
U
U Cu
§
A
C
A
158

Principal Component 2
4
3 -
2 -
1 -
0 -
-1 -
-2 -
-3 -
-4 -
-3-11 3
Principal Component 1
Super Regular Gasolines
A Amoco
S Shell
Figure 45. Plot of principal component scores for Shell and Amoco super regular
gasolines.

160
Plots of the principal components for three brand pairs are
illustrated in Figures 46 to 48.
Amoco and Shell gasolines (Figure 46) were well
separated for most of the samples. Shell samples clustered
for all grades, while the Amoco samples showed much
diversity. Amoco gasoline composition was more variable
than other brands since the processing involved the addition
of toluene at varying levels.
Most Chevron and Union gasoline samples were separated
by the first principal component (Figure 47). The three
Chevron samples that had negative first principal component
scores were two regular leaded samples and one super premium
sample. Therefore, the lack of separation of these three
samples did not represent a factors specific to one
particular grade.
Gulf and Phillips gasolines (Figure 48) did not cluster
or separate from each other. Some of the differences were a
result of the differences in the grades of these gasolines.
The principal component plots are a visual comparison
of the similarities and differences of the gasoline brands
and grades evaluated in this data set. The gasoline samples
for a single brand and grade did not usually form close
clusters due to variability resulting from differences in
sampling source and time of collection. The principal
component scores between two brands may provide a way of
describing the differences in those brands for selected

PRIN
4

3 -
2 -
1 -
O
-1 -
-2 -
-3
-3
+
+

X
7
-1
PRIN 2
Amoco
Shell
RU
a RU
+ SR
x SR
SP
<¡
00
T)
T
1
T
3
Figure 46. Principal component plot of all grades of Amoco and Shell gasolines.

Principal Component 2
4
3
2
1
0
-1
-2
-3
-4
-6 -4 -2 0 2 4 6
Principal Component 1
Figure 47, Principal component plot of all grades of Chevron and Union gasolines.
C Chevron IJ IJ
U Union y
(All Grades)
U u
o
o
u
C c
u
C u 1
J
c
u
c
c
T
162

Principal Component 2
4
3
2
1
0
-1
-2
-3
-4
-6 -4 -2 0 2 4 6
Principal Component 1
Figure 48. Principal component plot of all grades of Gulf and Phillips gasolines.
G Gulf
P Phillips
(All Grades)
G
G
P
P
P G
p P G
P
P
G
i 1 1 1 1 1 r
163

164
gasolines. Although all gasoline brand pair combinations
did not separate, some brands like Amoco and Shell appear to
have specific processing which provide a unique composition
of major constituents in the water soluble extracts.
Summary
A data base consisting of 21 of major components
detected in GC/FID gas chromatograms of water extracts of 65
gasoline samples was compiled and evaluated. The gasoline
samples represented six brands and four grades. The samples
were collected over a period of six months from different
sources (gas stations or terminals) to maximize the
variation within any particular brand and grade. Specific
name brands were evaluated, the problem would become more
complex for independant stations which may purchase gasoline
from a variety of sources.
The average aqueous equilibrium concentration (mg/1)
was 42.6 for benzene, 69.4 for toluene and approximately 17
for the xylenes. The concentrations of these particular
constituents may vary by as much as one order of magnitude.
There was some correlation among a number of the
aromatic components like ethylbenzene and xylenes. The
ratios of the concentrations of these components typically
lie within a particular range.
Stepwise discriminant analysis of these data show that
these parameters do not provide sufficiently unique
information about the gasolines to allow separation of all

165
17 types. Specific subsets of the data could be separated
on the basis of the concentrations of the 21 peaks used in
the analysis.
This was further illustrated with principal component
analysis. Differentiation of two brands can sometimes be
accomplished using the 21 components, however this was not
always successful.

SUMMARY AND CONCLUSIONS
This research focused on two major categories of ground
water contaminants, chlorinated solvents and gasoline.
Halogenated organic compounds were examined to determine
degradation mechanisms and and pathways. The partitioning
of gasoline into water and the variability in the water
extracts of gasolines were evaluated.
Chlorinated Solvents
The degradation rate of TCA was compared with the
degradation of other 1,1,1-trihaloethanes, 1,1-
dichloroethane and chloropropane. The brominated analogs of
TCA degraded 10-14 times faster than TCA. As the number of
bromines increased, the percentage of the elimination
product increased.
The mechanism for degradation of the trihaloethanes was
SN1. The formation of a reactive carbocation intermediate
on this type of halogenated alkane was supported by the
following evidence.
1. These molecules are sterically hindered making
nucleophilic attack by an SN2 mechanism difficult as shown
for chloroform.
2. An E2 elimination reaction requires a weakening of
166

167
the carbon-hydrogen bond by interaction with base, resulting
in increased rate with increasing pH. The hydrogens in
1,1,1-trihaloethanes are not acidic unless halogens or other
functional groups are present on the beta carbon. This
suggests that the elimination reaction for TCA is El and not
E2, which also suggests a corresponding SN1 mechanism.
3. The reaction is independent of the concentration
and strength of the nucleophile.
4. Degradation of 1-bromo-1,1-dichloroe thane in a 1 M
KC1 solution resulted in the formation of some TCA providing
evidence of the formation of an ion pair in an E1/SN1
degradation.
5. Halogens have been shown to assist in stabilizing a
carbonium ion in other compounds.
1,1,1-Trichloroethane degraded more rapidly in
distilled deionized water than chloropropane or 1,1-
dichloroethane. In the presence of a strong nucleophile
chloropropane degraded rapidly, DCA degradation rate
increased by approximately a factor of 10, while the TCA
rate increased less than a factor of 2. Chloropropane
reacts by an SN2 mechanism, and DCA appears to degrade by an
intermediate mechanism.
The chemical degradation rate of TCA is independant of
the presence and strength of nucleophiles. Therefore, the
abiotic degradation rate in strongly reducing environments
which contain strong sulfide nucleophiles would be

168
unaffected. Similarly, the degradation rate will not be
enhanced with increasing pH, as will many other alkyl
halides.
The degradation rate is affected by the overall ability
of the solution to create the ion pair. Increasing amounts
of ions in solution will tend to increase the rate, while
high concentrations of dissolved organics decrease the rate.
The effects on the rate would typically be less than a
factor of 2. The primary factor determining the degradation
rate is the temperature.
Halogenated ethenes are resistant to chemical
degradation in aqueous solution. At high pH, TCE and DCE
degradation occurs by elimination in the first step
resulting in formation of chloroacetylene intermediates.
Trichloroethene reacts most quickly due its acidic hydrogen.
Degradation of TCA produces approximately 20-25% DCE,
an elimination product more toxic than the parent compound.
The solubility of DCE is approximately twice that of TCA and
does not tend to accumulate in the zone where residual
solvent may occur. The overall rate of attenuation of TCA
will be decreased if other hydrophobic solvents are present
in the organic phase.
Gasoline
The dominant gasoline components which partition into
water are aromatic constituents, lower molecular weight
alkenes and oxygenated additives. Partition coefficients

169
for benzene and various substituted aromatic compounds were
measured. The coefficient of variation for the estimates as
determined in measurements of 31 gasolines of varying
composition was less than 30%.
Oxygenated additives did not enhance the concentrations
of the more hydrophobic constituents in water at the maximum
levels found in gasolines and a gasoline to water ratio of
less than 1:1. At a ratio of 1:20, no cosolvent effect was
measured for ethanol, t-butyl alcohol, or MTBE. The
partition coefficient for MTBE was approximately 15, and for
the alcohols the coefficient was less than 1.
Variability in the concentrations of aromatic
constituents partitioning into water for 65 gasolines of
varying brands, grades, or sampling locations/times, was
observed. Of the samples evaluated, Amoco was the most
unique and easily distinguished from other brands due to
unusually high concentrations of toluene in the higher
grades. Principal component analysis was not successful in
completely separating the 11 gasoline brand/grade
combinations.
Physical processes cause weathering in the zone
containing residual gasoline through volatilization and
solubilization. Both of these processes can result in
increasing concentrations of the higher molecular weight
aromatic constituents partitioning into water as the mole

170
fraction of these constituents in the solvent phase
increases.
Although differences in aqueous concentrations of
various gasoline constituents measured using GC/FID occur,
measurement of other parameters (lead, EDB, etc.) would
provide important additional information to aid in
identifying gasolines.

APPENDIX A
SOLUBILITY MEASUREMENTS BY LINDA LEE

03/18/88
LEE
Solubility Determination of 1,1 DCE
RPLC Analysis: Vaters Radial Compression Column C-18
Mobile Phase 8/70/22 Acetonitrile/Methanol/Water
Flow Rate- 1.5 ml/min
Vaters 490 UV Detector Wavelength-240 nm
AUF-0.02 Time Constant-1.Osee
Retention Time of 1,1 DCE 3.85min
(No impurities detected at operating conditions)
Equilibration Time 24 hours on platform shaker (Low speed)
Centrifuged at 2400rpm (6000 RCF) for 20 minutes
Inmiscible phase still present in all systems at end of equilibration
1,1 DCE received from Pat Cline with no further purification
1,1 DCE Standard prepared in MeOH according to EPA 601 for VOC
Standard Curve:
InJ. Vol.
Cone.
Mass
(uL)
(ug/ml)
ugx!0A3
Area
25
2067.2
51680
780800
15
2067.2
31008
490670
40
2067.2
82688
1176400
25
2067.2
51680
770230
25
2067.2
51680
796700
Samples:
HPLC
Water
1,1 DCE
InJ. Vol.
Cone.
#
Vol.(ml)
Vt. (g)
(uL)
Area
(ug/ml)
1
4.04
0.54
25
1102800
2982.9
2
4.02
0.49
25
1072800
2901.8
3
4.02
0.52
25
1139100
3081.1
Average
Std. Dev.
% Dev.
2988.6
73.3
2.45
Standard Regression Output:
Constant 0
Std Err of Y Est 33716.90
R Squared 0.980940
No. of Observations 5
Degrees of Freedom 4
X Coefficient(s) 14.78815
Std Err of Coef. 0.268141
172

APPENDIX B
FORTRAN PROGRAM FOR MODELING LOSS
OF RESIDUAL TCA

PROGRAM TCA2
2
3
4
PARAMETER
DIMENSION
DIMENSION
DIMENSION
CHARACTER
WRITE(* *) 'ENTER
EXTENSION ',
1 '(MAX 20 CHAR)'
WRITE(*,*) '=->
WRITE ( * ) '>'
READ(*,2) FNAME
WRITE(*,*) 'ENTER
EXTENSION) ',
1 (MAX 20 CHAR)'
WRITE(*,*) '- >
WRITE(*,*) '>'
READ(*,2) OUTNM
FORMAT (A)
OPEN(5,FILE-FNAME,
OPEN(8,FILE-OUTNM,
(NP=2)
CONC(NP,3),SOLUB(NP)
FRACM(NP),ERRCRI(2)
DEGRT(NP)
FNAME*2 0,OUTNM*2 0
DATA FILENAME
(OR PATH)
OUTPUT FILENAME(OR PATH)
STATUS-'OLD')
STATUS-'UNKNOWN')
WRITE(8
READ(5,
READ(5,
WRITE(8
READ(5,
READ(5,
WRITE(8
*)
*)
(CONC(1,1) ,
1=1,NP)
,3)
(CONC(1,1)
, 1=1,NP
*)
*)
(SOLUB(I),
1=1,NP)
,3)
(SOLUB(I),
1-1,NP)
*)
*)
(ERRCRI(I),
1=1,NP)
,4)
(ERRCRI(I)
, I1,NP
*)
*)
DT,DAYS
,3)
DT,DAYS
*)
*)
FLOWRT
A)
FLOWRT
*)
*)
(DEGRT(I),
1=1,NP)
WRITE(8,4) (DEGRT(I)
FORMAT (6F9.2)
FORMAT (3X,2F9.5)
WRITE(8,*)
WRITE(8,*)
WRITE(8,*) DAY
DCETOT' ,
1' DCESOL DCEWTR'
1=1,NP)
TCATOT
TCASOL
INCLUDING
(OPTIONAL
TCAWTR
174

o o o
175
C INITIALIZE MOLE FRAC & CONCS
CALL DENM(NP,CONC,1,DENOM)
DO 12 K-l.NP
FRACM(K)- CONC(K, D/DENOM
CONC(K,3) (FRACM(K)*SOLUB(K))
CONC(K,2) CONC(K,1)-(CONC(K,3))
12 CONTINUE
CALL DENM(NP,CONC,2,DENOM)
DO 14 K-l.NP
FRACM(K)- CONC(K,2)/DENOM
14 CONTINUE
WRITE(*,*) 'WORKING ON DAY '
WRITE(*,*)
C MAIN LOOP
DO 100 NT-0,DAYS,DT
WRITE(*,20) NT
20 FORMAT ('+',4X,14)
CALL ERRCHK(NP,SOLUB,CONC,FRACM,ERRCRI,ISERR)
IF(IS ERR.EQ.1) THEN
CALL
MATCON(NP,CONC,FRACM,SOLUB,DENOM,ERRCRI)
END IF
C SPECIFIC TO 2 CMPD SYSTEM
MUST REPLACE W/ DIFF CODE
FOR LARGER SYSTEM
C OUTPUT ROUTINES
WRITE(8 '(16,3X,6F9.3) ') NT ((CONC(I J ) J-1,3),
1-1,2)
DEGR2--0.2*DEGRT(l)*CONC(l,3)
CONC(l,l)=CONC(l,1)-CONC(l,3)*(DEGRT(l)+FLOWRT)*DT
CONC(2,1)-CONC(2,1)-(DEGR2+(CONC(2,3)*FLOWRT))*DT
C RECOMP MOLE FRAC & CHECK ERR
CALL DENM(NP,CONC,2,DENOM)
DO 60 K-l.NP
FRACM(K)= CONC(K,2)/DENOM
CONC(K,3) CONC(K,1)-(CONC(K,2))
60 CONTINUE
100 CONTINUE
CLOSE(5)
END

SUBROUTINE ERRCHK(NP,SOLUB,CONC,FRACM,ERRCRI,IS ERR)
PARAMETER (ND-2)
DIMENSION
CONC(ND,3),ERRCRI(2),FRACM(ND),ERRMAT(ND,2)
DIMENSION SOLUB(ND)
INTEGER ERRMAT
IS ERR 0
DO 10 I-l.NP
ABSERR- ABS(CONC(I,3)-SOLUB(I)*FRACM(I))
RELERR- ABSERR/ABS(CONC(I,3) )
IF (ABSERR.LE.ERRCRI(1)) THEN
ERRMAT(1,1)-1
ELSE
ERRMAT(I,l)-0
ENDIF
IF (RELERR.LE.ERRCRI(2)) THEN
ERRMAT(I,2)-l
ELSE
ERRMAT(I,2)-0
ENDIF
IF (ERRMAT(1,1).EQ.0.AND.ERRMAT(1,2).EQ.0) THEN
IS ERR1
RETURN
ENDIF
CONTINUE
RETURN
END
SUBROUTINE MATCON(NP,CONC,FRACM,SOLUB,DENOM,ERRCRI)
PARAMETER (ND-2)
DIMENSION CONC(ND,3),ERRCRI(2),FRACM(ND),SOLUB(2)
DIMENSION PRTDRV(ND,ND)
COMPUTE FUNCTION & INVMAT
NOTE IF NP>2 THEN REPLACE
INVERSION ROUTINE WITH GENERAL
METHOD
DO 50 N-1,50
COMPUTE PARTIALS
PRTDRV(1,l)--SOLUB(l)*FRACM(2)/DENOM
PRTDRV(1,2)-1+SOLUB(1)*FRACM(1)/DENOM
PRTDRV(2,2)--SOLUB(2)*FRACM(2)/DENOM
PRTDRV(2,1)-1+SOLUB(2)*FRACM(l)/DENOM

c¡ o
177
C COMPUTE DETERMINAT
D E T M
PRTDRV(1,1)*PRTDRV(2,2)-PRTDRV(1, 2)*PRTDRV(2,1)
C
COMPUTE INVERSE
PRTDRV(1,1)
PRTDRV(1,2)
PRTDRV(2,2)
PRTDRV(2,1)
PRTDRV(2,2)/DETM
-PRTDRV(2,1)/DETM
PRTDRV(1,1)/DETM
-PRTDRV(1,2)/DETM
SOLV LINEAR EQNS
AN INCREMENT CONC
DO 25 I-l.NP
DELTA-0
F C A L C
(CONC(I,2)+CONC(NP-I+l,2)+SOLUB(I))*FRACM(I)-CONC(I,1)
DO 20 J-l.NP
DELTADELTA-FCALC*PRTDRV(I,J)
20 CONTINUE
C0NC(I,2)-C0NC(I,2)+DELTA
25 CONTINUE
CALL DENM(NP,CONC,2,DENOM)
DO 30 K-l.NP
FRACM(K)- CONC(K,2)/DEN0M
CONC(K,3) CONC(K,l)-(CONC(K,2))
30 CONTINUE
CALL ERRCHK(NP,SOLUB,CONC,FRACM,ERRCRI,ISERR)
IF(ISERR.EQ.0) THEN
RETURN
ENDIF
C WRITE(*, (I4.6F9.4)') N, ((CONC(I J ) J = l,3),
1-1,NP)
50 CONTINUE
WRITE(*,*) 'ITERATED 50 TIMES WITHOUT CONVERGENCE
PAUSE
RETURN
END

SUBROUTINE DENM(NP,CONC,L,DENOM)
DIMENSION CONC(NP,3)
DENOM-O
DO 10 K-l.NP
DENOM-DENOM+CONC(K,L)
CONTINUE
RETURN
END

APPENDIX C
AREA COUNT DATA SET FOR STATISTICAL
ANALYSIS OF WATER EXTRACTS OF GASOLINES

Area counts for peaks used in statistical analysis of water
extracts of gasoline.
Brand
Grade
Date Symbol
P3
P4
P5
MTBE
1
Amoco
RU
12/15/86
A
27.83
23.82
256.14
2
Amoco
RU
1/05/87
A
25.82
24.70
713.67
3
Amoco
RU
2/04/87
A
31.12
30.01
614.49
4
Amoco
RU
2/04/87
A
33.26
31.39
606.22
5
Amoco
RU
2/04/87
A
30.91
27 68
587.94
6
Amoco
SR
1/05/87
B
35.28
37.81
729.93
7
Amo c o
SR
2/04/87
B
34.44
35.62
1315.03
8
Amo c o
SR
2/04/87
B
35.10
35.42
1236.45
9
Amoco
SP
12/15/86
C
15.19
28 39
21.56
10
Amoco
SP
12/15/86
C
13.83
27.06
20.31
11
Amoco
SP
1/05/87
C
25.92
26.81
37.10
12
Amoco
SP
2/04/87
C
30.04
34.70
55.08
13
Chevron
R
1/05/87
D
10.84
21.47
31.40
14
Chevron
R
1/05/87
D
17.26
25.34
33.92
15
Chevron
R
2/04/87
D
17.51
17.20
12.75
16
Chevron
SP
1/05/87
E
11.78
21.76
12.59
17
Chevron
SP
2/04/87
E
16.08
17.43
34.71
18
Chevron
RU
1/05/87
F
32.19
29.28
26.67
19
Chevron
RU
2/04/87
F
39.32
28.35
19.90
20
Chevron
RU
2/04/87
F
33.31
27.63
20.92
21
Gulf
SP
12/15/86
G
3.05
5 30
17.92
22
Gulf
SP
10/20/86
G
14.26
12.90
76.92
23
Gulf
SP
10/20/86
G
15.23
12.98
56.43
24
Gulf
RU
10/20/86
H
39.48
45.01
12.10
25
Gulf
RU
10/20/86
H
36.30
39.61
13.24
26
Gulf
RU
12/15/86
H
17.40
36.41
189.08
27
Phillips
R
10/20/86
I
29.79
31.09
57.71
28
Phillips
R
11/09/86
I
27.06
30.14
31.79
29
Phillips
RU
10/20/86
J
15.79
27.93
16.97
30
Phillips
RU
11/09/86
J
14.50
11 17
8.27
31
Phillips
RU
12/15/86
J
18.47
28.10
20.83
32
Phillips
RU
12/15/86
J
16.78
26.91
20.47
33
Shell
R
11/09/86
K
8.42
6 96
2.81
34
Shell
R
11/09/86
K
8.83
7.09
2.93
35
Shell
R
12/15/86
K
26.20
7.00
10.47
36
Shell
RU
11/09/86
L
10.42
9 11
8.37
37
Shell
RU
12/15/86
L
32.45
25.80
13.83
38
Shell
RU
12/15/86
L
32.96
26.90
4.35
39
Shell
RU
1/05/87
L
21.42
20.85
11.35
40
Shell
RU
2/04/87
L
21 79
21.31
20.01
41
Shell
RU
2/04/87
L
18.51
19.22
29.03
42
Shell
RU
2/04/87
L
20.54
20.91
21.41
43
Shell
SR
1/05/87
M
27.06
20.12
6 .82
44
Shell
SR
2/04/87
M
14.52
18.38
36.14
45
Shell
SR
2/04/87
M
15.98
14.59
16.25
180

181
Brand
Grade
Date Symbol
P3
P4
P5
MTBE
46
Shell
SR
2/04/87
M
15.17
13.11
17 14
47
Shell
SR
2/04/87
M
14.07
14.88
16 .68
48
Shell
SP
11/09/86
N
11.73
12.72
6 .25
49
Shell
SP
11/09/86
N
12.01
21.31
9.81
50
Shell
SP
11/09/86
N
12.21
20.13
8.10
51
Shell
SP
12/15/86
N
20.66
20.66
1164.99
52
Shell
SP
1/05/87
N
17.86
11.85
38 50
53
Shell
SP
2/04/87
N
17.37
10.70
23.73
54
Shell
SP
2/04/87
N
18 55
11.05
20.65
55
Shell
SP
2/04/87
N
18.05
10.77
21.52
56
Union
R
10/20/86
0
13.86
23.48
5.21
57
Union
R
11/09/86
0
16.96
15.47
18.40
58
Union
R
11/09/86
0
16.50
3.98
12.84
59
Union
SP
10/20/86
P
5 30
6.56
4.30
60
Union
SP
11/09/86
P
12.99
9.91
590.51
61
Union
SP
11/09/86
P
8.87
1.83
3 22
62
Union
RU
10/20/86
Q
28.09
29.53
10.26
63
Union
RU
10/20/86
Q
30.59
31.28
10.99
64
Union
RU
11/09/86
Q
30.19
24.27
9.08
65
Union
RU
11/09/86
Q
20.31
7.01
7.03

1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
182
P6
P7
P8
P9
P10
Pll
Benzene
Toluene
Ethylbz
m,p-Xylene
o Xylene
53
.97
210 .
.89
9 .
19
28
.38
13 .
65
1 .
10
40
. 82
80 .
. 53
6 .
71
17
. 02
14 .
46
1.
10
94
.83
183 .
.21
12 .
25
36
.73
17 .
59
1.
48
97
. 12
192 .
. 94
13 .
34
39
.52
18 .
93
6 .
88
87
. 03
165 .
. 59
10 .
95
34
. 00
16 .
42
1 .
68
57
. 36
250 ,
.77
8 .
55
27
. 44
13 .
23
1.
20
70
. 69
192 .
, 30
10 .
83
33
. 14
16 .
37
2 .
81
72
. 62
201.
.11
10.
67
34
.01
16 .
62
1.
54
53
. 23
490 .
. 17
6 .
62
17
.83
8 .
58
0 .
78
52
. 99
491 .
.95
6 .
57
18
. 06
8 .
64
0 .
69
38
.46
578 ,
.55
9 .
04
25
. 26
11 .
76
2 .
54
54
. 06
451 .
.66
10 .
31
30
.22
14 .
99
1 .
50
137
. 24
145 .
. 58
6 .
15
18
.52
9 .
31
0 .
56
139
. 47
143 .
. 06
5 .
53
17
.33
8 .
51
1 .
58
202
. 64
264.
.08
8 .
84
25
. 50
12 .
34
2 .
76
246
. 92
190 .
. 34
6 .
85
17
.57
12 .
24
0 .
59
278
.78
204 .
. 64
8 .
22
28
. 90
30 .
42
1.
02
137
. 98
141 .
. 54
5 .
04
18
.13
10 .
37
1.
08
175
. 51
198 .
. 28
8 .
12
30
. 50
16 .
79
0 .
97
171
. 62
185 .
.59
7 .
39
29
.58
16 .
34
1.
65
86
. 55
153 .
.75
4 .
78
17
.89
9 .
30
1.
74
67
. 22
267 ,
. 18
8 .
90
34
. 20
2 .
76
15 .
80
60
. 50
232 .
. 18
7 .
20
27
. 15
12 .
69
1.
06
85
. 39
123 .
. 54
6 .
30
24
.26
11 .
16
0 .
74
111
. 04
174.
.87
9 .
55
36
. 76
17 .
30
0 .
80
116
. 51
101 .
.27
6 .
78
27
. 70
13 .
33
0.
92
57
.07
115 .
. 26
7 .
03
25
. 35
12 .
16
0.
00
114
. 16
137 .
. 22
6 .
97
27
.33
12 .
55
0 .
98
77
.57
84 .
.83
6 .
20
22
. 80
10 .
14
0 .
62
96
. 12
155 ,
. 09
6 .
26
29
. 37
13 .
75
1 .
02
140
.65
202 ,
.81
10.
31
38
. 11
21 .
15
2 .
86
138
.68
196 .
.55
10 .
11
38
.86
21.
81
0 .
59
97
. 48
143 .
.95
8 .
53
34
.65
16 .
00
0 .
78
106
.91
171 .
.81
10 .
81
41
.52
19 .
21
0 .
94
133
.77
249 .
.17
13 .
00
45
. 84
20 .
64
0.8862
107
.11
153 .
. 26
10.
05
36
.42
2 .
35
16 .
57
122
.23
194 ,
. 64
12 .
27
35
. 00
15 .
86
0 .
88
124
.87
200 ,
. 44
12 .
47
34
. 82
16 .
12
0 .
80
69
.63
112 .
,41
9 .
05
20
. 75
12 .
39
0 .
97
107
.91
166 .
.31
9 .
82
35
.65
16 .
94
2 .
61
102
. 30
172 ,
.68
10 .
77
35
.53
16 .
50
1 .
75
98
.03
145 .
. 80
8 .
61
30
. 45
14 .
76
2 .
00
133
.16
206 .
. 11
11 .
70
44
. 17
19 .
59
2 .
56
146
.62
197 .
.51
11.
11
47
.22
21 .
54
3 .
15
168
. 84
238 .
.53
13 .
85
56
.83
25 .
34
2 .
95
162
. 24
238 .
.83
13 .
94
60
.26
26 .
84
1 .
51
129
.48
157 .
.83
17 .
94
45
.78
20 .
96
1 .
97

1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
183
PI 2 PI 3 P14 P15 P16 PI 7
0.82
4.82
2.68
1.91
6.93
0.77
2.30
5.45
4.47
2.79
8.07
1.89
0.89
5.60
2.95
1.85
7.08
2 56
9.25
8.07
8.81
4.91
27 61
0.00
0.86
5.40
2.47
1 74
6.53
2.21
1.39
4.90
4.25
2.16
7.56
1.06
0.00
5 .33
3.09
2.16
8.19
4.82
0.93
5.72
2.55
2.00
7.99
2 39
0.84
4.39
2.25
1.69
5.45
0.7653
0.80
4.39
2.26
1.68
5 56
0.76
1.12
7.04
5 60
3.09
8 24
0.86
0.79
4.68
2.08
1.72
6 17
2.49
1.17
2 .97
2.09
1.36
4.19
0.56
0.61
2.61
3.67
1.25
4.47
0.82
4.50
4.72
3.72
4.50
7.58
3.11
1.14
2.45
2.49
1.97
4.27
0.81
1.91
11.87
4.35
3.81
15 19
1.94
1.17
5.44
4.72
2.62
7.50
0.82
6.56
9.77
4.25
3.22
19.91
1.89
3 .88
8.97
4.23
3.08
11.17
3 .92
0.82
3 84
2.34
1.93
5.13
0.64
0.80
5 .53
3.97
1.99
7 58
0.80
0.62
3.71
2 74
1.59
5.66
0.90
0.61
4.06
2.48
1.46
5 .31
0.74
0.88
5.96
2.95
2.22
7.82
0.00
0.66
4.34
2.11
1.73
5.96
0.59
0.00
3 .31
2.06
1.36
4.42
0.00
0.78
3 64
2 .58
1.67
5.03
0.00
1 .81
3 .77
2 20
1.39
4.73
0.82
0.96
3.67
2.58
1 72
5.22
0.59
2.00
6.47
3 .85
2.76
9.22
0.79
1 59
6.02
2 60
2 34
8.29
0.52
0.49
3.81
2.19
1 .57
5.20
0.24
0.69
5.02
2.72
2 17
6.97
0.37
5 .11
0.74
2.41
1 .86
6.97
0.58
0.63
4.38
2.49
1 89
6.20
0.41
0.84
5.13
2.42
1.71
6 .37
0.62
0.85
5.11
2.42
1.73
6 19
0.57
0.95
2.44
1.63
1.43
3.25
0.75
1.33
4.92
2 89
2.00
9.21
4.60
0.79
4.92
2 17
1 79
8.05
2.37
0.77
4.12
2 64
1.88
8.41
4.58
1 80
5.59
5.65
2.52
9 34
0.00
1.22
5 .37
3 28
2.20
10.09
5 .95
0.90
6 .67
3 .97
2.61
10.63
4.95
1.13
7.02
3 14
2 32
10.56
2.31
0.67
5 .27
3.33
2 .18
9 80
4.59

48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
184
P6
P7
P8
P9
P10
Pll
Benzene
Toluene
Ethylbz
in p Xylene
o-Xylene
142.56
260.14
14.26
56.41
25.61
1.03
156.94
304.14
16.71
68.54
30.11
1.22
139.82
249.37
12 20
52 64
23.90
0.96
151.26
343.92
17 74
59.57
26.99
2.43
191.15
236.15
13.48
37.66
21.99
0.86
170.03
347.25
16.27
61.69
28.39
1.54
152.79
284.58
12.42
52.61
25 11
2.64
131.91
238.43
11.54
48.21
22.68
2.14
233.39
145.14
0
8.08
13 .88
0.61
127.39
127.21
8.10
32.07
15.68
4.07
202.15
305.79
13.38
71 70
30.43
1. 16
406.30
267.47
13.77
47.19
23.58
0.00
258.05
275.13
7.10
24.71
15.87
1.07
227.50
307.59
11.73
59.77
29 26
1.26
186.40
156.11
8 35
31.15
15.07
0.62
197.03
167.29
9.26
34.49
16.65
0.75
150.22
184.64
10.27
38.04
19.62
1.62
183.63
245.32
9.9998
49.32
23.28
1.04
PI 2
PI 3
P14
PI 5
PI 6
P17
0.84
5 .35
2.88
2.22
7 34
0.43
0.99
6 .22
3.45
2.31
9.01
0.00
0.58
4.46
2 52
1.85
5.82
0.36
1.32
6.94
4.14
2.63
8.02
5 .85
1.81
3.80
2 88
1.89
4.24
0.60
0.89
6.44
3.05
2 .29
10.36
2.49
1.03
5.56
3 .26
2.32
9.89
4.46
1.03
5.24
3.12
2.17
9.80
4.54
0.59
3.15
1.90
1.35
4.21
0.00
0.74
4.75
4.10
1 84
6.69
0.92
0.80
5 .56
3.51
2.04
8 68
0.40
1 .29
5 .75
2.58
2.38
7 .58
0.00
1.24
6.01
3.71
2 74
7.56
0.00
0.74
5.23
3 39
1.90
7 18
0.49
0.68
4.15
2.43
1.86
5.37
0.68
0.87
5.01
2.70
1.93
6.58
0.00
0.83
4.89
3.26
2 36
7 .26
0.00
0.77
4.30
2 73
1 .71
6 12
0.37

