determination of the degradation mechanism for

University of Central Florida University of Central Florida

STARS STARS

Electronic Theses and Dissertations, 2004-2019

2008

Determination Of The Degradation Mechanism For Determination Of The Degradation Mechanism For

Polychlorinated Biphenyl Congeners Using Mechanically Alloyed Polychlorinated Biphenyl Congeners Using Mechanically Alloyed

Magnesium/palladium Magnesium/palladium

Robert DeVor University of Central Florida

Part of the Chemistry Commons

Find similar works at: https://stars.library.ucf.edu/etd

University of Central Florida Libraries http://library.ucf.edu

This Doctoral Dissertation (Open Access) is brought to you for free and open access by STARS. It has been accepted

for inclusion in Electronic Theses and Dissertations, 2004-2019 by an authorized administrator of STARS. For more

information, please contact [email protected].

STARS Citation STARS Citation DeVor, Robert, "Determination Of The Degradation Mechanism For Polychlorinated Biphenyl Congeners Using Mechanically Alloyed Magnesium/palladium" (2008). Electronic Theses and Dissertations, 2004-2019. 3471. https://stars.library.ucf.edu/etd/3471

https://stars.library.ucf.edu/



https://stars.library.ucf.edu/etd

http://network.bepress.com/hgg/discipline/131?utm_source=stars.library.ucf.edu%2Fetd%2F3471&utm_medium=PDF&utm_campaign=PDFCoverPages

https://stars.library.ucf.edu/etd

http://library.ucf.edu/

mailto:[email protected]

https://stars.library.ucf.edu/etd/3471?utm_source=stars.library.ucf.edu%2Fetd%2F3471&utm_medium=PDF&utm_campaign=PDFCoverPages



DETERMINATION OF THE DEGRADATION MECHANISM FOR POLYCHLORINATED BIPHENYL CONGENERS USING MECHANICALLY ALLOYED

MAGNESIUM/PALLADIUM IN METHANOL

by

ROBERT WILLIAM DEVOR B.S. University of Central Florida, 2003

M.S. University of North Carolina at Chapel Hill, 2005

A dissertation submitted in partial fulfillment of the requirements for the degree of the Doctor of Philosophy

in the Department of Chemistry in the College of Sciences

at the University of Central Florida Orlando, Florida

Spring Term 2008

Major Professor: Cherie L. Geiger

ii

© 2008 Robert William DeVor

iii

ABSTRACT

Polychlorinated biphenyls are a ubiquitous environmental contaminant that can be found

today throughout the world in soils and sediments, lakes and rivers, and flora and fauna. PCBs

have percolated throughout the food chain, so that almost every human being has a detectable

amount of the contaminant within their blood stream. Existing remediation methods include

incineration, dredging and landfilling, and microbial degradation, but all of these methods have

drawbacks that limit their effectiveness as treatment options. Recently, the use of zero-valent

metals as a means of reductive dechlorination has been explored. Using a combination of zero-

valent magnesium and catalytic palladium, a successful bimetallic system capable of degrading

PCBs has been created and optimized.

Determining the mechanism for the reductive dechlorination has proven to be an arduous

task, but experimental evidence has suggested three possible radical-type mechanisms for the use

Mg/Pd specifically in methanol (as compared to aqueous systems). These possible mechanisms

differ in the type of hydrogen species that replaces the chlorine atom on the PCB.

Thermodynamic information has also aided in narrowing down which of the suggested pathways

is most likely. It appears likely that the hydrogen involved in the dechlorination has the form of

a “hydride-like” radical, which is a form of electron-rich atomic hydrogen. According to the

literature, Pd catalysts create this species within the first few subsurface layers of the palladium

in the presence of molecular hydrogen. Further work will be necessary to confirm that the

“hydride-like” radical is actually the species involved in the dechlorination.

iv

To Mom who was the mountain of support that allowed me to be able to accomplish this goal

To Dad

who showed me through his own actions and work ethic how anything was possible

To My Family who were always there for me and always believed in me, even when I couldn’t

v

ACKNOWLEDGEMENTS

I would like to express my eternal gratitude to both my dissertation advisor, Dr. Cherie

Geiger, and her fellow committee member, Dr. Christian Clausen. Without their patience,

guidance and support this research would not have been possible, nor would I have been capable

of completing it. In addition, Dr. Kathy Carvalho-Knighton, also a committee member, has also

been an invaluable resource and a source of guidance throughout this entire process. I would

also like to thank Dr. Seth Elsheimer, who was instrumental in helping to define the mechanism

proposed in this dissertation. I’m also very grateful to Dr. Elsheimer and Dr. Michael Hampton,

who along with Dr. Geiger, Dr. Clausen, and Dr. Knighton were part of my dissertation

committee. Finally, I have to thank my research laboratory colleagues who have helped me

achieve all that I have thus far in my scientific career: Rachel Calabro, Rebecca Fidler, Michael

Gittings, Erin Holland, Phillip Maloney, Debbie Maxwell, Amy Stout, Lucasz Talalaj and

especially Brian Aitken.

vi

TABLE OF CONTENTS

LIST OF FIGURES ..................................................................................................................... viii

LIST OF EQUATIONS ................................................................................................................. xi

LIST OF ABBREVIATIONS....................................................................................................... xii

CHAPTER ONE: INTRODUCTION............................................................................................ 1

Polychlorinated Biphenyls .......................................................................................................... 2

Current Remediation Technologies ............................................................................................ 7

Zero-Valent Metals ................................................................................................................... 10

Catalytic Palladium................................................................................................................... 12

Bimetallic Reactive Metals ....................................................................................................... 13

Communition Theory................................................................................................................ 16

Possible Degradation Mechanisms ........................................................................................... 21

Research Objectives.................................................................................................................. 24

CHAPTER TWO: MECHANICAL ALLOYING IN USE FOR THE PREPARATION OF A

PALLADIZED MAGNESIUM BIMETALLIC COMPOUND FOR THE REMEDIATION OF

PCBS............................................................................................................................................. 26

Introduction............................................................................................................................... 26

Materials and Methods.............................................................................................................. 30

Results and Discussion ............................................................................................................. 36

Conclusion ................................................................................................................................ 48

vii

CHAPTER THREE: DECHLORINATION COMPARISON OF MONO-SUBSTITUTED

PCBS WITH Mg/Pd IN DIFFERENT SOLVENT SYSTEMS ................................................... 50

Introduction............................................................................................................................... 50

Experimental ............................................................................................................................. 53


Conclusion ................................................................................................................................ 65

Acknowledgements................................................................................................................... 66

CHAPTER FOUR: MECHANISM OF THE DEGRADATION OF INDIVIDUAL PCB

CONGENERS USING MECHANICALLY ALLOYED Mg/Pd IN METHANOL.................... 67

Introduction............................................................................................................................... 67

Methods..................................................................................................................................... 70


Acknowledgements................................................................................................................... 84

CHAPTER FIVE: CONCLUSION.............................................................................................. 85

APPENDIX A SUPPORTING INFORMATION FOR CHAPTER TWO .................................. 89

APPENDIX B SUPPORTING INFORMATION FOR CHAPTER THREE............................. 105

APPENDIX C SUPPORTING INFORMATION FOR CHAPTER FOUR............................... 120

REFERENCES ........................................................................................................................... 129

viii

LIST OF FIGURES

Figure 1: Picture of the milling canisters, templates and mill ..................................................... 33

Figure 2: GC-FID chromatogram for 40-ppm Arochlor 1260 control ........................................ 38

Figure 3: GC-FID chromatogram for 40-ppm Arochlor 1260 sample after 72 hr exposure to 1.00

g Mg/Pd................................................................................................................................. 39

Figure 4: GC-FID chromatogram for 40-ppm Arochlor sample after 96 hr exposure to 1.00 g

Mg/Pd.................................................................................................................................... 39

Figure 5: Chromatogram of 10.0-ppm 1254 Arochlor in 9:1 water:methanol control ................ 40

Figure 6: Chromatogram of 10.0-ppm Arochlor 1254 in 9:1 water:methanol after 24 days

exposure to 0.100 g of Mg/Pd............................................................................................... 41

Figure 7: Plot to determine activity of Mg/Pd using parent:product signal ratio ........................ 41

Figure 8: Plot of the activity of Mg/Pd for the optimization of the %Pd loading........................ 42

Figure 9: GC-ECD chromatograms of 10 mL of 10.0-ppm PCB-151 solution reacted with (a)

Mg/Pd, (b) Mg, and (c) Pd/graphite...................................................................................... 43

Figure 10: Plot of the activity of Mg/Pd for the optimization of the milling time ...................... 44

Figure 11: Plot of the activity of Mg/Pd for the optimization of the # of milling balls............... 44

Figure 12: Plot of the activity of the Mg/Pd for the optimization of canister loading................. 45

Figure 13: SEM micrograph of an Mg/Pd particle (2700x magnification). White line (scale)

represents 1 um ..................................................................................................................... 46

Figure 14: Pseudo 1st order rate constants for three bimetals with equivalent Pd content .......... 47

Figure 15: Structures of PCB-001, PCB-002, and PCB-003 ....................................................... 53

ix

Figure 16: Pseudo 1st Order kinetic plot of the degradation of PCB-001 with Mg/Pd in methanol

............................................................................................................................................... 56

Figure 17: Pseudo 1st order kinetic plot of the degradation of PCB-002 with Mg/Pd in methanol

............................................................................................................................................... 56


............................................................................................................................................... 57

Figure 19: Pseudo 1st order kinetic plot of the degradation of PCB-001 with Mg/Pd in 9:1

water:methanol...................................................................................................................... 58


water:methanol...................................................................................................................... 58


water:methanol...................................................................................................................... 59

Figure 22: Pseudo 1st order kinetics plot of the degradation of PCB-001 with Mg/Pd from "Lag"

study...................................................................................................................................... 61

Figure 23: Degradation of biphenyl in methanol w/ 10% Mg/Pd................................................ 62

Figure 24: Degradation of biphenyl in 9:1 water:methanol w/ 10% Mg/Pd................................ 63

Figure 25: Gas chromatogram and mass spectrum of PCB-001 degraded with Mg/Pd in MeOH

............................................................................................................................................... 64

Figure 26: Gas chromatogram and mass spectrum of PCB-001 degraded with Mg/Pd in MeOD

............................................................................................................................................... 65

Figure 27: Degradation kinetics of PCB-151 in Mg/Pd in methanol........................................... 73

Figure 28: Comparison of PCB-151 and PCB-93/95 .................................................................. 74

x

Figure 29: GC-MS analysis of the degradation of PCB-151 using Mg/Pd in methanol.............. 75

Figure 30: Pseudo 1st order kinetics plot of the degradation of PCB-151 with Mg/Pd in methanol

............................................................................................................................................... 75

Figure 31: Degradation Plot of PCB-151 using reactivated Mg/Pd in methanol ........................ 76

Figure 32: Pseudo 1st order kinetics plot of the degradation of PCB-151 using reactivated Mg/Pd

in methanol............................................................................................................................ 77

Figure 33: Pseudo 1st order kinetics plot of the degradation of PCB-151 using reactivated Mg/Pd

in methanol............................................................................................................................ 78

Figure 34: Balanced mechanism for the declorination of PCBs by Mg/Pd in methanol............. 81

Figure 35: Proposed mechanism for the dechlorination of PCBs by Mg/Pd in methanol by (a)

atomic hydrogen, (b) "hydride-like" radicals, and (c) hydride (H* denotes both hydrogen

and "hydride-like" species) ................................................................................................... 83

xi

LIST OF EQUATIONS

Equation 1: Chemical equations relevant to the reductive dehalogenation of alkyl halide via Fe

............................................................................................................................................... 11

Equation 2: Oxidation/reduction of iron and palladium .............................................................. 14

Equation 3: Standard reduction potentials for Mg, Zn, and Fe..................................................... 15

Equation 4: Equilibrium size of milled particle (Devaswithin et al., 1988) ................................ 18

Equation 5: aij term (Devaswithin et al., 1988) ........................................................................... 18

Equation 6: Bij term (Devaswithin et al., 1988)........................................................................... 18

Equation 7: S term (Devaswithin et al., 1988)............................................................................. 18

Equation 8: Determination of change in surface area of milled material (Tamura and Tanaka,

1970) ..................................................................................................................................... 19

Equation 9: Modified corrosion based scheme for Mg/Pd bimetal particles............................... 23

Equation 10: Standard reduction potential for Mg and Fe........................................................... 28

Equation 11: Production of atomic hydrogen on Pd and replacement of a Cl atom.................... 28

Equation 12: Overall reaction for dechlorination of PCBs to biphenyl....................................... 28

Equation 13: Generation of H2 from Fe and H2O........................................................................ 52

Equation 14: Standard reduction potential for Mg, Fe, and Zn ................................................... 52

Equation 15: Standard reduction potentials for Mg, Fe, and Zn.................................................. 69

xii

LIST OF ABBREVIATIONS

BZ Ballschmiter and Zell

DDT Diphenyl trichloroethane

df Thickness of stationary phase in capillary column

DNAPL Dense non-aqueous phase liquid

Eº Combined standard reduction potential of a right and left half-cell

ECD Electron capture detector

EPA Environmental protection agency

EZVM Emulsifed zero-valent metal

FID Flame ionization detector

GC Gas chromatography

i.d. Inner diameter

MCLG Maximum contaminant level goal

MeOH Methanol

MeOD Deuterated methanol (methanol-d)

MNA Monitored natural attenuation

MNR Monitored natural recovery

MS Mass spectrometer

NMR Natural monitored recovery

PBDE Polybrominated diphenyl ether

PCB Polychlorinated biphenyl

xiii

PCDD Polychlorinated dibenzodioxins

PCE Polychlorinated ethene

PCDF Polychlorinated dibenzofurans

ppb Part-per-billion

ppm Part-per-million

PTFE Polytetrafluoroethylene

RDX Cyclotrimethylenetrinitramine

RPM Revolutions-per-minute

SEM Scanning electron microscope

SHE Standard hydrogen electrode

SRN1 Radical nucleophilic substitution

STAR Science to achieve results

TCE Trichloroethylene

TOF Time-of-flight

TSCA Toxic Substance Control Act

UHP Ultra high purity

XPS X-Ray photoelectron spectroscopy

ZVI Zero-valent iron

ZVM Zero-valent metal

1

CHAPTER ONE: INTRODUCTION

The past century has introduced many technological and scientific marvels which at the

time appeared to be only beneficial to humanity. Time, along with further advances in

instrumentation and analysis techniques, has gone on to show that this has not always been the

case. Many of these advances have turned out to have unanticipated and unpredictable

consequences, some of which the world is only now beginning to attempt to rectify. Many

chemical compounds created in mankind’s past fall into this category.

Halogenated compounds are a class of environmental contaminants which have received

much attention over the years, due to their prevalent use in various industries throughout the past

century. Aliphatic chlorinated compounds have been used for a variety of purposes, such as

trichloroethylene (TCE) which was an industrial solvent with a wide variety of uses including:

an extraction solvent, a dry cleaning solvent, and as a metal degreaser. Chemicals of these types

have been implicated in a variety of health effects such as liver and kidney damage, spontaneous

abortions, and many are considered carcinogenic (Moran et al., 2007).

Of special interest are halogenated aromatics, which have proven to be far more difficult

to remediate than other halogenated aliphatics due to the stability of the aromatic system

involved. Ironically, this stability is one of the reasons this type of chemical was so often used

for industrial purposes. Of these, chlorinated aromatics such as dichloro diphenyl

trichloroethane (DDT) and polychlorinated biphenyls (PCBs) have received the most attention,

due to the fact that compounds of this type have regulatory limits and legislation which require

contaminated zones to undergo cleanup and treatment (Engelmann et al., 2001; Ross, 2004).

Other halogenated aromatics, such as polybrominated diphenyl ethers (PBDEs) have no such

2

legislation, so although the effects of the contaminants are similar in scope, much less research

into the removal and destruction of these types of chemicals has been done (Solomon and

Huddle, 2002). The focus of this dissertation is the attempt to create a technology for the

degradation of PCBs, and to then explain the mechanism by which the dechlorination reaction

occurs.

Polychlorinated Biphenyls

Polychlorinated biphenyls consist of a biphenyl backbone which can contain as few as

one or as many as ten chlorine atoms. They are a class of synthetic chemicals; there is no known

natural source of PCBs (Erickson, 1992). There are 209 distinct PCB congeners, each depending

upon the number of chlorine atoms and the substitution pattern exhibited (Voorspoels et al.,

2007). These congeners are named according to the Ballschmiter and Zell convention from BZ-

1 to BZ-209 (Ballschmiter and Zell, 1980). The congeners are then often broken down even

further by the degree of chlorination into homolog classes (i.e. dichloro and trichloro

homologues) to help ease analysis efforts (Erickson, 1992; Wiegel and Wu, 2000).

Originally manufactured in 1929, PCBs were highly used for industrial purposes in the

United States throughout the mid 20th century until the production and use were banned in 1976

by the Toxic Substances Control Act (TSCA). Useful physical properties of these compounds

include: high thermal conductivity, high dielectric constants, inert chemical structure, low

flammability, extreme hydrophobicity, and a high flash point. Because of these properties, PCBs

found themselves being used in a wide variety of applications throughout their existence (prior to

being banned). These applications included (but are not limited to): capacitor and transformer

oils, lubricating oils, sealants, adhesives, fungicides, plasticizers in paints, flame-retardant

3

materials, cooling/dielectric fluids, and even in carbonless copy paper. Many of the PCBs

created before the production ban are still in use today and are becoming a problem due to

weathering and aging that is occurring, which allow the PCBs to be released into the

environment as a continuing point source of contamination (Alonso et al., 2002; Erickson, 1992;

Jones et al., 2003; Ross, 2004).

Mass production of PCBs began in 1929, and these were most commonly produced as

complex mixtures which were then marketed under various trade names, such as Arochlor from

Montsanto (the largest producer of PCBs in the world). The manufacturing process for these

complex mixtures was a direct chlorination of biphenyl using chlorine gas. By changing the

various reaction conditions, it was possible to control the average degree of chlorination within

the mixtures. Various Arochlor mixtures were produced via this method, for example Arochlor

1248, Arochlor 1254, and Arochlor 1260 (Erickson, 1992). The numbering system for the

mixtures has two components. 12 corresponds to the number of carbons in the biphenyl system,

and the second pair of numbers (48 vs 54 vs 60) refers to the % chlorine by mass used in the

manufacturing process. These complex mixtures contained anywhere from 60 to 90 individual

congeners, but each mixture had its own distinctive “fingerprint”, or relative ratio of the

individual congeners making them easily identifiable from one another (Wiegel and Wu, 2000).

Individual congeners were not widely produced, due to the difficulty in isolating each congener

from the complex mixtures; however using individual congeners has become more popular for

use in research today because of the relative ease of single congener analysis.

Concerns over the use of PCBs were recognized as early as 1936, when information

regarding occupational exposures causing health complications first were reported. Due to this,

4

workplace/occupational exposure limits were put into place by the early 1940’s; however

regulation of PCB use did not occur for several more decades. The first environmental samples

of PCBs were discovered in 1966 in Sweden. PCBs were found in tissue samples from eagles

and herrings, giving rise to the first indication that PCBs were traveling up the food chain

(Erickson, 1992; Ross, 2004). Environmental samples in the United States were discovered in

1968 (Ross, 2004). PCBs were known to be lipophilic and thus prone to bioaccumulation,

causing concern as to the possible effect they could have upon human beings. Today, PCBs are

considered a universal environment pollutant, and can be found in most flora, fauna, and human

beings at detectable levels (Erickson, 1992). They have also exhibited a tendency to biomagnify

within the food chain, meaning that accumulation is higher the further up the food chain that is

examined (Wiegel and Wu, 2000).

Humans can be exposed to PCBs in a variety of ways. However, the primary route of

contact in dietary in nature. Over 90% of all PCB exposure falls into this category. Primarily,

dietary risks are primarily due to contaminated fish products (Voorspoels et al., 2007). Fish are

exposed to PCBs through sediment and water contamination, and bioaccumulation allows for

higher levels of contamination (even though the aqueous solubility of PCBs is very low) (Wiegel

and Wu, 2000). The PCBs are then passed up through the food chain to human beings. Other

routes of exposure include inhalation and skin adsorption, although these are considered minor

risks when compared to dietary routes.

Currently, there is an ongoing debate as to the actual danger PCBs present in terms of

toxicological effects and health risks. The majority of toxicological data for the possible heath

risks is extrapolated from animal studies, because of the ethical concern of testing these

5

compounds on humans. The only large enough pools of data for analysis on humans comes from

occupational exposures (both chronic and acute), which is not easily comparable to

environmental exposure (much lower level). The data from occupational exposures is also

limited due to the fact PCBs are no longer manufactured or used in industry as they were prior to

the TSCA regulation. However, based on data from exposure to rodents, the Environmental

Protection Agency classified PCBs as probable human carcinogens, and the International Agency

for Research on Cancer, the National Toxicology Program, and the American Conference of

Governmental Industrial Hygienists have all concluded that studies have shown PCBs are known

animal carcinogens. Several studies have reported a variety of exposure-related effects of PCBs

in humans, including endocrine disruption, non-specific reproductive effects, dermal

abnormalities including chloroacne, and abnormal neurobehavioral effects in children (Ross,

2004). This last effect concerning children is of great importance due to the ability of PCBs to

be passed from mother to child through consumption of breast milk. The lipophilic nature of

PCBs makes transmission of this nature a significant issue, and several studies have shown

detectable levels of PCBs in human breast milk (Erickson, 1992; Ross, 2004).

On the other hand, several studies have cited no discernable health effects even from

chronic occupational exposure within a PCB manufacturing plant, or stated that the observed ill-

health effects can not be directly linked to PCBs specifically. This is due to the fact that many of

the accidental exposures which triggered rising concerns in the world did not contain PCBs

alone. One specific example is the contaminated rice oil incident in Yusho, Japan in 1968. This

was a case of rice oil being contaminated with thermally degraded PCBs that came from faulty

and leaking equipment during the food manufacturing process. Because the PCBs had been

6

partially degraded due to thermal exposure, the mixture also contained more toxic compounds

such as polychlorinated dibenzo furans (PCDFs). These compounds are now considered to have

had a major role in the toxic effects observed in people affected by this exposure. This clouds

the issue of which contaminant is truly to blame (Erickson, 1992; Ross, 2004). A similar

incident occurred a year later in Taiwan, with many of the same issues obscuring the true cause

of observed health effects. In addition, several reports argue that with production ban enacted by

Congress in the 1970’s, environmental levels of PCBs have seriously declined to a point at which

the low level exposure currently experienced today in the environment poses no real danger to

humans (Ross, 2004).

Whether or not PCBs truly are a danger to people in the world today is seemingly

irrelevant, due to the fact current government regulations require that any release of PCBs into

the environment must be reported and regulated, depending upon the type of release and where it

occurs (referring to the availability of general population having access to the site and being

exposed). The Toxic Substances Control Act enacted by Congress in 1976 gave the

Environmental Protection Agency the responsibility and power to ban all production and use of

PCBs in any system that was not totally closed. Certain applications that are completely closed,

such as fenced in transformers, are still allowable under current regulations until the end of their

useful lifetime (Ross, 2004). Applications such as these require regular reports to the EPA on

the safety, use, and disposal practices throughout the lifetime of the equipment. The current

action levels described in the Clean Water Act for drinking water are 0.5 parts per billion, while

the maximum contaminant level goal (MCLG) is zero (Jones et al., 2003; Ross, 2004).

