Preparation of sic-based fibers from organosilicon polymers

MISSING IMAGE

Material Information

Title:
Preparation of sic-based fibers from organosilicon polymers (I) Effects of polyvinylsilazane on the characteristics and processing behavior of polycarbosilane-based solutions and (II) synthesis, characterization, and processif of polymethylsilanes
Physical Description:
xxii, 435 leaves : ill. ; 29 cm.
Language:
English
Creator:
Saleem, Mohamed, 1965-
Publication Date:

Subjects

Subjects / Keywords:
Materials Science and Engineering thesis, Ph.D   ( lcsh )
Dissertations, Academic -- Materials Science and Engineering -- UF   ( lcsh )
Genre:
bibliography   ( marcgt )
non-fiction   ( marcgt )

Notes

Thesis:
Thesis (Ph.D.)--University of Florida, 1998.
Bibliography:
Includes bibliographical references (leaves 425-434).
Statement of Responsibility:
by Mohamed Saleem.
General Note:
Typescript.
General Note:
Vita.

Record Information

Source Institution:
University of Florida
Rights Management:
All applicable rights reserved by the source institution and holding location.
Resource Identifier:
aleph - 029546497
oclc - 40154079
System ID:
AA00017705:00001


This item is only available as the following downloads:


Full Text










PREPARATION OF SIC-BASED FIBERS FROM ORGANOSILICON POLYMERS:
(I) EFFECTS OF POLYVINYLSILAZANE ON THE CHARACTERISTICS AND
PROCESSING BEHAVIOR OF POLYCARBOSILANE-BASED SOLUTIONS AND
(II) SYNTHESIS, CHARACTERIZATION, AND PROCESSING OF
POLYMETHYLSILANES













By


MOHAMED SALEEM


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



UNIVERSITY OF FLORIDA


1998














ACKNOWLEDGMENTS


I am grateful to Dr. M.D. Sacks for his invaluable guidance and support. I am

thankful to Drs. C.D. Batich, J.H. Simmons, E.D. Whitney, R. Drago, and D. Talham for

serving on my committee. I would like to thank Dr. S. Bates, Dr. A. Morrone, R.

Crockett, W. Acree, and E. Lambers for their help in use of various analytical

instruments.

I would also like to thank G.W. Schieffele, J.H. Dow, Y.J. Lin, K. Wang, R.

Raghunathan, T.J. Williams, and G. Staab for their assistance in carrying out various

experiments in this study. I am grateful to Drs. Rajiv Bendale and Priya Bendale for

many useful technical discussions throughout the course of this study. Thanks are also

due to U. Mahajan, S. Lathi, M. Lakshmipathy, and J. DePuy for their help in compiling

this dissertation. I would aiso like to thank P. Raghunathan, R. Raghunathan, J.

Sethuraman, J.H. Dow, V. Shenoy, U. Shenoy, N. Srinivasa, A. Srinivasa, H. Kannan,

D. Kuruvilla, E. Naveen, R. Parikh, C. Parikh, and V. Srinivas for their friendship

throughout the course of my stay in Gainesville.

Finally, my special gratitude goes to Anuradha and Prakash Krishnans, for their

friendship, support, and encouragement, and for providing an atmosphere of home

away from home.















TABLE OF CONTENTS

AC KNO W LEDG M ENTS ................................................. ........ ........ ........ ii

LIS T O F TA B LE S ............................................. .......................... ................. vii

LIST OF FIGURES ....... .............................................. .............. ................. xi

A B S T R A C T ............................................. ......................................................... xxi

1. INTR O D U C TIO N ................................................. ....................................... 1

2. LITERATURE REVIEW .............................................................. ................ 5
2.1. Background ........... ................................................ ...... ......... ...... 5
2.2. Polysilane synthesis ............................ ................... 7
2.2.1. Wurtz-coupling of dichlorosilanes with alkali metal ......................... 7
2.2.1.1. M echanism ....................................... ... .............. 7
2.2.1.2. Mode of addition of reagents .................. ...................... 10
2.2.1.3. Effect of alkali m etal ......................................... .. ............. 12
2.2.1.4. Solvent effects ........................................ ..................... 14
2.2.1.5. Temperature effects ....................................... ................ 23
2.2.2. Ultrasonically activated Wurtz-coupling reactions ........................... 25
2.2.3. Polymerization of monoalkylchlorosilanes ...................................... 26
2.2.4. Polysilane copolym ers .................................... ......................... 28
2.2.5. Dehydrocoupling ............................... ................. 28
2.2.6. Redistribution/substitution reactions ...................................... 37
2.3. Pyrolysis behavior ............................................. ....................................... 42
2.4. Cross-linking of polysilane polymers ....................... ........................ 60
2.4.1. Oxidative cross-linking ........................................... ............. 60
2.4.2. Room temperature vulcanization ..................................... .... 60
2.4.3. Photo-cross linking ........................................ ......................... 62
2.5. Applications of Polysilane Polymers ..................................... ....... 63
2.5.1. Precursor for p-SiC ................... ................... ...................... 63
2.5.2. Photoinitiators for radical polymerization ..................................... 68
2.5.3. Photoresists in microelectronics .................................... ..... 69

3. EXPERIMENTAL PROCEDURE .................................................................... 72
3.1. Role of polyvinylsilazane as a spinning aid for polycarbosilane ..................72
3.1.1. Polymer synthesis ............................................... ................ 72
3.1.2. Spin dope preparation, fiber spinning, and fiber heat treatment ...... 74
3.1.3. Characterization of PCS polymer solutions .................................. 77
3.1.4. Characterization of fibers ............................................ ........... 82
3.2. Synthesis and characterization of polymethylsilane (PMS) polymers ....... 84








3 .2 .1. S tarting m materials .............................................. ....................... 84
3.2.2. Procedure for polymerization ..................................................... 84
3.2.3 Determination of polymer yield .................... .......... .......... 89
3.2.4. General procedure for heat treatment of PMS-based polymer
solutions .................................. ............. ................... 90
3.2.5. Procedure for fractional precipitation of PMS polymers .................. 91
3.2.6. Characterization of polymers and samples prepared by heat
treatment of the polymers ................................................... 92
3.3. Spinning and characterization of fibers prepared from PMS-based
po lym e rs ........................................ ........ .......... .................... ... ....... 9 5
3.3.1. Spin dope preparation, fiber spinning, and fiber heat treatment ..... 95
3.3.2. Fiber characterization .......................................... ............... 98

4. RESULTS AND DISCUSSIO N ....................................................................... 100
4.1. Effect of PSZ as a cross-linking/processing aid for spinning of fibers
from P C S solution ................................................................. ................. 100
4.1.1. Fiber spinning characteristics .................................................. 100
4.1.2. Polymer solution characteristics................................................... 107
4.1.2.1. Molecular weight and intrinsic viscosity measurements..... 107
4.1.2.2. Studies on rate of evaporation of solvents from PCS
and PCS+PSZ solutions ..................... ............. 111
4.1.2.3. Contact angle measurements ........................................ 114
4.1.2.4. Surface tension measurements ....................................... 126
4.1.3. Characterization of fibers ............................................... 137
4.1.3.1. Characterization of fibers by FTIR .................................. 137
4.1.3.1.1. FTIR spectra of PCS fibers during heat
treatment in nitrogen ........................................ 143
4.1 3.1.2. FTIR spectra of PCS fibers heat-treated in air..... 151
4.1.3.1.3. FTIR spectra of air-heat treated PCS fibers
during heat treatment in nitrogen ...................... 155
4.1.3.1.4. FTIR spectra of PSZ during heat treatment
in nitrogen ............................. ............................ 162
4.1.3.1.5. FTIR spectra of PCS+PSZ fibers during heat
treatment in nitrogen ......................................... 172
4.1.3.1.6. FTIR spectra of air-heat treated PCS+PSZ
fibers during heat treatment in nitrogen .............. 184
4.1.3.2. Mechanical properties of fibers ........................................ 195
4.2. Synthesis and characterization of polymethylsilane polymers ............... 216
4.2.1. Effect of synthesis conditions on molecular weight .................... 220
4.2.2. Effect of synthesis conditions on polymer yield ............................. 229
4.2.3. Characterization of PMS polymers and ceramic residues resulting
from pyrolysis .................. ..... ................... ................ 225
4.2.3.1. Weight loss behavior ............................................... 233
4.2.3.2. FTIR spectroscopy studies on PMS polymers .............. 236
4.2.3.3. XRD characteristics ................................................ 252
4.2.3.4. EMA analysis .................................................. 256
4.2.4. Sensitivity of PMS polymers to oxygen contamination ............... 260
4.3. Preparation of silicon carbide fibers from polymethylsilane polymers...... 265
4.3.1. Methods to increase molecular weight of PMS polymers .............. 268








4.3.1.1. Heat treatment of PMS polymers ................................... 268
4.3.1.2. Fractional precipitation ................... .................... 289
4.3.2. Spinning of fibers from PMS-based polymers ............................. 299
4.3.2.1. Spinning of fibers from as-prepared blends of PMS:PCS
polym ers ........ ..... ... .............................................. 299
4.3.2.2. Spinning of heat-treated PMS polymers and PMS:PCS
polymer blends .............................................................. 306
4.3.2.3. Spinning of fibers from fractionally-precipitated PMS
polym ers ............. ....... ..... ....... ............ ..... .... ............ 331
4.3.3. Fiber extension experiments on PMS polymer spin dopes
containing PSZ .................................................... 339

5. SUMMARY AND CONCLUSIONS .................................. 346


APPENDIX A


APPENDIX B


APPENDIX C


APPENDIX D


APPENDIX E

APPENDIX F


APPENDIX G



APPENDIX H


APPENDIX I



APPENDIX J



APPENDIX K


RHEOLOGICAL CHARACTERIZATION OF PCS AND
PCS+PSZ POLYMER SPIN DOPES................................... 351

FIBER SPINNING CHARACTERISTICS FOR PCS AND
PCS+PSZ SPIN BATCHES ................. ... ................... 357

FIBER EXTENSION DISTANCES FOR PCS AND PCS+PSZ
SPIN DOPES ...................................... ..................... 362

INTRINSIC VISCOSITY CALCULATIONS FOR PCS,
PCS+PSZ AND PMS POLYMERS ................ .............. 365

RESULTS OF SURFACE TENSION MEASUREMENTS ..... 371

TEMPERATURE AND WEIGHT GAINS FOR HEAT
TREATMENT OF PCS, PCS+PSZ FIBERS IN AIR ........... 376

WEIGHT LOSS DATA FOR PCS, PCS+PSZ FIBERS
(AIR-HEAT TREATED AND NON-AIR HEAT TREATED)
AFTER PYROLYSIS AT 1150C IN NITROGEN .............. 378

TENSILE STRENGTH DATA FOR PYROLYZED PCS AND
PCS+PSZ FIBERS ...............................................................380

GPC MOLECULAR WEIGHT DISTRIBUTIONS FOR PMS
POLYMERS FRACTIONALLY-PRECIPITATED BY ADDITION
OF ALCOHOLS ...................................... .................. 400

GPC MOLECULAR WEIGHT DISTRIBUTIONS FOR PMS
POLYMERS FRACTIONALLY-PRECIPITATED BY ADDITION
OF ACETONE ............... ................. .... ................... 403

CALCULATIONS FOR EXCESS SILICON AND CARBON
IN SIC FIBERS ........ ....... ................................... .. 413








APPENDIX L


APPENDIX M


APPENDIX N


CHARACTERISTICS OF PMS POLYMERS SYNTHESIZED
FROM MDCS MONOMER .................................................415

CHARACTERISTICS OF PMS POLYMERS SYNTHESIZED
FROM MDCS AND MTCS MONOMER MIXTURE ..............417

CHEMICAL FORMULAS FOR MONOMERS USED IN
WURTZ-COUPLING POLYMERIZATION ........................ 424


REFERENCES ............................................... ..................................................426














LIST OF TABLES


Table Page

2.1. Effect of diglyme and heptane additions on polymerization of some
dichlorosilane m onom ers .............................................. ............. 17

2.2. Effect of temperature on polymerization of methylphenyldichlorosilane ....... 24

2.3. Effect of sonication time on molecular weights and polydispersities
of polymethylphenylsilane ............. ............................. ................ 26

2.4. Summary of methylsilane polymerization by catalytic dehydrogenation
reactions ...... ........... ....... ..... .................... .................. 30

2.5. Effect of time and temperature on polymerization of phenylsilane in the
presence of a lanthanoid complex ................................ ...... ............. 35

2.6. Ceramic yield and chemical compositions upon pyrolysis of polysilane
homopolymers, copolymers, and terpolymers ........................................ 43

2.7. Synthesis conditions and characteristics of PMS polymers prepared by
W ood .................. ..................... ............................................ .... 48

2.8. Pyrolysis results for catalytically cross-linked polysilane polymers .............. 50

2.9. Peak assignments for IR absorption spectra of vinylic polysilanes .............. 56

2.10. Ceramic yield characteristics and decomposition temperatures for
polysilane polymers synthesized by Abu-Eid et al. ................................... 59

2.11. List of polysilanes that can be used in radical polymerization .................. 69

3.1. Properties of reagents used in synthesis of polymethylsilane polymers ...... 85

4.1. Fiber spinning characteristics for PCS and PCS+PSZ spin dopes spun
under identical conditions ............. ............................................... 104

4.2. Average fiber extension distances for PCS and PCS+PSZ spin dopes at
the same viscosities used in the fiber spinning experiments.................... 108

4.3. Surface tension values for PCS and PCS+PSZ solutions in toluene at
different solids-loadings ....... ..... .................................................. 128








4.4. FTIR peak assignments for polycarbosilane (PCS) fibers ............................ 139

4.5. FTIR peak assignments for polydimethylsilane (PDMS) polymer ............... 141

4.6. FTIR peak assignments for PCS fibers (batch 65s) heat-treated in air
at 1870C .................................... ...................... .... .. ....... 153

4.7. FTIR peak assignments for polyvinylsilazane (PSZ) polymer ....................... 166

4.8. FTIR peak assignments for PCS+PSZ fibers (batch 70s)......................... 174

4.9. FTIR peak assignments for PCS+PSZ fibers (batch 70s) heat-treated in air
a t 17 7 C ...................................................................................................... 18 6

4.10. Average tensile strengths and rupture strains for PCS and PCS+PSZ fibers
(green, heat treatment in air at 180 10C, heat treatment in nitrogen at
400C, and heat treatment in air at 180 10C followed by heat
treatment in nitrogen at 4000C) .............................. ................. 196

4.11. Tensile properties of as-spun and air-heat treated (187C) PCS fibers,
heat-treated to various temperatures between 200 and 1150C in nitrogen .203

4.12. Tensile properties of as-spun and air-heat treated (177C) PCS+PSZ
fibers, heat-treated to various temperatures between 200 and 1150C
in nitrogen ................. ............... .................... .................... 204

4.13. Properties of SiC fibers spun from PCS ................................ ................. 211

4.14. Properties of SiC fibers spun from PCS+PSZ ........................................ 211

4.15. Synthesis conditions and characteristics for PMS polymers ...................... 219

4.16. FTIR peak assignments for polymethylsilane polymer PMS-F (batch
PMS-256) ....... ...................... ............ ........ .............. .. 239

4.17. FTIR peak assignments for polymethylsilane polymer PMS-C (batch
PMS-263) ...... .................. ....... ..................... .. ................ 240

4.18. d-spacings and 20 Bragg angles for Si and 3-SiC .................................... 257

4.19. Crystallite sizes for Si and SiC calculated by Scherrer's formula for various
polymers pyrolyzed at 1350C in nitrogen at 10C/min with no hold ........... 258

4.20. Results of Electron Microprobe Analysis (EMA) on pyrolyzed ceramic fibers
from PM S polym ers ...................................................... ................... 259

4.21. Conditions for heat treatment for PMS polymers containing PSZ, DCP,
and D B ....... .... .................................................................... .................. 270








4.22. Nomenclature of PMS polymers used in the heat treatment experiments ... 271

4.23. Conditions and results of heat treatment for PMS/PCS polymer blends ...... 282

4.24. Molecular weight distributions for PCS polymers used in the heat treatment
of PM S/PC S blends ........................................................ ................ 283

4.25. Conditions for fractional precipitation of PMS polymers ............................. 290

4.26. Results of EMA analysis on fractionally precipitated 1150C-pyrolyzed
(nitrogen) polym ers .................................................................................... 295

4.27. Conditions for fiber spinning experiments from as-prepared PMS/PCS
polymer blends (non-heat treated) ....................................... ................ 300

4.28. Tensile strengths of SiC fibers spun from as-prepared PMS:PCS polymer
b le nd s ... .. ...................... ..................... ........... ....... ............................... 30 5

4.29. Conditions and qualitative results for fiber spinning experiments from
heat-treated PMS and PMS/PCS polymer blends .................................... 307

4.30. Tensile strengths of SiC fibers spun from heat-treated PMS polymers and
PMS/PCS polymer blends .............................. ........................ 323

4.31. Elemental analysis by EMA for SiC fibers prepared from heat-treated PMS/
PCS polymer blends .................... ............... ..................................... 332

4.32. Conditions and qualitative results of fiber spinning experiments for
fractionally-precipitated PMS polymers...................... ............... 333

4.33. Tensile strengths of SiC fibers spun from fractionally-precipitated PMS
polym ers ...... ................................................... 336

4.34. Elemental analysis by electron microprobe for SiC fibers prepared from
fractionally-precipitated PMS polymers .............................................. 336

4.35. Results of fiber extension experiments for PMS polymers containing PSZ. 341

D-1. Intrinsic viscosity calculations for PCS and PCS+PSZ solutions in toluene 365

D-2. Intrinsic viscosity calculations for PMS polymer in toluene ....................... 368

D-3. Intrinsic viscosity calculations for PMS polymer in toluene-1,4 dioxane
mixture (50:50 by volume) ....... .............. ..................... 369

E-1. Results of surface tension measurements ....................................... 371

F-1. Air-heat treatment temperatures and weight gains for PCS and PCS+PSZ
fi be rs .............................................................. .............................. ........ 3 7 6










G-1. Weight loss data for PCS, PCS+PSZ fibers (air-heat treated and non-air
heat treated) after pyrolysis at 1150C in nitrogen ............................ 378

