W&M ScholarWorks
Dissertations, Theses, and Masters Projects Theses, Dissertations, & Master Projects
1995
Mechanisms of Poly(Vinyl Chloride) Pyrolysis in the Presence of Transition Metals
William Stephen Bryant College of William & Mary - Arts & Sciences
Follow this and additional works at: https://scholarworks.wm.edu/etd
Part of the Physical Chemistry Commons, and the Polymer Science Commons
Recommended Citation Bryant, William Stephen, "Mechanisms of Poly(Vinyl Chloride) Pyrolysis in the Presence of Transition Metals" (1995). Dissertations, Theses, and Masters Projects. Paper 1539626994. https://dx.doi.org/doi:10.21220/s2-3ya0-ks60
This Thesis is brought to you for free and open access by the Theses, Dissertations, & Master Projects at W&M ScholarWorks. It has been accepted for inclusion in Dissertations, Theses, and Masters Projects by an authorized administrator of W&M ScholarWorks. For more information, please contact [email protected]. MECHANISMS OF POLY(VINYL CHLORIDE) PYROLYSIS IN THE
PRESENCE OF TRANSITION METALS
A Thesis
Presented to
The Faculty of the Department of Chemistry
The College of William and Mary in Virginia
In Partial Fulfillment
Of the Requirements for the Degree of
Master of Arts
by
William Stephen Bryant
1995 APPROVAL SHEET
This thesis is submitted in partial fulfillment of the
requirements for the degree of
Master of Arts
Author
Approved, August 4, 1995
W. H. Starnes, Jr. TABLE OF CONTENTS
Acknowledgments vi
List of Tables vii
List of Figures viii
List of Schemes xi
Abstract xiii
I. Introduction
A. PVC Smoke and Fire 2
B. Thermal Degradation of PVC 4
C. Mechanisms for Benzene Formation 11
D. Current Smoke Suppression Additives 14
E. Reductive Coupling Mechanism 19
F. Low-Valent or Zero-Valent Metal Additives for Smoke
Suppression 22
II. Experimental
A. Instrumentation
1. Gas Chromatography-Mass Spectroscopy (GC-MS) 24
2. Nuclear Magnetic Resonance (NMR) 25
3. Infrared Spectroscopy (IR) 25
iii 4. Melting Point Apparatus 25
5. Thermogravimetric Analysis (TGA) 25
B. Model-Compound Reactions
1. General 27
2. Open System (External Flame) 28
3. Open System (Oil Bath) 29
4. Closed System 30
C. PVC Gel Reactions
1. General 32
2. Solid State Gel Reactions 32
3. IR Analysis Reactions 33
4. Solvated Gel Reactions 33
D. Synthesis of 3,-4-Dimethyl-1,5-hexadiene and Its Isomers 34
E. Synthesis of 4-Chloromethylbiphenyl 35
F. Synthesis of 4,4’-Diphenylbibenzyl 42
G. Synthesis of Cobalt(II) Formate 47
III. Results and Discussion
A. Model Compounds 49
1. 3-Chloro-l-butene 50
2. Benzyl Chloride 62
iv 3. Cinnamyl Chloride 71
4 .4-Chloromethylbiphenyl 80
B. Solid State Degradation of PVC 91
C. Gelation of PVC in Solution 105
IV. Conclusions 106
References 108
v ACKNOWLEDGMENTS
The author wishes to express his appreciation to Professor William H.
Starnes, Jr., under whose guidance this investigation was conducted, for his patient counsel and criticism. Appreciation is also expressed to Professors Robert D. Pike and Michael A. G. Berg for their careful review and criticism of the manuscript.
Particular gratitude goes to my parents, Mr. and Mrs. Julian A. Bryant, Jr., without whom this experience could not have been made possible.
Finally, the author is especially grateful to his wife, Kerry L. Bryant, for her love, sacrifice, and caring encouragement during this research.
VI List of Tables
1. Coupling of 3-chloro-1 -butene (1) with metal additives. 60
2. GC area percentages for pyrolysis products from benzyl chloride and metal
. additives. 64
3. Gel yields. 92
vii List of Figures
1. Weight loss vs. temperature profile for PVC. 4
2. Most thermally unstable defect sites in PVC. 5
3. Model-compound reactions with molybdenum-containing Lewis acids. 16
4. Experimental setup for the open system (external flame) reactions. 28
5. Experimental setup for the open system (oil bath) reactions. 29
6. Experimental setup for the closed system reactions. 30
7 . Experimental setup for the solid state gel reactions. 32
8. GC-MS results for 4-biphenylmethanol. 39
9. *H NMR spectrum of 4-chloromethylbiphenyl. 40
10. 13C NMR spectrum of 4-chloromethylbiphenyl. 41
11. GC-MS results for 4-chloromethylbiphenyl. 41
12. *11 NMR spectrum of 4,4'-diphenylbibenzyl. 44
13. 13C NMR spectrum of 4,4'-diphenylbibenzyl. 45
14. GC-MS results for 4,4'-diphenylbibenzyl. 46
15. IR spectrum for cobalt(II) formate. 48
16. GC-MS data for the coupled products from 3-chloro-1 -butene. 51 -57
17. Isomers of 3-chloro-1 -butene reductive coupling product. 58
18. Major contributing structures of the cis radical. 59
viii 19. GC-MS results for bibenzyl. 63
20. GC chromatogram of pyrolysis products from Mo(CO)6 and benzyl
chloride. 65
21. GC chromatogram of pyrolysis products from copper powder and benzyl
chloride. 67
22. GC chromatogram of pyrolysis products from (methylcyclopentadienyl)-
manganese tricarbonyl and benzyl chloride. 68
23. GC chromatogram of pyrolysis products from tetrakis(acetonitrile)-
copper(I) hexafluorophosphate and benzyl chloride. 69
24. GC-MS results for pyrolysis products from copper(II) formate and benzyl
chloride. 70
25. GC chromatogram of “95%” cinnamyl chloride. 72
26. GC chromatogram of “97%” cinnamyl chloride. 73
27. GC chromatogram of “97%” cinnamyl bromide. 74
28. ]H NMR spectrum of “97%” cinnamyl chloride. 75
29. Partial GC chromatogram of “97%” cinnamyl chloride. 77
30. GC-MS data for products from the lithium coupling reaction of cinnamyl
chloride. 78
31. GC-MS data for control run with neat 4-chloromethylbiphenyl. 82
32. GC-MS data for pyrolysis products from iron nonacarbonyl and 4-
ix chloromethylbiphenyl. 84
33. GC-MS data for pyrolysis products from copper powder and 4-chloro
methylbiphenyl. 85
34. GC-MS data for products resulting from copper(H) formate
decomposition followed by addition of 4-chloromethylbiphenyl. 87
35. GC-MS data for pyrolysis products from copper(II) formate and
4-chloromethylbiphenyl. 89
36. IR spectra for PVC degraded in the presence of copper additives. 95
37. IR spectra for PVC degraded in the presence of a tungsten additive. 96
38. IR spectra for PVC degraded in the presence of molybdenum additives. 97
39. IR spectra for PVC degraded in the presence of cobalt additives. 98
40. IR spectra for PVC degraded in the presence of nickel additives. 99
41. IR spectra for PVC degraded in the presence of a manganese additive. 100
42. IR spectra for PVC degraded in the presence of iron additives. 101
x List of Schemes
1. Combustion of PVC. 3
2. Ion-pair mechanism. 6
3. Radical mechanism. 7
4. Chloronium cation intermediate mechanism. 8
5. Four-membered cyclic transition state mechanism. 9
6. Thermal degradation of PVC. 10
7. Hexatriene mechanism for formation of benzene. 11
8. Octatetraene mechanism for formation of benzene. 12
9. PVC pyrolysis scheme. 15
10. Mo(VI)-catalyzed olefin dimerization and chloroalkylation. 17
11 Possible mechanism for Mo(VI)-catalyzed crosslinking in pyrolyzing
PVC. 17
12. Metal-catalyzed reductive coupling mechanism. 19
13. Allylic and alkyl site coupling. 20
14. Homocoupling of alkyl halides via activated zero-valent copper. 23
15. Synthesis of 3,4-dimethyl-1,5-hexadiene and its isomers. 34
16. Conversion of 4-biphenyl carboxyl ic acid into the corresponding alcohol. 35
17. Conversion of 4-biphenylmethanol into 4-chloromethylbiphenyl. 36
xi 18. Synthesis of 4,4-diphenylbibenzyl. 42
19. Synthesis of cobalt formate. 47
20. Formation of radical intermediate from 3-chloro-1-butene. 59
21. Possible fragmentation routes for cinnamyl chloride coupled product. 79
22. Possible fragmentation route for 4,4'-diphenylbibenzyl. 81
23. Lewis-acid-catalyzed dehydrochlorination with subsequent crosslinking. 93
24. Crosslinking of PVC chains via Friedel-Crafts alkylation. 103
25. Crosslinking of PVC chains via reductive coupling. 103
xii ABSTRACT
Many metal additives used as smoke suppressants and fire retardants for the thermal degradation of poly (vinyl chloride) (PVC) utilize a Lewis acid catalyzed crosslinking mechanism. However, this mechanism involves cracking reactions that generate volatile hydrocarbons which, in turn, increase flame spread. An alternative mechanism involving reductive coupling has been studied using zero-valent and low-valent transition metal additives (e.g., carbonyls, formates, and other compounds).
The studies with low-molecular-weight models have shown that the reductive coupling mechanism is possible. Many of the metal carbonyls coupled 3-chloro-1- butene when decomposed prior to the introduction of the model compound. Copper- containing additives were the most effective for the coupling of the other model compounds studied. The solid-state PVC studies have also indicated the possible occurrence of a reductive coupling mechanism during the pyrolysis of the polymer. Several of the additives caused significant polymer gelation but did not cause the extensive double bond formation that would have indicated the operation of a Lewis- acid-catalyzed process. MECHANISMS OF POLY(VINYL CHLORIDE) PYROLYSIS IN THE
PRESENCE OF TRANSITION METALS 2
I. INTRODUCTION
A. PVC Smoke and Fire
The industrial production of poly(vinyl chloride) (PVC), (-CH2CHCl-)n, is
second only in terms of tonnage to that of polyethylene.1 It is considered to be the most important bulk polymer to the plastics industry today.2 It is well-known that
PVC is widely used in the construction market as insulation for electrical and communication wiring, water pipes, home interior furnishings, and many other construction applications.3,4 In 1992, 5.6 billion pounds of PVC were utilized in construction in the U.S. alone.3
In a typical dwelling fire, excessive amounts of smoke and gases are produced. More than 80% of deaths directly associated with fires come not from bums but from asphyxiation by smoke and toxic gases.5 The necessity for fire retardance and smoke suppression of PVC is abundantly evident from its excessive use in construction materials.
Although pure PVC is inherently fire resistant,6' 9 when subjected to sufficient thermal energy, it does evolve smoke and toxic gas as shown in Scheme
1.10 A part of the toxic gas consists of hydrogen chloride (HC1).11 The presence of plasticizers, processing additives, stabilizers, and other additives can increase the 3
SMOKE A
pvc -^!:olIsV TOXIC GAS , lgmt,on „ FLAME ------0 2(Air) A Heat
Schem e 1 Combustion of PVC.10
flammability of PVC.7 Extensive research has been conducted to reduce the fire hazard from PVC through the use of flame retardants and smoke suppressants.
Mechanistic studies of the effectiveness of fire-retardant additives have been carried out and have contributed to the overall understanding of the thermal degradation of this polymer. 4
B. Thermal Degradation of PVC
The thermal degradation of PVC occurs mainly in a two-stage process. The first stage involves the evolution of hydrogen chloride (HC1) gas and the formation of conjugated polyene sequences.12' 15 Using thermal gravimetric analysis (TGA),
Boettner, Ball, and Weiss found that PVC begins to experience rapid weight loss as it reaches 300 °C in air (Figure l).16 They determined that approximately 60% of the original weight was lost as HC1. This loss corresponds to almost all of the chlorine content of the polymer. The second stage of degradation is the production of unsubstituted aromatics (benzene, naphthalene, etc.).17 Benzene is considered to
Figure 1 Boettner et al. TGA
Figure 1 Weight loss vs. temperature profile for PVC.16 5 be the major source of combustion and smoke in the burning of PVC.7
Michelson et al. studied the degradation of PVC using a TGA instrument coupled to a mass spectrometer (MS).18 During an isothermal run at 240 °C (helium purge) they detected both HC1 and benzene by focusing the detector at m/z = 78
(benzene) and m/z = 36 (HC1). They concluded that HC1 gas and benzene evolution occurred simultaneously. This belief is held today.17,19 21
The principal source of the thermal instability of PVC is the labile defect sites.12,22 The formation of these sites is a direct result of the preparation of the polymer.23 The most thermally unstable defect structures are shown in Figure 2.