7

Current Remediation Technologies

Since releases of PCBs into the environment are heavily regulated, much research has

been done in an attempt to create a technology that is capable of remediating contaminated areas.

Depending on the type of contamination encountered (sediments/soil, groundwater, atmospheric,

enclosed systems) different remediation options exist. Techniques currently in use or under

investigation include incineration, dredging, landfilling, soil washing/extraction, microbial

degradation, capping, monitored natural recovery (MNR), chemical reduction, chemical

oxidation, photolytic/radiolytic degradation, and others. However, many of these techniques are

hampered by specific limitations or drawbacks that make them impractical for use in the field or

on a large scale. Some of the more frequently used options will now be discussed.

One of the most commonly used remediation technologies is incineration of PCB wastes.

This technology can be applied to most any type of contamination or hazardous waste, as long as

it has been isolated (i.e. extracting wastes from soils or sediments via solvent washing). This is

an ex-situ technique which converts the contaminant by thermal dechlorination to harmless

products; however incomplete combustion is a large problem with this option. This can lead to

the production of even more toxic byproducts, such as polychlorinated dibenzo-p-dioxins and

polychlorinated dibenzo-furans (PCDFs) , which are commonly called dioxins (Chuang et al.,

1995;Erickson, 1992, Wu et al., 2005). Incomplete combustion occurs when the incineration

process occurs at low temperatures, which is why incineration of PCBs is performed at

temperatures of 1200°C or greater. This calls for a large increase in fuel consumption which

leads to increased costs for this remediation technique (Wu et al., 2005). Also, an additional

concern associated with incineration (and all other off-site ex-situ processes) is the transport of

8

the hazardous materials to the incineration site (Flilippis et al., 1995). Costs for this technique

have been on the rise due to the limited facilities capable of performing the high temperature

incineration (Jones et al., 2003).

A more recent development in the remediation of PCBs is degradation using microbial

agents. Two distinct types of microbial degradation exist, consisting of using either anaerobic or

aerobic microbes to reductively dechlorinate or to oxidize the PCBs. Aerobic degradation causes

the oxidation of PCBs through a series of intermediates, ultimately destroying the contaminant.

Unfortunately, research has shown that aerobic microorganisms will only degrade the lower

chlorinated congeners, which are not the congeners primarily used in the mixtures that are found

in the environment. In addition, aerobic micro-organisms are only found in the top few

millimeters of soils and sediments, limiting the effectiveness as an in-situ remediation option.

Anaerobic microbes, on the other hand, degrade PCBs through reductive dechlorination carried

out via the removal of chlorine atoms as halogen ions. Higher chlorinated congeners can be

degraded than when utilizing anaerobic microbes, however, this is normally limited to the

dechlorination of meta and para congeners. Ortho-substituted chlorines are much more resistant

to this type of attack, which can lead to incomplete degradation of the contaminants (Erickson,

1992; Wiegel and Wu, 2000).

Sediment contamination, while prevalent in the United States, can be difficult to deal

with because of the difficulty in accessing the sites of pollution. Several options exist, although

the most often used is dredging of the contaminated sediments, followed by a second ex-situ step

such as landfilling. Dredging has a major drawback in that disturbing the contaminated

sediments can lead to re-release of the PCBs into the surrounding water supply. This allows for

9

the mobilization of PCBs, and can allow for the further spread of the problem into surrounding

areas that were previously free of contaminants. Even if the contaminants do not spread to new

uncontaminated areas, release into the water system can cause increased contamination of local

biota (Rice and White, 1986; Schmidt, 2001). Another serious limitation to dredging is that it

does not eradicate the existing problem; it simply moves it from one location to another (in the

case of landfilling) or requires an additional remediation technique (i.e. incineration) for the

degradation of the PCBs.

Another common technique for remediation of sediments is called “capping” which is

often then followed by monitored natural recovery. Capping refers to placing an inert chemical

substance above contaminated sediments to prevent the pollutants from interacting or mobilizing

into the surrounding water system. Again, this does not act to degrade the PCBs themselves; it is

simply a means to immobilize the immediate threat presented by the contaminant. Another

problem associated with sediment capping is the possibility of breaches occurring in the “cap”,

allowing release of the PCBs into the surrounding environment. MNR (also called monitored

natural attenuation, or MNA) refers to letting natural processes to reduce the magnitude and

bioavailability of hazardous materials. A variety of biological (biotransformation,

biodegradation), chemical (sorption, oxidation, reduction) and physical (volatilization,

dispersion, dilution) processes are considered pertinent in MNR. The time frame for MNR is

much longer than other current remediation options, but it does have the advantage of being

more cost-effective (since an active in-situ technology is not utilized) than other techniques.

However, one risk is that nature can upset the process, in the form of sediment disturbances from

phenomena like flooding, hurricanes, earthquakes, etc… (Kremer et al., 2006)

10

Zero-Valent Metals

Zero-valent metals have proven to be a promising avenue for remediation of a variety of

halogenated contaminants. This is primarily done through a process of reductive dehalogenation,

however some studies have shown ZVMs can catalyze oxidation in a Fenton-type reaction, (the

oxidiation of contaminants using a solution of Fe2+ and hydrogen peroxide). Several different

ZVMs (zinc, magnesium) have been examined for remediation properties; however the majority

of research has focused on zero-valent iron (ZVI) due to its low cost, availability, and its non-

toxic nature. ZVI has shown promise not only for the remediation of halogenated organics

(primarily chlorinated systems), but also with nitrates, perchlorates, RDX, and metals such as

lead and chromium (Manning et al., 2007; McDowall, 2005; Schrick et al., 2004).

Several studies have shown the capability of ZVI to degrade halogenated organics, and

some of the earliest studied contaminants were chlorinated aliphatics including carbon

tetrachloride and TCE. Studies performed by Tratnyek and Matheson have shown the

degradation of these compounds to exhibit pseudo 1st order kinetics, and to procede via a

stepwise dechlorination. In the case of carbon tetrachloride, ZVI sequentially produces

methylene chloride through an intermediate of chloroform. Each additional dehalogenation step

proceeds slower than previous steps. The dechlorination of TCE proceeded at a much slower

rate than that of carbon tetrachloride, and the reaction products were not determined within this

study (Matheson and Tratnyek, 1994).

The reaction of chlorinated aliphatics with ZVI is the result of the redox couple of the

zero-valent iron particles (Fe0) and dissolved solubilized iron (Fe2+). The reduction potential of

11

this redox couple (-0.44 V vs. SHE) is capable of reducing a variety of alkyl halides through

reductive dehalogenation. The relevant chemical equations are given below:

Fe2+ + 2e- Fe0

RX + 2e- + H+ RH + X-

Fe0 + RX + H+ Fe2+ + RH +X-

Equation 1: Chemical equations relevant to the reductive dehalogenation of alkyl halide via Fe

The overall equation for this type of reaction can be seen as the corrosion of the ZVI due to the

ability of the alkyl halide to act as an oxidizing agent. This reaction is thermodynamically

favorable under most conditions because the standard reduction potentials for many alkyl halides

are between +0.5 V to 1.5 V at a neutral pH. Since the surface of the iron is playing a role in the

reaction (being oxidized), changes in its condition should affect the kinetics of the

dechlorination, which has been shown experimentally. ZVI samples prepared with dilute acid

washes showed greater dehalogenation kinetics than unwashed metal. This is likely due to the

removal of unreactive iron oxide from the surface of the ZVI, increasing the available surface

area for the reaction to occur (Matheson and Tratnyek, 1994).

Published literature has shown very little success using ZVI alone for the degradation of

PCBs. Only one published study has shown ZVI to be capable of this, and it required extreme

reaction conditions, including a temperature of over 200ºC (Chuang et al., 1995). Use of various

catalysts in combination with ZVM was the next logical step in order to make the dechlorination

of PCBs feasible.

12

Catalytic Palladium

Several different catalytic metals have been studied for possible use in combination with

ZVMs (including palladium, platinum, ruthenium, rhodium, nickel, etc…). Since ZVI was

proven capable of degrading PCBs, even though it was at a high temperature, the idea was to

provide a lower energy pathway which would make the use of this technology in the field

feasible. Palladium is the most well-known hydrodehalogenation catalyst used in organic

reactions, so a majority of research has centered on its use with ZVMs (Alonso et al., 2002).

Other common hydrodehalogenation catalysts, such as platinum, have shown the capability of

dechlorinating halogenated aromatics (i.e. PCBs, chlorophenols, etc…) but at a much slower rate

when compared to studies using palladium. It has been proposed that this is due to palladium’s

unique ability to absorb large amounts of hydrogen within its lattice sites. Hydrogen is capable

of residing within the interstitial sites of bulk palladium and of bonding to subsurface sites of the

palladium lattice. These two types of absorbed hydrogen, along with dissociated adsorbed

atomic hydrogen on the surface of palladium, represent three distinct types of hydrogen that may

be involved in the reductive dechlorination of PCBs (Cybulski and Moulijn; 2006).

One question that is of relevance to the study of Pd as a catalytic metal in conjunction

with ZVM systems is what is the oxidation state of the metal? A recent study performed by

Arcoya et al. attempted to answer this question in relation to Pd/C systems used for the reductive

dechlorination of carbon tetrachloride. X-ray photoelectron spectroscopy (XPS) was used to

elucidate the oxidation state of the catalyst pre- and post-reaction for a variety of Pd/C systems.

Active catalysts systems contained both Pd0 and Pdn+ species, while systems only one or the

other were inactive. The proportions of both species were shown to depend on the nature of the

13

palladium compounds used and the reduction temperature at which the reaction was carried out

(Gomez-Sainero et al., 2002). This suggests the palladium is an electron source in the reductive

dechlorination of chlorinated contaminants at least at some level.

Studies have demonstrated that ZVMs are not necessary for Pd/C to degrade halogenated

organics if molecular hydrogen is made part of the reaction system, so why are they even

necessary for an applicable remediation technology? Pressurizing a reaction vessel with

hydrogen is relatively simple to accomplish within the laboratory setting, however it becomes

much difficult with regard to realistic applications in the world of environmental remediation. It

may be feasible in use with ex-situ remediation techniques, but it much more problematic when

dealing with hazardous wastes associated with groundwater and sediments where an in-situ

option is preferable. This leads to the combination of catalytic metals with ZVMs as the basis

for an in-situ. The ZVM, with an appropriate proton donor source, becomes the source of the

hydrogen gas in the reaction scheme.

Bimetallic Reactive Metals

A bimetallic reactive metal particle consists of a zero-valent metal substrate (i.e. iron or

magnesium) and a hydrogenation/hydrodechlorination catalyst (i.e. palladium, planinum, or

nickel) for the purpose of increasing the activity towards the dehalogenation of various

contaminants. A variety of bimetallic particles have been studied for use in the dechlorination of

different halogenated contaminants, including Fe/Pd, Fe/Ni, Mg/Pd, Zn/Pd, etc… Fe/Pd, Zn/Pd

and Mg/Pd have been shown to be capable of degrading PCBs at room temperature and normal

reaction conditions (Doyle et al., 1998; Engelmann et al., 2003; Engelmann et al., 2001 ; Grittini

14

et al., 1995; Kim et al., 2004; Korte et al., 2002; Liu et al., 2001; Muftikian et al., 1996; Wang

and Zhang, 1997; Zhang et al., 1998).

The most common preparation technique for bimetallic reactive particles is the deposition

of the catalytic metal onto the ZVM substrate through a reductive precipitation reaction of the

catalyst onto the surface of the ZVM. Due to the reduction potentials involved, no external

current is necessary for the plating/precipitation process, giving rise to the term electrode-less or

electrodeposition. This is accomplished by simply immersing the ZVM being used into a

solution containing the catalyst of interest. Wang and Zhang published a procedure for the

preparation of palladized nanoscale iron particles in which the wet nanoiron is placed into an

ethanol solution containing [Pd(C2H3O2)2]3, which causes the reduction and deposition of Pd

onto the iron surface (Wang and Zhang, 1997). Hexachloropalladate can also be used, and is

available in commercial solutions for this purpose (i.e. Pallamerse). This technique can also be

applied to commercially available iron particles (both micro- and nano-scale). The following is

the general reduction reaction for electrodeposition onto the surface of ZVI:

Pd2+ + Fe0 Pd0 + Fe2

Equation 2: Oxidation/reduction of iron and palladium This procedure has been used with addition ZVMs including magnesium and zinc (Kim

et al., 2004), although magnesium specifically presents several advantages when compared to

iron. The standard reduction potential for magnesium is much higher than that of iron (or zinc)

as can be seen below:

15

Mg2+ + 2e- Mg0 E0 = -2.37 V vs. SHE

Zn2+ + 2e- Zn0 E0 = 0.76 V vs. SHE

Fe2+ + 2e- Fe0 E0 = -0.44 V vs. SHE

Equation 3: Standard reduction potentials for Mg, Zn, and Fe

The increased reduction potential for magnesium equates to an increased thermodynamic driving

force in the dechlorination reaction, and to an increase in the production of hydrogen gas which

is necessary for hydrodehalogenation to occur. When exposed to oxygen, magnesium forms a

thin permeable oxide layer which is capable of allowing the reaction to proceed, however

prevents the entire ZVM “core” from becoming completely oxidized and rendered useless

(Agarwal et al., 2007). Iron lacks this capability to limit the corrosion of the core of the metallic

particle. In fact, due to highly reactive nature of nano-scale iron, the preparation of palladized

iron particles by electroless deposition requires it be conducted under an inert atmosphere,

complicating the procedure. Unfortunately, preparation of Mg/Pd particles through

electrodepostion has been shown to require relatively large amounts of palladium (0.5% by

weight) to achieve reasonable kinetics for the degradation of PCBs (Aitken et al., 2006). Due to

the high cost of this noble metal, this is unacceptable when contemplating using these particles

for large scale field applications. It is possible this is due to uneven distribution of the catalyst

onto the surface of the ZVM during the preparation of the bimetal, whereby the majority of the

palladium is being deposited on only a very small amount of the magnesium. This is a

consequence of it not being able to sufficiently mix the palladium into the plating solution before

deposition onto the surface of the ZVM occurs. The reactivity of micro- and nano-scale metal

particles is such that this is certainly feasible, although at the present time this supposition

16

remains unproven. Regardless as to why it occurred, this still presents the problem that it is

prohibitively expensive to produce this bimetal particle on a large-scale.

A more cost effective and easily accomplished preparation procedure was required for the

large-scale production of the reactive Mg/Pd bimetallic particles for use in field applications. To

that end, mechanical alloying was explored as the possible solution to this problem, which leads

to the discussion of communition (milling) theory.

Communition Theory

Communition theory is principally concerned with reducing the average size of particles

in a sample of crystalline or metallic solid; however, it can also be used to understand

mechanical alloying of particles. To accomplish either of these tasks the most commonly used

processes involve ball milling, vibrator milling, attrition, and roller milling, however the focus of

the remainder of this discussion will be on vibrator milling since it was the method of choice for

preparation of the target bimetal (Lü and Lai, 1998; Suryanarayana, 2004).

Vibrator milling is a process in which a milling vessel, containing the material to be

milled and some grinding material (usually stainless steel balls), is vigorously shaken in a back

and forth motion, or in a back and forth motion in conjunction with a lateral motion that

produces a figure 8 type path. The process relies solely on the high-energy collisions between

rapidly moving milling balls. If some particulate material is pinched between the participants in

one of these collisions, and the collision is of adequate energy, the particles are fractured into

smaller particles. Since vibrator mills can shake canisters at rates of up to 1200 RPM, often

producing ball speeds of upwards of 5 m/s, vibrator milling commonly yields the desired

17

reduction in particle size at a rate of about one order of magnitude faster than that of any other

type of milling process (Lü and Lai, 1998; Suryanarayana, 2004; Tamura and Tanaka, 1970).

For all milling types, the reduction of particle size relies on stresses induced in individual

particles caused by collisions within the milling vessel. This process reduces the average particle

size until equilibrium is reached, at which point no further size reduction is observed. This

phenomenon can be explained if one considers the competing processes of fracturing and cold

welding (Austin, 1973; Devaswithin et al., 1988; Lü and Lai, 1998; Suryanarayana, 2004;

Tamura and Tanaka, 1970).

Initially, the particles involved in the milling are ductile (in ductile-ductile or ductile-

brittle systems) and undergo plastic deformation. This allows for cold welding to occur between

the (relatively) soft particles. The overall particle size increases, and a broad range of particles

sizes are produced, at least for a time. Eventually, the continued deformation causes the particles

to become work hardened and then to fracture into smaller particles. This fracturing occurs by a

fatigue failure mechanism as well as by fragmentation of the relatively fragile flakes produced

during the milling process. Eventually, theses competing processes (cold welding and

fracturing) reach a steady-state equilibrium in which the average particle size remains constant.

Eventually, very fine (small) particles reach a size where it is difficult to achieve enough energy

in the ball mill to cause fracturing, and most often cold welding will occur. Larger particles are

still capable of being fracturing into smaller particles, and the combination of these two

processes drive the equilibrium particle size distribution to an intermediate range

(Suryanarayana, 2004).

18

The equilibrium size of the particles in a milling batch has been the topic of much

research and one of the more straightforward methods for calculating this size, based on a

number of variables, follows. This equation applies to all types of milling.

W = mass of particles with surface area < S

∑=

−=i

j

tSiji

jeatW1

)(

Equation 4: Equilibrium size of milled particle (Devaswithin et al., 1988)

⎥⎥⎥⎥⎥⎥⎥

⎦

⎤

⎢⎢⎢⎢⎢⎢⎢

⎣

⎡

>−

=−

<

=

∑

∑−

=

−

=

jiforabSSS

jiforaW

jifor

ai

jkkjikk

ji

i

kikij

1

1

1

1

)0(

0

Equation 5: aij term (Devaswithin et al., 1988)

( ) ( ) ( )( ) ( )vijjijjij xxxxxxxxB //1// 11δβδ ψψ −+=

Equation 6: Bij term (Devaswithin et al., 1988)

( )α1/ xxKS jj =

Equation 7: S term (Devaswithin et al., 1988)

To use this equation the desired mass of reduced size particles and the size of these particles is

specified to yield the milling time required for achieving these parameters. Solving this series

however, involves the use of many constants, which must be determined through calibration

experiments using materials for which the constants are already known (Devaswithin et al.,

1988).

In other areas of research, such as mechanical activation of reactants in a milling vessel,

researchers are focused more on the rate of particle size reduction as opposed to the average

19

particle size of the end product. In general, the rate at which a mill reduces the average size of

the particles being milled is a function of the probability of any particle, at a given time, being

trapped between the participants in the aforementioned type of collisions when those collisions

possess the energy necessary to fracture a particle. This focus has produced a number of

functions, which have been supported empirically, to determine the rate of particle size

reduction. One of the more general examples is shown below for vibrator milling (Tamura and

Tanaka, 1970).

In terms of the change in total surface area of materials

5.12/1

22

'1'1222''

1 23016.0⎪⎭

⎪⎬⎫

⎪⎩

⎪⎨⎧

⎟⎟⎠

⎞⎜⎜⎝

⎛−= +−

ωασαωθ

ρβ

β

dYxUkUxdLDJKk

dtdSM q

Equation 8: Determination of change in surface area of milled material (Tamura and Tanaka, 1970)

where: =U volume of particles in mill/(volume of mill – volume of balls)

== − Uxdn 334.0 number of particles per ball, dimensionless

=S specific surface of particles, cm2 /g.m

=φ ratio of mill speed to critical speed, dimensionless

=L length of mill, cm

=J fractional ball filling of mill, dimensionless

=d diameter of ball, cm

=D diameter of mill, cm

=', xx particle sizes, cm

=Y modulus of elasticity of materials, g.f/cm2

All other terms relate to the material being milled (Tamura and Tanaka, 1970).

20

Since these equations are often very difficult to use, empirical studies of milling variables

may prove to be more useful. Such studies have been published on in the past. In general, the

rate increases with ball density and is greatest when the mill filling ratio (volume of material to

be milled/volume of mill) is approximately 10%-20%, while the volume occupied by milling

balls is approximately 40%-60% of the total mill volume (Austin, 1973).

In the case where one wishes to mechanically alloy materials, the previously discussed

theories apply, however, several additional topics must also be considered. Mechanical alloying

is a high-energy milling process for producing composite materials with an even distribution

(though not homogeneous in the rigorous sense) of one material into another (Suryanarayana,

2001). By definition at least one of the materials must be metallic to be considered an alloy;

however, the topics discussed here can be applied to non-metallic materials as well. Two

systems will be discussed, brittle-malleable and malleable-malleable.

In a malleable-malleable system such as the milling of two soft materials, like sodium

and gold, ball-powder-ball collisions initially reduce the size of both materials until the average

surface area/mass ratio of each particle in the sample is large enough for re-welding to occur.

When this critical surface area is achieved in a particle, re-welding can occur between two

similar particles or two dissimilar particles. If re-welding occurs between two similar particles

the net process results in no change of the material nature. If re-welding occurs between two

dissimilar particles an alloy particle is created. This alloy particle can then undergo further

fragmentation along alternative planes and subsequently be re-welded multiple times. The

longer this process is allowed to take place the more dissolved one material becomes in the other

(Suryanarayana, 2001; Suryanarayana, 2004).

21

In a brittle-malleable system, such as palladium on graphite and magnesium, the more

ductile magnesium is initially flattened while fragmentation of the more brittle

palladium/graphite occurs. With further milling, the ductile material occludes brittle fragments

leading to an individual particle composition of the starting mixture (Suryanarayana, 2001;

Suryanarayana, 2004). Depending on the solubility of the brittle phase, continued milling may

result in near chemical homogeneity, an undesirable result for a system such as this, in which the

target reaction relies on events that occur strictly on the particle surface. In an attempt to

produce Mg/Pd particles at scaled up quantities that would most rapidly dechlorinate PCBs, the

three aforementioned factors which affect vibrator milling (mill filling ratio, volume occupied by

the milling balls, and milling time) were optimized so as to produce the most active Mg/Pd

bimetal.

Possible Degradation Mechanisms

Several papers have discussed possible mechanisms for the reaction of bimetallic systems

with halogenated aromatic compounds. Many of these are in reference to particles prepared

through electrodeposition, which may not be truly relevant to mechanically alloyed Mg/Pd.

Milled Mg/Pd contains carbon as part of the system, whereas electrodeposited Mg/Pd does not,

possibly affecting the mechanistic pathway of the reaction. One of the first papers to address a

bimetallic system was by Cheng et al. for the electrochemical dechlorination of 4-chlorophenol

using palladized graphite electrodes. This study proposed three facts that are key to the to

degradation of chlorinated aromatics by palladized ZVM: 1) hydrogen gas must be evolved by

the ZVM due to the reduction of the solvent, 2) adsorption of the molecular hydrogen forms

powerful reducing species (authors speculate palladium hydride is formed), and 3) reduction of

22

the chlorinated organic occurs after adsorption of the contaminant to the bimetallic surface.

