Atomistic modelling of the structure and kinetics of silica-based sol-gel processes

A to m istic M odelling

o f the S tru ctu re a n d K in etics

o f S ilica-based Sol-Gel Processes

ProQuest Number: 10055871

All rights reserved

INFORMATION TO ALL USERS

The quality of this reproduction is dependent upon the quality of the copy submitted. In the unlikely event that the author did not send a complete manuscript and there are missing pages, these will be noted. Also, if material had to be removed,

a note will indicate the deletion.

uest.

ProQuest 10055871

Published by ProQuest LLC(2016). Copyright of the Dissertation is held by the Author. All rights reserved.

This work is protected against unauthorized copying under Title 17, United States Code. Microform Edition © ProQuest LLC.

ProQuest LLC

789 East Eisenhower Parkway P.O. Box 1346

The Royal Institution of Great Britain^

Davy Faraday Laboratory,

London.

University College London,

Department of Geological Sciences,

London.

Instituto Superior Tecnico,

Departamento de E ngenhaiia de M ateriais,

Lisboa.

Thesis submitted for the degree of Ph.D.

University of London.

1/fou mcuf Icuf. 9'm a dneame^

I

U 9'm 4 ^ th e (mli^ one...

^okn Jlen w in

C o n ten ts

A b stra ct 1

A cknow ledgem ents 2

I S o l-G e l P r o c e s s e s 3

1 H isto ry 4

2 Technological im portance 8

2.1 Advantages and d isa d v a n ta g e s... 8

2.2 A p p lic a tio n s ... 10

2.2.1 M o n o lith s... 11

2.2.2 C o a t i n g s ... 14

2.2.3 M e m b ra n e s... 16

2.2.4 P o w d ers... 16

2.2.5 Aerogels ... 17

2.2.6 B io log ical... 17

2.2.7 P r e s e r v a ti o n ... 19

3 B asic principles 20 3.1 G e n e ra l... 20

3.2 P r e c u r s o r s ... 21

3.3 H y d r o ly s is ... 28

3.3.1 Inductive e f f e c t s ... 28

3.3.2 C a t a l y s i s ... 29

3.3.3 S o lu b ility ... 31

3.3.4 W ater/ alkoxide ratio ... 32

3.3.5 Isotopic s tu d ie s ... 33

3.3.6 Chain e f f e c t s ... 33

3.3.7 Transestérification and reesterification... 36

3.4 C o n d e n s a tio n ... 36

3.4.1 S o lu b ility ... 37

3.4.2 E le c tro sta tic s... 38

3.4.3 P o ly m e risa tio n ... 40

3.4.4 C atalyst r o l e ... 44

C O N T E N T S iv

3.4.5 Tem perature e f f e c t s ... 45

3.4.6 Pressure effects... 46

3.4.7 Chemical a d d itiv e s... 47

3.5 K inetic m o d e ls... 47

3.5.1 Branching Theory m o d e l ... 49

3.5.2 Functional Group model ... 50

3.5.3 Bicomponent m o d e l ... 51

3.5.4 Nearest Neighbour m o d e l... 53

3.5.5 Statistical m o d e l... 55

3.6 Growth m o d e l s ... 56

3.6.1 Kinetic m o d e ls ... 56

3.6.2 Monomer-cluster models ... 57

3.6.3 Cluster-cluster m o d e ls ... 58

4 Sol-G el bibliography 61 4.1 General referen ces... 61

4.2 Topical re fe re n c e s ... 62

4.3 C urrent referen ces... 63

II A t o m is t ic M o d e llin g 64 5 H isto ry 65 6 A b -In itio theory 67 6.1 Density Functional t h e o r y ... 67

6.2 F u nction als... 71

6.3 Basis S e t s ... 73

6.4 Atomic charges ... 74

6.5 O p tim isatio n ... 75

6.6 COSMO m o d e l ... 77

7 M olecular M echanics th eory 80 7.1 Molecular Dynamics ... 80

7.2 Long range in te ra c tio n s... 86

7.3 Pressure and tem perature ... 88

7.4 NVT and N PT ensembles ... 89

7.5 P r o p e r t i e s ... 92

8 A tom istic M odelling bibliography 95 8.1 General references... 95

C O N T E N T S V

I I I A b - I n i t i o S t u d i e s 9 7

9 S m a ll c lu s te r s 98

9.1 I n tr o d u c tio n ... 98

9.2 H2O, O H - and H2O - - H 2O 100

9.3 C H i O H and C H3C H2O H ... 106

9.4 S i {0 H) 4 and S i2 0 (0 H)e ...109

9.5 S i{ O C H3 ) 4 and S i{ O C H2C H3 ) 4 ...114

9 .6 A l { O H ) ï , A h O { O H ) l - a n d A l S i O { O H ) ^...117

9.7 C o n c lu s io n s ... 119

10 D if f e r e n t c o n f o r m a tio n s 121 10.1 I n tr o d u c t io n ... 121

10.2 Intram olecular e f f e c ts ...122

10.3 Cyclisation e f f e c ts ...128

10.4 Ring effects ... 134

10.5 C o n c lu s io n s ... 140

11 S ilic a te c lu s te r s 141 11.1 I n tr o d u c tio n ... 141

11.2 Branching effects ... 143

11.3 Branched r in g s ...147

11.4 Double r i n g s ... 152

11.5 Large r i n g s ...157

11.6 Silicate clusters d a t a ... 159

11.7 C o n c lu s io n s ... 164

12 S ilic a te r e a c tio n s 166 12.1 I n tr o d u c tio n ... 166

12.2 Silicate clusters p r e c u r s o r s ... 167

12.3 K inetic m o d e l ... 168

12.4 Silicate clusters e v o lu tio n ...175

12.5 C o n c lu s io n s ...180

13 S o l-G e l c lu s te r s 182 13.1 I n tr o d u c tio n ... 182

13.2 S tr u c tu r e ... 183

13.3 Charge d is tr ib u tio n ...189

13.4 Reaction e n e r g i e s ...190

13.5 C o n c lu s io n s ...194

14 S o lv a te d c lu s te r s 195 14.1 I n tr o d u c tio n ...195

14.2 Solvated w ater and m e th a n o l... 196

14.3 Solvated T M O S ... 197

14.4 Solvated m o n o m e r... 198

C O N T E N T S vi

14.6 E n e r g y ... 200

14.7 C o n c lu s io n s ... 200

15 R ea ctio n m echanism s 203 15.1 I n tr o d u c t io n ... 203

15.2 D a ta on protonated sp ecie s... 204

15.3 Form ation of protonated s p e c ie s ... 206

15.4 Reactions of protonated s p e c i e s ...209

15.5 C ondensation 3 ^ 2 and Lateral Attack m e ch an ism s... 209

15.6 Hydrolysis iS'at2 m ech an ism ... 214

15.7 C o n c lu s io n s ...217

I V M o le c u la r M e c h a n ic s stu d ie s 219 16 C utoff-based liquids 220 16.1 I n tr o d u c t io n ... 220

16.2 D ensity versus partial c h a r g e s ... 221

16.3 Dipoles versus partial c h a r g e s ...223

16.4 Energy versus p artial charges ... 224

16.5 R D F versus partial c h a rg e s... 227

16.6 MSD versus p artial ch arg es...228

16.7 Intram olecular e q u ilib r a tio n ...235

16.8 Forcefield dependence of d e n s ity ... 243

16.9 Density versus c u t o f f ...244

16.10 D ensity versus equilibration t i m e ... 244

16.11 D ensity versus system s i z e ... 245

16.12 Energy fluctuation versus timestep ...246

16.13 E xecution tim e versus c u t o f f ...246

16.14 C o n c lu s io n s ... 248

17 C utoff-based solution s 249 17.1 I n tr o d u c t io n ... 249

17.2 F o rcefield ...250

17.3 M ixture e ffects...251

17.4 Aggregation e f f e c t s ... 252

17.5 D ensity and e n e r g y ... 259

17.6 C o n c lu s io n s ... 261

18 Ewald sum -b ased liquids 264 18.1 I n tr o d u c t io n ... 264

18.2 F orcefield...265

18.3 Ewald s u m ... 267

18.4 D e n s ity ... 270

18.5 P oten tial and kinetic e n e r g y ...271

18.6 E nthalpy of v a p o r is a tio n ... 275

C O N T E N T S vii

18.8 Radial distribution f u n c t i o n ...282

18.9 Mean square d is p la c e m e n t... 290

18.10 Diffusion c o efficien t... 294

18.11 C o n c lu s io n s ... 294

19 Ew ald sum -based solutions 296 19.1 I n tr o d u c tio n ... 296

19.2 Reactive in t e r a c t io n s ... 297

19.3 Monomer and dimer ag g reg atio n ... 301

19.4 Density and e n e r g y ... 304

19.5 Short d a ta collection t i m e s ... 308

19.6 Short ru n t i m e s ... 308

19.7 Solvation e f f e c ts ...308

19.8 C o n c lu s io n s ... 313

V F in a l R e m a rk s 31 4 20 D iscussion 315 20.1 Chem istry of silica in s o l u t i o n ... 315