1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
185
P18
P19
P20
P21
P22
P23
2 10
2.66
1.33
1.87
2 39
2 36
3 57
2 .51
1.54
2.41
2.14
3.13
2 79
2 .83
7.32
2.31
4.16
5.53
7.40
6.14
11.32
1.72
24.08
0.00
2.48
2.45
2.84
2 54
4.16
1.31
1.92
2.27
1.73
2.46
2 .79
2.80
2.92
5.47
6.63
2.94
8.72
3.20
2 55
3.87
3.62
2.83
4.03
1.38
1.36
1.57
1.23
1.53
2.17
1.76
1.38
0.91
1.19
1.50
2.03
1.79
3.12
1.83
3.74
8 .93
11.10
10.63
1.88
2.64
3.21
1 69
4.04
1.59
1.35
1.07
1.02
1.43
1.94
1.36
1.10
1.17
1.49
3.15
3.42
3.09
2.35
1.74
0.00
33.57
33.57
11.36
1 .75
1.26
0.98
0.73
1.77
1.69
4.14
2.13
2 .83
2.43
4.34
1.12
2.15
1.35
6.48
2 80
8.66
10.51
3 34
2.97
3.01
2 .75
4.64
4.99
3 .83
2.41
5.55
4.68
7.61
1.95
2.11
0.95
2.85
1.08
4.21
2.83
2 63
1.85
0.00
2.63
15.61
12 96
1 87
1.11
0.00
2.49
0.00
4.03
1.70
1.21
0.00
1.58
0.00
2.09
2.60
2.12
0.00
2.71
0.00
2.11
1.95
2.08
1.32
2 .55
2.19
1.39
1.41
1.60
0.00
2.47
0.00
1.57
1.43
1.61
0.44
2 50
1.45
2 38
1 54
1.34
0.00
2.19
0.00
1.83
1.60
1.41
0.40
2.31
0.27
1.91
2 .63
2 .38
2.00
3 .35
3.14
3 .35
2 .67
2 .23
1.15
2.41
1.92
1.29
1.61
1.62
0.26
2.10
0.22
1.52
1.94
2.12
0.54
2.76
0.33
1.82
1.98
2.17
1.08
1.97
2.09
1.65
1 .75
2.05
0.42
2.75
0.24
1.61
1 79
2.58
1.19
2.01
1.98
1.44
1 .76
2.60
1.18
1.91
1.97
1.34
1.43
1.96
1.48
1.24
1.79
4.02
3 .35
2.94
6 60
3 .53
8.42
2 50
2 75
2 71
3 39
3.15
3.99
1.45
2.43
3.61
6.49
3.03
8.22
2.47
2 14
2 15
1 .58
2.29
4.54
2.84
3.97
2 .36
7.57
3.40
9.64
3.22
3 18
5 .23
6.49
3 54
9 54
2 80
3.42
2 .62
3.32
2 .52
3.93
1 34
2 .79
3.42
6.27
2 .83
8.29
2.59

186
P18
P19
48
1.93
1.25
49
2.04
1.18
50
1 59
0.94
51
3.23
2 30
52
1.90
1.33
53
3.43
1.81
54
2.95
3.15
55
3.79
1.82
56
1.41
1.18
57
1.84
2.04
58
2.01
1.21
59
2 60
1.56
60
2.38
1.30
61
1.88
0.79
62
1.85
1.46
63
2.29
1.68
64
2.07
1.83
65
1.71
1.01
P21
P22
P23
1.75
0.37
1.63
1.84
0.40
2.16
1.53
0.00
1.58
2.30
7.23
3.49
1.02
1.95
1.92
1.93
4.39
1.48
2 58
8 13
2.35
2.32
8.30
2.40
1.73
0.00
0.62
3 80
0.57
4.09
1.76
0.66
1.92
2.17
0.00
2.11
2 52
0.66
3.03
1.61
0.00
2.04
2.20
0.00
1.67
2.05
0.00
1.74
3.00
0.33
3 24
1.56
1.87
1 .87
P20
O .36
O 36
0.00
5.27
1.13
3.47
7.09
7.13
0.00
0.00
0.00
0.00
0.00
0.00
0.00
0.00
0.78
0.00

REFERENCES
Banerjee, S., Solubility of Organic Mixtures,
Environmental Science and Technology. 18(8): pp 587- 591,
1984 .
Bentley, T. W., and Schleyer, P. von R., Medium Effects
on the Rates and Mechanisms of Solvolytic Reactions, in
Advances in Physical Organic Chemistry. edited by Gold, V.
and Bethell, D. Academic Press, New York, pp. 1-67, 1977.
Bossert, I., and Bartha, R., The Fate of Petroleum in
Soil Ecosystems, in Petroleum Microbiology, edited by Atlas,
R. M., Macmillan Publishing Co., New York, 1984.
Bouwer, E.J., and McCarty, P.L., Transformations of 1-
and 2-Carbon Compounds under Methanogenic Conditions,
Applied and Environmental Microbiology. 45: pp. 1286-1294,
1983 .
Brookman, G. T., Flanagan, M. and Kebe, J. 0. Literature
Survey: Hydrocarbon Solubilities and Attenuation Mechanisms,
American Petroleum Institute Report, pp. 1-101, 1985a.
Brookman, G. T., Flanagan, M. and Kebe, J. 0. Laboratory
Study on Solubilities of Petroleum Hydrocarbons in
Groundwater. TRC Project Report No. 2663- N31-00. TRC
Environmental Consultants, Inc., East Hartford, CT., 1985b.
Carey, F. A., and Sundberg, R. J., Advanced Organic
Chemistry: Part A: Structure and Mechanisms. Plenum Press,
New York, 1984.
Clark, H. A., and Jurs, P. C., Qualitative Determination
of Petroleum Sample Type from Gas Chromatograms Using
Pattern Recognition Techniques, Analytical Chemistry.
47(3): pp. 374-378, 1975.
Cline, P.V., Abiotic Degradation of 1.1.1-
Trichloroethane: Formation of 1,1-Dichloroethene. M. S.
Thesis, University of Florida, 1987.
Cline, P.V., Delfino, J.J., and Cooper, W.J., Hydrolysis
of 1,1,1-Trichloroethane; Formation of 1,1-Dichloroethene
in Proceedings of NWWA/API Conference on Petroleum
187

188
Hydrocarbons and Organic Chemicals in Ground Water:
Prevention. Detection and Restoration. Dublin, Ohio
National Water Well Association, pp. 239-247, 1986.
Cohen, Y., and Ryan, P. A., Multimedia Modeling of
Environmental Transport: Trichloroethylene Test Case,
Environmental Science and Technology. 19(5): pp. 412-417,
1985 .
Coleman, W. E., The Identification and Measurement of
Components in Gasoline, Kerosene and No. 2 Fuel Oil that
Partition into the Aqueous Phase After Mixing, Archives of
Environmental Contaminants and Toxicology. 13: pp. 171-178,
1984.
Csikos, R., Pallay, I., Laky, J., Radcsenko, E. D.,
Englin, B. A., and Robert, J. A., Low-Lead Fuel with MTBE
and C4 Alcohols, Hydrocarbon Processing. 55: pp. 121-125,
1976 .
Dilling, W. L., Tefertiller, N. B., and Rallos, G. J.,
Evaporation Rates and Reactivities of Methylene Chloride,
Chloroform, 1,1,1-Trichloroethane, Trichloroethylene,
Tetrachloroethylene, and Other Chlorinated Compounds in
Dilute Aqueous Solutions, Environmental Science and
Technology. 9: pp. 833-838, 1975.
Duewer, D. L., Source Identification of Oil Spills by
Pattern Recognition Analysis of Natural Elemental
Composition. Analytical Chemistry. 47(9): pp. 1573- 1583,
1975 .
Dynes, K., and Burns, D. T., Identification of Weathered
Petrol Residues by High-Resolution Gas Chromatography with
Dual Flame Ionisation Detector-Hall Electrolytic
Conductivity Detector, Journal of Chromatography. 396: pp.
183-189, 1987.
Environmental Protection Agency, Water Related
Environmental Fate of 129 Priority Pollutants. U.S.EPA
Report 440/4-79-0296, Vol II, 1979.
Environmental Protection Agency, Health Assessment
Document for Vinylidene Chloride. U.S. EPA Report 600/8-
8 3/03 IF, 1985.
Flanigan, G. A. and Frame, G. M. Oil Spill
"Fingerprinting" with Gas Chromatography, Research and
Development. 28(9): pp. 28-26, 1977.
Florida Department of Environmental Regulations, Florida
Sites List. Tallassee, Florida, 1985.

189
Fu, J.K. and Luthy, R. G., Aromatic Compound Solubility
in Solvent/Water Mixtures. Journal of Environmental
Engineering, 112(2): pp. 328-345, 1986.
Gordon, A. D. Class ification. London, England, Chapman
and Hall, 1981.
Graf, C.G. "VOC's in Arizona's Groundwater; A Status
Report," paper presented at the NWWA FOCUS Conference on
Southwestern Ground Water Issues, Tempe, Arizona, September
1986 .
Greim, H., Wolff, T., Holfler, M., and Lahaniatis, E.,
Formation of Dichloroacetylene from Trichloroethylene in the
Presence of Alkaline Material Possible Cause of
Intoxication after Abundant Use of Chloroethylene-Containing
Solvents. Archives of Toxicology. 56: pp. 74-77, 1984.
Groves, F. R. Jr., Effect of Cosolvents on the
Solubility of Hydrocarbons in Water, Environmental Science
and Technology. 22: pp. 282-286, 1988.
Haag, W.R., and Mill, T., Effect of Subsurface Sediments
on Hydrolysis of Haloalkanes and Epoxides, Environmental
Science and Technology. 22: pp. 658-663, 1988.
Harder, A. M., A Study on Gasoline and its Behavior in a
Soil/Water Environment. M.S. Thesis, University of Florida,
1987 .
Hie, J., Carbon Dichloride as an Intermediate in the
Basic Hydrolysis of Chloroform, A Mechanism for Substitution
Reactions at a Saturated Carbon Atom, Journal of the
American Chemical Society. 72: pp. 2438-2445, 1950.
Hosenfeld, J. M., and Bauer, K. M., Application of
Pattern Recognition to High-Resolution Gas Chromatographic
Data Obtained from and Environmental Survey, in
Environmental Applications of Chemometrics. edited by Breen,
J. J. and Robison, R. E., Washington, D.C., American
Chemical Society, pp. 69-82, 1985.
Jones, D. M., Douglas, A. G., Parkes, R. J., Taylor, J.,
Giger, W., and Schaffner, C., The Recognition of Biodegraded
Pe troleum-Derived Aromatic Hydrocarbons in Recent Marine
Sediments, Marine Pollution Bulletin. 14: pp. 103-108,
1983.
Jones, S. C., Some Surprises in the Transport of Miscible
Fluids in the Presence of a Second Immiscible Phase,

190
Society of Petroleum Engineers Journal. 25(1): pp. 101-112,
1985 .
Kinzelbach, W. K. H., Modelling of the Transport of
Chlorinated Hydrocarbon Solvents in Groundwater: A Case
Study, Water Science and Technology. 17: pp. 13-21, 1985.
Lane, J.C., Gasoline and Other Motor Fuels. In:
Kirk-Othmer Encyclopedia of Chemical Technology edited by
Mark, H.F., McKetta, J. J., Othmer, D. F. and Standen, A.,
New York, Interscience Publishers, 11: pp. 664-671, 1977.
Leinonen, P. J., and Mackay.D., The Multicompoment
Solubility of Hydrocarbons in Water. Canadian Journal of
Chemical Engineering. 51: pp. 230- 233, 1973.
Levenspiel, 0., Chemical Reaction Engineering. New York,
Wiley Publishing Co., 1962.
Lyman, J. W., Reehl, W. F., and Rosenblatt, D. H.,
Handbook of Chemical Property Estimation Methods. San
Francisco, McGraw Hill Publishing Co., 1982.
Lysyj, I., and Newton, P. R., Multicomponent Pattern
Recognition and Differentiation Method, Analytical
Chemistry. 44(14): pp. 2385-2386, 1972.
Mabey, W., and Mill, T., Critical Review of Hydrolysis
of Organic Compounds in Water Under Environmental
Conditions, Journal of Physical Chemical Reference Data. 7:
pp. 383-415, 1978.
March, J., Advanced Organic Chemistry. New York, John
Wiley & Sons, 1985.
Mochida, I., Jun-ichiro, T, Saito, Y., and Yoneda, Y.,
Linear Free-Energy Relationships in Heterogeneous Catalysis.
VI. Catalytic Elimination Reaction of Hydrogen Chloride from
Chloroethanes on Solid Acids and Bases, Journal of Organic
Chemistry. 32: pp. 3894-3899, 1967.
Nkedi-Kizza, P., Rao, P. S. C., and Hornsby, A. G.,
Influence of Organic Cosolvents on Sorption of Hydrophobic
Organic Chemicals by Soils, Environmental Science and
Technology. 19: pp. 975-979, 1985.
Noller, H., and Kladnig, W., Elimination Reactions over
Polar Catalysts: Mechanistic Considerations, Catalvsis
Reviews. 13(2): pp. 149-207, 1976.
Novak, J. P., Matous, J., and Pick, J., Liquid-Liquid
Equilibria. New York, Elsevier, 1987.

191
Olah, G. A., Carbocatlons and Electrophilic Reactions.
John Wiley & Sons, New York, 1974.
Parsons, F., and Lage, G. B., Chlorinated Organics in
Simulated Groundwater Environments, Journal of the American
Water Works Association. 77(5): pp. 52- 59, 1985.
Pearson, C. R., and McConnell, G., Chlorinated Cl and C2
Hydrocarbons in the Marine Environment, Proceedings of the
Roval Society of London. B.. 189: pp. 305-332, 1975.
Pielou, E. C., The Interpretation of Ecological Data.
John Wiley & Sons, New York, 1984.
Prausnitz, J. M., Anderson, T. F., Grens, E. A., Eckert,
C. A., Hsieh, R., and O'Connell, J. P., Computer
Calculations for Multicomponent Vapor Liquid and Liquid-
Liquid Equilibria. Englewood Cliffs, New Jersey, Prentice-
Hall, Inc, 1980.
Queen, A., and Robertson, R. E., Heat Capacity of
Activation for the Hydrolysis of 2,2-Dihalopropanes.
Journal of the American Chemical Society. 88(7): pp. 1363-
1365, 1966.
Quemeneur, F., Bariou, B., and Kerfanto, M., Kinetic
Study of the Hydrolysis of p-Substituted alpha alpha
Dichloro and a,a,a-Trichlorotoluenes in a 50% Water-Acetone
Medium. Cometes Rendus de L'Academie des Sciences. Serie C.
272(5),497-9, 1971, in Chemical Abstracts. CA74(17):87067b.
Quemeneur, F., Bariou, B., and Kerfanto, M., Kinetics of
the Hydrolysis of Some p-Substituted alpha-Halo-, alpha
alpha -Dihalo- and alpha alpha alpha -Trihalotoluenes in
Variable Proportions of Water-Acetone Solvent. Comptes
Rendus de L'Academie des Sciences. Serie C. 278(4), 299-301,
1974, in Chemical Abstracts. CA81 (3) : 12761 j .
Rappaport, Z., Nucleophilic Vinylic Substitution.
Advances in Physical Organic Chemistry, edited by Gold, V.,
New York, Academic Press, Vol 7: pp. 1-111, 1969.
Roberts, P. V., Field Observations of Organic Contaminant
Behavior in the Palo Alto Baylands, Artificial Recharge of
Groundwater. edited by Asano, T., Butterworth Publishers,
New York, pp. 647-679, 1985.
Roberts, A. J., and Thomas, T. C., Characterization and
Evaluation of JP-4, Jet A and Mixtures of these Fuels in
Environmental Water Samples. Environmental Toxicology and
Chemistry. 5: pp 3 -11, 1986.

192
Rubino, J. T., and Yalkowsky, S. H., Cosolvency and
Deviations from Log-Linear Solubilization. Pharmaceutical
Research. 4(3): pp. 231-236, 1987.
Saltzman, S., and Mingelgrin, U., Nonbiological
Degradation of Pesticides in the Unsaturated Zone, in
Pollutants in Porous Media. edited by Yaron, B., Dagan, G.,
Goldshmid, J. Springer- Verlag, New York, NY. 1984.
Sanemasa, I., Miyazaki, Y., Arakawa, S., Kumamaru, M.,
and Degughi, T., The Solubility of Benzene Hydrocarbon
Binary Mixtures in Water. Bulletin of the Chemical Society
of Japan. 60: pp. 517-523, 1987.
SAS Institute Inc., SAS User's Guide: Statistics. Cary,
NC: SAS Institute Inc., 1985.
Schwille, F., Migration of Organic Fluids Immiscible
with Water in the Unsaturated Zone. In: Pollutants in Porous
Media. Yaron, B., Dagan, J., Goldshmid, J. (eds.),
Springer-Verlag, New York, 1984.
Schwille, F., Dense Chlorinated Solvents in Porous and
Fractured Media. Lewis Publishers, Chelsea, Michigan, 1988.
Senn, R. B., and Johnson, M. S., Interpretation of Gas
Chromatographic Data in Subsurface Hydrocarbon
Investigations, Ground Water Monitoring Review. 7(1): pp.
58-63, 1987.
Stengle, T. R., and Taylor, R. C., Raman Spectra and
Vibrational Assignments for 1,1,1-Trihaloethanes and Their
Deuterium Derivatives, Journal of Molecular Spectroscopy.
34: pp. 33-46, 1970.
Swain, C. G., and Scott, C. B., Journal of the American
Chemical Society. 75: pp. 141-143, 1953.
Verschueren, K., Handbook of Environmental Data on
Organic Chemicals. Van Nostrand Reinhold Co., New York,
1977 .
Vogel, T.M., and McCarty, P. L., Rate of Abiotic
Formation of 1,1-Dichloroethylene from 1,1,1-Trichloroethane
in Groundwater, Journal of Contaminant Hydrology. March
1987 .
Walraevens, R., Trouillet, P., and Devos, A., Basic
Elimination of HC1 from Chlorinated Ethanes, Internation
Journal of Chemical Kinetics 6: pp. 777-786 (1974).

193
Watts, G., Groundwater Monitoring Parameters and
Pollution Sources, Ground Water Training Workshop,
Department of Environmental Regulation, Orlando, FL., April
2-3, 1986.
Westrick, J. J., Mello, J. W., and Thomas, R. F., The
Groundwater Supply Survey, Journal of the American Water
Works Association. 76: pp. 52-59 (1984).
Wilson, J.L., and Conrad, S.H., Is Physical Displacement
of Residual Hydrocarbons a Realistic Possibility in Aquifer
Restoration?, in Proceedings of NWWA/API Conference on
Petroleum Hydrocarbons and Organic Chemicals in Ground
Water: Prevention. Detection and Restoration. Dublin, Ohio
National Water Well Association, pp. 274-298, 1985.
Wolfe, N. L., Zepp, R.G., Paris, D. F., Baughman, G. L.,
and Hollis, R. C., Methoxychlor and DDT Degradation in
Water: Rates and Products. Environmental Science and
Technology. 11(12): pp. 1077-1081, 1977.
Wolff, D. D., and Parsons, M. L., Pattern Recognition
Approach To Data Interpretation. New York: Plenum Press,
1983 .
Youngless, T. L., Swansiger, J. T., Danner, D. A., and
Greco, M., Mass Spectral Characterization of Petroleum Dyes,
Tracers, and Additives, Analytical Chemistry. 57: pp.
1894-1902, 1985.

BIOGRAPHICAL SKETCH
Patricia VanOvermeer Cline obtained a B.S in chemistry
from the University of Michigan in May, 1970. She worked as
an Honorarium Instructor in chemistry at the University of
Colorado, Denver Extension, and received a degree in
secondary education in May, 1974 from the University of
Colorado in Boulder. She taught secondary science in
Longmont, Colorado, for four years.
In Madison, Wisconsin, she directed analytical and
field services for Warzyn Engineering, Inc. She served as a
project manager and environmental chemist for studies of
waste treatment and characterization as well as ground water
contamination investigations.
In 1984 she moved with her husband, Ken, and her son,
Brendan, to Gainesville, Florida. She received her Masters
of Science degree in water chemistry from the Department of
Environmental Engineering Sciences at the University of
Florida in May, 1987 and is currently receiving her Ph.D. in
water chemistry. She plans to continue working in
Gainesville as an environmental consultant.
194

I certify that I have read this study and that in my
opinion it conforms to acceptable standards of scholarly
presentation and is fully adequate, in scope and quality,
as a dissertation for the degrea^-trfS Doctor^o&j Philosophy.
fseph J f Delfiny, Chairman
rofessor of Environmental
Engineering Sciences
I certify that I have read this study and that in my
opinion it conforms to acceptable standards of scholarly
presentation and is fully adequate, in scope and quality,
as a dissertation for the degree of Doctor of Philosophy.
Paul A. Chadik
Assistant Professor of
Environmental Engineering
S cience s
I certify that I have read this study and that in my
opinion it conforms to acceptable standards of scholarly
presentation and is fully adequate, in scope and quality,
as a dissertation for the degree of-'iTO'etp-K off'PTT'ilosophy.
>hn) Dorsey
Asferrciate Professor
Chemistry

I certify that I have read this study and that in my
opinion it conforms to acceptable standards of scholarly
presentation and is fully adequate, in scope and quality,
as a dissertation for the degree of-Doctor of Philosophy.
Rao
Ifessor of Soil Science
I certify that I have read this study and that in my
opinion it conforms to acceptable standards of scholarly
presentation and is fully adequate, in scope and quality,
as a dissertation for the degree of Doctor of Philosophy.
Richard A. Yost(J
Associate Professor of
Chemistry
This dissertation was submitted to the Graduate Faculty
of the College of Engineering and to the Graduate School and
was accepted as partial fulfillment of the requirements for
the degree of Doctor of Philosophy.
August 1988
Dean
liege
Engineering
Dean, Graduate School

* VV"V UN IVI RSITY of
yr FLORIDA
Internet Distribution Consent Agreement
In reference to the following dissertation:
AUTHOR: Cline, Patricia
TITLE: Behavior of partially miscible organic compounds in simulated ground
water systems / (record number: 1116907)
PUBLICATION DATE: 1988
I, Patricia Cline a as copyright holder for the aforementioned dissertation, hereby grant
specific and limited archive and distribution rights to the Board of Trustees of the University of
Florida and its agents. I authorize the University of Florida to digitize and distribute the
dissertation described above for nonprofit, educational purposes via the Internet or successive
technologies.
This is a non-exclusive grant of permissions for specific off-line and on-line uses for an
indefinite term. Off-line uses shall be limited to those specifically allowed by "Fair Use" as
prescribed by the terms of United States copyright legislation (cf, Title 17, U.S. Code) as well as
to the maintenance and preservation of a digital archive copy. Digitization allows the University
of Florida to generate image- and text-based versions as appropriate and to provide and enhance
access using search software.
This grant of permissions prohibits use of the digitized versions for commercial use or profit.
Signature of Copyright Holder
Patricia Cline
Printed or Typed Name of Copyright Holder/Licensee
Personal Information Blurred
Printed or Typed Phone Number and Email Address of Copyright Holder/Licensee
/A, y
T)ate of Signature



72
Table 11. Matrix Effects for Degradation Rates of
1-Chloropropane and 1,1,1-Trichloroethane at 70C.
Linear Regression Output for the Plot of In C (ug/1)
vs. Time (hours).
Chloropropane
Regression Output: MQ
Constant (C) 7.20
Std Err of Y Est 0.12
R Squared 0.99
No. of Observation 12
Degrees of Freedom 10
X Coeff. (Rate) -0.0102
Std Err of Coef. 0.0004
Relative rate 1.00
(MQ 1)
Clay
Limes tone
S and
Silica
7 59
7.44
7.14
7.11
0.26
0.07
0.13
0.23
0.94
0.99
0.99
0.96
9
7
11
12
7
5
9
10
-0.0094
-0.0098
-0.0113
-0.0107
0.0009
0.0004
0.0004
0.0007
0.92
0.96
1.11
1.06
1.1.1-Trichloroethane
Regression Output:
MQ
Clay Limestone
Silica Gel
S and
Constant (C^)
5 .86
6.26
5.23
5.91
6.01
Std Err of Y Est
0.11
0.12
0.07
0.18
0.28
R Squared
0.99
0.99
1.00
0.98
0.96
No. of Observations
11
7
6
7
9
Degrees of Freedom
9
5
4
5
7
X Coeff. (Rate)
0.027
-0.020
-0.038
-0.034
-0.040
Std Err of Coef.
0.0008
0.0009
0.0010
0.0020
0.0030
Relative rate
1.00
0.74
1.38
1 25
1.45
(MQ = 1)


169
for benzene and various substituted aromatic compounds were
measured. The coefficient of variation for the estimates as
determined in measurements of 31 gasolines of varying
composition was less than 30%.
Oxygenated additives did not enhance the concentrations
of the more hydrophobic constituents in water at the maximum
levels found in gasolines and a gasoline to water ratio of
less than 1:1. At a ratio of 1:20, no cosolvent effect was
measured for ethanol, t-butyl alcohol, or MTBE. The
partition coefficient for MTBE was approximately 15, and for
the alcohols the coefficient was less than 1.
Variability in the concentrations of aromatic
constituents partitioning into water for 65 gasolines of
varying brands, grades, or sampling locations/times, was
observed. Of the samples evaluated, Amoco was the most
unique and easily distinguished from other brands due to
unusually high concentrations of toluene in the higher
grades. Principal component analysis was not successful in
completely separating the 11 gasoline brand/grade
combinations.
Physical processes cause weathering in the zone
containing residual gasoline through volatilization and
solubilization. Both of these processes can result in
increasing concentrations of the higher molecular weight
aromatic constituents partitioning into water as the mole


4
according to a survey by the US Environmental Protection
Agency (Westrick et al., 1984). Vinylidene chloride (1,1-
DCE) is a highly reactive, flammable liquid which is
primarily used in the production of copolymers with vinyl
chloride or acrylonitrile. Emissions occur during
manufacturing, shipping and production; however, these
emissions represent less than 1% of the total 1,1-DCE
produced (Environmental Protection Agency, 1985). The
common occurrence of this compound as a ground water
contaminant cannot be entirely explained by its production
and usage patterns.
One source of 1,1-DCE develops during the abiotic
degradation of 1,1,1-trichloroethane (TCA). The production
of TCA is more than three times the production of 1,1-DCE,
and unlike 1,1-DCE, it is an end-use product indicating that
emission to the environment is essentially equivalent to the
production (Environmental Protection Agency, 1985). The
presence of 1,1-DCE is typically associated with the
presence of other alkyl halides. Since 1,1-DCE is more
toxic than TCA, the conversion to 1,1-DCE in ground water
can increase the toxicity of the water supply.
The association of 1,1-DCE with TCA can be seen more
dramatically in field data from sites which show high levels
of chlorinated solvents in ground water. A summary of
volatile organic compounds (VOC's) in Arizona's ground water
(Graf, 1986) states that, of the six most commonly detected


60
chloride degraded approximately 3000 times faster than 2,2-
dichloropropane and 10^ faster than TCA.
There were two major differences between my results and
those reported by Queen and Robertson (1966). First, they
reported a rate nearly four times higher for 2-bromo-2-
chloropropane than for 2,2-dibromopropane, while the rate
coefficients I measured for the trihaloethanes containing at
least one bromine were approximately equal (within 20%).
Secondly, they report only formation of the elimination
product for all 2,2-dihaloethanes, while the percent
elimination in my experiments was a function of the number
of bromines and was always less than 60%. The percent
elimination for t-butyl chloride was less than the value
obtained for the trihaloethanes.
The effect of alpha halogen is complex, "combining a
negative inductive effect and an electron-releasing
resonance effect" (Queen and Robertson, 1966, p. 1364).
Based on my results and the results for t-butyl chloride,
elimination was not expected as the primary pathway nor the
large difference in rates observed for the two
dihalopropanes which contained a bromine. The rate data
were determined for the dihalopropanes by a conductance
method. Extraction of the products of solvolysis of 2,2-
dibromoethane with CCI4 and analysis by vapor phase
chromatography (GC) and nmr showed 2-bromopropene was the
only product in other than trace amounts. It may be that


120
Table 18. Comparison of Kfw and Kow.
Kfw
Kow
Benzene
350
135
Toluene
1250
620
Ethylbenzene
4480
2190
(Lyman et al.
, 1982)
Prediction of Kfy for Other Gasoline Components
An average value for Kfw could be used to describe the
partitioning into water of gasoline components which have
been quantified in this research (e.g. benzene, toluene,
etc.). For other gasoline constituents, the Kfw's may be
estimated from a regression equation. A least squares
regression analysis of the log of the fuel/water partition
coefficient (Log Kfw )and log of the solubility (Log S) was
examined to determine the degree of correlation of those
parameters. If they were highly correlated, the equation
could be used to predict the Kfw for various constituents
which were not measured in this study. Therefore, the
aqueous solution concentration of other components could be
estimated from their solubility in pure water and their
weight percentage in the fuel (Cw = Cf/Kfw).
Brookman et al. (1985b) applied this regression model
to data generated in a study of partitioning for a well
characterized reference gasoline supplied by API. The
following equation was derived based on concentrations of


1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
183
PI 2 PI 3 P14 P15 P16 PI 7
0.82
4.82
2.68
1.91
6.93
0.77
2.30
5.45
4.47
2.79
8.07
1.89
0.89
5.60
2.95
1.85
7.08
2 56
9.25
8.07
8.81
4.91
27 61
0.00
0.86
5.40
2.47
1 74
6.53
2.21
1.39
4.90
4.25
2.16
7.56
1.06
0.00
5 .33
3.09
2.16
8.19
4.82
0.93
5.72
2.55
2.00
7.99
2 39
0.84
4.39
2.25
1.69
5.45
0.7653
0.80
4.39
2.26
1.68
5 56
0.76
1.12
7.04
5 60
3.09
8 24
0.86
0.79
4.68
2.08
1.72
6 17
2.49
1.17
2 .97
2.09
1.36
4.19
0.56
0.61
2.61
3.67
1.25
4.47
0.82
4.50
4.72
3.72
4.50
7.58
3.11
1.14
2.45
2.49
1.97
4.27
0.81
1.91
11.87
4.35
3.81
15 19
1.94
1.17
5.44
4.72
2.62
7.50
0.82
6.56
9.77
4.25
3.22
19.91
1.89
3 .88
8.97
4.23
3.08
11.17
3 .92
0.82
3 84
2.34
1.93
5.13
0.64
0.80
5 .53
3.97
1.99
7 58
0.80
0.62
3.71
2 74
1.59
5.66
0.90
0.61
4.06
2.48
1.46
5 .31
0.74
0.88
5.96
2.95
2.22
7.82
0.00
0.66
4.34
2.11
1.73
5.96
0.59
0.00
3 .31
2.06
1.36
4.42
0.00
0.78
3 64
2 .58
1.67
5.03
0.00
1 .81
3 .77
2 20
1.39
4.73
0.82
0.96
3.67
2.58
1 72
5.22
0.59
2.00
6.47
3 .85
2.76
9.22
0.79
1 59
6.02
2 60
2 34
8.29
0.52
0.49
3.81
2.19
1 .57
5.20
0.24
0.69
5.02
2.72
2 17
6.97
0.37
5 .11
0.74
2.41
1 .86
6.97
0.58
0.63
4.38
2.49
1 89
6.20
0.41
0.84
5.13
2.42
1.71
6 .37
0.62
0.85
5.11
2.42
1.73
6 19
0.57
0.95
2.44
1.63
1.43
3.25
0.75
1.33
4.92
2 89
2.00
9.21
4.60
0.79
4.92
2 17
1 79
8.05
2.37
0.77
4.12
2 64
1.88
8.41
4.58
1 80
5.59
5.65
2.52
9 34
0.00
1.22
5 .37
3 28
2.20
10.09
5 .95
0.90
6 .67
3 .97
2.61
10.63
4.95
1.13
7.02
3 14
2 32
10.56
2.31
0.67
5 .27
3.33
2 .18
9 80
4.59


APPENDIX C
AREA COUNT DATA SET FOR STATISTICAL
ANALYSIS OF WATER EXTRACTS OF GASOLINES


76
Vadose Zone
T
Ground y
Water
FlOW
tFFFFFFFFXtF TCA
1,1-DCE
Water
;; Saturated
with Solvent
Chlorinated Solvent Pool
at impermeable layer
Figure 18. Equilibrium model for the attenuation of
residual TCA present beneath the water table.