Three possible mechanisms were discussed within this dissertation. The first was the direct

reduction of the contaminant by the zero-valent metal. However, this was discarded because if

this were the case, palladizing the ZVM should have little or no effect on the reaction kinetics,

which is not experimentally observed. Without the palladium catalyst in the system, degradation

did not occur. A second possible mechanistic pathway involves the dechlorination at the surface

of the palladium catalyst, but this too was dismissed as unlikely because experiments conducted

with palladized gauze saturated with H2 gas failed to produce dechlorination when held at an

appropriate current. The third and final possibility was that degradation occurs at the interface

between the substrate and the catalytic islands. Experiments support this mechanism, showing

degradation for palladized iron, palladized graphite, and palladized carbon electrodes. Because

degradation occurs only when both the substrate and catalyst are present, it seems both are

essential for the dechlorination to occur. Cheng et al. postulated that the dechlorination occurs

directly at the interface between the substrate the catalyst, and nothing in the literature seems to

contradict this (Cheng et al., 1997).

This study also showed the importance of hydrogen adsorption to the degradation

process. Similar experiments were conducted using electrodes coated with platinum, however

rate constants were far lower than when palladium is used. Both of these catalysts are well know

for the electrochemical production of H2, however it appears that simply the production of H2 is

not enough for dechlorination to occur. One of the major differences between palladium and

platinum is the ability of palladium to adsorb large amounts of hydrogen within its lattice

23

structure. For platinum, the hydrogen appears to escape quickly once it is evolved, removing it

from the reaction system and stymieing degradation of the contaminant (Cheng et al., 1997).

A more current study was published by Agarwal et al. specifically for Mg/Pd bimetallic

systems which expands upon the work done by Cheng et al. In this work, Mg/Pd bimetallic

systems created by electrodeposition are considered different from Mg/Pd where palladium is

deposited on the magnesium surface by doping. Mechanically alloyed Mg/Pd systems may not

be exactly similar, but this study is deemed an appropriate reference point at least for comparison

purposes. Data from this work are used to suggest a modified mechanism from the one

postulated by Cheng et al. to take into account the self-limiting corrosion behavior of

magnesium. This modified mechanism is based upon the following chemical equations:

Mg Mg2+(aq) + 2e- (Corrosion of Mg)

2H2O + 2e- H2(g) + 2OH- (Electrolysis of water)

H2(g) PdH2 (Intercalation of H2 in Pd lattice)

PdH2 + Ar2Cl Ar2H + PdClH (Hydrodechlorination of organics)

Mg2+ + 2OH- ↔ Mg(OH)2(s) (Dynamic portioning of soluble Mg species)

Mg(OH)2(s) Mg(OH)2 + Mg(OH)2 (Anodic repair and precipitate)

Equation 9: Modified corrosion based scheme for Mg/Pd bimetal particles

The implications of this scheme are that the shelf life of Mg/Pd bimetallic particles in an aqueous

solvent will be extremely long, due to the self-repairing nature of the zero-valent metal substrate.

This is very advantageous when considering possible field applications, such as permeable

reactive barriers (Agarwal et al., 2007).

Neither of these mechanisms explains how the hydrogen (hydride) is replacing the

chlorine atom, and both are conducted in aqueous solvent systems. Nevertheless, they present a

24

starting point for the elucidation of the mechanism for dechlorination by milled Mg/Pd in

methanol solvent.

Research Objectives

The objectives of this dissertation are threefold. 1) Produce an effective, optimized,

mechanically alloyed bimetallic reactive particle capable of being manufactured on a large-scale

both efficiently and cost-effectively, 2) determination of the reaction kinetics, order, and

pathway of the mechanically alloyed Mg/Pd in the dechlorination of various PCB congeners, and

3) proposal of possible mechanism for the degradation of PCBs using Mg/Pd in methanol.

The first objective was completed by first inventing a method of mechanically alloying

Mg/Pd particles on a small-scale using a high energy vibrator mill. These milled particles had an

equivalent reactivity towards the dechlorination of PCBs using 0.08% Pd loading as compared to

a 4% Pd loading required by electroless deposition. This was then scaled-up to accommodate

larger batch sizes to increase the efficiency of the production process. The milling procedure

was optimized for dechlorination activity via several parameters including canister loading,

milling times, %Pd loading, and ball-to-mass ratio. This produced an optimized bimetallic

particle even more reactive than that created on the small scale high energy milling device. The

final optimized procedure for the milling process corresponded to a canister loading of 85 g of

Mg/Pd (78 g magnesium and 7 g 1% palladium on graphite) and milling for 23 minutes using 16

milling balls (ball-to-mass ratio of 4.20).

Kinetic studies were performed on each of the mono-chlorinated PCB congeners in both

methanol and water:methanol (9:1) solvent systems in order to complete the second research

objective. Pseudo 1st order kinetics were seen in agreement with published literature for both

25

solvent systems. When more highly chlorinated congeners were used in methanol, stepwise

dechlorination was also observed. Significant differences were seen in the results between the

two solvent systems, indicating there may be solvent specific mechanisms for the degradation of

PCBs using Mg/Pd. Relative degradation for ortho, meta, and para substitution differed between

the two solvents. Also, the final products of the degradation that were detected were different

depending upon the solvent system being used.

The third and final objective was completed by specifically studying PCB-151, a hexa-

chlorinated congener corresponding to a chlorine substitution pattern of 2, 2’, 3, 5, 5’, 6-

chlorobiphenyl. Several possible mechanisms were considered and discarded for the use of

milled Mg/Pd particles, including a benzyne intermediate and nucleophilic aromatic substation.

Isotopic studies using deuterated methanol pointed to C-H bond formation playing a role in the

rate-limiting step of the mechanism. It has been reported that palladium can produce three

distinct types of reactive hydrogen species when exposed to molecular hydrogen, each with a

different enthalpy of desorption. These correspond to chemisorbed surface atomic hydrogen,

bulk hydride, and “hydride-like” atomic hydrogen bonded within the first layers of the

subsurface palladium lattice (Cybulski and Moulijn, 2006). Combining these ideas, three

possible mechanistic schemes involving the replacement of a chlorine atom via either radical

hydrogen or a hydride species is postulated. At this point, research has not been done to

distinguish which of these species is responsible for the dechlorination of PCBs.

26

CHAPTER TWO: MECHANICAL ALLOYING IN USE FOR THE PREPARATION OF A PALLADIZED MAGNESIUM BIMETALLIC

COMPOUND FOR THE REMEDIATION OF PCBS

Reproduced with permission from Langmuir, submitted for publication. Unpublished work copyright 2008 American Chemical Society.

Introduction

The class of 209 aromatic chlorinated molecules, resulting from the attachment of up to

ten chlorine atoms to biphenyl, are collectively known as polychlorinated biphenyls (PCBs).

Among the properties of these synthetic colorless liquids are high chemical stability, low

flammability, low thermal and electrical conductivity, and low solubility in water (EPA, 1983).

Twenty-nine years following the 1976 Toxic Substances Control Act ban on their manufacture,

PCBs remain a continued environmental threat, their persistence owing to the very high chemical

stability of these molecules. Prior to the TSCA ban, these favorable properties were exploited in

a variety of applications including paint stabilizers, transformer oils, capacitors, printing inks,

flame retardants, anti-fungicides and pesticides (EPA, 1983; Erickson, 1992).

While toxicity evidence was found adequate to justify TSCA regulation, the debate over

the extent of PCB toxicity on organisms remains heated (Erickson, 1992). PCBs are known to

bioaccumulate and concentrate in fatty tissues (Weigel and Wu, 2000). Additionally, studies

suggest that increased incidences of cancer are found among people that have experienced long-

term exposure to PCBs, however these studies are arguably inconclusive as they involve the

simultaneous analysis of multiple congeners and other environmental contaminates. Further

complications arise from the potential for contamination of commercial mixtures with other more

27

toxic chlorinated compounds, such as polychlorinated dibenzodioxins (PCDDs) and

polychlorinated dibenzofurans (PCDFs) (Erickson, 1992).

Until recently only one economical option was available for the treatment of PCB

contaminated materials, incineration. This, however, can be more detrimental to the

environment than the PCBs themselves due to the potential for formation of PCDDs which have

been shown to exist at abnormally high levels in proximity to incinerators burning chlorine-

contaminated materials, especially those burning PCBs. Cancer rates have also been well

correlated to both dioxins and proximity to chlorine-contaminated waste burning incinerators

(Pirard et al., 2005). Other remediation options in use or currently under investigation include

dredging followed by landfilling, or microbial degradation. Dredging and landfilling are not

ideal treatment options however, since the contaminant itself is not remediated, and the

hazardous material is still a problem. Microbial degradation currently suffers from low rate

constants and incomplete degradation, so it is also not a successful remediation option at this

time (Weigel and Wu, 2000).

Recent literature, however, reports the rapid and complete reductive dechlorination of

PCBs using palladium coated iron (Fe/Pd) or magnesium (Mg/Pd) in aqueous medium (Doyle et

al., 1998; Engelmann et al., 2003; Engelmann et al., 2001 ; Grittini et al., 1995; Korte et al.,

2002; Liu et al., 2001; Muftikian et al., 1996; Wang and Zhang, 1997; Zhang et al., 1998). The

use of these bimetallic particles for dechlorination relies on the reduction potentials of the zero-

valent metal coupled with the hydrodehalogenation-type catalytic activity of palladium, although

the exact mechanism for the dechlorination is not known at this time.

28

This research focuses on the Mg/Pd bimetallic system as it has several advantages over

Fe/Pd. One advantage is the ability of Mg/Pd to dechlorinate in the presence of oxygen, in

which micro- or nano-scale iron is fully corroded. This is possible because magnesium forms

thin, oxygen impermeable, but slightly water soluble, oxide layers. Additionally the magnesium

or iron acts as a reductant (electron donor) during the removal of chlorine from PCBs, thus

another advantage arises from the greater thermodynamic driving force of magnesium versus

iron, as demonstrated by a comparison of reduction potentials (Doyle et al., 1998):

Mg2+ + 2e- Mg0 E0 = -2.20 V vs. SHE

Fe2+ + 2e- Fe0 E0 = -0.44 V vs. SHE

Equation 10: Standard reduction potential for Mg and Fe

Classically, the palladium catalyst acts as a hydrodehalogenation catalyst by dissociating

hydrogen gas (in this case, formed from the reaction of Mg0 or Fe0 with water or another proton

donor) adsorbed onto the palladium surface to produce atomic hydrogen, which can then react

via radical mechanisms with chlorine atoms attached to an organic molecule. The following is

the generally accepted dechlorination reaction for Mg/Pd:

M0 + 2H2O M2+ + H2 + 2OH-

H2 2H· (due to Pd catalyst)

2RCl + 2H· (dissociated on catalyst surface) 2RH + Cl2

Equation 11: Production of atomic hydrogen on Pd and replacement of a Cl atom

The overall reaction for biphenyl may be expressed as followed

C12HxCly (aq) + (x + y) M0(s) + (x + y) H+(aq) C12H10 (aq) + (x + y) M2+

(s) + (x + y) Cl-(aq)

Equation 12: Overall reaction for dechlorination of PCBs to biphenyl

29

In this scheme the reaction products are biphenyl and chloride ions, however, while these are the

observed products of this reaction, the mechanism for hydrodehalogenation of PCBs to biphenyl,

by these bimetallic particles, has not yet been fully elucidated. Recent studies, though, have

concluded that the dechlorination process is step-wise (Korte et al., 2002). Additionally, some

studies have shown that decomposition of biphenyl is possible, depending upon the solvent

system being studied.

Initially the bimetal was prepared by deposition of palladium onto the magnesium surface

by reaction of zero-valent magnesium with hexachloropalladate or palladium acetate, however,

to produce reasonable kinetics for remediation, at least at 0.5% palladium coating was required.

It was hypothesized that this was due to the majority of the palladium coating only a small

portion of the magnesium powder during deposition, presumably a result of the increased activity

of a magnesium particle that is coated in palladium, thereby reducing: the dispersion of

palladium through the material, the active surface area, and, thus, the overall activity of the

metal. While this hypothesis has been neither refuted nor supported by surface analysis, the

simple fact is that a large quantity of palladium was required to produce an active bimetal. Since

this was entirely non-cost effective, due to the extremely high cost of palladium, an alternate

process for producing the bimetal, mechanical alloying through the use of vibrator milling, was

attempted.

The first attempts at producing active mechanically alloyed Mg/Pd were carried out using

a SPEX 8000M Mixer/Mill high-energy vibrator mill in which 4-μm magnesium was milled with

palladium impregnated on graphite, chosen instead of pure palladium because the highly

dispersed palladium on graphite is much more catalytically active than palladium powder and is

30

much easier to handle. Ball-to-mass ratios and loading levels were not considered while semi-

optimizing the process; instead recommended loading levels and ball to mass ratios were used.

Milling time and percent palladium were the only variables considered. It was then determined,

after several attempts at producing active material, that 5.55 g of bimetal (0.0909% Pd, 8.18% C,

99.1% Mg) milled for three minutes with two 10-g steel balls, in the Spex CentiPrep Tungsten-

Carbide milling vessel, matched the reactivity of a (4% Pd, 96% Mg) bimetal produced by the

electrodeless deposition method. This was much more economical, however, mass production of

the bimetal was impossible using a mill that produced only 5.55 g of material at one time. Since

an efficient large-scale mechanical process for preparation of the bimetal was necessary, a nearly

infinite combination of variables, which were disregarded in initial mechanical alloying attempts,

had to be reconsidered, understood, and narrowed to a reasonable range before optimization of

any upscale process could begin. The optimization and experimental procedures for the analysis

of the reactivity of the bimetal is in the following section.

Materials and Methods

Chemicals. Neat PCB single congener standards were obtained from Accustandard while

Arochlor mixtures were obtained from Supelco. Optima© grade methanol and toluene and

HPLC-grade hexane were all obtained from Fisher Scientific. Potassium hexachloropalladate

(99%) was obtained from Acros Organics. Magnesium (~4-μm) was obtained from Hart Metals,

Inc. and used as received. 1% palladium (on graphite) was obtained from Engelhard and 10%

palladium (on graphite) was obtained from Aldrich Chemicals, and both were used as received.

Sodium sulfate was also obtained from Aldrich Chemicals.

31

Preparation of Mg/Pd bimetal by electroless deposition. Mg/Pd bimetals were prepared

under an argon atomosphere by first placing 20 g of 4-μm magnesium powder and 100 mL of

deoxygenated deionized water, prepared by bubbling argon gas through deionized water, in a

250-mL Erlenmeyer flask equipped with a magnetic stirring bar. An appropriate quantity of 20

wt. % hexachloropalladate in water was then pipetted into the flask while rapidly stirring the

slurry, so as to produce 0.01, 0.05, 0.1, 0.5, 1, 2, 3, 4, 5, and 10% Pd bimetals. The slurry was

allowed to stir for an additional 1 minute, at which point it was transferred to a filter funnel,

filtered, and copiously rinsed with ethanol to prevent further hydroxide formation from the

presence of water. The powder was then thoroughly dried by continued filter suction in the

argon atmosphere.

Preparation of Mg/Pd bimetal on SPEX 8000M Mixer/Mill. Mg/Pd samples were prepared

under an argon atmosphere by first placing 5.50 g of 4-μm magnesium powder and varying

quantities of 10 wt. % palladium on activated carbon in a 55.4-mL tungsten carbide canister

purchased from SPEX CentiPrep. Two 10.00-g stainless steel balls, also purchased from SPEX

CentiPrep, were then placed in the canister. After loading, the canister was closed, removed

from the inert atmosphere box, and clamped into the SPEX 8000M Mixer/Mill. Milling times

were 0, 1, 2, 3, 5, 7, 10, and 30 minutes and the quantities of palladium on carbon used were

0.010 g, 0.020 g, 0.050 g, 0.070 g, 0.100 g, 0.200 g, 0.300 g, 0.500 g, 1.00 g, 1.50 g, and 2.00 g.

After milling was completed, the canister was returned to an argon atmosphere for re-opening.

Sample Analysis and Experimental Setup for Mg/Pd prepared on SPEX 8000M Mixer/Mill

and by electroless deposition. Sample vials were prepared in an inert atmosphere box by

placing 1.00 g Mg/Pd, 10.00 mL of deoxygenated deionized water (prepared by bubbling argon

32

gas through deionized water), and 80.0 μL of 5000-μg/mL Arochlor 1260 in methanol (supplied

by Supelco) in 20-mL crimp top vials, yielding a 40-ppm PCB solution. Control vials were

prepared in an identical fashion without the addition of the bimetal. After crimping, vials were

removed from the argon atmosphere and placed on a shaker table. At the required reaction time,

vials were placed in a VWR Scientific Aquasonic Model 750D ultrasound bath for thirty minutes

(at full power) prior to opening. 5.00 mL of hexane was then added; the vials were re-crimped,

and placed in the ultrasound bath for an additional thirty minutes. Following ultrasound, the

contents of each vial were transferred to 15-mL centrifuge tubes and centrifuged for five

minutes. The hexane layer was then transferred to a clean screw top vial, 1 g of Na2SO4 was

added, and the dried solution was decanted.

Separation and analysis of the components were performed on an HP-5890 Series II Plus

GC-FID. The initial oven temperature was 100° C and was held for four minutes then ramped

15° C/min. to 280° C at which point it was held for fourteen minutes. The injector and detector

temperatures were set to 260° C. A purge was set to begin at 0.60 minutes and stop at 1.50

minutes. The column flow rate was 1.26 mL/min. The column used was a RTX-5 column (30

m, 0.32 mm i.d., 0.25 µm df) purchased from Restek Corporation.

Preparation of Pd/Mg bimetals on scaled up mill. Mg/Pd bimetals were prepared similarly as

when prepared on the SPEX 8000M Mixer/Mill except that galvanized steel pipes, with an

internal diameter of 5.03-cm and a length of 17.80-cm, fitted with end caps were used as milling

vessels and 1.6-cm3 stainless steel ball bearings, weighing 16.32-g each, were used in place of

the grinding material provided by SPEX CentiPrep. Additionally, a Red Devil 5400 twin arm

33

paint shaker, fitted with custom plates to hold the milling canisters, was used as the mill (as

shown in Figure 1).

Figure 1: Picture of the milling canisters, templates and mill

Optimization of the milling procedure was carried out by preparing bimetals with

varying: numbers of milling balls used, relative (to magnesium) quantities of palladium on

graphite used, total loading quantities (mass of bimetal milled), and lengths of time that the mill

was run. Each variable was initially isolated and varied while leaving all other parameters

constant and set at the middle point of each variable range. For instance, to determine the most

effective milling time, the canister was filled 50% with ball bearings and 15% with the palladium

and magnesium mixture. The material was milled for varying periods of time then each material

produced was tested for effectiveness at degrading PCB’s. Using this method, the optimum

milling time was initially found to be 30 minutes. Other variables were then individually varied

in a similar manner while keeping the milling time constant at 30 minutes. Once rough estimates

of the optimum values for each parameter were obtained, the process was repeated, in a

successive approximations type method, using the best known values for each parameter in place

34

of the middle point of each variable range. This procedure was repeated three times for each

variable and values of the parameters used in the last optimization study are listed in Tables 1-3.

Table 1: Palladium loading optimization parameters

Mass of 1% Pd/C (g) % Pd in Bimetal 1 0.0117 5 0.0588 7 0.0824 9 0.106 14 0.165

Table 2: Ball to mass ratio optimization parameters

# of Milling Balls Used per 85 g Mg/Pd Canister Loading Ball To Mg/Pd Mass Ratio 4 1.049411765 8 2.098823529 12 3.148235294 16 4.197647059 20 5.247058824 24 6.296470588 32 8.395294118

Table 3: Additional Mg/Pd optimization parameters

Total Bimetal Loading per Canister (g) Milling Times (min) 15.91 3 26.54 8 31.81 15 42.43 23 58.31 30

85 38 127.19 45

Sample Analysis and Experimental Setup for Mg/Pd prepared on scaled up mill. Initially,

optimization studies were carried out Pd loading (relative to Mg). For this, sample vials were

prepared in an inert atmosphere box by placing 0.100 g Pd/Mg and 10.00 mL of 6.0-ppm 1254

Arochlor in 90% deoxygenated deionized water/10% methanol (prepared by bubbling argon gas

through the mixture for one minute) in 20-mL screw cap vials. Control vials were prepared in an

35

identical fashion both without the addition of the bimetal and as zero hour samples with the

addition of bimetal (extracted after 5 minutes of equilibration time). After closing, vials were

removed from the argon atmosphere and placed on a shaker table. At the required reaction time

10.00 mL of toluene was added to each vial and placed in a VWR Scientific Aquasonic Model

750D ultrasound bath for sixty minutes. Following ultrasound, the contents of each vial were

transferred to a 15-mL centrifuge tube and centrifuged for five minutes. The toluene layer was

then transferred to a clean screw top vial, 1 g of Na2SO4 was added, and the dried solution was

decanted.

Experimental conditions were altered for the remaining optimization parameters, to

simply the analysis process. Sample vials were prepared in an inert atmosphere box by placing

0.250 g Mg/Pd and 10.00 mL of 10.0-ppm PCB-151 (a hexa-chlorinated PCB congener) in

methanol in 20-mL screw cap vials. Control vials were prepared in an identical fashion both

without the addition of the bimetal and as zero hour samples with the addition of bimetal

(extracted after 5 minutes of equilibration time). After closing, vials were removed from the

argon atmosphere and placed on a shaker table. At the required reaction time 10.00 mL of

toluene was added to each vial. The vials were then returned to the shaker table for two minutes,

after which a 5-mL lure lock SGE brand gas tight syringe, fitted with a nylon Millipore Millex

(model-HN) 0.45-μm filter, was used to remove 4 mL of the solution. The solution, containing

50% Methanol, 50% Toluene, and PCBs, was then transferred to a 15-mL centrifuge tube and 2

mL of deionized water was added. The mixture was shaken for one minute then centrifuged for

one minute and the toluene layer, containing the PCBs, was removed. The toluene layer was

36

then transferred to a clean screw top vial, 1 g of Na2SO4 was added, and the dried solution was

decanted.

Experimental setup for extraction efficiency/adsorption study. Sample vials were prepared

in an inert atmosphere box by placing, in four vials each (two to be used as zero hour samples

and two to be extracted after 24 hours), 0.250 g Mg/Pd, 0.229 g Mg, or 0.021 g Pd on graphite.

It should be noted that these masses were chosen because they are representative of the

respective quantities that would be found in 0.250 g of the optimized Mg/Pd. 10.00 mL of 10.0

ppm PCB-151 in methanol was then added to each of the twelve vials and the vials were closed,

removed from the inert atmosphere box, and placed on a shaker table. The extraction procedure

used for these samples was identical to that which was described in the preceding section.

Separation and analysis of the components were performed on an Autosystem XL GC-

ECD. The detector and injector temperatures were 325° C and 275° C respectively. The initial

oven temperature was 120° C and was held for one minute then ramped 20° C/min. to 200° C, at

which point it was immediately ramped 2° C/min, to 270° C. It was then ramped 20° C/min, to

300° C at which point it was held for 10 minutes. The column flow rate was 1.3 mL/min., the

sample volume was 1 μL, the attenuation was 0, and the offset voltage was 5 mV. The column

used was a RTX-5 (30-m, 0.32-mm i.d., 0.25-μm df) purchased from Restek Corporation.