20.2 Atomistic m o d ellin g...323

List o f Figures

1.1 Lim eplaster ancient s c u l p t u r e ... 5

1.2 Precious O p a l ... 6

2.1 Tem perature d ia g ra m ... 9

2.2 Sol-gel c r a c k i n g ... 10

2.3 Gel-silica m i r r o r ... 12

2.4 U ltraviolet transm ission ... 12

2.5 Infrared tr a n s m is s io n ... 13

2.6 T herm al e x p a n s io n ... 13

2.7 Window c o a t in g ... 14

2.8 Refractive index versus T i0 2 and p o r o s i t y ... 15

2.9 Reflectance d ia g ra m ... 15

2.10 Tricom ponent a e r o g e l s ... 17

2.11 Aorts Si c o n t e n t... 18

3.1 Drying and s i n t é r i s a t i o n ... 21

3.2 S tructure of m etal a lk o x id e s ... 22

3.3 Alkoxide o lig o m e riz a tio n ... 24

3.4 DMDES, M TES and TEGS b e h a v i o u r ... 26

3.5 Organic-inorganic c o p o ly m er... 28

3.6 Inductive e f f e c ts ... 29

3.7 pH rate p r o f ile ... 29

3.8 Hydrolysis of different silanes ... 30

3.9 TEO S-w ater-ethanol phase d ia g r a m ... 31

3.10 Solvent e f f e c t s ... 33

3.11 G elation tim e vs water and alcohol c o n c e n tr a tio n ... 34

3.12 Hydrolysis of TEOS species ... 35

3.13 Silification of w o o d ... 37

3.14 T em perature and pH dependence on s o lu b ility ... 38

3.15 Size dependence on solubility ... 39

3.16 Ionised species in s o l u t i o n ... 39

3.17 Silica p o ly m e r is a tio n ... 40

3.18 Silicate s p e c ie s... 41

3.19 Silicate a n i o n s ... 42

3.20 TMOS and TEO S polymer growth ... 42

3.21 Stober silica p o w d e r s ... 44

L I S T OF FIG U R E S ix

3.22 pH condensation t r e n d s ... 45

3.23 Sol-gel pressure t r e n d s ... 46

3.24 Drying control a d d i t i v e s ... 47

3.25 DCCAs m icrostructural in f lu e n c e ... 48

3.26 Sol-gel s p e c ie s ... 48

3.27 Branching Theory kinetic m o d e l... 50

3.28 Functional Group kinetic m o d e l ... 52

3.29 Bicomponent kinetic m o d e l... 53

3.30 Statistical kinetic m o d e l ... 56

3.31 Fractal growth models ... 57

7.1 MD energy c o n s e r v a tio n ... 83

9.1 H2O, i Î3Û+, O H - and H2O - - H 2O c lu s te r s ... 102

9.2 M ethanol and Ethanol Molecules ...106

9.3 S i{ 0 H )4 and S i 2 0 { 0 H ) e c lu s te r s ...109

9.4 S iiO C H s)^ and S i( O C H2C H z) 4 A lkoxides...116

9.5 A l{ O H )^ , A l 2 0 { 0 H ) l - and AlSiO{OH)Q C lu s te rs ... 117

10.1 D F-BH L/D N P d a ta for S i{0 H )4 co n fo rm a tio n s...124

10.2 D F-B L Y P/TN P d ata for S i{ 0 H )4 c o n fo rm a tio n s ... 125

10.3 D F-BH L/D N P d a ta for Si20{0H )Q c o n fo rm a tio n s ... 126

10.4 D F-B L Y P/TN P d a ta for S i2 0 { 0 H ) e conform ations... 127

10.5 D F-BH L/D N P d ata for S iz 0 2 { 0 H ) z co n fo rm atio n s... 129

10.6 D F-BLY P/D N P d a ta for S iz 0 2 { 0 H ) z c o n fo rm a tio n s... 130

10.7 D ata for S i4 0z{0 H)\Q co n fo rm a tio n s... 132

10.8 D ata for S i z0 4 {0 H ) i2 co n fo rm a tio n s... 133

10.9 D ata for S'23(9 3(0 FZ’)6 c o n f o rm a tio n s ... 135

10.10 D ata for S i4 0 4{0 H )z c o n f o r m a tio n s ... 137

10.11 D ata for S i s O u i O H ) ^ co n fo rm atio n s... 138

11.1 TM OS-based system gas chromatographic analysis ...142

11.2 TEOS-based System ^^Si NMR A n a ly s is ... 143

11.3 D F-BH L/D N P d a ta for Q'^Q^ and g j Q ? ... 144

11.4 D F-BH L/D N P d a ta for Q l Q i Q lQ lQ l and Q \ Q f... 146

11.5 D F-BH L/D N P d a ta for Q2Q1Q1 and Q ^ Q i Q i...148

11.6 D F-BH L/D N P d a ta for Q I Q I Q I Q I Q lQ ^Q f, Q \Q \Q \c and Q \ Q \ Q \ l 150 11.7 D F-BH L/D N P d a ta for Q \Q \, Q IQ IQ IQ \ and Q l Q f Q l Q l...153