134
Figure 36.
gasoline.
Retention Time
Comparison of chromatograms of water extracts of


30
were not significantly different; however, the rate in the
seawater matrix was higher than these at the p<0.01 level.
The 10-14% increase in reaction rate observed in the
seawater matrix at this temperature may be due to the
catalytic influence of some component of that matrix, or to
the increase of ionized species concentration in the
solution.
The relationship between the rate coefficient, k, and
temperature is expressed by the Arrhenius equation,
In k In A E^/RT, where is the Arrhenius activation
energy, R is the gas constant, T is the temperature and A is
the Arrhenius pre-exponential factor. The plot of the data
from this and other studies is shown in Figure 6. The plot
includes rates for a variety of matrices including seawater
and sodium hydroxide solutions. Since two products were
formed, the degradation process was complex, but the overall
linearity of the Arrhenius plot implies that a single rate
determining step is involved in the degradation. Based on
these results, an activation energy of 119+/-3 kJ/mol and an
Arrhenius (A) factor of 2.0x10^ s" ^ were calculated.
Extrapolated rate constants and estimated half-lives are
shown in Table 5.
Table 5. Extrapolated Half-Lives for the Degradation of TCA
Temperature (C) Half-life (years)
15
4.5
+/-
0.8
20
2.0
+/-
0.3
25
0.85
+/-
0.13


113
components and eluted in a region of the chromatogram where
less peak overlap was observed. The partition coefficient
for toluene was not a function of concentration for these
samples.
Water Soluble Blending Agents
The most water-soluble constituent detected in some of
the gasoline samples, MTBE, had no apparent effect on the
aqueous phase concentrations of aromatic compounds measured
for these gasoline-water mixtures. However, the
partitioning of MTBE could not be estimated from the data
generated for the four samples containing this additive
because we could not accurately quantitate the area percent
in the neat gasoline due to peak overlap and decreased FID
sens itivity.
Separate experiments were performed to estimate Kfw for
MTBE, tert-butyl alcohol (t-BA) and ethanol and to evaluate
the effect of these compounds on the partitioning of other
constituents. The aqueous solubility of MTBE is 48,000
mg/1 while ethanol and t-BA are completely miscible in
water.
A Shell regular unleaded gasoline which did not contain
oxygenated compounds was spiked with MTBE at concentrations
as high as 11% by weight, which is slightly in excess of the
maximum allowable concentration in the state of Florida.
Water extracts of the gas o1ine-blend mixtures were obtained


97
The use of oxygenated blending agents in gasoline, e.g.
methyl tertiary-butyl ether (MTBE) is increasing. The
maximum permitted volume percent of oxygenated compounds is
10% in the state of Florida. This level is regulated under
Chapter 525 of the Florida Statutes (Department of
Agriculture and Consumer Services, Bureau of Petroleum
Inspection, personal communication, 1988). The levels may
vary among states, although EPA sets a limit of 2.0 weight
percent of oxygen in unleaded gasoline. Methanol is
specifically restricted and not sold in Florida.
Gasoline composition will vary in different parts of
the country and at different times of the year. Product
composition may change daily depending on refinery
operations and continues to change in response to changes in
regulations, for example the phase-out of lead. In
addition, independent service stations may obtain product
from different suppliers depending on market conditions.
All of these factors contribute to variability in gasoline
composition and changes will continue to occur in the future
(Coleman et al., 1984)
Multicomponent Liquid-Liquid Equilibria
Fluid-phase equilibria have been extensively studied
and numerous texts, reviews and data compilations have been
published (Prausnitz et al., 1980; Brookman et al. 1985b;
Novak et al., 1987). Basically, mixtures of structurally
related hydrophobic liquids have activity coefficients


1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
185
P18
P19
P20
P21
P22
P23
2 10
2.66
1.33
1.87
2 39
2 36
3 57
2 .51
1.54
2.41
2.14
3.13
2 79
2 .83
7.32
2.31
4.16
5.53
7.40
6.14
11.32
1.72
24.08
0.00
2.48
2.45
2.84
2 54
4.16
1.31
1.92
2.27
1.73
2.46
2 .79
2.80
2.92
5.47
6.63
2.94
8.72
3.20
2 55
3.87
3.62
2.83
4.03
1.38
1.36
1.57
1.23
1.53
2.17
1.76
1.38
0.91
1.19
1.50
2.03
1.79
3.12
1.83
3.74
8 .93
11.10
10.63
1.88
2.64
3.21
1 69
4.04
1.59
1.35
1.07
1.02
1.43
1.94
1.36
1.10
1.17
1.49
3.15
3.42
3.09
2.35
1.74
0.00
33.57
33.57
11.36
1 .75
1.26
0.98
0.73
1.77
1.69
4.14
2.13
2 .83
2.43
4.34
1.12
2.15
1.35
6.48
2 80
8.66
10.51
3 34
2.97
3.01
2 .75
4.64
4.99
3 .83
2.41
5.55
4.68
7.61
1.95
2.11
0.95
2.85
1.08
4.21
2.83
2 63
1.85
0.00
2.63
15.61
12 96
1 87
1.11
0.00
2.49
0.00
4.03
1.70
1.21
0.00
1.58
0.00
2.09
2.60
2.12
0.00
2.71
0.00
2.11
1.95
2.08
1.32
2 .55
2.19
1.39
1.41
1.60
0.00
2.47
0.00
1.57
1.43
1.61
0.44
2 50
1.45
2 38
1 54
1.34
0.00
2.19
0.00
1.83
1.60
1.41
0.40
2.31
0.27
1.91
2 .63
2 .38
2.00
3 .35
3.14
3 .35
2 .67
2 .23
1.15
2.41
1.92
1.29
1.61
1.62
0.26
2.10
0.22
1.52
1.94
2.12
0.54
2.76
0.33
1.82
1.98
2.17
1.08
1.97
2.09
1.65
1 .75
2.05
0.42
2.75
0.24
1.61
1 79
2.58
1.19
2.01
1.98
1.44
1 .76
2.60
1.18
1.91
1.97
1.34
1.43
1.96
1.48
1.24
1.79
4.02
3 .35
2.94
6 60
3 .53
8.42
2 50
2 75
2 71
3 39
3.15
3.99
1.45
2.43
3.61
6.49
3.03
8.22
2.47
2 14
2 15
1 .58
2.29
4.54
2.84
3.97
2 .36
7.57
3.40
9.64
3.22
3 18
5 .23
6.49
3 54
9 54
2 80
3.42
2 .62
3.32
2 .52
3.93
1 34
2 .79
3.42
6.27
2 .83
8.29
2.59


Cl
I
HvC-C-CI
J I
Cl
1,1,1 -Trichloroethane
F
h3c-q>
ci
+ ci
\
H
Cl
\
/
c = c
/
\
H Cl
1,1 -Dichloroethene
ELIMINATION PATHWAY
r ~
0
i
'/
h3c-c-ci
1

h3c-c-ci
Cl __J
h3c-c-oh
Acetic Acid
SUBSTITUTION PATHWAY
Figure 1. Abiotic degradation pathways for 1,1,1-trichloroethane.


168
unaffected. Similarly, the degradation rate will not be
enhanced with increasing pH, as will many other alkyl
halides.
The degradation rate is affected by the overall ability
of the solution to create the ion pair. Increasing amounts
of ions in solution will tend to increase the rate, while
high concentrations of dissolved organics decrease the rate.
The effects on the rate would typically be less than a
factor of 2. The primary factor determining the degradation
rate is the temperature.
Halogenated ethenes are resistant to chemical
degradation in aqueous solution. At high pH, TCE and DCE
degradation occurs by elimination in the first step
resulting in formation of chloroacetylene intermediates.
Trichloroethene reacts most quickly due its acidic hydrogen.
Degradation of TCA produces approximately 20-25% DCE,
an elimination product more toxic than the parent compound.
The solubility of DCE is approximately twice that of TCA and
does not tend to accumulate in the zone where residual
solvent may occur. The overall rate of attenuation of TCA
will be decreased if other hydrophobic solvents are present
in the organic phase.
Gasoline
The dominant gasoline components which partition into
water are aromatic constituents, lower molecular weight
alkenes and oxygenated additives. Partition coefficients


115
Table 16.
Partitioning
of Oxygenated
Blending Agents
Compound
Original
Equi1ib rium
Kf W
Fuel
Fue 1
water
%
g/L
mg/L
MTBE
2 2
7.5
430
17
MTBE
4 3
17.8
730
24
MTBE
6.5
19.8
1450
14
MTBE
8.6
27.2
1910
14
MTBE
10.8
33.0
2460
13
MTBE
10.8
29.4
2640
11
t-BA
4
<1
1410
<1
t-BA
4
<1
1540
<1
t-BA
6
<1
2310
<1
t-BA
6
<1
2300
<1
t-BA
8
<1
2440
<1
t-BA
10
<1
3160
<1
t-BA
10
<1
3750
<1
Ethanol
4
<1
1540
<1
Ethanol
8
<1
3050
<1
Ethano1
8
<1
3050
<1
E thanol
10
<1
3880
<1


Abstract of Dissertation Presented to the Graduate School
of the University of Florida in Partial Fulfillment of the
Requirements for the Degree of Doctor of Philosophy
BEHAVIOR OF PARTIALLY MISCIBLE ORGANIC COMPOUNDS
IN SIMULATED GROUND WATER SYSTEMS
By
Patricia V. Cline
August 1988
Chairman: Joseph J. Delfino
Major Department: Environmental Engineering Sciences
Serious ground water contamination problems result from
leaks or spills of organic liquids which are partially
miscible in water. Two important categories of these
liquids include low molecular weight chlorinated solvents
and gasoline.
1,1,1-Trichloroethane (TCA) abiotically degrades in
water forming approximately 17-25% 1,1-dichloroethene (1,1-
DCE) via an elimination reaction. The substitution product
is acetic acid. The Arrhenius activation energy is 119 +/-
3 kj/mol with an Arrhenius factor of 2 X 10^ s"^, which
results in an estimated half-life for the degradation at
25C of 10.2 months.
Brominated analogs of TCA hydrolyze 11-13 times faster
v


106
Analyses of the water extracts produced simplified
chromatograms compared to the analysis of the neat gasoline
as shown in an example in Figure 28. The chromatogram of a
neat gasoline may show as many as 180 peaks, whereas a water
extract typically showed 40 to 80 peaks under the analytical
conditions employed. The water extracts usually contained
about 10 peaks that had area percentages greater than 1.0.
Four of the gasoline samples contained MTBE, but these
four did not represent either a single brand or grade. One
of the four Shell premium gasolines and one of the two Union
premium gasolines contained MTBE. These samples illustrated
the changes in composition that may occur for a single brand
and grade of gasoline, and also indicated that MTBE was not
a clear "marker compound" for any particular brand or grade.
Methyl t-butyl ether is used as an octane enhancer.
Other octane enhancers include ethanol, methanol,
tertiary-butyl alcohol (TBA), "reformate", "alkylate" or
extra amounts of toluene and/or xylenes. The concentrations
of toluene or the xylene compounds may be lower where a
gasoline contains MTBE than in gasolines without MTBE. The
presence of this additive was also postulated to increase
the water solubility of other components due to its high
pure compound aqueous solubility of 48000 mg/1 (Csikos et
al., 1976). Concentrations of aromatics in the water
extracts for samples containing MTBE, however, were not


83
increase in solubility of TCA at lower temperatures was not
observed.
The linearity of the change in solubility with
increasing mole fraction for these two compounds suggested
that 1,1-DCE and TCA form a near-ideal solution in the
solvent phase. Based on these measured data, I assumed that
mole fraction in the solvent phase multiplied by the aqueous
solubility of the pure compound provided a reasonable
estimate of aqueous phase concentration of TCA and 1,1-DCE.
Advec tion
Loss of TCA from this hypothetical contaminated zone
occurs via advection and degradation, both of which are a
function of the aqueous phase concentration. The relative
importance of these two mechanisms is a function of the flow
velocity (advection) and the temperature (solubility and
degradation rate). Observations of selected field data
suggest higher concentrations of 1,1-DCE appear in southern
state aquifers where the ground water temperatures are
higher. The model therefore, assigns a temperature of 25C.
The volume of water exchanged through the contaminated
zone is a function of the ground water flow velocity and the
length of the contaminated zone. Fresh water upgradient of
the spill enters the contaminated zone while an equal volume
of water at equilibrium saturation of the contaminants is
displaced. Velocities for the model are expressed as the
per cent of the volume of contaminated water exchanged per


67
The degradation of 1-chloropropane was enhanced by more
than a factor of 100 in the thiosulfate solution, 1,1-
dichloroethane degraded approximately 22 times faster, and
TCA degradation rate increased less than a factor of 2. The
differences in rate enhancement among these compounds is
attributed to differences in mechanism. Part of the
increase in rate of degradation of TCA in thiosulfate is
attributed to the increasing ionic strength, and TCA
degradation rate was clearly less affected by the presence
of thiosulfate than the other compounds. The rate
enhancement for 1,1-dichloroethane was similar to the type
of rate increase which would be observed for secondary
halides which react by an intermediate mechanism.
The thiosulfate solution was used as a matter of
convenience as a strong nucleophile to assist in
demonstrating how knowledge of mechanism may be necessary in
estimating degradation rates as matrices change. Greatest
changes in rates in the presence of sulfur nucleophiles may
be expected for simple primary alkyl halides, and the least
effect occur with compounds which react via an SNl or El
mechanism.
Sediment Matrix Effects
There is considerable interest in possible effects of
solid surfaces on rates of hydrolysis. Most hydrolysis
experiments are performed in simple buffered aqueous
solution. Contaminants in the vadose zone or ground water


Principal Component 2
2.5
2
1.5
1
0.5
0
-0.5
-1
-1.5
-2
-2.5
-5-3-11 35
Principal Component 1
Super Premium Gasolines
Figure 44. Plot of principal component scores for 6 brands of super premium gasolines.
A Amoco
C Chevron
G Gulf
P Phillips
S Shell
U Union
U
U Cu
§
A
C
A
158


77
in media of lower permeability (Schwille, 1984). Additional
factors which influence whether the NAPL will reach the
water table include the spilled volume and infiltration
process.
If sufficient volume of dense NAPL reach the water
table, it will sink into the saturated zone and continue to
migrate downward as long as the retention capacity of the
zone is exceeded. Wilson and Conrad (1985) reported
residual hydrocarbon occupying 15-40% of the pore space in
the saturated zone.
Water continues to flow laterally through the water
saturated zone containing residual NAPL. The globules of
NAPL provide a large interface with the water providing a
solution zone, where the initial concentration of a given
component is proportional to its aqueous solubility as
determined by the NAPL composition. These globules are
generally trapped in the larger pore spaces and are being
prevented from entering the smaller pores due to the high
capillary entrance pressure. There is a reduction in
permeability to water where the residual NAPL is present, as
the largest channels become blocked at several places by
discontinuous solvent ganglia. This forces water to flow
around the solvent in fairly thin films and/or be diverted
into the smaller channels whose carrying capacity
(conductivity) is low (Jones, 1985).


45
These experiments provided evidence that the abiotic
degradation of 1,1,1-trihaloethanes occurred by SN1/E1
rather than SN2 and E2 mechanisms. The trihaloethanes
containing one or more bromine atoms degraded at similar
rates, approximately a factor of 11-13 faster than TCA,
reflecting that bromine was a better leaving group. As the
number of bromines present on the trihaloethanes increased,
the percent of the degradation occurring through the
elimination pathway increased.
Degradation of Haloeenated Ethenes
One of the primary objectives of examining the behavior
of halogenated ethenes was to provide an accurate evaluation
of their formation and stability during degradation of the
corresponding ethanes. The literature provided some
evidence that slow degradation of these ethenes may occur at
a rate of interest for ground water studies.
Supporting the possiblity of degradation, Billing et
al ( 1975) reported half-lives for the abiotic degradation
of trichloroethene (TCE) of 10.7 months (0.002 day~^) and
for tetrachloroethene (PCE) of 9.9 months at 25C.
Molecular oxygen was present and the degradation rates were
suggested to result from oxidation as well as hydrolysis.
In this often referenced work, it was suggested that
mechanisms of degradation at lower temperatures may differ
from rates extrapolated from studies at higher temperatures.


APPENDIX B
FORTRAN PROGRAM FOR MODELING LOSS
OF RESIDUAL TCA


65
hydrolysis in aqueous solution, and the response to an
increase in concentration of a strong nucleophile whose
effect would be a function of the mechanism.
Based on the literature, simple primary alkyl halides
like 1-chloropropane are expected to degrade by an SN2
mechanism. Therefore, 1-chloropropane should show an
increase in degradation rate in the presence of a strong
nucleophile, since the nucleophile is involved in the rate
determining step.
Predicting the degradation rate of 1,1-dichloroethane
is more difficult. Secondary chlorides, like isopropyl
chloride, have been shown to degrade more quickly than the
primary alkyl halides, possibly by an intermediate
mechanism. Chloride can contribute somewhat to the
stability of a carbocation, however, it is not as effective
as a methyl group as discussed previously. In addition, the
presence of a halogen can increase the steric hindrance at
the alpha carbon.
Comparisons of the degradation rates of these compounds
were made at elevated temperature (65C) in pH 7 buffer
solution, and in a 1 M thiosulfate solution. In the buffer
solution the degradation of TCA was approximately 6 times
faster than the hydrolysis of 1-chloropropane. Degradation
of 1,1-dichloroethane was less than 6% of the rate of 1-
chloropropane degradation. This rate comparison is
illustrated in Figure 15.


50
40 -
30 -
ro
CL
20 -
10 -
0 -
2 6 10 14 18 22 26 30
m,p-Xy!ene
Figure 38. Bivariate plot of area counts of m,p-xylene and peak P3, a C5 hydrocarbon,
for aqueous extractions of 65 gasoline samples.
r=-0.181
G
S
G
C
C
A
c
A
A
P A
p A U
C
G
AJ A
A
S
C G
G
C
e
U
u
s
s
A
U
S S
Pf§S
SS
u
s
u
A Amoco
C Chevron
G Gulf
P Phillips
S Shell
U Union
S
S
S S
S
S
S
d
S
u
144


48
the chemical studies show presence of electron withdrawing
groups like chlorine increases the susceptibility of an
olefin to nucleophilic attack, experiments to evaluate
possible reactions were performed.
References to possible hydrolysis reactions of 1,1-DCE
or its brominated analogs were not found upon review of the
literature. The 1,1-dihaloethenes would be less susceptible
to nucleophilic attack than TCE since fewer electron
withdrawing groups are present. The pure compounds however,
are very reactive and polymerize readily. Their
reactivities in dilute aqueous solution have not been
examined.
The focus of my research with halogenated ethenes was
to examine the stability of these compounds in relatively
dilute aqueous solutions and to determine their
susceptibility to nucleophilic attack. Autooxidation or
other reactions of the pure liquid compounds which may be
present in the vadose zone following a spill could occur,
but these reactions are not addressed here.
There were two major purposes for the examination of
the degradation behavior of halogenated ethenes. First, the
stability of the ethene products formed during the
transformation of the geminal trihalides needed to be
determined to accurately describe the kinetics of the
appearance of these elimination products. Secondly,
previous studies which indicated that halogenated ethenes


101
ratio of a chemicals concentration in a two-phase
octanol/water system. Kow is measured using low solute
concentrations and is a weak function of solute
concentration (Lyman, 1982). The concentrations of
particular gasoline components are variable and may be as
high as 20% of the fuel layer. The second major factor
affecting the fuel/water partition coefficient is the
overall composition of the gasoline. Gasolines can differ
in molar volume (number of moles/liter), which affects
partitioning because the solubility is a function of the
mole fraction of the component in the solvent phase. Also,
certain gasoline components may be cosolvents, or change the
activity coefficients in the solvent phase.
The hydrocarbons which partition into the aqueous phase
are predominantly aromatics, including benzene, toluene and
xylenes. Methyl-tert-butyl ether (MTBE) and other
oxygenated additives are highly water soluble additives
which can be identified in the aqueous phase if they are
present in a given brand of gasoline.
A literature survey on hydrocarbon solubilities
summarized several factors which have been found to affect
solubility (Brookman et al., 1985a). These include
temperature, salinity, and dissolved organic matter. Minor
increases in solubility were noted at 0C as compared to
25C, while hydrocarbon concentration decreased with an
increase in salt concentration. Dissolved organic matter


188
Hydrocarbons and Organic Chemicals in Ground Water:
Prevention. Detection and Restoration. Dublin, Ohio
National Water Well Association, pp. 239-247, 1986.
Cohen, Y., and Ryan, P. A., Multimedia Modeling of
Environmental Transport: Trichloroethylene Test Case,
Environmental Science and Technology. 19(5): pp. 412-417,
1985 .
Coleman, W. E., The Identification and Measurement of
Components in Gasoline, Kerosene and No. 2 Fuel Oil that
Partition into the Aqueous Phase After Mixing, Archives of
Environmental Contaminants and Toxicology. 13: pp. 171-178,
1984.
Csikos, R., Pallay, I., Laky, J., Radcsenko, E. D.,
Englin, B. A., and Robert, J. A., Low-Lead Fuel with MTBE
and C4 Alcohols, Hydrocarbon Processing. 55: pp. 121-125,
1976 .
Dilling, W. L., Tefertiller, N. B., and Rallos, G. J.,
Evaporation Rates and Reactivities of Methylene Chloride,
Chloroform, 1,1,1-Trichloroethane, Trichloroethylene,
Tetrachloroethylene, and Other Chlorinated Compounds in
Dilute Aqueous Solutions, Environmental Science and
Technology. 9: pp. 833-838, 1975.
Duewer, D. L., Source Identification of Oil Spills by
Pattern Recognition Analysis of Natural Elemental
Composition. Analytical Chemistry. 47(9): pp. 1573- 1583,
1975 .
Dynes, K., and Burns, D. T., Identification of Weathered
Petrol Residues by High-Resolution Gas Chromatography with
Dual Flame Ionisation Detector-Hall Electrolytic
Conductivity Detector, Journal of Chromatography. 396: pp.
183-189, 1987.
Environmental Protection Agency, Water Related
Environmental Fate of 129 Priority Pollutants. U.S.EPA
Report 440/4-79-0296, Vol II, 1979.
Environmental Protection Agency, Health Assessment
Document for Vinylidene Chloride. U.S. EPA Report 600/8-
8 3/03 IF, 1985.
Flanigan, G. A. and Frame, G. M. Oil Spill
"Fingerprinting" with Gas Chromatography, Research and
Development. 28(9): pp. 28-26, 1977.
Florida Department of Environmental Regulations, Florida
Sites List. Tallassee, Florida, 1985.


BEHAVIOR OF PARTIALLY MISCIBLE ORGANIC COMPOUNDS
IN SIMULATED GROUND WATER SYSTEMS
BY
PATRICIA V. CLINE
A DISSERTATION PRESENTED TO THE GRADUATE SCHOOL
OF THE UNIVERSITY OF FLORIDA IN
PARTIAL FULFILLMENT OF THE REQUIREMENTS
FOR THE DEGREE OF DOCTOR OF PHILOSOPHY
UNIVERSITY OF FLORIDA
fflWSITY OF FLORIDA LIBRARIES
1988


51
Table 7. Summary of Experimental Conditions for which
Halogenated Ethenes were Stable.
Temp .
C
Cmpd.
Time
Days
No .
Obs .
Average
Concentration
C V.
Matrix
27
BCE
160
12
mg/L
4.2
7%
pH 10
27
BCE
160
10
4.2
9%
pH 7
27
BCE
160
10
4.2
6%
pH 4
27
BCE
54
12
88
4%
DW
37
BCE
126
18
80
4%
DW
65
BCE
6
13
392
11%
Thio
27
DBE
160
12
32
4%
pH 10
27
DBE
160
10
32
7%
pH 4
27
DBE
160
10
28
10%
pH 7
27
DBE
54
12
120
2%
DW
37
DBE
126
18
110
11%
DW
65
DBE
6
13
690
12%
Thio
27
DCE
274
15
2.7
7%
Nutrient
27
DCE
386
26
1 1
13%
pH 4
27
DCE
386
27
1.1
14%
pH 7
27
DCE
386
25
1 1
12%
pH 8.5
37
DCE
386
16
2.4
9%
pH 4
37
DCE
386
17
2.4
8%
pH 7
37
DCE
386
26
2.4
10%
pH 8.5
55
DCE
15
8
2.1
16%
DW
55
DCE
15
15
2 2
6%
pH 7
55
DCE
15
12
2 3
9%
Seawater
80
DCE
14
8
2.3
12%
pH 4
80
DCE
14
8
2.3
9%
pH 7
80
DCE
14
8
2 3
14%
pH 8.5
65
DCE
6
13
.4
12%
Thio
27
TCE
274
14
3.0
4%
Nutrient
27
TCE
386
10
1 2
9%
pH 4
27
TCE
386
11
1.2
10%
pH 7
27
TCE
386
10
1.2
8%
pH 8.5
37
TCE
386
16
1.7
12%
pH 4
37
TCE
386
16
1.7
9%
pH 7
37
TCE
386
16
1 7
16%
pH 8 5
80
TCE
14
9
1.7
8%
pH 4
80
TCE
14
8
1.7
9%
pH 7
80
TCE
14
10
1.7
9%
pH 8.5
DW Distilled organic free water
Thio 1 M Sodium thiosulfate solution


116
into the aqueous phase. The Kfw measured for MTBE of 15.7
was closer to the solvent/water partition coefficient for
the hexane/water (14.9-15.5) system than for the
benzene/water (23.3) system which suggests the fuel behaved
more like hexane than benzene in describing the partitioning
of MTBE.
Since the water solubility of the oxygenated additives
was so high relative to other fuel constituents, the
possibility of a cosolvent effect existed. This effect
would increase the solubility or partitioning of other fuel
constituents into the aqueous phase. The concentrations of
benzene and toluene measured in the aqueous phase in the
presence of these alcohol additives showed typical
analytical variations in concentration and were not enhanced
at higher percentages of additives (Figure 31).
The solubility of a solute in an aqueous cosolvent
mixture is a function of the mole fraction of the cosolvent
(oxygenated additive) in the aqueous phase. The mole
fraction of oxygenated compound in my extraction experiments
was a function of the volume percent of the additive in the
fuel, and the fuel to water ratio used in the extraction.
The fuel to water extraction ratio was less critical for the
equilibrium concentrations of the less soluble hydrocarbon
components since negligible amounts partition into the water
and a large excess of fuel was present. The addition of
oxygenated compounds to gasoline at a regulated maximum


25
polarity has less effect on the reaction rate than is
observed for SN1 reactions, but the rate is more sensitive
to changes in concentration or strength of nucleophiles.
The E2 reaction occurs when base attacks the hydrogen
at the carbon adjacent to the carbon containing the leaving
group (beta carbon). This reaction occurs at higher pH and
is more rapid for molecules containing a more acidic
hydrogen.
Degradation of 1 1 1-Trichloroethane (TCA)
The abiotic degradation of TCA was the subject of my
master's thesis (Cline, 1987) which included a detailed
discussion of related degradation studies and illustrations
of the first order decay of TCA in aqueous solution.
Additional data were collected subsequent to those studies.
This included additional concentration measurements in long
term degradation studies and measurements of rate
coefficients in additional matrices. In this section, a
concise comprehensive summary of these data are presented.
A brief synopsis of previous degradation studies of TCA
which have been reported in the literature is summarized
here. Dilling et al. (1975) performed reactivity studies on
selected chlorinated solvents, including TCA. Estimated
rate coefficients were based on four measurements over a
period of one year for each of two sets of reaction ampules;
one set was maintained in the laboratory and a second set
kept outdoors in Midland, Michigan. The same estimated rate


104
applied to various petroleum problems. Jet fuels (JP-4 or
Jet A fuel) were distiguishable by GC/FID, even if the fuels
were partially weathered (Roberts and Thomas, 1986).
Characterization of oil spills (Flanigan and Frame, 1977)
was successful using a nitrogen-sensitive detector rather
than with the FID detector.
Jones et al. (1983) reported that aromatic constituents
in crude oil aerobically degrade more quickly than the
normal alkanes, resulting in formation of an unresolved
complex mixture in the GC/MS scan of the aromatic fractions.
The latter could act as a marker for environmental
contamination caused by crude oil leakage or spills.
Dynes and Burns (1987) showed it was possible to detect
and identify petrol burned on cotton wool and weathered for
12 days using by GC with a Hall electrolytic conductivity
detector to obtain the sulfur chromatograms. The GC-FID
interpretations of these samples were inconclusive in
distinguishing the type of fuel product.
Gasoline, kerosene, and heavier oils represent
different boiling ranges of petroleum fuels. For example,
gasoline generally contains constituents having carbon
numbers less than C9, while diesel fuel ranges from Cn to
C20* These differences in composition are easily
distinguishable by GC-FID. The gas chromatogram of fuel for
different grades of gasolines can also show distinguishing
characteristics since higher grades frequently contain


P3
Figure 41. Bivariate plot of area counts of peaks P3 and P4 for aqueous extractions
of 65 gasoline samples.
147


SOLUBILIZATION AND DEGRADATION OF RESIDUAL TCA
A computational model was constructed to describe the
attentuation of TCA beneath the water table in the presence
of multiple phases. This simplified scenario for a TCA
spill considered the chemical transformation of TCA to 1,1-
DCE along with advective transport resulting from ground
water flow, of TCA and 1,1-DCE out of this zone containing
the residual solvent. Biodegradation of TCA in this highly
contaminated zone was considered negligible.
The major objective in developing this model was to
describe the relative concentrations of the major
constituents and how their concentrations may change with
time. These trends are illustrated for various ground water
flow rates, change in initial concentrations, and initial
composition.
Behavior of Residual Solvent
The migration pattern of chlorinated hydrocarbons
following a spill is illustrated in Figure 18. These dense
nonaqueous phase liquids (NAPL) will infiltrate the porous
media, with some of the NAPL retained in residual
concentration. The retention capacity for these NAPL in the
unsaturated zone may range from 5 L m"^ (approximately 12
mL/L of pore space) in highly permeable media to 30-50 L mJ
75


38
natural log of the concentrations versus time. All rate
constants were based on reactions showing a minimum of 75%
degradation.
The results of the degradation of TBA, DBCA and BDCA at
65C are illustrated in Figure 8. The differences in
slopes for the degradation of these compounds were not
statistically significant indicating that the rate
determining step was similar for each compound.
The formation of products (Figure 9) was calculated as
discussed previously for the formation of 1,1-DCE. The
percent elimination (ke/k) was the slope of the regression
line divided by the initial concentration of the parent
product. The smaller slope for 1,1-DCE, and its lower
maximum concentration, was a function of both lower initial
concentration of reactant (BDCA) and lower percent of BDCA
degradation which occurred through the elimination pathway.
The Arrhenius plot for TCA as determined in this study
is compared in Figure 10 with that of the brominated
compounds, TBA and DBCA. The Arrhenius plot for the two
brominated compounds was represented by a single regression
line. The regression line for TCA was essentially parallel
to that of the brominated compounds. The Arrhenius
activation energy (E^) for all of these compounds was almost
identical, since E^ is a function of the slope of this line.
The rate of degradation of TCA at 25C was
approximately a factor of 11 to 13 times slower than for the


54
aqueous solution. A broader perspective on hydrolysis /
elimination reactions can be obtained by comparisons with
other haloalkanes reported in the literature. The
obj ectives are
1. To compare degradation rates measured for
trihaloethanes of other simple alkyl halides which react by
an SN1/E1 mechanism.
2. To compare degradation rates of trihaloethanes with
other geminal trihalides reported in the literature to
determine structure/activity relationships with changes in
the substituents on the beta carbon, and describe shifts in
mechanisms which may occur for these trihalides.
3. To compare degradation rates and pathways of 1-
chloropropane and 1,1-dichloroethane with TCA to show the
effects of increasing number of chlorines on the alpha
carbon.
The classic reaction mechanisms for substitution and
elimination reactions are SN1, SN2, El and E2, as previously
discussed. The presence of various functional groups can
effect the rate and pathway of degradation of an alkyl
halide. For example, rates of hydrolysis are greater for
alkyl halides containing Br rather than for Cl by a factor
of 5 to 10. The rates also increase as the alkyl group goes
from primary to secondary to tertiary in the ratio of
1:10:1000 for chloride. Allyl groups enhance the rate of
hydrolysis of a primary halide by a factor of 5 to 100,


APPENDIX A
SOLUBILITY MEASUREMENTS BY LINDA LEE


5
4-
3-
CM
CL
2-
1 -
0-
14 16 18 20 22 24 26 28 30
P3
Phillips Gasoline
Regular
+ Regular Unleaded
-i 1 1 1 1 1 1 1 1 1 1 1 1 1 r
in
CL
4 i
3
2-
1 -
0
Union Gasolines
* Regular
+ Super Premium
Regular Unleaded
i 1 1 1 1 1 1 1 1 1 1 1 1
4 8 12 16 20 24 28 32
P3
Figure 42. Bivariate plots based on peaks selected using
stepwise discriminant procedure to distinguish grades of
Phillips or Union gasolines.


