Title: Phase separation of metal or metal-oxide microparticles in solid polymer matrices /
CITATION PDF VIEWER THUMBNAILS PAGE IMAGE ZOOMABLE
Full Citation
STANDARD VIEW MARC VIEW
Permanent Link: http://ufdc.ufl.edu/UF00099485/00001
 Material Information
Title: Phase separation of metal or metal-oxide microparticles in solid polymer matrices /
Physical Description: Book
Language: English
Creator: Flenniken, Cindy Lou, 1952-
Publisher: s.n.
Copyright Date: 1984
 Subjects
Genre: bibliography   ( marcgt )
non-fiction   ( marcgt )
 Notes
Statement of Responsibility: by Cindy Lou Flenniken.
 Record Information
Bibliographic ID: UF00099485
Volume ID: VID00001
Source Institution: University of Florida
Holding Location: University of Florida
Rights Management: All rights reserved by the source institution and holding location.
Resource Identifier: alephbibnum - 001088019
oclc - 19299339
notis - AFH3392

Downloads

This item has the following downloads:

phaseseparationo00flen ( PDF )


Full Text

PHASE SEPARATION OF METAL OR METAL-0XIDE
MICROPARTICLES IN SOLID POLYMER MATRICES






BY



CINDY LOU FLENNIKEN


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



UNIVERSITY OF FLORIDA

1984


1.





























To my Family for the wealth of understanding
and encouragement which they provided.















ACKNOWLEDGEMENTS


The author wishes to thank all her committee members for their

suggestions and guidance, especially Dr. Eugene P. Goldberg for his

constructive criticisms and direction throughout this project and Dr. J.

J. Hren for his technical dicussions. Acknowledgement is also made of

Dr. R. Tannenbaum's contribution to interpretation of the organometallic

chemistry involved in this study. The use of MAIC (Major Analytical

Instrumentation Center) facilities and the cooperation of its staff were

essential to the research presented in this dissertation.

Warm thanks are also extended to Dr. Henry Catherino in Detroit,

Michigan, for his encouragement to pursue graduate studies; to Or. J. S.

Lin for his assistance with the SAXS studies at Oak Ridge National

Laboratories; to Ms. Peg Mochel at the Materials Research Laboratory at

the University of Illinois for her willing assistance on the electron

beam decomposition studies; and to Dr. R. C. Stoufer for his assistance

in evaluating the magnetic characteristics of these systems.

A special thank you is extended to Stuart Cheshire and Amy Smith

for providing constant encouragement and a warm home environment during

the writing of this manuscript. The author wishes to express gratitude

and appreciation to Bill Longo, Jan Stacholy, John Sheets, David Osborn,

Jeff Larson, Mindy Donenfeld, Ching-Wang Luo, and Richard Robinson for

their friendship, understanding, and assistance.









Thanks go to the faculty and staff for their varied assistance: E.

J. Jenkins, Drs. Joe Newkirk, H. C. Aldrich and R. H. Berg III for their

assistance with TEM sample preparations and studies; to Guy LaTorre and

Frank Sodek, for their varied skills and help; to Mildred Neal for her

assistance in literature searches at the Chemistry Library; to the

Engineering Publications Department for their excellent graphic works;

to and Michelle Smith for her patience in typing this manuscript.









































iV

















TABLE OF CONTENTS


ACKNOWLEOGEMENTS .. .. .. .. .. . .. . .. iii

LIST OF TABLES ... .. ... . . .. .. . . .. xi

LIST OF FIGURES. ... .. .. .. .. ... .. .. .. xiii

ABSTRACT ... .. . ... ... .. .. .. .. .. .. xix

1. INTRODUCTION .. .. .. .. .. .. .... .. .. 1

2. BACKGROUND .. .. ... .. . .... .. . .. .. 6

2.1 Conventional Metal-Polymer Composites .. .. .. .. 6

2.2 Metal-Polymer Composites by Solid
State Phase Separation .. .. ... .. ... 8

2.3 Chemistry of Metal Carbonyls ... .. .. .. .. 10

2.3.1 Disproportionation Reactions of
Metal Carbonyls. . .... .. .. ... 10

2.3.2 Metal Carbonyl Reactions with Monoolefins. .. 11

2.3.3 Metal Carbonyl Reactions with Conjugated
and Nonconjugated Oienes . .. .. .. . 13

2.3.4 Metal Carbonyl Reactions with Haldes.. .. 16

2.4 Related Studies . ... ... .... .. .. 19

2.4.1 Ferrofluids and Other Colloidal Dispersions 19

2.4.2 Organometallic Complexes as Catalysts .. 25

2.4.3 Metal-Polymer Composites Prepared by
Adsorption of Metal Carbonyl Solution
onto Films. .. .. . . .. ..... .. 27

2.4.4 Metal-Polymer Composites Prepared
by Casting of Metal Carbonyl-Polymer
Solutions .. ... .. .. .. .. .. 31

2.5 Metal-Polymer Composites by Phase Separation. . .. 37









2.5.1 Polymer Matrices Selected for Preparation
of Metal-Polymer Composites . .. ... .. 37

2.5.2 Organometallic Complexes Selected
for Preparation of Metal-Polymer Composites 39

3. EXPERIMENTAL .. .. .... .. .. ... . ... 42

3.1 Preparation of Iron Pentacarbonyl-Polymer Films .. 42

3.1.1 Polycarbonate (PC)/Fe(CO)5/Methylene
Chloride. .. .. .. ... . .. ... 42

3.1.2 Polystyrene (PS)/Fe(CO)S/Methyl ene Chloride 43

3.1.2.1 Polystyrene/Fe(CO)5/methylene
chloride preparation in air. . .. 43

3.1.2.2 Polystyrene/Fe(CO)5/methylene
chloride preparation in H2 . ... 43

3.1.3 Polysulfone (PSF)/Fe(CO)5/Methylene
Chloride. . . . . . . . . . 45

3 .1 .3.1 Poly sul fone/Fe(CO) 5/Nethyl ene
Chloride Preparation in Air. . .. 45

3.1.3.2 Polysulfone/Fe(CO)5/Methylene
Chloride Preparation in H2 .. .. 46

3.1.4 Polydimethylsiloxane (PSi)/Fe(CO)5/
Methylene Chloride.............. 47

3.1.5 Polymethylmethacrylate
(PMMA)/Fe(CO)5/Benzene. .. .. . . 47

3.1.6 Polyvinylidene Fluoride (PVF2)/Fe(CO)5/
Dimethylformamide (DMF) . .. . . 48

3.2 Preparation of Triiron Dodecacarbonyl-Polymer Films 49

3.2.1 Polystyrene/Fe3(CO) 12/M~ethylene Chloride. . 49

3.2.2 Polycarbonate/Fe3(CO)12/Methylene Chloride. 49

3.2.3 Polysulfone/Fe3(CO)12/Methylene
Chloride With Added Surfactant. .. .. .. 51

3.2.4 Polystyrene/Fe3(CO) 12/Methylene Chloride
With Added Surfactant. .. .. .. . 52

3.3 Preparation of Dicobalt Octacarbonyl-Polymer Films 52





3.3.1 Yaried Loadings of Co2(CO)g i
Polystyrene/Methylene Chloride
Prepared Under CO. .. ... ... .. 52

3.3.2 Co,(CO)Q in Polystyrene/Methylene
Chloride Prepared Under H2 .. .. ... 54

3.3.3 Polystyrene/Co2!(CO3)l/Methylene Chloride
Sandwich Preparatiron (to Evaluate Oxygen
Diffusion Rate). .. .. .. .. .. .. 54

3.3.4 Preparation of Co2(CO)g in
Polyvinylidene Fluoride (PVF2) Dimethyl
Formamide (DMF). 56

3.4 Preparation of Iron and Cobalt
Carbonyl-Polymer Films ... .. .. .. .. .. 56

3.4.1 Polyvinylidene Fluoride
(PVF2)/Fe(CO) -o(O8/mehfoaid
Prepared in Ca~LO) . .. .. .. .. 56

3.4.2 Polysulfone (PSF)/Fe(CO)5-Co2(CO)8
Methylene Chloride Preparation in CO . .. 57

3.5 Preparation of Control Films .. ... .. .. 57

3.6 Quantitative Metal Analysis of Metal
Carbonyl-Polymer Films ... ... .. .. .. 58

3.6.1 Standard Curve For Yogells Iron Analysis . 58

3.6.2 Vogells 1,10-Phenanthroline Method
for Iron Analysis. .. .. ... .. .. 59

3.6.3 Pyrolytic Technique for Atomic
Absorption to Quantitative Iron and
Cobalt Content .. .. ... ... .. 61

3.7 Determination of Extinction Coefficient
of Netal Carbonyls in a Polymer or in a Solution . 62

3.8 Solid State Decomposition Methods for
Metal Carbonyl-Polymer Systems .. .. .. . 62

3.8.1 Thermal Decomposition of Metal
Carbonyl-Polymer Systems .. .. .. .. 62

3.8.2 Photolytic Decomposition of Metal
Carbonyl-Polymer Systems . .. ... .. 63

3.8.3 Gamma Radiation Decomposition of
Metal Carbonyl-Polymer Systems ... .. 63









3.8.4 Electron Beam Decomposition of
Metal Carbonyl-Polymer Systems .. .. .. 66

3.9 Kinetics for Decomposition of Metal
Carbonyls in Solid Polymer Matrices. . .. .. .. 67

3.10 Kinetics of Carbonyl Decomposition in Solution .. 69

3.10.1 Thermal Decomposition of Fe(CO)5
in Ethyl Benzene Solution. .. .. ... 69

3.10.2 Thermal Decomposition of Fe3(CO)12
in Ethyl Benzene Solution. 69

3.10.3 Photolytic Decomposition of Fe(CO)5
in Ethyl Benzene Solution. .. .. ... 70

3.11 Microscopic Characterization of
Composite Film Morphology. ... .. .. .. .. 70

3.12 Small Angle X-Ray Scattering Experiments
to Measure Particle Size Distribution. .. .. .. 72

3.13 Determination of Average Magnetic
Susceptibility of Metal-Polymer Composites
by the Guoy Method . ... .. .. .. .... 75

3.13.1 Sample Preparation for Magnetic
Susceptibility Measurements. .. .. .. 77

3.13.2 Guoy Balance for Measurements of
Magnetic Properties. . .. .. .. ... 78

3.14 Effect of Fe(CO)5 Decomposition on
Polymer Molecular Weight . ... .. .. .. .. 78

4. RESULTS AND DISCUSSION. .. .. .. .. ... .. .. 80

4.1 Preparation of Metal Carbonyl-Polymer
Solid Solutions .. .. ... .. .. .. .. .. 80

4.1.1 Carbonyl-Solvent Interactions
Affecting Polymer Films. . .. ... . 80

4.1.2 Photolytic and Oxidative Decomposition
Factors Affecting Incorporation of
Metal Carbonyls Into Polymer Films . .. .. 92

4.2 Spectroscopic Analysis of Carbonyl-Polymer
Compositions .. .. .. .. .. ... .. . .. . 94

4.2.1 UV-Vis Spectroscopic and Atomic
Absorption Analysis to Determine Metal
Concentrations In Films. .. .. .. .. 97





4.2.2 Determination of Infrared Extinction
Coefficients to Correlate Metal and Metal
Carbonyl Concentrations. .. .. .. . 98

4.3 Decomposition of Metal-Carbonyls
in Polymer Matrices. . .. .... . ... ... 101

4.3.1 Kinetics of Thermal Decomposition
for Iron Pentacarbony1 Solid Polymeric
Matrices ... .. .. ... .. .. .. 102

4.3.2 Kinetics of Thermal Decomposition for
Triiron Dodecacarbonyl in Solid Polymer
Matrices . . . ... .. .. .. 109

4.3.3 Kinetics of Thermal Decompositions
of Cobalt Carbonyls in Solid Polymer
Matrices .. .. .. .. .. .. ... 114

4.3.4 Kinetics of Photolytic Decompositions
of Iron Carbonyls in Solid Polymer Matrices. 122

4.3.5 Kinetics of Gamma Radiation Decomposi-
tions of Iron Carbonyls in Solid
Polymer Matrices .. .. .. .. .... 126

4.4 Characterization of Metal-Polymer Composites .. .. 129

4.4.1 Microparticles Observed Within Polymeric
Matrices .. .. .. .. .. ... . ... 130

4.4.1.1 Transmission electron microscopy with
electron energy dispersive spectra to
characterize chemical composition of
particles formed by solid state decom-
position of metal carbonyls in polymer
matrices. .. .. .. .. . .. ... 130

4.4.1.2 Small angle x-ray scattering analyses to
determine range of particle sizes in
metal-polymer composites. .. . ... 152

4.4.2 Determination of Average Mlagnetic
Susceptibility of Metal-Polymer
Composites by the Guoy Method. .. ... 159

4.4.3 Electron Beam Decompositions of Iron
Carbonyls in Solid Polymer Matrices. .. .. 164

4.5 Effect of Carbonyl Decomposition Upon Polymeric
Matrix .. .. .. .. .. .. .. ... . ... 171

4.6 Mechanistic Aspects. ... .. .. .. .. ... 178









5. CONCLUSIONS .. .... .. .. .. ... .. .. 184

6. FUTURE RESEARCH .. .. .. .. .. . .. . .. .. 192

6.1 Preparation Techniques to Limit Metal
Carbonyl Decomposition Prior to Incorporation
into Polymer Thus Controlling Product Purity .. .. 192

6.2 Metal Carbonyl-Polymer Systems .. .. .. .. .. 192

6.3 Decomposition Treatments . .. .... .. ... 193

6.4 Analytical Techniques Recommended. . ... . ... 194

APPENDIX A CALCULATION OF ENERGY ABSORBED BY SAMPLE
DURING ELECTRON BEAM DECOMPOSITION .. .. .. 196

APPENDIX B INTRINSIC VISCOSITY MEASUREMENT AND
CALCULATION OF YISCOSITY-AVERAGE MOLECULAR
WEIGHT. .... .. .. .. .. . .. 198

APPENDIX C DETERMINATION OF EXTINCTION COEFFICIENTS
IN A POLYMER MATRIX. .. .. .. .. ... 202

APPENDIX D CALCULATION OF RATE CONSTANTS FOR DECOM-
POSITION REACTION OF METAL CARBONYLS IN
POLYMER MATRIX .. .. .. .. .. .. .. .. 204

APPENDIX E CALCULATION OF MAGNETIC SUSCEPTIBILITY OF
COMPOSITES BY THE GU0Y METHOD. . .. .... 206

REFERENCES. .. ... .... .. .. .. .. .. .. 208

BIOGRAPHICAL SKETCH .. .. .. .. .. .. ... .. .. 215















LIST OF TABLES


TABLE

1 Preparation of Dilution Series for Standard lon
Concentration Curve .. .... .. ... .. 60

2 Metal Carbonyl-Polymer Systems Studied. . ... .. 81

3 Casting Methods for Fe(CO)5/PVF2/DMF Systems. .. .. 88

4 IR Extinction Coefficients for Metal Carbonyls
In Solution .. .. .. .. .. ... .. .. .. 99

5 IR Extinction Coefficients for Metal Carbonyls
in Polymer Matrices .. .. .. .. .. .. ... 100

6 Observed Rate Constants for the Thermal
Decomposition of Fe(CO)5-Polymer Composites . ... 106

7 Thermal Activation Energies For The Decom-
position of Fe(CO)5-Polymer Composites .. .. .. 108

8 Observed Rate Constants For The Thermal Decom-
postion In Air of Fe3(CO)12-Polymer Composites
and Solutions. ... . . .... ... ... 112

9 Observed Rate Constants for the Thermal
Decomposition of Co2(CO)B in Solution
and in Polystyrene ......... ........ 118

10 Observed Rate Constants for the Photolytic
Decompostion of Iron-Polymer Systems. .. .. .. .. 126

11 Observed Rate Constants for Gamma Radiation
Decomposition of Fe3(CO)12-Polymer Composites .. .. 128

12 Electron Diffraction Analyses for FeF2an
FeS in PVF2 and PSF Composites, Respectively. . ... 138

13 Electron Diffraction Analysises for a-Co Formed
in Co2(CO]8-PS Films Thermally Treatin N2 .. .. .. 150

14 Particles Obtained in various Systems . ... ... 153









15 Distribution in Particle Radii of Gyration
for Metal-Polymer Composite Systems .. .. .. .. 158

16 Magnetic Susceptibility, x, for Metal-Polymer
Composites. ................... ... 162

17 Viscosity-Molecular Weight Averages for Iron
Carbonyl-Polymer Composites. .. .. .. .. ... 172















LIST OF FIGURES


FIGURE

1 Concept for Producing Metal-Polymer Composite
Films. . .. .. ... .. .. .. .. 9

2 Infrared Spectra From An Experiment With
W(CO)6 in a PVC Film (Cast From THF). . ... ... 33

3 Photoreactions of Gruop VI M~etal
Hexacarbonyls in PVC Films. .. ... . ... .. 34

4 Polymers Selected For Metal-Polymer Composite
Studies . .. .. ... .. . .. . 38

5 Proposed Chemical Configuration of (a) Fe(CO)5*
(b) Fe3(CO)12 and (c) Co2(CO)8. .. .... . .. 40

6 Schematic for Preparation of Composites in
a Closed Environment. .. .. .. ... .. ... 44

7 "Sandwich" 02 Barrier Concept for
Polystyrene-Co2(CO)8 Films. . .. .. ... .. .. 55
8 Iron Standard Curve in Dimethyl Formamide .. .. .. 60

