synthesis, purification and micronisation of copper

SYNTHESIS, PURIFICATION AND MICRONISATION

OF

COPPER INDOMETHACIN

USING

DENSE GAS TECHNOLOGY

by

Barry Warwick, M.Sc. (Chem)

A Thesis Submitted to The School of Chemical Engineering and Industrial Chemistry

in Partial Fulfillment of the Requirements for the Degree of Doctor of Philosophy

THE UNIVERSITY OF NEW SOUTH WALES

February 2001

ABSTRACT

II

The primary aim of this work was to provide an alternative method of synthesis of the

non-steroidal anti-inflammatory drug copper indomethacin (Cu-Indo) and to produce

alternative forms of the drug to increase its marketability. Dense gases as anti-solvents

were used to achieve these aims. The study involved the synthesis, purification,

micronisation and co-precipitation of Cu-Indo with polyvinylpyrrolidone (PVP) using

dense carbon dioxide as an anti-solvent.

Initially the volumetric and solubility behaviours of the solvent—anti-solvent systems

were investigated to determine the optimum processing conditions. The solubility of

Cu-Indo in an expanded solution was found to be a complex function of the solvent and

other solutes.

Copper indomethacin was successfully synthesised and purified in a single vessel using

dense carbon dioxide as an anti-solvent. Drug yields of 98 % and purities near 100 %

were achieved at optimum conditions with the advantages of less residual solvent in the

drug, less solvent waste, reduced processing time and increased yields over the

conventional synthesis process.

Copper indomethacin was produced in a variety of morphologies and particle sizes

using dense carbon dioxide as an anti-solvent. An investigation of the effect of process

parameters on the particle characteristics showed that solute concentration was the

dominant variable. Spherical particles with diameters less than 8 µm were obtained at

optimum conditions. The immediate benefit of micronising Cu-Indo was demonstrated

with an eight fold increase in dissolution rate when compared to the conventionally

produced drug.

Polyvinylpyrrolidone was successfully co-precipitated with Cu-Indo using dense carbon

dioxide as an anti-solvent. The PVP—Cu-Indo co-precipitates were found to increase

the solubility of the drug in ethanol with a 36 fold solubility enhancement at optimum

conditions.

ABSTRACT

III

The use of dense carbon dioxide as anti-solvent in this work demonstrates the potential

of the GAS and ASES processes in the pharmaceutical industry. Copper indomethacin

was synthesised, purified and micronised in a single vessel at a substantial saving in

terms of time and solvent usage. The micronisation of Cu-Indo and the formation of the

PVP—Cu-Indo co-precipitate provided alternative forms of the drug substantially

increasing its marketability.

IV

“The fear of the Lord is the beginning of knowledge,

but fools despise wisdom and discipline.”

Proverbs 1 vs. 7

To Lucille, Sarah and Anna

ACKNOWLEDGEMENTS

V

First and foremost I would like to thank Prof. Neil R. Foster for his guidance,

encouragement and support throughout the course of this work. I am especially grateful

for his ability to see opportunity where others see failure. It has been a most valuable

experience working under his supervision.

Many thanks to Dr. Fariba Dehghani for her advice and encouragement at every stage of

my studies. Her patience and selfless efforts are appreciated. I would especially like to

thank her for the many hours she has spent editing this thesis.

I would like to thank Dr. Ray Biffin and Mr. Hubertus Regtop for making this project

possible. Their generous donation of time and equipment is much appreciated.

I extend my gratitude to my colleagues. To Dr. Emma Coen, Dr. Linda Sze Tu, Dr.

Russell Thiering, Keivan Bezanehtak, Rana Bustami, Kiang Charoenchaitrakool, Gary

Combes and Raffaella Mammucari for stimulating conversation and much comic relief.

I would like to thank Louise Stanton for her efforts in the co-precipitation experiments.

Thank you to Mr. Bong van Dang for all your administrative duties. Your efforts are

appreciated.

Thank you to Dr. Jane E. Weder for providing the background information on Cu-Indo

and for the FTIR and EPR analyses. Your expertise has been most valuable.

Thank you to my mother-in-law Sylvia McNicoll for proof reading I appreciate it very

much.

Lastly I would like to thank my wife, Lucille, for her selfless giving, love,

encouragement and support. None of this work would have been possible without her.

TABLE OF CONTENTS

VI

ABSTRACT................................................................................................................. II

CHAPTER 1 1. BACKGROUND ....................................................................................................1

1.1. Introduction............................................................................................................1

1.2. Copper Indomethacin.............................................................................................4

1.3. References............................................................................................................13

CHAPTER 2

2. DENSE GASES AS ANTI-SOLVENTS — THEORY ...........................17 2.1. Introduction..........................................................................................................17

2.2. Thermodynamic Considerations ..........................................................................19

2.2.1. Pure Anti-Solvent Phase Behaviour.............................................................19

2.2.2. Anti-Solvent—Solvent Phase Behaviour.....................................................23

2.2.3. Anti-Solvent—Solvent—Solute Phase Behaviour.......................................35

2.2.4. Modelling GAS and ASES Phase Behaviour ..............................................40

2.2.4.1. Modelling Anti-Solvent—Solvent Phase Behaviour ...........................40

2.2.4.2. Modelling Anti-Solvent—Solvent—Solute Phase Behaviour .............44

2.3. Crystallisation Theory Considerations ...............................................................48

2.4. Hydrodynamic Considerations ............................................................................56

2.5. Mass Transfer Considerations .............................................................................58

2.6. References............................................................................................................63

CHAPTER 3

3. DENSE GASES AS ANTI-SOLVENTS — APPLICATIONS ............68 3.1. Introduction..........................................................................................................68

3.2. Micronisation/Recrystallisation...........................................................................68

3.2.1. Applications of Micronised Pharmaceuticals ..............................................70

3.2.1.1. Increased Bioavailability......................................................................70

TABLE OF CONTENTS

VII

3.2.1.2. Intravenous Delivery............................................................................71

3.2.1.3. Oral Delivery........................................................................................71

3.2.1.4. Inhalation Delivery ..............................................................................72

3.2.1.5. Ocular Delivery....................................................................................72

3.2.1.6. Nasal Delivery......................................................................................73

3.2.1.7. Drug Delivery to the Central Nervous System ....................................74

3.2.2. Methods of Micronising Pharmaceuticals....................................................74

3.2.3. GAS as a Micronisation/Recrystallisation Technique .................................77

3.2.4. ASES as a Micronisation/Recrystallisation Technique ...............................81

3.3. Co-Precipitation/Encapsulation...........................................................................86

3.3.1. ASES as an Encapsulation Technique .........................................................87

3.4. Purification/Fractional Crystallisation.................................................................89

3.4.1. GAS as a Purification / Fractional Crystallisation Technique ....................90

3.4.2. ASES as a Fractionation Technique.............................................................93

3.5. Experimental Techniques ....................................................................................95

3.5.1. GAS Experimental Techniques ....................................................................95

3.5.2. ASES Experimental Techniques ..................................................................97

3.6. References..........................................................................................................101

CHAPTER 4

4. DENSE GASES AS ANTI-SOLVENTS — PARTICLE

FORMATION MECHANISMS AND THE INFLUENCE OF

PROCESS PARAMETERS ...........................................................................114 4.1. Introduction........................................................................................................114

4.2. GAS – The Influence of Process Parameters.....................................................115

4.3. ASES The Influence of Process Parameters ......................................................117

4.3.1. The Formation of Microspheres.................................................................120

4.3.1.1. One Droplet — One Particle Theory .................................................121

4.3.1.2. One Droplet — Many Particles Theory.............................................121

4.3.1.3. No Droplet Formation, Nucleation and Growth Theory....................122

4.3.2. Fibre Formation..........................................................................................122

4.3.3. The Effect of Solute ...................................................................................123

TABLE OF CONTENTS

VIII

4.3.4. The effect of Anti-Solvent Density (Pressure)...........................................125

4.3.5. The Effect of Temperature.........................................................................132

4.3.6. The Effect of Solution/Anti-Solvent Flow Rate ........................................133

4.3.7. The Effect of Solvent Type........................................................................136

4.3.8. The Effect of Solute Concentration ..........................................................137

4.3.9. The Effect of Nozzle Diameter/Type.........................................................140

4.3.10. Pre-Addition of Anti-Solvent...................................................................144

4.3.11. The Effect of Stirring Rate.......................................................................145

4.4. Effect of Process Parameters on Encapsulation/Co-Precipitation.....................145

4.5. References..........................................................................................................148

CHAPTER 5

5. SOLUBILITY OF SOLUTES IN SOLUTIONS EXPANDED

WITH CO2 ...........................................................................................................155

5.1. Introduction........................................................................................................155

5.2. Experimental......................................................................................................156

5.2.1. Experimental Apparatus and Procedure.....................................................156

5.3. Volumetric Expansion of the CO2—Solvent Systems.......................................160

5.4. Solid Solubility in Expanded Solution ..............................................................165

5.5. Conclusions........................................................................................................177

5.6. References..........................................................................................................179

CHAPTER 6

6. SYNTHESIS AND PURIFICATION OF COPPER

INDOMETHACIN USING THE GAS AND ASES

PROCESSES .......................................................................................................180 6.1. Introduction........................................................................................................180

6.2. Experimental......................................................................................................183


6.3. Synthesis of Cu-Indo by the GAS Process ........................................................187

6.4. Synthesis and Purification of Cu-Indo by the ASES Process ............................192

TABLE OF CONTENTS

IX

6.5. Characterisation of the GAS and ASES Processed Cu-Indo .............................194

6.6. Comparison between CO2 and Ethanol as Antisolvents....................................200

6.7. Conclusion .........................................................................................................201

6.8. References..........................................................................................................203

CHAPTER 7

7. MICRONISATION OF COPPER INDOMETHACIN

USING THE GAS AND ASES PROCESSES .........................................205 7.1. Introduction........................................................................................................205

7.2. Experimental......................................................................................................206


7.3. Micronisation by the GAS Process....................................................................208

7.3.1. The Effect of Expansion Rate and Stirring................................................209


7.3.3. The Effect of Solvent .................................................................................214

7.3.4. The Effect of Concentration.......................................................................217

7.3.5. Summary of the effects of GAS Process Variables ...................................219

7.4. Micronisation of Cu-Indo by the ASES Process ...............................................220

7.4.1. The Effect of Concentration.......................................................................222

7.4.2. The Effect of Nozzle Diameter ..................................................................228

7.4.3. The Effect of Antisolvent Density .............................................................230


7.4.5. The Effect of Solution Flow Rate ..............................................................238

7.4.6. The Effect of Solvent .................................................................................240

7.4.7. Summary of the effects of ASES Process Variables..................................244

7.5. Particle Size Distribution...................................................................................244

7.6. Dissolution Studies ............................................................................................248

7.7. Conclusions........................................................................................................250

7.8. References..........................................................................................................252

TABLE OF CONTENTS

X

CHAPTER 8

8. CO-PRECIPITATION OF COPPER INDOMETHACIN

AND POLYVINYLPYRROLIDONE BY THE ASES

PROCESS .............................................................................................................256 8.1. Introduction........................................................................................................256

8.1.1. Polyvinylpyrrolidone .................................................................................257

8.2. Experimental......................................................................................................260


8.3. Precipitation of PVP Using Dense CO2 as Anti-Solvent...................................263

8.3.1. The Effect of PVP Concentration ..............................................................265

8.3.2. The Effect of Anti-Solvent Density ...........................................................267


8.4. Co-Precipitation of PVP and Cu-Indo by ASES ...............................................271

8.5. Drug Content .....................................................................................................280

8.6. Solubility of the PVP—Cu-Indo Co-Precipitates..............................................281

8.7. Dissolution of the PVP—Cu-Indo Co-Precipitates ...........................................288

8.8. Conclusions........................................................................................................290

8.9. References..........................................................................................................292

APPENDIX I

I. LITERATURE REVIEW OF THE USE OF DENSE GASES

AS ANTI-SOLVENTS .....................................................................................295

I.1. Review of GAS Applications .............................................................................295

I.2. Review of ASES Applications ...........................................................................305

I.3. References ..........................................................................................................317

APPENDIX II

II. VOLUMETRIC EXPANSION EXPERIMENTAL

RESULTS — BINARY DATA .....................................................................329

II.1. CO2—DMF.......................................................................................................329

TABLE OF CONTENTS

XI

II.2. CO2—NMP.......................................................................................................331

II.3. CO2—DMSO....................................................................................................332

APPENDIX III

III. VOLUMETRIC EXPANSION EXPERIMENTAL

RESULTS — TERNARY DATA .................................................................333 III.1. CO2—DMF—Cu-Indo....................................................................................333

III.2. CO2—DMF—Cu-Acetate................................................................................335

III.3. CO2—DMF—Indomethacin............................................................................336

APPENDIX IV

IV. VOLUMETRIC EXPANSION EXPERIMENTAL

RESULTS — QUATERNARY DATA ......................................................337

IV.1. CO2—DMF—Cu-Indo—Cu-Acetate..............................................................337

IV.2. CO2—DMF—Cu-Indo—Indomethacin ..........................................................338

IV.3. CO2—DMF—Cu-Indo—Acetic acid ..............................................................338

APPENDIX V

V. DISSOLUTION EXPERIMENTAL RESULTS ...................................339

V.1. Dissolution of Micronised Cu-Indo..................................................................339

V.2. Dissolution of PVP—Cu-Indo Co-Precipitates ................................................340

APPENDIX VI

VI. SEM IMAGES OF COPPER INDOMETHACIN

PARTICLES PRODUCED BY THE GAS AND ASES

PROCESSES .......................................................................................................341

VI.1. Particles Produced by the GAS Process ..........................................................341

VI.2. Particles Produced by the ASES Process ........................................................347

TABLE OF CONTENTS

XII

APPENDIX VII

VII. PARTICLE SIZE ANALYSES RESULTS .........................................351

APPENDIX VIII

VIII. SEM IMAGES OF PVP AND PVP—Cu-INDO CO-

PRECIPITATES ................................................................................................354

VIII.1. PVP Particles Produced by the ASES Process .............................................354

VIII.2. PVP—Cu-Indo Particles Produced by the ASES Process ............................355

LIST OF PUBLICATIONS ................................................................................358

LIST OF FIGURES

XIII

Figure 1.1 Molecular structure of [Cu2(indomethacin)4(DMF)2].....................................5

Figure 1.2 The inflammation cascade. .............................................................................8

Figure 1.3 COX I and COX II sites of action.................................................................10

Figure 2.1 Phase diagram for a pure substance..............................................................20

Figure 2.2 Reduced density of a pure substance in the vicinity of its critical

point............................................................................................................................22

Figure 2.3 Diffusivity behaviour of CO2........................................................................24

Figure 2.4 Viscosity behaviour of CO2. ........................................................................25

Figure 2.5 Solubility of ethane and ethene in organic solvents......................................27

Figure 2.6 Solubility of CO2 in organic solvents. ..........................................................28

Figure 2.7 Volumetric expansion of ethyl acetate pressurised with CO2.......................30

Figure 2.8 Variation of density of organic solvents upon pressurisation with

CO2. ............................................................................................................................31

Figure 2.9 ∆V vs CO2 mole fraction for various organic solvents. ................................33

Figure 2.10 ∆V* vs CO2 mole fraction for various organic solvents. ............................34

Figure 2.11 Ternary phase diagram for CO2—water—ethanol at constant

temperature and pressure. (a) 35°C, 6.9 MPa, (b) 35°C, 10.2 MPa...........................36

Figure 2.12 Ternary anti-solvent—solvent—solute phase behaviour............................37

Figure 2.13 Ternary phase diagram for CO2—toluene—naphthalene at

temperatures between 315 and 325 K and pressure between 8.2 and 9.8

MPa.............................................................................................................................39

Figure 2.14 The relationship between supersaturation and particle growth...................52

LIST OF FIGURES

XIV

Figure 2.15 Ternary phase diagram showing different polymer precipitation

schemes for the ASES process. ..................................................................................61

Figure 3.1 Diagram of a typical GAS experimental apparatus......................................96

Figure 3.2 Diagram of a typical ASES experimental apparatus.....................................98

Figure 4.1 Common ASES processing conditions in terms of anti-solvent

phase behaviour. .......................................................................................................126

Figure 4.2 Common ASES processing conditions in terms of anti-solvent

density.......................................................................................................................127

Figure 4.3 Ternary phase diagram showing the effect of solvent quality upon

the two phase envelope and the mass transfer pathway for a polystyrene

solution precipitated into compressed CO2...............................................................138

Figure 4.4 Different nozzle designs used in ASES. .....................................................141

Figure 5.1 Experimental apparatus used to determine volumetric expansion

and solute solubility..................................................................................................157

Figure 5.2 Volumetric expansion of DMF—CO2 as a function of pressure at

25, 30 and 40°C. .......................................................................................................161

Figure 5.3 Volumetric expansion of NMP—CO2 and DMSO—CO2 as a

function of pressure..................................................................................................162

Figure 5.4 Comparison between the DMF—CO2, NMP—CO2 and DMSO—

CO2 systems at 25°C.................................................................................................163

Figure 5.5 Volumetric expansion of the DMF—CO2, NMP—CO2 and

DMSO—CO2 systems as a function of CO2 mole fraction. .....................................164

Figure 5.6 Comparison between literature results and this work for the

volumetric expansion of DMF with CO2 at 25°C.....................................................166

Figure 5.7 Mole fraction of Cu-Indo dissolved in the liquid phase..............................167

LIST OF FIGURES

XV

Figure 5.8 Concentration of Cu-Indo in DMF (CO2 free basis)...................................168

Figure 5.9 Solubility of Cu-Acetate in expanded DMF at 25°C..................................170

Figure 5.10 Solubility of indomethacin in expanded DMF at 25°C.............................171

Figure 5.11 Mole fraction of CO2 as a function of pressure at 25°C in the

presence of solute. ....................................................................................................173

Figure 5.12 Volumetric expansion of DMF containing acetic acid at 25°C. ...............175

Figure 5.13 Solubility of Cu-Indo in quaternary systems at 25°C...............................176

Figure 5.14 Solubility of the second solute in the CO2—DMF—Cu-Indo—

Solute systems at 25°C.............................................................................................178

Figure 6.1 Comparison between the conventional, GAS and ASES processes

for the synthesis of Cu-Indo.....................................................................................182

Figure 6.2 GAS experimental apparatus......................................................................185

Figure 6.3 ASES experimental apparatus.....................................................................185

Figure 6.4 SEM image of the Cu-Indo obtained from the conventional

synthesis....................................................................................................................189

Figure 6.5 SEM images of the precipitates collected after GAS processing at

25°C. (a) A 5 mg.g-1 solution of Cu-Indo in DMF expanded slowly without

stirring. (b) A 200 mg.g-1 solution of Cu-Indo in DMF expanded rapidly

with stirring...............................................................................................................190

Figure 6.6 SEM images of Cu-Indo precipitates collected after ASES

processing at 25°C and 6.89 MPa using a solution flow rate of 0.2 mL.min-1

and a 1020 µm nozzle. (a) 20 mg.g-1, (b) 200 mg.g-1 solution of Cu-Indo in

DMF..........................................................................................................................195

Figure 6.7 Comparison between the Infrared spectra of the conventionally

synthesised Cu-Indo and the GAS and ASES processed Cu-Indo. (a)

LIST OF FIGURES

XVI

Conventional Synthesis, (b) GAS Process, 5 mg.g-1 solution, (c) GAS

Process, 200 mg.g-1 solution, (d) ASES Process, 20 mg.g-1 solution.......................197

Figure 6.8 Comparison between the X-band EPR spectra of the conventionally




Figure 6.9 Comparison between the DSC spectra of the conventionally




Figure 7.1 SEM images of Cu-Indo particles produced by the GAS process at

25°C from a 5 mg.g-1 solution of Cu-Indo in DMF. (a) slow expansion, (b)

rapid expansion.........................................................................................................211


25°C from a stirred 5 mg.g-1 solution of Cu-Indo in DMF. (a) slow

expansion, (b) rapid expansion.................................................................................213


40°C from a stirred 5 mg.g-1 solution of Cu-Indo in DMF. (a) slow

expansion, (b) rapid expansion.................................................................................215


25°C. (a) A stirred 5 mg.g-1 solution of Cu-Indo in DMSO expanded

rapidly. (b) A stirred 5 mg.g-1 solution of Cu-Indo in DMSO expanded

rapidly.......................................................................................................................216


25°C. (a) A stirred 200 mg.g-1 solution of Cu-Indo in DMF expanded

rapidly. (b) A stirred 50 mg.g-1 solution of Cu-Indo in DMF expanded

slowly........................................................................................................................218

LIST OF FIGURES

XVII

Figure 7.6 SEM images of Cu-Indo particles produced by the ASES process at

25°C and 6.89 MPa with a solution flow rate of 0.2 mL.min-1 and a nozzle

diameter of 1020 µm, (a) 5 mg.g-1, (b) 20 mg.g-1,(c) 100 mg.g-1, (d) 200

mg.g-1 solution of Cu-Indo in DMF..........................................................................223

Figure 7.7 SEM images of Cu-Indo particles by the ASES process from 200

mg.g-1 solutions of Cu-Indo in DMF at 25°C and 6.89 MPa with a solution

flow rate of 0.2 mL.min-1 and a nozzle diameter of 1020 µm. .................................224

Figure 7.8 Photographs of DMF solutions sprayed into liquid CO2 at 25°C and

6.89 MPa at a solution flow rate of 0.2 mL.min-1 through a 1020 µm nozzle,

(a) Pure DMF, (b) 5 mg.g-1, (c) 100 mg.g-1, (d) 200 mg.g-1 Cu-Indo solution.........227



diameter of 229 µm. (a) 5 mg.g-1, (b) 20 mg.g-1, (c) 100 mg.g-1, (d) 200

mg.g-1 solution of Cu-Indo in DMF..........................................................................229

Figure 7.10 SEM images of Cu-Indo particles produced by the ASES process

at 25°C with a solution flow rate of 0.2 mL.min-1 and a nozzle diameter of

229 µm. (a) 5 mg.g-1 solution of Cu-Indo in DMF at 13.79 MPa, (b) 20

mg.g-1 solution of Cu-Indo in DMF at 10.34 MPa, (c) 100 mg.g-1 solution of

Cu-Indo in DMF at 13.79 MPa, (d) 200 mg.g-1 solution of Cu-Indo in DMF

at 10.34 MPa.............................................................................................................231


mg.g-1 solutions of Cu-Indo in DMF at 25°C and 6.55 MPa with a solution

flow rate of 0.2 mL.min-1 and a nozzle diameter of 229 µm. ...................................233


at 40°C and 14.48 MPa with a solution flow rate of 0.2 mL.min-1 and a

nozzle diameter of 229 µm. (a) 5 mg.g-1, (b) 20 mg.g-1, (c) 100 mg.g-1

solution of Cu-Indo in DMF.....................................................................................235

LIST OF FIGURES

XVIII

Figure 7.13 X-ray diffraction patterns of Cu-Indo particles produced from the

conventional synthesis process and by the ASES process from 100 mg.g-1

solutions of Cu-Indo in DMF with a solution flow rate of 0.2 mL/min and a

nozzle diameter of 229µm. (a) Conventional synthesis process. (b) 25 °C

and 6.89 MPa. (c) 25 °C and 13.79 MPa. (d) 40 °C and 14.48 MPa........................237


from 100 mg.g-1 solutions of Cu-Indo in DMF at 25°C and 6.89 MPa using

a nozzle with a diameter of 229 µm. Solution flow rate of (a) 0.1 mL.min-1,

(b) 0.2 mL.min-1, (c) 0.5 mL.min-1. ..........................................................................239


from DMSO and NMP at a Cu-Indo concentration of 5 mg.g-1 with a

solution flow rate of 0.2 mL.min-1 and a nozzle diameter of 1020 µm. (a)

Cu-Indo in DMSO at 40°C and 14.48 MPa. (b) Cu-Indo in NMP at 25°C

and 6.89 MPa............................................................................................................241


from 100 mg.g-1 Cu-Indo solutions in NMP at 25°C and 6.89 MPa with a

solution flow rate of 0.2 mL.min-1 and a nozzle diameter of 1020 µm. ...................243

Figure 7.17 Particle size distributions of Cu-Indo particles produced from

DMF solutions by the ASES process. (a) 200 mg.g-1 Cu-Indo solution at

25°C and 10.34 MPa using a 1020 µm nozzle and a solution flow rate of 0.2

mL.min-1, (b) 20 mg.g-1 Cu-Indo solution at 25°C and 6.89 MPa using a 229

µm nozzle and a solution flow rate of 0.2 mL.min-1. ...............................................246

Figure 7.18 Particle size distributions of Cu-Indo particles produced from

DMF solutions at a concentration of 100 mg.g-1 by the ASES process using

a 229 µm nozzle, 0.2 mL.min-1 solution flow rate. (a) 25°C and 6.89 MPa,

(b) 25°C and 13.79 MPa, (c) 40°C and 14.48 MPa..................................................247

Figure 7.19 Dissolution of Cu-Indo produced by ASES. 100 mg.g-1 solution of

Cu-Indo in DMF using a 229 µm nozzle and a solution flow rate of 0.2

mL.min-1. ..................................................................................................................249

LIST OF FIGURES

XIX

Figure 8.1 Types of polymer—drug composites..........................................................257

Figure 8.2 Polyvinylpyrrolidone chemical structure....................................................257

Figure 8.3 SEM image of PVP. ....................................................................................264

Figure 8.4 SEM images of PVP particles produced from DMF by the ASES

process at 25°C and 14.0 MPa using solution flow rate of 0.1 mL.min-1

through a 229 µm nozzle. (a) 50 mg.g-1, (b) 100 mg.g-1 solution of PVP in

DMF..........................................................................................................................266

Figure 8.5 SEM images of PVP particles produced from DMF solutions of

PVP at a concentration of 50 mg.g-1 by the ASES process at 25°C and 7.0

MPa using solution flow rate of 0.1 mL.min-1 through a 229 µm nozzle. ...............268

Figure 8.6 SEM images of PVP particles produced from DMF solutions of

PVP at a concentration of 100 mg.g-1 by the ASES process at 25°C and 6.6

MPa using solution flow rate of 0.1 mL.min-1 through a 229 µm nozzle. ...............270

Figure 8.7 SEM images of PVP particles produced by the ASES process from

a 50 mg.g-1 solution of in DMF at 10°C and 5.5 MPa using a solution flow

rate of 0.1 mL.min-1 through a 229 µm nozzle.........................................................272

Figure 8.8 SEM images of the PVP—Cu-Indo particles produced from DMF

solutions with a solute concentration of 100 mg.g-1 by the ASES process at

25°C using a solution flow rate of 0.1 mL.min-1 through a 229 µm nozzle.

PVP to Cu-Indo ratio (a) 10.39 MPa and a 70:30 PVP:Cu-Indo ratio, (b)

14.0 MPa and a 50:50 PVP:Cu-Indo ratio. ...............................................................274

Figure 8.9 SEM images comparing the particles produced from ASES using

100 mg.g-1 solutions of Cu-Indo, PVP and PVP—Cu-Indo in DMF at 25°C

and 14.0 MPa. (a) Cu-Indo, (b) PVP, (c) 50:50 PVP:Cu-Indo ratio.........................275

Figure 8.10 Photograph of the PVP—Cu-Indo precipitate formed from the

ASES process............................................................................................................277

LIST OF FIGURES

XX


solutions with a solute concentration of 100 mg.g-1 by the ASES process at

25°C and 10.39 MPa using a solution flow rate of 0.1 mL.min-1 through a

229 µm nozzle. PVP to Cu-Indo ratio (a) 70:30, (b) 50:50......................................278

Figure 8.12 Photograph of the precipitation of PVP—Cu-Indo co-precipitate

by ASES. The solution concentration was 100 mg.g-1 in a 90:10 ratio of

PVP:Cu-Indo. The system was at 25°C, 10.39 MPa with a solution flow rate

of 0.1 mL.min-1 through a 229 µm nozzle................................................................279

Figure 8.13 PVP—sulfathiazole complex. ...................................................................283

Figure 8.14 Comparison between the DSC spectra of (a) PVP, (b) Cu-Indo, (c)

50:50 physical mix of PVP and Cu-Indo, (d) 50:50 PVP—Cu-Indo co-

precipitate. ................................................................................................................284

Figure 8.15 Comparison between the X-band EPR spectra of (a) Cu-Indo and

(b) 50:50 PVP—Cu-Indo co-precipitate...................................................................286

Figure 8.16 Proposed PVP—Cu-Indo complex. (a) Proposed dimeric complex,

(b) proposed monomeric complex............................................................................287

Figure 8.17 Dissolution of Cu-Indo..............................................................................289

LIST OF TABLES

XXI

Table 1.1 FDA guidelines for levels of residual solvents in pharmaceuticals.

Permitted Daily Exposure (PDE)..................................................................................3

Table 1.2 Non-steroidal anti-inflammatory drugs. ...........................................................7

Table 2.1 Critical pressures and temperatures of anti-solvents used in GAS and

ASES...........................................................................................................................21

Table 3.1 Conventional methods of micronisation.........................................................75

Table 5.1 Chemicals and reagents. ...............................................................................156


Table 6.2 The yield and purity of the Cu-Indo synthesis in DMF expanded

solution at 25°C. (GAS Process). .............................................................................188

Table 6.3 The yield and purity of the Cu-Indo synthesis in DMF expanded

solution at 25°C in the presence of excess reactant. (GAS process)........................192

Table 6.4 The purity of the Cu-Indo synthesis in DMF expanded solution at 25°C.

(ASES Process).........................................................................................................193

Table 6.5 Comparison between CO2 and Ethanol as antisolvents for the Cu-

Indo—DMF system at 25°C.....................................................................................201

Table 6.6 Comparison between GAS and the conventional process in the

synthesis of Cu-Indo at 25°C....................................................................................201


Table 7.2 GAS micronisation results............................................................................210

Table 7.3 Experimental conditions and results for the micronisation of Cu-Indo

using the ASES technique. The solution and CO2 flow rates were 0.2 and 4–5

mL.min-1 respectively unless otherwise stated.........................................................221

LIST OF TABLES

XXII

Table 7.4 Particle size distribution results....................................................................245

Table 8.1 K values and molecular weights of PVP. .....................................................258


Table 8.3 Typical properties of PVP............................................................................263

Table 8.4 Results of the precipitation of PVP by ASES...............................................265

Table 8.5 Results of the co-precipitation of PVP and Cu-Indo by ASES. ...................273

Table 8.6 Drug content of PVP—Cu-Indo co-precipitates...........................................280

Table 8.7 Solubility of Cu-Indo and PVP—Cu-Indo co-precipitate in ethanol at

25°C..........................................................................................................................281

CHAPTER 1

1

1. BACKGROUND

1.1. Introduction

In many countries demands are being placed on the production process by regulatory

authorities due to environmental and health concerns. In these countries most organic

solvents are banned for use in the food industry, or the tolerances of residual solvent in

the final product are very low.1 In the future these demands will become greater as

worldwide concern for the environment and personal health increases.

A recent book entitled “Green Chemistry”2 highlighted the following facts regarding

environmental concerns and the chemical industry:

i. Despite the technical advances that have been made in the last

century, the manufacture, processing, use, and disposal of many

chemical products has had a negative impact on human health and the

environment.

ii. Public opinion at present is that all chemicals are toxic or hazardous,

and should be treated with suspicion. Public opinion has been

translated into tougher government regulation of the chemical

industry. As an example, in the United States in 1950 less than 20

environmental laws were in place, but by 1995 these laws had

increased to over 120.

iii. The cost of complying with environmental regulations through

remediation activities (i.e. waste treatment, control, and disposal

costs) in the United States is in the range 100 to 150 billion US

dollars per year.

iv. The remediation approach to environmental protection is not

sustainable. These costs should be reclaimed for use in the research

and development of environmentally friendly alternatives.

v. Areas of investigation into environmentally friendly processes should

encompass:

CHAPTER 1

2

a. Finding alternative chemical feedstocks that are less hazardous and

are renewable.

b. Selecting reagents that are less hazardous, produce less waste, are

more selective and have greater efficiency.

c. Finding alternative chemical syntheses that produce less waste.

d. Finding alternative solvents that are less hazardous to the

environment and to human health. Solvents present one of the

greatest hazards in the chemical industry because they are used in

such high volumes.

It is apparent that the survival of the chemical industry is dependent on the discovery of

environmentally friendly alternatives to existing processes. Nowhere is this truer than in

the pharmaceutical industry where product specifications are especially stringent.

The Food and Drug Administration (FDA) guidelines of levels of solvent residues

allowed in pharmaceuticals are given in Table 1.1.3 Solvents are classified as Class 1, 2

or 3 depending on the nature of the solvent in terms of toxicity. Class 1 solvents should

be avoided, Class 2 solvents should be limited and Class 3 solvents are not considered

hazardous to humans at levels normally accepted in pharmaceuticals. The Permitted

Daily Exposure (PDE) is the maximum amount of solvent that is allowed in an

equivalent daily dose of the pharmaceutical.

The FDA guidelines for residual solvents state that all steps should be taken by the

manufacturer of the pharmaceutical to reduce the use and levels of solvents in the final

product. The onus is on the manufacturer to prove that the levels of residual solvent in

the pharmaceutical meet the requirements in Table 1.1. The use of solvents in the

pharmaceutical industry is unavoidable, and the aim becomes to choose the most

acceptable solvent that will give the desired results, and to reduce the residues of this

solvent to acceptable limits. The reduction of solvent residues is non-trivial and can

become a time consuming and expensive part of the chemical process.

CHAPTER 1

3

Table 1.1 FDA guidelines for levels of residual solvents in pharmaceuticals.3

Permitted Daily Exposure (PDE)

Solvent PDE / mg/day Solvent PDE / mg/day

Class 1 Solvents

Benzene 0 1,1-Dichloroethane 0

Carbon Tetrachloride 0 1,1,1-Trichloroethane 0

1,2-Dicloroethane 0

Class 2 Solvents

Acetonitrile 4.1 Hexane 2.9

Chlorobenzene 3.6 Methanol 30

Chloroform 0.6 2-Methoxyethanol 0.5

Cyclohexane 38.8 Methylbutyl ketone 0.5

1,2-Dichloroethane 18.7 Methylcyclohexane 11.8

Dichloromethane 6 N-Methylpyrrolidone 48.4

1,2-Dimethoxyethane 1 Nitromethane 0.5

N,N-Dimethylacetamide 10.9 Pyridine 2

N,N-Dimethylformamide 8.8 Sulfolane 1.6

1,4-Dioxane 3.8 Tetralin 1

2-Ethoxyethanol 1.6 Toluene 8.9

Ethyleneglycol 6.2 1,1,2-Trichloroethane 0.8

Formamide 2.2 Xylene 21.7

Class 3 Solvents

Acetic acid 50 Heptane 50 Acetone 50 Isobutyl acetate 50

Anisole 50 Isopropyl acetate 50

1-Butanol 50 Methyl acetate 50

2-Butanol 50 3-Methyl-1-butanol 50

Butyl acetate 50 Methylethyl ketone 50

tert-Butylmethyl ether 50 Methylisobutyl ketone 50

Cumene 50 2-Methyl-1-propanol 50

Dimethyl sulfoxide 50 Pentane 50

Ethanol 50 1-Pentanol 50

Ethylacetate 50 1-Propanol 50

Ethyl ether 50 2-Propanol 50

Ethyl formate 50 Propyl acetate 50

Formic Acid 50 Tetrahydrofuran 50

CHAPTER 1

4

The trend towards greater environmental and health concerns has resulted in the

development of innovative technologies that may prove to be far superior to the

technology that they replace. Supercritical fluid technology is such a technology that

has advanced rapidly in recent years.

It has long been the goal of supercritical fluid researchers to discover alternative

processing methods that are environmentally friendly as well as commercially feasible.

This has recently been referred to as developing a “Sustainable Technology”.4 It is no

coincidence that as the demand for environmentally friendly alternative processes

increased, so the research effort into supercritical fluid technology increased. The

reason for this is mainly due to the use of carbon dioxide. Carbon dioxide, is able to

replace more expensive and toxic substances such as chlorofluorocarbons and carbon

tetrachloride at near or supercritical conditions.5 When carbon dioxide, which is

environmentally benign, is used as the solvent, the issues of solvent residue and solvent

waste are eliminated.

1.2. Copper Indomethacin

Copper indomethacin is a novel non-steroidal, anti-inflammatory drug (NSAID) that has

been developed and manufactured by Biochemical Veterinary Research (BVR).6-8 A

representation of the drug [Cu2(indomethacin)4L2] [Cu-Indo; indomethacin = 1-4-

(chlorobenzoyl)-5-methoxy-2-methyl-1H-indole-3-acetic acid, L = N,N-

dimethylformamide (DMF)] is shown in Figure 1.1.

The structure of Cu-Indo has been determined previously.9,10 The molecule consists of

two Cu atoms linked by four indomethacin molecules. A DMF molecule is bound to

each Cu atom. If the drug is synthesised in another solvent, that solvent will replace the

DMF.10,11 Prior to the release of Cu-Indo by BVR, there had been no clinical

application of the copper complex despite a number of reports on the preparation of Cu-

Indo complexes in the literature.12-14

CHAPTER 1

5

Cu

Cu

O

O

O O

C

C

C

R

DMF

R

R

R

DMF

CCH3

CH2

O

= R

N

O

C

O

O

O

Figure 1.1 Molecular structure of [Cu2(indomethacin)4(DMF)2].

CHAPTER 1

6

It has been stated that NSAIDs are the most frequently prescribed drugs in the world

with over 100 million prescriptions per annum.15,16 Historically, the use of synthetic

NSAIDs as anti-inflammatory, analgesic and antipyretic agents began in 1899 with the

introduction of aspirin.17 By the end on the nineteenth century a number of other drugs

were discovered that showed behaviour similar to aspirin.17 Of these only derivatives of

acetaminophen are in use today.17 More recently, new generations of aspirin-like drugs

have been introduced which are listed in Table 1.2.

The drugs listed in Table 1.2 are antipyretic, analgesic and anti-inflammatory. The

activities of these drugs are significantly different. However, there is no clear

explanation for the different activities.17 Generally the drugs similar to aspirin possess

side effects, such as gastric or intestinal ulceration.18-23 Despite the side effects, the use

of NSAIDs continues to expand, particularly in the treatment of chronic diseases in the

elderly, in whom the gastrointestinal side effects are the most common and serious.24-27

It has been reported that up to 40% of patients taking NSAIDs suffer from

gastrointestinal side effects.25 Much research has been performed to find compounds

with the same therapeutic effects and lower side effects. The search for safer NSAIDs

has resulted in over 35 NSAIDs being available on the market today.18

The current understanding of inflammation in the body is that it is due to the response

of the body to an antigen such as gamma globulin.28 Once entering the body the antigen

combines with antibodies from the defense system releasing a number of chemical

mediators such as prostaglandins and leukotrines. The release of these mediators results

in heat, redness, swelling and pain.11 The inflammation cascade and the sites of action

of NSAIDs are shown in Figure 1.2.

CHAPTER 1

7

Table 1.2 Non-steroidal anti-inflammatory drugs.17

NSAID Year Comment

Aspirin 1899 Most widely prescribed analgesic.

Phenylbutazone 1949 Limited long term use due to toxicity.

Acetominophen 1949 Weak anti-inflammatory activity.

Indomethacin 1963 Toxicity limits its use.

Sulindac Half as potent as indomethacin with common

untoward reactions.

Fenamates 1950 Frequent side effects.

Tolmetin 1976

Potency between asprin and indomethacin.

Causes gastric erosions and prolongs bleeding

time.

Ibuprofen

—

Better tolerated than asprin and indomethacin.

Still suffer from detrimental features of the other

NSAIDs.

Naproxen — Ibid.

Fenoprofen — Ibid.

Piroxicam —

Better tolerated than asprin and indomethacin.

Long half life which means fewer doses. Causes

gastric erosion and prolongs bleeding time.

CHAPTER 1

8

Arachidonic acid

Cyclic endoperoxides

Prostaglandins; Thromboxanes; Prostacyclin

Inflammatory & Immune SystemResponse; Gastric Protection

Leukotrienes

Tissue Injury to Cell Membrane Phospholipid(e.g. Chemical, Physical, Biological)

Phospholipidase A2(Inhibited in part by NSAIDs)

Lipooxygenase(Variable inhibition byNSAIDs)

Cyclooxygenase I & II(Inhibited by NSAIDs)

Inflammatory & Immune SystemResponse.

Figure 1.2 The inflammation cascade.11

CHAPTER 1

9

The anti-inflammatory action of NSAIDs is thought to be the inhibition of prostaglandin

synthesis by the inhibition of the cyclooxygenase (COX) enzyme.29 There are two

forms of COX, which are known as COX I and COX II. The two forms are similar

except that COX II has a larger active site.30 COX I is thought to be involved in the

maintenance of essential physiological functions, such as platelet aggregation, gastric

protection and maintenance of renal function.31 The activation of COX I leads to the

production of prostacyclin, which is anti-thrombogenic and when expressed by the

gastric mucosa is gastric protective.32 The roles of COX I and COX II are illustrated in

Figure 1.3.

COX II is induced under inflammatory conditions33,34 and results in the inflammatory

response of the body. Most modern NSAIDs preferentially inhibit COX I over COX

II.31,35,36 This selective inhibition of COX I is thought to be responsible for the

gastrointestinal toxicity of NSAIDs.23,35,37 Two selective COX II inhibitors that are

currently available are celecoxib and rofecoxib.11 These compounds have still been

found to have deleterious effects with celecoxib having being linked to 10 deaths and 11

cases of gastro-intestinal hemorrhage after three months on the US market and 2.5

million prescriptions filled.11 Finding alternatives is therefore of serious importance.

Copper has been used as a traditional drug remedy for arthritis for many years.38

Interest in copper complexes as anti-inflammatory drugs was recently inspired by

Sorensen39 who suggested that the active forms of anti-inflammatory drugs were in fact

the Cu-complexes. The Cu-complex was found to be more active than either the parent

Cu(II) salt or NSAID with fewer side effects.39 Since this report there have been

numerous reports on the potential benefits of Cu(II)-complexes.40

Two copper complexes of asprin were developed in Australia as topical application for

human use in the 1970s and 1980s.41,42 Production of these two complexes has since

ceased and they are no longer available. The release of Cu-Indo by BVR for the

treatment of inflammation and pain in animals is the first widely available Cu-NSAID

complex.

CHAPTER 1

10

Normal G.I.Tract

Site of Inflammation

Arachidonic Acid

NSAIDs

NSAIDs

COX - Iconstitutive

COX - IIinducible

Physiological prostaglandin,e.g., gastric protection

Pathological prostaglandin,

e.g., inflammation

Cytokines , Endotoxins

Corticosteroids

COX - II inhibitors

(+)

(-)

(-)

(-)(-)

(-)(-)

Figure 1.3 COX I and COX II sites of action.11

CHAPTER 1

11

The exact reason why Cu-Indo has none of the side effects of traditional NSAIDs and

enhanced potency is still unknown at present. It is presumed to be uninhibiting to COX

I while inhibiting COX II. Another advantage of Cu-Indo is its free radical scavenging

activity. Copper indomethacin has been found to be ten times more effective at

dismutasing superoxide than superoxide dismutase12 whereas indomethacin has no

activity against superoxides.

A clear example of the anti-ulcerogenic activity of Cu-Indo is the fact that indomethacin

cannot be used in dogs because of fatal gastrointestinal side effects,19,21,22 while Cu-

Indo can be used without these side effects.43 Since the introduction of Cu-Indo in

Australia over 7 million dog doses have been sold capturing a sizable portion of the

domestic market.

At present Cu-Indo is available in the form of tablets, granules and a paste. The tablets

are used for the treatment of inflammation and pain in dogs while the granules and

pastes are used for horses. Future directions in the development of Cu-Indo are to

increase the range of formulations and to introduce the drug into the human market. The

motivation for the research described in this thesis was to accelerate the development of

Cu-Indo.

To enable Cu-Indo to be marketed on a worldwide basis in both animals and humans a

method of manufacture that produces a pure solvent free drug and is environmentally

friendly is needed. Conventional methods of synthesis of the drug are inadequate

because the processes needed to purify the drug are lengthy and costly. Organic solvents

and anti-solvents are used in the synthesis and purification steps, resulting in the

generation of solvent waste as well as the problem of residual solvent in the drug. A

more efficient method of manufacture would be to have the reaction, drug recovery and

purification all occurring in a single vessel. The applicability of using carbon dioxide as

an alternative anti-solvent is investigated in this thesis to demonstrate its applicability as

a tool in the synthesis of pharmaceuticals.

CHAPTER 1

12

The second motivation for the research described in this thesis was to produce Cu-Indo

in novel forms that would enable it to be marketed on a worldwide basis for both human

and animal use. A difficulty associated with developing new formulations of Cu-Indo is

the low solubility of the drug in solvents which are acceptable for use in animals or

humans. Further difficulties arise from the pH dependent stability of the drug. The

simplest and safest solvent to use in drug formulation is water. Many drugs, including

Cu-Indo, are only sparingly soluble in water and additives such as surfactants and

additional co-solvents are needed to increase the solubility of the drug in water. To date,

attempts to develop alternative formulations such as for ophthalmic, intravenous or

inhalation delivery have failed due to instability of the drug, or difficulty in producing

the drug with the correct physical properties. To overcome these problems, the potential

of using carbon dioxide as an anti-solvent in producing different forms of Cu-Indo were

investigated. The aims were to micronise Cu-Indo for possible use in inhalation

delivery, or as a suspension for application in ocular drug delivery and to co-precipitate

Cu-Indo with a suitable polymer to increase its solubility in solvents suitable for

intravenous applications.

In Chapters 2, 3 and 4 a detailed discussion of the theoretical and practical aspects of

using gases as anti-solvents is presented. A thorough review of the applications of gases

as anti-solvents is also included. In Chapter 5 the volumetric expansion behaviour of the

high pressure systems involving carbon dioxide and the solvents and solutes used in the

synthesis and processing of Cu-Indo are presented. The use of gases as anti-solvents in

the synthesis of Cu-Indo is presented in Chapter 6. The synthesis using gases as anti-

solvents is compared to the conventional method of synthesis using organic liquid anti-

solvents. The use of gases as anti-solvents in the recrystallisation and micronising of

Cu-Indo is demonstrated in Chapter 7. The effect of process parameters on the

precipitates produced are examined and discussed. The use of gases as anti-solvents to

form blends of drug and polymer with the formation of a Cu-Indo—polyvinyl

pyrrolidone co-precipitate is demonstrated in Chapter 8. The co-precipitate so formed is

compared to the pure drug in terms of solubility and dissolution rate.

CHAPTER 1

13

1.3. References

1. Perrut, M.; "Supercritical Fluid Applications : Industrial Developments and

Economic Issues," Proceedings of the 5th International Symposium on Supercritical

Fluids, Atlanta, Georgia, 2000.

2. Anastas, P. T.; Williamson, T. C.; "Frontiers in Green Chemistry," Green Chemistry,

Anastas, P. T. and Williamson, T. C., Ed.; Oxford University Press: Oxford, 1998, 1.

3. Hubbard, K. W.; "Guidence on Impurities: Residual Solvents," International

Conference on Harmonisation, 1997.

4. Eckert, C. A.; Bush, D.; Brown, J. S.; Liotta, C. L.; "Tuning Solvents for Sustainable

Technology," Proceedings of the 5th International Symposium on Supercritical


5. Hutchenson, K. W.; Foster, N. R.; Innovations in Supercritical Fluid Science and

Technology; Hutchenson, K. W. and Foster, N. R., Ed.; American Chemical Society:

Washington, DC, 1995; 608, 1.

6. Regtop, H. L.; Biffin, J. R.; "Divalent Metal Complexes of Indomethacin,

Compositions and Medical Methods of their Use," U.S. 5,466,824, 1995.

7. Regtop, H. L.; Biffin, J. R.; "Preparation of Divalent Metal Salts of Indomethacin,"

U.S. 5,310,936, 1994.

8. Regtop, H. L.; Biffin, J. R.; "Divalent Metal Salts of Indomethacin as Anti-

Inflammatory and Analgesic Agents," Wo. 9,014,337, 1990.

9. Weser, U.; Schubotz, L. M.; Catalytic Reaction of Copper Complexes with

Superoxide; Rainsford, K. D., Winlesham, K. and Whitehouse, M. W., Ed.;

Birkhauser Verlag: Basel, Boston, Stuttgart, 1981, 103.

10. Weder, J. E.; Hambley, T. W.; Kennedy, B. J.; Lay, P. A.; MacLachlan, D.;

Bramley, R.; Delfs, C. D.; Murray, K. S.; Moubaraki, B.; Warwick, B.; Biffin, J. R.;

Regtop, H. L.; "Anti-Inflammatory Dinuclear Copper(II) Complexes with

Indomethacin. Synthesis, Magnetism and EPR Spectroscopy. Crystal Structure of the

N,N-Dimethylformamide Adduct," Inorganic Chemistry. 1999, 38, 1736.

11. Weder, J. E. "Characterisation of Copper(II) Dimers of the Non-Steroidal Anti-

Inflammatory Drug Indomethacin," University of Sydney; 2000

CHAPTER 1

14

12. Weser, U.; Sellinger, K. H.; Lengfelder, E.; Werner, W.; Strahle, J.; "Structure of

Cu2(Indomethacin)4 and the Reaction with Superoxide in Aprotic Systems,"

Biochimica et Biophysica Acta 1980, 631, 232.

13. David, L.; Cosar, O.; Chis, V.; Negoescu, A.; Vlasin, I.; "ESR Study of Cu(II)-

Indomethacin and Its Pyridine and DMF Adducts," Applied Magnetic Resonance

1994, 6, 521.

14. Sorenson, J. R. J.; Oberley, L. W.; Crouch, R. K.; Kensler, T. W.; Kishore, V.;

Leuthauser, S. W. C.; Oberley, T. D.; Pezeshk, A.; "Pharmacological Activities of

Copper Compounds in Chronic Diseases.," Biol. Trace Elem. Res. 1983, 5, 257.

15. Gabriel, S. E.; Matteson, E. L.; "Economic and Quality-of-Life Impact on NSAIDs

in Rheumatoid Arthritis," PharmacoEconomics 1995, 8, 479.

16. Davies, N. M.; "Non-steroidal Anti-inflammatory Drug-Induced Gastro-intestinal

Permeablity," Alimentary Pharmacology and Therapeutics 1998, 12, 303.

17. Goodman, L. S.; Gilman, A.; The Pharmacological Basis of Therapeutics; 7th ed.;

Macmillan Publishing Co., Inc.: New York, 1985, 1839.

18. Reynolds, J. E. F.; Martindale. The Extra Pharmacopoeia.; 31st ed.; The

Pharmaceutical Press: London, 1996.

19. Adams, H. R.; "Veterinary Pharmacology and Therapeutics," 7th ed.; Iowa State

University Press 1995, 443.

20. Dukes, M. N. G.; "Meyler's Side Effects of Drugs An Encyclopedia of Adverse

Reactions and Interactions," 13th ed.; Elsevier: Amsterdam, 1996.

21. Ewing, G. O.; "Indomethacin-Associated Gastrointestinal Haemorrhage in a Dog,"

Journal. American Veterinary Medical Association 1972, 161, 1665.

22. Menguy, R.; Desbaillets, L.; "Role of Inhibition of Gastric Mucous Secretion in the

Phenomenon of Gastric Mucosal Injury by Indomethacin," American Journal of

Digestive Diseases 1967, 12, 862.

23. Wallace, J. L.; "Mechanisms of Non-Steroidal Anti-Inflammatory Drugs (NSAID)

Induced Gastrointestinal Damage - Potential for Development of Gastrointestinal

Tract Safe NSAIDs," Canadian Journal of Physiology and Pharmacology 1994, 72,

1493.

24. Thomas, J.; "Australian Presciption Products Guide," 26th ed.; Australian

Pharmaceutical Publishing Co. Ltd.: Hawthorn, Victoria, 1997; Vol. 1, 1327.

CHAPTER 1

15

25. Rodriguez, L. A. G.; "Non-Steroidal Anti-Inflammatory Drugs, Ulcers and Risk: A

Collabarative Meta-Analysis," Seminars in Arthritis and Rhematism 1997, 26, 16.

26. Hogan, D. B.; Campbell, N. R. C.; Crutcher, R.; Jennett, P.; MacLeod, N.;

"Presciption of Non-Steroidal Anti-inflammatory Drugs for Elderly People in

Alberta," Canadian Medical Association. Journal 1994, 151, 315.

27. Griffin, M. R.; Piper, J. M.; Daugherty, J. R.; Snowden, M.; Ray, W. A.; "Non-

Steroidal Anti-inflammatory drug Use and Increased Risk for Peptic Ulcer Disease in

Elderely Persons," Annals of Internal Medicine 1991, 114, 257.

28. Davidson, L. S. P.; "Davidson's Principles and Practice of Medicine," Davidson's

Principles and Practice of Medicine, 17th ed.; Edwards, C. R. W., Bouchier, I. A. D.,

Haslett, C. and Chilvers, E. R., Ed.; Churchill Livingstone: Edinburgh, 1995.

29. Kurumbail, R. G.; Stevens, A., M.; Gierse, J. K.; McDonald, J. J.; Stegeman, R. A.;

Pak, J. Y.; Gildehaus, D.; Miyashiro, J. M.; Penning, T. D.; Seibert, K.; Isakson, P.

C.; Stalling, W. C.; "Structural Basis for Selective Inhibition of Cyclo-Oxygenase-2

by Anti-inflammatory Agents.," Nature 1996, 384, 644.

30. Hawkey, C. J.; "COX-2 Inhibitors," The Lancet 1999, 353, 307.

31. Brideau, C.; Kargman, S.; Liu, S.; Dallob, A. L.; Ehrich, E. W.; Rodger, I. W.;

Chan, C. C.; "A Human Blood Assay for Clinical Evaluation of Biochemical

Efficacy of Cyclo-Oxygenase Inhibitors," Inflammation Research 1996, 45, 68.

32. Frohlich, J. C.; "A Classification of NSAIDs According to the Relative Inhibition

of Cyclo-Oxygenase Isoenzymes," JIPS 1997, 18, 30.

33. Kojubu, D. A.; Fletcher, B. S.; Varnum, B. C.; Lim, R. W.; Herschman, H. R.;

"TIS10, a Phorbol Ester Tumor Promoter-Inducible mRNA from Swiss 3T3 Cells,

Endcodes a Novel Prostaglandin Synthase/Cyclo-Oxygenase Homolog," J. Biol.

Chem. 1991, 266.

34. Xie, W.; Robertson, D. L.; Sunmons, D. L.; "Mitogen-Inducible Prostaglandin G/H

Synthase: A New Target for Non-Steroidal Anti-Iinflammatory Drugs," Drug Dev.

Res. 1992, 25, 249.

35. Donnelly, M. T.; Hawkey, C. J.; "Review Article: Cox-11 Inhibitors - A New

Generation of Safer NSAIDs?," Alimentary Pharmacology and Therapeutics 1997,

11, 227.

36. de Brum Fernandes, A. J.; "New Perspectives for Non-Steroidal Anti-Inflammatory

Therapy," Journal of Rheumatology 1997, 24, 246.

CHAPTER 1

16

37. Wallace, J. L.; "Non-Steroidal Anti-Inflammatory Drugs and Gastro-Eneropathy:

The Second Hundred Years," Gastroenterology 1997, 112, 1000.

38. Walker, W. R.; Keats, D. M.; "An Investigation of the Therapeutic Value of the

Copper Braclet. Demal Assimilation of Copper in Arthritis/Rheumatoid Condition,"

Agents and Actions 1976, 6/4, 454.

39. Sorenson, J. R. J.; "Copper Chelates as Possible Active Forms of the Arthritic

Agents," Journal of Medicinal Chemistry 1976, 19, 135.

40. Sorenson, R. J.; "Copper Complexes Offer a Physiological Approach to Treatment

of Chronic Diseases," Progress in Medicinal Chemistry, ; Ellis, G. P. and West, G.

B., Ed.; Elsevier Science Publishers, B.V. (Biomedical Division): New York, 1989;

Vol. 26, 437.

41. Walker, W. R.; Beveridge, S. J.; Whitehouse, M. W.; "Anti-inflammatory Activity

of a Dermally Applied Copper salicylate Preparation (Alcusal)," Agents and Actions

1980, 10, 38.

42. Beveridge, S. J.; Walker, W. R.; Whitehouse, M. W.; "Anti-Inflammatory Activity

of Copper Aslicylates Applied to Rats Percutaneously in Dimethylsulfoxide with

Glycerol," J. Pharm. Pharmacol. 1980, 32, 425.

43. "MIMS Australia, IVS Annual," Index of Veterinary Specialists Annual, ; MIMS

Publishing, Crows Nest, N.S.W.: Sydney, 1997, 145 & 276.

CHAPTER 2

17

2. DENSE GASES AS ANTI-SOLVENTS — THEORY

2.1. Introduction

The concept of using supercritical fluids to precipitate solids has been known for over a

century.1 The concept of precipitating solids from solutions using dense gases as anti-

solvents however, is a relatively recent development, with the first experiments being

reported by Gallagher and co-workers2 The ability of dense gases such as carbon

dioxide to dissolve in and expand organic solvents was exploited in order to precipitate

solids from organic solution. The process reported by Gallagher and co-workers has

now become known as the gas anti-solvent process and has sparked off a whole new

direction in research. The number of recent publications relating to the potential of

using dense gases as anti-solvents is evidence of this.3-9

Two experimental techniques have been designed to take advantage of the gas anti-

solvent procedure. The first type of process, known as the Gas Anti-Solvent (GAS)

process, involves the addition of the gas anti-solvent to a solution containing a solute at

a certain temperature. As the pressure is increased, the gas anti-solvent dissolves in the

solution, reducing the dissolving power of the solvent and precipitating the solute. The

second type of process involves the injection of a solution into a precipitation chamber

that has been pressurised with gas anti-solvent. The second process has been referred to

by many different names depending on the conditions under which the gas anti-solvent

is being used. The more common names are Precipitation with a Compressed Anti-

Solvent (PCA),10 Supercritical Anti-Solvent precipitation (SAS),11 Solution Enhanced

Dispersion by Supercritical fluids (SEDS),12 and the Aerosol Solvent Extraction System

(ASES). In this text the acronym ASES will be used in reference to the latter process.

For both processes precipitation occurs in an extremely short time span and precipitated

particles have narrow particle size distribution.

The GAS process is a batch precipitation technique that utilises near critical, or

supercritical fluid anti-solvents to precipitate solutes from solution. By definition the

process relies on the solute being relatively insoluble in the anti-solvent, but the anti-

CHAPTER 2

18

solvent being miscible with the solvent. As the concentration of anti-solvent mixed into

the liquid solution increases, the liquid solvent expands and the solvent power of that

phase decreases. At a critical anti-solvent concentration, or expanded volume, the

solution becomes saturated and, at higher anti-solvent concentrations, the solute

precipitates. The precipitate can then be filtered out and washed with pressurised anti-

solvent to prevent re-dissolution into the solvent. The nature of the final precipitate in

this process is simply a function of the pressurisation rate, temperature and the type of

solvent and anti-solvent used.

