REMOVAL OF OIL FROM WATER USING FUNGAL BIOMASS A …

REMOVAL OF OIL FROM WATER USING FUNGAL BIOMASS

A Thesis

Submitted to the Faculty of Graduate Studies and Research

In Partial Fulfillment of the Requirements

For the degree of

Doctor of Philosophy

In

Environmental Systems Engineering

University of Regina

By

Asha Srinivasan

Regina, Saskatchewan

February 2012

Copyright 2012: A. Srinivasan

REMOVAL OF OIL FROM WATER USING FUNGAL BIOMASS

A Thesis

Submitted to the Faculty of Graduate Studies and Research

In Partial Fulfillment of the Requirements

For the degree of

Doctor of Philosophy

In

Environmental Systems Engineering

University of Regina

By

Asha Srinivasan

Regina, Saskatchewan

February 2012

Copyright 2012: A. Srinivasan

I 1 Library and Archives Canada

Published Heritage Branch

Bibliotheque et Archives Canada

Direction du Patrimoine de ('edition

395 Wellington Street 395, rue Wellington Ottawa ON KlA ON4 Ottawa ON MA ON4 Canada Canada

NOTICE: AVIS:

The author has granted a non-exclusive license allowing Library and Archives Canada to reproduce, publish, archive, preserve, conserve, communicate to the public by telecommunication or on the Internet, loan, distrbute and sell theses worldwide, for commercial or non-commercial purposes, in microform, paper, electronic and/or any other formats.

The author retains copyright ownership and moral rights in this thesis. Neither the thesis nor substantial extracts from it may be printed or otherwise reproduced without the author's permission.

L'auteur a accord permettant a la 131 Canada de repro( sauvegarder, con par telecommunic distribuer et vend monde, a des fine support microforn autres formats.

L'auteur conserve et des droits more la these ni des ex ne doivent etre irr reproduits sans sl

1*1 Library and Archives Bibliotheque et Canada Archives Canada

Published Heritage Direction du Branch Patrimoine de I'edition

395 Wellington Street 395, rue Wellington Ottawa ON K1A 0N4 Ottawa ON K1A 0N4 Canada Canada

NOTICE:

The author has granted a nonexclusive license allowing Library and Archives Canada to reproduce, publish, archive, preserve, conserve, communicate to the public by telecommunication or on the Internet, loan, distrbute and sell theses worldwide, for commercial or noncommercial purposes, in microform, paper, electronic and/or any other formats.

AVIS:

L'auteur a accord permettant a la Bi

Canada de reproc sauvegarder, con partelecommunic distribuer et vend monde, a des fins support microforn autres formats.

The author retains copyright ownership and moral rights in this thesis. Neither the thesis nor substantial extracts from it may be printed or otherwise reproduced without the author's permission.

L'auteur conserve et des droits more la these ni des ex ne doivent etre irr reproduits sans s<

UNIVERSITY OF REGINA

FACULTY OF GRADUATE STUDIES AND RESEARCH

SUPERVISORY AND EXAMINING COMMITTEE

Asha Srinivasan, candidate for the degree of Doctor of Philosophy in Environmental Systems Engineering, has presented a thesis titled, Removal of Oil from Water Using Fungal Biomass, in an oral examination held on December 16, 2011. The following committee members have found the thesis acceptable in form and content, and that the candidate demonstrated satisfactory knowledge of the subject material.

External Examiner:

Co-Supervisor:

Co-Supervisor:

Committee Member:

Committee Member:

Committee Member:

Committee Member:

Chair of Defense:

*Not present at defense

Dr. Jian Peng, University of Saskatchewan

Dr. Tsun Wai Kelvin Ng, Envirnmental Systems Engineering.

Dr. Thiruvenkatachari Viraraghavan, Adjunct

Dr. Yee-Chung Jin, Envirnmental Systems Engineering

Dr. Dena McMartin, Environmental Systems Engineering

Dr. Amr Henni, Industrial Systems Engineering

Dr. Harold Weger, Department of Biology

Dr. Dongyan Blachford, Faculty of Graduate Studies & Research

UNIVERSITY OF REGINA

FACULTY OF GRADUATE STUDIES AND RESEARCH

SUPERVISORY AND EXAMINING COMMITTEE

Asha Srinivasan, candidate for the degree of Doctor of Philosophy in Environmental Systems Engineering, has presented a thesis titled, Removal of Oil from Water Using Fungal Biomass, in an oral examination held on December 16, 2011. The following committee members have found the thesis acceptable in form and content, and that the candidate demonstrated satisfactory knowledge of the subject material.

External Examiner: Dr. Jian Peng, University of Saskatchewan

Co-Supervisor: Dr. Tsun Wai Kelvin Ng, Envirnmental Systems Engineering.

Co-Supervisor: Dr. Thiruvenkatachari Viraraghavan, Adjunct

Committee Member: Dr. Yee-Chung Jin, Envirnmental Systems Engineering

Committee Member: Dr. Dena McMartin, Environmental Systems Engineering

Committee Member: Dr. Amr Henni, Industrial Systems Engineering

Committee Member: Dr. Harold Weger, Department of Biology

Chair of Defense: Dr. Dongyan Blachford, Faculty of Graduate Studies & Research

*Not present at defense

Abstract

The presence of oil in water is of major concern due to its impact on the

environment. Various materials, used to remove oil from water, have either exhibited low

removal efficiencies or are not selective. The use of biomaterials such as bacteria, fungi,

or plant biomass, to adsorb organic substances, has been examined to a limited extent; the

biosorption of oil from water by fungal biomass has not been investigated, so far.

The present study evaluated the potential of non-viable Mucor rouxii biomass

with respect to the removal of three representative oils (Standard Mineral Oil (SMO),

Canola Oil (CO) and Bright-Edge 80 cutting oil) from water via a series of batch and

column adsorption experiments. A preliminary batch adsorption study was conducted to

evaluate oil removal capacities of two non-viable fungal biomasses, Mucor rouxii and

Absidia coerulea. Non- viable M rouxii biomass was found to be more effective than A.

coerulea biomass in removing oil from water. A fractional factorial design analysis was

conducted to screen significant factors influencing the removal of three oils from water

using M rouxii biomass. pH of the solution was observed to be the most influencing

parameter. Temperature had an effect on SMO and Bright-Edge 80 removal while the

adsorbent dose was found to influence the removal of SMO.

Detailed batch adsorption studies were therefore conducted to remove oil from

water using M rouxii biomass by varying the solution pH, the adsorbent dosage, the oil

concentration and the temperature. Adsorption of the three oils on to the M rouxii

biomass followed the pseudo second-order model. On further analysis, adsorption process

was found to have followed the intra-particle diffusion mechanism along with boundary

ii

Abstract

The presence of oil in water is of major concern due to its impact on the

environment. Various materials, used to remove oil from water, have either exhibited low

removal efficiencies or are not selective. The use of biomaterials such as bacteria, fungi,

or plant biomass, to adsorb organic substances, has been examined to a limited extent; the

biosorption of oil from water by fungal biomass has not been investigated, so far.

The present study evaluated the potential of non-viable Mucor rouxii biomass

with respect to the removal of three representative oils (Standard Mineral Oil (SMO),

Canola Oil (CO) and Bright-Edge 80 cutting oil) from water via a series of batch and

column adsorption experiments. A preliminary batch adsorption study was conducted to

evaluate oil removal capacities of two non-viable fungal biomasses, Mucor rouxii and

Absidia coerulea. Non- viable M. rouxii biomass was found to be more effective than A.

coerulea biomass in removing oil from water. A fractional factorial design analysis was

conducted to screen significant factors influencing the removal of three oils from water

using M. rouxii biomass. pH of the solution was observed to be the most influencing

parameter. Temperature had an effect on SMO and Bright-Edge 80 removal while the

adsorbent dose was found to influence the removal of SMO.

Detailed batch adsorption studies were therefore conducted to remove oil from

water using M. rouxii biomass by varying the solution pH, the adsorbent dosage, the oil

concentration and the temperature. Adsorption of the three oils on to the M. rouxii

biomass followed the pseudo second-order model. On further analysis, adsorption process

was found to have followed the intra-particle diffusion mechanism along with boundary

ii

layer diffusion. The Langmuir and Freundlich adsorption models were able to adequately

describe the equilibrium isotherms at different temperatures (5, 15, 22, and 30 °C).

Thermodynamic analysis showed adsorption to be spontaneous and endothermic. The

activation parameters indicated that adsorption was likely diffusion controlled. Chemical

modifications of the biomass and the FTIR analysis showed that carboxyl and amino

groups, present on the M rouxii cell surface were involved in oil sorption.

A continuous column study was carried out using immobilized M rouxii biomass

beads as a biosorbent for the removal of the three oils from water. The Thomas, Yan and

Yoon-Nelson models were found to be suitable in describing column behavior for all

three studied oils. Following column regeneration using de-ionized water, the beads

could be reused to remove oil to that of its initial capacity. Investigations on the

breakdown mechanisms and flow characteristics indicated the possible sequential

occurrence of coalescence and filtration in the immobilized M rouxii biomass bed.

In summary, non-viable M rouxii biomass was found to be an effective medium

for oil removal. This research improves our understanding of the mechanisms

contributing to adsorption of oil by the biomass. Immobilized biomass can be employed

in packed bed columns and re-used to increase their economic attractiveness. The

fundamental understanding of the breakthrough curve behavior is important for process

scale-up under realistic conditions. Further, the study provides a better knowledge on oil-

in-water emulsion flow in coalescing beds. These findings are significant for future

development of filtration systems in coalescing oil droplets that can be used for emulsion

separation.

iii

layer diffusion. The Langmuir and Freundlich adsorption models were able to adequately

describe the equilibrium isotherms at different temperatures (5, 15, 22, and 30 °C).

Thermodynamic analysis showed adsorption to be spontaneous and endothermic. The

activation parameters indicated that adsorption was likely diffusion controlled. Chemical

modifications of the biomass and the FTIR analysis showed that carboxyl and amino

groups, present on the M. rouxii cell surface were involved in oil sorption.