11.8 D F-BH L/D N P d a ta for Q j Q l Q lQ iQ je and Q l Q i Q f c... 156

11.9 D F-B H L/D N P d a ta for and Q \ ... 158

12.1 Small clusters kinetic evolution ...176

12.2 Non-cyclic clusters kinetic e v o lu tio n ... 177

12.3 Trimer rings kinetic e v o l u tio n ...178

12.4 Large rings kinetic e v o lu tio n ... 179

L I S T OF F IG U R E S x

13.2 [X Y 1][0 3 1] c lu s te rs ... 184

13.3 [X Y 1][X Y 1] c l u s t e r s ... 185

13.4 [2 1 1][2 1 1] c l u s t e r s ... 188

15.1 Lasaga hydrolysis mechanism ... 204

15.2 Condensation Sn^ m e c h a n is m ... 210

15.3 Lateral Attack m e c h a n ism ...211

15.4 S]\f2 and Lateral A ttack energy e v o l u t i o n ...213

15.5 Lateral Attack energy across the v a lle y ... 214

15.6 Condensation energy ev o lu tio n ... 215

15.7 F irst hydrolysis energy ev o lu tio n ... 217

15.8 Last hydrolysis energy e v o lu tio n ... 218

16.1 Reference RDFs in w a t e r ... 227

16.2 Reference RDFs in m e th a n o l... 227

16.3 Reference RDFs in e t h a n o l ... 228

16.4 Cutoff-based RDFs in w a t e r ...229

16.5 Cutoff-bcLsed RDFs in m e th a n o l... 230

16.6 Cutoff-based RDFs in m ethanol (Cont.) ...231

16.7 Cutoff-based RDFs in e t h a n o l ... 232

16.8 Cutoff-based RDFs in ethanol (C o n t.)... 233

16.9 Cutoff-based RDFs in ethanol (C o n t.) ... 234

16.10 Cutoff-based RDFs in ethanol (C o n t.)... 235

16.11 CutoflF-based RDFs in TMOS ... 236

16.12 Cutoff-based RDFs in TMOS ( C o n t .) ... 237

16.13 Cutoff-based MSDs in w a t e r ...238

16.14 Cutoff-based MSDs in m e th a n o l... 238

16.15 Cutoff-based MSDs in e t h a n o l ... 239

16.16 Cutoff-based MSDs in T M O S ... 239

16.17 Intram olecular energy in w a te r ... 240

16.18 Intram olecular energy in m ethanol ... 241

16.19 Intram olecular energy in e th a n o l... 241

16.20 Intram olecular energy in T M O S ... 242

16.21 Intram olecular energy in T E O S ... 242

16.22 Energy fluctuation versus tim estep ...247

17.1 Initial c o n f ig u r a tio n s ...253

17.2 M ethanolic solution before m i x t u r e ...254

17.3 M ethanolic solution after m i x t u r e ...255

17.4 E thanolic solution before m i x t u r e ...256

17.5 E thanolic solution after m i x t u r e ...257

17.6 M onomer aggregation in te ra c tio n s ...258

17.7 Trim er ring aggregation in te ra c tio n s...259

17.8 Aggregation e f f e c t s ... 260

18.1 Ew ald sum cutoff e f f e c t s ... 267

L I S T OF F IG U R E S xi

18.3 Ewald sum cutoff effects ( c o n t . ) ...269

18.4 Ewald sum pressure and tem perature e f f e c t s ...271

18.5 Ewald sum pressure and tem perature effects ( c o n t . ) ...272

18.6 Ewald sum pressure and tem perature effects ( c o n t . ) ...273

18.7 Ewald sum-bcised RDFs in water ...283

18.8 Ewald sum-based RDFs in heavy w a t e r ... 283

18.9 Ewald sum-based RDFs in m e t h a n o l ... 284

18.10 Ewald sum-based RDFs in ethanol ...285

18.11 Ewald sum-based RDFs in ethanol ( C o n t .) ...286

18.12 Ewald sum-based RDFs in T M O S ... 287

18.13 Ewald sum-based RDFs in T E O S ...288

18.14 Ewald sum-based MSDs in w a t e r ...291

18.15 Ew ald sum-based MSDs in heavy w a te r ... 291

18.16 Ewald sum-based MSDs in m e t h a n o l ... 291

18.17 Ewald sum-based MSDs in e t h a n o l ...292

18.18 Ewald sum-based MSDs in T M O S ...292

18.19 Ewald sum-based MSDs in T E O S ... 292

18.20 W ater phase d ia g r a m ... 293

19.1 Alkoxide reactive i n t e r a c t io n s ... 299

19.2 Monomer reactive in te ra c tio n s ... 300

19.3 Dimer reactive in te ra c tio n s ...302

19.4 Monomer and dimer ag greg ation ... 303

19.5 Aggregation in te r a c tio n s ...305

19.6 Solvation before h y d ro ly s is ...311

19.7 Solvation between hydrolysis and co nd en sation... 312

List o f Tables

2.1 Sol-gel advantages and d is a d v a n ta g e s... 8

2.2 Silica O p t i c s ... 11

2.3 Gel optics a d v a n ta g e s ... 11

3.1 T e tra a lk o x id e s ... 25

3.2 O rgano alk ox id es... 25

3.3 Sol-gel solvents ... 32

3.4 Long chain e f f e c ts ... 34

3.5 Branched chain effects ... 35

3.6 G elation characteristics for S i 0 2... 36

3.7 Transestérification e ffe c ts ... 36

3.8 Role of c a ta ly s ts ... 44

3.9 Sol-gel tem perature t r e n d s ... 46

9.1 H2O stru ctu re and charge d i s t r i b u t i o n ...101

9.2 H2O H2O structure and charge d is tr ib u tio n ...103

9.3 O H ~ stru ctu re and charge d is tr ib u tio n ... 105

9.4 structure and charge d i s t r i b u t i o n ... 105

9.5 W ater auto-ionisation e n e rg y ... 105

9.6 M ethanol structu re and charge d is trib u tio n ... 107

9.7 E th ano l stru ctu re and charge distribution ... 108

9.8 S 'i(0 iJ)4 structu re and charge d is tr ib u tio n ...110

9.9 S i2 0 {0 H )Q structure and charge d i s t r i b u t i o n ... I l l 9.10 S i{O H )^ condensation e n e r g y ... 114

9.11 TE O S and TMOS structure and charge d is trib u tio n ... 115

9.12 A l { O H ) ^ , A l 2 0 { 0 H ) l ~ and AlSiO{OH)Q structure and charge distribution . . 118

11.1 [Silicate clusters bond l e n g t h s ...160

11.2 Silicate clusters bond a n g l e s ...161

11.3 Silicate clusters charge d i s tr i b u tio n ...162

11.4 Silicate clusters condensation energy ...163

12.1 Silicate clusters p r e c u r s o r s ... 169

12.2 Silicate clusters p r e c u r s o r s ... 170

13.1 [X Y 0] clusters structural d a t a ...186

13.2 [X Y 1][0 3 1] clusters structural d a t a ...187

L I S T OF T A B L E S xiii

13.3 [X Y 1][X Y 1] clusters structu ral d a t a ... 187

13.4 [2 1 1][2 1 1] conformations d a t a ...188

13.5 [X Y 0] clusters charge d i s tr i b u tio n ...189

13.6 [X Y 1][0 3 1] clusters charge d is trib u tio n ... 190

13.7 [X Y 1][X Y 1] clusters charge d is tr ib u tio n ...190

13.8 Monomer h y d ro ly sis... 191

13.9 Dimer hydrolysis ... 191

13.10 W ater forming asymm etrical c o n d e n s a tio n ...192

13.11 W ater forming symmetrical c o n d e n sa tio n ... 192

13.12 M ethanol forming asymm etrical c o n d e n s a tio n ... 193

13.13 M ethanol forming symmetrical c o n d e n s a tio n ...193

14.1 Solvated w ater d a t a ... 197

14.2 Solvated m ethanol d a t a ...198

14.3 Solvated TMOS d a t a ...199

14.4 Solvated monomer d a t a ...199

14.5 Solvated dimer d a t a ...201

14.6 Cosmo reaction energies ... 201

15.1 S i{0 H )3 {H 2 0 )+ and S i2 0 H { O H ) ^ d a t a ...205

15.2 S i( O M e h { H O M e )+ , S i ( 0 H ) 3 { H 0 M e ) + a n d S i{ 0 M e )3 { H 2 0 )+ d s L ta ....207

15.3 W ater ionisation environment d e p e n d e n c e ...207

15.4 P rotonated species fo rm atio n ...208

15.5 P rotonated species r e a c tio n s ...209

15.6 Condensation S'//2 attack e v o lu tio n ...211

15.7 Condensation Lateral Attack e v o lu tio n ... 212

15.8 Protonated species r e a c tio n s ...216

16.1 Cutoff-based simple liquids d e n s i t y ...222

16.2 Cutoff-bcised partial c h a r g e s ...222

16.3 Molecular electric dipole m o m e n t ...223

16.4 Cutoff-based MD energy in liquids at high p r e s s u r e ... 224

16.5 Cutoff-based MD energy in liquids at ambient c o n d itio n s ... 225

16.6 Cutoff-based MD energy in liquids at boiling t e m p e r a t u r e ... 225

16.7 Cutoff-based internal e n e r g y ...226

16.8 Cutoff-based diffusion coefficient... 240

16.9 Density versus f o r c e f ie ld ... 243

16.10 Density versus forcefield w ith Mulliken ch arg es... 244

16.11 Density versus c u t o f f ...245

16.12 Density versus equilibration t i m e ...245

16.13 Density versus size ... 246

16.14 Energy fluctuation versus tim estep ...246

16.15 Execution tim e versus c u t o f f ...247

16.16 Execution tim e versus buffer w i d t h ...248

17.1 Cutoff-based p artial c h a r g e s ...251

L I S T OF T A B L E S xiv

17.3 Ethanolic solutions d a t a ... 262

18.1 Ewald sum -based partial charges ... 266

18.2 Ewald sum -based simple liquids d e n s ity ... 274

18.3 Ewald sum -based w ater e n e rg y ...276

18.4 Ewald sum -based m ethanol energy ...277

18.5 Ewald sum -based ethanol e n e rg y ... 278

18.6 Ewald sum -based TM OS e n e r g y ... 279

18.7 Ewald sum -based TEO S e n e r g y ...280

18.8 E nthalpy of vaporisation in liquids ...281

18.9 Specific heat in l i q u i d s ... 282

18.10 Self-diffusion coefficient in l i q u i d s ... 295

19.1 Sol-gel solutions density ...306

19.2 Sol-gel solutions e n e r g y ... 307

19.3 Short collecting tim e d e n s ity ... 309

19.4 Short collecting tim e e n e r g y ... 310

A b stra ct

This thesis concerns the study of the structural and kinetic aspects of silica-based sol-gel processes, which we investigate using ab-initio and molecular mechanics com puter modelling techniques. It is divided into five parts.