Results and Discussion

The complications involved in quantitative analysis of PCBs are well documented

(Erickson, 1992; Doyle et al., 1998; Engelmann et al., 2003; Zhang et al., 1998). These

complications are due largely to the multitude of possible congeners present in the PCB samples.

Commercial mixtures, those found contaminating the environment, generally contain one

37

hundred or more congeners all possessing similar physical and chemical properties. These

similarities lead to difficulties in separation. Variations in the degree of specific congener

volatility and response to methods of analysis also contribute significant errors in quantification

(Engelmann et al., 2003). Further skewing PCB estimation, is the step-wise nature of the

mechanism by which hydrodehalogenation occurs, that leads to the appearance of congeners not

present in the original mixture. On chromatograms, PCB mixtures appear as a cluster of peaks in

a given range known as the "PCB envelope" as shown in Figure 2. This characteristic pattern is

utilized by EPA method 8082 'Polychlorinated Biphenyl by Gas Chromatography' and is

dependent on the analyst's aptitude for pattern recognition. Unpredictable degradation pathways

and the aforementioned differences in volatility and solubility among congeners make the pattern

recognition approach all the more unreliable (Engelmann et al., 2003). Another option, EPA

method 508A, 'Screening for Polychlorinated Biphenyls by Perchlorination and Gas

Chromatography', utilizes perchlorination to convert all congeners to decachlorobiphenyl prior to

analysis but this method still produces up to a 25% error in quantification (Engelmann et al.,

2003).

To simplify the analysis of PCB degradation in the studies conducted on metals prepared

on the SPEX CentiPrep 8000 and electrodeless deposited Pd/Mg, production of the biphenyl

byproduct was monitored; however, it was observed that the biphenyl peak disappeared over

time along with the peaks in the PCB envelope. Engelmann et al. encountered the same

phenomenon and reported that ring degradation occurs as biphenyl reacts with the bimetallic

reductants, thus, the appearance of biphenyl on chromatograms served only as an indicator that

the bimetallic system was active and was not useful for accurate quantification purposes

38

(Engelmann et al., 2003). Nevertheless, the relative rates at which biphenyl was produced, as

determined by observing the time at which the maximum biphenyl concentration was reached

prior to its complete degradation, were noticeably different for bimetals prepared with differing

quantities of palladium or those prepared under different milling conditions. An ordering of the

relative activity of all the bimetals prepared could then be arranged and, in this way, it was

determined that the optimum bimetal could be prepared by milling 5.50 g of magnesium with

0.05 g of 10% palladium on graphite for 3 minutes, using the technique described in the

“Materials and Methods” section of this chapter paper. Using this bimetal, the cluster of peaks

within the PCB envelope does not appear on chromatograms after 4 days (Figure 3) of reaction

time and the biphenyl peak disappears within two weeks (Figure 4). This procedure for

determination of the optimum palladium content and milling time was repeated four times and

the results were found to be reproducible.

Figure 2: GC-FID chromatogram for 40-ppm Arochlor 1260 control

39

Figure 3: GC-FID chromatogram for 40-ppm Arochlor 1260 sample after 72 hr exposure to 1.00 g Mg/Pd

Figure 4: GC-FID chromatogram for 40-ppm Arochlor sample after 96 hr exposure to 1.00 g Mg/Pd

As mentioned in the “Introduction” section of this chapter, another method for the

preparation of these bimetallic particles, relying on electroless deposition of zero-valent

palladium onto the particle surface, appears in recent literature. Complete degradation of PCBs

within seventeen hours of reaction time was reported with bimetallic particles prepared this way

(Gritinni et al., 1995; Wang and Zhang, 1997). However, for comparison, this experiment was

repeated sixteen times, failing to ever reproduce the reported results. In fact, when a quantity of

Pd equivalent to that used in the optimized Mg/Pd prepared on the SPEX CentiPrep was

electroless deposited by this method, the bimetal caused very little PCB degradation even after

40

ten days of reaction time. To produce kinetics equivalent to the mechanically alloyed material a

4% Pd coating, via electroless deposition, was required.

As stated previously, the analysis of PCB degradation by direct observation of a PCB

envelope observed for industrial mixtures such as Arochlors is virtually impossible, thus, the

analysis of the Arochlor 1254 mixtures used in studies conducted on bimetals prepared on the

scaled up mill were also simplified as follows. Since PCB degradation has been shown to occur

via a stepwise mechanism in which chlorine atoms are removed from higher chlorinated

congeners to produce lower chlorinated congeners before being completely dechlorinated to

biphenyl, the relative concentrations of parent congeners

(higher chlorinated congeners) and their dechlorination product congeners (lower chlorinated

congeners) will change according to reaction progress, as shown in Figure 5 and Figure 6 (Korte

et al., 2002).

Figure 5: Chromatogram of 10.0-ppm 1254 Arochlor in 9:1 water:methanol control

41

Figure 6: Chromatogram of 10.0-ppm Arochlor 1254 in 9:1 water:methanol after 24 days exposure to 0.100 g of Mg/Pd

This fact was used to monitor dechlorination reaction progress for Arochlor 1254

mixtures by observing the relative ECD signals of parent:byproduct congener pair,

corresponding to one of the major hexa-chlorinated congeners in the Arochlor 1254 mixtures and

one of its single chlorine removal products, a penta-chlorinated congener. The ratio of the

analytical signals of these two congeners was plotted vs. reaction time (see Figure 7) and the

slopes of these plots were used to quantify, in a strictly relative sense, the reactivity of the

bimetals tested (see Figure 8).

Parent:Product Signal Ratio From Degradation of Arochlor 1254 vs Time for Determination of Optimum Palladium Loading

(0.05%Pd)

0.50

0.55

0.60

0.65

0.70

0.75

0.80

0 50 100 150 200 250

Time (hours)

Pare

nts

vs. P

rodu

ct

Con

gene

r Sig

nal R

atio

Figure 7: Plot to determine activity of Mg/Pd using parent:product signal ratio

42

Activity vs. %Pd in Bimetallic for the Optimization of Pd Mass Loading

0

0.0002

0.0004

0.0006

0.0008

0.001

0 0.05 0.1 0.15 0.2

%Pd in Bimetallic

Act

ivity

(slo

pe fr

om

degr

datio

n of

125

4 pl

ots)

Figure 8: Plot of the activity of Mg/Pd for the optimization of the %Pd loading

As can be seen from Figure 8, the bimetal that had the highest reactivity towards

dechlorination of the PCBs found in the Arochlor 1254 solution was the one containing 0.083%

Pd. While one might expect that the reactivity of a Mg/Pd bimetal would increase constantly

with increasing palladium content, one explanation for the observed plateau effect that occurs in

the later half of the plot is that, at some point, the graphite content of the bimetal becomes so

high that the magnesium particles become thoroughly coated with graphite, thereby isolating the

magnesium surface, thus preventing the necessary oxidation of the particle. No work has yet

been conducted, however, to validate this hypothesis.

The remainder of the work completed for optimization of milling parameters (ball to

mass ratio, canister loading, and milling time) was carried using a single congener (PCB-151) for

ease of analysis. However, since this analytical method relies on the measurement of the

concentration of PCB-151 in the reaction mixture and since the bimetals contained activated

43

carbon, a common sorbing agent for many organic molecules, the efficiency of the extraction

procedure had to be verified. This was done by comparing the measured concentration of PCB-

151 at different times in separate samples that were allowed to contact the bimetal and each of its

individual components. Figure 9 shows chromatographic results of this study. As can be seen,

the extraction procedure used throughout this work was shown to be highly efficient. This

coupled with the fact that reaction products are only seen for the ball milled material,

conclusively proves that the results reported in this paper are not due to adsorption of PCBs by

the activated carbon but are, in fact, due to true degradation of PCBs. Figures 10, 11 and 12

show the results of the additional milling parameter optimization studies.

Figure 9: GC-ECD chromatograms of 10 mL of 10.0-ppm PCB-151 solution reacted with (a) Mg/Pd, (b) Mg, and (c) Pd/graphite

44

Pseudo 1st Order Rate Constants From the Degradation of PCB-151 vs. Milling Time Plot for the Optimization of Milling

Parameters

0

0.001

0.002

0.003

0.004

0.005

0.006

0.007

0.008

0 5 10 15 20 25 30 35 40 45 50Milling Time (min)

Pse

udo

Firs

t Ord

er R

ate

Con

stan

t (1/

min

)

Figure 10: Plot of the activity of Mg/Pd for the optimization of the milling time

Pseudo 1st Order Rate Constant From the Degradation of PCB-151 vs. The Number of Milling Balls Used Plot for the

Optimization of Milling Parameters

0

0.001

0.002

0.003

0.004

0.005

0.006

0.007

0.008

0 5 10 15 20 25

# Balls per 85g Mg/Pd

Pseu

do F

irst

Ord

er R

ate

Cons

tant

(1/m

in)

Figure 11: Plot of the activity of Mg/Pd for the optimization of the # of milling balls

45

Pseudo 1st Rate Constant From the Degradation of PCB-151 vs. Mass of Mg/Pd Milled Plot for the Optimization of

Milling Parameters

0

0.001

0.002

0.003

0.004

0.005

0.006

0.007

0.008

0 20 40 60 80 100 120 140

Mass of Pd/Mg Milled (g)

Pse

udo

Firs

t Ord

er R

ate

Cons

tant

(1/m

in)

Figure 12: Plot of the activity of the Mg/Pd for the optimization of canister loading

As can be seen, the milling parameters for maximum bimetal reactivity correspond to

milling 85 g of Mg/Pd (78 g magnesium and 7 g 1% palladium on graphite) for 23 minutes with

16 milling balls. An SEM micrograph of an Mg/Pd particle produced under these optimized

conditions, showing that the mechanical alloying process successfully embedded palladium

fragments (small white particles) in the magnesium surface (grey), is given below.

46

Figure 13: SEM micrograph of an Mg/Pd particle (2700x magnification). White line (scale) represents 1 um

While nothing remarkable was observed for optimization of the ball to mass ratio (Figure

11) or the milling time (Figure 10), the data obtained for the optimization of bimetal mass

loading (Figure 12) was unexpected in that no simple correlation between the parameter and

reactivity was observed, though no explanation for this strange behavior has been put forward.

The study was, however, repeated and statistically equivalent results were obtained, thus, it can

be reasonably concluded that the observed behavior is, in fact, the result of some physical

process. It should be noted, though, that while the goal of this work was to determine the

optimum parameters for preparing the most reactive Mg/Pd bimetal; for economical reasons, the

amount of palladium used could be reduced from 0.0823% to 0.0118% of the bimetal mass and

the mass of bimetal milled in each canister could be increased from 85 g to almost 130 g while

reducing the bimetal’s rate of PCB degradation by only about 50%. For large scale application

47

of this bimetal, these parameters may be used as a starting place for optimization, from an

economical perspective, of the milling parameters.

Finally, to determine if the up scaling process was successful in producing equivalently

reactive bimetal, the degradation rates of PCB-45 (chosen simply because it was available at the

time that these results were needed), for the materials prepared on the scaled up mill and on the

SPEX 8000M Mixer/Mill, were measured and compared. The results of this study as well as the

results for degradation of PCB-45 by material containing the same percentage of palladium as

the bimetal prepared on the scaled up mill, prepared by electrodeless deposition, are given in

Figure 14.

Pseudo First Order Rate Constants for Three Bimetals With Equivalent Pd Content

00.0020.0040.0060.008

0.010.0120.0140.0160.018

Scaled Up Mill SPEX 8000M Mixer/Mill Electrodeless Deposition

Technique of Bimetal Preparation

Psue

do F

irst O

rder

Rat

e Co

nsta

n(1

/s)

Figure 14: Pseudo 1st order rate constants for three bimetals with equivalent Pd content

As can be seen, the rate constant (0.0157 s-1) for the material prepared on the scaled up mill is

actually greater than the rate constant (0.0112 s-1) for the material prepared on the SPEX 8000M

Mixer/Mill. This study also shows the massive increase in reactivity that is observed for material

48

prepared by this mechanical alloying process relative to material prepared by electroless

deposition (rate constant = 0.000217 s-1).

Conclusion

Simplification of the analysis of PCB degradation, by monitoring biphenyl production

rather than reduction in PCB concentration, allowed for the determination of the optimum

milling parameters for preparation of a particulate Mg/Pd bimetal on a SPEX 8000M Mixer/Mill

high energy figure 8 vibrator mill. These parameters corresponded to milling 5.5 g of 4-μm

magnesium powder with 0.05 g of 10 wt. % palladium on graphite, obtained from for three

minutes, under an argon atmosphere, with two 10-g stainless steel milling balls, in a 55.4-mL

tungsten carbide milling vessel. It was determined that this mechanically alloyed bimetal caused

degradation of PCBs several orders of magnitude faster than bimetal prepared by electrodeless

deposition of palladium onto magnesium, possibly due to more even dispersion of the catalyst

throughout the metal.

This mechanical alloying process, however, was capable of producing only 5.55 g of

material during one 3 minute milling procedure, thereby making it unsuitable as an industrial

process for large scale production. A scaled-up method of producing the bimetal was therefore

necessary, and required re-optimization of the bimetal to confirm reactivity wasn’t lost during

the scaling up process. The following milling parameters were optimized: Pd loading, ball to

mass ratio, total bimetal mass loading, and milling time. It was determined, through many

variations of the milling parameters, that the optimum bimetal could be prepared by milling 7 g

of 1 wt. % palladium on graphite with 78 g magnesium, for 23 min, with sixteen 16.32-g

stainless steel ball bearings (3.07 ball to mass ratio). This bimetal was proven more reactive than

49

the small-scale bimetal, and far more reactive than bimetal prepared using electroless deposition.

In summary, a large-scale method for the production of Mg/Pd has proven more efficient and

effective than that of electroless deposition.

50

CHAPTER THREE: DECHLORINATION COMPARISON OF MONO-SUBSTITUTED PCBS WITH Mg/Pd IN DIFFERENT SOLVENT SYSTEMS

Reproduced with permission from Chemosphere, submitted for publication. Unpublished work copyright 2008 Elsevier.

Introduction

Polychlorinated biphenyl (PCB) is an overarching term used to denote the family of 209

congeners having the generic formula C12HnCl10-n. Before governmental regulation of PCB’s in

the 1970’s, mixtures of multiple PCB congeners, commonly known as Arochlors, were used in a

myriad of industrial applications because of their high boiling points, high degree of stability,

low flammability, antifungal properties and low electrical conductivity. These applications

included; use in transformers and capacitors as a dielectric fluid, heat transfer and hydraulic

fluids, dye carriers in carbonless copy paper, paints, adhesives and caulking compounds, and

fillers in investment casting wax (EPA, 1983). Over time many of the containers and/or

compounds to which PCBs were added have themselves broken down allowing the highly stable

PCBs to enter into the environment where they may be dispersed mainly through atmospheric

transport (Eisenreich et al., 1983).

Reductive dechlorination is the primary means of PCB remediation. In this reaction

chlorine atoms on the biphenyl are replaced (one at a time) by hydrogen atoms, producing a

mixture of daughter products whose structure is determined by the position of the chlorine

removed from the parent PCB. Although ideally the reductive dechlorination reactions will lead

to the non-toxic, final product of biphenyl, each stepwise reaction is important since less

chlorinated PCBs have been shown to be less toxic (Quensen et al., 1998; Mousa et al., 1996),

have lower bioaccumulation factors, and are more susceptible to aerobic metabolism such as

51

mineralization and ring opening (Mousa et al., 1996; Bedard et al., 1987). The most common

remediation technique in use today is high temperature incineration; however the costs

associated with fuel and the production of highly toxic byproducts such as polychlorinated

dibenzo-p-dioxins and polychlorinated dibenzo-furans (commonly referred to as dioxins) (Wu et

al., 2005) which may result from incomplete incineration make this process less than ideal.

Dredging and subsequent land filling of the contaminated soils and materials has also been used

as a means of site decontamination, however due to the highly stable nature of PCBs this

technique merely displaces the point source of the PCB contamination. A third technique

currently being investigated is microbial degradation but initial results show slow reaction rates

and the inability of the microbes to completely dechlorinate PCBs.

A preferable technique would be one which completely dechlorinates PCBs, could be

applied in situ, and is low cost. To this end, much recent work has focused on the use of zero-

valent metals such as magnesium, zinc and iron. Initial studies showed that zero-valent iron

dechlorinated PCBs at temperatures above 200°C (Chuang et al., 1995) and subsequent research

has shown that coating zero-valent iron with palladium, a known hydrodehalogenation catalyst

(Wang and Zhang, 1997; Cheng et al., 1997; Li and Farrell, 2000; Kovenklioglu et al., 1992;

Lowry and Reinhard, 1999), allows the dechlorination of various chlorinated organic compounds

to proceed both at a faster rate and at ambient temperatures. Studies conducted by Muftikian and

co-workers showed rapid dechlorination of polychlorinated ethene (PCE) using a palladium/iron

(Pd/Fe) bimetal (Muftikian et al., 1995) and Grittini and co-workers (1995) were one of the first

groups to demonstrate the effectiveness of the Pd/Fe bimetal for the dechlorination of PCBs,

however they did not quantify their results (Grittini et al., 1995). In this same article, Grittini

52

proposed that the greater reactivity seen with the addition of the palladium catalyst is due to the

palladium’s ability to adsorb molecular hydrogen generated from the reaction of iron and water.

Fe0 + 2H2O → Fe2+ + H2 + 2OH-

Equation 13: Generation of H2 from Fe and H2O

In the studies conducted for this chapter, palladium is used as the hydrodehalogention catalyst

similar to the studies referred to thus far, however magnesium was chosen as the zero-valent

metal component of the bimetal for the following reasons. Unlike iron, magnesium forms a self-

limiting oxide layer upon exposure to oxygen. This allows the Mg/Pd bimetal to be used in

normal atmospheric conditions as opposed to the Fe/Pd bimetal which can only be used in an

inert atmosphere environment. Secondly, magnesium provides a greater thermodynamic force

when compared to other zero-valent metals such as iron and zinc:

Mg2+ + 2e- Mg0 E0 = -2.37 V

Fe2+ + 2e- Fe0 E0 = -0.44 V

Zn2+ + 2e- Zn0 E0 = -0.76 V

Equation 14: Standard reduction potential for Mg, Fe, and Zn

The studies discussed in this chapter were designed for the purpose of elucidating the

mechanism of PCB dechlorination since a fundamental mechanistic understanding of the

reaction is vital to tailoring a bimetal system for maximum effectiveness. Because previous

studies were conducted in water (Kim et al., 2004), the –ortho, -meta and –para (see Figure 15)

congeners of monochlorinated biphenyl were studied in both a water/methanol (9:1) mixture and

pure methanol to determine solvent effects on the dechlorination mechanism. Additionally the

final composition of the products was monitored to see if further degradation of the biphenyl

53

product was occurring. Lastly, kinetic irregularities seen during approximately the first 30

minutes of PCB degradation studies conducted in methanol were investigated.

Cl

Cl

Cl

2-Monochlorobiphenyl

3-Monochlorobiphenyl 4-Monochlorobiphenyl Figure 15: Structures of PCB-001, PCB-002, and PCB-003

Experimental

Materials and Chemicals: Neat PCB standards were obtained from Accustandard and Optima©

grade methanol and toluene were obtained from Fisher Scientific. 99.0% Methanol-d was

obtained from Acros Organics. Magnesium (~ 4-μm) was obtained from Hart Metals, Inc. and

used as received. 1% palladium (on graphite) was obtained from Engelhard and was used as

received. 10% palladium (on graphite) was obtained from Acros Organics and was used as

received. A ~0.08 wt % palladium-magnesium mixture was prepared by ball-milling 78 g Mg

with 7 g of 1% palladium on carbon in a stainless steel canister (inner dimensions 5.5-cm by 17-

cm) with a 16 steel ball bearings at a total mass of 261.15 g (~16.32-g per ball bearing). The

54

material was milled for 30 minutes using a Red Devil 5400 series paint mixer. A ~0.8 wt %

palladium magnesium mixture was prepared in a similar fashion.

In order to “reactivate” the bimetallic compound, the already prepared material (85 g)

was milled for an additional 30 minutes.

Solutions with 50-μg/mL concentration (in methanol) and 5-μg/mL (in 9:1

water:methanol) of PCB-001, PCB-002, and PCB-003 respectively were prepared with

corresponding calibration standards.

Experimental rate constants were normalized using a ρ constant (g of Mg/Pd per L of

solution) to allow comparison between various studies.

Experimental Procedure: Vial studies using 0.25 g of Mg/Pd and 10 mL of individual PCB

solution in 20-mL vials (with PTFE lined caps) were conducted. Samples were placed on Cole-

Parmer Series 57013 Reciprocating Shaker table (speed 7) until appropriate extraction time.

Each extraction was performed as follows: Exactly, 10 mL of toluene was placed into the vial.

The resulting mixture was then shaken by hand for two minutes. Next, 4 mL of this miscible

solution was pulled with a glass syringe with a Millex® 0.45-μm nylon syringe filter attached and

placed into a centrifuge tube with a PTFE lined cap (to prevent evaporation). The mixture was

then shaken by hand for two minutes followed with centrifugation for two minutes. The top

layer of the extract was collected for further analysis. Studies in 9:1 water:methanol were

conducted similarly, however the starting weight of Mg/Pd was 50 mg and after the 2 minutes of

shaking, a five minute ultrasound treatment was performed, to more fully extract the PCBs from

the surface of the bimetal. The extraction efficiency for this separation technique was ~97%

with a relative standard deviation of less than 4%.

55

Analysis of the extracted samples were performed on a Shimadzu GC-2014 w/TOF MS

and a Thermo Finnigan Trace GC/DSQ, both with a RTX-5 column containing 5% diphenyl-

95% dimethyl polysiloxane with the temperature was ramped from 120°C to 270°C.

Identification of each of the mono-chlorinated PCBs was based upon the retention times of

known standards.


Zero-valent magnesium coated with 1% palladium on graphite showed 50%

dechlorination of a 10 mL 50-μg/mL methanol solution of PCB-001 within 30 minutes and had a

normalized pseudo-1st order rate constant of k = 0.0011 L min-1 g-1 . The reaction with PCB-002

and PCB-003 with Mg/Pd showed 50% dechlorination slightly after 2 hours and the normalized

pseudo-1st order rate constants of k = 0.00045 L min-1 g-1 and k = 0.00052 L min-1 g-1,

respectively. Pseudo-1st order kinetics plots of each reaction are found in Figures 16-18.