The Aerosol Solvent Extraction System (ASES) process is a continuous precipitation

technique that, like the GAS process, utilises near-critical, or supercritical fluid anti-

solvents to precipitate solutes from solution. The main difference between the GAS and

ASES processes is that in the latter the solution containing the solute/s is sprayed into

an anti-solvent environment, allowing the solution to expand and induce precipitation of

the solute. The ASES process causes a dramatic drop in the solvation power of the

solvent and therefore the supersaturation at which precipitation occurs can be

exceptionally high. Particles with diameters less than 1 µm can be precipitated in this

manner. The efficiency of ASES precipitation is further enhanced by the use of spraying

devices such as fine nozzles and orifices that are able to uniformly disperse the solution

as fine droplets with large surface area to improve the diffusion processes involved in

the precipitation. The time scale for the ASES precipitation process has been measured

to be within 10-5s.13

The work presented in this thesis involved the use of both the GAS and the ASES

process using CO2 as the anti-solvent. To provide a framework of understanding,

theoretical aspects such as thermodynamics, hydrodynamics, mass transfer and

nucleation and crystallisation will be presented in relation to the GAS and ASES

processes. It must be kept in mind that, in terms of industrial processes, the GAS and

ASES concepts are in their infancy and, as such, the fundamental data required for a

rigorous description of the processes is unavailable.

CHAPTER 2

19

2.2. Thermodynamic Considerations

The ability of a dense gas to behave as an anti-solvent depends on its ability to dissolve

in the solution containing the solute to be precipitated. An understanding of the phase

behaviour is therefore essential to performing a successful precipitation using the GAS

or ASES processes.

The following discussion will focus on the phase behaviour of systems relevant to the

GAS and ASES processes. The phase behaviour of the pure anti-solvent will be

presented first and the sections that follow will discuss the more complex systems that

contain solvent and one or more solutes. The various models that have been reported in

the literature to predict the phase behaviour of these systems will also be discussed.

2.2.1. Pure Anti-Solvent Phase Behaviour

Essentially any gas can be used as an anti-solvent in the GAS or ASES processes with

the only requirement being that the gas must be sufficiently soluble in the solution of

interest at the desired pressure and temperature. Carbon dioxide, however, is by far the

most commonly used anti-solvent due to its inert nature, low cost, availability and mild

operating conditions.

The phase diagram of a pure component is depicted in Figure 2.1. The three curves

represented by a solid line indicate the co-existence of two phases and are boundaries

for the single phase regions. Line A—B is referred to as the sublimation curve and

indicates the region where the gas and solid phases exist in equilibrium. Line B—C is

the fusion curve and indicates the region where solid and liquid phases exist in

equilibrium. Line B—D is the vapourisation curve and indicates the region where gas

and liquid phases exist in equilibrium. Point D, called the critical point, marks the end

of the vapour—liquid equilibrium curve. At pressures and temperatures above the

critical point, indicated by the shaded area, the gas and liquid phases become identical

and the system is said to be supercritical.

CHAPTER 2

20

��

Supercritical FluidRegion

Liquid

Gas

SolidPres

sure

Temperature

A

B

C

D

Figure 2.1 Phase diagram for a pure substance.

CHAPTER 2

21

The critical pressures and temperatures of a number of compounds are listed in Table

2.1. A trend that can be noted is that most hydrocarbons have critical pressures close to

5.0 MPa. The critical temperatures for the light hydrocarbons are around room

temperature and carbon dioxide has a mild critical temperature and slightly elevated

critical pressure.14

As a fluid approaches its critical point, unusual effects are observed. The isothermal

compressibility and the heat capacity (at constant pressure) of the fluid become

infinitely large. The density and the dielectric constant of the fluid increases sharply as

the pressure increases through the critical point. The density behaviour of a pure

component in the vicinity of the critical point is depicted in Figure 2.2.14

Table 2.1 Critical pressures and temperatures of anti-solvents used in GAS and

ASES.14

Anti-Solvent Critical Temperature / °C Critical Pressure / MPa

Carbon Dioxide 31 7.1

Ethane 32 4.8

Ethylene 10 5.1

Propane 96.7 4.19

Propylene 91.9 4.56

Ammonia 132.5 11.13

Water 374.2 21.76

Benzene 289.0 4.83

Trifluoromethane 26 4.7

Chlorotrifluoromethane 29 3.9

Xenon 16 5.8

Nitrous Oxide 36 7.2

Isopropanol 235.2 4.7

At reduced temperatures (T/Tc) between 0.9 and 1.2 and reduced pressures greater than

1.0 the reduced density of the fluid changes from gas-like to liquid-like. The slope of

CHAPTER 2

22

Tr = 0.8

0.9

1.0

1.1

1.2

1.55

1.0

1.0

Red

uced

Den

sity

Reduced Pressure

Figure 2.2 Reduced density of a pure substance in the vicinity of its critical point.14

CHAPTER 2

23

the isotherms (the compressibilities) are very high, close to the critical point, implying

that very small changes in pressure can induce large changes in density. It is this density

characteristic of fluids near their critical point that has made them attractive as solvents

for extracting components from mixtures.

Another important feature of a supercritical fluid is that although the densities approach

those of liquids the transport properties such as diffusivity and viscosity remain gas-

like. The self-diffusivity of carbon dioxide with those of solutes in organic liquids is

compared in Figure 2.3.14 The self-diffusivity of CO2 is about 1-2 orders of magnitude

higher than the diffusivity of solutes in liquids.

The change in viscosity of CO2 with pressure at various temperatures is shown in Figure

2.4. The viscosity increases rapidly around the critical point, but even at high pressures,

the values are around 0.09 cp, which is an order of magnitude below the typical

viscosities of liquid organic solvents.14

The combined properties of liquid-like densities, gas-like diffusivities and viscosities

and zero surface tension of fluids such as CO2 near their critical point have been the

catalyst for the research conducted with supercritical fluids.

2.2.2. Anti-Solvent—Solvent Phase Behaviour

The ability of gases such as CO2 to dissolve in organic solvents has been

demonstrated15 long before the introduction of the use of dense gases as anti-solvents.2

For example the equilibrium phase properties of the toluene—CO2 system were

reported 20 years earlier.15 However, since the report by Gallagher and co-workers,2

studies specific to the GAS and ASES systems have appeared.

CHAPTER 2

24

Diff

usiv

ity /

cm2.s-1

Temperature / oC0 20 40 60 80 100

10-5

10-4

10-3

10-2

��Diffusivities of solutes in Normal Liquids

Pressure (MPa)78

101520

Saturated Vapour

Critical Point

Saturated Liquid

Figure 2.3 Diffusivity behaviour of CO2.14

CHAPTER 2

25

Visc

osity

/ cp

Pressure / MPa4.0 10.0 100.0

0.01

0.07

0.09

0.11

0.03

0.05

77 oC47 oC

37 oC

Figure 2.4 Viscosity behaviour of CO2.14

CHAPTER 2

26

Kordikowski and co-workers16 have reported the results on the pressurisation of a

number of organic solvents with CO2, ethane and ethene. The solubility of the anti-

solvents in the solvents as a function of pressure and temperature are depicted in Figure

2.5 and Figure 2.6.

The results reported by Kordikowski and co-workers showed that the selection of anti-

solvent for a particular solution depends on the degree of polarity of the solvent. The

pressurisation of solutions of polar solvents such as acetonitrile with ethane or ethene

resulted in liquid—liquid immiscibility. Liquid—Liquid immiscibility did not occur

when acetonitrile was pressurised with CO2. Similar results were observed for the

solvents N-methyl pyrrolidone (NMP) and N,N-dimethylformamide (DMF) where the

formation of a second liquid phase occurred upon pressurisation with ethylene and

ethane. Liquid—liquid immiscibility did not occur when DMF and NMP were

pressurised with CO2, chlorodifluoromethane, or dichlorodifluoromethane.2 The

observations were correlated to the differences in polarity of the solvents and anti-

solvents. Ethane, ethene and ethylene are non-polar compounds, whereas DMF, NMP

and acetonitrile are polar molecules. Interestingly, the possibility of CO2 dissolving a

solute is often correlated to the ability of hexane to dissolve the solute. In the case of

these systems, CO2 was able to dissolve in the polar solvents that are immiscible with

hexane.2 The ability of CO2 to dissolve in polar solvents has been attributed to the

relatively large quadropole moment of the molecule, which results in the existence of a

small polarity.16

For the GAS and ASES processes the occurrence of immiscibility is undesirable as the

processes of precipitation of the solute and the subsequent removal of solvent relies on

the anti-solvent being miscible with the solvent. It is therefore essential that the phase

behaviour of the binary anti-solvent—solvent system be investigated. Although

immiscibility may occur for a particular system at certain conditions, the system may

still be operable if the process parameters are manipulated such that the region of

immiscibility is avoided.

CHAPTER 2

27

0

0.1

0.2

0.3

0.4

0.5

0.6

0.7

0.8

0.9

0 5.0 10.0 15.0 20.0 25.0Pressure / MPa

Mol

e Fr

actio

n (A

ntis

olve

nt)

Ethane—AcetonitrileEthane—Ethyl AcetateEthane—1,4-Dioxane

0

0.1

0.2

0.3

0.4

0.5

0.6

0.7

0.8

0.9

1

0 2.0 4.0 6.0 8.0Pressure / MPa

Mol

e Fr

actio

n (A

ntis

olve

nt) Ethene—Ethyl Acetate

Ethene — 1,4-Dioxane

Figure 2.5 Solubility of ethane and ethene in organic solvents.16

CHAPTER 2

28

0

0.1

0.2

0.3

0.4

0.5

0.6

0.7

0.8

0.9

1

0 2 4 6 8Pressure / MPa

Mol

e Fr

actio

n (A

ntiso

lven

t)

Acetonitrile

Ethyl Acetate

1,4-DioxaneEthanol

Figure 2.6 Solubility of CO2 in organic solvents.16

CHAPTER 2

29

In describing the phase behaviour of the anti-solvent—solvent system, the expansion of

the liquid phase has often been used. The classical definition of volumetric expansion

(∆V) was given by the following equation.2

∆VV

= VL (T,P,x1)-V2 (T,P0)

V2 (T,P0) 2.1

where VL is the total volume of the liquid phase and V2 is the total volume of the pure

solvent at the same temperature and atmospheric pressure P0, and x1 is the mole fraction

of the anti-solvent in the liquid phase. At low pressures the expansion of the liquid

phase is low and follows Henry's law.2 The volumetric expansion of ethylacetate upon

pressurisation with CO2 is depicted in Figure 2.7.16

As the pressure approached the vapour pressure of the anti-solvent, the expansion

became asymptotic and eventually the solvent and anti-solvent became totally miscible.

The point at which the two fluids become miscible is termed the threshold pressure, or

critical point, of the mixture. Below the point of mutual miscibility, the solvent has little

solubility in the anti-solvent.

The density of the liquid phase of various organic solvents upon pressurisation with

CO2 is shown in Figure 2.8. The density passes through a maximum as the pressure

increases. The increase in density indicates that at low pressures CO2 could act as a

solvent rather than an anti-solvent.16

It has been shown that for a given anti-solvent mole fraction, the volumetric expansion

was the same for all temperatures and organic solvents studied (Figure 2.9).16 The

inability to distinguish between solvents using the classical definition of expansion has

led to the following definition being proposed for solvent expansion.17 :

CHAPTER 2

30

0

100

200

300

400

500

600

700

800

900

0 2.0 4.0 6.0Pressure / MPa

∆V

/ %

25 °C

30 °C

40 °C

Figure 2.7 Volumetric expansion of ethyl acetate pressurised with CO2.16

CHAPTER 2

31

700

750

800

850

900

950

1000

1050

1100

0 1 2 3 4 5 6Pressure / MPa

Den

sity

/ kg.

m-3

AcetonitrileEthyl Acetate1,4-DioxaneEthanol

Figure 2.8 Variation of density of organic solvents upon pressurisation with CO2.16

CHAPTER 2

32

∆V*

V* = V% L (T,P,x1) — V% 2 (T,P0)

V% 2 (T,P0) 2.2

where V% L is the molar volume of the liquid mixture and V% 2 is the molar volume of the

pure solvent. The expansion behaviour of a number of solvents pressurised with CO2 is

shown in Figure 2.10.

When Equation 2.2 is used to describe the expansion, different solvents show different

expansion behaviour. These differences include ∆V* passing through a minimum with

increasing mole fraction of anti-solvent. The precipitation of solute was found to

coincide with the minimum in ∆V*. It was therfore proposed that the optimum

expansion for a given solvent and anti-solvent was the pressure where the volumetric

expansion passed through a minimum.18

In some anti-solvent processes a mixture of solvents are used in order to increase the

solubility of a specific solute or for co-precipitation purposes. Hydrophilic compounds

that are sparingly soluble in organic solvent, such as proteins, require water to be used

as a solvent. Common anti-solvents with moderate critical temperatures are insoluble in

water and as such are unable to act as an anti-solvent. In these cases an organic solvent

that is miscible in both the anti-solvent and water may be added to improve the

miscibility of water and the anti-solvent.19-21

The three component CO2—water—ethanol systems at 35°C and at 6.9 MPa and 10.2

MPa respectively are shown in Figure 2.11.5 At the lower pressure CO2 and ethanol

exists in two phases and is seen by the V—L boundary intersecting the CO2—ethanol

axis. At both pressures water and CO2 exist in two phases and water and ethanol are

miscible. The mixture exists in two phases and is bounded by the saturated liquid and

vapour curves. As the pressure is increased above the critical pressure of the binary

CO2—ethanol system, CO2 and ethanol become totally miscible.

CHAPTER 2

33

0

200

400

600

800

1000

1200

0 0.2 0.4 0.6 0.8 1CO2 Mole Fraction

∆V /

%

Acetonitrile

Ethylacetate

1,4-Dioxane

Ethanol

Figure 2.9 ∆V vs CO2 mole fraction for various organic solvents.16

CHAPTER 2

34

15

5

-5

-15

-25

-35

-45

-55

∆V

* / %

CO2 Mole Fraction0 0.2 0.4 0.6 0.8 1

Acetonitrile

Ethanol

Ethyl Acetate1,4-Dioxane

Figure 2.10 ∆V* vs CO2 mole fraction for various organic solvents.17

CHAPTER 2

35

If the system shown in Figure 2.11 were to be used in the GAS or ASES process the

conditions chosen would need to be where all three components are miscible. The

conditions where all three components are miscible are shown by the shaded area in

Figure 2.11.

2.2.3. Anti-Solvent—Solvent—Solute Phase Behaviour

The pure anti-solvent and binary anti-solvent—solvent phase behaviour, although

important in determining the suitability of a system for application of the GAS or ASES

processes, can be far removed from the actual phase behaviour of a ternary anti-

solvent—solvent—solute system. To date, no study has presented the complete phase

behaviour of an organic solid in the presence of an organic solvent expanded with a

supercritical fluid.5 A simple representation of the phase behaviour of a ternary system

at a pressure and temperature above and below the anti-solvent—solvent critical point is

shown in Figure 2.12 a and b. Below the critical point of the anti-solvent—solvent

mixture (Figure 2.12 a) the following phases can co-exist: liquid, vapour—liquid,

liquid—solid, vapour—liquid—solid, vapour—solid and vapour. Above the critical

point of the anti-solvent—solvent mixture (Figure 2.12 b) the number of co-existing

phases reduces to vapour, vapour—solid, liquid and liquid—solid.

If the concentration of solute in solution is relatively low, the ternary system displays

similar phase behaviour to the binary anti-solvent—solvent system. At higher

concentrations of solute in solution, in many instances, the ternary system cannot be

related to the phase behaviour of the three corresponding binary systems.18 At higher

concentrations, the presence of solute in the expanded solution can result in complex

phase behaviour.17,18,22,23 For example the CO2—toluene—naphthalene system gives a

liquid phase split upon pressurisation24 and an ethanol solution of citric acid does not

undergo expansion when pressurised with CO2.25 Tai and co-workers25 suggested that

this was due to the strong cohesive forces between the solute and solvent preventing the

solvent molecules from interacting with the anti-solvent molecules.

CHAPTER 2

36

��

��

Ethanol

CO2Water

L + V

L

6.9 MPa

V

a

Ethanol

CO2Water

L + V

L

10.2 MPa

b

Figure 2.11 Ternary phase diagram for CO2—water—ethanol at constant

temperature and pressure.5 (a) 35°C, 6.9 MPa, (b) 35°C, 10.2 MPa

CHAPTER 2

37

Solid

CO2Organic

L

L + S

V + S V

b

Solid

CO2Organic

V + L + S

V

L + S

L

V + L

V + S

a

Figure 2.12 Ternary anti-solvent—solvent—solute phase behaviour.5

CHAPTER 2

38

The phase behaviour of the CO2—toluene—naphthalene system has been determined by

means of a thermodynamic model at pressures and temperatures of relevance to the

GAS and ASES processes.22 An example of the complex phase behaviour possible

upon the addition of a solute to the anti-solvent—solvent system is shown in Figure

2.13. A second liquid phase appears and, depending on the overall composition, either

L1—L2—V or S—L2—V equilibrium may occur.

Polymers are a class of compounds that have complex phase behaviour when

pressurised with anti-solvents such as CO2. Liquid—liquid phase separation can occur

before the polymer vitrifies.8 Dissolved gases have a plasticising effect on amorphous

polymers due to dissolution of the gas in the polymer, resulting in a lowering of the

glass transition temperature. This lowering of glass transition temperature can go below

that of operating temperatures of GAS and ASES26 and can influence the morphology

and particle size of the particles produced.

The solubility of solute in an expanded solution at the operating conditions also needs to

be considered. Depending on the concentration of the solute in solution and the nature

of the solute, precipitation will occur at different expansion levels.24,27 520 This is the

basis of fractionation and can play a significant role in the GAS and ASES processes.

As the concentration of anti-solvent increases in the expanded solution, a solute will

tend to its solubility in pure anti-solvent.8 If the solute to be precipitated is too soluble

at the final conditions, or if the process is performed at a condition where the expansion

of the solvent is below the precipitation threshold of the solute, then no precipitation

will occur. In some studies microparticle formation has been unsuccessful due to the

solubility of the drug at process conditions28-30 and encapsulation experiments have

failed because of the drug being extracted instead of precipitated.31

The solubility of the substance to be precipitated in a suitable organic solvent is another

thermodynamic criteria. For most organic solids this criterion is met, but for most

inorganic and hydrophilic materials the situation is different. They are usually soluble in

water or sulfuric acid which are not expanded by simple gases such as CO2.32 A

CHAPTER 2

39

Naphthalene (50 %)

Toluene (50 %)CO2

L1 + L2 + V

S + L

LL1 + VV

S + V

L2 + V

S + L2 + V

Figure 2.13 Ternary phase diagram for CO2—toluene—naphthalene at

temperatures between 315 and 325 K and pressure between 8.2 and 9.8 MPa.22

CHAPTER 2

40

technique called hydrophobic ion pairing can be used to increase the solubility of a

hydrophilic solute in an organic solvent to a level that is practical for ASES.31,33,34

This involves pairing the counter ion of the drug with a hydrophobic counter ion. Care

must be taken that the additive used does not increase the solubility of the solute in the

anti-solvent to an extent that it is extracted rather than precipitated. The addition of an

amphiphilic material has also been used to increase the solubility of hydrophilic

molecules such as proteins in organic solvents.35 An alternative method of precipitating

hydrophilic solutes with CO2 is known as Solution Enhanced Dispersion (SED). In the

SEDs process, aqueous solutions are injected into a CO2 phase that has been modified

with an organic solvent that is miscible in both CO2 and water, to improve the

miscibility of water and CO2.19,20

2.2.4. Modelling GAS and ASES Phase Behaviour

2.2.4.1. Modelling Anti-Solvent—Solvent Phase Behaviour

Often the phase behaviour data is unavailable for a particular anti-solvent—solvent

system, or the data does not extend into the region desired. It is also often not

convenient to determine the phase behaviour experimentally. The phase behaviour can

be predicted using an equation of state model such as the Peng-Robinson equation of

state (PREOS).36 The PREOS has been demonstrated to give good correlation with

experimental data for a number of anti-solvent—solvent systems.16

The PREOS is as follows:

P = RT

υ - b —

a(T)υ (υ + b)+b (υ - b)

2.3

where P is the system pressure, R is the gas constant, T is the system temperature, a is

the attraction parameter, b is the van der Waals co-volume and υ is the molar volume.

Equation 2.3 can be expressed as

Z3 - (1 - B) Z2 + (A - 3B2 - 2B) Z - (AB - B2 - B3) = 0 2.4

CHAPTER 2

41

where

A = aP

R2T2 2.5

B = bPRT

2.6

Z = PυRT

2.7

Equation 2.4 yields one to three roots depending on the number of phases in the system.

For a two phase system, the largest value of Z corresponds to the vapour phase and the

smallest positive root to that of the liquid phase.

At the critical point:

b(Tc) = 0.07780 RTc

Pc 2.8

a(Tc) = 0.45724 R2T2

c

Pc 2.9

where Tc is the critical temperature and Pc is the critical pressure of the component in

question. At temperatures other than the critical point:

a(T) = a(Tc) α(Tr,ω) 2.10

b(Tc) = b(T) 2.11

where α(Tr,ω) is a dimensionless function of reduced temperature and the accentric

factor ω. Expressions for k and α are as follows:

CHAPTER 2

42

k = 0.37464 + 1.54226ω - 0.26992ω2 2.12

α = [1 + k

1 - Tr ]2 2.13

a(T) = 0.45724 R2T2

c

Pc

1 + (0.37464 + 1.54226ω - 0.26992ω2)

1 - Tr

2 2.14

The mixture parameter a is defined as follows:

a = ∑i

∑j

xixjaij 2.15

aij = (1 - kij) aiaj 2.16

where xi and yi are the mole fractions in the liquid and vapour phase of component i. kij

is the unlike pair interaction parameter which accounts for non-ideal interactions

between unlike molecules. For non-polar solvents kij is sufficient to model the data and

the mixture parameter b is defined as follows:

b = ∑i

xibi 2.17

Most of the solvents used in the GAS or ASES processes are polar and two binary

interaction parameters, kij and lij, are needed to model the data accurately.16 The

expression for the mixture parameter b then becomes :

b = ∑i

∑j

xixjbij 2.18

bij = (1 - lij) bi + bj

2 2.19

For a binary system the following iso-fugacity criteria hold:

CHAPTER 2

43

y1

x1 =

φV1

φL1

2.20

y2

x2 =

φV2

φL2

2.21

where φLi and φ

Vi are the liquid and vapour phase fugacities of component i respectively.

The following constraints on the two fluid phases applies:

∑i = 1

2 xi = 1 2.22

∑i = 1

2 yi = 1 2.23

If the pressure and temperature are fixed Equations 2.20 to 2.23 represent four non-

linear equations and four unknowns, which are the molar compositions of the liquid and

vapour phases. The fugacity co-efficients of the liquid and vapour phase constituents

can be calculated using the following equation:

ln φi = bNB

(Z - 1) - ln (Z - B) — A

2 2B

2∑

j xjaij

a ln

Z + 2.414B

Z - 0.414B 2.24

bN = 2∑k

N xkbik — ∑

j

Nx j

2bjj — 2∑j=1

N-1 ∑

i=j+1

N xixi-jbij 2.25

The PREOS has been applied to a number of anti-solvent—solvent systems.16 Using

temperature independent binary interaction parameters, good correlation with

experimenatal data were obtained for most of the anti-solvent—solvent systems studied.

CHAPTER 2

44

The PREOS was unable to model the systems such as ethane—acetonitrile in which

liquid—liquid immiscibility occurred.

2.2.4.2. Modelling Anti-Solvent—Solvent—Solute Phase Behaviour

A few models have been proposed to predict the solubility of solids in expanded

solutions.22,24,37,38 For a ternary anti-solvent—solvent—solute system, the following

isofugacity equation can be added to those given by Equations 2.22 and 2.23:

y3

x3 =

φV3

φL3

2.26

with the following constraints on the two fluid phases:

∑i=1

3 xi = 1 2.27

∑i=1

3 yi = 1 2.28

The subscripts 1,2 and 3 refer to the anti-solvent, solvent and solute respectively. In

most cases the PREOS, with one or two interaction parameters, has been used to

calculate the fugacity co-efficients of each component in the liquid and vapour phase.

The PREOS is unable to represent the behaviour of the solid phase and it is in the

description of the solid phase that the following models differ.

Chang and Randolph used the decrease in the partial molar volume (υ2) of the solvent

upon expansion to calculate the solubility of the solute in the expanded solvent.37 Only

Equations 2.22 and 2.23 are considered in the model. The partial molar volumes were

calculated using the following analytical differentiation of the PREOS:

CHAPTER 2

45

υi = RTP

Z + ( )1 - xi

δZ

δxi T,P i = 1,2 2.29

The solubility of the solute in the liquid phase was calculated in terms of the partial

molar volumes of the solvent using the following expression:

S3(T,P) = υ2 (T,P,x) υ2 (T,1,0)

S3(T,1) 2.30

where S3(T,P) is the equilibrium solubility of the solute in the liquid phase, S3(T,1) is

the solubility of the solute in the solvent at atmospheric pressure and υ2(T,1,0) is the

partial molar volume of the pure solvent at atmospheric pressure. Using this approach

only the binary data was needed to predict the ternary phase behaviour, and a good

correlation between experimental and predicted results were obtained for the CO2—

toluene—β-carotene and CO2—n-butanol—acetominophen systems at 25 °C up to

pressures of 9.0 MPa. The model is based on the assumption that the anti-solvent—

solute interactions are negligible compared to the anti-solvent—solvent interactions and

only provides results based upon the rough approximation of Equation 2.30.11 Despite

the approximations, the model may be useful when data such as critical constants are

unavailable for the solid.

Other models that have been proposed make use of all three iso-fugacity Equations:

2.22, 2.23 and 2.26. The difference in these models is the manner in which the

equilibrium between the solute in the solid and liquid phases is dealt. As all the models

treat the vapour—liquid equilibrium in the same way, only the manner in which the

solid—liquid equilibrium is dealt with will be described below.

Dixon and Johnston used a combination of regular solution theory and the PREOS to

describe the ternary phase behaviour of the CO2—toluene—phenanthrene and CO2—

toluene—naphthalene systems.38 The solid—liquid equilibrium was described by the

following relationship:

CHAPTER 2

46

f3 0S

= x3γ

3 f

3 0L

2.31

where f3 0S

and f3 0L

represent the fugacities of the pure solid solute and the pure subcooled

liquid solute respectively, and γ3 is the solute activity co-efficient. At the system

temperature and at a reference pressure (P0) the fugacity of the solute in the liquid phase

was given by:

f3 L = x

3γ

3 (P

0,{x

i

0}) f

3 0L

φ3 (P,{x

i})P

φ3 (P0,{x

i

0})P

0 2.32

where {xi

0} is the set of mole fractions at the reference pressure. The fugacity of the

solid at pressure P was given by:

f3 S = f

3 0S

exp

⌡⌠

P0

P

υ3

S

RT dP 2.33

where υ3

S is the solid molar volume. By inserting Equations 2.32 and 2.33 into 2.31 the

following expression was obtained:

f3 0S

exp

⌡⌠

P0

P

υ3

S

RT dP = x

3γ

3 (P

0,{x

i

0}) f

3 0L

φ3 (P,{x

i})P

φ3 (P0,{x

i

0})P

0 2.34

The assumptions that the solid phase contained pure solute and that the mole fraction of

solute in the gas phase was negligible were made using this model. The latter

assumption reduces 2.26 to zero and only Equations 2.22, 2.23 and 2.34 are solved to

calculate the phase behaviour of the ternary system. Good correlations between the

calculated and experimental data were obtained at 25°C and pressures up to 7 MPa

using only binary interaction parameters. 38

CHAPTER 2

47

Another description of the solid—liquid equilibrium has been given by Kikic and co-

workers.11,22 The assumption was made that the solid phase is pure and the equilibrium

equation for the solute between the solid and liquid phases was given by:

f3 0S

= x3 φ L 3 P 2.35

The fugacity of the solid state was calculated using an equation developed by Prausnitz

and co-workers39 which relates the solid state to a fictitious liquid state. The equation is

strictly valid at the triple point pressure of the solute and is of the following form:

ln f

3 0S

f3 0L =

∆h3 f

RT3 f

1 - T

3 f

T 2.36

where f3 0S

and f3 0L

are the fugacities of the solute in the solid and subcooled liquid phase

and ∆h3 f and T

3 f are the heat of fusion and the melting temperature of the solute. The

effect of pressure on the fugacities is taken into account using the following two

expressions:

f3 0S

= f3

0S(T,P0) exp

⌡⌠

P0

P

υ3 0S

RT dP 2.37

f3 0L

= f3 0L

(T,P0) exp⌡⌠

P0

P

υ3

0L

RT dP 2.38

The ratio of Equations 2.37 and 2.38 gives:

f

3 0S

f3 0L =

f3 0S

f3 0L (T,P0) exp

⌡⌠

P0

P

υ3

0S - υ

3 0L

RT dP 2.39

CHAPTER 2

48

Inserting Equation 2.36 into Equation 2.39 gives:

f3 0S

= f3

0L(T,P) exp

⌡⌠

P0

P

υ3 0S

- υ3 0L

RT dP +

∆h3 f

RT3 f

1 - T

3 f

T 2.40

The difference between the solid and liquid molar volumes is usually negligible and the

first addendum to the exponential term in Equation 2.40 can be dropped without loss of

accuracy. Good correlation between the experimental and calculated results for the

CO2—toluene—phenanthrene, CO2—toluene—naphthalene and CO2—toluene—β-

carotene were obtained at a temperature of 25 °C and pressures up to 7.0 MPa using this

model.

De la Fuenta and co-workers17 have proposed an alternative equation to calculate the

fugacity of the solid phase. The fugacity of the solid phase was given by:

ln φ3 = ln

PSubl

P —

VT,P

RT ( )P-PSubl 2.41

where TT,P, PT,P and VT,P are the triple point temperature, triple point pressure and molar

volume of the solute at the triple point respectively. PSubl is given by:

PSubl = PT,P exp

∆HSubl

T,P

R

1

TT,P -

1T

2.42

where ∆HSublT,P is the heat of sublimation.

2.3. Crystallisation Theory Considerations

In spite of the fact that the crystallisation of solids is one of the oldest chemical

processes, the process is not fully understood and has been referred to as an art rather

CHAPTER 2

49

than a science.40,41 Part of the reason for this is that crystallisation is a complex

process.42 As a consequence, the theoretical foundations of the crystallisation process

are not well developed and only yield acceptable results when applied to systems that

differ only slightly from ideality.42

Crystallisation from solution can be divided into three processes: supersaturation,

nucleation and crystal growth. The interaction between these processes is important in

determining the particle size and particle size distribution of the precipitate collected

from the crystallisation vessel. The aim of this section is to briefly review the

application of crystallisation theory to the GAS and ASES processes. For a detailed

description of the theoretical and practical aspects of crystallisation the following

references can be used as a starting point.40-42

Supersaturation is often quoted as being the dominant factor that determines the size

and shape of particles in the GAS and ASES processes.8 Conventionally supersaturation

in solutions is achieved by cooling, evaporation of solvent, chemical reaction, or by the

introduction of a second solute. The latter method is often referred to as salting out

precipitation. In the GAS and ASES processes, supersaturation is achieved by the

dissolution of gas anti-solvent into the solution. In such systems pressure, or rate of

diffusion of anti-solvent into the solution, becomes a unique variable that determines the

concentration of anti-solvent.

Expressions that are commonly quoted to express supersaturation in the GAS or ASES

process are the supersaturation ratio :

S = CC* 2.43

and the concentration driving force:

∆C = C - C* 2.44

CHAPTER 2

50

where C is the solution concentration at a certain pressure and C* is the equilibrium

saturation at atmospheric pressure.

The importance of supersaturation in the GAS and ASES process has been qualitatively

demonstrated using equations derived from the thermodynamic approach of Becker and

Doring43,44 and McCabe and Smith45 for the rate of homogenous nucleation.2,7,32,46,47

Equations of the following type have been quoted:

J = Z exp

∆Gmax

RT 2.45

where J is the rate of production of nuclei, Z is the collision frequency and ∆Gmax is the

Gibbs free energy expression. ∆Gmax is given by

∆Gmax = B

(RT lnS)2 2.46

where B is a constant which includes the physical properties of the system, and S is the

supersaturation ratio.

The growth of nuclei from the GAS process has been related to the mass transfer

correlation.2,32,46:

flux of material to the surface = kA ∆C 2.47

where k is the mass transfer co-efficient, A is the surface area of the particle at any

particular instant and ∆C is the concentration driving force at any particular instant

given by the concentration of the solute in the expanded solution minus its equilibrium

concentration.

Equations 2.45, 2.46 and 2.47 have not been used for quantitative purposes with

Equations 2.45 and 2.46 being strictly valid for the homogenous condensation of a

vapour to liquid droplets. However, these Equations have some qualitative value in

CHAPTER 2

51

determining the relevant experimental parameters governing crystal formation in the

GAS and ASES systems.32

Equations 2.45, 2.46 and 2.47 show that the important parameters in the GAS process

that determine particle size and morphology are: initial concentration of the solution,

operating temperature and pressure, rate of creation of supersaturation as determined by

the rate of addition of anti-solvent, maximum level of supersaturation as determined by

the rate of addition of anti-solvent and the growth period determined by the amount of

solute remaining after primary nucleation.

The precipitate characteristics in terms of particle size and size distribution are

determined by the interaction between the nucleation and growth rates and the rate of

creation of supersaturation.2 In the GAS process all three are influenced by the rate of

addition of anti-solvent. The relationship between solution concentration and solvent

expansion is shown in Figure 2.14. The rate of nucleation and growth of precipitate is

also shown in relation to solution expansion. The graph is divided into two regions;

solutions located below the solid line are stable unsaturated solutions (S < 1), and

solutions above the solid line are saturated (S ≥ 1). Saturated solutions may be described

as either metastable, where particle growth dominates (below the dashed line) or

unstable where nucleation is most rapid (above the dashed line), that is the shaded

region in Figure 2.14. The solid line, S = 1, is the binodal curve, at which point

dissolution of a solid occurs when depressurising an expanded solution at an infinitely

slow rate. The dashed line, S = Smax, represents the spinodal curve.

If an anti-solvent is injected into a solution, as in the GAS process, such as that defined

by the point A in Figure 2.14 a, the solution expands and the solvation power of the

solution dramatically decreases. When enough carbon dioxide has been dissolved, the

solution becomes saturated, S =1 (point B). Further expansion causes supersaturation

and in the vicinity of the unstable region (point C), S =Smax, nucleation results. The

critical supersaturation ratio is typically within the range of 2 to 546 and may or may not

correspond with the visual observations of particulate aggregation, called the point of

catastrophic nucleation. It would be expected the injection of a solution into an anti-

CHAPTER 2

52

S = Smax

S = 1D

BA

Solu

tion

Con

cent

ratio

n

Solution Expansion

Supersaturation Ratio

Rat

e

Growth Rate

Nucleation Rate

1 Smax

C

Figure 2.14 The relationship between supersaturation and particle growth.42

CHAPTER 2

53

solvent that is already at pressure, as in the ASES process, would result in an extremely

high rate of expansion. As a result even higher levels of supersaturation would be

expected in the ASES process and have been estimated to be in the order of 2630.47 At

these relatively high values of supersaturation the rate of particle nucleation is rapid

(Figure 2.14). The onset of nucleation forces solute out of solution and the solute

concentration falls back to a slightly supersaturated concentration (point D), close to the

binodal line. As the supersaturation drops the relative rate of nucleation decreases whilst

the particle growth becomes the dominant mechanism. In the case where solvent

expansion is sufficiently rapid to hold the solution saturation in the vicinity of the

unstable region, the nucleation of particles will occur continuously. The precipitate is

therefore polydisperse as each nucleus grows for a different period of time. In practice,

however, the rate of nucleation is so rapid that there is a significant drop in

supersaturation and it is very difficult to maintain supersaturation ratios close to Smax,

even with rapid solution expansion or cooling.

The width of the metastable region is a very important variable in crystallisation, and is

a strong function of the kinetics of the system. A rapid increase in the supersaturation of

a solution leads to a broadening of the metastable region, therefore delaying

precipitation to a higher supersaturation concentration. As nucleation rate is a function

of the supersaturation ratio, this results in more rapid nucleation and the generation of a

greater number of nucleation sites, which grow to a smaller size due to the finite mass

of solute in solution. At a slower rate of supersaturation increase, the precipitate exists

as larger and therefore fewer particles.

Solvent strength and the extent to which the precipitation system is mixed also effect

the width of the metastable region. For strong solvents, or highly concentrated solutions,

the number of molecules in a critically sized stable nucleus decreases and therefore the

metastable region narrows.42 As a result, the frequency of formation of critically sized

aggregates increases. Mixing, or agitation, of a solution again decreases the width of the

metastable zone and therefore the extent to which a solution is supersaturated before

precipitation.

CHAPTER 2

54

A number of saturated solvent—solute systems have been studied in terms of nucleation

and growth when expanded by CO2 at low expansion rates.25 It was discovered that not

all systems behaved the same when pressurised with CO2. The solutes could be divided

into three groups depending on the type of nucleation observed. Some systems

experienced catastrophic nucleation even at very low levels of expansion and others did

not crystallise at all even at high levels of expansion. The majority of the systems

studied underwent both nucleation and growth. Of the solutes studied, most inorganic

solutes precipitated by heterogeneous nucleation and a few by catastrophic nucleation,

metal—organic solutes precipitated by both heterogeneous and catastrophic nucleation

and organic solutes precipitated by heterogeneous nucleation or not at all. Tai and

Cheng have used supersaturation to explain the behaviour of different solutions when

expanded by the GAS process.25,48 A function for supersaturation in the GAS process

was given :

∆C = ⌡⌠

-

∂Ce

∂Cg T,P dCg = ⌡⌠

0

Cg

αdCg 2.48

α = -

∂Ce

∂Cg 2.49

where Ce is the solute solubility in the expansion course and Cg is the concentration of

dissolved anti-solvent on a solute free basis.48 Equation 2.49 was used to explain the

observation that not all solutes are precipitated from solution upon addition of anti-

solvent in the GAS process. The α factor is the change in solute concentration with

dissolution of anti-solvent. For a solution to remain in equilibrium the α factor can be

affected in two ways. The anti-solvent can reduce the dissolving power of the solvent by

decreasing the partial molar volume of the solvent, or the anti-solvent can act as an

additional solvent. A positive value for α implies the solute is insoluble in the anti-

solvent and nucleation and growth will occur. A negative α implies the anti-solvent

behaves as an additional solvent and undersaturation occurs. A large α implies that the

solute is insoluble in anti-solvent and the sensitivity of the solvent—solute dissolving

CHAPTER 2

55

power to the concentration of dissolved anti-solvent is high. At moderate values of α

nucleation and growth will occur.

Berends and co-workers50 adapted Nývlt theory49 to predict the pressure profile needed

to maintain constant supersaturation and growth rate in a batch GAS crystalliser. The

following expression was used:

weq

* (0) - weq* (t)

weq* (0) - weq

* (τ) =

t

τα 2.50

where weq* is the equilibrium solubility of the solute in the liquid phase and is calculated

by Equation 2.30, weq* (0) and weq

* (t) are the equilibrium solubility at time 0 and t

respectively, α is a correction factor and τ is the growth time. The pressure profiles for

unseeded and seeded growth were predicted and the results applied to the crystallisation

of phenanthrene from toluene using CO2 as anti-solvent.

The growth rates of the individual faces of a crystal are affected to different extents by

the supersaturation of the solution. As a consequence, in some cases it is possible to

alter the morphology of the forming crystals by changing the level of supersaturation at

which the crystallisation occurs. As a general rule the morphology of crystals is

disturbed at high levels of supersaturation, while at low levels of supersaturation more

regular crystals are formed due to the decrease in crystal growth rates. The effect of

supersaturation on crystal morphology has been investigated using the GAS process by

Yeo and co-workers.51 The effect of the rate of expansion on the crystal morphology

formed from DMSO solutions of NH4Cl and BaCl2 has been investigated by Yeo and

co-workers.51 It was observed that the rate of expansion altered the crystal morphology

and size. Generally slow expansion conditions produced equant morphologies, while

rapid expansion conditions resulted in needle-like or tabular morphologies which is

consistent with growth rates of individual crystal faces being dependent on

supersaturation.

Tai and Cheng investigated growth of single crystals of a number of solutes from

expanded solutions utilising the GAS process.48 In all cases the solutions were

CHAPTER 2

56

expanded slowly. The observations were that crystals generally undergo facet-growth in

the GAS process as occurs in conventional crystallisation processes. The crystal growth

rates ranged from 0.2 to 1.2 x 10-7 m.s-1, which was in the same order as growth rates in

aqueous solutions. The conclusions reached from the study were that crystal growth, at

conditions of slow expansion, in the GAS process is similar to that observed in

conventional processes of solution crystallisation.

2.4. Hydrodynamic Considerations

Hydrodynamic considerations in the dense gas precipitation of solutes has been

associated mainly with the ASES process. When a solution is injected into an anti-

solvent through a capillary, hydrodynamic forces act on the jet produced and tend to

cause the liquid jet to breakup. The breakup of the liquid stream occurs as a result of

shear, surface tension, and viscosity forces created by rapid expansion of a liquid stream

in a fluid. For low viscosity solutions the Weber number (NWe) indicates the extent of

break-up of a liquid droplet emerging from a nozzle into a flowing fluid. The Weber

number is the ratio of the deforming external pressure forces to the reforming surface

tension forces experienced by a liquid droplet.52 It is numerically defined by the

following equation:

NWe = ρU2D

σ 2.51

where ρ is the anti-solvent density, U the relative velocity, D the droplet diameter, and

σ is the interfacial tension. When external pressure forces are large compared to the

surface tension forces, NWe is large, causing jet breakup into droplets.

Large Weber numbers are expected in the ASES process compared to conventional

techniques such as spray drying and liquid anti-solvent processes. In the ASES process,

a number of contributing processes aid in breakup of the liquid jet into fine droplets.

Small values of interfacial tension, high anti-solvent density, and large mass transfer

driving forces are all expected to aid in jet breakup.

CHAPTER 2

57

The jet breakup length for Newtonian fluids can be related to jet velocity and viscosity

through the Weber and Reynolds numbers. An example of such an expression is:

Ld

= C NWe (1+3Z) 2.52

where L is the jet breakup length, d is the jet diameter, and C is a constant. The

Ohnesorge number Z is defined as follows:

Z = NWe

NRe =

µ

ρdσ 2.53

where µ is the jet viscosity, d is the jet diameter and NRe is the Reynolds number. Z

relates viscosity to surface tension forces. Jet breakup by atomisation will occur when Z

and NRe are large and by Rayleigh instabilities when Z and NRe have lower values.

Lengsfeld and co-workers23 proposed that when a dilute organic solvent is injected into

an anti-solvent in which the solvent is totally miscible (above the critical point of the

mixture), jet breakup does not proceed as described by conventional atomisation theory.

The formation of droplets is negated by the rapid disappearance of surface tension as the

solvent is sprayed into the anti-solvent. Dispersion of the jet then proceeds by means of

gas-like mixing. The basis of this proposition is that surface tension disappears before

the time scale of jet breakup. To test this hypothesis a model was developed based on

the modified Weber theory for jet breakup, which included time dependent surface

tension. The calculated jet breakup length was compared to the experimentally

measured jet breakup lengths. The results indicated that for partially miscible systems

the model could predict the experimentally observed jet breakup lengths reasonably

well when using dynamic surface tension. When the model was applied to totally

miscible systems, it was found that at conditions in the two phase region, the predicted

and observed jet breakup lengths were once again in agreement. In the single phase

region, the calculated jet breakup length was far shorter than the visually observed jet

breakup length. Both these lengths were far greater than the calculated length for the

disappearance of surface tension. The conclusion reached from the analyses was that the

CHAPTER 2

58

observed jet breakup in the single phase region was not a phase boundary, but a Kelvin-

Helmholtz instability driven by a density and velocity jump. These conclusions were

supported by the observation that the same jet shape was observed when injecting

supercritical trifluromethane into supercritical CO2.

For highly viscous solutions, jet breakup is more complex. Changing solution viscosity

or interfacial tension, affects jet breakup with increased viscosity stabilising the jet, and

increased interfacial tension destabilising the jet. The following function has been

proposed to predict jet breakup lengths for highly viscous solutions.53

λ ~ 13

v2ρa3

σ 2.54

where λ is the wavelength, v is the kinematic viscosity, ρ is the solution density, a is the

jet radius and σ is the interfacial tension. Equation 2.54 does not consider any

viscoelastic effects but it is thought to be useful in predicting jet breakup lengths for

highly viscous solutions in vapour over liquid ASES systems.53

2.5. Mass Transfer Considerations

As with hydrodynamic considerations, the discussion of mass transfer in the literature

has mainly dealt with the ASES process. When a solution is injected into a bulk anti-

solvent, mass transfer can occur in two directions. The anti-solvent can diffuse into the

liquid solution and the solvent can diffuse into the bulk anti-solvent.

A spray drying model has been used to describe the mass transfer processes occurring

on a methylene chloride droplet containing dissolved poly(L-lactic acid) in compressed

CO2.54 At pressures of 85 bar and temperatures of 29 °C, the flying time of the droplets

was found to be shorter than the time taken for the droplets to dry. The mass transfer

was found to be a function of droplet size with larger droplets taking longer to dry. The

conclusions reached from this study were that the required flying time of droplets is an

important consideration and should be considered in crystalliser design, and that the

growth of particles, rather than nucleation, was the dominant process.

CHAPTER 2

59

Another model has been developed to describe the mass transfer processes occurring on

a droplet that forms when injecting toluene into CO2.55 A 50 µm droplet was chosen as

a starting point and the initial interfacial flux, droplet radius, and droplet saturation were

calculated at various pressures and temperatures below the mixture critical point.

Calculations were performed to determine the initial interfacial flux, the droplet radius

with respect to time, and the composition of the droplet with respect to time. The results

showed that the initial net mass transfer was always into the droplet, which caused the

droplet to swell. The extent of swelling was a function of pressure and temperature with

the droplets undergoing a four-fold increase in size at certain conditions. The time to

reach maximum size was less than 0.3 seconds. Droplet lifetimes were found to vary

with pressure and temperature with shorter lifetimes at high values of pressure and low

values of temperature. The lifetimes of the droplets were between 1 and 2.5 seconds.

The composition of the droplet did not follow the same trend as the lifetime of the

droplet. Droplets that had longer lifetimes were found to saturate relatively early while

droplets with short lifetimes were found to saturate relatively late. The implications of

this last result are that, depending on the conditions, a droplet could saturate at a stage

when it was still larger than its initial size, an important consideration when forming

fine particles.

A number of researches have described the time for mass transfer to occur for a given

droplet radius. For example the time scale for diffusion loss of toluene from spherical

droplets can be given by:

t½ ~ 0.02182a2

D 2.55

where t½ is the time taken to lose half of the toluene from the droplet and D is the

diffusivity of toluene at the conditions in question.53,56 By assuming a diffusivity of 10-

5 cm2.s-1 the time for half the toluene to diffuse out of the droplet can be estimated for a

given droplet radius. For a droplet of 0.5 µm a half time of 10-6 s can be calculated. For

a droplet of 10 µm the half time is of approximately 4 x 10-4 s and for a droplet of 330

µm the half time is 0.6 s.

CHAPTER 2

60

For a droplet of solvent in vapour phase CO2 the diffusion of solvent into the vapour

phase is very low. The mass transfer of CO2 into such a droplet can be estimated using

the following equation57,58:

XCO2,tXCO2,∞

= 6d

Dtπ 2.56

where D is the diffusion co-efficient of CO2 into the liquid phase, d is the droplet

diameter, XCO2,t .and XCO2,∞ are the CO2 concentrations inside the droplets at time t and

at equilibrium respectively. The equation is strictly valid for short times only. For D = 2

x 10-9 m2.s-1 and d = 35 µm the time required for 90% dissolution into the droplet is 40

ms.

The mass transfer pathways in regards to polymer solution have been used to interpret

the morphology of the precipitates produced. Possible mass transfer pathways for a

polymer solution being injected into a compressed anti-solvent are depicted in Figure

2.15. Each pathway can be thought of as a change in composition with time at a certain

location in the liquid solution, or at a particular time the change in composition from the

center of the liquid solution phase to the interface.59 Two boundary curves which mark

the transitions from binodal to spinodal behaviour converge at the plait point (C*). The

region between the solvent—polymer axis and the binodal curve is the stable region. In

this region no precipitation will occur. The area between the binodal and spinodal curve

is the metastable region in which precipitation primarily occurs by nucleation and

growth. The area between the spinodal and anti-solvent axis is an unstable region where

precipitation occurs by spinodal decomposition. The concentration of the polymer

solution will determine where the mass transfer path crosses the binodal and spinodal

curves and ultimately determine the morphology of the polymer particles. On the

solvent side of the plait point in the metastable region polymer discrete domains

nucleate and grow in a solvent continuous phase. On the polymer side of the plait point,

solvent voids will nucleate and grow in a polymer rich phase forming fibres. Near the

plait point the distance to the binodal curve is short and the binodal and spinodal curves

CHAPTER 2

61

A

A’

B’

B

CD

C’

D’

Polymer

AntisolventSolvent

GlassyRegion

C*

Figure 2.15 Ternary phase diagram showing different polymer precipitation

schemes for the ASES process.10,53,59

CHAPTER 2

62

are close together. Phase separation will occur primarily by spinodal decomposition

which is characterised by the presence of bicontinuous phases, often with fine

structure.59 The spinodal curve can be crossed at any concentration but is most likely at

the plait point. The closer the end point of the mass transfer path is to the anti-solvent

vertex, the greater the number of voids in the polymer.10 The transition from discrete to

continuous polymer morphologies is thought to occur at approximately 3C*.13,60

CHAPTER 2

63

2.6. References

1. Hannay, B. J.; Hogarth, J.; ""On The Solubility Of Solids In Gases."," Proc. Roy.

Soc. 1879, 324.

2. Gallagher, P. M.; Coffey, M. P.; Krukonis, V. J.; Klasutis, N.; Gas Anti-Solvent

Recrystallization: New Process To Recrystallize Compounds Insoluble in

Supercritical Fluids; Johnston, K. P. and Penninger, J. M. L., Ed.; American

Chemical Society: Washington, DC, 1989; 406, 334.

3. Subramaniam, B.; Rajewski, R. A.; Snavely, K.; "Pharamaceutical Processing with

Supercritical Carbon Dioxide," J. Pharm. Sci. 1997, 86, 885.

4. Subra, P.; Jestin, P.; "Powders Elaboration in Supercritical Media: Comparison with

Conventional Routes," Powder Technol. 1999, 103, 2.

5. Palakodaty, S.; York, P.; "Phase Behavioral Effects on Particle Formation Processes

Using Supercritical Fluids," Pharm. Res. 1999, 16, 976.

6. Jarzebski, A. B.; Malinowski, J. J.; "Potentials and Prospects for Application of

Supercritical Fluid Technology in Bioprocessing," Process Biochem. (Oxford) 1995,

30, 343.

7. Bungert, B.; Sadowski, G.; Arlt, W.; "New Processes With Compressed Gases,"

Chem. Ing. Tech. 1997, 69, 298.

8. Bungert, B.; Sadowski, G.; Arlt, W.; "Separations And Material Processing In

Solutions With Dense Gases," Ind. Eng. Chem. Res. 1998, 37, 3208.

9. Reverchon, E.; "Supercritical Anti-Solvent Precipitation of Micro- and Nano-

Particles," J. Supercrit. Fluids 1999, 15, 1.

10. Dixon, D. J.; Johnston, K. P.; "Formation of Microporous Polymer Fibers and

Oriented Fibrils by Precipitation with a Compressed Fluid Anti-Solvent," J. Appl.

Polym. Sci. 1993, 50, 1929.

11. Kikic, I.; Bertucco, A.; Lora, M.; "A Thermodynamic Description Of Systems

Involved In Supercritical Anti-Solvent Processes," The 4th International Symposium

on Supercritical Fluids, Sendai, Japan, 1997, A, 39.

12. York, P.; Hanna, M.; "Particle Engineering by Supercritical Fluid Technologies for

Powder Inhalation Drug Delivery," Respiratory Drug Delivery, 1996; Vol. V, 231.

13. Mawson, S.; Kanakia, S.; Johnston, K. P.; "Metastable Polymer Blends by

Precipitation with a Compressed Fluid Anti-Solvent," Polymer 1997, 38, 2957.

CHAPTER 2

64

14. McHugh, M. A.; Krukonis, V. J.; "Supercritical Fluid Extraction," ; Butterworth:

Stoneham, MA, 1986.

15. Ng, H.J.; Robinson, D. B.; "Equlibrium Phase Properties of the Toluene-Carbon

Dioxide System," J. Chem. Eng. Data 1978, 23, 325.

16. Kordikowski, A.; Schenk, A. P.; Van Nielen, R. M.; Peters, C. J.; "Volume

Expansions and Vapor-Liquid Equilibria of Binary Mixtures of a Variety of Polar

Solvents and Certain Near-Critical Solvents," J. Supercrit. Fluids 1995, 8, 205.

17. de la Fuenta Badilla, J. C.; Peters, C. J.; de Swaan Arons, J.; "Volume Expansion in

Relation to the Gas-Anti-Solvent Process," J. Supercrit. Fluids 2000, 17, 13.

18. Peters, C. J.; Kordikowski, A.; Wilmes, B.; "Phase Behaviour of Selected Systems

of Interest for the GAS Process," Proceedings of the 5th International Symposium on

Supercritical Fluids, Atlanta, Georgia, 2000.

19. Nesta, D. P.; Elliott, J. S.; Warr, J. P.; "Supercritical Fluid Precipitation of

Recombinant Human Immunoglobulin from Aqueous Solutions," Biotechnol.

Bioeng. 2000, 67, 457.

20. Palakodaty, S.; York, P.; Pritchard, J.; "Supercritical Fluid Processing of Materials

from Aqueous Solutions: The Application of SEDS to Lactose as a Model

Substance," Pharm. Res. 1998, 15, 1835.

21. Thiering, R.; Dehghani, F.; Dillow, A.; Foster, N. R.; "Solvent Effects On The

Controlled Dense Gas Precipitation Of Model Proteins," J. Chem. Technol.

Biotechnol. 2000, 75, 29.

22. Kikic, I.; Lora, M.; Bertucco, A.; "A Thermodynamic Analysis of Three-Phase

Equilibria in Binary and Ternary Systems for Applications in Rapid Expansion of a

Supercritical Solution (RESS), Particles from Gas-Saturated Solutions (PGSS), and

Supercritical Anti-Solvent (SAS)," Ind. Eng. Chem. Res. 1997, 36, 5507.

23. Lengsfeld, C. S.; Delplanque, J. P.; Barocas, V. H.; Randolph, T. W.; "Mechanism

Governing Microparticle Morphology during Precipitation by a Compressed Anti-

Solvent: Atomization vs Nucleation and Growth," J. Phys. Chem. B 2000, 104, 2725.

24. Bertucco, A.; Lora, M.; Kikic, I.; "Fractional Crystallization by Gas Anti-Solvent

Technique: Theory and Experiments," AIChE J. 1998, 44, 2149.

25. Tai, C. Y.; Cheng, C.S.; "Effect of CO2 on Expansion and Supersaturation of

Saturated Solutions," AIChE J. 1998, 44, 989.

CHAPTER 2

65

26. Cannon, C. S.; Falk, R. F.; Randolph, T. W.; "Role of Crystallinity in Retention of

Polymer Particle Morphology in the Presence of Compressed Carbon Dioxide,"

Macromolecules 1999, 32, 1890.

27. Chang, C. J.; Randolph, A. D.; Craft, N. E.; "Separation of β-Carotene Mixtures

Precipitated from Liquid Solvents with High-Pressure Carbon Dioxide," Biotechnol.

Prog. 1991, 7, 275.

28. Bleich, J.; Müller, B.; "Production Of Drug Loaded Microparticles By The Use Of

Supercritical Gases With The Aerosol Solvent Extraction System (ASES) Process,"

J. Mocroencapsulation 1996, 13, 131.

29. Bodmeier, R.; Wang, H.; Dixon, D.; Mawson, S.; Johnston, K.; "Polymeric

Microspheres Prepared By Spraying Into Compressed Carbon Dioxide," Pharm. Res.

1995, 12, 1211.

30. Steckel, H.; Thies, J.; Müller, B.; "Micronising of Steroids for Pulmonary Delivery

by Supercritical Carbon Dioxide," Int. J. Pharm. 1997, 152, 99.

31. Falk, R.; Randolph, T. W.; Meyer, J. D.; Kelly, R. M.; Manning, M. C.; "Controlled

Release of Ionic Compounds from Poly(L-Lactide) Microspheres Produced by

Precipitation with a Compressed Anti-Solvent," J. Controlled Release 1997, 44, 77.

32. Gallagher, P. M.; Krukonis, V.; Botsaris, G. D.; "Gas Anti-Solvent (GAS)

Recrystallization: Application to Particle Design," AIChE Symposioum Series 1991,

87, 96.

33. Falk, R. F.; Randolph, T. W.; "Process Variable Implications for Residual Solvent

Removal and Polymer Morphology in the Formation of Gentamicin-Loaded Poly(L-

Lactide) Microparticles," Pharm. Res. 1998, 15, 1233.

34. Meyer, J. D.; Falk, R. F.; Kelly, R. M.; Shively, J. E.; Withrow, S. J.; Dernell, W.

S.; Kroll, D. J.; Randolph, T. W.; Manning, M. C.; "Preparation and in Vitro

Characterization of Gentamycin-Impregnated Biodegradable Beads Suitable for

Treatment of Osteomyelitis," J. Pharm. Sci. 1998, 87, 1149.

35. Manning, M. C.; Randolph, T. W.; Shefter, E.; Falk, R. F., III; "Solubilization of

Pharmaceutical Substances in an Organic Solvent and Preparation of Pharmaceutical

Powders Using the Same," Us 5770559, 1998.

36. Peng, D.; Robinson, D. B.; "A New Two-Constant Equation of State," Ind. Eng.

Chem., Fundam. 1976, 15, 59.

CHAPTER 2

66

37. Chang, C. J.; Randolph, A. D.; "Solvent expansion and solute solubility predictions

in gas-expanded liquids," AIChE J. 1990, 36, 939.

38. Dixon, D. J.; Johnston, K. P.; "Molecular Thermodynamics of Solubilities in Gas

Anti-Solvent Crystallization," AIChE J. 1991, 37, 1441.