T he first is a general introduction to sol-gel processes. It contains four chapters, which describe th e historical evolution of sol-gel processes, their technological im portance, the current scientific background in the chemistry of silica-based processes and a guide to the principal sol-gel bibliography.

T he second p art is a general introduction to atomistic modelling. It is divided in four chapters. T he first is a brief historical overview focusing on the crescent m otivation to apply com puting m ethods. T he second reviews ab-initio theory. The th ird is a review of molecular mechanics theory and the fourth is a guide on the main atom istic modelhng bibliography.

T he th ird p a rt contains all density functional and Hartree-Fock studies, aiming to elucidate the stru ctu ra l growth and reaction mechanisms occurring in sol-gel processes. It contains seven chapters. Small species are analysed in chapter one to investigate the accuracy of the methods. T he role of different silicate conformations is analysed in chapter two. In chapter three, the study of the silicate clusters is completed. All information is gathered in chapter four to create a grow th kinetic model. The influence of the alkoxy groups is studied in chapter five. Solvation effects are investigated in chapter six. Hydrolysis and condensation reaction mechanisms are studied in chapter seven.

The fourth p art contains all molecular dynamics studies, aiming to create a realistic model to sim ulate sol-gel processes in solution. It contains four chapters. In the first, a model is developed to sim ulate liquids in a large range of therm odynam ic conditions. This model is used in chapter two to stu dy m ixture and aggregation effects in diluted sol-gel solutions. The methodology developed in chapter one is enhanced in chapter three. This model is applied in chapter four to study aggregation and solvation effects in more realistic sol-gel solutions.

The fifth p a rt is an overall discussion w ith two sections. The first summarises the implications of this work and considers possible directions for future studies. The second contains general considerations about the present and future of com puter modelling.

A ck n ow led gem en ts

T he au th o r is greatly indebted to In stitu to Superior Tecnico, Dept. Eng. M ateriais, Lisboa, and JN IC T , P rogram a Ciência and program a Praxis XXI, Lisboa, for their financial support.

T he author wishes to acnowledge the computer tim e generously given by The Royal Insti­ tu tio n , London, on Convex 0220, SG Power Challenge Faraday and Rumford and SO Indigo w orkstations, by EPSR C , Edinburgh, on Cray T3D Darwin, by Daresbury Laboratories, Dares- bury, on Intel 86001, by University College London, Dept, of Geological Sciences, on SG Indigo

w orkstation, by Dr. Jo h n Brodholt, UCL, on SG Power Challenge Slamdunk, by ULCC, London,

on Convex Neptune an d Pluto, by In stitu to Superior Tecnico, Civil Eng. Dept., Lisboa, on IBM 6000 w orkstations and by Molecular Simulations Inc., San Diego, on SG Indy workstations. The au th o r is also grateful to INESC, Lisboa. The author wants to acnowledge Dr. Bill Sm ith and Dr. T im Forester at Daresbury, who wrote DLpoly, Dr. Julian Gale at RI, who wrote Gulp and the MSI reenter staff a t San Diego, who support Dmol and Discover. A special acknowledgement to MSI, for a very enrichning working experience at San Diego.

T he first personal acknowledgement is for Prof. Richard Catlow and Prof. David Price, the m ain supervisors of this work. The author wishes to thank, in particular, their inspiring scientific advice, encouragement, patience and eternal good humour. The author would also like to thank Prof. Rui Alm eida for his co-supervisition. A work of this dimension is not possible w ithout the help of m any coleagues. The author is particularly indebted to Prof. Benedito Cabral at Faculdade de Ciencias, Lisboa, Mr. Alexei Sokol, Dr. Alex Lifshivtz, Dr. Ashley George, Dr. A drian Smith, Dr. Robin Grimes, Dr. Dewi Lewis, Dr. David Gay, Dr. R obert Bell, Mr. Furio Cora, Mr. Chris N u ttal, Dr. Simon Carling, Mr. Andy Davids, Dr. Richard Oldroyd, Dr. Yuei Mori, Dr. Josef Breu, and Dr. M artin Exner at RI, London, to Mr. P atricia Carvalho, Dr. Luis Santos, Prof. A m aral Fortes, Prof. F atim a Vaz and Prof. Fernanda M argarido a t 1ST, D ept. Eng. M ateriais, Lisboa, to Dr. Vessal Behnam, Dr. Clive Freeman, Dr. Ton Van Dalen, Dr. Zhengwei Peng and Dr. K it Lau, at MSI California, to Dr. David K itson at MSI UK, to Dr. Steve M arshall at D aresbury Laboratory, to Dr. Julie A ltm ann at ULCC, London, to Prof. M oitinho de Alm eida at 1ST, D ept. Eng. Civil, Lisboa and Antonio Costa a t CIIST, 1ST, Lisboa. A word of g ratitu d e for Dr. Sasha Shluger at RI, for his encouragment.

Part I

C h ap ter 1

H isto ry

The ability to make ceramic objects, in a controlled way, is the first technological achievement of m ankind, evidenced by the earliest fired ceramics discovered to date, the molded figures found in Dolne Vestonice, former Czechoslovakia, and made about 22,000 B.C. [127].

In spite of the evolution in the manufacturing processes, particularly in the Middle East, about 10,000 B.C. (see Figure 1.1), producing limeplaster from limestone, and later during the classic civilisations, ceramics and glasses remained essentially as non noble materials, made essentially to fulfill functional purposes. Only after mastering the technique of producing porce­ lain, in China, during the Tang dynasty (618-907 A. D.) and much later in Europe, in 1708, in Germany, under the direction of J. F. Bdtger [127], ceramic technology received renewed interest and ceramic pieces acquired enormous esthetical and social value.

M any advances have occurred since then in the preparation of powders, forming and m ethods of densification. Powders may have a crystal chemistry controlled to produce specific physical properties. Glass-ceramics may contain a unique combination of m icrostructure and phase assemblages not obtainable by traditional ceramic processes. However, they are limited by the constraints on traditional glciss-maJdng prior to ceraming and the m ethods used to transform the powders in commercial ceramic products are not very different from those used by the Chinese, more th an a thousand years ago. As pointed out by L. L. Hench [95], for millennia, ceramics have been made w ith basically the same technology: powders, either n atural or man-made, have been shaped into objects and subsequently densified at tem peratures close to their tem perature liquidus. The technology of making glass has also remained fundam entally the same since prehistory: particles are melted, homogenised and shaped into objects from the liquid.

The lim itations of such traditional processing are certainly not im portant for many con­ ventional applications, b u t they impose severe constraints on developing sophisticated new ma­ terials for electronic, structural, space or medical applications. M odern ceramics obtained by these m ethods are prim arily the products of applied physics and physical metallurgy, relating physical behaviour and m icrostructure. M ajor advances in ceramic and glass technology need “u ltrastru ctu ral control” , requiring an application of chemical principles unprecedented in the history of ceramics.

C H A P T E R 1. H IS T O R Y

? - J L

Figure 1.1: Limeplaster sculpture, found in Jericho, about 7,000 B.C. [127].

and control the function of cellular membranes and consequently cells [95], thereby reducing the distance between molecular biochemistry and molecular materials chemistry. The emphasis in ultrastructure processing and in sol-gel processes is on controlling the structure and microstruc­ ture of the m aterial at the molecular level, namely its homogeneity and its internal surfaces, to obtain materials with unique combinations of physical and chemical properties.

Silicon constitutes approximately 20 at. % of the earth’s crust in the form of Si02 or the silicates in rocks, minerals and soil constituents. Many natural systems show evidence of silicate hydrolysis and condensation to form polysilicates gels and particles. The precious opal, for example, is formed by amorphous silica particles glued together by a lower-density silicate gel [42]. Repeated hydrolysis and condensation of silica in water, lead to the formation of aqueous polysilicate species th a t evolve, under appropriate chemical conditions, to form spherical particles of essentially anhydrous silica.