9 Apparatus for Thermal Decomposition of Metal
Carbonyl-Polymer Systems in a Closed Environment. .. 64

10 UV Irradiation Assembly . .. .. . .... .. 64

11 Schematic Drawing of Co-60 Source, Department of
Radiation Biology, University of Florida . .. .. .. 65

12 Apparatus Used for Thermal Decomposition of
Fe(CO)5 in Ethyl Benzene Solution .. .. . 71

13 Apparatus for Photolytic Decomposition of
Fe(CO)5 in Ethyl Benzene Solution .. .. .. 71

14 Photograph of Small Angle X-Ray Scattering
Facility at Oak Ridge National Laboratory . ... .. 73

15 Schematic of SAXS Facility Used at Oak Ridge
National Laboratory ......... ... 74


X11









16 The Guoy Balance. .. .. ... .. .. .. ... 76

17 Infrared Spectrum of Fe(CO)5 in DMF and in Benzene. . 84

18 Infrared Spectrum of Co2(CO)8 in DMF. .. .. .. .. 86

19 Infrared Spectrum of Co?(CO)8 nHxn fe
Mixing 30 Units and in CO .... ..... .... 86

20 Disproportionation Reaction of Iron and
Cobalt Carbonyl in DMF . .. ... .. .. .. ... 87

21 Infrared Spectrum of a 30 wt. % Co (CO)8"
Polystyrene Film (Cast from (CH2C1 ) .... ..... 87

22 Infrared Spectrum of Fe(CO)5 i
Methylene Chloride .............. .... 89

23 Infrared Spectra of Fe3(CO)12 nMFadi
Methylene Chloride .................. 91

24 Photomicrograph of Fe3(CO)12-Polycarbonate
Film Showing Fe3(CO)12 Crystals. ............ 91

25 Infrared Spectrum of Fe(CO)5 in Benzene . ... ... 95

26 Infrared Spectrum of Fe(CO)5-Polystyrene
Film Before Decomposition of Carbony1. ... .. .. 95

27 Infrared Spectrum of Fe3(CO)12 in Polycarbonate
(Cast from CH2C12). .......... ... .. 96

28 Thermal Decomposition Fe(CO)5 in PSF at 118"C. .. .. 104

29 Thermal Decomposition of Fe(CO)5 in PVF2 at 140*C. .. 104

30 Decomposition Kinetics of Fe(CO)5-P 2
Composites at 140"C................... 105

31 Decomposition Kinetics of Fe(CO)5-
Polymer Systems at 132*C ................ 105

32 Arrhenius Plot for Thermal Decomposition of
Iron Pentacarbonyl-Polymer Composites .. .. .. .. 108

33 Infrared Spectra Showing Thermal Decomposition
of Fe3(CO)12 in Polystyrene Prepared in
Air (a) and in CO (b) .......... ....... 110

34 Thermal Decomposition of Fe3(CO)12 in Ethyl Benzene
at 82"C in Air ..................... 113





35 Decomposition in Air of 4.76 x 10-2M/L
Co2(CO)8 In Hexane ... .. .. .. .. .. ... 115

36 Decomposition in N2 of 5.53 x 10-2M/L Co2(CO)8
in Toluene . . . . . . . . . . . 115

37 Second-Order Decomposition of Co2(CO)8
to Co4(CO)12 in Toluene at 90"C In Inert
Atmosphere ...................... .116

38 Infrared Spectra of the Decomposition of Cop(COig
Polystyrene (Cast from CH2C12 in H2) at 36C nAr...17

39 Infrared Spectra of the Decomposition of Cop(Og
Polystyrene (Cast from CH2Cl2 in CO) at 36C in ir .11

40 Decomposition of "Protected" Sample: Co2(CO)g i
Polystyrene Effect of Increasing 02 BrirFl
Thickness .. .. . .. .... . .. ... . 120

41 First-Order Decomposition in Air of
Co2(CO)4-Polystyrene Film With Polystyrene
Protect ve Layers at Room Temperature. .. .. .. .. 120

42 Second-Order Decomposition of Co2(CO)8 to
Co41CO)12 in Polystyrene Matrix in Inert
Atmosphere at 90*C. .. .. .. .. ... .. .. 121

43 Photolytic Decomposition of Fe(CO)5-Polymer
Systems at 38 Nlanoamperes ...... ....... 125

44 TEM Photomicrograph Showing Particles Formed by
Thermal Decomposition of Fe(CO)5 in a Polycarbonate
Matrix ... .. .. ... .. .. .. .. .. 131

45 TEM Photomicrograph (Higher Magnification
and Higher Carbonyl Loading) Showing Particles
Formed by Thermal Decomposition of Fe(CO)5
in a Polycarbonate Matrix ............... 131
46 SEM Photomicrograph of a Liquid Npoe Frctr
Surface Showing Iron Clusters Encosdb
Polycarbonate Matrix . .. ... .. ... . 132

47 Electron Diffraction Pattern for Iron-PC
Composite Indicative of y-Fe20g and 8-Fe00H. . ... 132

48 Electron Energy Dispersive Spectra for Polycarbonate
Control Film Showing Chloride Peak (i.e. Retained
CH2Cl2) .. .. .. ... .. .. ... .. .. 134
49 Electron Energy Dispersive Spectra for Iron-
Polycarbonate Film Showing Chloride Peak (i.e.
Retained CH2CI2) . . . . . . . . . 134





50 TEM Photomicrograph of Chain-like Particle
Morphology of "cold cast" Iron-PVF2 Fl
Following Thermal Decomposition. 135

51 TEM Photomicrograph of Polysulfone Film
Thermally Treated (Lighter Regions Represent
Thinner Film) .. .. .. .. .. . . .. 135

52 TEM Photomicrograph of Iron-PSF Films
Thermally Treated .. .. .. .. .. . 136

53 Higher Magnification TEM Photomicrograph of
Iron-PSF Film Thermally Treated Showing Particle
Morphology in Thin Film Region ... .. .. .. 136

54 Electron Diffraction Pattern of Hexagonal
Crystalline Morphology in Iron-PSF Composite. .. 137

55 Electron Energy Dispersive Spectra of Crystalline
Morphology in fron-PSF Composite Indicative of
iron (Copper Peak is Due to Cu Grid, S Peak to PSF) 137

56 TEM Photomicrograph of Fe(CO)5-PS Film Thermally
Treated . .. . .. .. ... .. .... .140

57 TEM Photomicrograph of Fe(CO)5-PMMA Film
Thermally Treated . . ..... .. .. .. .. 140

58 Diffraction Pattern of Polydimethylsiloxane Film. . 141

59 TEM Photomicrograph of Polydimethylsiloxane
Which Shows Diffracting Morphology . ... .. .. 141

60 TEM Photomicrograph of Fe(CO)5-PSF Film Prepared
in H2-Argon and Thermally treated in Air . .. ... 142

61 TEM Photomicrograph of Fe(CO)5-PS Film Prepared
in H2-Argon and Thermally Treated in Air ..,,... 142
62 Dark Field TEM Photomicrograph Image of Iron
Cluster in Fe(CO)5-PSF Film Prepared in H2-Argon,
Thermally Treated in Air ...... ..... 143

63 TEM Photomicrograph Showing Spherical Morphologies
of Sulfur and iron in Fe3(CO)12-PSF Composite
(Prepared with DMF). 146

64 TEM Photomicrograph of Iron, Iron Oxide Clusters
in Fe3(CO)12-PS Composites (Prepared with DMF) . .. 146

65 TEM Photomicrograph of Low Concentration of
Fe3(CO)12-PSF With Added Pluronic Surfactant. .. .. 147









66 TEM Photomicrograph of High Concentration
Fe3(CO)12-PSF With Added Pluronic Surfactant.. .. 147
67 TEM Photomicrograph of Fe3(CO)12-PSF Showing
Segregation of Phases. .. .. .. .. .. .. 48

68 TEM Photomicrograph of Fe3(CO)12-PSF Indicating
Colloidal Nature of Aggregates . ... .. .. .. 148

69 TEM Photomicrograph of Co2(CO) -PS Film
(Air Decomposition of Carbonyl) 151

70 TEM Photomicrograph of Thermally Treated
Co2(CO]8-PVF2 ... .. .. .. ... .. .. 151
71 SAXS Intensity Versus Angle Plot for PVF
Film Showing d Spacing of 63A. . ... .. .. .. 154

72 SAXS Intensity Versus Angle Plot for Iron-PVF2
Composite . .. .. .. .. .. .. .. ... 154

73 2 D Contour Map for SAXS Experiment in Iron-
Polyvinylidene Fluoride Films . .. .. .. .... 156

74 2D Contour Map for SAXS Experiment in Iron-
Polyvinylidene Fluoride Films .. . ... .. ... 156
75 Guinier Plot for Fe (CO)12L-Polystyrene Films,
Thermally Treated (SAXS Data Corrected for
Background and Polystyrene Matrix Scattering) .. .. 157

76 Optical Photomicrographs of Electron Beam
Decompositions of PC-Fe(CO)5 Film at 30 KeV for
500 Seconds (a and b) ................. 167

77 Control Polysulfone Sample Viewed at 100 and
60 KeV in Selected Area (Dark Regions and Dense
Layers of Carbon Support Film) . .. .. .. ... 168

78 Fe(CO)5-Polysulfone Sample Viewed at 100 KeV in
Selected Area (Arrow Indicates Decomposition
Region on Sample) .. .. .... .. .. .. .. 168

79 Fe(CO)5;-PSF Sample Line Decomposition at
100 KeV and 100,000x Magnification (Selected
Area Decomposition Shown Above Line Decomposition) . 170

80 Thermally Treated Fe(CO)5-PC Film Showing CO
Evolution in PC Matrix .... .. .. . .. 173

81 SEM Photomicrograph of Untreated Fe(CO)5-P
Film Solvent Etched in Benzene 75









82 SEM Photomicrograph of UV Treated Fe(CO)5-PC
Film Solvent Etched in Benzene. 175

83 Proposed Mechanism for Decomposition Reaction
of Fe(CO)5 in a Polycarbonate Matrix. . ... .. 179

84 Proposed Mechanism for Decomposition Reaction of
Fe(CO)5 in a Polyvinylidene Fluoride Matrix . ... 179


XV111





Abstract of Disserration Presented to the Graduate Council
of the University of Florida in Partial Fulfillment of the
Requirements for the Degree of Doctor of Philosophy
PHASE SEPARATION OF METAL OR METAL-0XIDE MICROPARTICLES
IN SOLID POLYMER MATRICES



Cindy Lou Flenniken

December 1984

Chairperson: Eugene P. Goldberg
Major Department: Materials Science and Engineering

The aim of this study was the synthesis and characterization of

polymer composites in which microscopic metal or metal-oxide particles

were incorporated by thermal, photolytic, or electron beam

decompositions of solid solutions of organometallic complexes in

polymers. This study involved the preparation of metal carbonyl solid

solutions in polymers and the phase separation decomposition of the

organometallic complex to form metal or metal-oxide dispersions of very

small particle size (15-250/).

This approach to the preparation of metal-polymer composites may

afford unique opportunities for investigating fundamental aspects of

nucleation and growth of metal and metal-oxide clusters in the solid

state as well as the properties of such microparticles as a function of

cluster size, concentration, and environment. Many interesting





catalysis, microelectronics and imaging applications appear possible for

metal-polymer composites made by this phase-separation method.

As model systems Fe(CO)5, Fe3(CO)12, and Co2(CO)8 were studied as

the organometallic complexes in the following polymer matrices:

bisphenol polycarbonate (PC), polyvinylidene fluoride (PVF2),

polydimethylsiloxane (PSi)], polystyrene (PS), aromatic polysulfone

(PSF), and polymethylmethacrylate (PMMA). Polymers with varied

molecular structures, morphologies, and mechanical properties were

selected to assess interactions with active metal species and evaluate

effects upon physical properties.

The hydrocarbon solvents used to form the homogeneous metal

carbonyl-polymer solutions and the chemical structure of the polymer

were shown to significantly affect chemical interactions with the

organometallic complexes and to influence the morphologies and

properties of the composites. The microparticles produced were observed

to be a function of matrix polymer interactions as evidenced by the

formation of FeF2 and FeS in PVF2 and PSF. The reactive carbonyl

species formed during the solid state decomposition was shown to

interact with the matrix to crosslink the polymer in the case of PVF2 or

even promote degradation in the case of polycarbonate. Solid state

decomposition of carbonyls were faster than in solution. IR extinction

coefficients were also found to decrease in solid polymer matrices

presumably due to more restricted vibrational freedom.

By clarification of the variables which govern the solid state

decomposition chemistry, it is anticipated that this phase separation

technique may be improved to afford more precise control of the

decompositions and the resulting composite compositions.





1. INTRODUCTION


Electronic devices and materials derived from unique polymers will

become increasingly important technologically (1). Current electronic

materials applications often involve metals, chalcogenides, ceramics,and

other inorganic substances combined with polymers which function

primarily in an electrically passive manner as insulators or

dielectrics. Depending upon the application, the utility of available

electronically active materials may be limited by factors such as

weight, fragility, fabrication problems, corrosion, scarcity, and high

cost. The trend in high mechanical strength materials has been toward

filled or composite polymers. Advantages commonly associated with

polymeric materials include high strength-to-weight ratios, toughness,

low cost, molecular tailoring of desired properties, and the ease of

fabrication into complex shapes. Our objective in this research has

therefore been to explore opportunities for preparing novel metal-

polymer composites with special interest in obtaining materials with

unusual electronic or mechanical properties.

Metal-polymer composites are conventionally prepared by physically

mixing metal or metal oxide powders with a polymer and fabricating

mixtures by extrusion or molding into materials containing large metal

or metal oxide particles dispersed in the polymer matrix. Particle size

and shape, volume loading of filler, fabrication technique, specific









gravity of the metal versus the polymer, and compatibility (surface

interactions) of the polymer with the metal are important factors

affecting the properties of such conventional metal-polymer composite

systems.

Fine dispersions of metals in a polymer matrix could combine the

interesting electrical properties of metals with the useful mechanical

and thermal properties (such as strength, rigidity, high softening

temperature, and good corrosion resistance) of polymers to produce a

variety of new technologically significant materials (2). Applications

for such compositions include (a) photorecording and microcircuitry

materials--by selective formation of free metallic species in polymer

matrices by special control of phase separations; (b) magnetic materials

for magnetic recording or magneto-optical switching devices; (c)

polymers of high mechanical modulus and/or electrical conductivity;

(d) metal-polymer catalysts.

In contrast to conventional composites or filled polymer

compositions, which are prepared by dispersion of particles or fibers in

a polymer matrix, this study has involved the preparation of metal

carbonyl solid solutions in polymers and the phase separation

decomposition of the organometallic complex to form metal or metal oxide

dispersions of very small particle size. This novel preparative method

has been shown to yield microscopic metal or metal oxide particle

dispersions by thermal, photolytic or electron beam decompositions of

the solid solutions of organometallic complexes in polymers. The

nascent metal species or organometallic intermediates created by UV

photolysis, thermal, or electron beam treatments in a polymeric matrix

are highly reactive species. This study has been aimed at





understanding chemical reaction pathways as well as the structure and

properties of the complex composite products.

The growing number and variety of organometallic complexes that can

serve as reagents in selective organic synthesis reactions is a very

exciting aspect of contempory synthetic chemistry (3). Many metal

carbonyl cluster complexes have been prepared in recent years (4), and

extensive studies of the intramolecular chemistry and ligand exchange

reactions (5) of these complexes have been carried out. Less is known

about the details of reactions in which metal-metal bonds in clusters

are formed and broken, but the Information which is available (4)

indicates that such reactions may be quite useful in developing new

transition metal systems for use in synthetic and catalytic processes.

Most metal carbonyls produce very labile complexes upon photolytic

or thermal treatments. Frozen gas matrices have most commonly been used

to study the labile intermediates (metal carbonyl fragments) generated

by insitu photolysis of parent molecules (6, 7). It has been proposed

that the driving forces for cluster complex decomposition arises from

the expulsion of a molecule of relatively stable mononuclear metal

carbonyl, leaving behind a coordinatively unsaturated intermediate which

(in the absence of other reagents) oligomerizes to form a higher

coordinatively saturated cluster (8). This suggests that in the

presence of organic ligands such decompositions might give rise to

active catalysts. In fact, numerous metal complex catalysts have been

attached to crosslinked polymer supports (9). Though the insoluble

catalysts are easily separated from reaction mixtures (unlike their

homogeneous analogs), difficulties with such polymer-bound catalysts





systems occur. Changes in activity and selectivity due to the altered

ligand environment have been observed.