The tertiary chlorides are considered by many researchers to be the largest contributors to the dehydrochlorination and subsequent degradation of PVC.24 25
Cl Cl Cl Cl Cl
Cl
Internal Allylic Chloride Tertiary Chloride
Figure 2 Most thermally unstable defect sites in PVC. 6
However, others believe that allylic chlorides are even more influential in the degradation.26"28
As noted above, the first stage of PVC degradation involves the loss of HC1 and the concurrent formation of conjugated polyene sequences. Currently there are at least four proposed mechanisms for this process. The first is the ion-pair mechanism (Scheme 2).12,22 This involves the loss of a chloride ion from the polymer backbone, usually at a defect site, and subsequent loss of a neighboring proton to form HC1. This process results in the formation of an allylic chloride group. Further dehydrochlorination can continue along the chain and create polyene sequences.
Conjugated Polyene Growth
Scheme 2 Ion-pair mechanism.12,22 The second mechanism involves radicals (Scheme 3).12,22 A free radical, R,
abstracts a hydrogen atom from the polymer chain, thus creating RH and a polymer
chain radical. This reaction is followed by the loss of the neighboring chlorine atom
as a radical and the formation of a double bond. The chlorine radical is then able to
abstract another hydrogen atom to form HC1 and an allylic radical. This process
continues in order to create a conjugated polyene sequence.
RH +
+ HC1 ^ Cl- +
+ C1* etc.
Scheme 3 Radical mechanism.12,22
The third mechanism (Scheme 4)22 includes a cyclic chloronium cation intermediate. This mechanism again forms HC1 and initiates sequential conjugated polyene formation. 8
cyclic chloronium cation intermediate
-HC1
-HC1
-HC1 ^ continued conjugated polyene formation
Scheme 4 Chloronium cation intermediate mechanism.22
The final mechanism (Scheme 5)12 involves the loss of a molecule of HC1 through a four-membered cyclic transition state. This process results in the formation of an internal allylic chloride group. Further loss of HC1 can occur through the same type of four-membered transition state to create a polyene sequence.
Recent evidence indicates that dehydrochlorination and the formation of polyene sequences occur by mechanisms involving polar concerted transition states rather than by a mechanism initiated by free radicals.29 Boughdady et al.30 have recently reported model-compound evidence to show that the thermal 9
-HC1
-HC1 continued conjugated polyene formation
Scheme 5 Four-membered cyclic transition state mechanism.22
dehydrochlorination of PVC involves a polar transition state. Also, Starnes et a l29 have reported that dehydrochlorination rates for allylic and tertiary model compounds increase with increasing solvent polarity. The same dehydrochlorin ation rate increase was observed for PVC itself.29
Possible pathways for the thermal degradation of PVC are shown in Scheme
6.21,31 After the initial stage of dehydrochlorination occurs, the resulting polyene chain can either undergo intramolecular cyclization to form unsubstituted aromatic rings (I),21,32 or experience crosslinking (2).21,32 As stated previously, benzene is the major source of combustion and smoke in a PVC fire7 and is formed through pathway 1. Pathway 2 can lead to either the formation of alkyl-substituted aromatics and volatile aliphatics (3),9 or to a thermally stable char (4).8 A fire would spread 10
Unsubstituted Increased Flame Spread Aromatics and Smoke
Linear PVC Alkylaromatics and -HCl Polyene Aliphatics Crosslinked Polyene Chains Char
Scheme 6 Thermal degradation of PVC 21
rapidly in the presence of volatile aliphatics because they bum more efficiently than aromatics7 The formation of char (4) would be desired for fire retardance because it reduces the temperature of the substrate, lowers the oxygen intake necessary for combustion, and can evolve little or no toxic gases or smoke4 11
C. Mechanisms for Benzene Formation
Several studies have reported benzene as the major pyrolysis product from the combustion of PVC7,9’33 and the largest contributor to smoke.7 As seen above in
Scheme 6, benzene forms after the generation of conjugated polyene sequences in the polymer.7 32 The formation of benzene has been proven by O’Mara32 to occur solely by intramolecular cyclization of the polyene sequences. His research revealed that when a mixture of protio PVC and perdeuterio PVC was thermally degraded there were only trace amounts of mixed isotopic benzene formed.32 O’Mara’s results have been verified by other researchers.34"37
Attempts to understand the intramolecular cyclization of polyene sequences to form benzene have resulted in the following two proposed mechanisms (Schemes
7 and 8).7 The first mechanism is the hexatriene mechanism.7,38 It involves the
R,* +
Rj, R2 = -CH2-,-CHCl-, -CH=CH-
Scheme 7 Hexatriene mechanism for formation of benzene7 12
intramolecular cyclization of three conjugated double bonds in the polymer chain,
with the internal double bond being cis while the other two can be either cis or trans.
This process results in a cyclohexadiene structure which can be converted into
benzene by two C-C homolyses.
The second mechanism is the octatetraene mechanism7 This mechanism,
like the hexatriene mechanism, involves an intramolecular 2+2 cyclization,
f ~ \ / ~ \ X J + r 1h c= chr2
R l R2 Rl Rj
Rb R2 = -CH2-, -CHC1-, -CH=CH-
Scheme 8 Octatetraene mechanism for formation of benzene.7
involving four conjugated double bonds. Also, the two internal double bonds must be d s while the other two can be either ds or trans. Benzene is formed as in the hexatriene mechanism from two successive C-C bond cleavages.
Three important observations can be made from the above discussion. First, both mechanisms must involve at least one cis double bond to allow the 13 intramolecular cyclization to occur. Second, Scheme 7 involves the cleavage of the polymer backbone, which would result in a lower molecular weight. This lowering was observed in dilute solution when benzene formation was detected from PVC degradation.39 Scheme 8, however, does not result in polymer backbone cleavage.7
Third, there exists a strong resonance stabilization driving force for the cleavage of the C-C bonds in both mechanisms to form benzene.7
Furthur evidence supports the formation of benzene from polyene sequences in thermally degraded PVC. First, when PVC degrades, HC1 gas is evolved. That
HC1 gas can further accelerate dehydrochlorination of the polymer to give polyene sequences.7,40,41 Also, the rate of benzene formation is enhanced by HC1.38,42
However, Neiman, et al. 42 observed increased benzene evolution when HC1 was removed from the reaction mixture. Based on this evidence, it can be assumed that benzene evolution occurs regardless of the HC1 concentration. The presence of HC1 gas only affects the rate of the benzene evolution. Second, under a constant HC1 concentration, the rate of benzene evolution was autoaccelerating.39,42,43 Third, lida, et al.,33 observed increased benzene evolution at higher temperatures from previously dehydrochlorinated PVC. This result indicates that under mild thermolysis conditions, intermediate structures are formed that can later convert to benzene upon further thermolysis.7 All of the above evidence indicates the presence of polyene intermediates in the formation of benzene. 14
D. Current Smoke Suppression Additives
A significant amount of research has been conducted to reduce the amount of smoke and flame spread during the burning of PVC. Research has been primarily focused on transition metal additives, which are considered to be the most effective smoke suppressants.8,9,44 The purpose of the metal additive is to divert the degradation of PVC to form a stable, non-toxic char (paths 2 and 4, Scheme 6).
Lattimer and Kroenke45 have reported that transition metal additives affect rigid
PVC in three general ways: (1) smoke formation is reduced; (2) char formation is enhanced; (3) evolution of volatile aromatic compounds is reduced.
Although the mechanism of the smoke-suppressing action of transition metal additives is still under study, Starnes and Edelson7 have proposed a Lewis-acid mechanism (Scheme 9) for the role of Mo03 in the thermal degradation of PVC.
The effects of Mo03 on the pyrolysis of PVC7,19>46*48 include: (1) decreased smoke,
(2) decreased flammability at low AH, (3) increased flammability at high AH, (4) decreased benzene yield, (5) accelerated dehydrochlorination, and (6) action only in the condensed phase.
Path 2 allows for the formation of benzene. According to Starnes and
Edelson,7 addition of the Lewis acid, Mo03 (which is subsequently converted into
Mo 0 2C12, a stronger Lewis acid, by HC134) suppresses the formation of benzene 15
FLAME SMOKE /
CIS/TRANS POLYENE CROSSLINKED PVC
ALL-TRANS POLYENE CHAR
Scheme 9 PVC pyrolysis scheme.7
by the following competing acid-catalyzed reactions: (1) the crosslinking of polyene sequences having both d s and trans double bonds (path 3), (2) the formation of trans-only polyene sequences (path 4), and (3) the isomerization of cis-alkene moieties into the more thermodynamically stable trans arrangement (path 5).
Experimental evidence that supports the Lewis-acid-catalysis mechanism for
Mo 0 3 comes from model-compound reactions summarized in Figure 3.34 In reaction
1, Mo 0 2C12 acts as a Lewis acid to convert cis-5-decene into trans-5-decene. a result which supports the cis-trans isomerization in the Lewis-acid theory. Reactions 2-6 are examples of reactions that resulted in degradation products consistent with the occurrence of a Lewis-acid catalysis process. Scheme 10 illustrates the possible
Lewis-acid-catalyzed olefin dimerization and chloroalkylation (Friedel-Crafts) 16
Mo02C1; 100 °c
M003 or tridecene (several isomers) Mo02 or 2 . M02C, ^ 4 * ^ 200 °C
Mo02C12 , 300 °C ^ C26H52 + C 39H78 or M0 O3 ,260 °C + > C6 alkanes + > C8 alkenes
complex mixture Mo02C1: 4. ci 160 °C
ci + CjoHi6 + CioH17 C1 (several isomers M0 O3 , 100 °C or + C 15H2 5 C1 % Mo02C12 , 25 °C \ y \ /
Cl C 1 qH 16 + C 1 qH ! 7 CI (several isomers) Mo02 or % Mo^C, 100 °C
Figure 3 Model-compound reactions with molybdenum-containing Lewis acids.34 17
Mo(VI) Mo(VI) -HC1
Mo(VI)
-HC1 Mo(VI) Mo(VI) i r
Scheme 10 Mo(VI)-catalyzed olefin dimerization and chloroalkylation49
ci Mo(VI) + \ A A ^ \ a /V
-IT + C1
+ isomers
+ isomers ci
Scheme 11 Possible mechanism for Mo(VI)-catalyzed crosslinking in pyrolyzing PVC 49
reactions of 4-chloro-2-pentene.6 The formation of char would occur in analogous reactions of pyrolyzing PVC via the crosslinking of polymer chains (Scheme 11),6,49
The use of MoQ3 as a smoke suppressant in PVC combustion does have one 18 major drawback. However, as the temperature is increased, Mo 0 2C12 can promote a
“cracking” process which forms hydrocarbon fragments (mostly alkanes) from the crosslinked polymer.34,50 Flame spread is then increased by the presence of these aliphatic hydrocarbons.16 “Cracking” has been observed during the pyrolysis of 7- chlorotridecane in the presence of Mo03 at temperatures as low as 200-250 °C.51
Copper additives have been extensively studied as smoke-suppressant additives for PVC.6,49> 52 Kroenke investigated several metal additives as smoke suppressants and concluded that copper additives were the most effective.8 Copper chlorides and oxides have been studied with model compounds and PVC itself by
Starnes and Huang.51,52 Their results suggested that copper additives can function as mild Lewis acids and catalyze the dehydrochlorination and early crosslinking of
PVC through Friedel-Crafts oligomerization and haloalkylation. The copper additives were shown not to accelerate the cis-trans isomerization of
£h>-5-decene.51,52 Therefore, path 5 of Scheme 9 is not a viable mechanistic pathway for copper chlorides and oxides.51,52 However, copper additives do not promote the cationic cracking seen with molybdenum additives.51,52 The above results explain, in part, why copper additives are superior to molybdenum additives for smoke suppression in PVC. 19
E. Reductive Coupling Mechanism
In another effort to account for the smoke-suppressing ability of Mo03 in
PVC, Lattimer and Kroenke have suggested a catalytic reductive coupling mechanism to supplant the Lewis-acid theory.9 This mechanism is illustrated in
Scheme 12, where the metal ligands are not specified. The metal joins allylic or alkyl chloride groups by acting as a coupling agent (Scheme 13)6 This coupling
2 RC1 + 2 M+n ______► R -R + 2 Mn+1C]
2 M"+1C1 + -CH=CH------► -CH=CC1- +HC1 +2M +n
[ RC1 = PVC; M = a metal]
Scheme 12 Metal-catalyzed reductive coupling mechanism.9 would result in “early crosslinking” of the polymer. However, many recent model - compound studies34,37,49,50 have shown that reductive coupling plays only a minor role, if any, in forming crosslinks during the pyrolysis of molybdenum-containing
PVC. The same result was obtained in Starnes and Huang’s investigation of copper chlorides and oxides.51,52
Further investigation of the reductive coupling mechanism has reestablished 20
a) Allylic Site Coupling
[MJ + 2 [MCI]
b) Alkyl Site Coupling
[M] + 2 [MCI]
Scheme 13 Allylic and alkyl site coupling.6
the potential importance of copper additives in PVC.6 As mentioned previously, copper additives do not promote cationic cracking in model compounds. They also are known to promote reductive coupling reactions of organohalides under certain conditions.53'55 Furthermore, during the burning of PVC, high-valent copper is readily reduced to the zero-valent metal.9,45 This metal is produced in a highly active state and should effectively promote a reductive coupling reaction similar to that of the first equation in Scheme 12.