A continuous column study was carried out using immobilized M. rouxii biomass

beads as a biosorbent for the removal of the three oils from water. The Thomas, Yan and

Yoon-Nelson models were found to be suitable in describing column behavior for all

three studied oils. Following column regeneration using de-ionized water, the beads

could be reused to remove oil to that of its initial capacity. Investigations on the

breakdown mechanisms and flow characteristics indicated the possible sequential

occurrence of coalescence and filtration in the immobilized M. rouxii biomass bed.

In summary, non-viable M. rouxii biomass was found to be an effective medium

for oil removal. This research improves our understanding of the mechanisms

contributing to adsorption of oil by the biomass. Immobilized biomass can be employed

in packed bed columns and re-used to increase their economic attractiveness. The

fundamental understanding of the breakthrough curve behavior is important for process

scale-up under realistic conditions. Further, the study provides a better knowledge on oil-

in-water emulsion flow in coalescing beds. These findings are significant for future

development of filtration systems in coalescing oil droplets that can be used for emulsion

separation.

iii

Acknowledgements

I am grateful to my supervisor, Dr. T. Viraraghavan, for his guidance,

encouragement, and understanding that he extended throughout the course of this work. I

thank my co-supervisor, Dr. Kelvin Ng, for his valuable suggestions and assistance in the

preparation of this thesis. I also wish to thank the following members of the dissertation

committee, Dr. Y-C. Jin, Dr. Dena McMartin, Dr. Harold Weger, Dr. Amr Henni and Dr.

Nader Mahinpey for their valuable input over the course of my study.

I would like to acknowledge the assistance of Mr. S. R. Dhanushkodi, University

of Waterloo, who conducted surface area, FTIR and SEM analyses, Dr. Raimar

LObenberg, University of Alberta, who measured the zeta potential of the samples and

Ms. Lauren Bradshaw, University of Regina, who conducted surface area analysis of the

biomaterials. I thank Mr. Herald Berwald who provided the free Bright-Edge 80 cutting

oil used in the study.

I would like to thank the Natural Sciences and Engineering Research Council of

Canada for their financial support to this study by way of a grant to Dr. T. Viraraghavan.

I would like to thank the Faculty of Graduate Studies and Research, University of Regina,

for the partial financial support through graduate scholarships and teaching assistantships.

I also wish to thank the Faculty of Engineering and Applied Science, University of

Regina, for partial financial support.

Above all, I love and appreciate my family for their understanding and moral

support throughout this research.

iv

Acknowledgements

I am grateful to my supervisor, Dr. T. Viraraghavan, for his guidance,

encouragement, and understanding that he extended throughout the course of this work. I

thank my co-supervisor, Dr. Kelvin Ng, for his valuable suggestions and assistance in the

preparation of this thesis. I also wish to thank the following members of the dissertation

committee, Dr. Y-C. Jin, Dr. Dena McMartin, Dr. Harold Weger, Dr. Amr Henni and Dr.

Nader Mahinpey for their valuable input over the course of my study.

I would like to acknowledge the assistance of Mr. S. R. Dhanushkodi, University

of Waterloo, who conducted surface area, FTIR and SEM analyses, Dr. Raimar

Lobenberg, University of Alberta, who measured the zeta potential of the samples and

Ms. Lauren Bradshaw, University of Regina, who conducted surface area analysis of the

biomaterials. I thank Mr. Herald Berwald who provided the free Bright-Edge 80 cutting

oil used in the study.

I would like to thank the Natural Sciences and Engineering Research Council of

Canada for their financial support to this study by way of a grant to Dr. T. Viraraghavan.

I would like to thank the Faculty of Graduate Studies and Research, University of Regina,

for the partial financial support through graduate scholarships and teaching assistantships.

I also wish to thank the Faculty of Engineering and Applied Science, University of

Regina, for partial financial support.

Above all, I love and appreciate my family for their understanding and moral

support throughout this research.