99
expected solubility of a component in a water cosolvent
mixture to be
log Sm = f log Sc + (1-f) log Sw
where Sm is the solubility in the water cosolvent mixture, f
is the volume fraction of cosolvent, Sc is the solubility in
neat cosolvent, and Sw is the solubility in water.
Groves (1988) reported there was no cosolvent effect
for MTBE concentrations up to mole fractions of nearly 0.2
in the hydrocarbon phase, while methanol and ethanol had an
effect at sufficiently high concentrations of alcohol.
Prausnitz et al. (1980) stated that liquid-liquid
equilibria were much more sensitive than vapor-liquid
equilibria to small changes in the effect of composition on
activity coefficients. Therefore, calculations for liquid-
liquid equilibria should be based, whenever possible, at
least in part, on experimental liquid-liquid equilibrium
data. Calculations become increasingly difficult for larger
numbers of components.
The difficulties in applying the above partitioning
calculations to commercial gasoline mixtures include the
large number of components involved and the problem of
determining the moles of each in theoretically infinite
combinations. One approach has been the estimation of
fuel/water partition coefficients which are based on weight
percent or concentration in the fuel rather than mole
fraction.


141
Table 23. Pearson Correlation Coefficients
N -
P 3
P4
P5
P3
1.00
P4
0.739
1.00
P 5
0.360
0.324
1.00
P6
-0.327
-0.397
-0.227
P7
-0.120
-0.120
-0.007
P8
-0.080
-0.299
0.133
P9
-0.181
-0.448
-0.025
P10
-0.219
-0.467
-0.011
Pll
0.077
-0.031
0.242
P12
0.281
0.157
0.118
PI 3
0.235
0.023
0.166
P14
0.241
0.057
0.229
P15
0.145
-0.004
0.173
P16
0.239
-0.000
0.117
P17
0.081
-0.007
0.319
PI 8
0.206
0.032
0.293
P19
0.408
0.262
0.478
P20
0.239
0.123
0.327
P21
0.179
0.125
0.026
P22
0.103
-0.002
0.166
P23
0.069
-0.007
-0.007
P10
Pll
P12
P10
1.00
Pll
0.147
1.00
P12
0.012
0.503
1.00
PI 3
0.522
0.303
0.503
P14
0.199
0.706
0.681
P15
0.288
0.557
0.769
PI 6
0.459
0.498
0.693
P17
0.225
0.423
0.086
Pi 8
0.369
0.683
0.671
P1 9
0.127
0.630
0.408
P 2 0
0.148
0.625
0.362
P21
-0.087
0.348
0.078
P22
0.056
0.462
0.069
P23
-0.226
0.218
0.091
PI 7
PI 8
P19
P17
1.00
P18
0.490
1.00
P19
0.554
0.650
1.00
P20
0.757
0.735
0.718
P21
0.280
0.267
0.229
P22
0 700
0.446
0.439
P 2 3
0.089
0.028
-0.073
65
P6
P7
P8
P9
1.00
-0.011
1.00
0.258
0.266
1.00
0.292
0.250
0.825
1.00
0.493
0.198
0.783
0.899
-0.140
0.056
0.256
0.205
0.094
-0.037
0.013
-0.052
0.256
0.233
0.310
0.340
0.038
0.146
0.154
0.149
0.257
0.178
0.191
0.150
0.191
0.104
0.317
0.346
-0.068
0.063
0.318
0.237
0.085
0.047
0.329
0.268
-0.212
-0.100
0.324
0.142
-0.147
0.037
0.280
0.155
-0.156
0.209
-0.059
-0.051
-0.178
0.214
0.136
0.079
-0.126
0.198
-0.181
-0.174
P13
P14
PI 5
P16
1.00
0.586
1.00
0.727
0.782
1.00
0.870
0.670
0.751
1.00
0.285
0.125
0.233
0.386
0.709
0.709
0.757
0.817
0.397
0.413
0.415
0.560
0.387
0.460
0.404
0.558
0.339
0.376
0.331
0.255
0.332
0.341
0.267
0.377
0.081
0.313
0.299
0.047
P20
P21
P 2 2
P23
1.00
0.262
1.00
0.673
0.487
1.00
0.058
0.487
0.541 1.00


96
thermal and catalytic cracking, reforming, isomerization).
Gasoline may also include a number of additives (dyes,
antiknock agents, lead scavengers, anti-oxidants, metal
deactivators, corrosion inhibitors, and volatility/octane
enhancers) (Lane, 1977; Youngless et al 1985).
The approximate volume percent composition is given in
Table 13. Unleaded gasoline generally has a higher fraction
of aromatic hydrocarbons than leaded brands. These "lead-
free brands contain no more than 0.05 gram of lead per
gallon. The use of tetraethyl lead as an antiknock agent
has been phased-out for environmental and health reasons.
Table 13. Volume Percent Composition of Gasoline
(Watts, 1986)
Compound
Unleaded
Leaded
Normal/iso hydrocarbons
55%
59%
isopentane
9-11%
9-11%
n-butane
4-5%
4-5%
n-pentane
2.6-2.7%
2.6-2 7%
Aromatic Hydrocarbons
34%
26%
Xylenes
6-7%
6-7%
Toluene
6-7%
6-7%
Ethylbenzene
5%
5%
Benzene
2-5%
2 5%
Naphthalene
0.2-0.5%
0.2 -0.5%
Benzo(b)fluoranthene
3.9 mg/1
3.9 mg/1
Anthracene
1.8 mg/1
1.8 mg/1
Olefins
5%
10%
Cyclic Hydrocarbons
5%
5%
Additives
Tetraethyl lead
-
600 mg/1
Tetramethyl lead
-
5 mg/1
Dichloroe thane
-
210 mg/1
Dibromoe thane
-
190 mg/1


126
new concentration of the component in the fuel layer. The
equations were
cw = Cf(initial) / Kfw
Cf(after extraction) = (Cf(initial)*Vf Cw*Vw)/Vf2
where Vf2 was the new fuel volume adjusted for loss of all
major components. Assuming the volume of the fuel was not
decreased significantly after contacting with water the
relative decrease in concentration after n extractions is
Cw,n/Cw,o Cf>n/Cf/o [1/(1 4- Vw/(Vf*Kfw))]n
The results of these calculations are illustrated in
graphic form in Figure 33. The shape of the plot of the
theoretical decrease in concentration in fuel or water was a
function of the partition coefficient. For highly soluble
components like MTBE or benzene, an exponential decay curve
was projected. Compounds with a high Kfw partition into
water at an essentially steady rate over the model range of
the theoretical 100 extractions because the concentration in
the fuel remains essentially constant.
For most of the gasolines evaluated in this study, the
initial aqueous extract concentration was lower for benzene
than toluene as illustrated for a sample of Shell SU2000
gasoline (Figure 34). The benzene concentration will
decrease more rapidly than the toluene concentration from
solubilization because of its lower Kfw, reaching
concentrations lower than for xylenes.


35
Brominated analogs of TCA were not commercially
available. Therefore, TBA was synthesized according to the
methods reported by Stengle and Taylor (1970). The
procedure for the synthesis of 1,1,1-tribromoethane (TBA)
produced a mixture of brominated analogs of TCA. The
primary components were TBA and 1,1-dibromo -1-chloroethane
(DBCA), while smaller quantities of l-bromo-1,1-
dichloroethane (BDCA) were present. Kinetic data for
abiotic degradation of TBA and DBCA were measured for
several temperatures while data for BDCA were obtained in
only selected experiments conducted at higher overall
concentrations. Compound structures are illustrated in
Figure 7. The elimination pathway involved loss of HBr to
form the corresponding alkene, the dominant elimination
product was the ethene formed by loss of a bromine. The
substitution pathway forms acetic acid.
Initial degradation experiments involving the
synthesized brominated mixture were conducted in reagent
grade (Milli-Q) water to obtain preliminary data on the
transformation process. Subsequent experiments were
conducted in buffer solutions at pH 4, 7, and 10. The
results of these experiments are summarized in Table 6.
First-order kinetics of degradation were observed, as
were also seen for TCA. Rate constants were calculated
from the linear regression analysis of the plots of the


26
was reported for each experiment, with half-lives of
approximately six months. Reaction products were not
measured.
The hydrolysis of TCA in seawater was reported by
Pearson and McConnell (1975). A half-life of 39 weeks (9
months) was estimated for TCA at 10C with the predominant
reaction being dehydrochlorination to 1,1-DCE. Walraevens
et al. (1974) examined the degradation of TCA in 0.5, 1.0
and 2.0 M sodium hydroxide solutions. The elimination
reaction was not observed, and sodium acetate was shown by
infrared analysis to be the sole reaction product. The
elimination product, 1,1-DCE, was assumed to be stable under
all experimental conditions.
Vogel and McCarty (1987) monitored the degradation of
TCA and formation of 1,1-DCE in water at pH 7 and a
temperature of 20C. The TCA half-life at 20C was
estimated to be between 2.8 and 19 years. Haag and Mill
(1988) report approximately 22% conversion of TCA to the
elimination product, with an extrapolated half-life of 350
days (11.5 months) at 25C.
Degradation experiments were performed at various
temperatures and in different sample matrices. The results
of these experiments are summarized in Table 4. First order
degradation kinetics were observed (Figure 5) in the
data as verified by plotting In [TCA] versus time. Linear
regression analyses were performed on each data set. All


11
study in spite of differences which exist in the age of the
spills and various physical and biological factors. The
time component for the weathering of gasoline at the source
is dependent on many site-specific factors. Even the
relative contributions of volatilization and solubilization
will depend on conditions like the depth of the water table
at the time of the spill and subsequent water table
fluctuations.
A simplification of the complex problem of determining
patterns of gasoline constituent concentrations following a
spill is to initially focus on the partitioning of gasoline
components from the fuel to water. This allows estimations
of equilibrium concentrations of different components from a
fresh spill in contact with water. Different brands and
grades of gasolines may then be evaluated to determine if
differences among the source types are distinguishable, and
how differences in composition affect the partitioning
behavior.
The major objectives of the gasoline study include
determination of the variability in the fuel/water partition
coefficients for aromatic constituents. Factors which may
affect the partitioning (concentration, cosolvents) will be
evaluated. Chemometric analyses on hydrocarbon components
present in the aqueous solution in equilibrium with gasoline
will be performed to evaluate similarities and differences
in various brands and grades of gasolines.


36
H Br
HC-CBr
H Br
Hx /Br
C = C
xBr
1,1,1 Tribromoethane (TBA)
1,1 Dibromoethene (DBE)
H Br
HC-CBr
H Cl
Hx /Br
C = C
XCI
1,1 Dibromo-1 chloroethane (DBCA) 1 Bromo-1 -chloroethene (BCE)
H Br
HC-CCI
H Cl
Hx /Cl
C = C
H7 XCI
1 -Bromo1,1 -dichloroethane (BDCA)
1,1-Dichloroethene (DCE)
Figure 7. Brominated analogs of 1,1,1 -trichloroethane and
corresponding elimination products. Since bromine is a
better leaving group than chlorine, the predominant pathway
is elimination of HBr.


88
converted to 1,1-DCE, the effect of the accumulation is not
observed until substantial degradation has occurred. If all
the TCA degraded in this closed system, 20 mmoles of 1,1-DCE
would be produced, which is 60% of the pure component
aqueous solubility of 1,1-DCE. Therefore, for the initial
conditions of the model, a residual NAPL will exist only
when excess TCA is present.
A comparison of different initial conditions for a
constant flow (0.25%) is shown in Figure 22. With an
increase in amount of residual TCA, the same zero-order
decay rate is observed, indicating that doubling the amount
of TCA in the solvent phase doubles the time needed for
removal of the residual.
In addition, Figure 22 illustrates the rate of loss of
TCA when the initial 100 mmoles is mixed with another
solvent, a hypothetical mixture in which the mole fraction
of the "inert" compound remains at 0.5 in the solvent phase.
This represents a case where a compound with solubility
similar to TCA (like TCE) is present in the residual. The
presence of this other compound causes a 50% reduction in
the aqueous phase concentration of TCA, and therefore the
rate of loss of TCA, doubling the time to remove the TCA
from the residual phase.
The patterns of change in mass of 1,1-DCE in the
solvent or aqueous phase over time are more complex when
there is advection from the system Figure 23. The aqueous


71
Similar trends were observed for both compounds (Table
11). The slowest rates relative to water were obtained for
both compounds in the sample containing clay, while the
fastest rates were observed in the sand.
The data generally showed greater variability in the
samples containing the solids as compared to the DW system
(Figures 16 and 17) as evidenced by correlation coefficients
less than 0.99. However, the rates of 1-chloropropane
degradation in ampules containing solids differed by less
than 10% of the rate obtained for Milli-Q water.
The relative degradation rates for TCA differed more as
a function of matrix than observed for chloropropane,
however, there was also greater variability as evidenced by
the correlation coefficients. In the case of TCA, the
formation of 1,1-DCE was similar in all matrices suggesting
the ratio of products was not affected by the presence of
the s e solids.
The relatively small differences in rates measured in
these matrices may be due to a variety of factors including
sorption, however significant surface catalysis was not
observed. For this type of saturated system, the amount of
alkyl halide in contact with the surface would be small.
Differences may be attributed to normal variability and
differences in ionic strength or composition of the aqueous
phase in contact with the solids.


REFERENCES
Banerjee, S., Solubility of Organic Mixtures,
Environmental Science and Technology. 18(8): pp 587- 591,
1984 .
Bentley, T. W., and Schleyer, P. von R., Medium Effects
on the Rates and Mechanisms of Solvolytic Reactions, in
Advances in Physical Organic Chemistry. edited by Gold, V.
and Bethell, D. Academic Press, New York, pp. 1-67, 1977.
Bossert, I., and Bartha, R., The Fate of Petroleum in
Soil Ecosystems, in Petroleum Microbiology, edited by Atlas,
R. M., Macmillan Publishing Co., New York, 1984.
Bouwer, E.J., and McCarty, P.L., Transformations of 1-
and 2-Carbon Compounds under Methanogenic Conditions,
Applied and Environmental Microbiology. 45: pp. 1286-1294,
1983 .
Brookman, G. T., Flanagan, M. and Kebe, J. 0. Literature
Survey: Hydrocarbon Solubilities and Attenuation Mechanisms,
American Petroleum Institute Report, pp. 1-101, 1985a.
Brookman, G. T., Flanagan, M. and Kebe, J. 0. Laboratory
Study on Solubilities of Petroleum Hydrocarbons in
Groundwater. TRC Project Report No. 2663- N31-00. TRC
Environmental Consultants, Inc., East Hartford, CT., 1985b.
Carey, F. A., and Sundberg, R. J., Advanced Organic
Chemistry: Part A: Structure and Mechanisms. Plenum Press,
New York, 1984.
Clark, H. A., and Jurs, P. C., Qualitative Determination
of Petroleum Sample Type from Gas Chromatograms Using
Pattern Recognition Techniques, Analytical Chemistry.
47(3): pp. 374-378, 1975.
Cline, P.V., Abiotic Degradation of 1.1.1-
Trichloroethane: Formation of 1,1-Dichloroethene. M. S.
Thesis, University of Florida, 1987.
Cline, P.V., Delfino, J.J., and Cooper, W.J., Hydrolysis
of 1,1,1-Trichloroethane; Formation of 1,1-Dichloroethene
in Proceedings of NWWA/API Conference on Petroleum
187


53
incubated at 20C in a solution at pH 12.5. No degradation
was observed during four months of incubation (Table 7).
The experiments demonstrate the resistance of the
halogenated ethenes to degradation in dilute aqueous
solution. Reports of the degradation of these compounds
with half-lives of less than 1 year appear to represent a
process other than abiotic degradation in water. In the
same way as billing et al. (1975), my experiments were
conducted in sealed ampules containing a headspace, however
degradation was not observed as reported in their study. I
believe their results may be a result of analytical error.
The half-lives for each experiment were based on four
measurements. The results showed a chemically diverse group
of compounds had similar decreases in concentration and
temperature had little effect on these decreases. A
possible explaination for these results would be a decrease
in instrument response over the year of the study.
The halogenated ethenes generally showed very little
degradation, with the exception of the rapid degradation of
TCE at high pH and temperature. It appears that any
degradation of these compounds in aqueous solution which
occurs, does so under rather extreme conditions and is not
expected to be a dominant process.
Struc ture/Rate Relationships of Alkyl Halides
In the previous sections degradation patterns and
kinetics were evaluated for various 1,1,1-trihaloethanes in


84
day. These values include the "no flow" or "low flow" (0.1%
per day) cases, in which the dominant loss occurs through
degradation. At 0.25% per day, the rate of advection is
comparable to the rate of degradation. Finally, a flow rate
of 0.5% per day represented the case in which the loss of
TCA is primarily due to advection. At flows greater than
0.5% per day the losses would be dominated by the advective
term. These volume exchange rates represent slow flows
and/or very large spill areas. An exchange of 0.5% per day
represents an approximate flow through 5 meters of
contaminated porous media at a rate of 2.5 cm/day.
Degradation Rate
The solubility of TCA affects not only its rate of
advection from the contaminated zone, but also the total
mass of TCA degraded per unit time. The first-order rate
constant at 25C is approximately 0.00226 day'^ as measured
in this study. In a contaminant plume, the half-life for
the degradation of TCA is approximately 10.2 months.
Although the first-order rate coefficient remains constant,
the mass of TCA converted per unit time decreases as the
concentration of TCA in the aqueous phase decreases.
In the model, it was assumed that the TCA concentration
remained at saturation within the zone containing residual
solvent since the TCA that degraded was replaced by
dissolution of the residual solvent. The amount of TCA
degraded per unit time follows zero-order kinetics. The


91
Figure 24. Model results: Increase in aqueous
concentration of 1,1-DCE forming from degradation of TCA as
a function of flow.
a>
C/5
o
-C
a.
c
<1>
>
O
to
c/5
_cu
o
E
E
CJ
Q
Figure 25. Model results: Pattern of accumulation of 1,1-
DCE in the solvent phase as TCA degrades.


153
which of two possible sources was more likely responsible
for a petroleum fuel spill.
There were two ways in which the output from this
procedure could be used to clarify the differences seen
among various samples. The first and simplest option exists
when only two or three peaks were selected by the SAS
procedure, or if those first two or three peaks explain most
of the variation between the classes. Simple plots of the
two or three peaks should provide visual assistance in
seeing the differences represented in the samples.
To illustrate, only two peaks were selected to
distinguish between two grades of Phillips gasoline (P3 and
P12) or among three grades of Union gasoline (P3 and P15).
Simple bivariate plots of these peaks reveal the separation
of grades (Figures 42).
Although graphic representations of two or three peaks
are possible, it is not easy to graphically represent the
difference that exist when a larger number of peaks need to
be included in the analysis. Therefore, one additional
statistical tool, principal component analysis, was used to
provide insight into the differences in the peaks of the
GC/FID chromatograms used in this analysis. The information
from the stepwise discriminant analysis determines the peaks
to include in the principal component analysis and the ASCC
following the selection of the final peak indicates how
successful the separation may be.


156
indicates that the peaks selected will explain only about
32% of the variation in the data. Principal component
analysis will not improve the separation of gasoline types,
but simply provides a visual presentation of the
differences.
Principal component analyses were performed on various
subsets of the data. Each grade (regular unleaded, super
regular and super premium) was evaluated separately (Figures
43 to 45). The ASCC values for the discriminant analyses
for regular and super premium grades were about 0.8 and the
plots reveal incomplete separation. Amoco samples,
particularly for the super premium grades, were frequently
separated from other brands. Shell super premium gasolines
showed more clustering (similarities) than other brands.
Since there were only two super regular gasoline
brands, Amoco and Shell, and the discriminant ASCC value was
one, these two sample types were separated in the principal
component plot.
These principal component plots do not reveal adequate
separation of all brands included in the analysis. Specific
pairs of brands were examined to determine if they could be
separated in a principal component plot. The data subsets
included all grades of each the two gasolines brands. This
type of analysis, if successful, could be used in a specific
spill situation where two possible sources are present.


125
Table 19. Multiple Aqueous Extractions
of Gasoline (mg/L)
Compound
1
Extract Number
2
3
MTBE
69
36
18
Benzene
43
40
35
Toluene
57
55
51
Ethylbenzene
3.4
3.4
3
m,p,-Xylene
12
12
12
o-Xylene
6.4
6 3
6
Total Hydrocarbons
198
168
150


100
The concentration of a component present in the aqueous
phase has been estimated based on the solubility and weight
percentage in the gasoline. This approach was examined in a
recent laboratory study (Brookman et al., 1985b) on the
solubility of petroleum hydrocarbons in groundwater, where a
reference regular unleaded gasoline (API PS-6) was
equilibrated with organic free, deionized water.
The partitioning of components into water is affected
by the solubility of each compound in pure water and the
gasoline composition. The partitioning of fuel oil
components can be described using a partitioning coefficient
based on the following equation (Brookman et al. 1985b):
Kfw Cf/Cw
where Kfw fuel/water partition coefficient
Cf concentration of component in fuel (g/1)
Cw concentration of component in water (g/1)
The concentration of a particular component in the fuel
was based on the area percent determined in the analysis
which was assumed to approximate the weight percent of that
compound. This was converted to concentration (g/1) based
on an average density for gasoline of 0.74 g/ml (Brookman et
al. 1985a).
There are greater sources of variation in values of
fuel/water partition coefficients to describe gasoline
component partitioning than found for measurements of
octanol/water partition coefficients (Kow). Kow
i s the


10
downgradient from the spill occur as a result of transport
from the source, and therefore show higher concentrations of
the more mobile constituents.
The downgradient aqueous concentrations are dependent
on the initial partitioning of the gasoline components into
water at the source. The presence of the residual
hydrocarbon will dominate the partitioning process, with
soils playing an increasing role as the residual hydrocarbon
is depleted. Field data are complex to interpret. This is
due to many factors, including site heterogeneities, well
construction and sampling variables, and lack of detailed
information which can provide estimates of the rates of
partitioning and transport. However, patterns resulting
from physical processes, i.e. partitioning and transport,
may be observed. In Table 2 are summarized the highest
concentrations of BTX components measured in monitoring
wells at various gasoline spill/leak sites in Florida.
Table 2. Maximum concentrations (mg/L) of BTX components
in monitoring wells at selected gasoline
contamination sites in Florida.
County
Benzene
Toluene
Xylenes
Hi11sbo rough
24
64
16
11
46
15
Volusia
10
28
11
8
46
9
Desoto
0.8
60
9
These concentrations are similar to those measured in
laboratory gasoline-water partitioning experiments in this


47
incubation with moderately alkaline material such as
concrete (Greim et al., 1984). They concluded
dehydrohalogenation can occur under these relatively mild
conditions resulting in toxicity from exposure to the
dichloroacetylene.
Many substitution and addition reactions of TCE have
been carried out in the presence of base. What initially
appeared to be a direct substitution reaction may in fact
have been multistep processes involving intermediates like
carbanions, chloroacetylenes, or carbenes. Rappaport (1969)
reviewed the mechanisms for nucleophilic vinylic
substitution processes in alkaline solutions at elevated
temperatures.
Mechanisms may differ for chemical studies performed
under extreme conditions of temperature and high pH compared
to reactions occurring under more typical environmental
conditions. The possibility of slow nucleophilic attack in
aqueous solution was considered because March (1985) reports
that although vinyl halides are generally considered
resistant to nucleophilic attack, the presence of electron-
withdrawing groups like halogen lower the electron density
of the double bond enhancing nucleophilic substitution or
addition reactions.
In ground water, even very slow degradation may be an
important attenuation mechanism. Since environmental
studies report slow degradation of TCE or PCE in water and


Area counts for peaks used in statistical analysis of water
extracts of gasoline.
Brand
Grade
Date Symbol
P3
P4
P5
MTBE
1
Amoco
RU
12/15/86
A
27.83
23.82
256.14
2
Amoco
RU
1/05/87
A
25.82
24.70
713.67
3
Amoco
RU
2/04/87
A
31.12
30.01
614.49
4
Amoco
RU
2/04/87
A
33.26
31.39
606.22
5
Amoco
RU
2/04/87
A
30.91
27 68
587.94
6
Amoco
SR
1/05/87
B
35.28
37.81
729.93
7
Amo c o
SR
2/04/87
B
34.44
35.62
1315.03
8
Amo c o
SR
2/04/87
B
35.10
35.42
1236.45
9
Amoco
SP
12/15/86
C
15.19
28 39
21.56
10
Amoco
SP
12/15/86
C
13.83
27.06
20.31
11
Amoco
SP
1/05/87
C
25.92
26.81
37.10
12
Amoco
SP
2/04/87
C
30.04
34.70
55.08
13
Chevron
R
1/05/87
D
10.84
21.47
31.40
14
Chevron
R
1/05/87
D
17.26
25.34
33.92
15
Chevron
R
2/04/87
D
17.51
17.20
12.75
16
Chevron
SP
1/05/87
E
11.78
21.76
12.59
17
Chevron
SP
2/04/87
E
16.08
17.43
34.71
18
Chevron
RU
1/05/87
F
32.19
29.28
26.67
19
Chevron
RU
2/04/87
F
39.32
28.35
19.90
20
Chevron
RU
2/04/87
F
33.31
27.63
20.92
21
Gulf
SP
12/15/86
G
3.05
5 30
17.92
22
Gulf
SP
10/20/86
G
14.26
12.90
76.92
23
Gulf
SP
10/20/86
G
15.23
12.98
56.43
24
Gulf
RU
10/20/86
H
39.48
45.01
12.10
25
Gulf
RU
10/20/86
H
36.30
39.61
13.24
26
Gulf
RU
12/15/86
H
17.40
36.41
189.08
27
Phillips
R
10/20/86
I
29.79
31.09
57.71
28
Phillips
R
11/09/86
I
27.06
30.14
31.79
29
Phillips
RU
10/20/86
J
15.79
27.93
16.97
30
Phillips
RU
11/09/86
J
14.50
11 17
8.27
31
Phillips
RU
12/15/86
J
18.47
28.10
20.83
32
Phillips
RU
12/15/86
J
16.78
26.91
20.47
33
Shell
R
11/09/86
K
8.42
6 96
2.81
34
Shell
R
11/09/86
K
8.83
7.09
2.93
35
Shell
R
12/15/86
K
26.20
7.00
10.47
36
Shell
RU
11/09/86
L
10.42
9 11
8.37
37
Shell
RU
12/15/86
L
32.45
25.80
13.83
38
Shell
RU
12/15/86
L
32.96
26.90
4.35
39
Shell
RU
1/05/87
L
21.42
20.85
11.35
40
Shell
RU
2/04/87
L
21 79
21.31
20.01
41
Shell
RU
2/04/87
L
18.51
19.22
29.03
42
Shell
RU
2/04/87
L
20.54
20.91
21.41
43
Shell
SR
1/05/87
M
27.06
20.12
6 .82
44
Shell
SR
2/04/87
M
14.52
18.38
36.14
45
Shell
SR
2/04/87
M
15.98
14.59
16.25
180


68
have considerable contact with a variety of aquifer
materials which could potentially affect degradation rate.
Hydrolysis reactions may be affected by factors like acid or
base catalysis, sorption and ionic strength. Since
compounds which react by different mechanisms may be
impacted differently by these solid surfaces, both 1-
chloropropane and TCA were used in degradation experiments
performed in various matrices.
Catalysis of hydrolysis or elimination reactions of
alkyl halides by saturated aquifer materials has not been
demonstrated. Because high concentrations of 1,1-DCE have
been observed in Florida and Arizona at solvent spill sites
contaminated with TCA, the role of sand or other materials
which may influence the degradation of TCA was evaluated.
The nonbiological degradation of pesticides in the
unsaturated zone was shown to play an important role for a
few groups of pesticides, mainly organophosphates and s-
triazines. Clay mineral surfaces have shown catalytic
activity, correlated to their acid strength. This catalytic
process is most important at low moisture content, and
therefore is more important in the vadose zone than beneath
the water table (Saltzman and Mingelgrin, 1984).
Haag and Mill (1988) did not observe significant
differences in the kinetics or products of TCA in contact
with sediment pore water. Epoxide hydrolysis was


135
area percentages of aromatic constituents in the fuel would
result in higher aqueous phase concentrations of those
components. Qualitative descriptions of similarities and
differences were not able to simultaneously consider
fluctuations for any single brand and grade.
Preparation of the Data Base for Statistical Analysis
The pattern recognition approach to data analysis
involved a series of steps that attempted to identify
patterns in the experimental data. First, specific peaks
in the chromatograms of the water extracts were selected for
inclusion in the data base. These data were then evaluated
using various statistical procedures.
Twenty-three peaks with the highest area counts and
most frequent appearance in the water extracts of the 65
gasoline samples, were selected from the chromatograms. The
peaks were identified by a peak number, which represented a
particular retention time for the chromatographic elution.
A summary of the peaks, retention times and identification
where available, is shown in Table 21.
The statistical analysis did not require complete
separation nor identification of the peak. As stated
earlier, this type of data interpretation has been done
without prior knowledge of chromatogram peak identity for
either compound class or type (Hosenfeld and Bauer, 1985).
However, most of the peaks which are identified in Table 21
consisted predominantly of the component named. This was


129
A few Union gasolines had higher initial concentrations
of benzene as compared to toluene in the water extracts
(Figure 35). As the sample ages, the benzene concentration
would decrease more rapidly than the concentration of other
aromatic components. Since high equilibrium concentrations
of benzene in water were less commonly measured, and benzene
concentration would decrease rapidly, high concentrations of
benzene measured in water in equilibrium with residual
gasoline in field sites would represent relatively
unweathered gasoline and may suggest particular sources.
The Kfw was measured for many constituents like toluene
which were in high concentrations in the fuel and water.
The fuel/water partition coefficient should be reevaluated
for low solute concentrations which will develop as ageing
occurs. The values presented in this study provide a
reasonable estimate for many of these compounds over a
fairly wide range of concentrations (e.g. toluene
concentrations in water ranging from 23 to 133 mg/1). The
slight increase in Kfw as concentration decreased, was
primarily noted for the more soluble constituents, benzene
and MTBE, but may occur for other constituents.
Differences in Water Extracts of Gasolines
The objective of this section is to describe and
evaluate similarities and differences in water extracts of
gasolines in part by using selected routine statistical
procedures.