The utilization of the reductive coupling mechanism in PVC smoke suppression is extremely attractive for several reasons.56 First, it would facilitate early crosslinking of the polymer chains.36 This process would result in reduced 21 emissions of flammable volatiles (and smoke) and enhance char formation. Second, subsequent polyene formation could be stopped by crosslinking of the active allylic chloride segments,56 because this type of chloride is needed for polyene propagation.
Third, there would be no acid-catalyzed cracking of the pyrolysis products at high temperatures. 22
F. Low-Valent or Zero-Valent Metal Additives for Smoke Suppression
The overall mechanism for the reductive coupling of PVC still evades
researchers. However, Lattimer and Kroenke45 observed that when metal-containing
PVC was thermally degraded and a high yield of char resulted, there was a
significant yield of low-valent metal in the char. For example, in the thermal
degradation of a PVC-copper(I) sulfide mixture, copper metal and copper(I) oxide
were detected in the char.9 This finding would indicate that the presence of low-
valent and/or zero-valent copper was important in the formation of char.
The use of high-surface-area (HSA) copper powder (purity between 99.995%
and 99.999%) has been studied recently.49,51,52 Huang56 and Jeng6 both concluded
that copper powder that was 99.995% pure had undergone enough surface oxidation
to inhibit the reductive coupling mechanism. Jeng6 later used higher-purity copper
powder (99.999% pure) and showed that reductive coupling did occur for model
compounds of PVC. He also produced HSA copper metal via (1) the reduction of the Cul P(n-Bu)3 complex by lithium naphthalenide in an ethereal solvent (Cu°
slurry)57,58 and (2) the pyrolysis of copper(II) formate, which produced a highly reactive copper mirror.59 Both HSA copper products were effective coupling agents for model compounds. A possible mechanism for the coupling is illustrated in
Scheme 14.6,58 It was first proposed by Ginah, et al.,58 for the homocoupling of 23
RX + 2Cu' RCu + CuX
2RCu RCu + RX RR + CuX
Scheme 14 Homocoupling of alkyl halides via activated zero-valent copper.6’58
alkyl halides via activated zero-valent copper.
The practicality of the aforementioned additives for PVC smoke suppression
is low for several reasons. First, the additives are colored, a feature usually not
desired for commercial polymers. Second, the degradation of copper(II) formate
occurs at 195-200 °C. This temperature may be too close to the processing
temperature of PVC. Third, the need for an anaerobic atmosphere in order to
maintain the activity of HSA copper powder obviously precludes its addition to
PVC. For these reasons, other transition metal complexes (particularly those
containing Cu, Ni, Co, Fe, Mn, Mo, or W) are being explored which may promote
the reductive coupling of model compounds for PVC and of PVC itself. This was the overall research objective of the work described in this dissertation. 24
II. EXPERIMENTAL
A. Instrumentation
1 Gas Chromatographv-Mass Spectroscopy fGC-MSV
Gas chromatography-mass spectroscopy analysis was performed on:
A. A Hewlett-Packard 5890 Series II GC instrument coupled to a Hewlett-
Packard 5971A MSD apparatus. The data were analyzed with Hewlett-Packard
G1034B software for ChemStation running on a Hewlett-Packard Vectra 386/25 microprocessor.
The GC column was an ULTRA-1 crosslinked methyl silicone gum fiised- silica capillary column containing a 95:5 dimethykdiphenylpolysiloxane mixture
(12 m x 0.2 mm ID x 0.33 pm film thickness). The carrier gas was helium. Sample introduction was through a capillary direct interface with split injection.
B. A Hewlett-Packard 5890 GC instrument coupled to a Hewlett-Packard
5970 MSD apparatus. The data were analyzed with Hewlett-Packard 59970C Chem
Station software running on a Hewlett-Packard 9000 Series 300 microprocessor.
The GC column was an HP-1 crosslinked methyl silicone gum fused-silica capillary column containing a 95:5 dimethyl-:diphenylpolysiloxane mixture (12 m x
0.2 mm ID x 0.33 pm film thickness). Helium was the carrier gas. The sample was 25 introduced through a capillary direct interface with split injection.
2. Nuclear Magnetic Resonance (NMRY.
The NMR spectra were acquired using a Cryomagnet (Oxford) 300 MHz spectrometer controlled by a General Electric QE-300 Aquarius Tecmag instrument.
Data processing was performed by an Apple Power Macintosh microprocessor.
Chemical shifts are reported in ppm (5) with Me4Si as an internal reference (5 =
0 .00).
3. Infrared Spectroscopy (IR):
All IR samples were examined as KBr pellets on a Perkin-Elmer 1600 Series
FTIR instrument. The KBr (Fisher) was IR Grade.
4. Melting Point Apparatus:
Melting and decomposition points were determined by using a UniMelt
(Thomas Hoover) Capillary Melting Point Apparatus. Samples were placed in capillary tubes and heated using a silicone oil bath. All melting points were uncorrected.
5. Themogravimetric Analysis (TGAV
The TGA analyses were performed on a Shimadzu TGA-50H apparatus with 26 an appropriate gas flow. The instrument was interfaced with a WIN 486 microprocessor via a Shimadzu Thermal Analyzer TA-501 apparatus, and data processing was performed by TASystem software. 27
B. Model-Compound Reactions
1- General:
Metal additives used for model compound reactions were either purchased or synthesized. They are as follows: copper(II) formate (Cu(02CH)2, Pfaltz & Bauer), nickel(II) formate (Ni(02CH)2, Alfa), iron(II) formate (Fe(02CH)2, synthesized by P.
Kourtesis), cobalt(II) formate (Co(02CH)2, synthesized by W. S. Bryant), molybdenum hexacarbonyl (Mo(CO)6, Strem), tungsten hexacarbonyl (W(CO)6,
Strem), chromium hexacarbonyl (Cr(CO)6, Strem), dicobalt octacarbonyl (Co2(CO)8,
Strem), dimanganese decacarbonyl (Mn2(CO)10, Strem), iron pentacarbonyl
(Fe(CO)5, Aldrich), diiron nonacarbonyl (Fe2(CO)9, Strem), tetrakis(acetonitrile) copper(I) hexafluorophosphate (Cu(NCCH3)4PF6, synthesized by R. D. Pike), iron(II) chloride (FeCl2, Fisher), copper(II) chloride (CuCl2, Aldrich, 99.999%), nickel(H) chloride (NiCl2, Fisher), cobalt(II) chloride (CoCl2, Fisher), iron(III) chloride (FeCl3, Aldrich), molybdenum trioxide (Mo03, ICN Biomedicals, ACS
Reagent Grade), (methylcyclopentadienyl)manganese tricarbonyl
(C5H4CH3Mn(CO)3,Strem), copper powder (Aldrich, 99.999%), and cobalt powder
(Fisher, 99%).
Model compounds also were either synthesized from known procedures or purchased. They are as follows. 3-chloro-l-butene (Aldrich, 98%), benzyl chloride 28
(Aldrich, 99%), cinnamyl chloride (Plfaltz & Bauer, 97%), cinnamyl chloride
(Aldrich, 95%), cinnamyl bromide (Aldrich, 97%), 4-biphenylcarboxylic acid
(Aldrich, 95%), 4-chloromethylbiphenyl (synthesized by W. S. Bryant).
2. Open System ^External Flamel:
A picture of the experimental setup is shown in Figure 4. A three-neck round-bottom flask was equipped with a vertical drying column packed with Drierite
argon out
Drierite > glass wool argon in ^ _ model compound
metal additive
Figure 4 Experimental setup for the open system (external flame).
and glass wool, and then flame-dried. The metal additive was added, and dry argon gas was purged through the system. An external flame from a Bunsen burner was 29 used to heat the metal additive well above its decomposition temperature. Complete
decomposition was evidenced by the formation of a metal mirror and/or complete loss of original color. The reaction flask was then cooled to room temperature, and the model compound was injected using a glass syringe. The residue was
immediately diluted with distilled THF and analyzed by GC-MS.
3. Open System (Oil BathV
A picture of the experimental setup is shown in Figure 5. A pear-shaped single-neck flask was equipped with a Vigreux distilling column capped with an
argon in
argon out
Vigreux distilling column
ofl bath metal additive + model t compound
I> stir bar
Figure 5 Experimental setup for the open system (oil bath) reactions. 30 inlet gas adaptor. Both 4-chloromethylbiphenyl and metal additive (1:1) were then added to the flask. Argon gas was purged throughout the setup. After the argon flow was stopped, the flask was lowered into a silicone oil bath (200 ± 2 °C). After heating under argon for 10 min, the flask was raised and cooled to room temperature. The residue was extracted with a 10-mL portion of methylene chloride/hexadecane (internal standard) stock solution (7000 ng hexadecane/mL methylene chloride) and analyzed by GC-MS.
4. Closed System:
A picture of the experimental setup is shown in Figure 6. The metal additive
metal additive + model compound
Figure 6 Experimental setup for the closed system reactions. and model compound were added to a sealable ampule. All ampules were degassed under vacuum using liquid nitrogen and sealed. The ampules were placed in an oven for 1 h at 200 ± 5 °C. After cooling to room temperature, the ampules were 31 opened, and the residue was dissolved in THF, filtered and analyzed by GC-MS. 32
C. PVC Gel Reactions
1- General:
Poly(vinyl chloride) used for gelation reactions was purchased from Aldrich and had a nominal inherent viscosity of 1.02.
2. Solid State Gel Reactions:
Finely ground metal additive was mixed with PVC (1:10 w/w). The mixture was then heated at 200 + 2 °C for a specified amount of time while purging continuously with dry argon. The residual solid was transferred to a cellulose thimble and extracted with hot THF in a Soxhlet apparatus for 24 h The samples were allowed to air dry and then were vacuum dried overnight at 60 + 2 °C. A picture of the experimental setup is shown in Figure 7.
argon in argon out
metal additive + PVC
stir bar
Figure 7 Experimental setup for the solid state gel reactions. 3. IR Analysis Reactions:
Mixtures of metal additive and PVC were prepared as above and heated under like conditions. The samples, however, were not extracted with THF, but were directly analyzed by IR using KBr pellets.
4. Solvated Gel Reactions:
Poly(vinyl chloride) was dissolved in 100 mL of phenyl ether followed by addition of copper(II) formate (2:1 PVC:copper(II) formate, w/w). The setup was equipped with a condenser and thermometer and purged with dry argon for 30 min
The solution was then stirred and heated at -250 °C until apparent gel formation occurred and stirring could no longer take place. The residue was transferred to a cellulose extraction thimble and extracted with hot THF in a Soxhlet apparatus. 34
D. Synthesis of 3,4-DimethyI-l,5-hexadiene and Its Isomers
In a three-neck round-bottom flask equipped with a reflux condenser, lithium wire (1.6 g, 231 mmol) and naphthalene (3.13 g, 24.4 mmol) were dissolved by stirring in 40 mL of diglyme overnight under dry argon. Using a glass syringe, 3- chloro-l-butene (2.00 g, 22.1 mmol) was injected into the flask, and the resulting mixture was stirred for 1 h and analyzed by GC-MS. This procedure resulted in a
52.9% area yield by GC. Scheme 15 shows the reaction.
diglyme Li 5 + 2Li *♦5” ■+ LiCl \A
Li 5 + LiCl \ / \ 5 + isomers
Scheme 15 Synthesis of 3,4-dimethyl-1,5-hexadiene and its isomers. 35
E. Synthesis of 4-Chloromethylbiphenyl
The synthesis was based on two different published procedures starting with the reduction of 4-biphenylcarboxylic acid to 4-biphenylmethanol60 (Scheme 16).
The alcohol was then converted into 4-chloromethylbiphenyl61 (Scheme 17). All reaction flasks were oven-dried overnight at 110 °C and later flame-dried when needed. The THF was distilled over sodium metal/benzophenone under nitrogen.
4-Biphenyl-carboxylic acid (12.00 g, 60.54 mmol) was dissolved in 80 mL of distilled THF. In a separate flask, 100 mL of borane-THF complex (1.0 M,
Aldrich) was added to 70 mL of distilled THF under dry nitrogen. The THF solution containing 4-biphenylcarboxylic acid was added dropwise using a steel
THF trimerization 3RCOOH + 3BH3 ------o[r^^n2UDUj t.o n 2 R = p-PhC6H4- R c h 2 O 6H20 ------► 3RCH2-OH + 3B(OH)3 0 0 hydrolysis 1 1 o - V R -C U ° 'C -R h2 h2
Scheme 16 Conversion of 4-biphenylcarboxylic acid into the corresponding alcohol. 36 ...... ROH © © A ------^ M^NCHO + SOCl2 ------►- [Me2N=CHCl][Cl] + S02 T
© © [Me2N=CHOR][Cl] * ------M^NCHO + RCl
R = ^-PhC6H4-CH2-
Scheme 17 Conversion of 4-biphenylmethanol into 4-chloromethylbiphenyl.
transfer needle, and the resulting mixture was stirred for 30 min. The solution was then neutralized by adding 10 mL of a solution prepared from 100 mL of 2 N H2S04 and 100 mL of ethylene glycol (Aldrich, 99+%). The neutralized solution was transferred to a separatory funnel. After the addition of deionized water (75 mL), the mixture was extracted with several portions of ether (150 mL x 3), and the combined extracts were washed repeatedly with deionized water (150 mL x 3). The ether solution was dried with magnesium sulfate and concentrated under vacuum using a Rotovap apparatus. The resulting solid was dried under vacuum overnight.