iv

Table of Contents

Abstract ii

Acknowledgements iv

Table of Contents v

List of Figures x

List of Tables xvi

List of Abbreviations, Symbols, Nomenclature xxi

Chapter 1 Introduction 1

1.1 Background 1

1.2 Sources and Quantity of Generated Oily Waters 3

1.3 Existing Treatment Technologies for Oily Water 4

1.4 Environmental Legislation and Concerns 4

1.5 Rationale for using M rouxii Biomass as an Adsorbent for Oil Removal 6

1.6 Objectives and Scope of this Study 8

Chapter 2 Literature Review 11

2.1 Emulsions 11

2.1.1 Classification of Emulsions 11

2.1.2 Physical Characteristics of Emulsions 12

2.1.3 Emulsion Droplet Characteristics 13

2.1.4 Emulsion Stability 14

2.2 Removal of Oil from Water 16

2.2.1 Dissolved Air Flotation 16

v

Table of Contents

Abstract ii

Acknowledgements iv

Table of Contents v

List of Figures x

List of Tables xvi

List of Abbreviations, Symbols, Nomenclature xxi

Chapter 1 Introduction 1

1.1 Background 1

1.2 Sources and Quantity of Generated Oily Waters 3

1.3 Existing Treatment Technologies for Oily Water 4

1.4 Environmental Legislation and Concerns 4

1.5 Rationale for using M. ronxii Biomass as an Adsorbent for Oil Removal 6

1.6 Objectives and Scope of this Study 8

Chapter 2 Literature Review 11

2.1 Emulsions 11

2.1.1 Classification of Emulsions 11

2.1.2 Physical Characteristics of Emulsions 12

2.1.3 Emulsion Droplet Characteristics 13

2.1.4 Emulsion Stability 14

2.2 Removal of Oil from Water 16

2.2.1 Dissolved Air Flotation 16

v

2.2.2 Aerobic Treatment of Oily Wastewater 19

2.2.3 Anaerobic Treatment of Oily Wastewaters 23

2.2.1 Enzymes for the Treatment of Oily Wastewaters 27

2.2.2 Microbial Degradation of Oils 32

2.2.3 Removal of Oil by Various Sorbents 39

2.2.4 Summary 45

2.3 Adsorption Processes in Environmental Engineering 45

2.3.1 Theoretical Background 45

2.3.2 Adsorption Kinetics 46

2.3.3 Adsorption Isotherm 48

2.3.4 Thermodynamic and Activation Parameters 49

2.3.5 Fixed Bed Column in Adsorption Studies 51

2.4 Breakdown Mechanisms 59

2.4.1 Filtration Mechanisms 59

2.4.2 Coalescence Mechanisms 61

Chapter 3 Materials and Methods 66

3.1 Glassware Preparation 66

3.2 Experimental Materials 66

3.3 Preparation of Oil-in-water Emulsions 67

3.4 Characterization of Oil-in-water Emulsions 67

3.5 Preparation of Fungal Seed 68

3.6 Preparation of Non-viable Fungal Biomass 68

vi

2.2.2 Aerobic Treatment of Oily Wastewater 19

2.2.3 Anaerobic Treatment of Oily Wastewaters 23

2.2.1 Enzymes for the Treatment of Oily Wastewaters 27

2.2.2 Microbial Degradation of Oils 32

2.2.3 Removal of Oil by Various Sorbents 39

2.2.4 Summary 45

2.3 Adsorption Processes in Environmental Engineering 45

2.3.1 Theoretical Background 45

2.3.2 Adsorption Kinetics 46

2.3.3 Adsorption Isotherm 48

2.3.4 Thermodynamic and Activation Parameters 49

2.3.5 Fixed Bed Column in Adsorption Studies 51

2.4 Breakdown Mechanisms 59

2.4.1 Filtration Mechanisms 59

2.4.2 Coalescence Mechanisms 61

Chapter 3 Materials and Methods 66

3.1 Glassware Preparation 66

3.2 Experimental Materials 66

3.3 Preparation of Oil-in-water Emulsions 67

3.4 Characterization of Oil-in-water Emulsions 67

3.5 Preparation of Fungal Seed 68

3.6 Preparation of Non-viable Fungal Biomass 68

vi

3.7 Characterization of Fungal Biomass 69

3.7.1 Surface Area Analysis and Surface Charge Measurement 70

3.7.2 Scanning Electron Microscope (SEM) Studies 70

3.7.3 Fourier Transform Infrared (FTIR) Analysis 71

3.8 Oil Concentration Measurement 71

3.9 Batch Biosorption Experiments 74

3.9.1 Preliminary Studies 75

3.9.2 Factorial Design of Experiments 75

3.9.3 Effect of pH 78

3.9.4 Effect of Concentration 78

3.9.5 Batch Kinetic Studies 79

3.9.6 Batch Isotherm Studies 79

3.9.7 Batch Desorption Studies 80

3.9.8 Modification of Functional Groups and Lipid Extraction 80

3.10 Use of Immobilized Biomass in Oil Removal 82

3.10.1 Procedure for Immobilization of Biomass 82

3.10.2 Characterization of Immobilized M rouxii Beads 82

3.11 Column Studies 83

3.11.1 Continuous Breakthrough Studies 83

3.11.1 Column Regeneration and Reuse 85

3.11.2 Coalescence/ Filtration Experiments 85

Chapter 4 Results and Discussion 89

vii

3.7 Characterization of Fungal Biomass 69

3.7.1 Surface Area Analysis and Surface Charge Measurement 70

3.7.2 Scanning Electron Microscope (SEM) Studies 70

3.7.3 Fourier Transform Infrared (FTIR) Analysis 71

3.8 Oil Concentration Measurement 71

3.9 Batch Biosorption Experiments 74

3.9.1 Preliminary Studies 75

3.9.2 Factorial Design of Experiments 75

3.9.3 Effect of pH 78

3.9.4 Effect of Concentration 78

3.9.5 Batch Kinetic Studies 79

3.9.6 Batch Isotherm Studies 79

3.9.7 Batch Desorption Studies 80

3.9.8 Modification of Functional Groups and Lipid Extraction 80

3.10 Use of Immobilized Biomass in Oil Removal 82

3.10.1 Procedure for Immobilization of Biomass 82

3.10.2 Characterization of Immobilized M. rouxii Be ads 82

3.11 Column Studies 83

3.11.1 Continuous Breakthrough Studies 83

3.11.1 Column Regeneration and Reuse 85

3.11.2 Coalescence/ Filtration Experiments 85

Chapter 4 Results and Discussion 89

vii

4.1 Data and Supplemental Figures 89

4.2 Characterization of Oil 89

4.3 Characterization of M rouxii Biomass and Other Adsorbents 94

4.4 Biosorption of Oil using Different Biomaterials 94

4.5 Factorial Design of Experiments 107

4.5.1 Pareto Plot of Effect 113

4.5.2 Main Effects Plot 113

4.5.3 Interaction Effects Plot 120

4.5.4 Normal Probability Plot of Residuals 124

4.6 Effect of pH on Biosorption 125

4.7 Effect of Concentration on Biosorption 130

4.8 Batch Kinetic Studies 132

4.9 Batch Isotherm Studies 143

4.10 Thermodynamics and Activation Parameters 147

4.11 Batch Desorption Studies 149

4.12 Surface Functional Groups on M rouxii Biomass 149

4.13 Role of Surface Functional Groups, Lipids and Surface Charge on Oil Biosorption

154

4.14 Immobilization of Fungal Biomass and its use in Biosorption 162

4.14.1 Batch Studies using Immobilized Biomass Beads 163

4.15 Column Breakthrough Studies 166

4.15.1 The Thomas Model 166

viii

4.1 Data and Supplemental Figures 89

4.2 Characterization of Oil 89

4.3 Characterization of M. rouxii Z?iomass and Other Adsorbents 94

4.4 Biosorption of Oil using Different Biomaterials 94

4.5 Factorial Design of Experiments 107

4.5.1 Pareto Plot of Effect 113

4.5.2 Main Effects Plot 113

4.5.3 Interaction Effects Plot 120

4.5.4 Normal Probability Plot of Residuals 124

4.6 Effect of pH on Biosorption 125

4.7 Effect of Concentration on Biosorption 130

4.8 Batch Kinetic Studies 132

4.9 Batch Isotherm Studies 143

4.10 Thermodynamics and Activation Parameters 147

4.11 Batch Desorption Studies 149

4.12 Surface Functional Groups on M. rouxii Biomass 149

4.13 Role of Surface Functional Groups, Lipids and Surface Charge on Oil Biosorption

154

4.14 Immobilization of Fungal Biomass and its use in Biosorption 162

4.14.1 Batch Studies using Immobilized Biomass Beads 163

4.15 Column Breakthrough Studies 166

4.15.1 The Thomas Model 166

viii

4.15.2 The Yan Model 171

4.15.3 The Belter and Chu Models 173

4.15.4 The Yoon and Nelson Model 180

4.15.5 The Outman Model 181

4.15.6 The Wolborska Model 181

4.16 Column Regeneration and Reuse 185

4.17 Coalescence/Filtration Mechanism 189

Chapter 5 Conclusion and Recommendations 209

5.1 Conclusion 209

5.2 Practical Research Applications 211

5.3 Contribution to the Field 214

5.4 Recommendations for Further Study 216

References 218

Appendix A Data and Supplementary Figures 248

Appendix B Design of Batch and Column Adsorber System 284

ix

4.15.2 The Yan Model 171

4.15.3 The Belter and Chu Models 173

4.15.4 The Yoon and Nelson Model 180

4.15.5 The Oulman Model 181

4.15.6 The Wolborska Model 181

4.16 Column Regeneration and Reuse 185

4.17 Coalescence/Filtration Mechanism 189

Chapter 5 Conclusion and Recommendations 209

5.1 Conclusion 209

5.2 Practical Research Applications 211

5.3 Contribution to the Field 214

5.4 Recommendations for Further Study 216

References 218

Appendix A Data and Supplementary Figures 248

Appendix B Design of Batch and Column Adsorber System 284

ix

List of Figures

Figure 3.1: Schematics of the experimental set up used for breakthrough studies 84

Figure 3.2: Schematics of the experimental set up used for breakdown studies 86

Figure 4.1: Zeta potential of autoclaved M rouxii biomass and oil-in-water emulsions 91

Figure 4.2: Scanning electron micrographs 96

Figure 4.3: Plot of diameter versus % walnut shell media passing the related sieve 98

Figure 4.4: Adsorption capacity versus time for SMO 99

Figure 4.5: Adsorption capacity versus time for CO 100

Figure 4.6: Adsorption capacity versus time for Bright-Edge 80 101

Figure 4.7: Pareto chart for standardized effects for the removal of SMO 114

Figure 4.8: Pareto chart for standardized effects for the removal of CO 115

Figure 4.9: Pareto chart for standardized effects for the removal of Bright-Edge 80 116

Figure 4.10: Main effects plot for the removal of SMO 117

Figure 4.11: Main effects plot for the removal of CO 118

Figure 4.12: Main effects plot for the removal of Bright-Edge 80 119

Figure 4.13: Interaction effects plot for the removal of SMO 121

Figure 4.14: Interaction effects plot for the removal of CO 122

Figure 4.15: Interaction effects plot for the removal of Bright-Edge 80 123

Figure 4.16: Normal probability plot of the residuals for removal of SMO 126

Figure 4.17: Normal probability plot of the residuals for removal of CO 127

Figure 4.18: Normal probability plot of the residuals for removal of Bright-Edge 80 128

List of Figures

Figure 3.1: Schematics of the experimental set up used for breakthrough studies 84

Figure 3.2: Schematics of the experimental set up used for breakdown studies 86

Figure 4.1: Zeta potential of autoclaved M. rouxii biomass and oil-in-water emulsions 91

Figure 4.2: Scanning electron micrographs 96

Figure 4.3: Plot of diameter versus % walnut shell media passing the related sieve 98

Figure 4.4: Adsorption capacity versus time for SMO 99

Figure 4.5: Adsorption capacity versus time for CO 100

Figure 4.6: Adsorption capacity versus time for Bright-Edge 80 101

Figure 4.7: Pareto chart for standardized effects for the removal of SMO 114

Figure 4.8: Pareto chart for standardized effects for the removal of CO 115

Figure 4.9: Pareto chart for standardized effects for the removal of Bright-Edge 80 116

Figure 4.10: Main effects plot for the removal of SMO 117

Figure 4.11: Main effects plot for the removal of CO 118

Figure 4.12: Main effects plot for the removal of Bright-Edge 80 119

Figure 4.13: Interaction effects plot for the removal of SMO 121

Figure 4.14: Interaction effects plot for the removal of CO 122

Figure 4.15: Interaction effects plot for the removal of Bright-Edge 80 123

Figure 4.16: Normal probability plot of the residuals for removal of SMO 126

Figure 4.17: Normal probability plot of the residuals for removal of CO 127

Figure 4.18: Normal probability plot of the residuals for removal of Bright-Edge 80 128

x

Figure 4.19: Effect of pH on biosorption of oils and zeta potential of autoclaved M rouxii

biomass and three oil-in-water emulsions 129

Figure 4.20: Percentage removal of oil at various initial concentrations 131

Figure 4.21: SMO concentration versus time for different temperatures 133

Figure 4.22: CO concentration versus time for different temperatures 134

Figure 4.23: Bright-Edge 80 concentration versus time for different temperatures 135

Figure 4.24: Rate of SMO biosorption predicted by Lagergren and Ho models at 22°C 140

Figure 4.25: Rate of CO biosorption predicted by Lagergren and Ho models at 22°C 141

Figure 4.26: Rate of Bright-Edge 80 biosorption predicted by Lagergren and Ho models

at 22°C 142

Figure 4.27: Desorption plot for M rouxii biomass using water as an eluent 150

Figure 4.28: FTIR spectra of M rouxii biomass before and after oil adsorption 151

Figure 4.29: FTIR spectra of biomass residue (B1) after methanol and hydrochloric acid

treatment before and after oil adsorption 155

Figure 4.30: FTIR spectra of biomass residue (B2) after formic acid and formaldehyde


Figure 4.31: FTIR spectra of biomass residue (B3) after nitromethane and

triethylphosphite treatment before and after oil adsorption 157

Figure 4.32: FTIR spectra of biomass residue (B4) after acetone treatment before and

after oil adsorption. 158

Figure 4.33: FTIR spectra of biomass residue (B5) after benzene treatment before and

after oil adsorption 159

xi

Figure 4.19: Effect of pH on biosorption of oils and zeta potential of autoclaved M. rouxii

biomass and three oil-in-water emulsions 129

Figure 4.20: Percentage removal of oil at various initial concentrations 131

Figure 4.21: SMO concentration versus time for different temperatures 133

Figure 4.22: CO concentration versus time for different temperatures 134

Figure 4.23: Bright-Edge 80 concentration versus time for different temperatures 135

Figure 4.24: Rate of SMO biosorption predicted by Lagergren and Ho models at 22°C 140

Figure 4.25: Rate of CO biosorption predicted by Lagergren and Ho models at 22°C 141

Figure 4.26: Rate of Bright-Edge 80 biosorption predicted by Lagergren and Ho models

at 22°C 142

Figure 4.27: Desorption plot for M. rouxii biomass using water as an eluent 150

Figure 4.28: FTIR spectra of M. rouxii biomass before and after oil adsorption 151

Figure 4.29: FTIR spectra of biomass residue (Bl) after methanol and hydrochloric acid

treatment before and after oil adsorption 15 5

Figure 4.30: FTIR spectra of biomass residue (B2) after formic acid and formaldehyde


Figure 4.31: FTIR spectra of biomass residue (B3) after nitromethane and

triethylphosphite treatment before and after oil adsorption 157

Figure 4.32: FTIR spectra of biomass residue (B4) after acetone treatment before and

after oil adsorption. 15 8

Figure 4.33: FTIR spectra of biomass residue (B5) after benzene treatment before and

after oil adsorption 15 9

XI

Figure 4.34: Diameter size versus percent of immobilized beads passing the sieve 164

Figure 4.35: Breakthrough curves for SMO predicted using Thomas model 167

Figure 4.36: Breakthrough curves for CO predicted using Thomas model 168

Figure 4.37: Breakthrough curves for Bright-Edge 80 predicted using Thomas model 169