H-1. Tensile strength data for pyrolyzed PCS fibers (batch 63s) .................... 380

H-2. Tensile strength data for pyrolyzed PCS fibers (batch 64s) ................... 381

H-3. Tensile strength data for pyrolyzed PCS fibers (batch 65s) .................. 384

H-4. Tensile strength data for pyrolyzed PCS fibers (batch 69s) .................. 387

H-5. Tensile strength data for pyrolyzed PCS+PSZ fibers (batch 67s)............. 390

H-6. Tensile strength data for pyrolyzed PCS+PSZ fibers (batch 68s) .......... 394

H-7. Tensile strength data for pyrolyzed PCS+PSZ fibers (batch 70s) .......... 397

L-1. Results of characterization of PMS polymers synthesized from MDCS ...... 415

M-1. Results of characterization of PMS polymers synthesized from MDCS and
M TC S (70:30 w t% ) .......................................................... ................. 417














LIST OF FIGURES


Figure Page

2.1. Schematic showing mechanism of Wurtz coupling polymerization............ 8

2.2. Effect of reactant addition rate on molecular weight distribution of phenyl
m ethylsilane ....................... .................................... .............. ........... 11

2.3. Effect of sodium surface area on the rate of consumption of hexylmethyl-
dichlorosilane ................................ .. ............................................ 13

2.4. UV-Vis Diffuse reflectance spectrum of purple solid isolated during Wurtz
polymerization ....................................... .......................................... 15

2.5. Chemical formulas of polar solvents used in Wurtz polymerization .......... 16

2.6. Schematic illustration of the influence of solvent on the polymer/sodium
particle interaction during Wurtz polymerization ............................... 20

2.7. Rate of disappearance of monomer n-hexylmethyldichlorosilane as a
function of time and weight percent of 15-Crown-5 ether ....................... 22

2.8. GPC of polymethylsilanes synthesized by Mu and Harrod ..................... 32

2.9. Scheme for redistribution/substitution reactions of chlorodisilanes ......... 39

2.10. Structure of methylchloropolysilane polymer .......................................... 40

2.11. IR spectral changes during pyrolysis of a polysilane polymer ................. 45

2.12. Changes in intensities of pendant groups based on IR spectra for polysilane
polymer ... ........................................ ......... .................................... 46

2.13. TGA plots for polymethylsilane polymers, prepared by Zhang et al. .......... 51

2.14. FTIR spectra of PMS polymer, prepared by Zhang et al. ..................... 53

2.15. TGA and DTA plots for a VPS polymer heated in nitrogen at 200C/min to
12000C ...... ............. .... .... .......... .......................... ...................... 55

2.16. IR spectra of VPS polymer:(a) 25C (b) 250C (c) 400C (d) 650C
(e) 1000C ................................................................. .... ...................... 57










2.17. FTIR spectrum of polymethylsilane polymer, prepared by Abu-Eid et al. ... 61

2.18. Scheme for photo cross-linking reactions of polysilane polymers ............ 64

2.19. Comparison of single layer photoresist process vs. multilayer photoresist
process ....... ............................................. 71

3.1. Structure of 1,3,5-trimethyl-1,3,5-trivinylcyclotrisilazane .......................... 72

3.2. Schematic of reaction assembly for PSZ synthesis .................................. 73

3.3. Definition of terms in Young's equation and schematic illustration of the
geometry for determination of the contact angle by the sessile drop
method.................... ............... .............. ... 79

3.4. Schematic of reaction assembly for synthesis of polymethylsilane ............ 87

4.1. Plots of (A) shear stress vs. shear rate (B) viscosity vs. shear rate for a
PCS spin dope (solids concentration ~68 wt%) ...................................... 101

4.2. Plots of (A) shear stress vs. shear rate (B) viscosity vs. shear rate for a
PCS+PSZ spin dope (solids concentration ~70 wt%) ............................ 102

4.3. Schematic illustration of globule formation during spinning of fibers from
PCS spin dope .............................................. ................. 106

4.4. GPC molecular weight distributions: (A) PCS (B) PSZ ............................. 109

4.5 Plots of Ti,/c vs. c for (A) PCS (B) PCS+PSZ (C) PSZ ............................ 110

4.6. (a) Percentage change in weight of PCS and PCS+PSZ spin dope as a
function of time due to evaporation of toluene, and (b) Absolute weight
change of PCS and PCS+PSZ spin dopes as a function of time due to
evaporation of toluene ......... .. ........ ........... ................. 112

4.7. Advancing and receding contact angles for water on: (A) teflon substrate
(B) stainless steel substrate as a function of cumulative drop volume ...... 115

4.8. Advancing and receding contact angles for toluene on teflon substrate
as a function of cumulative drop volume ................................. ............ 117

4.9. Advancing and receding contact angles for toluene on stainless steel
substrate as a function of cumulative drop volume............................... 118

4.10. Advancing and receding contact angles for PCS (33 wt%)/toluene solution
on stainless steel substrate as a function of cumulative drop volume ...... 119








4.11. Advancing and receding contact angles for PCS+PSZ (33 wt%)/toluene
solution on stainless steel substrate as a function of cumulative drop
volume ...... ............................ .. .................... 120

4.12. Advancing and receding contact angles of PCS (33 wt%)/toluene solution
on teflon substrate as a function of cumulative drop volume .................... 121

4.13. Advancing and receding contact angles of PCS+PSZ (33 wt%)/toluene
solution on teflon substrate as a function of cumulative drop volume ......... 122

4.14. Advancing and receding contact angles for PCS (33 wt%)/toluene solution
on a stainless steel substrate coated with PCS as a function of cumulative
drop volume .................................. ... ............ ................. 124

4.15. Advancing and receding contact angles for PCS+PSZ (33 wt%)/toluene
solution on a stainless steel substrate coated with PCS+PSZ as a function
of cumulative drop volume ... ............................. .............. 125

4.16. Plots of: (A) shear stress vs. shear rate (B) viscosity vs. shear rate for a
33 wt% PCS solution used in surface tension measurement ................... 129

4.17. Plots of: (A) shear stress vs. shear rate (B) viscosity vs. shear rate for a
50 wt% PCS solution used in surface tension measurement ................... 130

4.18 Plots of: (A) shear stress vs. shear rate (B) viscosity vs. shear rate for a
66 wt% PCS solution used in surface tension measurement ................... 131

4.19. Plots of: (A) shear stress vs. shear rate (B) viscosity vs. shear rate for a
33 wt% PCS+PSZ solution used in surface tension measurement ............ 132

4.20. Plots of: (A) shear stress vs. shear rate (B) viscosity vs. shear rate for a
50 wt% PCS+PSZ solution used in surface tension measurement ............ 133

4.21. Plots of: (A) shear stress vs. shear rate (B) viscosity vs. shear rate for a
66 wt% PCS+PSZ solution used in surface tension measurement ............ 134

4.22. Surface tension as a function of concentration for PCS and PCS+PSZ
so lutio n s ............................. ................................... .................. ..... 13 5

4.23. Surface tension as a function of viscosity for PCS and PCS+PSZ
solutions ...................................................................... ........ 136

4.24. Room temperature FTIR spectra of green PCS fibers ............................. 138

4.25. Room temperature FTIR spectra of polydimethylsilane (PDMS) .............. 140

4.26. FTIR spectra of PCS fibers during heat treatment to 6000C at 1*C/min in
nitrogen atm osphere ...................................................... ................ 144








4.27. Intensity vs. temperature from FTIR spectra for PCS (green) fibers ........ 145

4.28. Subtraction spectra for PCS fibers (69s) heat-treated in nitrogen,
600-40C ...... ............. ............. ............................................ 149

4.29. Comparison of FTIR spectra of PCS fibers before and after heat treatment
at 600C in nitrogen ........................... ....... ................................. 150

4.30. Room temperature FTIR spectra for PCS fibers heat-treated in air at
18 7 .C .................................................... ........................ 152

4.31. Subtraction spectra for PCS fibers heat-treated in air at 187C (65s)
and PCS green fibers (69s) ...... ..................... ..... ............... 154

4.32. Comparison of FTIR spectra of PCS fibers at 40C before and after heat
treatm ent in air ....................... ............................. ............................. 156

4.33. FTIR spectra for air-heat treated (187C) PCS fibers during heat
treatment to 600C at 1 C/min in nitrogen ............................................ 157

4.34. Intensity vs. temperature from FTIR spectra for PCS (air-heat treated
at 187C) fibers during heat treatment in nitrogen to 6000C ................... 158

4.35. FTIR spectra of air-heat treated Nippon PCS fibers during heat treatment
in nitrogen to 500C (from Ichikawa et al.) ........................................ 160

4.36. Comparison of FTIR spectra of air-heat treated (187C) PCS fibers
before and after heat treatment at 6000C in nitrogen ............................ 163

4.37. Subtraction spectra for air-heat treated (187C) PCS fibers heat treated
in nitrogen (600-40 C) .................. ... ........................................ 164

4.38. Room temperature FTIR spectra for PSZ polymer (batch 0831A)............. 165

4.39. FTIR spectra for PSZ polymer during heat treatment to 600C at 1C/min
in nitrogen ...................................... ............ ............................. 167

4.40. Intensity vs. temperature from FTIR spectra of PSZ ............................. 169

4.41. Room temperature FTIR spectra of PCS+PSZ green fibers (batch 70s).. 173

4.42. FTIR spectra of PCS+PSZ green fibers during heat treatment to 600C in
nitrogen ........................................... 175

4.43. Intensity vs. temperature from FTIR spectra of PCS+PSZ fibers ............ 176

4.44. Comparison of FTIR spectra of PCS+PSZ fibers (70s) before and after
heat treatment at 6000C in nitrogen ....................................... .......... 179








4.45. Subtraction spectra for PCS+PSZ fibers (70s), 600-40C ........................ 180

4.46. Subtraction spectra for PCS+PSZ and PCS fibers at 40C ...................... 181

4.47. Subtraction spectra for PCS+PSZ and PCS fibers at 6000C .................... 182

4.48. Room temperature FTIR spectra of PCS+PSZ fibers heat-treated in air at
177 C ............... .... .. ........... ................................. ....... ... ............. 185

4.49. FTIR spectra of air-heat treated PCS+PSZ fibers during heat treatment
to 600C in nitrogen ...................................... ................................... 188

4.50. Intensity vs. temperature from FTIR spectra of air-heat treated PCS+PSZ
fibers ....... ....................... .... ........... .... 189

4.51. Comparison of spectra for air-heat treated PCS+PSZ fibers before and
after heat treatment at 600C in nitrogen ................... ................. 191

4.52. Subtraction spectra for air-heat treated PCS+PSZ fibers, 600-40C ......... 192

4.53. Comparison of spectra for air-heat treated PCS fibers (batch 65s)
and PCS+PSZ fibers (batch 70s) at 400C ...................... ............... 193

4.54. Comparison of spectra for air-heat treated PCS fibers (batch 65s)
and PCS+PSZ fibers (batch 70s) after heat treatment in nitrogen ............ 194

4.55. Average tensile strength for PCS, PCS+PSZ fibers, as-spun and after
heat treatment in (i) nitrogen at 400C (ii) air at 180 10C (iii) air at
180 10C, followed by nitrogen at 400C .................................... 197

4.56. Average rupture strain for PCS, PCS+PSZ fibers, as-spun and after
heat treatment in (i) nitrogen at 400C (ii) air at 180 10C (iii) air at
180 10C, followed by nitrogen at 400C ............................. ......... 198

4.57. Plots of (a) tensile strength vs. heat treatment temperature (b) %
elongation vs. temperature for polyacrylonitrile (PAN) fibers .................. 200

4.58. Schematic of structural changes taking place in PCS during heat
treatment in air at 180 10 C ............................ ..... .............. 202

4.59. Average rupture strain vs. temperature for: (A) PCS (batch 69s) and
(B) PCS+PSZ fibers (batch 70s) ..................................................... 206

4.60. Average rupture strain vs. temperature for fibers heat-treated in air:
(A) PCS (batch 65s/ 187C air heat treatment), and (B) PCS+PSZ (batch
70s/ 177C air heat treatm ent) ................. ................ ................... 207

4.61. Average tensile strength vs. temperature for (A) PCS (batch 69s), and
(B) PCS+PSZ fibers (batch 70s) ...................................................... 208








4.62. Distribution of tensile strengths for fibers after pyrolysis at 1150C in
nitrogen (A) PCS (B) PCS+PSZ ................ .......................................... 212

4.63. Distribution of diameters for fibers after pyrolysis at 1150C in nitrogen
(A) PCS (B) PCS+PSZ ...................................... .............. 213

4.64. Average tensile strength vs. temperature for fibers heat-treated in air:
(A) PCS (batch 65s/ 187C air heat treatment), and (B) PCS+PSZ (batch
70s/ 177C air heat treatm ent) ................................ .... ............... 214

4.65. Schematic for Wurtz polymerization of (a) MDCS (b) MDCS and MTCS
(70:30 wt%) .. ................. ........................... 217

4.66. (a) Gel permeation chromatograms for polymers prepared from MDCS
(A) toluene (B) Toluene:THF (95:5 vol%) (C) Toluene, dioxane (50:50 vol%)
(b) Gel permeation chromatograms for polymers prepared from
MDCS:MTCS(70:30 wt%) (A) toluene (B) Toluene:THF (95:5 vol%)
(C) Toluene, dioxane (50:50 vol% ) .................................................. 221

4.67. Effect of cosolvents on molecular weight of PMS polymers A,B, and C
(prepared using 100% MDCS) ....................... ......... .............. 223

4.68. Effect of cosolvents on molecular weight of PMS polymers D,E, and F
(prepared using MDCS/MTCS (70/30 wt%)) ........................ .............. 224

4.69. Plot of ns,/c vs. C for polymer F (batch PMS-256) (MDCS/MTCS (70/30
wt%); toluene/1,4-dioxane (50/50)) in toluene .......... ....... ............... 226

4.70. Plot of ri,/c vs. C for polymer F (batch PMS-256) (MDCS/MTCS (70/30
wt%); toluene/1,4-dioxane (50/50)) in a mixture of toluene and 1,4-dioxane
(50:50 vol% ) ...................................... ........................... .. 227

4.71. Schematic illustration for solvent effects in polymerization of MDCS (a) in
toluene (b) in a 50:50 mixture of toluene:1,4-dioxane .............................. 228

4.72. Effect of cosolvents on yield for PMS polymers D,E, and F (prepared
from MDCS/MTCS (70/30 wt% )) .................................. ................ 230

4.73. Effect of cosolvents on yield for PMS polymers A,B, and C (prepared
from 100% MDCS) .......................... ... .......... ..... ............... .. 232

4.74. TGA plots (% weight vs. temperature) for polymers prepared with 100%
MDCS ............ .... ............... .... .. .............................. 234

4.75. TGA plots (% weight vs. temperature) for polymers prepared with 70:30
wt% MDCS:MTCS ................... .................... ..................... 235

4.76. Room temperature FTIR spectra of PMS polymer C (batch PMS-263)
(prepared from 100% MDCS in toluene/dioxane solvent) ..................... 237








4.77. Room temperature FTIR spectra of PMS polymer F (batch PMS-256)
(prepared from 70:30 wt% MDCS:MTCS in toluene/dioxane solvent)...... 238

4.78. FTIR spectra of PMS polymer C (100% MDCS; toluene, dioxane (50:50
vol%)), 40 to 600C at 5C/min in nitrogen ............... ....... ............... 242

4.79. Intensity vs. temperature from FTIR spectra for PMS polymer C ............. 244

4.80. FTIR spectra of PMS polymer C, 750 to 1150C at 50C/min in nitrogen .... 247

4.81. FTIR spectra of PMS polymer F (MDCS, MTCS (70:30 wt%)), toluene,
dioxane (50:50 vol%)), 40 to 600C at 50C/min in nitrogen .................. 248

4.82. Intensity vs. temperature from FTIR spectra for PMS polymer F ............. 250

4.83. FTIR spectra of PMS polymer F, 750 to 1150C at 50C/min in nitrogen ... 253

4.84. XRD patterns for PMS polymers prepared from monomer MDCS: (A) 100%
toluene, (B) toluene:THF (95:5 Vol%), and (C) toluene:dioxane (50:50
vol%) .......................... ...... ............... .................. 254

4.85. XRD patterns for PMS polymers prepared from monomers MDCS:MTCS
(70:30 wt%): (A) 100% toluene, (B) toluene:THF (95:5 Vol%), and (C)
toluene:dioxane (50:50 vol% ) ................................. .............. 255

4.86. FTIR spectra of PMS polymer C (100% MDCS; toluene/dioxane
(50:50 vol%)) exposed to air, shown as a function of time ................... 261

4.87. Intensity vs. time of exposure to air from FTIR spectra for PMS polymer C. 262

4.88. Gel permeation chromatograms for PMS-231 polymer: (A) after 3 days
of storage, (B) after 260 days of storage, and (C) after heat treatment
(PMS-231-H) ........... ............ .......... ........... 273

4.89. Polydispersity index vs. molecular weight for PMS polymers containing
5-14.5 wt% PSZ (and 0.5-1.5 wt% DCP) as additives ........................... 274

4.90. GPC molecular weight distributions for: (A) PMS-214-A (B) PMS-214-A-H 275

4.91. GPC molecular weight distributions for: (A) PMS-216-A (B) PMS-216-A-H...277

4.92. GPC molecular weight distribution for PMS-216-A2-H ............................ 278

4.93. GPC molecular weight distributions for: (A) PMS-219-A (B) PMS-219-A-H.. 280

4.94. GPC molecular weight distributions for: (A) PMS-223-AD2-A
(B) PM S-223-AD2-H ........................................... ........................... 281

4.95. GPC molecular weight distribution for PMS-216-AP-H ............................ 285








4.96. GPC molecular weight distribution for PMS-217-AP-H .............................. 285

4.97. GPC molecular weight distribution for PMS-217-AP2-H ............................ 286

4.98. GPC molecular weight distribution for PMS-218-AP-H ............................ 286

4.99. GPC molecular weight distribution for PMS-220-AP2-H .......................... 288

4.100. GPC molecular weight distribution for PMS-221-AP2-H ....................... 288

4.101. GPC molecular weight distribution for PMS-240 polymer (A) before and
(B) after fractional precipitation with alcohol mixture................................. 293