This procedure resulted in 10.40 g of the alcohol (93.2% yield). Spectral analysis was as follows: MS (Figure 8), m/z 184 (M); purity > 99% by GC.
The alcohol was dissolved in 150 mL ofN,N-dimethylformamide (DMF)
(Mallinckrodt, Analytical Reagent Grade). In a separate flask immersed in an ice bath, 5.0 mL (67.8 mmol) of thionyl chloride (Aldrich, 99+%) was added slowly 37
C Z Z ) 5c an Z . o 5 3 m m . of DRT5 : W5B2 . D [55 400yy 1 | 04 -i I 'J 30000 ■] j 20000 5 4 1i 1i • Cj i 0 01 j ij 4
y 23 40 60 6 y 1 m 20 140 160 180 fi a s s / L h a r g c
TiC of QftTH : 14'6 Bd . Q b- . 0fc. - ^ + •j 1-i to Jj 0 l . 5 E+5 i
1 1.0E+5] JlL ^ cn 5 . 0 h> 4 q 3 Q.0E>0J=’ 4.0 6.0 y. W Time (m i n. )
Figure 8 GC-MS results for 4-chloromethylbiphenyl. 38
with stirring to 5.5 mL of DMF. This addition was done while purging with dry
argon. The solution of DMF and alcohol was then added dropwise under argon, with stirring, to the reaction flask using a steel transfer needle. The resulting mixture was heated at 85-90 °C for 30 min. After cooling to room temperature, the solution was transferred to a separatory funnel and worked up with deionized water and ether as explained above. Following the evaporation of the solvent, the resulting solid residue was dried overnight under vacuum. This resulted in 7.53 g of 4- chloromethylbiphenyl (65.8% yield). A melting point determination yielded a value of 65-67 °C (literature value: 68-70 °C)62. Spectral analysis was as follows: !H NMR
(300 MHZ, CDCl3)(Figure 9), 6 7.63-7.50 (multiplet, 4H, aromatic H), 7.47-7.38
(multiplet, 4H, aromatic H), 7.38-7.25 (multiplet, 1H, aromatic H), and 4.61 ppm
(singlet, 2H, -CH2-C1); 13C NMR (75 MHz, CDCl3)(Figure 10), 5 141.74, 140.85,
136.78, 129.02, 128.90, 127.60, 127.50, 127.28(aromatic carbons), and 45.96 ppm
( -CH2C1); MS (Figure 11), m/z 202, 204 (M +), 167 (M-35 and M-37); purity > 99% by *H NMR and GC. Nucleus hydrogen Obs. freq 300.£200050 Spec, width 3600.0 Points Acq. 8192 ooooa:n_Q_jtn<_j(jou_Q «— o Points FT Q3 8192 LOLOO —r TCO* * O O J ■*— *— N O ( C 'T O C r— «— uU> ao raw > U u u o o— w (l s. » •*-> a» . -s w ’ O (fl w — ro ro o o>§ ut ^ C ^ t? u § > ro ro c c c c ua ^agw uiaw _ O o O O CD 0 3 CM o o o o O O - <° o-—Eu
r— o _ .CM <£) ao E CL CL
Figure 9
'H NMR spectrum of 4-chloromethylbiphenyl. 39 40 _o
o o _f\j o LO o o D ro o Q. E •C O O o L O . rsj CNJ cE o Cl ln o OsJ o r i Ao v c w o ow o jo ‘ Um ■gre uoocdvd^-tj-o . o m n o ono o w «»— o ■tjcn^r_uCirs/2 o n u r ^ c j.—*— loldpocdcoi —o c o » —o a t o
O ^ v 03 cr^of-a-I crJS S g 3to 5: r— O O) a, oj cr- O g W.S.E re S > ° u ^ o - c r - > C « J3 -Q 0 . 0 O U U U > a; re 43 OCO Or= ™ Z 0 coa~o_cocoaca_ 0 _ ]c o < _ j c o u -0
too
.
o o NMR 4-chloromethylbiphenyl. spectrum C of
coo
cr.
E Q. a. 41
( 4 4 J bc ■!n 4 , 6 2 -i rn i n . o f D A I R : 145BCLJ . D i
ti . c-j r-ic * .j . Oj 1 . “ 1 ' J> 1 . 0 E+S H 1 5 3 ------■! ^ •j . id L + 4 1 o o 52 , cn
Id . Id E + 0 “*—r V fW i-* I ...... I JjL 5 0 1 0 0 M a s s / L h a r c| e
IC of DhTh :W5BCL3. D
9 3.0E+5: (A 2 ,0 E ^5 :
cn 1 . 0 E + 5
0 . 0E+0 -i r -i—i— r~ d I 0 Time (m i n . J
Figure 11 GC-MS results for 4-chloromethylbiphenyl. 42
F. Synthesis of 4,4'-Diphenylbibenzyl
The synthesis (Scheme 18) was modeled after a method for synthesizing 2,2’- diphenylbibenzyl .63 While stirring, 4-chloromethylbiphenyl (2.44 g, 12.0 mmol) was added to dry acetonitrile (distilled from calcium hydride) and followed by the addition of sodium metal (1.01 g) under argon. The resulting mixture was allowed to reflux overnight under nitrogen and then was quenched with 50 mL of ethanol.
Upon addition of 200 mL of water, the acetonitrile layer was separated, washed with deionized water (3 x 50 mL), and dried over magnesium sulfate. After concentraton of the solution using a Rotovap apparatus, the resulting precipitate was recrystallized from hexanes. The final product was dried overnight under vacuum.
This procedure resulted in 0.40 g of 4,4’-diphenylbibenzyl (9.9% yield). A melting
CH2CI + 2N a
+ 2 NaCl
Scheme 18 Synthesis of 4,4’-diphenylbibenzyl. 43 point determination resulted in a melting point range of 200-201 °C (mp: 202 ,64
20565 °C). Spectral analysis was as follows: 'H NMR (300 MHZ, 1 ,1,2,2- tetrachloroethane-d,)( Figure 12), 8 7.21-7.57(complex multiplet, 18H, aromatic H) and 2.99 ppm (singlet, 4H, -CH2->; 13C NMR (75 MHZ, 1,1,2, 2-tetrachloroethane- d2)(Figure 13) 8 141.08, 140.93, 138.86, 128.93, 128.77, 127.10, 127.07, 126.97
(aromatic C’s) and 37.20 ppm (-CH:-); MS (Figure 14), m/z 334 (M+), 167 (M+-
167); purity > 99% by ’H NMR and GC. L 44
f1 0 I—r h r i h CO ^ ! c uS | i-~ Si i h “ J s , “t\f ] b i *-
- CJ
IS I CO I ^PU no s3_
_
o o Q> Figure 12 oln © o c§ w _ o o XTLO^cr>un t t O o — 01 S a j o O 1) w UNN o ^ oojcvj ^ in© (Nlroc- ■nOOCIO) O © «CnJXK%I >,010 O .O J C O '.j: in • .cmm coaotDtocoTrco.—*-— — ^ o do v> c\jo »— ?co *H NMR spectrum NMR spectrum 4,4'-diphenylbibenzyl. of *H -C ■ ^ ^ CT c - a> ? ^ cr^yi—aa: > ■*£ aD-- S ot ; <-> E = •£: . w w uiu a;-u:Q «EU - u w 1),£,E ra U in O' _c « ^ d-q clo o u uU > v n oi yen 0 = « Z OC-O Q. O. LOLO C£ C l O-J t o < —J L/CU_ O VD
00 _ o 45
o
O ;co
aoo Figure 13
O LO O o oa fM o I*£ C O O o r - L T J o "UCTJ cijo-^-Tr.< \J, CM o JJ'V. olrocooooo;r*-, m *-> r*~ -g.ommooo o O O i^Csl a r— C - s nmo«3ioooo ao « 9 2 4,4'-diphenylbibenzyl. NMR spectrum C of UNfVJr-l-r-r-(NJlOmr-u c o *-o ZDcn o o
- c r Q 3~ c c > o •^r £ a a. 46 ( 133) Scan 15.323 min. of DATR:BILLI 3 .D 5 000: : i s ? 4 000 : i; 34 / : 1 1 5 207 448 00 0 I *\ J X i . ■ -y |- ■ .1 ■ 1 , t ■ -f- 0 •in----1---- 1----r-1-!----1----» ' X - 4 **. i , ? * > i " »— r ■ i i ■! ■i'-i t i"' v^ 100 150 200 2 50 300 359 4 0Q 450 M a 3 s / C h a r g a TIC of DfiTfi:BILL13.D 10000 o' 6 0 0 0 : 4 000: 2 000 i 0 ^ —?~~i—? -p ~i1 t r i i—>■ .1 8 10 14 16 18 2 w Time (m i n. ) TIC of DATA:8 ILL I 3.0 I integration peaks found P a g e Figure 14 GC-MS results for 4,4-diphenylbibenzyl. 47 G. Synthesis of Cobalt(II) Formate The synthesis (Scheme 19) was based on a published procedure for the synthesis of iron(H) formate 66 To a flask containing 240 mL of distilled water, cobalt powder (6.00 g, 102 mg-atom) (99%) was added and followed by 20 mL of 45% formic acid. The mixture was stirred and heated under nitrogen until hydrogen gas evolution ceased (2 h) then filtered by vacuum into a flask containing 10 mL of 45% formic acid. The contents of the flask were heated and stirred until crystallization occurred (2 h); then they were then cooled in an ice bath and air-dried overnight. The crystals were recrystallized from 95% ethanol, and the resulting product was dried overnight under vacuum. This procedure resulted in 3.01 g of cobalt formate (16% yield). The decomposition temperature of the product, as determined by TGA, was 277 + 2 °C. Spectral analysis was as follows: IR (neat)(Figure 15) 3302.6, 3220.5, 1574.0, 1394.9, 1374.4, 1353.9, and 764.1 cm 1. o I Co + 2 h—c —oh - ► Co(02CH)2 + H2 I Scheme 19 Synthesis of cobalt formate. 40.52-J Figure 15 1R spectrum for cobalt(II) formate. 48 49 III. RESULTS AND DISCUSSION A. Model Compounds A series of model compounds was chosen in order to investigate the possibility of the reductive coupling mechanism occurring during the thermal degradation of PVC. The model compounds were chosen to represent “active” sites in PVC suitable for reductive coupling. Model compounds containing allylic chloride sites were chosen not only for their high reactivity but also for their resemblance to the allylic sites in PVC. Allylic moieties are known to be present in both the virgin and the thermally degraded polymer .6 22 Benzylic-chloride- containing compounds were used as models because their chemical properties are similar to those of allylic chlorides and because the ones selected had higher molecular weights. Higher-molecular-weight compounds were needed for reactions run at temperatures near or above the decomposition temperatures of the metal additives. Unless otherwise noted, the transition metal additives used for all model compound pyrolysis experiments were Mn 2(CO)10, Co 2(CO)8, Mo(CO)6, Fe(CO)5, Fe2(CO)9, Cr(CO)6, W(CO)6, (C5H4CH3)Mn(CO)3, Cu(NCCH 3)4PF6, Cu(0 2CH)2, and copper powder. Although cobalt(II) formate was synthesized, its decomposition 50 temperature was too high (277 + 1 °C as determined by TGA) for the reaction conditions. Chromium hexacarbonyl was not used since it is explosive above 210 °C 67 1. 3-Chloro-l -butene: In a previous paper, 3-chloro-1-butene was reported to couple in the presence of high-surface-area copper produced from the reduction of CuI PBu 3 by lithium naphthalenide.6 The ability of this compound to couple was the basis for its choice as a model compound of allylic chloride structures in PVC. The homocoupled products were first synthesized by using lithium metal in diglyme as the coupling agent. This was done to determine the retention times for the GC column (Figure 16). The seven labeled peaks in the chromatogram represent the seven homocoupled product isomers (MW =110 g/mol). Possible isomers are shown in Figure 17. The relative stability of the isomers can be determined from the fragmentation patterns of each product. The first four peaks (0.659, 0.673, 0.751, and 0.794 min) are the least stable because of the absence of the parent ion peak (m/z 110). The heaviest ion peak is 95 m/z, which represents the loss of a methyl group. The MS peak at 81 m/z represents a rearrangement and loss of an ethylene end group (29 m/z). The last three peaks (0.971, 1.007, and 1.057 min) are more 51 Abundance TIC: WSB1 450000 - 400000 - 350000 - 300000 - 250000 200000 - 150000 - 100000 - 50000 - Time -> [Abundance Scan 37 (0.659 WSBl. D 140000 - 120000 - 100000 - 80000 - 60000 - 40000 - 81 95 20000 - 18 109 14 H 5 3 279 294 M/Z - > Figure 16a GC-MS data for the coupled products from 3-chloro-1-butene. 52 Abundance WSB1.D 450000 400000 350000 300000 250000 - 200000 150000 - 100000 - 50000 - Time -> 0 . 60 1 . 00 1 . 60 Abundance Scan : WSB1.D 160000 140000 - 1 2 0 0 0 0 - 100000 - 80000 60000 - 40000 - 81 39 95 20000 - 18 109 177 264 286300 - > Figure 16b GC-MS data for the coupled products from 3-chloro-1-butene. 53 Abundance TIC: WSB1.D 450000 400000 350000 300000 250000 200000 - 150000 - 100000 - 50000 Time -> 0.60 0 .80 1 . 00 1.20 1.40 1 . 600.40 Abundance Scan 100 (0.751 min): WSB1.D 100000 - 90000 80000 70000 60000 50000 40000 30000 95 20000 39 81 10000 - 13 110 136 162177 205 226240 259273 291 - > 20 40 60 80 100 120 140 160 180 200 220 240 260 280 300 Figure 16c GC-MS data for the coupled products from 3-ch loro-1-butene. 54 Abundance TIC: WSB1.D 450000 - 400000 350000 - 300000 - 250000 - 200000 - 150000 - 100000 - 50000 - T im e - > 0 .40 0 . 60 0 .80 1 . 00 1.20 1. 40 1 . 60 Abundance Scan 106 (0.794 min): WSB1.D 140000 - 120000 - 100000 - 80000 - 60000 - 40000 - 95 39 81 20000 - 18 110 145 16.4> 9 139 208 2 4 5 2 6 8 2 829 4 Figure 16d GC-MS data for the coupled products from 3-chloro-1-butene. 