Figure 4.38: Breakthrough curves for SMO predicted using Yan model 174

Figure 4.39: Breakthrough curves for CO predicted using Yan model 175

Figure 4.40: Breakthrough curves for Bright-Edge 80 predicted using Yan model 176

Figure 4.41: Breakthrough curves for SMO predicted using Belter and Chu models 177

Figure 4.42: Breakthrough curves for CO predicted using Belter and Chu models 178

Figure 4.43: Breakthrough curves for Bright-Edge 80 predicted using Belter and Chu

models 179

Figure 4.44: Breakthrough curves for SMO predicted using Yoon and Nelson model 182

Figure 4.45: Breakthrough curves for CO predicted using Yoon and Nelson model 183

Figure 4.46: Breakthrough curves for Bright-Edge 80 using Yoon and Nelson model 184

Figure 4.47: Desorption profile for SMO using de-ionized water 186

Figure 4.48: Desorption profile for CO using de-ionized water 187

Figure 4.49: Desorption profile for Bright-Edge 80 using de-ionized water 188

Figure 4.50: Breakthrough curve for SMO for the second run 190

Figure 4.51: Breakthrough curve for CO for the second run 191

Figure 4.52: Breakthrough curve for Bright-Edge 80 for the second run 192

Figure 4.53: Linearized plot of single-phase flow pressure drop 193

Figure 4.54: Predicted versus actual headloss for single-phase flow 196

xii

Figure 4.34: Diameter size versus percent of immobilized beads passing the sieve 164

Figure 4.35: Breakthrough curves for SMO predicted using Thomas model 167

Figure 4.36: Breakthrough curves for CO predicted using Thomas model 168

Figure 4.37: Breakthrough curves for Bright-Edge 80 predicted using Thomas model 169

Figure 4.38: Breakthrough curves for SMO predicted using Yan model 174

Figure 4.39: Breakthrough curves for CO predicted using Yan model 175

Figure 4.40: Breakthrough curves for Bright-Edge 80 predicted using Yan model 176

Figure 4.41: Breakthrough curves for SMO predicted using Belter and Chu models 177

Figure 4.42: Breakthrough curves for CO predicted using Belter and Chu models 178

Figure 4.43: Breakthrough curves for Bright-Edge 80 predicted using Belter and Chu

models 179

Figure 4.44: Breakthrough curves for SMO predicted using Yoon and Nelson model 182

Figure 4.45: Breakthrough curves for CO predicted using Yoon and Nelson model 183

Figure 4.46: Breakthrough curves for Bright-Edge 80 using Yoon and Nelson model 184

Figure 4.47: Desorption profile for SMO using de-ionized water 186

Figure 4.48: Desorption profile for CO using de-ionized water 187

Figure 4.49: Desorption profile for Bright-Edge 80 using de-ionized water 188

Figure 4.50: Breakthrough curve for SMO for the second run 190

Figure 4.51: Breakthrough curve for CO for the second run 191

Figure 4.52: Breakthrough curve for Bright-Edge 80 for the second run 192

Figure 4.53: Linearized plot of single-phase flow pressure drop 193


xii

Figure 4.55: Linearized plot of two-phase flow pressure drop

Figure 4.56: Predicted versus actual headloss for single-phase flow

Figure 4.57: Coalescence efficiency versus Reynolds number

197

198

202

Figure 4.58: Plot of the ratio of the drop diameter to the immobilized M rouxii biomass

bead diameter versus bed depth for various flow rates 205

Figure 4.59: Average holdup versus Reynolds number 206

Figure 4.60: Coalescence kinetics for SMO predicted by the Crickmore model 208

Figure A.1: Rate of SMO biosorption by Lagergren and Ho kinetic models at 5°C 266



Figure A.4: Rate of CO biosorption by Lagergren and Ho kinetic models at 5°C 267



Figure A.7: Rate of Bright-Edge 80 biosorption by Lagergren and Ho models at 5°C 269



Figure A.10: Rate of SMO biosorption by intra-particle diffusion model at 5°C 270




Figure A.14: Rate of CO biosorption by intra-particle diffusion model at 5°C 272


Figure 4.55 : Linearized plot of two-phase flow pressure drop 197


Figure 4.57: Coalescence efficiency versus Reynolds number 202

Figure 4.58: Plot of the ratio of the drop diameter to the immobilized M. rouxii biomass

bead diameter versus bed depth for various flow rates 205

Figure 4.59: Average holdup versus Reynolds number 206

Figure 4.60: Coalescence kinetics for SMO predicted by the Crickmore model 208

Figure A. 1: Rate of SMO biosorption by Lagergren and Ho kinetic models at 5°C 266









Figure A. 10: Rate of SMO biosorption by intra-particle diffusion model at 5°C 270




Figure A. 14: Rate of CO biosorption by intra-particle diffusion model at 5°C 272


Figure A.16: Rate of CO biosorption by infra-particle diffusion model at 22°C 273


Figure A.18: Rate of Bright-Edge 80 biosorption by intra-particle diffusion model at 5°C

274

Figure A.19: Rate of Bright-Edge 80 biosorption by intra-particle diffusion model at

15°C

Figure A.20: Rate of Bright-Edge 80 biosorption by infra-particle diffusion model at

22°C


30°C

275

275

276

Figure A.22: The Langmuir and Freundlich isotherm for biosorption of SMO at 5°C 276



Figure A.25: The Langmuir and Freundlich isotherm for biosorption of CO at 5°C 278



Figure A.28: The Langmuir and Freundlich isotherm model plots for biosorption of

Bright-Edge 80 at 5°C 279


Bright-Edge 80 at 22 °C 280



xiv



Figure A. 18: Rate of Bright-Edge 80 biosorption by intra-particle diffusion model at 5°C

274

Figure A. 19: Rate of Bright-Edge 80 biosorption by intra-particle diffusion model at

15°C 275


22°C 275


30°C 276








Bright-Edge 80 at 5°C 279





xiv

Figure A.31: Breakthrough curve predicted by Oulman model for SMO 281

Figure A.32: Breakthrough curve predicted by Oulman model for CO 281

Figure A.33: Breakthrough curve predicted by Oulman model for Bright-Edge 80 282

Figure A.34: Breakthrough curve predicted by Wolbroska model for SMO 282

Figure A.35: Breakthrough curve predicted by Wolbroska model for CO 283

Figure A.36: Breakthrough curve predicted by Wolbroska model for Bright-Edge 80 283

xv

Figure A.31: Breakthrough curve predicted by Oulman model for SMO 281

Figure A.32: Breakthrough curve predicted by Oulman model for CO 281

Figure A.33: Breakthrough curve predicted by Oulman model for Bright-Edge 80 282

Figure A.34: Breakthrough curve predicted by Wolbroska model for SMO 282

Figure A.35: Breakthrough curve predicted by Wolbroska model for CO 283

Figure A.36: Breakthrough curve predicted by Wolbroska model for Bright-Edge 80 283

xv

List of Tables

Table 1.1: Chitosan content in fungi 7

Table 2.1: Biological treatment 28

Table 2.2: Microbial degradation 40

Table 2.3: Oil sorption capacity of different media 43

Table 2.4: Design criteria for packed bed columns 53

Table 2.5: Design details of lab scale columns used in biosorption studies 54

Table 3.1: Oil in water measurement methods 72

Table 3.2: Coded and uncoded values of the factors 76

Table 4.1: Characteristics of oils used for the study at 20°C 90

Table 4.2: Characteristics of the solvent used in the measurement of oil 92

Table 4.3: Emulsion classification based on the diameter of oil droplets 93

Table 4.4: Characteristics of M rouxii biomass 95

Table 4.5: Characteristics of other adsorbents used in the preliminary study 97

Table 4.6: Residual oil concentration obtained using different biomaterials 102

Table 4.7: Oil removals by biomaterials 104

Table 4.8: Uncoded design table for the factors and response 108

Table 4.9: Estimated effects and coefficients for the removal of SMO (% coded units) 109

Table 4.10: Estimated effects and coefficients for the removal of CO (% coded units) 110

Table 4.11: Estimated effects and coefficients for the removal of Bright-Edge 80 (%

coded units) 111

xvi

List of Tables

Table 1.1: Chitosan content in fungi 7

Table 2.1: Biological treatment 28

Table 2.2: Microbial degradation 40

Table 2.3: Oil sorption capacity of different media 43

Table 2.4: Design criteria for packed bed columns 53

Table 2.5: Design details of lab scale columns used in biosorption studies 54

Table 3.1: Oil in water measurement methods 72

Table 3.2: Coded and uncoded values of the factors 76

Table 4.1: Characteristics of oils used for the study at 20°C 90

Table 4.2: Characteristics of the solvent used in the measurement of oil 92

Table 4.3: Emulsion classification based on the diameter of oil droplets 93

Table 4.4: Characteristics of M. rouxii biomass 95

Table 4.5: Characteristics of other adsorbents used in the preliminary study 97

Table 4.6: Residual oil concentration obtained using different biomaterials 102

Table 4.7: Oil removals by biomaterials 104

Table 4.8: Uncoded design table for the factors and response 108

Table 4.9: Estimated effects and coefficients for the removal of SMO (% coded units) 109

Table 4.10: Estimated effects and coefficients for the removal of CO (% coded units) 110

Table 4.11: Estimated effects and coefficients for the removal of Bright-Edge 80 (%

coded units) 111

xvi

Table 4.12: Comparison of oil removal efficiencies of M rouxii biomass and other

sorbents obtained in batch studies

Table 4.13: Parameters calculated using kinetic models

Table 4.14: Isotherm model constants

Table 4.15: Separation factor, RL based on the Langmuir equation

Table 4.16: Thermodynamic and activation parameter

137

138

145

146

148

Table 4.17: Functional groups of autoclaved biomass of M rouxii biomass, oil adsorbed

biomass and the corresponding infrared absorption wavelengths (Lin-Vien 1991) 152