In order to relate unstable species observed in a rigid matrix

frozenn gas matrix) to possible photochemical mechanisms in solution, it

has been very useful to study reactions in a variety of matrix media and

at different temperatures. Polymers have provided organic photochemists

with convenient and inexpensive matrices for trapping relatively

unstable luminescent intermediates and free radicals at temperatures

between 80-350"K (10, 11). However, the use of polymer matrices for

studying decompositions of metal carbonyls has been limited probably

because solute IR bands are generally much broader than those exhibited

by the frozen gas matrices. Nevertheless, the chemical structure of a

polymer may be readily varied to study the interaction of reactive metal

carbonyls in an inert matrix or one In which functional groups of the

polymer may react with the labile carbonyl fragments.

In spite of the many studies on a great variety of ligands much

fundamental quantitative kinetic and thermodynamic information is still

lacking regarding the simplest interconversions of the metal carbonyls

and the active metallic species they produce on decomposition. This is

also the case for reactions of polynuclear metal carbonyl cluster

compounds with CO and for reactions in which carbonyls act as

homogeneous catalysts or catalyst precursors (12). In this study we

have sought to understand how such active species may react with (and in

the presence of) a polymeric matrix, solvents, the atmospheric

environments, and with each other. Reaction with polymer chains may lead

to degradation or crosslinking of a polymer matrix whereas interactions

of active species with each other and the atmosphere may create small





metal or metal oxide clusters. A complex relationship must exist

between such active organometallic species and the molecular structure

of the polymer matrix.

As model systems for studying the synthesis and properties of

metal-polymer composites prepared by solid state phase separation,

Fe(CO)5, Fe3(CO)12 and Co2(CO)B were used as the organometallic

complexes with the following polymers: bisphenol polycarbonate (PC),

polyvinylidene fluoride (PVF2), polydimethylsil oxane (PSi), polystyrene

(PS), aromatic polysulfone (PSF) and polymethylmethacryl ate (PMMA).

Polymers with varied molecular structures, morphologies and mechanical

properties were selected to assess interactions with active metal

species and to evaluate effects upon physical properties.

This approach to the preparation of metal-polymer composites, as

already suggested, may afford unique opportunities for investigating

some fundamental aspects of nucleation and growth of metal and metal

oxide clusters in the solid state as well as the properties of such

microparticles. The resulting metal containing polymers may prove

interesting as catalysts, magnetic or conducting polymers, semi-

Insulators, and microelectronic materials (13).


















2.1 Conventional Metal-Polymer Composites

Polymeric composite materials play an increasingly important role

in meeting the materials needs of today's technological society because

of their light weight strength, flexibility, toughness,and ability to be

readily formed into intricate shapes (1). The fillers may be of a

variety of materials (glass, metal, cellulose) and structures (fibers,

powders, beads).

Metal-polymer composites are conventionally prepared by physically

mixing metal or metal oxide powders with a polymer and fabricating

mixtures by extrusion or molding into materials containing relatively

large metal or metal oxide particles dispersed in the polymer matrix.

Particle size and shape, volume loading of filler, fabrication

technique, specific gravity of the metal versus the polymer, and

compatibility (surface interactions) of the polymer with the metal are

important factors affecting properties of such conventional composite

systems. It is not easy to control dispersion homogeneity and

aggregation in such blended composites.

The concept of polymer reinforcement using fibers is important, but

there are relatively few examples of metal-polymer composites of this

type. Fiberglass or graphite fiber-reinforced plastics are two

commonplace examples. More currently, boron-reinforced epoxy composites

are in development for jet engine and fuselage components and even as

complete aircraft wings (14).


2. BACKGROUND





Another type of composite technology is paints made conductive by

polymer suspension of small metallic particles in a polymer latex or

solution (acting as carrier fluid and binder). The silver or carbon

conductive paints used in electron microscopy to conduct the electron

beam current away from the sample to prevent beam damage are examples of

this form of metal-polymer composite. Magnetic recording tapes also

represent metal-polymer composite technology. Modern recording tapes

call for the coating of fine uniform dispersions of oriented particles

(which exhibit high coercivity) onto thin polymeric films. Particles of

iron oxide, v-Fe203, were first used in the 1930's as magnetic recording

materials (15). A list of the most recently improved particles used for

this purpose include Fe203-Fe304 compositions, cobalt-modified fron

oxides and chromium dioxide. The desire to increase the coercivity of

these materials corresponds to more attractive magnetic properties but

seem also invariably to be associated with thermal, mechanical, or

chemical instability (16).

The focus of this research has been upon the preparation and

characterization of metal-polymer composites; more specifically, a novel

phase separation method has been investigated to create a homogeneous

dispersion of small metal or metal-oxide particles in a solid polymer

matrix. Such fine dispersions of metals in polymer matrices could

combine the useful electronic and magnetic properties of metals with the

mechanical and thermal properties (such as strength, rigidity, high

softening temperature, and good corrosion resistance) of polymers to

produce a variety of new and interesting materials (2).





2.2 Metal-Polymer Composites by Solid State Phase Separation
The aim of this study was to prepare metal-polymer composites by a

phase separation method to create a homogeneous dispersion of metal or
metal oxide in a solid polymer matrix. The approach used was to

dissolve organometallic complexes in polymer solutions, cast films and

then decompose the organometallic-polymer films using thermal,

photolytic or electron beam energy (see Figure 1). This method of

preparation should overcome many of the difficulties associated with

mechanically blended composites and provide a better understanding of

organometallic complex chemistry.

Nascent metallic species created by UV photolysis or by thermal or

electron beam decomposition in a polymer matrix should be highly

reactive (17, 18). The primary process is the photochemistry of simple

carbonyls is the formation of a coordinatively unsaturated species

M(CO)n-1 (19).


M(CO), --k-e [M(CO) ]* ---- M(CO)n-1+ CO


Continued irradiation often leads to further carbonyl abstraction.

Stabilization of these intermediates may be brought about:



a) by recombination with CO to form M(CO),

b) by reaction with a metal carbonyl:


M(CO)n-1 + m Mx(CO)y ----- M(CO)n-1[Mx~(CO) ]m





Polymer Soln.


Organometallic


S l.Ha m seneous

Film Cost


UV, Thermal, or
Electron Beam


Polymer Matrix
with Metal or
Metal Oxide
Domolns


Figure 1. Concept for Producing Metal-Polymer Composite
Films.





c) by addition of a ligand L having n- or rr-donor properties:

M(CO)n-1 + L ------ LM(CO)n-_1

d) by reaction with a molecule X-Y according to


M(CO]n-1 + X-Y -----) (CO)n-_1M,



The emphasis of this research has been directed to process c where

the ligand is a polymeric system (polymer and solvent). In this case

the metal ions/radicals may follow several pathways of reaction: (a)

interaction with remnant solvent to produce further photosensitive

complexes (20), (b) attachment to the polymer, a process that can lead

to various polymer degradation, crosslinking, or metal attachment

processes, and (c) aggregation of metal species to form very small
microcluster.

These processes may occur and be significantly influenced by

reducing or oxidative atmospheres. The chemistry of the system (i.e.

the relative importance of each process) Is determined by the structure

of the solid polymer matrix (molecular weight, crystallinity and

polarity), remnant solvent and its structure, and by the rate of the

decomposition process as compared to the diffusion time of the active

metal species and gases through the solid polymer matrix (21).


2.3 Chemistry of Metal Carbonyls

2.3.1 Disproportionation Reactions of Metal Carbonyls

The me-tals of group VII and VIII (iron and cobalt) are known to

react with strong n-donars to form ionic metal carbonylates in a





so-called "base reaction" (19, 22, 23). This reaction is shown below

for both Fe(CO)5 and Co2(CO)B with dimethylformamide (DMF).



(1) 5 Fe(CO)5 + 6 DMF-----) [Fe(DMF)6][Fe4(CO)131 + 12 CO
850C 1


(2 C2CO)8I + 12 DMF --- Co (DOMF)6[Co(CO)4 2 + 8 CO



Fe3(CO)12 is not observed to undergo this disproportionation
reaction with DMF. The base reaction is of interest to our study

because PYF2 is soluble only in DMF for our film casting technique. If
the resultant cation-anion complex is treated with an acid (hydrogen

donor), metal hydrides are formed. Compounds of transition metals (M)

with M-H bonds are of critical Importance in many catalytic reactions

(24). The first known complexes with M-H bonds were in fact the hydride

carbonyls H2Fe(CO)4 and HCo(CO)4 made in the 1930's by W. Heiber. The

hydride cobalt tetracarbonyl has been demonstrated to react with

styrenes having various alkyl substituents (bound either to the a- or B-

position of styrene) in hydroformylation and hydrogenation reactions

(25). These reactions appear to be subject to steric effects of the
substituents.

2.3.2 Metal Carbonyl Reactions with Monoolefins
One aspect of metal carbonyl chemistry pertinent for consideration

here is the interaction with Isolated double bonds. Molecular orbital

theory suggests that a double bond will coordinate with the metal in a

manner similar to carbon monoxide (24, 26). The substitution of one CO

by a double bond would therefore produce a complex containing the C=C

group: i.e. Fe(C=C)4. Indeed, in recent literature, there are several





reports concerning the formation of olefin-iron tetracarbonyl

complexes. The first example was the preparation of acrylonitrile iron

tetracarbonyl from acrylonitrile and Fe2(CO)9 (27). The same complex

may also be obtained with Fe(CO)5 upon irradiation (28). This suggests

that Fe(CO)5 is first converted to Fe2(CO)9 and then reacts with the
double bond. The x-ray data for this complex shows that the iron atom

is bonded to the C=C group rather than to the -CE-N or the nitrogen atom

(29).

Kirch and Orchin (30) observed the oxo reaction of dicobalt

octacarbonyl with olefins to form aldehydes proceeds by an olefin cobalt

hydrocarbonyl intermediate. It has further been established that the

rate of the oxo reaction is inversely proportional to the carbon

monoxide partial pressure. As discussed earlier in this text, the

cobalt hydrocarbonyl may be prepared under CO by disporportionation with

dimethylformamide according to the equation



3 Co2(CO)8 + 12 DMF ---) 2 Co(DMF)6 (Co(CO)4 2 + 8 CO


Co(DMF)6[Co(CO)4 2 + 2 HC1 ---) 2 HCo(CO)4 + CoC12 + 6 DMF


Dubois and Garrou (31) observed that Co2(CO)B catalyzed reaction of

acrylonitrile with CO/H2 in methanol is both temperature and ligand
sensitive. The selectivity and activity of cobalt catalytic complexes

can be altered dramatically by the presence of ancilliary ligands such

as amines and phosphines (32). It has also been shown that when

monoolefins of various structures are heated with iron carbonyl,

geometrical and positional isomerization occurs (33) and that small









amounts of polar substances appear to promote this isomerization (34).

Through variation in solvent systems, it was concluded that this

isomerization also occurred via the iron carbonyl hydrides. It is well

known that iron carbonyl hydrides are active catalysts for hydrogenation

and isomerization of olefins (35, 36) as are the cobalt carbonyls (37,

38).

2.3.3 Metal Carbonyl Reactions with Conjugated and Nonconjugated Dienes

The most common feature of Fe(CO)5 chemistry is its interaction

with conjugated or nonconjugated dienes, with the formation of the

Fe(CO)3 moiety. Diene-iron carbonyl complexes were first prepared by

two entirely different synthetic methods. In 1930, Reihlen and

coworkers (39) obtained butadiene-iron tricarbonyl by reaction of

butadiene with iron pentacarbonyl, and in 1953 Reppe and Yetter (40, 39)

reported the formation of organoiron compounds (since shown to be diene-

iron carbonyl complexes) following reaction of acetylene with iron

carbonyls. The diene-iron carbonyl complex was further investigated by

Hallam and Pauson in 1958, and was found to resist hydrogenation and not

to undergo Diels-Alder type reactions (41). Spectroscopic and chemical

evidence led to the suggestion that the butadiene molecule remained

essentially intact in the organoiron complex. Furthermore, it was shown

that an analogous compound, 1,3-cyclohexadiene-iron-tricarbonyl could be

prepared in a similar manner from 1,3-cyclohexadiene and Fe(CO)5. It
was therefore concluded that the diene system adopted a :is arrangement





of double bonds within the complex rather than trans. The following

structure was proposed:






~2 CH CH--

Fe(CO)3





The butadiene structure is assumed to be nearly planar, with the

iron atom lying below this plane approximately equidistant from the four

carbon atoms of the diene system. The nature of the diene-iron bonding

was presumed to involve interaction of the Fe atomic orbitals with

molecular orbitals of the diene system as a whole. The structure is

therefore more analogous to a-bonding in ferrocene (24) than to 6-type

interactions implied by Reiblen (39). The conjugated diene system was

considered to be essential for the formation of iron derivatives of this

type.

It is now known that conjugated dienes and a,8-unsaturated

carbonyls add hydrogen under oxo conditions (42) and were further

demonstrated with isoprene by Kirch and Orchin. In 1961 Arnet and

Pettit (43) and Stone et al. (44) reported a reaction involving the

rearrangement of non-conjugated dienes to corresponding conjugated

isomers of diene-iron tricarbonyl compounds following treatment with

iron carbonyls [Fe(CO)5 and Fe3(CO)12 respectively]. This was

confirmed by Whitesides and Neilan (45) and postulated to proceed by

metal hydride intermediates. Butadiene-Fe(CO)4 and butadiene-[Fe(CO)4 2





have also been prepared from butadiene and Fe2(CO)9 (46). In these

complexes, one of the carbon double bonds is n-bonded to the Fe atom.

The butadiene-Fe(CO)4 complex reacts with HC1 to form the 1-methyl-n-

allylchloroiron tricarbony1 complex, possibly through intermediate

formation of a methyl-n-ally1-Fe(CO )4 cation (46,47).

Interesting correlations exist between the electronic structure of

the carbonyl groups for various diene-iron tricarbonyl complexes.

Cationic ligands shift the C=0 Infrared absorption frequencies to higher

values (48,49,50,51). In butadiene-iron tricarbonyl itself, there are

two regions of carbonyl absorption (52), a narrow Intense band at 2053

cm-1 and a broader band which is resolved into two maxima at 1985 and

1975 cm-1 (53). In analogous derivatives the gross structure of these

bands is retained. However, positions shift according to the nature of

the ligand (24,54).

The majority of diene-iron tricarbonyl complexes have been made by

direct reaction of dienes with one of the three common iron carbonyls

[Fe(CO)5 Fe2CO)9 and Fe3(CO)12], either by simply heating the reagents

together or by photochemical reaction. Using Fe(CO)5, a temperature of
120-160"C is required and reactions are conducted in sealed tubes. In

equivalent reactions using Fe3(CO)12, the products isolated are (with

very few exceptions) diene-iron tricarbonyls (53).

The previously mentioned work with diene systems has demonstrated

the reaction chemistry of iron carbonyls and opened the door to further

interesting work for incorporating organometallics into polymers and for

polymer synthesis. Indeed, reactions of chromium, molydenum, tungsten,

and iron polybalogenated organic compounds have been used to initiate

the free radical polymerization of vinyl monomers (34).





2.3.4 Metal Carbonyl Reactions With Halides

Halogenated hydrocarbons have been found to be involved in

reactions with metal carbonyls either as oxidizing agents (55,56) or

entering as additional molecules (57,58) besides producing substitution

products (59). Metal carbonyl chemistry in the presence of halides is

of great importance for our consideration as we have selected methylene

chloride as a polymer solvent. Iron pentacarbonyl reacts with organic

halides if two conditions are met. First, the halide must be activated

by at least one, and preferably two, groups such as cyano, phenyl or

halogen (in decreasing order of effectiveness). Second, there must be

at least two halogens on the same carbon atom or in very close proximity

to each other. For example, dichloradiphenylmethane 1(C6HS)2CCl2]

reacts readily with Fe(CO)5 (55) according to the following reaction



2 Fe(CO)5 + 2 (C6H5)CC12 ---4 10 CO + 2 FeC12 + (C6H5 2C=C(C6HS 2


Furthermore, 1,2-dichloro-1,1,2,2-tetraphenylethane, even though it does

not have geminal halogen atoms, reacts readily with Fe(CO)5 according to

the following reaction



Fe(CO)5 + (C6HS)C1C-CC1(C6HS )2 --) 5 CO + FeC12 + (C6HS 2C=C(C6H5 2


Molecular models of this halide reveal that the halogens (in the rotamer

with the closest proximity of halogens) are as close as in a geminal

dihalide and it is perhaps not surprising that reaction occurred. In no

case does reaction occur with monohalogenated compounds, even when the

halide is as strongly activated as that in triphenylmethyl chloride.