39. Prausnitz, J. M.; Lichtenthaler, R. N.; de Azevedo, E. G.; "Molecular

Thermodynamics of Fluid-Phase Equilibria," ; Prentice-Hall Inc.: Englewood Cliffs,

NJ, 1986.

40. Khamski, E. V.; "Crystallization From Solutions," ; Olenum Publishing

Coorporation: New York, 1970.

41. Mullin, J. W.; "Crystallization," 3Rev. ed.; Butterworth-Heinemann: Oxford, 1997.

42. Nývlt, J.; "Industrial Crystallisation From Solutions," ; Butterworths: London, 1971.

43. Becker, R.; Doring, W.; Ann. Physik 1935, 24, 719.

44. Adamson, A. W.; "Physical Chemistry of Surfaces," ; Interscience Publishers:

Easton, P.A., 1963.

45. McCabe, W. L.; Smith, J. C.; "Unit Operations of Chemical Engineering," 3 ed.;

McGraw-Hill: New York, 1976.

46. Gallagher, P. M.; Coffey, M. P.; Krukonis, V. J.; Hillstrom, W. W.; "Gas Anti-

Solvent Recrystallization of RDX: Formation of Ultra-Fine Particles of a Difficult-to-

Comminute Explosive," J. Supercrit. Fluids 1992, 5, 130.

47. Schmitt, W. J.; Salada, M. C.; Shook, G. G.; Speaker, S. M., III; "Finely-Divided

Powders by Carrier Solution Injection into a Near or Supercritical Fluid," AIChE J.

1995, 41, 2476.

48. Tai, C. Y.; Cheng, C.-S.; "Supersaturation and Crystal Growth in Gas Anti-Solvent

Crystallization," J. Cryst. Growth 1998, 183, 622.

49. Nývlt, J.; "Batch Crystalliser Design," Advances in Industrial Crystallization, ;

Garside, J., Davey, R. J. and Jones, A. G., Ed.; Butterworth-Heineman Ltd: Oxford,

1991.

50. Berends, E. M.; Bruinsma, O. S. L.; de Graauw, J.; van Rosmalen, G. M.;

"Crystallization of Phenanthrene from Toluene with Carbon Dioxide by the GAS

Process," AIChE J. 1996, 42, 431.

51. Yeo, S.; Choi, J.; Lee, T.; "Crystal Formation of BaCl2 and NH4Cl Using a

Supercritical Fluid Anti-Solvent," Proceedings of the 5th International Symposium

on Supercritical Fluids, Atlanta, Georgia, 2000.

CHAPTER 2

67

52. Lefebvre, A. H.; "Atomization and Sprays," ; Hemisphere: New York, 1989.

53. Dixon, D. J.; Luna-Bárcenas, G.; Johnston, K. P.; "Microcellular Microspheres and

Miroballoons by Precipitation with a Vapour-Liquid Compressed Fluid Anti-

Solvent," Polymer 1994, 35, 3998.

54. Rantakyla, M.; Jantti, M.; Jaarmo, S.; Aaltonen, O.; "Modeling Droplet - Gas

Interaction and Particle Formation in Gas-Anti-Solvent System (GAS)," Proceedings

of the 5th Meeting on Supercritical Fluids, Nice, France, 1998, 333.

55. Werling, J.; Debenedetti, P.; "Numerical Modeling of Mass Transfer in the

Supercritical Anti-Solvent Process," J. Supercrit. Fluids 1999, 16, 167.

56. Crank, J.; "The Mathematics of Diffusion," 2nd ed.; Oxford University Press:

London, 1975.

57. Mugele, R. A.; AIChE J. 1960, 6, 3.

58. Wubbolts, F.; Bruinsma, O.; van Rosmalen, G.; "Dry-Spraying of Ascorbic Acid or

Acetaminophen Solutions with Supercritical Carbon Dioxide," J. Cryst. Growth

1999, 198/199, 767.

59. Dixon, D. J.; Johnston, K. P.; Bodmeier, R. A.; "Polymeric Materials Formed by

Precipitation with a Compressed Fluid Anti-Solvent," AIChE J. 1993, 39, 127.

60. Luna-Bárcenas, G.; Kanakia, S. K.; Sanchez, I. C.; Johnston, K. P.; "Semicrystalline

Microfibrils and Hollow Fibers by Precipitation with a Compressed-Fluid Anti-


CHAPTER 3

68

3. DENSE GASES AS ANTI-SOLVENTS — APPLICATIONS

3.1. Introduction

The reasons for applying the GAS or ASES processes to a substance can be to alter its

morphology, change its particle size or particle size distribution, fractionate it from a

mixture, precipitate solutes simultaneously to form a blend, or a combination of the

above reasons.

The potential benefits of the GAS process have been shown to have application in the

crystallisation of drugs, polymers, explosives, purification of organic acids and in the

fractionation and purification of polymers such as polymeric absorbents. An up to date

summary of all work investigating the GAS process is presented in Appendix I, Table

I.1.

The ASES process was first developed by Müller and coworkers as an alternative

microencapsulation technique for the formation of polymeric microparticle drug

delivery systems.1 Since then the process has been extended to the encapsulation of

pharmaceuticals.2-5 In addition, the ASES process has been employed for the

micronisation or recrystallisation of a wide range of compounds including

superconductor precursors6, pigments7, and for the micronisation of fine

pharmaceuticals such as proteins for aerosol drug delivery systems.8-11 The process is

also capable of precipitating polymeric fibres12 and metastable polymer blends13 and

fractionating solute mixtures.14 A summary of the research in the area of ASES

precipitation is presented in Appendix I, Table I.2.

3.2. Micronisation/Recrystallisation

Micronised precipitates are a requirement of many applications. Micronised powders

are required for use as explosives, chromatography, absorbents and catalyst supports,

dyes, catalysts, and superconductors. The most common reason quoted for applying the

ASES or GAS process as a micronisation technique is to micronise compounds for use

CHAPTER 3

69

in the pharmaceutical industry. In view of this, the application of micronisation in the

pharmaceutical industry and current methods of micronisation will be briefly discussed

before the use of GAS and ASES as a micronisation technique is reviewed.

The fact that two formulations of a specific drug meet certain chemical or physical

criteria does not necessarily mean that they are equivalent in their biological activity.15

Two drug formulations that are chemically equivalent may differ in their bioavailability

due to differences in physical factors such as crystal form and particle size.15 The

implication is that physical differences in a particular drug, a result of limitations in the

manufacturing process, can have consequences in the therapeutic activity of the drug.

The physical properties of the drug (such as particle size and particle size distribution)

can have significant implications depending on the method of drug delivery. The strive

towards improvements in drug therapy often focuses on the discovery of new drugs with

the delivery system used in the application of the drug receiving less attention.16

However, innovative drug delivery systems are needed to keep up with drug discovery

technologies and are an essential aspect of future innovation and growth.16,17 Drug

delivery development is an exciting avenue of research that at present the larger

pharmaceutical companies have not addressed.16 New drug delivery systems can help

cope with cost containment in healthcare, reduce adverse effects, improve patient

compliance, shorten the drug development time and extend the life of products with

expiring patents.17

There are many methods of delivering a drug to a patient. The task of the

pharmaceutical researcher is to discover the most appropriate method of delivering the

desired drug to achieve the most beneficial therapeutic activity with the least discomfort

to the patient. Often the latter two requirements are difficult to achieve simultaneously

and can pose a considerable challenge in drug development. For instance, the delivery

of vaccines is routinely performed by injection due to gastrointestinal destruction when

administered orally.18 Oral delivery is preferred due to the advantages of ease of

delivery, patient comfort , ease of administration, improved safety and less stringent

quality control.18 The challenge then becomes one of developing a delivery system for

vaccines that enables oral administration while maintaining therapeutic activity.

CHAPTER 3

70

The micronisation of pharmaceuticals is frequently utilised in the development of

advanced drug delivery systems17 and it is current trends in these systems that will be

discussed below.

3.2.1. Applications of Micronised Pharmaceuticals

The aim of micronising a pharmaceutical is to enhance bioavailability, increase the rate

of action, target delivery, give a predictable therapeutic response and provide greater

efficacy/safety.17 Micronised pharmaceuticals have found application in various

delivery systems including inhalation aerosols, injectable suspensions, controlled

release dosage forms and topical applications.19 The aim of this discussion is not to

provide a thorough review of the applications of micronised pharmaceuticals, but to

demonstrate and give examples of the usefulness of micronisation in advanced drug

delivery systems.

3.2.1.1. Increased Bioavailability

For a drug to be of use it has to be absorbed by the body. The absorption of a drug

involves its passage across cell membranes with the plasma membrane being the

common barrier to all drugs. The characteristics of a drug that determine its ability to

cross cell membranes are its molecular size and shape, solubility at the site of

absorption, degree of ionisation and relative lipid solubility of its ionised and

nonionised forms.15 The barriers to the absorption of a drug may be a single layer of

cells such as the intestinal epithelium or multiple layers of cells such as the skin.

Solubility is one of the important considerations in the absorption of drugs. Generally

drugs given in aqueous solution are more readily absorbed than those given in oily

solution, suspension, or solid form because they mix more readily with the aqueous

phase at the absorptive site.15

CHAPTER 3

71

The solubility of poorly water soluble drugs may be enhanced by micronisation. The

rate of dissolution of the micronised drug increases due to a larger surface area being

available to the dissolution medium. The increase in dissolution rate for micronised

pharmaceuticals has been translated to increased rate of absorption, level of absorption

and activity of the drug when administered orally.20-23 The enhanced bioavailability of

a drug as a consequence of micronisation may enable the use of a lower therapeutic

dose whilst maintaining the same level of activity. Reducing the required dose of drug

translates to the benefits of less chance of toxic side effects and a commercial saving in

terms of drug usage.

3.2.1.2. Intravenous Delivery

Ideally an intravenous drug formula is water based. In many instances drugs are

insufficiently soluble in water to enable an effective dose to be delivered intravenously.

A method of enabling the intravenous delivery of water insoluble drugs is to deliver

them as aqueous suspensions with drug particle sizes less than 5 µm.24 The upper limit

on particle size is 5 µm to avoid capillary blockage.

3.2.1.3. Oral Delivery

The application of micronised drugs in oral delivery systems is not limited to increased

bioavailability through increased dissolution. If the drug in question is unstable when

administered orally due to gastro-intestinal destruction, it is possible to administer the

drug orally by using a microparticulate carrier to protect the drug until it is absorbed as

a microparticle through the gastro-intestinal tract. The drug is then released after

absorption of the microparticle in an environment that does not destroy the drug.

Examples of this are illustrated in a number of recent reviews on the oral delivery of

peptides, proteins and vaccines.18,25-31

Microparticles are thought to be absorbed through the region of the gut-associated

lymphoid tissue known as Peyers Patches.18,25,26,29 It is generally accepted that particle

size is a key factor that determines uptake.18 Particles in the 0.05 to 3 µm size range are

CHAPTER 3

72

capable of uptake and decreasing particle size increases the extent of uptake.18 Particles

greater than 10 µm are not taken up and pass through the body.18,25

3.2.1.4. Inhalation Delivery

The respiratory system is an attractive delivery route due to the large surface area of the

alveolar region (50 to 100 m2) for absorption.15 Furthermore the rate of blood flow is

high and the blood is in close proximity to the alveolar air (10 µm).15 Advantages with

drug delivery to the lung as opposed to other routes include, almost instantaneous

absorption of drug into the blood, avoidance of hepatic first-pass loss and, in the case of

pulmonary disease, local application.15 Inhalation of a drug is also less invasive than

other delivery routes such as injection thereby increasing patient compliance to

treatment.

Inhalation delivery is not only applicable to the delivery of drugs that treat asthma or

chronic obstructive pulmonary diseases, such as peptides, proteins and hormones, but

can be used for the delivery of narcotics, ACE-inhibitors, insulin, calcitonin, heparin, α-

1 antitrypsin, interferon, FSH, vaccines and gene therapy.17

For effective adsorption of a drug after inhalation, it has to reach the alveolar zone of

the lung. In the past, pressurised metered dose aerosols have been the most common

form for inhalation but dry powder inhalation is gaining increased importance.17 In the

case of the inhalation of suspensions and powders, the particle sizes of the drug should

fall into the 1 to 5 µm range to reach the alveolar region.32 Particles larger than 5 µm

are typically deposited in the upper airway where they are expelled through sneezing,

blowing, wiping or carried into the gastrointestinal tract through the swallowing of

mucus.15 Particles that are less than 1 µm remain suspended and are exhaled.33

3.2.1.5. Ocular Delivery

The classic drug delivery system to the eye is the aqueous eye drop.34 Besides the

problem of delivering poorly water soluble drugs to the eye using this mechanism, rapid

CHAPTER 3

73

elimination of the drug occurs due to rapid tear turnover and precorneal loss.34 The

result is that the half life of drugs applied as aqueous drops to the eye is about 1 to 3

minutes and only 1 to 3% of the drug is absorbed through the intraocular tissue.34

Furthermore the possibility of systematic drug exposure due to nasolacrimal duct

drainage can lead to unwanted side effects and toxicity.34,35

Alternatives to the aqueous eye drop are drug formulations that involve ointments or

large inserts. These delivery systems suffer the disadvantages of patient discomfort

through blurred vision or difficulty in application.

Two alternative ocular drug delivery formulations are nano- and microparticulates.35

The drug is combined with a polymer carrier as a nano- or microparticle and a

suspension of these particles are administered to the eye in much the same way as the

aqueous eye drop. Upon application the particles reside in the occular cul-de-sac, and

the drug is released through diffusion, chemical reaction or polymer degradation.35 To

prevent discomfort the particle sizes should not exceed 10 µm.34 Upon administration

the particulate formulation has a longer residence time in the eye as it is not removed as

easily as aqueous formulations by the action of tears.

Of all the ocular delivery systems, microparticulate technology is the most promising as

it has the advantage of improved patient compliance because the microparticles can be

topically administered as an eye drop.35 To date only one microparticulate product,

Betopic S, has been released on the US market.35 The limitation to forming

microparticulate drug systems for ocular drug delivery is the difficulty and expense of

manufacture of sterile drug-polymer particles.35

3.2.1.6. Nasal Delivery

The nose is an attractive drug delivery route because of the convenience of application

and the ability of the nasal mucosa to absorb various drugs.36 The administration of (20

to 45 µm) microparticulate particles of insulin encapsulated with various resins via the

nasal route of rabbits has been found to result in the effective absorption of insulin into

CHAPTER 3

74

the blood stream.36 Encapsulation of the drug was necessary as polar drugs such as

insulin are not well absorbed by the nasal mucosa.

3.2.1.7. Drug Delivery to the Central Nervous System

Drug delivery to the central nervous system poses a problem due to the exclusion of

many drugs from blood transfer to the brain owing to the negligible permeability of the

brain capillary endothelial wall, which makes up the blood brain barrier. Strategies that

have been used to transfer drugs to the brain include osmotic disruption of the blood

brain barrier, infusion pumps delivering drugs to the cerebrospinal fluid, intravenous

injection of surfactant coated nanoparticles, implantation of tissues or cells and gene

therapy.37 The newest strategies involve drug delivery polymeric devices which are

inserted into the target area by invasive surgery after which the drug is slowly released

from the polymer to the surrounding area.37 Another strategy of drug delivery to the

brain involves the formation of encapsulated microparticles of the drug.37 A suspension

of these microparticles is then delivered to the precise target area by stereotaxy which is

far less invasive than open surgery.

3.2.2. Methods of Micronising Pharmaceuticals

In the formulation of suspensions for intravenous delivery or dry powders for

inhalation, particle size determines the effectiveness of the drug. It is therefore

important in the micronising of pharmaceuticals that the process be predictable and

reproducible.

As a unit operation, the theory of size reduction has focused on energy relationships,

equipment and grinding limits.38-40 In terms of pharmaceuticals, the compounds to be

micronised are often costly, sensitive and unstable and the application of grinding

technologies from other industries is not suitable.19 The availability of equipment

designed specifically for the pharmaceutical industry is, however, lacking and often the

pharmaceutical industry has no choice but to adopt these unsuitable technologies. Some

CHAPTER 3

75

of the more common micronising techniques that have been adopted by the

pharmaceutical industry are listed in Table 3.1.

Table 3.1 Conventional methods of micronisation.19,41,42

Method Size Distribution

(µm)

Disadvantages

Fluid Energy Mill 1-5 High energy input, temperature increase,

electrostatic charging.

Spray Drying ∼ 5 Operation above ambient temperature.

Lyophilisation < 1 Poor control over size distribution, energy

intensive.

Solution

Preparation

< 1 Poor control over size distribution, solvent

recovery.

Freeze Drying Applicable to only a few substances.

More recently Rapid Expansion of Supercritical Solution (RESS) and Precipitation

from Gas Saturated Solution (PGSS) technologies based on the use of supercritical

fluids have been developed for use in the micronising of pharmaceuticals. These

technologies were developed to enable the micronising of pharmaceuticals at low

temperatures using non-toxic solvents such as supercritical CO2.

The RESS process involves dissolving the solute of interest in a supercritical fluid and then rapidly (< 10-6 s) depressurising the solution. Very high levels of supersaturation can result thereby precipitating the solute as a fine precipitate. The very rapid

depressurisation and high supersaturation ratios are expected to provide uniform conditions of nucleation and therefore narrow particle size distributions and small

particle sizes respectively.43 The RESS process offers a number of advantages over

conventional micronisation methods such as grinding or precipitation from solution. Solids which are sensitive to shock, thermal degradation, or chemical degradation from exposure to the atmosphere can be micronised by RESS which takes place at mild

operating conditions in an inert atmosphere of supercritical solvent. The precipitate and supercritical solvent are also conveniently separated during the decompression stage,

CHAPTER 3

76

unlike a conventional precipitation from solution which requires a solvent removal step.

The process has been demonstrated to be suitable for the micronisation of polymers44,

dyes45, pharmaceuticals46 and inorganic substances.47 A significant disadvantage of the RESS process however is the limited solubility of most pharmaceuticals in supercritical

fluids such as CO2. The low solubility in many cases makes the RESS process unsuitable as a micronisation technique due to the very high gas to solute ratios.

The PGSS process is similar to the RESS process but in this case a compressed gas such

as CO2 is dissolved under pressure in a melted solid.48-50 The solution so formed is then rapidly expanded to a lower pressure through a nozzle and the evaporation of the

compressed medium lowers the temperature below that of the solidification temperature of the solute. Micronised solid particles are formed during the spray process. The advantages of PGSS are much the same as those for RESS with the additional

advantages of lower pressure requirements and lower consumption of gas compared to RESS. As with the RESS process the PGSS process is limited in the substances that can be processed. A requirement is that the solute has a low melting point and that the

compressed gas dissolves in the melt. These conditions are not met by most pharmaceuticals. In terms of the processing of pharmaceuticals there is much scope for the development

of alternative micronisation techniques. Although the techniques described above have been used to process pharmaceuticals, they all suffer from serious drawbacks. The use of compressed gases as anti-solvents is presented in this thesis as an alternative

technique that may enable the processing of a broader range of not only pharmaceuticals but many other chemicals.

Dense gas processing occurs in an inert atmosphere at near ambient temperatures,

without the application of high shear forces (in the case of GAS), which is in contrast to

some of the traditional “high shear” techniques listed above. It is therefore

advantageous to use dense gas precipitation in the processing of thermally, chemically,

or physically unstable material.

CHAPTER 3

77

3.2.3. GAS as a Micronisation/Recrystallisation Technique

In terms of some explosives, the requirement on powders is that the particle size should

be less than 40 µm51 and of a regular morphology.52 These two requirements are

important in terms of the combustion process and sensitivity of the material. The GAS

process has been demonstrated to be an effective tool in the recrystallisation of the

explosives nitroguanidine, cyclotrimethylenetrinitramine and 1,3,5,7-tetranitro-1,3,5,7-

tetraazacyclooctane (HMX),51-55 molecules which are sensitive to thermal and

mechanical stress. It was observed that the size of the particles produced was related to

the rate of the expansion of the solution. Rapid expansion resulted in smaller particles

than slow expansion. The particle size distribution and quality of the precipitate was

superior to the conventional crystallisation precipitate.

The GAS process has been demonstrated as being an effective micronisation technique

for the precipitation of both organic and inorganic materials.56 Cobalt chloride and 2-

hydroxybenzylalcohol (saligenin) were precipitated from acetone using a carbon

dioxide anti-solvent. In the case of saligenin it was discovered that particles formed

were independent of the initial concentration of solution. Monodisperse, irregular

shaped particles were produced in all cases. Slow volumetric expansion of the solution

produced very large particles (1000 µm) whereas faster expansion produced small

particles (20 µm). For cobalt chloride two distinct crystal morphologies, having

different colours, were observed - "purple needles" and "magenta chunks". For rapidly

expanded solutions, small particles were formed with the crystal structure depending on

the initial concentration and temperature. Higher concentration at higher temperature

resulted in a purple precipitate, whilst a lower concentration at room temperature,

resulted in a precipitate that was more pink in colour. Slow expansion produced crystals

only slightly larger than those produced by rapid expansion (100 µm).

Benedetti and co-workers studied the micronisation of hyaluronic acid ethyl ester

(HYAFF-11) by expanding a DMSO polymer solution with carbon dioxide.57 Spheres

with an average size of 0.4 µm and a very narrow size distribution were obtained. These

were smaller by an order of magnitude than the particles obtained from the ASES

CHAPTER 3

78

process. It was found that temperature did not have any effect on the particle

morphology within a range of 35°C to 60°C. The feed concentration was found to have

the largest effect on the particle size distribution. Larger particles that tended to

aggregate formed from dilute solutions. As the solutions became more concentrated,

smaller more discrete particles were formed.

Differing crystal morphologies of a substituted para linked polyamide were formed by

expanding DMSO solutions at different rates with carbon dioxide.58 At a rapid rate of

expansion (2.06 MPa.min-1) a 0.03 (w/w)% solution formed a non-spherulitic

precipitate, whereas at a slow pressurisation (0.08 MPa.min-1) rate a 0.12 (w/w)%

solution resulted in the formation of a spherulitic network of lamellar crystals. This

difference in crystal morphology was attributed to agitation in the expanding solution.

Slow injection of carbon dioxide provides a stable environment for the formation of

crystals. Fast addition of carbon dioxide, on the other hand, provides a turbulent

environment for crystal growth. This leads to particles with a lower crystalline content.

The precipitation of insulin from a carbon dioxide expanded DMSO solution again

showed that expansion rate influences particle size.11

GAS as a micronisation technique was further extended into pharmaceutical

applications by the recrystallisation of a variety of steroidal compounds from acetone

and ethanol using a carbon dioxide anti-solvent..11,59,60 The aim was to produce

particles approximately 10 µm in size for use in respiratory inhalers. Particles of 1 to 2

µm of dexamethasone were produced from acetone and flat platelets were produced

from expanded ethanol solutions.

The GAS process has been applied to the model compounds phenanthrene and

naphthalene to determine the effect of process parameters. The effects of stirrer speed,

different pressurisation profiles and growth times for phenanthrene precipitated from a

toluene and carbon dioxide system have been studied.61 It was discovered that the

average particle size of the phenanthrene crystals could be influenced by a factor of

three (160 to 540 µm) by varying process conditions. Berends and co-workers61 also

CHAPTER 3

79

measured the residual solvent in the product. It was found that the residual solvent

concentration could be reduced to 12 ppm by washing with supercritical carbon dioxide.

The GAS process was found to produce smaller particles of β-Carotene with a narrower

particle size distribution than those obtained from a conventional micronisation

technique.62 Particle sizes in the range from 2 to 10 µm were obtained. The advantages

of low oxygen exposure and rapid recrystallisation were noted to reduce the oxidation

and degradation of the product.

Micron sized particles of bilirubin, a heat labile oxidation medicine, have been produced

using the GAS process.63 Crude bilirubin was dissolved in DMSO and then precipitated

by expanding the solution with supercritical carbon dioxide. Particles of less than 0.5

µm diameter and 1 µm length were obtained.

The size of sulfathiozole crystals could be controlled by varying the expansion rate.64 A

slow stepwise expansion of the sulfathiozole saturated ethanol solutions with CO2

resulted in large pillar like crystals (2 to 6 mm) attached to the walls of the container.

By rapidly expanding the solution similar crystals were obtained that were reduced in

size by two to three times. These also grew on the walls of the container. Using a two

step pressurisation much smaller crystals (45 - 280 µm) precipitated from the bulk

solution and did not grow on the sides of the container. Consequently it was concluded

that the pressure pulse caused homogeneous nucleation and precipitate growth occurred

in the suspension without aggregation. A very high pressure was initially applied to the

system and held until a desired expansion was reached at which time a second

depressurisation step was introduced to hold the solution at that expansion level.

Wubbolts and co-workers observed that the temperature in the precipitation chamber

increased upon pressurisation.65 This increase in temperature was due to the heat of

condensation that occurred when gaseous CO2 condensed into the solvent. The GAS

process was modified slightly by introducing carbon dioxide into the system as a liquid

rather than as a vapour which prevented the temperature increase. The system was first

pressurised with helium before liquid carbon dioxide was introduced at the elevated

CHAPTER 3

80

pressure. Using this pre-pressurisation technique, acetaminophen was recrystallised

from ethanol. Varying the rate of addition of carbon dioxide controlled expansion of the

solution. Crystals with well developed faces with very little agglomeration were formed.

High yields were obtained from concentrated solutions. The particles formed from the

higher concentration solutions were smaller.

Thiering and co-workers studied the effect of various parameters on the precipitation of

para-hydroxybenzoic acid (p-HBA) from methanol, acetone and ethyl acetate solutions

expanded with carbon dioxide at 35°C. The type of solvent used in the process was

found to have the greatest effect on crystal morphology as opposed to changes in rate of

expansion and initial solute concentration. Precipitation of p-HBA from methanol

solutions resulted in 1 - 2 µm spheres, whilst precipitation from acetone and ethyl

acetate caused products up to 3 mm in length to form. This was explained in terms of

the hydrogen bonding that forms in methanol solutions between the methanol and p-

HBA. Greater expansion rates and therefore higher supersaturation are required before

these bonds can be broken and precipitation can occur. The corresponding rapid

nucleation and growth rate resulted in an amorphous precipitate. Changes to the initial

solute concentration of methanol solutions caused changes in particle morphology and

size. Larger particles were formed at higher concentrations. The formation of larger

particles was explained by the fact that there is a greater mass of solute to build upon

the same number of existing nuclei. Changes in the rate of addition of carbon dioxide

also resulted in changes to particle morphology. At slow rates of carbon dioxide

addition, and therefore slow expansion, dendritic shaped needles of up to 1 mm were

formed. At fast addition and therefore high expansion rates, amorphous particles were

produced.

The use of an alternative anti-solvent to CO2 has been demonstrated with the

precipitation of polystyrene from toluene using HFC-134a as anti-solvent.66 A lower

pressure was needed than when using CO2 to precipitate the polymer. In the same

experiments the morphology of the polymer obtained could be varied by changing the

concentration of the initial solution.

CHAPTER 3

81

The micronisation of the proteins lysozyme, insulin, and myoglobin has been

achieved.11,67,68 Monodisperse microspheres with an average diameter of 0.1 to 1.5 µm

were obtained when lysozyme or insulin was precipitated from DMSO with CO2.

Importantly lysozyme showed no loss of biological activity after processing. To develop

a more attractive solvent system, insulin was precipitated from mixtures of water and

ethanol or methanol. Protein particles of 1.5 µm were obtained from water using

ammonia as an anti-solvent.

Solutes that are sparingly soluble in organic solvents such as α-chymotrypsin,

imipramine, insulin, ribonuclease, cytochrome C and pentamidine have been

successfully micronised with the GAS process in combination with the hydrophobic ion

pairing technique to increase the solubility.69 By adding an amphiphilic material such as

sodium dodecyl sulfate or bis-(2-ethylhexyl) sodium sulfosuccinate to the solution the

solubility of the solute can be increased to levels that are practical for GAS processing.

The crystal size and morphology of BaCl2 and NH4Cl were altered by varying the rate

of addition of CO2.70,71 A slow injection rate produced cubic crystals of BaCl2 while a

fast injection rate produced flattened needles. For NH4Cl the cubic shape was also

obtained for slow expansion but a tabular habit was obtained with fast expansion. For

NH4Cl a size reduction from 200 to 1.8 µm was observed with increasing expansion

rate.

3.2.4. ASES as a Micronisation/Recrystallisation Technique

The ASES process has been shown to be able to precipitate particles of a wide range of

sizes and morphologies including nanospheres as small as 0.1 µm72, to porous fibres

averaging 150 µm in size12, microfibrils73 and larger wafer structures.74 Many of the

encouraging results to date are discussed with respect to the types of solutes that have

been precipitated.

The ASES process has been applied to the proteins catalase and insulin.75-77 Catalase

precipitated as 1 µm spherical and rectangular particles whereas insulin formed both

CHAPTER 3

82

aggregated nanospheres and 1 µm thick needles 5 µm in length. The two types of

particles were found in separate areas of the collection filter device.

Yeo and co-workers conducted a more detailed study of the precipitation of insulin from

DMSO and DMFA.11,78 Eighty percent of the powder produced was between 1 µm to 4

µm in diameter. Temperature, solute concentration or the nature of the solvent did not

significantly affect particle size. The biological activity of the insulin powder was also

shown to be unchanged from in vivo studies. Raman spectroscopy further confirmed

these findings.78 Similar results for other proteins such as trypsin and lysozyme were

also recorded.9,10

The micronisation of proteins is a difficult, yet desirable, objective in the

pharmaceutical industry for the development of alternative drug delivery systems such

as inhalation. The low solubility and incompatibility of proteins in organic solvents

provides a challenge when using the ASES technique due to the low solubility of CO2 in

water. Methods that have been used to enable the ASES process to be used as a

micronisation technique for proteins are to either inject a solution of the protein

dissolved in water simultaneously with an organic solvent into CO2, or to inject the

aqueous solution of protein directly into CO2 which has been modified with an organic

solvent. The presence of the organic solvent enables the water to be extracted by the

CO2 and the protein to precipitate. Human immunoglobulin79, lysozyme80,81, albumin,

rhDNase insulin81, trypsin80, a therapeutic peptide, antibody Fv and Fab and Plasmid

DNA pSVβ82 have been processed using this method. Trypsin and human

immunoglobulin were micronised, but were found to lose biological activity on

processing. A more positive result was obtained for lysozyme, albumin and insulin

which produced micro-sized particles which maintained biological activity.

Substantially denaturation of rhDNase occurred during the process.

The ASES technique has also been applied to a number of organic compounds and

pharmaceuticals to determine the feasibility of the process as a micronisation technique.

Naproxen5, prednisolone acetate83, amoxicillin, griseofulvin, ampicillin, tetracycline84-

86, methylprednisilone87, paracetamol88, hydrocortisone, ibuprofen, camptothecin89,90,

CHAPTER 3

83

salmeterol xinafoate91, hydroquinone, acetominophen ascorbic acid92,93 and p-

hydroxybenzoic acid94,95 have all been processed using ASES. The success of the

micronisation has been found to be solute-solvent dependent with some systems such as

griseofulvin - N-methylpyrolidone not precipitating due to solvent-solute interactions.86

A similar result was obtained for the steroids betamethasone-17-valerate, and

beclomethasone-17,21-dipropionate with no precipitate forming on processing by

ASES.96 In some cases particle morphology was found to be solvent dependent.91

Salmeterol xinofoate precipitated from acetone exhibited accretion-forms while those

precipitated from ethanol exhibited a thin blade-like morphology. Generally once the

optimum conditions were found microparticles were produced for most of the

pharmaceuticals studied.

The micronisation of pharmaceuticals that are not normally soluble in organic solvent

suitable for ASES such as streptomycin, has been performed using hydrophobic ion

pairing to increase the solubility of the solute in the organic solvent.69 The homogenous

organic solutions containing the ion paired compound were injected into CO2 through a

sonicated nozzle producing spheroidal 0.4 to 1 µm particles.

A number of polymers have been processed at various conditions using the ASES

process with varying results. Polystyrene formed a wide range of morphologies ranging

from microspheres, to porous fibres and microballoons.12,74,97 Microspheres were

precipitated from a 1 wt% solution of polystyrene in toluene through a 100µm nozzle

and at higher concentrations fibres were precipitated. Microparticles of polystyrene

could be obtained for higher concentrations of polystyrene by injecting the toluene

solution into gaseous HFC-134a instead of CO2.98 When injecting into liquid HFC-134a

chunks of polystyrene were obtained.98 For the semicrystalline polymer

polyacrylonitrile hollow fibres and highly oriented microfirils were produced even at

low concentrations.73,99 Fibre formation was also observed for a polyamide with a rigid

backbone.58

Since the successful development of the ASES process as a micronisation technique for

the production of fine pharmaceuticals, its application has been extended to the

CHAPTER 3

84

precipitation of FDA approved bioerodible polymers as encapsulating agents. These

have included slow-degrading types of high molecular weight such as poly(lactic

acid)57,100-103 and polycaprolactone102 as well as faster degrading copolymers such as

poly(lactide-co-glycolic acid).104

Randolph and co-workers first precipitated L-poly(lactic acid) in 1993 from methylene

chloride solution.100 Sub-micrometer-sized particles were produced using near-critical

or supercritical carbon dioxide as the anti-solvent. Various other authors have since

consistently precipitated L-poly(lactic acid) in the form of discrete microspheres

throughout a broad range of operating conditions and solution concentrations.57,95,101-

103

Bodmeier and co-workers attempted to find polymers that would not agglomerate

during the ASES process.102 A range of biodegradable polymers, including ethyl

cellulose, poly(methyl methylacrylate), poly(E-caprolactone), DL-poly(lactic acid), L-

poly(lactic acid) and DL-poly(lactide-co-glycolic acid) were studied. The amorphous

polymers with a low glass transition temperature were agglomerated even at low

processing temperatures such as –10°C. The formation of discrete microparticles was

favoured at moderate temperatures, low polymer concentrations, high pressures and

high flow rates of carbon dioxide. The semi-crystalline polymer, L-poly(lactic acid),

which has a high glass transition temperature, exhibited negligible plasticisation effects.

L-poly(lactic acid) was found to commonly precipitate as discrete microspheres, the size

of which increased from less than 1 µm up to 5 µm with a temperature increase of 0 to

32°C, which may be due to effects on the system hydrodynamics. At 40°C the

semicrystalline biodegrdable polymers L-poly(lactic acid) and poly(β-hydroxy butyric

acid) formed microspheres when processed by ASES but the amorphous polymers

poly(DL-lactide) and poly(DL-lactide-co-glycolide) were extracted and gave no

precipitate.104

Mawson and co-workers modified the solution injection device by developing a coaxial

nozzle in an attempt to reduce particle agglomeration.105 Polystyrene and L-poly(lactic

acid) particles were precipitated from toluene and methylene chloride respectively. A

CHAPTER 3

85

stabiliser (poly(1,1-dihydroperfluorooctyl acrylate)) was added into the solutions to

further reduce particle flocculation and agglomeration.106,107 The addition of the

stabiliser was found to significantly reduce flocculation.

A study of the ASES processing of a range of biopolymers including dextran, inulin,

poly-L-lactic acid, poly(hydroxypropylmetacrylamide) and poly-hyaluronic acid has

been performed.108,109 Nano- and microparticles were obtained for all polymers tested

with some fiber formation occurring at certain conditions in the processing of poly-

hyaluronic acid. In the same study networked particles of polyvinyl alcohol were

obtained while the polymer poly caprolactone was not precipitated at all. The lack of

precipitation of poly caprolactone was attributed to the high solubility of the polymer in

CO2.

Epoxy resin has been precipitated from acetone and methyl ethyl ketone using the

ASES process.110,111 By comparing particle morphology with and without surfactant,

and with co- and counter-current operation, it was concluded that mass transfer and

hence nucleation and growth rates are more significant than jet break-up in controlling

the particle size distribution.

The very fast precipitation that occurs in the ASES process has been exploited to form a

polymer blend of polycarbonate and poly(styrene-co-acrylonitrile).13 The success of

blend formation was found to be a function of temperature and concentration with

optimum conditions being temperatures below 25°C and concentrations below 3 wt %.

Various other compounds have also been precipitated such as highly uniform

nanospheres of several kinds of pigments,7,112,113 nanoparticles of superconductor

precursors for the textile and electronic industries6,114,115 and various morphologies of

buckminsterfullerene.116 Phospholipids such as soy lecithin have been successfully

micronised by the ASES process for application in the elaboration of liposomes.117,118

Amorphous spherical and agglomerated particles in the range 1 to 40 µm were obtained

when injecting ethanol solution of the phospholipids into CO2. A very promising result

CHAPTER 3

86

from this study was the absence of residual solvent in the processed product which

meant that more toxic solvent could be used in processing.

The ASES process has been demonstrated to be an effective micronisation technique for

a number of hydrophilic compounds such as lactose and sucrose.119-122 The

micronisation was achieved by introducing the solute dissolved in water, an organic

modifier such as methanol and CO2 simultaneously through a multiple channel nozzle.

Extraction of water was possible due to the presence of methanol and the solute

precipitated as microparticles.

The potential of the ASES process was evaluated in terms of scale-up using

acetominophen and lysozyme as model compounds.123 For both solutes the

microparticles produced in the pilot plant were essentially the same as those produced at

lab scale. The biological integrity of lysozyme was also maintained after pilot plant

processing.

3.3. Co-Precipitation/Encapsulation

The formation of co-precipitates or the encapsulation of particles with a second solid is

an important process, especially in relation to the pharmaceutical industry. A successful

encapsulation or co-precipitation does not necessarily have to result in microparticles. In

some applications the formation of microparticles is essential as in the case of inhalation

formulations. The formation of macrostructures of biodegradable polymer such as

hollow fibres or balloons that contain the relevant drug are useful in other applications

such as delayed release where the drug is taken orally.

To date there are no reported studies where the GAS process has been used as an

encapsulation technique. In a fundamental study on co-precipitation it was observed that

as different solutes show different precipitation pressures, from a thermodynamic point

of view, they cannot be precipitated together.124 To obtain the simultaneous

precipitation of two solutes the process must be carried out at kinetically limited

CHAPTER 3

87

conditions. The ASES technique is a very fast precipitation process and as such is

attractive as an encapsulation or co-precipitation process.

3.3.1. ASES as an Encapsulation Technique

At present ASES technology for microencapsulation purposes is still in its early stages

with many areas yet to be explored. Studies to date have involved the simultaneous

precipitation of neat core and coating solute solutions or the precipitation of pre-mixed

solutions with various types of coaxial nozzles. The technique has been applied with

some success and typically some percentage of the microspheres produced are bound up

in a network structure.

The potential of the ASES process for microencapsulation was demonstrated with the

microencapsulation of the model drug, hyoscine butylbromide, with L-poly(lactic acid)

by co-precipitation from a mixture of methanol and methylene chloride.2 In this case a

large percentage of the drug was on the surface of the polymer. Other model drugs,

including indomethacin, piroxicam and thymopentin, have also been co-precipitated

with L-poly(lactic acid) from methylene chloride.3 A maximum drug loading of 19.8 wt

% was obtained with hyoscine butylbromide while a loading of 4.9% was achieved with

thymopentin. The drug loading of the particles was related to the solubility of the drug

in the anti-solvent phase. The more soluble drug, indomethacin, was totally extracted

and was not present in the polymer after processing.

A detailed study on the microencapsulation of compounds with L-poly(lactic acid) from

methylene chloride was published by Falk and co-workers.4,125 The ionic compounds

considered were gentamycin, naloxone, and naltrexone. Drug loading efficiencies of up

to 37.4 wt % were achieved with rifampin/L-poly(lactic acid) particles while others

such as naltrexone/L-PLA were as low as 1.7 wt %. The process was unable to

effectively incorporate highly lipophilic drugs into the polymer particles. Drug release

profiles were also studied. For gentamycin/L-poly(lactic acid) microspheres almost no

burst release was observed and the drug release profile indicated release by matrix-

controlled diffusion.

CHAPTER 3

88

A successful co-precipitation of naproxen and L-poly(lactic acid) from acetone has been

achieved using the ASES process.5 A relatively uniform mixture of the drug and

polymer was obtained.

The potential of ASES to encapsulate proteins has been investigated in an attempt to

encapsulate lysozyme with L-poly(lactic acid) or D,L-poly(lactic-co-glycolic acid) from

a dichloromethane solution.126 As in the pure protein precipitation studies, the

biological integrity and solubility of the protein in a suitable solvent for ASES are

important issues. Initially, a suspension of lysozyme in the polymer solution was

injected into the carbon dioxide anti-solvent present as a vapour to partially solidify the

droplets before they fell into a carbon dioxide liquid phase. The integrity of the polymer

coating or the biological activity of the lysozyme was not investigated. Particle

aggregation, however, was noted to be significantly reduced when operating at a

temperature of -20°C. At this temperature the polymer was sufficiently rigid to prevent

coalescence and the rate of mixing between the carbon dioxide and organic solvent is

still rapid enough to disperse the polymer.

The possibility of forming protein polymer composites has been investigated further

with the formation of micro-capsules of insulin, lysozyme and chimotrypsin with L-

poly(lactic acid).127 Alternative methods to solubilise the protein in the organic solvent-

polymer solution were tried to enable the preparation of a homogenous mixture of the

polymer and protein in an organic solvent. Methods of hydrophobic ion pairing for

chimotrypsin, protein conjugation for insulin and a binary solvent mixture for lysozyme

were investigated. The two former methods did not solubilise the protein enough for an

ASES study and the latter method was used for further investigation. A high protein

yield of over 80% was obtained with the biological activity of insulin being maintained.

The biological activity of lysozyme decreased by 70% during processing. Importantly,

dissolution studies showed that the protein had been encapsulated with slow release of

the protein occurring.

CHAPTER 3

89

Combinations of urea/chloramphenicol and ascorbic acid/paracetamol have been co-

precipitated from ethanol at 40°C and 9.0 MPa using the ASES process.128 The quality

of the products produced each time were not quantified, although problems with particle

agglomeration and fractionation below 8.5 MPa were identified for the co-precipitation

of paracetamol with ascorbic acid.

The need to precipitate multiple solutes, with differing solubilities, in the ASES process

has lead to the development of multiple channel nozzles that enable drug and polymer

solutions to be mixed near, or at the point of contact, with the anti-solvent.95 Using the

coaxial nozzle the co-precipitation of the model drug p-hydroxybenzoic acid with the L-

poly(lactic acid) or L-poly(lactide-co-glycolide) was investigated. The solution of the

drug was injected through the inner nozzle and the polymer through the outer nozzle.

The co-precipitation appeared to be successful when analysed visually.

The effect of polymer crystallinity and thermal behaviour in the encapsulation of two

model drugs albumin and estriolm have been investigated.129 The block copolymers b-

poly-L-lactide-co-D,L-lactide-co-glycolide and poly-D,L-lactide-co-glycolide were co-

precipitated with the model drugs. The two polymers gave the same encapsulation

results, which were that the drug was primarily coated on the surface.

The coating of sugar and glass beads with poly(D,L,-lactide-glycolide) or

hydrocortisone has been achieved by spraying the coating solution into a vessel

containing the beads pre-pressurised with CO2.90 The beads were coated with

microspheres of the polymer or a thin film of the drug.

3.4. Purification/Fractional Crystallisation

Purification and fractional crystallisation using the GAS and ASES processes exploit

the sharp change in solvent power that occurs as a solvent expands. Solutes of differing

solubility will therefore precipitate at different stages of expansion. This separation

process is analogous to conventional liquid phase fractional crystallisation.

CHAPTER 3

90

The main advantage of fractionation with GAS and ASES over conventional techniques

is that the use of organic solvents is minimised and the recovery of a dry product is

more easily attainable.

3.4.1. GAS as a Purification / Fractional Crystallisation Technique

The first reported work using the GAS process for purification was a solubility study of

β-carotene in toluene or n-butanol expanded with CO2.14 The solubility of β-carotene

was found to decrease with increasing pressure (ie. increasing expansion). A binary

phase equilibrium model was used to predict the experimental results. Fractional

crystallisation was investigated by the separation of β-carotene from oxidation products.

Butanol, toluene and cyclohexanone were chosen as solvents due to the differential

solubility of β-carotene and its oxides. The ratio of solids salting out to solids in

solution decreased exponentially as the feed concentration was reduced. The separation

factor also decreased as the amount of β-carotene in the feed increased. Best separation

was obtained under supercritical conditions rather than at subcritical conditions. Using

two consecutive GAS crystallisations β-carotene content was increased from 73.1% to

90.1%. This was an improvement over conventional liquid recrystallisation techniques

that achieved enrichment from 88.7% to 94.5% in a four-stage process.

Liou and Chang studied the separation of anthracene from crude anthracene using

acetone as the solvent and carbon dioxide as the anti-solvent.130 Anthracene was the

least soluble constituent in the crude acetone mixture, also consisting of phenanthrene,

carbazole and naphthalene. Different feed concentrations were used at pressures of 5.2

and 10.3 MPa. The results showed that total yield of precipitated solids increased with

an increasing feed concentration and at increased pressure. The low solubility of

anthracene in acetone implied that it had the highest degree of saturation in the solution.

Supersaturation and precipitation could therefore be achieved at a lower expansion than

for the other solids. This implied that the purification of anthracene is possible. At 103

atm the purity and yield of anthracene was as high as 90%.

CHAPTER 3

91

Similarly, the separation of a mixture of anthracene and anthraquinone dissolved in

cyclohexanone using carbon dioxide as the anti-solvent has been studied.131,132 The

anthracene and anthraquinone solutions were both prepared with the same degree of

saturation to allow equal opportunity for supersaturation upon expansion. The solubility

of these solutions was measured at 18°C and 40°C. It was discovered that for feed

concentrations having the same level of saturation, the yield of anthracene was always

higher and increased with increasing degree of saturation. A parameter termed the

minimum solubility was defined as the hypothetical feed concentration required for a

zero yield at a fixed temperature. The minimum solubility of anthracene was always

lower than that for anthraquinone at a fixed temperature. It was therefore implied that

anthracene needs less expansion to reach supersaturation and therefore could be

separated by keeping the pressure below the precipitation pressure of anthraquinone.

This was tested by expanding solutions containing mixtures of the two at different feed

concentrations. It was found that higher temperature gave rise to better separation than

lower temperature. After precipitation of solid at the correct conditions, pure anthracene

could be obtained by simple filtration.

Shishikura investigated the separation of citric acid from by-product organic acids in

acetone.133 The separation of citric acid from oxalic acid was studied at the same

concentrations as those expected in a fermentation broth. Citric acid was recovered at a

99.8% to 96.4% purity, whilst oxalic acid remained in solution and did not precipitate

out.

The separation of fractions of licorice root extract was performed by means of a two

step extraction using ethanol as the solvent and carbon dioxide as the anti-solvent.134

Licorice extract was dissolved in ethanol and fed into an extraction column that was

maintained at a low supercritical carbon dioxide to solution ratio (ie. low expansion).

The polar substances that were least soluble precipitated out in the column. The

supernatant was then fed into a second extraction column that had a higher supercritical

carbon dioxide to solution ratio. In the second column the poorly polar fractions were

extracted with the carbon dioxide and the middle polar fractions were concentrated in

solution. In this study the carbon dioxide was acting as an anti-solvent and a poor

CHAPTER 3

92

solvent. The purification of fermented citric acid was further investigated using acetone

as the solvent and carbon dioxide as the anti-solvent. Citric acid was extracted from the

condensed broth with acetone and the residual impurities were removed by carbon

dioxide induced precipitation. The acetone solution was filtered to remove the

impurities and the citric acid was crystallised by further expansion with carbon dioxide.

The residual acetone concentration was below commercially acceptable levels, but still

had a deleterious affect on taste.

The purification of bilirubin extended the use of GAS as a purification process to

medicinal compounds.63 Dimethylsulfoxide was used as the solvent and carbon dioxide

as the anti-solvent. Temperatures ranging from 35°C to 60°C and pressures between

15.0 MPa to 30.6 MPa were used in the experiments. Temperatures above 35°C were

chosen because of the ability of supercritical carbon dioxide to rapidly dry precipitates.

Process time was therefore shortened by operating at supercritical conditions. The

purity of bilirubin in the feed was 30.1%. The precipitated bilirubin had a purity of

greater than 90%. The optimum operating conditions were 40°C and 100 – 15.0 MPa.

The GAS process has been used for the separation of isomers. In a fundamental study a

mixture of ortho- and para-hydroxybenzoic acid (HBA) was separated by dissolution in

methanol and expanding with carbon dioxide.135 Both isomers possess similar

solubilities in methanol, but the solubility of p-HBA in carbon dioxide expanded

methanol is two orders of magnitude lower than for o-HBA. This implies that p-HBA

should precipitate out first in expanded methanol solutions. A precipitate containing

99% p-HBA was obtained after a single step from a 50:50 o-HBA : p-HBA methanol

feed solution. It was also observed that the presence of a solute in the solvent-anti-

solvent system significantly changed the volume expansion of the solvent. The volume

expansion of methanol-carbon dioxide system did not significantly change upon

addition of o-HBA acid into the system. However, when p-HBA was added a second

liquid phase formed.

Phenanthrene and naphthalene were fractionally precipitated from an equimolar toluene

solution with carbon dioxide.136 Phenanthrene could be collected as 98.5% pure

CHAPTER 3

93

precipitate. Under no conditions was the precipitation of naphthalene noted and it was

only recovered as an 87% pure liquid phase. In order to explain the contrasting phase

behaviour of these two solutes, the quaternary system was modelled using the structure

outlined by Kikic and co-workers.137,138

The liquid solutes lecithin and coriander essential oil have been separated from soya oil

and coriander triglycerides respectively using the GAS process.139 The GAS process

was implemented to reduce the large number of refining steps normally required to

obtain pure product conventionally. Using the GAS process 90% separation could be

achieved in a single step.

The fractionation of the protein systems lysozyme-ribonuclease, alkaline-phosphatase

and trypsin-catalse dissolved in DMSO and precipitated with CO2 have been studied

using the GAS process.140 Pure lysozyme, alkaline phosphatase and trypsin could be

obtained from their respective mixtures in a single step using the GAS process. Gas

anti-solvent processing of ribonuclease, lysozyme and trypsin resulted in 20 to 30% loss

of biological activity. Almost total loss of the biological activity of alkaline phosphatase

occurred.

In a precipitation study on polyamides it was found that the polyamide could be

separated from LiCl, a salt used to solubilise the polyamide, using the GAS process.141

LiCl was found to precipitate at a higher pressure than the polyamide and by remaining

at pressures below this pressure pure polyamide could be obtained.

The inorganic salts BaCl2 and NH4Cl could be separated by dissolving a mixture of the

two solutes in DMSO and precipitating with CO2.70,71

3.4.2. ASES as a Fractionation Technique

The first application of ASES as a fractionation technique was the separation of trans-β-

carotene and total β-carotene from raw β-carotene.14 After processing, the trans-β-

carotene content of the powder increased from 74.8 to 81.6%.

CHAPTER 3

94

Catchpole and Bergmann used the ASES process to fractionate a mixture of lecithin

from soya oil in hexane.142 A one hundred percent efficient separation was achieved at

operating conditions of only 25°C and 5.0 MPa. The separation process was conducted

continuously for up to 7 hours. The degree of separation was a function of pressure,

temperature, and to a lesser extent, the ratio of solvent to carbon dioxide. Higher

extraction of lecithin was found at higher pressure.

Racemic mixtures of (R)-2,2'-Binaphthyl-1,1'-diamine and (S)-2,2'-Binaphthyl-1,1'-

diamine were separated by crystallising with (R)-camphorsulfonic acid in CO2.143 The

efficiency of the separation was found to decrease with increasing pressure. The

relationship between resolution and temperature was more complex with an increase in

resolution with increasing temperature observed at lower pressures and the opposite

effect observed at higher pressures. At optimum conditions a resolution of 93% could

be obtained in a single crystallisation.

Nylon was successfully separated from carpet waste using the ASES technique.144 A

slight increase in molecular weight of the recovered nylon was observed compared to

the original nylon. The increase in molecular weight was attributed to the lower

molecular weight fractions being washed out with the expanded solution.

Lecithin has been separated from de-oiled egg yolk by injecting a hexane solution of the

egg yolk into liquid CO2.145 Sub-critical processing was found to be more successful

than supercritical conditions. The potential of scaling up the technique was

demonstrated from a cost point of view, with the cost of the fractionation being well

within the limits for nutritional uses.

In a co-precipitation study, the use of ASES as a fractionation technique was

demonstrated with the separation of paracetamol and ascorbic acid occurring at lower

pressures.128 The goal of the study was to form the composite, but at lower pressure,

only paracetamol precipitated.

CHAPTER 3

95

3.5. Experimental Techniques

3.5.1. GAS Experimental Techniques

A schematic diagram of a typical GAS set up is shown in Figure 3.1. A GAS apparatus

typically consists of an anti-solvent supply (A), from which preheated (C) anti-solvent

is delivered into the precipitation vessel (F) by means of a pump (B). The flow of anti-

solvent is controlled by a needle valve (D). Valve (E) permits anti-solvent to be added

from the bottom or the top of the precipitation vessel. Once precipitation is complete

filtration of the solution through a filter (G) allows the precipitate to be collected. The

flow of solution out of the precipitation vessel is controlled by a valve (H). The solvent

and anti-solvent are separated in the solvent trap (I) and the mass of anti-solvent may be

measured by a gas flow meter (J).

The standard GAS experimental procedure begins with the loading of organic solution

into the crystallisation vessel. A filter at the bottom of the vessel is able to hold the

solution within the chamber. The crystallisation vessel is then charged with gaseous

anti-solvent through the filter or a capillary tube submerged in the fluid. The pressure is

controlled by a back-pressure regulator, adjustment of the gas flow using the micro-

metering valve, or the pump. Mixing of the gas and solution is achieved by mechanical

stirrer, typically magnetic, or by using the frit as a gas sparger. The expansion of the

solution and subsequent precipitation is observed through a view glass. The level of

expansion can be measured by observing the solution level against calibrated markings

on the glass, or by using a travelling telescope or cathetometer. Once crystallisation is

complete, anti-solvent is delivered at constant pressure from the top of the vessel to

force the expanded organic solution through the filter. Once all of the liquid has been

filtered off, the precipitate is washed with fresh anti-solvent to remove residual solvent.

The mass of dense gas used in this stage of the experiment can be measured using the

gas flow meter. The pressure is reduced and the precipitate collected for analysis.

Most of the experimental designs rely on the gas bubbling through the solution for

mixing. This method of mixing has been proved to be as efficient as a magnetic stirrer

for small volume crystallisation vessels.51 In larger vessels, such as those exceeding 1L,

CHAPTER 3

96

F

A

D

EG

H

I

J

C

TP

B

Figure 3.1 Diagram of a typical GAS experimental apparatus.

CHAPTER 3

97

an efficient mixer is required in order to provide a homogenous mixture. Dead volume

should be avoided in the design of the precipitation chamber.

It has been observed that significant temperature changes can occur during the

pressurisation of the precipitation vessel due to the exothermic nature of gas

compression.51,64,65 Temperature increases of up to 20°C have been noted in the GAS

process during the rapid addition of carbon dioxide. To minimise these temperature

fluctuations, the crystallisation vessel may be initially pressurised with an inert gas.65

The anti-solvent may then be added from the bottom of the crystalliser at this elevated

pressure as a compressed gas or liquid. The pressure is kept constant by releasing the

inert gas through a micro-metering valve or backpressure regulator.

A GAS apparatus used in purification or fractionation is very similar to that previously

described for micronisation. The experiments are conducted by loading the crystalliser

with the solution to be extracted and then pressurising the solution with the anti-solvent

to the desired pressure. The system is then left to equilibrate until precipitation has

ceased. The liquid phase is then removed at constant pressure, either by forcing it

through the filter or by drawing it off with a capillary. The liquid phase is separated

from the gas anti-solvent in the cold trap. The precipitate is then washed with anti-

solvent. The desired compounds may be in the precipitated product or in the liquid. A

multi-stage GAS separator such as that used for the concentration of the antimicrobial

substances in licorice and in the purification of citric acid134,146 consists of a series of

precipitation vessels that sequentially step up in pressure, precipitating different solutes

in each vessel.

3.5.2. ASES Experimental Techniques

A schematic diagram of a typical ASES set up is shown in Figure 3.2. A solution, held

in a solution reservoir (A), is delivered by pump (B) into the ASES precipitation vessel

(C) via a dispersing device such as a nozzle or orifice (D). The precipitation chamber is

pre-pressurised with anti-solvent (E) by a secondary pump (F), which also controls the

system pressure during the course of an experiment. Precipitate is usually collected by

CHAPTER 3

98

J

A

BD

E

H

I

T

T

P

G

CF

Figure 3.2 Diagram of a typical ASES experimental apparatus.

CHAPTER 3

99

means of a filtration device (G) through which the solvent and anti-solvent solution can

pass before being depressurised across a valve or orifice (H). The anti-solvent and liquid

solvent may be separated (J) and recycled at this point. The whole apparatus is

immersed in a temperature-controlled water bath or oven to maintain constant operating

temperatures (I).

In most ASES research the anti-solvent phase is present as a single phase. Spraying a

solution into a two phase anti-solvent environment has been investigated in attempts to

reduce particle agglomeration.97,126

The basic experimental apparatus has been modified to tailor the process to specific

applications. For micronisation applications, nozzles of different diameters and designs

have been employed to control and/or reduce particle size and distribution of

precipitated products. Nozzle and orifice sizes employed in the studies to date have

ranged from 20 µm to 500 µm. Ultrasonic69,100, high-energy89,90, double81,95,105,123

and triple80,82,119,120 nozzle arrangements have been incorporated to further improve

the quality of micronised products by increasing jet breakup, reducing particle

agglomeration and controlling the extent of mixing and solute-solvent interaction at the

point of precipitation.

Multiple nozzle assemblies have not only been employed for micronisation applications,

but also for microencapsulation applications. Double and triple nozzle arrangements

have allowed more control over the order of precipitation of each solute to improve

encapsulation of pharmaceuticals precipitated with the process.81,91,95

The design of the precipitation vessel has also been modified to increase the efficiency

and control of the process. Agitated precipitation vessels with a baffle cone87, co-

current and countercurrent systems104,110,111 have been investigated to aid in mixing

and product separation.

The experimental scale of the majority of ASES studies has involved high pressure

precipitation vessels of about 50-100 ml in volume with sight gauge to enable visual

CHAPTER 3

100

observations during operation. The process has been scaled up to a 50 litre precipitation

vessel that consumed up to 40 kg.h-1 anti-solvent and had the ability to recycle

solvent.103

CHAPTER 3

101

3.6. References

1. Müller, B. W.; Waßmus, W.; "ASES - A New Production Technique for Polymeric

Microparticles," Acta Pharm. Technol. 1990, 36, 35 S.