56

Pseudo 1st Order Degradation Plot of PCB-1 in MeOH w/ MgPd

y = -0.0276x + 0.7545R2 = 0.9219

-1

-0.8

-0.6

-0.4

-0.2

030 35 40 45 50 55 60

Reaction Time (min)

Ln([C

]/[C

]o)

Figure 16: Pseudo 1st Order kinetic plot of the degradation of PCB-001 with Mg/Pd in methanol


y = -0.0113x + 0.2291R2 = 0.86

-1

-0.8

-0.6

-0.4

-0.2

030 40 50 60 70 80 90

Reaction Time (min)

Ln([C

]/[C

]o)


57

1st Order Kinetics Degradation Plot of PCB-3 (MeOH) w/MgPd

y = -0.013x + 0.2322R2 = 0.9879

-1.25

-1.05

-0.85

-0.65

-0.45

-0.25

-0.05 25 35 45 55 65 75 85 95 105 115 125

Reaction Time

Ln([C

]/[C

]o)


A similar set of experiments were conducted using a solvent system comprised of 90%

water and 10% methanol (used to increase solubility). In these studies, 0.05 g of 1% Mg/Pd was

used instead of 0.25 g, due to greater reactivity of the Mg/Pd in water. Also, the samples were

exposed to 5 minutes of ultrasound to assist in extraction of any adsorbed PCBs from the surface

of the catalytic metal prior to syringe filtering. Starting concentrations were also lower (5-μg/mL

to 10-μg/mL) due to solubility issues of PCBs in water. Pseudo 1st order kinetic plots are shown

in Figures 19-22. Pseudo 1st order rate constants obtained were k = 0.00226 L min-1 g-1, k =

0.00486 L min-1 g-1, and k = 0.00716 L min-1 g-1 for PCB-001, PCB-002, and PCB-003,

respectively.

58

Pseudo 1st Order Degradation Plot of PCB-1 in 90:10 Water:MeOH using MgPd

y = -0.0113x + 0.0335R2 = 0.9455

-0.3

-0.25

-0.2

-0.15

-0.1

-0.05

00 5 10 15 20 25 30

Reaction Time (min)

Ln([

C]/[

C]o

)

Figure 19: Pseudo 1st order kinetic plot of the degradation of PCB-001 with Mg/Pd in 9:1 water:methanol


y = -0.0243x + 0.0602R2 = 0.9255

-0.7

-0.6

-0.5

-0.4

-0.3

-0.2

-0.1

00 5 10 15 20 25 30

Reaction Time (min)

Ln([

C]/[

C]o

)


59


y = -0.0358x + 0.0836R2 = 0.9842

-1-0.9-0.8-0.7-0.6-0.5-0.4-0.3-0.2-0.1

00 5 10 15 20 25 30

Reaction Time (min)

Ln([

C]/[

C]o

)


Kim et al. performed a similar study on monochlorinated biphenyls in water using Pd/Zn

and Pd/Fe (Kim et al, 2004). They achieved results similar to what was observed in the

experiments conducted in the water:methanol solvent system, but there were some differences

when compared to the pure methanol system. Kim et al. observed that the rate of dechlorination

for monochlorinated congeners in water was PCB-003 > PCB-002 > PCB-001, which is in

agreement with the experimental results presented above. The studies conducted in pure

methanol solvent show a completely different preference for dechlorination, PCB-001 > PCB-

003 > PCB-002. The differences presented above could be due to a multitude of factors. In the

methanol study, the reversal in the dechlorination trend may indicate that bimetallic degradation

is dependent upon the proton donor that is used. In addition, the previous study used both Zn/Pd

and Fe/Pd, rather than Mg/Pd, which may indicate different bimetallics can undergo different

mechanistic pathways. Finally, the bimetallics were prepared differently which may also affect

60

the mechanism of degradation. Kim et al. (2004) used electrodeposition to prepare the Zn/Pd

and Fe/Pd, whereas the Mg/Pd was prepared by mechanically alloying.

One interesting data point that appeared in all three studies performed using methanol as

the solvent system was that there appeared to an initial “lag time” before significant degradation

would begin to occur. In all three of the monochlorinated studies performed in methanol that are

presented in this paper, that lag time was approximately 30 minutes. Higher chlorinated

congeners have exhibited similar “lag” times as well. One possible explanation for this is that it

requires a certain amount of time for enough molecular/atomic hydrogen to be generated before

degradation can occur. To test this theory, a study was initiated in which pure methanol and

~0.25 g of 1% Mg/Pd were allowed to react for ~30 minutes, at which point the samples were

spiked with PCB-001 to a concentration consistent with prior studies. The samples were then

extracted and analyzed as before. The data are included in Figure 22.

61

Lag Study: Pseudo 1st Order Degradation Plot of PCB-1 in MeOH w/ MgPd

y = -0.027x + 0.9766R2 = 0.8622

-2.5

-2

-1.5

-1

-0.5

00 20 40 60 80 100 120

Reaction Time (min)

Ln([C

]/[C]

o)

Figure 22: Pseudo 1st order kinetics plot of the degradation of PCB-001 with Mg/Pd from "Lag" study

This study produced a rate constant almost identical to previous study without the

additional “lag” time. This indicates that the reason for the difference in kinetics is due to

generation/adsorption of hydrogen rather than the adsorption of the contaminant itself, other wise

the same ~30 minute “lag” time would have been observed in this study. Interestingly enough,

this trend is not seen in studies with water as the solvent, most likely due to water’s greater

ability to donate a proton, thus creating molecular hydrogen more quickly.

Another important question to answer is the final fate of the contaminant, whether or not

biphenyl is the end product or is it degraded further. Studies were conducted in both methanol

and 9:1 water:methanol solvent systems where more active palladized magnesium was used

(10% Mg/Pd vs. 1% Mg/Pd) in order to answer this question. The results are shown below in

Figure 23 and 24. There is no significant degradation of biphenyl apparent with pure methanol

62

as the solvent in more than 30 days. However, the studies with 9:1 water:methanol show

significant degradation within the first six hours, and near complete degradation within 3 days.

This is in agreement with other published studies where the primary solvent is water (Kim et al.,

2004).

Degradation of Biphenyl w/10% MgPd in MeOH

0

0.2

0.4

0.6

0.8

1

0 50 100 150 200 250

Reaction Time (hrs)

[C]/[

C]o

Figure 23: Degradation of biphenyl in methanol w/ 10% Mg/Pd

63

Degradation of Biphenyl w/10% MgPd in 90:10 Water:Methanol

0

0.2

0.4

0.6

0.8

1

1.2

0 20 40 60 80 100 120 140 160

Reaction Time (Hours)

[C]/[

C]o

Figure 24: Degradation of biphenyl in 9:1 water:methanol w/ 10% Mg/Pd

The source of hydrogen in this reaction was investigated to determine if it was coming

from the alcohol or methyl moiety of the solvent. Gas-phase studies show that the methyl C-H

bond requires less energy to break then the O-H bond in methanol (Blanksby and Ellison, 2003).

However, whether or not this is true in this reaction isn’t clear, due to the role that solvent effects

can play in bond strengths. In order to determine the source of the hydrogen, a reaction study

involving isotopically-labeled solvent was performed using PCB-1. Methanol-d (CH3OD) was

used as the solvent, and the reaction products were determined by GC-MS. If the proton was

coming from the alcohol group, than the m/z ratio of the biphenyl product will be 1 amu higher

than when the same study is run using pure methanol as the solvent. The mass spectra of both of

these studies are given in Figure 25 and Figure 26. As can be seen in the corresponding mass

spectra, when pure methanol is used as the reacting solvent, the parent ion of the biphenyl

product is 154 m/z. When isotopically labeled methanol-d is used, however, the parent ion of

64

biphenyl is 155 m/z. This conclusively demonstrates that the proton being donated in the

reaction is coming from the alcohol group of the methanol, rather than from the methyl group (as

gas-phase data suggests). Additionally, this information rules out the possibility of the

dechlorination going through a benzyne intermediate, which was being considered as a possible

mechanistic pathway for Mg/Pd dechlorination. However, once the intermediate C-C triple bond

is formed, two protons (or deutons) would have to add across the triple bond. The above data

show only one additional deuton was seen in the degradation of PCB-1 to biphenyl, eliminating

the benzyne mechanism as a possible mechanistic pathway.

PCB_1 4day MeOD A 1_5_071003012354 10/3/2007 1:23:54 AM

RT: 0.00 - 15.64

0 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15Time (min)

0

20

40

60

80

100

Rel

ativ

e A

bund

ance

10.54

10.95 15.5614.8814.0711.97 12.9810.269.619.018.207.62

NL:1.85E7TIC F: MS PCB_1 4day MeOD A 1_5_071003012354

PCB_1 4day MeOD A 1_5_071003012354 #898 RT: 10.53 AV: 1 NL: 8.97E6T: + c SIM ms [ 147.00-157.00, 183.00-193.00]

150 155 160 165 170 175 180 185 190m/z

0

20

40

60

80

100

Rel

ativ

e A

bund

ance

155.00

154.00

152.99

156.02151.98

149.97 156.82 190.89186.35183.15

Figure 25: Gas chromatogram and mass spectrum of PCB-001 degraded with Mg/Pd in MeOH

65

L:\Removable Disk (G)\...\PCB_1 335 MeOH 10/3/2007 4:36:08 AM

RT: 0.00 - 15.63

0 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15Time (min)

0

20

40

60

80

100

Rel

ativ

e A

bund

ance

11.31

10.54

15.3414.8413.9913.4911.5810.429.127.12 8.62

NL:2.94E6TIC F: MS PCB_1 335 MeOH

PCB_1 335 MeOH #905 RT: 10.56 AV: 1 NL: 7.53E5T: + c SIM ms [ 147.00-157.00, 183.00-193.00]

150 155 160 165 170 175 180 185 190m/z

0

20

40

60

80

100

Rel

ativ

e A

bund

ance

153.99

152.98

151.98

155.01150.97156.00 190.96189.03185.04183.30

Figure 26: Gas chromatogram and mass spectrum of PCB-001 degraded with Mg/Pd in MeOD

Conclusion

The research discussed above is highly indicative that the mechanism for degradation of

polychlorinated biphenyls varies depending on the solvent system that is used. Previous research

published indicates that the para substituted congener has the highest rate constant, followed by

the meta and then the ortho monochlorinated congeners in when performed in water solvent

systems. Preliminary studies confirm this experimental observation with palladized magnesium

when the experiments are performed in water:methanol (9:1), but when pure methanol is used as

the solvent, the results are different. The order of increasing rate constants was found to be

ortho>para>meta with pure methanol solvent systems. Additionally, studies performed in

methanol do not begin degradation immediately (as is seen in the water:methanol studies). There

66

is an initial period of no degradation, which seems to be due to the need for hydrogen to be

created/adsorbed to the surface of the metal.

Mechanistically, it appears there may be solvent specificity as well. Not only are the

relative rates of dechlorination different when comparing ortho, meta, and para, but the final

products also are different. In methanol systems, it doesn’t appear that biphenyl is capable of

being degraded any further. Conversely, in water:methanol systems, biphenyl is degraded in the

presence of Mg/Pd. At this time, the final products of that breakdown are not known. In both

solvent systems, the proton being donated is coming from hydrogen bound to oxygen (rather

than coming from CH3 in methanol). The exact mechanism of degradation using Mg/Pd is still

not known at this time, but a benzyne intermediate has been eliminated as a possibility.

Acknowledgements

This research has been supported by a grant from the U.S. Environmental Protection

Agency's Science to Achieve Results (STAR) program. Although the research described in the

article has been funded in part by the U.S. Environmental Protection Agency's STAR program

through grant X832302, it has not been subjected to any EPA review and therefore does not

necessarily reflect the views of the Agency, and no official endorsement should be inferred.

67

CHAPTER FOUR: MECHANISM OF THE DEGRADATION OF INDIVIDUAL PCB CONGENERS USING MECHANICALLY ALLOYED

Mg/Pd IN METHANOL

Reproduced with permission from Environmental Science and Technology, submitted for publication. Unpublished work copyright 2008 American Chemical Society.

Introduction

Polychlorinated biphenyls (PCBs) are a family of 209 chemical compounds for which

there are no known natural sources. They have a heavy oil like consistency (single congeners can

exist as solids), high boiling points, a high degree of chemical stability, low flammability, low

electrical conductivity, and a specific gravity between 1.20 and 1.44. Because of the above-

mentioned characteristics, were used in a variety of applications such as: heat transfer and

hydraulic fluids; dye carriers in carbonless copy paper; plasticizer in paints, adhesives, and

caulking compounds; and fillers in investment casting wax (EPA, 1983). PCBs can volatilize

from sources and are capable of resisting low temperature incineration. This makes atmospheric

transport the primary mode of global distribution (Eisenreich et al., 1983). PCBs are subject to

reductive dechlorination, even though they are generally considered recalcitrant in the

environment (Bedard and Quensen, 1995). The process of PCB reductive dechlorination replaces

chlorines on the biphenyl ring with hydrogen, reducing the average number of chlorines per

biphenyl in the resulting product mixture. This reduction is important because the less

chlorinated products are less toxic (Quensen et al., 1998; Mousa et al., 1996), have lower

bioaccumulation factors, and are more susceptible to aerobic metabolism, including ring opening

and mineralization (Mousa et al., 1996, Bedard et al., 1987).

68

Currently, the most common remediation technique is incineration, but this procedure is

not without its problems. Incineration requires a large amount of fuel and can lead to the

formation of highly toxic by-products, including polychlorinated dibenzo-p-dioxins and

polychlorinated dibenzo-furans (commonly referred to as dioxins) (Wu et al., 2005). Another

traditional remediation technique for PCB contamination is dredging of contaminated soils and

sediments followed by land filling of the resulting hazardous waste. Land filling is undesirable

because of the permanent and persistent nature of the PCBs. Microbial degradation is another

treatment option currently being investigated, but slow reaction rates and incomplete degradation

have hindered the use of this approach in the field. Two different approaches exist for microbial

degradation: aerobic and anaerobic. Aerobic processes proceed via oxidative destruction of the

PCBs, although dechlorination is limited to the lighter congeners which have five or less

chlorines present on the biphenyl ring. Anaerobic microbial degradation occurs via a reductive

dehalogenation pathway which can typically only remove chlorines from the meta or para

position (Wiegel and Wu, 2000).

A more promising technique that has been studied in recent years is the use of zero-valent

metals (including magnesium, zinc, and iron) for the in situ remediation of chlorinated

compounds including PCBs. Dechlorination of polychlorinated biphenyls (PCBs) by zero-valent

iron has been demonstrated at high temperatures (Chuang et al., 1995) but at 200°C or below,

little dechlorination of PCBs occurred. However, rates of dechlorination by iron have been

increased by using palladium, a known hydrodechlorination catalyst (Wang and Zhang, 1997;

Cheng et al., 1997; Li and Farrell, 2000; Kovenklioglu et al., 1992; Lowry and Reinhard, 1999),

as a coating on the zero-valent iron surface yielding biphenyl (a non-chlorinated, innocuous

69

product). Muftikian and co-workers demonstrated rapid degradation of PCE with Fe/Pd

(Muftikian et al., 1995) and Grittini showed that the Fe/Pd bimetallic system can degrade PCBs

but did not quantify the degradation (1995). While the Fe/Pd has shown high levels of

degradation in laboratory studies, the bimetal must be prepared under inert atmosphere after

rigorous acid-wash of the iron metal (Doyle et al., 1998). A report by Fernando and co-workers

proposed that the enhanced reactivity of Pd/Fe might be due to the adsorption of hydrogen (H2),

generated by iron corrosion, on palladium (Gritinni et al., 1995).

The disappearance of chlorinated organic compounds from aqueous solutions contacting

ZVMs may be due to dechlorination reactions or sorption to ZVM-related surfaces. This

investigation utilizes mechanically alloyed magnesium and palladium (1% on graphite) as the

bimetallic system which has proven effective in previous research in the degradation of PCBs

(Aitken et al., 2006). Magnesium has several advantages over other previously used zero valent

metals. Due to the self-limiting oxide layer that forms on the surface of magnesium, it is capable

of dechlorination even after exposure to oxygen (unlike iron, which will completely oxidize upon

exposure to oxygen and become deactivated). Another advantage in using magnesium is that it

has a greater thermodynamic driving force when compared to both iron and zinc, as shown

below:

Mg2+ + 2e- Mg0 E0 = -2.37 V

Fe2+ + 2e- Fe0 E0 = -0.44 V

Zn2+ + 2e- Zn0 E0 = -0.76 V

Equation 15: Standard reduction potentials for Mg, Fe, and Zn

An understanding of the underlying mechanism for the dechlorination is a necessary next

step in order to fine-tune the bimetallic system for maximum possible effectiveness. Several

70

studies are discussed within this paper to help elucidate and to propose a mechanism by which

PCBs are degraded by palladized magnesium (Mg/Pd). Possible mechanistic pathways under

investigation included degradation through a benznyne intermediate, nucleophilic aromatic

substitution, and the use of hydrogen (in the form of a radical or hydride) in removing the

chlorine atom. Experimental evidence suggests three different possible mechanistic pathways,

all of which include the removal of the chlorine atom by a hydrogen atom as the rate-limiting

step.

Methods

Materials and Chemicals. Neat PCB standards were obtained from Accustandard (New

Haven, CT) and Optima© grade methanol/toluene were obtained from Fisher Scientific

(Pittsburgh, PA). 99.0% Methanol-d was obtained from Acros Organics (Morris Plain, NJ).

Magnesium (~ 4-μm) was obtained from Hart Metals, Inc (Tamaqua, PA). 1% palladium on

graphite was obtained from Engelhard (Iselin, NJ), while 10% palladium on graphite was

obtained from Acros Organics. All metals and catalysts listed above were used as received.

A ~0.08 wt% palladium-magnesium mixture was prepared by ball-milling 78 g Mg with

7 g of 1% palladium on graphite in a stainless steel canister (inner dimensions 5.5 cm by 17 cm)

with 16 steel ball bearings (1.5 cm diameter, at a total mass of 261.15 g). The material was

milled for 30 minutes using a Red Devil 5400 series paint mixer. A ~0.8 wt% palladium-

magnesium mixture was prepared in a similar fashion using 10% palladium on graphite.

Some experiments were conducted on bimetal material that had been milled up to three

years earlier. In order to “reactivate” the bimetallic compound, the already prepared material (85

g) was milled for an additional 30 minutes on the Red Devil 5400 series paint shaker.

71

Individual PCB solutions were prepared by diluting the neat standards with Optima©

grade methanol (or 99.0% methanol-d) to the desired concentration. Further dilution was done

on several mono-chlorinated congener solutions with deionized water to bring the final solvent

ratio to 9:1 water: methanol.

Experimental rate constants were normalized using a ρ constant (g of Mg/Pd/L of

solution) to allow comparison between various studies.

Experimental Procedure: Vial studies using 0.25 g of Mg/Pd and 10 mL of individual

PCB solution in 20-mL vials (with PTFE lined caps) were conducted. Samples were placed on

Cole-Parmer Series 57013 Reciprocating Shaker table (speed 7) until the appropriate extraction

time. Samples were extracted using 10 mL of toluene (Optima©) which was added to the sample

vial. The resulting mixture was then shaken by hand for two minutes. Next, 4 mL of this

miscible solution was filtered with a Puradisc® 25-mm (0.45-μm pore size) nylon syringe filter

attached to a glass syringe. This mixture was then placed into a centrifuge tube with a PTFE

lined cap (to prevent evaporation). The mixture was then shaken by hand for two minutes

followed with centrifugation for two minutes. The top layer of the extract was collected for

further analysis. Studies in 9:1 water:methanol were conducted similarly, however the starting

weight of Mg/Pd was 50 mg and after 2 minutes of shaking, a five minute ultrasound treatment

was performed, to ensure complete extraction of PCBs from the surface of the bimetal. The

extraction efficiency for this separation technique was ~97% with a relative standard deviation of

less than 4%.

Analysis of the extracted samples were performed on a Perkin Elmer Autosystem XL gas

chromatograph equipped with an electron capture detector (GC-ECD) and a Thermo Finnigan

72

Trace GC/DSQ, both using a RTX-5 column (30-m, 0.25-mm i.d., 0.25-μm df). UHP nitrogen

was used as the ECD makeup gas at a flow of 30 mL/min. Helium acted as the carrier gas in

both instruments, and was held at a constant flow of 1.3 mL/min. On the GC-ECD, the injector

port temperature was held at 275ºC and the detector was at 325ºC. On the GC/DSQ, the injector

temperature was 220ºC, and the ion source temperature was 250ºC. Both instruments were

equipped with autosamplers. An initial oven temperature of 100ºC was used, and then ramped

up to 270ºC. Identification of each of the single congener PCBs was based upon the retention

times of known standards.


Zero-valent magnesium coated with 1% palladium on graphite was shown to be capable

of degrading 80% of 10 mL of a 20-μg/mL PCB-151 in methanol solution within 24 hours

(Figure 27).

73

Zero Order Kinetics Plot of PCB-151 Using MgPd in MeOH

0

0.2

0.4

0.6

0.8

1

1.2

0 500 1000 1500 2000 2500

Reaction Time (min)

[C]/[

C]o

Figure 27: Degradation kinetics of PCB-151 in Mg/Pd in methanol

In the preliminary 2,2’,3,5,5’,6-PCB (PCB-151) study, after the initial rapid

dechlorination subsequent degradation of this hexa-chlorinated congener was halted. A possible

explanation is the competition of the daughter compounds for the (relatively) few active

palladium catalytic sites on the magnesium. It is possible that the contaminant is bound to the

surface of the bimetal during the dehalogenation process. If this is the case, the newly created

byproduct may not leave the surface once the first chlorine atom is removed. This would not

allow non-degraded PCB-151 to interact with the active site. If there is not an excess of active

sites in the reaction system (as is the case here), then competition will favor the dechlorination of

the newly formed lower chlorinated byproduct, halting the degradation of the parent congener

(Figure 28). The following experiment was performed to test this theory. A PCB-151 solution

was spiked with biphenyl, so that the final concentrations of both analytes were equal (5-μg/mL).

This solution was exposed to the Mg/Pd as was done in the previous experiments. When

74

compared to a control study consisting of only PCB-151 exposed to Mg/Pd, the degradation of

the parent congener is severely inhibited. This evidence supports the concept that competition is

occurring at the active sites on the Mg/Pd surface.

Comparison Plot of the Degradation of PCB-151 and the Appearance of the PCB-93/95

0

2

4

6

8

10

12

14

0 500 1000 1500 2000 2500

Reaction Time (min)

Con

c (μ

g/m

L)

PCB-151

PCB93/95

Figure 28: Comparison of PCB-151 and PCB-93/95

The disappearance of PCB-151 coincides with the production of PCB-93/PCB-95. The

degradation of the parent congener slows dramatically indicating that the active sites are

occupied with the newly formed lower chlorinated byproducts. Subsequently, the degradation of

the newly formed byproducts slows down and halts as lower chlorinated byproducts are created

and occupy the active sites. The degradation of PCB-151 indicates a stepwise dechlorination due

to the sequential production of degradation byproducts with the initial reaction pathway PCB-

151 PCB-93/PCB-95 (major products) + PCB-92 (minor products) PCB-45 (Figure 29).

75

Figure 29: GC-MS analysis of the degradation of PCB-151 using Mg/Pd in methanol

Due to the change in the kinetics of the parent congener once there is competition for the active

sites, pseudo rate constants and reaction order was determined for the first six hours of the study.

These data were fit to a pseudo 1st order plot, and is shown below in Figure 30.

Pseudo 1st Order Degradation Plot for PCB-151 (in MeOH) w/ MgPd

y = -0.0043x + 0.1816R2 = 0.9901

-1.4

-1.2

-1

-0.8

-0.6

-0.4

-0.2

0

0.2

0 50 100 150 200 250 300 350

Reaction Time (min)

Ln(C

/Co)

Figure 30: Pseudo 1st order kinetics plot of the degradation of PCB-151 with Mg/Pd in methanol

83hr 24hr 8hr 3hr 0.5hr 0hr

76

From this plot, a pseudo 1st order rate constant of k = 1.72E-4 L min-1 g-1 (normalized by volume

of solution and mass of Mg/Pd used) was obtained for PCB-151 degraded by Mg/Pd in methanol.