Another type of reaction leads to the formation of iron complexes

of the general form FeC12L2 or FeC13L, where L stands for ligands such

as formamide, n-methylformamide, aniline, benzamide, acetamide,

triphenylphosphine and triphenylarsine (59,60). All three iron

carbonyls [Fe(CO)5, Fe2(CO)9, Fe3(CO)12] react with these various

ligands in the presence of chloroform to form halide adducts of the type

FeC12L2 or FeCI3L. Solvents such as carbon tetrachloride,
tetrachloroethane, and benzylchloride lead to similar oxidation

reactions. Dichloroethane and methylene chloride lead to partial

decomposition giving rise to impure compounds. The organic ligands

which form FeC12L2 complexes are formamide, methylformamide and

acetamide, whereas those which form the FeC13L complex are aniline,

benzamide, triphenylphosphine and triphenylarsine. All ligands which

have phenyl groups give rise to an iron complex of higher oxidation

state.

The formation of complexes of the type FeC12L2 or FeC13L by the

reaction of iron carbonyl with various ligands in the presence of

halogenated hydrocarbons may be described by the following reaction

mechanism. A substituted carbonyl complex of the type Fe(CO)4L forms

and reacts with halogenated hydrocarbons to give rise to a halide

complex. To verify this, Fe(CO)4L was allowed to react with chloroform,
and the halide complex was obtained. In order to establish the reaction

mechanism, Singh and Rivest (59) tried unsuccessfully to isolate the

product formed from the oxidation reaction in chloroform. Therefore,

they replaced the chloroform with diphenyldichloromethane and allowed

this to react with Fe(CO)4(AsPh3). They could then isolate





tetraphenylethylene and the halide complex. A possible mechanism for

the reaction, therefore, can be suggested:



Fe(CO)5 + 2 L --) Fe(CO)3L2 + 2 CO



2 Fe(CO)3L2 + 2 (C6HS 2CC12 ---) 2 FeC12L2 (6H5 2C=C(C6HS 2 + 6 CO


In view of the foregoing, the decomposition of metal carbonyls in

polymers of differing molecular structures (and using different mutual

film casting solvents) may be expected to lead to significant

differences in polymer interactions and decomposition chemistry.

Polymers with different backbone structures and side chain groups were

therefore studied to help explain the different physicochemical

interactions which might occur with iron pentacarbonyl and other

organometallic complexes.

Present electronically active materials are composed primarily from

a variety of metals, semi-metals, ceramics and other inorganic

substances with polymers used in the the electrically passive manner as

insulators or dielectrics. Though synthetic polymers are generally

regarded as electrically passive, the last two decades have seen the

development of plastic conductive fibers (61,62). This has been

achieved by mixing a high-modulus aramid polymeric solution with an

organic metal solution from which fibers are spun. To increase

conductivity, the fibers have been doped with iodine.

Yet another area of development is the use of metal polymer

composites in novel circuitry applications. Printed circuit boards are

commonly made by the photoetching process. Material losses are great





and the process is extremely complicated (high cost, low yield). Tabei

et al. (63) investigated a new, much simpler process in which metal

salts of organic acids are coated on an insulating substrate. When

irradiated with UY light in a pattern of the desired circuit, metal ions

(silver) are liberated so as to deposit the metal atoms to form

aggregates. Electroless plating on the pattern is carried out with the

deposited silver metal functioning as the nuclei for plating. This is a

simple process to form printed circuits and can be applied to other

fields such as photographic material, inlaid work, and photomasks.


2.4 Related Studies

Aside from the more conventional metal-polymer composites, there

are other functions which compositions of metals and polymers May

serve. There is also an active interest in the magnetic properties of

particles small enough to contain single domains when there Is no

applied magnetic field. Magnetic fluids, or ferrofluids, exhibiting

this characteristic are colloidal suspensions of ferro- or ferri-

magnetic particles. This base of knowledge Is of interest in our

studies in that it provides an understanding of the interactions that

may occur between small magnetic particles and the carrier fluid (the

polymer matrix). In addition, carbonyls decomposed in the presence of

organic ligands have been demonstrated to be active catalysts for

acetylene trimerizations. Polymers also provide organic photochemists

with a convenient matrix for Isolating unstable and air sensitive

organometallic complexes and clusters for study.

2.4.1 Ferrofluids and Other Colloidal Dispersions

There is an active interest in magnetic properties of stable

dispersions of subdomain magnetic particles--ferrofluids. Particles





that are small enough, on energetic grounds, to contain but one domain

without an applied magnetic field are termed subdomain (64). Ferrofluid

chemistry is of interest in this metal-polymer composite study in that

it provides an understanding of the interaction between the small

magnetic particles and their dispersant in terms of concentration,

dispersant chemical structure and surfactants.

A magnetic fluid or ferrofluid is a dispersion of ferro- or ferri-

magnetic particles in a carrier fluid. The particles are coated with

long chain surfactant molecules which prevents formation of agglomerates

via the short range van der Waals interactions (65). Ultrafine

particles (100 A diameters) are used to reduce the longer range

Interactions between magnetic particles. The magnetic response of a

ferrofluid is due to the coupling of response in individual magnetic

particles with a substantial volume of the surrounding carrier liquid

(66). Thus, when a magnetic field is applied to a ferrofluid this force

is transmitted not only to the particle but also to the associated

li quid phase. Coupling of these magnetic particles to the bulk liquid

phase is accomplished by the use of surfactants (compounds with

functional groups which adsorb onto the particle and still are solvated

by the carrier liquid). A typical compound is 01eic acid--a molecule

containing a polar carboxylic end group that absorbs on the particle

surface and a hydrocarbon moiety that Is similar to the dispersing

medium in chemical structure (66). By proper choice of stabilizing

agents, magnetic properties may be conferred to a wide variety of liquid

mediums. Depending on the structure and molecular weight of the

dispersing agent and selected solvent, the effective thickness of the

sheath may be varied from 30 to 1000 A (67,68). It is the solvated









sheath which is responsible for the stability of very small particle

suspensions.

The concept for preparing uniform dispersions of metals by phase

separation in media of low dielectric constants is suggested by some

metal colloidal literature. For example, thermolysis of transition-

metal carbonyls in fluids under an inert atmosphere is a well known

technique for the preparation of pure metal powders and a widely used

process for the preparation of ferromagnetic cobalt particle dispersion

(21,69). In 1966, Hess and Parker (21) thermally decomposed dicobalt

octacarbonyl in solutions of polymer dispersants to form stable colloids

of discrete particles which were separated by polymer coatings. In

their work, they studied a large number of purified, uninhibited polymer

and solvent systems. High concentrations of dicobalt octacarbonyl to

polymer were used (typically 75 wt%). Dicobalt octacarbonyl (23 g) was

dissolved in an organic solvent (188 ml). This solution was placed in a

three-necked round-bottomed flask equipped with stirrer and condenser.

Sufficient polymer was added to yield desired metal-to-polymer ratios

and the system closed and stirred 5 minutes. The contents were then

rapidly heated to achieve decomposition of the carbonyl. Carbon

monoxide evolution was monitored and when it ceased the reaction mixture

was cooled with continued stirring. These solutions were then film cast

onto Mylar to 0.1-0.2 mm thickness and dried with infrared lamps.

Magnetic properties were measured with a B-H meter with a 2000 gauss

field strength. Electron photomicrographs of a 1:10 dilution of the

cobalt particle sample to solvent were used to measure particle size.

Preparation of single-domain ferromagnetic cobalt particles with

good magnetic properties depends on a delicate balance between





dispersant polymer, solvent and the growing metal particle. Hess and

Parker concluded that variation of polymer composition, molecular weight

and solvent affects particle size and colloid stability. They

specifically observed that the most successful polymers are linear

addition polymers of high molecular weight having relatively nonpolar

backbones with pendent polar groups (i.e. amides) every 200 or so

backbone carbon atoms. Stabilization of the cobalt particles results by

adsorption of the polar group of the polymer to the metal particles to

form a film which sterically hinders carbonyl decomposition at the metal

surface, thereby retarding particle growth. The solvents for these

systems should be less polar than the most polar groups in the polymer

to minimize competition with the polymer-metal particle interaction.

Though a number of condensation and addition polymers can act as the

dispersant, Thomas (69) observed that polymers containing a relatively

large percentage of highly polar groups promote the growth of smaller

particles and that higher polymer concentrations produce a similar

trend.

Hess and Parker's method was successfully applied to the

preparation of small iron particles by Griffiths et al. (70) in polymer

solutions. The macromolecules reportedly act as dispersing and

stabilizing agents for the iron. Stable colloidal dispersions

(50-100 A) of zero valent iron were obtained by thermolysis of low

concentrations of Fe(CO)5 in polymer solutions. All specimen

preparation and manipulation were carried out in a nitrogen

atmosphere. Magnetic measurements were made on a vibrating sample

magnetometer. Diffraction and morphological characterization were

accomplished using a electron microscope. Exposure of dispersions to









the atmosphere decreased the observed magnetic moment due to formation

of an oxide film on the particles. Chlorinated solvent based

dispersions were observed to generate chlorine which destroyed the

u-Fe203 spinel film passivity and promoted reaction with water to give
8-Fe00H.

In 1979, Smith and Wychick (71) prepared stable colloidal

dispersions (50-150 A) of zero-valent iron by the thermolysis at

130-160'C under argon of Fe(CO)5 in dilute solutions of functional

polymers. The kinetics of the decomposition were monitored by following

rates of CO evolution. Infrared spectral analysis was used to follow

the progress of the reaction and GLC to determine either relative

concentrations of Fe2(CO)9 and Fe(CO)5. Completion of the thermolysis

reaction yielded a strongly superparamagnetic Fe0 dispersion. The

percentage of iron in final dispersions was determined colorimetrically

or by atomic absorption. Magnetic properties were evaluated using a

vibrating sample magnetometer and transmission electron microscopy used

to determine particle size. A "locus control" formalism of particle

nucleation and growth was proposed to describe the formation of these

colloidal dispersions (71). It was suggested that the polymer served as

a catalyst for the decomposition of the metal carbonyl and induced

particle nucleation. Selected polymer systems were termed "active"

because the initial and overall rate of decomposition of Fe(CO)5, as

determined by the rate of CO evolution, was much faster in the presence

of the active polymers (those with an amine function) than in solvent

saone.









Using the refluxing technique of Hess and Parker, Berger and Manuel

(72) noted that treatment of polybutadienes with iron carbonyl resulted

in formation of polymers containing tricarbonyl complexes. Hampton's

infrared analysis technique (73) was used to show the distribution of

free double bonds remaining. The infrared spectra exhibited a pair of

carbonyl stretching frequencies in the 2000 cm-1 region characteristic

of tricarbonyl (conjugated dine) iron complexes. A combustion analysis

for iron showed 20 wt% Fe(CO)3. The completing reactions for the

polybutadiene were done with either iron pentacarbonyl or triiron

dodecacarbonyl as the reagent in xylene or benzene, respectively. They

observed substantial molecular weight breakdown in the recovered polymer

when a hydrocarbon alone is used as solvent and that this could be

avoided by addition of small amounts of polar nonacidic solvents

(alcohol, ketone,or ether). Basic solvents and high temperature favored

this incorporation of the iron carbonyl groups into the polymer to

produce ferromagnetic products with enhanced thermal stability.

Chantrell et al. (65,74) studied the effect of dilution on the

stability of a ferrofluid consisting of cobalt particles in toluene.

The cobalt/toluene ferrofluid was prepared by the thermal decomposition

of dicobalt octacarbonyl (using the method of Hess and Parker) with

sodium diocty1 sulphosuccinate as surfactant (to stabilize the resultant

colloidal dispersion and control particle size). Magnetic measurements

were made at room temperature on a vibrating sample magnetometer. This

data indicates the superparamagnetic character of the cobalt/toluene

ferrofluid. The particle size distribution parameters were calculated

from the low and high field magnetic behavior and found to be 50 A.





The long term gravitational stability of the original

cobalt/toluene fluid and commercial ferrofluid was tested and the change

in concentration profile found to be negligible for both over a 50 day

period. In order to test the dilution stability of the cobalt/toluene

fluids, several toluene dilutions were prepared from the original

fluid. The concentration profile clearly shows apparent instability of

the fluid with the formation of larger aggregates of particles on

dilution of the ferrofluid. Chantrell et al. proposed that this

instability is due to a disturbance of the balance which must exist

between the adsorbed surfactant on the surface of the particles In order

to maintain the equilibrium between adsorbed and free surfactant. They

concluded, therefore, that when making measurements on ferrofluids in

large applied fields the magnitude of the field induced aggregation

should be measured.

2.4.2 Organometallic Complexes as Catalysts

Another aspect of interest in composite chemistry is the use of

organometallic complexes that serve as reagents in selective organic

synthesis reactions (3). The feasibility of this was demonstrated by

Vollhardt (8) in 1975 with the photochemical decomposition of

n5-cyclopentadienyl-cobalt dicarbonyl in the presence of organic ligands

to produce active catalysts for acetylene trimerization at temperatures

lower than that normally required. Further information is available

which indicates such reactions may be quite useful in developing new

transition metal systems for use in synthetic and catalytic processes.

However, with various ligands much fundamental quantitative kinetic and

thermodynamic Information concerning metal carbonyls is still lacking in

the literature (12,19,60).





Dispersed metal crystallites have been formed by thermal

decomposition of metal carbonyl complexes anchored on both organic and

inorganic supports (75). Derouane and Nagy (76) report that adsorption

of Ni(CO)4 on various oxide materials followed by thermal decomposition

leads to the formation of stable, porous metallic crystallites. The

interaction of the carbonyl with the support, the nature of the

decomposition mechanism and resulting catalytic activity of the products

warrant further investigation. These systems are capable of leading to

improvements with respect to preparation of multimetallic catalysts by

direct clustering of metals in a zero-valent state.

Metal-containing polymers offer potential utility as catalysts.

Organometallice polymers for this purpose may be prepared by derivatizing

preformed organic polymers with organometallic functions (77), or by

preparing monomers which contain organometallic units and polymerizing

these monomers (13,78). Many transition metal complex catalysts have

been attached to crosslinked polymer supports. Difficulties with

polymer-bound catalyst systems include changes in activity and

selectivity due to the altered lIgand environment (79); steric

constraints imposed by the presence of the polymer structure in the

vicinity of the attached complex (79,80); reduced activity due to the

thermodynamic exclusion of reactant molecules from the solvent-swollen

polymer support phase (81); and intraparticle mass transport effects on

the reaction rate since the reactants must diffuse through the solvent-

swollen polymer matrix to reach active complexes therein (79,81).

With regard to hydrocarbon media, it has been found that

appropriate modification of alky1 methacrylates can result in

outstanding dispersants for use in automotive lubricants (82). A better





understanding of the chemistry of these systems is provided in the

literature and is relevant to this study. Fontana and Thomas (83) found

that Incorporation of 17 mole I N-viny1-2-pyrrolidone (VP) into

polylauryl methacrylate (PLMA) increased the adsorbed surface layer on

silica to 200 A (relative to 30 A for samples without VP). They

proposed that this thicker film results from preferential adsorption of

the smaller number of more strongly polar pyrrolidone groups.

In a similar study, Fontana (84) was able to quantitatively

demonstrate this surface adsorption phenomenon (via infrared

spectroscopy) for a 20:1 mole ratio of alky1 methacrylate-polyglycol

methacrylate copolymer (PAM-PG). The copolymer molecular weight was

approximately 310,000 as determined from intrinsic viscosity

measurements. Polyglycol methacrylate was the acting dispersant.

Details of the determination of the pertinent extinction coefficients

and calculations of adsorbed ester segments are described in a previous

publication (83). The data demonstrate the exclusion of ester segments

from attachment to the surface because of preferential adsorption of the

polyglycol ether segments. Thus, the resultant polymer configuration is

more extended away from the surface and aids in stabilizing the

colloidal dispersion by virtue of its polar substiltuents.

2.4.3 Metal-Polymer Composites Prepared byAdsorption of Metal
Carbonyl Solution onto ims

De Paoli (10,11) found that by incorporating the unstable metal

carbonyls by adsorption into polymers he could work with them at ambient

conditions without the inconvenience of cryogenic temperatures. He also

noted that, in addition, the active carbonyl intermediates exhibited

higher mobility and a higher local concentration of reagents in a

polymer matrix. He used PTFE and PE (0.2 and 1.0 mm thick,





respectively) as matrices to study (via infrared spectroscopy) the

photochemical products and reaction of Fe(CO)5 with olefinic ligands
(such as ethylene, acrylic acid, methylacrylate, butadiene, isoprene,

norbornadiene). De Paoli felt that saturated hydro- or fluorocarbon

polymers would not absorb the radiation used to induce the photochemical

carbonyl process, and thus would not hinder the reaction nor themselves

be degraded.