2. Bleich, J.; Kleinebudde, B. W.; Müller, B. W.; "Influence Of Gas Density And

Pressure On Microparticles Produced With The ASES Process," Int. J. Pharm. 1994,

106, 77.







5. Chou, Y.; Tomasko, D. L.; "GAS Crystallization of Polymer-Pharmaceutical

Composite Particles," The 4th International Symposium on Supercritical Fluids,

Sendai, Japan, 1997, A, 55.

6. Reverchon, E.; Della Porta, G.; Di Trolio, A.; Pace, S.; "Supercritical Anti-Solvent

Precipitation of Nanoparticles of Superconductor Precursors," Ind. Eng. Chem. Res.

1998, 37, 952.

7. Gao, Y.; Mulenda, T.; Shi, Y.; Yuan, W.; "Fine Particles Preparation of Red Lake C

Pigment by Supercritical Fluid," J. Supercrit. Fluids 1998, 369.

8. Debenedetti, P. G.; Lim, G. B.; Prud'homme, R. K.; "Formation of Protein

Microparticles by Anti-Solvent Precipitation," Ep 542314, 1993.

9. Winters, M. A.; Knutson, B. L.; Debenedetti, P. G.; Sparks, H. G.; Przybycien, T.

M.; Stevenson, C. L.; Prestrelski, S. J.; "Precipitation of Proteins in Supercritical

Carbon Dioxide," J. Pharm. Sci. 1996, 85, 586.

10. Winters, M. A.; Debenedetti, P. G.; Carey, J.; Sparks, H. G.; Sane, S. U.;

Przybycien, T. M.; "Long-Term and High-Temperature Storage of Supercritically-

Processed Microparticulate Protein Powders," Pharm. Res. 1997, 14, 1370.

11. Yeo, S. D.; Lim, G. B.; Debenedetti, P. G.; Bernstein, H.; "Formation of

Microparticulate Protein Powders Using a Supercritical Fluid Anti-Solvent,"

Biotechnol. Bioeng. 1993, 41, 341.

CHAPTER 3

102



Polym. Sci. 1993, 50, 1929.





Prog. 1991, 7, 275.

15. Goodman, L. S.; Gilman, A.; The Pharmacological Basis of Therapeutics; 7th ed.;

Macmillan Publishing Co., Inc.: New York, 1985, 1839.

16. Breimer, D. D.; "Future Challanges for Drug Delivery Research," Adv. Drug Deliv.

Rev. 1998, 33, 265.

17. Jain, K. K.; "Strategies and Technologies for Drug Delivery Systems," TiPS 1998,

19, 155.

18. Hillery, A. M.; "Microparticulate Delivery Systems: Potential Drug/Vaccine

Carriers Via Mucosal Routes," PSTT 1998, 1, 69.

19. Thibert, R.; Tawashi, R.; "Micronization of Pharmaceutical Solids," Microspheres

Microencapsules Liposomes 1999, Volume : 1 (Preparation and Chemical

Applications), 327.

20. Atkinson, R. M.; Bedford, C.; Child, K. J.; E.G., T.; "The Effect of Griseofulvin

Particle Size on Blood Levels in Man," Antibiotics and Chemotherapy 1962, XII,

232.

21. Johnson, B. F.; O'Grady, J. O.; Bye, C.; "The Influence of Digoxin Particle Size on

Absorption of Digoxin and the Efefct of Propantheline and Metoclopromide," Br. J.

Clin. Pharmac. 1978, 5, 465.

22. Asbury, M. J.; McInnes, G. T.; Ramsay, L. E.; Shelton, J. R.; "Improvement in the

Bioavailability and Activity of Spironolactone by Micronisation," Proceedings of the

B.P.S., 1981, 270P.

23. Kondo, N.; Iwao, T.; Masuda, H.; Yamanouchi, K.; Ishihara, Y.; Yamada, N.; Haga,

T.; Ogawa, Y.; Yokoyama, K.; "Improved Oral Absorption of a Poorly Water-

Soluble Drug, HO-221 by Wet-Bead Milling Producing Particles in Submicron

Region," Chem. Pharm. Bull. 1993, 41, 737.

CHAPTER 3

103

24. Grau, M. J.; Kayser, O.; Müller, R. H.; "Nanosuspensions of Poorly Soluble Drugs -

Reproducibility of Small Scale Production," Int. J. Pharm. 2000, 196, 155.

25. Chen, H.; Langer, R.; "Oral Particularte Delivery: Status and Future Trends," Adv.

Drug Deliv. Rev. 1998, 339.

26. Andrianov, A. K.; Payne, L. G.; "Polymeric Carriers for Oral Uptake of

Microparticulates," Adv. Drug Deliv. Rev. 1998, 34, 155.

27. Brayden, D. J.; O'Mahoney, D. J.; "Novel Oral Drug Delivery Gateways for

Biotechnology Products: Polypeptides and Vaccines," PSTT 1998, 1, 291.

28. Ponchel, G.; Montisci, M.; Dembri, A.; Durrer, C.; Duchêne, D.; "Mucoadhesion of

Colloidal Particulate Systems in the Gastro-Intestinal Tract," Eur. J. Pharm.

Biopharm. 1997, 44, 25.

29. O' Hagan, D. T.; "Microparticles and Polymers for the Mucosal Delivery of

Vaccines," Adv. Drug Deliv. Rev. 1998, 34, 305.

30. Yeh, P.; Ellens, H.; Smith, P. L.; "Physiological Considerations in the Design of

Particulate Dosage Forms for Oral Vaccine Delivery," Adv. Drug Deliv. Rev. 1998,

34, 123.

31. Allémann, E.; Leroux, J.; Gurny, R.; "Polymeric nano- and Microparticles for the

Oral Delivery of Peptides and Peptidomimetics," Adv. Drug Deliv. Rev. 1998, 34,

171.

32. Moren, F.; "Aerosol Dosage Forms and Formulations," Aerosols in Medicine:

Principles, Diagnosis and Therapy, ; Moren, F. and Newhouse, M. T., Ed.; Elsevier:

New York, 1985, 261.

33. Patton, J. S.; "Pulmonary Delivery of Drugs for Bone Disorders," Adv. Drug Deliv.

Rev. 2000, 42, 239.

34. Zimmer, A.; Kreuter, J.; "Microspheres and Nanoparticles Used in Ocular Delivery

Systems," Adv. Drug Deliv. Rev. 1995, 16, 61.

35. Ding, S.; "Recent Developments in Ophthalmic Drug Delivery," PSTT 1998, 1, 328.

36. Takenaga, M.; Serizawa, Y.; Azechi, Y.; Ochiai, A.; Kosaka, Y.; Igarashi, R.;

Mizushima, Y.; "Microparticle Resins as a Potential Nasal Drug Delivery System for

Insulin," J. Controlled Release 1998, 52, 81.

37. Benoit, J.; Faisant, N.; Venier-Julienne, M.; Menei, P.; "Development of

Microspheres for Nurelogical Disorders: From Basics to Clinical Applications," J.

Controlled Release 2000, 65, 285.

CHAPTER 3

104

38. Austin, L. G.; Rogers, R. S. C.; "Powder Technology in Industrial Size Reduction,"

Powder Technol. 1985, 42, 90.

39. Lowrison, G. C.; "Crushing and Grinding: The Size Reduction of Solid Materials,"

Butterworth: London, 1974.

40. Snow, R. H.; "Annual Review of Size Reduction - 1975," Powder Technol. 1975,

23, 31.

41. Jarzebski, A. B.; Malinowski, J. J.; "Potentials and Prospects for Application of

Supercritical Fluid Technology in Bioprocessing," Process Biochem. (Oxford) 1995,

30, 343.

42. Weidner, E.; Knez, Z.; Novak, Z.; "Process for the Production of Particles or

Powders," U.S. 6,056,791, 2000.

43. Tom, J.; Debenedetti, P.; "Particle Formation with Supercritical Fluids - A Review,"

J. Aerosol Sci. 1991, 22, 555.

44. Bush, P. J.; Pradhan, D.; Ehrlich, P.; "Lamellar Structure and Organization in

Polyethylene Gels Crystallized from Supercritical Solution in Propane,"




46. Tom, J. W.; Debenedetti, P. G.; "Formation of Bioerodible Polymeric Microspheres

and Microparticles by Rapid Expansion of Supercritical Solutions," Biotechnol.

Prog. 1991, 7, 403.

47. Matson, D. W.; Peterson, R. C.; Smith, R. D.; "Production of Fine Powders by the

Rapid Expansion of Supercritical Fluid Solutions," Advances in Ceramics 1987, 21,

109.

48. Weidner, E.; Knez, Z.; Novak, Z.; "PGSS (Particles from Gas Saturated Solutions) -

A New Process for Powder Generation," Proceedings of the 3rd International

Symposium on Supercritical Fluids, Strasbourg, France, 1994, Tome 3, 229.

49. Weidner, E.; Steiner, R.; Knez, Z.; "Powder Generation from Polyethyleneglycols

with Compressed Fluids," Process Technol. Proc. 1996, 12, 223.

50. Weidner, E.; "Powder Generation by High Pressure Spray Processes," Proccedings

of the International Meeting of the GVC-Fachausschuß

,,Hochdruckverfahrenstechnik'', Karlsruhe, Germany, 1999, 217.

CHAPTER 3

105


Solvent Recrystallization of RDX: Formation of Ultra-Fine Particles of a Difficult-

to-Comminute Explosive," J. Supercrit. Fluids 1992, 5, 130.

52. Krukonis, V. J.; Gallagher, P. M.; Coffey, M. P.; "Gas Anti-Solvent

Recrystallization Process," U.S. 5,360,478, 1994.





54. Cai, J.-G.; Liao, X.-C.; Zhou, Z.-Y.; "Mocroparticle Formation and Crystallization

Rate of HMX Using Supercritical Carbon Dioxide Anti-Solvent Recrystallization,"

The 4th International Symposium on Supercritical Fluids, Sendai, Japan, 1997, A,

23.

55. Teipel, U.; Gerber, P.; Foerter-Barth, U.; Niehaus, M.; Krause, H.; "Formation of

Particles of Explosives with Supercritical Fluids," Int. Annu. Conf. ICT 1996, 27th,

7.1.



87, 96.

57. Benedetti, L.; Bertucco, A.; Pallado, P.; "Production Of Micronic Particles Of

Biocompatible Polymer Using Supercritical Carbon Dioxide," Biotechnol. Bioeng.

1997, 53, 232.

58. Yeo, S. D.; Debenedetti, P. G.; Radosz, M.; Schmidt, H. W.; "Supercritical Anti-

Solvent Process for Substituted Para-Linked Aromatic Polyamides: Phase

Equilibrium and Morphology Study," Macromolecules 1993, 26, 6207.

59. Gallagher-Wetmore, P.; Coffey, M. P.; Krukonis, V.; "Application of Supercritical

Fluids in Recrystallization: Nucleation and Gas Anti-Solvent (GAS) Techniques,"

Respiratory Drug Delivery 1994, IV, 287.

60. Gallagher-Wetmore, P.; Coffey, M. P.; Krukonis, V.; "Recrystallization Using

Supercritical Fluids: Novel Techniques for Particle Modification," 1994, 162.



Process," AIChE J. 1996, 42, 431.

CHAPTER 3

106

62. Cocero, M. J.; Ferrero, S.; Vicente, S.; "GAS Crystallization of β-Carotene from

Ethyl Acetate Solutions Using CO2 as Anti-Solvent," Proceedings of the 5th

International Symposium on Supercritical Fluids, Atlanta, Georgia, 2000.

63. Jianguo, C.; Zhongwen, Y.; Zhanyun, Z.; "Purification of Bilirubin and Micro-

Particle Formation with Supercritical Fluid Anti-Solvent Precipitation," Chinese J. of

Chem. Eng. 1996, 4, 257.

64. Kitamura, M.; Yamamoto, M.; Yoshinaga, Y.; Masuoka, H.; "Crystal Size Control

of Sulfathiazole Using High Pressure Carbon Dioxide," J. Cryst. Growth 1997, 178,

378.

65. Wubbolts, F. E.; Kersch, C.; van Rosmalen, G. M.; "Semi-Batch Precipitation of

Acetaminophen from Ethanol with Liquid Carbon Dioxide at a Constant Pressure,"

Proceedings of the 5th Meeting on Supercritical Fluids, Nice, France, 1998, 249.

66. Tan, C.-S.; Chang, W.-W.; "Precipitation of Polystyrene from Toluene with HFC-

134a by the GAS Process," Ind. Eng. Chem. Res. 1998, 37, 1821.




68. Thiering, R.; Dehghani, F.; Dillow, A.; Foster, N. R.; "The Influence of Operating

Conditions on the Dense Gas Precipitation of Model Proteins," J. Chem. Technol.





70. Yeo, S.-D.; Choi, J.-H.; Lee, T.-J.; "Crystal Formation of BaCl2 and NH4Cl Using a

Supercritical Fluid Anti-Solvent," J. Supercrit. Fluids 2000, 16, 235.




72. Reverchon, E.; Porta, G. D.; Sannino, D.; Lisi, L.; Ciambelli, P.; "Supercritical Anti-

Solvent Precipitation: a Novel Technique to Produce Catalyst Precursors. Preparation

and Characterization of Samarium Oxide Nanoparticles," Stud. Surf. Sci. Catal. 1998,

118, 349.

CHAPTER 3

107





Microballoons by Precipitation with a Vapour-Liquid Compressed Fluid Anti-


75. Tom, J. W.; Lim, G.; Debenedetti, P. G.; Prud'homme, R. K.; "Applications of

Supercritical Fluids in the Controlled Release of Drugs," Supercritical Fluid

Engineering Science Fundamentals and Applications, ; Kiran, E. and Brennecke, J.

F., Ed. 1993; Vol. 514, 238.

76. Debenedetti, P. G.; Tom, J. W.; Yeo, S.-D.; Lim, G.-B.; "Application of

Supercritical Fluids for the Production of Sustained Delivery Devices," J. Controlled

Release 1993, 24, 27.

77. Debenedetti, P. G.; Lim, G.; Prud'Homme, R. K.; "Preparation of Protein

Microparticles by Supercritical Fluid Precipitation," U.S. 6,063,910, 2000.

78. Yeo, S.-D.; Debenedetti, P. G.; Patro, S. Y.; Przybycien, T. M.; "Secondary

Structure Characterization of Microparticulate Insulin Powders," J. Pharm. Sci. 1994,

83, 1651.



Bioeng. 2000, 67, 457.

80. Sloan, R.; Hollowood, M. E.; Humpreys, G. O.; Ashraf, W.; York, P.; "Supercritical

Fluid Processing: Preparation of Stable Protein Particles," Proceedings of the 5th

Meeting on Supercritical Fluids, Nice, France, 1998, 301.

81. Bustami, R. T.; Chan, H.; Dehghani, F.; Foster, N. R.; "Generation of Protein

Micro-Particles Using High Pressure Modified Carbon Dioxide," Proceedings of the

5th International Symposium on Supercritical Fluids, Atlanta, Georgia, 2000.

82. Sloan, R.; Tservistas, M.; Hollowood, M. E.; Sarup, L.; Humphreys, G. O.; York,

P.; Ashraf, W.; Hoare, M.; "Controlled Particle Formation of Biological Material

Using Supercritical Fluids," Proceedings of the 6th Meeting on Supercritical Fluids,

Nottingham, United Kingdom, 1999, 169.

CHAPTER 3

108

83. Kulshreshtha, A. K.; Smith, G. G.; Anderson, S. D.; Krukonis, V. J.; "Process for

Sizing Prednisolone Acetate using a Supercritical Fluid Anti-Solvent," US 5803966,

1998.

84. Reverchon, E.; Della Porta, G.; Flaivene, M. G.; "Process Parameters and

Morphology in Amoxicillin Micro and Submicro Particles Generation by

Supercritical Anti-Solvent Precipitation," J. Supercrit. Fluids 2000, 17, 239.

85. Reverchon, E.; Della Porta, G.; Falivene, M. G.; "Process Parameters Controlling

the Supercritical Anti-Solvent Micronisation of Some Antibiotics," Proceedings of

the 6th Meeting on Supercritical Fluids, Nottingham, United Kingdom, 1999, 157.

86. Reverchon, E.; Porta, G. D.; "Production of Antibiotic Micro- and Nano-Particles by

Supercritical Anti-Solvent Precipitation," Powder Technol. 1999, 106, 23.



1995, 41, 2476.

88. Shekunov, B. Y.; Hanna, M.; York, P.; "Crystallization Process in Turbulent

Supercritical Flows," J. Cryst. Growth 1999, 198/199, 1345.

89. Subramaniam, B.; Saim, S.; Rajewski, A.; Stella, V.; "Methods for Particle

Micronization and Nanonization by Recrystallization from Organic Solutions

Sprayed into a Compressed Anti-Solvent," U.S. 5,874,029, 1999.

90. Subramaniam, B.; Saim, S.; Rajewski, R. A.; Stella, V.; "Methods for a Particle

Precipitation and Coating Using Near-Critical and Supercritical Antisolvents," US

5833891, 1998.



92. Wubbolts, F. E.; Bruinsma, O. S. L.; de Graauw, J.; van Rosmalen, G. M.;

"Continuous Gas Anti-Solvent Crystallisation of Hydroquinone from Acetone Using

Carbon Dioxide," The 4th International Symposium on Supercritical Fluids, Sendai,

Japan, 1997, A, 63.



1999, 198/199, 767.

94. Thiering, R.; Charoenchaitrakool, M.; Sze-Tu, L.; Dehghani, F.; Dillow, A. K.;

Foster, N. R.; "Crystallization of Para-Hydroxybenzoic Acid by Solvent Expansion

CHAPTER 3

109

with Dense Carbondioxide," Proceedings of the 5th Meeting on Supercritical Fluids,

Nice, France, 1998, 291.

95. Sze Tu, L.; Dehghani, F.; Dillow, A. K.; Foster, N. R.; "Applications of Dense

Gases in Pharmaceutical Processing," Proceedings of the 5th Meeting on

Supercritical Fluids, Nice, France, 1998, 263.





98. Tan, C.-S.; Lin, H.-Y.; "Precipitation of Polystyrene by Spraying Polystyrene-

Toluene Solution into Compressed HFC-134a," Ind. Eng. Chem. Res. 1999, 38, 3898.

99. Johnston, K. P.; Luna-Barcenas, G.; Dixon, D.; Mawson, S.; "Polymeric Materials

by Precipitation with a Compressed Fluid Anti-Solvent," Proceedings of the 3rd

International Symposium on Supercritical Fluids, Strasbourg, France, 1994, 359.

100. Randolph, T. W.; Randolph, A. D.; Mebes, M.; Yeung, S.; "Sub-Micrometer-Sized

Biodegradable Particles of Poly(L-Lactic Acid) via the Gas Anti-Solvent Spray

Precipitation Process," Biotechnol. Prog. 1993, 9, 429.

101. Ruchatz, F.; Kleinebudde, P.; Müller, B. W.; "Residual Solvents in Biodegradable

Microparticles. Influence of Process Parameters on the Residual SOlvent in

Microparticles Produced by the Aerosol Solvent Extraction System (ASES) Process,"

J. Pharm. Sci. 1997, 86, 101.



1995, 12, 1211.

103. Thies, J.; Müller, B. W.; "Size Controlled Production of Biodegradable

Microparticles with Supercritical Gases," Eur. J. Pharm. Biopharm. 1998, 45, 67.

104. Bleich, J.; Müller, B. W.; Waßmus, W.; "Aerosol Solvent Extraction System - A

New Microparticle Production Technique," Int. J. Pharm. 1993, 97, 111.

105. Mawson, S.; Kanakia, S.; Johnston, K. P.; "Coaxial Nozzle for Control of Particle

Morphology in Precipitation with a Compressed Fluid Anti-Solvent," J. Appl. Polym.

Sci. 1997, 64, 2105.

CHAPTER 3

110

106. Mawson, S.; Johnston, K. P.; Betts, D. E.; McClain, J. B.; DeSimone, J. M.;

"Stabilized Polymer Microparticles by Precipitation with a Compressed Fluid Anti-

Solvent. 1. Poly(fluoro acrylates)," Macromolecules 1997, 30, 71.

107. Mawson, S.; Yates, M. Z.; O'Neill, M. L.; Johnston, K. P.; "Stabilized Polymer

Microparticles by Precipitation with a Compressed Fluid Anti-Solvent. 2.

Poly(propylene oxide)- and Poly(butylene oxide)-Based Copolymers," Langmuir

1997, 13, 1519.

108. Reverchon, E.; Della Porta, G.; De Rosa, I.; Subra, P.; Letourneur, D.;

"Biopolymers Micronisation by Supercritical Anti-Solvent Precipitation: the

Influence of Some Process Parameters," Fifth Conference on Supercritical Fluids

and their Applications, Garda (Verona), 1999, 473.

109. Reverchon, E.; De Rosa, I.; Della Porta, G.; "Effect of Process Parameters on the

Supercritical Anti-Solvent Precipitation of Microspheres of Natural Polymers,"

Dipartimento Ingegneria Chimica Alimentare,Universita Salerno,Fisciano,Italy.,

1999.

110. Heater, K. J.; Tomasko, D. L.; "Processing of Epoxy Resins Using Carbon Dioxide

as an Anti-Solvent," J. Supercrit. Fluids 1998, 14, 55.



112. Gao, Y.; Mulenda, T. K.; Shi, Y.-F.; Yuan, W.-K.; "Fine Particles Preparation of

Red Lake Pigment by Supercritical Fluid," The 4th International Symposium on

Supercritical Fluids, Sendai, Japan, 1997, A, 31.

113. Hong, L.; Bitemo, S. R.; Gao, Y.; Yuan, W.; "Precipitation of Microparticulate

Organic Pigment Powders by Supercritical Anti-Solvent (SAS) Process,"

Proceedings of the 5th International Symposium on Supercritical Fluids, Atlanta,

Georgia, 2000.

114. Reverchon, E.; Celano, C.; Della Porta, G.; Di Trolio, A.; Pace, S.; "Supercritical

Anti-Solvent Precipitation: A New Technique for Preparing Submicronic Yttrium

Powders to Improve YBCO Superconductors," J. Mater. Res. 1998, 13, 284.

115. Reverchon, E.; Della Porta, G.; Sannino, D.; Ciambelli, P.; "Supercritical Anti-

Solvent Precipitation of Nanoparticles of a Zinc Oxide Precursor," Powder Technol.

1999, 102, 127.

CHAPTER 3

111

116. Chattopadhyay, P.; Gupta, R. B.; "Supercritical CO2 Based Production of

Fullerene Nanoparticles," Proceedings of the 5th International Symposium on


117. Magnan, C.; Commenges, N.; Badens, E.; Charbit, G.; "Fine Phospholipid

Particles Formed by Precipitation with a Compressed Fluid Anti-Solvent,"

Proccedings of the International Meeting of the GVC-Fachausschuß

,,Hochdruckverfahrenstechnik'', Karlsruhe, Germany, 1999, 223.

118. Magnan, C.; Badens, E.; Commenges, N.; Charbit, G.; "Soy Lecithin

Micronization by Precipitation with a Compressed Fluid Anti-Solvent - Influence of

Process Parameters," Fifth Conference on Supercritical Fluids and their

Applications, Garda (Verona), 1999, 479.

119. Hanna, M.; York, P.; Yu. Shekunov, B.; "Control of the Polymeric Forms of a

Drug Substance by Solution Enhanced Dispersion by Supercritical Fluids (SEDS),"

Proceedings of the 5th Meeting on Supercritical Fluids, Nice, France, 1998, Tome 1.

120. Hanna, M.; York, P.; "Method and Apparatus for the Formation of Particles," U.S.

6,063,138, 2000.

121. Palakodaty, S.; York, P.; Hanna, M.; Pritchard, J.; "Crystallization of Lactose

Using Solution Enhanced Dispersion by Supercritical Fluids (SEDS) Technique,"

Proceedings of the 5th Meeting on Supercritical Fluids, Nice, France, 1998, Tome 1,

275.




123. Gilbert, D. J.; Palakodaty, S.; Sloan, R.; York, P.; "Particle Engineering for

Pharmaceurical Applications - A Process Scale Up," Proceedings of the 5th


124. Bertucco, A.; Pallado, P.; "Understanding Gas Anti-Solvent Processes of

Biocompatible Polymers and Drugs with Supercritical CO2," Proccedings of the

International Meeting of the GVC-Fachausschuß ,,Hochdruckverfahrenstechnik'',

Karlsruhe, Germany, 1999, 231.



CHAPTER 3

112



126. Young, T. J.; Johnston, K. P.; Mishima, K.; Tanaka, H.; "Encapsulation of

Lysozyme in a Biodegradable Polymer by Precipitation with a Vapor-over-Liquid

Anti-Solvent," J. Pharm. Sci. 1999, 88, 640.

127. Elvassore, N.; Bertucco, A.; Caliceti, P.; "Production of Protein-Polymer Micro-

Capsules by Supercritical Anti-Solvent Techniques," Proceedings of the 5th


128. Weber, A.; Tschernjaew, J.; Kummel, R.; "Co-Precipitation with Compressed

Antisolvents for the Manufacture of Microcomposites," Proceedings of the 5th

Meeting on Supercritical Fluids, Nice, France, 1998, Tome 1, 243.

129. Engwicht, A.; Girreser, U.; Müller, B. W.; "Critical Properties of Lactide-co-

Glycoloid Polymers for the Use in Microparticle Preparation by the Aerosol Solvent

Extraction System," Int. J. Pharm. 1999, 185, 61.

130. Liou, Y.; Chang, C. J.; "Separation of Anthracene from Crude Anthracene Using

Gas Anti-Solvent Recrystallization," Sep. Sci. Technol. 1992, 27, 1277.

131. Chang, C. J.; Liou, Y.; "Purification of Polycyclic Aromatic Compounds Using

Salting-Out Separation in High-Pressure Carbon Dioxide," J. Chem. Eng. Jpn. 1993,

26, 517.

132. Chang, C. J.; Liou, Y.; Lan, W. J.; "Relative Supersaturation Ratio and Separation

Factor in Crystallization with High Pressure CO2," Can. J. Chem. Eng. 1994, 72, 56.

133. Shishikura, A.; "Purification of Organic Acids by Gas Anti-Solvent

Crystallization," Dev. Food Eng., Proc. Int. Congr. Eng. Food, 6th 1994, Pt. 2, 858.

134. Shishikura, A.; "Applications of Compressed Carbon Dioxide in the Separation

Process of Foodstuffs as a Poor and Anti-Solvent.," The 4th International Symposium


135. Foster, N. R.; Yun, S. L. J.; Dillow, A.; Wells, P. A.; Lucien, F. P.; "A

Fundamental Study of the Gas Anti-Solvent Process," The 4th International

Symposium on Supercritical Fluids, Sendai, Japan, 1997, A, 27.



CHAPTER 3

113








139. Catchpole, O. J.; Hochmann, S.; Anderson, S. R. J.; "Gas Anti-Solvent

Fractionation of Natural Products," Process Technol. Proc. 1996, 12, 309.

140. Winters, M. A.; Frankel, D. Z.; Debenedetti, P. G.; Carey, J.; Devaney, M.;

Przybycien, T. M.; "Protein Purification With Vapor-Phase Carbon Dioxide,"


141. Yeo, S.-D.; Debenedetti, P. G.; Radosz, M.; Giesa, R.; Schmidt, H.-W.;

"Supercritical Anti-Solvent Process for a Series of Substituted Para-Linked Aromatic

Polyamides," Macromolecules 1995, 28, 1316.

142. Catchpole, O. J.; Bergmann, C.; "Continuous Gas Anti-Solvent Fractionation Of

Natural Products," Proceedings of the 5th Meeting on Supercritical Fluids, Nice,

France, 1998, 257.

143. Kordikowski, A.; York, P.; "Chiral Separation Using Supercritical CO2,"

Proceedings of the 6th Meeting on Supercritical Fluids, Nottingham, United

Kingdom, 1999, 163.

144. Griffith, A. T.; Park, Y.; Roberts, C. B.; "Separation and Recovery of Nylon from

Carpet Waste using a Supercritical Fluid Anti-Solvent Technique," Polymer Plastics

Technology & Engineering 1999, 38, 411.

145. Weber, A.; Nolte, C.; Bork, M.; Kummel, R.; "Recovery of Lecithin from Egg

Yolk-Extracts by Gas Anti-Solvent Crystallization," Proceedings of the 6th Meeting

on Supercritical Fluids, Nottingham, United Kingdom, 1999.

146. Shishikura, A.; Kanamori, K.; Takahashi, H.; Kinbara, H.; "Separation and

Purification of Organic Acids by Gas Anti-Solvent Crystallization," J. Agric. Food.

Chem. 1994, 42, 1993.

CHAPTER 4

114

4. DENSE GASES AS ANTI-SOLVENTS — PARTICLE

FORMATION MECHANISMS AND THE INFLUENCE OF

PROCESS PARAMETERS

4.1. Introduction

The ASES and GAS processes have been shown to be able to produce precipitates of a

variety of particle sizes and morphology. In the majority of circumstances the aim is to

produce fine powders. In a few cases, particularly encapsulation, this is not the case and

larger, alternative morphologies may be preferred. It is not possible at present to predict

particle size and morphology for a system. The result is that numerous preliminary

experiments have to be performed on each system to determine the optimum process

conditions. In part, this is due to the complexity of the physical processes governing the

system. Thermodynamics, hydrodynamics, mass transfer nucleation and crystallisation

thermodynamics change throughout the process and all need to be considered

simultaneously.1 Despite the fact that the processes governing GAS and ASES are

complex, some information regarding the mechanisms involved have been determined

by studying the effect of process parameters on the particles produced. These

mechanisms will be discussed below.

Supersaturation is the key parameter that determines particle size in both the GAS and

ASES processes.2 In the GAS process the rate of creation of supersaturation and the

level of supersaturation is manipulated by the rate of pressurisation. The ASES process

on the other hand is performed at a static pressure normally, at conditions of asymptotic

expansion or in the single-phase miscible region. At conditions of asymptotic expansion

the very high levels of supersaturation are generated. For this reason, the ASES process

has been referred to as an optimised GAS process.2 The GAS technique has been said to

be able to offer more control of the final particle characteristics due to the ability to

control the rate of expansion and therefore supersaturation.3

CHAPTER 4

115

In the GAS process the anti-solvent is added to the solution and is closely related to a

conventional salting out crystallisation. The rate of supersaturation is essentially a

function of the diffusion of the anti-solvent into the solution. The ASES process is more

complex with the level of supersaturation being a function of the diffusion of the anti-

solvent into the solution and the solvent into the anti-solvent. The precipitation is a

combination of both a salting out and evaporation crystallisation. The operating

conditions of the ASES process will determine which diffusion process dominates.

4.2. GAS – The Influence of Process Parameters

Process parameters that influence particle formation in GAS precipitation include the

expansion (pressurisation) rate, seeding by pressure pulse, or the addition of seed

crystals, solvent and anti-solvent type, feed concentration, temperature and agitation.

Each of these parameters can influence the size, size distribution, shape and crystallinity

of the particles formed.

The most influential parameter over particle size is the rate of expansion of the solution.

The rate of expansion determines supersaturation concentration at which nucleation

occurs, therefore effecting the rate of precipitation. The effect of rate of expansion has

been observed in a number of experiments4-8 and some researchers have reported a

twofold decrease in particle size with even relatively small increases in the rate of

expansion. The effect of expansion rate has been described in terms of supersaturation.

The faster the rate of expansion, the greater the supersaturation will be at the point of

nucleation. Thus the rate of nucleation and the number of nuclei will be larger. The final

precipitate size will therefore be small. The size of the particles precipitated by GAS, at

a controlled supersaturation, have been found to increase with a longer growth time.9

Kitamura and co-workers showed that by using a more rapid initial pressurisation rate to

nucleate seed crystals and then returning to lower pressures and slower rates of

pressurisation a sharp reduction in precipitate size could be achieved.5 These smaller

crystals mainly precipitated from the bulk solution and did not grow on the sides of the

container due to the very rapid nucleation. Bimodal size distributions have been

observed for intermediate expansion rates.10,11 At intermediate expansion rates the

CHAPTER 4

116

creation of supersaturation by gas dissolving in the solution can be reduced by the

growth of crystals. During the expansion the supersaturation line may be crossed a

number of times as the two competing effects of nucleation and growth take place. A

number of distinct nucleation events results in a multimodal particle size distribution.11

It has also been observed that the rate of expansion can have an effect on particle

morphology.7,12-14 The uniformity of crystal morphology is dependent on the stability

of the growth environment during expansion. A slow expansion rate provides a stable

environment for crystal growth resulting in a uniformly crystalline solid. Conversely a

fast expansion rate results in a turbulent environment giving rise to particles of less

crystalline nature. Another phenomenon that results in particles of differing

morphology, and is directly related to expansion rate, is the formation of particles at the

gas liquid interface.8,9 At the liquid gas interface, the concentration of anti-solvent in

the liquid phase will be greater resulting in a higher level of supersaturation than in the

bulk liquid phase. At the interface the rate of nucleation is greater resulting in

kinetically favoured crystal forms.

Seeding has been observed to affect particle size and morphology.9 Seeds were induced

by a pressure pulse during the expansion profile which resulted in a period of high

supersaturation generating nucleus particles. The seeds then formed the basis of further

particle growth. It was found that seeding resulted in larger, better packed particles with

a narrower size distribution than precipitate obtained from non-seeded experiments.

Solvent type has been found to have a dramatic effect on particle morphology.3,7,8,13,15-

17 Many differing particle morphologies are obtained for the same material crystallised

from differing solvents. The different solvents provide different environments for

particle growth and, for the reasons listed above, lead to different growth mechanisms

and therefore morphologies. The influence of anti-solvent type has not been explored in

depth. Changing anti-solvent had a minimal effect on the particle size and morphology.6

The concentration of the solid in the organic solvent has been found to have an

inconsistent effect on particle size and morphology and it is not possible to generalise

CHAPTER 4

117

results. Increases in particle size with increasing solute concentration,18 and decreases

in particle size with increasing solute concentration,13,14,19 have been reported. The

morphology of precipitate has been observed to change with solute concentration.7,8,12

Increased solute concentration has resulted in more crystalline precipitate14, and also

less crystalline precipitate.13 A decrease in precipitate aggregation at higher solute

concentrations was also reported.19 Gallagher and co-workers7 noted that concentration

of the solute had little effect when the solution was expanded rapidly.

The effect of temperature on the characteristics of precipitated material has also been

found to be variable. Different behaviour for the effect of temperature has been

observed7,18,20 with some workers reporting no effect with changing temperature.8,19

The inconsistency in these results may be due to the fact that temperature has a series of

effects on the dense gas precipitation mechanism. Solvent/anti-solvent miscibility and

therefore solution expansion, as well as the crystallisation mechanism, are both strong

functions of temperature.

Stirrer speed or solution agitation is an experimental parameter that has been

disregarded by many researchers, since equilibrium conditions have been assumed.

Often mixing is achieved by sparging less dense anti-solvent through the solvent. Both

magnetic stirrers and spargers have been used to more vigorously mix solutions under

pressure and rapidly bring about equilibrium. Note that mixing is most difficult at low

anti-solvent densities. Stirrer speed, though, has also been observed to have an effect on

particle size and size distribution.9 Faster stirrer speeds were found to give smaller

particles with a narrow size distribution due to attrition. At very slow stirrer speeds

inefficient mixing of solvent and anti-solvent results in non-equilibrium particles giving

a larger size distribution.

4.3. ASES The Influence of Process Parameters

There are numerous adjustable parameters associated with the ASES process for the

precipitation of a particular solute. Pressure, temperature, flow rate of anti-solvent, flow

rate of solvent, nozzle type and diameter, concentration, type of solvent and type of

CHAPTER 4

118

anti-solvent are parameters that can be manipulated in ASES. Having multiple process

variables makes the ASES process attractive in that it is possible to manipulate particle

size and morphology by altering process parameters. However, it is a non-trivial task to

compare experimental results from different studies in the literature. It is also

impossible to limit the formation of particles to one type of mechanism as the

thermodynamics, hydrodynamics, mass transfer and ultimately nucleation and particle

growth are strong functions of these variables.

As an introduction it is convenient to start with the simplest experiment, which is the

injection of pure solvent into the anti-solvent. The injection of pure toluene into liquid

phase CO2 was found to result in jet breakup and the formation of small suspended

liquid droplets when using a single channel nozzle by Mawson and co-workers.21 The

droplets were observed as a refractive index discontinuity. By changing to a coaxial

nozzle and injecting toluene into CO2 at the same conditions, the refractive index

gradients were not observed and the conclusion was reached that no droplets were

formed due to rapid mixing.

When pure dichloromethane was injected into supercritical CO2 the droplets formed

were similar in size to the diameter of the capillary (100 µm).1 Rapid mass transfer of

CO2 into the droplet and solvent out of the droplet caused the droplet to swell, and

eventually evaporate, before reaching the bottom of the vessel.

When pure ethanol was sprayed into vapour phase CO2 a jet of fine solvent droplets was

observed.22 These droplets filled the whole vessel and collected as a liquid at the

bottom of the vessel. When injecting ethanol into CO2 at conditions above the mixture

critical point, no mist was seen. Instead the jet could be observed extending 1.5 cm from

the nozzle tip after which it disappeared.

Pure dichloromethane has been injected into a vapour over liquid CO2 system.23 At the

flow rate studied, the dichloromethane did not atomise but fell as large droplets through

the vapour phase and into the liquid phase and quickly dispersed. At temperatures less

CHAPTER 4

119

than –30°C the drops did not disperse immediately but remained intact in the liquid

phase for about 1 cm before dispersing.

These observations indicate that jet breakup is a strong function of the density of the

anti-solvent. At low density, when the anti-solvent is in the vapour phase, the flow rate

and nozzle geometry determine the hydrodynamic mechanism. At high enough flow

rates, and large Weber numbers, atomisation of the liquid jet occurs and small droplets

form in the vapour phase which then fall to the bottom of the vessel and collect as a

liquid phase. At low flow rates, and low Weber numbers, atomisation does not occur

and the solvent falls through the vapour phase as large droplets. At low anti-solvent

density the solubility of the anti-solvent in the solvent is limited by thermodynamics.

Mass transfer of anti-solvent into the solvent droplet is limited. Negligible mass transfer

of solvent into the anti-solvent occurs and the integrity of the droplet is maintained until

the droplet is destroyed upon hitting the bottom of the vessel or the liquid phase at the

bottom of the vessel. It has been shown that for partially miscible systems the surface

tension of the jet rapidly approaches the equilibrium surface tension of the mixture.24

As the density of the anti-solvent is increased, and the single phase region of the

mixture is approached, the hydrodynamics of the system changes. Mass transfer of anti-

solvent into the solvent increases and once the single phase region is reached the solvent

and anti-solvent become totally miscible. The high density of the anti-solvent and the

reduction in surface tension between the solvent and the anti-solvent is expected to

produce very high Weber numbers and therefore very small droplets. If the surface

tension between the solvent and anti-solvent decreases rapidly enough, Weber number

theory is no longer applicable and dispersion of the solvent is by gaslike mixing.24

Whatever the mechanism of dispersion, whether by droplet formation or gas-like

mixing, it is evident that the rate at which the phase boundary between the solvent and

anti-solvent disappears is dependent on the density of the bulk anti-solvent. At low

density, when the anti-solvent is in the vapour phase, the dispersed liquid solvent

remains as droplets until these droplets are destroyed by impact with the vessel or a

liquid phase. At high density when the anti-solvent is liquid or supercritical, the solvent

phase boundary disappears rapidly which implies that the mass transfer processes

CHAPTER 4

120

operating are occurring rapidly. Any form of mixing enhancement such as the use of a

coaxial nozzle speeds up the dispersion process.

The issue of whether the liquid is dispersed as droplets or by gas-like mixing has

generated much discussion in the literature. The most common explanation for the

results that have been observed are based on the premise that droplets are formed when

the liquid solution is injected into the anti-solvent through a capillary.1,4,21,22,25-42 The

mechanism of particle formation has then been explained in terms of the processes that

occur on these droplets.

4.3.1. The Formation of Microspheres

In a large number of studies the predominant morphology is the formation of

microspheres. In some cases microspheres have formed even when the solute used is

crystalline, which is unusual because microspheres are indicative of an amorphous

morphology.4

The change from crystalline to amorphous has been observed in a number of systems

and has resulted in the conclusion that the ASES process tends to reduce the

crystallinity of the solute.30-32,43,44 The amorphous change is thought to be due to the

very fast precipitation that does not enable organisation of the compound into a

crystalline form.

Descriptions of possible mechanisms of particle formation in the ASES process were

first proposed by Dixon and co-workers to explain the processes involved in the

microspheres obtained when injecting toluene solutions of polystyrene into CO2.25 The

mechanisms proposed have been termed the One Droplet - One Particle and One

Droplet - Many Particle theories.24

CHAPTER 4

121

4.3.1.1. One Droplet — One Particle Theory

The one droplet — one particle theory proposes that when a solution is injected into a

bulk anti-solvent, hydrodynamic forces break up the liquid stream into discrete droplets.

Mass transfer processes occurring on these droplets result in the formation of a single

particle. The formation of hollow spheres is thought to proceed by this mechanism. As

the anti-solvent enters the droplet of solution, a film forms on the surface of the droplet.

Mass transfer of solvent out of the droplet then causes the outer shell to harden.45 A

hollow core is thought to be a result of CO2 entering the droplet through the glassy skin.

More anti-solvent enters the droplet than solvent leaves, resulting in a decrease in

concentration of solute in the core. The concentration of solvent in the core is sufficient

to prevent vitrification. The solvent rich droplets in the core grow and coalesce. As this

occurs the droplets rupture and collapse against the inner surface resulting in a hollow

core.45

4.3.1.2. One Droplet — Many Particles Theory

The One Droplet — Many Particle theory was proposed in conjunction with the One

Droplet — One Particle theory, as a result of experimental observation that indicated

that a more complex process than droplet formation was at work.25 In this theory the

initial formation of droplets is proposed, but the mass transfer of anti-solvent into and

solvent out of the droplet results in high levels of supersaturation in the droplet. The

high level of supersaturation causes many nuclei to form in the droplet which results in

many particles, which are smaller than the initial droplet size.

A similar mechanism of particle formation has been proposed by Reverchon and co-

workers.31,33,46 Droplet formation is once again the method of solution dispersion in

the bulk anti-solvent. Once a droplet has formed, rapid diffusion of anti-solvent into the

droplet causes the droplet to expand and solute inside the droplet to precipitate.

Precipitation begins at the surface of the droplet where the supersaturation is the

greatest. Solute inside the droplet migrates to the surface and an empty shell made up of

many primary particles is formed. If the expansion level is low enough, the formed shell

CHAPTER 4

122

remains intact as a single particle made up of many primary particles forms from each

droplet. If expansion of the droplet continues once the shell has formed, it explodes

producing many small particles. The formation of these empty shells is similar to the

formation of the hollow spheres described in the One Droplet — One Particle theory,

however the distinction is that in the case of the empty shells, they are made up of many

smaller particles whereas the hollow spheres are essentially a single particle.

4.3.1.3. No Droplet Formation, Nucleation and Growth Theory

The conclusion drawn from a number of studies is that the formation of droplets is not

very important in the ASES process:.24,44,47-50 Rather the mechanism of particle

formation is predominantly one of nucleation and growth. The basis for this conclusion

has been the relative insensitivity of the particle size and morphology to adjusting

process variables that are expected to increase Weber number and produce smaller

particles. If this mechanism holds then the ASES process becomes a GAS process with

very high rates of pressurisation.

Evidence for the particles produced by ASES being predominantly a nucleation and

growth mechanism rather than a droplet formation mechanism is the similarity of

particles produced by the GAS and ASES processes. Microspheres of 1-2 µm have been

produced for polystyrene25 and various proteins,16,17 when processing by the GAS and

ASES processes.

4.3.2. Fibre Formation

The formation of fibres is observed particularly when solutions of polymer in high

concentration are injected into liquid or supercritical anti-solvent.23,25,27,51,52 Fibre

formation is due to vitrification of the solute before breakup of the liquid jet occurs. In

terms of hydrodynamics an increase in concentration results in two competing effects.

Stabilisation of the jet occurs due to an increase in viscosity and destabilisation of the

jet can occur due to an increase in viscoelasticity.25 Stabilisation of the jet tends to

delay jet disintegration which enables vitrification of the solute while the jet is intact,

CHAPTER 4

123

resulting in the formation of fibres.25,27 Jet destabilisation due to increased

viscoelasticity is offset by the stabilisation due to increased viscosity.

Hollow fibres can form by a similar mechanism to the formation of hollow spheres.

Mass transfer of anti-solvent into the liquid solution results in skin formation around the

jet. Diffusion of anti-solvent occurs through the skin faster than solvent can diffuse out

of the skin. The interior of the jet remains rich in solvent and anti-solvent rich-voids

grow. These voids coalesce before vitrification occurs due to the high level of solvent

present in the core and leads to a hollow fibre.51

As mentioned previously, there are numerous process parameters that can be adjusted

when carrying out the ASES process. It is difficult to isolate each parameter and this has

been highlighted by Dixon and co-workers.25 The density of the anti-solvent

environment was noted to effect both jet break-up and mass transfer through a complex

mechanism, which is complicated by the fact that the density of the liquid solution

changes throughout the spraying process. Nevertheless the effect of process parameters

on the morphology and size of precipitates is important in terms of understanding the

ASES process.

4.3.3. The Effect of Solute

Strictly speaking the solute cannot be considered a process parameter, but in terms of

the ASES process the nature of the solute has a significant influence on the particle size

and morphology.

A few generalisations can be made about the precipitates formed by ASES based on the

nature of the solute. Compounds with simple molecular structures generally form

primary particles, while complex molecules produce more complex morphologies.32,46

Examples of this will be given below.

It has been found that as the size of a molecule increases it becomes more difficult to

form crystalline material.53 The reason for this is that the high turbulence generated

CHAPTER 4

124

during the process affects both nucleation and growth of the resulting particles. The

regular arrangement of molecules is interrupted and the crystallinity of the final

precipitate is reduced. The reduction in crystallinity with increasing molecular weight

results in macromolecular compounds such as proteins and polymers precipitating in the

form of microspheres and/or other amorphous structures such as fibres and films.

The differences between low and high molecular weight compounds has been attributed

to differences between the rates of mass transfer compared to the rates of nucleation and

growth.54 The mass transfer process will play a more significant role for high molecular

weight compounds which have slower nucleation kinetics. In these cases the solute

concentration with respect to time can determine the phase behaviour which influences

particle morphology. In the case of low molecular weight compounds, nucleation

kinetics are much faster than the mass transfer process implying that the rate of

nucleation plays a more significant role.54

Carbon dioxide can have a considerable influence on the morphology of polymer

precipitates produced by the ASES process. A significant difference between molecules

of low molecular weight and polymers is the solubility of CO2 in the solute. The

solubility of CO2 in low molecular weight compounds is in most cases negligible,

whereas it can be significant in high molecular weight compounds such as polymers.55

The gas solubility in the solute has the effect of altering physical properties such as

lowering the glass transition temperature and causing plasticisation.23,27 Particle-

particle and particle-wall collisions of plasticised polymers result in primary particles

forming flocculates and agglomerates.51,56 When two plasticised particles collide they

initially flocculate, and if they are soft enough, surface area minimisation may occur by

coalescence forming a single larger particle.56

Amorphous polymers tend to flocculate and agglomerate as a result of plasticisation,23

while more crystalline polymers have been found to give particles with less

agglomeration due to their better thermal behaviour and greater crystallinity.23,57

CHAPTER 4

125

The issue of agglomeration has led to attempts to prevent agglomeration of polymer

particles by the addition of surfactants.48,58 Poly(methyl methacrylate) (PMMA), which

is highly plasticised by CO2 was successfully precipitated as individual microparticles

by ASES with the additions of triblock co-polymers as stabilisers to the polymer

solution.58 The inclusion of a similar surfactant to a solution of epoxy resulted in the

formation of microspheres but agglomeration was not prevented.48 In the case of the

epoxy precipitation, the agglomeration was thought to occur during the washing stage

rather than the precipitation stage of the process. The stabilisers are thought to operate

by absorbing onto the formed particle surface in the form of a micelle, thus reducing

interfacial tension between CO2 and the dispersed phase.

A class of polymers called aramids (para-linked aromatic polyamides), which have

useful applications as high modulus and heat resistant fibres, has been found to form

fibres due to their rigid backbone when processing by ASES.12 When a dilute solution

of polymer was injected concurrently with carbon dioxide, shear caused by the different

flow rates of the two phases resulted in an orientation of the polymer chains in the

direction of the applied stress. The rigidity of the backbone encouraged fibre formation.

4.3.4. The effect of Anti-Solvent Density (Pressure)

The adjustment of the density of the gas anti-solvent is one of the major variables in the

ASES process and can effect jet break-up, phase behaviour, and mass transfer

pathways.21,51 The effect of density on particle size and morphology has resulted in

conflicting results in the literature.24,46 Part of the reason for these contradicting results

may be due to the experiments being performed with the anti-solvent in different states.

The conditions that are commonly used when carrying out the ASES process in terms of

phase and density respectively are shown in Figure 4.1 and Figure 4.2. The anti-solvent

can exist in three phases once the temperature is set. At low pressure and low density

the anti-solvent exists as a vapour, as the pressure is increased a liquid anti-solvent

phase will form and the vapour and liquid phases coexist. As the pressure is increased

further the vapour and liquid phases will converge and become a single liquid-phase, or

if the temperature is high enough a supercritical fluid phase. The

CHAPTER 4

126

��

Supercritical FluidRegion

Liquid

Gas

SolidPres

sure

Temperature

Figure 4.1 Common ASES processing conditions in terms of anti-solvent phase

behaviour.

CHAPTER 4

127

0.0

0.2

0.4

0.6

0.8

1.0

0.0 10.0 20.0 30.0 40.0Pressure / MPa

Den

sity

/ g.m

L-1

20 °C30 °C40 °C50 °C60 °C70 °C

Figure 4.2 Common ASES processing conditions in terms of anti-solvent density.

CHAPTER 4

128

pressure and temperature of the anti-solvent determines the equilibrium conditions of

the process mixture. At low pressures the system is at low expansion conditions and as

the pressure is increased the system moves towards the miscible or asymptotic

expansion conditions. The particle sizes and morphologies that have been produced by

ASES for various solutes have been strongly influenced by the phase of the anti-solvent.

It is convenient to divide the discussion on the effect of density into three sections based

on the phase of the anti-solvent.

Vapour Phase

There are a number of consequences of injecting a solution into low-density anti-

solvent. Firstly, the vapour phase corresponds to systems where the solvent and anti-

solvent will be partially miscible. Diffusion of anti-solvent into the solution occurs up to

the thermodynamically set limit but extraction of the solvent into the anti-solvent does

not occur to any great extent. The lower bulk density means that Weber numbers are

smaller and jet break-up is more difficult to achieve. The result of the limited diffusion

of the anti-solvent and dispersion of the solution is expected to result in lower levels of

supersaturation. Low precipitation yields can be expected.40 In the worst case, the

density is such that enough anti-solvent can diffuse into the solution to cause

supersaturation and precipitation.

As the density is increased the diffusion of anti-solvent into the solution will increase

and the solute will begin to precipitate. Depending on the mass transfer rates, diffusion

of anti-solvent into the solution may not be rapid enough to cause precipitation in the

droplets. In this case precipitation will occur in the liquid layer which collects at the

bottom of the vessel.33,34 Large crystals of organics22 and irregular films for polymers

have been obtained due to precipitation occurring in the liquid layer.1 The liquid layer is

rich in solvent which implies that the level of supersaturation will be low resulting in

low levels of nucleation and larger particles.

Increasing the density further but remaining in the vapour phase has resulted in

precipitation occurring in the liquid droplets before they collect as a liquid layer. Skin

formation around droplets has been observed for polymer solutions at these conditions.1

CHAPTER 4

129

The particles formed had a cusped morphology due to slow drying in the low-density

environment.

Bodmeir and co-workers27 found that the smallest particles of poly(d,l-lactide) were

obtained when a dichloromethane solution was sprayed into vapour phase CO2 They

found that there was hardly any change in particle size with increasing density. They

concluded that for their system atomisation was sufficient at low density and any

increase in atomisation with increased density did not make any difference to the final

particle size.

Spraying a dilute solution of polymer into vapour CO2 did not result in atomisation of

the liquid jet.25 The jet broke into large droplets. A second liquid phase formed at the

bottom of the cell and significant agglomeration occurred.

Vapour Over Liquid

At higher densities the vapour over liquid condition occurs where the anti-solvent exists

in both the liquid and vapour phase in the vessel. When a solution is injected at these

conditions mass transfer and hydrodynamics will be similar to the lower density system.

However the increase in density is expected to increase Weber numbers and result in

smaller droplets. The major difference with the vapour over liquid compared to the

vapour phase system is that for the former the droplets fall into a liquid phase where the

solvent and anti-solvent are miscible.

As before, depending on the rates of mass transfer, precipitation can either occur in the

liquid droplets or in the liquid phase at the bottom of the vessel. In this case however,

the liquid phase is rich in anti-solvent and high rates of supersaturation are expected.

Microspheres were produced when polymer solutions of low concentration were

injected into vapour over liquid CO2.23 The solution did not atomise but fell as large

droplets into the liquid phase. Little precipitation occurred before the droplet hit the

liquid phase after which rapid mass transfer produced intense nucleation resulting in

CHAPTER 4

130

small microspheres. The results from this study are interesting in that microspheres

were formed from large droplets.

Hollow microspheres and microballoons have been formed for toluene solutions of

polystyrene sprayed into a vapour over liquid CO2.45 The formation of these particles

was attributed to jet break-up occurring in the vapour phase to form discrete droplets.

Skin formation is thought to occur in the vapour phase due to diffusion of CO2 into the

droplet. Once the droplet enters the liquid phase, toluene on the surface diffuses out and

the outer shell is hardened. Jet break-up in the vapour phase was essential to the

formation of these microspheres as evidenced by fibre formation for the same solution

when injected into liquid CO2. Microspheres of polystyrene have formed at similar

conditions by spraying into vapour over liquid HFC-134a.40 Atomisation of the liquid

jet was again important with fused and coalesced particles being formed when the flow

rate was insufficient to cause jet break-up.

Fibre formation was prevented for a toluene solution of polystyrene in high

concentration by injecting at vapour over liquid conditions25 The reason for the change

in morphology was attributed to the low level of solvent dissolution in the gaseous anti-

solvent which prevents the jet viscosity increasing. Jet break-up is able to occur before

vitrification and droplets form. These droplets fall into the liquid phase anti-solvent and

result in the formation of discrete microspheres.

Microspheres of L-PLA were obtained when spraying into vapour over liquid CO2.27

The solution was seen to break up into droplets and then fall into the liquid phase. Rapid

precipitation occurred and microparticles formed.

Liquid/Supercritical Phase

The majority of ASES studies have been conducted at conditions where the anti-solvent

is in the liquid or supercritical phase. At these conditions, an increase in density is

expected to increase Weber numbers resulting in jet break-up into smaller droplets.

Smaller droplets will provide greater surface area for mass transfer to occur thereby

CHAPTER 4

131

increasing the rate of mass transfer. Higher levels of supersaturation should be

obtainable, resulting in greater rates of nucleation and smaller particles.

Mass transfer effects are expected to be more significant at higher densities. The

diffusion of solvent into the anti-solvent is much greater than at vapour phase conditions

and it should be decreased by an increase in density.48 The rate of mass transfer of anti-

solvent into the solvent decreases with increasing density due to the increase in

viscosity of the anti-solvent.

The influence of increasing pressure and therefore density has been found to produce

conflicting results. Increasing density has resulted in a decrease in particle size25,27,59-

61, an increase in particle size1 and to have little effect on particle

size.12,22,25,27,31,33,41,46,47,62,63

The reasoning for the above observations has been based on the effects of

hydrodynamics and mass transfer rates. The decrease in particle size with increasing

density occurs in systems where atomisation is important. At higher densities the Weber

numbers are larger which means that jet breakup will produce smaller droplets

producing smaller particles. For the systems that showed little change with varying

density, the reasoning was that atomisation was sufficient at low density to produce

very small droplets and an increase in density had little effect on the droplet size and

correspondingly on the particle size. An increase in particle size as density is increased

has been explained in terms of mass transfer and nucleation kinetics, which are more

dominant than the hydrodynamic influences at the conditions studied.1,64 At higher

pressures in the miscible region the diffusivity of anti-solvent into the solution will

decrease. This will produce lower supersaturation levels and correspondingly a lower

rate of nucleation and larger particles. The decrease in droplet size due to increased

density is offset by the reduction in anti-solvent diffusivity.

Two different morphologies of amoxicillin dissolved in DMSO were obtained when

spraying at low density supercritical conditions.30 At these conditions the solvent and

anti-solvent were only partially miscible and a solvent rich liquid phase collected at the

CHAPTER 4

132

bottom of the vessel. Spherical particles were obtained from the upper supercritical

phase, and a film was obtained in the lower liquid phase. At similar conditions a film of

tetracycline hydrochloride was produced.32 A solute film was thought to form as a

result of the low solvent expansion conditions and formation of the second solvent rich

liquid phase at the bottom of the vessel. The solute forms as a film as the liquid solution

dries.

4.3.5. The Effect of Temperature

As with adjusting density, varying the temperature of the anti-solvent has

increased1,25,27,30,31,33,46,59,62,65, decreased60 and had no effect22,66 on particle size.

The influence of temperature on the ASES system is once again a result of competing

effects. In terms of hydrodynamics an increase in temperature will reduce the viscosity

of the solution and therefore increase the velocity of the solution through the nozzle.25

Weber numbers are expected to increase and therefore reduce the droplet size producing

smaller particles. Mass transport rates are expected to increase, although only slightly,

aiding in the reduction in particle size.1 However an increase in temperature will

increase particle sizes by reducing the density of the anti-solvent resulting in less

effective atomisation and increasing the growth rates of the particles produced.1,31,33,46

The effect on particle size with increasing temperature depends on whether

hydrodynamic or growth effects dominate.

The majority of studies have shown that the particle size increases with increasing

temperature. The indication is therefore that atomisation is less dominant than particle

growth at elevated temperatures. In the two phase region, changing the temperature had

no affect on the particle size or morphology.25 The latter observation reinforces the

observation from the density variation experiments that indicate that at low density

conditions atomisation is more dominant.

Manning and co-workers came to the conclusion that the smallest particles are generally

observed just above the critical temperature of the anti-solvent with the particles

increasing in size as the temperature is lowered or raised.67

CHAPTER 4

133

An important consideration regarding temperature and the processing of polymers is

plasticisation. A number of studies have shown that agglomeration of polymer particles

occurred resulting in an increase in particle size. At increased temperatures

plasticisation of the polymer can occur due to the lowering of the glass transition

temperature in the presence of compressed CO2. The formed particles coalesce to form

larger particles by interparticle collision..25,27,47,51

The agglomeration of a polymer has been reduced by decreasing the temperature of the

anti-solvent to -20°C.23 Interestingly, by reducing the temperature further to –30°C

large agglomerates formed as a result of the decreased mass transfer rates. The slow

dispersion of the polymer resulted in the particles agglomerating.