90
concentration of 1,1-DCE continues to increase for some time
as the mass of 1,1-DCE in the solvent phase begins to
decrease because its mole fraction continues to increase in
the solvent phase.
The total mass of 1,1-DCE in the zone of residual
contamination increased over time, reaching a maximum as the
TCA mass in the solvent phase approached zero. Increasing
the flow rate not only shortened the time in which 1,1-DCE
was accumulating, but decreased the maximum amount of 1,1-
DCE present in that zone. This is true for the aqueous
phase concentrations (Figure 24) and amount in the solvent
phase (Figure 25). The maximum concentration of 1,1-DCE in
the aqueous phase for a flow of 0.5% per day is
approximately 1 mmole/L (100 mg/L) at the point where some
residual phase is still present. The concentration of TCA
at that time is nearly at saturation (approximately 1500
mg/L).
The changes in aqueous concentration of 1,1-DCE for
larger amounts of TCA originally present or in the presence
of an inert solvent as previously discussed, are shown in
Figure 26. The changes in the amount of 1,1-DCE in the
solvent phase is shown in Figure 27. The inert solvent
increases partitioning into the organic phase, keeping the
aqueous concentration low.
The model illustrates factors which affect the time for
removal of a residual phase under varying conditions, and


Principal Component 2
4
3
2
1
0
-1
-2
-3
-4
-6 -4 -2 0 2 4 6
Principal Component 1
Figure 47, Principal component plot of all grades of Chevron and Union gasolines.
C Chevron IJ IJ
U Union y
(All Grades)
U u
o
o
u
C c
u
C u 1
J
c
u
c
c
T
162


39
n 1 1 i 1 1
0 2 4 6
Time (Hours)
Figure 8. First order kinetic data for the abiotic
degradation of TBA, DBCA and BDCA in water at 65C.
Figure 9. Formation of the elimination products (BCE, DBE,
DCE) in water at 65C from the abiotic degradation of the
corresponding 1,1,1-trihaloethanes.


191
Olah, G. A., Carbocatlons and Electrophilic Reactions.
John Wiley & Sons, New York, 1974.
Parsons, F., and Lage, G. B., Chlorinated Organics in
Simulated Groundwater Environments, Journal of the American
Water Works Association. 77(5): pp. 52- 59, 1985.
Pearson, C. R., and McConnell, G., Chlorinated Cl and C2
Hydrocarbons in the Marine Environment, Proceedings of the
Roval Society of London. B.. 189: pp. 305-332, 1975.
Pielou, E. C., The Interpretation of Ecological Data.
John Wiley & Sons, New York, 1984.
Prausnitz, J. M., Anderson, T. F., Grens, E. A., Eckert,
C. A., Hsieh, R., and O'Connell, J. P., Computer
Calculations for Multicomponent Vapor Liquid and Liquid-
Liquid Equilibria. Englewood Cliffs, New Jersey, Prentice-
Hall, Inc, 1980.
Queen, A., and Robertson, R. E., Heat Capacity of
Activation for the Hydrolysis of 2,2-Dihalopropanes.
Journal of the American Chemical Society. 88(7): pp. 1363-
1365, 1966.
Quemeneur, F., Bariou, B., and Kerfanto, M., Kinetic
Study of the Hydrolysis of p-Substituted alpha alpha
Dichloro and a,a,a-Trichlorotoluenes in a 50% Water-Acetone
Medium. Cometes Rendus de L'Academie des Sciences. Serie C.
272(5),497-9, 1971, in Chemical Abstracts. CA74(17):87067b.
Quemeneur, F., Bariou, B., and Kerfanto, M., Kinetics of
the Hydrolysis of Some p-Substituted alpha-Halo-, alpha
alpha -Dihalo- and alpha alpha alpha -Trihalotoluenes in
Variable Proportions of Water-Acetone Solvent. Comptes
Rendus de L'Academie des Sciences. Serie C. 278(4), 299-301,
1974, in Chemical Abstracts. CA81 (3) : 12761 j .
Rappaport, Z., Nucleophilic Vinylic Substitution.
Advances in Physical Organic Chemistry, edited by Gold, V.,
New York, Academic Press, Vol 7: pp. 1-111, 1969.
Roberts, P. V., Field Observations of Organic Contaminant
Behavior in the Palo Alto Baylands, Artificial Recharge of
Groundwater. edited by Asano, T., Butterworth Publishers,
New York, pp. 647-679, 1985.
Roberts, A. J., and Thomas, T. C., Characterization and
Evaluation of JP-4, Jet A and Mixtures of these Fuels in
Environmental Water Samples. Environmental Toxicology and
Chemistry. 5: pp 3 -11, 1986.


DCE (ug/l) In [JCA] ug/l
1 e kt
Figure 5. First order kinetic data for the degradation of
1,1,1-trichloroethane at 28C and pH 4.5, with the
corresponding data for the formation of the elimination
product, 1,1-dichloroethene.
28


[TCA] in water (mmoles/l) Total TCA (mmoles)
87
YEARS
20. Model results: Decrease in total TCA mass in
idual zone as a function of flow.
YEARS
Figure 21. Model results: Change in aqueous concentration
of TCA as a function of flow.


86
calculations are stopped when amounts of TCA in the residual
NAPL are less than 10 mmoles. At lower levels of residual
NAPL, the process may become diffusion limited as the NAPL
is trapped in regions of the soil matrix removed from the
aqueous flow. The results of the model are shown in Figures
20-26.
The total mass of TCA in the NAPL showed zero-order
decay with flows from 0.1-0.5% per day (Figure 20). As the
flow rate decreases, slight nonlinearity is observed. This
reflects the slow accumulation of 1,1-DCE in the solvent
phase which begins to decrease the aqueous concentration of
TCA .
The decrease in aqueous concentration of TCA (Figure
21) as the total mass of TCA in the system goes from 100
mmoles to approximately 15 mmoles (slightly in excess of the
solubility) is dependant on the flow. The larger decrease
is observed for the case of no-flow, which results in a 45%
decrease in the aqueous phase concentration after 10 years.
The major reason 1,1-DCE fails to accumulate
significantly in the solvent phase is its higher water
solubility. Having a solubility twice that of TCA, 1,1-DCE
is advected from the zone containing residual solvent more
readily. In the special case of no flow through the system,
1,1-DCE is not advected and begins to accumulate in the
solvent phase affecting the aqueous phase concentration of
TCA. However, since only approximately 20% of the TCA is


than TCA. As the number of bromines increase, the percent
of elimination products increases.
These geminal trihalides degrade by a unimolecular
mechanism (E1/SN1). The rate coefficient for TCA
degradation in buffered water at elevated temperature is
approximately six times greater than hydrolysis of 1-
chloropropane (SN2 mechanism) and more than 100 times
greater than 1,1 -dichloroethane. In the presence of sodium
thiosulfate, the 1-chloropropane degradation rate increased
by more than a factor of 100, 1,1-dichloroethane rate by 22
and TCA degradation by approximately two.
Halogenated ethenes are stable at various temperatures
and reaction conditions. Trichloroethene degrades in
alkaline solution at elevated temperature.
1,1,1-Trichloroe thane and 1,1-DCE form a near ideal
solution in the solvent phase. The solubility of 1,1-DCE at
24C is 3200 mg/1 and the solubility of TCA is approximately
1580 mg/1.
The range of concentrations for major components of
gasoline which partition into water was determined for 65
gasoline samples. Benzene concentrations in the water
extracts ranged from 12.3-130 mg/1 and toluene
concentrations ranged from 23-185 mg/1.
Fuel/water partition coefficients of seven major
aromatic constituents were measured for 31 gasoline types
and showed a standard deviation of 10-30%. These
vi


119
to water ratio and an equilibration time of 2 hours for
three samples and 96 hours for two samples in a study of API
PS-6 Reference gasoline. A comparison of Kfw 's for
selected constituents obtained by Brookman et al. (1985a)
with results obtained in this study are presented in Table
17 .
Table 17.
Comparison of Kfw
Kf w
Compound
This Study Brookman et al., 1985a
Benzene
350
Toluene
1250
Ethylbenzene
4480
m,p- Xylene
4360
0-Xylene
3630
245
1050
3440
3550
2430
The Kfw values reported by Brookman et al. (1985a) were
lower than my average result by 18 to 30%, but within two
standard deviations of my average value. Their results,
which indicated more partitioning into the water phase, may
be due to differences in analytical and extraction protocols
or gasoline composition.
The average fuel/water partition coefficients for
benzene, toluene, and ethylbenzene were approximately twice
the Kow's for those compounds as shown in Table 18.


ni.PVC-'TBP.'KCt.Me'i IHTI .4S-4S8,2 :88P .24M0V87 ,Ut10 HrHI 13S66, 31
SBS, 36M, 1UN,8#306S-2SG,5.2D,3Sft,SCRYO 874 SCOMS < 874 SC8HS, IS.88 MIHS1
1.8 | Mars RftMCCs 44.8, 269.8 TOTOL 8SUMB 3235689.
I I
*!!
ES
110
DEE
I JL
i TBA
16S 220 276 331 384
I
439
I
494
SSO60S
1
660
r"
7lS
778 825
OJC.)
u
ra/z
Figure 3. Total ion chromatogram (center) for
degraded geminal trihalide mixture, with mass
1,1,1-tribromoethane and 1,1-dibromoethene.
partially
spectra for


9
Aerobic biodegradation will be an important attenuation
mechanism provided that sufficient oxygen and nutrients are
present, and these components typically become limiting
after a spill or leak. Attempts to stimulate aerobic
biodegradation of underground petroleum need to remedy both
nutrient and oxygen deficiencies. In addition, hydrocarbons
in the C5-C9 range (which are typical of gasoline) have
relatively high solvent-type membrane toxicity which will
reduce the number of microorganisms and therefore, decrease
the amount of biodegradation following a gasoline spill
(Bossert and Bartha, 1984).
Sites which have been contaminated by gasoline spills
occasionally report results of the analysis of the "floating
layer." Recovery wells to remove the residual organic
liquid are typically installed as an early remediation
measure. Ground water is typically analyzed for benzene,
toluene, and the xylenes (BTX) and more recently for the
oxygenated gasoline additive methyl tertiary butyl ether
(MTBE).
Concentrations of the BTX or oxygenated constituents
will vary spatially and temporally. At the source, changes
in relative concentrations of hydrocarbon components occur
through weathering, primarily volatilization and
solubilization of the liquid residual organic constituents,
resulting in increasing concentrations of the least mobile
constituents. Compounds detected in ground water


150
Table 24. Selection of Major Discriminating Peaks
to Detect Differences in Water Extracts of Gasoline
SAS STEPWISE DISCRIMINANT ANALYSIS
65 OBSERVATIONS
17 CLASS LEVELS
SIGNIFICANCE
SIGNIFICANCE
21 VARIABLE(S) IN THE ANALYSIS
0 VARIABLE(S) WILL BE INCLUDED
LEVEL TO ENTER 0.1500
LEVEL TO STAY 0.1500
STEPWISE SELECTION: SUMMARY
VARIABLE
NUMBER
PARTIAL
F
PROB
STEP
ENTERED
IN
R**2
STATISTIC
F
1
P6
1
0.8359
15.284
0.0001
2
P 7
2
0.8336
14.715
0.0001
3
P8
3
0.7741
9.849
0.0001
4
P4
4
0.7442
8.181
0.0001
5
P5
5
0.6671
5.511
0.0001
6
P 3
6
0.5751
3.637
0.0004
7
P9
7
0.5053
2.682
0.0053
8
PI 3
8
0.4567
2.154
0.0244
9
P10
9
0.4829
2.335
0.0151
VARIABLE
NUMBER
WILKS'
PROB <
STEP
ENTERED
IN
LAMBDA
LAMBDA
1
P6
1
0.16407433
0.0001
2
P 7
2
0.02730376
0.0001
3
P8
3
0.00616909
0.0001
4
P4
4
0.00157832
0.0001
5
P5
5
0.00052543
0.0001
6
P3
6
0.00022327
0.0001
7
P9
7
0.00011044
0.0001
8
PI 3
8
0.00006000
0.0001
9
P10
9
0.00003103
0.0001
AVERAGE
SQUARED
VARIABLE
NUMBER
CANONICAL
PROB >
STEP
ENTERED
IN
CORRELATION
AS C C
1
P6
1
0.0522
0.0001
2
P7
2
0.1033
0.0001
3
P8
3
0.1480
0.0001
4
P4
4
0.1878
0.0001
5
P5
5
0.2286
0.0001
6
P3
6
0.2538
0.0001
7
P9
7
0.2710
0.0001
8
PI 3
8
0.2973
0.0001
9
P10
9
0.3217
0.0001


149
This analysis was performed on the entire data set (21
variables, 65 observations), as well as various subsets of
the data. An example of the summary table (Table 24) shows
the information provided during the analysis of the entire
data set when the procedure was separating data by "type,"
which was a specific brand and grade. Seventeen types
(class levels) of gasoline were included in the data set.
The peaks included by this procedure were the earliest
eluting, P3 through P10, and P13. Therefore, these peaks
provided the most information on the differences between all
the various types of gasolines. The average squared
canonical correlation (ASCC) is close to one if all groups
are well separated and if all or most directions in the
discriminant space show good separation for at least two
groups. For this analysis the ASCC was 0.32 after addition
of the final variable, which indicates relatively poor
separation of the groups.
Typically, the smaller the number of types of gasolines
to be distinguished, the fewer number of peaks were included
by the procedure and the greater the resolution obtained
among the gasoline types. While this approach cannot
provide definitive ways to distinguish among all brands and
grades of gasoline, it is possible that selections among
only a limited number of gasoline types may show clear
relationships.


41
brominated analogs. The observed rate constants for the
various pH values were not significantly different, as was
also observed for the degradation of TCA.
Experiments were performed to determine the reaction
mechanism for these 1,1,1-trihaloethanes. First, the
degradation experiment was conducted in a 1 molar sodium
thiosulfate solution. Sodium thiosulfate is a much stronger
nucleophile than water or hydroxide (Swain and Scott, 1953)
and a dramatic increase in degradation rate in this solution
is indicative of an SN2 reaction in which the nucleophile is
directly involved in the rate determining step. The
degradation rates measured for the brominated 1,1,1-
trihaloethanes increased less than a factor of 2 in the
thiosulfate solution, which may be attributed to the
increased ionic strength of the solution. The percent
elimination was also unaffected by this sample matrix.
To further characterize the mechanism for these
degradation reactions, the brominated geminal trihalides
were placed in a 1 M KCl solution at 37C. High ionic
strength solutions generally increase the rate of SN1 or El
reactions. When a common ion is present, the rate of the
reverse reaction is enhanced. In the presence of high
concentrations of chloride, chloride may be exchanged for
bromide when an ion pair forms. If BDCA forms an ion pair
and chloride is exchanged, TCA will be formed (Figure 11)
providing evidence of a carbocation intermediate. Even


62
Table 10. Half-lives for Abiotic Degradation of
Geminal Trihalides
COMPOUND
STRUCTURE HALF-LIFE REFERENCE
25C
Chloroform
Bromo form
CHCI3 3500 yr Mabey and Mill,
1978
CHBr3 690 yr Mabey and Mill,
1978
1,1,1-Trichloroethane CH3CCI3 10.2 mo This dissertation
1,1,1-Tribromoethane CH3CBr3
1 mo This dissertation
DDT
12 yr Wolfe et al., 1977
(PH5, 27C)
Me thoxychlo r
CH CCI-
1 yr Wolfe et al., 1977
(PH5, 2 7 C)
a a a T r ichlo r o to luene // \
CCL
19 s Lyman et al., 1982


98
approximately equal to one in the solvent phase. The
concentration of solute in the aqueous phase follows
Raoult's ideal solution law, and is expected to be
proportional to the mole fraction of solute in the mixture
(Banerjee, 1984).
Solute concentrations in water resulting from contact
with an immiscible mixture containing components which
interact in the solvent phase show deviations from ideal
behavior. The activity coefficients to explain these
deviations could be estimated by the UNIFAC model (Banerjee,
1984). Specific deviations from ideal behavior were
reported for mixtures of aromatic hydrocarbons and saturated
paraffins (Leinonen and Mackay, 1973; Sanemasa et al.,
1987). These deviations resulted in higher aqueous
concentrations (10-20%) than predicted from the ideal
behavior.
The addition of polar organic compounds which are
miscible or highly soluble in water (e.g. ethanol, tert-
butyl alcohol, MTBE) to a mixture of hydrocarbons and water
has a potential cosolvent effect, resulting in an increased
aqueous concentration of hydrocarbon (Groves, 1988; Rubino
and Yalkowsky, 1987; Fu and Luthy, 1986). Among the
methods for estimating the solubility of a solute in a
water cosolvent mixture is one based on a log linear
relationship. Rubino and Yalkowsky (1987) report the


37
Table 6. Summary of Brominated Compound Degradation
Rate Coefficients and Product Formation
Temp .
Matrix
Compound
108 k
s" 1
ke/k
%
20
DW
TBA
14
50.9
20
pH 4
TBA
10
60.7
20
pH 7
TBA
9
58.5
20
pH 10
TBA
11
61.8
28
DW
TBA
42
64.1
30
pH 4
TBA
65
61.8
30
pH 7
TBA
64
56.3
30
pH 10
TBA
71
63.1
37
1 M KC1
TBA
492
38.0
65
DW
TBA
7700
51.6
65
Na2S203
TBA
13000
60.1
20
DW
DBCA
17
35.6
20
pH 4
DBCA
14
33.9
20
pH 7
DBCA
11
32.5
20
pH 10
DBCA
15
33.4
28
DW
DBCA
51
45.6
30
pH 4
DBCA
81
39.8
30
pH 7
DBCA
69
31.9
30
pH 10
DBCA
73
38.5
37
1 M KC1
DBCA
492
24.4
65
DW
DBCA
7860
33.8
65
Na2 S 2*7 3
DBCA
14000
40.6
65
DW
BDCA
5350
29.6
DW, Distilled organic free water
Na2S2C>3, 1 M Sodium thiosulfate
Extrapolated Half-Lives for Degradation of TBA and DBCA
Temp .
108 k
e-1
T 1/2
Days
108 k
e-1
T 1/2
Days
15 5.54
20 12.83
25 28.92
145 6.90
62 15.56
28 34.14
116
52
23
TBA
DBCA


28
Figure 40. Bivariate plot of area counts of peak P16 and 3- or 4-ethyltoluene for
aqueous extractions of 65 gasoline samples.
146


110
one source of Kfw measurement variability due to the gas
chromatographic coelution of some of the constituents. This
was an example of another problem which would not occur with
systems where composition could be closely controlled. This
complexity factor primarily caused difficulty in obtaining
precise area percentages from gas chromatograms for those
neat gasolines where baseline resolution was not
accomplished. For example, 1,2,4-trimethylbenzene coeluted
with a low solubility hydrocarbon in the analysis of the
neat gasoline, which resulted in a high value for the
partition coefficient.
Among the factors that could affect Kfw was the solute
concentration. The relationship between Kfw and constituent
concentrations in the fuel or aqueous phase was examined and
is illustrated for benzene and toluene in Figure 29 and m,p-
xylene and 1,3,5-trimethylbenzene in Figure 30. Generally,
the Kfw did not show any significant trends with
concentration. Benzene showed a somewhat higher Kfw for low
concentrations which may be due to the coelution of
nonaromatic hydrocarbon compounds during the gas
chromatographic quantification of benzene in the neat
gasolines. The effect of the presence of lower solubility
coeluting hydrocarbons on the estimate of Kfw for benzene
would become more pronounced as the overall area percent of
the benzene decreased. Toluene was generally in much higher
concentration in the neat gasoline than most other


34
elimination pathway was measured in the strongest sodium
hydroxide (pH 13) solution.
Qualitative observations (GC and GC/MS) of TCA
degradation at approximately 60C in 0.5, 1.0 and 2.0 molar
sodium hydroxide solutions, showed the presence of 1,1-DCE,
and separate experiments indicated that 1,1-DCE also slowly
degraded under those conditions. These findings contradict
the results reported by Walraevens et al. (1974) in which
1.1-DCE was not detected in TCA degradation experiments at
high pH. This may be due to differences in analytical
methods, or the slow degradation of 1,1-DCE under their
reaction conditions.
Degradation of Brominated Ethanes
The degradation rates of brominated versus chlorinated
1.1.1-trihaloethanes were compared to provide insight into
the mechanisms and overall behavior of these compounds.
Since bromine is a better "leaving group" than chlorine,
brominated compounds typically degrade faster than their
chlorinated counterparts. In reviewing hydrolysis
degradation processes, Mabey and Mill (1978) concluded that
Br is more reactive than Cl by a factor of 5 to 10.
In a search of Chemical Abstracts, fewer than 20
references were reported for the brominated analog of TCA,
1.1.1-tribromoethane (TBA). Most of the papers addressed
spectra and bond energy studies, while no information on the
hydrolysis of this compound was reported.


170
fraction of these constituents in the solvent phase
increases.
Although differences in aqueous concentrations of
various gasoline constituents measured using GC/FID occur,
measurement of other parameters (lead, EDB, etc.) would
provide important additional information to aid in
identifying gasolines.


19
extracts included a 13-minute hold time at 35C, temperature
ramping of 3C/min to 90C, then 5C/min to 200C. The
helium flow rate was 3.0 mL/min.
Between August and December 1986, subsamples of
gasolines were obtained from the Department of Agriculture
and Consumer Services (DACS) Petroleum Laboratory in
Tallahassee, Florida. These samples were originally
collected by field inspectors and shipped for analysis to
assess compliance with ASTM guidelines and represent various
terminals in northern and central Florida. These samples
represented both summer and winter blends. Subsamples were
collected into 40 mL VOA screw cap vials with Teflon lined
septa and stored on ice prior to analysis.
Local samples were also collected from selected gas
stations in Gainesville. Samples were obtained from the
pump in gasoline safety containers, then a subsample was
transferred to a VOA vial and cooled.
Procedures for evaluating the partitioning of gasoline
into the aqueous phase were reported by Coleman (1984) and
Brookman et al. (1985a). Brookman et al. (1985a) measured
concentrations of aromatic compounds in the aqueous phase
with varying rotation contact times and found a maximum
concentration after two hours. Samples were then
centrifuged to separate the two phases. Coleman et al.
(1984) determined that a rotation contact time of 30 minutes
and an equilibration period of approximately 1 hour produced


Table 25. Stepwise Discriminant Analysis
Peaks to distinguish grades for each brand
152
Step
Amoco
Chevron
Gulf
Phillips
Shell
Union
1
P7
P 3
P4
P3
P10
P 3
2
P5
P6
P6
PI 2
P13
PI 5
3
P13
P7
PI 3
P 7
4
P14
P18
P21
P14
5
P3
P20
P14
P4
6
P17
P22
Pll
7
P9
P21
8
P6
9
P17
10
P23
11
P15
ASCC*
0.98
1.00
1.00
0.98
0.72
0.90
# Ob s
12
8
6
6
23
10
#Grades
3
3
2
2
4
3
Average squared canonical correlation, following the
final step.
Table 26. Stepwise Discriminant Analysis
Peaks to distinguish brands for each grade
Re guiar
Super
Super
Step
Regular
Unleaded
Regular
Premium
1
P4
P 5
P4
P7
2
P9
P6
PI 2
P6
3
P20
P8
P22
P 8
4
P12
P13
P7
P4
5
P23
P23
P14
6
P7
P10
P23
7
P4
8
P10
9
P3
10
P22
11
P9
ASCC*
0.64
0.83
1.00
0.81
# Ob s
11
26
8
19
# BRANDS
4
6
2
5


SUBROUTINE DENM(NP,CONC,L,DENOM)
DIMENSION CONC(NP,3)
DENOM-O
DO 10 K-l.NP
DENOM-DENOM+CONC(K,L)
CONTINUE
RETURN
END


131
Eaul1ibrium Concentrations of Major Constituents
The equilibrium concentrations of selected major
components in the water soluble fractions were quantitated
(Table 20). Water in equilibrium with a relatively fresh
spill of gasoline would be expected to show similar
concentrations. Benzene and toluene typically represented
approximately 55-65% by weight of the hydrocarbons detected
in the aqueous extracts.
Visual Comparison of Water Extracts of Gasoline
Appropriate dilutions for the purge and trap analysis
of the water extracts were, to a large extent, determined by
the presence of the major peaks, benzene, toluene and
xylenes. For this reason, during a typical ground water
investigations, these BTX constituents are the primary
reported analytes.
As discussed in the previous section, the aqueous
equilibrium concentration of a constituent was based on the
solubility of the individual components and the gasoline
composition. Given these critera, the other major peaks
detected in the water extracts were lower molecular weight
hydrocarbons which eluted early in the chromatogram, and
various aromatic constituents which eluted after toluene.
The attenuation settings selected during the printing
of chromatograms determined the visual appearance of the
chromatograms. My analytical protocol generated
chromatograms which could be visually compared, because


167
the carbon-hydrogen bond by interaction with base, resulting
in increased rate with increasing pH. The hydrogens in
1,1,1-trihaloethanes are not acidic unless halogens or other
functional groups are present on the beta carbon. This
suggests that the elimination reaction for TCA is El and not
E2, which also suggests a corresponding SN1 mechanism.
3. The reaction is independent of the concentration
and strength of the nucleophile.
4. Degradation of 1-bromo-1,1-dichloroe thane in a 1 M
KC1 solution resulted in the formation of some TCA providing
evidence of the formation of an ion pair in an E1/SN1
degradation.
5. Halogens have been shown to assist in stabilizing a
carbonium ion in other compounds.
1,1,1-Trichloroethane degraded more rapidly in
distilled deionized water than chloropropane or 1,1-
dichloroethane. In the presence of a strong nucleophile
chloropropane degraded rapidly, DCA degradation rate
increased by approximately a factor of 10, while the TCA
rate increased less than a factor of 2. Chloropropane
reacts by an SN2 mechanism, and DCA appears to degrade by an
intermediate mechanism.
The chemical degradation rate of TCA is independant of
the presence and strength of nucleophiles. Therefore, the
abiotic degradation rate in strongly reducing environments
which contain strong sulfide nucleophiles would be


69
accelerated by a factor of four in sediment as compared to
rates in buffered water.
Mabey and Mill (1978) indicated that acid promotion of
the aqueous hydrolysis of halogenated aliphatic hydrocarbons
has not been observed. March (1985) stated that gem-
dihalides can be hydrolyzed in water with either acid or
basic catalysis to give aldehydes or ketones, although the
strength of acid was not addressed.
In a review of elimination reactions in the presence of
polar catalysts, Noller and Kladnig (1976) stated that
"interaction of X with an acid is probably as indispensable
as the reaction of H with base in liquid-phase elimination
reactions, but this function is probably taken over by the
solvent and is less pronounced than the base promoted
process."
Clarification of interactions with polar surfaces may
provide insight into possible effects of sediments or soil
on reaction rates. Clays, for example, contain polar
surfaces which have been shown to catalyze degradation of
some pesticides (Saltzman and Mingelgrin, 1984).
Noller and Kladnig (1976) illustrated elimination
reaction products were a function of the specific catalyst
with 1,1,2-trichloroethane
Cl H
Cl Ci C2 Cl
H H


24
attack at the alpha carbon (carbon containing the halogen
leaving group) occurs before the transition state in the
rate determining step, not the extent to which the bond to
the leaving group is broken. Clear cut differences in
substitution reaction mechanisms are apparent in many
reactions. In practice, there is a spectrum of SN2
mechanisms involving varying amounts of nucleophilic attack,
with SN1 being the limiting case where nucleophilic attack
does not occur before the transition state of the rate
determining step.
Unimolecular (SN1 or El) processes are favored by
systems that form stable carbocations^. A classic example
would be the hydrolysis of t-butyl bromide. The more polar
the solvent, the faster the reaction. An increase in ionic
strength will typically increase the reaction rate, unless
the anion is the leaving group ion (common ion effect). The
reaction is independent of the concentration of nucleophile.
The classic SN2 case occurs in molecules with low
steric hindrance and low carbocation stability. Simple
primary halides react by the SN2 mechanism, while secondary
halides react by an SN2 or intermediate mechanism. Solvent
1 For years these were called "carbonium ions".
Recently, it was determined that the term "carbonium ions"
more accurately refers to pentacoordinated positive ions
(e.g. CH5+) and the more typical positive ion intermediates
(R3C+) are "carbenium ions". The term "carbocation"
includes either type and is generally used to describe any
of these intermediates (March, 1985, p. 141-142).


50
A summary of the results of these experiments is
presented in Table 7. No significant degradation of these
compounds was found in the experimental matrices during the
indicated reaction times, as evidenced by the slopes of In C
vs time which were not significantly different from zero.
The overall coefficient of variation for the observations is
similar to values obtained for simple replicate analyses.
Experiments were also conducted to evaluate the overall
behavior of these compounds under more rigorous conditions.
The literature indicated that halogenated ethenes such as
TCE can undergo elimination to form chloroacetylenes at
elevated pH (Rappaport, 1969). This reaction was verified
by using GC/MS to confirm the formation of dichloroacetylene
from TCE and also chloroacetylene from 1,1-DCE by analysis
of the headspace vapor above an alkaline (1 M NaOH) aqueous
solution of the halogenated ethene which was warmed to
approximately 60C.
The rate of degradation of components in a mixture of
1,1-DCE, TCE and PCE in sodium hydroxide solutions was
examined at 60C (Table 8). These were the matrices used by
Walraevens et al. (1974) in their examination of the
degradation of TCA, wherein they did not observe formation
of 1,1-DCE. One objective was to establish if the
elimination product was stable under their reaction
conditions.