Subsequent testing of the original mechanically alloyed material had shown that the metal

had become deactivated after six months. In an attempt to reactivate the metal, the inactive

bimetal was subjected an additional 30 minutes of mechanical alloying (under identical

conditions to the initial milling). The re-milled bimetal was tested on PCB-151 to determine the

effectiveness of reactivation process. This study was done using identical conditions to the

original PCB-151 degradation study, and is shown in Figure 31.

Zero Order Degradation Plot of PCB-151 Using Reballmilled MgPd in MeOH

0

0.2

0.4

0.6

0.8

1

1.2

0 50 100 150 200 250 300 350 400 450 500

Reaction Time (min)

[C]/[

C] o

Figure 31: Degradation Plot of PCB-151 using reactivated Mg/Pd in methanol

A similar degradation profile is observed for the reactivated metal as seen with the original

Mg/Pd bimetal, although the degradation of the parent congener appears to happen at a faster

rate. A pseudo-first order kinetics plot is shown in Figure 32.

77

Pseudo 1st Order Kinetics Plot of PCB-151 Using Reballmilled MgPd in MeOH

y = -0.0136x - 0.1729R2 = 0.9611

-1.8

-1.6

-1.4

-1.2

-1

-0.8

-0.6

-0.4

-0.2

00 20 40 60 80 100 120

Reaction Time (min)

Ln([C

]/[C]

o)

Figure 32: Pseudo 1st order kinetics plot of the degradation of PCB-151 using reactivated Mg/Pd in methanol

A normalized pseudo 1st order rate constant of 5.44E-4 L min-1 g-1 was obtained, which is more

than 3x faster than the original bimetal. Not only is the reactivated metal more reactive, it

appears to be less selective as well. Additional lower chlorinated byproducts were detected,

including all possible penta-chlorinated congeners.

As a mechanistic probe, kinetic isotope effects were explored using methanol-d (CH3OD)

with PCB-151. Samples were prepared in an inert atmosphere glove box to prevent hydrogen

contamination in the solvent. A plot showing the pseudo 1st order kinetics is shown in Figure 33.

78

Pseudo 1st Order Degradation Plot of PCB-151 Using Reballmilled MgPd in MeOH

y = -0.0118x + 0.0578R2 = 0.9961

-2

-1.5

-1

-0.5

0

0.5

0 20 40 60 80 100 120 140 160

Reaction Time (min)

Ln(C

/Co)

Figure 33: Pseudo 1st order kinetics plot of the degradation of PCB-151 using reactivated Mg/Pd in methanol

Applying mass normalization of the bimetal to the observed rate constants determined from

Figure 32 and Figure 33 gives a kH/kD = 2.31, indicative of a primary kinetic isotope effect.

Several possible reaction mechanisms have been eliminated. It has been experimentally

determined that PCBs are not passing through a benzyne intermediate during degradation.

Deuterium labeled studies have shown only a single deuterium being added to the biphenyl

structure upon the removal of a chlorine atom, rather than the two deuterium’s expected if the

degradation pathway did include a benzyne intermediate. Nucleophilic aromatic substitution is

also unlikely, due to the fact that the rate constant decreases as the as the degree of chlorination

increases. Nucleophilic aromatic substitution requires electron-withdrawing substituents to

activate the aromatic system, which means a higher degree of chlorination should increase the

rate constant of the reaction (depending on the pattern of the substituents, ortho/para and meta).

79

Having a kH/kD greater than 2 indicates that the hydrogen bond is being either broken or

formed in the rate-limiting step of the reaction. It is unlikely that the rate-limiting step involves a

hydrogen bond being broken, since it has already been removed from the methanol to form

molecular hydrogen. It is a more likely possibility that the formation of a new carbon-hydrogen

bond is the rate-limiting step, occurring after the contaminant has been adsorbed on the

bimetallic system. There are several mechanistic pathways which can account for the observed

kinetic data presented in the paper, all of which contain aryl radical intermediates.

One possible mechanism (Mechanism A) is the formation of adsorbed atomic hydrogen

(H·) at the surface of the palladium (formed from the dissociation of H2 on the Pd surface),

although this seems unlikely to occur due to the less favorable thermodynamics involved in

atomic hydrogen desorption from the surface of the palladium (compared to other

hydrogen/palladium species) (Cybulski and Moulijn, 2006). The chlorine atom would be

abstracted in a homolytic bond cleavage, forming HCl and an aryl radical. The intermediate aryl

radical would then react with a second H· to terminate the radical process.

A thermodynamically more reasonable approach would be that the hydrogen is absorbed

within the first few sublayers of the palladium (Mechanism B). Electron density would be added

to the hydrogen via the zero-valent magnesium, increasing the nucleophilicity of this species.

Previous research has suggested that “hydride like” species would then be capable of degrading

halogenated species (Cwiertny et al., 2007). The role of the magnesium is also clearer in this

case. Not only does it produce the molecular hydrogen necessary for the reaction, it is also used

as a means to help create and store this absorbed hydride-like species. This more activated

80

nucleophilic species could then react with the contaminant and would then proceed in by the

same mechanism as the atomic hydrogen.

The first two possible mechanisms discussed are similar to an SRN1 type of reaction, in

which a nucleophilic substitution occurs involving a radical intermediate. The proposed

mechanisms differ from the SRN1 type in that the initiation step proceeds via the generation of

the aryl radical by the abstraction of the chlorine by a hydrogen radical or hydride-like radical in

a homolytic bond cleavage, in which HCl(aq) is formed. Normally, SRN1 mechanisms involving

halides proceed by the unimolecular decomposition of radical anion (which is obtained from the

substrate) formed by an attacking nucleophile (Carey and Sundberg, 2004), which is seen in the

third proposed pathway. In this case, the attacking nucleophile abstracts the leaving group as the

aryl radical is formed, rather than imparting a negative charge to the aryl system (which would

then be followed by the expulsion of the anion halide, leaving an aryl radical). A second

nucleophile (atomic hydrogen or hydride-like radical) is then able to react with the aryl radical in

a termination step. Subsequent dechlorination can continue as long as there are available

chlorine atoms on the biphenyl ring to initiate the process. Dechlorination will stop once all

chlorine atoms have been removed, leaving biphenyl as the final product for the degradation

mechanism. Of these two, Mechanism B is more energetically favorable. Hydrogen is more

easily able to bind to the subsurface sites than surface sites, the enthalpies of desorption are 32

kJ/mol and 80 kJ/mol, respectively (Cybulski and Moulijn, 2006).

A second pathway is the formation of hydride (H-) moieties within bulk palladium, which

could then react with the PCB substrate in a SRN1 reaction. The hydride acts as a nucleophile

which can transfer an electron to the contaminant substrate. This causes the expulsion of a

81

chlorine atom, leaving an aryl radical which can quickly react with another H-. The charged

biphenyl species can transfer an electron to another PCB substrate, so that process can continue

to propagate (Carey and Sundberg, 2004).

A balanced chemical equation is given in Figure 34, which holds true for the above

possible mechanistic pathways.

2CH3OH + Mg0 Mg(OCH3)2 +Pd/C

HCl

HCl

+ +

Figure 34: Balanced mechanism for the declorination of PCBs by Mg/Pd in methanol

There are several reasons mechanisms of the proposed type are possible. Palladium

catalysts are well known to produce both atomic hydrogen and hydride species when molecular

hydrogen is present, as is the case here, so there would be no shortage of the initiating

nucleophile. Also, this type of mechanism would explain the lack of dimerization and additional

chlorinated byproducts. Radical chlorine is never produced as a reactive species in this reaction

scheme; it is bound up with the atomic hydrogen in a covalent bond or as an anion species

capable of abstracting a proton from the solvent as soon as it is removed from the biphenyl

structure. This helps to explain the lack of chlorinated adducts. Additionally, the aryl radical

would be unlikely to come into contact with a second aryl radical, which explains the lack of

dimerization. It is unknown at this time if the reactive intermediate remains at the surface of the

bimetal during the reaction, but is seems likely when looking at prior experiments involving

chlorinated aromatics and Pd/C systems (Cheng et al., 1997). These schemes require adsorption

of the PCB onto the surface of the bimetallic compound, then reaction at the interface of the

palladium and graphite, limiting the mobility of the aryl radical. This limited mobility and the

82

overabundance of atomic hydrogen, hydride-like radicals, and hydrides on the surface of the

catalyst almost certainly allows for the reaction of the aryl radical and second nucleophilic

hydrogen, rather than two separate aryl radicals coming into contact. A visual representation

these mechanistic schemes are shown in Figure 35.

83

H

Mg0

Graphite

Mg0 Mg0

Pd

Surface

Adsorption

Cl

2CH3OH

2e-

+Mg(OCH3)2

2H2H

HHH

Surface

Desorpt

ion

Cl

HCl+

**

e-e-

Scheme A/B

H

Mg0

Graphite

Mg0 Mg0

Pd

Surface

Adsorption

Cl

HHH-

H-

Surf

ace

Des

orpt

ion

Cl

HCl+

e-e-

Scheme C

-HCl

-Cl-

H

+

Cl

Figure 35: Proposed mechanism for the dechlorination of PCBs by Mg/Pd in methanol by (a) atomic hydrogen, (b) "hydride-like" radicals, and (c) hydride (H* denotes both hydrogen and "hydride-like" species)

This work has confirmed mechanically alloyed Mg/Pd to be an effective remediation

technique for the degradation of PCBs. PCB-151 undergoes a stepwise dechlorination

84

substantiated by observance of sequential production of degradation byproducts, and exhibits

pseudo 1st order kinetics. Three possible types of hydrogen species have been proposed as the

intermediate reactant responsible for the abstraction of chlorine in the reductive dechlorination

process.

Acknowledgements

This research has been supported by a grant from the U.S. Environmental Protection

Agency's Science to Achieve Results (STAR) program. Although the research described in the

article has been funded in part by the U.S. Environmental Protection Agency's STAR program

through grant X832302, it has not been subjected to any EPA review and therefore does not

necessarily reflect the views of the Agency, and no official endorsement should be inferred.

85

CHAPTER FIVE: CONCLUSION

Mechanically alloyed Mg/Pd bimetallic particles have successfully been created that are

capable of completely degrading PCBs to biphenyl and even further, depending upon what

solvent system is used. The implications for the use of this in environmental remediation have

not been fully realized at this point, as technology capable of utilizing them to their fullest

potential is still in development. However, a first step in the understanding of the mechanism of

dechlorination has been completed.

Existing technology for the preparation of such particles was deemed both inefficient and

costly, so a new preparation technique was devised and implemented. The preparation process

involved mechanically alloying zero-valent magnesium and palladium on graphite to create the

reactive particles, and it proved to be more effective in terms of both cost and reactivity than the

use of the current industry preparation technique of electrodeposition. It is also more easily

scaled up for the production of large amounts of the bimetallic particles, making it particularly

suitable for use in field applications. The bimetal particles were first milled on a small-scale

using a high energy vibrator milling device to determine dechlorination effectiveness. Once

proven effective, larger scale production took place using a commercially available paint shaker.

The milling process was optimized for a variety of parameters, including %Pd loading, milling

time, canister loading, and ball-to-mass ratio, all of which contribute in vibrator milling in

determining final particle size and reactivity. The final parameters for the mechanical alloying

of Mg/Pd particles correspond to a batch size of 85 g (78 g of zero-valent magnesium and 7 g of

1% Pd on graphite) being milled for 23 minutes and using 16 stainless steel ball bearings. The

preparation and milling process can be done within one hour, and using custom made templates

86

for the paint shaker, up to six canisters can be milled at one time. This allows for an effective

batch size of more than half a kilogram within 2 hours (including cleaning and prep time).

The effectiveness of mechanically alloyed Mg/Pd has been proven experimentally,

however the mechanism behind the degradation process is not completely understood. A more

complete understanding of the reaction mechanism will be useful in tailoring the bimetal for

specific in-situ field applications, such as for the remediation of PCBs in sediments or in material

coatings (paints, caulking, etc…). The majority of research with bimetallic particles such as

Mg/Pd has been accomplished using aqueous solvents, but the high reactivity of this catalytic

system suggests a less labile proton donor may be more appropriate. Alcoholic solvent systems,

such as methanol and ethanol, are much safer to use when handling Mg/Pd. Additionally, PCBs

are much more soluble within these types of solvents, making the research into the mechanism

more easily accomplished. As a first step towards understanding the mechanism, relative rates of

degradation for the ortho, meta, and para mono-chlorinated congeners were determined in both

methanol and 9:1 water:methanol solvent systems. Previous work on similar bimetals (Fe/Pd

and Zn/Pd) had been conducted by Kim et al. using an aqueous solvent system. The bimetal

used in the study by Kim et al. were prepared using electroless deposition, but were useful as a

reference point for the studies using milled Mg/Pd. Significant differences were detected in the

results between the two solvents, suggesting that the mechanism may be dependent upon which

proton donor is used. Previous published work has stated that the order of degradation when

using water is para > meta > ortho, and these experiments confirm this. However, when

methanol is used, the relative rates of degradation change to ortho > para > meta. Additionally,

experiments in both solvent systems have shown that the final products of the degradation are

87

different. Literature has reported that zero-valent metals are capable of breaking down biphenyl,

and in the case of the 9:1 water:methanol system, this is shown to be true, with the final products

unknown at this time. This does not appear to be the true when methanol is used. Studies have

shown Mg/Pd is incapable of degrading biphenyl, even in experiments where a more active form

of the palladium catalyst is used. Experiments in the 9:1 water:methanol solvent system agree

with published literature, showing biphenyl is broken down in the presence of Mg/Pd. This is a

further indication that a solvent specific dechlorination mechanism exists for mechanically

alloyed Mg/Pd. Additional work on the mono-chlorinated congeners using deuterated methanol

has proven that the hydrogen involved in the reduction is coming from the alcohol moiety, rather

than the methyl group. It has also ruled out the possibility of a benzyne intermediate, which was

being considered as one of the possible reaction pathways.

The next step in the determination of the mechanism was to work with a more highly

chlorinated congener, such as PCB-151 (2, 2’, 3, 5, 5’, 6-chloro biphenyl). Studies using this

congener (and others) have proven that degradation occurs in a stepwise manner until biphenyl is

produced, at least when methanol is used as the proton donor. In an aqueous solvent,

degradation continues past biphenyl, to a currently unknown set of final products. Inhibition of

the degradation of the parent congener was seen once the initial reaction products reached a

steady-state concentration. This is attributed to the preferential degradation of the lighter

chlorinated byproducts. Lastly, isotopic studies using deuterated methanol were performed with

PCB-151. These studies produced a value for the ratio of kH/kD equal to 2.31. This value is

strongly indicative of a primary kinetic isotopic effect exhibited by hydrogen, suggesting that

hydrogen is involved in the rate-limiting step of the reaction in either a bond making or bond

88

breaking step. With this in mind, it seems likely that hydrogen is acting as a nucleophilic or

radical species produced by the palladium catalyst (from molecular hydrogen) which can then

react with the PCBs. A mechanism of this type would account for the kinetic isotope effect, and

also explain the lack of dimerization and additional chlorinated products, since chlorine radicals

would not be formed in this type of reaction.

From this research, three different mechanisms have been postulated as being responsible

for the dechlorination of PCBs by mechanically alloyed Mg/Pd in a methanol solvent system.

Each of these mechanisms involve radical species, as many published studies suspected it might,

however the major difference is in the nature of the hydrogen species which replaces the chlorine

atom in the PCB. These correspond to atomic hydrogen at the surface of the palladium catalyst,

hydride species residing within the bulk palladium, and an intermediate species of atomic

hydrogen that “hydride-like” in character, meaning it is more electronegative in comparison to

chemisorbed atomic hydrogen. This last “hydride-like” species is located within the first few

sublayers of the palladium lattice structure. At this time, which of the species is directly

responsible for the dechlorination is unknown; however use of other indirect chemical

information can help to narrow it down. Thermodynamically, it appears the “hydride-like”

species is most likely responsible for the dechlorination of PCBs. It has an enthalpy of

desorption much more realistic than that of the atomic hydrogen (32 kJ/mol vs 80 kJ/mol), and

unlike hydride located within the bulk palladium, is more accessible for interaction with

contaminant species adsorbed to the surface of the palladium on graphite. Further work will be

necessary to confirm that the “hydride-like” radical is actually the species involved in the

dechlorination.

89

APPENDIX A SUPPORTING INFORMATION FOR CHAPTER TWO

90

Calibration Data for PCB-151

Conc (μg/mL) Area Count 0.1 302283.72 0.5 1460732.48 1 2805441.81

Calibration Curve for PCB-151 y = 2830187.636508xR2 = 0.999023

0

500000

1000000

1500000

2000000

2500000

3000000

0 0.2 0.4 0.6 0.8 1 1.2

Concentration (μg/mL)

Are

a C

ount

Calibration Data for PCB-45

Conc (μg/mL) Area Count 0.1 75704.63 0.5 288116.91 1 541511.49

1.5 807920.55

91

Calibration Curve for PCB-45y = 521219x + 24369

R2 = 0.9999

0

100000

200000

300000

400000

500000

600000

700000

800000

900000

0 0.2 0.4 0.6 0.8 1 1.2 1.4 1.6


Are

a C

ount

Optimization Data for Large-Scale Mechanically Alloyed Mg/Pd Results for %Pd Loading

Sample ID

Reaction Time Peak 1 (mv)

Peak 2 (mv)

Peak1/Peak2 Ratio

0.01%

Loading 1a 0 93316.11 77377.36 1b 0 202519.43 178242.15 0.88 1c 4 191517.96 166539.05 0.87 1d 4 198779.04 171895.94 0.86 1f 26.5 193704.62 169745.3 0.88 1g 52.5 182359.31 159329.44 0.87 1h 52.5 203474.02 179202.41 0.88 1i 76.5 189301.52 164589.4 0.87 1j 76.5 173350.39 147111.65 0.85 1k 218.25 131831.28 100281.08 0.76 1l 336.5

1m 510.5 155167.47 126690.49 0.82

0.05%

Loading 5a 0 89946.96 72874.96 0.78 5b 0 186079.9 145529.05 0.78 5c 4 192442.6 149819.41 0.78 5d 4 217997.32 166219.03 0.76 5e 0.77

92

Sample ID

Reaction Time Peak 1 (mv)

Peak 2 (mv)

Peak1/Peak2 Ratio

0.05%

Loading 5f 26.5 193206.86 146682.01 0.76 5g 52.5 200450.53 153415.93 0.77 5h 52.5 193046.15 140893.21 0.73 5i 76.5 192657.8 139166.8 0.72 5k 218.25 166831.57 107059.21 0.64 5l 336.5

5m 510.5 175631.36 114651.07 0.65

0.08%

Loading 7a 0 129486.23 112536.28 0.87 7b 0 195820.99 156845.06 0.8 7c 4 195513.06 150400.21 0.77 7d 4 83574.12 62409.54 0.75 7f 26.5 183133.63 131666.16 0.72 7g 52.5 187009.49 129683.22 0.69 7h 52.5 198678.54 131700.82 0.66 7i 76.5 177814.22 118086.67 0.66 7j 76.5 177986.01 111948.94 0.63 7k 218.25 165567.71 99578.83 0.6 7l 336.5

7m 510.5 167191.5 86838.85 0.52

0.10%

Loading 9a 0 113493.81 93697.99 0.83 9b 0 89184.22 70272.57 0.79 9c 4 191519.8 145956.64 0.76 9d 4 197435.01 150617.36 0.76 9f 26.5 182773.07 128476.09 0.7 9g 52.5 190744.51 138602.23 0.73 9h 52.5 196659.6 141450.01 0.72 9i 76.5 184517.82 119937.62 0.65 9j 76.5 184517.82 119937.62 0.65 9k 218.25 169766.11 102410.91 0.6 9l 336.5

9m 510.5 149347.13 74889.91 0.5

0.16%

Loading 14a 0 113118.75 91650.25 0.81 14b 0 131502.54 102092.62 0.78 14c 4 172134.8 128388.07 0.75 14d 4 186120.4 141261.69 0.76 14f 26.5 183479.8 126413.51 0.69 14g 52.5 177467.09 120341.43 0.68 14h 52.5 181331.35 123996.43 0.68

93

0.16%

Loading 14i 76.5 183490.36 122608.74 0.67 14j 76.5 182128.65 117683.14 0.65 14k 218.25 165948.36 97510.9 0.59 14l 336.5

14m 510.5 166288.82 84426.71 0.51 Pd Loading Data g of Pd/C (per 85 g batch) 1 5 7 9 14

% Pd 0.011765 0.058824 0.082353 0.105882 0.164706 Rate Constant 0.000509 0.00062 0.000827 0.000805 0.000798

Activity vs. %Pd in Bimetallic for the Optimization of Pd Mass Loading

0

0.0002

0.0004

0.0006

0.0008

0.001

0 0.05 0.1 0.15 0.2

%Pd in Bimetallic

Act

ivity

(slo

pe fr

om

degr

datio

n of

125

4 pl

ots)

Results for Optimization of Canister Loading

Sample ID Rxn Time (min) Area Counts Ln(Area Counts) 16A 6 686278.13 13.43903826 16B 45 302796.73 12.620817 16C 51 242826.19 12.4001012 16D 90 188318.84 12.14589176 16E 122 160564.61 11.9864517 16F 152 135552.31 11.8171129 16G 180 116822.11 11.66840763 16H 232 95784.51 11.46985626

26.5A 6 26.5B 32 350528 12.76719587

94

Sample ID Rxn Time (min) Area Counts Ln(Area Counts) 26.5C 51 271099.41 12.51024086 26.5D 90 225857.41 12.32765915 26.5E 122 191897.23 12.16471525 26.5F 152 175649.84 12.07624775 26.5G 180 179089.23 12.09563945 26.5H 242 13037.86 9.475612712 31.8A 10 640263.75 13.36963548 31.8B 37 319759.21 12.67532352 31.8C 58 208352.22 12.24698529 31.8D 95 185829.72 12.13258605 31.8E 129 160481.7 11.9859352 31.8F 158 127788.22 11.75812964 31.8G 185 139651.57 11.84690581 31.8H 247 107155.74 11.58203857 42.5A 10 711944.08 13.47575465 42.5B 37 488095.02 13.09826538 42.5C 58 427265.24 12.96516027 42.5D 95 295786.13 12.59739194 42.5E 129 259598.35 12.46689091 42.5F 158 254394.75 12.44664247 42.5G 185 227684.3 12.3357153 42.5H 247 197112.82 12.19153153 58.3A 14 614085.67 13.32788973 58.3B 40 377927.55 12.84245779 58.3C 66 255503.85 12.45099276 58.3D 100 215600.39 12.28118193 58.3E 134 182284.77 12.11332541 58.3F 162 178854.02 12.09432522 58.3G 190 164829.51 12.01266695 58.3H 252 145287.19 11.88646768 85A 14 685488.03 13.43788632 85B 40 534989.93 13.1900032 85C 66 391837.04 12.87860132 85D 100 339279.09 12.73457832 85E 134 259903.29 12.46806488 85F 162 251885.86 12.43673133 85G 190 195288.94 12.18223548 85H 252 129246.96 11.76948027