4.102. Plot of non-solvent to polymer ratio vs. final Mw for polymers precipitated
using acetone as non-solvent ................................... 296

4.103. GPC molecular weight distribution for PMS-250 polymer (A) before and
(B) after fractional precipitation with acetone ........................................ 298

4.104. SEM micrographs of as-spun fibers (batch 24s) prepared from PMS/PCS
blends (non-heat treated) showing necking....................... ........ ........ 303

4.105. SEM micrographs of pyrolyzed fibers (batch 26s) prepared from
PMS/PCS blends (non-heat treated) showing necking ............................ 304

4.106 Plots of (A) shear stress vs. shear rate and (B) viscosity vs. shear rate for
U F-33s spin dope ......................................................... ....................... 3 11

4.107. Plots of (A) shear stress vs. shear rate and (B) viscosity vs. shear rate for
UF-35s spin dope ............ ..................................................... 312

4.108. Plots of (A) shear stress vs. shear rate and (B) viscosity vs. shear rate for
UF-36s spin dope ......................................... ................. 313

4.109. Plots of (A) shear stress vs. shear rate and (B) viscosity vs. shear rate for
UF-45s spin dope ........................... .. ............................ 316

4.110. Plots of (A) shear stress vs. shear rate and (B) viscosity vs. shear rate for
UF-42s spin dope ....................... .............................. 318

4.111. Plots of (A) shear stress vs. shear rate and (B) viscosity vs. shear rate for
UF-56s spin dope ................................................................ ................. 319

4.112. Plots of (A) shear stress vs. shear rate and (B) viscosity vs. shear rate for
UF-54s spin dope ................................. ................. 321

4.113. Plots of (A) shear stress vs. shear rate and (B) viscosity vs. shear rate for
UF-52s spin dope .......................................................... 322


xviii








4.114. SEM micrographs of pyrolyzed fibers (batch 42s) prepared from
heat-treated PMS/PCS polymer blends showing necking between fibers ... 324

4.115. SEM micrographs of UF-35s fibers after heat treatment at 1700*C in argon:
(A) and (B) surfaces, (C) fiber cross section ....................... .................. 326

4.116. SEM micrographs of fracture surfaces of UF-35s fibers after heat treatment
at 1700 C in argon ........................ ........................... ............................ 328

4.117. GPC molecular weight distribution for PMS-242 polymer: (A) before and
(B) after fractional precipitation .......... ........................... 342

4.118. Average extension for fibers drawn from PMS-based polymers as a
function of amount of PSZ 0908A added .......................................... 344

A-1. Plots of (A) shear stress vs. shear rate and (B) viscosity vs. shear rate for
64s spin dope (PCS) (solids concentration -67 wt%) ............................... 351

A-2. Plots of (A) shear stress vs. shear rate and (B) viscosity vs. shear rate for
65s spin dope (PCS) (solids concentration ~66 wt%) ............................... 352

A-3. Plots of (A) shear stress vs. shear rate and (B) viscosity vs. shear rate for
69s spin dope (PCS) (solids concentration -66 wt%) ............................... 353

A-4. Plots of (A) shear stress vs. shear rate and (B) viscosity vs. shear rate for
67s spin dope (PCS+PSZ) (solids concentration -69 wt%) ...................... 354

A-5. Plots of (A) shear stress vs. shear rate and (B) viscosity vs. shear rate for
68s spin dope (PCS+PSZ) (solids concentration -70 wt%) ...................... 355

C-1. Fiber extension distances for PCS spin dope ....................................... 362

C-2. Fiber extension distances for PCS+PSZ spin dope ................................. 363

I-1. GPC molecular weight distributions for PMS-241 (A) before and (B) after
fractional precipitation with alcohols .................... ................. ... ............ 400

1-2. GPC molecular weight distributions for PMS-243 (A) before and (B) after
fractional precipitation with alcohols .................................... ................. 401

J-1. GPC molecular weight distributions for PMS-245 (A) before (B) and (C)
after two fractional precipitations with acetone ........................................ 403

J-2. GPC molecular weight distributions for PMS-246 (A) before and (B) after
fractional precipitation with acetone ........................................................ 404

J-3. GPC molecular weight distributions for PMS-247 (A) before and (B) after
fractional precipitation with acetone .................................... .................. 405








J-4. GPC molecular weight distributions for PMS-248 (A) before and (B) after
fractional precipitation with acetone ..................................................... 406

J-5. GPC molecular weight distributions for PMS-249 (A) before and (B) after
fractional precipitation w ith acetone .......................................................... 407

J-6. GPC molecular weight distributions for PMS-251 (A) before and (B) after
fractional precipitation with acetone ..................................................... 408

J-7. GPC molecular weight distributions for PMS-252 (A) before and (B) after
fractional precipitation with acetone ..................................................... 409

J-8. GPC molecular weight distributions for PMS-253 (A) before and (B) after
fractional precipitation with acetone .................................... 410

J-9. GPC molecular weight distributions for PMS-254 (A) before and (B) after
fractional precipitation w ith acetone ......................................................... 411














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


PREPARATION OF SIC-BASED FIBERS FROM ORGANOSILICON POLYMERS:
(I) EFFECTS OF POLYVINYLSILAZANE ON THE CHARACTERISTICS AND
PROCESSING BEHAVIOR OF POLYCARBOSILANE-BASED SOLUTIONS AND
(II) SYNTHESIS, CHARACTERIZATION, AND PROCESSING OF
POLYMETHYLSILANES

By

Mohamed Saleem

August 1998

Chairman: Dr. Michael D. Sacks
Major Department: Materials Science and Engineering

The effect of the addition of polyvinylsilazane (PSZ) on the characteristics (i.e.,

spinnability, theological behavior, wetting behavior, evaporation behavior, etc.) of

polycarbosilane (PCS) solutions was investigated. Spinnability of PCS solution was

characterized by number of breaks occurring during spinning and amount of fibers

formed after spinning. PCS and PCS+PSZ solutions were characterized by measuring

surface tension, contact angles and rate of solvent evaporation. Effect of PSZ on

mechanical properties of SiC fibers prepared from PCS and PCS+PSZ solutions was

also investigated. Chemical changes taking place in PCS and PCS+PSZ fibers during

heat treatment from 40-600C in nitrogen were studied by Fourier transform infrared

spectroscopy (FTIR).

Addition of PSZ to PCS greatly improved spinnability of PCS solutions.

Significant differences in wetting characteristics were observed for PCS solutions and








PCS+PSZ solutions deposited on stainless steel and teflon substrates, as well as on

PCS-coated and PCS+PSZ-coated stainless steel substrates. The rate of evaporation of

solvent was higher for PCS solution than for PCS+PSZ solution at identical polymer

concentrations. As-spun PCS and PCS+PSZ fibers developed similar tensile strengths

and rupture strains. After heat treatment at 400C in nitrogen, PCS+PSZ fibers showed

higher tensile strength and rupture strain compared to PCS fibers. Based on FTIR

spectra of PCS and PCS+PSZ fibers during heat treatment from 40-600C, it is

suggested that PSZ acts as a cross-linking aid for PCS. PCS+PSZ fibers developed

higher tensile strengths than PCS fibers at all heat treatment temperatures between

200-1150C.

SiC fibers were fabricated from polymethylsilane (PMS) and PMS/PCS polymer

blends. PMS polymers were synthesized by a Wurtz-coupling polymerization of

methyldichlorosilane (MDCS) and methyltrichlorosilane (MTCS) in 70:30 wt% proportion

with sodium in refluxing toluene. The addition of polar solvents (i.e., THF and 1,4-

dioxane) to toluene improved yields and increased the molecular weight of PMS

polymers. As-prepared PMS polymers with additives were heat-treated to increase

molecular weight to permit fiber spinning. As-prepared PMS polymers were also

fractionally-precipitated to isolate higher molecular weight fractions suitable for fiber

spinning. The heat treatment approach was ineffective in obtaining molecular weight

increases reproducibly. Fractionally- precipitated PMS polymers were useful in

preparation of high strength pyrolyzed SiC fibers. These fibers, however, exhibited poor

thermal stability at high temperatures. More investigations will be needed to address this

problem.














CHAPTER 1
INTRODUCTION



There has been much interest in the preparation of ceramic materials from

organosilicon-based preceramic polymers. Polycarbosilane (PCS) polymers, in

particular, have been shown to be useful for producing SiC-based ceramic materials,

particularly fibers. Continuous SiC fibers with fine diameter and high strength are of

considerable interest for the development of ceramic-matrix composites for high

temperature applications. Commercially available fibers (e.g., NicalonTM, Nippon Carbon

Company, Tokyo, Japan and TyrannoTM, Ube Industries, Tokyo, Japan) are not pure

stoichiometric SiC, i.e., the fibers contain relatively high concentrations of excess carbon

and oxygen. (In addition, TyrannoTM fibers are actually Si-Ti-C-O fibers.) As a

consequence, these fibers degrade extensively at high temperatures. This degradation

is associated with the presence of oxygen (-8-15 wt%) and excess carbon (-15 wt%) in

the fibers. At high temperatures, carbothermal reduction reactions occur between

carbon and siliceous material in the fibers, leading to large weight losses and

degradation in mechanical properties. The fibers contain excess carbon as a result of

the high C:Si ratios in the starting materials. (PCS has a high C:Si ratio, as it is formed

by pressure pyrolysis of polydimethylsilane (PDMS) which has a C:Si ratio of -2. (PDMS

is produced from dimethyldichlorosilane, (CH3)2SiCI2, which, in turn, also has a C:Si

ratio of 2.) The high C:Si ratio in the preceramic polymer, PCS, leads to excess C in the

SiC-based fibers after pyrolysis.) The oxygen in SiC-based fibers, such as NicalonTM and








Tyranno", is a result of an oxidative curing step that is used to render the fibers

infusible during pyrolytic conversion to SiC.

There has been considerable interest in recent years in producing SiC-based

fibers with improved thermomechanical properties. Much of the effort in this area has

been directed toward fabrication of fine-diameter, high-strength, polymer-derived fibers

which have low oxygen content [Lip91A; Lip91B; Lip94A; Lip94B; Lip95; Tak94; Tak95;

Tor92A; Tor92B; Tor94]. The development of such fibers with both carbon-rich and

near-stoichiometric compositions have been reported by several research groups,

including those at Nippon Carbon Co. (Hi-NicalonTM and Hi-NicalonTM Type S fibers

[Tak94; Tak95], Dow Corning Co. (SylramicTM fibers) [Lip95], and the University of

Florida (UF and UF-HM fibers) [Tor94; Sac95A; Sac95B]. All of these fibers show

improved thermochemical stability and thermomechanical properties compared to fibers

which contain large amounts of oxygen, such as NicalonTM and TyrannoTM fibers.

The approach developed at the University of Florida is based on using a high-

molecular-weight PCS polymer which is infusible. Hence, an oxidative curing step is

unnecessary during pyrolysis. In producing SiC-based fibers from the high-molecular-

weight PCS polymers, Toreki et al. [Tor90] reported that the addition of a

polyvinylsilazane (PSZ) polymer to a PCS-based spinning solution (in amounts up to

14.5 wt%) improved fiber spinnability and produced fibers with improved mechanical

properties. The reasons for these improvements were not investigated.

There are two major areas of investigation described in this dissertation. The first

area of investigation concerned the effects of PSZ on both spinnability and mechanical

properties of the SiC-based fibers prepared using the University of Florida method.

Fibers were spun from PCS solutions by dry spinning with and without PSZ. Mechanical

properties of these fibers were evaluated after heat treatment at various temperatures








up to 1150C. Fourier transform infrared spectroscopy (FTIR) was used to study the

chemical changes occurring in these fibers during pyrolysis. In an effort to understand

the effect of PSZ on spinnability of PCS solutions, polymer solutions were characterized

using several methods, including measurements of surface tension, contact angles,

theological characteristics (e.g., intrinsic viscosity), and the rate of evaporation of

solvent from the solutions.

The second major area of investigation was the synthesis and processing of

PMS (polymethylsilane) polymers for the fabrication of SiC-based fibers. There has

been limited work on the preparation of SiC fibers from organosilicon polymer blends.

SiC fibers may be fabricated by using PMS polymers and PMS/PCS polymer blends.

PMS polymers generally produce an excess of elemental Si (in addition to SiC) upon

pyrolysis. As indicated earlier, PCS polymers form an excess of elemental C upon

pyrolysis. Therefore, a combination of these two polymers might potentially be used to

form SiC fibers with controlled stoichiometry.

PMS polymers were synthesized in this study by Wurtz-coupling polymerization

of methyldichlorosilanes (MDCS) and methyltrichlorosilanes (MTCS) with sodium (Na) in

refluxing solvent/solvent mixtures. One of the major disadvantage of this method is poor

polymer yields. Polymer yields and molecular weight distributions are quite sensitive to

substituents (pendant groups) in the monomers, order of reagent addition, solvent

additives, reaction temperatures, etc. It has been reported that addition of polar solvents

promote anionic polymerization (such as Wurtz-coupling polymerization) and increase

polymer yields [Mil93; Gau89]. In this study, effects of addition of polar solvents THF

and 1,4-Dioxane on polymer yield and molecular weight were investigated.

One of the main drawbacks of PMS polymers for use in fiber fabrication is that

they are liquids at room temperature and generally have low molecular weight (M, <








1,000 and M_ < 2,500). In order to form fibers from these polymers, an increased

molecular weight and an increased extent of cross-linking are needed so that the

polymers are solids at room temperature (and remain solids during pyrolysis).

Investigations were carried out to increase the molecular weight/cross-linking of these

polymers as well as produce a solid polymer with sufficiently high molecular weight to

permit fiber spinning. Two approaches were utilized to raise the molecular weight/cross-

linking of the polymer: (1) polymerization and cross-linking by heat treatment with

additives and (2) fractional precipitation of higher molecular weight fractions by addition

of nonsolvents. The additives used consisted of polyvinylsilazane (PSZ), dicumyl

peroxide (DCP), and decaborane (DB). The nonsolvents used were a mixture of

methanol and 2-propanol, and acetone. Fibers were spun from heat-treated PMS,

PMS/PCS polymer blends, and fractionally-precipitated PMS polymers, and converted to

SiC fibers by pyrolysis at 1000-1150C in a nitrogen atmosphere.













CHAPTER 2
LITERATURE REVIEW




2.1 Background

There has been much interest in recent years in the preparation of ceramic

materials by pyrolysis of organometallic polymers. A wide range of ceramic materials

can be produced by this method, such as SiC, Si3N4, B4C, BN, SiO2, A1203 and AIN. In

this review, only silicon-based preceramic polymers, viz., polysilanes are discussed.

Polysilanes are a class of polymers with Si-Si backbone in their main chain. The interest

in polysilanes stems from a number of commercially attractive applications such as

precursors to p-SiC fibers, photoresists in multilayer microlithography and photoinitiators

in radical polymerization. Despite the commercial significance of these polymers,

polysilane technology suffers from the lack of well-controlled and reproducible methods

for synthesis of polysilanes with high yields, high molecular weight, and narrow

polydispersity.

A number of factors need to be considered when selecting a polysilane polymer

for a specific application: (1) nature of the ceramic material (chemical composition and

crystal structure) produced after further processing (e.g., heat treatment), (2) elemental

composition of the starting polymer (which influences the final stoichiometry of the

ceramic produced), (3) molecular architecture of the polymer (linear vs. cross-linked

polymer, which strongly influences the ceramic yield), (4) sensitivity to air (i.e., oxygen

and water vapor) of the polymer, (5) starting molecular weight, (6) capacity of the








polymer to be cross-linked at some stage in processing, and (7) solubility in common

organic solvents. As indicated in (1) above, pyrolysis conditions (temperature,

atmosphere, and heating rate) play a very important role in determining the

characteristics (yield, elemental composition, and crystal structure) of the ceramic

produced. The ceramic yield (i.e., weight percentage retained after polymer-to-ceramic

conversion) is an important consideration when discussing the suitability of a polysilane

polymer as a precursor for silicon carbide. As indicated in (3) above, molecular

architecture also has great impact on the ceramic yield of the polymer synthesized.

Cross-linked or branched polymers give much higher ceramic yield than their linear

counterparts. However, excessive cross-linking is generally not desirable, as it would

make the processing of the polymer difficult (i.e., the polymer will be less likely to melt or

to be soluble in common solvents). It is sometimes desirable that the final composition

of the ceramic produced be that of near-stoichiometric silicon carbide. A convenient

method of achieving this is to start with a polymer which has a 1:1 Si:C ratio, such as

polymethylsilanes. This can be contrasted to polyphenylsilanes, for example, which

have Si:C ratios of 1:6, and, therefore, result in SiC/C mixtures upon pyrolysis.

Polysilane polymers were first synthesized by Kipping [Kip21] in the early 1920's

by condensation reaction of diphenyldichlorosilane with sodium. This polymer was not

useful in practical applications since it was intractable (i.e., not processable into useful

articles because of poor solubility and infusibility). Subsequently, in 1949, Burkhard

[Bur49] reacted dimethyldichlorosilane with sodium to produce poly(dimethylsilane),

which also was insoluble in common organic solvents and infusible. In 1975, Yajima and

coworkers [Yaj75; Yaj78A; Yaj78B] were able to convert poly(dimethylsilane) to a

tractable form of polycarbosilane (PCS) by pressure pyrolysis in an autoclave at 450*C.

The polycarbosilanes were then melt spun into fibers which were subsequently heat








treated to form SiC-based ceramic fibers. The conversion process of polydimethylsilane

to polycarbosilane takes place by Kumada rearrangement reactions and is described in

detail elsewhere [Shi58].



2.2 Polysilane Synthesis

The prominent methods of synthesis of polysilanes are:

(i) Wurtz-coupling (dehalocoupling) reactions of chlorosilanes with alkali metals.

(ii) Dehydrocoupling of primary organosilanes in the presence of a catalyst.

(iii) Redistribution/substitution reactions of chlorosilanes.