Increasing temperature has been found to reduce the yield of precipitate due to

increased solubility of the solute at elevated temperatures in the anti-solvent and

solvent.62 The yield is reduced due to the some of the solute remaining in solution and

being extracted with the solvent and anti-solvent. In this same study an increase in

temperature was found to increase the particle size distribution.62

Precipitation of polystyrene at 0°C resulted in a change in morphology from a hollow

fibre at higher temperature to an interconnected network of polymer voids.51 At lower

temperatures mass transfer is slowed down thus delaying nucleation. Phase separation

could therefore occur by spinodal decomposition rather than by nucleation and growth

resulting in the change in morphology.

4.3.6. The Effect of Solution/Anti-Solvent Flow Rate

The size and morphology of particles produced by the ASES process can be influenced

by varying the anti-solvent and the solution flow rate. The effect of adjusting these flow

rates is threefold. Firstly, the thermodynamic conditions in the vessel will change to

solvent lean conditions if the anti-solvent flow rate is increased or solution flow rate

decreased. Secondly, the hydrodynamics will change due to the change in the velocity

CHAPTER 4

134

of the anti-solvent or solution. Lastly, the mixing conditions in the vessel will change

due to changes in turbulence in the vessel.

Changing the ratio of anti-solvent to solution flow rates can influence the mass transfer

rates in the system. Mass transfer rates can be improved if a high relative velocity of

anti-solvent to solvent is maintained. The improvement is a consequence of the

increased turbulence in the vessel increasing dispersion of the solution.68

Reducing the anti-solvent flow rate has been found to increase the crystallinity of

polymer.43 The increase in crystallinity was attributed to the increase in solvent

concentration in the vessel at steady state conditions which increases crystallinity in two

ways. Firstly, the presence of a higher concentration of solvent, at low anti-solvent flow

rates, increases the mobility of the polymer chains enabling them to rearrange in a more

crystalline. Secondly, the precipitation process is slowed down with a higher level of

solvent being present allowing time for the polymer molecules to rearrange into a more

crystalline form.43

A change in morphology of polymer particles has been observed when changing the

flow rate of anti-solvent and solution. Microparticles instead of fibres were observed

when the flow rate of anti-solvent was increased.27 The change in morphology was

attributed to increased jet break-up as a result of the increased flow rate through the

nozzle. At faster flow rates the jet is broken into droplets and microparticles form.

An increase in solution flow rate has produced more aligned polymer fibrils.52 This

alignment was attributed to an increase in shear with increasing solution flow rate,

which in turn elongates the polymer coils and reduces the viscosity of the solution.

Diffusion into and out of the jet is increased due to the reduced viscosity and nucleation

is rapid. The microfibrils can be formed in the jet before the coils undergo relaxation

resulting in a more aligned morphology in the direction of the jet.

Increasing the solution flow rate resulted in a sintering of amoxicillin particles.30 The

sintering was attributed to the higher level of solvent in the vessel which encouraged

CHAPTER 4

135

coalescence of particles. A similar but reverse result was observed where increasing

anti-solvent flow rate reduced agglomeration.49 The decrease in agglomeration was

attributed to the reduction in the level of residual solvent.

Changing the direction of anti-solvent flow from co-current to countercurrent has

resulted in less agglomeration of polymer particles.48 This was attributed to two effects.

Firstly, countercurrent flow increased the relative velocity between the anti-solvent and

the liquid jet resulting in increased shear forces and increased jet break-up. Secondly,

mixing was improved, thus increasing mass transfer and delaying the settling of the

particles allowing them to dry while suspended.

The effect of decreasing anti-solvent flow rate has been to decrease particle sizes in the

SEDS process.66 In the SEDS process the level of organic modifier in the vessel

determines the extent of mixing between the CO2 anti-solvent and water. At high CO2

flow rates the selectivity of the organic solvent towards the CO2 is high and the organic

solvent is selectively removed. The liquid phase becomes mainly water and since

Reynolds numbers are high due to the high flow rate, the limiting factor becomes the

diffusion coefficient of water in CO2. The rate of formation of thermodynamic

equilibrium is longer than the nucleation rate and the achievable level of supersaturation

remains low. Growth is able to occur resulting in larger particles. At low flow rates the

aqueous phase remains rich in organic solvent and remains miscible with CO2. Water is

extracted in a similar time scale to nucleation. Crystallisation is rapid and particles are

smaller.66

An increased solution flow rate of polymer resulted in larger particles due to the

increased level of solvent in the vessel.1 The driving force for mass transfer is reduced

and the level of supersaturation is reduced resulting in fewer nuclei and larger particles.

Solution flow rate was found to influence morphology in the fractionation of nylon from

carpet waste.41 At very high upstream pressures a change in morphology from spherical

to hollow half-spheres occurred. The change in morphology was attributed to the

increased mass transfer, as a result of smaller droplets, and increased shear at higher

CHAPTER 4

136

flow rate causing the forming spheres to rupture. Particle sizes were not influenced to

any great extent with varying flow rate.

A few workers have found that altering the anti-solvent and solvent flow rate did not

affect the particles produced in the miscible region.22,32,49 The supersaturation in the

miscible region is very large and forming smaller droplets makes little difference. The

conclusion was that parameters that affect mixing of the solvent and anti-solvent have

little effect on the particles, because it works on a shorter time scale than the

crystallisation process.22

An increased solution flow rate has been found to change the phase separation of

polymer from a predominantly nucleation and growth mechanism to one of spinodal

decomposition.51 The transition into spinodal decomposition is thought to be a result of

increased rates of mass transfer with increased flow rate.

The level of residual solvent has been found to be a function of anti-solvent flow

rate.43,49 Higher anti-solvent flow rates were found to result in lower levels of residual

solvent in the precipitate.

4.3.7. The Effect of Solvent Type

The solvent used can affect the processes occurring in the ASES system in a number of

ways. Solvents can have different viscosities, vapour pressures, ability to dissolve solute

and different mass transfer rates, all of which can be important in determining the final

particle characteristics.

In the precipitation of the steroid methylprednisilone, it was observed that the more

volatile or soluble with carbon dioxide the solvent was, the smaller the size of the

particles that were generated.62 The more miscible the anti-solvent and the solvent the

higher will be the supersaturation levels achievable and consequently smaller particles

should be produced.

CHAPTER 4

137

The solute solubility in the solvent was important in the production of microparticles of

polystyrene.21 Smaller particles were produced using tetrahydrofuran - a poor solvent

for polystyrene as opposed to toluene which is a good solvent for polystyrene. The

observation was explained by comparing the ternary phase diagrams of the two systems

as in Figure 4.3. The width of the metastable region on the ternary phase diagram of the

tetrahydrofuran-polystyrene-carbon dioxide system is narrower due to the lower

solubility of polystyrene in tetrahydrofuran. The mass transfer pathway during the

precipitation process shows that the nucleation and growth rate was faster for

tetrahydrofuran than toluene, as indicated by a narrower metastable region. The polymer

therefore precipitates sooner in the jet and smaller particles are produced.

Epoxy resin precipitation from methyl ethyl ketone and acetone produced different

morphologies.48 The particles produced from methyl ethyl ketone were smaller and

more spherical. The change in size and morphology was attributed to the lower vapour

pressure of methyl ethyl ketone which resulted in a lower driving force into CO2. The

lower driving force alters the nucleation mechanism, particularly at the surface of the

droplet, resulting in the changed morphology and size. It was also theorised that solute-

solvent interactions may play a significant part in final particle morphology.

N-methyl pyrrolidone (NMP) has been found to produce coalesced particles of zinc

acetate as opposed to those formed from dimethylsulfoxide.31 The coalescence was

attributed to the increased solubility of the solute in NMP, resulting in a co-solvent

effect and solubilising the solute slightly at final conditions. N-methyl pyrrolidone has

been found to totally extract the antibiotic griseofulvin, as opposed to dimethylsulfoxide

and dichloromethane which produced precipitates at the same conditions.32 Strong

interactions between the solvent and the solute are thought to result in the formation of a

solvato complex. The anti-solvent is unable to break up the solvato complex. The solute

remains soluble and is extracted from the vessel with the solvent.

CHAPTER 4

138

Polystyrene

CO2Solvent

GlassyRegion

Tetrahydrofuran

Toluene

Figure 4.3 Ternary phase diagram showing the effect of solvent quality upon the

two phase envelope and the mass transfer pathway for a polystyrene solution

precipitated into compressed CO2.21

CHAPTER 4

139

4.3.8. The Effect of Solute Concentration

The morphology of polymer particles has been found to be a strong function of

concentration.21,25,51,52,69 The most dramatic changes in polymer morphology have

been observed when the concentration was increased from the dilute to semi-dilute

regime. In the dilute regime discrete morphologies were observed such as

microspheres1,21,25,48,51,69 and microfibrils.52 At higher concentrations continuous

morphologies such as fibres25,51,52,69 and rods48 have been formed.

The transition from polymer discrete to polymer continuous morphologies occurs at

approximately three times the plait point concentration.52,69 The change in morphology

has been attributed to the increased polymer chain entanglement that occurs at higher

concentration. A dilute polymer solution is one in which the polymer chains are far

apart and do not overlap in solution. A semi-dilute polymer solution is one in which the

polymer chains are close enough together to overlap in solution. Overlap of polymer

chains can have a significant influence on the hydrodynamics of the system due to

increased solution viscosity. Increased concentration also means that precipitation

occurs sooner due to the lower level of anti-solvent needed to form a supersaturated

solution. The combination of these effects and the change in the mass transfer pathway

is thought to cause the changes in morphology.

Polystyrene and poly(L-lactic acid) were found to produce microparticles which

decreased in size as the concentration was increased below the plait point.21 The trend

of decreasing size with increasing concentration was attributed to the shortening of the

distance between the binodal and spinodal curve as the concentration approaches the

plait point. The metastable region is narrower and the mass transfer pathway crosses the

spinodal curve before much growth has had time to occur in the metastable region.

Time scales of polymer—solvent phase separation and polymer—polymer phase

separation were used to explain the results obtained at different concentrations from

attempts to form a polymer blend.69 Below the plait point, atomisation of the solution

occurs and the diffusion of CO2 into the solution is rapid enough that the polymer

CHAPTER 4

140

solution phase separation occurs before polymer-polymer phase separation and a blend

was formed. At concentrations three times above the plait point concentration,

stabilisation of the jet due to increased viscosity resulted in skin formation. The

formation of a skin decreased the mass transfer of CO2 into the solution and polymer-

polymer phase separation could occur before polymer-solvent phase separation.69

The concentration of polystyrene in toluene has been found to influence the morphology

of hollow microspheres produced by spraying into a vapour over liquid regime45 and

the morphology of hollow fibres produced when spraying in the miscible region.51 As

the concentration increased, the cavity in the spheres or fibres became smaller until they

were no longer hollow.

The effect of concentration of precipitation involving molecules of lower molecular

weight has also been studied. Most observations have been that the size of the particles

increases with increasing concentration.30,32,33,42,62 The increase in particle size has

been attributed to the change in the time for supersaturation to be reached. A dilute

solution reaches saturation relatively late, thus enabling atomisation to occur before

nucleation and smaller particles are formed, while concentrated solutions nucleate

before atomisation can occur resulting in larger particles.42

No effect with changing concentration41 and smaller particles with increasing

concentration has been observed.22 The reduction in particle size with increasing

concentration was attributed to higher supersaturation at higher concentration resulting

in a greater nucleation rate and less opportunity for particle growth.

4.3.9. The Effect of Nozzle Diameter/Type

The approach that has been made to nozzle design in the literature has been closely

related to the particle formation mechanism that is thought to be in operation. As

mentioned previously, the most common interpretations have been based on the

formation of droplets in the anti-solvent phase. If this is the mechanism in operation,

then the objective in making small particles becomes one of decreasing droplet size.

CHAPTER 4

141

Most of the nozzle adaptations and designs are made with the objective of increasing

atomisation and forming smaller droplets of liquid solution in the bulk anti-solvent.

Typical nozzle designs that have been reported as being used in ASES processes in the

literature are shown in Figure 4.4. These designs range from a single capillary to

multichannel and vibrating nozzles.

The simplest nozzle is a single capillary through which the liquid solution flows (Figure

4.4 a). The nozzle diameter and flow rate determine the Weber number once all other

parameters are set. Different results have been produced from studies in which the effect

of nozzle diameter on the hydrodynamics of the process and ultimately its effect on

particle size have been examined. An increase in nozzle diameter, from 30 to 200 µm,

was found to increase the particle size of Red Lake C pigment by an order of

magnitude.59 Changing the nozzle diameter from 300 to 500 µm had no effect on the

size of L-poly(lactic acid).38 Independence of particle size on changing nozzle diameter

has also been observed in other studies.41,48 The effect of nozzle diameter can be

explained in terms of the particle formation mechanism in operation. If atomisation and

droplet formation is the dominant process determining particle size, then reducing the

nozzle diameter will most likely result in a reduction in particle size. If mass transfer is

very rapid and supersaturation, and nucleation levels are not altered to any significant

extent by a reduction in droplet size, then changing the nozzle diameter is not expected

to change particle design.

Another nozzle design which is described in the literature is the coaxial nozzle which

consists of an inner nozzle surrounded by one or more larger outer nozzles (Figure 4.4

b). The anti-solvent usually flows through the outer and the liquid solution through the

inner nozzle. The effect of using a coaxial nozzle is twofold.21 Firstly, the relative

velocities of the two flows will be reduced thus reducing atomisation and mass transfer.

Secondly, mixing is improved by the increased turbulence created by the anti-solvent

flowing at higher velocity in close proximity to the liquid solvent increasing dispersion

of the liquid solution.

CHAPTER 4

142

(a) (b)

(d)

EnergisingGas

(c)

Figure 4.4 Different nozzle designs used in ASES.

CHAPTER 4

143

A coaxial nozzle with CO2 flowing in the annular region has been used to improve

circulation of the droplets and reduce agglomeration compared to the standard

nozzle.21,23 The coaxial nozzle improved mixing and this was observed by the

precipitate being recirculated through the vessel. Mass transfer was slower in the jet for

the coaxial nozzle but faster in the suspension outside the jet leading to improved drying

of the suspended particles.

A coaxial nozzle design and the presence of a stabiliser has been found to reduce

flocculation of poly(methyl methacrylate) compared to the particles produced with

stabiliser and the standard nozzle design.56 The reason for the reduced flocculation was

attributed to the altering of the position of the majority of mass transfer occurring from

inside to outside of the jet with the coaxial design. As absorption of the stabiliser is

thought to occur mainly outside of the jet, transferring the mass transfer process outside

of the jet enabled more time for diffusion and absorption of the stabiliser.

Multi-channel nozzles containing three concentric nozzles have been employed to

enable the micronisation of hydrophilic molecules (Figure 4.4 c).70,71 The solubility of

water in the anti-solvent was improved by delivering an organic solvent via one of the

nozzle channels.

The standard nozzle design has been noted to have limitations in producing micron

sized particles due to premature nucleation occurring before secondary atomisation.

Relatively large droplets form which results in larger particles. Coalescence of droplets

can also occur before nucleation due to low interphase mass transfer.42 Premature

nucleation can become a major hindrance to atomisation for concentrated solutions.42

Nozzle variations aimed at improving atomisation have included the generation of high

frequency waves through sonication1,29,67 and the inclusion of a high energy gas

flowing proximal to the solvent (Figure 4.4 c).37,42 Smaller particles are thought to be

achievable with the introduction of high frequency waves speeding up the atomisation

process so that it occurs before nucleation and to increase mass transfer so that

nucleation occurs before coalescence of droplets.42 As with the standard nozzle, the

effect of the vibrating nozzles on particle size will depend on the mechanism of particle

CHAPTER 4

144

formation in operation. Correspondingly, the vibrating nozzle has been found to make

little difference to particle size1,67 and to decrease particle size.37,42

Subramaniam and co-workers found that the use of an energising gas other than CO2 in

their high energy nozzle can delay precipitation of solutions of high concentration by

the second gas acting as a buffer.42 The delay in nucleation allows atomisation to occur

and subsequently smaller particles to be produced.

4.3.10. Pre-Addition of Anti-Solvent

A variation of the ASES process is to pressurise the solution to be injected with anti-

solvent before injection into the bulk anti-solvent. The level of pre-added anti-solvent is

kept below the saturation level of the solute to prevent precipitation in the lines before

entering the bulk anti-solvent phase. A consequence of pre-adding anti-solvent to the

solution is to reduce the viscosity of the solution and bring the solution closer to

supersaturation conditions. Less anti-solvent diffusion into the liquid solution is

required for precipitation which means that precipitation should occur sooner.

Pre-addition of CO2 to the solution has produced porous fibres when spraying a semi-

dilute concentration of polystyrene into CO2.25 It was hoped that the pre-addition of

CO2 would aid jet break-up and prevent fibre formation. However, the effect of

breaking up the jet was offset by the reduction in the time scale of vitrification. Pre-

addition of CO2 had no effect for dilute polymer solutions.51

Pre-addition of CO2 to a polymer solution and spraying into a vapour over liquid system

produced microballoons.45 These microballoons had a more uniform porous structure

than those produced without the pre-addition of CO2. Pre-addition of CO2 to the

solution results in rapid supersaturation due to a lower level of CO2 being needed to

diffuse into the droplet to cause vitrification. The rapid generation of supersaturation

means that there was less time for the growth and rupture of cells resulting in a more

uniform particle.

CHAPTER 4

145

The pre-addition of CO2 to a solution of RG503H had little effect on particle size but

caused the formation of bubbles on the surface of the particles.37,42

Pre-addition of CO2 reduced the viscosity of the polymer solution, resulting in a slight

decrease in spraying pressure.38 A reduction in particle size was observed compared

with the conventional spray. The decrease in agglomeration was most likely the result

of a rapid hardening of the droplet surface. There was no increase in particle porosity by

pre-addition of CO2.

4.3.11. The Effect of Stirring Rate

The introduction of stirring into the ASES process is expected to increase the rate of

mixing of the anti-solvent and solution. An increase in mixing should increase the mass

transfer rates resulting in higher levels of supersaturation and consequently greater

nucleation rates. Smaller particles are therefore expected. Schmitt and co-workers62

investigated the effect of stirring while performing the ASES process. It was observed

that an increase in stirring rate produced smaller particles up to 900 rpm and then no

further reduction in particle size was observed by increasing the rate.

4.4. Effect of Process Parameters on Encapsulation/Co-Precipitation

The possibility of encapsulating drugs with polymers, forming drug polymer

composites, or coating surfaces with drugs or polymers by ASES is arguably one of the

most promising applications of the process. The encapsulation or coating of

pharmaceuticals or other materials requires that the particles to be encapsulated are

totally encased in a second compound. Many of the issues that were discussed

previously for the ASES process also apply to the encapsulation of compounds and will

not be repeated here. Extra issues that are important when producing encapsulated or

composite drugs are the drug loading and encapsulation efficiency. Combining a drug

and polymer and precipitating them by ASES does not necessarily result in an effective

encapsulation. For example, the drug and polymer may precipitate in separate domains,

or the drug may precipitate mostly on the surface of the polymer both resulting in

CHAPTER 4

146

ineffective encapsulation. Although the potential of the ASES process as an

encapsulation technique has been demonstrated, little information has been given

regarding the effect of process parameters and few generalisations can be made.

Various authors have shown that drug loading is a function of the solubility of the drug

in the anti-solvent and therefore is a function of the pressure or density of the

system.26,29,47 It was generally found that the higher the solubility of the drug at higher

pressure, the lower the drug loading due to the enhancement in solubility and thus the

extraction of the drug into the anti-solvent phase. Drug loadings of non-polar drugs such

as piroxicam in L-poly(lactic acid) particles were relatively low and found to decrease

significantly as the pressure was increased from 90 to 200 bar at 40°C.26 Experiments

on more polar drugs such as hyoscine-butylbromide achieved drug loadings of 19.5

wt% which remained unchanged, even at higher pressures, due to their negligible

change in solubility in the carbon dioxide anti-solvent.

Falk and co-workers found that an increase in the drug to polymer ratio increased the

drug loading of rifampin, gentamycin, and naltrexone particles in L-poly(lactic acid).29

The effect was least significant with the most lipophilic drug, rifampin, which had

highest solubility in carbon dioxide.

The concentration of polymer solution has been found to influence encapsulation

efficiency.23,37 Subramaniam and co-workers found that at lower concentrations of

polymer solution sugar beads were more uniformly coated than at higher polymer

concentrations.37 The reduction in encapsulation with increasing concentration was

attributed to larger polymer particles forming at higher concentration which did not coat

the beads as well. Young and co-workers,23 however, found that when spraying a

lysozyme suspension containing dissolved polymer into vapour over liquid CO2, an

increase in polymer concentration resulted in improved encapsulation efficiencies. The

vapour over liquid regime was used to enable large droplets of solution to form in the

vapour phase and fall into the liquid phase. The suspended lysozyme particles would be

contained in these droplets and the probability of encapsulation would be increased due

to the close proximity of the precipitating polymer to the suspended particles. At low

CHAPTER 4

147

polymer concentrations a low encapsulation efficiency was obtained. At low polymer

concentrations large droplets dispersed as they entered the anti-solvent liquid phase. The

polymer precipitated as microspheres which were too small to effectively encapsulate

the suspended lysozyme particles. At higher polymer concentrations the large droplets

were not dispersed as they entered the anti-solvent liquid phase. Larger polymer

particles were formed which were capable of encapsulating the lysozyme particles

resulting in higher encapsulation efficiencies.

Subramaniam and co-workers investigated the effect of anti-solvent flow rate and

temperature on encapsulation efficiencies.37 An increase in anti-solvent flow rate was

found to reduce the efficiency of coating of non-pareil sugar beads. The reduction in

polymer coating at higher anti-solvent flow rates is thought to be as a result of the

precipitating polymer being entrained in the flowing anti-solvent and not adhering to the

surface of the sugar beads. Raising the temperature to 40°C increased the coating

efficiency and the polymer precipitated as a film which totally covered the surface of

the beads.

CHAPTER 4

148

4.5. References












5. Kitamura, M.; Yamamoto, M.; Yoshinaga, Y.; Masuoka, H.; "Crystal Size Control of

Sulfathiazole Using High Pressure Carbon Dioxide," J. Cryst. Growth 1997, 178,

378.







87, 96.


Solvent Recrystallization of RDX: Formation of Ultra-Fine Particles of a Difficult-to-

Comminute Explosive," J. Supercrit. Fluids 1992, 5, 130.



Process," AIChE J. 1996, 42, 431.

10. Müller, M.; Meier, U.; Kessler, A.; Mazzotti, M.; "Precipitation of a Pharmaceutical

Using High Pressure Carbon Dioxide as Anti-Solvent," Proceedings of the 5th


CHAPTER 4

149

11. Muhrer, G.; Dörfler, W.; Mazzotti, M.; "Gas Anti-Solvent Recrystallization of

Speciality Chemicals : Effect of Process Parameters on Particle Size Distribution,"


Georgia, 2000.



















18. Cai, J.-G.; Liao, X.-C.; Zhou, Z.-Y.; "Microparticle Formation and Crystallization



23.



1997, 53, 232.



Chem. Eng. 1996, 4, 257.

CHAPTER 4

150



Sci. 1997, 64, 2105.



1999, 198/199, 767.






Solvent: Atomization vs Nucleation and Growth," J. Phys. Chem. B 2000, 104, 2725.








1995, 12, 1211.



Release 1993, 24, 27.








Solvent Precipitation of Nano-Particles of a Zinc Oxide Precursor," Powder Technol.

1999, 102, 127.

CHAPTER 4

151




Precipitation of Nano-Particles of Superconductor Precursors," Ind. Eng. Chem. Res.

1998, 37, 952.






36. Subra, P.; Jestin, P.; "Powders Elaboration in Supercritical Media: Comparison with

Conventional Routes," Powder Technol. 1999, 103, 2.


Precipitation and Coating Using Near-Critical and Supercritical Antisolvents," US

5833891, 1998.






F., Ed. 1993; Vol. 514, 238.


Toluene Solution into Compressed HFC-134a," Ind. Eng. Chem. Res. 1999, 38, 3898.










CHAPTER 4

152










106, 77.




Microparticles. Influence of Process Parameters on the Residual Solvent in


J. Pharm. Sci. 1997, 86, 101.

50. Chattopadhyay, P.; Gupta, R. B.; "Supercritical CO2 Based Production of Fullerene

Nano-Particles," Proceedings of the 5th International Symposium on Supercritical




Polym. Sci. 1993, 50, 1929.






54. Werling, J.; Debenedetti, P.; "Numerical Modeling of Mass Transfer in the

Supercritical Anti-Solvent Process," J. Supercrit. Fluids 1999, 16, 167.

55. Connon, C. S.; Falk, R. F.; Randolph, T. W.; "Role of Crystallinity in Retention of



CHAPTER 4

153










1997, 13, 1519.



60. Robertson, J.; King, M. B.; Seville, J. P. K.; Merrifield, D. R.; Buxton, P. C.;

"Recrystallisation of Organic Compounds Using Near Critical Carbon Dioxide," The

4th International Symposium on Supercritical Fluids, Sendai, Japan, 1997, A, 47.





1995, 41, 2476.





64. Polakodaty, S.; York, P.; Hanna, M.; Pritchard, J.; "Crystallization of Lactose Using

Solution Enhanced Dispersion by Supercritical Fluids (SEDS) Technique,"








CHAPTER 4

154



Powders Using the Same," US 5770559, 1998.

68. Palakodaty, S.; York, P.; "Phase Behavioral Effects on Particle Formation Processes

Using Supercritical Fluids," Pharm. Res. 1999, 16, 976.







6,063,138, 2000.

CHAPTER 5

155

5. SOLUBILITY OF SOLUTES IN SOLUTIONS EXPANDED WITH

CO2

5.1. Introduction

There have been many examples of the application of the GAS and ASES processes

reported in the literature. However, there is relatively little data available regarding the

solubility of solutes in the expanded solutions of the systems that have been studied.

Knowledge of the solubility of the reactants and products in a reaction medium is an

important parameter to be verified before a synthesis is carried out. In a conventional

synthesis reaction, variables such as the volume of organic solvent, or the concentration

of one reactant, can be easily varied. However, in the case of an expanded solution

which is under pressure, and in many cases with no means of visual observation,

alteration of these variables would be difficult.

Knowledge of the thermodynamics of a system is of equal importance when

micronising compounds by GAS or ASES. The condition at which precipitation of

solute occurs in a solution expanded with a gas anti-solvent depends on many factors

such as the nature of the solvent and solute. It is therefore important to study the

behaviour of the system to ensure that the micronisation process is operated at optimum

conditions.

The volumetric expansion and solute solubility of systems relevant to the synthesis and

micronisation of Cu-Indo using CO2 as an antisolvent were investigated and the results

are reported below. Firstly, the volumetric expansion of N,N-dimethylformamide

(DMF), N-methylpyrrolidone (NMP) and dimethylsulfoxide (DMSO) upon

pressurisation with CO2 will be presented. Secondly, the solubility of copper

indomethacin (Cu-Indo), copper acetate (Cu-Acetate), indomethacin and acetic acid in

DMF expanded with CO2 will be presented to determine the applicability of using CO2

as an antisolvent for the synthesis and micronisation of Cu-Indo. Thirdly, the solubility

of Cu-Indo in DMF expanded with CO2 in the presence of Cu-Acetate, indomethacin or

CHAPTER 5

156

acetic acid will be presented and compared to the solubility of Cu-Indo in the ternary

CO2—DMF—Cu-Indo system.

5.2. Experimental

All chemicals and reagents used are listed in Table 5.1. All chemicals and reagents were

used as supplied.

Table 5.1 Chemicals and reagents.

Chemical/Reagent Purity (%) /

Grade

Supplier

Copper Indomethacin 90 Biochemical Veterinary

Research

Copper (II) Acetate 98 Sigma-Aldrich

Indomethacin 99 Shanghai Shon Long

Pharmaceutical Factory

Glacial Acetic Acid 99.7 Ajax Chemicals

N,N-dimethylformamide 99.9 Burdick and Jackson

N-methylpyrrolidone 99.5 ISP Technologies Inc.

Dimethylsulfoxide 99.8 Ajax Chemicals

Carbon Dioxide > 99.9 BOC Gasses

Tetramethylammonium Chloride 97 Aldrich Chemical Co.

Acetonitrile 99.8 EM Science

5.2.1. Experimental Apparatus and Procedure

The experimental apparatus used in the volumetric expansion and solubility studies is

shown schematically in Figure 5.1. The apparatus consists of a high pressure vessel

(Jerguson sight gauge series No. 32) with an internal volume of 60 mL. The high

pressure vessel enabled observation of the system throughout each experiment. A

magnetic stirrer was designed for the high pressure vessel to ensure the solution was

well mixed before and during sampling. A 0.5 µm stainless steel filter frit was situated

at the bottom of the reaction vessel to sparge the CO2 in the liquid phase during

CHAPTER 5

157

P

CO2 Source

Syringe PumpHeating Coil

Magnetic StirrerThermostated Heater

Reaction Vessel

Water bath

CO2Reservoir

Filter

T

Sample Loop

Figure 5.1 Experimental apparatus used to determine volumetric expansion and

solute solubility.

CHAPTER 5

158

pressurisation and to trap any precipitated solid in the washing step. A 500 mL stainless

steel surge tank was placed before the high pressure vessel to maintain constant pressure

while a sample was taken. An ISCO LC-500 syringe pump was used to generate the

required pressure. Pressure monitoring was made possible by the use of a Druck

Pressure transducer (Model DPI 260 ± 0.007 MPa). The temperature was controlled to

within 0.1oC by submerging the apparatus in a water bath heated with a Thermoline

Unistat 130 water heater. The flow of CO2 through the vessel was from the bottom

during the pressurisation step or from the top during the washing step and was

controlled by two three way ball valves (Whitey SS-41XS2). Two 2-way ball valves

(Whitey SS-41S2) separated by a filter housing (Swagelok ANUPROA Filter) were

used as a sample loop. The sample loop had an internal volume of 1.85 mL (calibrated

with water).

A volumetric or solubility experiment consisted of loading the high pressure vessel with

approximately 20 mL of pure solvent or solution. In the systems that involved solid, an

excess of solid was added to ensure that the solution was saturated throughout the

experiment. The vessel was then placed in a controlled temperature water bath in order

to reach the desired temperature. The CO2 was passed through a heating coil to maintain

the CO2 at operating temperature, and then sparged through the solution via a frit

located at the bottom of the Jerguson. The system was allowed to equilibrate with

stirring after which a sample was taken at constant temperature and pressure. During

sampling, the CO2 was passed through the vessel from the top of the high pressure

vessel and a sample was taken by pushing the solution through the filter and into a

sample loop at constant pressure. To ensure that there were no bubbles in the sample

loop the pressurised liquid was slowly let into the sample loop from the bottom by

allowing the vapor to purge from the top of the sample loop. Once liquid started

escaping from the top of the valve, sampling was ceased and the sample valve

disconnected. Each experimental point was repeated until consistent results within 5%

RSD were obtained.

The mole fraction of each component was determined by gravimetric means. The mass

of CO2 was calculated by measuring mass of the sample loop before and after

depressurisation. The sample loop was depressurised by releasing the pressure over 10

CHAPTER 5

159

minutes and collecting any liquid that was extracted with the CO2 into a cold trap. The

liquid in the cold trap was sonicated for five minutes to remove residual CO2. The mass

of the depressurised liquid and the empty sample loop was then measured and recorded.

The mass of CO2 in the sample loop was then calculated by subtracting the mass of the

empty sample loop and collected solution from the full sample loop.

The mass of Cu-Indo and indomethacin was determined by High Pressure Liquid

Chromatography (HPLC). The HPLC used consisted of a Waters 600 pump, a 996

Photo Diode Array detector and a 717Plus Autoinjector. The samples were separated

using a Symmetry C18 column with a mobile phase made up of 70% acetonitrile and

30% 0.01M tetramethyl ammonium chloride in 0.1% acetic acid. A flow rate of 1

mL.min-1 was used and the analytes were detected at a wavelength of 320 nm.

The mass of Cu-Acetate was determined by atomic absorption (AAS) using a Perkin

Elmer Analyst 300. The analyses conditions were as follows: air-acetylene flame, 324.8

nm wavelength and a 0.7 nm slit width.

The mass of organic solvent was calculated by subtracting the mass of solute from the

mass of collected solution. The mole fraction of each component was calculated from

the measured masses.

The volumetric expansion of the liquid phase was calculated from the mass of organic

solvent in the sample loop as follows:

mP0

S

mPS — 1 5.1

where mS is the mass of organic solvent in the sample loop and P0 and P represent the

pressures at atmospheric and the pressure at sampling respectively.

CHAPTER 5

160

5.3. Volumetric Expansion of the CO2—Solvent Systems

The volumetric expansion behaviour of the CO2—DMF system at 25, 30 and 40°C was

examined and is illustrated in Figure 5.2 (See Appendix II for data). The volumetric

expansion was calculated using Equation 5.1. The volume of the liquid phase was found

to increase with increasing pressure. At a given pressure, there was a reduction in the

volumetric expansion of the solution with increasing temperature as shown in Figure

5.2.

The volumetric expansion behaviour of the CO2—NMP and CO2—DMSO systems at

25°C and 40°C was examined and is illustrated in Figure 5.3 (See Appendix II for data).

The trend of increasing expansion with increasing pressure and decreasing temperature

was observed for both NMP and DMSO.

The volumetric expansion behaviour of DMF, NMP and DMSO when pressurised with

CO2 at 25°C is compared in Figure 5.4. At pressures above 3 MPa, CO2 was more

soluble in DMF and the DMF solution undergoes a greater degree of expansion than the

other two solvents. NMP and DMSO had similar expansion profiles over the pressure

range tested. The solubility of CO2 in a solvent is expected to increase with decreasing

organic solvent polarity. The dipole moment, which can be used as a measure of

polarity, of DMF, NMP and DMSO is 13.0, 13.6 and 13.7 Cm respectively.1 DMF is

less polar than NMP and DMSO and correspondingly the solubility of CO2 in DMF is

greater for a given temperature and pressure than for NMP and DMSO. The polarity of

NMP and DMSO is similar and is reflected by CO2 having a similar solubility in these

solvents.

The solubility of CO2 in the liquid phase was measured for the DMF, NMP and DMSO

systems and is shown in Figure 5.5. The data in Figure 5.5 reveal an interesting

observation. For all isotherms, and for a given mole fraction of CO2, the volumetric

expansion of the liquid phase was the same.2 Kordikowski and co-workers observed a

similar trend for a number of different solvents pressurised with CO2. The inability to

distinguish between different solvents when pressurised with CO2 led Kordikowski and

CHAPTER 5

161

Pressure / MPa

% E

xpan

sion

0

200

400

600

800

1000

0.0 1.0 2.0 3.0 4.0 5.0 6.0 7.0 8.0

25 °C

30 °C

40 °C

Figure 5.2 Volumetric expansion of DMF—CO2 as a function of pressure at 25, 30

and 40°C.

CHAPTER 5

162

Pressure / MPa

% E

xpan

sion

0

50

100

150

200

250

0.0 1.0 2.0 3.0 4.0 5.0 6.0 7.0 8.0

NMP 25°C

NMP 40°C

DMSO 25°C

DMSO 40°C

Figure 5.3 Volumetric expansion of NMP—CO2 and DMSO—CO2 as a function of

pressure.

CHAPTER 5

163

Pressure / MPa

% E

xpan

sion

0

200

400

600

800

1000

0.0 1.0 2.0 3.0 4.0 5.0 6.0 7.0 8.0

DMF

NMP

DMSO

Figure 5.4 Comparison between the DMF—CO2, NMP—CO2 and DMSO—CO2

systems at 25°C.

CHAPTER 5

164

Mole Fraction CO2

% E

xpan

sion

0

100

200

300

400

500

0.0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9 1.0

NMP 25°C

NMP 40°C

DMSO 25°C

DMSO 40°C

DMF 25°C

DMF 30°C

DMF 40°C

Figure 5.5 Volumetric expansion of the DMF—CO2, NMP—CO2 and DMSO—

CO2 systems as a function of CO2 mole fraction.

CHAPTER 5

165

co-workers to suggest that the solute—solvent interactions were more important in the

GAS process than the solute—CO2 interactions.2

The expansion results obtained for the CO2—DMF system at 25°C are compared to the

results given in the literature in Figure 5.6.2 The two sets of experimental results show a

similar liquid phase volumetric expansion for a given pressure as shown in Figure 5.6a.

A significant difference between the two sets of data is observed when the liquid phase

volumetric expansion is plotted as a function of CO2 mole fraction, as shown in Figure

5.6b. In this work the mole fraction of CO2 was determined by gravimetric means as

described in the experimental section. The data in the literature was obtained by

measuring the density of the expanded solution using a densiometer and then measuring

the pressure increase upon depressurising a known volume of solution into a sampling

vessel of known volume. The data reported in this study are the result of repeated

experiments within a 5 % standard deviation. The differences in the measured mole

fraction of CO2 at a particular expansion are most likely due to the difference in

experimental technique used.

5.4. Solid Solubility in Expanded Solution

In the synthesis of Cu-Indo there are two reactants, Cu-Acetate and indomethacin, and

two possible products, Cu-Indo and acetic acid. All components may be present during

the synthesis. The solubility of each component in DMF expanded solution was

determined (See Appendix III for data). As the presence of other solutes may influence

the solubility of reactants and products,3,4 the solubility of each solute and a binary

mixture of these compounds in DMF expanded solution was also investigated.

The solubilities of Cu-Indo in DMF expanded with CO2 at 25, 30 and 40°C are shown in

Figure 5.7 and Figure 5.8 (See Appendix III for data). The mole fraction of Cu-Indo

decreased with increasing pressure and increased with temperature. As the temperature

increased, the pressure required to precipitate the Cu-Indo was increased due to the

reduction in the solubility of CO2 in the DMF at higher temperature. At each

temperature the solubility of Cu-Indo approached a minimum with increasing pressure.

CHAPTER 5

166

0

200

400

600

800

1000

0.0 1.0 2.0 3.0 4.0 5.0 6.0 7.0 8.0Pressure / MPa

% E

xpan

sion

This Work 25°CKordikowski 25°C

0

200

400

600

800

1000

0.0 0.2 0.4 0.6 0.8 1.0Mole Fraction CO2

% E

xpan

sion

This Work 25°CKordikowski 25°C

(a)

(b)

Figure 5.6 Comparison between literature2 results and this work for the

volumetric expansion of DMF with CO2 at 25°C.

CHAPTER 5

167

Pressure / MPa

Mol

e Fra

ctio

n x

10-4

0.0

0.5

1.0

1.5

2.0

2.5

3.0

3.5

4.0

4.5

0.0 2.0 4.0 6.0 8.0

25°C

30°C

40°C

Figure 5.7 Mole fraction of Cu-Indo dissolved in the liquid phase.

CHAPTER 5

168

Pressure / MPa

Con

cent

ratio

n / m

g.g-1

0.0

2.0

4.0

6.0

8.0

10.0

12.0

0.0 2.0 4.0 6.0 8.0

25°C

30°C

40°C

Figure 5.8 Concentration of Cu-Indo in DMF (CO2 free basis).

CHAPTER 5

169

This minimum is expected to coincide with the solubility of the solute in CO2.5 A

decrease in mole fraction of Cu-Indo was due to the increased mass capacity of the

solution (the dilution effect) as CO2 dissolved in the liquid phase and was not

necessarily indicative of precipitation. The concentration of Cu-Indo, where the

concentration of Cu-Indo is expressed in terms of a CO2 free basis, is shown in Figure

5.8 and indicates the pressure at which precipitation occurred in the system. The

solubility of Cu-Indo decreased rapidly as the pressure was increased, even at low

pressures. The effectiveness of CO2 as an antisolvent for precipitating Cu-Indo was

demonstrated with a greater than 95% yield at 25°C and 5.9 MPa.

The solubility of Cu-Acetate in DMF expanded with CO2 at 25°C is shown in Figure 5.9

(See Appendix III for data). The mole fraction of Cu-Acetate decreased with increasing

mole fraction of CO2 in the liquid phase. The concentration of Cu-Acetate (CO2 free

basis) increased slightly as the pressure was increased up to 3 MPa and then decreased

when the pressure was increased further. The initial increase in solubility may be due to

the density increase of the liquid phase that occurs at low pressures.2 This density

increase results in CO2 acting as a co-solvent at pressures below 3 MPa. The increased

solubility of Cu-Acetate in the CO2—DMF—Cu-Acetate system may be due to the high

solubility of Cu-Acetate in DMF at atmospheric conditions. Cu-Acetate was two orders

of magnitude more soluble in DMF than Cu-Indo at atmospheric conditions. The

solute—solvent interaction between DMF and Cu-Acetate was greater than between

DMF and Cu-Indo. The initial density increase in the liquid phase outweighed the

antisolvent effect of CO2 and this led to a solubility increase for Cu-Acetate. The

solubility increase continued until enough CO2 dissolved in the liquid phase and the

antisolvent effect dominated the solute—solvent interactions. In the case of Cu-Indo,

which was far less soluble in the organic solvent, this initial density increase did not

result in greater solubility because the solute—solvent interaction was dominated by the

antisolvent effect of CO2.

The solubility of indomethacin in DMF expanded with CO2 at 25°C is depicted in

Figure 5.10 (See Appendix III for data). The mole fraction of indomethacin decreased

gradually and at 5.5 MPa it dramatically decreased. The concentration of indomethacin

(CO2 free basis) remained constant at pressures below 5.5 MPa and then decreased to 46

CHAPTER 5

170

Pressure / MPa

Mol

e Fra

ctio

n

Con

cent

ratio

n / m

g.g-1

0.000

0.005

0.010

0.015

0.020

0.025

0.0 2.0 4.0 6.0 8.00

10

20

30

40

50

60

70

80

Figure 5.9 Solubility of Cu-Acetate in expanded DMF at 25°C.

CHAPTER 5

171

Pressure / MPa

Mol

e Fra

ctio

n

Con

cent

ratio

n / m

g.g-1

0.000

0.050

0.100

0.150

0.200

0.250

0.0 2.0 4.0 6.0 8.00

100

200

300

400

500

600

Figure 5.10 Solubility of indomethacin in expanded DMF at 25°C.

CHAPTER 5

172

mg.g-1 of DMF at 6 MPa. A second liquid phase appeared between 5.5 MPa and 6 MPa.

The second liquid phase converged with the first liquid phase once precipitation of the

indomethacin occurred.

The mole fraction of CO2 dissolved in the liquid phase for the DMF, Cu-Indo + DMF,

Cu-Actetate + DMF and indomethacin + DMF systems at 25°C is compared in Figure

5.11. At pressures below 5.5 MPa, the mole fraction of CO2 in the liquid phase was

lower for DMF solutions containing indomethacin than for the other systems. The

CO2—DMF expansion behaviour (the amount of CO2 dissolved in DMF) did not

change when solutes of low solubility such as Cu-Indo and Cu-Acetate were present in

the system. The solubility of Cu-Indo and Cu-Acetate in DMF at atmospheric pressure

and 25°C was 5.4 and 60 mg.g-1 respectively. A similar observation was made for the

expansion profile of the DMSO—CO2 system when protein with low solubility was

added to the system.6 However, the data in Figure 5.11 show that highly soluble

compounds such as indomethacin, which has a solubility of 571 mg.g-1 in DMF at

atmospheric pressure and 25°C, dramatically influences the phase behaviour of the

system. The lower solubility of CO2 in DMF in the presence of indomethacin may be

due to intermolecular forces between the solute and solvent. An analogous observation

has been made for the citric acid—ethanol system.7 A high solubility suggests that there

are strong cohesive forces between the solute and the solvent and the system no longer

behaves as pure DMF. The strong cohesive forces, which leave little solvent available to

interact with CO2, resulted in a second liquid phase forming at higher pressures. A

splitting of the liquid phase occurred in the CO2—naphthalene—toluene system in

which one liquid phase was rich in CO2 and the other rich in the solute.8 As the pressure

increased the amount of CO2 increased which resulted in an increased competing force

for DMF between CO2 and indomethacin. Eventually at pressures greater than 5.5 MPa,

the DMF—indomethacin and DMF—CO2 phases split. A further pressure increase

caused the indomethacin to precipitate out of solution. Once precipitation occurred and

the indomethacin concentration in solution decreased, the two liquid phases merged and

the phase behaviour of the system approached that of the CO2—DMF system.

CHAPTER 5

173

Pressure / MPa

Mol

e Fr

actio

n

0.0

0.2

0.4

0.6

0.8

1.0

0.0 2.0 4.0 6.0 8.0

DMF

Cu-Indo

Cu-Acetate

Indomethacin

Figure 5.11 Mole fraction of CO2 as a function of pressure at 25°C in the presence

of solute.

CHAPTER 5

174

The phase behaviour of the DMF—CO2—acetic acid (36 mg.g-1) system at 25°C is

shown in Figure 5.12. The presence of acetic acid did not influence the expansion

profile of the DMF—CO2 system over the pressure range tested.

The data from the ternary systems containing a solute indicate that Cu-Indo could be

separated from Cu-Acetate and indomethacin at 25°C by pressure manipulation. The

solubility of Cu-Indo in DMF and DMF expanded solutions at all pressures examined

was lower than that of Cu-Acetate and indomethacin. The saturation concentrations of

Cu-Acetate and indomethacin at various pressures can be obtained using the data in

Figure 5.9 and Figure 5.10. The synthesis process would be performed at conditions

where the concentration of the reactants, Cu-Acetate and indomethacin, are far from

saturation. If the initial concentration of a solute in solution is less than saturation,

precipitation will not occur until the solid—liquid equilibrium line is crossed.8 The

solubility studies show that at each pressure, the solubility of Cu-Indo was lower than

Cu-Acetate and indomethacin, hence it crossed the solid—liquid equilibrium line before

the others.

The solubility of a compound in an expanded solution may be influenced by the

presence of another solute. The solubility of Cu-Indo at 25°C in DMF expanded

solutions containing acetic acid, Cu-Acetate or indomethacin are depicted in Figure 5.13

(See Appendix IV for data). The initial concentrations of acetic acid, Cu-Acetate and

indomethacin in DMF were 60, 8 and 6.2 mg.g-1, respectively. These concentrations

were chosen close to the conditions that are to be used during the Cu-Indo synthesis.

The presence of acetic acid resulted in a 10 fold increase in the solubility of Cu-Indo.

Precipitation only occurred at a pressure above 5 MPa, corresponding to a CO2 mole

fraction of 0.7. Above these pressures pure Cu-Indo precipitated from solution. Cu-Indo

is dissociated in the presence of acetic acid to indomethacin and Cu-Acetate. A driving

force is needed to push the equilibrium toward the formation of Cu-Indo. Carbon

dioxide is able to induce precipitation of Cu-Indo at pressures above 5 MPa. At these

conditions CO2 drives the equilibrium toward the formation of Cu-Indo by its

antisolvent action.

CHAPTER 5

175

Pressure / MPa

% E

xpan

sion

00.0 2.0 4.0 6.0 8.0

200

400

600

800

1000

Figure 5.12 Volumetric expansion of DMF containing acetic acid at 25°C.

CHAPTER 5

176

Pressure / MPa

Con

cent

ratio

n / m

g.g-1

0.0

10.0

20.0

30.0

40.0

50.0

60.0

70.0

0.0 2.0 4.0 6.0 8.0

Indomethacin

Cu-Acetate

Acetic Acid

Ternary Cu-indo

Figure 5.13 Solubility of Cu-Indo in quaternary systems at 25°C.

CHAPTER 5

177

The presence of Cu-Acetate resulted in a four fold increase in the initial concentration

of Cu-Indo. In this case, however, the precipitation of Cu-Indo followed a similar trend

to that of the ternary CO2—DMF—Cu-Indo system, where precipitation occurred

throughout the pressurisation process. The presence of indomethacin resulted in a two

fold decrease in the initial concentration of Cu-Indo. Precipitation of Cu-Indo occurred

throughout the pressurisation process shown by the solubility decrease in Figure 5.13.

In the ternary CO2—DMF—Cu-Acetate system, an initial Cu-Acetate concentration of

8 mg.g-1 would result in precipitation occurring at pressures above 5 MPa. However, in

the presence of Cu-Indo the precipitation of Cu-Acetate began at 2 MPa (Figure 5.14). It

is suggested that Cu-Indo lowered the solubility of Cu-Acetate in the DMF expanded

solution. The concentration of indomethacin in the expanded solution is illustrated in

Figure 5.14. The concentration of indomethacin remained constant at all pressures,

which implies that Cu-Indo could be separated from indomethacin simply by increasing

the pressure of the system. It is interesting to note that at a concentration of 6.2 mg.g-1

of indomethacin in DMF, the second liquid phase did not appear in the system.

5.5. Conclusions

The volumetric expansion and solubility studies of the binary, ternary and quaternary

systems demonstrate that the solubility of a compound such as Cu-Indo in an expanded

solution depends not only on temperature and pressure, but also on the concentration

and presence of other compounds dissolved in the system. The effect of pressure and

temperature can be related to CO2 density and is predictable. However, the phase

behaviour and hence the solubility of Cu-Indo is a complex function of the solvent and

other solutes.

CHAPTER 5

178

Pressure / MPa

Con

cent

ratio

n / m

g.g-1

0.0

1.0

2.0

3.0

4.0

5.0

6.0

7.0

8.0

9.0

0.0 2.0 4.0 6.0 8.0

Cu-Acetate

Indomethacin

Figure 5.14 Solubility of the second solute in the CO2—DMF—Cu-Indo—Solute

systems at 25°C.

CHAPTER 5

179

5.6. References

1. Reichardt, C.; "Solvent and Solvent Effects in Organic Chemistry," 2 ed.; VCH:

Weinheim, Germany, 1990.

2. Kordikowski, A.; Schenk, A. P.; Van Nielen, R. M.; Peters, C. J.; "Volume

Expansions and Vapor-Liquid Equilibria of Binary Mixtures of a Variety of Polar

Solvents and Certain Near-Critical Solvents," J. Supercrit. Fluids 1995, 8, 205.



4. Foster, N. R.; Yun, S. L. J.; Dillow, A.; Wells, P. A.; Lucien, F. P.; "A Fundamental

Study of the Gas Anti-Solvent Process," The 4th International Symposium on







7. Tai, C. Y.; Cheng, C.-S.; "Effect of CO2 on Expansion and Supersaturation of




CHAPTER 6

180

6. SYNTHESIS AND PURIFICATION OF COPPER

INDOMETHACIN USING THE GAS AND ASES PROCESSES

6.1. Introduction

The major emphasis of the GAS and ASES processes has been to develop fine

particles.1,2 Some work has been published on the use of these techniques to purify

chemicals by selectively precipitating impurities and removing the substance of interest,

or by precipitating the desired substance and leaving the impurities behind in solution.3-

12

A new application of GAS or ASES is in the area of chemical reactions. In most

chemical reactions, it is desirable to obtain product of the highest possible purity, in the

least number of steps and using as little solvent as possible. This becomes increasingly

important in the pharmaceutical industry, where the presence of impurities in the final

product can make the batch unsuitable for use. Removing these impurities is often more

complex than the initial reaction and is very costly in terms of equipment required and

time taken to carry out the purification steps. These impurities come from a number of

sources namely: unreacted reactants, side reactions, unwanted isomers and residual

solvents. The use of dense gases as antisolvents can provide a convenient way of

removing these impurities.

A further advantage of using dense gases in chemical reactions is the ability to control

solvent parameters such as density, viscosity and diffusivity. Changes in these

parameters have been well documented for reactions undertaken in supercritical

fluids.13,14 The ability of the solvent to dissolve the product can be altered by varying

pressure and temperature. Reactions that are equilibrium controlled can then be

controlled by selectively removing product from the reaction mixture through

precipitation. One example is the reaction of isoprene with maleic anhydride in

supercritical CO2.14 Undesirable side reactions can be avoided by removing the product

from the reaction as it forms.

CHAPTER 6

181

In this study, the feasibility of using CO2 as an antisolvent in the synthesis and

processing of Cu-Indo (introduced in the background section) is examined. Cu-Indo

may be synthesised by the following reaction:

4 Indomethacin + 2 Cu-Acetate ↔ Cu-Indo + 4 Acetic Acid 6.1

In this reaction, Cu(II) acetate monohydrate (Cu-Acetate) is reacted with indomethacin

in an organic solvent such as N,N-dimethylformamide (DMF) to yield Cu-Indo.

The conventional process for the synthesis of Cu-Indo involves multiple stages as

shown in Figure 6.1. Organic solvents, such as DMF as the primary solvent and ethanol

as the antisolvent, are used in both the synthesis and purification of Cu-Indo, which

results in the generation of solvent waste as well as solvent residues in the drug. The

crystallisation step, which commences once the ethanol is added to the synthesis

mixture, takes longer than 24 hours. The organic solvents are removed by filtration,

which is a slow process mainly due to the blockage of filters. The larger the batch size

the more complex the problem becomes. A single filtration step is not sufficient to

purify Cu-Indo and a washing step is needed to remove impurities, which requires

further filtration and product loss. The Cu-Indo is then dried to remove the wash

solvent. The drug is unstable at elevated temperatures, which means that lower

temperatures need to be used for drying thus requiring long drying times.

Utilisation of the GAS or ASES processes in the Cu-Indo synthesis enables the entire

process to be performed in a single vessel and at room temperature. With the GAS

process, once reaction is complete, the expanded solution is filtered and the precipitated

drug washed with antisolvent. When the synthesis is performed by means of the ASES

process the antisolvent—solvent mix is continuously removed from the precipitated

drug by filtration. Once precipitation is complete, washing of the precipitate is

performed by simply stopping the flow of solution thereby allowing pure antisolvent to

flow through the vessel. The reduced viscosity and increased diffusivity of the dense

gas results in a solution that is easier to filter thus speeding up filtration time. The

antisolvent is then removed from Cu-Indo by a simple depressurisation.

CHAPTER 6

182

DMF

antisolvent

reactantsFiltration Step Filtration Stepresidual

Cu-Indo

DMF

++

+

Wash andCu-Indo

+

residualwash solvent

Drying StepCu-Indo

Conventional Process

DMF

antisolvent

reactants

Filtration Step+

+

Solvent Recovery Solvent Recovery

Heat

Solvent RecoveryDepressurisation

Depressurisationto giveCu-Indo

GAS / ASES Process

Figure 6.1 Comparison between the conventional, GAS and ASES processes for the

synthesis of Cu-Indo.

CHAPTER 6

183

Another advantage of performing the synthesis by GAS or ASES is that, at all times, the

process is under an atmosphere of CO2 with no exposure to the outside environment.

The potential of drug oxidation or contamination by foreign substances is reduced,

which is particularly important in pharmaceutical manufacture.

One of the objectives in the synthesis of Cu-Indo is to maximise yield and, at the same

time, reduce solvent usage. Both reactants need to be soluble to enable the reaction to

take place, therefore the limiting reactant in the Cu-indo synthesis is Cu-Acetate with a

solubility of approximately 60 mg.g-1 (0.3 mmol/g) in DMF at 25°C. Accordingly, the

concentration of indomethacin for complete reaction would need to be 215 mg.g-1 (0.6

mmol/g). Reaction of Cu-Acetate and indomethacin at this concentration in solution

yields Cu-Indo and acetic acid at concentrations of 255.5 mg.g-1 (0.15 mmol/g) and 72

mg.g-1 (0.12 mmol/g) respectively.

6.2. Experimental

All chemicals and reagents used in the synthesis of Cu-Indo are listed in Table 6.1. All

chemicals and reagents were used as supplied.



Grade

Supplier

Copper Indomethacin 90 Biochemical Veterinary Research




Glacial Acetic Acid 99.7 Ajax Chemicals


Ethanol 99.8 CSR Distilleries




CHAPTER 6

184


The GAS experimental apparatus used in the synthesis experiments is shown

schematically in Figure 6.2. The apparatus consists of a high pressure vessel (Jerguson

sight gauge series No. 32) with an internal volume of 60 mL. The reaction vessel

enabled observation of phase behaviour throughout each experiment. A magnetic stirrer

was designed for the high pressure vessel to ensure the solution was well mixed before

and during sampling. A 0.5 µm stainless steel filter frit was situated at the bottom of the

reaction vessel to sparge the CO2 in the liquid phase during pressurisation and to trap

any precipitated solid in the washing step. A 500 mL stainless steel surge tank was

placed before the reaction vessel to maintain constant pressure while a sample was

taken. An ISCO LC-500 syringe pump was used to generate the required pressure.

Pressure monitoring was made possible by the use of a Druck Pressure transducer

(Model DPI 260 ± 0.007 MPa). The temperature was controlled to within 0.1°C by

submerging the apparatus in a water bath heated with a Thermoline Unistat 130 water

heater. The flow of CO2 through the reaction vessel was from the bottom during the

pressurisation step or from the top during the washing step and was controlled by two

three way ball valves (Whitey SS-41XS2). A solvent trap was placed after the three way

ball valves.

Syntheses by GAS experiments were conducted by mixing Cu-Acetate and

indomethacin in a 1:2 molar ratio in DMF. Sufficient DMF was added to adjust the

synthesis solution to the desired concentration. The reaction vessel was then charged

with a 5 to 10 mL volume of the mixture and brought to the desired pressure by passing

CO2 through the filter. Once reaction was complete, the liquid was removed at constant

pressure by passing CO2 from the top of the vessel and pushing the liquid through the

filter. Cu-Indo that had precipitated was washed and dried by removing the solvent rich

liquid phase at a constant pressure between 5.7 and 5.9 MPa and a CO2 flow rate of 2 to

4 mL.min-1. In each batch 200 to 400 mL of CO2 was used to minimise the residual

solvent. The system was then depressurised and a sample taken for analysis.

A schematic of the ASES experimental apparatus is shown in Figure 6.3. The apparatus

is similar to the GAS apparatus with the following exceptions: The magnetic stirrer was

CHAPTER 6

185

P

CO2 Source


Magnetic StirrerThermostatedHeater

Reaction Vessel

Water bath

CO2Reservoir

Filter

T

Solvent Trap

Figure 6.2 GAS experimental apparatus.

P

CO2 Source


Reaction Vessel

Water bath

CO2Reservoir

Filter

T

Reactant Solution

HPLC Pump

Nozzle

Solvent Trap

ThermostatedHeater

Figure 6.3 ASES experimental apparatus.

CHAPTER 6

186

removed. The flow of CO2 through the reaction vessel was from the top and was

controlled by a needle valve located before the solvent trap (Whitey SS-41XS2). A

HPLC pump (Waters 500) was used to feed the reaction solution into the high pressure

vessel through the nozzle. The nozzles used consisted of stainless steel tubes about 10

cm long with internal diameters of 0.229 mm or 1.02 mm.