The sorption of Fe(CO)5 in the polymer was carried out by soaking
the film in a 10% solution of it in degassed hexane. This was followed

by an ethanol wash under an inert atmosphere to prevent oxidation of

iron pentacarbonyl on the surface. Sorption of the second reagent was

achieved 1) for liquids, by soaking for the time required to saturate

it, and 2) for gases, irradiation in an immersion well filled with the

gas.

De Paoli noted that once iron pentacarbonyl was trapped inside the

matrix it did not oxidize during the time required for IR experiments.

Photochemical decompositions of the Fe(CO)5 composites was done using a

Philips HPK-125 W Iamp with Vycor filters and an adapted Phillips

HPLN-125 W lamp with Pyrex filters. Reactions were followed with a

double beam IR spectrophotometer with control films used as

references. To study the photo-fragments of Fe(CO)5, irradiated films

directly in the IR spectrophotometer. De Paoli observed that after 1 to

2 hours of visible light irradiation PTFE films treated with Fe(CO)5 and

butadiene produced spectra corresponding to butadiene iron tricarbonyl

(VCO= 2060, 1995 cm" ) and bislbutediene) ironmonocarbonyl

(VCO= 1985 cm-1). Experiments with isoprene produced similar results,
He observed that the IR spectra of these films did not change after one





week at ambient conditions or even after pumping air though them. Even

heating the films containing dieneiron tricarbonyl compounds to 160"C

produces no spectra changes. PTFE films pretreated with Fe(CO)5 and
irradiated in an immersion well filled with ethylene for 45 minutes show

no absorptions due to Fe(CO)5, only four peaks at 2092, 2029, 2014 and
1993 cm-1 (corresponding to substitution of one carbon monoxide by an

ethyl ene).

Experiments with low density polyethylene (which is highly

amorphous) allowed for higher degrees of carbonyl sorption. Irradiation

for three minutes of PE films pretreated with Fe(CO)5 and norbornadiene

produced four new IR bands: 2098, 2032, 2020 and 2010 cm-1 as well as a
band at 1730 cm-1. Continued irradiation led to an increase in the 1730

cm-1 band while the others vanished. These four initial bands

correspond to (n2-NBD)Fe(CO)4 while the 1730 cm'l absorption corresponds
to a carbonyl inserted into a dimer of NBD.

Photolysis of PE-Fe(CO)5-MMA for one minute shows complete

disappearance of the vC0 band of Fe(CO)5 and the formation of four new
bands at 2098, 2032, 2020 and 1996 cm-1. These bands correspond to the

compound methylacrylate-iron tetracarbonyl, a compound of the type

Fe(CO)4L. Further photolysis does not produce additional changes.
In solution the photochemical substitution of carbon monoxide in

Fe(CO)5 is assumed to follow an SN1 type mechanism (19) and to decay to
the very reactive Fe(CO)4 species. De Paoli assumed this species to

also to be formed in the polymer matrices. UY treatment of a PE-Fe(CO)5
film showed twro new absorptions at 2063 and 1973 cmn-1 and the still

remaining Fe(CO)5 band at 1998 cm- Kinetic measurements indicated a
first order rate constant of 6.5 and 2.5 x 10-3 sec-1, respectively.





The bands at 1973 and 1998 cm-1 are assigned to the photo-fragment

Fe(CO)4 and the other to a product of further fragmentation, probably

Fe(CO)3 (11).

De Paoli suggested that the molecules of iron pentacarbonyl are

lodged in close proximity within the amorphous sites of the polymers,

yet are sufficiently mobile for reaction to occur between the carbonyl

and methylacrylate (or other second reagent) and to allow relatively

rapid desorption of the CO out. De Paoli, however, assumed that there

was no chemical effect of irradiation upon the polymer and hence that

there was no Interaction between the matrix and reactive carbonyl

species. Literature (2,85,86,87), however, states that photolytic

degradation of polymers does occur--thus UV stabilizers are often added

to polymers to be exposed to an external environment. This suggests

that radical species may have been formed within the polymers themselves

which would also be open to reaction with the carbonyl.

Galembeck (88) also introduced metal carbonyls into polymer films

by soaking the latter in ethanol solutions containing 10%1 iron

pentacarbonyl for a period depending on the degree of crystallinity of

the polymer. He observed that sorption of Fe(CO)5 by

polytetrafl uorinated ethylene (PTFE) films 0.2 mm thick followed by

insitu oxidation gives iron oxfde-PTFE and that upon exposure to

sunlight the Fe(CO)5 treated PTFE forms Fe2(CO)12 dimer (89). He

determined the original PTFE films to be 65% crystalline using x-ray

diffraction. He irradiated samples left soaking for various periods of

time in the carbonyl In the beam of a slide projector under which

conditions the nonvolatile Fe2(CO)12 was formed. Complete oxidation of

samples was then allowed to occur in air for three days. This was





followed by heating at 105"C to eliminate unreacted traces of Fe(CO)5

(as indicated by IR). Following a wash in IN sulfuric acid and ethanol

and air drying, iron oxide content was determined gravimetrically. Iran

oxide content ranged from 0.34 to 1.50% (w/w).

Particle size was determined using transmission electron

microscopy. Samples were mechanically sliced and electron bombarded

under vacuum in the TEM to achieve adequately thin samples for

transmission. This procedure led to the depolymerization of the PTFE

with evaporation of monomeric C2F4. Particles were observed to be

rather uniformly distributed and spherically shaped with diameters of 15

to 80 A. Size of the particles did appear to increase slightly for

higher concentrations of iron oxide. Optical adsorption measurements

were done using a photoacoustic technique. These measurements indicate

the formation of two characteristic volume distributions of iron oxide

particles in these PTFE composite materials, namely the 36 and 44 A,

average diameter sizes.

2.4.4 Metal-Polymer Composites Prepared by Casting of Metal
Carbonyl-Polymer Solutions

Cassen et al. (90) and Hitam et al. (91) used a solvent-casting

technique to prepare metal carbonyl-polymer films. Hooker and Rest also

tried the sorption technique but observed that a more uniform

concentration of the carbonyl could be obtained throughout the polymer

matrix by using this solvent-casting technique (92). They noted that

chromium, molybdenum, and tungsten hexacarbonyls in polyvinyl chloride

(PYC) showed broader and shifted carbonyl bands (between 1980 and 1974

cm-1) suggesting a polymer-solute (THF) interaction. The generation of

pentacarbonyl fragments and photoreversal are analogous to reactions

observed in hydrocarbon glasses and frozen gas matrices.





They prepared 0.8 and 2.7 wt% solutions of PVC (0.5 g each) in

1,2-dichloromethane and tetrahydrofuran (THF), respectively. To each

solution 4 mg of hexacarbonyl was added and the resulting solution

poured into petri dishes. Samples were stored in the dark and allowed

to dry overnight. This produced films approximately 40 u thick. A

water cooled mercury arc lamp was used as the photolysis source with

samples at a distance of 12 cm. Filters were used to achieve

irradiation at specific wavelengths: filter A, x < 350 nm; filter B, X >

400 nm.

Both IR and UY-visible spectra were used to follow the photolysis

of these films. Figure 2 shows the IR spectra of UV irradiated W(CO)6-

PVC films. The photoreactions of the Group VI metal hexacarbonyls

observed in PVC films were studied by Hooker and Rest. New IR bands

appeared at 2074, 1929 and 1887 cm-1 with reduction in the parent

band. A very slow reversal of the reaction occurred when the polymer

was left in the dark for several days. W(CO)6 films cast from

dichloromethane also showed the same effect but with IR bands at a

higher wavenumbers (2080, 1932, and 1888 cm-1). In contrast, however,

thermal reversal of the reaction in this film occurred after only a few

minutes. These reactions are summarized in Figure 3. Similar results

were observed for the UV irradiation of Cr(CO)6 and Mo(CO)6 films,

although these hexacarbonyls were regenerated in the dark more quickly

than the tungsten analogue.

Evidence for residual THF in films cast from this solvent was found

when polymer bands were subtracted out of the spectra. Films cast from

dichloroethane solution showed only a band at 1728 cm-1 for ketones.

These bands did not disappear with irradiation (see Figure 2).








2100 1900
100-

(before) *?
80~ -


2 60
E- 10

S 4 3 1
80 (15 min.) A< h~ 350 nm











100I I 2000 1800



80 2

S10
2 2

1 2


a: 2
C 60


40


Figure 2. Infrared Spectra from an Experiment with W(CO)6
in a PVC Film (Cast From THF): (a) Before
Photolysis, at 12*K; (b) after 15 min. Photolysis,
(n <,350 nm); (c) as (b) then Wlarming to 120*K and
(d) as (c) then Further Warming to Room Temperature
Bands Marked 1, 2, 3, and 4 are for W(CO) (CO) (THF),
W(CO)5 and Free CO, Respectively. Bands adrked ( ) and
(+) are for Oxidized F Residues and Chloroketones,
Respectively (92).






















(i) m(o)b
M (CO) ,"M(C)bC
(ii), (iii)

(vi)aTHF, -H :H
(V) -H H
-CO
CO, (i) (v
-THF


M (CO)5 (THF)


SReaction observed in solution
b
Unstable species observed in other
matrices


low temperature


Figure 3. Photoreactions of Group VI Metal Hexacarbonyls in PVC
Films: (i) X<350 nm at 12*K; (ii) hv X>400 nm at 12*K;
(iii) Warming to 2000K; (iv) Warming Between 100 and
200"K; (v) Thermal Reaction Above 200"K; (iv) X<350 nm at
298 K*(92).









respectively and a broad but well resolved vCO band at higher
wavenumbers. The bands are broad even when only 0.2 wt% of carbonyl is

present in the polymer matrix, suggesting that significant polymer

solute interactions exist. Lowering the temperature of the polymer

films made the IR spectra of the hexacarbonyls slightly sharper, but no

significant shift in band positions was noted. The vCO band at higher
wavenumbers, which has been observed for metal hexacarbonyls in various

hydrocarbon glasses at 77"K, probably results from site symmetry effects

or from some distortion of the carbonyl molecules trapped in amorphous

sites of the polymer matrix. For Group VIB metal hexacarbonyls,

McIntyre (93) suggested the mechanism of the photochromic reaction to

be





M(CO)6 + hv --, M(CO)6 --- M(CO)5 + CO (1)
(yellow)



M(CO) + 0~ M(CO) 0 (2)


M(CO)5 + CO -- M(CO)6 (3)
D = charge donor


Experiments with both degassed and nondegassed solutions

demonstrated the sensitivity of the rate of reaction (3) to traces of

oxygen. Cyclohexane was checked with mass spectroscopy for impurities
which would interfere with the reaction. Polystyrene films (0.05 cm

thick) containing 1 wt% Cr(CO)6 were prepared by casting from a









methylethyl ketone solution onto quartz plates in the dark. Photolytic

decompositions were performed. McIntyre assumed a second-order rate law

for reaction (3), the fading reaction, on the basis of the good linear

plots of 1/At vs time. Exact extinction coefficients were not
determined due to the extensive overlap of all absorption bands of

Cr(CO)6 and Cr(CO)5, but a minimum value of e = 1000 liter/mole~cm was

suggested. The values of k/E (rate constant/extinction coefficient)
were 0.317 and 0.00025 for cyclohexane and polystyrene as solvent,

respectively. This large difference is attributed to the greatly

decreased mobility of CO in polystyrene relative to cyclohexane.

Miassey and Orgel (94) dissolved M(CO)6 (M =- Cr, Mo or W) in a

polymethylmethacrylate solution. With evaporation of the solvent and UV

irradiation, the polymer assumed the color of the M(CO)5 fragment

(yellow). The M{CO)6 complex was observed to regenerate with time in
the dark. They further observed that the M(CO)5 fragments in the PMMA

matrix appeared to be stabilized by the oxygen donating groups in the

polymer itself. This is supported by the work of Hooker and Rest (92)

wherein the "naked" pentacarbonyl intermediates were not observed in a

PVC matrix which contains no oxygen donator. The same technique was

used by Mclntyre (50) to study the flash-photolysis of Cr(CO)6 in a

polystyrene matrix. He also observed the thermal back reaction to

regenerate Cr(CO)6 but noted that it occurred at a rate much slower than
in solution.





2.5 Metal-Polymer Composites by Phase Separation
2.5.1 Polymer Matrices Selected for Preparation of Metal-Polymer
Composites

As model systems for studying the preparation of metal-polymer

composites by solid state phase separation, we selected Fe(CO)5*

Fe3(CO)12 and Co2(CO)8 as the organometallic complexes and the following

polymers: bispheno1 polycarbonate (PC), polyvinylidene fluoride (PVF2)*

poly sty rene ( PS ), aromati c poly sulfone (PSF ), polymethylmethacryl ate

(PMMA) and polydimethylsiloxane rubber (PSi), (see Figure 4). Some of

the criteria for selecting these polymers were



1. Commercially available and soluble thermoplastic polymers to

facilitate sample preparation and analysis.

2. Mutual solubility of polymer and organometallic complex in

common organic solvents.

3. Polymers exhibiting high melting and glass transition

temperatures, preferably above the decomposition temperature of

the metal carbonyl to be used.

4. Polymers with varied molecular structures, morphologies, and

mechanical properties to assess interactions with active metal

species and evaluate effects upon physical properties.



Polycarbonate (LEXAN*, 131-111, AV= 323,000) Is soluble in benzene

and polyvinylidene fluoride (Polyscience, Inc., Ry= 119,000) is soluble
in dimethylformamide. We knew from preliminary studies (95,96) that in

the PC system both thermal and photolytic decompositions Fe(CO)5 led to

significant polymer-Fe interactions which were accompanied by severe


















POLYCARBONATE (PC)


CH 0


POLYSULFON~E (PSF)


O( 0~O



POLYSTYRENE (PS)







POLYMETHYLMETHACRYLATE(PMMA)





OCH3
POLYVINYLIDENE FLUORIDE (PVF )







POLYDIMETHYLSILOXANE (PSI)



1 H3S O





Figure 4. Polymers Selected For Metal-Polymer
Composite Studies.








degradation of the polymer. The PC system may in fact be of interest

for photoresist applications.

PVF2-Fe(CO)5 composites were initially studied in some detail in
this laboratory by S. Reich and found to produce ferrimagnetic materials

upon thermolysis due to formation of submicroscopic a-Fe and u-Fe203

particles .(97). The PVF2-iron composites may be of interest for

magnetic recording applications. The decomposition of Fe(CO)5 in PVF2
leads to some fluorine abstraction by Fe and slight crosslinking of the

polymer yielding insoluble but tough, strong, magnetic polymer composite

films. The PC and PVF2 systems represent extremes of Fe(CO)5-polymer
chemistry. Other polymer media produce various degrees of Fe-Fe versus

Fe-polymer bond formation. Investigation of polymers of differing
molecular structures and the analysis of the products obtained with each

polymer were undertaken to help establish mechanistic pathways for

decompositions of Fe(CO)5 and other organometallic complexes in

polymeric matrices.

2.5.2 Organometallic Complexes Selected for Preparation of
Metal- olymer Composites

Sufficient literature is available concerning the incorporation of

both Group VII and Group VIII metal carbonyls into various polymers to

make the prospects of further studies interesting. However, for this

research project we have limited our work to include Fe(CO)5, Fe3(CO)12
and CO2(CO)8 (see Figure 5). These carbonyls were selected because of

their availabili-ty, lower toxicity relative to other carbonyls (i.e.,

nickel) and because there exists a large body of knowledge concerning

their thermal and photolytic decomposition reactions, though less
information is available on the kinetics of these reactions. It was of

interest to study the solid-state decomposition of mono- and trinuclear





O
C


OC


OC


Figure 5. Proposed Chemical Configuration of (a) Fe(CO)5'
(b) Fe3(CO)12 and (c) Co2(CO)S.





fron carbonyls in order to compare differences in decomposition products

and consequent interaction with the matrix polymer. Both Fe(CO)5 and

Co2(CO)8 exhibit good solubility in organic solvents, a necessity in
these polymer systems. The study and comparison of the metal carbonyl

decompositions in solid polymers were pursued to allucidate the

mechanistic pathways and kinetics of these reactions.





3. EXPERIMENTAL


3.1 Preparation of Iron Pentacarbonyl-Polymer Films

3.1.1 Polycarbonate (PC)/Fe(CO) E/Methylene Chloride

Polycarbonate pellets (10.0 g) and 43.9 ml methylene chloride

(Fisher Reagent) were mixed 24 hours at room temperature and ambient

atmosphere with a Teflon stir bar on a magnetic stir plate in a 50 ml

glass-stoppered erlenmeyer flask. Methylene chloride used for solvation

of the polymer was dried using molecular sieve pellets (Matheson,

Coleman & Bell). Iron pentacarbonyl (Alfa Products, Thiokol/Ventron

Division) was filtered through filter paper circles (Whatman 7.0 cm,

qualitative 4) into a foil covered test tube. Then 1.0 ml was added

dropwise to the above erlenmeyer, now completely covered with aluminum

foil. This solution of iron pentacarbonyl and polymer in methylene

chloride was mixed for approximately five minutes and then film cast

onto sheets of glass using a 0.005" or 0.020" steel doctor blade.