2.2.1 Wurtz-coupling of dichlorosilanes with alkali metal

2.2.1.1 Mechanism

Wurtz-coupling of dichlorosilanes with an alkali metal is a strongly exothermic

and heterogeneous reaction. The reaction can be represented as follows:



R1 R2 SiCI2 + 2Na solvent. Reflux (R1 R2 Si)n + 2NaCI (2.1)

where R1, R2 can be H, CH3, CHs5, C6H13, etc.



The detailed reaction chemistry is shown in Figure 2.1 fBen91]. Although many

polymers are produced by this route, it suffers from the disadvantages of poor

reproducibility, polymodal molecular weight distributions, and low polymer yields. Poor

reproducibility arises because it is difficult to control the exothermic and heterogeneous

reaction (i.e., the reaction is heterogeneous in that it involves liquid and solid reagents

[Zei86A]). In addition, some reaction variables, such as the purity of the chlorosilanes,

and the state of dispersion of sodium, are difficult to control [Mar92].













Initiation


Cl-Si-CI
12
R


+ 2Na -0


Cl-Si Na+
I-,


Propagation


R1

-Si-Na+ +
12
R


R1 R1 R1
1 1 1
I I I
CI-Si-CI -- Si-Si-CI
12 12 12
R RR


(Rate determining)


1 1
RR
I I
- Si- Si- CI
1 2 1 2
122
RR


+ 2Na -


1 1
R R

- Si- Si Na+
I 2R
R R


(Fast)


Figure 2.1. Reaction scheme for Wurtz-coupling polymerization reaction


+ NaCI


(2.2)


+ NaCI


(2.3)


+ NaCI


(2.4)








Polymer yields and molecular weight distributions are quite sensitive to

substituents (pendant groups) in the monomers, order of reagent addition, solvent

composition, reaction temperatures, etc. The reaction is usually carried out at

elevated temperatures (~1000C) using a suitable alkali metal dispersion. Sodium,

potassium or lithium could be chosen as alkali metals but sodium is usually preferred

because potassium and lithium are relatively more flammable and hazardous. In the

Wurtz-coupling reaction, sodium is normally employed as a dispersion in an appropriate

solvent such as toluene, xylene, THF, etc. As discussed in section 2.2.1.4, the choice of

solvents plays an important role in determining the polymer molecular weight

distribution and polymer yields.

Equation (2.1) is indicative of the fact the polymerization reaction proceeds by

condensation type mechanism. However, Worsfold [Wor88] and Miller et al. [Mil91]

reported, based on the characteristics of the polymers produced during the reaction, that

it proceeds by an addition type mechanism. In an addition type polymerization reaction,

high molecular weight polymer fractions form very early in the reaction and the formation

of high molecular weight polymer is not affected by the stoichiometry of the reagents

(i.e., high molecular weight polymer forms even when one of the reagents is in excess).

Worsfold demonstrated these characteristics for Wurtz-coupling polymerization of

hexylmethyldichlorosilane (carried out in the "normal mode" (see section 2.2.1.2) in

which monomers are added to molten sodium) by isolating high molecular weight

polymer (-105) in the early stages of reaction. The rate determining step in Wurtz-

coupling polymerization is the reaction between silyl radical and monomer, as shown by

equation (2.3) in Figure 2.1. The reaction between chlorine-ended chain and sodium

takes place rapidly (equation (2.4) in Figure 2.1). Weyenberg [Wey69] et al. have








demonstrated by gas chromatography that molecules containing sequences of Si atoms

react faster than molecules containing single atoms.

2.2.1.2 Mode of addition of reagents

At the beginning of the reaction, molten sodium (melting point of sodium =

98.5C) can be added to dichlorosilanes dissolved in a suitable solvent at the reflux

temperature of the solvent ('inverse' mode of addition) or dichlorosilanes dissolved in a

small amount of solvent could be added to the molten sodium dispersed in the inert

solvent ('normal' mode of addition). The inverse mode of addition usually leads to higher

molecular weight polymers with lower polymer yields compared to the normal mode of

addition. The former method is also more hazardous (i.e., due to handling of sodium)

and difficult to control.

Zeigler [Zei86A; Zei87] investigated the effect of rate of monomer addition (in the

normal mode) and sodium addition (in the inverse mode) on the polymodality of the

molecular weight distribution in the synthesis of polymethylphenylsilane. Zeigler

concluded that the rate of reagent addition (monomer or sodium) and the mode of

addition had an important role in determining the molecular weight distribution because

of its influence in controlling the rate of diffusion of reactive species to and from the

sodium surface. When the rates of addition of Na or monomer were kept constant (for a

range of addition rates 80-640 meq/min (i.e., moles equivalent per min)), the molecular

weight distributions were nearly monomodal and the average molecular weights

remained approximately constant at 600,000 for inverse mode of addition and 4000 for

normal mode of addition. However, when the addition rate was varied, there was a

tendency to form a bimodal molecular weight distribution. Figure 2.2 shows the effect of

the rate of Na addition (inverse mode) on the (PhMeSi)n molecular weight distribution.




































106105 104 103102
MW


Effect of reactant addition rate on (PhMeSi)n molecular weight
distribution [Zei86].


Figure 2.2.









(Zeigler et al. did not provide plots of the molecular weight distributions obtained by

using different monomer addition rates (normal mode)).

2.2.1.3 Effect of alkali metal

As indicated earlier, sodium, potassium or lithium could be chosen as the alkali

metal for the polymerization reaction. Based on ease of handling (for example, sodium

is available as 3-5 mm pellets where as potassium and lithium are available as blocks of

materials and need to be cut into smaller sizes for accurately weighing) and flammability

considerations, sodium is normally preferred over the other two. Alternatively, alloys of

sodium and potassium of varying composition could be used but these alloys often

cause degradation of polymer molecular weight and form cyclic oligomers at elevated

temperatures (~1000C) (Na/K alloy promotes hydrogen abstraction from solvent and

causes chain transfer) [Mil93].

The polymerization reactions take place very close to the alkali metal surface

and, hence, the surface area of the alkali metal plays a very important role in

determining the molecular weight distribution of the polymer formed. Worsfold [Wor88]

studied the effect of sodium surface area on the molecular weight of polymers formed

during the polymerization of hexylymethyldichlorosilane and found that rate of

consumption of monomers increased as the sodium surface area increased (Figure 2.3)

(The monomer consumption was monitored by removing small amounts of the reaction

contents periodically during the course of reaction and analyzing the samples by gas

chromatography (GC).) Therefore, to obtain good polymer yield in a reasonable time, it

is important to use a fine dispersion of sodium in the reaction solvent.

The plots in Figure 2.3 show a sigmoidal behavior. The incubation period is

interpreted as the time during which initiation occurs, i.e., according to equation (2.2) in






13










1.0



S0.8-

ro

0.6-
.-


0
E



S0.64
0
o





E 0.2




0 10 20 30 40
Time, min







Figure 2.3. Effect of sodium surface area on the rate of consumption of
hexylmethyldichlorosilane [Wor88]; A: 0.20 m2, 0 : 0.67 m2, o: 4.64 m2
per mole of dichlorosilane








Figure 2.1. It might be expected, based on equation (2.2), that the rate of initiation would

be dependent upon the available sodium surface area. The results in Figure 2.3 are

consistent with this interpretation in that the incubation period decreases with increasing

sodium surface area.

The Wurtz polymerization reaction is strongly exothermic and the initially clear

reaction mixture changes to purple or dark blue color quickly. Miller et al. [Mi191]

attributed this blue color to 'defects' in sodium chloride, i.e., color centers produced by

incorporation of sodium ions in the interstitials of precipitating sodium chloride. Benfield

et al. [Ben91] suggested, based on spectroscopic studies, that the blue color is due to

colloidal Na particles (submicron) formed during the reaction. They collected diffuse

reflectance ultraviolet-visible (UV Vis) spectra of the products formed during the reaction

and found that there were two absorption bands for polysilanes, a sharper band below

400 nm and a broader band centered around 560 nm (Figure 2.4). Comparing this

spectra with published results of sodium colloids and defect centers of sodium chloride,

the authors concluded that the absorptions in the UV-Vis spectra were due to colloidal

sodium formed during the reaction.

2.2.1.4 Solvent effects

The influence of types of solvents on Wurtz-coupling reactions of dichlorosilanes

with sodium was first noted by Miller et al.[Mil93] in the preparation of

polycyclohexylmethylsilane (normal mode). They reported that when diglyme

(diethyleneglycol dimethylether, see Figure 2.5) was added to the reaction mixture of

sodium and monomers, overall polymer yield and molecular weight distribution are

affected (see Table 2.1). When diglyme was added in low concentrations (-10 vol%) in

the polymerization of cyclohexylmethylsilane, there was a significant increase in polymer































30 ';5 s O 700 900
Wavelength (nm)




Figure 2.4. UV-Vis Diffuse Reflectance spectrum of purple solid isolated during Wurtz
polymerization [Ben91].










/o
0 0



15-Crown-5 Ether (CH2)1005
(1,4,7,10,13-Pentaoxacyclopentadecane)


/o\


H2C


H2C

0


CH2

CH2


1,4-Dioxane


H2C CH2

H2 rorn H2



Tetrahydrofuran (THF)


(CH30CH2CH2)20
Diethyleneglycol dimethylether (Diglyme)


Figure 2.5. Chemical formulas of polar solvents used in Wurtz-coupling polymerization








Table 2.1: Effect of diglyme and heptane
dichlorosilane monomers [Mil91].


additions on polymerization of some


Polymer toluene:diglyme Yield, % Mw x103 R
(vol%)
(c-HexSiMe)n 100:0 20 804, 4.5 b 8.7

(c-HexSiMe)n 90:10 35 1477, 24.8 b 0.12

(c-HexSiMe)n 75:25 32 23.1 a

(c-HexSiMe)n 25:75 33 16.5a ...

(n-dodecylSiMe), 100:0 8 1345, 8.4 b 2.73

(n-dodecylSiMe), 70:30 33 476, 40.7 b 0.74

(n-Hex2Si), 100:0 5.9 1982, 1.2 b 3.12

(n-Hex2Si), 95:5 34 1008, 22.3 b 3.4

(n-Hex2Si), 90:10 36 1358, 26.6 b 2.6

(n-HexSi)n 70:30 37 1073, 31.7 b 1.42

(n-dodecyl2Si)n 100:0 3 521, 14.1 b 5.2

(n-dodecylSi), 70:30 34 570, 27.4 b 2.3

(PhMeSi)n 100:0 25 383, 16 b.

Polymer toluene:heptane Yield, % M, x10-3 R
(vol%)
(n-Hex2Si), 84:16 27 1386, 1.1 b

(PhMeSi), 85:15 9 1390, 10.5 b ...


" monomodal; b bimodal; c ratio of amounts of high molecular weight to low molecular weight fractions









yield and average molecular weight. At higher concentration of diglyme (-25 vol%),

polymer molecular weight distribution became monomodal and overall molecular weight

of the polymer decreased drastically. The polymer yield remained relatively high (32%).

The effect on polymer yield of diglyme additions was particularly significant in the case

of polysilane polymers derived from symmetric dialkylsilane monomers (e.g., dicholoro-

di-n-hexylsilane, dichloro-di-n-dodecylsilane) (see Table 2.1). In the case of poly(di-n-

hexylsilane), a typical dialkyldichlorosilane derived polymer, the yield of the polymer was

only 5.9% when synthesized in toluene alone as a solvent The polymer yield increased

approximately six fold (~34-37%) when diglyme was added in amounts of 5, 10, and

30% by volume of toluene. The addition of diglyme also resulted in lower average

molecular weight of the polymer. In the case of dichloro-di-n-dodecylsilane, the addition

of 30% diglyme resulted in large increase in the polymer yield (from 3 to 34%), while the

average molecular weight increased slightly for each mode in the distribution. Also, the

high molecular weight proportion of the polymer decreased significantly.

Table 2.1 also shows the effect of addition of a non-polar solvent (i.e., heptane)

on polymerization of diakyldichlorosilane monomers (dicholoro-di-n-hexylsilane) and

arylakyldichlorosilane monomers (phenylmethyldichlorosilane). Addition of heptane

caused an increase in polymer yield for the case of polymerization of dicholoro-di-n-

hexylsilane and a decrease in polymer yield for the case of polymerization of

phenylmethyldichlorosilane. The heptane addition resulted in lower average molecular

weight in the former case and higher average molecular weight in the latter case. Miller

et al. explain that polymerization of arylalkyldichlorosilanes is highly exothermic and

takes place rapidly because arylalkyldichlorosilanes are not sterically hindered unlike

dialkyldichlorosilanes. Solvent effects on polymerization of arylalkyldichlorosilanes are








not well understood because of the difficulty in obtaining controlled kinetic data (due to

the rapid polymerization rates).

Miller et al. attempted to explain the observation of increased polymer yield for

dialkyldichlorosilanes in the presence of polar cosolvents such as diglyme, crown ethers

etc., by suggesting the presence of silyl anion radical intermediates (such as shown in

equation (2.2) in Figure 2.1) as the main propagating species in the polymerization. It is

well known that polar solvents aid in the transfer of electrons from metal to monomer,

promoting formation of silyl anion radicals [Gau90; Mi193; Bil83]. A large number of

radicals formed would mean a large number of initiation sites and this would favor an

increased polymer yield. However, Miller et al. [Mil91] reported the same beneficial

effect (i.e., improved yield) when 16 vol% of non-polar solvent (i.e., heptane) was added

to toluene in the polymerization of dialkyldichlorosilane (such as dichloro-di-n-

hexylsilane).

Zeigler et al. [Zei87] have developed a model concerning the bulk solvent effects

in the polymerization of dialkyldichlorosilanes when monomer is present in excess

compared to sodium (inverse mode of addition). According to their model, yield and

molecular weight are determined by effective monomer concentration at the sodium

surface. This depends on the rate of diffusion of monomer to the sodium surface, which,

in turn, depends on degree of coverage of the sodium by growing polymer chains (see

Figure 2.6). In a "good" solvent (i.e., in which the difference in polymer and solvent

solubility parameters, A6 = 8p-s,, approaches zero), polymer-solvent contacts are highly

favored and the polymer coils are relatively extended in the solvent. Thus, polymers

tend to remain in the solvent phase and tend not to adsorb on the sodium particle

surfaces. The monomer continues to have easy access to sodium surface and






















"GOOD SOLVENT"





(R1R2S 2 ]


(I- Spl I SMALL)


"POOR SOLVENT"



Op

C(RR2SCW2 18


(ISs- SpI LARGE)


Figure 2.6. Schematic illustration of the influence of solvent on the polymer/sodium
particle interaction during Wurtz polymerization [Zei87].








new initiation reactions can occur readily. This tends to result in high polymer yield

(because monomer is "consumed" readily) and low average molecular weight. (Because

there are a large number of chains, the amount of chain extension is limited since the

supply of monomer is fixed.) In a poor solvent on the other hand, polymer coils are

contracted and there is a much greater tendency for the polymer chains to absorb on

the sodium particle surfaces. Since the direct path of monomer to the sodium surface

(through the solvents) is impeded now, the monomer is forced to diffuse through

polymer chains. This tends to promote propagation reactions (at the reactive chain

ends) (i.e., causes formation of longer chain polymers) and leads to lower polymer yield

and higher overall polymer molecular weight. In the extreme case when the solvent is

'too poor', the polymer would tend to precipitate out of solution, which is not desirable.

Thus, Zeigler et al. model suggests that there is an optimum A8 for a given polysilane

polymer-solvent system which would dictate the yield and overall molecular weight of

the polymer.

Gauthier and Worsfold [Gau89] investigated the influence of cosolvent 15-crown-

5 ether ('phase-transfer catalyst') on the Wurtz-coupling polymerization of n-

hexylmethyldichlorosilane (Figure 2.5 shows structure of 15-crown-5 ether). The primary

solvent used was toluene and the amount of 15-crown-5 ether used was in the range of

0.25-4 mol% (of hexylmethyldichlorosilane). They also found that in the presence of the

cosolvent, the polymer yield becomes high, the overall molecular weight of the polymer

decreases, and the molecular weight distribution changes from bimodal to monomodal.

Figure 2.7 shows the amount of monomer consumed as a function of time in the

study by Gauthier and Worsfold. They suggested that silyl anionic intermediates (shown

























I-
U-


Z-
60 -



40


0.25%\
1%
20 -




0 20 40
REACTION TIME (min)





Figure 2.7. Rate of disappearance of monomer n-hexylmethyldichlorosilane as a
function of time and weight percent of 15-crown-5 ether [Gau89].








by equation (2.2) in Figure 2.1) are involved in the polymerization of n-

hexylmethyldichlorosilane and claimed that 15-crown-5 ether accelerated the

occurrence of initiation reactions. Although data is limited, it is evident that the rate of

monomer consumption is increased with small additions of 15-crown-5 ether.

2.2.1.5 Temperature effects

Miller et al. [Mil93] investigated the effect of temperature on the molecular weight

distribution and yield for polymers produced from diaryl and dialkyl substituted

chlorosilanes. They found in the case of polymerization of methylphenyldichlorosilane in

toluene that lowering the reaction temperature to 65C (from the refluxing temperature

of 110C) decreased the total yield from 25% to 10%, while causing an increase in

molecular weight of the polymer. In addition, the molecular weight distribution changed

from bimodal to monomodal (Table 2.2). However, in the presence of a polar solvent

(i.e., 15% diglyme), lowering the reaction temperature to 65C (instead of the reflux

temperature) resulted in the opposite trends from what is noted above, i.e., the

polymer yield increased slightly and the polymer molecular weight decreased. (The latter

changes may have been within the limits of experimental error.)

When polymerization was carried out in a blend of toluene/15% heptane,

lowering the reaction temperature to 65C had no effect on polymer yield or overall

polymer molecular weight. Miller et al. also reported that low temperature (650C)

polymerization of alkyl substituted chlorosilanes (such as dichloro-di-n-hexylsilane) took

place sluggishly. The typical change in color to purple or dark blue was conspicuously

absent. In addition, the yield for such a polymerization was less than a percent (i.e.,

essentially no polymerization occurred).