70
as reactant. Basic catalysts (e.g., KOH-SO2) attack the
most acidic H, that at C^ forming more 1,1- than 1,2-
dichloroethene. Acidic catalysts (e.g., si1ica-alumina)
attack Cl on because the formation of the carbocation is
facilitated by the other Cl on that carbon resulting in the
formation of much more of the 1,2-dichloroethene isomer.
The choice of catalyst will determine the predominant
product giving selectivity to the reaction.
Mochida et al. (1967) reported that the reactivity of
TCA on solid acids was greater than that for other
chlorinated ethanes (mono-, di-, tri- and tetra- chloro
compounds). On solid bases it was less reactive than
penta-, tetra-, and 1,1,2-tri- chloroethanes. The shift in
reactivity of the ethanes with change in catalyst showed
enhanced ability of TCA to form a carbocation by
accelerating the reaction on an acid surface as compared to
the other chlorinated ethanes. There was also the lack of
an acidic beta hydrogen to permit catalysis by base.
Possible catalysis would be compound- and mechanism-
specific. Degradation experiments were performed on 1-
chloropropane (SN2) and TCA (SN1,E1) at 65C in 5 ml
distilled deionized water, with a final concentration of
approximately 2 mg/1. Separate ampules were prepared with
the addition of 0.4 g bentonite clay, 1 g limestone, 1 g
sand, and 0.2 g silica gel.


61
the substitution product, acetone, would not have
partitioned and been measured using that analytical
protocol.
Comparisons of Geminal Trihalides
A number of compounds in the literature contain a
geminal trihalide group (R-CX3), and many of these compounds
have environmental implications. My experiments on 1,1,1-
trihaloethanes indicated that the -CX3 group was sterically
hindered and resistant to attack by an SN2 mechanism, and
that the halogens could help to stabilize the formation of a
carbocation. The overall rate of degradation of other
geminal trihalides will increase if R also stabilizes the
carbocation. If the beta carbon contains an acidic hydrogen
the mechanism may shift to E2 at elevated pH.
A summary of degradation rates (expressed as reaction
half-lives) of various geminal trihalides is presented in
Table 10. The simplest compounds, trihalomethanes, were
very resistant to hydrolysis. The R- consists only of
hydrogen, which was inadequate to stabilize a carbocation.
The mechanism for this degradation has been determined to be
a base catalyzed process with a carbanion intermediate
(Hine, 1950). The extremely low reactivity also suggests
that steric hindrance may prevent SN2 attack.
By contrast alpha,alpha,alpha-trichlorotoluene has a
half-life of 19 seconds at 25C, which corresponds to a rate
of a factor of 10^ greater than for TCA. Therefore, the


I certify that I have read this study and that in my
opinion it conforms to acceptable standards of scholarly
presentation and is fully adequate, in scope and quality,
as a dissertation for the degrea^-trfS Doctor^o&j Philosophy.
fseph J f Delfiny, Chairman
rofessor of Environmental
Engineering Sciences
I certify that I have read this study and that in my
opinion it conforms to acceptable standards of scholarly
presentation and is fully adequate, in scope and quality,
as a dissertation for the degree of Doctor of Philosophy.
Paul A. Chadik
Assistant Professor of
Environmental Engineering
S cience s
I certify that I have read this study and that in my
opinion it conforms to acceptable standards of scholarly
presentation and is fully adequate, in scope and quality,
as a dissertation for the degree of-'iTO'etp-K off'PTT'ilosophy.
>hn) Dorsey
Asferrciate Professor
Chemistry


82
solubility of each with varying compositions of the binary
mixture. Mixtures were at room temperature, approximately
24C .
The pure component solubility of TCA (1580 mg/L or 11.8
mmoles/L) and the solubility of 1,1-DCE in the aqueous phase
(3200 mg/L or 33 mmoles/L) measured at 24C were within the
concentration range listed by Verscheuren (1977) who
reported solubilities at 20 and 30C. This is significantly
higher than solubilities reported by Lyman (1981) and
Pearson and McConnell (1975). The solubility for 1,1-DCE
reported in this dissertation was verified independently by
solubility measurements performed using high performance
liquid chromatography (HPLC) (Linda Lee, University of
Florida, Personal communication, 1988). She measured an
average for the solubility of 1,1-DCE at 24C as 2990 mg/L.
Her report is included in Appendix B.
Verscheuren (1977) reported that the solubility of TCA
at 20C was four times greater than at 30C, a value
approximately three times greater than our result at 24C.
Since the mass lost per unit time from degradation is a
function of aqueous concentration and the first-order
degradation rate coefficient, higher aqueous concentrations
at lower temperatures could compensate for the lower
degradation rate. The solubility of TCA at 4C was measured
to verify this trend. As shown in Table 12, a significant


33
The elimination product, 1,1-DCE, was measured to
establish the factors which influenced the reaction pathway
(substitution versus elimination). Degradation of 1,1-DCE
was observed only at very high pH and even under those
conditions the rate was slow compared with the degradation
of TCA. Therefore, the ratio of the rate for elimination
(ke) to the total rate of degradation (k) was estimated by
plotting the concentration of 1,1-DCE versus (l-e'^t) where
t is time. The slope of the line equals ([TCA]c (ke/k)),
where [TCA]0 is the concentration of TCA at time zero.
This calculation required an estimate for the starting
concentration of TCA. For most experiments, multiple
analyses were performed for the estimate of the initial
concentration. Other authors (eg. Vogel and McCarty, 1987)
have used the intercept in the regression analysis of the
degradation, and this value was used as the estimate of
initial concentrations in this study.
Increases in pH and/or temperature theoretically favor
elimination over substitution. The elimination pathway
(Table 4) ranged from 17 to 38% of the total degradation
rate of TCA. Higher temperatures showed slightly more
transformation to 1,1-DCE over the temperature range
evaluated in these experiments. The percent of TCA
degradation due to elimination was not affected by matrix in
the pH range of 4.5 to 8.5. Seawater had no apparent effect
on the relative proportion of products. The highest percent


Aqueous Concentration (mg/I) Sc Concentration (mg/I)
128
e 34. Theoretical concentration changes from multiple
us extractions of a Shell super unleaded gasoline.
EXTRACTION NUMBER
Figure 35. Theoretical concentration changes from multiple
aqueous extractions of a Union super unleaded gasoline.


o o o
175
C INITIALIZE MOLE FRAC & CONCS
CALL DENM(NP,CONC,1,DENOM)
DO 12 K-l.NP
FRACM(K)- CONC(K, D/DENOM
CONC(K,3) (FRACM(K)*SOLUB(K))
CONC(K,2) CONC(K,1)-(CONC(K,3))
12 CONTINUE
CALL DENM(NP,CONC,2,DENOM)
DO 14 K-l.NP
FRACM(K)- CONC(K,2)/DENOM
14 CONTINUE
WRITE(*,*) 'WORKING ON DAY '
WRITE(*,*)
C MAIN LOOP
DO 100 NT-0,DAYS,DT
WRITE(*,20) NT
20 FORMAT ('+',4X,14)
CALL ERRCHK(NP,SOLUB,CONC,FRACM,ERRCRI,ISERR)
IF(IS ERR.EQ.1) THEN
CALL
MATCON(NP,CONC,FRACM,SOLUB,DENOM,ERRCRI)
END IF
C SPECIFIC TO 2 CMPD SYSTEM
MUST REPLACE W/ DIFF CODE
FOR LARGER SYSTEM
C OUTPUT ROUTINES
WRITE(8 '(16,3X,6F9.3) ') NT ((CONC(I J ) J-1,3),
1-1,2)
DEGR2--0.2*DEGRT(l)*CONC(l,3)
CONC(l,l)=CONC(l,1)-CONC(l,3)*(DEGRT(l)+FLOWRT)*DT
CONC(2,1)-CONC(2,1)-(DEGR2+(CONC(2,3)*FLOWRT))*DT
C RECOMP MOLE FRAC & CHECK ERR
CALL DENM(NP,CONC,2,DENOM)
DO 60 K-l.NP
FRACM(K)= CONC(K,2)/DENOM
CONC(K,3) CONC(K,1)-(CONC(K,2))
60 CONTINUE
100 CONTINUE
CLOSE(5)
END


I certify that I have read this study and that in my
opinion it conforms to acceptable standards of scholarly
presentation and is fully adequate, in scope and quality,
as a dissertation for the degree of-Doctor of Philosophy.
Rao
Ifessor of Soil Science
I certify that I have read this study and that in my
opinion it conforms to acceptable standards of scholarly
presentation and is fully adequate, in scope and quality,
as a dissertation for the degree of Doctor of Philosophy.
Richard A. Yost(J
Associate Professor of
Chemistry
This dissertation was submitted to the Graduate Faculty
of the College of Engineering and to the Graduate School and
was accepted as partial fulfillment of the requirements for
the degree of Doctor of Philosophy.
August 1988
Dean
liege
Engineering
Dean, Graduate School


181
Brand
Grade
Date Symbol
P3
P4
P5
MTBE
46
Shell
SR
2/04/87
M
15.17
13.11
17 14
47
Shell
SR
2/04/87
M
14.07
14.88
16 .68
48
Shell
SP
11/09/86
N
11.73
12.72
6 .25
49
Shell
SP
11/09/86
N
12.01
21.31
9.81
50
Shell
SP
11/09/86
N
12.21
20.13
8.10
51
Shell
SP
12/15/86
N
20.66
20.66
1164.99
52
Shell
SP
1/05/87
N
17.86
11.85
38 50
53
Shell
SP
2/04/87
N
17.37
10.70
23.73
54
Shell
SP
2/04/87
N
18 55
11.05
20.65
55
Shell
SP
2/04/87
N
18.05
10.77
21.52
56
Union
R
10/20/86
0
13.86
23.48
5.21
57
Union
R
11/09/86
0
16.96
15.47
18.40
58
Union
R
11/09/86
0
16.50
3.98
12.84
59
Union
SP
10/20/86
P
5 30
6.56
4.30
60
Union
SP
11/09/86
P
12.99
9.91
590.51
61
Union
SP
11/09/86
P
8.87
1.83
3 22
62
Union
RU
10/20/86
Q
28.09
29.53
10.26
63
Union
RU
10/20/86
Q
30.59
31.28
10.99
64
Union
RU
11/09/86
Q
30.19
24.27
9.08
65
Union
RU
11/09/86
Q
20.31
7.01
7.03


121
six aromatics: benzene, toluene, ethylbenzene, o-, m-, and
p- xylenes.
Log Kfw -0.884 Log S + 0.975 r2 0.99
where S is the solubility of the pure hydrocarbon in water
(moles/1) and r2 is the correlation coefficient. This
equation would apply to the specific gasoline, water ratio
(1:10) and temperature of 13C used to generate the values
of Kfw. The regression analysis provided a conservative
equation which would be most accurately applied only to
aromatic compounds.
Brookman et al. (1985b) then applied this type of
regression model to several components including 2-butene,
2-pentene, n-butane, 1,2,4-trimethylbenzene, 2 me thylbutane
and n-pentane in addition to the BTX components and ethyl
benzene that were originally included in the first equation.
The log plot of the twelve components' solubilities and
partition coefficients resulted in the linear equation:
Log Kfw -1.018 log S + 0.706 r2 = 0.87
The correlation coefficient was lower when paraffins
and olefins were included with aromatic compounds in the
regression analysis, but suggested that Kfw was correlated
to solubility for a variety of components in addition to the
aromatic constituents. Therefore, solubility may be used to
provide an estimate for the partitioning behavior of a
broader spectrum of gasoline constituents. Brookman et al.
(1985b) stated one limitation of this model was the caveat


aqueous phase concentrations at equilibrium. The total
number of moles of TCA in a unit volume of porous media is
the sum of the moles present in the aqueous and solvent
80
phase s.
The model describes changes for TCA spilled on a high
permeability material like sand. As TCA degrades and forms
1,1-DCE, the degradation product partitions into the NAPL
affecting the aqueous phase concentration of TCA (and DCE).
Both the individual solubilities and the solubility of
a mixture of TCA and 1,1-DCE are required in the model and
it was also necessary to assess if mixtures of TCA and 1,1-
DCE deviate significantly from ideality. Literature values
for the solubilities of these constituents vary widely
(Table 12). The solubility data for 1,1-DCE reported by
Lyman (1981), showed as much as a 700% error from a
predicted concentration based on regression relationships.
That estimated concentration is much closer to the
concentrations reported by Verschueren (1977).
Table 12. Solubilities of TCA and 1,1-DCE (mg/L)
TeniD (C)
TCA
1 1
20
480
400
20
4400
2640
30
1088
3675
25
273
4
1700
4200
24
1580
3200
Source
Pearson and McConnell (1975)
Verschueren (1977)
Verschueren (1977)
Lyman (1982)
This study.
This study.
Measurements (Figure 19) were made on the solubility of
the individual components (TCA and 1,1-DCE) and on the


YEARS (Flow, 0.25%/Day)
Figure 26. Model results: Change in aqueous concentration
of 1,1-DCE as a function of initial mass of TCA and
composition of the solvent phase.
YEARS
Figure 27. Model results: Change in total mass of 1,1-DCE
in the residual zone as a function of initial mass of TCA
and composition of the solvent phase.


48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
184
P6
P7
P8
P9
P10
Pll
Benzene
Toluene
Ethylbz
in p Xylene
o-Xylene
142.56
260.14
14.26
56.41
25.61
1.03
156.94
304.14
16.71
68.54
30.11
1.22
139.82
249.37
12 20
52 64
23.90
0.96
151.26
343.92
17 74
59.57
26.99
2.43
191.15
236.15
13.48
37.66
21.99
0.86
170.03
347.25
16.27
61.69
28.39
1.54
152.79
284.58
12.42
52.61
25 11
2.64
131.91
238.43
11.54
48.21
22.68
2.14
233.39
145.14
0
8.08
13 .88
0.61
127.39
127.21
8.10
32.07
15.68
4.07
202.15
305.79
13.38
71 70
30.43
1. 16
406.30
267.47
13.77
47.19
23.58
0.00
258.05
275.13
7.10
24.71
15.87
1.07
227.50
307.59
11.73
59.77
29 26
1.26
186.40
156.11
8 35
31.15
15.07
0.62
197.03
167.29
9.26
34.49
16.65
0.75
150.22
184.64
10.27
38.04
19.62
1.62
183.63
245.32
9.9998
49.32
23.28
1.04
PI 2
PI 3
P14
PI 5
PI 6
P17
0.84
5 .35
2.88
2.22
7 34
0.43
0.99
6 .22
3.45
2.31
9.01
0.00
0.58
4.46
2 52
1.85
5.82
0.36
1.32
6.94
4.14
2.63
8.02
5 .85
1.81
3.80
2 88
1.89
4.24
0.60
0.89
6.44
3.05
2 .29
10.36
2.49
1.03
5.56
3 .26
2.32
9.89
4.46
1.03
5.24
3.12
2.17
9.80
4.54
0.59
3.15
1.90
1.35
4.21
0.00
0.74
4.75
4.10
1 84
6.69
0.92
0.80
5 .56
3.51
2.04
8 68
0.40
1 .29
5 .75
2.58
2.38
7 .58
0.00
1.24
6.01
3.71
2 74
7.56
0.00
0.74
5.23
3 39
1.90
7 18
0.49
0.68
4.15
2.43
1.86
5.37
0.68
0.87
5.01
2.70
1.93
6.58
0.00
0.83
4.89
3.26
2 36
7 .26
0.00
0.77
4.30
2 73
1 .71
6 12
0.37


29
rate constants were based on reactions showing a minimum of
75% degradation of the initial concentration of TCA.
Statistical analyses were performed to assess if the
slopes measured at any given temperature were significantly
different, thus determining the extent to which the sample
matrix, or pH affected the rate constant. The reaction
rates in the buffer solutions (pH 4.5, 7 and 8.5) were not
significantly affected by pH (p < 0.01). In addition, the
rates measured in ground water matrices at 70C (GW1, GW2)
were not significantly different from rates measured in the
buffer solutions at the same temperature.
The spiking solutions typically were prepared with
methanol, which resulted in approximately 0.1% methanol in
the final solution. Separate experiments were conducted
without the use of methanol with no apparent affect on the
rates. The use of methanol decreased the variability in
concentrations observed among ampules, apparently due to the
decreased volatility of TCA in the methanol spiking
solution.
Reaction rates at 53C in seawater, distilled deionized
water and 0.05 M phosphate buffer solutions showed that the
ionic matrix affected the rate of reaction. The fastest
rate was observed for seawater, while the rate in distilled
deionized water (DW) was 14% lower and those in the buffer
solutions were approximately 10% lower. The rates measured
in the distilled deionized water and the buffer solutions


151
The stepwise discriminant analysis was therefore
applied to various subsets of the data. In the first case,
each brand was evaluated separately to determine the peaks
that would distinguish among its various grades. Some
brands had only two classes (grades) to be distinguished,
while others had as many as four for my data set. The peaks
selected by this statistical procedure varied in number,
type and order of selection for the different brands (Table
25). The greatest number of peaks included by this
procedure were selected to distinguish four grades of Shell
gasolines, and clear separation was not obtained (ASCG of
0.92).
For a second subset, each grade was evaluated to select
peaks which would best distinguish between brands (Table
26). Similar to the previous subset, each grade did not
have the same number of classes (brands) or observations.
This approach could be further applied in a variety of
ways, for example, selecting only two brands of interest, or
possibly two specific types. The more limited the number of
classes to be distinguished, the better the possibility of
separating the samples.
The obvious advantage of this statistical procedure was
the focus it provided for subsequent evaluation of the
information. The procedure was applied here in a general
sense, however, its real utility would be in response to a
more specific question; for example, how does one determine


49
like trichloroethene (TCE) and tetrachloroethene (PCE) may
undergo slow abiotic degradation in water at room
temperature with a half-life of less than one year were
reevaluated. The question of possible nucleophilic attack
by water, hydroxide ion, or other nucleophiles must be
addressed to understand the stability of these commonly
detected ground water contaminants.
The stability of 1,1-DCE was evaluated in experiments
that were performed concurrently with the evaluation of TCA
degradation. In the buffer solutions, seawater, and
distilled deionized water, no significant degradation of
1,1-DCE occurred during the course of the evaluation of the
degradation of TCA.
The formation of ethenes containing bromine was
monitored during the degradation studies of the brominated
ethanes, and their concentrations were continually monitored
for some time after the ethane degradation was completed.
Trichloroethene was studied in separate experiments
performed at various temperatures selected to repeat the
experiments conducted by Dilling et al. (1975). In addition
to buffer solutions, one set of ampules was prepared with a
nutrient solution which was not autoclaved, and to which
ground water known to show biological activity was added.
This was done to determine if any degradation which might
have occurred during the long term studies could have been
due to biological activity.


Principal Component 2
4
3 -
2 -
1 -
0 -
-1 -
-2 -
-3 -
-4 -
-3-11 3
Principal Component 1
Super Regular Gasolines
A Amoco
S Shell
Figure 45. Plot of principal component scores for Shell and Amoco super regular
gasolines.


160
Plots of the principal components for three brand pairs are
illustrated in Figures 46 to 48.
Amoco and Shell gasolines (Figure 46) were well
separated for most of the samples. Shell samples clustered
for all grades, while the Amoco samples showed much
diversity. Amoco gasoline composition was more variable
than other brands since the processing involved the addition
of toluene at varying levels.
Most Chevron and Union gasoline samples were separated
by the first principal component (Figure 47). The three
Chevron samples that had negative first principal component
scores were two regular leaded samples and one super premium
sample. Therefore, the lack of separation of these three
samples did not represent a factors specific to one
particular grade.
Gulf and Phillips gasolines (Figure 48) did not cluster
or separate from each other. Some of the differences were a
result of the differences in the grades of these gasolines.
The principal component plots are a visual comparison
of the similarities and differences of the gasoline brands
and grades evaluated in this data set. The gasoline samples
for a single brand and grade did not usually form close
clusters due to variability resulting from differences in
sampling source and time of collection. The principal
component scores between two brands may provide a way of
describing the differences in those brands for selected


H Cl
HC-CCI
H Cl
H Cl
HC-CBr
H Cl
BDCA
TCA
^C-C
h' XCI
DCE
H
HC-C
H
Acetic Acid
Figure 11. Reaction pathways for BDCA in 1 M KC1 solution. The exchange reaction of Cl
with the ion pair forms TCA, which degrades more slowly than the brominated compound.
P-
co


PRIN
4

3 -
2 -
1 -
O
-1 -
-2 -
-3
-3
+
+

X
7
-1
PRIN 2
Amoco
Shell
RU
a RU
+ SR
x SR
SP
<¡
00
T)
T
1
T
3
Figure 46. Principal component plot of all grades of Amoco and Shell gasolines.


190
Society of Petroleum Engineers Journal. 25(1): pp. 101-112,
1985 .
Kinzelbach, W. K. H., Modelling of the Transport of
Chlorinated Hydrocarbon Solvents in Groundwater: A Case
Study, Water Science and Technology. 17: pp. 13-21, 1985.
Lane, J.C., Gasoline and Other Motor Fuels. In:
Kirk-Othmer Encyclopedia of Chemical Technology edited by
Mark, H.F., McKetta, J. J., Othmer, D. F. and Standen, A.,
New York, Interscience Publishers, 11: pp. 664-671, 1977.
Leinonen, P. J., and Mackay.D., The Multicompoment
Solubility of Hydrocarbons in Water. Canadian Journal of
Chemical Engineering. 51: pp. 230- 233, 1973.
Levenspiel, 0., Chemical Reaction Engineering. New York,
Wiley Publishing Co., 1962.
Lyman, J. W., Reehl, W. F., and Rosenblatt, D. H.,
Handbook of Chemical Property Estimation Methods. San
Francisco, McGraw Hill Publishing Co., 1982.
Lysyj, I., and Newton, P. R., Multicomponent Pattern
Recognition and Differentiation Method, Analytical
Chemistry. 44(14): pp. 2385-2386, 1972.
Mabey, W., and Mill, T., Critical Review of Hydrolysis
of Organic Compounds in Water Under Environmental
Conditions, Journal of Physical Chemical Reference Data. 7:
pp. 383-415, 1978.
March, J., Advanced Organic Chemistry. New York, John
Wiley & Sons, 1985.
Mochida, I., Jun-ichiro, T, Saito, Y., and Yoneda, Y.,
Linear Free-Energy Relationships in Heterogeneous Catalysis.
VI. Catalytic Elimination Reaction of Hydrogen Chloride from
Chloroethanes on Solid Acids and Bases, Journal of Organic
Chemistry. 32: pp. 3894-3899, 1967.
Nkedi-Kizza, P., Rao, P. S. C., and Hornsby, A. G.,
Influence of Organic Cosolvents on Sorption of Hydrophobic
Organic Chemicals by Soils, Environmental Science and
Technology. 19: pp. 975-979, 1985.
Noller, H., and Kladnig, W., Elimination Reactions over
Polar Catalysts: Mechanistic Considerations, Catalvsis
Reviews. 13(2): pp. 149-207, 1976.
Novak, J. P., Matous, J., and Pick, J., Liquid-Liquid
Equilibria. New York, Elsevier, 1987.


124
performed for a gasoline sample (Phillips regular unleaded)
which contained approximately 5% MTBE. After the first
extraction was performed using the methods described
previously, the water fraction was sampled and the remaining
water layer was removed. Forty ml of distilled deionized
water were added to the vial to reextract the fuel layer.
The procedure was repeated a third time. The experiment was
performed in triplicate, and average concentrations of
selected constituents are summarized in Table 19.
Concentrations of MTBE showed a dramatic decrease over the
three extractions, while the other components remained
essentially constant. The sum of the area counts of all
components also showed substantial decreases over the three
extractions, reflecting primarily the losses of benzene and
MTBE .
The changes in composition with sequential extractions,
which represent changes with time, can be calculated from
the fuel/water partition coefficients. For each extraction,
the concentration of the component in the aqueous phase was
estimated using the Kfw. The amount of the compound
partitioning into the water was then subtracted from the
fuel layer. The new concentration of the component in the
fuel layer was calculated with a correction made for the
change in volume (based on the sum of the major
constituents). The calculation was then repeated using the


SUBROUTINE ERRCHK(NP,SOLUB,CONC,FRACM,ERRCRI,IS ERR)
PARAMETER (ND-2)
DIMENSION
CONC(ND,3),ERRCRI(2),FRACM(ND),ERRMAT(ND,2)
DIMENSION SOLUB(ND)
INTEGER ERRMAT
IS ERR 0
DO 10 I-l.NP
ABSERR- ABS(CONC(I,3)-SOLUB(I)*FRACM(I))
RELERR- ABSERR/ABS(CONC(I,3) )
IF (ABSERR.LE.ERRCRI(1)) THEN
ERRMAT(1,1)-1
ELSE
ERRMAT(I,l)-0
ENDIF
IF (RELERR.LE.ERRCRI(2)) THEN
ERRMAT(I,2)-l
ELSE
ERRMAT(I,2)-0
ENDIF
IF (ERRMAT(1,1).EQ.0.AND.ERRMAT(1,2).EQ.0) THEN
IS ERR1
RETURN
ENDIF
CONTINUE
RETURN
END
SUBROUTINE MATCON(NP,CONC,FRACM,SOLUB,DENOM,ERRCRI)
PARAMETER (ND-2)
DIMENSION CONC(ND,3),ERRCRI(2),FRACM(ND),SOLUB(2)
DIMENSION PRTDRV(ND,ND)
COMPUTE FUNCTION & INVMAT
NOTE IF NP>2 THEN REPLACE
INVERSION ROUTINE WITH GENERAL
METHOD
DO 50 N-1,50
COMPUTE PARTIALS
PRTDRV(1,l)--SOLUB(l)*FRACM(2)/DENOM
PRTDRV(1,2)-1+SOLUB(1)*FRACM(1)/DENOM
PRTDRV(2,2)--SOLUB(2)*FRACM(2)/DENOM
PRTDRV(2,1)-1+SOLUB(2)*FRACM(l)/DENOM


155
Principal Component Analysis
Principal component analysis is a simple ordination
procedure for projecting a multidimensional data set into
two or three dimensions to reveal intrinsic patterns.
In essence, the data are projected, without any
differential weighting, onto a differently oriented s-space.
The axes of the original coordinate frame in which the data
points are plotted are rotated rigidly around their origin.
This is done in such a way that the pattern of the data is
simplified. Principal component scores are then determined,
being weighted sums of the quantities after they have been
centered by the species means. The first principal axis is
oriented to make the variance of the first principal
component scores as great as possible, and the second is
oriented to make the spread in the data as great as possible
with the restriction that the second axis must be
perpendicular to the first axis. This approach is continued
to the final axis which is equal to the number of variables.
The effort is considered successful when a large proportion
of the total dispersion of the data is parallel with the
first two or three principal axes; then this large
proportion of the information contained in the original,
unvisualizable, s-dimensional data swarm can be plotted in
two-space or three-space and examined (Pielou, 1984).
The stepwise discriminant analysis of the total data
set had an ASCC of 0.32 after the final step. This


5
VOC's, only three (1,1,1-trichloroethane (TCA),
trichloroethene (TCE), and tetrachloroethene (PCE)) are used
in large quantities at the industrial facilities. The
presence of 1,2-dichloroethene isomers and 1,1-
dichloroethane, particularly with frequent detections of
vinyl chloride, suggest anaerobic biodegradation (Parsons
and Lage 1985; Bouwer and McCarty, 1983 ). Selected
locations show very high levels of 1,1-DCE in association
with TCA, and frequently little evidence of biodegradation
(Table 1). The primary source of 1,1-DCE at these locations
appears to be the chemical degradation of TCA, prompting
questions as to the rate of formation of the 1,1-DCE and its
stability in ground water.
Table 1. Maximum Concentrations (/g/L) of VOC's Detected
at Selected Sites in Arizona (Graf, 1986)
Site
TCA
1,1-DCE
TCE
1,2-DCE
1
630
3320
13000
20
2
490
1320
9
-
3
9800
10400
410
933
4
98
206
139
106
Two products are formed during the abiotic degradation
of TCA. The elimination product is 1,1-DCE, while the
substitution or hydrolysis product is acetic acid (Figure
1). Previous research (Cline, 1987) described the rate of
degradation of TCA and formation of 1,1-DCE in dilute buffer
solutions (pH 4-10) at temperatures from 27 to 70C.


TABLE OF CONTENTS
ACKNOWLEDGEMENTS ii
ABSTRACT v
INTRODUCTION ....... 1
Chemistry of Alkyl Halides 3
Gasoline Partitioning 8
MATERIALS AND METHODS 12
Alkyl Halides 12
Gasoline 18
DEGRADATION OF ALKYL HALIDES 22
Introduction 22
Degradation of 1,1,1-Trichloroethane 25
Degradation of Brominated Ethanes 34
Halogenated Ethenes 45
Structure/Rate Relationships of Alkyl Halides . 53
Simple SN1/E1 Reactions 55
Comparisons of Geminal Trihalides 61
Effect of Additional Halogens on the Alpha
Carbon 64
Sediment Matrix Affects 67
SOLUBILIZATION AND DEGRADATION OF RESIDUAL TCA .... 75
Behavior of Residual Solvent 75
Aqueous Phase Concentrations 79
Advection 83
Degradation Rate 84
Model Parameters and Procedures 85
Limitations of the Model Assumptions .... 94
i i i


13
Seawater samples were obtained from the coastal
Atlantic Ocean near Ormond Beach, Florida. Samples were
filtered and subsequently handled similar to the phosphate
solutions.
Ground water samples from two monitoring wells were
obtained from a site in Orlando, Florida, which had been
contaminated by chlorinated solvents. These samples were
purged to remove existing solvents and interfering
substances, then filter (10 pm) sterilized.
Approximately 6.6 mL of the phosphate solutions,
seawater or distilled deionized water were added to 5 mL
(nominal volume) glass ampules (Wheaton Scientific). The
ampules were plugged with cotton and autoclaved for 15
minutes at 121C.
These ampules were then aseptically spiked with 10 pL
of the stock solution of TCA in methanol and flame sealed
using a Model 524PS sealing unit manufactured by O.I.
Corporation. Final concentrations were approximately 1-3
mg/1. Approximately 0.5 to 1 mL of air space was present in
the ampules after sealing.
Ampules were incubated at 28C (Precision Scientific
Model 6) and at 37C (Precision Scientific Model 4).
Experiments at higher temperatures were performed in a
Magna-Whirl constant temperature water bath (Blue M).
Samples were analyzed using a purge and trap device
(Tekmar LSC-2), interfaced with a Perkin Elmer Model 8410 GC


122
that it was applicable to only the specific gasoline for
which it was derived.
The results of my research indicate that the Kfw may be
more widely applied to estimate fuel/water partitioning for
a range of fuel compositions i.e. different brands and
grades. The relationship between log of the average
fuel/water partition coefficient and the log of the
solubility was determined for MTBE and eight aromatic
compounds; benzene, toluene, ethylbenzene, m,p-xylene,
o-xylene, n-propylbenzene and 1,2,3-trimethylbenzene. The
plot of the log S versus log Kfw is show on Figure 32, and
the regression equation was
Log Kfw -0.84 log S + 1.29 r2 0.98
The plot showed a strong linear relationship for log
Kfw and log S. The variability was due, as stated earlier,
to a variety of factors including differences in composition
of the fuels and peak overlap.
Changes in Concentrations with Time
An ageing effect occurs when complex fuel mixtures
remain in contact with air or water. The more soluble and
volatile components will be differentially removed from the
fuel, producing a time-dependent composition for the
hydrocarbon layers, and also the waters in contact with
them.
To evaluate the patterns of the changes in composition
with time due to solubilization, multiple extractions were


8
Gasoline Partitioning
Gasoline contamination of ground water has become a
major environmental concern. Documented cases of
contamination from underground storage tanks (Florida
Department of Environmental Regulation, 1985) have prompted
enactment of additional legislation, the "State Underground
Petroleum Environmental Response Act of 1986" (SUPER Act),
to protect the ground water and surface waters of the state
of Florida. The SUPER act was designed to maximize ground
water protection, encourage early detection, reporting, and
clean-up of leaking underground storage tanks.
Issues relating to the behavior of gasoline components
in ground water are diverse and complex. Gasoline itself is
a complex mixture of hydrocarbons and some of the factors
which affect the concentration of these constituents in the
subsurface environment (vadose zone and ground water)
include solubility, biodegradability, volatility, soil
sorptive capacity, and dilution.
Components of gasoline may undergo abiotic chemical or
photochemical oxidations through free radical formation.
Thermal degradation is negligible at environmental
temperatures below 80C. Since free radicals are limited in
the subsurface environment, chemical degradation is not
expected to play a significant role there (Bossert and
Bartha, 1984).


