determination of the degradation mechanism for
TRANSCRIPT
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
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
Results and Discussion ............................................................................................................. 55
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
Results and Discussion ............................................................................................................. 72
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
Figure 18: Pseudo 1st order kinetic plot of the degradation of PCB-003 with Mg/Pd in methanol
............................................................................................................................................... 57
Figure 19: Pseudo 1st order kinetic plot of the degradation of PCB-001 with Mg/Pd in 9:1
water:methanol...................................................................................................................... 58
Figure 20: Pseudo 1st order kinetic plot of the degradation of PCB-002 with Mg/Pd in 9:1
water:methanol...................................................................................................................... 58
Figure 21: Pseudo 1st order kinetic plot of the degradation of PCB-003 with Mg/Pd in 9:1
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.
Results and Discussion
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
Pseudo 1st Order Degradation Plot of PCB-2 in MeOH w/ MgPd
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)
Figure 17: Pseudo 1st order kinetic plot of the degradation of PCB-002 with Mg/Pd in methanol
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)
Figure 18: Pseudo 1st order kinetic plot of the degradation of PCB-003 with Mg/Pd in methanol
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
Pseudo 1st Order Degradation Plot of PCB-2 in 90:10 Water:MeOH using MgPd
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
)
Figure 20: Pseudo 1st order kinetic plot of the degradation of PCB-002 with Mg/Pd in 9:1 water:methanol
59
Pseudo 1st Order Degradation Plot of PCB-3 in 90:10 Water:MeOH using MgPd
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
)
Figure 21: Pseudo 1st order kinetic plot of the degradation of PCB-003 with Mg/Pd in 9:1 water:methanol
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.
Results and Discussion
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
Concentration (μg/mL)
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
Calibration Curve for PCB-2 in 90:10 Water:MeOH
y = 204111x - 21817R2 = 0.9979
0100000200000300000400000500000600000700000800000900000
0 1 2 3 4 5
Conc(μg/mL)
Are
a C
ount
108
Calibration Curve for PCB-3 in 90:10 Water:MeOH
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
Calibration Curve for PCB-2 in MeOH
y = 112996x + 34779R2 = 0.9981
0
1000000
2000000
3000000
4000000
5000000
0 10 20 30 40 50
Conc(μg/mL)
Are
a C
ount
Calibration Curve for PCB-3 in MeOH
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
Pseudo 1st Order Degradation Plot of PCB-1 in MeOH w/ MgPd
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)
Pseudo 1st Order Degradation Plot of PCB-2 in MeOH w/ MgPd
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
Pseudo 1st Order Degradation Plot of PCB-1 in 90:10 Water:MeOH using MgPd
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)
Pseudo 1st Order Degradation Plot of PCB-2 in 90:10 Water:MeOH using MgPd
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
Pseudo 1st Order Degradation Plot of PCB-3 in 90:10 Water:MeOH using MgPd
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
Concentration (μg/mL)
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)
Rxn Time (min) Response
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)
Rxn Time (min) Response
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.