Jones et al. [Jon94] investigated polymerization of methylphenylsilane in THF at

low temperatures (-79"C) using a sodium/electron acceptor complex. The electron











Table 2.2. Effect of temperature on polymerization of methylphenyldichlorosilane [Mil93].


( PhMeSiCI2 + Na


--(PhMeSi)n)


Solvent Temperature,OC Additive Yield, % Mw x 10-3 M x 10-3

Toluene Reflux 25 383, 16 a 267, 8.1
a
Toluene 65 10 1073b 377b
Toluene/15% Reflux 9 1390,10.5 375, 6"
Heptane a
Toluene/15% 65 9 1367 b 580 b
Heptane
Toluene/15% Reflux Diglyme 25 23.8 b 9.7 b
Heptane (15%)
Toluene/15% 65 Diglyme 28 14.2 b 6.7 b
Heptane (15%)


a bimodal; b monomodal








acceptors used in the polymerization were naphthalene, anthracene and tetraphenyl

ethene, and were used in stoichiometric excess to disperse sodium. The disadvantages

of typical Wurtz-type polymerization (e.g., low polymer yields, poor reproducibility)

persisted, but polydispersities of the polymer produced were much lower (1.5-3) than

that of typical Wurtz-type polymerization (> 5).

2.2.2 Ultrasonically-activated Wurtz-coupling reactions

Matyjaszewski et al. [Mat88; Mat91; Kim88] pioneered the use of ultrasonic

energy in the Wurtz-coupling synthesis of polysilanes (derived from aryl-substituted

monomers) having high molecular weight and monomodal distributions. Use of

ultrasonic energy enabled reactions to be performed at low or ambient temperatures.

The principle of ultrasonic polymerization is based on the implosive collapse of cavities

with very high pressures and temperatures existing locally for short duration of times

[Pri94]. The ultrasonic energy is generated using an immersion-type probe or ultrasonic

bath. The reasons for obtaining monomodal and high-molecular-weight distributions for

the polymers synthesized by ultrasonic method is attributed to the formation of high

quality sodium dispersions which are continuously regenerated during the coupling

process with continuous removal of sodium chloride byproduct from the sodium

surfaces.

Matyjaszewski et al. report that polymer molecular weight distribution becomes

broader and the average molecular weight decreases as the reaction temperature

increases. It was also observed that prolonged sonication (both during and after the

addition of monomer) results in degradation of high-molecular-weight components, as

shown in Table 2.3. (This results in polymers with lower average molecular weight and

lower polydispersity.)








Table 2.3: Effect of sonication time on molecular weights and polydispersities of
polymethylphenylsilane [Kim88].

Molecular Sonication time (min) Sonication time (min)
Weight During addition of monomers After addition of monomers
5 10 15 30 60 80 120
Mn. 10-5 3.8 2.24 2.30 1.82 1.48 1.06 0.40
Mw. 10- 17.3 6.68 6.35 3.73 2.57 1.57 0.47
MJMn 4.5 2.98 2.71 2.05 1.73 1.48 1.17



Matyjaszewski et al. also observed that ultrasonic polymerization of dialkyl-

substituted chlorosilanes occur sluggishly when compared with that of diaryl-substituted

dichlorosilanes and that cosolvent additions (e.g., diglyme) and higher temperatures

were required to obtain any meaningful yield of polymer. In general, polymer yields for

ultrasonic synthesis of polysilanes are low when compared to those of classic Wurtz-

coupling reactions performed at high temperatures (i.e., the reflux temperatures of the

solvents). Furthermore, the ultrasonic synthesis method is amenable only for aryl-

substituted dichlorosilanes. However, it has better potential for control of polymer

molecular weight and is less hazardous to perform, since the reactions are carried out at

or near ambient temperatures.

2.2.3 Polymerization of monoalkylchlorosilanes

Although a lot of information is available on the Wurtz-coupling reactions of aryl-

substituted and dialkyl-substituted chlorosilanes, information is scarce on the

polymerization of monoalkylchlorosilanes. Seyferth et al. [Sey92] have synthesized

polymethylsilanes by condensation of methyldichlorosilanes with sodium (normal mode)

in a solvent mixture of hexane and THF (7:1 by volume) at reflux for 16 h. The polymer

was a liquid with a composition of ((CH3SiH)0,8(CH3Si)0o2) (elucidated by NMR








spectroscopy) and was produced in high yields (i.e., 60 to 70%). Average molecular

weights for these polymers were reported to be low (620-690). When the polymerization

was carried out in THF instead of a mixture of THF and hexane, the resulting polymer

apparently had higher molecular weight (absolute numbers were not reported) and more

cross-linked structure. The latter conclusions were based on NMR studies (showing a

lower concentration of Si-H bonds) and thermogravimetric analysis (TGA) (showing

higher ceramic yield). When the reaction was carried out in xylene (under refluxing

conditions), the yellow-colored polymer was produced with a yield of 40%, molecular

weight in the range of 520-600 (Mw), and structure of ((CH3SiH)04(CH3Si)o36)

(determined from NMR).

Qiu and Du [Qiu89A; Qiu89B] prepared polymethylsilane polymers by

condensation of methyldichlorosilane with sodium (normal addition mode) in a blend of

toluene and dioxane (33:67 vol% ratio). It is expected that dioxane, a dipolar solvent, will

promote polymerization of MeHSiCI2 (i.e., higher reaction rate). (The effect of polar

solvents on Wurtz-coupling reaction of dichlorosilanes with sodium was discussed in

section 2.2.1.4) The polymerization was carried out at the reflux temperatures of the

toluene-dioxane solvent mixture. The end point of polymerization was determined by

testing for the acidic nature of the reaction contents. The reaction was stopped when the

reactions contents did not test acidic (pH=6-7). The reaction contents were separated

from the NaCI precipitates by filtering and the polymer was isolated by evaporation of

the polymer solution under vacuum. The polymer was fractionated by adding, drop by

drop, a mixture of methanol and 2-propanol with vigorous stirring. The precipitate was

collected and dried in a vacuum oven at room temperature for 2 h. The polymers

synthesized by this method were used in studies involving oxidative cross-linking, photo








cross-linking, and room temperature vulcanization. The polymer was produced in ~45%

yield and had an appearance of pale yellow waxy solid with M, -1,800.

2.2.4 Polysilane copolymers

As discussed in section 2.1, when dialkyldichlorosilane is reacted with sodium in

a refluxing solvent, polydimethylsilane polymer is formed which is infusible and insoluble

in common organic solvents (e.g., toluene, benzene, xylene, etc.). West et al. [Wes81;

Wes86A] discovered that when phenylmethyldichlorosilane was added to

dialkyldichlorosilane in a 1:1 proportion by volume, and the reaction was carried out

under same conditions, the resultant polysilane copolymer was highly soluble in

common organic solvents (e.g., toluene, xylene, etc.). West et al. referred to this

copolymer as "poly(silastyrene)" (PSS). The copolymerization reaction can be

represented as:



CH3 C6H5
>100C I I
(CH3)2SSiCII + 6 5C3Si2 Na -(--Si Si- (2.5)

CH3 CH3




The PSS copolymer had a bimodal molecular weight distribution with modal values of

-15,000 and -300,000.

2.2.5 Dehydrocoupling

Harrod et al. [Har88; Mu91A; Mu91B; Ait89; Ait87; Ait85] were the first to report

a catalyst-based synthetic route for polysilanes prepared from primary organosilanes








(e.g., RSiH3, where R is an alkyl or aryl group) with the evolution of hydrogen. The

reaction can be represented as



R
Catalyst I
nRSiH3 -- -(- Si -- + H2 (2.6)
20-65C n

H


The catalysts used were early transition metal complexes of titanium and zirconium,

namely, bis(rl-cyclopentadienyl) dimethyltitanium (CpTiMe2)(Dimethyl Titanocene,

DMT) and bis (r5-cyclopentadienyl) dimethylzirconium (Cp2ZrMe2) (Dimethyl

Zirconocene, DMZ). Mu and Harrod [Mu91A] have investigated polymerization of

methylsilane by dehydrocoupling in the presence of DMT catalyst and reported

significantly higher yield of polymer in the form of a glassy solid in comparison to that

produced by classic Wurtz-coupling reactions. Table 2.4 shows polymerization

conditions (temperature, catalysts, solvents, time, amount of monomer used) and

characteristics of the polymers produced (yield, molecular weight, etc.). Their method,

however, suffers from the following disadvantages: (i) the reaction must be performed at

9-10 atm at 500C because methylsilane is a gas at room temperatures; enhancing

reaction rate would require working at higher pressures, which in turn requires

sophisticated instrumentation in order to perform the experiments safely and (ii)

handling methylsilane is dangerous since it is spontaneously flammable in air.

The monomer, methylsilane, was synthesized from methyltrichlorosilane.

Methyltrichlorosilane was reacted with a suspension of lithium aluminum hydride in THF

at 50C for 3 h and then reaction products were cooled under liquid N2 temperature to









Table 2.4. Summary of methylsilane polymerization by catalytic dehydrogenation reactions [Mu91A].


Run Solvent Catalyst MeSiH3 T, "C Time, Amount of Yield, % Mw C Mn MJ/Mn %
# (psi x Lb) days PMS, g Cyclicsc
1 cyclohexene DMT 120 x 0.12 20 6 1.52 90 1590 790 2.01 0.4
toluenea (50 mg)
2 cyclohexene DMT 130 x 0.12 20 9 1.82 -100 6350 1200 5.30 1.6
+toluene" (50 mg)
3 cyclohexene DMT 140 x 0.12 20 12 1.96 -100 10100 1250 8.10 3.4
+toluene" (50 mg)_
4 cyclohexene DMT 120 x0.12 45 4 1.68 -100 7890 1240 6.36 0.5
+toluene" (50 mg)
5 cyclohexene DMT 110 x0.12 65 1 1.53 -100 12990 1260 10.30 0.5
+toluene" (50 mg)
6 Toluene DMT 100 x0.12 20 9 0.37 26 830 560 1.48 4.6
(50 mg)
7 cyclohexene DMZ 110 x0.12 20 5 1.43 -92 1730 800 2.16 ...
+toluene" (60 mg)
8 cyclohexene DMZ 130 x0.12 20 7 1.81 -100 6010 1080 5.56 0.8
+toluene" (60 mg)
9 cyclohexene DMZ 125 x0.12 20 9 1.75 -100 9990 1350 7.40 2.5
+toluene" (60 mg)_
10 cyclohexene DMZ 100 x0.12 65 1 1.40 -100 Insoluble
+toluene" (60 mg)
11 Toluene DMZ 110 x0.12 20 7 0.99 64 1020 620 1.65 3.8
(60 mg)
* Cyclohexene:toluene proportion 70:30 (Vol %); b Volume of silane gas used
c Based on the assumption that the low molecular weight species in the gel permeation chromatograms are cyclic oligomers.








trap methylsilane. The subsequent polymerization of polymethylsilane was carried

out in cyclohexene solvent which helps to avoid build up of hydrogen produced during

reaction by promoting hydrogenation of cyclohexene to cyclohexane.

Molecular weight characteristics of some of the polymers synthesized by Mu and

Harrod are illustrated by the GPC profiles in Figure 2.8. (The chromatograms are for

samples associated with the entries in Table 2.4. The "A" chromatograms, from top-to-

bottom, correspond to run #s 3,2,1, and 6, respectively, and the "B" chromatograms,

from top-to-bottom, correspond to run #s 9,8,7, and 11, respectively.) The bimodal

distributions that develop with longer reaction times are reported to be typical of

dehydrocoupling of primary organosilanes. The lower molecular weight peak (appearing

as a shoulder in most chromatograms in Figure 2.8) is attributed due to cyclic oligomers.

The high polydispersity of these polymers can be attributed to branching/cross-linking

that occurs at residual SiH3, SiH2 and SiH groups during prolonged reaction.

Mu and Harrod also studied polymerization of phenylsilane by dehydrocoupling

using the aforementioned catalysts The polyphenylsilane polymer synthesized with

DMT and DMZ had average degrees of polymerization of 1,000 and 2,000, respectively

(corresponding to molecular weights of 46,000 and 92,000, respectively). In both cases,

gel permeation chromatograms did not suggest the presence of cyclic oligomers (i.e.,

low molecular weight oligomers were not observed). Harrod et al. also report that

secondary organosilanes (e.g., phenylmethylsilane) do not polymerize easily under

similar reaction conditions and form only dimers and trimers. Brown-Wensley [Bro85;

Bro87] has shown that a good catalyst for conversion of secondary silanes (R2SiH2) to

dimeric silanes (HR2Si-SiR2H) is a (Ph3P)3RhCI complex.

While Harrod's work on the synthesis of poly(arylsilanes) (e.g.,

poly(phenylsilane)) indicated that cyclic oligomers do not form, Campbell and Hilty

















































3060 2M :20 1 87 0.16S .-:'


F'-40 :63 22.0 1.A7 0.16 xtlC


Molecular Weight






Figure 2.8. GPC of polymethylsilanes synthesized by Mu and Harrod [Mu91A]. A: DMT
catalyst B: DMZ catalyst








[Cam89] have shown by gas chromatography that they are the main products in the

polymerization of alkylsilanes (e.g., n-butylsilanes) catalyzed by DMZ. This is attributed

to the ability of DMZ to promote reversible reactions for both primary and secondary

silanes. In the case of polymerization of methylsilane, Campbell et al. reported the

presence of a small amount of cyclic oligomers (n=5 to 10), in agreement with the

results of Mu and Harrod. Thus, it can be concluded that formation of cyclic oligomers is

inevitable in the polymerization of alkylsilanes by dehydrocoupling.

In general, DMZ and DMT catalysts are effective for polymerization of primary

organosilanes. However, as shown by Mu and Harrod [Mu91A], polymerization rates for

primary silanes are about ten times faster using DMZ compared to using DMT.

Nevertheless, molecular weight characteristics of the polymers synthesized with the

two catalysts are essentially identical. The DMT-catalyzed polymerization exhibited a

pronounced induction period and a complete reduction of titanium to Ti (III). In contrast,

an induction period was absent for DMZ catalyzed reactions and a slight auto-

acceleration in the reaction rates was observed in the reactions at low catalyst or

monomer concentrations.

The mechanism of catalytically activated dehydrocoupling of organosilanes is

complex. Harrod [Har88] suggested (based on NMR studies) that the mechanism

involved formation of titanium (IV) silylhydride ( Cp2Ti(H)(SiH2R) ) which decomposed by

elimination of a-hydride from the SiH2R group, followed by release of H2 from the

complex to give Cp2Ti=SiHR (silylene) complex. A number of metallocene (catalyst)

derivatives are formed during polymerization that can be isolated and these compounds

are presumed inactive in the polymerization cycle. Propagation then occurred by

repetitive insertion of the silylene into a Ti-Si bond (a rapid addition mechanism) in which

the intermediates are not observable because they are short-lived or because they are








spectroscopically 'silent' (i.e., absent in NMR because of paramagnetic nature). At

present, a mechanism for chain termination is not elucidated, although catalyst-induced

chain scission has been observed in the polymerization of cyclohexasilanes.

Since early transition metal complexes are not effective for dehydrocoupling of

secondary silanes, other catalysts have been investigated. Corey et al. [Cor91]

synthesized disilanes through pentasilanes by using a Cp2ZrCI2-nBuLi mixture as a

catalyst for dehydrocoupling of phenylmethylsilanes in toluene at 90C. (This

temperature is higher than that used for dehydrocoupling of primary silanes.) This

condensation reaction of secondary silanes is sensitive to steric effects (i.e., steric

hindrance) of the substituents, as observed by the sluggish reaction of Ph2SiH2

compared to PhMeSiH2 [Cor91].

Sakakura et al., [Sak91; Sak93] have developed a method for producing

polysilanes by dehydrocoupling of primary organosilanes using a lanthanoid complex

(1.5 wt%) as a catalyst. They reported that lanthanoid complexes have higher activity

and selectivity than early transition metal complexes used by Harrod et al. In this case,

the dehydrogenative reactions of phenylsilanes were performed at temperatures

ranging from 200C to 1600C and in the presence of a solvent such as toluene or

benzene, with reaction times extending from several hours to several days. The authors

reported that higher polymerization temperatures and longer reaction times lead to

higher polymer molecular weight. The effect of the above variables in the polymerization

of phenylsilane in the presence of hydrobis(pentamethylcyclopentadienyl)-neodymium

catalysts (lanthanoid complex) is illustrated in Table 2.5.












Table 2.5: Effect of time and temperature on polymerization of phenylsilane in the
presence of a lanthanoid complex [Sak93].

Temperature, C Time, days Product M" Mn
_Appearance
25 15 oil 520 1.26
80 2 gum 780 1.37
100 2 gum 990 1.54
130 2 solid 1600 1.91
130, 160 a 2, 7 a solid 4380 3.09
" 2 days at 130"C followed by 7 days at 160"C.








Berris [Ber92] has also developed a process for synthesizing polysilane

polymers with a M, of -1000-1500 by dehydrogenative coupling of primary

organosilanes in the presence of (1-1.7 wt%) dimethyldialkylphosphine nickelhalide

(e.g., 1,2-bis(dimethylphosphine) ethanenickel(ll)chloride, dmpe NiCI2) at temperatures

of 20C to 500C in the presence of an inert solvent. The reaction time varied from 1 h to

10 days, depending on the temperature used (i.e., lower reaction times were used at

higher temperatures). The dmpe NiCI2 catalyst was reported to have a much higher

activity than the early transition metal complexes used by Harrod et al.[Har88].

Seyferth et al. [Sey88; Sey90; Sey92; Sey93] have used dehydrogenative

coupling to cross-link low molecular weight polymethylsilanes (containing multiple

secondary or tertiary Si-H bonds) which had been synthesized by the Wurtz-coupling

reaction of methyldichlorosilane with sodium in hexane/THF. This resulted in polymers

that could be pyrolyzed to produce near-stoichiometric SiC with high yield (in the range

of 95-98%). The low molecular weight polymethylsilanes were reacted with ~3 wt%

cyclopentadienyl zirconium hydride catalyst in an inert solvent (such as hexane) at reflux

temperatures. (Hexane was chosen because it readily dissolves the catalyst and it has a

low reflux temperature.) Seyferth et al. [Sey93] observed that the products of the

dehydrogenative coupling reaction of low-molecular-weight polysilanes with

cyclopentadienyl zirconium hydride catalyst ranged from oil to solid (both orange

in color) depending upon time-temperature conditions of the reaction. In order to impart

infusibility to articles prepared from these cross-linked polymers (e.g., fibers), photolysis

is required which can be accomplished by UV irradiation in hexane for 2 hours.