Table 4.18: Oil removal by raw and chemically modified M rouxii biomass 160

Table 4.19: Characteristics of immobilized M rouxii biomass 165

Table 4.20: Parameters calculated using kinetic models 170

Table 4.21: Comparison of Thomas constants for other oil sorbents 172

Table 4.22: Summary of coalescence and filtration data 195

Table 4.23: Average saturation values of the immobilized M rouxii biomass bed 200

Table 4.24: Coalescence efficiency 201

Table 5.1: Cost of adsorbents required to treat 100 m3 per day of oily water based on

SMO data 212

Table 5.2: Design of batch adsorber system for M rouxii biomass and SMO with a flow

rate of 100 m3 per day 213

Table 5.3: Design of immobilized M rouxii filter for SMO with a flow rate of 100 m3 per

day 215

Table A.1: Batch kinetic data for SMO 248

xvii

Table 4.12: Comparison of oil removal efficiencies of M. rouxii biomass and other

sorbents obtained in batch studies 137


Table 4.14: Isotherm model constants 145

Table 4.15: Separation factor, RL based on the Langmuir equation 146

Table 4.16: Thermodynamic and activation parameter 148

Table 4.17: Functional groups of autoclaved biomass of M. rouxii biomass, oil adsorbed

biomass and the corresponding infrared absorption wavelengths (Lin-Vien 1991) 152

Table 4.18: Oil removal by raw and chemically modified M. rouxii biomass 160

Table 4.19: Characteristics of immobilized M. rouxii biomass 165


Table 4.21: Comparison of Thomas constants for other oil sorbents 172

Table 4.22: Summary of coalescence and filtration data 195

Table 4.23: Average saturation values of the immobilized M. rouxii biomass bed 200

Table 4.24: Coalescence efficiency 201

Table 5.1: Cost of adsorbents required to treat 100 m3 per day of oily water based on

SMOdata 212

Table 5.2: Design of batch adsorber system for M. rouxii biomass and SMO with a flow

rate of 100 m3 per day 213

Table 5.3: Design of immobilized M rouxii filter for SMO with a flow rate of 100 m3 per

day 215

Table A. 1: Batch kinetic data for SMO 248

xvii

Table A.2: Batch kinetic data for CO 249

Table A.3: Batch kinetic data for Bright-Edge 80 250

Table A.4: Batch isotherm data for SMO 250

Table A.5: Batch isotherm data for CO 251

Table A.6: Batch isotherm data for Bright-Edge 80 251

Table A.7: Desorption data for SMO 251

Table A.8: Desorption data for CO 252

Table A.9: Desorption data for Bright-Edge 80 252

Table A.10: Batch study with immobilized biomass beads for 6 h 253

Table A.11: Batch kinetic studies with immobilized biomass beads at pH 3.0 253

Table A.12: Column breakthrough data 254

Table A.13: Column desorption data for SMO 255

Table A.14: Column desorption data for CO 255

Table A.15: Column desorption data for Bright-Edge 80 256

Table A.16: Column breakthrough second run data for SMO 256

Table A.17: Column breakthrough second run data for CO 257

Table A.18: Column breakthrough second run data for Bright-Edge 80 257

Table A.19: Experimental and predicted head loss for single-phase flow 258

Table A.20: Experimental and predicted head loss for two-phase flow 259

Table A.21: Drop diameters (gm) for 12 mL/min 260

Table A.22: Drop density (no./ cm3) for 12 mL/min 260

Table A.23: Coalescence efficiency for 12 mL/min 261

xviii

Table A.2: Batch kinetic data for CO 249

Table A.3: Batch kinetic data for Bright-Edge 80 250

Table A.4: Batch isotherm data for SMO 250

Table A.5: Batch isotherm data for CO 251

Table A.6: Batch isotherm data for Bright-Edge 80 251

Table A.7: Desorption data for SMO 251

Table A.8: Desorption data for CO 252

Table A.9: Desorption data for Bright-Edge 80 252

Table A. 10: Batch study with immobilized biomass beads for 6 h 253

Table A. 11: Batch kinetic studies with immobilized biomass beads at pH 3.0 253

Table A. 12: Column breakthrough data 254

Table A. 13: Column desorption data for SMO 255

Table A. 14: Column desorption data for CO 255

Table A. 15: Column desorption data for Bright-Edge 80 256

Table A. 16: Column breakthrough second run data for SMO 256

Table A.17: Column breakthrough second run data for CO 257

Table A. 18: Column breakthrough second run data for Bright-Edge 80 257

Table A. 19: Experimental and predicted head loss for single-phase flow 258

Table A.20: Experimental and predicted head loss for two-phase flow 259

Table A.21: Drop diameters (|im) for 12 mL/min 260



xviii




Table A.27: Drop diameters (µm) for 20 mL/min 262






Table A.33: Drop diameters (µm) for 28 mL/min 264






Table A.39: Data used to fit Crickmore model 265

Table Bl: Design of batch adsorber system for M rouxii biomass and SMO with a flow

rate of 100 m3/d 284

Table B2: Design of batch adsorber system for M rouxii biomass and CO with a flow


Table B3: Design of batch adsorber system for M rouxii biomass and Bright-Edge 80

with a flow rate of 100 m3/d 285

xix

Table A.24: Drop diameters (urn) for 16 mL/min 261






Table A.30: Drop diameters (urn) for 24 mL/min 263






Table A.36: Drop diameters (jam) for 32 mL/min 265



Table A.39: Data used to fit Crickmore model 265

Table Bl: Design of batch adsorber system for M. rouxii biomass and SMO with a flow


Table B2: Design of batch adsorber system for M. rouxii biomass and CO with a flow


Table B3: Design of batch adsorber system for M. rouxii biomass and Bright-Edge 80


xix

Table B4: Design of immobilized M rouxii biomass column filter for SMO with a flow


Table B5: Design of immobilized M rouxii biomass column filter for CO with a flow rate

of 100 m3/d

Table B6: Design of immobilized M rouxii biomass column filter for Bright-Edge 80

with a flow rate of 100 m3/d

xx

287

288

Table B4: Design of immobilized M. rouxii biomass column filter for SMO with a flow


Table B5: Design of immobilized M. rouxii biomass column filter for CO with a flow rate

of 100 m3/d 287

Table B6: Design of immobilized M. rouxii biomass column filter for Bright-Edge 80


xx

List of Abbreviations, Symbols, Nomenclature

A Arrhenius factor

a Yan constant denoting the slope of the function.

b Langmuir constant

Bo Specific permeability coefficient

BOD Biological oxygen demand

C Concentration of solute in solution at equilibrium

C Effluent concentration

Co Influent concentration

CA Fraction of fluid emulsified

CA0 Fraction of fluid emulsified at entrance to the packed bed

Ce Equilibrium solute concentration

CO Canola oil

COD Chemical oxygen demand

Cs Concentration of solute in solution

d Throughput volume

DAF Dissolved air flotation

df Average diameter of immobilized biomass beads

di Average particle size of the distribution

Ead Activation energy for adsorption

EPA Environmental Protection Agency

erf(x) Error function of x

xxi

List of Abbreviations, Symbols, Nomenclature

A Arrhenius factor

a Yan constant denoting the slope of the function.