55 Abundance TIC; WSB1. 220000 - . 200000 180000 - 160000 - 140000 - 120000 100000 - 80000 60000 A Time -> 1 . 00 Abundance Scan 131 (0.971 min): WSB1.D 7000 6000 55 5000 4000 18 3000 - 2000 • 81 1000 - 95 110 138 165C67 02 2238 254 28297 jiU il.iiiw ii U i.LiiiiJlW lJ il.iW i^n tiil.iliil. i^lliifiL llh(H lllljUli il| M/Z 50 100 150 200 250 300 Figure 16e GC-MS data for the coupled products from 3-chloro-l-butene. 56 Abundance. TIC: WSB1 450000 - 00000 - 50000 00000 50000 00000 150000 - 100000 - 50000 - Time -> 1 . 00 Abundance Scan 13 6 . 007 min) : WSB1 12000 - 10000 - 8000 - 28 6000 - 4000 - 18 81 2000 - 1|° 134149 17 0312 &&2 2 9 5 liil.Lj ■ in.h'jllti III iih n-yuiai^l I f M/ Z - > Figure 16f GC-MS data for the coupled products from 3-chloro-l-butene. 57 Abundance TIC: WSB1.D 220000 - 2G0000 - 180000 - 160000 - 140000 - 120000 - 100000 - 30000 - 60000 A Time -> 0 . 600.40 0 .80 1 . 00 1-20 1.40 1. 60 Abundance iScan 143 (1.057 min): WSB1.D 6000 - 55 5000 18 4000 3000 2000 31 1000 95 110 143157 175138 218 236 263 2% 0 ,M/Z - > Figure 16g GC-MS data for the coupled products from 3-chloro-l-butene. 58 1,2 y y X v \ V\ one meso isomer cis- trans- and one racemic isomer pair 6 7 \ = / cis-trans trans-trans cis-cis Figure 17 Isomers of 3-chloro-l-butene reductive coupling product. stable than the four earlier peaks because the parent ion peak 110 m/z can be clearly seen. Further peak assignments can be made based on the stability of the radical intermediate(s) involved in the coupling (these intermediates may be incipient rather than actual). The two possible major resonance structures of the trans radical intermediates formed during reductive coupling are shown in Scheme 20. 59 ci -Cl- V \/\ \ y \ Scheme 20 Formation of radical intermediate from 3-chloro-l-butene. '\= / -— - V _/ V w Figure 18 Major contributing structures of the cjs radical. Also shown (Figure 18) are the major contributing structures of the isomeric ds radical. The reactive forms involved in the creation of structures 1 and 2 in Figure 13 are resonance contributors X and W. The homocoupling of these forms results in doubly branched diene isomers. Structures 3 and 4 in Figure 13 are formed from the coupling of radical intermediates X and W with the corresponding cis-trans isomeric radical (V or Y) to give mono-branched dienes. Structures 5-7, which are straight- chained dienes, represent the three possible coupling products of the less stable primary radicals V and Y. Peaks E-G of Figure 12 can be assigned to structures 5- 7. Peak F is assigned to structure 5. Since there are two ways of forming the cis- 60 trans isomer and only one way to form the cis-cis and trans-trans isomers, the peak for the cis-trans isomer should be approximately two times as large by area. To determine whether the model compound would couple in the absence of "coupling additives", 3-chloro-l-butene was injected into the GC-MS apparatus as a control. None of the coupled products was seen. Positive results for reductive coupling from reduced metal additives are listed in Table 1. All metal additives were tested in the open (external flame) and Table 1 Coupling of 3-chloro-l-butene (1) with metal additives. Metal Additive Integration of Peak [mmol/(mmol 1)] System (as determined by GC) j Mn 2(CO)10 [0.14] open (external flame) 7.2 j Co 2(CO)8 [0.22] open (external flame) 5.5 j Mo(CO )6 [0.25] open (external flame) 6.0 j C5H4CH3Mn(CO )3 [0.50] sealed 1.3 Fe(CO)5 [0.50] sealed 3.2 Cr(CO)6 [0.29] sealed trace j the closed (sealed tube) systems. Copper(II) formate and copper powder were not used in either experimental setup. 61 When transition metal carbonyls are decomposed, they are converted into carbon monoxide gas and zero-valent metal. Some of the metals produced shiny mirrors on the glass surface of the flask [in the case of Co2(CO\, Mo(CO)6, and Fe2(CO)9]. Previous experiments have shown that a highly reactive copper mirror formed from the decomposition of copper(II) formate couples 3-chloro-l-butene .6 The mechanism for the reductive coupling of the metal may be similar to that of Scheme 14 proposed by Ginah et al .58 for zero-valent copper. According to their research, at temperatures above 0 °C the organocopper compounds formed have low stability, and coupling may occur via steps 2 and 3 .58 Coupling has also occurred when using zero-valent copper formed from the reduction of Cul PBu 3 by lithium naphthalenide.6 Although further research is needed to confirm the mechanism for the reductive coupling by the transition metal carbonyls, this mechanism certainly may involve zero-valent transition metals. In summary, the results of the model-compound study of transition metals in the presence of 3-chloro-l-butene indicated that reductive coupling did occur. These results support those obtained in Jeng’s study of the reductive coupling of allylic chlorides via highly activated zero-valent copper metal .6 Since allylic chloride structural defects are known to occur in both virgin and thermally degraded PVC and are now shown to couple reductively, this coupling reaction could lead to extensive crosslinking in PVC. The addition of other low-valent transition metals to 62 the list of possible reductive coupling agents allows a broader range of compounds to be studied. 2. Benzvl Chloride: Benzyl chloride (Aldrich, 99%) was chosen as a model compound because the chemical reactivity of benzylic chlorides is similar to that of allylic chlorides. Benzyl chloride also has a high boiling point (179 °C ).68 This allows experiments to be conducted at temperatures higher than those used with 3-chloro-1 -butene. The GC retention time for the coupled product of benzyl chloride was determined by injection of bibenzyl (Aldrich, 99%) into the GC-MS instrument. Figure 19 shows the GC-MS results. The mass spectrum shows a parent ion peak of 182 m/z. The base peak is the stable 91 m/z ion which represents half of the coupled product. The reactions with benzyl chloride were conducted by using the closed system (sealed tube), as described in the Experimental section. The mole ratio of metal additive to model compound was 1:10, and the GC-MS analyses were performed after diluting the residue, in the sealed tubes with THF. The GC chromatogram of the pyrolysis products from benzyl chloride and the metal additives consisted primarily of three groups of peaks. A representative 63 Abundance TIC: WSB6. D 7 . 48 2 e+07 1 .8 e+0 7 1 . 6e+07 1 . 4e+Q7 1 . 2e+07 le+07 8000000 6000000 - 4 0 0 0 0 0 0 : 2000000 - W -ffisD 1 1 .41 0 i . i I i i i ' T | ! T 1 ' '1 1 , . , . 1 Time -> 2.00 4.00 6.00 8.00 10.00 12.00 Abundance rain): WSB6. D 6000000 - 5000000 - 4000000 - 182 3000000 - 65 2000000 - 1000000 - 51 77 104 165 27 15 115 128 139 1^2 195 - > 20 40 60 80 100 120 140 160 180 200 Figure 19 GC-MS results for bibenzyl. 64 chromatogram is shown in Figure 20 (for Mo(CO )6 and benzyl chloride). The three groups can be categorized by retention times: A) 1-3.5 min, B) 5.5-9 min, and C) 11-13.5 min. Table 2 lists area percentages of these groups for each metal additive. Group A contains low-molecular-weight molecules such as toluene (~ 1.0 - 1.2 min) Table 2 GC area percentages for pyrolysis products from benzyl chloride and metal additives .3 A B C Metal Additive (1-3.5 min) (5.5-9 min) (11-13.5 min) ) Fe(CO)5 35.7 40.6 23.7 Fe2(CO)9 25.9 24.0 50.1 | Co 2(CO)8 25.0 25.1 49.9 W(CO)6 52.5 20.5 27.0 Mo(CO )6 21.5 43.6 34.9 [ Mn 2(CO)10 25.7 31.0 43.3 j (C5H4CH3)Mn(CO)3b 81.2 3.0 o j Cu(NCCH3)4PF6 40.2 7.8 51.9 j Cu(0 2CH)2 41.3 43.1 5.0 copper powder 32.0 68.0 a Percentages do not add up to 100% because some chromatograms had peaks that were not in any of the three groups. b Average values from two runs. Average deviation was ±1.8%. 65 Abundance TIC: WSB3.D 80 240000 - 220000 200000 - 180000 - 160000 - 12 11. 07 140000 - 120000 - 100000 - 80000 11.38 60000 40000 - 65 7 . 45 10 . 3!) I ll 4 5 20000 - Time -> 4 . 00 6 . 00 10 . 00 12 . 00 14 . 00 Figure 20 GC chromatogram of pyrolysis products from Mo(CO)6 and benzyl chloride. 66 and the starting material itself (~ 3.5 - 3.8 min). Group B contains dimers formed by Friedel-Crafts chloroalkylation and reductive coupling dimerization. Group C contains heavy products resulting from Lewis-acid catalyzed oligomerization. The grouping of the compounds is by retention time only and does not imply that the peaks in each group have similar mass spectra. In fact, most of the reactions gave product mixtures with unique and complex spectra. The chromatogram for the products obtained with copper powder (Figure 21) could not be easily divided into the groups in Table 2. There was no apparent division of peaks above 3.14 min. The starting material peak was seen for only two additives in Table 2, (methylcyclopentadienyl)manganese tricarbonyl and tetrakis(acetonitrile)copper(I) hexafluorophosphate (Figures 22 and 23 respectively). As can be seen from the chromatograms, the yields for the starting material were large, 82.6 and 79.5 area percent (two separate runs) for (methylcyclopentadienyl)manganese tricarbonyl and 37.9 area percent for tetrakis(acetonitrile)copper(I) hexafluorophosphate. These results would indicate that the low-valent metal centers were not as reactive as the zero-valent metals formed by decomposition. Also, (methylcyclopentadienyl)- manganese tricarbonyl did not produce any heavy products, giving further evidence for its passiveness in reacting with benzyl chloride when it is decomposed. It is interesting to note that the only positive test for the coupled product was obtained when copper(II) formate was used. Figure 24 shows the results obtained 67 Abundance TIC: WSB1.D 200000 i 190000 - 180000 - 170000 i 160000 - 150000 140000 - 130000 - 1 2 0 0 0 0 -1 n o o o o A 100000 A 90000 J 80000 i 70000 - 60000 - 50000 - 40000 30000 20000 - 10000 Time -> 2.00 4 . 00 10 . 00 12 . 00 14 . 00 Figure 21 GC chromatogram of pyrolysis products from copper powder and benzyl chloride. 68 Abundance i TIC: WSB10.D 1 . 3e+07 3 31 1 . 2e+07 ^ _ starting material 1 . le+07 le + 07 - 9000000 - 8000000 - 7000000 - 6000000 - 5000000 - 4000000 - 5 . 59 3000000 - ^ — unknown 2000000 - 1000000 - n 1 Hi u« —:: | : : I ‘ :-----1 : :, , |-----;-----,-----,-----;----- ■ 1-----| ;-----,-----, , |----- ! ,----- ‘ ,----- 1 , 1 , * | i :: • • :■ ■ ■ | ! •r- iTime - > 2.00______4 . 00______6 . 00______8.00______10.00 12.00 14 . 00 Figure 22 GC chromatogram of pyrolysis products from (methylcyclopentadienyl)manganese tricarbonyl and benzyl chloride. 69 Abundance TIC: WSB13.D 31 8500000 - starting material 8000000 - 7500000 - 7000000 - 6500000 - 6000000 5500000 5000000 - 4500000 - 4000000 - 3500000 3000000 2500000 - 2000000 - 1500000 - 1000000 - 8.90 13 . 15 500000 - JLii Time - > 2 . 00 4 . 00 6.00 8 . 00 10 . 00 12.00 14 . 00 Figure 23 GC chromatogram of pyrolysis products from tetrakis(acetonitrile)copper(I) hexafluorophosphate and benzyl chloride. 