105
higher ratios of aromatic constituents (Senn and Johnson,
1987) .
Background information on statistical procedures is
available from a variety of sources (Pielou, 1984; Wolff and
Parsons, 1983; SAS Institute, Inc., 1985; Gordon, 1981).
The basic procedures used in my analysis and described in
these sources included simple statistics, correlations,
stepwise discriminant analyses, and principal component
analyses.
The interpretation of complex chromatograms using
pattern recognition techniques has been demonstrated
(Hosenfeld and Bauer, 1985). Statistical procedures were
applied without prior knowledge of chromatogram peak
identity for either compound class or type. Peaks were
identified by relative retention time, with principal
component analysis used to define the patterns in the
complex data set. This approach was used to interpret the
gasoline chromatograms in my study.
Partitioning of Gasoline Components into Water
Fuel/Water Partition Coefficients
To examine the partitioning behavior of individual
gasoline components in various grades and brands of
gasolines, neat gasoline and water extracts of gasoline were
analyzed for 31 gasoline samples. Commercial brands
included Amoco, Gulf, Shell, Phillips and Union.


93
the different concentrations of 1,1-DCE which would result.
Given a constant initial mass of TCA, the maximum
concentration of 1,1-DCE in the aqueous phase occurs at the
lowest flow rates. For flow rates higher than the 0.5%
volume exchange per day the advective term is dominant and
concentrations of 1,1-DCE in the residual zone remain
negligible.
As long as a residual NAPL is present, aqueous
concentrations are dominated by TCA. Equal concentrations
of TCA and 1,1-DCE in the water from monitoring well data
from various sites would occur according to the model
primarily in the plume of dissolved constituents
downgradient from the residual zone, or in the original
spill area after all residual solvent was dissolved or
degraded. The presence of a low solubility compound in the
solvent phase with the TCA will considerably slow TCA rate
of advection and degradation.
First-order degradation will continue in the ground
water plume downgradient from the source and this process
could be modeled (Kinzelbach, 1985). Evidence of the
formation of 1,1-DCE would support the assignment of a
degradation rate. Assuming similar retardation factors for
TCA and 1,1-DCE, equal concentrations of TCA and 1,1-DCE
would occur after approximately 3 half-lives, approximately
2.5 years at 25C.


58
Table 9. Summary of Degradation Rate Coefficients and
Pathways for Tertiary
and Secondary
Halides
Compound
k (sec"^)
ke/^t
Reference
2 5C
%
t-Butyl Bromide
5
1
t-Butyl Chloride
2.98xl02
2
2-Bromo-2-chloropropane
1.78xl0'4
100
2
2,2 Dibromopropane
4.6 2x105
100
2
2,2-Dichloropropane
9.09xl0'6
100
2
1,1,1-Tribromoe thane
2.89x10 7
58
3
1,1-Dibromo-l-chloroethane
3.41xl0'7
37
3
1,1 1-Trichloroe thane
2.62xl08
23
3
2-Bromopropane
3.82xl06
0
2
2-Chloropropane
2.11x10'7
0
2
References:
1. March, 1985 .
2. Queen and Robertson, 1966.
3. This dissertation.


139
Table 22. Basic Statistics on Parameter Area Count Data
for 21 Gas Chromatography Peaks Identified in the
Water Extracts of Gasolines.
VARIABLE
N
MEAN
STD DEV
SUM
MINIMUM
MAXIM
P 3
65
21.0
9.0
1365
3.05
39 5
P4
65
21.2
9.9
1378
1.83
45.0
P 5
65
139
305
9091
2.81
1315
P6
65
133
65.0
8634
38.5
406
P7
65
217
95.4
14100
80.5
578
P8
65
10.0
3.07
652
4.78
17 9
P9
65
35.6
13.7
2316
8.08
71.7
P10
65
17.6
5.85
1146
8 51
30.4
Pll
65
1 54
1.04
100
0.56
6.88
P12
65
1 30
1.39
84 .
6
0.49
9 .25
P13
65
5.18
1.61
337
2.44
11.9
P14
65
3 14
1.11
204
1.63
8.81
Pi 5
65
2 14
0.67
139
1.25
4.91
P16
65
7.58
2.91
493
3.25
19 9
P17
65
1.48
1.63
96 .
, 2
0.00
5 .95
P18
65
2 .38
0.97
155
1.10
7.40
PI 9
65
2 .11
1.05
137
0.79
6 14
P20
65
2.29
2 .65
149
0.00
11.3
P21
65
2.45
1.09
159
0.73
8.93
P 2 2
65
3 .31
3.42
215
0.00
15.6
P 2 3
65
2 .79
2.42
181
0.00
13.0


136
Table 21. Peaks Used to Statistically Characterize
Water Extracts of Gasoline
Peak
Approximate
Retention Time (rain)
Compound
P3
5.85
C5 Hydrocarbon
P4
6.26
C5 Hydrocarbon
P5
8.12
MTBE, C5 Hydrocarbon
P6
14.06
Benzene
P7
23.08
Toluene
P 8
30.09
Ethylbenzene
P9
30.69
m p- Xylene
P10
32.19
o-Xylene
Pll
32.78
unidentified
P12
35.69
n-Propylbenzene
P13
36.03
3- and 4- Ethyltoluene
P14
36.39
1,3,5-Trimethylbenzene
P15
36.89
2-Ethyltoluene
P16
37.58
unidentified
P17
38.27
unidentifie d
P18
38.81
1,2,4-Trimethylbenzene
P19
39.31
1,2,3-Trimethylbenzene
P20
44.06
unidentified
P21
44.66
Napthalene
P22
49.21
unidentified
P23
53.21
unidentified


2
Decreases in the concentration of contaminants measured
in environmental samples can occur as a result of various
attenuation mechanisms. These include biodegradation,
volatilization, photooxidation, and dispersion. In the
subsurface, losses from pathways like photoxidation are not
important. Other pathways like volatilization occur at
rates slower than those measured from exposed surfaces.
Aerobic biodegradation can occur in the subsurface providing
adequate oxygen and nutrients are available and that the
contaminants are not present in concentrations which are
toxic for microorganisms.
The major objectives of this research include
determining fuel/water partitioning patterns and measuring
chemical degradation rates to aid in the interpretation of
data from contaminated ground water sites. Field
investigations of sites contaminated by gasoline or
chlorinated solvents typically analyze and report the
presence of constituents which are regulated by the state or
federal government (e.g. priority pollutants). These
components are emphasized in my research.
Many chlorinated organic compounds will degrade in
water by hydrolysis or elimination mechanisms. Due to the
extended residence times of organic pollutants in ground
water, this typically slow abiotic degradation within months
or years can be a significant attenuation mechanism. The
focus of my research on the halogenated solvents is on the


5
73
<
O
I
O
o
\
o
c

Clay
&
+
Silica Gel
4
4 -
0
Milli Q Water

3-
A
Sand
+
6
l 1
2-
s

A

1 -
o
* .* .
o
0 <
1

, 1 1
i 1
~1 1 1 1 1 1
O 20 40 60 80 100 120 140
Time (hours)
Figure 16. Effect of the presence of solid material on the
rate of hydrolytic degradation of TCA at 65C.
6
5
4
3
2
1
0
0 100 200 300 400
Time (hours)
Clay
+ Silica Gel
Limestone
A MilliQ Water
X

* Sand

0
Â¥
t h 1
J.:
Â¥
A

i 1 1 1 1 1 r
Figure 17. Effect of the presence of solid material on the
rate of hydrolytic degradation of 1-chloropropane at 65C.


94
Limitations of the Model Assumptions
This type of model does not address the slower rate of
removal of residual NAPL which would occur from a pool of
excess solvent reaching bedrock. In that case, the surface
area is much smaller and the process is limited by the rate
of diffusion from the surface of the pool. This is expected
to considerably lengthen the time that residual solvent is
present at the source (Schwille, 1988).
In an actual site, complete mixing and equilibrium
conditions will not be maintained over time. The NAPL may
be trapped in pores which have minimal contact with the
water. Although much of the NAPL may be removed through
dissolution and advection, some of the residual NAPL will
remain out of the major flow pathways for the water. Losses
would be limited by diffusion out of these regions and
trends in this zone may represent the "no flow" conditions.
This model addresses relatively small percentage of the
pore space occupied by residual NAPL while theoretically, up
to 40% of the pore space may be occupied by a NAPL. In that
case, the slower degradation of TCA in the NAPL may
contribute to the overall attenuation.


118
level of 10% for ethanol and t-BA would result in a volume
fraction cosolvent in the aqueous phase of 0.005 (aqueous
mole fraction of less than 0.002) for the 1:20 extraction
procedure used in these experiments. Groves (1988) measured
an increase in benzene solubility of about 18% when the
ethanol mole fraction in water was approximately 0.025
compared to the benzene solubility in water. Therefore,
using my experimental conditions a cosolvent effect would
not be significant. Assuming the only source of alcohol was
from the gasoline, an alternate water/gasoline extraction
ratio of less than 1:1 would be required to show a
measurable effect of cosolvency for the alcohols.
Additives which are only partially miscible in water,
like MTBE, will have lower aqueous concentrations. The
solubility of MTBE in water is 48 g/1 or a maximum volume
fraction cosolvent less than 0.065. In addition, it will
partition into the fuel or solvent layer reducing the
aqueous phase concentration and decreasing the effectiveness
as a cosolvent.
The fuel/water partition coefficients developed during
this study provided a reasonable estimate of the
partitioning behavior of specific constituents regardless of
the gasoline composition. The selection of the 1:20 ratio
for the water extraction procedure followed the work of
Coleman et al. (1984); however, other ratios were reported
in the literature. Brookman et al. (1985a) used a 1:10 fuel


127
Figure 33. Theoretical change in relative aqueous
concentration of selected gasoline constituents with
repeated aqueous extractions of the fuel.


ACKNOWLEDGMENTS
I sincerely appreciate the technical and editorial
assistance provided by my research director, Dr. J. Delfino.
I also thank Dr. P. S. C. Rao and Dr. P. Chadik for their
advice and for providing opportunities for challenging
discussions, and Dr. J. Dorsey and Dr. R. Yost for serving
on my committee and reviewing this dissertation. Each
member of my committee has contributed to my graduate career
through excellent teaching and creating a positive
intellectual environment.
This work was funded by grants from the Florida
Department of Environmental Regulations. Special thanks are
extended to Dr. Geoffrey Watts for his role in securing
funds and providing technical support and comments.
I am grateful to Dr. M. Battiste for discussions of
reaction mechanisms, and for providing the use of his
laboratory for the synthesis of brominated ethanes.
Special thanks go to Angie Harder for her hard work,
Linda Lee for her generosity with analyses and information,
Tom Potter for unselfish computer and mathematical
assistance, and Bill Davis for technical support.
I extend warmest and deepest thanks to my husband Ken
for technical assistance and emotional support, and my son
Brendan for giving me joy.


193
Watts, G., Groundwater Monitoring Parameters and
Pollution Sources, Ground Water Training Workshop,
Department of Environmental Regulation, Orlando, FL., April
2-3, 1986.
Westrick, J. J., Mello, J. W., and Thomas, R. F., The
Groundwater Supply Survey, Journal of the American Water
Works Association. 76: pp. 52-59 (1984).
Wilson, J.L., and Conrad, S.H., Is Physical Displacement
of Residual Hydrocarbons a Realistic Possibility in Aquifer
Restoration?, in Proceedings of NWWA/API Conference on
Petroleum Hydrocarbons and Organic Chemicals in Ground
Water: Prevention. Detection and Restoration. Dublin, Ohio
National Water Well Association, pp. 274-298, 1985.
Wolfe, N. L., Zepp, R.G., Paris, D. F., Baughman, G. L.,
and Hollis, R. C., Methoxychlor and DDT Degradation in
Water: Rates and Products. Environmental Science and
Technology. 11(12): pp. 1077-1081, 1977.
Wolff, D. D., and Parsons, M. L., Pattern Recognition
Approach To Data Interpretation. New York: Plenum Press,
1983 .
Youngless, T. L., Swansiger, J. T., Danner, D. A., and
Greco, M., Mass Spectral Characterization of Petroleum Dyes,
Tracers, and Additives, Analytical Chemistry. 57: pp.
1894-1902, 1985.


SUMMARY AND CONCLUSIONS
This research focused on two major categories of ground
water contaminants, chlorinated solvents and gasoline.
Halogenated organic compounds were examined to determine
degradation mechanisms and and pathways. The partitioning
of gasoline into water and the variability in the water
extracts of gasolines were evaluated.
Chlorinated Solvents
The degradation rate of TCA was compared with the
degradation of other 1,1,1-trihaloethanes, 1,1-
dichloroethane and chloropropane. The brominated analogs of
TCA degraded 10-14 times faster than TCA. As the number of
bromines increased, the percentage of the elimination
product increased.
The mechanism for degradation of the trihaloethanes was
SN1. The formation of a reactive carbocation intermediate
on this type of halogenated alkane was supported by the
following evidence.
1. These molecules are sterically hindered making
nucleophilic attack by an SN2 mechanism difficult as shown
for chloroform.
2. An E2 elimination reaction requires a weakening of
166


137
supported by comparison of mass spectra and concentrations
measured by GC/MS analysis. Area counts were obviously
related to concentrations and were used without conversion
since not all response factors were available. The data
evaluation presented in this section is, therefore,
dependent on the protocol used for the analysis. Focus
could be placed on only the peaks which have been identified
and which show very little evidence of coeluting peaks;
however, this may eliminate from the analysis peaks which
may distinguish the gasoline types.
After further examination the data from the 23 peaks,
the two earliest eluting peaks were deleted from the
statistical analysis. These earliest eluting peaks showed
more variation than some of the later peaks, and the
concentrations measured were not as reliable because
(1.) there was considerable peak overlap in this area of the
chromatogram; (2.) methanol and acetone, common laboratory
solvents which were used for rinsing sampling equipment,
elute in this region; (3.) these early eluting components
represent the most volatile constituents, making
reproducible extraction and quantification more difficult;
and (4.) the area counts of these peaks were random and
highly variable for each brand and grade. The final data
base consisted of 21 peaks, designated P3 to P23, which were
evaluated for the 65 gasoline samples.


coefficients were highly correlated with the pure component
solubilities .
Chemometric techniques were applied to 20 peaks
measured in the aqueous extracts of the 65 gasolines.
Bivariate plots and principal component analyses show
selected brands have distinguishing equilibrium
concentrations, but complete separation of brands was not
observed.
vi i


Ln k (sec
1000/T (K)
Figure 6. Arrhenius plot for the abiotic degradation of 1,1,1-trichloroethane.


189
Fu, J.K. and Luthy, R. G., Aromatic Compound Solubility
in Solvent/Water Mixtures. Journal of Environmental
Engineering, 112(2): pp. 328-345, 1986.
Gordon, A. D. Class ification. London, England, Chapman
and Hall, 1981.
Graf, C.G. "VOC's in Arizona's Groundwater; A Status
Report," paper presented at the NWWA FOCUS Conference on
Southwestern Ground Water Issues, Tempe, Arizona, September
1986 .
Greim, H., Wolff, T., Holfler, M., and Lahaniatis, E.,
Formation of Dichloroacetylene from Trichloroethylene in the
Presence of Alkaline Material Possible Cause of
Intoxication after Abundant Use of Chloroethylene-Containing
Solvents. Archives of Toxicology. 56: pp. 74-77, 1984.
Groves, F. R. Jr., Effect of Cosolvents on the
Solubility of Hydrocarbons in Water, Environmental Science
and Technology. 22: pp. 282-286, 1988.
Haag, W.R., and Mill, T., Effect of Subsurface Sediments
on Hydrolysis of Haloalkanes and Epoxides, Environmental
Science and Technology. 22: pp. 658-663, 1988.
Harder, A. M., A Study on Gasoline and its Behavior in a
Soil/Water Environment. M.S. Thesis, University of Florida,
1987 .
Hie, J., Carbon Dichloride as an Intermediate in the
Basic Hydrolysis of Chloroform, A Mechanism for Substitution
Reactions at a Saturated Carbon Atom, Journal of the
American Chemical Society. 72: pp. 2438-2445, 1950.
Hosenfeld, J. M., and Bauer, K. M., Application of
Pattern Recognition to High-Resolution Gas Chromatographic
Data Obtained from and Environmental Survey, in
Environmental Applications of Chemometrics. edited by Breen,
J. J. and Robison, R. E., Washington, D.C., American
Chemical Society, pp. 69-82, 1985.
Jones, D. M., Douglas, A. G., Parkes, R. J., Taylor, J.,
Giger, W., and Schaffner, C., The Recognition of Biodegraded
Pe troleum-Derived Aromatic Hydrocarbons in Recent Marine
Sediments, Marine Pollution Bulletin. 14: pp. 103-108,
1983.
Jones, S. C., Some Surprises in the Transport of Miscible
Fluids in the Presence of a Second Immiscible Phase,


59
groups. The rate coefficient at 25C is approximately 10^
faster than for TCA.
Queen and Robertson (1966) examined the hydrolysis of
2,2-dihalopropanes. These compounds form carbocation
intermediates with two methyl and one halogen group. The
rate coefficients for the degradation of 2,2-dihalopropanes
are intermediate between t-butyl chloride and the 1,1,1-
trihaloethanes. The mechanism was reported to be SN1/E1
based on results of experiments with deuterated gem-
dihalides. The degradation rates of these compounds were
10-50 times higher than of the corresponding secondary
halides (e.g., 2-chloropropane).
The degradation rates were affected by the leaving
group, bromine or chlorine. Also, the structure and
stability of the resulting carbocation affected the rate and
pathway (elimination and/or substitution) of the reaction.
Since bromine was a better leaving group than chlorine,
there was a rate increase when bromine was present as
compared to the corresponding chlorinated compound. 2,2-
Dibromopropane degraded 19 times faster than 2,2-
dichloropropane, while 2 bromo 2-chloropropane degraded 5
times faster than the dichloro compound (Queen and
Robertson, 1966). The 1,1,1-trihaloethanes containing
bromine degraded 11-13 times faster than TCA.
Rates were also increased as the number of methyl
groups present on the carbocation increased. The t-butyl


c¡ o
177
C COMPUTE DETERMINAT
D E T M
PRTDRV(1,1)*PRTDRV(2,2)-PRTDRV(1, 2)*PRTDRV(2,1)
C
COMPUTE INVERSE
PRTDRV(1,1)
PRTDRV(1,2)
PRTDRV(2,2)
PRTDRV(2,1)
PRTDRV(2,2)/DETM
-PRTDRV(2,1)/DETM
PRTDRV(1,1)/DETM
-PRTDRV(1,2)/DETM
SOLV LINEAR EQNS
AN INCREMENT CONC
DO 25 I-l.NP
DELTA-0
F C A L C
(CONC(I,2)+CONC(NP-I+l,2)+SOLUB(I))*FRACM(I)-CONC(I,1)
DO 20 J-l.NP
DELTADELTA-FCALC*PRTDRV(I,J)
20 CONTINUE
C0NC(I,2)-C0NC(I,2)+DELTA
25 CONTINUE
CALL DENM(NP,CONC,2,DENOM)
DO 30 K-l.NP
FRACM(K)- CONC(K,2)/DEN0M
CONC(K,3) CONC(K,l)-(CONC(K,2))
30 CONTINUE
CALL ERRCHK(NP,SOLUB,CONC,FRACM,ERRCRI,ISERR)
IF(ISERR.EQ.0) THEN
RETURN
ENDIF
C WRITE(*, (I4.6F9.4)') N, ((CONC(I J ) J = l,3),
1-1,NP)
50 CONTINUE
WRITE(*,*) 'ITERATED 50 TIMES WITHOUT CONVERGENCE
PAUSE
RETURN
END


64
DMDE. The hydrolysis products formed were anisoin and
anisil, which were explained by phenyl group rearrangement
after the formation of the carbocation.
Mochida et al. (1967) showed 1,1,1,2-tetrachloroethane
and pentachloroethanes reacted more slowly than TCA under
lower pH conditions, which indicated that chlorines on the
beta carbon decrease the stability of the carbocation. The
presence of these chlorines on the beta carbon however,
increased the acidity of the hydrogens, with enhanced
degradation rates for the tetra and pentachloroethanes by an
E2 mechanism at elevated pH.
There is considerable evidence that geminal trihalides
can form carbocations in the presence of an appropriate
neighboring group. Subsequent reaction pathways may vary
according to the structure of the carbocation resulting in
elimination, substitution, or rearrangements. An E2
reaction may also occur for compounds containing an acidic
hydrogen on the beta carbon.
Effect of Additional Halogens on the Alpha Carbon
The hydrolysis of a simple primary halide, 1-
chloropropane, was compared with the reactivity of 1,1-
dichloroethane and TCA in experiments I performed at
elevated temperature. As the number of hydrogens on the
alpha carbon decrease, steric hindrance can increase and
result in a shift in reaction mechanism. The experiments
were designed to demonstrate the relative rates of


PROGRAM TCA2
2
3
4
PARAMETER
DIMENSION
DIMENSION
DIMENSION
CHARACTER
WRITE(* *) 'ENTER
EXTENSION ',
1 '(MAX 20 CHAR)'
WRITE(*,*) '=->
WRITE ( * ) '>'
READ(*,2) FNAME
WRITE(*,*) 'ENTER
EXTENSION) ',
1 (MAX 20 CHAR)'
WRITE(*,*) '- >
WRITE(*,*) '>'
READ(*,2) OUTNM
FORMAT (A)
OPEN(5,FILE-FNAME,
OPEN(8,FILE-OUTNM,
(NP=2)
CONC(NP,3),SOLUB(NP)
FRACM(NP),ERRCRI(2)
DEGRT(NP)
FNAME*2 0,OUTNM*2 0
DATA FILENAME
(OR PATH)
OUTPUT FILENAME(OR PATH)
STATUS-'OLD')
STATUS-'UNKNOWN')
WRITE(8
READ(5,
READ(5,
WRITE(8
READ(5,
READ(5,
WRITE(8
*)
*)
(CONC(1,1) ,
1=1,NP)
,3)
(CONC(1,1)
, 1=1,NP
*)
*)
(SOLUB(I),
1=1,NP)
,3)
(SOLUB(I),
1-1,NP)
*)
*)
(ERRCRI(I),
1=1,NP)
,4)
(ERRCRI(I)
, I1,NP
*)
*)
DT,DAYS
,3)
DT,DAYS
*)
*)
FLOWRT
A)
FLOWRT
*)
*)
(DEGRT(I),
1=1,NP)
WRITE(8,4) (DEGRT(I)
FORMAT (6F9.2)
FORMAT (3X,2F9.5)
WRITE(8,*)
WRITE(8,*)
WRITE(8,*) DAY
DCETOT' ,
1' DCESOL DCEWTR'
1=1,NP)
TCATOT
TCASOL
INCLUDING
(OPTIONAL
TCAWTR
174


1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
182
P6
P7
P8
P9
P10
Pll
Benzene
Toluene
Ethylbz
m,p-Xylene
o Xylene
53
.97
210 .
.89
9 .
19
28
.38
13 .
65
1 .
10
40
. 82
80 .
. 53
6 .
71
17
. 02
14 .
46
1.
10
94
.83
183 .
.21
12 .
25
36
.73
17 .
59
1.
48
97
. 12
192 .
. 94
13 .
34
39
.52
18 .
93
6 .
88
87
. 03
165 .
. 59
10 .
95
34
. 00
16 .
42
1 .
68
57
. 36
250 ,
.77
8 .
55
27
. 44
13 .
23
1.
20
70
. 69
192 .
, 30
10 .
83
33
. 14
16 .
37
2 .
81
72
. 62
201.
.11
10.
67
34
.01
16 .
62
1.
54
53
. 23
490 .
. 17
6 .
62
17
.83
8 .
58
0 .
78
52
. 99
491 .
.95
6 .
57
18
. 06
8 .
64
0 .
69
38
.46
578 ,
.55
9 .
04
25
. 26
11 .
76
2 .
54
54
. 06
451 .
.66
10 .
31
30
.22
14 .
99
1 .
50
137
. 24
145 .
. 58
6 .
15
18
.52
9 .
31
0 .
56
139
. 47
143 .
. 06
5 .
53
17
.33
8 .
51
1 .
58
202
. 64
264.
.08
8 .
84
25
. 50
12 .
34
2 .
76
246
. 92
190 .
. 34
6 .
85
17
.57
12 .
24
0 .
59
278
.78
204 .
. 64
8 .
22
28
. 90
30 .
42
1.
02
137
. 98
141 .
. 54
5 .
04
18
.13
10 .
37
1.
08
175
. 51
198 .
. 28
8 .
12
30
. 50
16 .
79
0 .
97
171
. 62
185 .
.59
7 .
39
29
.58
16 .
34
1.
65
86
. 55
153 .
.75
4 .
78
17
.89
9 .
30
1.
74
67
. 22
267 ,
. 18
8 .
90
34
. 20
2 .
76
15 .
80
60
. 50
232 .
. 18
7 .
20
27
. 15
12 .
69
1.
06
85
. 39
123 .
. 54
6 .
30
24
.26
11 .
16
0 .
74
111
. 04
174.
.87
9 .
55
36
. 76
17 .
30
0 .
80
116
. 51
101 .
.27
6 .
78
27
. 70
13 .
33
0.
92
57
.07
115 .
. 26
7 .
03
25
. 35
12 .
16
0.
00
114
. 16
137 .
. 22
6 .
97
27
.33
12 .
55
0 .
98
77
.57
84 .
.83
6 .
20
22
. 80
10 .
14
0 .
62
96
. 12
155 ,
. 09
6 .
26
29
. 37
13 .
75
1 .
02
140
.65
202 ,
.81
10.
31
38
. 11
21 .
15
2 .
86
138
.68
196 .
.55
10 .
11
38
.86
21.
81
0 .
59
97
. 48
143 .
.95
8 .
53
34
.65
16 .
00
0 .
78
106
.91
171 .
.81
10 .
81
41
.52
19 .
21
0 .
94
133
.77
249 .
.17
13 .
00
45
. 84
20 .
64
0.8862
107
.11
153 .
. 26
10.
05
36
.42
2 .
35
16 .
57
122
.23
194 ,
. 64
12 .
27
35
. 00
15 .
86
0 .
88
124
.87
200 ,
. 44
12 .
47
34
. 82
16 .
12
0 .
80
69
.63
112 .
,41
9 .
05
20
. 75
12 .
39
0 .
97
107
.91
166 .
.31
9 .
82
35
.65
16 .
94
2 .
61
102
. 30
172 ,
.68
10 .
77
35
.53
16 .
50
1 .
75
98
.03
145 .
. 80
8 .
61
30
. 45
14 .
76
2 .
00
133
.16
206 .
. 11
11 .
70
44
. 17
19 .
59
2 .
56
146
.62
197 .
.51
11.
11
47
.22
21 .
54
3 .
15
168
. 84
238 .
.53
13 .
85
56
.83
25 .
34
2 .
95
162
. 24
238 .
.83
13 .
94
60
.26
26 .
84
1 .
51
129
.48
157 .
.83
17 .
94
45
.78
20 .
96
1 .
97


32
Included In the Arrhenius plot are the degradation rate
coefficient for TCA in a pH 13 buffer and also the rate
coefficients calculated by Walraevens et al. (1974) for the
sodium hydroxide solutions. The rates for these high pH
solutions were within the confidence interval for the
regression line, indicating the reaction rate was not
significantly accelerated in alkaline media. The lack of
change in the rate in the presence of a high concentration
of a strong nucleophile (i.e. OH") suggested that the
reaction with the nucleophile occurs after the rate
determining step, characteristic of SN1 reactions.
Similarly, the increase in base strength did not shift the
elimination to an E2 mechanism through a large rate increase
and/or increase in formation of the elimination product.
The rate data which exceeded the confidence interval of
the regression line (Figure 6) were from studies (Vogel and
McCarty, 1987; Pearson and McConnell, 1975) which estimated
the rates of the slow reactions with less than 50%
degradation of the parent compound occurring. Rate
constants calculated for low conversion are more variable
than rates established based on higher amounts of conversion
(Levenspiel, 1972, p. 85). The strong linear Arrhenius
relationship between temperature and rate observed between
25 and 80C, regardless of sample matrix, suggests that
reaction rates at temperatures below 25C can be estimated
by extrapolation.