b Langmuir constant

Bo Specific permeability coefficient

BOD Biological oxygen demand

C Concentration of solute in solution at equilibrium

C Effluent concentration

Co Influent concentration

CA Fraction of fluid emulsified

CAO Fraction of fluid emulsified at entrance to the packed bed

Ce Equilibrium solute concentration

CO Canola oil

COD Chemical oxygen demand

Cs Concentration of solute in solution

d Throughput volume

DAF Dissolved air flotation

df Average diameter of immobilized biomass beads

dj Average particle size of the distribution

Ead Activation energy for adsorption

EPA Environmental Protection Agency

erf(x) Error function of x

xxi

FTIR Fourier transform infrared analysis

GAC Granular activated carbon

gc Acceleration due to gravity

K Rate constant of adsorption used in Arrhenius equation

K Adsorption rate coefficient

ko Shape factor

ki Lagergren rate constant for adsorption

k1 Carman-Kozeny constant for single-phase flow

k2 Pseudo-second order adsorption rate constant

k2 Carman-Kozeny constant for two-phase flow

kc Crickmore rate constant for coalescence

KF Freundlich equilibrium constant indicative of adsorption capacity

Intra-particle diffusion rate constant

kJ Kilo joule

Kr Thomas rate constant

KyN Yoon Nelson rate constant

L Bed length

m Mass of the adsorbent

n Freundlich adsorption equilibrium constant indicative of adsorption

intensity

N Adsorption capacity coefficient

Na Saturation concentration in the Wolborska model

FTIR Fourier transform infrared analysis

GAC Granular activated carbon

gc Acceleration due to gravity

K Rate constant of adsorption used in Arrhenius equation

K Adsorption rate coefficient

ko Shape factor

k\ Lagergren rate constant for adsorption

ki Carman-Kozeny constant for single-phase flow

ki Pseudo-second order adsorption rate constant

k2 Carman-Kozeny constant for two-phase flow

kc Crickmore rate constant for coalescence

Kp Freundlich equilibrium constant indicative of adsorption capacity

ki Intra-particle diffusion rate constant

kJ Kilo joule

Kj Thomas rate constant

KYN Yoon Nelson rate constant

L Bed length

m Mass of the adsorbent

n Freundlich adsorption equilibrium constant indicative of adsorption

intensity

N Adsorption capacity coefficient

N0 Saturation concentration in the Wolborska model

xxii

NRe Reynolds number

OMW Olive mill wastewater

POME Palm oil mill effluent

q Amount of adsorbate adsorbed per unit mass of adsorbent

Q Volumetric flow rate

Qo Langmuir constant

ge Amount of adsorbate adsorbed at equilibrium

qt Amount of adsorbate adsorbed at time t

R Universal gas constant

RBC Rotating biological contactor

RL Separation factor

Sd Average saturation

SEM Scanning electron microscope

SMO Standard mineral oil

t Residence time inside the column

T Tortuosity

to Temporal parameter

tin Time required for 50% adsorbate breakthrough

Te Absolute temperature

U Superficial velocity

UASB Upflow anaerobic sludge blanket digestion

UASFF Upflow anaerobic sludge-fixed film

Np_e Reynolds number

OMW Olive mill wastewater

POME Palm oil mill effluent

q Amount of adsorbate adsorbed per unit mass of adsorbent

Q Volumetric flow rate

Qo Langmuir constant

qe Amount of adsorbate adsorbed at equilibrium

q, Amount of adsorbate adsorbed at time t

R Universal gas constant

RBC Rotating biological contactor

RL Separation factor

Sa Average saturation

SEM Scanning electron microscope

SMO Standard mineral oil

t Residence time inside the column

T Tortuosity

to Temporal parameter

ti/2 Time required for 50% adsorbate breakthrough

Te Absolute temperature

U Superficial velocity

UASB Upflow anaerobic sludge blanket digestion

UASFF Upflow anaerobic sludge-fixed film

xxiii

V Throughput volume

VFA Volatile fatty acids

VORW Vegetable oil refinery wastewater

X„, Amount of solute adsorbed in forming a complete monolayer

Z Height of the column

AG° Gibbs free energy change

AFP Heat of adsorption or enthalpy change

API Pressure drop across the bed for single-phase flow

AP2 Pressure drop across the bed for two-phase flow

AS° Entropy change

13a Kinetic coefficient of the external mass transfer

Ye Dynamic viscosity of continuous phase

E Porosity of immobilized biomass bed in single-phase flow

Et Porosity of immobilized biomass bed in two-phase flow

ric Overall coalescence efficiency

a Standard deviation

t Residence time

'E' Modified residence time

(I)H Average holdup

xxiv

V Throughput volume

VFA Volatile fatty acids

VORW Vegetable oil refinery wastewater

Xm Amount of solute adsorbed in forming a complete monolayer

Z Height of the column

AG0 Gibbs free energy change

A/f° Heat of adsorption or enthalpy change

APi Pressure drop across the bed for single-phase flow

AP2 Pressure drop across the bed for two-phase flow

AS0 Entropy change

Pa Kinetic coefficient of the external mass transfer

yc Dynamic viscosity of continuous phase

e Porosity of immobilized biomass bed in single-phase flow

et Porosity of immobilized biomass bed in two-phase flow

r|c Overall coalescence efficiency

a Standard deviation

x Residence time

x' Modified residence time

4>H Average holdup

xxiv

Chapter 1

Introduction

1.1 Background

Rapid urbanization creates an annual increase in the discharge of oil-containing

wastewater into the environment. Oil and grease in wastewater constitute a complex,

heterogeneous matrix and their sources range from the hydrocarbons (petroleum based)

to fatty matter from animal and vegetable sources (Franson and Eaton 2005). Oils can

cause environmental pollution during various stages of production, transportation,

refining and utilization. Oils found in contaminated waters can be fats, lubricants, cutting

liquids, heavy hydrocarbons such as tars, grease, crude oils, diesel oil, and light

hydrocarbons such as kerosene, jet fuel and gasoline. A primary component of oil

contaminants are crudes and its derivatives. Principally, crudes include paraffin, olefin,

naphthene and aromatic hydrocarbons; oxygen, sulfur and nitrogen are present in the

form of compounds containing these elements (Pushkarev et al. 1983). Mineral oils

consist of mixtures of high molecular paraffins, naphthene and aromatic hydrocarbons

with a certain admixture of tar and asphaltene substances (Pushkarev et al. 1983). Light

mineral oil is a paraffin oil that may contain mixtures of alkanes in the range of C8 to

C15 carbon atoms. Cutting oils such as Bright-Edge 80 are made of 85 — 95%

hydrotreated naphthenic oil with chlorinated fatty esters and sulfitrized hydrocarbon.

Vegetable oils are essentially triglycerides consisting of straight chain fatty acids

attached, as esters, to glycerol (Wakelin and Forster 1997). The component fatty acids of

edible oil vary considerably and can differ in chain length, may be saturated or

1

Chapter 1

Introduction

1.1 Background

Rapid urbanization creates an annual increase in the discharge of oil-containing

wastewater into the environment. Oil and grease in wastewater constitute a complex,

heterogeneous matrix and their sources range from the hydrocarbons (petroleum based)

to fatty matter from animal and vegetable sources (Franson and Eaton 2005). Oils can

cause environmental pollution during various stages of production, transportation,

refining and utilization. Oils found in contaminated waters can be fats, lubricants, cutting

liquids, heavy hydrocarbons such as tars, grease, crude oils, diesel oil, and light

hydrocarbons such as kerosene, jet fuel and gasoline. A primary component of oil

contaminants are crudes and its derivatives. Principally, crudes include paraffin, olefin,

naphthene and aromatic hydrocarbons; oxygen, sulfur and nitrogen are present in the

form of compounds containing these elements (Pushkarev et al. 1983). Mineral oils

consist of mixtures of high molecular paraffins, naphthene and aromatic hydrocarbons

with a certain admixture of tar and asphaltene substances (Pushkarev et al. 1983). Light

mineral oil is a paraffin oil that may contain mixtures of alkanes in the range of C8 to

C15 carbon atoms. Cutting oils such as Bright-Edge 80 are made of 85 - 95%

hydrotreated naphthenic oil with chlorinated fatty esters and sulfurized hydrocarbon.

Vegetable oils are essentially triglycerides consisting of straight chain fatty acids

attached, as esters, to glycerol (Wakelin and Forster 1997). The component fatty acids of

edible oil vary considerably and can differ in chain length, may be saturated or

1

unsaturated, and may contain an odd or even number of carbon atoms. Canola oil

comprises of oleic, linoleic, linolenic and erucic acids. Palm oil has a fatty acid profile

including palmitic, linoleic, oleic and steric acids while olive oil has palmitoleic,

palmitic, linoleic and oleic acids (Ayorinde et al. 2007).

Oil can be characterized in three ways: by polarity, biodegradability and physical

characteristics. Polar oils normally are derived from animal and vegetable material such

as wastes from food processing operations. It is generally acknowledged that polar oils

are biodegradable and therefore, become part of the organic load that must be handled in

a biological treatment process. Non-polar oils are usually derived from petroleum or

mineral sources. The physical characteristics of oils are usually designated as being non-

floatable or dispersed (emulsified) and non-dispersed (Young 1979).

Most industrial wastewaters contain oil-in-water emulsions among their basic

contaminants. Oily wastewater may contain, in addition to oil, metal shavings, silt,

surfactants, cleansers, soaps, solvents and other residue. The most common and widely

used emulsifying agents for cutting oil in cold rolling operation at metallurgical plants

and cold cutting operations at metal working industry are soaps and sulfonates. The soaps

used contain fatty acids and oleic acid, stearic acid and palmitic acids are most widely

used. One of the most accepted emulsifiers for cutting oil is a mixture of both oleic acid

and amines (Biswas, 1973). Many emulsified oils can be removed from wastewater by

primary sedimentation, skimming or adsorption. Nevertheless, chemically or physically

stabilized oil-water emulsions should be managed in an appropriate manner. The

presence of emulsified oil in wastewaters is of serious concern as it often results in

2

unsaturated, and may contain an odd or even number of carbon atoms. Canola oil

comprises of oleic, linoleic, linolenic and erucic acids. Palm oil has a fatty acid profile

including palmitic, linoleic, oleic and steric acids while olive oil has palmitoleic,

palmitic, linoleic and oleic acids (Ayorinde et al. 2007).

Oil can be characterized in three ways: by polarity, biodegradability and physical

characteristics. Polar oils normally are derived from animal and vegetable material such

as wastes from food processing operations. It is generally acknowledged that polar oils

are biodegradable and therefore, become part of the organic load that must be handled in

a biological treatment process. Non-polar oils are usually derived from petroleum or

mineral sources. The physical characteristics of oils are usually designated as being non-

floatable or dispersed (emulsified) and non-dispersed (Young 1979).

Most industrial wastewaters contain oil-in-water emulsions among their basic

contaminants. Oily wastewater may contain, in addition to oil, metal shavings, silt,

surfactants, cleansers, soaps, solvents and other residue. The most common and widely

used emulsifying agents for cutting oil in cold rolling operation at metallurgical plants

and cold cutting operations at metal working industry are soaps and sulfonates. The soaps

used contain fatty acids and oleic acid, stearic acid and palmitic acids are most widely

used. One of the most accepted emulsifiers for cutting oil is a mixture of both oleic acid

and amines (Biswas, 1973). Many emulsified oils can be removed from wastewater by

primary sedimentation, skimming or adsorption. Nevertheless, chemically or physically

stabilized oil-water emulsions should be managed in an appropriate manner. The

presence of emulsified oil in wastewaters is of serious concern as it often results in

2

fouling of process equipment and creates problems during biological treatment of such

wastewaters. Oils that pass through physical-chemical processes contribute Biological

Oxygen Demand (BOD) and Chemical Oxygen Demand (COD) in effluents (Chao and

Yang 1981; Keenan and Sabelnikov 2000; Chang et al. 2001).

1.2 Sources and Quantity of Generated Oily Waters

Oil is considered to be a principal energy source. The average production of

Canadian crude oil in the year 2010 is estimated to be 474335 m3/day (IEA 2011). Global

oil demand in 2012 is expected to rise by 1.5 million barrels/day, year-on-year, up to 91.0

million barrels/day (IEA 2011). Petroleum products are used as raw materials in a wide

variety of industries, generating large quantities of hydrocarbon-containing oily

wastewater from various industrial sources.

Major industrial sources of oily wastewater include petroleum refineries, metal

• manufacturing and machining, and food processors. Plant and animal oils are handled in

industries such as olive oil and palm oil mills, butter, paints, polishes, detergent and soap

manufacturing units (Stams and Oude 1997). Industries such as slaughterhouses, dairies,

meatpacking and food processing operations are also known to produce oily wastewaters.

Sources of oil in municipal wastewater are kitchen and human wastes (Quemnuer and

Marty 1994) and constitute one of the major types of organic matter found in municipal

wastewater (Quemeneur and Marty 1994; Raunkjaer et al. 1994).