Resultant films contained 13.0 wt. % added Fe(CO)5. Films containing

9.1, 4.6, 2.3,and 1.1 wt. % added Fe(CO)5 in polycarbonate were prepared

to determine the extinction coefficient of Fe(CO)5 in a polycarbonate

matrix. All films were analyzed to determine the actual Fe(CO)5

retained and found to contain 9.3, 6.8, 2.6, 1.3,and 0.44 wt. %, Fe(CO)5,

respectively.





3.1.2 Polystyrene (PS)/Fe(CO)r;/Methy lene Chloride

3.1.2.1 Polystyrene/Fe(CO)cg/methylene chloride preparation in air

Polystyrene beads (5.0 g) and 21.6 ml methylene chloride were mixed

24 hours in a 50 ml glass-stoppered erlenmeyer flask. Methylene

chloride was dried as for the polycarbonate films. Iron pentacarbonyl

was filtered through Whatman filter paper circles into a fail covered

test tube. Then 0.6 ml was added dropwise to the above (now foil

covered) erlenmeyer. This solution of iron pentacarbonyl and polymer in

methylene chloride was mixed approximately five minutes and then cast

twice onto glass using a 0.005" steel doctor blade (with a brief drying

between layers). Resultant films contained 10.7 wt. % added Fe(CO)5 in

PS. Films containing graduated levels of Fe(CO)5 were also prepared to

determine an extinction coefficient for Fe(CO)5 in a PS matrix. Added

versus actual wt. % (respectively) iron pentacarbonyl in PS were shown

to be 10.7 vs. 1.6, 9.2 vs 1.41, 7.3 vs. 1.31, 5.9 vs. 1.03, 3.8 vs.

0.99, and 2.1 vs. 0.71 wt. % Fe(CO)5'

3.1.2.2 Polystyrene/Fe(CO ),/methylene chloride preparation in H,

Polystyrene beads (8.03 g) and 34.7 ml methylene chloride were

mixed for 24 hours in a 50 ml glass-stoppered erlenmeyer flask. This

produced a 15 wt. % solution weighing 53.7 g. This solution was divided

into two portions of 24.3 g each for use as the composite polymer

solution (A) and as the control polystyrene solution (B). The remaining

5.1 g was used for preparation of TEM grids by the dilution technique.

Solution A plus 10 ml methylene chloride was placed in a 1000 ml round-

bottom three-necked flask fitted with a 2-way gas adapter and equipped

with a magnetic stir bar and placed over a magnetic stir plate (see

Figure 6). This closed gas-vacuum system was set-up within a hood.





Hood Vent


Magnetic Stir Plate


Figure 6. Schematic for Preparation of Composites in a
Closed Environment.









Solution A was then mixed under a 10% H2-argon environment for

approximately 3 hours. Iron pentacarbonyl was filtered through a

Whatman filter paper circle into a foil covered test tube and 0.22 ml

added to solution A while positive gas pressure was maintained. This

Fe(CO)5-PS-methylene chloride solution was then mixed under the H2

atmosphere foil covered for approximately 15 minutes. The magnetic stir

bar was then removed. The flask was next turned at an angle and the

solution rotated to coat the interior of the flask walls. This was

continued with Hp-argon gas flow (~2 psi) for 20 to 30 minutes until a

sufficiently viscous film was cast upon the walls. Gas flow was

continued for 2 to 3 hours and then this reaction vessel placed in a

glove box purged with the same H2-argon environment. The film was then
removed from the 1000 ml flask and stored within the glove box in

darkness until needed. Control films were prepared from solution B in

the same manner but without the addition of Fe(CO)5*

3.1.3 Polysulfone (PSF )/Fe(CO~r;/Methyl ene Chloride

3.1.3.1 Polysul fone/Fe(CO)r;/methyl ene chloride preparation in air

Polysulfone pellets (7.5 g) were added to methylene chloride

(32.6 ml) in a 50 ml glass-stoppered erlenmeyer flask. This solution

was mixed 24 hours with a Teflon stir bar on a magnetic stir plate. The

methylene chloride was dried with molecular sieve (as for the poly-

carbonate system). Iron pentacarbonyl was filtered as before and 0.50

ml added dropwise to the foil-covered polysulfone solution and stirring

continued for approximately 5 minutes. Thin films were cast from this

solution onto sheet glass using an 0.005" doctor blade. Films

containing 16.7 added wt. "n Fe(CO)5 were produced. Films





containing 9.1, 7.7, 5.9, 4.0 and,2.0 added wt. % Fe(CO)5 in PSF were

prepared to determine the extinction coefficient of Fe(CO)5 in a PSF

matrix. All films were analyzed to determined the actual Fe(CO)S
retained and found to contain 2.1, 1.8, 0.4, 0.3, 0.26, and 0.11 wt %

Fe(CO)5, respectively.

3.1.3.2 Polysulfone/Fe(CO),/methylene chloride preparation in H7

Polysulfone beads (3.77 g) and 16.2 ml methylene chloride were

mixed for 24 hours in a 50 ml glass-stoppered erlenmeyer flask. This

produced a 15 wt. % solution weighing 25.1 g. This solution was divided

into two portions of 13.7 (A) and 9.0 g (B) for use as the composite

polymer solution and as the control polysulfone solution,

respectively. The remaining 2.4 g was retained for preparation of TEM

grids. Solution A and 10 ml methylene chloride was placed in a 100 ml

round-bottom three-necked flask equipped with a magnetic stir bar and

placed over a magnetic stir plate (refer to Figure 6). Solution A was

then mixed under 10% H2-argon gas for approximately 3 hours. Iron

pentacarbonyl was filtered through a Whatman filter paper circle into a

foil covered test tube and 0.15 ml added to solution A while positive

gas pressure was maintained within the flask. This Fe(CO)5-PSF-

methylene chloride solution was mixed under the H2 environment for

approximately 15 minutes. At this point, it was observed that very

thick films would be produced by casting all this composite solution in

one 100 ml round-bottom flask so 6.7 g was transferred to another 100 ml

round-bottom flask (equipped as before) via a 90" fritted glass elbow

adapter. Some sample solution (1.4 g) was lost in the tube during

transfer. The magnetic stir bar was then removed and the flasks rotated

while at an angle to coat the interior wall. This was continued until





films were cast and then allowed to dry under a slow flow of H2 for two
to three hours. Gas inlet and outlet valves were then closed and the

reaction vessel placed in a glove box purged with the same gas

atmosphere and films removed and stored under H2. Control films were

prepared from solution B in the same manner but without the addition of

Fe(CO)5. All films were dried and stored in the dark within the glove
box.

3.1.4 Polydimethylsiloxane (PSi)/Fe(CO),/Methylene Chloride

Uncrosslinked polydimethylsloxane elastomer (3.02 g) and 10 ml

methylene chloride were vortexed (speed control No. 6) for 5 minutes in

a 50 ml glass-stoppered flask. Vigorous and short mixing times were

used because the polysiloxane is a one component room-temperature

vulcanizing elastomer. Filtered iron pentacarbonyl (0.5 ml) was added

dropwise to the above (now foil covered) erlenmeyer. This solution was

mixed 5 minutes and poured into glass petri dishes and allowed to dry

and vulcanize in the dark. All petri dish bottoms used for the

polysiloxane studies were coated with Sigma Cote@ (Sigma Chemical Co.)

to prevent adhesion of the elastomer to the glass. Resultant films

contained 20 wt. % added Fe(CO)5 in polysiloxane. Films were analyzed

for iron following pyrolysis (since crosslinking renders it Insoluble)

and found to contain < 0.1 t. %b Fe(CO)5 in PSi as evidenced by the low

absorption in the IR carbony1 stretching region and by its lack of the

characteristic brown-yellow carbonyl color.

3.1.5 PolymethylImethacry late (PMMA)/Fe(CO);/ Benz ene

PMMA pellets (5.0 g) and 32.2 ml benzene (Fisher Reagent) were

mixed for 48 hours at room temperature in an erlenmeyer flask. The









benzene was dried using molecular sieve pellets. Higher loadings of

iron pentacarbonyl were used for the PMMA studies, primarily because

iron analysis of composite PMMA films indicated that less iron

pentacarbonyl was retained in the PMMA film (in comparison to

polycarbonate composites). This may be due to loss of the carbonyl with

the benzene solvent during drying of the films. Filtered iron

pentacarbonyl was added dropwise to a foil covered flask of PMMA-benzene
solution and mixed for five minutes. This solution was used for film

casting. PMMA composite films with initial loadings of 43, 37, 31, 23

and 11 wt. % Fe(CO)5 were prepared by dropwise addition of 2.5, 2.0, 1.0

and 0.4 ml Fe(CO)5, respectively. As before, all films were analyzed

for iron content to determine the actual Fe(CO)5 concentration

retained. The final concentrations were found to be 0.8, 0.5, 0.4, and

0.2 wt. % Fe(CO)5 retained, respectively.

3.1.6 Polyvinylidene Fluoride (PVF?)/Fe(CO),/Dimethylformamide (DMF)

Two methods were used to produce PVF2 films for 1) TEM studies and

2) kinetic and SAXS studies. Films for TEM were prepared as follows:

PVF2 pellets (10.0 g) and 94.8 ml N,N-dimethylformanide (Fisher Reagent)
were mixed for six days in a flask. Filtered iron pentacarbonyl

(4.1 ml) was added dropwise to a covered flask of PVF2-DMF. This
solution was cast onto hot glass (110'C) and held at this temperature

for 1 hour. This was followed by heating for 24 hours at 140'C. This

procedure simultaneously volatilized the DMF and decomposed the

carbonyl. For kinetic studies, solutions of Fe(CO)5-PVF2-DMF were

prepared and film cast at room temperature and dried for 48 hours.

Though poor quality films resulted, they were adequate for kinetic





studies. Actual Fe(CO)5 content was determined to be approximately

5 wt. % by pyrolysis of the composite films followed by iron analysis.


3.2 Preparation of Triiron Dodecacarbonyl-Polymer Films

3.2.1 Polystyrene Feg(CO)19/Methylene Chloride
Polystyrene beads (1.0 g) and 10.0 ml methylene chloride were

placed inside a CO purged glove box (LABCONCO) and mixed 12 hours at

room temperature in Teflon stoppered 16 X 125 mm test tubes. Molecular

sieve pellets were used to dry the organic solvent. Triiron

dodecacarbonyl (0.0796 g) was solvated with methylene chloride in a 25

ml glass-stoppered volumetric flask (carbon monoxide gas was bubbled

through solvent prior to carbonyl addition). This Fe3(CO)12 solution

was mixed in the CO atmosphere of the glove box for approximately one-

half hour. Then 2.5 ml of the Fe3(CO)12 solution was added to the PS

solution and mixed on a Vortex Genie under CO. Good quality films were

cast by pouring this entire solution into a petri and drying under CO.

Resultant films contained 0.79 added wt. % Fe3(CO)12 but iron analysis
indicates a 0.72 actual wt. %. In a similar manner, four more

polystyrene-methyl ene chloride solutions were prepared and 2.0, 1.5, 1.0

and 0 ml of the same Fe3(CO)12-methylene chloride solution added to

obtain 0.63, 0.48, 0.32 and 0 wt. % added Fe3(CO)12 in films. Analyses

indicates 0.57, 0.45, 0.28,and 0 wt. % actual Fe3(CO)12 in resulting
films, respectively.

3.2.2 Polycarbonate- Fel(CO)1?/Methylene Chloride
Polycarbonate pellet (1.25 g) and 15.0 ml methylene chloride were

mixed 24 hours at room temperature with a Teflon stir bar on a magnetic

stir plate in a 50 ml glass-stoppered erlemneyer flask. Molecular sieve





pellets were used to dry the solvent prior to mixing with PC pellets.

Triiron dodecacarbonyl (0.06 g) was solvated in 7.5 ml methylene

chloride in a fail covered erlenmeyer prior to addition to the

PC-methylene chloride solution. Following a five minute mixing period,

this composite solution was film cast twice onto glass with an 0.005"

steel doctor blade. Films were dried and stored in the dark. Resultant

films contained 4.6 added wt. % Fe3(CO)12 but crystals were observed

within these films even prior to decomposition. To prevent the

formation of the triiron dodecacarbonyl crystals within films,

concentrations were lowered and preparation done under CO.

Polycarbonate beads (1.0 g) and 10.0 ml ethylene chloride were

placed inside a CO purged glove box (LABCONCO) and mixed 12 hours at

room temperature in Teflon stoppered 16 x 125 mm test tubes. Molecular

sieve pellets were used to dry the organic solvent. Trifron

dodecacarbonyl (0.0796 g) was solvated with methylene chloride in a

25 ml glass-stoppered volumetric flask (carbon monoxide gas was bubbled

through solvent prior to carbonyl addition). This Fe3(CO)12 solution

was mixed in the CO atmosphere of the glove box for approximately one-

half hour. Then 1.5 ml of the Fe3(CO)12 solution was added to the PS

solution and mixed on a Vortex Genie under CO. Good quality films were

cast by pouring this entire solution into a petri and drying under CO.

Resultant films contained 0.48 added wt. % Fe3(CO)12 (iron analysis
indicates a 0.52 actual wt. %). In a similar manner, three more

polycarbonate-methylene chloride solutions were prepared and 1.0, 0.5

and 0 ml of the same Fe31CO)12-methylene chloride solution added to

obtain 0.32, 0.16,and 0 wt. % added Fe3(CO)12 in films. Analyses

indicate- 0.33, 0.24,and 0 wt. 5 actual Fe3(CO)12 in resulting films,









respectively. Evidently, the solution was more concentrated than

assumed possibly due to volatllization of the methylene chloride.

3.2.3 Polysulfone/FeCQ GG id~ethytonClrer- Chierl~t- dedWt-de
Surfactant

The solubility of Fe3(CO)12 (Aldrich Reagent) in methylene chloride

was determined to be about 3.6 g/liter at room temperature. If this

concentration was used in the initial methylene chloride solution added

to the polymer, crystals were observed to form during the evaporation of

solvent and resulting increase In concentration of the Fe3(CO)12*
Surfactants were used in an attempt to remedy this problem and thus lead

to the formation of homogeneous Fe3(CO)12-polymer films. Fe3(CO)12

(0.18759 g) and 0.01 q Pluranic F68 (a nonionic polyethylene oxide-

polypropylene surfactant) were brought up to a 50 ml volume with

methylene chloride and mixed in a foil covered volumetric. A 25 ml

portion of this solution was added to an erlenmeyer flask and 1.0 g

polysulfone added. This composite solution was foil covered and mixed

approximately 1 hour. Following this, the dark green mixture was film

cast with a 0.005" doctor blade four times onto glass (i.e. 0.020"

thickness prior to drying). These films were dried and stored in the

dark. Apparently homogeneous green films resulted. Control polysulfone

films were the same as those prepared for polysulfone-fe(CO)5 studies as

Pluranic F68 surfactant does not show an absorbance in the IR carbonyl

absorption region.

An identical composite solution was prepared as discussed above but

with the addition of 2.0 ml DMF in place of Pluronic. This was done

also In an attempt to affect the solubility of the Fe3(CO)12. This

technique enabled higher loadings of the Fe3(CO)12 to be achieved

without precipitation of fron carbonyT crystals, though heating









necessary to volatilize the DMF also volatilized some of the Fe3(CO)12

added. Of 51 wt. % added Fe3(CO)12, only 6 wt. % was retained.

3.2.4 Polystyrene Fegg0D)17/Methylene Chloride With Added
Surfactant

A 25 ml solution of Fe3(CO)12 in methylene chloride was added to a

50 ml erlenmeyer flask and 1.0 g of polystyrene and 2 ml DMF added.

Again the solution was mixed and then cast four times onto glass with
0.005" doctor blade. It was then dried and stored in the dark. Control

films were prepared in the same'manner as those prepared for the

polystyrene-Fe(CO)5 studies.

3.3 Preparation of Dicobalt Octacarbonyl-Polymer Films

3.31 Vrie Lodins o Co(CO)R in Polystyrene/M~ethylene Chloride
Prepared Under CO
Polystyrene beads (1.2 g) and 5.4 ml methylene chloride were mixed

24 hours at room temperature with a Teflon stir bar on a magnetic stir

plate in a 50 ml glass-stoppered erlenmeyer flask. Molecular sieve

pellets were used to dry the organic solvent. Dicobalt octacarbonyl

(0.38 g) was solvated In 10 ml methylene chloride under a CO atmosphere

in a three necked, round bottomed flask (as shown in schematic 3.1.2.2

for H2 atmosphere). The PS-methylene chloride solution was then added

to this reaction flask while positive CO pressure was maintained. This

composite solution was mixed under the CO atmosphere for approximately

15 minutes and then transferred to a glove box purged with CO. Films

were cast by pouring approximately 6 ml of the composite solution into

petri dishes and drying under CO (exchanged 2 time daily). However,
better quality films were obtained by allowing a constant CO flow in the

reaction flask while rotating it. The composite solution became









increasingly more viscous as solvent volatized and was deposited upon

the walls of the flask to produce good quality films of 23.1 added wt. %

Co2(CO)g for our study. The CO gas flow was continued until films were
dry. They were then sealed under this CO environment until analyzed.