INTRODUCTION
Liquids organic compounds ar frequent causes of ground
water contamination. Nonaqueous-phase liquids (NAPL) fall
into two broad categories based on their migration patterns
upon reaching ground water. Mineral oils, including crude
oils as well as various refined products like gasoline, are
less dense than water and move vertically through the
unsaturated (vadose) zone and tend to spread laterally upon
reaching the water table. The majority of spills involving
organic fluids which contaminate ground water result from
this group of compounds (Schwille, 1984).
In many industrialized countries, serious threats to
ground water supplies result from low molecular weight
chlorinated solvents. These anthropogenic substances are
more dense than water and vertical rather than lateral
movement dominates upon reaching the water table. The more
common solvents detected in ground water include 1,1,1-
trichloroethane (TCA), trichloroethene (TCE),
tetrachloroethene or perchloroethene (PCE), and various
dichloroethene isomers. In addition to common usage, the
high frequency of detection is attributed to the compounds'
high mobility and relatively high resistance to degradation.
1


78
In laboratory experiments, the initial concentration of
chlorinated solvent was at saturation concentration even
when the layer of sand with residual solvent was thin
(Schwille, 1988). The concentration gradually decreased
until the levels in the water were sufficiently low that
further removal of solvent was slow. At this point
approximately 86% of the residual had been removed.
My model was developed assuming that equilibrium
saturation was maintained, the dissolution of residual
solvent being faster than the degradation or advective
transport of components. Diffusion or hydrodynamic
dispersion was not considered to be a limiting factor in
maintaining equilibrium. The solvent contaminated zone was
then treated similar to a well-mixed flow reactor.
Interactions of the solutes in the water with the solid
matrix of the saturated zone were considered minimal
providing residual solvent was present; the porous medium
was assumed to provide a matrix in which the residual
solvent was retained.
Once the flow of the NAPL stopped, the subsequent
losses were assumed to occur through degradation or
advection of the compound in the aqueous phase.
Hydrolysis/elimination of TCA occurs much faster in dilute
aqueous solution than would occur for water dissolved in the
TCA solvent phase (Walraevens et al., 1974). Ground water
continues to flow through this zone, although at somewhat


14
with flame ionization detector (FID) which employed a 30 m
J&W DB-1, 0.53 mm i.d. wide bore capillary column with a 3
un stationary film thickness. The temperature program
included a 10 minute hold time at 30C and temperature
ramping of 5C/min to 80G. The helium flow was 2.5 mL/min.
Selected analyses were performed by gas chromatography/mass
spectrometry (GC/MS) for quantification and confirmation of
the formation of 1,1-DCE.
The brominated analogs of TCA were not commercially
available. These compounds were synthesized according to
the protocol described by Stengle and Taylor (1970). Two
hundred and fifty milliliters of carbon disulfide (CS2) were
added to a 500 mL, 3-neck flask that was saturated with HBr
vapors at 0C. Excess vapors were trapped over aqueous KOH.
Five milliliters of TCA were added. Five grams of aluminum
bromide (AlBr3) were added to 100 mL anhydrous CS2, placed
in a dropping funnel, and gradually added to the TCA/CS2/HBr
solution over a period of one hour.
This solution was extracted with ice water made basic
with ammonium hydroxide. The solvent was then removed by
distillation and the residue was filtered. An aliquot of
the mixture was added to methanol and spiked into ampules
containing water. Analysis by purge and trap GC showed two
primary peaks and a secondary peak. The major peaks were
determined by GC/MS to be tribromoethane and
dibromochloroethane. A smaller peak was shown to be


GASOLINE IN GROUND WATER 95
Background 95
Composition of Gasoline 95
Multicomponent Liquid-Liquid Equilibria ... 97
Statistics and Pattern Recognition
Applications 102
Partitioning of Gasoline Components into Water . 105
Fuel/Water Partition Coefficients 105
Water Soluble Blending Agents 113
Prediction of Kfw for Other Components . 120
Changes in Concentrations with Time 122
Differences in Water Extracts of Gasolines .... 129
Equilibrium Concentrations of Major
Constituents 131
Visual Comparison of Water Extracts of
Gasoline 131
Preparation of the Data Base for Statistical
Analysis 135
Basic Descriptive Statistics 138
Bivariate Plots 141
Stepwise Discriminant Analysis 148
Principal Component Analysis 155
Summary 164
SUMMARY AND CONCLUSIONS 166
APPENDIX A. SOLUBILITY MEASUREMENTS BY LINDA LEE . 171
APPENDIX B. FORTRAN PROGRAM FOR MODELING LOSS OF
RESIDUAL TCA 173
APPENDIX C. AREA COUNT DATA SET FOR STATISTICAL ANALYSIS
OF WATER EXTRACTS OF GASOLINE 179
REFERENCES 187
BIOGRAPHICAL SKETCH 194
iv


DEGRADATION OF ALKYL HALIDES
Introduction
In this section the degradation kinetics for 1,1,1-
trichloroethane (TCA) and other 1,1,1 -trihaloethanes will be
presented and discussed. These compounds degrade in water
forming both elimination and substitution products.
Specific experiments were performed to determine the
mechanism of this reaction and to describe factors which may
effect the rate or pathway of the degradation.
Mechanisms of hydrolysis/elimination have been studied
for many years and numerous reviews, textbook chapters and
empirical concepts have been developed to describe the
chemical degradation of alkyl halides in water. The
following review provides the framework for subsequent
discussions of alkyl halide structure and reaction
mechanisms where specific examples will be presented. The
information was synthesized from several sources (March,
1985; Carey and Sundberg, 1984; Mabey and Mill, 1978;
Bentley and Schleyer, 1977).
Classical SN1, SN2, El and E2 mechanisms have been
defined as early as 1933 (Figure 4). The distinction
between SN1 and SN2 is whether or not the nucleophilic
22


140
Identified. The Pearson product-moment correlation is
calculated by the formula
rxy S (x x)(y-y)/ J S(x-x)2 S(y-y)2
The correlation matrix is shown in Table 23. This type of
intervariable correlation indicated the strength of linear
relationships between any two variables. Values of 1 or -1
indicated complete correlation, and as values approached
zero, more scatter in a plot of the two variables was
expected. If two variables were highly correlated then one
variable could be expressed as a linear function of the
other variable, although this does not necessarily imply a
cause and effect. This approach did not address nonlinear
relationships that may exist. Values midrange between
complete correlation or total scatter were difficult to
interpret and would require further evaluation. Higher
values for the correlations (greater than 0.75) are
presented in bold type.
This analysis revealed fairly high correlation (0.78 to
0.90) between peaks P8, P9, and P10 which represent
ethylbenzene, m,p-xylene, and o-xylene respectively.
Similar correlations were not observed for benzene and
toluene (P6 and P7). Although the ratios of benzene and
toluene may be fixed for a particular type of gasoline,
differences in processing may result in varying composition
for different types of gasolines.


Table 15. Fuel Water Partition Coefficients as
measured for 31
Average
gasolines.
Standard
Deviation
S olubi1ity
(mg/L)
Benzene
350
75
1740
Toluene
1250
180
530
Ethylbenzene
4500
600
160
m,p-Xylene
4350
530
146,156
o-Xylene
3630
420
170
n-Propylbenzene
18500
5600
55
3-, 4-Ethy1 toluene
12500
2350
40
1,3,5-Trime thylbenzene
10200
2350
48
2 Ethyl toluene
10300
2100
40
1,2,3-Trimethylbenzene
13800
3980
75
*
Brookman et al., 1985a


23
Step 1.
Step 2.
Step 2.
OH- +
OH
Unimolecular Mechanisms
-C-C-X
i i
i i
-C-C +
I I
+ X
C C+ + OH
i i
i i
-C-C-OH
i i
SN1
i i
-C-C +
I
H
\
/
/
c = c
\
El
Bimolecular Mechanisms
C-X
\ /
HO C X
HO-C- + X
SN2
Figure 4. Classical substitution and elimination reaction
mechanisms for degradation of alkyl halides in water.


8
112
7-
6-
5-
4-
3-
2-
1 -
O
i 1 1 1 1 1 1 1 1 1 1 1
4 8 12 16 20 24
m.p-Xylene (mg/I) in aqueous solution
80
70
60
if)
> 1 50
§
^ I 40
30
20
10
10 14 18 22 26 30 34 38
1,3,5-Trimethylbenzene (g/l) neat gas
Figure 30. Fuel/Water partition coefficient (Kfw) as a
function of concentration of 1,3,5-trimethylbenzene and
m,p-xylene.


15
bromodichloroethane. Trichloroethane was below detection
levels ( <30 ng/L )in these analyses.
Some of the spiked ampules were heated for a few hours
to determine if halogenated ethenes would be formed, and if
so, to subsequently determine their corresponding retention
times. Two major peaks were identified by GC/MS to be 1,1-
dibromoethene and 1-bromo-1 -chloroethene. A sample GC
chromatogram containing reactants and products is shown in
Figure 2, with mass spectra of TBA and DBE in Figure 3.
The same analytical conditions were used for the
brominated compounds as were used for TCA, although the
final temperature was slightly increased.
Pure standards of these compounds were not available
for quantification. The degradation rate was determined
directly from the decrease in area counts, since the
response of the external standard remained consistent during
the time of the experiments. However, determination of
molar concentrations was required to determine the percent
transformation to the elimination product.
The response on an FID is generally related to the
number of carbons and can be affected by functional groups.
To determine if the molar response on the FID was affected
by the type of halogen on the molecule (bromine or
chlorine), I examined the response of the trihalomethane
series for which standards were available (Table 3). The
molar response on the FID was the same for this series of


o+
X
+
X = C X *+ x c
I 1
R R
Figure 13.
Stabilization of carbocations by halogen (Olah, 1974).
+x


63
rate increase was much greater than the factor of 50
reported by Mabey and Hill (1978).
Quemeneur et al. (1971) determined that tri-chloro
compounds of the type P-RC6H4-CCI3 (R is OMe, Me, H, Cl, or
NO2) were hydrolyzed in neutral or acidic medium via a
cationic transition state for all types of R substituents.
The hydrolysis of the p-substituted alpha,alpha-
dichlorotoluenes reacted via a cationic mechanism when R is
an electron-donor, and a bimolecular mechanism when R is an
electron-attracting group. These results also supported the
observation that halogens contributed to the stability of
the carbocation. Monochlorotoluene reacts nearly 3000 times
more slowly by an SN2 mechanism than the trichlorotoluene
reacts by the SN1.
Methoxychlor and DDT are two environmentally important
pesticides which contain a geminal trihalide functional
group. Wolfe et al. (1977) provided an in depth examination
of the degradation of these compounds. There is a beta
hydrogen on each of these compounds. At elevated pH the
degradation rate increased as a function of pH and the
elimination products were dominant, suggesting these
structures were more susceptible to degradation by the E2
mechanism than is TCA. While the elimination product, DDE,
was the major product of DDT hydrolysis even at lower pH,
the major product of methoxychlor at pH 7 was the hydrolysis
product, with minor amounts of the elimination product,


BIOGRAPHICAL SKETCH
Patricia VanOvermeer Cline obtained a B.S in chemistry
from the University of Michigan in May, 1970. She worked as
an Honorarium Instructor in chemistry at the University of
Colorado, Denver Extension, and received a degree in
secondary education in May, 1974 from the University of
Colorado in Boulder. She taught secondary science in
Longmont, Colorado, for four years.
In Madison, Wisconsin, she directed analytical and
field services for Warzyn Engineering, Inc. She served as a
project manager and environmental chemist for studies of
waste treatment and characterization as well as ground water
contamination investigations.
In 1984 she moved with her husband, Ken, and her son,
Brendan, to Gainesville, Florida. She received her Masters
of Science degree in water chemistry from the Department of
Environmental Engineering Sciences at the University of
Florida in May, 1987 and is currently receiving her Ph.D. in
water chemistry. She plans to continue working in
Gainesville as an environmental consultant.
194


192
Rubino, J. T., and Yalkowsky, S. H., Cosolvency and
Deviations from Log-Linear Solubilization. Pharmaceutical
Research. 4(3): pp. 231-236, 1987.
Saltzman, S., and Mingelgrin, U., Nonbiological
Degradation of Pesticides in the Unsaturated Zone, in
Pollutants in Porous Media. edited by Yaron, B., Dagan, G.,
Goldshmid, J. Springer- Verlag, New York, NY. 1984.
Sanemasa, I., Miyazaki, Y., Arakawa, S., Kumamaru, M.,
and Degughi, T., The Solubility of Benzene Hydrocarbon
Binary Mixtures in Water. Bulletin of the Chemical Society
of Japan. 60: pp. 517-523, 1987.
SAS Institute Inc., SAS User's Guide: Statistics. Cary,
NC: SAS Institute Inc., 1985.
Schwille, F., Migration of Organic Fluids Immiscible
with Water in the Unsaturated Zone. In: Pollutants in Porous
Media. Yaron, B., Dagan, J., Goldshmid, J. (eds.),
Springer-Verlag, New York, 1984.
Schwille, F., Dense Chlorinated Solvents in Porous and
Fractured Media. Lewis Publishers, Chelsea, Michigan, 1988.
Senn, R. B., and Johnson, M. S., Interpretation of Gas
Chromatographic Data in Subsurface Hydrocarbon
Investigations, Ground Water Monitoring Review. 7(1): pp.
58-63, 1987.
Stengle, T. R., and Taylor, R. C., Raman Spectra and
Vibrational Assignments for 1,1,1-Trihaloethanes and Their
Deuterium Derivatives, Journal of Molecular Spectroscopy.
34: pp. 33-46, 1970.
Swain, C. G., and Scott, C. B., Journal of the American
Chemical Society. 75: pp. 141-143, 1953.
Verschueren, K., Handbook of Environmental Data on
Organic Chemicals. Van Nostrand Reinhold Co., New York,
1977 .
Vogel, T.M., and McCarty, P. L., Rate of Abiotic
Formation of 1,1-Dichloroethylene from 1,1,1-Trichloroethane
in Groundwater, Journal of Contaminant Hydrology. March
1987 .
Walraevens, R., Trouillet, P., and Devos, A., Basic
Elimination of HC1 from Chlorinated Ethanes, Internation
Journal of Chemical Kinetics 6: pp. 777-786 (1974).


89
YEARS (Flow, 0.25%/Day)
Figure 22. Model results: Change in total mass of TCA as a
function of initial mass of TCA and composition of the
solvent phase.
YEARS (Flow, 0.25%/Doy)
Figure 23. Model results: Pattern of 1,1-DCE formation and
advection as 100 mmoles of TCA in the residual zone
degrades.


1000/T (K)
Figure 10, Arrhenius plot for the abiotic degradation of 1,1,1-trihaloethanes.
p'
o


117
cn
E
o
120
110-
100O
90-
80 -
70-
60-
50-
40 Jl
30-
20-
10
0
0
Benzene
Toluene
mpXylene
+
+
800 1600 2400 3200
Ethanol in aqueous phase (mg/l)
4000
140
120
100
\
I 80
60
40
20
0
Figure 31. Equilibrium concentrations of major gasoline
constituents in water as a function of the concentration of
oxygenated additive (ethanol or t-BA) in the aqueous phase.

Benzene
+
Toluene
o
mp-Xylene
-
+ +
+

+ + +
-
+
Jl

-

o
o
o
o
o
o
o
0
lili
1000
I I 1 1 1 1
2000 3000
4000
TBA in aqueous phase (mq/l)


-Xylene
Ethylbenzene
Figure 37. Bivariate plot of area counts of m,p-xylene and ethylbenzene for aqueous
extractions of 65 gasoline samples.
143


108
significantly different (either higher or lower) than
samples without this additive.
Fuel/water partition coefficients were expected to show
greater variation than parameters like octanol/water
partition coefficients because of differences in the
gasoline compositions as previously discussed. The
usefulness of the measured Kfw to estimate the partitioning
of specific gasoline components into water was evaluated.
The 31 measurements of Kfw for the gasoline samples
were examined to evaluate the variability in the coefficient
(Table 15). The variability would determine the general
usefulness for estimating partitioning behavior. The
reported measurements were based on my extraction protocol,
room temperature (ca.22C) and a 1:20 fuel to water ratio.
The relative deviation in Kfw for the 31 samples varied
between 11.5 and 30.0%, while the Kfw 's for the 10
components varied over two orders of magnitude. This is
rather consistent considering the wide variations in the
compositions of the gasolines. To put this in perspective,
Log Kow has been more precisely defined with only a three-
component system and fixed low solute concentration and
Lyman et al. (1982) state "it is frequently possible to
estimate log Kow with an uncertainty of no more than plus or
minus 0.1-0.2 log Kow units."
There were analytical factors which affected the
results. The complex nature of the gasoline mixture created


165
17 types. Specific subsets of the data could be separated
on the basis of the concentrations of the 21 peaks used in
the analysis.
This was further illustrated with principal component
analysis. Differentiation of two brands can sometimes be
accomplished using the 21 components, however this was not
always successful.


55
while benzyl groups enhance the rate by a factor of 50.
(Mabey and Mill, 1978)
The formation of stabilized carbocations by electron
donation from the non-bonded electron pairs of halogens
adjacent to the cationic carbon center have been reported
(Olah, 1974). The stabilizing effect was enhanced when two
or even three electron-donating heteroatoms coordinate with
the eleetron-deficient carbon atom as illustrated in Figure
13. Specific examples, designated as "chlorocarbenium
ions" by Olah (1974), have been identified and are
illustrated in Figure 14.
Simple SN1/E1 Reactions
My data suggested 1,1,1-trihaloethanes form carbocation
intermediates. The intermediate would contain two halogens
and one methyl group. The observed rates and pathways are
compared (Table 9) to compounds containing two methyl groups
and one halogen (2,2-dihalopropanes) and three methyl groups
(t-butyl chloride).
Degradation of tertiary halides like t-butyl chloride
occurs with a carbocation intermediate and these compounds
are resistant to bimolecular nucleophilic displacement. The
half-life for the aqueous degradation of t-butyl chloride is
approximately 23 seconds at 25C with about 19% of the
degradation occurring through the elimination pathway. The
carbocation intermediate is stabilized by the three methyl


164
gasolines. Although all gasoline brand pair combinations
did not separate, some brands like Amoco and Shell appear to
have specific processing which provide a unique composition
of major constituents in the water soluble extracts.
Summary
A data base consisting of 21 of major components
detected in GC/FID gas chromatograms of water extracts of 65
gasoline samples was compiled and evaluated. The gasoline
samples represented six brands and four grades. The samples
were collected over a period of six months from different
sources (gas stations or terminals) to maximize the
variation within any particular brand and grade. Specific
name brands were evaluated, the problem would become more
complex for independant stations which may purchase gasoline
from a variety of sources.
The average aqueous equilibrium concentration (mg/1)
was 42.6 for benzene, 69.4 for toluene and approximately 17
for the xylenes. The concentrations of these particular
constituents may vary by as much as one order of magnitude.
There was some correlation among a number of the
aromatic components like ethylbenzene and xylenes. The
ratios of the concentrations of these components typically
lie within a particular range.
Stepwise discriminant analysis of these data show that
these parameters do not provide sufficiently unique
information about the gasolines to allow separation of all


Figure 14. Examples of Mchlorocarbenium ions" (Olah, 1974)
Ul


MATERIALS AND METHODS
Alkvl Halides
Reagent grade chemicals (Fisher Scientific) were used
to prepare buffers and standard solutions. Phosphate
solutions (0.05 M) were prepared at pH 4.5, 7.0 and 8.5 by
mixing stock solutions and monitoring the pH with a Fisher
Accuiet model 230A pH meter. Solutions of 0.05 M potassium
dihydrogen phosphate and 0.05 M potassium hydrogen phosphate
were prepared using distilled deionized water. Equal molar
volumes were used for the pH 7.0 buffer. The phosphate
solutions at pH 4.5 (potassium dihydrogen phosphate) and pH
8.5 (potassium hydrogen phosphate) required minor pH
adjustment using 0.05 M phosphoric acid or potassium
hydroxide solutions.
Stock standard solutions of TCA and 1,1-DCE were
prepared in methanol at concentrations of approximately 1
mg/mL. Working standards were prepared by spiking
approximately 5 iL of the stock standard solution into 10 mL
of distilled deionized water. Aliquots of 100-500 pL of.the
working standards were used to prepare standard curves for
the response of the gas chromatograph (GC) to the
concentration of analyte.
12


102
enhanced the solubility of higher molecular weight
hydrocarbons in seawater, however the application of this to
gasoline components in lower ionic strength solutions was
not reported. The solubilities of selected gasoline
constituents in water are summarized in Table 14.
Statistics and Pattern Recognition Applications
A large number of GC analyses of water extracts of
gasolines provided a large data base which became difficult
to characterize by graphical means alone. Various
statistical procedures were employed to describe
similarities and differences among gasoline brands
(commercial or trade name, e.g., Shell, Chevron, etc.) and
grades (e.g., regular, unleaded, super unleaded, etc.) and
to generally improve understanding of the data base.
The simplest solution to gasoline source identification
would be to find a single unique compound which repeatedly
and consistently identified a specific source. "Active
tagging", where a known chemical or physical label is added
to a gasoline, would be a superior method for making
absolute identification, but such procedures have not yet
been implemented in the gasoline market, although crude oils
are often tagged for identification purposes. "Passive
tagging" is a procedure that attempts to identify the fuel
source based on the natural composition of the product.
Fingerprinting, or the identification of fuel type
based on chromatographic patterns, has been successfully


42
though TCA will degrade, it is more stable than the
brominated compounds and it may accumulate to detectable
levels.
The BDCA compound had the lowest concentration in the
mixture of the three geminal trihalides in reaction
solution, and TCA concentration was less than 40 ug/1.
After three days of incubation at 37C, the concentration of
TCA rose to approximately 200 ug/1. This was a minor
pathway (less than 5% of the BDCA degraded forming
detectable TCA) in the overall degradation process. 1,1,1-
Trichloroethane was not detected in other sample matrices
during the degradation experiments of the brominated
compounds, indicating that its presence in this solution was
a result of the reverse reaction of carbocation with the
chloride in the solution.
Increasing the extent of bromination increased the
percent of the degradation resulting in the elimination
product (Table 6). The proportion of the total degradation
which resulted in elimination for BDCA at 65C was within
the error estimate for the percent elimination of TCA at
elevated temperatures, and both of these parent compounds
produced 1,1-DCE. The highest percent elimination was
observed for TBA which formed approximately 60% 1,1-
dibromoethene (Figure 12). This may be due to an increase
in steric hindrance in carbocations containing bromine
rather than chlorine, slowing the substitution pathway.


52
Table 8. Second Order Degradation Rates (1 mole'^ hr"^)
of Halogenated Ethenes at 60C
in Sodium Hydroxide Solutions
NaOH
ncentration
TCE
1,1-DCE
PCE
0.1 M
0.6
0.02
nd
0.5 M
0.28
0.01
nd
1.0 H
0.17
0.01
nd
2.0 M
0.12
0.004
nd
nd no significant degradation occurred after 260 hours.
The rate of degradation of TCE was the greatest among
the tested compounds due to the presence of an acidic
hydrogen (a hydrogen present on a carbon containing a
halogen). The elimination reaction was also an available
pathway for the degradation of 1,1-DCE, although the rate of
degradation was approximately 30 times slower than for TCE
in all solutions except for the 1.0 M NaOH.
Tetrach1oroethene (PCE) did not degrade since the
dehydrohalogenation reaction could not occur, and apparently
conditions were not favorable for an addition process.
For environmental applications, there are concerns with
the mildest conditions (temperature and pH) which may still
result in degradation of these compounds. The pathway for
the degradation of ethenes at elevated temperature could
differ from reactions at lower temperatures where the
elimination reaction would be less favorable and a possible
addition reaction could occur instead.
Therefore, TCE was


GASOLINE IN GROUND WATER
The overall goal of this part of the dissertation
research was to describe the patterns and the variability in
the partitioning of gasoline components into water.
Specific objectives were to
1. Evaluate concentration ranges of major components
in water extracts of various gasolines.
2. Measure fuel/water partition coefficients.
3. Assess the behavior of oxygenated additives and
their effect on partitioning of hydrocarbons.
4. Describe the changes in gasoline which would occur
through weathering (changes in composition resulting from
environmental exposure).
5. Determine if specific fuel sources could be
identified by the concentrations of components measured in
water extracts of gasoline.
Background
Composition of Gasoline
Gasoline is a complex mixture of volatile hydrocarbons.
The major components are branched-chain paraffins,
cycloparaffins and aromatics. The specific composition will
vary depending on the source of the petroleum as well as the
production method (e.g. distillation or fractionation,
95


186
P18
P19
48
1.93
1.25
49
2.04
1.18
50
1 59
0.94
51
3.23
2 30
52
1.90
1.33
53
3.43
1.81
54
2.95
3.15
55
3.79
1.82
56
1.41
1.18
57
1.84
2.04
58
2.01
1.21
59
2 60
1.56
60
2.38
1.30
61
1.88
0.79
62
1.85
1.46
63
2.29
1.68
64
2.07
1.83
65
1.71
1.01
P21
P22
P23
1.75
0.37
1.63
1.84
0.40
2.16
1.53
0.00
1.58
2.30
7.23
3.49
1.02
1.95
1.92
1.93
4.39
1.48
2 58
8 13
2.35
2.32
8.30
2.40
1.73
0.00
0.62
3 80
0.57
4.09
1.76
0.66
1.92
2.17
0.00
2.11
2 52
0.66
3.03
1.61
0.00
2.04
2.20
0.00
1.67
2.05
0.00
1.74
3.00
0.33
3 24
1.56
1.87
1 .87
P20
O .36
O 36
0.00
5.27
1.13
3.47
7.09
7.13
0.00
0.00
0.00
0.00
0.00
0.00
0.00
0.00
0.78
0.00


46
In a study of hydrolytic decomposition by Pearson and
McConnell (1975), volatilization was extrapolated to zero
and a degradation half-life for TCE of 30 months was
e s timate d.
Roberts (1985) examined field evidence for the
degradation of various chlorinated organics and estimated
rate constants for both TCE and PCE of approximately 0.003
day-1, which may be due to a variety of factors including
sorption and dilution.
Wilson et al. (1985) studied the aerobic degradation of
TCE, PCE and other compounds in actual aquifer materials
from two sites in Oklahoma and Louisiana. No detectable
biodegradation of these compounds was observed under the
experimental conditions. Since degradation was noted in
autoclaved samples, the authors postulated that TCE and PCE
degradation was likely due to abiotic processes with rates
similar to those reported by Dilling et al. (1975).
The dehydrochlorination reaction of TCE occurs under
basic conditions and generates dichloroacetylene and
hydrogen chloride. This reaction of TCE with base is
spontaneous at room temperature and was responsible for
dichloroacetylene intoxication observed in patients inhaling
TCE-containing air in closed systems equipped with alkali
absorbers (Environmental Protection Agency, 1979).
Dichloroacetylene was detected in the gas phase above
aqueous alkaline solutions with pH 11 to 13 and upon


20
consistant results and that longer periods had little effect
on the final concentrations.
Saturated, equilibrated solutions of neat gasolines in
contact with distilled, deionized, organic-free water were
prepared. Two mL of gasoline were added to 40 mL water in
VOA vials having Teflon septa. Samples were mixed on a
rotating disk apparatus for 30 minutes at room temperature
(generally 21-23C). The vials then sat undisturbed for one
hour, in an inverted position. Each separated water phase
was removed through the septum at the bottom of the VOA
bottle using a 5 mL syringe. A separate needle was inserted
to allow air to enter the vial so that a vacuum did not form
preventing withdrawal of the water.
Triplicate samples of each water phase were then sealed
in 2 mL crimp-seal vials and refrigerated until the GC
analysis was performed, typically within 2 days. Replicate
extractions, and replicate analyses of extracts were
performed for quality control.
Some overlap or incomplete peak resolution occurred in
the early eluting compounds for both the neat gasoline
samples and the water extracts. Enhancement of the more
water soluble components occurred following aqueous
extraction, making it easier to identify compounds like
benzene and MTBE in the water extract. Toluene was easily
identified in both the neat and water fractions.


18
compounds. Therefore, the molar response factor for TCA was
used to quantify the ethanes containing bromine and the
molar response factor for 1,1-DCE used to quantify the
brominated ethenes.
Table 3. Relative response of trihalomethanes on GC/FID
Trihalomethane
ng
nmoles
Area
Counts
rf*
1
Chloroform
616
5.15
24.18
0.21
2
Bromoform
924
3.65
15 19
0.24
3
Chloroform
924
7 73
43.98
0.18
4
Bromoform
1386
5.48
22.02
0.25
5
Bromodichlorome thane
502
3.06
18.47
0.17
6
Dibromochloromethane
386
1 .85
9 .89
0.19
7
Bromodichloromethane
1255
7 .65
38 79
0.20
8
Dibromochloromethane
965
4.63
21.56
0.21
Average 0.21
S td. Dev. 0.03
Re 1. Dev. 13.5%
Response Factor, nmoles/area counts.
Gasoline
Analyses for gasoline consituents were also performed
by GC/FID, using a Perkin-Elmer Model 8410 gas chromatograph
with a 30 m wide bore capillary column (J&W, DB-1) having a
3 pm film thickness. The neat gasoline samples were
analyzed by direct injection of 0.05 pL of the fuel.
Gasoline components dissolved in water were determined by
sparging volatiles from water using a Tekmar LSC-2 Purge and
Trap instrument interfaced to the Perkin-Elmer GC. The
temperature program for both neat gasolines and water


114
as described earlier. This procedure was repeated for
ethanol and t-BA.
The Kfw for MTBE was found to be 15.7 +/- 4.3 (Table
16). The standard deviation for the measured Kfw's is 0.05
log units, while the Kfw was about a factor of 20 less than
benzene. The partition coefficients for the alcohols were
all found to be less than one. Accurate calculations of the
partition coefficients for the highly water soluble
components, ethanol and t-BA, were difficult using the
analytical protocols and fuels described in this
dissertation. Most of these alcohols were detected in the
aqueous phase, and coelution with other fuel components made
quantitation of the small amount of alcohol in the fuel
layer difficult to accurately perform.
These Kfw results can be compared with those reported
by Groves (1988) by calculating a solvent water partitioning
coefficient based on concentrations, similar to Kfw, instead
of ratios of mole fractions as reported. The solvent/water
coefficients as recalculated from Groves (1988) were
Compound Solvent system K
MTBE
MTBE
E thano1
Me thanol
Benzene/water
Hexane/water
Benzene/water
Hexane/water
23.3
14.9-15.5
0.05
0.002-0.004
These solvent/water partition coefficents correlate with the
Kfw experimental results. A partially miscible cosolvent
like MTBE concentrated in the solvent or fuel phase, while
the completely miscible alcohols predominantly partitioned


142
The early eluting peaks, P3 and P4, showed some
correlation (0.739) while various combinations of peaks P12
to P20 showed some linear relationship. A high correlation
between peaks indicated some redundancy in the information
provided by those peaks in distinguishing among brands or
grades of gasolines.
Bivariate Plots
Bivariate plots of selected components provided a
better visualization of the relationship between variables
than did a single Pearson correlation coefficient. Figures
37 through 41 illustrate patterns in the data, with the
various brands identified by letter. Variables with higher
correlation coefficients, e.g., mp-xylene and ethylbenzene
(Figure 37), showed a linear relationship, while a plot of
P3 (C5 hydrocarbon) and m,p-xylene (Figure 38) demonstrated
considerable scatter. A bivariate plot of variables which
were not highly correlated may show random scatter or
perhaps a nonlinear relationship. The data in plot of P5
versus m,p-xylene (correlation coefficient of -0.025) showed
random scatter. Some clustering of the brands of gasoline
can be observed in this plot, particularly with some of the
Shell gasoline samples.
A high linear correlation between two variables implies
that the ratio of the two concentrations is fairly constant
for all the grades and brands evaluated in this study. The
ratio of m,p-xylene to ethylbenzene will not be a


LOG K
6
5 -
4 -
3 -
2 -
1 -
0
-4 -3 -2 -1 0
LOG Solubility
Figure 32. Correlation between fuel/water partition coefficient and pure component solubility.
1,2,3-Trimethylbenzene
MTBE


138
Basic Descriptive Statistics
Basic statistics on the peak area counts (Table 22)
summarize the range of values (as area counts) obtained
during the analyses using identical protocols. The results
of the application of these procedures can be used to
establish the consistency or variability in the data for
particular components and may be used in quality control by
highlighting anomolous data points. The concentrations in
the water extracts were approximately 0.3 times the area
count based on the sample volume (0.1 ml) and average
response factor determined by analysis of standards.
The large standard deviation for P5 suggests highly
variable area counts. The peak at this retention time most
commonly consists of an alkene with area counts typically
less than 30. Eight samples with MTBE had area counts at
this retention time ranging from 580-1320.
For peaks P11-P23, the maximum values were much higher
than the average values. The data were examined to
determine if only one sample was responsible for the high
values for each of these components. These maximum values
represented several different samples, although many of the
samples with a high value for one of these components showed
elevated concentrations of a number of these constituents.
Correlations between constituent concentrations were
calculated using the SAS procedure CORR, so linear
relationships between various constituents would be


Aqueous Solubility (mg/l)
Equilibrium Mole Fraction of TCA
(Solvent Phase)
Figure 19. Aqueous solubilities of a binary mixture of TCA and 1,1-DCE as a function of
mole fraction composition in the solvent phase (24 C).


Retention Time
16
u
16 .
20
J.B0N
FID Response
17.33
23.2*
25. ?3
Figure 2. Sample chromatogram of partially degraded geminal
trihalides.
Compound Retention Time
1.1.1-Tribromoethane
1.1-Dibromo-l-chioroethane
l-Bromo-1,1-dichloroethane
25.93
23.24
19.47
1.1-Dibromoethene
1-Bromo-l-chloroethene
1.1-Dichloroethene
17.33
12.30
7.80


100
90 -
80 -
70 -
TBA DBCA BDCA
1 r
TCA
Figure 12. Comparison of the percent of the elimination pathway for 1,1,1-trihaloethanes.
p-