Tilley [Til91; Til93] developed a method of producing cross-linked, high-

molecular-weight, silicon-rich polymers by dehydrogenative coupling reactions of

organosilanes. The reaction of more than one Si-H group per silicon center caused








cross-linking of chains and increased molecular weights. A wide range of homopolymers

with different structures were prepared by modifying the reaction conditions to vary the

degree of branching (cross-linking) or chain extension. For example, when 1,3-

disilylbenzene(1,3-(H3Si)2C6H4) or 1,3-dimethylsilyl benzene (1,3-(CH3H2Si)2C6H4) was

reacted in the presence of cyclopentadienyl zirconium hydride catalyst (Cp2(ZrH2)2), the

resulting polymer was highly cross-linked and high in molecular weight. The disilyl

monomers (1,3-disilyl benzene or 1,3-dimethylsilylbenezene) developed by Tilley were

prepared by reacting tetraethoxysilanes (Si(OEt)4) or methyl triethoxysilanes

(CH3Si(OEt)3) with dibromobenzene and magnesium in an inert solvent, followed by

reduction of the intermediate compound (1,3-di(triethoxylsilyl)-benzene or di-

(trimethoxysiloxy)-benzene) with lithium aluminum hydride. The dehydrogenative

polymerization was then carried out by adding organosilanes (containing multiple Si-H

groups per silicon center) drop by drop to a benzene solution containing the catalyst and

stirring for 24 hours at 20-650C under nitrogen. Many of the polymers prepared by

Tilley's method were highly cross-linked and were insoluble in common solvents (e.g.,

toluene), indicating difficulty in controlling cross-linking reactions. The soluble polymers

exhibited Mw ranging from 5,500 to 90,000 and Mn ranging from 1,300 to 2,200.

2.2.6 Redistribution/substitution reactions:-

Baney et al. [Ban82; Ban83; Ban85] prepared another class of polysilane

polymers, methylpolysilanes (MPS polymers) by catalytic redistribution reactions

involving Si-Si/Si-CI bonds of methylchlorodisilane mixtures' The methylchlorodisilane

mixtures, comprising 55 wt% [MeCI2Si]2, 35 wt% Me2CISiSiMeCI2, and 10 wt%



1 Methylpolysilane (MPS) polymers, as described by Baney et al. have a structure of [((CH3)2Si)x(CH3Si),],.
They are different from polymethylsilane (PMS) polymers discussed earlier, which have a structure of
[(CH3SiH),(CH3Si),Ln-








(Me3CI3Si2), were obtained as fractions from an industrial process for manufacture of

methylchlorosilanes. The catalyst used for the redistribution/substitution reactions of

methylchlorodisilane mixtures was tetrabutylphosphonium chloride. The proposed

reaction scheme for these reactions is shown in Figure 2.9.

The rearrangement of disilanes into a monomer/polymer mixture occurred when

the disilane mixture was heated to 250C. Additional monomeric methylchlorosilanes

formed even after the starting disilanes reacted completely. This occurred by the

reaction of any Si-CI bond with a terminal Si-Si bond in the polysilane backbone. The

amount of monomers formed and the extent of polymerization were controlled by

manipulating the heating schedule and final reaction temperature. The resulting yellow-

colored methylchloropolysilane polymers (MCPS) were soluble in toluene and had

polycyclic structures with seven rings per molecule ((Me2Si)3(MeSi),17CI)) as determined

by gas chromatography (see Figure 2.10). The Si-CI bonds in the

methylchloropolysilane polymers were highly reactive and permitted easy chemical

modification (such as reaction with Grignard reagents (alkyl-magnesium halide or

phenyl-magnesium halide) to form methylpolysilane polymer (MPS). According to

Baney et al., MCPS polymer reacts readily with Grignard reagents to replace the

reactive Si-CI groups with more stable Si-R groups. The modified polymers can be melt

spun to form fibers which can be subsequently pyrolyzed to SiC as discussed in section

2.4.

The chemical modification of methylchloropolysilanes (MCPS) can also be

accomplished by reducing MCPS over a slurry of lithium aluminum hydride under an

inert blanket in a refluxing solvent such as toluene [Ban83]. The excess reducing agent

is neutralized by adding water and aqueous NaOH and the solution is subsequently

filtered to give a yellow-colored polymer of composition ((CH3)2Si)o0(CH3Si)04),).
















Si- CI + Me- Si- Si- Me ~250
I I Catalyst
Cl CI


CI

- SSi-S Me + MeSiCI
CI
CU


CI Me CI
25I
&Si-CI + Me-Si- Si-Me CaaSi-Si-Me + MeSiCL

CI CI CI

CI Me Me
I I
~250"C
Si CI + Me- Si- Si-Me Clyst Si-Si-Me + MeSiCI
I I Catalyst
CI CI CI


Figure 2.9. Scheme for redistribution/substitution reactions of chlorodisilanes.


(2.7)





(2.8)





(2.9)



















Me


*MeSi


Figure 2.10. Structure of methylchloropolysilane polymer [Ban83].








(Information on oxygen incorporation into the polymer due to the addition of water and

NaOH was not reported.)

The main disadvantage of Baney et al. MPS polymers from the point of view of

subsequent processing for ceramic articles (e.g., fibers) is their poor oxidative stability

(MPS polymers are pyrophoric). Burns [Bur90] attributed this tendency for spontaneous

oxidation to a large number of Si-H groups in the polymer structure. The polymer

develops a cross-linked structure upon oxidation due to the formation of Si-O-Si

networks. Burns developed a remedy to the problem of oxidative instability in these

MPS polymers by inserting multiple unsaturated bonds (such as acetylene or phenyl

acetylene or diene compounds). According to Burns, by selectively inserting multiple

unsaturated bonds in the Si-Si backbone, the final Si:C stoichiometry can also be

controlled (unmodified MPS polymers typically yield silicon-rich ceramic residue). The

insertion reaction can be carried out by reacting MPS polymers with ~8 wt%

unsaturated compounds (e.g., phenyl acetylene) in the presence of a transition metal

catalyst (e.g., tetrakis (triphenylphosphine) palladium or tris(tri-phenylphosphine)

rhodium chloride) in an inert solvent such as toluene at reflux temperatures for ~20

hours. In addition to better oxidative stability and control of stoichiometry, Burns' method

presents opportunities to synthesize polycarbosilanes by introducing unsaturated

moieties between Si-Si bonds. An example of such a synthesis is reported as the

reaction between 1,3-butadiene with a linear polysilane polymer [Bur90].

Bujalski et al. [Buj90] developed an alternate method of synthesis of chlorine-

containing polysilanes. These polymers were prepared by reacting a mixture of 70-99

wt% of one or more of chlorine-containing disilanes (e.g., ((CH3)2CISi)2,

(CH3Si)2CISiSiCI2CH3, (CH3CI2Si)2 etc.) with one or more of monoorganosilanes (e.g.,

C6H5SiCI3). The reaction required 0.1 to 2 wt% rearrangement catalyst (e.g., quaternary








ammonium halide, quaternary phosphonium halide, etc.) at temperatures ranging from

100C to 340C. The polysilanes (solid at room temperature) were easily converted to

silicon carbide by pyrolysis at elevated temperatures (at >1000C). According to Bujalski

et al., the use of a monoorganosilane (of structure R'SiX where R' is methyl, phenyl or

octyl groups and X is chlorine) (or a mixture of monoorganosilanes) permits control of

the glass transition temperature T, of the polysilanes as well as the stoichiometry of the

silicon carbide produced. Their work suggests using monoorganosilanes with silyl

groups R'Si (where R'=n-octyl) allows a greater reduction in Tg of the polymer compared

to using monoorganosilanes R'SiX (where R'=phenyl). Bujalski et al. reported that all the

n-octyl groups are lost as olefins upon pyrolysis and this results in a carbon-deficient

ceramic. In contrast, polymers with pheny! groups produce a carbon-rich ceramic after

pyrolysis. However, the final Si:C stoichiometry also depends significantly on the

presence of methyl radicals in the polysilane (CH3Si or (CH3Si)2Si), which are generally

not lost upon pyrolysis. Bujalski et al. indicated that presence of the n-octyl or phenyl-Si

units enabled "fine tuning" of the silicon and carbon contents in the ceramic.



2.3. Pyrolysis Behavior

Carlsson et al. [Car90] have studied the pyrolysis behavior of various silicon

backbone polymers such as polyphenylsilanes, poly-n-2-hexylsilanes, and

polydimethylsilanes by means of thermogravimetric analysis (TGA) and Fourier

transform infrared spectroscopy (FTIR). Polymers for the TGA study were heated at

10C/min with 10 min isothermal holds at 200C, 400C, 600*C and 1000C. Polymers

for FTIR study were heated at 3600C, 454C, 650C, and 1200C, with a 90 min hold at

each temperature. Their results are summarized in Table 2.6. The effect of pendant

groups (such as phenyl) on pyrolysis yield is particularly noticeable. For example,











Table 2.6. Ceramic yields and chemical compositions of polysilane homopolymers,
copolymers and terpolymers [Car90].

Batch Polymer" Yield (wt%) Inorganic Residue Analysis (wt%)
Theor.b Observed' Cd Si d Calculated SiC Yield
SiC Yielde (%Theor)'

Homopolymers
I (CH3-Si-CH3)n 69 1.0 42 58 0.8 1.2
II (C6Hs-Si-CH3)n 33 24.6 55 39 13.7 41.5
III (CeH13-Si-CH3)n 31 5.8 32 67 5.6 17.5

Copolymers
IV (C6Hs-Si-CH3)10
(CH3-Si-CH3)10 51 13 62 51 7.1 15.4
V (C6Hs-Si-CH3)1.0
(C6H13-Si-CH3)1.0 32 8 49 49 5.9 17.9
VI (CH2=CH-Si-CH3)1.0
(C6Hs-Si-CH3)9o 36 39.1 65 38 19.5 54.5
Terpolymers (5:5:1)
Vila (C6Hs-Si-CH3)
(C6H13-Si-CH3)
(CH2=CH-Si-CH3)9 36 7 47 54 5.3 15.6
VIIb (C6Hs-Si-CH3)
(CH 13-Si-CH3)
(CH2=CH-Si-CH3)h 36 22 53 42 14.8 41.2
VIII (C6Hs-Si-CH3)
(CH3-Si-CH3)
(CH2=CH-Si-CH3) 52 27 61 45 15.1 29.1
IX (C6Hs-Si-CH3)
(C6H,3-Si-CH3)
(CH2=CH-CH2-Si-CH3) 51 21 61 47 11.6 22.8

a The functional groups shown are attached to Si as side groups.
b Theoretical conversion to SiC (for e.g., (CH3-Si-CH3)n --- SiC is 69% conversion).
c Observed ceramic yield.
d C+Si total in some cases exceed 100%. The totals appeared as such in the paper and presumably reflect
experimental error in the measuring technique.
e The ceramic residue consists of SiC and C. The percentage of SiC is calculated by multiplying the
experimentally observed yield by the factor (100-C)/70 wt% where C is the carbon content of the residue.
(Note that the theoretical composition of SiC is 70 wt% Si/ 30 wt% C.)
S[Calculated SiC yieldd/theoretical SiC yieldb]xl00.
g low MW viscous oil.
" high MW fraction.








polydimethylsilane with methyl groups as substituents gives high theoretical yield

(69%) but low pyrolysis yield ( ~1%). (Experiments by Wood [Woo84] and West et al.

[Wes81] also confirm this observation.) Replacement of a methyl substituent by a bulky

phenyl group (as in polyphenylmethylsilane) results in improved pyrolysis yield, ~25%

(presumably due to the retention of some phenyl groups during pyrolysis). (It is also

possible that high yield depends on the ability to develop a Si-C-Si-C backbone with

sufficient cross-linking.)

Figure 2.11 shows the FTIR spectra of a polysilane terpolymer (polymer Vllb

shown in Table 2.6) during pyrolysis to 12000C. The polymer was cast as a thin film on a

silicon wafer and heated under argon atmosphere and spectra were collected at

different temperatures. The typical absorptions for the as-prepared polysilane occurred

at 3050 cm'1 (due to the C-H stretching vibration of Si-CH=CH2), 1428 cm-' (due to the

CH2 bending vibration of Si-C6Hs), 1468 cm-' ( due to the CH bending vibration of Si-

CH,1) and 2100 cm-1 (due to stretching vibration of Si-H), and 1247 cm'1 (due to the

rocking vibration of Si-CH3). Figure 2.12 shows the changes in concentrations in residual

pendant organic groups calculated based on the IR spectra, allowing for reduction in

thickness of the film which occurred during pyrolysis. The slight increase in the Si-H

group intensity is attributed to methylene insertion reactions taking place between 200C

and 450C. (This observation was also confirmed by Schilling [Sch84; Sch88] and

Schmidt [Sch91] by NMR studies and resembled reactions occurring during the

conversion of polydimethylsilane to polycarbosilane [Has83].) Carlsson et al. indicated

that no significant changes occurred during the pyrolysis up to 300C but rapid

elimination of -CH3, -C6H, and -C6H13 groups occurred between 300C and 450C. The

increase in the absorption intensity at 1030 cm-1 corresponded to formation of Si-(CH2)n-

Si linkages and network.



















o


Uj


C1


LU

















4000


1000 600


Figure 2.11.


IR spectral changes during pyrolysis of a polysilane polymer (Vllb in Table
2.6) [Car90]; A: initial film; B: 360"C (1.5 h hold); C: 454C (1.5 h); D:
650C (1.5 h); E:1200C (1.5 h); F: dispersion of single crystal SiC
whiskers in KBr for comparison.


3000 2000 1500
WAVENUMBER (cm-t)






46












70



60
"E



E
O
U










20 -
Z 00 30 cm 0
0U 0
CJ
S1800 cm w,





+
0 200 400 600 800 1000 1200
FIRING TEMPERATURE (C)








Figure 2.12. Change in intensities of pendant groups based on IR spectra for polysilane
polymer (Vllb in Table 2.6) [Car90] : o: p(CH3) from Si-CH3; 0: 5 (CH2) of
Si-C6Hs; l: 8 (CH) of Si-C6H13; +: 8 (CH2) of Si-(CH2)n-Si ; A:v (Si-H) ; V: v
(Si-C)








Wood [Woo84] studied the pyrolysis behavior of three different polymethylsilanes

(designated as PMS-I, PMS-II and PMS-III) prepared by Wurtz-coupling reactions of

dichlorosilane with sodium in the presence of hexane, hexane/THF (7:1 volume)

mixture and THF, respectively. Table 2.7 gives a summary of the synthesis conditions

and characteristics for the three polymers. PMS-I showed very low ceramic yield (~25%)

upon pyrolysis to 1000C. The low ceramic yield was attributed to the loss of Si by

volatilization of low molecular weight components (which was confirmed by mass

spectral analysis of the pyrolysis species). X-ray Diffraction (XRD) analysis of the

ceramic residue obtained from pyrolysis of PMS-I showed peaks due to excess Si as

well as SiC peaks. The pyrolyzed ceramic had an overall composition of 67% SiC and

33 wt% Si (calculated based on the Si/C ratio determined by elemental analysis).

Polymer PMS-II showed a ceramic yield of ~27% The XRD analysis of the pyrolyzed

ceramic residue showed no Si peaks (for unknown reasons) although elemental

analysis revealed silicon-rich composition (77 wt% SiC and 23 wt% Si). PMS-III showed

a much higher ceramic yield of 60% compared to the other two PMS polymers and had

an elemental composition of 75 wt% SiC and 25 wt% Si. (XRD analysis showed both Si

and SiC peaks.) The differences in the pyrolysis yields were attributed to differences in

cross-linking in the three polymers. Based on NMR data, both PMS-I and PMS-II

contained higher number of Si-H functionalities (which are potential cross-linking sites)

compared to PMS-III. (This suggested that cross-linking was more extensive in PMS-III

due to consumption of Si-H moieties by condensation reactions. Recall that a polar

solvent such as THF aids in the anionic polymerization of methylchlorosilanes with

sodium and leads to formation of polymers which are rich in Si-H groups; these Si-H

groups undergo condensation causing extensive cross-linking in the polymer).













Table 2.7. Synthesis conditions
Wood [Woo84].


and characteristics for PMS polymers prepared by


Polymer Synthesis Molecular Ceramic Composition
Designation Conditions Weight Yield ,%"
PMS-I Reflux, Hexane 520 25 67 wt% SiC,

33 wt% Si
PMS-Il Reflux, 620-690 27 77 wt% SiC,

Hexane:THFa 23 wt% Si

PMS-ll Reflux, THF ---b 60 75 wt% SiC,

25 wt% Si


" 7:1 Volume proportion
b Insoluble in benzene, hence cryoscopic determination of molecular weight could not be carried out.
c Pyrolyzed to 1000"C in nitrogen atmosphere at 10C/min








Seyferth at al. [Sey93] have significantly enhanced pyrolysis yields of PMS polymers

prepared by Wood's method by dehydrogenatively cross-linking the polymers in the

presence of early transition metal complex catalysts (zirconocene and titanocene) (see

section 2.3). The ceramic yield of PMS polymers increased from ~25% (for

unmodified polymer) to 74% (for cross-linked polymer). The ceramic residue after

pyrolysis for a typical cross-linked polysilane polymer had an elemental composition of

98% SiC, 1.6% ZrC and traces of elemental Si (as opposed to 74% SiC and 26% Si

for an unmodified polymer). Table 2.8 shows pyrolysis results for a number of cross-

linked polymethylsilane polymers prepared under different processing conditions (i.e.,

varying catalyst concentration, solvent, and reflux time). It is evident from the table that

the type of solvent used for the catalytic cross-linking plays an important role in

determining the ceramic yield of the polymer produced. For example, it appears that

hexane and benzene are good solvents for cross-linking in comparison with

polar solvents ether and THF. (This is not to be confused with the effect of polar

solvents on Wurtz-coupling of dichlorosilanes. For example, Wood reported higher

ceramic yield using THF and lower yield using hexane. Seyferth et al.'s results indicated

that non-polar solvents are useful in dehydrogenative cross-linking of low molecular

weight polysilanes prepared by the Wurtz-coupling method.) Furthermore, there

appears to be a level of catalyst concentration above which the pyrolysis yield does not

change significantly, but below which the pyrolysis yield decreases.