As water is a unique liquid, so is amorphous silica a unique solid [105]. Apart from their tech­ nological importance, silica-based sol-gel processes are particularly fascinating for their unique insight into the chemistry of silicon. Manmade synthesis of polysilicate gels from alkoxide precur­ sors closely followed the first preparation of silicon tetrachloride in 1824 [42]. In 1845 Ebelman was able to synthesise S i{O C H2C H z) 4 by reacting SiCU with ethanol, and in the next years he

reported the hydrolysis and subsequent condensation of Si{O C H2C H s)4, forming silicate solu­

tions from which fibres could be drawn and gels that could be used to produce optical monoliths, if dried for long enough (about one year...) [96].

C H A P T E R 1. H ISTO R Y

F ig u re 1.2: P recious O pal, from F ichtelgebirge/R F A , 35x 25 m m [163].

preserve th e d e te rio ra tin g stonew ork of th e Houses of P arliam en t, in London, proposed by von H offm ann, in 1861 [245]), these m ateria ls rem ained of in tere st only for chem ists, for alm ost one h u n d re d years. D uring the w hole n in etee n th century, noted chem ists like O stw ald an d Lord R ayleigh in vestigated th e pro b lem of th e periodic p re cip ita tio n phen o m en a th a t lead to th e fo rm a tio n of Liesegang rings an d th e grow th of crystals from gels [96]. In 1864 G ra h a m showed th a t th e w ater in silica gel could be exchanged for organic solvents, th u s s u p p o rtin g th e th eo ry th a t th e gel consisted of a solid netw ork w ith continuous porosity.

T h e 1930s becom e an im p o rta n t tu rn in g p oint for sol-gel processes, in p a rtic u la r reg ard in g th e ir technological ap p licatio n s a n d re latio n w ith industry. T h e process of su p e rc ritic a l drying, w hich d u rin g th e last 60 years proved to be fun d am en tal in produ cin g aerogels in a reaso n ab ly sh o rt tim e, w ith o u t cracking, was invented by K istler in 1932. In 1939 W . Geffcken recognised th a t oxide films could be p re p a re d from alkoxides, and for th e first tim e, th e co rresp o n d in g in­ d u stria l ap p lic a tio n was developed by a glass com pany. S chroeder developed a th in films physics for th is process by using m any single-oxide and m ixed-oxide layers [66]. T h e c h a ra c te risa tio n of silica gels as form ed by a netw ork s tru c tu re was already generally accepted in th e 1930s, largely due to th e work of G. B. H urd, w ho showed th a t th ey m ust consist of a entire polym eric netw ork form ed by polysilicic acid, enclosing a continuous liquid phase.

Since 1935, m ineralogists have becom e interested in using sol-gel processes to p re p a re pow ­ ders w hich could be used w ith ad v an tag e in stu d ies of solid s ta te equilibria, w here hom ogeneous sam ples are essential [227]. In th e 1960s, R. J. Roy recognised th e p o te n tia l of th e m e th o d for achieving very high levels of chem ical hom ogeneity in colloidal gels and used it to synthesise a large n um ber of new ceram ic oxide com positions, involving m ainly Al, Si, T i an d Zr, th a t could n ot b e m ade using tra d itio n a l ceram ic pow der m ethods [96].

C H A P T E R 1. H I S T O R Y 7

Ludox spheres. The m ethod was extended in 1968 by Stober, creating the so-called Stober spherical silica powder, w ith a uniform distribution of particles sizes (see Figure 3.21). At the same time, im portant b u t classified scientific and technological work was being carried out by the nuclear-fuel industry, to prepare small spheres of radioactive oxides and thus avoiding the generation of the dangerous dust produced by conventional ceramic methods.

T he ceramics industry began to show interest in gels in the sixties. Large scale production of autom otive rear-view windows started in 1959 and continued w ith anti-refiection coatings in 1964 and w ith sunshielding windows in 1969, based on the S iÜ2 — T i0 2 system [66]. In 1969,

L. Levene and I. M. Thomas, and H. Dislich and P. Hinz, working independently, established the chemical basis for the preparation of m ulticomponent glasses, glass ceramics and crystalline substances, through controlled hydrolysis and condensation of alkoxides. Since then, any type of m ulticom ponent oxide can be synthesised by the sol-gel process, using the alkoxides of the different elements [66]. In the beginning of the 1970s, ceramic fibres were already being made from metalorganic precursors by several companies, but only in 1975-77, when B.E. Yoldas and M. Yamane dem onstrated th a t indeed monoliths can be made by carefully drying of gels [42], the enthusiasm for sol-gel processes exploded and im portant financial resources were allocated for research in this area, all over the world.

C h ap ter 2

T echnological im p ortan ce

2.1

A dvantages and disadvantages

The m ain advantages and disadvantages of sol-gel processes, when compared w ith other techno­ logical m ethods, like high tem perature melting, vapour deposition, ceramics from metal-organic polym ers or preparation of ultrafine powders of controllable size [157], are presented in Ta­ ble 2.1 [153].

As the raw m aterials are liquids at working conditions and after mixing they form low viscosity liquid solutions, homogenisation can usually be achieved, at a molecular level, in a short time. Since the reactants are so well mixed in the solutions, they are likely to be equally well-mixed, at the molecular level, when the gel is formed.

Using synthetic chemicals rath er th an minerals, as is usually the case in th e ceramics indus­ try, guarantees high purity and reproducibility of the final products, an im portant point when considering the quality control of the whole industrial process.

As hydrolysis and condensation occur even at ambient tem perature and pressure, th e high tem peratures (above Liquidus tem perature T l) required to melt and mix th e raw materials in conventional melting processes are not needed. As can be seen in Figure 2.1, even the densification, the stage requiring higher tem peratures (close to the glass tran sitio n tem perature

tg), is usually done at a lower tem perature th a t the annealing required in m elting techniques to reduce the strain and defects introduced during the quench of the melt. Low tem perature processing is a key advantage, because it allows a much b etter control of th e whole chemical and physical transform ations.

From the technological point of view, low tem perature saves energy, increases the safety,

Advantages Disadvantages

B e tte r h om ogen eity Expensive raw materials

Better purity Undesirable residuals

Lower temperature of preparation Large shrinkage during processing Better chemical and physical control Long processing times

Better glass products Insufficient scientific understanding

C H A P T E R 2. TECHNOLOGICAL IMPORTANCE

M elt 1 5 0 0

Q u en ch _{M e lt - f o r m e d} G la s s

A n n e a l

, D en sify

Ul

I G e l- 4 D e riv e d ; Amorphous

*

P y r o ly s is

t y i n g G e llin g

T I M E

F ig u re 2.1: C om parison of m elt-d eriv ed and gel-derived processing as a fu n c tio n o f te m p e ra tu re , from M ackenzie [156].

m inim ises air p o llu tio n a n d reduces th e com plexity of th e in d u stria l p la n t an d co n seq u en tly th e financial investm ent needed. F rom th e scientific p oint of view, it te n d s to m inim ise e v a p o ra tio n losses a n d reactio n s w ith co n tain ers, therefore increasing th e re p ro d u cib ility , a n d avoids u n d e ­ sirable p h ase sep a ratio n a n d cry sta llisa tio n processes tak in g place d u rin g th e quench in g an d an n e alin g step s in m elt-form ed glasses. New non-cry stallin e solids can th u s be p ro d u c e d o u tsid e th e range of n o rm al glass fo rm a tio n , from w hich new crystallin e phases can su b se q u e n tly be form ed. O rganic m onom ers th a t decom pose or ev ap o rate a t high te m p e ra tu re s can be com bined w ith in o rg an ic p recursors to c re a te new organic-inorganic m ateria ls w ith u n iq u e p ro p e rtie s [212].

A large range of e x p e rim e n ta l p aram eters, including precursors, solvents, ad d itiv es, te m ­ p e ra tu re , pressu re, tim e, have a direct influence in th e s tru c tu re , m ic ro stru c tu re a n d g eneral p ro p e rtie s of th e final p ro d u c ts, allow ing a enorm ous freedom to design new m ethodologies a n d to co n tro l b e tte r th e final p ro p e rtie s required.