Weight per cent loadings of 16.7, 13.0,and 9.1 were also prepared in

this manner with the addition of 1.0, 0.75,and 0.5 g Co2(CO)8,

respectively.

Polystyrene beads (1.0 g) and 10.0 ml methylene chloride were

placed inside a CO purged glove box and mixed 12 hours at room

temperature in a Teflon stoppered 16 X 125 mm test tube. Molecular

steve pellets were used to dry the organic solvent. Dicobalt

octacarbonyl (0.18 g) was solvated in 15.0 ml methylene chloride under a

CO atmosphere in a 25 ml volumetric flask. Carbon monoxide was slowly

bubbled through the solvent for approximately one-half hour prior to

addition of Co2(CO)8. The solution of methylene chloride and Co2(CO)8

was mixed in the CO purged glove box for approximately one-half hour.

It was noted that not all the Co2(CO)8 went into solution, possibly due

to decomposed dicobalt octacarbony1. Then 4 ml of this Co2(CO)8

solution was added to the PS-methylene chloride solution and mixed under

CO. Films were cast by pouring this entire solution into a petri dish

and drying under CO. Resultant films contained 4.5 wt. % added Co2(CO)8

but cobalt analysis indicated a 1.1 actual wt. %. In a similar manner,

four more polystyrene-methylene chloride solutions were prepared and 3,

2, 1 and 0 ml of the same Co2(CO)8-methylene chloride solution added to

obtain 3.5, 2.3, 1.2, and 0 wt. X added Co2(CO)g. Analysis indicated

0.9, 0.7, 0.3, and 0 wt. % actual Co2(CO)8 in resulting films,

respectively.









3.3.2 Co? (CO ) i n Pol ysty rene/Methy lene Chl ori de Prepared i n H?

Sample preparation was done as for the PS-CO2(CO)a composites cast

In CO with the exception that 10% H2-argon was slowly passed through the

system. Films of theoretical wt. % Co2(CO)8 were cast by rotating the

flask while H2-argon gas was slowly passed through the closed system.

Films were dried and stored in this H2 atmosphere.
3.3.3 Poly sty rene/Co7 (CO )p/Methylene Chl ori de Sandwi ch Preparation (to
Evaluate Uxygen Diffusion kate)

Polystyrene beads (5.0 g) and 22.0 ml methylene chloride

(Mallinkrodt Reagent) were mixed 24 hours at room temperature with a

Teflon stir bar on a magnetic stir plate in a 50 ml glass-stoppered

erlenmeyer flask. As for the polycarbonate systems, molecular sieve

pellets (Matheson, Coleman & Bell) were used to dry the methylene

chloride for solvation of polymers. This solution was divided into two

equal portions, both in stoppered 50 ml erlanmeyers with Teflon stir

bars. Both polystyrene-methylene chloride solutions were placed in a

glove box (LABCONCO) purged with carbon monoxide following evacuation.

Co2(CO]8 (0.6361 g) was weighed into a test tube (16 x 125 mm) in air
and immediately purged with CO in the glove box. The Co2(CO)8 was added

to one of the PS-methylene chloride solutions to give 20.3 wt. %

Co2(CO)8 in polystyrene. This solution was mixed approximately 5
minutes.

Mi xi ng and casti ng of the poly styrene and Co2(CO)B-poly styrene

solutions were done in a carbon monoxide environment. The Co2(CO)8-PS

sandwich films were prepared by first casting a polystyrene film onto

glass, then a layer of Co2(CO]8-PS sandwich films were prepared by first
casting a polystyrene film onto glass, then a layer of Co2(CO)8-PS and
finally another polystyrene layer (see Figure 7). An 0.005" steel





1


=-/


varying thickness


I = Polystyrene protective layer of
II= Polystyrene containing 20% Co


Figure 7. "Sandwich" 02 Barrier Concept for Polystyrene-Co2(CO)3
Films.





doctor blade was used for each layer and a five minute drying period was

allowed between casting of layers. This sandwich was then removed from

the glass and stored in a 3-necked round bottom glass flask in a CO

environment. A control polystyrene film was prepared in the same manner

but with a center layer of polystyrene. A similar sandwich of thicker

polystyrene "bread" was prepared by using a 0.020" doctor blade for both

outer polystyrene layers. This Co2(CO)B-PS film, however, dried more

slowly so was not removed from the glass but left under CO in the glove

box overnite to dry. Good quality, yellow-colored films were obtained

in both cases.

3.3.4 Preparation of Co (CO)A in Polyvinylidene Fluoride (PVF3)/
ulmethyl Formamlde (bm,

Dicobalt octacarbonyl (~0.2 g) was solvated in 10 ml DMF. A pink

color developed with the evolution of gas (probably CO) on the glass

interface and formation of black precipitate. To this still bubbling

solution, approximately 1.0 g of PVF2 beads was added and mixed. This

solution was cast into petri dishes, covered and stored under CO for 5

days. Drying was not complete at this time so the sample was placed in

air to more rapidly volatilize the remaining DMF.


3.4 Preparation of Iron and Cobalt Carbonyl-Polymer Films

3.4.1 Polyvinylidene Fluoride Fe(CO)S- o2 CO)p/Dimethylformamide
prepared In CU
Polyvinylidene fluoride (4.00 g) and 38.3 ml DMF were mixed for six

days in a 50 ml erlenmeyer flask. Co2(CO]8 (0.5 g) was added to 15 ml
DMF In a 100 ml round bottom three necked flask with 2 psi CO gas flow

through (see Figure 6) and stirred. Filtered iron pentacarbonyl (0.34

ml) was added dropwise with mixing to this Co2(CO)8-DMF solution. Next

30 g of the PVF2-DMF solution (3.0 g PVF2) was added with mixing and gas









flow continued. The remaining PVF2-DMF solution was used to prepare

control samples. This solution of PYF2-cobalt/iron carbonyl in DMF was

then heated to 160"C to volatilize the solvent. By angling and rotating

the flask a film was produced on the interior wall. Films theoretically

contained 12.5 wt. % Fe(CO)5 and 12.5 wt. % Co2(CO)8 in PVF2. Films

were stored in CO and In the dark until analyzed.

3.4.2 Polysulfone Fe(CO)g Co?(CO)pMethylene Chloride Preparation
in CO

Polysulfone (8.03 g) and 34.7 ml methylene chloride were mixed for

24 hours In a 50 ml erlenmeyer flask. Co2(CO)8 (0.33 g) was added to

15 ml methylene chloride in a 100 ml round bottom flask with CO gas

flowing through and stirring. Filtered iron pentacarbonyl (0.22 ml) was

added dropwise to the cobalt solution. Next, 20.1 g of the

PSF-methylene chloride solution (3.0 g PSF) was added to this CO gas

purged system and mixed for approximately 10 minutes. The gas flow was

then slightly increased and the flask angled and rotated to cast a film

on the interior wall. Films of 10 added wt. % Fe(CO)5 and 10 added wt.

% Co2(CO)8 were formed. They were stored in the dark and in CO until
analyzed.


3.5 Preparation of Control Films

Appropriate polymer controls were prepared by eliminating the

addition of iron pentacarbonyl to identical polymer-solvent solutions.

All cast films were fail covered during drying and storage to prevent

photolytic decomposition of the carbonyl in composite films and for

experimental consistency in the control samples.





3.6 Quantitative Metal Analysis of Metal Carbonyl-Polymer Films

3.6.1 Standard Curve For Vogel's Iron Analysis

A variation of Yogells 1,10-phenanthroline iron analysis technique

(98) was used for the composite iron analysis, thus a comparable

standard curve was prepared. Dimethyl formamide (DMF) was selected as

the solvent for the composite and control polymers because it is not

only a good polymer solvent, but also water soluble as are the reagents

for Vogel's method. For the standard curve, two solutions were

prepared: an Fe solution (A) and a control (or reference) solution

(8). A 2.5 ml volume of the Fe standard (prepared according to Yogel)

was added to a 50 ml volumetric flask A. No Fe standard was added to

the control B. The standard iron solution was prepared (as directed

by Vogel) from 0.7030 g ammonium iron sulfate (Matheson, Coleman & Bell)

dissolved in 100 ml water. Five ml of 1:5 sulfuric acid was added and a

dilute solution of potassium permanganate (0.604 g/300 ml) was run in

cautiously until a slight pink color remained after stirring well. This

was diluted to 1 liter and mixed thoroughly (final solution: 1 ml =

0.1001 ag Fe).

Next, 5.0 ml of 10% hydroxylammonium chloride, (Fisher Reagent)

0.7 ml 0.2 M sodium acetate (Fisher Reagent) and 4.0 ml of 1,10-phenan-

throline (Fisher Reagent) were added to both flasks. Solutions were

vortexed after each addition. Finally, both flask A and B were brought

up to 50 ml volume with DMF and again mixed. After standing for one

half to one hour both solutions were filtered through Whatman glass

fiber filters. The Fe solution A contained 5.01 x 10-3 mg Fe/ml

solution.





Next, a dilution series of the Fe solution A in DMF was prepared as

shown in Table 1 and the UV absorbance at 396 nm measured in quartz

curvettes. Solution B was used as the reference. These data are

plotted in Figure 8. A linear curve fitting for the data shows 0.999

correlation.

3.6.2 Vogel's 1,10-Phenanthroline Method for Iron Analysis

The actual weight percent of iron in the untreated films was

determined by Vogel's 1,10-phenanthroline method. This method measures

the UV absorption at 396 nm of dissolved samples. Absorption was

compared to a standard curve (see section 3.6.1) to quantify total

iron. Sample for this analysis were prepared by dissolving 0.02-0.10 g

of the composite film in 15 ml dimethylformamide in a 25 ml volumetric

flask. After solvation of the polymer, the following were added: 2.5

ml of 10% hydroxylammonium chloride 2.5 ml 0.2M sodium acetate to adjust

pH to 3.51 1.0, and 2.0 m l1,10-phenanthroline solutions were vortexed

(Vortex-Genie, Scientific Industries) after each addition. This

solution was then brought up to 25 ml with dimethylformamide and again

mixed. After standing 1 hour (to develop full color), the solution was

filtered through a Whatman 934-AH glass fiber filter to remove the

precipitated polymer and the UV absorption at 396 nm of the filtrate was

measured in quartz cuvettes. Appropriate control polymer films prepared

in the same manner were used as a reference.

Selected iron polymer samples were also analyzed by atomic

absorption (Perkin-Elmer 460 AA Spectrophotometer with iron hollow

cathode lamp and air-C2H2 flame) to verify values obtained by Yogel's

method. The use of Soil Science's AA Spectrophotometer and Frank

Sodek's assistance is gratefully acknowledged. Samples were prepared












Fe Soln. Control Soln. Added*
a b DMF Corresponding
ml ml ml mg Fe/ml AUV(396 nm)


0 2 0 0 0
2 0 0 5.01 x 10-3 0.281
2 0 1 3.34 x 10-3 0.180
2 0 2 2.50 x 10- 0.132
2 0 0 1.57 x 10- 0.091
2 0 5 1.25 x 10- 0.064


*vortexing after additions






0.4



Euggs=53.7 L/g Feecm
ao 0.3-



O,


Table 1. Preparation of Dilution Series for
Standard Iron Concentration Curve.


1.sD 2.0 3.0 4.0 5.0
mg Fe/rn xi x03


Figure 8. Iron Standard Curve in Dimethyl Formamide.









both by 1) solvation in DMF and also 2) as for Vogel's iron analysis.

Iron standards of 0, 5 and 10 ppm Fe were also prepared in a similar

fashion; however, a commercial iron standard (Scientific Products--

1000 ppm) was used with the respective additions of 0, 0.125, 0.250 ml

of this standard as measured with a microburette. Comparison of these

values (from the DMF solvation and from Yogel's chemical medium)

indicate that only slight amounts of iron are retained in the

precipitated polymer thus similar values were obtained in both

instances.

3.6.3 Pyrolytic Technique for Atomic Absorption to Quantitate Iron and
Coat content

The PVF2 iron carbonyl composites were observed to be insoluble in

OMF and other solvents. Thus, an ashing technique suggested by Smith

and Wychick (71) was used to liberate the iron from the crosslinked

polymer matrix. Samples of composite and control films (0.03 g) were

weighed into platinum crucibles. This was followed by ashing at 600"C

for one hour In a muffle furnace (Sybron Thermolyne, Model 1-6028).

After cooling, samples for colorimetry were prepared by extracting the

iron from the hydrocarbon medium with 100 ul (Clay Adams Autopipettor)

of concentrated HCI (Fisher Reagent). This was then followed by Vogel's

1,10-phenanthroline iron analysis. To verify this value, a wet

digestion procedure (99) followed by atomic absorption analysis (100)

was done by Galbraith Laboratories. Similar values for iron in PVF2

films were obtained from both techniques. The polysiloxane composites

were also insoluble and were therefore analyzed by Galbraith


Laboratories in this same manner.









Cobalt-polystyrene films were also analyzed by Galbraith

Laboratories. Again samples were wet digested in a Carius furnace

followed by atomic absorption analysis for cobalt.


3.7 Determination of Extinction Coefficient of Metal
abols in a rolymer or In a solution

The IR carbonyl stretching absorption at approximately 2000 cm-1 of

composite films was measured prior to metal analysis. Iron or cobalt

concentrations in the metal carbonyl-polymer samples were then

determined either by Vogel's 1,10-phenanthroline method or by atomic

absorption as previously described. Using the IR absorption observed at

the 2000 cm-1 carbonyl stretching region, the iron concentration (as

determined either from atomic absorption or from Vogel's

1,10-phenanthroline method), and Beer's Law (Abs = Edc), the extinction

coefficient, E, f0P OSCh metal Carbonyl was calculated as a function of

solid polymer and carbonyl concentration.


3.8 Solid State Decomposition Methods for Metal
Carbonyl-Polymer Systems

3.8.1 Thermal Decomposition of Metal Carbonyl-Polymer Systems

Thermal decompositions were carried out in a temperature controlled

vacuum oven (National Appliance) with controlled atmosphere

capability. Decomposition times and temperatures were varied for

kinetic measurements. Temperatures were varied from 118"C to 142*C with

measurement times dependent upon rate of decomposition (a function of

temperature).

For selected systems, it was of interest to follow the thermal

decomposition kinetics under a 10% H2-argon atmosphere. For these





studies, small (100 ml) three-necked round-bottomed flasks were

equipment as shown in Figure 9. This flask was placed in an oil bath

within a hood. Films were placed in vials and these vials placed into

the round bottom flasks and decompositions carried out at various

temperatures. Decompositions of the Co2(CO)8 samples under an N2
atmosphere could be done simply by placing them within the closed

chamber (purged with N2) of the Nicolet MX-1, FTIR.

3.8.2 Photolytic Decomposition of Metal Carbonyl-Polymer Systems

Photolytic (UY) decompositions were performed using a 100 watt

mercury arc lamp (Ealing Optics Corp.) or a 75 watt xenon lamp (Ealing)

mounted on an optical bench with a quartz water filter between the

sample and source to avoid heating by infrared radiation (see Figure

10). UV irradiance and time were varied. A WG 305 filter (Schott

Optical) was used to limit the UV radiation to wavelengths greater than

3052 nm. Irradiance at the sample (for distances of 30 to 50 cm from

the source) were measured with a radi ometer/photometer system (EG & G,

Electro-Optics Model 550).

3.8.3 Gamma Radiation Decomposition of Metal Carbonyl-Polymer Systems
In this study, gamma radiation was explored as a means to decompose

metal carbonyl complexes within the solid polymer matrix. A cobalt

(measured to emit 600 curie in 1975) source was selected for this

purpose. Film samples (0.016 cm thick) were placed on microscope slides

within Coplin jars (see Figure 11). These containers were placed 2

inches from the gamma -radiation source and times of exposure varied from

0 to 72 hours. Samples treated in this manner were Fe(CO)5-PC,

Fe3(CO)12-PC and Fe3(CO)12 *S













Hood Vent


Magnetic Stir Plate


Figure 9. Apparatus for Thermal Decomposition of Metal
Carbonyl-Polymer Systems in Closed Environment.


Figure 10. UV Irradiation Assembly 1. Sample, 2. WG 305 Filter,
3. Quartz Water Filter, 4. Collimating/Focusing Lens,
5. UV Source and 6. Radiometer/Photometer.

