Zhang et al. [Zha91; Zha94A; Zha94B] studied the pyrolysis behavior of

polymethylsilane polymers prepared by dehydrocoupling of methylsilane in the presence

of a DMZ (Dimethyl Zirconocene) catalyst. Ceramic yields were ~60% when the polymer

did not contain any processing additives and ~75% when 5-20% processing additives

were added (Figure 2.13). (The chemistry of the processing additives was not specified.)













Table 2.8. Pyrolysis results for catalytically cross-linked polysilane polymers [Sey93]a.

% Catalyst Solvent Reaction Polymer Pyrolysis yield, %
(DMZ) mol% condition appearance

1.0 Hexaneb 30 min/reflux Orange solid 62
0.57 Hexaneb 30 min/reflux Orange solid 74
0.54 Hexaneb 30 min/reflux Orange solid 81
0.14 Hexaneb 30 min/reflux Yellow wax 40

0.56 Benzenec 30 min/reflux Orange solid 72
0.37 Etherd 30 min/reflux Yellow Oil 23
0.38 THFe 30 min/reflux Yellow Oil 31
0.48 Hexane 16 h/ 25" C Yellow solid 67
0.47 Hexaneb 2 h/reflux Orange solid 62

Reflux temperatures: b 68.7"C; c 80.1*C; d 34.9*C; 0 66*C
a The polymethylsilane polymer used for these experiments was PMS-I, prepared by Wood by Wurtz-
coupling reaction of methyldichlorosilane with sodium in hexane. The ceramic yield for this polymer was
-25 wt%.


































Temperature (oC)


Figure 2.13. TGA plots for polymethylsilane polymer prepared by Zhang et al. [Zha94]








The pyrolyzed ceramic residue had an elemental composition of 69 wt% Si/ 31wt% C

(i.e., close to stoichiometric composition). The bulk of the weight loss is shown to occur

between 200C and 600C in a gradual manner. This weight loss behavior is

different from that of Wood's polymethylsilane polymers (prepared by Wurtz coupling

reaction), where weight loss is reported to occur between 130C and 410C. (The

heating rates were comparable, i.e., 10C/min to 10000C.) Zhang et al. also studied the

chemical evolution of the polymethyslilane polymer during pyrolysis to 1100C using

diffuse reflectance infrared Fourier transform spectroscopy (DRIFTS). Figure 2.14

shows DRIFT spectra of polymethylsilane polymer heated to selected temperatures

at 1 C/min in nitrogen and held at temperature for 0.5 h. The 400C spectra shows

the appearance of a strong peak at ~1350 cm-1 attributed to the bending

vibration of Si-CH2-Si group. (This is analogous to the thermal rearrangement taking

place during the conversion of polydimethylsilane to polycarbosilane.) Schmidt et al.

[Sch91] have reported similar observations. At 600C, the polymer is shown to lose well-

defined molecular structure with the only peaks remaining attributed to v (C-H) of CH3

(2896 cm-1), v (Si-H) (-2068 cm-'), and 6 (CH2) of Si-CH2-Si (at 1354 cm-'). These peaks

disappeared at temperatures >800C, and the spectra showed absorptions in the

region of 400 cm-1 to 1000 cm-' corresponding to 3-SiC.

Recall that linear polysilanes prepared by the Wurtz-coupling reaction of dialkyl

or monoalkyl chlorosilanes have ceramic yields of only up to < 25% [Bur49; Qiu89;

Woo84]. Schilling and Kanner [Sch88] reported that when olefinic halosilanes are

used as monomers in the Wurtz-coupling reaction with sodium, the resultant polysilanes

contain olefinic groups which act as backbone branching sites and cause in-situ cross-

linking. This resulted in relatively higher ceramic yields (38-50%) for these polymers.












































1_,
1 400 OC




200 0C



RT




4000 3000 2000 1000
400.0
Wavenumbers


Figure 2.14. DRIFT Spectra of PMS polymer, prepared by Zhang et al. [Zha94].








The olefin groups in olefinic halosilanes do not react with sodium and, therefore, are

retained in large amounts in the polymer.

Schmidt et al. [Sch91] have studied the pyrolysis characteristics of vinylic

polysilane (VPS) (manufactured by Union Carbide Corporation, Tarrytown, NY based on

Schilling and Kanner's patent [Sch88]). The polymer was prepared using Me3SiCI,

Me2SiClI, CH2=CHSiMeCI2 monomers in 0.85 :0.3: 1 proportion under refluxing

conditions in a xylene/THF mixture (7:1 wt ratio). The TGA profile of the VPS polymer is

shown in Figure 2.15. The pyrolysis process can be divided into three distinct regions

based on the TGA of profile: (i) ~ 50-300*C, where thermal cross-linking occurred

without much loss of weight, (ii) ~300-750C, where major weight loss occurred due to

polymer degradation, and (iii) above 750C, where small weight losses were observed.

The ceramic yield was ~58%, which is slightly higher than that reported by Schilling and

Kanner. The DTA showed a strong exotherm at around 2500C, corresponding to cross-

linking reactions, and a weaker exotherm (at about 4500C) during regime of large

weight loss. Schmidt et al. suggested that the exotherm at about ~1100C may be

indicative of partial crystallization of silicon carbide. Elemental analysis of the ceramic

formed after pyrolysis at 1000C showed a composition of 55% Si, 40% C, 2.7% 0, and

less than a percent each of H and N. The presence of excess carbon (-17 wt%) in the

ceramic is not surprising, considering the fact that vinylic groups are retained in the

polymer backbone due to early cross-linking reactions at temperatures less than 300*C.

The transmission IR spectra of as-received VPS and VPS heat-treated at temperatures

of 2500C, 4000C, 650C and 10000C in nitrogen atmosphere are shown in Figure 2.16.

Table 2.9 [Qiu89B; Col64] lists the IR peak assignments for the VPS polymer. The VPS

polymer undergoes following changes upon heating to 250C: (i) decrease in






















105


-5

4

3

-2

o ?
0
2





-1


1500


Temperature (oc)









Figure 2.15. TGA and DTA plots for a VPS polymer heated in N2 at 20C/min to 1200C
[Sch91].








Table 2.9. Peak assignments for IR absorption spectra of vinylic polysilanes [Qiu89B;
Col64]


Peak (cm-')

3048 (m)

2952 (s)

2894 (s)

2078 (s)

1732 (w)

1582 (w)

1397 (s)

1246 (vs)

1000-1100 (s)

937 (m)

750-850


Assignment

v (CH=CH2) of Si-CH=CH2

v,, (C-H) of CH3

v, (C-H) of CH3

v (Si-H)

v (C=O)

v (C=C) of Si-CH=CH2

8 (Si-CH=CH2)

(8 (CH3) from Si-CH3

o (CH2) from Si CH2- Si,v (Si-O-Si)

5 (Si-H)

v., (Si-C)


v = stretching ; 8 = bending ; (o = bending ; p = rocking; vs = very strong; s = strong;
m =medium; w = weak











4000


w
z
O




Z

I- ,


3000 2000 100 1200 800 400
. 1 I I I 1


4000 3000 2000 1800 1200 800 400

WAVENUMBER (cm-')



Figure 2.16. IR spectra of VPS polymer; (a) room temperature (b) 250C (c) 400*C (d)
6500C (e) 10000C [Sch91].








asymmetric stretching of CH2 band (3048 cm-" (ii) decrease in intensity and broadening

of Si-CH=CH2 deformation band (1397 cm-1), (iii) slight decrease in both the intensity

and area, and broadening of the Si-H stretching band (2078 cm'1), and (iv)

appearance of bending vibration at 1640 cm-', due to C-H bending vibration of Si-

CH=CH2. The decrease in intensity and broadening of Si-CH=CH2 band is attributed to

loss of vinyl groups due to cross-linking reactions. The increase in Si-H absorption band

at 4000C could be attributed to methylene insertion reactions, similar to the reactions

occurring in the conversion of polydimethylsilane to polycarbosilane [Yaj78B]. The

absorption band at 1642 cm-' due to the C-H bending vibration of Si-CH=CH2

disappears beyond 250C due to the loss of vinylic groups.

Abu-Eid et al. [Abu92] have also studied the pyrolysis behavior of polysilane

polymers prepared by Wurtz-coupling of monoorganosilanes ( R'R2SiCI where R' =

CH3 and R2=H, C2H, C3H7, C4H9, CH,,, C8H17 or C6H5 ) with sodium in an inert solvent.

Table 2.10 shows information regarding the polymer characteristics and the pyrolysis

behavior. According to Abu-Eid et al., the high ceramic yield for polymethylsilane (

(CH3SiH), ) is due to an early onset of intermediate carbosilane formation and cross-

linking of Si-H functionalities (e.g., reaction with moisture to form Si-OH, and

subsequent condensation to form Si-O-Si networks). The relatively high ceramic yield

for polydimethylsilane (-25%) is not consistent with values (-1%) reported by Wood

[Woo84] and Carlsson et al. [Car90]. For other dialkyl or alkylaryl polysilanes listed in

the table, the low ceramic yields are consistent with previously reported values.


















Table 2.10. Ceramic yield characteristics and decomposition temperatures for polysilane
polymers synthesized by Abu-Eid et al. [Abu92].

Polymer Type Theor. SiC Actual % of Onset of End of
yield,% ceramic Theor. decomp. decomp.
yield,% yield temp, *C temp, C
[(CH3)2Si], 69.0 25.0 36.0 240 610
[(C2Hs)2Si], 46.5 7.0 15.0 170 600
[(CH3SiC3H7)]n 46.5 7.8 17.0 330 650
[(CH3Si-n- 40.0 2.5 6.3 280 500
C4H,)],
[(CH3Si-n- 25.6 11.2 43.8 260 540
C.H17)]n
[(CH3SiH)], 91.0 60.0 66.0 240 750
[(CH3SiC6H5)], 33.3 21.5 64.6 270 520








2.4. Cross-linking of Polysilane Polymers

Polysilane polymers exhibit a wide range of properties based on the pendant

substituent groups in the polymer chain and the degree of cross-linking. The physical

appearance of the polymer could range from that of a viscous liquid (e.g.,

polymethylsilane) to a solid (e.g., polymethylphenylsilane) depending on the molecular

architecture of the polymer (cross-linking, molecular weight, side groups etc.). The

polymers can be cross-linked by oxidation, room temperature vulcanization, and

photolysis.

2.4.1 Oxidative cross-linking

Oxidative cross-linking of polysilane polymers can be accomplished by

converting Si-H groups in the polymer to Si-O-C or Si-O-Si groups by reacting with air or

moisture. These oxidatively cross-linked polymers are insoluble in organic solvents and

do not melt (i.e., infusible) during pyrolysis to silicon carbide. The degree and rate of

cross-linking depends on the amount of Si-H groups present in the polymer. Figure 2.17

shows an FTIR spectrum of a polymethylsilane polymer which clearly demonstrates the

air sensitivity of these polymers as indicated by the broad absorption band around 3450

cm1. (This absorption is due to Si-OH stretching which arises from conversion of Si-H.)

The oxygen sensitivity of the polysilane polymers is attractive for some applications

(e.g., multilayer lithography), but the incorporation of oxygen is sometimes not desirable

if the polymer is used as a SiC precursor.

2.4.2 Room temperature vulcanization

The Si-H groups present in the polysilane polymers have been exploited in the

preparation of highly cross-linked polymers by catalytic dehydrogenation (see section

2.3).







61




















0.6873



0.6145



0.5417



s 0.4689



0.3961



0.3233



0.2505 I I I
4000 3600 3200 2800 2400 2000 1600 1200 800 400

Wavenumber (cmn')










Figure 2.17. FTIR spectrum of polymethylsilane polymer prepared by Abu-Eid et al.
[Abu92]








West et al. [Wes86B] have cross-linked Si-H containing polyphenylsilane polymers by

using a vinylic silane monomer (e.g., trivinylphenylsilane, trivinylmethylsilane) as the

cross-linking agent in the presence of traces of chloroplatinic acid as a catalyst. During

the reaction, the initially viscous reaction mixture transforms to a solid polymer, which is

insoluble and infusible (a process analogous to room temperature vulcanization of

silicone elastomers). The cross-linking reaction can be represented as:


Ph Ph Ph

-Sii _Si

H H H


- n


-Si -

Vi -- Si-Ph
I H2PtCI6 I I
+ Vi-Si-Ph ------ -Si--Si-Ph + C2H
I I
V, Ph -Si-Ph

_Si-
1


(2.10)


Vi CH=CH2

Ph C6Hs





2.4.3 Photo-cross linking

Qiu and Du [Qiu89B] have shown that when polysilane polymers such as

polymethylsilane and polyphenylsilane were irradiated with UV light of wavelength 254








nm in nitrogen or vacuum, cross-linking of the polymer occurred with the formation of

insoluble material. One of the disadvantages of photo-cross linking is that some

degradation of polymer molecular weight (due to photo-scission) always accompanies

photo-cross linking (as shown by equations (2.11), (2.12), and (2.13)). However, West et

al. [Wes86B] observed that when a cross-linking agent containing C=C double bonds

(e.g., tetravinylsilane) was mixed with the polymer and then irradiated with UV light, no

degradation in molecular weight occurred and all of the polymer converted to an

insoluble material. The photo cross-linking takes place by cleavage of polysilane chains

to form radicals and addition of these radicals to C=C double bonds of vinylic silanes

(polyunsaturated additives), causing formation of cross-links and generation of new

carbon radicals. These new carbon radicals sustain further cross-linking reactions

(equation (2.14)). The reactions can be represented as given in Figure 2.18.



2.5. Applications of Polysilane Polymers

There are three main technological applications of polysilane polymers: (i)

precursors for P-SiC, (ii) photoinitiators for radical polymerization reactions, and (iii)

photoresists in microelectronics.

2.5.1 Precursor for B-SiC

Yajima and Hasegawa [Yaj78A; Yaj78B; Has83A; Has83B; Has86; Ich86;

Has89] pioneered research in the preparation of 3-SiC from polysilane-based

preceramic polymers. They first synthesized polydimethylsilane (PDMS) by Wurtz-

coupling of dimethyldichlorosilane with sodium in xylene at 135C and then converted

the insoluble PDMS to a soluble polycarbosilane by pressure pyrolysis at ~450C. The










fR Rf hv
------- -Si -Si------- 2

RI I
R1 R1


R2

-Si

R


I I I
R1 R1R




-------Si -Si -Si-------

SRI I
R1 R1R


hv
---


Si: + 2

R1


R2 RI
hv R2
S--Si: + -------Si
R/ RI
"1 R,


R2

- Si.

RI


Rf
Si -------

Ri


where R1= n-C4H11, n-C6H13, or c-C6H,;
R2= C-C4H11 or CH3


R2

------- Si. +

RI
R1


R2
hv I
t rated v p--- Nei C o a
Polyunsaturated vhyl corrpound I New Carbon radical
RI


Figure 2.18. Scheme for photo cross-linking reactions of polysilane polymers


(2.11)


(2.12)


(2.13)


(2.14)








conversion of PDMS to PCS takes place by Kumada rearrangement [Shi58], in which

insertion of CH2 groups into the main chain Si-Si takes place, leaving a hydrogen bound

to silicon, as shown below:




H3 H3 Argon CH3 H
I I Argon I I
--Si-Si --Si-C-- (2.15)
0I I 450 I I
CH3CH3 H H
n-


PDMS PCS



Yajima et al. melt spun the polycarbosilane polymers (Me,3,000) into fibers and

subsequently converted them to SiC by pyrolysis. The fibers required an air-curing step

(~200*C, 2-4 h) to render them infusible during pyrolysis to silicon carbide. These fibers,

commercially produced as NicalonTM fibers (by Nippon carbon company, Tokyo, Japan),

degrade rapidly at temperatures in excess of 1400C due to carbothermal reduction

reactions between siliceous material and carbon. This leads to evolution of volatile

species (primarily CO and SiO) which results in large weight losses, formation of

porosity, and growth of SiC grains and other strength degrading flaws in the fiber

structure. In recent years, a lot of attention has been directed toward improving the

thermomechanical stability of fibers derived via organosilicon polymer route. A method

developed at University of Florida involves preparation of fibers ("UF Fibers") by dry

spinning of high-molecular-weight polycarbosilanes. SiC fibers were produced with low

oxygen content and either carbon-rich (non-stoichiometric) or near-stoichiometric

composition [Tor92A; Tor92B; Tor94; Sac95A; Sac95B]. Non-stoichiometric UF fibers








have room temperature mechanical properties similar to that of NicalonM fibers, with

average tensile strengths ~3 GPa. In addition, UF fibers showed significantly improved

thermomechanical stability compared to Nicalon, as indicated by lower weight losses,

lower specific surface areas, and improved strength retention after heat treatment to

1700C. Near-stoichiometric UF fibers ("UF-HM Fibers") have high tensile strengths

(2.1-3.4 GPa), fine grain sizes (mostly -0.1-0.2 pm), high bulk densities (-3.1-3.2 g/cm3)

and small residual pore sizes (mostly 0.1 pm). These fibers retained -93% of their

initial strength after heat treatment in argon at 18000C.

It is also possible to convert polysilane polymers to silicon carbide fibers directly

without resorting to the preparation of intermediate polycarbosilane polymers. As

discussed in section 2.1.6, West et al. [Wes81; Wes86A] have prepared

phenylmethylsilane-dimethylsilane copolymers, (Polysilastyrene (PSS)), which afford

improved processability over PDMS. However, the fibers prepared from these polymers

require a cross-linking (curing) step to make them infusible in order to survive pyrolysis.

Since the polymers lack Si-H groups (eliminating the possibility of air-curing), the only

operative cross-linking mechanism is by UV irradiation. Thermomechanical data on the

fibers prepared by this method have not been reported.