70 Abundance TIC: WSB7 6 Jf725.79 34 140000 - 120000 - 100000 - 80000 - 60000 - 40000 - 40 20000 - Time -> 5.50 7 . 50 Abundance Scan 946 (7.338 min): WSB7. D 90000 - 80000 70000 60000 50000 40000 30000 182 20000 65 10000 28 3914 104 152 1?5 178 M/Z-> 20 6040 100 120 140 160 180 Figure 24 GC-MS results for pyrolysis products from copper(II) formate and benzyl chloride. 71 by GC-MS, which gave an area percent of 16.7 for the coupled product. Besides the formation of the reductively coupled product, Friedel-Crafts chloroalkylation products (Group 2) were produced. However, there was little Lewis-acid catalyzed oligomerization or other heavy product (Group 3) formation. 3. Cinnamvl Chloride: To study further the reductive coupling mechanism of PVC, a model compound which contained an allylic chloride group and had a high boiling point was needed. Cinnamyl chloride was a suitable choice. Its boiling point (125-6 °C, 22 mm Hg)69 allowed reactions to be conducted at temperatures closer to the decomposition temperatures of the metal additi ves. When determined by GC-MS, the purities of neat cinnamyl chloride (Aldrich, 95%, and Pfaltz and Bauer, 97%) and cinnamyl bromide (Aldrich, 97%) were not as expected. The chromatograms showed several peaks for all three compounds (Figures 25, 26, and 27 respectively)3 and gave purities of 92.0, 81.0, and 51.3%, respectively. Also, the JH NMR spectrum obtained for cinnamyl chloride (97%) (Figure 28) does not account for all the impurities seen in the corresponding gas chromatogram. This evidence indicates that a reaction occurred during the GC a The GC chromatograms in Figures 25 and 26 showed different retention times for the two isomers because the injection port temperatures were different. 72 Abundance TIC: WSBM.D 67 2200000 - 2000000 - 1800000 - 1600000 - trans-cinnamvl chloride 1400000 - 1 2 0 0 0 0 0 - 1000000 - 800000 - 600000 - 400000 - cis-cinnamyl chloiide 10 .01 200000 - 14 10. 34 : 00 47 Time -> 5 . 00 6 . 00 7 . 00 8 . 00 9 . 00 10.00 11.00 12.00 Figure 25 GC chromatogram of “95%” cinnamyl chloride. 73 Abundance TIC: PETERI. D 2.04 1100000 1050000 1000000 950000 900000 850000 800000 750000 700000 650000 - trans-cmnamvl chloride 600000 - 550000 - 500000 450000 400000 350000 300000 250000 200000 cis-cinnamyl chloride 150000 100000 : .63 50000 - Time •-> 1. 00 2 . 00 4.00 5.00 9 . 00 Figure 26 GC chromatogram of “97%” cinnamyl chloride. 74 Abundance TIC: WSB2.D 1600000 6.60 1500000 - 1400000 - 1300000 - 1200000 - trans-cinnamyl bromide nooooo - 1000000 900000 800000 - 700000 600000 500000 400000 - cis-cinnamyi bromide 300000 - 6.05 200000 100000 - 8 . 00 10 . 00 12 . 00 14 . 00 16 . 00 Figure 27 GC chromatogram of “97%” cinnamyl bromide. 75 CL CL Q. O CN oo ro o CN m C\J 'Ooo U~\ CN Q. CU o CO CN O h Figure 28 Figure OX H NMR spectrum of 97% NMR chloride. 97% cinnamyl spectrum of H o ox _o 76 analysis (presumably in the injection port) that reduced the purity of the compounds. Attempts to eliminate this problem by altering the GC analysis conditions failed, even when the injection port temperature was lowered to 100 °C. Also, neat 97% cinnamyl chloride sometimes gave a peak that corresponded to the coupled product (m/z 234) (Figure 29). However, this result was not available until after the closed- system-model compound reactions were run with 95% cinnamyl chloride. To determine the GC retention times and mass spectra for the coupled product and its isomers, the lithium coupling reaction described in the Experimental section was performed using cinnamyl chloride (Aldrich, 95%). Figures 30a-d show the GC-MS results obtained for the product mixture. Possible MS fragmentation routes for the coupled product isomers are shown in Scheme 21. The reaction appears to be extremely selective for one of the three possible isomers. Gas chromatographic yields for the three isomer peaks were 46.6% at 8.47 min, 1.4% at 9.06 min, and 22.0 % at 9.41 min. Since the trans-trans isomer is the most thermodynamically stable of the three isomers, the peak at 9.41 minutes can be tentatively assigned to this isomer. All reactions with additives were done in a closed system as explained in the Experimental section (the amounts of metal additive and model compound were unmeasured because of the small scale of the reaction). The model compound was heated at 200 °C for 1 h as a control experiment to see if coupling occurred without 77 Abundance TIC: WSB3.D 600000 - 500000 400000 300000 - 200000 8 . 19 100000 - 4.51 6 . 10 7 . 77 Abundance Scan 1070 (7.181 min): WSB3.D l i 7 50000 - 45000 - 40000 35000 30000 11 25000 20000 15000 91 10000 - 28 5000 - 14 234 152 20 40 60 80 100 120 140 160 180 200 220 Figure 29 Partial GC chromatogram of “97% cinnamyl chloride. 78 TIC: WSB2.D ■17 90 0 0 0 - 9 .0 6 8 0 0 0 0 - 7 0 0 0 0 60000 50000 - 9 .02 4 0 0 0 0 - 30000 - . 08 3.0 . 06 20000 - . 14 100QQ - T im e -> 8 . 00 9 . 00 10 . 00 1 0 . 5 0 b Abundance S c a n 1265 ( 8 . 46S a m } : w $ S 2.0 k b u n d a n c a S c a n 1354 (9 .059 a m i : » S 82.0 S I 55 77 65 77 14 37 -1.1 .I j L 41LS 2 165 169 202 215 , ll L. i / Z 100 120 L 4Q 160 I S O 200 2 20 i/z -> 130 140 160 I S O 200 220 A s u n d a n c i S c a n .406 (9 d 1.7 J 30000 300000 250000 - 115 200000• 150000 91 100000 30000 * 51 «5 77 o J _ 14 27 1 i i J . i l L ! L4 115 2 1 S3 18 9 202 215 [ V n -> 20 40 60 SO 100 120 140 160 ISO 200 220 Figure 30 GC-MS data for products from the lithium coupling reaction of cinnamyl chloride. 79 -2e m/z 234 1,3 H shift + - ' V \ M / m/z 117 m/z 91 Scheme 21 Possible fragmentation routes for cinnamyl chloride coupled product. the metal additives. No coupling product was observed. Characterization of the products resulting from the pyrolysis of cinnamyl chloride and the metal additives proved to be very difficult. This difficulty was due to the complexity of the chromatograms obtained for all reaction mixtures. General 80 screening for the presence of the coupled product could not be achieved due to the later finding that the chromatogram of neat 97% cinnamyl chloride sometimes indicated the presence of that product. Therefore, the pyrolysis study of cinnamyl chloride and transition metal additives was abandoned. 4. 4-Chloromethvlbiphenvl: Since the reactivity of benzylic chlorides is similar to that of allylic chlorides, 4-chloromethylbiphenyl was chosen as a model compound for PVC. Its high molecular weight allowed for further study of the reductive coupling mechanism at temperatures closer to the decomposition temperatures of the metal additives. Unfortunately, 4-chloromethylbiphenyl proved to be as difficult to work with as cinnamyl chloride. Attempts to make the coupled product, 4,4'-diphenylbibenzyl, via the lithium reaction used for 3-chloro-l-butene were unsuccessful. Therefore, the coupled product was synthesized via the Wurtz reaction by using sodium metal. Once the coupled product was isolated, it was injected into the GC-MS apparatus to determine the retention time and MS fragmentation pattern (Figure 10). A possible fragmentation pathway is shown in Scheme 22. 81 + e" \ r ^ \ /“8r8r\ /-\ / 2 e m/z 234 - H,C \ / \ / + CH, \ r \ / / m/z 167 Scheme 22 Possible fragmentation route for 4,4'-diphenylbibenzyl. The method used to test for the reductive coupling ability of the metal additives with respect to 4-chloromethylbiphenyl was as described in the Experimental section [open system (oil bath)]. The low boiling point of Fe(CO)5 (102.8 °C, 740 mm Hg)71 made it unsuitable for study under the reaction conditions. A control reaction run with neat 4-chloromethylbiphenyl at 200 °C showed no coupling product and little decomposition of the starting material (Figure 31). The 82 3GG0G 25 000 4 4b TIC of Oh i h : BILLd'B.D T J Figure 31 GC-MS data for control run with neat 4-chloromethylbiphenyl. 83 largest peak (at 3.48 min) corresponds to hexadecane (internal standard). Another control, which was run at 260-265 °C, showed a chromatogram which was very similar to that for the run at 200 °C. The control runs indicate that 4- chloromethylbiphenyl is a stable compound even when heated near its boiling point. Analysis of the reaction mixtures obtained from the pyrolysis of 4- chloromethylbiphenyl in the presence of metal additives proved to be difficult. Since a large amount of solvent was needed to dissolve the coupled product (0.8 g was soluble in 10 mL of CH2C12), the reaction mixture was dilute. Unfortunately, the analysis of each mixture yielded extremely small GC peaks, if any. The only peak of significant size was that of the internal standard. A representative chromatogram is shown in Figure 32. One can see that the peaks in Figure 32 are also present in the chromatogram for the control run. The two metal additives that were the exceptions to these generalized observations were copper(II) formate and copper powder. The only metal additive that coupled the model compound under the experimental conditions described above was copper powder (4.5% yield, by GC area). When the same reaction mixture was heated again at 200 °C for another hour, the yield improved to 20.2% (Figure 33). Several products resulted from the pyrolysis other than the coupled product. The first peak in Figure 33 ( 2.37 min) is for the reduced product, 4-methylbiphenyl (m/z 168). This material apparently 84 (2.9) Scan 3.39? min. of DRTR: BILL 2 i . D 95 “Or*3 u ,~l 14 1 cr 4 4b 50 1 00 150 200 250 300 350 4 00. 4 50 M a s s / C h a r g e TIC of DRTR: BILL2 1 .D 0 000: 28000: c J 100005 Time (min.) Figure 32 GC-MS data for pyrolysis products from iron nonacarbonyl and 4-chloromethyl biphenyl. 85 TIC of DRTR:8ILL24 .D d . UE+5 BE+5: .0E+4: Q.-wE + U -I—i—r~i—'—■—r 8 10 T i me ( min. .1 (IS) Sc an i.373 (4 4 ) Scan 4.59? mi of DRIfl:BILL24 . 0 (185) Sean IS.SSI »in. of DOT 0: 6 1LL24 . D 4 000 15000 67 3000 10000 202 2 000 180 / ? 7 IIS 1 8 5 / ' 239 211 3 9 * S000 82 / 446 I 000 0 A -.lA i. L 1 . \ 0 ^ It i ■: - ll ,4^m4—i. SO 100 ISO 200 250 300 3S0 4 00 4 50 100 ISO 200 250 300 Mas*/Charge ______Maaa/Charge______ f I 90) Scan I S . 058 rnin. of QRTB:8ILL24 .0 (194) Scan I S .375 min. of OPTO:8 ILL24.0 SQ009-] \ 33 4 1500 40000- 165 = 30000- I 000 152 c 2 0 0 0 0 - 241 257 500 / ^ cc 1 0 0 0 0 - V 0 -* x il-,lJiulL ti jhil-lji ■ .1 j i / ... i 50 100 150 200 250 300 350 400 450 100 150 200 250 300 350 400 450 ______H ^ ta /C h tf ge______H* s j /C h arg e Figure 33 GC-MS data for pyrolysis products from copper powder and 4-chloromethyl biphenyl. 