3

fouling of process equipment and creates problems during biological treatment of such

wastewaters. Oils that pass through physical-chemical processes contribute Biological

Oxygen Demand (BOD) and Chemical Oxygen Demand (COD) in effluents (Chao and

Yang 1981; Keenan and Sabelnikov 2000; Chang et al. 2001).

1.2 Sources and Quantity of Generated Oily Waters

Oil is considered to be a principal energy source. The average production of

Canadian crude oil in the year 2010 is estimated to be 474335 m3/day (IEA 2011). Global

oil demand in 2012 is expected to rise by 1.5 million barrels/day, year-on-year, up to 91.0

million barrels/day (IEA 2011). Petroleum products are used as raw materials in a wide

variety of industries, generating large quantities of hydrocarbon-containing oily

wastewater from various industrial sources.

Major industrial sources of oily wastewater include petroleum refineries, metal

manufacturing and machining, and food processors. Plant and animal oils are handled in

industries such as olive oil and palm oil mills, butter, paints, polishes, detergent and soap

manufacturing units (Stams and Oude 1997). Industries such as slaughterhouses, dairies,

meatpacking and food processing operations are also known to produce oily wastewaters.

Sources of oil in municipal wastewater are kitchen and human wastes (Quemnuer and

Marty 1994) and constitute one of the major types of organic matter found in municipal

wastewater (Quemeneur and Marty 1994; Raunkjaer et al. 1994).

3

1.3 Existing Treatment Technologies for Oily Water

The best available technologies for oil removal include chemical treatment,

gravity separation, parallel-plate coalescers, gas flotation, cyclone separation, granular

media filtration, membrane processes, biological processes and adsorption (Yang et al.

2002). Although many advanced technologies such as microfiltration and ultra filtration

(Lipp et al. 1998) have been developed and applied to oily water treatment, expensive

initial operating costs prohibit the wide application of such membrane filtration

technologies. Deep-bed filtration is an attractive method to separate immiscible liquid

from polluted wastewater. Many natural and synthetic media are available to treat oily

waters. The media can be classified as filtering media (sand, coal, and diatomaceous

earth), coalescing media (fiberglass, polypropylene) and adsorption media (activated

carbon and peat). Among several chemical and physical methods, adsorption process is

one of the most widely used methods in wastewater systems (Ahmad et al. 2005b).

1.4 Environmental Legislation and Concerns

Numerous standards and regulations exist regarding the discharge of oily waters.

Factors affecting the standards and regulations include the type of industry, the quantity

of waste generated, and the environmental significance of the discharge area. In Canada,

the Fisheries Act recommends a hydrocarbon discharge of less than 10 mg/L. The

Canadian Council of Ministers of the Environment does offer a recommended practice,

which requires storm water runoff to be treated at 15 mg/L or less (Environment Canada

1976). With regard to the shipping industry, Canadian marine regulations require the

4

1.3 Existing Treatment Technologies for Oily Water

The best available technologies for oil removal include chemical treatment,

gravity separation, parallel-plate coalescers, gas flotation, cyclone separation, granular

media filtration, membrane processes, biological processes and adsorption (Yang et al.

2002). Although many advanced technologies such as microfiltration and ultra filtration

(Lipp et al. 1998) have been developed and applied to oily water treatment, expensive

initial operating costs prohibit the wide application of such membrane filtration

technologies. Deep-bed filtration is an attractive method to separate immiscible liquid

from polluted wastewater. Many natural and synthetic media are available to treat oily

waters. The media can be classified as filtering media (sand, coal, and diatomaceous

earth), coalescing media (fiberglass, polypropylene) and adsorption media (activated

carbon and peat). Among several chemical and physical methods, adsorption process is

one of the most widely used methods in wastewater systems (Ahmad et al. 2005b).

1.4 Environmental Legislation and Concerns

Numerous standards and regulations exist regarding the discharge of oily waters.

Factors affecting the standards and regulations include the type of industry, the quantity

of waste generated, and the environmental significance of the discharge area. In Canada,

the Fisheries Act recommends a hydrocarbon discharge of less than 10 mg/L. The

Canadian Council of Ministers of the Environment does offer a recommended practice,

which requires storm water runoff to be treated at 15 mg/L or less (Environment Canada

1976). With regard to the shipping industry, Canadian marine regulations require the

4

discharges to meet 15 mg/L or less. However, the discharges from the inland waters of

Canada must meet 5 mg/L or less. In the United States, permits granted under the

National Pollutant Discharge Elimination System program are generally administered by

the various state environmental agencies under the supervision of the US Environmental

Protection Agency (EPA). Most states and localities require discharges to contain 15

mg/L or less oil and grease, based on a 24-hour composite sample. Oil and grease may

include petroleum hydrocarbons as well as animal and vegetable oils. The technology-

based oil and grease limit, established by the effluent limitations guidelines for

agriculture and wildlife, use sub-category produced water with a maximum concentration

of 35 mg/L. The offshore sub-category effluent guidelines limit oil and grease in

produced water discharges to an average of 29 mg/L and a maximum of 42 mg/L (Wilson

2007).

Over the past several years, various governments of the European community have

enacted different legal requirements regarding the discharge of oil in water but the

European Committee for Standardization (CEN) has been working on a unified standard

for separator systems for oil and petrol (European Committee for Standardization 2002).

An effluent quality of 5 mg/L or less is required for separators for the processing of

rainwater where the discharge is released into surface water. An effluent quality of 100

mg/L or less is required for separators when processing industrial streams or rainwater

into sewer systems.

5

discharges to meet 15 mg/L or less. However, the discharges from the inland waters of

Canada must meet 5 mg/L or less. In the United States, permits granted under the

National Pollutant Discharge Elimination System program are generally administered by

the various state environmental agencies under the supervision of the US Environmental

Protection Agency (EPA). Most states and localities require discharges to contain 15

mg/L or less oil and grease, based on a 24-hour composite sample. Oil and grease may

include petroleum hydrocarbons as well as animal and vegetable oils. The technology-

based oil and grease limit, established by the effluent limitations guidelines for

agriculture and wildlife, use sub-category produced water with a maximum concentration

of 35 mg/L. The offshore sub-category effluent guidelines limit oil and grease in

produced water discharges to an average of 29 mg/L and a maximum of 42 mg/L (Wilson

2007).

Over the past several years, various governments of the European community have

enacted different legal requirements regarding the discharge of oil in water but the

European Committee for Standardization (CEN) has been working on a unified standard

for separator systems for oil and petrol (European Committee for Standardization 2002).

An effluent quality of 5 mg/L or less is required for separators for the processing of

rainwater where the discharge is released into surface water. An effluent quality of 100

mg/L or less is required for separators when processing industrial streams or rainwater

into sewer systems.

5

1.5 Rationale for using M. rouxii Biomass as an Adsorbent for Oil Removal

The use of biomaterials or biomass such as bacteria, fungi, yeast or even some

plant biomass to adsorb organic substances has been limited (Volesky 2003). Biosorption

is the term used to describe this phenomenon, which is based on passive sequestration by

a non-metabolizing, non-living biomass, while the term "bioaccumulation" is used to

refer to the active, metabolically mediated transport and deposition of chemical species.

Numerous studies exist which show that the fungal biomass of the order of Mucorales

represents a good biosorbent material (Volesky and Holan 1995). The biosorption

capacity of fungi is associated with their cell wall structure, which contains mainly chitin

and chitosan.

It has been observed that chitosan from crab shells has been effective in removing

oil from vegetable oil industry wastewater (Ahmad et al. 2005a). Chitosan is a natural

deacetylated polysaccharide, which has been used in a variety of practical fields

including wastewater treatment (Ravi Kumar 2000). It is therefore considered likely that

fungi, having chitosan in their cell wall, may prove to be an excellent sorbent to remove

oil from wastewaters. Table 1.1 provides the chitosan content in the cell wall of various

fungi.

Mucor rouxii and Absidia coerulea cell walls have a high chitosan content (Table

1.1), hence they may be good candidates for this study. The biomass of fungus M rouxii

is a type of biomaterial, which has been used in many applications of separation

technology. M rouxii is a filamentous fungus in which chitosan is the most abundant

component (33%) of the cell wall.

6

1.5 Rationale for using M. rouxii Biomass as an Adsorbent for Oil Removal

The use of biomaterials or biomass such as bacteria, fungi, yeast or even some

plant biomass to adsorb organic substances has been limited (Volesky 2003). Biosorption

is the term used to describe this phenomenon, which is based on passive sequestration by

a non-metabolizing, non-living biomass, while the term "bioaccumulation" is used to

refer to the active, metabolically mediated transport and deposition of chemical species.

Numerous studies exist which show that the fungal biomass of the order of Mucorales

represents a good biosorbent material (Volesky and Holan 1995). The biosorption

capacity of fungi is associated with their cell wall structure, which contains mainly chitin

and chitosan.

It has been observed that chitosan from crab shells has been effective in removing

oil from vegetable oil industry wastewater (Ahmad et al. 2005a). Chitosan is a natural

deacetylated polysaccharide, which has been used in a variety of practical fields

including wastewater treatment (Ravi Kumar 2000). It is therefore considered likely that

fungi, having chitosan in their cell wall, may prove to be an excellent sorbent to remove

oil from wastewaters. Table 1.1 provides the chitosan content in the cell wall of various

fungi.

Mucor rouxii and Absidia coerulea cell walls have a high chitosan content (Table

1.1), hence they may be good candidates for this study. The biomass of fungus M. rouxii

is a type of biomaterial, which has been used in many applications of separation

technology. M. rouxii is a filamentous fungus in which chitosan is the most abundant

component (33%) of the cell wall.