Syntheses by ASES experiments were conducted by first pressurising the reaction

vessel to the required pressure and then allowing the CO2 to flow through the vessel

from the top down. The CO2 flow rate was controlled by adjusting the needle valve just

before the solvent trap until the pump was flowing at the desired flow rate. Once the

system was in steady state, a mixture of Cu-Acetate and indomethacin in a 1:2 molar

ratio in DMF was pumped into the reaction vessel through the nozzle using the HPLC

pump. The reaction solution flow rate was controlled by the HPLC pump. During the

synthesis the CO2 flow rate was maintained by adjusting the needle valve. Once enough

precipitate had been collected, the flow of reaction solution was stopped. The CO2 flow

rate was maintained until the precipitate was sufficiently washed. The volume of CO2

used for washing the precipitate was measured using the CO2 pump. Once washing was

complete, a small amount of CO2 was purged through the nozzle to remove any

remaining solution and the reaction vessel was depressurised and samples of precipitate

were taken for analyses.

The masses of Cu-Indo and indomethacin were determined by High Pressure Liquid

Chromatography (HPLC). The HPLC used consisted of a Waters 600 pump, a 996

Photo Diode Array detector and a 717Plus Autoinjector. The samples were separated

using a Symmetry C18 column with a mobile phase made up of 70% acetonitrile and

30% 0.01M tetramethyl ammonium chloride in 0.1% acetic acid. A flow rate of 1

mL.min-1 was used and the analytes were detected at a wavelength of 320 nm.

The mass of Cu-Acetate was determined by atomic absorption spectroscopy (AAS)

using a Perkin Elmer Analyst 300. The analysis conditions were as follows: air-

acetylene flame, 324.8 nm wavelength and a 0.7 nm slit width.

CHAPTER 6

187

Differential Scanning Calorimetry (DSC) (2010 TA Instruments) was conducted by

heating 10 mg of sample in aluminium pans from 25 °C to the desired upper limit at 10

°C/min.

Room temperature X-band Electron Paramagnetic Resonance (EPR) spectroscopy was

used to determine the integrity of the Cu-Indo dimer. EPR spectra were measured using

a Bruker EMX EPR at X-band frequencies of approximately 9.5 GHz and ambient

temperatures.

Solid-state infrared spectroscopy (400-4000 cm-1, KBr matrix and background) spectra

were recorded using a Bio-Rad FTS-40 spectrophotometer fitted with a DRIFTS

attachment (resolution 4 cm-1 at 4000 cm-1, 2 cm-1 at 2000 cm-1 and 1 cm-1 at 1000 cm1).

A KBr background was recorded prior to each measurement, with a 30 s CO2 purge of

the chamber prior to each series of scans (16 scans per sample). All spectra were

normalised on their most intense peak.

6.3. Synthesis of Cu-Indo by the GAS Process

The results of the GAS synthesis experiments are given in Table 6.2. The target Cu-

Indo concentration was varied from 5 mg.g-1, below the ternary saturation concentration

to 200 mg.g-1 which is close to the maximum possible synthesis concentration.

The behaviour of the system during the synthesis was dependent on the concentration of

solute. In the dilute solutions (e.g. 5, 10, and 20 mg.g-1) Cu-Indo precipitated from

solution at 5.2 MPa. The higher concentration solutions (e.g. 50, 100 and 200 mg.g-1)

resulted in a second liquid phase, which appeared as liquid droplets. This second liquid

phase dried into solid Cu-Indo upon further pressurisation. The extent of formation of

the second liquid phase increased with increasing concentration of reactant and hence

Cu-Indo. The second liquid phase was hardly noticeable when 50 mg.g-1 of Cu-Indo

was produced, while at 200 mg.g-1 it was significant. The nature of the precipitates was

a function of the Cu-Indo concentration produced during the synthesis. Free flowing

precipitates made up of individual particles were collected from dilute solutions, while

agglomerated particles were formed from solutions of higher concentration.

CHAPTER 6

188

Table 6.2 The yield and purity of the Cu-Indo synthesis in DMF expanded solution

at 25°C. (GAS Process)

Max.

Pressure/

MPa

Wash

Volume/

mL

Time to

Max.

Expansion/

min.

Initial

Conc. of

Cu-Indo/

mg.g-1

% Yield % Purity

5.8 200 120 5 85 —

5.8 200 120 10 91 —

5.9 200 10 20 98 96.3

5.9 200 10 20 98 93.9

5.9 200 120 50 98 93.5

5.9 400 120 200 96 93.5

5.9 200 120 200 96 91.0

5.9 400 10 200 96 100

5.9 400 120 200 96 100

The morphology of the precipitate obtained after synthesis using the conventional

process is shown in Figure 6.4. The primary crystals have agglomerated into large

irregular clumps. The reason for the aggregation is most likely due to caking that occurs

during the filtration, washing and drying steps. Examples of the precipitate morphology

obtained from the GAS process synthesis experiments are shown in Figure 6.5. At low

concentrations the precipitate consists of single disperse rhombic crystals of uniform

size. The benefit of using a low viscosity dense gas as the antisolvent is clearly

demonstrated by the fact that the crystals have not been damaged even after extensive

washing. At a concentration of 200 mg.g-1 the morphology of the crystals changed from

rhombic to irregular bi-pyramids. A thorough discussion of the effect of process

parameters on the precipitate morphology will be given in Chapter 7.

CHAPTER 6

189

100 µm

Figure 6.4 SEM image of the Cu-Indo obtained from the conventional synthesis.

CHAPTER 6

190

(b)

(a)

10 µm

20 µm

Figure 6.5 SEM images of the precipitates collected after GAS processing at 25°C.

(a) A 5 mg.g-1 solution of Cu-Indo in DMF expanded slowly without stirring. (b) A

200 mg.g-1 solution of Cu-Indo in DMF expanded rapidly with stirring.

CHAPTER 6

191

The yield of each synthesis experiment was calculated by comparing the initial

concentration of Cu-Indo to the concentration in the liquid removed once precipitation

had ceased. The yield was greater than 90%, except at lower concentrations where the

yield was 85%. The amount of Cu-Indo that remained soluble in the expanded DMF

was a function of the acetic acid concentration. Quaternary data for the CO2—DMF—

Cu-Indo—acetic acid system indicates that the solubility of Cu-Indo increased in the

presence of acetic acid. During the synthesis four moles of acetic acid are formed for

each mole of Cu-Indo formed, therefore, the yield decreases due to the increased

concentration of acetic acid. However, a yield of 90% is quite reasonable.

The purity of the precipitates collected was determined by HPLC and was found to vary

between 91 and 100%. The purity of Cu-Indo depends on a number of factors such as

residual solvent, co-precipitants and molecules of solvation. A thorough characterisation

of Cu-Indo indicated that it is not unusual for molecules such as water and DMF to be

included in the Cu-Indo crystals.15,16 The water most likely comes from Cu-Acetate

which has a water molecule included in its crystal. Water and DMF do not form part of

the Cu-Indo structure but are included in the crystal as molecules of solvation. These

molecules are difficult to remove as they are quite strongly bound in the crystals. A

purity of greater than 94% is therefore considered acceptable. Purities greater than 94%

will occur when molecules of solvation are not included in the crystal.

The volume of CO2 used for washing was an important factor in removing residual

solvent. Cu-Indo washed with 400 mL of CO2 possessed less residual solvent than that

washed with 200 mL of CO2. The rate of expansion has little effect on the purity of Cu-

Indo. Similar purity results were obtained for both rapid and slow expansion

experiments.

Two Cu-Indo synthesis experiments were conducted at 25°C, one with excess Cu-

Acetate and the other with excess indomethacin, to determine what effect this would

have on the Cu-Indo precipitation. The results are given in Table 6.3. A 5% excess of

Cu-Acetate or indomethacin was added in each case.

CHAPTER 6

192

Table 6.3 The yield and purity of the Cu-Indo synthesis in DMF expanded solution

at 25°C in the presence of excess reactant. (GAS process)

Max.

Pressure/

MPa

Wash

Volume/

mL

Time to

Max.

Expansion/

min.

Reactant in

Excess

% Yield % Purity

5.9 400 120 Indomethacin 99 92.9

5.9 400 120 Cu-Acetate 96 94.8

In the presence of excess Cu-Acetate, the yield of Cu-Indo was 96% and the purity of

the precipitate collected was 94%. An excess of Cu-Acetate does not have any effect on

the purity of the Cu-Indo produced. This is contrary to the data for the CO2—DMF—

Cu-Indo—Cu-Acetate system presented in Chapter 5, which indicated that Cu-Acetate

would precipitate with Cu-Indo. This result demonstrates the complex interactions that

occur in multicomponent mixtures. The presence of the acetic acid has resulted in an

increased solubility of Cu-Acetate in expanded DMF.

The presence of excess indomethacin resulted in a Cu-Indo yield of 99% and a Cu-Indo

purity of 93%. In this case the yield has improved which can be explained by the

quaternary data for the CO2—DMF—Cu-Indo—indomethacin system. The solubility of

Cu-Indo was reduced in the presence of indomethacin, which resulted in more Cu-Indo

precipitating and an increase in the yield of the synthesis. The purity remained the same

implying that the excess indomethacin has remained in solution.

6.4. Synthesis and Purification of Cu-Indo by the ASES Process

A limitation of the GAS process is that the total volume of solution that can be

processed at one time is limited by the size of the vessel. At final conditions of

expansion when performing a GAS precipitation, the liquid solution has expanded by as

much as 900%. The vessel has to be approximately ten fold larger in volume than the

quantity of solution to be processed. A more efficient method, in terms of the quantity

of precipitate that could be obtained for a single run, would be to perform the synthesis

using the ASES process. For example, consider a solution of Cu-indo in DMF at a

CHAPTER 6

193

concentration of 200 mg.g-1 to be expanded in a 100 mL vessel with CO2. In order to

allow for 900% expansion, a maximum of 10 mL of solution could be added per batch.

The approximate yield of Cu-Indo, assuming 100% precipitation, would be 2 g. If the

synthesis was performed using the ASES process, the yield of precipitate that could be

collected in one batch is theoretically limited to the volume of precipitate that could fit

into the 100 mL vessel.

Experiments were performed to determine the applicability of the ASES process to the

synthesis of Cu-Indo and the results of the experiments are given in Table 6.4. The

target Cu-Indo concentration was varied from 20 mg.g-1 to 200 mg.g-1. In all

experiments, a 10 cm stainless steel tube with an internal diameter of 1020 µm was used

as a nozzle. The solution flow rate was 0.2 mL.min-1, the solution flow time was 10

min., and the CO2 flow rate was maintained at between 4 and 5 mL.min-1 during the

precipitation and at 2.5 mL.min-1 during the washing step. A wash volume of liquid CO2

of 200 mL at the pressure and temperature of the experiment was used for each

experiment.

Table 6.4 The purity of the Cu-Indo synthesis in DMF expanded solution at 25°C.

(ASES Process)

Pressure

/ Mpa

Temperature

/ °C

Conc. of

Cu-Indo/ mg.g-1

% Purity

6.89 25 200 98.5

10.34 25 200 102.4

14.48 40 200 100.8

6.89 25 20 97.9

10.34 25 20 101.2

14.48 40 20 103.1

The concentration of Cu-Indo in solution, as in the GAS experiments, affected the

behaviour of the system. In the case of the 200 mg.g-1 solutions, precipitation occurred

as soon as the solution entered the vessel through the nozzle and collected on the sides

and at the bottom of the vessel. After a few minutes, precipitate started collecting

CHAPTER 6

194

around the tip of the nozzle. The same result was observed for all conditions using the

200 mg.g-1 solution.

A different behaviour was observed at all pressures and temperatures examined when

the 20 mg.g-1 solutions were sprayed into the vessel. A cloud of precipitate formed as

soon as the solution entered the vessel which remained suspended in the vessel.

Towards the end of the spraying process fluffy aggregates of some of these particles

were seen to form, at which stage they fell to the bottom of the vessel.

For all experiments performed the Cu-Indo obtained was greater than 97.9% pure. The

purity results greater than 100% are interesting. They indicate that some of the solvent

molecules that are coordinated to the copper atoms may have been removed.

Examples of the precipitate morphology obtained from the ASES process synthesis

experiments are shown in Figure 6.6. At 20 mg.g-1, the precipitate consisted of

aggregates of microspheres. The primary particles were approximately 2 µm in

diameter. At 200 mg.g-1 concentrations, the morphology changed from uniform

microspheres to large spheres of various diameters. These examples are only given as an

indication of the type of precipitates that are obtained from ASES processing and a

thorough discussion of the effect of process parameters on the precipitate morphology

will be given in Chapter 7.

6.5. Characterisation of the GAS and ASES Processed Cu-Indo

The structure of the Cu-Indo complex is thought to be an integral part of the mechanism

of its action as an anti-inflammatory.16 It is therefore important to ensure that GAS or

ASES processing does not alter the basic Cu-Indo molecular structure. The SEM images

have shown that the morphology of the drug is significantly altered after GAS or ASES

processing, particularly at higher concentrations. In order to ascertain whether this

alteration had any effect on the molecular structure a number of analytical techniques

were used to characterise a few of the samples collected.

CHAPTER 6

195

(b)

(a)

100 µm

5 µm

Figure 6.6 SEM images of Cu-Indo precipitates collected after ASES processing at

25°C and 6.89 MPa using a solution flow rate of 0.2 mL.min-1 and a 1020 µm

nozzle. (a) 20 mg.g-1, (b) 200 mg.g-1 solution of Cu-Indo in DMF.

CHAPTER 6

196

The solid state infrared, X-band EPR and DSC spectra of the conventionally synthesised

Cu-Indo are compared to the spectra of the GAS and ASES processed Cu-Indo in

Figure 6.7 to 6.9 respectively.

Infrared spectroscopy reveals distinct bands for Cu-Indo that correspond to the stretches

of the carbonyl groups on the indomethacin ligand. Indomethacin contains two carbonyl

groups, a carboxylic acid and an amide group. These two groups give rise to three main

bands, namely υsym(COO) at 1410, υasym(COO) at 1620 and υ(C=O) amide at 1669 cm-

1.16 The band at 1410 cm-1 is of particular importance because it is a shift from 1300

cm-1 in free indomethacin due to coordination to copper. The shift, although not

definitive of coordination, does suggest that coordination has taken place. All the IR-

spectra of the precipitates shown in Figure 6.7 display similar patterns which suggests

that the structure of the Cu-Indo complex after GAS or ASES processing was the same

as that produced by the conventional process.

The X-band EPR spectra of Cu-Indo exhibit a broad resonance at geff ~ 2.1 (H⊥2 ~ 4720

G), with weak features at ~ 5890 (Hz2) and 500 G (Hz1) due to the spin-triplet state of

the dimeric complex and a small resonance at 3300 G due to a small quantity of Cu(II)

monomer impurity that is always present in the precipitate.16 X-band EPR reveals

whether the powder examined contains the complex in the form of a Cu—Cu dimer.

The spectra in Figure 6.8 all show resonances similar to the conventionally processed

material indicating that the dimeric nature of the drug has been maintained.

The DSC spectra of all the precipitates processed consist of a broad energy change

between 100 and 150°C, a narrow change at 205°C and then a broad peak above 210°C.

Thermogravimetric analysis, which has been reported previously for Cu-Indo16,

resulted in a weight loss upon heating from ambient to 130°C. The loss in weight was

attributed to the evaporation of solvent molecules. Correspondingly, the broad energy

change from 100 to 150°C in the DSC can be attributed to the loss of solvent. The

narrow peak at 205°C is due to the phase change on melting of Cu-Indo and the broad

peaks above 210°C are due to decomposition of the Cu-Indo. On comparison, the

conventionally processed material has a larger peak between 100 and 150°C compared

CHAPTER 6

197

Rel

ativ

e In

tens

ity

400 600 800 1000 1200 1400 1600 1800 2000

Wavenumber / cm-1

(a)

(b)

(c)

(d)

Figure 6.7 Comparison between the Infrared spectra of the conventionally

synthesised Cu-Indo and the GAS and ASES processed Cu-Indo. (a) Conventional

Synthesis, (b) GAS Process, 5 mg.g-1 solution, (c) GAS Process, 200 mg.g-1 solution,

(d) ASES Process, 20 mg.g-1 solution

CHAPTER 6

198

(a)

(b)

(c)

(d)

Magnetic Field / G0 2000 4000 6000 8000 10000

Figure 6.8 Comparison between the X-band EPR spectra of the conventionally

synthesised Cu-Indo and the GAS and ASES processed Cu-Indo. (a) Conventional

Synthesis, (b) GAS Process, 5 mg.g-1 solution, (c) GAS Process, 200 mg.g-1 solution,

(d) ASES Process, 20 mg.g-1 solution

CHAPTER 6

199

(a)

(b)

(c)

(d)

∆H

Temperature / °C50 100 150 200 250

Figure 6.9 Comparison between the DSC spectra of the conventionally synthesised

Cu-Indo and the GAS and ASES processed Cu-Indo. (a) Conventional Synthesis,

(b) GAS Process, 5 mg.g-1 solution, (c) GAS Process, 200 mg.g-1 solution, (d) ASES

Process, 20 mg.g-1 solution

CHAPTER 6

200

with the GAS and ASES processed materials with the precipitate produced by ASES

having the smallest peak. The indication is that synthesising Cu-Indo by the GAS and

ASES processes results in less solvent impurities being present in the final precipitate,

which supports the findings of the purity analyses above. The melting point is consistent

for all the precipitates indicating that GAS and ASES processing has not altered the

molecular structure of the molecule.

6.6. Comparison between CO2 and Ethanol as Antisolvents

The conventional synthesis for Cu-Indo involves the use of ethanol as an antisolvent.

The results of CO2 and ethanol as antisolvents for the DMF—Cu-Indo system in terms

of yield of precipitate relative to mole fraction of antisolvent at 25°C are given in Table

6.5. The yield of Cu-Indo precipitated was greater than 90% for mole fractions of

ethanol from 0.5 to 0.95. In the CO2 system the yield of Cu-Indo was greater than 90%

at mole fractions of CO2 greater than 0.8. Ethanol is a better antisolvent in terms of

yield at low mole fractions and is as effective as CO2 at higher mole fractions.

Carbon dioxide and ethanol are compared as antisolvents in the synthesis of Cu-Indo in

Table 6.6. Sufficient ethanol was added to each solution to give a mole fraction of 0.95.

The synthesis solutions contained target Cu-Indo concentrations of 5, 50 and 200mg.g-1.

The use of ethanol as an antisolvent only resulted in 40 and 75% yields for the 5 and 50

mg.g-1 solutions, respectively. A 91% yield was obtained for the 200 mg.g-1 solution. At

the same mole fraction CO2 gave yields of 85% for the 5 mg.g-1 solutions and greater

than 95% for the 50 and 200 mg.g-1 solutions. The data in Table 6.6 demonstrate that for

the Cu-Indo synthesis CO2 is more effective as an antisolvent, especially when using

low concentrations of Cu-Indo.

The benefits of replacing ethanol with CO2 are to give improved yields and to reduce

the time taken to crystallise Cu-Indo. Crystallisation was complete within 2 hours at

25°C when CO2 was used as the antisolvent, while at least 24 hours was required when

ethanol was used as an antisolvent. An added adavantage is the ease of recycling CO2.

Purification of the CO2, once it has been used as the antisolvent, is achieved by

CHAPTER 6

201

depressurising in a cold trap. Purification of ethanol is achieved by distillation which is

expensive in terms of energy used and time taken.

Table 6.5 Comparison between CO2 and Ethanol as antisolvents for the Cu-Indo—

DMF system at 25°C.

Antisolvent Pressure /

MPa

Antisolvent Mole

Fraction

% Yield

0.1 0.5 90

0.1 0.8 96

0.1 0.9 96 Ethanol

0.1 0.95 96

3.5 0.5 60

4.8 0.7 90 CO2

5.5 0.9 95

5.8 0.95 95

Table 6.6 Comparison between GAS and the conventional process in the synthesis

of Cu-Indo at 25°C.

Antisolvent Pressure/

MPa

Cu-Indo Conc./

mg.g-1

% Yield

Ethanol 0.1 40

CO2 5.8 5

85

Ethanol 0.1 75

CO2 5.8 50

98

Ethanol 0.1 91

CO2 5.8 200

96

6.7. Conclusion

The GAS and ASES techniques were successfully used for the synthesis and

purification of Cu-Indo. A high purity product can be achieved in a single step.

Characterisation of the precipitates obtained from the synthesis of Cu-Indo by the GAS

CHAPTER 6

202

and ASES processed indicated that the drug had maintained its molecular structure, the

only differences being that the GAS and ASES processed materials contained fewer

solvent impurities. The reactants and byproducts of the Cu-Indo synthesis had a marked

effect on the solubility of Cu-Indo in DMF expanded with CO2. Replacing CO2 with

ethanol as an antisolvent was found to increase the yield of the process.

Carbon dioxide provides an excellent alternative to ethanol in the synthesis of Cu-Indo,

giving the advantages of a single step synthesis, faster crystallisation rate, reduced

solvent requirement and controllable particle size.

CHAPTER 6

203

6.8. References



2. Bungert, B.; Sadowski, G.; Arlt, W.; "New Processes With Compressed Gases,"

Chem. Ing. Tech. 1997, 69, 298.



Chem. 1994, 42, 1993.





Prog. 1991, 7, 275.

6. Liou, Y.; Chang, C. J.; "Separation of Anthracene from Crude Anthracene Using Gas

Anti-Solvent Recrystallization," Sep. Sci. Technol. 1992, 27, 1277.



26, 517.





Chem. Eng. 1996, 4, 257.










CHAPTER 6

204

13. Savage, P. E.; Gopalan, S.; Mizan, T. I.; Martino, C. J.; Brock, E. E.; "Reactions at

Supercritical Conditions: Applications and Fundamentals," AIChE J. 1995, 41, 1723.

14. Subramaniam, B.; McHugh, M. A.; "Reactions in Supercritical Fluids-A Review,"

Ind. Eng. Chem. Process. Des. Dev. 1986, 25, 1.


Bramley, R.; Delfs, C. D.; Murray, K. S.; Moubaraki, B.; Warwick, B.; Biffin, J. R.;

Regtop, H. L.; "Anti-Inflammatory Dinuclear Copper(II) Complexes with

Indomethacin. Synthesis, Magnetism and EPR Spectroscopy. Crystal Structure of the

N,N-Dimethylformamide Adduct," Inorganic Chemistry. 1999, 38, 1736.



CHAPTER 7

205

7. MICRONISATION OF COPPER INDOMETHACIN USING THE

GAS AND ASES PROCESSES

7.1. Introduction

The parameters of purity and yield are not the only criteria that need to be considered

for many industrial syntheses. For materials such as explosives, catalysts, pigments

and pharmaceuticals, the physical properties such as particle size, particle morphology

and particle size distribution can significantly influence the product quality. This is

especially significant in the synthesis of pharmaceuticals as these physical properties

can have a significant effect on the applicability of a particular delivery system and/or

the pharmaceutical action of the drug. If the product obtained from the actual

synthesis does not meet the stipulated requirements, post synthesis processing such as

ball milling or spray drying is necessary for size reduction.

The GAS and ASES processes have been demonstrated to be effective for the

micronisation of various materials. The primary aim of the work described in this

chapter was to determine the applicability of the GAS and ASES processes as

techniques to micronise Cu-Indo. Of particular interest was the possibility of

micronising Cu-Indo during the synthesis, therefore eliminating the need for post

synthesis processing.

The ability to micronise Cu-Indo is expected to offer a number of advantages over the

particles produced from the conventional synthesis process. Micronisation of Cu-Indo

is expected to increase the dissolution rate of the drug in water. Cu-Indo is relatively

insoluble in water and any increase in dissolution rate is expected to increase

bioavailability. Increased bioavailability offers the possibility of a lower dose having

the required therapeutic response. A lower dose rate reduces the possibility of toxic

side effects and offers a commercial advantage of increased units for a quantity of

drug.

CHAPTER 7

206

The second objective of the work described in this chapter was to determine the GAS

and ASES process variables that affect Cu-Indo particle morphology. The process

conditions examined were chosen to be applicable to commercial processing of the

drug. As discussed in Chapter 4, conflicting results have been reported in the literature

regarding the effect of process parameters on drug particles formed when using dense

gases as anti-solvents. These conflicts have arisen primarily when micronising by

ASES which is indicative of the complexity of the particle formation mechanisms in

operation during the process. The results reported for the micronisation of Cu-indo are

therefore discussed in relation to particle formation mechanisms that are presented in

the literature.

7.2. Experimental

The chemicals and reagents used in the micronisation of Cu-Indo are listed Table 7.1.

All chemicals and reagents were used as supplied.


Chemical/Reagent Purity (%) Supplier






N-methylpyrrolidone 99.5 ISP Technologies Inc.

Dimethylsulfoxide 99.8 Ajax Chemicals


Iso-propyl Alcohol 99.9 Aldrich Chemical Co.


The GAS and ASES experimental apparatus used in the micronisation experiments

were identical to those described in Chapter 6 for the synthesis of Cu-Indo (Figures

6.2 and 6.3).

CHAPTER 7

207

GAS and ASES micronisation experiments were conducted with two types of

solution. The first type of solution was prepared by dissolving Cu-Indo in a suitable

solvent. The second type of solution examined was prepared by mixing Cu-Acetate

and indomethacin in a 1:2 molar ratio in a suitable solvent, as for the synthesis of Cu-

Indo. Copper indomethacin forms in the presence of acetic acid during the synthesis.

The presence of acetic acid enhances the solubility of Cu-Indo, therefore enabling

dissolution of up to 200 mg.g-1 of Cu-Indo in DMF.

Micronisation by GAS experiments was conducted by charging the vessel with 5 to 10

mL of solution. The vessel was then brought to the desired pressure by passing CO2

through the filter from the bottom. The rate of pressurisation was controlled by means

of a needle valve. Two pressurisation rates were examined. The slow pressurisation

rate experiments were conducted by increasing the pressure from atmospheric to the

desired maximum over 40 minutes. The fast pressurisation rate experiments were

conducted by increasing the pressure from atmospheric to the desired maximum over

1 minute. Once the precipitation was complete, the liquid was removed at constant

pressure by passing CO2 from the top of the vessel and pushing the liquid through the

filter. The Cu-Indo that precipitated was washed and dried by removing the solvent

rich liquid phase at a constant pressure between 5.7 and 5.9 MPa and a CO2 flow rate

of 2 to 4 mL.min-1. In each batch 200 to 400 mL of CO2 was used to minimise the

residual solvent. The system was depressurised and a sample taken for analysis.

Micronisation by ASES experiments was conducted by first pressurising the reaction


from the top down. The CO2 flow rate was controlled by adjusting the needle valve

just before the solvent trap until the pump was flowing at 4 to 5 mL.min-1. Once the

system was in steady state the solution was pumped into the reaction vessel through

the nozzle using the HPLC pump. The solution flow rate was controlled by the HPLC

pump. Once enough precipitate had been collected, the flow of solution was stopped.

The precipitate was washed by maintaining the flow of CO2 through the vessel until

200 mL of CO2, at the pressure in question had passed through the vessel. The volume

of CO2 used for washing the precipitate was measured using the CO2 pump. Once

CHAPTER 7

208

washing was complete, a small amount of CO2 was purged through the nozzle to

remove any remaining solution. The reaction vessel was then depressurised and

samples of precipitate were taken for analysis.

Precipitation of Cu-Indo by ASES with the anti-solvent existing as both a vapour and

a liquid were performed by pressurising the vessel until it was half full with liquid

CO2. The flow of CO2 was then stopped. In order to prevent a large accumulation of

solvent in the vessel, the solution was pumped through the nozzle for approximately 2

minutes and then stopped. Once solution flow was stopped the pressure was increased

to 6.89 MPa and the precipitate was washed with CO2 as described for the other ASES

experiments.

Particle morphology of the solid powders was determined using a scanning electron

microscope (SEM) (Hitachi S4500). Prior to use, samples were mounted on metal

plates and gold coated using a sputter coater under vacuum.

Particle size distributions of the powders were determined using laser diffraction

(Mastersizer, Malvern, UK). The samples were suspended in iso-propyl alcohol and

sonicated for one minute prior to analysis.

X-ray diffraction (XRD) was used to compare the level of crystallinity between the

powders formed. Measurements were made on a Siemans D-500 diffractometer using

Cu-K(alpha) radiation (λ = 1.54056 Å). Samples were placed in aluminium sample

holders and were scanned from 3 to 23° at a scanning rate of 2°.min-1.

7.3. Micronisation by the GAS Process

The size of particles produced by micronisation with GAS has been reported to be

dominated by the rate of expansion of the solution.1-5 The GAS process is essentially

one of nucleation and growth of particles from a supersaturated solution. The size of

particles is then reduced at high expansion rates, due to the high levels of

supersaturation generated, which result in rapid rates of nucleation and consequently

little particle growth.

CHAPTER 7

209

The morphology of particles formed from the GAS process has also been reported to

be affected by the rate of expansion of the solution.4,6-8 The relationship between

particle morphology and rate of expansion of solution arises from the stability of the

environment surrounding a growing particle. The uniformity of crystal morphology is

dependent on the stability of the growth environment during expansion. A slow

expansion rate provides a stable environment for crystal growth resulting in a

uniformly crystalline solid. Conversely, a fast expansion rate results in a turbulent

environment giving rise to particles of less crystalline nature.

The process parameters that have been examined when micronising Cu-Indo by the

GAS process are expansion rate, solvent, stirring, temperature and solute

concentration. The results of the GAS micronisation study are summarised in Table

7.2.

7.3.1. The Effect of Expansion Rate and Stirring

Examples of Cu-Indo particles produced from DMF solutions containing Cu-Indo at a

concentration of 5 mg.g-1 at 25°C and 5.8 MPa are shown in Figure 7.1. A

concentration of 5 mg.g-1 was chosen to ensure that the concentration of Cu-Indo was

below saturation. Precipitation that occurred was therefore a result of the addition of

CO2 to the solution. The particles produced by slow expansion (in Figure 7.1a) of the

solution had a rhombic morphology and particle sizes of 20 µm. Repeated Cu-Indo

crystallisation at the same conditions produced similar particles but the sizes varied

between 50 and 20 µm. The particles generated from the rapid expansion experiments

were not uniform and two different crystal morphologies were observed. The first type

of morphology was similar to that produced by slow expansion but was larger in the

order of 100 µm. The second type was bi-pyramidal in the order of 20 µm.

CHAPTER 7

210

Table 7.2 GAS micronisation results.

Solvent Temp

/ °C

Conc. of

Cu-Indo /

mg.g-1

Expansion

Rate

Morphology Approximate

Particle S ize / µm

DMF 25 5 Slow

No stirring Rhombic 50 — 20

DMF 25 5 Rapid

No Stirring

Rhombic

Bi-pyramidal

100

20

DMF 25 5 Slow Rhombic 50 — 10

DMF 25 5 Rapid Rhombic 50 — 10

DMF 40 5 Slow Rhombic 50

DMF 40 5 Rapid Rhombic 50

NMP 25 5 Slow Aggregates 100 — 50

NMP 25 5 Rapid Aggregates 100 — 50

DMSO 25 5 Slow Aggregates 100 — 50

DMSO 25 5 Rapid Aggregates 100 — 50

DMF 25 200* Slow Bi-pyramidal 2 — 20

DMF 25 200* Rapid Bi-pyramidal 2 — 20




Slow Expansion— 0 to 5.8 MPa over 40 minutes.

Fast Expansion— 0 to 5.8 MPa over 1 minute.

* — These solutions are formed during the synthesis of Cu-Indo and contain acetic acid.

The Cu-Indo particles formed from rapid expansion experiments with a bi-pyramidal

morphology may be the result of a second liquid phase that formed due to inefficient

mixing. Inefficient mixing of the solution resulted in the temporary formation of a

lower DMF—Cu-Indo rich liquid phase and an upper CO2 rich liquid phase. At the

interface of the two phases a precipitate formed that diffused into the DMF rich phase.

Further pressurisation of the solution with CO2 resulted in the two liquid phases

merging. Further precipitation occurred from the single expanded liquid phase. The

CHAPTER 7

211

(b)

(a)

20 µm

100 µm


25°C from a 5 mg.g-1 solution of Cu-Indo in DMF. (a) slow expansion, (b) rapid

expansion.

CHAPTER 7

212

Cu-Indo particles with a bi-pyramidal morphology are thought to occur at the interface

between the two liquid phases and the larger rhombic crystals from the system once

the two liquid phases merge. At the interface of the two liquid phases the level of

supersaturation will be greater due to the higher concentration of CO2 dissolved in the

DMF rich phase in the vicinity of the interface. Similar observations have been

reported in the literature when a liquid—liquid phase separation occurs during the

GAS process.5,9 The level of supersaturation is thought to be one of the main factors

that affect crystal habit.10,11 The change in habit with varying supersaturation levels

arises from the change in the relative growth rates of the various crystal faces. With

higher levels of supersaturation the nucleation and growth of crystals will be more

rapid and it is possible for a kinetically favoured crystal form to develop.

In order to determine whether the bi-pyramidal particles were the result of the second

liquid phase formation, the same experiments were conducted with stirring. Stirring

should prevent the formation of the second liquid phase by preventing the buildup of

CO2 at the liquid-vapour interface. Examples of the Cu-Indo particles produced from

DMF solutions containing Cu-Indo at a concentration of 5 mg.g-1 at 25°C and 5.8

MPa are shown in Figure 7.2. The Cu-Indo particles produced from both the slow and

rapid expansion experiments with stirring had a rhombic morphology with sizes

between 50 and 10 µm. A second liquid phase did not occur even when the solution

was expanded rapidly and there was no evidence of the bi-pyramidal particles that

were produced from non-stirred solutions. These results suggest that the bi-pyramidal

crystals were formed at the interface of the two liquid phases. The significant

difference between the particles produced with and without stirring is the degree of

aggregation and the particle size distribution. Two distinct crystal types were formed

in the agitated system. The particles produced were either single rhombic crystals of

approximately 50 µm in size or clusters of crystals of rhombic crystals less than 10

µm in size. It is known that stirring reduces the level of supersaturation achievable in

a liquid solution.12 The reduction in the level of supersaturation with agitation may

explain the similarity in the crystals formed at slow and rapid expansion rates. Even at

the low expansion rate the maximum level of supersaturation is reached. The low level

of supersaturation achievable with stirring is also evidenced by the fact that the

CHAPTER 7

213

(b)

(a)

100 µm

100 µm


25°C from a stirred 5 mg.g-1 solution of Cu-Indo in DMF. (a) slow expansion, (b)

rapid expansion.

CHAPTER 7

214

thermodynamically stable rhombic crystal form is produced for both the slow and

rapid expansion experiments.


In order to investigate the effect of temperature on the Cu-Indo particles produced

from the GAS process, a stirred solution of Cu-Indo in DMF at a concentration of 5

mg.g-1 was expanded with CO2 at 40°C and up to 7.5 MPa. Examples of the particles

collected from slow and rapid expansion experiments are shown in Figure 7.3 (See

also Appendix VI, Figure VI.1). Single rhombic crystals in the order of 50 µm were

obtained from both expansion rates at 40°C.

An increase in temperature is expected to have a number of effects on the crystallising

system. For an equivalent pressure at a higher temperature, the solubility of CO2 in the

solvent will be lower resulting in a lower level of expansion, the solubility of the

solute will increase and the rate of crystal growth is expected to increase. The first two

effects imply that supersaturation will take longer to achieve and that the level of

supersaturation will not be as high as at lower temperature. Lower levels of

supersaturation and increased growth rates imply that with an increase in temperature

larger particles are expected. In both the slow and fast expansion experiments at 40°C

the particles produced were only slightly larger than those produced at 25°C. The

minimal effect of temperature on particle size has been observed in other GAS

systems.5,11,13 The Cu-Indo particles produced at 40°C suggest that the rate of

nucleation is rapid enough to offset any increase in particle growth rate.

7.3.3. The Effect of Solvent

In order to examine the effect of changing the solvent on the size and morphology of

Cu-Indo particles produced from GAS, solutions of Cu-Indo in NMP and DMSO were

precipitated at both a slow and rapid rate of pressurisation. The solutions were stirred

to prevent second liquid phase formation. The pressure was increased from 0 to 5.8

MPa as described above for the DMF solutions. An example of the particles produced

are shown in Figure 7.4 (See also Appendix VI, Figures VI.2 to VI.7). The particles

CHAPTER 7

215

(b)

(a)

100 µm

100 µm


40°C from a stirred 5 mg.g-1 solution of Cu-Indo in DMF. (a) slow expansion, (b)

rapid expansion.

CHAPTER 7

216

(b)

(a)

100 µm

100 µm


25°C. (a) A stirred 5 mg.g-1 solution of Cu-Indo in NMP expanded rapidly. (b) A

stirred 5 mg.g-1 solution of Cu-Indo in DMSO expanded rapidly.

CHAPTER 7

217

produced from both NMP and DMSO consisted of clusters of needle like crystals.

These clusters ranged in size from 50 to 100 µm. As with the stirred DMF solutions,

the particle size was not a strong function of the rate of expansion.

Changing the solvent from DMF to DMSO and NMP has resulted in a change in Cu-

Indo morphology. Solvent type has been reported to have a dramatic effect on

morphology of particles produced by the GAS process.4,5,7,14-17 Different solvents

provide different environments for particle growth and lead to different growth

mechanisms and, therefore, morphologies.

7.3.4. The Effect of Concentration

In terms of solvent usage and processing steps it would be more economical to use

higher concentration solutions, such as those used in the synthesis of Cu-Indo. To

determine the effect of increasing the concentration of Cu-Indo and the presence of

acetic acid on the Cu-Indo particles produced from the GAS process, DMF solutions

of Cu-Indo at concentrations of 20, 50 , 100 and 200 mg.g-1 were expanded with CO2

at 25°C.

Examples of the Cu-Indo particles produced from the higher concentration DMF

solutions are shown in Figure 7.5 (See also Appendix VI, Figures VI.8 to VI.12). The

particles produced from all the solutions of higher concentration had a bi-pyramidal

morphology. Once again, the rate of expansion had little effect on the morphology or

size of the particles produced.

The change in crystal morphology from rhombic to bi-pyramidal, upon increasing the

concentration of Cu-Indo in DMF, can be explained in terms of the thermodynamic

behaviour of the system. Upon pressurisation with CO2, the solutions of higher

concentration formed a second liquid phase at approximately 4.8 MPa. The second

liquid phase could not be dispersed even with stirring. Some precipitation occurred in

the upper liquid phase but not in the lower solute rich phase. As the pressure was

increased above 4.8 MPa, rapid nucleation occurred in the lower solute rich liquid

phase. After precipitation the two liquid phases merged. An analogous observation has

CHAPTER 7

218

(b)

(a)

20 µm

20 µm


25°C. (a) A stirred 200 mg.g-1 solution of Cu-Indo in DMF expanded rapidly. (b)

A stirred 50 mg.g-1 solution of Cu-Indo in DMF expanded slowly.

CHAPTER 7

219

been made for a number of other systems containing solute in high concentration.18,19

The formation of a second liquid phase is thought to be the result of intermolecular

forces between the solute and solvent. The strong interactions, which leave little

solvent available to interact with CO2, resulted in a second liquid phase forming at

higher pressures. The hypothesis that the bi-pyramidal crystals are the kinetically

favoured morphology is supported by these high concentration experiments. The

formation of the second liquid phase results in CO2 being unable to dissolve in the

solute rich phase until the pressure reaches values where the Cu-Indo solutions of

lower concentration underwent exponential expansion. As the pressure increases, a

point is reached where the CO2-DMF interactions are favoured over the DMF-Cu-

Indo interactions. At this point DMF is extracted from the solute rich phase resulting

in very high levels of supersaturation in the solute rich phase. These very high levels

of supersaturation favour the formation of the kinetically favoured crystal form, which

in the case of Cu-Indo is the bi-pyramidal morphology.

7.3.5. Summary of the effects of GAS Process Variables

The dominant process variables that effect particle size and morphology when

precipitating Cu-Indo by GAS are solvent type and solute concentration. The change

in morphology from rhombic to bi-pyramidal occurred at conditions where levels of

supersaturation and rates of nucleation were greatest. The bi-pyramidal morphology is

therefore the kinetically favoured crystal form of Cu-Indo.

The size and morphology of Cu-Indo was relatively insensitive to the rate of

expansion and temperature. The insensitivity to the rate of expansion is surprising as

this is often reported to be the dominant variable in the GAS process. The reason for

this is most likely that the level of supersaturation achieved at both expansion rates

investigated was similar. Nucleation rates would then be similar and similar particle

sizes would be expected. The experiments conducted without stirring and at higher

concentration demonstrated that an increase in nucleation rate results in a change in

Cu-Indo morphology from rhombic to bi-pyramidal. The formation of rhombic

particles is therefore evidence of a lower rate of nucleation.

CHAPTER 7

220

From a micronisation perspective, the best particles of Cu-indo were produced from

DMF solution with concentrations greater than 20 mg.g-1. At these conditions the

smallest particles were produced with most particles being less than 20 µm in size.

7.4. Micronisation of Cu-Indo by the ASES Process

Although the ASES process has been used to micronise numerous compounds there is

at present little understood about the ASES particle formation mechanisms. Currently

there are two schools of thought regarding the particle formation mechanisms that

may be in operation.

The first mechanism proposed by Dixon and co-workers20 was based on the idea that

a solution, when injected through a nozzle into a dense gas environment, would

behave in a similar manner to existing spraying processes. The liquid jet, upon

entering the antisolvent environment, breaks up into droplets. Mass transfer processes

of anti-solvent into and solvent out of the droplets results in the formation of particles

within the confines of each droplet. The size of the droplets determines final particle

size.

The second mechanism proposed was by Lengsfeld and co-workers21 who suggested

that droplets do not form when the solution enters the antisolvent. At conditions where

the antisolvent and solvent are totally miscible a rapid decrease in surface tension

around the jet results in the solvent spreading out in a similar manner to a gaseous jet.

Nucleation and growth of particles occurs within this gaseous plume.

Both mechanisms are supported by results obtained in the literature and the exact

mechanism in operation appears to be system dependent. It is important to determine

which mechanism is in operation. Process parameters that determine particle size

differ depending on the particle formation mechanism in operation. Adjusting process

variables of the ASES process and determining their effect on the particles produced

can aid in determining the dominant particle formation mechanism.

CHAPTER 7

221

The process parameters that have been examined for the micronisation of Cu-Indo

using the ASES process include solution flow rate, temperature, nozzle diameter,

solvent type and antisolvent density. As mentioned previously two types of solution

were examined, pure Cu-Indo solutions and synthesis solutions. The results obtained

from the ASES process for both types of solutions are listed in Table 7.3.

Table 7.3 Experimental conditions and results for the micronisation of Cu-Indo

using the ASES technique. The solution and CO2 flow rates were 0.2 and 4–5

mL.min-1 respectively unless otherwise stated.

Solvent Conc. /

mg/g

T /

°C P/ Mpa

Nozzle

Diameter / µm Morphology

Size /

µm

DMF 5 25 6.89 1020 B-P < 5 DMF 5 25 6.89 229 B-P < 5 DMF 5 25 13.79 229 B-P < 5 DMF 5 40 14.48 1020 B-P < 5 DMF 5 40 14.48 229 B-P < 5 DMSO 5 25 6.89 1020 P – DMSO 5 25 13.79 1020 P – DMSO 5 40 14.48 1020 P – NMP 5 25 6.89 1020 I 5 — 2 NMP 5 25 13.79 1020 I 5 — 2 NMP 5 40 14.48 1020 I 5 — 2 DMF* 20 25 6.89 1020 S + B-P ∼ 5 DMF* 20 25 10.34 229 S + B-P ∼ 5 DMF* 20 25 6.89 229 S + B-P ∼ 5 DMF* 20 40 14.48 229 B-P ∼ 5 DMF* 100 25 6.89 229 S ∼ 5 DMF* 100 25 13.79 229 S ∼ 5 DMF*+ 100 25 6.89 229 S ∼ 5 DMF*% 100 25 6.89 229 B-P ∼ 5 DMF* 100 25 6.89 1020 S ∼ 5 DMF* 100 25 6.55 229 S ∼ 5 DMF* 100 40 14.48 229 I-S ∼ 5 NMP* 100 25 6.89 229 S ∼ 20 DMF* 200 25 6.89 1020 S + B-P 50 – 10 DMF* 200 25 10.34 229 S + B-P 50 – 10 DMF* 200 25 6.89 229 S + B-P 50 – 10

* — These solutions are formed during the synthesis of Cu-Indo and contain acetic acid.

+ — Solution flow rate of 0.1 mL.min-1. % — Solution flow rate of 0.5 mL.min-1.

B-P — Bi-pyramidal, S — Spherical, I — Irregular, I-S — Irregular spheres, P – Platelets

CHAPTER 7

222

7.4.1. The Effect of Concentration

Examples of Cu-Indo particles produced from DMF solutions containing Cu-Indo at

various concentrations are shown in Figure 7.6. The Cu-Indo particles produced from

solutions at a concentration of 5 mg.g-1 (Figure 7.6a) had a bi-pyramidal morphology

and were less than 5 µm in size. The Cu-Indo particles produced from solutions at a

concentration of 20 mg.g-1 (Figure 7.6b) had a predominantly spherical morphology

with diameters less than 5 µm. A few particles with a bi-pyramidal morphology were

also evident amongst the particles with a spherical morphology. The bi-pyramidal

particles were of a similar size to the particles with a spherical morphology. The Cu-

Indo particles produced from solutions with a concentration of 100 mg.g-1 (Figure

7.6c) had a spherical morphology with diameters less than 5 µm. At a Cu-Indo

concentration of 200 mg.g-1 (Figure 7.6d), the particles produced were spherical with

diameters ranging from 50 to 20 µm. Further examples of the particles produced from

200 mg.g-1 solutions are shown in Figure 7.7. In some cases, the surface of the spheres

were made up of smaller bi-pyramidal particles and in others the surface of the

spheres were smooth. The large spheres were porous and had hollow centres as shown

in Figure 7.7b. Single bi-pyramidal particles were also present amongst the larger

spheres.

The size and morphology of Cu-Indo particles produced at the conditions studied were

a strong function of solute concentration with an overall increase in particle size. An

increasing particle size with increasing concentration has been reported in the

literature and has been explained based on the jet breaking up into liquid droplets.22-26

An increase in solute concentration increases the viscosity of the solution which tends

to delay atomisation. An increase in solute concentration also enables supersaturation

to be achieved more rapidly. The combined effect of increasing viscosity and rapid

precipitation results in particles forming before the jet is able to atomise into small

droplets. Larger particles are therefore formed from larger droplets.

Alternatively, particle size is expected to decrease with increasing concentration if

nucleation is occurring within a gaseous plume. If the jet disperses as a gaseous

plume, the ASES process resembles the GAS process with extremely rapid rates of

CHAPTER 7

223

(c)

(a)

5 µm

5 µm

(b)

5 µm

(d)

10 µm



diameter of 1020 µm, (a) 5 mg.g-1, (b) 20 mg.g-1,(c) 100 mg.g-1, (d) 200 mg.g-1

solution of Cu-Indo in DMF.

CHAPTER 7

224

(b)

(a)

30 µm

50 µm


mg.g-1 solutions of Cu-Indo in DMF at 25°C and 6.89 MPa with a solution flow

rate of 0.2 mL.min-1 and a nozzle diameter of 1020 µm.

CHAPTER 7

225

solvent expansion. Wubbolts and co-workers reported a decrease in particle size with

increasing concentration and attributed their result to increased supersaturation and

rapid rates of nucleation enabling little opportunity for particle growth.27

In the case of the ASES processing of Cu-Indo at the above conditions the particle

formation appears to be a combination of both mechanisms. A negligible change in

particles size is observed when increasing Cu-Indo concentration between 5 and 100

mg.g-1, but an increase in particle size is observed when increasing the Cu-Indo

concentration to 200 mg.g-1. The bi-pyramidal particle morphology which is obtained

from the 5 mg.g-1 solutions is similar to the morphology obtained from the GAS

process at higher concentration conditions. The spherical particles formed at

concentrations between 20 and 100 mg.g-1 concentration of Cu-Indo appear to be

formed from liquid droplets. The formation of spherical particles is, however, not

always indicative of droplet formation. The formation of spherical particles is an

indication of a loss of particle crystallinity. Loss of particle crystallinity has been

attributed to extremely high levels of supersaturation and rapid nucleation rates

preventing the organisation of molecules into a crystalline form.22,23,28-30 It is

feasible that spherical particles can result from nucleation and growth within a

gaseous plume. The formation of microspheres of polystyrene20 and various

proteins16,17 when processing by the GAS processes is evidence of this.

The large spherical particles formed at Cu-Indo concentration of 200 mg.g-1 are

indicative of precipitation occurring from a liquid droplet. Hollow spherical particles

with diameters in the order of 10 to 100 µm have been produced using the ASES

process for solutions of polystyrene in toluene,31 yttrium acetate in DMSO32 and zinc

acetate in DMSO.29 The spheres in the latter two examples had surfaces that were

made up of many smaller primary particles. The proposed particle formation

mechanism for all of the examples given was based on the initial formation of droplets

of solution in the antisolvent. The presence of individual Cu-Indo bi-pyramidal

particles amongst the larger spheres may be a result of the disintegration of the larger

spheres. Reverchon and co-workers have proposed that the presence of such particles

was the result of “exploding balloons”.24,29,33 The proposed mechanism is based on

the formation of a droplet and the subsequent formation of many particles in each

CHAPTER 7

226

droplet. At high enough expansion rates, the droplets continue to expand until they

“explode” resulting in many individual particles.

Observation of the behaviour of the liquid jet when DMF solutions of Cu-Indo, at

various concentrations were sprayed into liquid CO2 was conducted to clarify the

particle formation mechanism in operation. Photographs of a jet of pure DMF and a 5,

100 and 200 mg.g-1 DMF solution of Cu-Indo being sprayed into liquid CO2 at 25°C,

6.89 MPa and a solution flow rate of 0.2 mL.min-1 through a 1020 µm nozzle are


For each solution examined the solution entered the vessel as a stream which had a

diameter similar to the internal diameter of the nozzle. After a few minutes of

spraying, the liquid in the vessel separated into two distinct regions which were

divided by a boundary about 1 cm from the tip of the nozzle. The existence of this

boundary, which is clearly evident in Figure 7.8b, is a result of the co-current flow of

CO2 and solution. The boundary occurs at the point where the solution and CO2

entering the vessel mix. The time delay before the boundary becomes evident is due to

the system reaching steady state conditions. Once steady state conditions are reached

the boundary remains constant.

At a CO2 pressure of 6.89 MPa and a temperature of 25°C, DMF and CO2 are totally

miscible. If jet break-up occurs at these conditions, as for a conventional spray

process, Weber numbers are expected to be large and the jet will atomise into small

droplets. Alternatively, at miscible conditions the rapid disappearance of surface

tension between the solution and the antisolvent results in the jet spreading out in a

gaseous plume and droplets never form.

The photographs shown in Figure 7.8 reveal that below 100 mg.g-1 (Figure 7.8 b and

c) of Cu-Indo in DMF the jet behaviour was similar to that of pure DMF (Figure 7.8a).

Jet break-up occurred for these solutions at approximately 5 mm from the tip of the

nozzle and DMF dissipated into the bulk antisolvent. Precipitation of Cu-Indo occurs

in the area of jet break-up. If jet break-up occurs by atomisation and droplet

formation, then precipitation is occurring within these droplets. If the jet is simply

CHAPTER 7

227

(b)(a)

(d)(c)

Figure 7.8 Photographs of DMF solutions sprayed into liquid CO2 at 25°C and

6.89 MPa at a solution flow rate of 0.2 mL.min-1 through a 1020 µm nozzle, (a)

Pure DMF, (b) 5 mg.g-1, (c) 100 mg.g-1, (d) 200 mg.g-1 Cu-Indo solution.

CHAPTER 7

228

spreading out as a gaseous plume, then precipitation is occurring by nucleation and

growth within this plume.

At a concentration of 200 mg.g-1 of Cu-Indo in DMF a change in jet behaviour

occurred. The jet extended further than the lower concentration solutions before

break-up occurred. Precipitate was seen to form at the tip of the nozzle and forming

particles fell to the bottom of the vessel from a point above the mixing boundary of

DMF and CO2. The jet did not spread out as for the lower concentration solutions.

These observations indicate that at concentrations of 200 mg.g-1 of Cu-Indo in DMF

the increase in solution viscosity results in the jet breaking up into liquid droplets. The

formation of large hollow spheres then results from mass transfer effects occurring on

each droplet.

In summary, increasing the concentration of Cu-Indo in DMF from 5 to 200 mg.g-1

results in a dramatic change in particle size and morphology. The particles produced

from the ASES process using 5, 20 and 100 mg.g-1 solutions of Cu-Indo in DMF are a

quarter of the size of those produced from the GAS process. The ASES particle

formation mechanism at 5 mg.g-1 Cu-Indo concentrations is most likely one of

nucleation and growth within a gaseous plume. At concentrations between 20 and 100

mg.g-1, the particle formation mechanism may be either droplet formation or gaseous

mixing. At Cu-Indo concentrations of 200 mg.g-1, the increased viscosity and surface

tension of the solution results in delayed jet break-up and particles forming from large

droplets.

7.4.2. The Effect of Nozzle Diameter

Cu-Indo solutions in DMF at concentrations of 5, 20, 100 and 200 mg.g-1 were

precipitated by ASES at 25°C and 6.89 MPa at a solution flow rate of 0.2 mL.min-1

through a 229 µm nozzle. Examples of the Cu-Indo particles produced using the

nozzle with reduced diameter are shown in Figure 7.9 (See also Appendix VI, Figures

VI.13 and VI.19).

CHAPTER 7

229

(c)

(a)

5 µm

10 µm

(b)

5 µm

(d)

10 µm



diameter of 229 µm. (a) 5 mg.g-1, (b) 20 mg.g-1, (c) 100 mg.g-1, (d) 200 mg.g-1

solution of Cu-Indo in DMF.

CHAPTER 7

230

For all Cu-Indo concentrations examined, the Cu-Indo particles were similar to those

produced from the corresponding solutions using the 1020 µm nozzle. A reduction in

nozzle diameter had a negligible effect on the morphology or size of the particles

produced.

A negligible change in particle size was reported when decreasing the nozzle diameter

used in the ASES process. This behaviour is indicative of a process where atomisation

is not the dominant process determining particle size.34-36 There are two reasons

which can be used to explain the insensitivity of particle size to nozzle diameter. The

first, based on droplet formation, is that Weber numbers are very large in the ASES

process and the atomisation process is not altered significantly when changing nozzle

diameter. The second reason, based on particles forming by nucleation and growth

within a gaseous like plume, is that reducing nozzle diameter does not alter the rate of

nucleation to any great extent.

7.4.3. The Effect of Antisolvent Density

Cu-Indo solutions in DMF at concentrations of 5, 20, 100 and 200 mg.g-1 were

precipitated by ASES at 25°C and pressures of 10.34 or 13.79 MPa at a solution flow

rate of 0.2 mL.min-1 through a 229 µm nozzle. The Cu-Indo particles produced at

higher CO2 pressure are shown in Figure 7.10.

The particles produced were similar to the particles produced at lower pressure for

each corresponding solution concentration. Changing the anti-solvent pressure from

6.89 MPa to 10.34 and 13.79 MPa had a negligible effect on the size or morphology

of the Cu-Indo particles produced.

At 25°C the density of CO2 at pressures of 6.89 and 13.79 MPa is 0.738 and 0.866

g.mL-1 respectively. An increase in antisolvent density is expected to increase the

Weber number and therefore favour the formation of smaller droplets. From a mass

transfer point of view, an increase in density from 0.738 to 0.866 mg.mL-1 at 25°C

should result in a decrease in the diffusivity of CO2 into the solution and of solvent

into the anti-solvent. The decrease in diffusivity results in a delay in supersaturation of

CHAPTER 7

231

(c)

(a)

5 µm

5 µm

(b)

5 µm

(d)

20 µm


25°C with a solution flow rate of 0.2 mL.min-1 and a nozzle diameter of 229 µm.

(a) 5 mg.g-1 solution of Cu-Indo in DMF at 13.79 MPa, (b) 20 mg.g-1 solution of

Cu-Indo in DMF at 10.34 MPa, (c) 100 mg.g-1 solution of Cu-Indo in DMF at

13.79 MPa, (d) 200 mg.g-1 solution of Cu-Indo in DMF at 10.34 MPa.

CHAPTER 7

232

the liquid solution, a decrease in nucleation rate and the subsequent formation of

larger particles. Enhanced atomisation favours the formation of smaller droplets which

results in an increase in the rate of mass transfer of CO2 into the droplet by increasing

the surface area of the liquid solution. An increase in anti-solvent density is therefore

expected to alter the mass transfer and hydrodynamics of the system in a complex

way.

In the experiments conducted with Cu-Indo in DMF, varying the anti-solvent density

between 0.738 and 0.866 g.mL-1 had a negligible effect on the size or morphology of

the particles produced for Cu-Indo concentrations between 5 and 200 mg.g-1. It is

evident that, for the density range examined, any increase in atomisation and reduction

in mass transfer rates was insufficient to alter the particles produced.

The effect of anti-solvent density was further explored for the Cu-Indo solutions at a

concentration of 100 mg.g-1. The pressure of CO2 was decreased at 25°C to 6.55 MPa.

At this pressure and temperature CO2 exists as both a liquid and vapour. The density

of CO2 is drastically reduced when it exists as a vapour. Particles produced when

injecting DMF solutions of Cu-Indo at a concentration of 100 mg.g-1 at 25°C and 6.55

MPa at a solution flow rate of 0.2 mL.min-1 through a 229 µm nozzle are shown in

Figure 7.11.

The Cu-Indo particles collected were spheres with diameters less than 5 µm. The Cu-

Indo particles were similar to the particles produced when injecting the same solution

into liquid CO2, keeping all other conditions the same.

By altering the phase of CO2 from liquid to vapour the behaviour of the jet changed

significantly. The solution fell through the vapour phase as large droplets which

dispersed as they hit the liquid layer. Precipitation occurred as soon as the droplet hit

the liquid layer. The formation of microspheres upon spraying a solution in the vapour

over liquid regime has been observed previously.37-39 Of these studies Bodmeier and

co-workers,37 and Tan and co-workers38 reported that atomisation of the solution in

the vapour phase was the important factor in forming micro-droplets and the

corresponding microparticles. In these experiments the solution flow rate and nozzle

CHAPTER 7

233

10 µm

Figure 7.11 SEM image of Cu-Indo particles by the ASES process from 100

mg.g-1 solutions of Cu-Indo in DMF at 25°C and 6.55 MPa with a solution flow

rate of 0.2 mL.min-1 and a nozzle diameter of 229 µm.