127.5A 18 604425.18 13.31203317 127.5B 46 534989.93 13.1900032 127.5C 127.5D 106 302142.56 12.61865424 127.5E 139 256138.01 12.45347168 127.5F 167 213252.12 12.27023041 127.5G 195 197773.65 12.19487847 127.5H 257 167869.71 12.03094342

95

Total Canister Loading per Batch Mass (g) 15.9 26.5 31.8 42.4 58.3 85 127.2

Rate Constants 0.005799 0.004457 0.004271 0.004585 0.004033 0.006031 0.005406

Pseudo 1st Order Rate Constant From the Degradation of PCB-151 vs. Mass of Mg/Pd Milled Plot for the Optimization

of Milling Parameters

0

0.001

0.002

0.003

0.004

0.005

0.006

0.007

0.008

0 20 40 60 80 100 120 140

Mass of Mg/Pd Milled (g)

Pse

udo

Firs

t Ord

er R

ate

Con

stan

t (1/

min

)

96

Results for the Optimization of Milling Times Sample ID Rxn Time (min) Area Counts Ln(Area Counts)

3a 5 725,568.72 13.49471107 3b 30 556,218.93 13.22891725 3c 61 409,882.38 12.92362552 3d 91 385,619.61 12.8626067 3e 121 372,166.66 12.82709704 3f 151 380,657.22 12.84965456 3g 181 314,155.86 12.65764451 3h 241 318,874.87 12.67255405 8a 5 804,296.10 13.59772276 8b 30 530,184.53 13.18098039 8c 61 421,294.57 12.95108756 8d 91 331,246.06 12.71061676 8e 121 330,254.80 12.70761976 8f 151 269,768.52 12.50531954 8g 181 257,762.96 12.45979568 8h 241 258,033.83 12.46084598 15a 45 5048695.2 15.43464039 15b 49 4603522 15.34233222 15c 74 3805568.6 15.15197597 15d 74 3755080.6 15.13862031 15e 138 3430473.15 15.04820875 15f 138 2819697.75 14.85214026 15g 202 2565230.55 14.75755892 15h 202 2408119.35 14.69435665 15i 293 1862142.1 14.43723805 15j 293 1911206.2 14.46324512 15k 391 1518653.7 14.23333478 15l 391 1420612.7 14.16659882

15m 1436 852217.7 13.65559729 15n 1436 918255 13.73023041 23a 11 718,460.91 13.48486658 23b 36 447,763.19 13.01201978 23c 66 315,406.49 12.66161753 23d 96 268,425.96 12.5003304 23e 126 226,056.62 12.32854078 23f 156 208,056.52 12.24556505 23g 186 170,417.58 12.04600706 23h 246 141,440.14 11.85963187 30a 45 4338873.4 15.28312529 30b 49 3852666.6 15.16427609 30c 74 3460597.6 15.05695185 30d 74 3254266 14.99547731 30e 138 2395379.85 14.68905238 30f 138 2389291.8 14.68650756

97

Sample ID Rxn Time (min) Area Counts Ln(Area Counts) 30g 202 1998771.9 14.5080435 30h 202 2199377.7 14.60368501 30i 293 1599341.6 14.2851026 30j 293 1612565.3 14.29333682 30k 391 1312976.5 14.08780726 30l 391 1710189.8 14.35211492

30m 1436 704723.2 13.46556038 30n 1436 594991.3 13.29630206 38a 11 700,600.79 13.45969352 38b 36 465,248.85 13.0503277 38c 66 292,498.00 12.58621311 38d 96 251,999.32 12.43718167 38e 126 249,282.74 12.42634303 38f 156 206,012.96 12.23569436 38g 186 177,890.75 12.08892488 38h 246 195,773.20 12.18471213 45a 43 3431238.6 15.04843186 45b 43 3380638.6 15.03357518 45c 66 2969355.6 14.90385552 45d 66 2945099.4 14.89565313 45e 132 2752251.9 14.82793001 45f 132 2472942 14.72091909 45g 193 2395524.3 14.68911268 45h 193 2551838.25 14.75232454 45i 283 1967343.6 14.49219476 45j 283 1945110.5 14.48082935 45k 383 1543649.7 14.24966011 45l 383 1816397.8 14.41236587

45m 1443 1181234.9 13.98207097 45n 1443 1212900.2 14.00852491

Optimization of Milling Times Milling Time (Min) 3 8 15 23 30 38 45

Rate Constants 0.003148 0.004563 0.0057 0.006373 0.0061 0.005217 0.001968

98

Pseudo 1st Order Rate Constants From the Degradation of PCB-151 vs. Milling Time Plot for the Optimization of Milling

Parameters

0

0.001

0.002

0.003

0.004

0.005

0.006

0.007

0.008

0 5 10 15 20 25 30 35 40 45 50Milling Time (min)

Pse

udo

Firs

t Ord

er R

ate

Con

stan

t (1/

min

)

Results for the Optimization of Ball:Mass Ratio

# of Balls Area Counts Rxn Time (min) Ln(Area Counts) 4a 585,299.07 11 13.27987823 4b 488,342.15 36 13.09877157 4c 450,414.52 66 13.01792359 4d 401,960.11 96 12.90410813 4e 406,522.57 126 12.91539473 4f 421,497.50 156 12.95156913 4g 413,937.04 186 12.93346916 4h 380,654.26 246 12.84964679 8a 5,987,045.20 31 15.60510856 8b 5,155,023.20 31 15.45548218 8c 4,296,796.80 60 15.27338037 8d 4,204,236.20 60 15.25160319 8e 3,962,297.25 125 15.19233453 8f 3,963,680.85 125 15.19268366 8h 2,742,201.90 186 14.82427177 8i 3,456,530.55 186 15.05577591 8j 1,598,203.90 276 14.28439099 8k 2,179,948.60 276 14.59481186 8l 1,770,651.50 376 14.38685812

8m 1,664,806.10 376 14.32521922 8n 945,236.70 1427 13.75919065 8o 945,236.70 1427 13.75919065 12a 563,794.76 11 13.24244556 12b 385,861.37 36 12.86323344 12c 282,483.23 66 12.55137446

99

# of Balls Area Counts Rxn Time (min) Ln(Area Counts) 12d 253,788.26 96 12.44425558 12e 231,687.26 126 12.35314372 12f 194,261.95 156 12.17696278 12g 176,542.99 186 12.0813197 12h 153,402.45 246 11.94082014 16a 4,338,873.40 45 15.28312529 16b 3,852,666.60 49 15.16427609 16c 3,460,597.60 74 15.05695185 16d 3,254,266.00 74 14.99547731 16e 2,395,379.85 138 14.68905238 16f 2,389,291.80 138 14.68650756 16g 1,998,771.90 202 14.5080435 16h 2,199,377.70 202 14.60368501 16i 1,599,341.60 293 14.2851026 16j 1,612,565.30 293 14.29333682 16k 1,312,976.50 391 14.08780726 16l 1,710,189.80 391 14.35211492

16m 704,723.20 1436 13.46556038 16n 594,991.30 1436 13.29630206 20a 527,708.73 5 13.17629976 20b 449,381.73 30 13.01562798 20c 312,454.89 61 12.65221539 20d 238,745.16 91 12.38315199 20e 230,648.12 121 12.34864854 20f 197,607.08 151 12.19403589 20g 181,497.52 181 12.10899727 20h 145,200.18 241 11.88586862 24a 4,275,339.00 49 15.26837396 24b 3,797,961.40 52 15.14997501 24c 3,367,710.20 74 15.02974361 24d 3,094,013.00 80 14.94497951 24e 3,039,789.00 138 14.92729866 24f 2,633,317.50 143 14.78375502 24g 2,418,359.10 202 14.69859981 24h 2,291,904.75 209 14.6448938 24i 1,642,701.30 293 14.31185258 24j 1,495,350.30 301 14.21787105 24k 1,383,161.90 399 14.13988267 24l 1,408,785.60 399 14.15823861

24m 784,391.60 1451 13.57266366 24n 784,391.60 1451 13.57266366 32a 588,609.40 5 13.28551808 32b 398,381.84 30 12.89516622 32c 278,313.47 61 12.53650335 32d 198,102.29 91 12.19653879 32e 150,516.77 121 11.92182979

100

# of Balls Area Counts Rxn Time (min) Ln(Area Counts) 32f 146,710.02 151 11.89621326 32g 130,580.17 181 11.77974265 32h 104,502.03 241 11.55696178

Optimization of Ball:Mass Ratio

Milling Balls 4 8 12 16 20 24 32 Rate Constants 0.001444 0.003461 0.005136 0.006031 0.005429 0.003041 0.007143

Pseudo 1st Order Rate Constant From the Degradation of PCB-151 vs. The Number of Milling Balls Used Plot for the

Optimization of Milling Parameters

0

0.001

0.002

0.003

0.004

0.005

0.006

0.007

0.008

0 5 10 15 20 25

# Balls per 85g Mg/Pd

Pseu

do F

irst

Ord

er R

ate

Cons

tant

(1/m

in)

101

Results for the Degradation of PCB-45 using Mg/Pd prepared by Electrodeposition, Spex (small-scale mill) and Red Devil paint shaker (large-scale mill)

ID Area Counts Dil Factor Conc (ppm) Avg Conc (ppm) Ln(Conc) Rxn Time (min) Electrodep a 590723.55 50 56.62074579 57.52 4.052116 45 Electrodep b 609451.37 50 58.41728621 45 Electrodep c 594952.44 50 57.02641884 57.03 4.043515 73 Electrodep d 553656.85 50 73 Electrodep e 632541.57 50 60.63230523 56.79 4.039301 133 Electrodep f 552363.72 50 52.94092694 133 Electrodep g 565458.77 50 54.19712156 55.5 196 Electrodep h 592632.82 50 56.80390009 196 Electrodep i 575219.29 50 55.13343815 55.17 4.010467 221 Electrodep j 576036.62 50 55.21184377 221 Electrodep k 619937.46 50 59.42320598 57.78 4.056699 0 Electrodep l 585746.35 50 56.14328814 0

Spex a 443226.9 10 4432269 15.3044222 3 Spex b 388554.11 10 3885541.1 15.1727728 8 Spex c 307186.76 10 3071867.6 14.9377963 20 Spex d 217717.87 10 2177178.7 14.5935404 23 Spex e 246286.15 10 2462861.5 14.7168344 41 Spex f 203910.34 10 2039103.4 14.5280208 46 Spex g 168725.32 10 1687253.2 14.3386124 61 Spex h 164957 10 1649570 14.3160252 67 Spex i 170152.4 10 1701524 14.3470349 83 Spex j 130870.72 10 1308707.2 14.0845503 88 Spex k 157850.14 10 1578501.4 14.2719865 101 Spex l 156028.2 10 1560282 14.2603771 106

Spex m 135466.64 10 1354666.4 14.1190658 119 Spex n 132547.61 10 1325476.1 14.0972823 125

Scale-up a 405458.55 10 4054585.5 15.215359 10 Scale up b 355908.41 10 3559084.1 15.0850138 20 Scale-up c 345079.9 10 3450799 15.0541164 30 Scale-up d 280136.25 10 2801362.5 14.8456165 39 Scale-up e 266208.58 10 2662085.8 14.7946205 50 Scale-up f 241369.03 10 2413690.3 14.6966674 60 Scale up g 200747.94 10 2007479.4 14.5123905 70 Scale-up h 207193.12 10 2071931.2 14.5439917 80 Scale-up i 207463.68 10 2074636.8 14.5452967 90 Scale-up j 197276.29 10 1972762.9 14.4949456 99 Scale-up k 198157.69 10 1981576.9 14.4994035 112 Scale-up l 185071.24 10 1850712.4 14.4310812 120

Preparation Rate Constant

Scaled Up Mill 0.0157 SPEX 8000M Mixer/Mill 0.0112 Electrodeless Deposition 0.000217

102

Kinetics Plot for the Degradation of PCB-45 (0.08% Pd bimetallic prepared with Pallamerse solution)

y = -0.00021x + 4.05978

4

4.01

4.02

4.03

4.04

4.05

4.06

4.07

0 50 100 150 200 250

Time(min)

Ln[C

onc

(μg/

mL)

]

103

Kinetics Plot Degradation of PCB-045 (High energy ball mill)

y = -0.0112x + 15.333R2 = 0.9737

14

14.2

14.4

14.6

14.8

15

15.2

15.4

15.6

15.8

16

0 10 20 30 40 50 60 70 80

Time (min)

Ln[C

onc

(μg/

mL)

]

104

Kinetics Plot (Degradation of PCB-045)

y = -0.0145x + 15.226R2 = 0.8738

14.2

14.4

14.6

14.8

15

15.2

15.4

0 10 20 30 40 50 60 70 80

Time (min)

Ln[C

onc

(μg/

mL)

]

105

APPENDIX B SUPPORTING INFORMATION FOR CHAPTER THREE

106

Calibration Data for PCB-1, PCB-2 , and PCB-3 in both MeOH and Water:MeOH (9:1) Congener Conc (μg/mL) Solvent PCB Area Count

1 0.8 Water 180769 1 1.6 Water 382914 1 2.4 Water 561668 1 3.2 Water 855237 1 4 Water 1052518 2 0.8 Water 145746 2 1.6 Water 299366 2 2.4 Water 458256 2 3.2 Water 650009 2 4 Water 786867 3 0.8 Water 207277 3 1.6 Water 444438 3 2.4 Water 660572 3 3.2 Water 949151 3 4 Water 1117959 1 8 MeOH 1428572 1 16 MeOH 2856472 1 24 MeOH 4254067 1 32 MeOH 5605864 1 40 MeOH 7165388 2 8 MeOH 2 16 MeOH 1871392 2 24 MeOH 2679015 2 32 MeOH 3700008 2 40 MeOH 4544298 3 8 MeOH 2235268 3 16 MeOH 5230899 3 24 MeOH 7476790 3 32 MeOH 10128218 3 40 MeOH 12748142

107

Calibration Curve for PCB-1 in 90:10 Water:MeOH

y = 276978x - 58125R2 = 0.9938

0

200000

400000

600000

800000

1000000

1200000

0 1 2 3 4 5

Conc(μg/mL)

Are

a C

ount


y = 204111x - 21817R2 = 0.9979

0100000200000300000400000500000600000700000800000900000

0 1 2 3 4 5

Conc(μg/mL)

Are

a C

ount

108


y = 290760x - 21944R2 = 0.9955

0

200000

400000

600000

800000

1000000

1200000

0 1 2 3 4 5

Conc(μg/mL)

Are

a C

ount

Calibration Curve for PCB-1 in MeOH

y = 177788x - 4834.6R2 = 0.9995

010000002000000300000040000005000000600000070000008000000

0 10 20 30 40 50

Conc(μg/mL)

Are

a C

ount

109


y = 112996x + 34779R2 = 0.9981

0

1000000

2000000

3000000

4000000

5000000

0 10 20 30 40 50

Conc(μg/mL)

Are

a C

ount


y = 324038x - 213057R2 = 0.9986

0

2000000

4000000

6000000

8000000

10000000

12000000

14000000

0 10 20 30 40 50

Conc(μg/mL)

Are

a C

ount

110

Results for the Degradation Studies of PCB-1, PCB-2, and PCB-3 in MeOH with Mg/Pd Sample ID Reaction Time (min) Mass of MgPd (g) Area Count

PCB 1A 4 0.2504 9093843 PCB 1B 7 0.25 9658702 PCB 1C 13 0.2508 9022665 PCB 1D 20 0.2505 9359558 PCB 1E 26 0.255 9050129 PCB 1F 31.5 0.2439 8486810 PCB 1G 38 0.2507 6634005 PCB 1H 45 0.255 5129397 PCB 1I 51 0.2511 4533087 PCB 1J 57 0.249 4363582 PCB 2A 6 0.2461 19925874 PCB 2B 6 0.2448 19838002 PCB 2C 15 0.246 19118834 PCB 2D 15 0.2441 18294968 PCB 2E 26 0.254 17611889 PCB 2F 26 0.2442 17894548 PCB 2G 39 0.243 18582417 PCB 2H 39 0.2568 17363753 PCB 2I 51 0.2451 13418768 PCB 2J 51 0.2573 13499024 PCB 2K 62 0.2479 11444077 PCB 2L 62 0.2539 11310696 PCB 2M 73 0.2532 12257755 PCB 2N 73 0.2526 11059370 PCB 2O 86 0.256 10073918

PCB-3 A1 4 0.2433 22277845 PCB-3 A2 4 0.2453 20286227 PCB-3 B1 14 0.2461 21661881 PCB-3 B2 14 0.2506 19948719 PCB-3 C1 26 0.2546 20170997 PCB-3 C2 26 0.2506 20739532 PCB-3 D1 37 0.253 17075353 PCB-3 D2 37 0.2471 17747457 PCB-3 E1 51 0.2566 13962528 PCB-3 E2 51 0.2541 14443253 PCB-3 F1 66 0.2486 12620314 PCB-3 F2 66 0.2584 10624662 PCB-3 G1 74 0.2437 10069805 PCB-3 G2 75 0.2548 9890061 PCB-3 H1 87 0.2495 8893582 PCB-3 H2 87 0.2515 7823747 PCB-3 I1 101 0.2474 6926877 PCB-3 I2 101 0.2481 7670174 PCB-3 J1 112 0.2433 6485853 PCB-3 J2 112 0.247 7029911

111

Sample ID Area Count (Norm) Conc (Norm) Ln(PCB Norm Conc) PCB 1A 36317264.38 44.62 3.80 PCB 1B 38634808 47.47 3.86 PCB 1C 35975538.28 44.20 3.79 PCB 1D 37363504.99 45.91 3.83 PCB 1E 35490701.96 43.60 3.78 PCB 1F 34796268.96 42.75 3.76 PCB 1G 26461926.61 32.51 3.48 PCB 1H 20115282.35 24.71 3.21 PCB 1I 18052915.17 22.18 3.10 PCB 1J 17524425.7 21.53 3.07 PCB 2A 80966574.56 46.05 3.83 PCB 2B 81037589.87 PCB 2C 77718837.4 43.40 3.77 PCB 2D 74948660.39 PCB 2E 69338145.67 40.54 3.70 PCB 2F 73278247.34 PCB 2G 76470851.85 40.96 3.71 PCB 2H 67615860.59 PCB 2I 54748135.45 30.48 3.42 PCB 2J 52464143.02 PCB 2K 46164086.33 25.79 3.25 PCB 2L 44547837.73 PCB 2M 48411354.66 26.21 3.27 PCB 2N 43782145.68 PCB 2O 39351242.19 22.92 3.13

PCB-3 A1 91565330.87 46.92 3.85 PCB-3 A2 82699661.64 PCB-3 B1 88020646.08 45.13 3.81 PCB-3 B2 79603826.82 PCB-3 C1 79226225.45 43.61 3.78 PCB-3 C2 82759505.19 PCB-3 D1 67491513.83 37.51 3.62 PCB-3 D2 71822974.5 PCB-3 E1 54413593.14 29.95 3.40 PCB-3 E2 56840822.51 PCB-3 F1 50765543.04 24.74 3.21 PCB-3 F2 41117113 PCB-3 G1 41320496.51 21.58 3.07 PCB-3 G2 38814996.08 PCB-3 H1 35645619.24 17.97 2.89 PCB-3 H2 31108337.97 PCB-3 I1 27998694.42 15.86 2.76 PCB-3 I2 30915654.98 PCB-3 J1 26657842.17 14.84 2.70 PCB-3 J2 28461178.14

112


y = -0.0276x + 4.5527R2 = 0.9219

3

3.1

3.2

3.3

3.4

3.5

3.6

3.7

3.8

30 35 40 45 50 55 60

Reaction Time (min)

Ln(μ

g/m

L)


y = -0.0113x + 4.0589R2 = 0.86

3

3.1

3.2

3.3

3.4

3.5

3.6

3.7

3.8

3.9

4

30 40 50 60 70 80 90

Reaction Time (min)

Ln(μ

g/m

L)

113

1st Order Kinetics Degradation Plot of PCB-3 (MeOH) w/MgPd

y = -0.013x + 4.0807R2 = 0.9879

2.5

2.7

2.9

3.1

3.3

3.5

3.7

3.9

25 35 45 55 65 75 85 95 105 115 125

Reaction Time

Ln(μ

g/m

L)

Results for the Degradation Studies of PCB-1, PCB-2, and PCB-3 in Water:MeOH (9:1) with Mg/Pd

Sample ID

Mass (mg)

Rxn Time (min) PCB

PCB Area

Average PCB

PCB Conc Ln(PCB)

1 50.7 3.22 1 1542338 1579402 5.912119 1.777 2 51.3 1 1616466 3 49.4 7.93 1 1501889 1495688 5.609879 1.72453 4 50.5 1 1489487 5 49.2 11.23 1 1465321 1457179 5.470844 1.69943 6 49.3 1 1449036 7 50.3 14.42 1 1359860 1383700 5.205556 1.64973 8 49.7 1 1407539 9 50.8 17.65 1 1276844 1257239 4.748984 1.55793 10 51 1 1237634 11 49 21.12 1 1395946 1296902 4.892181 1.58764 12 51.1 1 1197857 13 50.1 24.42 1 1246354 1218353 4.60859 1.52792 14 48.9 1 1190352 15 49.6 27.92 1 1235339 1197625 4.533752 1.51155 16 51.4 1 1159910 17 49.2 3.28 2 1176752 1186162 5.918243 1.77804 18 48.9 2 1195571 19 50.4 7.08 2 1117280 1113890 5.564164 1.71635

114

Sample ID

Mass (mg)

Rxn Time (min) PCB

PCB Area

Average PCB

PCB Conc Ln(PCB)

20 50.4 2 1110500 21 49.6 10.66 2 949496 931146.5 4.668849 1.54091 22 50.3 2 912797 23 49 13.83 2 838524 829988 4.173244 1.42869 24 49.6 2 821452 25 49.8 17.5 2 839240 772543 3.891804 1.35887 26 50.2 2 705846 27 51.3 20.75 2 821740 815682.5 4.103157 1.41176 28 49.2 2 809625 29 51.2 23.92 2 771989 737343 3.719349 1.31355 30 48.9 2 702697 31 49.9 27.3 2 666996 616424 3.126931 1.14005 32 50 2 616424 33 49.8 3 3 1906269 2133026 7.411508 2.00303 34 49.7 3 2359783 35 49.6 6.48 3 1944331 1944331 6.762536 1.9114 36 50.4 3 37 51.2 9.94 3 1675072 1545818 5.391944 1.68491 38 50.4 3 1416563 39 50.7 12.93 3 1488838 1390039 4.85618 1.58025 40 50.5 3 1291240 41 49.9 16.22 3 1152669 1237805 4.332606 1.46617 42 50 3 1322940 43 50 20.72 3 1057842 1104365 3.873672 1.3542 44 50.1 3 1150888 45 50.3 23.92 3 962644 980865 3.448924 1.23806 46 50.6 3 999086 47 49.4 27.1 3 878167 893614.5 3.148846 1.14704 48 50.7 3 909062

115


y = -0.0113x + 1.8105R2 = 0.9455

1.45

1.5

1.55

1.6

1.65

1.7

1.75

1.8

0 5 10 15 20 25 30

Reaction Time (min)

Ln(μ

g/m

L)


y = -0.0243x + 1.8382R2 = 0.9255

11.11.21.31.4

1.51.61.71.81.9

0 5 10 15 20 25 30

Reaction Time (min)