B e tte r glass p ro d u c ts can be m ad e exploring th e special p ro p e rtie s of gel [153], as th e sol-gel process h as th e unique ad v a n ta g e of allowing th e p re p a ra tio n of th e sam e com position, such as silica, in very different physical form s, fibres, coatings, m onoliths, ju s t by varying e x p e rim e n ta l conditions, essentially th o se contro llin g th e viscosity of th e system [96].

pro-C H A P T E R 2. TEpro-CHNOLOGIpro-CAL IMPORTANpro-CE 10

F ig u re 2.2: In itia tin g a crack in a sol-gel m onolith, from H ench et al. [97].

d u cin g w indow glass, for exam ple. In d u stria l ap p licatio n s of sol-gel processes need also to take in to account th e h e a lth h az ard s of o rganic solutions. It is unlikely therefore th a t sol-gel processes w ill be used in th e fu tu re in m a ss-p ro d u ctio n of o rd in ary glass an d ceram ics, w here q u an tity , im p ly in g sh o rt processing tim es a n d cheap raw m aterials, instead of quality, is th e key factor.

U ndesirable resid u al fine pores, hydroxyl an d organic groups, nam ely solvent, are difficult to rem ove a n d m ay in tro d u ce in te rn a l stra in s eventually leading to cracking d u rin g g elatio n an d d rying. In fact, large shrinkage d u rin g processing is p o ten tially a very serious p ro b lem in sol-gel processes, often leading to cracking d u rin g g elation or drying (see F ig u re 2.2). H owever, recent use of critica l p o in t d ry in g for aerogels an d d ry in g control chem ical ad d itiv es for xerogels, has solved som e of th e num erous p roblem s asso c ia te d w ith th e pore evacuation, m ak in g it possible to p ro d u c e large m onolithic d ried gels ro u tin ely an d rap id ly [98]. T h o u g h size scale-up is still one th e g re a te st u n c e rta in tie s of sol-gel processing, control of d rying ra te s, elim in a tio n o f ca stin g defects a n d rigorous atm o sp h ere control can yield large sol-gel o ptical com ponents.

A lth o u g h su b s ta n tia l effort in sol-gel fu n d a m e n ta l research has been carried o u t d u rin g th e la st 15 years, a lack of scientific u n d e rsta n d in g of th e m any com plexities asso ciated w ith th e process still su bsists. W ith o u t th is scientific fo u n d a tio n , ex p lo itatio n of sol-gel processes will be inefficient.

2.2

Applications

A p p licatio n s for sol-gel processing derive d irec tly from th e low te m p e ra tu re , c o m p o sitio n a l a n d m ic ro s tru c tu ra l control, com bined w ith th e various dim ensional sh ap es th a t can be o b ta in e d d irec tly from th e gel state: m on o lith s, films, fibres an d powders.

By th e end o f 1984, cry sta llin e a n d n o n cry sta llin e oxides from over 50 chem ical system s h a d been a lre a d y p re p a re d using th e sol-gel tech n iq u e and it was ex p ected th a t som e 200 oxide sy stem s w ould b e stu d ie d by th e sol-gel m e th o d d u rin g th e next two decades [155].

B o th glass a n d p o ly cry stallin e fibres have been m ade using th e sol-gel m eth o d . C o m p o sitio n s include T 2O 2 a n d Z r0 2~ S i0 2 glass fibres, high p u rity S i0 2 waveguide fibres a n d A /2O 3, Z r 0 2 ,

C H A P T E R 2. T EC H N O LO G IC AL IM P O R T A N C E 11

E xcellen t optical tra n sm issio n

fro m the ultraviolet (160n m ) to near infrared (3600n m ) E xcellen t refractive index hom ogeneity

Iso tro p ic optical pro p erties S m a ll strain birefringence

Very low coefficient o f th erm al expansion (0.55 x 1Q“ ® /°C )

Very high therm al sta b ility Very high chem ical durability

S m a ll num ber o f bubbles o r inclusions A b ility to be polished to high standards

Table 2.2: Silica optics properties, reviewed by L. L. Hench [97].

Im proved casting Com plex geom etries L ight w eight optics A sph eric optics Surface replication In tern a l structure Reduced grin din g Reduced polishing

Im proved p ro p erties L ow er coefficient o f th erm al expansion L ow er ultraviolet cu toff w avelength H igher optical tra n sm ission

N o absorption due to OH band L ow er solarisation Higher h om ogen eity F ewer defects

Im proved im pregn ation Im pregnation w ith organic p o lym ers Graded refractive index lenses Laser-enhanced densification

C ontrolled chem ical doping C ontrol o f variable oxidation sta te s o f dopants

Table 2.3: Potential advantages of gel-silica optics over fused quartz and fused silica components, from Hench et al. [97].

2 .2 .1 M o n o lith s

Silica gels containing wide ranges of alkalis, alkaline earths, transition m etals and rare earth elements can now be made using drying additives, allowing us to prepare m aterials w ith a wide range of optical absorption spectra, index of refraction and dispersion [98]. As reviewed by Hench et al. [97], silica optics are preferred for many optical systems, such as intracavity laser optics, because of the characteristics listed in Table 2.2.

The conventional m ethods to produce silica glass, fused quartz (naturally occurring quartz crystals are melted at high tem peratures) and fused silica (pure silicon tetrachloride is fused at high tem peratures), tend to let substantial amounts either of cation and hydroxyl im purities or w ater and Cl ion contents in the final glass. Furthermore, the nature of the processes involved makes th e direct m anufacture of near net shape optics impossible [97]. Sol-gel processes olfer th e p oten tial for improving many features of silica optics, as listed in Table 2.3.

C H A P T E R 2. TECHNOLOGICAL IMPORTANCE 12

\

F ig u re 2.3: L ightw eight sol-gel silica m irro r w ith in teg ral faceplate, as shown by H ench et al. [97].

100

.2

E

40

-T y p o III T y p o III 2 0 ■

TypoV

155 160 165 170 175 180 185 190 195 200

W a v e l e n g t h (n m )

CH A P T E R 2. TECHNOLOGICAL IMPORTANCE 13

100

8 0

-6 0

-c

o

in

E in c

2

t-40

-2 0

-T ype III Silica Type V S ilica (Gelsil™)

1000 1500 2000 2500 3000 3500 4000 4500

W a v e le n g th (nm)

F ig u re 2.5: N ear in frared tran sm issio n of gel silica, sta n d a rd silica, and two fused silica glasses, from H ench et al. [97].

370

-0,60

-U _3.50

-ul

O

N B S S ilic a R e fe re n c e T y p e III S ilica 0.30

-T y p e IV S ilic a T y p e V S ilic a (Gelsil™ )

020

200

0 100 3 0 0 4 0 0 5 0 0 600 7 00 8 0 0

T e m p e r a t u r e (°C)

C H A P T E R 2. TECHNOLOGICAL IMPORTANCE 14

F ig u re 2.7; Office b uildings w ith sol-gel coated glass w indow s, G erm any [42]. 2.2.2 C o a tin g s

O p tical coatin g s a lte r th e reflectance, tran sm issio n an d a b so rp tio n of th e su b stra te . A rchitec­ tu ra l glass, coated w ith a gel contain in g T i0 2, to control th e reflectance, an d P d , to control a b so rp tio n [42], is cu rre n tly applied to office buildings, as show n in F igure 2.7, w hich a p p e a r o u tw ard ly uniform ly reflective, w hile light tran sm issio n is controlled in accordance w ith sun exposure, to m inim ise su m m er cooling costs (different sides of th e bu ild in g m ight even have co atings w ith different tran sm issio n , due to different su n exposure). T h e S i 0 2 — T i 0 2 b in a ry sy stem is p a rtic u la rly su ita b le because th e refractive index for film s consolidated a t 500°C in­ creases co n tin u o u sly w ith T i 0 2 content from 1.4 to 2.2, as can b e seen in F igure 2.8. D ifferent a b so rp tio n s an d a v ariety of colored coatings can be p ro d u c ed in c o rp o ra tin g different tra n s itio n m etals.