CHAPTER 7

234

diameter were such that atomisation of the liquid jet occurred in the CO2 vapour

phase. Young and co-workers39 observed the same behaviour as seen for Cu-Indo.

The solution did not atomise but formed large droplets which fell through the vapour

phase into the liquid phase. Rapid mass transfer and intense nucleation resulted in the

formation of microspheres.

The Cu-Indo particle size and morphology produced from the vapour phase

experiments supports the hypothesis that the particle formation mechanism, at higher

pressures, is one of nucleation and growth within a gaseous plume. The formation of

microspheres is then a result of rapid rates of nucleation which occur due to the anti-

solvent rich environment of the process. For these solutions the ASES process

becomes an optimised GAS process.


Cu-Indo solutions in DMF at concentrations of 5, 20 and 100 mg.g-1 were sprayed into

supercritical CO2 at 40°C and 14.48 MPa at a solution flow rate of 0.2 mL.min-1

through a 229 µm nozzle. Examples of the Cu-Indo particles produced at 40°C are

shown in Figure 7.12 (See also Appendix VI, Figure VI.14).

The morphology of the particles produced at all concentrations had a bi-pyramidal

morphology that were less than 5 µm in size. The most significant observation to be

made from the experiments conducted at 40 °C is that the particles produced from the

DMF solutions with a Cu-Indo concentration of 20 and 100 mg.g-1 are no longer

spherical. At a pressure of 14.48 MPa and a temperature of 40°C, CO2 has a density of

0.772 g.mL-1 which falls between the densities studied at subcritical conditions.

An increase in temperature has a multiple effect on the ASES process. The viscosity

of the liquid solution decreases and therefore its velocity through the nozzle should

increase.20 An increase in solution velocity will give larger Weber numbers and

subsequently result in smaller droplets. Mass transfer rates are expected to increase

slightly with increasing temperature.40 Particle growth rates are expected to increase

with increasing temperature.20,24,29,33,40 An increase in temperature therefore results

CHAPTER 7

235

(c)

(a)

5 µm

5 µm

(b)

5 µm



diameter of 229 µm. (a) 5 mg.g-1, (b) 20 mg.g-1, (c) 100 mg.g-1 solution of Cu-Indo

in DMF.

CHAPTER 7

236

in competing effects on particle size. If hydrodynamics dominates, a decrease in

particle size is expected and if growth effects dominate, an increase in particle size is

expected.

In the case of the precipitation of DMF solutions of Cu-Indo at a concentration of 5

mg.g-1, the particle size and morphology was insensitive to changing temperature. For

the 20 and 100 mg.g-1 solution however, increasing the temperature resulted in more

crystalline particles. Increasing the temperature increases the solubility of the solute in

solution. An increase in solubility of Cu-Indo in DMF will tend to lower the level of

supersaturation achievable and a corresponding reduction in nucleation rate results. A

lowering in nucleation rate slows the crystallisation process and more crystalline

particles are able to form.22,23,28-30

The Cu-Indo particles produced at 40°C are further evidence in favour of the particle

formation mechanism being one of nucleation and growth within the expanding

gaseous plume. Interestingly, an increase in temperature did not significantly increase

the size of the Cu-Indo particles. The indication is that the rate of nucleation remained

rapid enough to offset any increase in particle growth.

To further investigate the increase in Cu-Indo crystallinity with increasing

temperature, X-ray diffraction (XRD) spectra were measured for the different types of

precipitates obtained from the ASES process for the 100 mg.g-1 solutions. These

spectra, as well spectra for the unprocessed Cu-Indo, are shown in Figure 7.13. The

XRD spectra of unprocessed Cu-Indo shows that the sample is reasonably crystalline

and is similar to the reported data for Cu-Indo.41 The XRD spectra of the ASES

processed Cu-Indo shows that much of the crystallinity is lost during ASES

processing. The degree of crystallinity of Cu-Indo processed by ASES was slightly

higher at supercritical conditions.

CHAPTER 7

237

7 11 15 19 232q

3

(a)

(b)

(c)

(d)

Rel

ativ

e In

tens

ity

Figure 7.13 X-ray diffraction patterns of Cu-Indo particles produced from the

conventional synthesis process and by the ASES process from 100 mg.g-1

solutions of Cu-Indo in DMF with a solution flow rate of 0.2 mL/min and a nozzle

diameter of 229µm. (a) Conventional synthesis process. (b) 25 °C and 6.89 MPa.

(c) 25 °C and 13.79 MPa. (d) 40 °C and 14.48 MPa.

CHAPTER 7

238

7.4.5. The Effect of Solution Flow Rate

Particles produced from injecting DMF solutions of Cu-Indo at a concentration of

100 mg.g-1, at different solution flow rates through a 229 µm nozzle at a temperature

of 25°C and a pressure of 6.89 MPa are shown in Figure 7.14.

At a solution flow rate of 0.1 mL.min-1 (Figure 7.14a), the particles produced were

spherical with diameters less than 5 µm, similar to those produced at 0.2 mL.min-1

(Figure 7.14b). When the flow rate was increased to 0.5 mL.min-1 (Figure 7.14c), the

morphology of the particles changed from smooth spheres to spheres with an irregular

surface. The particles with an irregular surface resemble the bi-pyramidal particles

produced from these solutions at 40°C (Figure 7.12c). The particles produced at all

flow rates were approximately the same size.

The insensitivity of Cu-Indo particle size to changing solution flow rate can be

attributed to two factors. In terms of hydrodynamics, increasing solution flow rate is

expected to decrease particle size by increasing atomisation of the liquid solution. If,

however, Weber numbers are already very large as in the ASES system, then

increasing solution flow rates will not result in a significant change in atomisation. As

a result, particle sizes are not altered significantly with changing solution flow rate.

The insensitivity of particle size to changing solution flow rate can also be attributed

to a particle formation mechanism that is not dependent on jet atomisation and droplet

formation. If the particle formation mechanism is dominated by the rate of nucleation,

then altering solution flow rates would not be expected to alter particle sizes

significantly.

The change in morphology with increasing solution flow rate can be attributed to the

increase in solvent concentration in the vessel. During an ASES precipitation, the

mole fraction of antisolvent and solvent in the vessel will reach an equilibrium

depending on the relative flow rates of each. In the case of CO2 and DMF this

equilibrium condition is evident by the appearance of a phase boundary just below the

nozzle after a few minutes of operation. If the solution flow rate is increased, or the

antisolvent flow rate decreased, the level of solvent in the vessel will increase. An

CHAPTER 7

239

(c)

(a)

10 µm

10 µm

(b)

10 µm


from 100 mg.g-1 solutions of Cu-Indo in DMF at 25°C and 6.89 MPa using a

nozzle with a diameter of 229 µm. Solution flow rate of (a) 0.1 mL.min-1, (b) 0.2

mL.min-1, (c) 0.5 mL.min-1.

CHAPTER 7

240

increase in solvent will reduce the mass transfer driving force in the vessel. The level

of supersaturation and the rate of nucleation is reduced. The slowing down of the

precipitation process favours the formation of more crystalline material as is observed

in these experiments. The degree of crystallinity of a number of solutes has been

reported to vary with changing solution or antisolvent flow rate when processing them

by ASES.22,28,40

7.4.6. The Effect of Solvent

The final parameter investigated for the precipitation of Cu-Indo was the effect of

changing solvent. Solutions of Cu-Indo in DMSO and NMP at a concentration of 5

mg.g-1 were precipitated by ASES at temperatures of 25 and 40°C and pressures of

6.89, 13.79 and 14.48 MPa. At the conditions investigated, CO2 and both DMSO and

NMP are totally miscible and the volumetric expansion behaviour of the systems are

similar. The solution flow rate was 0.2 mL.min-1 and the nozzle had a diameter of

1020 µm. Examples of the particles collected from 5 mg.g-1 solutions of Cu-Indo

dissolved in DMSO and NMP are shown in Figure 7.15 a and b, respectively (See also

Appendix VI, Figures VI.15 to VI.18).

The precipitate collected from the DMSO solutions (Figure 7.15a) consisted of

agglomerates of irregular platelets. Increasing the density or temperature of the

antisolvent did not alter the morphology of the precipitate. The particles of Cu-Indo

produced from NMP at all temperatures and pressures investigated were of sizes

between 2 and 5 µm and had an irregular morphology (Figure 7.15b).

The morphology of the particles produced when processing DMSO solutions of Cu-

Indo by ASES were of a similar morphology to those obtained when the same

solutions were precipitated by the GAS process. The particles obtained from the

NMP—Cu-Indo solutions were process dependent. The particles produced from the

ASES process were significantly smaller and less agglomerated than those produced

by the GAS process (Figure 7.4 c and d). The rounded edges of the particles produced

CHAPTER 7

241

(b)

(a)

200 µm

10 µm


from DMSO and NMP at a Cu-Indo concentration of 5 mg.g-1 with a solution

flow rate of 0.2 mL.min-1 and a nozzle diameter of 1020 µm. (a) Cu-Indo in

DMSO at 40°C and 14.48 MPa. (b) Cu-Indo in NMP at 25°C and 6.89 MPa.

CHAPTER 7

242

by ASES suggested that they were less crystalline, as was observed for DMF—Cu-

Indo solutions.

The effect of increasing the concentration of Cu-Indo in NMP was investigated by

processing a 100 mg.g-1 NMP—Cu-Indo solution by ASES. The particles obtained are


Two distinct Cu-Indo particle types were obtained when processing 100 mg.g-1

NMP—Cu-Indo solutions. Large spheres with diameters in the order of 20 µm and

irregular particles with sizes in the order of 5 µm were produced. The surfaces of the

large spheres were made up of many smaller particles (Figure 7.16b) which had a

similar morphology to the particles obtained from the 5 mg.g-1 NMP—Cu-Indo

solutions (Figure 7.15b).

Changing the solvent used for dilute solutions of Cu-Indo had a dramatic effect on the

morphology of the particles produced. The differences in morphology can be

attributed to the solvent—solute interactions. The solvent forms an integral part of the

Cu-Indo molecular structure in that the solvent molecule occupies a coordination site

on each copper atom. It would therefore be expected that the morphologies of the

crystals would differ upon changing the solvent. Once again, the process variables of

temperature and antisolvent density have had little effect on the particle size and

morphology. It can therefore be concluded that, as with DMF, the dominant particle

formation mechanism in operation at these concentrations is one of nucleation and

growth after the liquid droplets have dispersed.

Increasing the concentration of Cu-Indo in NMP to 100 mg.g-1 resulted in a change in

particle size and morphology. The particle formation mechanism appears to have

changed to one that may be dominated by droplet formation as discussed for the 200

mg.g-1 DMF solutions. The reason for the change in particle formation mechanism

with changing solvent is complex as the solvent has multiple effects in the ASES

process. Changing the solvent can alter the viscosity, vapour pressure, solubility of the

solute, mass transfer and phase behaviour in the system. A change in one or all of

these variables can alter the particle formation mechanism.

CHAPTER 7

243

(b)

(a)

50 µm

3 µm


from 100 mg.g-1 Cu-Indo solutions in NMP at 25°C and 6.89 MPa with a solution

flow rate of 0.2 mL.min-1 and a nozzle diameter of 229 µm.

CHAPTER 7

244

7.4.7. Summary of the effect of ASES Process Variables

The dominant process parameters which determine Cu-Indo particle size and

morphology when processing by ASES are solute concentration, solution flow rate,

temperature and solvent type. The size of the particles produced increased

significantly when the concentration of Cu-Indo reached 200 mg.g-1 in DMF and 100

mg.g-1 in NMP. At all other concentrations examined, the size of the Cu-Indo particles

was insensitive to process parameters. Particle morphology was found to change from

spherical to bi-pyramidal when increasing temperature or solution flow rate. The

change in morphology could be attributed to decreases in rates of crystallisation.

The change in particle size with increased solution concentration could be explained

by the particle formation mechanism in operation at the experimental conditions

studied. At solute concentrations below 100 mg.g-1, the dispersion of the jet upon

entering the vessel is by a process similar to that of gaseous jets. Droplets are not

formed and the particle formation mechanism is one of nucleation and growth within

the gaseous plume. At solute concentration of 200 mg.g-1 the increase in viscosity

results in increased surface tension and the particle formation mechanism becomes

one of precipitation within forming droplets.

The Cu-Indo particles produced from DMF solutions at concentrations below 100

mg.g-1 are most desirable. The particles produced were less than 5 µm and had a

narrow particle size distribution. Particle sizes below 5 µm are able to be used as

injectable suspensions, in inhalation therapies and as suspensions for use in

ophthalmic applications.

7.5. Particle S ize Distribution

The particle size distribution obtained from the ASES processing of DMF solutions of

Cu-Indo at 20, 100 and 200 mg.g-1 were analysed further using the Malvern

Mastersizer. The results are tabulated in Table 7.4.

CHAPTER 7

245

Table 7.4 Particle size distribution results.

Particle S ize Distribution Conc. /

mg.g-1

T / °C P / MPa Nozzle

Diameter

/ mm

D(v,10%) D(v,50%) D(v,90%)

200 25 10.34 1.02 2.7 4.7 10.1

20 25 6.89 0.229 3.0 5.3 8.6

100 25 6.89 0.229 2.7 (0.3) 4.9 (0.8) 8.5 (1.1)

100 25 13.79 0.229 2.9 (0.1) 5.5 (0.4) 9.7 (0.8)

100 40 14.48 0.229 2.6 (0.4) 4.8 (1.1) 9.8 (1.6)

100 25 6.89 1.02 2.6 (0.1) 4.1 (0.3) 7.1 (0.4)

The solution flow rate was 0.2 mL.min-1 for all experiments.

The results obtained for the particle size distribution support the observation from the

SEM images, with 90% of the volume of particles produced having diameters below

10 µm. The results suggest that little agglomeration of particles has occurred with the

mean particle diameters corresponding to the individual particles observed in the SEM

images.

The particle size distribution of the Cu-Indo precipitates analysed in Table 7.4 are

shown Figure 7.17 (See also Appendix VI, Figures VI.1 to 4). All the precipitates

examined show a similar size distribution between 1 and 10 µm. The formation of

larger spheres at solute concentration of 200 mg.g-1 are evident in Figure 7.17a with a

local maxima in particle diameter at around 25 µm.

The effect of prolonged spraying times on the particle size distribution was

determined by conducting experiments at three different conditions using 100 mg.g-1

solutions of Cu-Indo in DMF. For each experiment the solution was sprayed for 50

minutes. The particle size distribution from the prolonged spraying experiments are

shown in Figure 7.18. A prolonged spraying time has little effect on the particle size

distribution at all conditions examined, which implies that the production of bulk

quantities of precipitate is feasible - an important consideration in terms of scale up.

CHAPTER 7

246

(b)

(a)

0

5

10

15

20

25

30

2.0 2.7 3.6 4.9 6.6 9.0 12.2 16.6 22.5 30.5 48.3Particle Diameter / µm

In %

0

5

10

15

20

25

30

2.0 2.7 3.6 4.9 6.6 9.0 12.2 16.6 22.5 30.5 48.3Particle Diameter / µm

In %

Figure 7.17 Particle size distributions of Cu-Indo particles produced from DMF

solutions by the ASES process. (a) 200 mg.g-1 Cu-Indo solution at 25°C and 10.34

MPa using a 1020 µm nozzle and a solution flow rate of 0.2 mL.min-1, (b) 20

mg.g-1 Cu-Indo solution at 25°C and 6.89 MPa using a 229 µm nozzle and a

solution flow rate of 0.2 mL.min-1.

CHAPTER 7

247

(a)Particle Diameter / µm

In %

0

5

10

15

20

25

30

2.0 2.7 3.6 4.9 6.6 9.0 12.2 16.6 22.5 30.5 48.3

(b)Particle Diameter / µm

In %

0

5

10

15

20

25

30

2.0 2.7 3.6 4.9 6.6 9.0 12.2 16.6 22.5 30.5 48.3

(c)Particle Diameter / µm

In %

0

5

10

15

20

25

30

2.0 2.7 3.6 4.9 6.6 9.0 12.2 16.6 22.5 30.5 48.3

Figure 7.18 Particle size distributions of Cu-Indo particles produced from DMF

solutions at a concentration of 100 mg.g-1 by the ASES process using a 229 µm

nozzle, 0.2 mL.min-1 solution flow rate. (a) 25°C and 6.89 MPa, (b) 25°C and

13.79 MPa, (c) 40°C and 14.48 MPa.

CHAPTER 7

248

7.6. Dissolution Studies

One of the benefits of micronising a drug is the increase in bioavailability due to an

increase in the dissolution rate. The effect of micronisation on the dissolution rate of

Cu-Indo was investigated by comparing the dissolution rate of unprocessed Cu-Indo

with two micronised Cu-Indo precipitates. The first micronised precipitate used in the

study was produced at conditions of 25°C, 6.89 MPa, 0.2 mL.min-1, 229 µm nozzle

and using DMF as the solvent. The second micronised precipitate was produced at

conditions of 40°C, 14.48 MPa, 0.2 mL.min-1, 229 µm nozzle and using DMF as the

solvent. These two precipitates were chosen to determine whether the degree of

crystallinity as well as the reduction in particle size affected the dissolution rate.

As a basis for comparison, the dissolution rate coefficient (Kw) is defined as the

reciprocal of the time after which 63.2% of the original amount of drug has

dissolved.42

The dissolution profiles of the unprocessed and micronised Cu-Indo are shown in

Figure 7.19 (See Appendix V for data). The dissolution rate coefficient of the

unprocessed material was estimated to be 0.002 min-1 by assuming a linear dissolution

rate. The assumption of linear dissolution rate will give an over-estimation of KW but

is still useful for comparison purposes. The dissolution rate coefficient of the

amorphous and crystalline micronised Cu-Indo was found to be 0.015 and 0.016

respectively. The observed dissolution of the micronised drug is approximately 8

times higher than for the unprocessed material.

There was no significant difference in the dissolution rates between the two

micronised powders, which indicates that the increase in dissolution rate is a direct

consequence of the size reduction of the particles and not due to a change in the

degree of crystallinity.

CHAPTER 7

249

Time / min

% D

issol

utio

n

0

10

20

30

40

50

60

70

80

90

100

0 20 40 60 80 100 120 140 160 180 200

25°C, 6.89 MPa

40°C, 14.48 MPa

Unprocessed Cu-Indo

Figure 7.19 Dissolution of Cu-Indo produced by ASES. 100 mg.g-1 solution of Cu-

Indo in DMF using a 229 µm nozzle and a solution flow rate of 0.2 mL.min-1.

CHAPTER 7

250

7.7. Conclusions

The results obtained from the processing of solutions of Cu-Indo by GAS and ASES

indicate that both processes are dominated by a complex combination of process

parameters. The particles produced by the GAS process appear to be mostly

influenced by the concentration of Cu-Indo in solution. The particles produced by the

ASES process were also found to be a strong function of the solution concentration. In

this case, however, the particles produced at the lowest concentration resembled the

particles produced at the highest concentrations when processing by GAS. The similar

particle morphology produced at the opposite extremes of concentration for both

processes indicates that the statement by Bungert and co-workers,43 that the ASES

process is an optimised GAS process in terms of supersaturation, is valid. The

dominant particle formation mechanism at most of the conditions studied for both

processes was one of nucleation and growth within an homogenous liquid phase. At

the high concentration extreme examined using the ASES process however, a particle

formation mechanism based on precipitation occurring within liquid droplets,

appeared to be in operation.

The degree of crystallinity of the precipitate formed from the ASES process was

found to increase at supercritical conditions and at higher solution flow rates. This

result is significant since it provides a method of controlling the level of crystallinity

of a drug.

In terms of micronisation, the ASES produced particles of Cu-Indo were most suited

to drug delivery applications. The results from the particle size distribution study

showed that 90% of the mass of precipitate produced contained particles that were less

than 10 µm in diameter. The particle size distribution was not altered significantly

when altering the pressure, temperature or nozzle diameter, which makes the ASES

process well suited to the production of uniform batches of micronised Cu-Indo.

The micronisation of Cu-Indo was possible at the point of synthesis. The possibility of

micronising at the point of synthesis offers savings in terms of solvent usage, time and

cost, as secondary processing is avoided.

CHAPTER 7

251

The immediate benefit of micronising Cu-Indo was demonstrated with an eight fold

increase in the dissolution rate of the micronised drug when compared to the

unprocessed drug.

CHAPTER 7

252

7.8. References






378.







87, 96.









with Dense Carbondioxide," Proceedings of the 5th Meeting on Supercritical

Fluids, Nice, France, 1998, 291.






Process," AIChE J. 1996, 42, 431.

10. Pamplin, B. R.; "Crystal Growth," ; Permagon Press Ltd.: Oxford, 1975; Vol. 6.

CHAPTER 7

253

11. Yeo, S.-D.; Choi, J.-H.; Lee, T.-J.; "Crystal Formation of BaCl2 and NH4Cl Using

a Supercritical Fluid Anti-Solvent," J. Supercrit. Fluids 2000, 16, 235.

12. Khamski, E. V.; "Crystallization From Solutions," ; Olenum Publishing

Coorporation: New York, 1970.



1997, 53, 232.














19. Tai, C. Y.; Cheng, C.-S.; "Effect of CO2 on Expansion and Supersaturation of






Solvent: Atomization vs Nucleation and Growth," J. Phys. Chem. B 2000, 104,

2725.




23. Reverchon, E.; Porta, G. D.; "Production of Antibiotic Micro- and Nano-Particles

by Supercritical Anti-Solvent Precipitation," Powder Technol. 1999, 106, 23.

CHAPTER 7

254



1998, 37, 952.



1995, 41, 2476.




27. Wubbolts, F.; Bruinsma, O.; van Rosmalen, G.; "Dry-Spraying of Ascorbic Acid

or Acetaminophen Solutions with Supercritical Carbon Dioxide," J. Cryst. Growth

1999, 198/199, 767.


Removal and Polymer Morphology in the Formation of Gentamicin-Loaded

Poly(L-Lactide) Microparticles," Pharm. Res. 1998, 15, 1233.


Solvent Precipitation of Nanoparticles of a Zinc Oxide Precursor," Powder

Technol. 1999, 102, 127.



31. Dixon, D. J.; Luna-Bárcenas, G.; Johnston, K. P.; "Microcellular Microspheres

and Miroballoons by Precipitation with a Vapour-Liquid Compressed Fluid Anti-









35. Heater, K. J.; Tomasko, D. L.; "Processing of Epoxy Resins Using Carbon

Dioxide as an Anti-Solvent," J. Supercrit. Fluids 1998, 14, 55.

CHAPTER 7

255


Carpet Waste using a Supercritical Fluid Anti-Solvent Technique," Polymer

Plastics Technology & Engineering 1999, 38, 411.


Microspheres Prepared By Spraying Into Compressed Carbon Dioxide," Pharm.

Res. 1995, 12, 1211.


Toluene Solution into Compressed HFC-134a," Ind. Eng. Chem. Res. 1999, 38,

3898.




40. Randolph, T. W.; Randolph, A. D.; Mebes, M.; Yeung, S.; "Sub-Micrometer-

Sized Biodegradable Particles of Poly(L-Lactic Acid) via the Gas Anti-Solvent

Spray Precipitation Process," Biotechnol. Prog. 1993, 9, 429.


Bramley, R.; Delfs, C. D.; Murray, K. S.; Moubaraki, B.; Warwick, B.; Biffin, J.

R.; Regtop, H. L.; "Anti-Inflammatory Dinuclear Copper(II) Complexes with

Indomethacin. Synthesis, Magnetism and EPR Spectroscopy. Crystal Structure of

the N,N-Dimethylformamide Adduct," Inorganic Chemistry. 1999, 38, 1736.

42. Loth, H.; Hemgesberg, E.; "Properties and Dissolution of Drugs Micronized by

Crystallization from Supercritical Gases," Int. J. Pharm. 1986, 32, 265.



CHAPTER 8

256

8. CO-PRECIPITATION OF COPPER INDOMETHACIN AND

POLYVINYLPYRROLIDONE BY THE ASES PROCESS

8.1. Introduction

The formation of drug—polymer blends is an important process in the pharmaceutical

industry particularly when the drug in question is not suitable for use in its pure form.

Drug—polymer combinations may be formed to protect the drug from its surroundings,

increase the stability of the drug during storage, increase the solubility of the drug,

control the release of the drug in the body and for target delivery of the drug.

The combination of drug and polymer can occur in two solid forms as shown in Figure

8.1. Drug—polymer composites can exist in the form of a matrix where the drug and

polymer are distributed evenly throughout the solid matrix, or in a form where the drug

is totally encapsulated by the polymer. If the composite is of the matrix type, then either

the dissolution rate of the polymer, or the diffusion of the drug through the matrix,

controls the rate of drug release in the body. If the composite is of the encapsulated

type, then the rate of drug release is controlled by either the diffusion of drug through

the polymer wall, or by dissolution of the polymer wall.

Polymers that are commonly used for controlled delivery systems include polyamides

and polyesters.1 Of these, polyesters have received most attention because of their

predictable degradation properties with poly(lactic acid) being approved by the FDA for

use in humans. Water soluble polymers, such as polyethylene glycol and polyacrylic

acid, have also been used in the delivery of drugs. Water soluble polymers, when

combined with a drug, can facilitate rapid dissolution of the drug in the body, thus

providing faster therapeutic response.

CHAPTER 8

257

��

Encapsulation Matrix

Polymer

Drug

Figure 8.1 Types of polymer—drug composites.

8.1.1. Polyvinylpyrrolidone

Polyvinylpyrrolidone (PVP) is a synthetic water soluble polymer consisting of linear 1-

vinyl-2-pyrrolidone monomer, the molecular structure of which is shown in Figure 8.2.

The PVP polymers are characterised according to their viscosity in aqueous solution,

relative to that of water, expressed as a K-value. Various types of PVP polymers based

on K-values and approximate molecular weights are listed in Table 8.1.

CH2H2C

H2C C=ON

CH nCH2

Figure 8.2 Polyvinylpyrrolidone chemical structure.

CHAPTER 8

258

PVP is non-toxic, has no irritant effect on the skin and causes no sensitization. It has the

property of being soluble in both water and commonly used organic solvents, which

makes it attractive for use in many types of formulations. PVP has been used in the

pharmaceutical industry since the 1940s with its first use being as a plasma expander.2

It has since found widespread application in the pharmaceutical industry. PVP is

commonly used as a binder in wet granulation, a coating to improve tablet

characteristics, an additive to change the dissolution profile of slow release

formulations, a crystal growth retarder, an anti-irritant for topical and opthalmic

products, a stabiliser and protective colloid in suspensions, a solubiliser in injectable

products for veterinary use and a viscosity modifier.

Table 8.1 K values and molecular weights of PVP.

K-value Approximate Molecular Weight /

g.mol-1

12 2500

15 8000

17 10,000

25 30,000

30 50,000

60 400,000

90 1,000,000

120 3,000,000

PVP has been shown to enhance the therapeutic effect of a wide variety of drugs. For

example, the addition of PVP to an injectable prostaglandin was found to extend its

therapeutic effect by increasing its viscosity and decreasing its diffusion rate.3 PVP has

also been shown to be safe for use in injections with no adverse effects reported in a

study of 250 patients injected with oxytetracycline solubilised in PVP.4

The functions of enhanced therapeutic effect and increased solubility are thought to be

the most promising benefits of co-precipitating PVP and Cu-Indo. One of the objectives

of further development of Cu-Indo is to produce new formulations to increase the

CHAPTER 8

259

marketability of the drug. The aim of the work described in this chapter was to provide

a form of Cu-Indo that may be suitable for opthalmic and injectable applications by co-

precipitation with PVP.

The conventional method of preparing a PVP—drug co-precipitate is to solubilise both

the drug and the PVP in a common solvent. The solvent is then evaporated to give the

precipitate. The conventional method of co-precipitation presents limitations when the

solvents used have high boiling points and/or are toxic. In the case of Cu-Indo the

solvents that are suitable for solubilisation are methylene dichloride, DMF, NMP and

DMSO. All of these solvents present problems of either elevated boiling points and/or

toxicity. The conventional method of preparing PVP—Cu-Indo precipitates is,

therefore, unsuitable due to residual solvent levels left in the co-precipitate after

evaporation.

A way of overcoming these problems is to precipitate a solution of the polymer and

drug by the GAS or ASES processes. Of these two processes, ASES is more attractive.

The rapid precipitation time scale of ASES results in the drug and polymer precipitating

in close proximity, thus increasing the probability of forming drug—polymer

composites.

In this study the precipitation of DMF solutions of PVP (average molecular weight of

10,000) and Cu-Indo using CO2 as the anti-solvent was investigated. DMF was chosen

as the solvent because of the successful results achieved using DMF in the

micronisation of the pure drug by the GAS and ASES processes. Preliminary

investigations of the precipitation of PVP from DMF using CO2 as an anti-solvent were

conducted. The aim of these investigations was to demonstrate the applicability of the

GAS and ASES processes to precipitating PVP and to determine optimum conditions

for co-precipitation. The process parameters of solute concentration, temperature and

anti-solvent density were investigated. The effect of solute concentration, PVP to Cu-

Indo ratio and anti-solvent density on the co-precipitates formed was examined. The co-

precipitates formed were then compared to the pure drug in terms of solubility in

ethanol and dissolution rates in water.

CHAPTER 8

260

8.2. Experimental

All chemicals and reagents used in the co-precipitation of PVP and Cu-Indo are listed in

Table 8.2. All chemicals and reagents were used as supplied.



Grade

Supplier


Polyvinylpyrrolidone Average M w

ca. 10 000

Aldrich Chemical Co.


Carbon Dioxide > 99.9 BOC Gases




The GAS and ASES experimental apparatus used in the co-precipitation experiments

were identical to those described in Chapter 6 for the synthesis of Cu-Indo Indo

(Figures 6.2 and 6.3).

Precipitation experiments were conducted using solutions of PVP in DMF and mixtures

of PVP and Cu-Indo in DMF. The solutions that contained only PVP were prepared by

dissolving PVP in DMF to give the required concentration. Solutions used in the co-

precipitation experiments were formulated by weighing out PVP and Cu-Indo into a

beaker in the required ratio. Sufficient DMF was then added to the beaker to give the

desired concentration of solute in solution. The solutes were dissolved by sonication and

mixing.

Precipitation experiments using the GAS process were conducted by charging the vessel

with 5 to 10 mL of solution. The vessel was then brought to the desired pressure by

CHAPTER 8

261

passing CO2 through the filter from the bottom. The rate of pressurisation was

controlled by means of a needle valve. Once the precipitation was complete, the liquid

phase was removed at constant pressure by passing CO2 from the top of the vessel and

pushing the liquid through the filter. The precipitate that formed was washed and dried

by removing the solvent rich liquid phase at a constant pressure 6.8 MPa and a CO2

flow rate of 2 to 4 mL.min-1. In each batch 200 to 400 mL of CO2 was used to minimise

the residual solvent. The system was depressurised and a sample taken for analysis.

Precipitations by ASES experiments were conducted by first pressurising the pressure


from the top down. The CO2 flow rate was controlled by adjusting the needle valve just

before the solvent trap until the pump was flowing at 4 to 5 mL.min-1. Once the system

was in steady state, the solution was pumped into the pressure vessel through the nozzle

using the HPLC pump. The solution flow rate was controlled by the HPLC pump. Upon

completion of precipitation, the flow of solution was stopped. The precipitate was

washed with 200 mL of CO2 at the pressure in question.

Precipitation by ASES with the anti-solvent existing as both a vapour and a liquid were

performed by pressurising the vessel until it was half full with liquid CO2. The flow of

CO2 was then stopped. To prevent a large accumulation of solvent in the vessel, the

solution was pumped through the nozzle for approximately 2 minutes and then stopped.

Once solution flow was stopped the pressure was increased to 6.89 MPa and the

precipitate was washed with CO2 as described for the other ASES experiments.

Drug content of the PVP—Cu-Indo co-precipitates was determined by High Pressure

Liquid Chromatography (HPLC). The HPLC used consisted of a Waters 600 pump, a

996 Photo Diode Array detector and a 717Plus Autoinjector. The samples were

separated using a Symmetry C18 column with a mobile phase made up of 70%

acetonitrile and 30% 0.01M tetramethylammonium chloride in 0.1% acetic acid. A flow

rate of 1 mL.min-1 was used and the analytes were detected at a wavelength of 320 nm.

CHAPTER 8

262

Drug content was calculated using the following equation:

mD

mT × 100 8.1

where mD is the mass of Cu-Indo and mT is the total mass of the powder analysed.

The solubility of the co-precipitate powders was determined by adding an excess

quantity of co-precipitate to approximately 4 mL of ethanol at room temperature. The

mixtures were mixed for 20 min at room temperature. The excess solid was removed by

filtering the solution through a 0.45 µm syringe filter. The solution was analysed for

Cu-Indo content by HPLC as described above.

Dissolution rate experiments were conducted using the USP paddle method and the

Vanderkamp tester (VK6000, Vankel Industries Inc., Germany). In each experiment

enough powder was used to give approximately 10 mg of Cu-Indo. The powder was

mixed with 100 mg of lactose to aid in dispersion during dissolution. The mixture was

then added to 1 L of water (Milli-Q) which had been heated to 37°C. A stirring rate of

100 rpm was used for all dissolution experiments. Approximately 4 mL aliquots were

removed at regular intervals and were analysed for Cu-Indo concentration. The

concentration of Cu-Indo was determined by HPLC as described above.

Particle morphology of the solid powders was determined by scanning electron

microscope (SEM) (Hitachi S4500). Samples were mounted on metal plates and gold

coated using a sputter coater under vacuum prior to use.

Differential Scanning Calorimetry (DSC) (2010 TA Instruments) was conducted by

heating 10 mg of sample in aluminium pans from 25°C to the desired upper limit at 10

°C.min-1.

Room temperature X-band Electron Paramagnetic Resonance (EPR) spectroscopy was

used to determine the integrity of the Cu-Indo dimer. EPR spectra were measured using

CHAPTER 8

263

a Bruker EMX EPR at X-band frequencies of approximately 9.5 GHz and ambient

temperatures.

8.3. Precipitation of PVP Using Dense CO2 as Anti-Solvent

Previous reports on the micronisation of polymers by ASES have shown that particle

agglomeration was most likely when processing amorphous polymers with low glass

transition temperatures.5-7 At conditions commonly used in gas anti-solvent processes,

polymer glass transition temperatures are reduced to below operating conditions and

plasticisation occurs.

Typical properties of two PVP polymers are listed in Table 8.3 and an example SEM

image of unprocessed PVP (K = 17) is shown in Figure 8.3. Of particular relevance to

using PVP in gas anti-solvent processes are the relatively high glass transition

temperatures of 126 and 164°C.

Table 8.3 Typical properties of PVP.*

Property PVP K = 17 PVP K = 30

Appearance White powder White powder

Tgmax 126°C 164°C

Mean Particle Size 50 µm 30 µm

Polydispersity 3.12 3.36

Density (Bulk) 0.1 – 0.25 g/cc 0.25 - 0.4 g/cc

* (Information from GAF Chemicals Corporation technical brochure on Plasdone C-15 and Plasdone C-

30)

Tgmax — glass transition temperature

The feasibility of precipitating PVP from DMF by ASES was investigated. The process

parameters of PVP concentration, anti-solvent density and temperature were

investigated to determine their effect on the PVP precipitates produced. The process

conditions investigated and results obtained are listed in Table 8.4.

CHAPTER 8

264

20 µm

Figure 8.3 SEM image of PVP.

CHAPTER 8

265

Table 8.4 Results of the precipitation of PVP by ASES.

T / °C P / MPa Conc. / mg.g-1 Morphology Size / µm

10 5.5 50 Agglomerated spheres < 1





25 14.0 50 Spheres < 1

25 14.0 100 Spheres < 1

All experiments conducted using DMF as the solvent, a co-current flow of solution (0.1 mL.min-1) and

CO2 (4 to 5 mL.min-1), and a nozzle diameter of 229 µm.

Solution and CO2 flow rates of 0.1 and 4 mL.min-1 respectively were used for all

experiments to provide the greatest difference in the solvent to CO2 ratio in the vessel.

Decreasing the ratio of solution to anti-solvent flow rate resulted in an increase in the

level of solvent in the vessel. The increased concentration of solvent solubilised the

precipitate and resulted in extraction and agglomeration of PVP.

8.3.1. The Effect of PVP Concentration

To determine the effect of PVP concentration on the size and morphology of particles

produced by the ASES process, solutions of PVP in DMF at concentrations of 50 and

100 mg.g-1 were precipitated by ASES at 25°C and 14.0 MPa. The precipitation of

solutions with concentrations above 100 mg.g-1 was not attempted, as these

concentrations are beyond the range that was to be used in the co-precipitation

experiments. Solution concentrations below 50 mg.g-1 were not attempted, as higher

concentration solutions are more favourable from a processing throughput standpoint.

Examples of PVP precipitate formed at these conditions are shown Figure 8.4. The

precipitates produced from both solution concentrations consisted of micro-spheres of

less than 1 µm diameter.

CHAPTER 8

266

(b)

(a)

3 µm

1 µm

Figure 8.4 SEM images of PVP particles produced from DMF by the ASES process

at 25°C and 14.0 MPa using solution flow rate of 0.1 mL.min-1 through a 229 µm

nozzle. (a) 50 mg.g-1, (b) 100 mg.g-1 solution of PVP in DMF.

CHAPTER 8

267

Increasing the concentration of PVP in DMF from 50 to 100 mg.g-1 had no effect on the

size or morphology of the particles produced at 25°C and 14.0 MPa. The precipitates

shown in Figure 8.4 are typical of the types of precipitates that were produced when

dilute solutions of polymer are precipitated below the glass transition temperature of the

polymer.6,8-12 The fact that the spherical morphology did not change, even at solution

concentrations of 100 mg.g-1, indicates that the concentrations used are below the plait

point concentration of PVP. Above the plait point concentration morphologies such as

fibers can be formed by the ASES technique13,14 (See Chapter 2 and 4).

8.3.2. The Effect of Anti-Solvent Density

The effect of anti-solvent density on the size and morphology of PVP particles produced

by the ASES process was examined. Solutions of PVP in DMF were precipitated at

25°C and 7.0 MPa. The precipitates produced from both solutions consisted of

agglomerated micro-spheres, an example of which is shown Figure 8.5 (See also

Appendix VII, Figure VII.1).

The reduction in anti-solvent pressure from 14.0 to 7.0 MPa at 25°C corresponds to a

decrease in anti-solvent density from 0.868 to 0.743 g.mL-1. The reduction in density

has resulted in agglomeration of PVP particles. Agglomeration can occur due to

intermolecular forces, the presence of solvent and coalescence of particles15 or

particle—particle collisions of plasticised polymer particles. Plasticisation can occur

when the polymer is above its glass transition temperature and at conditions where the

rate of mass transfer of CO2 and solvent is insufficient to dry particles before particle—

particle collisions occur.

The high glass transition temperature of PVP and the absence of agglomeration at

higher pressures of 14.0 MPa indicate that the reason for agglomeration of PVP

particles at lower pressures was not plasticisation. Agglomeration of particles is,

therefore most likely a result of the fusing of spheres during the settling and washing

steps. At lower CO2 pressures the ratio of CO2 to DMF is lower which implies that the

concentration of DMF in the system is greater. Higher concentrations of DMF enable

CHAPTER 8

268

3 µm

Figure 8.5 SEM images of particles produced from DMF solutions of PVP at a

concentration of 50 mg.g-1 by the ASES process at 25°C and 7.0 MPa using

solution flow rate of 0.1 mL.min-1 through a 229 µm nozzle.

CHAPTER 8

269

some of the particles to dissolve and fuse with other particles. A method that may

minimise the agglomeration of particles when processing PVP by ASES is to increase

the ratio of CO2 to solution flow rates. This was not done in these experiments due to

the flow rate constraints of the CO2 and solution pumps.

At 25°C and 6.6 MPa both liquid and vapour phase CO2 was present in the system. An

example of the particles obtained from spraying a PVP solution into CO2 at these

conditions is shown in Figure 8.6. The precipitate produced consisted of spheres that

had coalesced into larger particles.

The types of particles that formed from spraying solution of polymer with the anti-

solvent in the two-phase vapour—liquid region include microspheres,6,7,10,16 hollow

microspheres, or microballoons.10 The transition from microspheres to hollow

microballoons has been attributed to polymer concentration with the higher

concentration solutions resulting in microballoons. In the case of the lower

concentration polymer solutions, the diffusion of anti-solvent into the liquid droplet as it

falls through the vapour phase is insufficient to cause precipitation and nucleation only

occurs when the droplet enters the liquid phase. The droplet disperses in the liquid

phase anti-solvent resulting in microspheres. In the case of higher concentration

polymer solutions, the low level of anti-solvent diffusing into the droplet as it falls

through the vapour phase is sufficient to result in precipitation on the surface of the

droplet and cause a skin to form around the droplet. As the droplet enters the liquid

phase, the droplet does not break up, but retains its shape and undergoes mass transfer

of anti-solvent into and solvent out of the droplet. In this way microballoons are

formed.11

Spraying 100 mg.g-1 solutions of PVP in DMF into vapour phase CO2 resulted in the

formation of droplets with diameters similar to the nozzle diameter. These droplets fell

through the vapour phase into the liquid CO2 layer. Upon entering the liquid phase the

droplets ruptured and a cloud of solid PVP formed in the liquid phase. The rapid mass

transfer of CO2 into the liquid droplets was evident by the observation that the droplets

turned white as they fell through the vapour phase. Precipitation in the droplet was not

CHAPTER 8

270

2.0 kV x 5.0 k 6 µm

Figure 8.6 SEM images of PVP particles produced from DMF solutions of PVP at

a concentration of 100 mg.g-1 by the ASES process at 25°C and 6.6 MPa using


CHAPTER 8

271

sufficient to result in skin formation and the droplet disintegrated upon entering the CO2

liquid phase. Rapid nucleation of PVP results in the formation of microspheres.

Agglomeration of particles most likely occurs due to the build-up of DMF in the liquid

phase. The smaller volume occupied by the liquid phase results in a higher

concentration of DMF in the liquid phase. The increased level of DMF causes the PVP

particles to soften and coalescence of particles results.

Agglomeration of polystyrene has been reported when precipitating by spraying into

vapour phase CO2.10 The formation of a solvent rich liquid phase at the bottom of the

vessel caused agglomeration of the polystyrene particles.

Spraying a solution of PVP into vapour over liquid CO2 can be viewed as similar to the

GAS process with extremely rapid rates of expansion. It would therefore be expected

that at best, precipitation by GAS would yield highly agglomerated particles as was

seen for the vapour over liquid ASES process. This was confirmed by the precipitation

of PVP from DMF by GAS yielding films and, in some instances, solubilisation and

extraction of the polymer.


The effect of temperature on the characteristics of PVP precipitated by the ASES

process was investigated. PVP particles were produced from 50 and 100 mg.g-1 DMF

solutions at 10°C using an anti-solvent pressure of 5.5 MPa examples of which are

shown in Figure 8.7 (See also Appendix VII, Figure VII.2).

Significant agglomeration of PVP particles occurred at 10°C. The results at 10°C

support the hypothesis that agglomeration was not due to plasticisation, since it was

expected to be decreased by reducing the temperature.

8.4. Co-Precipitation of PVP and Cu-Indo by ASES.

The preliminary investigations into the precipitation of PVP from DMF using ASES

showed that it was possible to form microspheres of PVP when operating at 25°C and

CHAPTER 8

272

3 µm

Figure 8.7 SEM images of PVP particles produced by the ASES process from a 50

mg.g-1 solution of in DMF at 10°C and 5.5 MPa using a solution flow rate of 0.1

mL.min-1 through a 229 µm nozzle.

CHAPTER 8

273

14.0 MPa. The aims of the co-precipitation experiments were to show the applicability

of ASES to produce polymer—drug co-precipitates and to investigate the effect of

process variables on the co-precipitate characteristics.

The ASES co-precipitation conditions investigated are listed in Table 8.5. All

experiments were conducted using DMF as the solvent, a solution flow rate of 0.1

mL.min-1, a nozzle with an internal diameter of 0.229 µm and a CO2 flow rate between

4 and 5 mL.min-1 co-current with the solution flow.

Table 8.5 Results of the co-precipitation of PVP and Cu-Indo by ASES.

T / °C P / MPa Conc. / mg.g-1 PVP:Cu-Indo

25 6.6 50 60:40

25 10.39 100 50:50

25 10.39 100 70:30

25 10.39 100 90:10

25 14.0 50 60:40

25 14.0 100 50:50

25 14.0 100 70:30

25 14.0 100 90:10

40 19.0 50 60:40

The particles produced from DMF solutions containing both PVP and Cu-Indo showed

no evidence of agglomeration even at low anti-solvent density conditions, examples of

which are shown in Figure 8.8. Variation of solute concentration, PVP to Cu-Indo ratio,

temperature and anti-solvent density had a negligible effect on the particles produced.

All the precipitates produced were pale green in colour and consisted of spheres with

diameters less than 1 µm.

The co-precipitation results demonstrate the system dependent nature of dense gas anti-

solvent processes. Examples of particles produced from ASES when precipitating

individual DMF solutions containing Cu-Indo, PVP and a mixture of PVP and Cu-Indo

are compared in Figure 8.9. The results obtained from experiments conducted using

CHAPTER 8

274

(b)

(a)

1.5 µm

1.5 µm


solutions with a solute concentration of 100 mg.g-1 by the ASES process at 25°C

using a solution flow rate of 0.1 mL.min-1 through a 229 µm nozzle. (a) 10.39 MPa

and a 70:30 PVP:Cu-Indo ratio, (b) 14.0 MPa and a 50:50 PVP:Cu-Indo ratio.

CHAPTER 8

275

(c)

(a)

3 µm

10 µm

(b)

1 µm

Figure 8.9 SEM images comparing the particles produced from ASES using 100

mg.g-1 solutions of Cu-Indo, PVP and PVP—Cu-Indo in DMF at 25°C and 14.0

MPa. (a) Cu-Indo, (b) PVP, (c) 50:50 PVP:Cu-Indo ratio.

CHAPTER 8

276

individual solutes are not necessarily indicative of the results to be expected from a

combination of the solutes. The particles produced from DMF solutions of pure Cu-

Indo have diameters approximately five times that of pure PVP and PVP—Cu-Indo

solutions.

The co-precipitate particles (Figure 8.9c) show no evidence of free Cu-Indo. This

suggests that the precipitate produced is an intimate mix of the two solutes rather than

each solute precipitating as discrete particles. Further support for an intimate mix of the

two solutes was the homogenous green colour (Figure 8.10) of the precipitate and the

drug content results that will be discussed later.

The precipitation experiments conducted using PVP—Cu-Indo solution in various ratios

with a solute concentration of 100 mg.g-1 were found to produce large, irregular

particles in addition to the micro-spheres, unless mechanical shock was applied to the

vessel during spraying. Examples of these larger particles are shown in Figure 8.11.

Particles of this type were large enough in some cases to be visible with the naked eye.

Observations of the jet during spraying of PVP—Cu-Indo solutions at a concentration

of 100 mg.g-1 revealed that, in the absence of mechanical shock, skin formation

occurred around the jet as shown in Figure 8.12. Once skin formation occurred,

precipitate was seen to build up near the tip of the nozzle until at some critical size the

large particle fell to the bottom of the vessel.

The formation of fibres by ASES is commonly observed when spraying polymer

solutions of high concentration.13,14 At high polymer concentrations, jet break-up is

prevented due to increased solution viscosity. At higher solute concentrations, less anti-

solvent is required to diffuse into solution to initiate nucleation and precipitation occurs

earlier further preventing jet break-up. As a result, a skin or precipitate is able to form

around the jet and continuous polymer morphologies such as fibres are formed.

In the co-precipitation of PVP—Cu-Indo at solution concentrations of 100 mg.g-1 skin

formation was avoided by tapping the vessel during spraying. Alternatively, a vibrating

nozzle could be used. A vibrating nozzle aids in jet break-up and may prevent fibre

formation.

CHAPTER 8

277

Figure 8.10 Photograph of the PVP—Cu-Indo precipitate formed from the ASES

process.

CHAPTER 8

278

(b)

(a)

10 µm

5 µm



and 10.39 MPa using a solution flow rate of 0.1 mL.min-1 through a 229 µm nozzle.

PVP to Cu-Indo ratio (a) 70:30, (b) 50:50.

CHAPTER 8

279

SkinFormation

Figure 8.12 Photograph of the precipitation of PVP—Cu-Indo co-precipitate by

ASES. The solution concentration was 100 mg.g-1 in a 90:10 ratio of PVP:Cu-Indo.

The system was at 25°C, 10.39 MPa with a solution flow rate of 0.1 mL.min-1

through a 229 µm nozzle.

CHAPTER 8

280

The co-precipitation results demonstrate the ability of the ASES technique to form solid

blends. The ability of ASES to form blends can be directly attributed to the rapid nature

of the precipitation. The formation of an intimate blend relies on the precipitation of

both solutes occurring on similar time scales in close proximity to each other.14 If the

solutes were to precipitate on different time scales, two discrete precipitates would form

rather than an intimate mix.

8.5. Drug Content

The experimental drug loading of co-precipitates produced by the ASES process at

different PVP to Cu-Indo ratios and at two different pressures are listed in Table 8.6. All

precipitates were formed at a temperature of 25°C and a solution concentration of 100

mg.g-1 in DMF.

In all cases the concentration of Cu-Indo in the precipitate is slightly higher than

expected. The reason for the slightly higher Cu-Indo content may be that some of the

PVP is extracted during the precipitation. As mentioned previously, PVP has a high

solubility in DMF and a small percent of DMF in the vessel would be enough to

dissolve some of the polymer.

Table 8.6 Drug content of PVP—Cu-Indo co-precipitates.

Sample/

PVP:Cu-Indo

P / MPa Cu-Indo

Expected / %

Cu-Indo

Obtained / %

50:50 10.4 51.3 53.5

70:30 10.4 30.1 31.3

90:10 10.4 10.3 10.3

50:50 14.0 47.1 51.4

70:30 14.0 27.6 29.3

90:10 14.0 8.9 10.2

CHAPTER 8

281

8.6. Solubility of the PVP—Cu-Indo Co-Precipitates

One of the main objectives of the co-precipitation of Cu-Indo and PVP was to increase

the solubility of Cu-Indo in solvents that can be used in injectable formulations,

particularly for administration to horses. A useful solvent for injectable applications is

ethanol. The solubility of a mixture of PVP and Cu-Indo and pure Cu-Indo in ethanol is

compared to the co-precipitate solubility results in Table 8.7.

A significant enhancement in the solubility of Cu-Indo in ethanol was observed for the

PVP—Cu-Indo co-precipitates obtained from ASES. The greatest enhancement of 36

was observed for the PVP—Cu-Indo co-precipitate in a ratio of 90:10. As the ratio of

PVP to Cu-Indo was decreased, the solubility enhancement decreased from 5.3 to 3.3

for the 70:30 and 50:50 co-precipitates respectively. The PVP—Cu-Indo co-precipitates

once dissolved in ethanol, remained soluble for approximately four hours. After four

hours Cu-Indo precipitated. No significant solubility enhancement was observed for the

PVP—Cu-Indo physical mixtures.

Table 8.7 Solubility of Cu-Indo and PVP—Cu-Indo co-precipitate in ethanol at

25°C.

Sample Solubility /

mg.g-1

Solubility

Enhancement

Cu-Indo 1.9 —

Physical mix PVP:Cu-Indo — 50:50 2.5 1.3



Co-precipitate PVP:Cu-Indo — 50:50 6.4 3.3

Co-precipitate PVP:Cu-Indo — 70:30 10.3 5.3

Co-precipitate PVP:Cu-Indo — 90:10 40.0 36

The typical dose of Cu-Indo is 2 mg per 10 kg of bodyweight. Taking the mass of a

horse to be 450 kg, at least 47 g of an ethanol solution of Cu-Indo would need to be

administered to give the required dose. By forming a PVP—Cu-Indo co-precipitate, the

CHAPTER 8

282

increase in solubility of the drug reduces the amount of solution to 14, 8.8 and 2.3 g for

the 50:50, 70:30 and 90:10 co-precipitates respectively. The reduction in the amount of

ethanol required per dose makes it possible to administer Cu-Indo by the injectable

route.

Two possible hypotheses have been presented to explain the solubility enhancement

observed when PVP is co-precipitated with compounds of lower molecular weight. The

first hypothesis explains the solubility enhancement in terms of an interaction between

the hydrocarbon segments of the polymer and similar portions of the co-solute.17,18

This interaction termed the "hydrophobic bond" concept has been used to explain the

increase in solubility observed for PVP—co-solute co-precipitates.19 The second

hypothesis explains the solubility enhancement in terms of the formation of a complex

between PVP and low molecular weight compounds.20 The formation of a complex has

been used to explain the solubility enhancement observed for a PVP—sulfathiazole co-

precipitate.20 The complex was proposed to have formed by hydrogen bonding between

the amino groups of the sulfathiazole and the oxygen groups of the PVP. A

representation of the proposed complex is shown in Figure 8.13

To investigate the nature of the PVP—Cu-Indo co-precipitate DSC was conducted on

PVP, Cu-Indo, a physical mix of PVP and Cu-Indo in a 50:50 ratio and the 50:50

PVP—Cu-Indo co-precipitate. The DSC spectra obtained are compared in Figure 8.14.

The DSC spectrum of PVP does not show a melting point at 300°C. The expected

melting point of PVP (10,000) is approximately 300°C. The absence of a melting point

peak is most likely due to the amorphous nature of the polymer and the correspondingly

low energy required for a phase transition. The broad peak from ambient to 100°C can

be attributed to a loss in water that has been observed for PVP previously.21-24 The

DSC spectra of the PVP—Cu-Indo physical mix (Figure 8.14 c) is identical to a

combination of the pure PVP and Cu-Indo spectra. The DSC spectra of the PVP–Cu-

Indo co-precipitate (Figure 8.14 d) is significantly different. The peak at 205°C

corresponding to the melting point of Cu-Indo shifted to a lower temperature. The broad

peak between 100 and 150°C corresponding to the loss of solvent from Cu-Indo is not

evident in the co-precipitate.

CHAPTER 8

283

O

SO2

NH

N

O

S

OH HO

N

Figure 8.13 PVP—sulfathiazole complex.20

CHAPTER 8

284

0 50 100 150 200 250 300 350 400Temperature / °C

∆H

(a)

(b)

(c)

(d)

Figure 8.14 Comparison between the DSC spectra of (a) PVP, (b) Cu-Indo, (c)

50:50 physical mix of PVP and Cu-Indo, (d) 50:50 PVP—Cu-Indo co-precipitate.

CHAPTER 8

285

Room temperature X-band EPR was conducted on the 50:50 PVP-Cu-Indo co-

precipitate to further investigate the nature of the co-precipitate. The X-band EPR

spectra of pure Cu-Indo and the PVP—Cu-Indo co-precipitate are compared in Figure

8.15. The spectra of the co-precipitate shows a large peak at 3300 G which corresponds

to Cu(II) monomer.25 The peak at 4720 G which corresponds to the Cu(II) dimer

appears in both spectra.25

The results from the DSC and X-band EPR analyses indicate that the PVP—Cu-Indo

co-precipitate is not simply a mixture of the two compounds. The results support the

hypothesis that the increase in Cu-Indo solubility in ethanol is due to a complex forming

between PVP and Cu-Indo. The complex may have a similar structure to the PVP—

sulfathiazole complex described previously, with hydrogen bonding between the two

molecules. In the case of PVP—Cu-Indo, however, it is more likely that the bonding is

occurring between the Cu of the Cu-Indo and the oxygen on PVP. The loss of the

solvent peak in the DSC spectra of the co-precipitate (Figure 8.14 d) indicates that the

coordination site on the Cu that is normally occupied by a solvent molecule may be

occupied by PVP. The presence of a larger Cu(II) monomer signal in the X-band EPR

spectra suggests that PVP may also be interfering with the Cu—Cu bond of the Cu-

Indo. A representation of the proposed PVP—Cu-Indo complex is shown in Figure

8.16.

The fact that a larger percentage of Cu(II) monomer is present in the PVP—Cu-Indo co-

precipitate is of concern as the active form of Cu-Indo is thought to be the dimeric

complex.25 As mentioned above, the co-precipitate was soluble in ethanol for

approximately four hours after which time Cu-Indo precipitated. The precipitation of

Cu-Indo may be due to PVP being replaced by ethanol in the Cu-Indo complex. Once

the solvent replaces PVP, the original Cu-Indo is formed which is slightly soluble in

ethanol and the excess solute precipitates. If this is the case, then the presence of

monomer in the co-precipitate does not pose a problem, as Cu-Indo should resort to its

original form once it is administered in the body.

CHAPTER 8

286

0 2000 4000 6000 8000 10000

Magnetic Field / G

(a)

(b)

Figure 8.15 Comparison between the X-band EPR spectra of (a) Cu-Indo and (b)

50:50 PVP—Cu-Indo co-precipitate.

CHAPTER 8

287

Cu

Cu

O

O

O O

C

C

C

RR

R

R

CCH 3

CH 2

O

= R

N

O

C

O

O

O

O

O

Cu

O

O

C

C

R

R

O

O

O

O

(a)

(b)

Figure 8.16 Proposed PVP—Cu-Indo complex. (a) Proposed dimeric complex, (b)

proposed monomeric complex

CHAPTER 8

288

8.7. Dissolution of the PVP—Cu-Indo Co-Precipitates

The dissolution rates of unprocessed Cu-Indo, micronised Cu-Indo, a physical mix of

PVP and Cu-Indo and PVP—Cu-Indo co-precipitates produced by ASES are compared

in Figure 8.17 (See Appendix V for data). The dissolution rate coefficient (Kw) was

used to compare dissolution rates. The dissolution rate coefficients of the 50:50, 70:30

and 90:10 PVP—Cu-Indo co-precipitates were calculated to be 0.029, 0.036 and 0.012,

respectively. No significant difference in dissolution rate was observed for a 50:50

physical mix of PVP and Cu-Indo. The 50:50 and 70:30 PVP—Cu-Indo co-precipitates

gave the greatest enhancement in dissolution rate with an increase of 18 and 14.5 times

respectively. Although this is a significant increase in dissolution rate compared to

unprocessed Cu-Indo, there is only a two fold increase in dissolution rate when

compared to the micronised Cu-Indo. The 90:10 PVP—Cu-Indo co-precipitate had a

similar dissolution rate to micronised Cu-Indo.

No significant increase in solubility was observed when the PVP—Cu-Indo co-

precipitates were dissolved in water. It is therefore unlikely that the increase in

dissolution rate was solely a result of the increased solubility of Cu-Indo in water. The

increase in dissolution rate may also be a result of the decrease in particle size of Cu-

Indo upon co-precipitation with PVP. The co-precipitate particles were generally less

than 1 µm in diameter, whereas the micronised Cu-Indo particles were between 2 and

10 µm. It is feasible that the reduction in particle size may be the reason for the increase

in dissolution rate. The fact that the 90:10 co-precipitate did not show as great an

enhancement in dissolution rate, indicates that a reduction in particle size is not the sole

reason for an increased rate of dissolution.