Ln(μ

g/m

L)

116


y = -0.0358x + 2.0866R2 = 0.9842

1

1.2

1.4

1.6

1.8

2

2.2

2.4

0 5 10 15 20 25 30

Reaction Time (min)

Ln([μg

/mL)

Results for the “Lag” Study examining PCB-1 in MeOH using Mg/Pd

Sample MgPd

(g) Rxn Time

(min) Area

Count Ln(Area Count)

PCB-1(Norm) Conc

Ln(PCB Norm Conc)

A 0.247 37 22915403 16.94731986 44.62 3.798182189 B 0.2462 37.5 18887587 16.75401549 36.77719008 3.604877818 C 0.2542 41.7 16234190 16.60263007 31.6105965 3.453492397 D 0.2547 45 16129644 16.59616938 31.40702851 3.447031706 E 0.2531 48.5 16048433 16.59112177 31.24889754 3.441984097 F 0.2495 52.25 16025681 16.58970306 31.20459571 3.440565382 G 0.254 56 13563113 16.42286439 26.40957709 3.273726713 H 0.2434 60.3 16244676 16.60327578 31.63101444 3.454138109 I 0.251 63.7 12021353 16.30219504 23.40752073 3.15305737 J 0.2562 67.3 11374609 16.24689415 22.14820545 3.097756475 K 0.2574 71.3 12268350 16.32253333 23.88846389 3.17339566 L 0.2457 75 7176776 15.78636082 13.9743449 2.637223142 M 0.2449 79 5759479 15.56635758 11.21463816 2.417219904 N 0.2463 83 7101284 15.77578617 13.82734976 2.626648497 O 0.2562 87.2 6721156 15.72077072 13.08717899 2.571633048 P 0.2571 90.7 5238436 15.47153354 10.20008308 2.322395865 Q 0.2535 94.5 2375725 14.68081321 4.625921242 1.531675539 S 0.2517 98.5 5371683 15.49665183 10.45953656 2.347514152 T 0.2542 102 3979690 15.19671448 7.749100804 2.047576811

117

Lag Study: Pseudo 1st Order Degradation Plot of PCB-1 in MeOH w/ MgPd

y = -0.027x + 4.7748R2 = 0.8622

0

0.5

1

1.5

2

2.5

3

3.5

4

0 20 40 60 80 100 120

Reaction Time (min)

Ln(μ

g/m

L)

Calibration Data for Biphenyl

Conc (μg/mL) Area Count 5 5415172

10 14172426 15 17969121 30 36229908 60 85534404

118

Calibration Curve for Biphenyl (MeOH)

y = 1438372.42x - 2656731.92R2 = 0.99

0

10000000

20000000

30000000

40000000

50000000

60000000

70000000

80000000

90000000

0 10 20 30 40 50 60 70

Conc(μg/mL)

Are

a C

ount

s

Results for the Degradation of Biphenyl in MeOH and Water:MeOH (9:1)

Degradation of Biphenyl w/10% MgPd in MeOH

0

0.2

0.4

0.6

0.8

1

0 50 100 150 200 250

Reaction Time (hrs)

[C]/[

C]o

119

Degradation of Biphenyl w/10% MgPd in 90:10 Water:Methanol

0

0.2

0.4

0.6

0.8

1

1.2

0 20 40 60 80 100 120 140 160

Reaction Time (Hours)

[C]/[

C]o

120

APPENDIX C SUPPORTING INFORMATION FOR CHAPTER FOUR

121

Calibration Data for PCB-151 in MeOH

Conc (μg/mL) Area Count 20 6550566 10 2726237 5 955952

2.5 456491 1.25 233146

Calibration Curve for PCB-151 in MeOH (Tol Extract) as Determined by GC-MS

y = 343387x - 476774R2 = 0.9912

-1000000

0

1000000

2000000

3000000

4000000

5000000

6000000

7000000

0 5 10 15 20 25


Are

a C

ount

122

Results for the Degradation of PCB-151 in MeOH using Mg/Pd (also included is the degradation data for the daughter byproduct PCB 93/95) PCB-151 Data

Sample ID Rxn Time (min) MgPd (g) Norm Area Count Conc Ln(Conc) PCB-151 A 31 0.2509 4798015.498 12.58 2.53 PCB-151 B 63 0.2543 4499261.726 11.71 2.46 PCB-151 C 67 0.2512 4581133.43 11.95 2.48 PCB-151 D 121 0.2522 3719708.469 9.44 2.25 PCB-151 E 127 0.2538 3434545.491 8.61 2.15 PCB-151 F 184 0.2566 2744917.693 6.61 1.89 PCB-151 G 192 0.2559 2764093.746 6.66 1.9 PCB-151 H 244 0.2582 2178001.452 4.95 1.6 PCB-151 I 252 0.2516 2315164.609 5.35 1.68 PCB-151 J 304 0.2566 1934748.367 4.25 1.45 PCB-151 K 311 0.2513 1863808.717 4.04 1.4 PCB-151 L 722 0.2505 1085557 1.77 0.57 PCB-151 M 729 0.2538 1189306.489 2.08 0.73 PCB-151 N 1444 0.2567 871253.0522 1.15 0.14 PCB-151 O 2177 0.2517 864476.7938 1.13 0.12 PCB-151 P 2915 0.2537 0 PCB-151 P 6443 0.2519 807411.5204

Byproduct PCB 93/95 Data

Sample ID Rxn Time (min) MgPd (g) Norm Area Count Conc Ln(Conc) PCB-151 A 31 0.2509 67951.49462 -1.190559 PCB-151 B 63 0.2543 278832.21 -0.576439 PCB-151 C 67 0.2512 326141.6262 -0.438667 PCB-151 D 121 0.2522 700627.2443 0.6518978 -0.4278674 PCB-151 E 127 0.2538 710713.3097 0.6812701 -0.3837964 PCB-151 F 184 0.2566 869743.6146 1.1443928 0.1348742 PCB-151 G 192 0.2559 824506.2134 1.012654 0.01257459PCB-151 H 244 0.2582 890601.2335 1.2051337 0.18659049PCB-151 I 252 0.2516 954503.5672 1.3912279 0.33018672PCB-151 J 304 0.2566 1033524.341 1.6213495 0.48325883PCB-151 K 311 0.2513 1014509.051 1.5659738 0.44850788PCB-151 L 722 0.2505 812940 0.9789712 -0.021253 PCB-151 M 729 0.2538 831223.747 1.0322166 0.03170849PCB-151 N 1444 0.2567 788281.6362 0.907162 -0.0974342 PCB-151 O 2177 0.2517 785931.0608 0.9003167 -0.1050087 PCB-151 P 2915 0.2537 0 PCB-151 P 6443 0.2519 779340.4049

123

Comparison Plot of the Degradation of PCB-151 and the Appearance of the PCB-93/95

0

2

4

6

8

10

12

14

0 500 1000 1500 2000 2500

Reaction Time (min)

Con

c (μ

g/m

L)

PCB-151

PCB93/95

New Calibration Curve for PCB-151 in MeOH for use with Reactivated Mg/Pd

Conc (μg/mL) Response 10 150816 2 36137 4 85206 6 104743 8 144635

124

Calibration Curve run on 8/27/07 at NASA for PCB-151 in MeOH

y = 14439x + 17671R2 = 0.9475

0

20000

40000

60000

80000

100000

120000

140000

160000

180000

0 2 4 6 8 10 12

Conc (μg/mL)

Area

Cou

nts

125

Results for Degradation of PCB-151 in MeOH using Reactivated Mg/Pd (Conc. Normalized to 10 ug/mL, corresponding to actual starting concentration of the solution)

ID MgPd

(g)

Rxn Time (min) Response

Conc (ppm)

Norm. Conc

Avg Conc Ln(Conc)

201 0.2536 2.5 1552678 106.41 107.94 9.78 2.280196 202 0.2464 1300215 88.91 87.63 203 0.2473 12.483 996275 67.84 67.11 6.51 1.873251 204 0.2516 921823 62.68 63.08 205 0.2486 22.5 874167 59.37 59.04 5.87 1.769552 206 0.2482 865213 58.75 58.33 207 0.2482 32.483 768941 52.08 51.71 5.04 1.617878 208 0.2516 722137 48.83 49.14 209 0.2467 42.4 655716 44.23 43.65 4.19 1.432991 210 0.2503 596609 40.13 40.18 211 0.248 52.417 571624 38.4 38.09 3.78 1.329194 212 0.2463 566315 38.03 37.47 213 0.2483 62.25 541930 36.34 36.09 3.47 1.243437 214 0.2531 491629 32.85 33.26 215 0.2533 72.583 447514 29.79 30.18 3.13 1.141818 216 0.2502 485623 32.44 32.47 217 0.2491 82.25 398758 26.41 26.31 2.49 0.912751 218 0.2542 351213 23.12 23.51 219 0.2488 92.5 362755 23.92 23.81 2.58 0.948757 220 0.2524 415661 27.58 27.84 221 0.2475 102.5 346251 22.77 22.54 2.14 0.762746 222 0.2535 307195 20.06 20.34 223 0.2458 112.5 436138 29 28.51 2.61 0.960798 224 0.2464 365554 24.11 23.76 225 0.2476 122.5 341632 22.45 22.23 2.05 0.717985 226 0.2483 290437 18.9 18.77 227 0.2467 132.417 385407 25.49 25.15 2.26 0.81372 228 0.2456 311026 20.33 19.97 229 0.2501 142.45 281166 18.26 18.27 1.83 0.601665 230 0.2475 283496 18.42 18.24 231 0.2465 152.5 385689 25.51 25.15 1.96 0.670918 232 0.2514 218079 13.89 13.97 233 0.251 162.5 264704 17.12 17.19 1.75 0.561156 234 0.2491 276341 17.93 17.87 235 0.2513 172.5 254666 16.42 16.51 1.46 0.381559 236 0.2547 198862 12.55 12.79 237 0.2524 182.417 230405 14.74 14.88 1.43 0.356847 238 0.2523 213451 13.57 13.69 239 0.249 192.5 222630 14.2 14.14 1.34 0.295623 240 0.2519 200165 12.64 12.74 241 0.2515 202.75 214616 13.65 13.73 1.33 0.286275 242 0.2519 202368 12.8 12.9

126

ID MgPd

(g)


Conc (ppm)

Norm. Conc

Avg Conc Ln(Conc)

243 0.2496 212.43 201542 12.74 12.72 1.46 0.37928 244 0.2533 252809 16.29 16.51 245 0.2515 222.67 194193 12.23 12.3 1.22 0.200934 246 0.2467 195360 12.31 12.15 247 0.2474 232.5 216981 13.81 13.67 1.47 0.384237 248 0.2466 247365 15.92 15.7 249 0.2464 242.5 239848 15.4 15.18 1.44 0.36397 250 0.2484 215186 13.69 13.6 251 0.2463 252.417 295699 19.27 18.98 1.59 0.462976 252 0.2512 201399 12.73 12.79 253 0.2506 262.57 213484 13.57 13.6 1.39 0.330863 254 0.2543 219771 14 14.24 255 0.249 273.7 169120 10.49 10.45 1.11 0.10487 256 0.2463 190001 11.94 11.76 257 0.252 282.55 147318 8.98 9.05 0.9 -0.10992 258 0.2539 143704 8.73 8.87 259 0.2483 294.67 240222 15.42 15.32 1.28 0.245358 260 0.2487 166328 10.3 10.25 261 0.248 302.517 170375 10.58 10.5 1.05 0.051847 262 0.2474 171864 10.68 10.57 263 0.2484 312.5 240815 15.46 15.36 1.31 0.270859 264 0.2491 174982 10.9 10.86 265 0.2524 322.983 160035 9.86 9.95 1.04 0.038612 266 0.2531 172175 10.7 10.83 267 0.2463 336.67 139392 8.43 8.31 0.93 -0.07308 268 0.2494 166471 10.31 10.29 269 0.2469 342.85 150260 9.18 9.07 0.9 -0.11049 270 0.2467 147019 8.96 8.84 271 0.2534 355.7 132085 7.92 8.03 0.89 -0.11737 272 0.2464 160616 9.9 9.76 273 0.249 363.517 112760 6.58 6.55 0.68 -0.3861 274 0.2479 120183 7.1 7.04 275 0.248 375.417 116969 6.88 6.82 0.98 -0.02428 276 0.2535 198371 12.52 12.7 277 0.2502 380.483 134148 8.07 8.08 0.79 -0.23052 278 0.2464 132060 7.92 7.81 279 0.2506 393 118188 6.96 6.98 0.67 -0.40178 280 0.2522 109430 6.35 6.41 281 0.2494 405.17 104461 6.01 6 0.59 -0.5233 282 0.2477 103036 5.91 5.86 283 0.2507 412.5 108615 6.3 6.32 0.54 -0.61647 284 0.2472 83184 4.53 4.48 285 0.2465 422.583 129164 7.72 7.61 0.73 -0.32094

127

ID MgPd

(g)


Conc (ppm)

Norm. Conc

Avg Conc Ln(Conc)

286 0.251 116819 6.87 6.9 287 0.251 432.5 108248 6.27 6.3 0.54 -0.61328 288 0.2498 83294 4.54 4.54 289 0.2495 442.67 93469 5.25 5.24 0.61 -0.49044 290 0.2539 117250 6.9 7.01 291 0.2486 452.916 89668 4.98 4.95 0.4 -0.92388 292 0.2473 61288 3.02 2.99 293 0.2505 462.3 82152 4.46 4.47 0.43 -0.83297 294 0.2486 79101 4.25 4.23 295 0.2472 472.17 109937 6.39 6.32 0.54 -0.61899 296 0.2484 82450 4.48 4.45

Results for the Degradation of PCB-151 in MeOD using Reactivated Mg/Pd (Calibration from previous data utilized)

Sample ID

MgPd (g)

Rxn Time (min) Solvent

PCB-151

PCB Average

Ln(PCB-151)

PCB Conc

1 0.2553 3.58 MeOD 1.3E+07 13850539.5 16.44383474 42.3087632 0.2556 MeOD 1.4E+07 3 0.2448 12.55 MeOD 1.2E+07 12110905 16.30961684 32.7105754 0.2455 MeOD 1.2E+07 5 0.2483 24.25 MeOD 11655478 16.27128684 30.19782 6 0.2489 MeOD 1.2E+07 7 0.2522 32.63 MeOD 1E+07 10081144 16.12617731 21.5116558 0.2485 MeOD 9715380 9 0.2535 42.75 MeOD 8755773 8708987.5 15.9798661 13.940975

10 0.2506 N/A 11 0.2471 MeOD 8662202 12 0.2536 52.92 MeOD 8171654 8035306.5 15.8993557 10.22403513 0.2524 MeOD 7898959 14 0.2485 72.58 MeOD 6456390 6313796.5 15.65824772 0.725846615 0.2541 MeOD 6171203 16 0.2463 102.67 MeOD 4553081 4146909 15.2378738 -11.22965 17 0.2531 MeOD 3740737 18 0.2488 142.58 MeOD 2773343 2750126.5 14.82715747 -18.9362 19 0.2497 MeOD 2726910

128

Pseudo 1st Order Degradation Plot of PCB-151 Using Reballmilled MgPd in MeOH

y = -0.0118x + 0.0578R2 = 0.9961

-2

-1.5

-1

-0.5

0

0.5

0 20 40 60 80 100 120 140 160

Reaction Time (min)

Ln(C

/Co)

129

REFERENCES

Agarwal, S.; Al-Abed, S. R.; Dionysiou, D. D. Environ. Sci. Technol. 2007, 41, 3722-3727.

Aitken, Brian; Geiger, C.L.; Clausen, C.A., Milum, K.M. and Brooks, K.. Variables Associated with Mechanical Alloying of Bimetals for PCB Remediation. In Remediation of Chlorinated and Recalcitrant Compounds, Proceedings of the 5th International Conference on Remediation of Chlorinated and Recalcitrant Compounds, Monterey, CA, May 22-25, 2006; Curran Associates, Inc: New York.

Alonso, F.; Beletskaya, I. P.; Yus, M. Chem. Rev. 2002, 102, 4009-4091.

Austin, L. G.; Industrial Engineering of Chem. Processes Design and Development. 1973, 12, No. 2.

Ballschmiter, K. H., Zell, M. Fresenius Z Anal Chem. 1980, 302, 20-31.

Bedard, D. L.; Wagner, R. E.; Brennan, M. J.; Haberl, M. L.; Brown, J. F., Jr.. App. Environ. Microbiol. 1987, 53, 1094.

Bedard, D. L.; Quensen, J. F., III. In Microbial Transformation and Degradation of Toxic Organic Chemicals; Young, L. Y.; Cerniglia, C., Eds.; John Wiley & Sons: New York, 1995; 127.

Blanksby, Stephen J.; Ellison, G. Barney. Bond Dissociation Energies of Organic Molecules. Acc. Chem. Res. 2003, 36, 255-263.

Carey, Francis A.; Sundberg, Richard J. Advanced Organic Chemistry Part A: Structure and Mechanisms.

Cheng, I. F.; Fernando, Q.; Korte, N. Environ. Sci. Technol. 1997, 31, 1074-1078.

Chuang, F.; Larson, R. A.; Wessman, M. S. Environ. Sci. Technol. 1995, 29, 2460-2463.

Cwiertny, David M.; Bransfield, Stephen J.; Roberts, A. Lynn. Environ Sci. Technol. 2007, 41, 3734-3740.

Cybulski, Andrzej; Moulijn, Jacob A. Structured Catalysts and Reactors. Taylor and Francis: Boca Raton, 2006, 585.

Devaswithin, A.; Pitchimani, B.; Silva, S. R.; Journal of the American Chemical Society.1988, 27, 723-726.

Doyle, J.G.; Miles, T.A.; Parker, E.; Cheng, I.F. Microchemical Journal. 1998, 60, 290-295.

130

Eisenreich S. J., Looney B. B. and Hollod G. J. In Physical Behavior of PCBs in the Great Lakes; Mackay, D.; Paterson, S., Eisenreich, S. J.; Simmons, M., Eds. Ann Arbor Science, Ann Arbor, Michigan, 1983, 115-125.

Engelmann, M.D.; Hutcheson, R.; Henschied, K.; Neal, R.; Cheng, I.F. Microchemical Journal. 2003, 74, 19-25.

Engelmann, M.D.; Doyle, J.G.; Cheng, I.F. Chemosphere. 2001, 43, 195-198.

EPA. Toxic Information Series (1983) Polychlorinated biphenyls. U.S. EPA Toxics Information Series, United States Environmental Protection Agency, Office of Toxics Substances, TSCA Assistance Office (TS-799) Washington, District of Columbia.

Erickson, M.D. Analytical Chemistry of PCBs; Lewis: Boca Raton, 1992; 5-45.

Gomez-Sainero, L. M.; Seonane, X. L.; Fierro, J. L.; Arcoya, A. Journal of Catalysis. 2002, 209, 279-288.

Grittini, C.; Malcomson, M.; Fernando, Q.; Korte, N. Environ. Sci. Technol. 1995, 29, 2898-2900.

Jones, C. G., Silverman, J., Al-Sheikly. Environ. Sci. Technol. 2003, 37, 5773-5777.

Kim, Young-Hun; Shin, Won Sik; Ko, Seok-Oh. Reductive Dechlorination of Chlorinated Biphenyls by Palladized Zero-Valent Metals. Environmental Sci. Health. 2004, A39(5), 1177-1188.

Korte, N.E.; West, O.R.; Liang, L.; Gu, B.; Zutman, J.L.; Fernando, Q. Waste Management. 2002, 22, 343-349.

Kovenklioglu, S.; Cao, Z.; Shah, D.; Farrauto, R. J.; Balko, E. N. Direct Catalytic Hydrodechlorination of toxic organics in wastewater. AIChE J. 1992, 38, 1003-1012.

Kremer, F., Nakles, D. V., Biksey, T. Remediation. 2006, Winter, 139-147

Li, T.; Farrell, J. Environ. Sci. Technol. 2000, 34, 173-179.

Liu, Y.; Yang, F.; Yue, P.L.; Chen, G.; Wat Res. 2001, 35, 1887-1890.

Lowry, G. V.; Reinhard, M. Environ. Sci. Technol. 1999, 33, 1905-1910.

Lü, L.; Lai, M. O. Mechanical Alloying. Kluwer Academic Publishers: Boston, 1998, 1-20.

Manning, B. A., Kiser, J. R., Kwon, H., Kanel S. R. Environ. Sci. Technol. 2007, 41, 586-592.

Matheson, L. J.; Tratnyek, P.G. Environ. Sci. Techno. 1994, 28, 2045-2053.

131

McDowall, L. Degradation of Toxic Chemicals by Zero-Valent Metal Nanoparticles – A Literature Review. Australian Department of Defense. Human Protection and Performance Division, Defense Science and Technology Organisation, Nov. 2005.

Mousa, M. A.; Quensen, J. F., III; Chou, K.; Boyd, S. A. Environ. Sci. Technol. 1996, 30, 2087.

Moran, M. J.; Zogorski, J. S.; Squillace, P. J. Environ. Sci. Technol. 2007, 41, 74-81.

Muftikian, R.; Fernando, Q.; Korte, N. Water Res. 1995, 29, 2434-2439.

Muftikian, R.; Nebesny, K.; Fernando, Q.; Korte, N. Environ. Sci. Technol. 1996, 30, 3593-3596.

Pirard, Catherine; Eppe, Guathier; Massart, A; Fierens, Sebastien; Pauw, Edwin; Focant, J.; Environ. Sci. Technol. 2005, 39, 4721-4728.

Quensen, J. F., III; Mousa, M. A.; Boyd, S. A.; Sanderson, J. T.; Fruese, K. L.; Giesy, J. P. Environ. Toxicol. Chem. 1998, 17, 806.

Rice, C. P.; White, D. S. Environ. Tox. Chem. 1987, 6, 259-274.

Ross, G. Ecotoxicology Environ. Safety. 2004, 59, 275-291.

Schmidt, C. Chemical Innovation. 2001, 31 (12), 48-52.

Schrick, B.; Hydutsky, B. W.; Blough, J. L.; Mallouk, T. E. Chem Mater. 2004, 16, 2187-2193.

Solomon, G. M., Huddle, A.M. J. Epidemiol. Community Health. 2002, 56, 826-827.

Suryanarayana, C; Progress in Materials Science. 2001, 46, 11-36

Suryanarayana, C; Mechanical Alloying and Milling. Marcel Dekker: New York; 2004, 35-47.

Tamura, Kunio; Tanaka, Tatsuo; Industrial Engineering of Chem. Processes Design and Development. 1970, 9, No. 2

Wang, C.B.; Zhang, W. Environ. Sci. Technol. 1997, 31, 2154-2156.

Wiegel, Juergen; Wu, Qingzhong. FEMS Microbiology Ecology. 2000, 32, 1-15.

Wu Wenhai; Xu Jie; Zhao Hongmei; Zhang Qing; Liao Shijian. Chemosphere, 2005, 60(7), 944-50.

Voorspoels, S., Covaci, A., Neels, H. Environ. Toxic. Pharm. 2007.

Zhang, W.; Wang, C.B.; Lien, H.L.; Catalysis Today. 1998, 40, 387-395.

determination of the degradation mechanism for

Documents