O xide co atin g s on glass a n d silicon su b stra te s have also been used extensively as antireflective surfaces, for exam ple in solar cells, laser optics a n d p h o to g rap h ic lenses, using m u ltip le coating tech n iq u es to decrease tran sm issio n in a wide sp e c tra l region, as show n in F ig u re 2.9. T h e reflectance of a surface can be alm ost com pletely elim in ated w hen th e refractive index of th e surface varies sm o o th ly from th e value of air to th e value of th e s u b s tra te [42]. T h e refractive index decreases linearly w ith porosity, as shown in F igure 2.8, a n d consequently careful control of th e m ic ro stru c tu re is necessary to o b ta in m ateria ls w ith th e p ro p e rtie s required.

Sol-gel film s can p ro te c t ag a in st corrosion or abrasio n , increase s tre n g th or provide p lan ari- s a t ion (d ecreasing diffuse a n d increasing sp ecu lar reflectance). In silicon-based m icroelectronics ap p lic a tio n s, sol-gel films are a lo w -tem p e ratu re a lte rn a tiv e to th e rm a l S i0 2 an d com p ete w ith C VD silica [42]. T h re e m a jo r draw backs of sol-gel films, from th e sta n d p o in t of a b rasio n an d corrosion re s is ta n t layers are: 1) T hick coatings, larger th a n 1 /r, are difficult to achieve w ith ­ o u t cracking; 2) Sol-gel films are in general q u ite b rittle ; 3) R elatively high te m p e ra tu re s are re q u ired to achieve good p ro p e rtie s.

C H A P T E R 2. T E C H N O LO G IC A L IM P O R T A N C E 15

UJ Û z lU >

< £

UJ

oc

2.2 _{1 .5 0}

1 .4 5

1 .4 0 2.0

X 1 .3 5 _{REFRACTIVE INDEX}

v s POROSITY Z 1 3 0

1.8

1 .2 5

1.6 1 .1 5

CC 1 .1 0

1 .0 5

1.4

0 20 40 60 SO

1.00

Mol %

VOLUME FRACTION POROSITY

Figure 2.8: Refractive index as a function of: a) x mol% for gel-derived films in the binaxy system x T i0 2 — (100 — x ) S i0 2 , densified at 450°C (circles) and 500°C (diamonds); b) volume

fraction porosity [42].

1.0

< 0.8

lJ 0.7

—I U.

LU 0.8

OC

_ l 0.5

<

2

oc

IÜ

I 0.3

OL — 0.2

I 0.1

U N CO A TED

AR CO ATED

0.0

300 400 500 600 700 800 900 1000 1100 1200 1300

W A V ELEN G TH (n m )

C H A P T E R 2. T E C H N O L O G IC A L IM P O R T A N C E 16

coatings on acrylic and polycarbonate windows and lenses [42].

O ther applications include electronic thin films, namely high tem perature superconductor films and titan ia films used as photoanodes [42]. Ferroelectric B a T iO z a-nd several other ti- tanates and zirconates, transition m etal oxide gels, namely T i0 2^ S n0 2, W O z and F2O5, w ith

a very high electrical conductivity, can also be currently fabricated [154].

Porous films explore specific features of gels, namely pore volume, surface area and surface reactivity, to achieve specific goals, namely in sensors and catalytic surfaces. As discussed by Brinker et al. [42], zeolite particles w ith pores between 5 and 10 Â, like ZSM-5, have been em bedded in silicate sol-gel matrices w ith pore size smaller th an 4Â, so all molecules larger th a n nitrogen can be adsorbed only w ithin the zeolite channels. A ZSM-5 zeolite/silicate com posite (ZSM-5 pore size = 6 Â) can easily separate propanol (kinetic size = 4.7 Â) from iso-octane (kinetic size = 6.2 Â).

2 .2 .3 M em b ra n es

High surface areas and small pore sizes characteristic of inorganic gels are properties unattain ab le by conventional ceramic processing methods. According to Brinker et al. [42], sol-gel mem branes also offer several advantages when compared w ith conventional organic polymer membranes: 1) They can be operated and sterilised at high tem peratures; 2) They do not swell or shrink in con­ tact w ith liquids; 3) They are much more abrasion-resistant. These properties can be exploited in applications such as filtration, separation, catalysis and chromatography. U ltrafiltration and reverse osmosis, m icrofiltration of water, wine, beverages and ultrafiltration of milk have also been discussed as possible applications [42].

2 .2 .4 P o w d er s

The fabrication of radioactive solid spheres is historically one of the most im portant and im­ pressive sol-gel applications. The technique can be processed easily by remote control and, due to liquid processing, it eliminates environm ental hazards associated w ith radioactive dusts, a m ajor problem in the nuclear industry. The process involves dropping spray-formed sol spheres into a column of a heated inert liquid. Internal gelation occurs during passage of the spheres down the column, after which they are collected, washed and fired at the bottom of the columns.

UO2 spheres w ith diam eters ranging from 30 // to 1200 /i and density in excess of 99% have

been routinely fabricated in this continuous process. UO2 — P u0 2 and C7e0 2 spheres have been

similarly fabricated and the process can be applicable to many other oxides [154]. This internal gelation m ethod has also been applied to prepare Z r0 2-based and other complex oxide spheres, containing high-level radioactive wastes, which are extremely leach resistant [154]. An alterna­ tive conventional sol-gel m ethod, using S i{0 E t) 4 and Al{O Et)z, was also able to produce highly

nonleachable pellets, after sintering at only 800°C [154]. The combination of low volatility of radionuclides, low processing tem peratures and low leach rates make the gel-derived glass pellets a potentially attractive high level waste form for long-term storage.

C H A P T E R 2. TECHNOLOGICAL IMPORTANCE 17

Figure 2.10: Monolithic aerogels in the S i0 2 — P2O5 binary system, from Woignier et al. [247].

2 .2 .5 A e r o g e ls

Since they were first produced, in the early 1930s, aerogels have been the subject of increas­ ingly im portant research, due to their unique properties; and nowadays specific conferences are dedicated exclusively to them. 5f0 2-based transparent tiles as large as 20 x 20 x 3 cm^ [76], and monoliths of complex systems such as S i0 2 - B2O3 - P2O5 [247], shown in Figure 2.10,

were produced more than ten years ago. As reviewed by Fricke [76, 75], due to their extremely high porosities (between 85% and 98%, in the porous range 1-lOOnm), aerogels have extremely low densities, between 0.08 and 0.3 p/cm^, very small index of refraction (between 1.015 and 1.06), extremely small Young’s modulus (E ~ 10® — lO^A/m^), sound velocities (100-300 m /s, compared with 5 x 10^ m /s in silica glass) and thermal conductivities (0.02 W /m K at ambient air pressure and 0.01 if evacuated, the lowest values ever found in powders and solid bodies for the same pressure and tem perature [76]).

Aerogels have been used in Cerenkov counters for detection of relativistic particles, at CERN/Geneva and DESY/Hamburg, as the small index of refraction allows the momentum of relativistic particles to be determined in a range which is not covered by compressed gases or by liquids [166]. Because aerogels have high solar transmission and thermal resistance, layers of granulated silica-aerogel between protective glasses have been used to insulate houses, as the radiation passes through b ut the heat subsequently produced is retained. Small high velocity particles in space can be captured by low density aerogels and gradually decelerated in their porous structure [75], this way acting as a soft wall and reducing considerably the potential damage of the impact. O ther possible applications for aerogels include gas filters in 20-100 nm region and substrates for catalytic materials.

2 .2 .6 B io lo g ic a l

C H A P T E R 2. TECHNOLOGICAL IMPORTANCE 18 100 80 60 40 2 0

E F F E C T S O F AGING ON Si IN T IS S U E S

^---RABBIT .

. a WITH . S C L E R O T IC ' DAM AGE

\ \

RAT SERUM PLA SM A

HUMAN A ORTA (NORM AL)

C E R E B R A L P L A Q U E S IN ALZHEIM ER S

D IS E A S E

HUMAN PLA SM A (BOUND SI)

(UNBOUND SI)

1 0' 1 0' 1 0" 1 1 0' d a y s

50

y r s .