86 resulted from the quenching of the reaction by the solvent, CH2C12. Fragmentation of the parent ion/base peak showed a loss of m/z 16 corresponding to the methyl group plus a hydrogen atom, to give m/z 152. The second peak (3.56 min) is for the internal standard, hexadecane (m/z 226). The third peak of interest (4.60 min) is that of the starting material (m/z 202, 204). The two peaks at 15.66 and 16.38 min both had the same parent ion peak of m/z 334 as the desired coupled product but did not match as well as the peak at 16.06 min in retention time and MS fragmentation pattern. Therefore, these peaks were not assigned to the coupled product. Both peaks may, however, be products formed by a dimerization mechanism other than reductive coupling. Although copper(II) formate did not couple the model compound when decomposed in situ, it did cause coupling when.decomposed prior to the introduction of the model compound. The decomposition of copper(II) formate was performed using the open system (external flame). Decomposition was evinced by a color change and deposition of a shiny copper-colored mirror on the reaction flask’s surface, yighly active copper metal produced in this way has been shown to promote rapid reductive coupling of primary and secondary allylic chlorides.6 Figure 34 shows the GC-MS results obtained from the above reaction. As in the reaction with copper powder, the only significant peaks observed were for the reduced product (4-methylbiphenyl), the internal standard, unchanged starting 87 TIC of DATA: BILL 18.0 0.0E+0 T i me ( m i n. ) Scan i.J / I nin. of tfH+M: BILL I*. H (30) Scon 3.4 74 . of OBTA:BILLI8 .D IS 800 5900 113 HI 200 __ 259 . 300 350 400 450100 ISO 200 Ha s a/Ch arg e Hi i i /C h arg e ( 4 4 ) 5 can 4. 5 7.9 . i n . ef 0ATA:8 ILL 18 . 0 ( 184) S e a n . 15.551 niin. ef OAf A: 9 1 LL 19 . 0 ISO 200 I 000 ' 100 150 200 250 300 350 400 4S0 ISO 350300 ______Haas/Charge (199) Scan 15.855 of DATA:BILL I 9 .0 152 445 77 150 350 Figure 34 GC-MS data for products resulting from copper(II) formate decomposition followed by addition of 4-chloromethylbiphenyl. 88 material, and the desired coupled product. When copper(II) formate was decomposed using the open system (oil bath) and the model compound was introduced while the reaction flask was still hot (-150 °C), two unexpected products were formed. Figure 35 shows the GC-MS results obtained for the reaction mixture. The two peaks observed at 5.06 and 5.46 min are of special interest. The first had a parent ion peak of m/z 212 and a base peak of m/z 167. This difference (m/z 45) can be assigned to a loss of a formate radical. This result would indicate that after the pyrolysis, formate ions or radicals either were not completely lost from the copper(II) formate or remained in the immediate vicinity of the decomposed additive. Upon the addition of the model compound, a substitution reaction occurred by which the chlorine atom was replaced by a formate group. It is possible, however, that substitution might have occurred at a highly reactive benzylic hydrogen instead of at the chloride. This may have been the case for the second GC peak of interest. The MS fragmentation pattern of this component showed a parent ion peak of m/z 246. The mass difference between this peak and the base peak of m/z 167 may have corresponded to a formate group and a chlorine radical. Unfortunately, further confirmation of the structure of this compound could not be obtained. The GC-MS data, however, do indicate that formate groups are involved in some type of substitution reaction under these conditions. 89 TIC of DfiTR: BILL 1 ?. D 0 6 oO 10 14 16 18 20 T i m e ( rii i n . ) (IS) 5cm 2.I7.1 .i of OATfl:BILL 17.0 of BHfH:BILLIt. B / 20000 25 00 168 71 Cl 2000 IS 000 43 c / T3 1 500 152 10000 3 1 000 •32 IIS cc \ 31 129 S0001 SO0 S3 / / / . . ii lh,. 11 ,. L , 11 It 0 , l i .1 I I00 ISO 200 SO 90 100 120 1 40 ISO H a * t / 0 li 1 r o e Ha at .'Charge Sc an 4 .030 .f OflTfl- BILL 17.0 (45) Scm 4.SS3 min. of OflTfl:BILL I 7.0 SOQ I S00 i 8 i 50000 167 ■100 - 40900 300- 30000 165 202 200 127 20000 100 7S 82 133 10000 51/ \ \ / / Vs \ 0 Ul i. ^ ____u L 50 100 ISO 200 250 300 3S0 400 ISO IOO 120 140 ISO 190 200 H u i •'Ch ar g e MajJ.'Charge (50) Sc. 5. OS I of OflTfl:BILL 17.0 15 b I 5 c an 5.45b of BMtM:BtLLIf.0 * 212 12000 1 3000 !G7 10000 67 2 000 9090 S0OO 1000 S3 4 000 82 202 2?6 446 \ IIS 194 2000 s 1 IS / / ' / 0 L -2*--- Q J U u4-,------N------—------eo ioo i 20 h o ISO 190 200 SO 100 ISO 200 250 300 350 400 4S0 ______H a*t .'Charge II aja/Charge ( I 99) Scan 15.994 min. of 00 10 : BILLI 7.0 N X 1000 IS 7 500 73 221 4 *16 is; 334 l 5S / / | 28 1 / ! \ i. 9 -M ------L j— ,.J- i .'C h ' Figure 35 GC-MS data for pyrolysis products from copper(II) formate and 4-chloromethyl- biphenyl. 90 The GC-MS results obtained for the reactions conducted using the other metal additives usually gave only the internal standard peak. In some cases, such as with Fe2(CO)9 and Co2(CO)8, the reduced product was observed. Also, upon heating the mixture of model compound and metal additive, vigorous effervescence was noticed for all the metal additives except copper powder. This occurrence can be attributed to the loss of metal ligands (mostly gases) during the decomposition and not to the boiling of the model compounds, since their boiling points were well above the temperature of the reaction mixture. Also, the products resulting from these reactions may have been to heavy to be analyzed by the method being used. 91 B. Solid State Degradation of PVC The ability of many of the metal additives to couple model compounds of PVC (see above) led to study of the metal additives with PVC itself. The metal additives were ground with an agate mortar and pestle and mixed with virgin PVC (1:10 weight ratio) in a flask. The resulting mixtures were heated in an oil bath under continuous argon flow at 200 ± 2 °C for one hour. The results of these experiments were recorded as percentage of polymer converted into gel, and these results are listed in Table 3. The metal chlorides were included in the solid state study for comparison between the Lewis-acid-catalyzed crosslinking (see below) and reductive coupling mechanisms of PVC. Carty and co-workers 72 had proposed that Lewis acids promote extensive dehydrochlorination and subsequent crosslinking during the thermal degradation of PVC, according to Scheme 23. Molybdenum trioxide was also included for comparison of Lewis-acid-catalyzed crosslinking to reductive coupling. A mechanism for the role of Mo0 3 in the thermal degradation of PVC was proposed by Starnes and Edelson (Scheme 11 ) 7 This mechanism involves accelerated heterolysis of the carbon-chlorine bond by the Lewis-acid metal to form an ion pair (see below), since the equilibrium constant for the ionization can be increased by the coordination of Cl to the metal center. As mentioned in the 92 Table 3 Gel Yields.* Run Avg. Metal Additive 1 2 3 4 5 Avg. Dev. none <5 0 Co 2(CO)8 69 73 76 73 2 Fe2(CO)q 76 8 8 67 77 7 Mn 2(CO) 10 71 6 6 6 6 6 8 3 Mo(CO), 69 58 60 83 70 64 5 W(CO), 94 93 72 65 69 69 2 Cu(0 2CH)9 84 97 73 95 92 5 N i(09CH)2 23 16 15 18 3 Fe(02CH)2 51 43 55 50 4 Cu(NCCH,)4PF, 87 89 72 83 7 FeCl, 45 64 63 58 62 2 CuCL, 83 75 87 82 4 NiCl, 21 2 0 8 16 6 CoCl? 66 50 39 25 42 44 4 FeCl, 79 57 75 61 49 6 8 9 MoO, 31 38 57 42 10 a Underlined values were not included in the calculations for the averages and average deviations from the mean. 93 Y 1 . y c p + n Y = Lewis acid decomposition HC1 + Y + \ n v n 2 . n v . n repeat crosslinking char Scheme 23 Lewis-acid-catalyzed dehydrochlorination with subsequent crosslinking .72 94 zcr -c h2ch- Z = M0O3, etc- Introduction, at very high temperatures, Lewis acids also cause extensive “cracking” of the char resulting from the thermal degradation of PVC, and they can increase the flame spread in this way. The formation of polyene sequences in the degraded PVC samples was detected by infrared spectra. Figures 36-42 show comparisons between a PVC control sample degraded without additives and those degraded with the various metal additives in Table 3. Concentrating on the absorption bands around 1600 cm'1, corresponding to conjugated alkene C=C asymmetric stretching, we can see differences between the Lewis-acid metal additives (chlorides and Mo03) and many of the other metal additives. When PVC was degraded alone in a control run, it showed no significant olefmic band. In general, the PVC samples containing Lewis acids showed distinct olefin bands. However, some of the metal carbonyls, the metal formates, and the copper® acetonitrile complex showed either a very small band or no band at all. The metal carbonyl that gave the most extensive double bond formation was Mn 2(CO)10. As can be seen in Table 3, all of the metal additives, with the exception of nickel(II) formate and NiCl2, caused substantial gel formation. Also, the metal chlorides and M 0 O3 generally caused less gel formation, Absorbance 0 0 0 4 IR spectra for PVC degraded in the presence of copper additives. copper of presence the in degraded PVC for spectra IR 0 0 5 3 V undegraded PVC 0 0 0 3 PVC CuCL 0 0 5 2 2000 Figure 36 Figure 0 0 5 1 1000 Absorbance 3000 2500 2000 0 0 2 0 0 5 2 0 0 0 3 - 0 0 5 3 0 0 0 4 fR spectra for PVC degraded in the presence of a tungsten additive. tungsten a of presence the in degraded PVC for spectra fR V undegraded PVC i ------W(CO)6 PVC 1 ------r Figure 37 Figure 1500 96 PVC undegraded PVC Mo(CO), 4 0 0 0 3 5 0 0 3 0 0 0 2 5 0 0 2000 1 5 0 0 1000 Figure 38 IR spectra for PVC degraded in the presence of molybdenum additives. Absorbance 0 0 0 4 IR spectra for PVC degraded in the presence of cobalt additives. cobalt of presence the in degraded PVC for spectra IR 0 0 5 3 V undegradedPVC 0 0 0 3 PVC 0 0 5 2 CoCL 2000 Figure 39 Figure 0 0 5 1 1000 Absorbance 0 0 0 4 IR spectra for PVC degraded in the presence of nickel additives. nickel of presence the in degraded PVC for spectra IR - 0 0 0 3 0 0 5 3 V undegraded PVC Ni( NiClr 0 2500 2CH): 2000 Figure 40 Figure PVC 1500 1000 Absorbance 0 0 0 4 IR spectra for PVC degraded in the presence of a manganese additive. manganese a of presence the in degraded PVC for spectra IR 0 0 5 2 0 0 5 3 V undegraded PVC 0 0 0 3 PVC 2000 Figure 41 Figure 0 0 5 1 1000 100 Absorbance 0 0 0 4 IR spectra for PVC degraded in the presence of iron additives. iron of presence the in degraded PVC for spectra IR 0 0 5 3 V undegraded PVC FeCL 0 0 0 3 0 0 5 2 2000 Figure 42 Figure 0 0 5 1 1000 I"*-' 102 on the average, than their corresponding carbonyl and/or formate complexes, with the exception of Fe(0 2CH)2. (It must be noted, however, that some of the average gel yields on which this conclusion is based actually are identical within the average deviations given). These results indicate that at least some of the metal carbonyls and formates, as well as the copper(I) acetonitrile complex, may not behave as Lewis acids in the crosslinking and subsequent gel formation of PVC. An explanation as to why the Fe(02CH)2 complex does not cause higher gel formation than either of the two iron chlorides may be the high decomposition temperature of this compound (> 200 °C). In order to obtain suitable IR spectra for detection of double bond formation, all three formate/PVC samples were degraded and then washed with 1.2 M HC1. This treatment effectively removed the formate IR absorption peak that overlapped with the C=C peak at 1600 cm'1. The presence of formate ligands in the degraded samples indicated incomplete decomposition of all three metal formates. The evidence collected by IR analysis supports the reductive coupling mechanism for the crosslinking of polymer chains. During the formation of polyene sequences, Friedel-Crafts alkylation may occur at an allylic chloride site to form a crosslink (Scheme 24).6 As seen in Scheme 24, this process stops the polyene growth in one polymer chain. The reductive coupling mechanism (Scheme 25), however, would terminate the growth of two polyene chains .6 This process would result in a lower number of polyene sequences in the PVC sample. Therefore, 103 -Cl' '/w\yxA /\y^~ "VWV" -H Scheme 24 Crosslinking of PVC chains via Friedel-Crafts alkylation. Scheme 25 Crosslinking of PVC chains via reductive coupling. if the crosslinking in the polymer were occurring via the reductive coupling mechanism, rather than the Friedel-Crafts mechanism, there would be less absorption at 1600 cm'1. The presence of only a small olefinic absorption band 104 gives further evidence for the role of reductive coupling with PVC samples degraded in the presence of formates, carbonyls, and copper® acetonitrile. 