CHAPTER 8

289

Time / min

% D

issol

utio

n

0

20

40

60

80

100

0 20 40 60 80 100 120

90:10 PVP:Cu-Indo50:50 PVP:Cu-Indo70:30 PVP:Cu-IndoMicronised Cu-Indo50:50 Physical MixUnprocessed

Figure 8.17 Dissolution of Cu-Indo.

CHAPTER 8

290

8.8. Conclusions

Polyvinylpyrrolidone was successfully processed by the ASES technique using CO2 as

the anti-solvent. The ratio of CO2 to solvent was found to be an important parameter in

the processing of PVP. If the ratio of CO2 to solvent was too low, PVP was extracted

rather than precipitated in the vessel. At optimum conditions microspheres with

diameters less than 1 µm were produced from the ASES process. Pressure and solution

concentration were found to be important parameters controlling particle characteristics.

The degree of agglomeration of the PVP precipitate was found to increase at lower

pressures. Agglomeration was attributed to the washing step where the presence of

DMF resulted in the coalescence of particles.

Polyvinylpyrrolidone and Cu-Indo were successfully co-precipitated using the ASES

process. The co-precipitates could be produced in a number of PVP to Cu-Indo ratios

and drug content was as expected. In all experiments conducted, the co-precipitates

produced had a spherical morphology with diameters less than 1 µm. The benefit of co-

precipitating Cu-Indo with PVP was demonstrated by a 36 fold increase in the solubility

of the drug in ethanol.

The co-precipitate was found to give different DSC and X-band EPR spectra from a

simple addition of the pure PVP and Cu-Indo spectra. These results indicate that the co-

precipitate is not simply an intimate mix of PVP and Cu-Indo. It is likely that a PVP—

Cu-Indo complex is forming. The exact nature of this complex needs to be confirmed by

further investigation.

The co-precipitate showed a two-fold increase in dissolution rate in water compared to

micronised Cu-Indo. Co-precipitating PVP and Cu-Indo did not enhance the solubility

of the drug in water. The increase in dissolution rate was, therefore, attributed to the

reduction in particle size rather than an enhanced solubility of the drug. The particles of

co-precipitate produced were approximately an order of magnitude smaller than the

ASES micronised particles of Cu-Indo. At PVP:Cu-Indo ratios of 90:10, the rate of

dissolution decreased to be equivalent to the pure micronised drug.

CHAPTER 8

291

The PVP—Cu-Indo co-precipitate produced by ASES provides opportunities to develop

new formulation of Cu-Indo that were not previously possible. These formulations will

increase the marketability and competitiveness of the drug.

CHAPTER 8

292

8.9. References

1. Deluka, P.; Mehta, R.; Hausberger, A.; Thanoo, B.; Biodegradable Polyesters for

Drug and Polypeptide Delivery, Polymeric Delivery Systems: Properties and

Applications; American Chemical Society: Washington, DC, 1993.

2. Wessel, W.; Schoog, M.; Winkler, E.; "Polyvinylpyrrolidone (PVP), its Diagnostic,

Therapeutic and Technical Application and Consequences Thereof,"

Arzneimittelforschung 1971, 21, 1468.

3. Wu, H.; Lin, Z.; Li, Z.; Zhong, D.; Liu, X.; "Preparations of Prostaglandins. III.

Studies on the Sustained-Release of dl-15-Methylprostaglandin F2α from

Injections," Acta. Acad. Med. Primae Shanghai 1982, 9, 450.

4. Mawad, J.; Dtsch. Med. J. 1972, 26, 453.

5. Cannon, C. S.; Falk, R. F.; Randolph, T. W.; "Role of Crystallinity in Retention of





1995, 12, 1211.

7. Young, T. J.; Johnston, K. P.; Mishima, K.; Tanaka, H.; "Encapsulation of Lysozyme

in a Biodegradable Polymer by Precipitation with a Vapor-over-Liquid Anti-

Solvent," J. Pharm. Sci. 1999, 88, 640.






Polym. Sci. 1993, 50, 1929.






CHAPTER 8

293












3898.

17. Molyneux, P.; Frank, H. P.; "The Interaction of Polyvinylpyrrolidone with Aromatic

Compounds in Aqueous Solution. Part I. Thermodynamics of the Binding Equilibria

and Interaction Forces," J. Am. Chem. Soc. 1961, 83, 3169.

18. Molyneux, P.; Frank, H. P.; "The Interaction of Polyvinylpyrrolidone with Aromatic

Compounds in Aqueous Solution. Part II. The Effect of the Interaction on the

Molecular Size of the Polymer," J. Am. Chem. Soc. 1961, 83, 3175.

19. Svoboda, G. H.; Sweeney, M. J.; Walkling, W. D.; "Antitumor Activity of an

Acronycine-Polyvinylpyrrolidone Co-Precipitate," J. Pharm. Sci. 1971, 60, 333.

20. Badawi, A. A.; El-Sayed, A. A.; "Dissolution Studies of Povidone-Sulfathiazole

Coacervated Systems," J. Pharm. Sci. 1980, 69, 492.

21. Suzuki, H.; "Influence of Water-Soluble Polymers on the Dissolution of Nifedipine

Solid Dispersions with Combined Carriers," Chem. Pharm. Bull. 1998, 46, 482.

22. Rodriguez-Espinosa, C.; Martin, C.; Goni, M. M.; Velaz, I.; Sanchez, M.;

"Dissolution Kinetics for Co-Precipitates of Diflunisal with PVP K30," European

Journal of Drug Metabolism and Pharmokinetics 1998, 23, 109.

23. Martinez, P.; Cantera, R. G.; Martin, C.; Dosi-Vieitez, C.; Martinez-Oharritz, C.;

"Preparation and Dissolution Rate of Gliquidone-PVP K30 Solid Dispersions,"

European Journal of Drug Metabolism and Pharmokinetics 1998, 23, 113.

24. Yagi, N.; Kenmotsu, H.; Sekikawa, H.; Takada, M.; "Dissolution Behaviour of

Probucol From Solid Dispersion Systems of Probucol-Polyvinylpyrrolidone," Chem.

Pharm. Bull. 1996, 44, 241.

CHAPTER 8

294



APPENDICE APPENDIX I : 295 APPENDIX II : 329 APPENDIX III : 333 APPENDIX IV : 337 APPENDIX V : 339 APPENDIX VI : 341 APPENDIX VII : 351 APPENDIX VIII : 354

295

APPENDIX I I.LITERATURE REVIEW OF THE USE OF DENSE GASES AS ANTI-SOLVENTS

I.1. Review of GAS Applications

Table I.1 Review of GAS applications.

MICRONISATION / RECRYSTALLISATION

Chemical Solvent Anti-Solvent T

(°C)

P

(MPa)

Particle

Size (µm)

Morphology

Reference

HYAFF-11 Hyaluronic acid

ethyl ester Dimethylsulfoxide CO2 35 10.0 0.4

aggregates

microspheres 1

Phenanthrene Toluene CO2 25 4.5 180 to 540 2 1,3,5,7-tetranitro-1,3,5,7-

tetraazacyclooctane (β-HMX) Acetone CO2 33

10.0 to

12.0 2 to 5 crystalline 3

β-Carotene Ethyl acetate CO2 43 to

82 2.5 to 6.0 2 to 10 Platelets 4

Nitroguanidine (NQ) N-Methylpyrrolidone,

Dimethylformamide ClF2CH, CO2

20 to

22

0.55 to

1.24 2 to 100

snowballs,

starbursts 5

Cobalt chloride

Saligenin Acetone

CO2

25 to

45 < 50

diamond, irregular

monodisperse

particles

6

APPEN

DIX

I

296




(°C)

P

(MPa)

Particle

Size (µm)

Morphology

Reference

Cyclotrimethylenetrinitramine

(RDX)

Acetone,

Cyclohexanone CO2

24 to

50 3.7 to 15.9 5 to 150 7

Prednisolone, Dexamethasone,

Flunisolide, Triamcinolone-

Acetonide

Acetone, Ethanol CO2 8

Dexamethasone Ethanol, Acetone CO2 1 to 2 large flat platelets 9

Bilirubin

Dimethylsulphoxide

CO2

35 to

60

10.0 to

30.0 < 0.5x1

powdered products

10

Sulfathiazole Ethanol CO2 25 5.8 10 to 6000 pillar like crystals 11 Cyclotrimethylenetrinitramine

(RDX)

Nitroguanidine (NQ)

Cobalt chloride

Acetone

Cyclohexanone

N-methylpyrrolidone

Dimethylformamide

CO2

Chloro-

difluoromethane

25 6.2

0.83 12

APPEN

DIX

I

297




(°C)

P

(MPa)

Particle

Size (µm)

Morphology

Reference

α-chymotrypsin

Imipramine

Insulin

Ribonuclease

Cytochrome C

Pentamidine

iso-Octane

Methylene chloride

Pyridine

Tetrahydrofuran

Methanol

Ethanol

CO2 28 to

36 7.6 to 8.55 0.1 to 50

spheroidal

fibre like

collapsed spheres

13

Polystyrene Toluene HFC-134a -20 to

60 0.5

microspheres

film 14

Cyclotetramethylene

Tetranitramine

Acetone

γ-Butyrolactone CO2 40 8.0 90 to 200 crystals 15,16

p-hydroxybenzoic acid

Methanol

Acetone

Ethyl acetate

CO2 35 1 to 2

clusters

amorphous

dendritic

spheres

17

APPEN

DIX

I

298




(°C)

P

(MPa)

Particle

Size (µm)

Morphology

Reference

Insulin

Lysozyme

Albumin

Dimethylsulfoxide

Dimethylformamide

Ethanol

Acetic acid

Ethyl acetate

Methanol

Ethanol + 2 to 15%

Water

CO2

NH4 0.01 to 1.8 spherical 18

Insulin

Lysozyme

Myoglobin

Methanol

Ethanol

Ethy lacetate

Dimethylsulfoxide

CO2 18 to

45 3.0 to 6.0 0.03 to 8

spherical

spherical-

agglomeration

19

Acetaminophen Ethanol CO2 20 to

25 7.1 20

Insulin Dimethylsulphoxide CO2 35 8.6 4 spheres 21

APPEN

DIX

I

299




(°C)

P

(MPa)

Particle

Size (µm)

Morphology

Reference

Substituted para-linked aromatic

polyamides

Dimethylsulphoxide,

Dimethylformamide CO2

23 to

40 10.3 10

spherulitic network

of lamellar crystals,

spheroidal star-

shaped

22

BaCl2

NH4Cl Dimethylsulfoxide CO2

25 to

35 9.5 1.8 to 400 crystals 23,24

APPEN

DIX

I

300


SEPARATION/PURIFICATION


(oC)

P

(MPa)

Purity

(%)

Reference

Naphthalene

Phenanthrene Toluene CO2 25 to 37 5.2 to 7.8 25

Lecithin from soya oil

Coriander essential from fixed oil Hexane CO2 25 to 40 4.0 to 7.5 100% to 0% 26

trans-β-Carotene from isomer and

oxide Acetaminophen

Cyclohexanone

Toluene

Butanol

CO2

25 to 35 7.0 to 9.7 > 99%

27

Anthracene from Anthraquinone Cyclo-hexanone CO2 18 to 40 3.5 to 12.4 28 Anthracene

Anthraquinone Cyclohexanone CO2 18 to 40 1.8 to 12.3 29

o-Hydroxybenzoic acid

p-Hydroxybenzoic acid Methanol CO2 25 7.0 99% 30

Bilirubin Dimethylsulfoxide CO2 35 to 60 10.0 to 30.0 > 90% 10

Anthracene from crude anthracene Acetone CO2 27 5.3 to 10.3 90% 31 Citric acid from byproduct organic

acids (oxalic and malic acids) Acetone CO2 30 2.5 to 2.7 > 99% 32

Fermented citric acid Acetone CO2 30 3.5 33

APPEN

DIX

I

301


SEPARATION/PURIFICATION


(oC)

P

(MPa)

Purity

(%)

Reference

Alkaline phosphatase from Insulin

Lysozyme from ribonuclease

Trypsin from catalase

DMSO CO2 34 to 44 0 to 8.0 100 34

Polyamides from LiCl N,N-Dimethylacetamide CO2 35 5.0 100 35

BaCl2 from NH4Cl DMSO CO2 25 4.35 100 23,24

APPEN

DIX

I

302


THEORETICAL


(oC)

P

(MPa)

Reference

Ethylcellulose

HP55

HYAFF 11

Salbuthanole

Pentamidine

Acetone

Ethylacetate

Dimthylsulfoxide

CO2 36

β-Carotene

Acetaminophen

Toluene

n-Butanol CO2 25 9.1 37

Naphthalene Toluene

CO2

Ethane

Ethene

Propane

38

Naphthalene

Phenanthrene

Anthracene

Benzoic Acis

Toluene

n-Octane

Ethanol

Acetone

CO2

Ethane

Ethene

Propane

39

Naphthalene

Phenanthrene

Mixture

Toluene CO2 25 6.4 40

APPEN

DIX

I

303


THEORETICAL


(oC)

P

(MPa)

Reference

NaCH3COO

LiCl

Acitominophen

Fructose

NH4Cl

NaSCN

Adipic acid

Naphthalene

Tartaric acid

Urea

NaClO3

NH4ClO4

Sucrose

NaCl

Monosodium glutamate

MgSO4•7H2O

Pentaerythritol

NH4H2PO4

K2SO4

Citric acid monohydrate

Ethanol

Acetone

Toluene

80 to 44% Methanol

CO2 14,41

APPEN

DIX

I

304


THEORETICAL


(oC)

P

(MPa)

Reference

Acetominophen

Paracetamol

Tylenol

Cholesterol

Ethanol

Diethyl ether CO2 19 to 44 7.0 to 15.0 42

Phenanthrene

Naphthalene

β-Carotene

Toluene CO2 25 7.0 43,44

Tetradecanoic acid

Stearic acid

Behenic acid

Ethyl acetate CO2 35 to 45 2.5 to 8.0 45

Naphthalene Toluene CO2 10 to 70 2.0 to 12.0 46

APPEN

DIX

I

305

I.2. Review of ASES Applications

Table I.2 Review of ASES applications.


Solute Solvent Anti-

Solvent

P

(MPa)

T

(°C)

Nozzle

(µm)

Particle Size

(µm) Morphology Reference

Hyaluronic acid benzylic ester

(HYAFF-11) Dimethylsulfoxide CO2 8.5 42 40 20 Microspheres 1

DL- Poly(lactic acid)

Poly(β-hydroxy butyric acid)

L- Poly(lactic acid)

DL-lactide-co-glycolide 50:50

Methylene chloride CO2 9.0/20.0 40 400 Microspheres 47

Poly(methyl methacrylate)

Ethyl cellulose

DL- Poly(lactic acid)

L-Poly(lactic Acid)

Poly(E-caprolactone)

DL-lactide-co-glycolide 50:50

Methylene chloride CO2 2.0 to 27.6 -10 to 40 100 Microspheres and

fibres 48

Lysozyme

Albumin

Dnase

Insulin

Water CO2 +

Ethanol 8.0 to 9.0 20 to 45

Coaxial nozzle

50 0.05 to 0.5 Microspheres 49

APPEN

DIX

I

306




Solvent

P

(MPa)

T

(°C)

Nozzle

(µm)

Particle Size


Buckminsterfullerene Toluene CO2 7.58 to

9.65 35 to 50 75 0.029 to 0.083

Porous spherical

balloons

Rods

50

Naproxen

L-Poly(Lactic Acid) Acetone CO2 6.8 25 180 3, 50

Fibres

Irregular particles 51

Catalase, Insulin Dimethylsulfoxide CO2 9.0 35 20 1 Crystals and

microspheres 52,53

Polystyrene Toluene CO2 4.65 to

8.03 0 to 30 100 0.1 to 20 Fibres and fibrils 54

Polystyrene Toluene CO2 3.96 to

22.47 0 to 40 100 0.1 to 20

Microspheres and

fibres 55

Polystyrene Toluene CO2 0.1 22 151 1400 Hollow

microspheres 56

Red Lake C

C. I. Pigment Blue

C. I. Pigment Yellow

Acetone CO2 6.0 to 25.0 40 to 150 5 to 500 Microspheres 57,58

Acetominophen

Lysozyme Water + Ethanol CO2

10.0 to

15.0 35 to 55 Coaxial nozzle

Acicular

Microspheres 59

APPEN

DIX

I

307




Solvent

P

(MPa)

T

(°C)

Nozzle

(µm)

Particle Size


Lactose

Maltose

Trehalose

Sucrose

Salmeterol xinafoate

R-TEM beta-lactamase

Water + Methanol

Water + Methanol

Water + Ethanol

Water + Ethanol

Actone + n-Hexane

Water + Ethanol

CO2 15.0 to

27.0 50 to 70

triple nozzle

assembly 60,61

Epoxy powder coating Acetone

Methyl ethyl ketone CO2 4.9 to 10.0 25 150 to 762

rods

microspheres 62

Bronze red Ethanol

Acetone CO2 5.5 to 12.0 35 to 75 50 1 to 14

Rhombic

Spheres

Needles

63

Sodium cromoglycate Methanol CO2 10.0 to

20.0 40 to 73 100 1 microspheres 64

Prednisolone acetate Acetone CO2 10.0 to

20.0 40 50 to 100 0.45 to 6.33 65

Polyacrylonitrile

Polystyrene

Dimethylformamide

Toluene CO2 40 50 to 100

0.1

100 to 3 cm

Microfibrils

Fibres 66,67

APPEN

DIX

I

308




Solvent

P

(MPa)

T

(°C)

Nozzle

(µm)

Particle Size


Soy lecithin Ethanol CO2 8.0 to 11.0 35 150 1 to 40

Agglomerated

spheres

Gel-like deposit

68,69

Streptomycin Methylene chloride CO2 8.8 to 8.9 35 to 36.8 sonicated

nozzle 0.1 to 0.4 Spheroidal 13

Polycarbonate

Poly(styrene-co-acrylonitrile) Tetrahydrofuran CO2 0, 23, 35 100 1

Aglgomerated

microspheres 70

L-Poly(lactic acid)

Polystyrene

Polyacrylonitrile

Methylene chloride

Toluene

Dimethylformamide

CO2 0 to 26 50 Microspheres 71

Polystyrene

Poly(methyl methacrylate)

Methyl ethyl ketone

Toluene CO2 12.4 23 50 1 to 50 Flocculates 72,73

Recombinant human immunoglobin G Water + Ethanol CO2 17.5 45 200 74

α-lactose Water + Ethanol

Water + Methanol CO2

15.0 to

30.0 50 to 90 75

Lactose monohydrate Water CO2 180 Irregular-shaped

particles 76

L-Poly(lactic acid) Methylene chloride CO2 6.2 to 9.65 31 to 40 75 1.5 Microspheres 77

APPEN

DIX

I

309




Solvent

P

(MPa)

T

(°C)

Nozzle

(µm)

Particle Size


Poly(L-lactic acid) Methylene Chloride CO2 8.5 to 17.0 28 to 60 Coaxial nozzle 13.4 Agglomerated

particles 78

Amoxicillin N-methylpyrolidone

Dimethylsulfoxide CO2 15.0 40 60 0.25 to 1.2 Spherical 79

Zinc Acetate Dimethylsulfoxide

N-methylpyrolidone CO2 8.0 to 15.0 40 22 to 60 0.05 to 0.137

Balloons

Nano-particles 80

Yttrium Acetate

Samarium Acetate

Neodymium Acetate

Dimethylsulfoxide CO2 6.0 to 16.0 35 to 70 22 0.1 to 20

Spherical

Balloon type

aggregates

81

Griseofulvin

Ampicillin

Amoxicillin

Tetracycline

N-methylpyrolidone

Dimethylsulfoxide

Ethanol

Methylene chloride

CO2 15.0 40 60 0.2 to 2000

No precipitation

Film

Needles

Irregular flat

crystals

Flat crystals

Spherical

aggregates

82,83

APPEN

DIX

I

310




Solvent

P

(MPa)

T

(°C)

Nozzle

(µm)

Particle Size


Samarium acetate

Yttrium acetate

Neodymium acetate

Dimethylsulfoxide CO2 6.0 to 16.0 35 to 70 22 0.266 Microspheres 84

Dextran

Inulin

Poly-L-lactic acid

Poly(hydroxypropylmetacrylamide)

Poly-hyaluronic acid (HYAFF 11)

Dimethylsulfoxide

Methylene chloride CO2 8.0 to 15.0 40 60 0.1 to 100

microspheres

fibres 85,86

Polyvinyl alcohol

Poly caprolactone

Dimethylsulfoxide

Acetone

N-methylpyrolidone

Methylene chloride

CO2 8.0 to 15.0 40 60 networked particles

no precipitation 86

Model pharmaceutical compounds Dimethylformamide CO2 10.0 to

25.0 12 to 60

aggregated needles

platelets

hollow spheres

87,88

L-Poly(lactic acid) Methylene chloride CO2 9.0 40 1 to 2 microspheres 89

APPEN

DIX

I

311




Solvent

P

(MPa)

T

(°C)

Nozzle

(µm)

Particle Size


Methylprednisilone

THF,

Dimethylacetamide,

MeOH/Methylene

chloride,

1,3-Dioxilane

Ethane 10.0 to

15.0 -6 to 56 500 2.8 Needles 90

Trypsin

Lysosyme

Therapeutic Peptide

Antibody Fv and Fab

Plasmid DNA pSVβ

Water and Ethanol CO2 20.0 55 3 channel

coaxial nozzle < 2

irregular particles

and microspheres 91,92

Paracetamol Ethanol CO2 8.0 to 25.0 35 to 85 200 coaxial

nozzle 3 to 300

spheres

faceted crystals

needles

93

8 Steroids/surface-acting agent Methylene chloride CO2 8.5 40 300 < 5 microspheres 94

APPEN

DIX

I

312




Solvent

P

(MPa)

T

(°C)

Nozzle

(µm)

Particle Size


Hydrocortisone

Poly(DL-lactide-glycolide)

RG503H

HYAFF-7

Ibuprofen

Camptothecin

Dimethylsulfoxide

Ethyl acetate CO2 10.3 35 to 40

High energy

nozzle 0.5 to 500

Spherical

Whiskers

Hollow

microspheres

Flakes

Resin

95,96

p-hydroxybenzoic acid

L-Poly(lactic acid)

MeOH

Methylene chloride CO2 7.6 to 14.1 25

180

1020

Rhombic crystals

Microspheres 97

Polystyrene Toluene HFC-134a 0.579 to

0.897 -20 to 60 0.5 to 6

Microparticles

Fibrils 98

p-Hydroxybenzoic acid

MeOH

Acetone

Ethyl acetate

CO2 7.6 7 to 36 50 Rhombic particles 17

L-Poly(Lactic Acid) Methylene chloride CO2 7.0 to 12.5 40 6 to 50 Microspheres 99 Catalase

Insulin Ethanol CO2 8.0 35 20

Crystals and

microspheres 100

Trypsin,

Lysozyme

Insulin

Dimethylsulfoxide CO2 7.34 to

14.2

26.6 to

46.5 30 to 50 Microspheres 101

APPEN

DIX

I

313




Solvent

P

(MPa)

T

(°C)

Nozzle

(µm)

Particle Size


Insulin

Trypsin

Lysozyme

Dimethylsulfoxide CO2 9.1 to 17.3 28 to 46 1 to 5 102

Hydroquinone Acetone CO2 5.5, 6.1 22 67 needles and

prismatic particles 103

Acetominophen

Ascorbic acid Ethanol CO2 6.2 to 11.5 25 to 45 67 1 to 200

crystals

small aggregates 104

Insulin Dimethylsulfoxide

Dimethylformamide CO2 8.62 25 30 2.5 microspheres 21

Polyamide Dimethylsulfoxide

Dimethylformamide CO2 10.34 23 to 40 30 1 fibrils 22

Insulin Dimethylsulfoxide CO2 8.62 25 to 35 30 30 105

Salmeterol xinafoate Acetone

Ethanol CO2 35 coaxial nozzle

irregular needles

and particles 106

APPEN

DIX

I

314


ENCAPSULATION / CO-PRECIPITATION


solvent

P

(MPa)

T

(°C)

Nozzle

(µm)

Particle Size

(µm)

Morphology Reference

Indomethacin with L-Poly(lactic acid)

Thymopentin with L-Poly(lactic acid)

Piroxicam with L-Poly(lactic acid)

Hyoscine butylbromide with L-

Poly(lactic acid)

Methylene chloride CO2 9.0 to 20.0 400 < 8.4 Microsphere and

fibre network 107

Hyoscine butylbromide with L-

Poly(lactic acid)

Methanol

Methylene chloride CO2 9.0 to 20.0 33 to 60 1.7 Microspheres 108

Naproxen with L-Poly(lactic acid) Acetone CO2 6.8 25 180 < 20 51 Insulin with L-Poly(lactic acid)

Lysozima with L-Poly(lactic acid)

Chimotrypsin with L-Poly(lactic acid)

Trifluroacetic acid

Dimethylsulfoxide

Methylene chloride

CO2 10.0 to

13.0 19 to 37 50 1 to 5 Microspheres 109

Estriol with DL-lactide-co-glycolide

50:50

Bovine serum albumin with DL-lactide-

co-glycolide 50:50

2,2,2-

Trifluoroethanol

Methylene chloride

CO2 1.0 34 150 to 5 Microspheres 110

APPEN

DIX

I

315


ENCAPSULATION / CO-PRECIPITATION


solvent

P

(MPa)

T

(°C)

Nozzle

(µm)

Particle Size

(µm)

Morphology Reference

Rifampin with L-Poly(lactic acid)

Naltrexone with L-Poly(lactic acid)

Gentamycin sulfate with L-Poly(lactic

acid)

Methylene chloride CO2 8.5 to 9.0 35 to 38 1000 0.2 to 1 Microspheres 111-113

Hyaluronic acid benzylic ester (HYAFF-

11) with two steroids or with a protein Dimethylsulfoxide CO2 35 to 50 < 1 66

Glass and sugar beads with

Hydrocortisone or RG503H

Dimethylsulfoxide

Ethyl acetate CO2 10.3 35 100

Coating of the beads

with microspheres of

the polymer and a

thin film of the drug.

95

p-Hydroxybenzoic acid with L-

Poly(lactic acid) or poly(lactide-co-

glycolide)

Methanol

Acetone

Methylene chloride

CO2 7.6 25 180 and 1020 microsphere and fibre

network 97

Paracetamol with Ascorbic acid

Chloramphenicol with Urea Ethanol CO2 4.5 to 12.0 40 100 to 300

10

2 x 30

30 x 100

Needles in deformed

prismatic crystals

Tubes

114

Lysozyme with L-Poly(lactic acid) or

Poly(lactic-co-glycolic acid) Methylene chloride CO2

vapour

over liquid 23 to -40 100 0.5 to 500 115

APPEN

DIX

I

316


FRACTIONATION


Solvent

P

(MPa)

T

(°C)

Purity

(%)

Reference

Lecithin from soya oil Hexane CO2 4.5 to 7.5 25 to 40 100 116 Trans-β-Carotene and total β-carotene from

raw β-carotene Toluene CO2 7.5 31 96.8 27

Nylon from Carpet Material Formic Acid CO2 4.0 to 15.0 40 117 (R)-2,2'-Binaphthyl-1,1'-diamine from (S)-

2,2'-Binaphthyl-1,1'-diamine Methanol CO2 8.5 to 30.0 35 to 75 93 118

Lecithin from de-oiled egg yolk Hexane

Ethanol CO2 4.5 to 5.0 15 95 119

Paracetamol and ascorbic acid Ethanol CO2 7.0 40 100 114

APPENDIX I

317

I.3. References



1997, 53, 232.



Process," AIChE J. 1996, 42, 431.

3. Cai, J.-G.; Liao, X.-C.; Zhou, Z.-Y.; "Mocroparticle Formation and Crystallization



23.

4. Cocero, M. J.; Ferrero, S.; Vicente, S.; "GAS Crystallization of β-Carotene from

Ethyl Acetate Solutions Using CO2 as Anti-Solvent," Proceedings of the 5th








87, 96.









APPENDIX I

318



Chem. Eng. 1996, 4, 257.



378.

12. Krukonis, V. J.; Gallagher, P. M.; Coffey, M. P.; "Gas Anti-Solvent

Recrystallization Process," U.S. 5,360,478, 1994.




14. Tan, C.-S.; Chang, W.-W.; "Precipitation of Polystyrene from Toluene with HFC-

134a by the GAS Process," Ind. Eng. Chem. Res. 1998, 37, 1821.

15. Teipel, U.; Foerter-Barth, U.; Krause, K. H.; "Formation of Particles with

Compressed Gases as Anti-Solvent," World Congress on Particle Technology 3,

Brighton, 1998, 189, 1.

16. Förter-Barth, U.; Teipel, U.; Krause, H.; "Formation of Particles by Applying the

Gas-Anti-Solvent (GAS) - Process," Proceedings of the 6th Meeting on Supercritical

Fluids, Nottingham, United Kingdom, 1999, 175.














APPENDIX I

319














26. Catchpole, O. J.; Hochmann, S.; Anderson, S. R. J.; "Gas Anti-Solvent Fractionation

of Natural Products," Process Technol. Proc. 1996, 12, 309.



Prog. 1991, 7, 275.



26, 517.






31. Liou, Y.; Chang, C. J.; "Separation of Anthracene from Crude Anthracene Using

Gas Anti-Solvent Recrystallization," Sep. Sci. Technol. 1992, 27, 1277.



Chem. 1994, 42, 1993.

APPENDIX I

320




34. Winters, M. A.; Frankel, D. Z.; Debenedetti, P. G.; Carey, J.; Devaney, M.;

Przybycien, T. M.; "Protein Purification With Vapor-Phase Carbon Dioxide,"


35. Yeo, S.-D.; Debenedetti, P. G.; Radosz, M.; Giesa, R.; Schmidt, H.-W.;

"Supercritical Anti-Solvent Process for a Series of Substituted Para-Linked Aromatic

Polyamides," Macromolecules 1995, 28, 1316.

36. Bertucco, A.; Pallado, P.; "Understanding Gas Anti-Solvent Processes of

Biocompatible Polymers and Drugs with Supercritical CO2," Proccedings of the





38. de la Fuenta Badilla, J. C.; Peters, C. J.; de swaan Arons, J.; "Selection of the

Appropriate Combination Solvent, Anti-Solvent and Process Conditions for the Gas-

Anti-Solvent Process," Proceedings of the 5th Meeting on Supercritical Fluids, Nice,

France, 1998, Tome 1, 237.

39. de la Fuenta Badilla, J. C.; Peters, C. J.; de Swaan Arons, J.; "Volume Expansion in

Relation to the Gas-Anti-Solvent Process," J. Supercrit. Fluids 2000, 17, 13.



41. Tai, C. Y.; Cheng, C.-S.; "Supersaturation and Crystal Growth in Gas Anti-Solvent

Crystallization," J. Cryst. Growth 1998, 183, 622.

42. Wubbolts, F. E.; Bruinsma, O. S. L.; van Rosmalen, G. M.; "Measurement and

Modelling of the Solubility of Acetaminophen and of Cholestrol in a Mixture of

Solvent and Carbon Dioxide at a Constant Pressure," Proceedings of the 5th






APPENDIX I

321




45. Liu, Z.; Li, D.; Yang, G.; Han, B.; "Solubility of Organic Acids in Ethyl Acetate

Expanded with CO2," Fluid Phase Equlibria 2000, 167, 123.

46. Peters, C. J.; Kordikowski, A.; Wilmes, B.; "Phase Behaviour of Selected Systems

of Interest for the GAS Process," Proceedings of the 5th International Symposium on






1995, 12, 1211.

49. Bustami, R. T.; Chan, H.; Dehghani, F.; Foster, N. R.; "Generation of Protein

Micro-Particles Using High Pressure Modified Carbon Dioxide," Proceedings of the

5th International Symposium on Supercritical Fluids, Atlanta, Georgia, 2000.

50. Chattopadhyay, P.; Gupta, R. B.; "Supercritical CO2 Based Production of Fullerene

Nanoparticles," Proceedings of the 5th International Symposium on Supercritical


51. Chou, Y.-H.; Tomasko, D. L.; "GAS Crystallization of Polymer-Pharmaceutical

Composite Particles," The 4th International Symposium on Supercritical Fluids,

Sendai, Japan, 1997, A, 55.



Release 1993, 24, 27.





Polym. Sci. 1993, 50, 1929.



APPENDIX I

322


Miroballoons by Precipitation with a Vapour-Liquid Compressed Fluid Anti-


57. Gao, Y.; Mulenda, T. K.; Shi, Y.-F.; Yuan, W.-K.; "Fine Particles Preparation of

Red Lake Pigment by Supercritical Fluid," The 4th International Symposium on




59. Gilbert, D. J.; Palakodaty, S.; Sloan, R.; York, P.; "Particle Engineering for

Pharmaceurical Applications - A Process Scale Up," Proceedings of the 5th


60. Hanna, M.; York, P.; Yu. Shekunov, B.; "Control of the Polymeric Forms of a Drug

Substance by Solution Enhanced Dispersion by Supercritical Fluids (SEDS),"

Proceedings of the 5th Meeting on Supercritical Fluids, Nice, France, 1998, Tome 1.


6,063,138, 2000.



63. Hong, L.; Bitemo, S. R.; Gao, Y.; Yuan, W.; "Precipitation of Microparticulate

Organic Pigment Powders by Supercritical Anti-Solvent (SAS) Process,"


Georgia, 2000.

64. Jaarmo, S.; Rantakyla, M.; Aaltonen, O.; "Particle Tailoring with Supercritical

Fluids: Production of Amorphous Pharmaceutical Particles," The 4th International

Symposium on Supercritical Fluids, Sendai, Japan, 1997, A, 263.

65. Kulshreshtha, A. K.; Smith, G. G.; Anderson, S. D.; Krukonis, V. J.; "Process for

Sizing Prednisolone Acetate using a Supercritical Fluid Anti-Solvent," Us 5803966,

1998.

66. Johnston, K. P.; Luna-Barcenas, G.; Dixon, D.; Mawson, S.; "Polymeric Materials

by Precipitation with a Compressed Fluid Anti-Solvent," Proceedings of the 3rd

International Symposium on Supercritical Fluids, Strasbourge, France, 1994, 359.

APPENDIX I

323




68. Magnan, C.; Commenges, N.; Badens, E.; Charbit, G.; "Fine Phospholipid Particles

Formed by Precipitation with a Compressed Fluid Anti-Solvent," Proccedings of the



69. Magnan, C.; Badens, E.; Commenges, N.; Charbit, G.; "Soy Lecithin Micronization

by Precipitation with a Compressed Fluid Anti-Solvent - Influence of Process

Parameters," Fifth Conference on Supercritical Fluids and their Applications, Garda

(Verona), 1999, 479.





Sci. 1997, 64, 2105.







1997, 13, 1519.



Bioeng. 2000, 67, 457.




76. Palakodaty, S.; York, P.; Hanna, M.; Pritchard, J.; "Crystallization of Lactose Using

Solution Enhanced Dispersion by Supercritical Fluids (SEDS) Technique,"

Proceedings of the 5th Meeting on Supercritical Fluids, Nice, France, 1998, Tome 1,

275.

APPENDIX I

324




78. Rantakyla, M.; Jantti, M.; Jaarmo, S.; Aaltonen, O.; "Modeling Droplet - Gas

Interaction and Particle Formation in Gas-Anti-Solvent System (GAS)," Proceedings

of the 5th Meeting on Supercritical Fluids, Nice, France, 1998, 333.





Solvent Precipitation of Nanoparticles of a Zinc Oxide Precursor," Powder Technol.

1999, 102, 127.



1998, 37, 952.

82. Reverchon, E.; Della Porta, G.; Falivene, M. G.; "Process Parameters Controlling

the Supercritical Anti-Solvent Micronisation of Some Antibiotics," Proceedings of

the 6th Meeting on Supercritical Fluids, Nottingham, United Kingdom, 1999, 157.






85. Reverchon, E.; De Rosa, I.; Della Porta, G.; "Effect of Process Parameters on the

Supercritical Anti-Solvent Precipitation of Microspheres of Natural Polymers,"

Dipartimento Ingegneria Chimica Alimentare,Universita Salerno,Fisciano,Italy.,

1999.





APPENDIX I

325

87. Robertson, J.; King, M. B.; Seville, J. P. K.; Merrifield, D. R.; Buxton, P. C.;

"Recrystallisation of Organic Compounds Using Near Critical Carbon Dioxide," The

4th International Symposium on Supercritical Fluids, Sendai, Japan, 1997, A, 47.

88. Robertson, J.; King, M. B.; Seville, J. P. K.; "Particle Production Using Near-

Critical Solvents," Proceedings of the 5th Meeting on Supercritical Fluids, Nice,

France, 1998, 339.


Microparticles. Influence of Process Parameters on the Residual SOlvent in


J. Pharm. Sci. 1997, 86, 101.



1995, 41, 2476.




92. Sloan, R.; Tservistas, M.; Hollowood, M. E.; Sarup, L.; Humphreys, G. O.; York,

P.; Ashraf, W.; Hoare, M.; "Controlled Particle Formation of Biological Material

Using Supercritical Fluids," Proceedings of the 6th Meeting on Supercritical Fluids,

Nottingham, United Kingdom, 1999, 169.

93. Shekunov, B. Y.; Hanna, M.; York, P.; "Crystallization Process in Turbulent

Supercritical Flows," J. Cryst. Growth 1999, 198/199, 1345.




Precipitation and Coating Using Near-Critical and Supercritical Antisolvents," Us

5833891, 1998.




97. Sze Tu, L.; Dehghani, F.; Dillow, A. K.; Foster, N. R.; "Applications of Dense

Gases in Pharmaceutical Processing," Proceedings of the 5th Meeting on

Supercritical Fluids, Nice, France, 1998, 263.

APPENDIX I

326



3898.






F., Ed. 1993; Vol. 514, 238.




102. Winters, M. A.; Debenedetti, P. G.; Carey, J.; Sparks, H. G.; Sane, S. U.;

Przybycien, T. M.; "Long-Term and High-Temperature Storage of Supercritically-

Processed Microparticulate Protein Powders," Pharm. Res. 1997, 14, 1370.

103. Wubbolts, F. E.; Bruinsma, O. S. L.; de Graauw, J.; van Rosmalen, G. M.;

"Continuous Gas Anti-Solvent Crystallisation of Hydroquinone from Acetone Using

Carbon Dioxide," The 4th International Symposium on Supercritical Fluids, Sendai,

Japan, 1997, A, 63.



1999, 198/199, 767.

105. Yeo, S.-D.; Debenedetti, P. G.; Patro, S. Y.; Przybycien, T. M.; "Secondary

Structure Characterization of Microparticulate Insulin Powders," J. Pharm. Sci.

1994, 83, 1651.








106, 77.

APPENDIX I

327

109. Elvassore, N.; Bertucco, A.; Caliceti, P.; "Production of Protein-Polymer Micro-

Capsules by Supercritical Anti-Solvent Techniques," Proceedings of the 5th





111. Falk, R.; Randolph, T. W.; Meyer, J. D.; Kelly, R. M.; Manning, M. C.;

"Controlled Release of Ionic Compounds from Poly(L-Lactide) Microspheres

Produced by Precipitation with a Compressed Anti-Solvent," J. Controlled Release

1997, 44, 77.








114. Weber, A.; Tschernjaew, J.; Kummel, R.; "Coprecipitation with Compressed

Antisolvents for the Manufacture of Microcomposites," Proceedings of the 5th

Meeting on Supercritical Fluids, Nice, France, 1998, Tome 1, 243.




116. Catchpole, O. J.; Bergmann, C.; "Continuous Gas Anti-Solvent Fractionation Of

Natural Products," Proceedings of the 5th Meeting on Supercritical Fluids, Nice,

France, 1998, 257.




118. Kordikowski, A.; York, P.; "Chiral Separation Using Supercritical CO2,"

Proceedings of the 6th Meeting on Supercritical Fluids, Nottingham, United

Kingdom, 1999, 163.

APPENDIX I

328

119. Weber, A.; Nolte, C.; Bork, M.; Kummel, R.; "Recovery of Lecithin from egg

Yolk-Extracts by Gas Anti-Solvent Crystallization," Proceedings of the 6th Meeting

on Supercritical Fluids, Nottingham, United Kingdom, 1999.

APPENDIX II

329

II. VOLUMETRIC EXPANSION EXPERIMENTAL RESULTS —

BINARY DATA

II.1. CO2—DMF

Table II.1 CO2—DMF experimental results at 25°C.

Mole Fraction Pressure /

MPa CO2 DMF

Expansion /

∆V %

0.00 0.000 1.000 0.0

0.77 0.097 0.903 2.2

1.38 0.179 0.821 8.8

2.12 0.285 0.715 17.8

2.76 0.381 0.619 28.4

3.48 0.486 0.514 49.1

4.16 0.587 0.413 71.6

4.81 0.700 0.300 121.6

5.49 0.869 0.131 384.6

5.85 0.939 0.061 942.5

APPENDIX II

330



MPa CO2 DMF

Expansion /

∆V %

0.00 0.000 1.000 0.0

0.82 0.122 0.878 4.8

1.52 0.216 0.784 12.2

2.10 0.290 0.710 17.5

2.74 0.379 0.621 29.1

3.45 0.466 0.534 41.8

4.12 0.552 0.448 61.3

4.83 0.653 0.347 97.9

5.54 0.768 0.232 177.8

6.21 0.896 0.104 500.8



MPa CO2 DMF

Expansion /

∆V %

0.00 0.000 1.000 0.0

0.72 0.102 0.898 2.4

2.14 0.251 0.749 14.0

3.45 0.382 0.618 28.5

4.19 0.465 0.535 42.4

4.85 0.539 0.461 57.5

5.50 0.613 0.387 79.3

6.21 0.675 0.325 106.0

6.89 0.757 0.243 167.5

7.58 0.873 0.127 394.5

APPENDIX II

331

II.2. CO2—NMP

Table II.4 CO2—NMP experimental results at 25°C.


MPa CO2 DMF

Expansion /

∆V %

0.00 0.000 1.000 0.0

0.81 0.113 0.887 5.1

1.42 0.194 0.806 9.5

2.11 0.287 0.713 16.0

2.79 0.379 0.621 26.1

3.50 0.463 0.537 34.7

4.19 0.578 0.452 50.8

4.88 0.624 0.376 77.8

5.52 0.725 0.275 121.1

5.85 0.801 0.199 197.0

Table II.5 CO2—NMP experimental results at 40°C.


MPa CO2 DMF

Expansion /

∆V %

0.00 0.000 1.000 0.0

2.10 0.167 0.833 8.9

3.38 0.277 0.723 18.3

4.85 0.406 0.594 34.8

5.61 0.478 0.522 48.7

6.25 0.537 0.463 62.9

6.89 0.612 0.388 89.3

7.59 0.718 0.282 151.0

APPENDIX II

332

II.3. CO2—DMSO

Table II.6 CO2—DMSO experimental results at 25°C.


MPa CO2 DMF

Expansion /

∆V %

0.00 0.000 1.000 0.0

0.79 0.076 0.924 3.7

1.39 0.132 0.868 6.8

2.16 0.206 0.794 14.3

2.98 0.291 0.709 25.1

3.48 0.353 0.647 33.5

4.19 0.435 0.565 45.3

4.90 0.514 0.486 71.3

5.55 0.608 0.392 104.2

5.96 0.730 0.270 176.6

Table II.7 CO2—DMSO experimental results at 40°C.


MPa CO2 DMF

Expansion /

∆V %

0.00 0.000 1.000 0.0

2.03 0.144 0.856 6.8

3.53 0.261 0.739 17.0

4.82 0.382 0.618 31.3

5.53 0.452 0.548 43.1

6.25 0.518 0.482 57.7

6.89 0.595 0.405 82.3

7.58 0.698 0.302 131.8

APPENDIX III

333

III. VOLUMETRIC EXPANSION EXPERIMENTAL RESULTS —

TERNARY DATA

III.1. CO2—DMF—Cu-Indo

Table III.1 CO2—DMF—Cu-Indo experimental results at 25°C.


MPa CO2 DMF Cu-Indo /

x 10-6

Cu-Indo

Conc. / mg.g-1

(CO2 free

basis)

0.00 0.000 0.999 23.3 5.39

0.69 0.080 0.920 21.4 5.32

1.38 0.197 0.803 17.1 4.93

2.07 0.282 0.718 11.2 3.61

2.76 0.387 0.613 8.05 3.05

3.45 0.487 0.513 4.75 2.15

4.14 0.582 0.418 2.11 1.17

4.83 0.697 0.303 0.69 0.53

5.52 0.871 0.129 0.14 0.26

5.86 0.941 0.059 0.08 0.25

APPENDIX III

334




x 10-6

Cu-Indo

Conc. / mg.g-1

(CO2 free

basis)

0.75 0.109 0.891 23.4 6.06

1.40 0.199 0.800 20.4 5.89

2.11 0.280 0.720 5.99 5.14

2.81 0.369 0.631 12.3 4.52

3.43 0.449 0.551 9.43 3.96

4.16 0.546 0.454 5.92 3.03

4.85 0.634 0.366 2.90 1.84

5.54 0.741 0.259 0.98 0.88

6.18 0.879 0.121 0.10 0.19




x 10-6

Cu-Indo

Conc. / mg.g-1

(CO2 free

basis)

0.69 0.076 0.924 42.2 10.5

2.10 0.226 0.773 32.5 9.69

3.47 0.370 0.630 21.2 7.76

4.16 0.443 0.557 15.0 6.22

4.84 0.514 0.486 10.9 5.18

5.52 0.585 0.415 5.64 3.16

6.23 0.669 0.331 2.57 1.80

6.89 0.750 0.250 1.16 1.08

7.63 0.879 0.121 0.24 0.47

APPENDIX III

335

III.2. CO2—DMF—Cu-Acetate

Table III.4 CO2—DMF—Cu-Acetate experimental results at 25°C.


MPa CO2 DMF Cu-Acetate /

x 10-3

Cu-Acetate

Conc. / mg.g-1

(CO2 free

basis)

0.00 0.000 0.981 18.7 51.9

0.79 0.103 0.877 20.6 60.4

1.37 0.184 0.796 19.8 63.5

2.10 0.294 0.688 18.04 66.8

2.79 0.400 0.584 15.6 68.0

3.59 0.466 0.521 12.1 59.8

4.14 0.555 0.439 5.90 35.4

4.83 0.665 0.332 3.14 25.2

5.52 0.854 0.147 4.03 7.45

5.86 0.946 0.054 1.46 7.38

APPENDIX III

336

III.3. CO2—DMF—Indomethacin

Table III.5 CO2—DMF—Indomethacin experimental results at 25°C.


MPa CO2 DMF Indomethacin

Indomethacin /

mg.g-1 (CO2

free basis)

0 0.000 0.786 0.214 571

0.77 0.036 0.758 0.206 571

1.43 0.124 0.689 0.188 571

2.05 0.177 0.647 0.176 571

2.76 0.239 0.598 0.163 571

3.50 0.318 0.536 0.146 571

4.14 0.393 0.477 0.130 571

4.83 0.464 0.421 0.115 571

5.52 0.578 0.332 0.090 571

5.93 0.951 0.048 0.000 46.0

APPENDIX IV

337

IV. VOLUMETRIC EXPANSION EXPERIMENTAL RESULTS —

QUATERNARY DATA

IV.1. CO2—DMF—Cu-Indo—Cu-Acetate

Table IV.1 CO2—DMF—Cu-Indo—Cu-Acetate experimental results at 25°C.

(Both solutes added in excess)

Mole Fraction Concentration / mg.g-1

(CO2 free basis) Pressure /

MPa CO2 DMF

Cu-Indo

/ x 10-5

Cu-Acetate

/ x 10 -5 Cu-Indo Cu-Acetate

0.00 0.000 0.983 57.7 1692 12.9 44.4

1.47 0.220 0.764 56.2 1508 16.0 50.4

3.33 0.474 0.516 38.3 1032 16.1 51.0

4.09 0.579 0.411 30.7 925 16.1 57.0

5.28 0.803 0.191 6.67 85.0 11.2 11.6

5.81 0.924 0.076 0.74 4.64 2.26 1.66

Table IV.2 CO2—DMF—Cu-Indo—Cu-Acetate experimental results at 25°C. (Cu-

Indo added in excess, Cu-Acetate initial concentration 8 mg/g)



MPa CO2 DMF

Cu-Indo

/ x 10-5

Cu-Acetate

/ x 10 –5 Cu-Indo Cu-Acetate

1.02 0.123 0.874 99.4 262 25.6 7.92

2.16 0.292 0.706 80.5 196 25.7 7.35

3.47 0.479 0.520 42.4 111 18.5 5.70

4.48 0.635 0.364 24.9 66.3 15.6 4.88

5.16 0.765 0.234 6.36 35.1 6.25 4.05

5.82 0.917 0.083 0.13 0.86 0.37 0.28

APPENDIX IV

338

IV.2. CO2—DMF—Cu-Indo—Indomethacin

Table IV.3 CO2—DMF—Cu-Indo—Indomethacin experimental results at 25°C.

(Cu-Indo added in excess, indomethacin initial concentration 6.2 mg/g)



MPa CO2 DMF

Cu-Indo

/ x 10-5

Indomethacin

/ x 10 -5

Cu-

Indo Indomethacin

0.90 0.132 0.867 10.2 111 2.72 6.2

2.34 0.334 0.666 5.28 63.5 1.83 6.2

3.45 0.495 0.504 2.48 49.1 1.14 6.2

4.48 0.666 0.333 0.75 36.3 0.52 6.2

5.17 0.789 0.211 0.06 26.2 0.07 6.2

5.86 0.956 0.044 0.01 5.70 0.04 6.2

IV.3. CO2—DMF—Cu-Indo—Acetic acid

Table IV.4 CO2—DMF—Cu-Indo—Acetic Acid experimental results at 25°C. (Cu-

Indo added in excess, acetic acid initial concentration 60 mg/g)

Pressure /

MPa

Cu-Indo /

mg.g-1 (CO2 free basis)

Acetic Acid /

mg.g-1 (CO2 free basis)

0.77 57.2 60

2.07 57.3 60

2.76 58.0 60

3.45 57.2 60

4.14 56.2 60

4.83 56.0 60

5.44 28.2 60

5.86 4.0 60

APPENDIX V

339

V. DISSOLUTION EXPERIMENTAL RESULTS

V.1. Dissolution of Micronised Cu-Indo

Table V.1 Dissolution of unprocessed and micronised Cu-Indo.

% Dissolution Time

/

min. Unprocessed Micronised

Amorphous

Micronised

Crystalline

10 3.57 12.55 20.10

20 3.19 30.52 38.00

30 5.45 43.62 48.01

40 — 52.15 56.26

45 8.46 — —

50 — 56.24 62.07

60 10.84 60.65 64.40

80 — 66.71 68.29

90 14.10 — —

100 — 67.42 72.97

120 18.32 72.75 75.82

150 19.85 — —

160 — 79.80 86.02

190 27.14 — —

APPENDIX V

340

V.2. Dissolution of PVP—Cu-Indo Co-Precipitates

Table V.2 Dissolution of PVP—Cu-Indo co-precipitates.

% Dissolution Time

/

min. 1:1

PVP:Cu-Indo

7:3

PVP:Cu-Indo

9:1

PVP:Cu-Indo

1:1

Physical Mix

10 27.81 41.67 27.82 2.81

15 — — — 5.68

20 46.98 54.39 34.98 6.54

25 — — — 6.54

30 61.41 65.85 41.80 7.40

40 — 68.48 47.15 —

45 — — — 10.26

50 69.18 72.71 50.48 —

60 74.19 75.40 56.93 15.99

80 80.62 79.17 63.19 —

100 84.71 83.16 71.25 —

120 87.44 86.12 73.29 —

APPENDIX VI

341

VI. SEM IMAGES OF COPPER INDOMETHACIN PARTICLES

PRODUCED BY THE GAS AND ASES PROCESSES VI.1. Particles Produced by the GAS Process

Figure VI.1 SEM image of Cu-Indo particles produced by the GAS process at 40°C

from a stirred 5 mg.g-1 solution of Cu-Indo in DMF with rapid expansion.

20 µm


from a stirred 5 mg.g-1 solution of Cu-Indo in NMP with slow expansion.

(b)

300 µm

APPENDIX VI

342


from a stirred 5 mg.g-1 solution of Cu-Indo in NMP with slow expansion.

10 µm


from a stirred 5 mg.g-1 solution of Cu-Indo in NMP with rapid expansion.

20 µm

APPENDIX VI

343


from a stirred 5 mg.g-1 solution of Cu-Indo in DMSO with slow expansion.

50 µm


from a stirred 5 mg.g-1 solution of Cu-Indo in DMSO with rapid expansion.

10 µm

APPENDIX VI

344


from a stirred 5 mg.g-1 solution of Cu-Indo in DMSO with rapid expansion.

4 µm


from a stirred 200 mg.g-1 solution of Cu-Indo in DMF with slow expansion.

5 µm

APPENDIX VI

345


from a stirred 200 mg.g-1 solution of Cu-Indo in DMF with slow expansion.

20 µm

Figure VI.10 SEM image of Cu-Indo particles produced by the GAS process at

25°C from a stirred 100 mg.g-1 solution of Cu-Indo in DMF with slow expansion.

50 µm

APPENDIX VI

346


25°C from a stirred 200 mg.g-1 solution of Cu-Indo in DMF with rapid expansion.

10 µm


25°C from a stirred 20 mg.g-1 solution of Cu-Indo in DMF with slow expansion.

75 µm

APPENDIX VI

347

VI.2. Particles Produced by the ASES Process

Figure VI.13 SEM images of Cu-Indo particles by the ASES process from 5 mg.g-1

solutions of Cu-Indo in DMF at 25°C and 6.89 MPa with a solution flow rate of 0.2

mL.min-1 and a nozzle diameter of 229 µm.

10 µm


solutions of Cu-Indo in DMF at 40°C and 14.48 MPa with a solution flow rate of

0.2 mL.min-1 and a nozzle diameter of 1020 µm.

5 µm

APPENDIX VI

348


solutions of Cu-Indo in DMSO at 25°C and 6.89 MPa with a solution flow rate of


10 µm


solutions of Cu-Indo in DMSO at 25°C and 13.79 MPa with a solution flow rate of


10 µm

APPENDIX VI

349


solutions of Cu-Indo in NMP at 25°C and 13.79 MPa with a solution flow rate of


10 µm


solutions of Cu-Indo in DMF at 40°C and 14.48 MPa with a solution flow rate of


5 µm

APPENDIX VI

350

Figure VI.19 SEM images of Cu-Indo particles by the ASES process from 100

mg.g-1 solutions of Cu-Indo in DMF at 25°C and 6.89 MPa with a solution flow rate

of 0.2 mL.min-1 and a nozzle diameter of 229 µm.

50 µm

APPENDIX VII

351

VII. PARTICLE SIZE ANALYSES RESULTS

Figure VII.1 Particle size distributions of Cu-Indo particles produced from DMF

solutions by the ASES process. 100 mg.g-1 Cu-Indo solution at 25°C and 6.89 MPa

using a 229 µm nozzle and a solution flow rate of 0.2 mL.min-1.

0

5

10

15

20

25

30

2.0 2.7 3.6 4.9 6.6 9.0 12.2 16.6 22.5 30.5 48.3

Particle Diameter / µm

In %

APPENDIX VII

352




0

5

10

15

20

25

30

2.0 2.7 3.6 4.9 6.6 9.0 12.2 16.6 22.5 30.5 48.3


In %




0

5

10

15

20

25

30

2.0 2.7 3.6 4.9 6.6 9.0 12.2 16.6 22.5 30.5 48.3


In %

APPENDIX VII

353




0

5

10

15

20

25

30

35

40

2.0 2.7 3.6 4.9 6.6 9.0 12.2 16.6 22.5 30.5 48.3


In %

APPENDIX VIII

354

VIII. SEM IMAGES OF PVP AND PVP—Cu-INDO CO-

PRECIPITATES VIII.1. PVP Particles Produced by the ASES Process

Figure VIII.1 SEM images of particles produced from DMF solutions of PVP at a



6 µm

APPENDIX VIII

355

Figure VIII.2 SEM images of particles produced from DMF solutions of PVP at a



4 µm VIII.2. PVP—Cu-Indo Particles Produced by the ASES Process

Figure VIII.3 SEM images of the PVP—Cu-Indo particles produced from DMF

solutions with a solute concentration of 50 mg.g-1 by the ASES process at 25°C and

6.6 MPa using a solution flow rate of 0.1 mL.min-1 through a 229 µm nozzle. PVP

to Cu-Indo ratio 60:40.

3 µm

APPENDIX VIII

356


solutions with a solute concentration of 50 mg.g-1 by the ASES process at 40°C and

19.0 MPa using a solution flow rate of 0.1 mL.min-1 through a 229 µm nozzle. PVP

to Cu-Indo ratio 60:40.

6 µm




PVP to Cu-Indo ratio 90:10.

5 µm

APPENDIX VIII

357





1 µm





1 µm

LIST OF PUBLICATIONS

358

The following list of publications have resulted from the work conducted in this thesis.

1. Warwick, B.; Dehghani, F.; Foster, N. R.; Biffin, J. R.; Regtop, H. L., “Synthesis,

Purification, and Micronization of Pharmaceuticals Using the Gas Anti-Solvent

Technique., Ind. Eng. Chem. Res., 2000, 39(12), 4571-4579.

2. Foster, N. R.; Bezanehtak, K.; Charoenchaitrakool, M.; Combes, G.; Dehghani, F.;

Sze Tu, L.; Thiering, R.; Warwick, B.; Bustami, R. T.; Chan, H-K.; Processing

Pharmaceuticals Using Dense Gas Technology, In the Proceedings of the 5th

International Symposium of Supercritical Fluids, Atlanta, April 9-11, 2000.

3. Warwick,, B., Dehghani, F., Foster, N. R.; Biffin, J. R.; Regtop, H. L.; Synthesis,

Purification and Micronisation of Pharmaceuticals Using the Gas Antisolvent

Technique, In the Proceedings of the 5th International Symposium of Supercritical

Fluids, Atlanta, April 9-11, 2000.

4. Combes, G.; Warwick, B.; Dehghani, F.; Lucein, F.; Dillow, A.; Foster, N. R.;

“Dense Gas Solvent as Reaction Media”, Proceeding of CISF99-5th Conference on

Supercritical Fluids and their applications, 13-16 June, 1999, Garda (Verona), Italy.

synthesis, purification and micronisation of copper

Documents