Figure 2.11: Time evolution of the Si content of rodent and human aorts, from Hench et al. [98].

fibre in tendon, a reinforcement for an elastic sheet in skin, and a mineral-reinforced matrix in bone. As Calvert wrote [47], biology seems to have displayed little versatility in the chemistry of structural materials, but shows great subtlety in processing to obtain different microstructures. Manufactured materials tend to follow the opposite trend. They have many different chemical compositions but most are used pure or with a simple reinforcement. Im portant research is currently being carried on to study how natural materials, like bone and mollusc shell, are actu­ ally made, in order to use this knowledge to produce new microstructures using the structural control potentially offered by sol-gel processes.

To understand the mechanisms of the low temperature organic-inorganic processes associated with silica sol and gel formation, aging and drying, might be potentially vital for understanding several biological and biomedical mechanisms, as im portant as the mineralisation of bone and dentim, plant metabolism and genetic manipulation, arteriosclerosis, tissue bonding of bioactive implants, and mechanisms of precyte evolution, as described by Hench et al. [98]. Figure 2.11 shows the logarithmic decrease in human and rodent aortic Si content with time. This drastic reduction of Si for individuals 10-20 years old, is attributed to the loss of Si bound to a mu­ copolysaccharide in the wall of the aorta. This loss appears to degrade the integrity of the wall, reduces its elasticity, increases lipid transport across the wall and increase the propensity for forming atherosclerotic plaques, hardening the arteries and in last instance conducting to heart attacks. Human aortas damaged by sclerotic lesions have been reported to have 1/3 less bound silica. The association of a hydroxylated polymerised silica gel surface with bonding of amino acids, collagen, mucopolysaccharides, soft and hard tissues has been studied extensively using special compositions of the glass system N a2 0 — CaO - P2O0 — S i0 2, which indeed develops

C H A P T E R 2. T E C H N O LO G IC A L IM P O R T A N C E 19

2 .2 .7 P r e s e r v a t i o n

Since the nineteenth-century, attem pts have been made to preserve stone and sculpture w ith alkoxysilanes [245] and in the last 20 years the subject has raised considerable research interest, w ith several books and congresses dedicated to this m atter. Although considerable advances have been m ade in the use of gels derived from alkoxysilanes els consolidants for stone, the results

C hapter 3

B asic principles

3.1

G eneral

Sol-gel processing is a general designation for a broad range of experim ental m ethods aiming to produce high quality amorphous and crystalline solids by preparation of colloidal solutions and subsequent transform ation in gels and homogeneous solid m aterials, as a result of several, carefully controlled, chemical and physical transformations. According to Hench et al. [96], the global process can be divided in several steps: hydrolysis and condensation, gelation, aging,

drying, stabilisation and densification.

Hydrolysis and condensation reactions of carefully chosen chemical precursors, in liquid so­ lution, lead to the formation of a colloid. A colloid is a two-phase suspension where the particles of the dispersed phase are so small, 1-1000 nm [42], th a t gravitational forces are negligible and interactions are dom inated by van der Walls and Coulombic forces. Due to their small mass and inertia, these particles exhibit Brownian motion, moving along random trajectories governed by m om entum transference in collisions. A sol is a colloidal suspension of solid particles in a liquid, an emulsion a colloidal suspension of liquid droplets in a liquid and an aerosol a colloidal suspension of particles in a gas: a fog if the particles are liquid and a smoke if they are solid. A sol formed by suspensions of branched macromolecules is usually called a polymeric sol, whereas a sol formed by dense particles is called a particulate sol.

Further condensation of the sol aggregates leads to the formation of larger molecules, until a single molecular system extends throughout the liquid solution, completing the gelation process. The gelation point, when the sol becomes a gel, can therefore be defined microscopically as the degree of reaction at which this single giant molecule is formed, whereas macroscopically it can be defined as the instant at which the system can support a stress elastically.

Several theories have been proposed to describe th e gelation process, namely the classical theory, developed by Flory [74], the percolation theory, reviewed by Zallen [250], where the

percolation threshold is the model equivalent of the gel point, and more recently, the fractal theory, introduced by M andelbrot [158]. W hen the chains of a polymeric sol grow random ly or the particles of a particulate sol aggregate, fractal structures may be formed.

The gel structure is best described as a two-phase system formed by the juxtaposition of continuous solid and liquid phases (in the sense th a t both phases extend continuously throughout the whole system) of colloidal dimensions (in the sense th a t everywhere in the system the distance between th e two phases never exceeds lOOOnm).

C H APT ER 3. B A SIC PRINCIPLES 21

Figure 3.1: Large dried gel plate and corresponding sintered glass [42].

The polycondensation process does not stop at the gelation point and it continues for hours or weeks, during the aging period. The cast gel remains completely immersed in liquid, this way allowing repeated solution and reprecipitation of the structure, to decrease the number of defects and the porosity and to increase the thickness of interparticle necks[96j. A relaxed aged gel must develop sufficient strength to resist cracking during drying.

During drying the liquid is removed from the interconnected pore network. Large capillary stresses can develop during drying when the pores are small, typically < 20 nm [96]. When a gel monolith is dried under hypercritical conditions, the network does not collapse and a low density aerogel is produced. Aerogels can have pore volumes as large as 98% and densities as low as 0.08 g/cm^. When the gel is dried at ambient pressure, shrinkage occurs, the volume is often reduced by a factor of 5 to 10 and the monolith is called a xerogel. To prevent cracking,

drying chemical controlling additives, DCCAs, are added. If the pore liquid is primarily alcohol based, the monolith is often called an alcogel Although sol-gel processes are usually carried out using a common solvent, ultrasonic irradiation can be used to promote hydrolysis and condensation [252], producing special gels called sonogels.

The subsequent removal of surface reactive groups from the pore network, dehydration or chemical stabilisation, results in a chemically stable ultraporous solid.

Finally the porous gel is sintered at high temperatures, to force its densification and the elimination of the pores. The densification temperature is about lOOO^C, though it depends considerably on the dimensions of the pore network, the connectivity of the pores and its surface area [96]. The glasses thus obtained (see Figure 3.1) ultimately become equivalent to the glasses obtained by conventional melting methods [153, 242, 175] (see also [87, 234], but with much higher degrees of purity and homogeneity.

3.2

Precursors

wa-C H A P T E R 3. B A S Iwa-C P R IN wa-C IP L E S 22

/? 7Ï

/° \

--.0 O ' ' 0 0-M M' 'M -0 0 -B e Al

(i) (iii) (iv)

[TUOMe)]jNa(OBu<)]^ [Cu(OBu^)]^ [B eC O R y ^ ROBe(>J-OR)2Al(OR) 2

X > S

k # :

o' (VI) (V«>

C u f o - O P r O j A l O P r i ] [ T K O E D J ^ [ N b ( O M . ^ ;

["=Ti,v,w]

Figure 3.2: Structure of some m etal alkoxides, (i) cubanes; (2) planar; (3) bridged trigonal and tetrahedral units; (5)-(viii) bridged tetra- and octa-hedral units, cls presented by R. C. M ehrotra [165]. The alkyl groups were omited for simplicity.

ter to produce the hydrolysed species required to continue the aggregation process. Furtherm ore, they can be easily purified by volatization or crystallisation.

Although metal alkoxides have been known for more than a century, it was not until 1978 th a t a monograph was published on this topic. In fact, the chemistry of alkoxides of alm ost all the elements in the periodic table has been elucidated only during the past few decades, due to the efforts of D. C. Bradley [39], R. C. M ehrotra [164, 165, 166] and others, who created a complete new family of chemical compounds, some of which are summarised in Figure 3.2.

A lthough several different routes have ben proposed to synthesize m etal oxides, the com­ mercial products, widely used in sol-gel research, are essentially m anufactured by: 1) reactions of m etals with alcohols, used to prepare alkoxides of alkali metals, alkaline earths and trivalent m etals [39]:

M + n R O H M (O R )n + n /2H2 t

2) reaction of metal halides w ith alcohols, used to prepare trivalent and higher valent metals:

MClfi + n R O H — ^ M {O R )n -I- n H C l ^ .

Even for those m etal halides reacting only partially w ith alcohols, the reaction can be pushed to completion, using a base like ammonia or alkali alkoxide [164]; 3) reaction of m etal halides w ith sodium alkoxides:

M C ln + n N a O R —y M {O R )n + n N a C l

3) alcohol interchange:

M {O R )n + n R 'O H — > M (O R ')n + n R O H