105 C. Gelation of PVC in Solution The solid state degradation of PVC in the presence of metal additives is limited by the mechanical mixing of the samples. Dissolution of the PVC sample creates a more homogeneous mixture between the polymer chains and the additive. Therefore, a heterogeneous mixture of copper(II) formate and virgin PVC, dissolved in phenyl ether, was heated at ~ 250 °C until a noticeable gel had formed. The gelation occurred within 20 min after the formation of the copper mirror. After the extraction of undegraded PVC and other organic material by hot THF, the yield of gel was calculated to be 69.0±2%. Since the control experiment yielded 0% of gel, the above result is of considerable importance. 106 rv. CONCLUSIONS Several allylic and benzyl chloride model compounds which represent structural segments of PVC were shown to couple reductively in the presence of some reduced transition-metal additives. The decomposition of the metal compound was done by pyrolysis, and experiments were conducted in both sealed and open systems. The transition-metal additives were also shown to cause substantial gelation of PVC when they were degraded in the presence of the polymer. In some cases the corresponding metal formate and/or carbonyl did not promote the extensive formation of conjugated double bonds and caused more gelation when compared to samples degraded in the presence of strong Lewis acids. Some of the metal additives studied are possible candidates for smoke suppression and fire retardance of PVC. Two additives of special interest are the previously studied copper(II) formate and tetrakis(acetonitrile)copper(I) hexafluoro- phosphate. The metal carbonyls, however, probably would not be useful additives for PVC because of their low decomposition temperatures. In light of the model-compound and solid-state PVC degradation studies, the reductive coupling mechanism could, in some cases, be an important mechanism for the crosslinking (and subsequent gelation) of the polymer. Reactions with other 107 metal additives, as well as direct microstructure analysis of the degraded PVC samples, might provide further support for this conclusion. 108 REFERENCES 1. Chem Eng. News. 62(26), 15 (1984). 2. D. Brown and E. Bezdadea, Encyclopedia of PVC (L. I. Nass and C. A. Heiberger; Eds.; Marcel Dekker: New York, 1986). 2nd ed., Vol. 1, Chap. 8 , and references cited therein. 3. Chem. Eng. News. 72(22), 21 (1994). 4. Encyclopedia of Polvmer Science and Engineering. 2nd ed., Suppl. Vol.; Wiley: New York, 1989. 5. W. Deborah, In the Mouth of the Dragon: Toxic Fires in the Age of Plastics. K. Price; Ed., Avery Publishing Group: Garden City Park, NY, 1990; p. 11. 6 . J. P. Jeng, “Mechanism of Poly(vinyl chloride) Fire Retardance and Smoke Suppression Induced by Copper Additives,” Ph.D. thesis, College of William and Mary, 1995. 7. W. H. Starnes, Jr. and D. Edelson, Macromolecules. 12, 797 (1979). 8 . W. J. Kroenke, J. Appl. Polvm. Sci.. 26, 1167 (1981). 9. R. P. Lattimer and W. J. Kroenke, J. Appl. Polvm. Sci.. 26, 1191 (1981). 10. D. H. Paul, J. Vinvl Technol.. 153 (1993). 11. Industrial Research. Feb. 1970, p. 24. 12. B. Ivan, T. Kelen, and F. Tudos, Degradation and Stabilization of Polymers. Vol. 109 2, H. H. Jellinek and H. Kachi, Eds.; Elsevier: Amsterdam, 1989; pp. 570-3, and references cited therein. 13. T. Kelen, Polymer Degradation. Van Nostrand Reinhold: New York, 1983; pp. 77-96, and references cited therein. 14.1. Mita, Aspects of Degradation and Stabilization of Polymers. H. H. Jellinek, Ed.; Elsevier: Amsterdam, 1978; p. 24. 15. C. David, Chemical Kinetics. C. H. Bamford and C. F. H. Tipper, Eds., Elsevier: New York, 1975; p. 1. 16. E. A. Boettner, G. Ball, and B. Weiss, J. Appl. Polvm. Sci.. 13, 377 (1969). 17. A. Ballistreri, S. Foti, G. Montaudo, and E. Scamporrino, J. Polvm. Sci.. Polvm. Chem. Ed.. 18, 1147 (1980). 18. R. W. Michelson, K. J. Voorhees, and P. W. Ryan, “Correlation Between Thermal Analysis and Effluent Gas Analysis,” Polymer Conference Series, University of Utah, July 7-11, 1975. 19. R. M. Lum, J. Appl. Polvm. Sci.. 23, 1247 (1979). 20. R. M. Lum, L. Seibles, D. Edelson, and W. H. Starnes, Jr., Org. Coat. Plast. Chem.. 43,176(1980). 21. A. Ballistreri, S. Foti, P. Mararvigna, G. Montaudo, and E. Scamporrino, L_ Polvm. Sci..Polvm. Chem. Ed.. 18, 3101 (1980). 22. W. H. Starnes, Jr., Dev. Polvm. Peg.. 3, 135 (1981), and references cited therein. 110 23. W. H. Starnes, Jr., F. C. Schilling, I. M. Plitz, R. E. Cais, D. J. Freed, R. L. Hartless, and F. A. Bovey, Macromolecules. 16, 790 (1983), and references cited therein. 24. T. Hjertberg and E. M. Sorvik, Polymer. 24, 673 (1983). 25. T. Hjertberg and E. M. Sorvik, and A. Wendel, Makromol. Chem.. Rapid Commun.. 4, 175 (1983). 26. J. Wypych, Polwinvl Chloride Degradation. Elsevier: New York, 1985; pp. 23- 29. 27. D. Braun and F. Weiss, Angew. Makromol. Chem.. 13, 55 (1970). 28. T. Suzuki and M. Nakamura, Jpn. Plastics. 4, 16 (1970). 29. W. H. Starnes, Jr. and S. Girois, Polvmer Yearbook, in press, and references cited therein. 30. N. M. Boughdady, K. R. Chynoweth, and D. G. Hewitt, Aust. J. Chem.. 44, 567 (1991). 31. G. Montaudo and C. Puglisi, Polvm. Peg, and Stab.. 33, 229 (1991). 32. M. M. O’Mara, Pure Appl. Chem.. 49, 649 (1977), and references cited therein. 33. T. Iida, M. Nakanishi, and K. Goto, J. Polvm. Sci.. Polvm. Chem. Ed.. 12, 737 (1974). 34. W. H. Starnes, Jr., L. D. Wescott, Jr., W. D. Reents, Jr., R. E. Cais, G. M. Villacota, I. M. Plitz, and L. J. Anthony, Org. Coat. Plast. Chem.. 46, 556 Ill (1982), and references cited therein. 35. R. P. Lattimer and W. J. Kroenke, J. Appl. Polvm. Sci.. 25, 101 (1980). 36. R. P. Lattimer and W. J. Kroenke, J. Appl. Polvm. Sci.. 27, 1355 (1982). 37. D. H. Edelson, R. M. Lum, W. D. Reents, W. H. Starnes, Jr., and L. D. Wescott, Proceedings. Nineteenth Symposium (International! on Combustion: The Combustion Institute: Pittsburgh, August 1982; p. 807. 38. B. B. Troitskii, L. S. Troitskaya, V. N. Myakov, A. F. Lepaev, J. Polvm. Sci.. Polvm. Svmp.. 42, 1347 (1973). 39. T. Kelen, J. Macromol. Sci.. Chem.. 12, 349 (1978). 40. E. Martinsson, T. Hjertberg, and E. Sorvik, Macromolecules. 21, 136 (1988). 41. T. Hjertberg and E. M. Sorvik, J. Appl. Polvm. Sci.. 22, 2415 (1978). 42. M. V. Neiman, R. A. Papko, and V. S. Pudov, Polvm. Sci. USSR (Engl. Transl. V 10, 975(1968). 43. F. Tudos, T. Kelen, T. T. Nagy, and B Turcsany, Pure Appl. Chem.. 38, 201 (1974). 44. R. P. Lattimer, W. J. Kroenke, and R. G. Getts, J. Appl. Polvm. Sci.. 29, 3783 (1984). 45. R. P. Lattimer and W. J. Kroenke, Analytical Pvrolvsis: Techniques and Applications: K. J. Voorhees, Ed.; Butterworths: Woburn, MA, 1984; p. 453. 46. G. A. Tsigdinos, F. W. Moore, V. J. Kuck, and E. Scalco, US Pat. 4,098,753 112 (1978). 47. V. J. Kuck and E. Scalco, “Smoke Suppression in PVC’” presented to the Society of Plastics Engineers, Atlantic City, NJ, 1976. 48. S. Kaufman and M. M. Yocum, International Conference on Plastics in Telecommunications II London, Sept. 18-20, 1978. 49. L. D. Wescott, W. H. Starnes, Jr., and A. M. Mujsce, J. Anal. Appl. Pvrolvsis. 8 , 163 (1985). 50. W. H. Starnes, Jr., L. D. Wescott, Jr., W. D. Reents, Jr., R. E. Cais, G. M. Villacorta, I. M. Plitz and L. J. Anthony, in Polvmer Additives: J. E. Kresta, Ed.; Plenum: New York, 1984, p. 237. 51. C.-H. O. Huang and W. H. Starnes, Jr., Proceedings. 2nd Beijing International Symposium on Flame Retardants: Geological Publishing House: Beijing, 1993; p. 168. 52. W. H. Starnes, Jr., and C.-H. O. Huang, Polvm. Prepr. (Am. Chem. Soc.. Div. PolvmChem.L 30(1), 527 (1989). 53. J. K. Kochi, Acc. Chem. Res.. 7, 351 (1974). 54. T. Cohen and I. Cristea, J. Org. Chem.. 40, 3649 (1975). 55. T. Kawaki and H. Hashimoto, Bull. Chem. Soc. Jpn.. 45, 3130 (1972). 56. C.-H. O. Huang, “Mechanistic Studies of Copper Additives as Smoke Suppressants for Poly(vinyl chloride),” Ph. D. thesis, Polytechnic University, 113 1990. 57. G. W. Ebert and R. D. Rieke, J. Org Chem.. 53, 4482 (1982). 58. F. O. Ginah, T. A. Donovan, Jr., S. D. Suchan, D. R. Pfenning, and G. W. Ebert, J. Org. Chem.. 55, 584 (1990). 59. F. Kordsy, Nature. 160, 21 (1973). 60. M. N. Yoon, C. S. Pak, H. C. Brown, S. Krishnamurthy, and T. S. Stocky, L_ Org. Chem.. 38, 2786 (1973). 61. H. R. Hudson and G. R. de Spinoza, J. Chem. Soc.. Perkin Trans. I. 1 , 104 (1976). 62. L. R. C. Barclay, H. R. Sonawane, and J. C. Hudson. Can. J. Chem.. 20, 2318 (1972). 63. P. J. Bunyan and D. H. Hey, J. Chem. Soc.. 1360 (1972). 64. K. M. Johnston and G. H. Williams, J. Chem. Soc.. 1168 (1960). 65. H. Hopf and K. Koulen, Chem. Ber.. 85, 897 (1952). 6 6 . R. N. Rhonda and A. V. Fraioli, Inorganic Synthesis. 4, 159 (1953). 67. Dictionary of Inorganic Compounds: J. E. Macintyre, Ed.; Chapman and Hall: London; 1992, p. 802. 6 8 . Dictionary of Organic Compounds: J. Buckingham, Ed.; Chapman and Hall: New York, Vol. 1, 1982; p. 1134. 69. Dictionary of Organic Compounds: J. Buckingham, Ed.; Chapman and Hall: 114 New York, Vol. 1, 1982; p. 1195. 70. J. A. Connor, H. A. Skinner, and Y. Virmani, J. Chem. Soc.. Faradav Trans. I. 6 8 , 1754(1972). 71. Dictionary of Qrganometallic Compounds: J. Buckingham, Ed.; Chapman and Hall: London, Vol. 1, 1984; p. 641. 72. P. Carty, E. Metcalfe, and S. White, Polvmer. 33, 2704 (1992). VITA William Stephen By rant was bom in Charleston, South Carloina on March 17, 1972. He attended high school in Virginia Beach, Virginia and graduated in 1990 from the First Colonial High School. He received his Bachelors of Science degree in Chemistry from the College of William and Mary in 1994. Upon the completion of the requirements for the Master of Arts degree in Chemistry in 1995, he enrolled at Virginia Polytechnic Institute and State University where he is currently pursuing a Doctor of Philosophy degree in Chemistry.