00
f/J


C1




r-








a




L.
oa



C ti


30




00

a C



Sr-l









3.8.4 Electron Beam Decomposition of Metal Carbonyl-Polymer Systems

Electron beam decompositions were done using both a JEOL JSM-35C

scanning electron microscope (SEM) and a VG-HB5 STEM both with a 100 KeV

tungsten field emission tip. Electron beam decomposition studies were

initially conducted on the JEOL JSM 35C SEM. Samples of polycarbonate

and 5 and 10 wt. % Fe(CO)5 polycarbonate films, typically 0.004 to 0.015

cm thick, were mounted on 1 inch aluminum SEM stumps with double sticky

tape and conductive carbon paint on the edges. No carbon or gold was

sputtered onto these samples to aid conduction of current off. A

variety of voltages ranging from 10 to 30 KeV, exposure times up to 10

minutes, magnifications, working distances, aperatures, and scanning

modes were tried as well as variations in iron carbonyl loadings.

Carbonyl decomposition was observed under the conditions discussed

below.

The filament was saturated and the 30 KeV electron beam focused on

a~ blank SEM stump with 0* tilt. The x and y coordinates were noted in

order to properly center sample later. Both 39 and 15 mm working

distances were used to vary magnification of the sample from 10 to 20

times, respectively. The voltage was then turned off, with the filament

condenser left in the saturation position, and the 5 wt. % added metal

carbonyl-polymer sample placed in the beam chamber. Aperatures were

pulled out, gun bias set to 3 and the beam blanking device then switched

on followed by the voltage. At onset of timed exposures, the beam

blanking device was turned off and sample scanned (irradiated) in Photo

mode 9/9. Specimen current and lens settings were recorded so that the

energy absorbed by the sample could be determined (see Appendix A for

calculation).









Electron beam decompositions were also conducted at the University

of Illinois Materials Research Laboratory. The willing and capable

assistance of Ms. Peg Mochel in these electron beam decomposition

studies was greatly appreciated. The VG-HBS STEM has the capability of

electron beam writing via a very slow scanning rate and small (12 A)

beam size. The electron focal spot on the sample in a scanning electron

microscope is quite small to provide good spatial resolution. As a

result, the current in the electron beam hitting the sample is small, of

the order of 10'9 ampere or less (101).

Thin samples of polysulfone and polysulfone with 30 wt. % loadings

of Fe(CO)5 were prepared by the dilution technique as described in

Section 3.1.1 and stored in darkness. A magnification of 100,000X and

applied voltages of 60 and 100 KeV were used to treat these thin

samples. With this STEM, a beam blanking device was not used. Initial

focusing was simply done at high magnification (to limit the area of

decomposition on the grid), then the grid region changed when

decomposition was to be followed. With these increased voltages,

shorter times could be used to achieve decomposition.


3.9 Kinetics for Decomposition of Metal Carbonyls in
Soi oymer natrices

Infrared absorption bands (Perkin-Elmer 283B Infrared

Spectrophotometer and Nicolet MX-1 FTIR) at 1996 and 645 cm-1 carbonyll

stretching and bending, respectively) were observed to change during

decomposition of the Fe(CO)5-polymer films. Th 96c-1badwsse

to follow the kinetics of the decompositions. Kinetics for thermal and

UV decompositions and activation energies for the thermal decomposition

reactions in composite films were determined. The willing cooperation









of Guy LaTorre was greatly appreciated in studies using the Nicolet MX-

1, FTIR.

Rate constants for each decomposition temperature (or specific

energy input from the UV source) were calculated based upon the

assumption that each absorbance measured could be correlated to a given

concentration of Fe(CO)5. A concentration could be evaluated using the

extinction coefficient as determined in section 3.8. Thus, the

remaining concentration of Fe(CO)5 in the polymer treated at a given

temperature T for time t was calculated. Assuming thermal decomposition

kinetics follows a first-order mechanism, the slope of a linear curve

fitting of x (time t) versus y (Ln concentration) will be the

decomposition rate constant k, at the temperature T. Furthermore, by

plotting the natural logarithm of each rate constant versus the

reciprocal of each respective temperature T, the slope is observed to be

the activation energy for the decomposition kinetics for Fe(CO)5 in a

polymeric matrix (102). For Co2(CO)g-polymer and solvent systems

decomposed under N2, second-order kinetics was observed. Thus, by

plotting:




10 K vs. time t, where A = concentration of
AoAo-A reactant A at time t
A, = initial concentration
of reactant A


One may determine the rate constant for this reaction from the slope of

this curve.

For photolytic decompositions, the kinetics of the decomposition

must be viewed differently. An activation energy was estimated by









plotting the rate versus energy of photolytic source at a given distance

or versus the inverse of this distance squared. The value for the

activation energy arrived at in this manner may still be compared with

that achieved for thermal decomposition data though the units differ.


3.10 Kinetics of Carbonyl Decomposition in Solution

3.10.1 Thermal Decomposition of Fe(CO)q in Ethyl Benzene Solution

Fifty ml of ethyl benzene (Matheson Coleman &~ Bell Reagent) were

placed in a foil-covered 100 ml three-neck flask. Filtered iron

pentacarbonyl (0.13 ml) was added to the ethyl benzene and stirred with

a magnetic stir bar. This reaction vessel was fitted with a condenser

(to limit volatilization) and stoppers (to allow for sampling) and

placed in a heated oil bath (see Figure 12). Decompositions were

measured at 70, 90, 110 and 130'C and followed by infrared spectroscopy

in sealed demountable NaC1 liquid cells. An ethyl benzene reference was

used. Noack's (103) extinction coefficient of 8000 1 mol-1cm-1 was used

for determination of Fe(CO)5 concentration in the ethyl benzene
solution.

3.10.2 Thermal Decomposition of Feg(CO)1, in Ethyl Benzene Solution

Fifty ml of ethyl benzene (Matheson Coleman & Bell Reagent) were

placed in a foil covered 100 ml three-necked flask. Triiron

dodecacarbonyl (0.12 g) was added to the ethyl benzene and stirred with

a magnetic stir bar. The reaction vessel was fitted with a condenser

(to limit volatilization) and stoppers (to allow for sampling) and

placed in a heated oil bath (see Figure 12). Decompositions were

followed at 82 and 110'C by infrared spectroscopy in sealed demountable

NaC1 liquid cells. An ethyl benzene reference was used. An extinction









coefficient of 12,650 1/mol cm was used for determination of Fe3(CO)12

concentration in the ethyl benzene solution.

3.10.3 Photolytic Decomposition of Fe(CO)F; in Ethyl Benzene Solution

Fifty ml of ethyl benzene were placed in a foil-covered 100 ml

beaker. Filtered iron pentacarbonyl (0.13 ml) was added to the ethyl

benzene and stirred with a magnetic stir bar. This beaker was then

covered with a quartz water filter to prevent evaporation of the solvent

yet allow the passage of UV light. It was then placed in line with the

UV source and the foil removed and shown below (Figure 13).

The UV source was placed above the solution to remedy interference of

carbonyl decomposition products deposited upon the container walls with

irradiation of the solution. The distance from source to sample was

varied and the respective photometric energies measured at each distance

to quantitate an activation energy.


3.11 Microscopic Characterization of Composite Film Morphology

Electron diffraction patterns from a field emission gun of a

transmission electron microscope (JEOL Model 200CX STEM and Philips

Model 301 TEM) and photomicrographs were used to study the compositions

and morphology of metallic particles. Samples were prepared by two

methods.

For those, TEM samples to be thermally decomposed in air the

dilution technique provided sample grids of satisfactory quality. A

0.1% solution of the polymer-carbonyl mixture in appropriate solvent was

applied dropwise to carbon coated (100 A thick) TEM grids. The thin

films produced on the grids were then treated to decompose them. Better

quality and more representative samples were found to be produced by

microtoming. For this procedure, composite films were embedded in epoxy

















UV)





C LL.





OOE

CIN
towr- C
3 th
cMO P








CL


cn










9-
O

E
O
*R



O*

U.C,



O
e--v1








o









~LO

r.O


r1 &

e--4


O
C
C V)









resin. The best embedding medium was found to be an equal ratio mixture

of solution A: solution 8 (Solution A and B prepared according to

instructions in Epon 812 Kit BU 013 055T-Baltzers). Embedded composites

were then ultramicrotomed (LKB Ultratome) at a speed of 5 mm/sec and

floated off on water onto formvar coated TEM grids. The assistance of

Drs. R. H. Berg III, and H. C. Aldrich was greatly appreciated in

developing a successful microtoming technique for these composites.

Samples were allowed to dry and then carbon coated (100 A). This

TEM sample preparation technique was found to be well suited to these

composite films decomposed in closed gas environments (CO or H2)*

Composite films were examined further by optical microscopy (Zeiss) and

scanning electron microscopy (JEOL Model 35C-SEM) with electron energy

dispersive spectra capabilities.


3.12 Small Angle X-ray Scattering Experiments to Measure
Particle bize Distribution

Quantitative small angle x-ray scattering (SAXS) experiments to

measure particle size and size distribution as well as changes in

polymer crystalline morphology were performed on the 10 meter SAXS

facility at the National Center for Small Angle Scattering Research

(NCSASR) at the Oak Ridge National Laboratory. This apparatus features

(Cu Ka radiation at 34.0 KY and 40 mA) pinhole collimation of the inci-

dent beam (1620 mm apart), a 2-dimensional position sensitive detector,

and a mini-computer based data acquisition and analysis system (see

Figures, 14 and 15). The distance between sample and the two-dimen-

sional position-sensitive proportional counter was 515 cm. All data

were corrected for background scattering, cosmic radiation, detector

sensitivity, variations in incident beam intensity, sample transmission












































Figure 14. Photograph of Small Angle X-Ray Scattering Facility at Oak
Ridge National Laboratory



















a

Eibl
e~I


aP


:g, a
:a
3I
;I" s
'- Z
% I:.
E:
: 21
m
ii ~
z


4 E~ ..rS IIIj a
i
;:i
:1 :




il'SI
c i
'II
-
if
--i1II
c=~
i I


3~4 ~1
~it 131
i, II
v
It ii
; i _I
-a tJl
i5 ~g
,,


3
r-









.i
=.
L(
Ci

fI


r. Z
'C s


I~
1!
Ci:
~
:Ir


B


7i


%f









coefficient and exposure time. Fully corrected data are presented as

two-dimensional contour plots of intensity versus scattering angle and

Guinier plots of log intensity versus scattering angle. Small angle

x-ray scattering studies required samples approximately 0.7 mm thick.

Metal carbonyl-polymer films cast to this thickness and then treated to

decompose the carbonyl were observed the bubble as CO was evolved.

Thus, thin films were prepared and treated to decompose the carbonyl.

This was followed by grinding of the samples (as done for magnetic

measurements--see Section 3.14) and then pressing at 10,000 pounds for

5-10 minutes in an infrared KBr pellet die (without KBr) at ambient

temperature to achieve the proper sample thickness necessary for SAXS

studies.

The assistance of Dr. J. S. Lin with SAXS studies and the use of

Oak Ridge National Laboratory facilities is gratefully acknowledged.

Studies were performed at the National Center for Small Angle Scattering

Research which is funded by National Science Foundation Grant No.

DMR-77-244-58 through Interagency Agreement No. 40-637-77 with the

Department of Energy (DOE) and was operated by the Union Carbide

Corporation under Contract No. W-7405-eng-26 with the DOE.


3.13 Determination of Average Magnetic Susceptibility of
Metal-Polymer Composites by the Guoy Method

A convenient way of making molar magnetic susceptibility

measurements is by means of a Guoy magnetic balance Illustrated in

Figure 16 (104). An electromagnet establishes a field H, and one

measures the resulting change in weight of a sample. The sample is

contained in a tube which is suspended between the pole pieces of a

magnet, with the bottom of the sample at centerline. This tube Is












































































~


N


So


To Balance


Figure 16. The Guoy Balance.





attached to a sensitive balance and one measures the change in pull that

occurs when the magnetic field is applied. If the sample is diamagnetic

(i.e. the induced field opposes the external magnetic field), the

magnetic field within it is less than Ho (M/Ho is negative and M =

magnetization). A paramagnetic sample shows the reverse effect (the

induced field aids the external one) so that M/Ho is positive.

Thus far, the possibility of an interaction between the mangetic

moments of adjacent atoms has been neglected. This is a reasonable

procedure to follow when the atoms have closed outer shells. However,

transition metal atoms (like Fe, Co,and NT) and ions have

characteristically incomplete d and f shells and these unpaired

electrons can interact with similar electrons in neighboring (adjacent)

atoms. This phenomenum is termed ferromaignetism and commonly results In

much larger intensities of magnetization. Ferromagnetism is rare In

nature though technologically important. Paramagnetism is common among

the transition group elements and diamagnetism is a universal property

of matter.

3.13.1 Sample Preparation for Magnetic Susceptibility Measurements

Thermally treated composite and control samples (approximately 5 g)

were first chopped in a Waring Commercial Blender (Model No. 31BL92)

followed by grinding (Arthur H. Thomas Gear-Cutter) to achieve small

particle sizes. This comminuated material was then sieved to select

particles of somewhat uniform size suitable for magnetic susceptibility

measurements. Particles passing through Tyler No. 28 but not through

Tyler No. 48 (595 mm and 300 mm, respectively) were found to be suitable

for this purpose. Selected polymer samples (i.e. PVF2) found to be too

tough for comminuation by this process were further ground in a mortor









and pestle with liquid nitrogen. To reduce sample loss due to static

charge, it was observed that ground samples could easily be collected if

wetted with ethanol. Samples were dried at 60"C after collection into

sample vials.

3,13.2 Guoy Balance for Measurements of Magnetic Properties

The magnetic properties of these composites were measured on a Guoy

Balance. Thanks are extended to Dr. Stoufer for his helpful discussions

and assistance as well as the use of his Guoy Balance in the Chemistry

Department here at the University of Florida.

A Mettler semi-microbalance (Model B6) was used to weigh these

sieved samples (0.5 g) in applied fields of 6830 gauss, 5740 gauss and

no field environments. This was achieved with a Varian Power Supply

(Model V-2300-A) and Current Regulator (Model V-2301-A). Triplicate

measurements were made at each field strength for all samples. Normally

samples of approximately 0.5 g were used and showed an average deviation

of 0.0004 g. Data were corrected for sample holder background in the

various fields as well as for matrix polymer effects. Metal content of

these samples was measured to directly correlate the magnetic

susceptibility of samples to metal content.


3.14 Effect of Fe(CO), Decomposition on Polymer Molecular Weight

Intrinsic viscosities ([n]) of polymers and composite compositions

were determined using a Ubbelohde OB viscometer. Solutions of the

composite and control polycarbonate, polysulfone and polydimethyl-

siloxane films were prepared and mixed in dioxane (Fisher Reagent)

overnite and then filtered through glass fiber filters just prior to

viscosity measurements. Solutions of polystyrene and

polymethylmethacrylate composite and control films were prepared in





benzene. Polyvinylidene fluoride composite and control viscosities were

measured prior to film casting as a function of time after addition of

Fe(CO)5 because the films became insoluble after film casting and

decomposition of the organometallic. Measurements were made at 30'C.

By plotting x (concentration C of solution in g/dl) versus y (reduced

viscosity = specific viscosi ty/concentration of solution) and

extrapolating to C = 0, one can determine the Intrinsic viscosity,

En]l. The [n]1 was used to calculate viscosity-average molecular weights,

R,, using the Mark-Houwink equation (105) and appropriate constants for
each polymeric system. This procedure is more clearly described in

Appendix 8,














4. RESULTS AND DISCUSSION


4.1 Preparation of Metal Carbonyl-Polymer Solid Solutions

The preparation, decomposition chemistry and properties for a

number of carbonyl-polymer compositions were studied (see Table 2). A

variety of metal carbonyls were selected to study any differences in

decomposition pathways and the products formed as well as their

interaction with the matrix polymer. By also selecting polymers with

various chemical structures, the chemistry of these solid state

decompositions may be better understood. The need for mutual solubility

of the carbonyl and polymer governs the organic solvent selected for the

preparation of films and the chemical interaction of the solvent with

the metal carbonyl and polymer was also found to affect the final

products. The following discussion summarizes the significant results

achieved in this investigation.

4.1.1 Carbonyl-Solvent Interactions Affecting- Polymer Films

To produce homogeneous solid solutions of a metal carbonyl in a

polymer, they must be comiscible in suitable film casting solvents. It

is important to understand any chemical interactions that may occur

between the metal carbonyl and solvent to produce active intermediates

which may in turn affect the final products. For example, in the system

PVF2-Fe(CO)5, PVF2 is soluble in DMF which also solvents Fe(CO)5*
However, the IR spectrum of Fe(CO)5 in DMF (Figure 17), shows the

appearance of a band at 1880 cm-1 corresponding to a carbonyl anion.




University of Florida Home Page
© 2004 - 2010 University of Florida George A. Smathers Libraries.
All rights reserved.

Acceptable Use, Copyright, and Disclaimer Statement
Last updated October 10, 2010 - - mvs