Lipowitz et al. [Lip89] have prepared SiC fibers from methylpolysilane (MPS)

polymers, synthesized based on Baney et al.'s redistribution/substitution reactions of

methylchlorodisilanes (as discussed in section 2.2.6). Fibers were melt spun, cross-

linked (cured) and converted to SiC by pyrolysis. By varying the ratio of alkyl to phenyl

Grignard reagents (used to react the intermediate methylchloropolysilane polymer),

fibers with composition ranging from silicon-rich through stoichiometric to carbon-rich

were produced. The method of cross-linking was not specified but the relatively low








oxygen content in the fibers (0.6-6.0 wt%) compared to Nicalon TM suggests that air-

curing step was not used. The low oxygen content in the fibers contributed to improved

thermomechanical stability of these fibers over that of NicalonTM fibers.

More recently, Lipowitz et al. [Lip91A; Lip91B; Lip94A; Lip94B; Lip95] developed

near-stoichiometric, polycrystalline SiC fibers using polycarbosilane and

methylpolydisilylazane polymers. Fibers were melt spun, oxidatively cross-linked, and

heat treated at temperatures above 16000C in argon in order to react excess carbon

and oxygen in the fibers. As noted earlier, PC-derived fibers normally become very

weak and develop a porous, large-grained microstructure during this type of heat

treatment. However, Lipowitz et al. incorporated a boron-based sintering additive in the

polymer which allowed fibers to be densified after the carbothermal reduction reactions

discussed earlier. The resulting fibers had fine diameter (8-10 pm), high relative density,

small average grain sizes (in the range -0.03-0.5 pm, depending on the Si:C ratio), low

oxygen content (<0.1%), high tensile strength (2.6 GPa), high elastic modulus (up to

420 GPa), and good strength retention after high temperature (18000C) heat treatment

in argon. The key limitation in this process was apparently a difficulty in producing

continuous fibers.

Takeda et al. [Tak94] have reported the development of low-oxygen-content (0.4

wt%), fine-diameter (-15 pm) SiC fibers. These fibers ('Hi-Nicalon') were prepared in a

similar manner as Nicalon (i..e., by melt spinning of polycarbosilane) except that cross-

linking was accomplished by electron beam irradiation instead of oxidation. The high

temperature stability of the fibers increased dramatically as the oxygen content of the

fibers decreased. Fibers with 0.5 wt% oxygen retained high strength (-2.4 GPa) and

high modulus (-250 GPa) after heat treatment at 1500C in argon. These fibers had a

chemical composition of 62% Si, 37.5% C, and 0.5 wt% O. The main drawback of this









method is that cross-linking of the polymer by electron beam irradiation is a slower and

expensive processing step. More recently, Takeda et al. [Tak95] produced near-

stoichiometric SiC fibers ('Hi-Nicalon Type S') by a modified Hi-Nicalon process. These

fibers had a chemical composition of 69% Si and 31% C, and exhibited better

thermomechanical than Hi-Nicalon fibers.

Zhang et al. [Zha91; Zha94A; Zha94B] have solution-spun fibers from

polymethylsilane polymers (see section 3) and converted them to SiC fibers by

pyrolysis. Since the precursor polymer was low in molecular weight, it required addition

of a cross-linking agent (unspecified chemistry) to render the fibers infusible. The

additive also acted as a spinning aid for the polymer, in addition to providing extra

carbon to adjust the stoichiometry of the ceramic produced to that of pure SiC. The SiC

fibers produced by Zhang et al. had near-stoichiometric composition. Dense fibers were

produced by adding a boron-based sintering additive. DRIFT spectra of the 1000C

pyrolyzed fibers showed the presence of a small amount of oxygen (exact amount not

determined). This was attributed to contamination during to handling, as PMS polymers

are very sensitive towards air. Thermomechanical stability data on these fibers indicate

that they are superior to commercially available NicalonTM fibers, although no directly

comparable data was reported.

Seyferth et al. [Sey92] have demonstrated the potential for production of near-

stoichiometric SiC fibers from polymethylsilane polymers which are catalytically cross-

linked (see section 2.2.2). However, the fibers need an additional curing step, which can

be brought about by UV irradiation. Information on thermomechanical properties on

these fibers is not available.








2.5.2 Photoinitiators for Radical Polymerization

The ability of polysilane polymers to form silyl radicals on photo-irradiation has

been exploited in the free radical polymerization of styrene, methyl methacrylate etc.

Table 2.11 lists a variety of monomers that can be polymerized by using polysilanes as

photoinitiators. One disadvantage with the use of polysilanes as photoinitiators is that

rate of polymerization is low when compared with conventional photoinitiators such as

benzoin methylether.



Table 2.11. List of polysilanes that can be used in radical polymerization [Wes88].

Polysilanes used Monomers polymerized
(PhC2H4SiMe), Styrene
[(PhC2H4SiMe)o8(Me2Si),.01n Ethyl acrylate
[(PhC2H4SiMe),. (PhMeSi)o.6]n Methyl methacrylate
(PhMeSi)n Isooctyl acrylate
[(PhMeSi)(Me2Si)l] Acrylic acid
[(CyHexSiMe)n] Phenoxyethyl acrylate
[(CyHexSiMe)o.7(Me2Si) ]ln 1,6-hexanediol diacrylate


However, this disadvantage is more than compensated by the fact that the silyl radicals

are insensitive to termination of polymerization by oxygen. This is beneficial because

less rigorous control over atmosphere is needed. Although, speculative in nature, this

oxygen insensitivity is attributed to scavenging of oxygen by secondary species formed

during photolysis [Wol88].

2.5.3 Photoresists in Microelectronics

The current technology in microelectronics is geared towards use of sub-micron

features in the integrated circuit chips, leading to an increase in aspect ratio








(height/width) of features in IC chips [Mil88; Mil90A; Mil90B]. The classical single layer

resist process used in microlithography has been found to be inadequate when dealing

with current trends and lithographers have resorted to multilayer resist processes to

meet the current requirements. Figure 2.19 shows a comparison of single layer

process vs multilayer process [Mil88]. The classical single layer resist process (wet

development) involves exposure of the resist and development of a pattern using a

suitable solvent. There is a loss in line width control for small features since wet

development processes are isotropic. In the case of multilayer photoresist process, the

wafer is covered with a thick, inert planarizing polymer layer followed by a thin layer of

photoresist. High resolution in line width can be achieved since the resist layer can be

thinner than what is acceptable in single layer resist process (0.05- 0.2 pm compared to

-1 pm for single layer resists). Subsequently, the pattern can be wet developed (without

loss in resolution) or dry developed during imaging ablativee exposure) down to the

planarizing layer and the image can be transferred through the planarizing layer by 02-

RIE (oxygen resistant ion etching). A necessary requirement for the multilayer resist

process is that the photoresist remaining after developing must be resistant to Oz-RIE, in

order to mask the underlying polymer effectively. Polysilanes have been found to be

ideally suited for use as photoresists since they are stable to 02-RIE by forming a thin

layer of inert SiO2. In addition, polysilanes possess excellent processing properties such

as good thermal stability, solubility for coatings, and imageability to light and ionizing

radiation.











Wet Development (single layer)


r*-- Resist
-ili Substrate


Coat


- Mask


Expose




Develop




Etch





Strip


Dry Development (multilayer)


,- Resist
/// Planarizing Layer
-- Substrate


-- Mask

Expose


Develop



Dry Etch
02-Plasma


Etch


Strip


Figure 2.19: Comparison of single layer photoresist process vs. multilayer photoresist
process [Mil88].


r- -


--M r'-1 '-













CHAPTER 3
EXPERIMENTAL PROCEDURES


3.1 Role of Polyvinylsilazane as a Spinning Aid for Polycarbosilane

3.1.1 Polymer synthesis

PSZ was synthesized according to the procedures developed by Toreki et al.

[Tor90]. This involved polymerization of a cyclic vinylsilazane (1,3,5-trimethyl-1,3,5-

trivinylcyclotrisilazanet) (see Figure 3.1) in the presence of a radical initiator, dicumyl

peroxide (DCP). The reaction assembly used for PSZ synthesis is shown in Figure 3.2.

In a typical PSZ synthesis, 20 g of cyclic vinylsilazane monomer was mixed with 0.090 g

of DCP and 7.5 g of toluene in a 50 ml double-necked flask. Polymerization was typically

carried out under nitrogen at a temperature of 125C for 18 h.

H
I
N
CH3\ /\ /CH3
CH2=CH Si Si- CH=CH2


H--N N --H

\Si/
CH2=CH CH3



Figure 3.1. Structure of 1,3,5-trimethyl-1,3,5-trivinylcyclotrisilazane


*Petrarch Systems, Bristol, PA.



















WATER OUTLET*--













TOLUENE+ MONOMER+ DCP


TO FUME
HOOD

BUBBLER



, 4--- WATER INLET


BATH


MANTLE


Figure 3.2. Schematic of reaction assembly for PSZ synthesis








The molecular weight of PSZ was controlled by selectively removing low molecular

weight fractions ('oils') by fractional distillation using an oil bath at temperatures ranging

from 125 to 150C.

Polycarbosilane (PCS) was synthesized according to the methods reported by

Toreki et al. [Tor92] and procedures developed at University of Florida. PCS was

synthesized by pressure pyrolysis of polydimethylsilanet in a stainless steel autoclave*

under a nitrogen atmosphere at ~4500C. PCS polymers were prepared and supplied for

this study by coworkers at University of Florida. PCS lots 268, 269, and 262 (combined

in 3:2:1 proportion) were used in this study. The average molecular weight was

~11,000. (The GPC molecular weight distribution for the combined PCS polymer is

discussed in section 4.1.1.) This relatively high molecular weight PCS not only enabled

preparation of highly concentrated solutions (e.g., typically ~70 wt%), but also allowed

fibers to be pyrolyzed without melting.

3.1.2 Spin dope preparation, fiber spinning, and fiber heat treatment

The Influence of polyvinylsilazane (PSZ) as a spinning aid and a cross-linking

agent in the low temperature heat treatment of PCS fibers was investigated in this study.

Two types of polymer solutions were prepared for fiber spinning: one containing 14.5

wt% PSZ and the other without any PSZ. The PCS polymers used in this study were

dried in a vacuum furnace for 12 h to remove any adsorbed moisture, traces of solvent,

etc. prior to solution preparation. The dried PCS polymers were then dissolved in

toluene at 33 wt% solids loading. For use in fiber spin batches, PSZ was dissolved in

toluene at 25 wt% solids loading, filtered through a 0.1 pm filter, and mixed with the PCS

solution.


tNisso Company, Tokyo, Japan
* Model 4651, Parr Instrument Company, Moline, IL.








The polymer solutions were filtered through 0.1 pm filter and concentrated in a

rotary evaporator at -50C until ~25-30 wt% solvent remained. A 'flow test' was

used as a rough indication that an appropriate viscosity for fiber spinning was attained.

The flow test was carried out by tilting the glass vial containing the concentrated

polymer solution at a 450 angle and measuring the time taken for the solution to travel

2.5 cm. (A fixed size of glass vial was used for concentrating the polymer solution and

carrying out the flow test.) Use of the flow test minimized the number of iterations

needed to reach the optimum viscosity for fiber spinning and enabled conservation of

spin dope material (i.e., by not making any theological measurements until just before

the concentrated polymer solution was ready for fiber spinning). The theological

characteristics of the final polymer solution were determined by using a cone-plate

viscometer'. Approximately 0.5 ml of the concentrated polymer solution was used for

the measurement. The measurements were made first by increasing the shear rate

from 1 to 40 s1' and then decreasing the shear rate back to 1 s'1. A toluene-soaked

paper tissue was wrapped around the inside periphery of the cylinder containing the

concentrated polymer solution in order saturate the local atmosphere with toluene and

thereby to minimize evaporation of toluene from the polymer solution during the

measurement. Care was taken to make sure that toluene-soaked paper tissue did not

come in contact with polymer solution or interfere in the measurement in any other way.

Fiber spinning was carried out inside a glove box5 The glove box was purged

with nitrogen three times prior to each spinning experiment. The spin dope was

transferred to a spinneret assembly inside the glove box. Four-hole spinnerets of -70

pm hole sizes were used for fiber spinning. Care was taken to clean the spinnerets


Model HBT, Brookfield Engineering Laboratories,Inc., Stoughton, MA.
Model 50001, Labconco Corporation, Kansas City, MO.








thoroughly before spinning to ensure that there were no particulates blocking the

spinneret holes. The face of spinneret was wiped clean with a toluene-soaked paper

tissue prior to commencement of spinning. Continuous 'green' fibers were formed by

winding on a wheel which was placed approximately 30 cm from the spinneret face.

The spinning conditions (winding speed, nitrogen pressure, and solution viscosity) were

kept constant to enable comparison of fibers produced with and without PSZ. The

solution viscosity was -35-40 Pa-s, the applied gas pressure during spinning was 400

psi, and the speed of the fiber collection wheel was -210 rpm (-440 linear ft/min). The

spinning behavior was documented by noting the number of fiber breaks occurring at

regular intervals of time. After fiber spinning, batches (typically < 0.5-1 g) were cut from

the wheel. These bundles were wrapped in aluminum foil as an ~ 24 cm bundle, then

cut into four ~ 12 cm long bundles, labeled, and stored in a vacuum desiccator for at

least 12 h (to remove some of the residual solvent from fibers) prior to pyrolysis. Some

fibers were pyrolyzed by heating in a tube furnace in nitrogen at 13C/min to 1150C (1

h hold at temperature). The flow rate of nitrogen used for pyrolysis was 30 std. atm

cc/min.

In order to study the effect of oxidative cross-linking on fiber mechanical

properties, several batches of PCS and PCS+PSZ fibers were heat-treated in flowing air

to temperatures of 180 10C in a tube furnace. The flow rate of air was 50 std. atm

cc/min. The heating schedule used was: 1C/min to 500C, 18 min hold at 50*C, 1C/min

to 650C, 18 min hold at 650C, 1C/min to 80C, 12 min hold at 80*C, 1C/min to 95C,

18 min hold at 950C, 1C/min to final temperature (180 100C), 1 h hold at temperature.

PCS and PCS+PSZ (green and air-heat treated) fibers were also heat-treated in flowing

nitrogen (30 std.atm cc/min) in a tube furnace for 1 h at temperatures in the range of








200-1000C. The heat treatment rate was 1lC/min to 1500C and 40C/min from 150C to

the final temperature.

3.1.3 Characterization of PCS polymer solutions

The molecular weight distributions of PCS polymers were determined by Gel

Permeation Chromatography (GPC)' using polystyrene columns and standards and

THF (tetrahydrofuran) as the solvent. PCS solutions for GPC were prepared by mixing

0.5 wt% of polymer in THF and filtering the solution through a 0.1 pm filter. Polymer

solutions were passed through 1000 A and 500 A columns connected in series. The

mobile phase for the columns was THF. GPC for PSZ polymers were analyzed using

10,000 A and 1,000 A columns connected in series. THF could not be used as the

mobile phase for PSZ since the chromatogram showed no clear elution peaks

corresponding to the different molecular weight species in the polymer (i.e., the

chromatogram showed a broad peak and a valley). For this reason, toluene was used

as the mobile phase and the PSZ polymers were dissolved in toluene (0.5 wt%) instead

of THF. The columns were conditioned by purging toluene through them for 24 h prior to

analysis.

The intrinsic viscosities of polymer solutions were determined according to

ASTM D-446 procedure by employing a Ubellohde Viscometer (type OC)1 The

measurements were carried out in a water bath maintained at 300C and the

concentrations of polymer solutions used ranged from 2 to 6 wt%. The efflux time t

required for the solution to pass through the capillary of the viscometer between marked





Waters600E Systems Controller, Waters410 Differential Refractometer and Waters707 Autosampler,
Millipore Corporation, Waters Chromatography Division, Milford, MA.
t Phenomenex Corporation, Torrance, CA.
11 Industrial Research Glasswares Ltd, Union, NJ.








lines was measured. The corresponding efflux time to for the pure solvent (toluene) was

also measured. The specific viscosity lsp was determined according to formula:

lsp = (t-to)/to (3.1)

The intrinsic viscosity [ri] was calculated according Huggins equation:

lsp/c = [1] + k' [1]2 c (3.2)

where c is the concentration of the polymer solution used in the measurement, in g/dl.

[re] was determined by plotting "rs/c vs. c, and extrapolating the straight line to c=0.

Contact angle measurements of polymer solutions were carried out using the

sessile drop method and a Contact Angle Goniometer5. In the sessile drop method, a

liquid droplet is deposited on a solid substrate, as illustrated in Figure 3.3.

Measurements were made of the angle formed by the intersection of a line along the

solid-liquid interface and a line tangent to the droplet surface, both of which pass

through the three-phase (solid-liquid-vapor) intersection point. The substrates used for

measurement were stainless steel, teflon, and stainless steel which were first coated

with PCS or PCS+PSZ. The concentrations of polymer solutions (PCS and PCS+PSZ)

were 33 wt%. The stainless steel (2 cm x 2 cm) and teflon substrates (3 cm x 3 cm)

were cleaned in an ultrasonicator bath followed by rinsing in acetone and drying in an

oven at 700C for 30 min. (The teflon substrate was polished on 1 pm diamond wheel for

20 min to obtain a smooth surface prior to measurement.)

Initially, contact angles of water and toluene were measured on teflon and

stainless steel substrates to enable comparison with reported values in literature. Both

advancing and receding contact angles were measured using the drop-buildup and

drop-withdrawal methods. A Hamilton syringe, calibrated up to 0.5 ml was used for


Rame-Hart, Inc., Mountain Lakes, NJ.
tt Model FS 28, Fisher Scientific, Pittsburgh, PA.




Full Text
xml version 1.0 encoding UTF-8
REPORT xmlns http:www.fcla.edudlsmddaitss xmlns:xsi http:www.w3.org2001XMLSchema-instance xsi:schemaLocation http:www.fcla.edudlsmddaitssdaitssReport.xsd
INGEST IEID EWYZUI8GD_NRGK75 INGEST_TIME 2013-10-24T18:43:45Z PACKAGE AA00017705_00001
AGREEMENT_INFO ACCOUNT UF PROJECT UFDC
FILES