6

Table 1.1: Chitosan content in fungi

Fungi Chitosan Yield Reference

Absidia coerulea 16.9% dry biomass Jawarska (2003)

10.4% /weight of dry fungal cells Miyoshi et al. (1992)

Absidia repens 15.2% of dry biomass Jawarska (2003)

Cunninghamella enchinulata

7% per mycelia dry weight

9.4% /weight of dry fungal cells

Tan et al. (1996)

Miyoshi et al. (1992)

Cunninghamella elegans

2% dry cell weight Amorim et al. (2001)

Gongronella butleri 20% of dry biomass Jawarska (2003)

Mucor racemosus 3.5% dry cell weight Amorim et al. (2000)

Mucor rouxianus 18.1% of dry biomass Jawarska (2003)

Mucor rouxii 30 — 40% of cell wall Arcdiacono and Kaplan (2004)

30 - 35% cell wall Arcdiacono et al. (1988)

33% of cell wall Bartnicki-Garcia and Nickerson (1962)

7.4 — 23.3% Knorr and Klein (1986)

20% of dry biomass Jawarska (2003)

12.5% of dry weight of mycelium Wu et al. (2005)

4 — 8 % of dry cell wall material White et al. (1979)

7.3% on dry basis in mycelia Synowiecki and Al-Khateeb (1997)

Phycomyces blakesleeanus

9.0 —12.2% Knorr and Klein (1986)

Rhizopus oryzae Up to 9.1% Yoshihara et al. (2003)

Rhizopus delamer 7.1% /weight of dry fungal cells Miyoshi et al. (1992)

7

Table 1.1: Chitosan content in fungi

Fungi Chitosan Yield Reference

Absidia coerulea 16.9% dry biomass Jawarska (2003)

10.4% /weight of dry fungal cells Miyoshi et al. (1992)

Absidia repens 15.2% of dry biomass Jawarska (2003)

Cunninghamella enchinulata

7% per mycelia dry weight

9.4% /weight of dry fungal cells

Tan etal. (1996)

Miyoshi et al. (1992)

Cunninghamella elegans

2% dry cell weight Amorim et al. (2001)

Gongronella butleri 20% of dry biomass Jawarska (2003)

Mucor racemosus 3.5% dry cell weight Amorim et al. (2000)

Mucor rouxianus 18.1% of dry biomass Jawarska (2003)

Mucor rouxii 30 - 40% of cell wall Arcdiacono and Kaplan (2004)

30 - 35% cell wall Arcdiacono et al. (1988)

33% of cell wall Bartnicki-Garcia and Nickerson (1962)

7.4-23.3% Knorr and Klein (1986)

20% of dry biomass Jawarska (2003)

12.5% of dry weight of mycelium Wu et al. (2005)

4 - 8 % of dry cell wall material White et al. (1979)

7.3% on dry basis in mycelia Synowiecki and Al-Khateeb (1997)

Phycomyces blakesleeanus

9.0-12.2% Knorr and Klein (1986)

Rhizopus oryzae Up to 9.1% Yoshihara et al. (2003)

Rhizopus delamer 7.1% /weight of dry fungal cells Miyoshi et al. (1992)

7

During the life cycle of M rouxii, a chemical defferenciation of cell walls takes

place, where it is evident that the highest proportion of chitosan is present during the

hyphae stage (Bartnicki-Garcia and Nickerson 1962). M rouxii contains large quantities

of chitosan, chitin, and phosphate on its cell wall (Bartnicki-Garcia and Nickerson 1962;

Muzzarelli 1977). Large quantities of positively charged chitosan and negatively charged

phosphate, and glucouronic acid on the cell wall of M. rouxii, have been found to present

extensive possibilities in binding heavy metals (Volesky 1993).

So far, work has not been conducted on the removal of oil from water by non-

viable fungal adsorbents, although, a few studies have been conducted regarding the

uptake of oil by live fungi (Srinivasan and Viraraghavan 2007). Therefore, this study

focuses on using Mucor rouxii biomass as a biosorbent for the removal of different types

of oil from water.

1.6 Objectives and Scope of this Study

The objectives of this research are as follows:

1. to evaluate the potential of M rouxii in the removal of oil from water;

2. to examine the major factors which affect the biosorption of oil by M rouxii

biomass as per set experimental conditions using a factorial design of

experiments;

3. to identify optimum conditions for the biosorption of oil by M rouxii biomass;

to identify the mechanisms involved in biosorption of oil by the M rouxii

biomass;

8

During the life cycle of M. rouxii, a chemical defferenciation of cell walls takes

place, where it is evident that the highest proportion of chitosan is present during the

hyphae stage (Bartnicki-Garcia and Nickerson 1962). M. rouxii contains large quantities

of chitosan, chitin, and phosphate on its cell wall (Bartnicki-Garcia and Nickerson 1962;

Muzzarelli 1977). Large quantities of positively charged chitosan and negatively charged

phosphate, and glucouronic acid on the cell wall of M. rouxii, have been found to present

extensive possibilities in binding heavy metals (Volesky 1993).

So far, work has not been conducted on the removal of oil from water by non

viable fungal adsorbents, although, a few studies have been conducted regarding the

uptake of oil by live fungi (Srinivasan and Viraraghavan 2007). Therefore, this study

focuses on using Mucor rouxii biomass as a biosorbent for the removal of different types

of oil from water.

1.6 Objectives and Scope of this Study

The objectives of this research are as follows:

1. to evaluate the potential of M. rouxii in the removal of oil from water;

2. to examine the major factors which affect the biosorption of oil by M. rouxii

biomass as per set experimental conditions using a factorial design of

experiments;

3. to identify optimum conditions for the biosorption of oil by M. rouxii biomass;

to identify the mechanisms involved in biosorption of oil by the M. rouxii

biomass;

8

4. to examine the removal of oil, under dynamic conditions, using an immobilized

biomass;

5. to study the regeneration of adsorbent; and

6. to identify various breakdown mechanisms involved in the removal of oil by M

rouxii biomass.

The scope of the study includes the following:

1. a review of literature regarding the removal of oil by various treatment methods,

properties and the classification of emulsions and various breakdown mechanisms

involved in the oil removal process;

2. a preliminary evaluation of the oil removal capacity of M rouxii biomass in

comparison with chitosan and walnut shell media;

3. characterization of the M rouxii biomass and various oil-in-water emulsions;

4. factorial design of experiments in order to identify major factors affecting the

biosorption of oil by M rouxii biomass;

5. batch adsorption kinetics and isotherm studies;

6. regeneration of adsorbed oil from the M rouxii biomass using de-ionized water;

7. batch experiments to identify mechanisms involved in the removal of oil by M

rouxii biomass;

8. Fourier transform infrared spectroscopy analysis of M rouxii biomass, before and

after oil adsorption, in order to understand the mechanisms involved in

biosorption;

9

4. to examine the removal of oil, under dynamic conditions, using an immobilized

biomass;

5. to study the regeneration of adsorbent; and

6. to identify various breakdown mechanisms involved in the removal of oil by M.

rouxii biomass.

The scope of the study includes the following:

1. a review of literature regarding the removal of oil by various treatment methods,

properties and the classification of emulsions and various breakdown mechanisms

involved in the oil removal process;

2. a preliminary evaluation of the oil removal capacity of M. rouxii biomass in

comparison with chitosan and walnut shell media;

3. characterization of the M. rouxii biomass and various oil-in-water emulsions;

4. factorial design of experiments in order to identify major factors affecting the

biosorption of oil by M. rouxii biomass;

5. batch adsorption kinetics and isotherm studies;

6. regeneration of adsorbed oil from the M. rouxii biomass using de-ionized water;

7. batch experiments to identify mechanisms involved in the removal of oil by M.

rouxii biomass;

8. Fourier transform infrared spectroscopy analysis of M. rouxii biomass, before and

after oil adsorption, in order to understand the mechanisms involved in

biosorption;

9

9. immobilization of the M rouxii biomass, in the form of spherical beads onto a

polymer matrix, and use of the beads in batch and column studies for the removal

of oil from water;

10. continuous column breakthrough studies;

11. elution of biosorbed oil from the immobilized biomass beads and the regeneration

and reuse of the bead in column studies; and

12. column experiments to identify various breakdown mechanisms such as filtration

and coalescence that may occur in the M rouxii biomass bed.

10

9. immobilization of the M. rouxii biomass, in the form of spherical beads onto a

polymer matrix, and use of the beads in batch and column studies for the removal

of oil from water;

10. continuous column breakthrough studies;

11. elution of biosorbed oil from the immobilized biomass beads and the regeneration

and reuse of the bead in column studies; and

12. column experiments to identify various breakdown mechanisms such as filtration

and coalescence that may occur in the M. rouxii biomass bed.

10

Chapter 2

Literature Review

2.1 Emulsions

2.1.1 Classification of Emulsions

An emulsion can be defined as a heterogeneous system consisting of one or more

immiscible liquids intimately dispersed within one another in droplet form. The material

within the emulsion droplets is usually referred to as "the dispersed phase" and the

material that makes up the surrounding liquid is usually referred to as "the continuous

phase" (Williams 2007). Typically, the range of diameter of the emulsion droplets lies

between 0.1 to 100 gm. Emulsions are a special type of colloidal systems in which the

droplets often exceed the size limit of the colloidal system (1 gm) (Schramm 1992).

Emulsions may be classified depending on the organization of the oil and aqueous phase

as: 1) a system consisting of oil droplets dispersed in an aqueous phase which is called

"oil-in-water (0/W) emulsions", and 2) a system consisting of water droplets dispersed in

an oil phase called "a water-in-oil (W/0) emulsion" (Williams 2007).

As a result of hydrophobicity, the contact between oil and water in the emulsions

is thermodynamically unfavorable. The stability of emulsions may be improved with the

addition of surfactants or emulsifiers (Becher 1977). Emulsifiers are surface—active

molecules that adsorb droplets in the emulsions at the surface (Williams 2007). The type

of oil, its concentration, and the hardness and the temperature of water affect the stability

of an oil-in-water emulsion. The stability of emulsions is determined by the surface

11

Chapter 2

Literature Review

2.1 Emulsions

2.1.1 Classification of Emulsions

An emulsion can be defined as a heterogeneous system consisting of one or more

immiscible liquids intimately dispersed within one another in droplet form. The material

within the emulsion droplets is usually referred to as "the dispersed phase" and the

material that makes up the surrounding liquid is usually referred to as "the continuous

phase" (Williams 2007). Typically, the range of diameter of the emulsion droplets lies

between 0.1 to 100 fim. Emulsions are a special type of colloidal systems in which the

droplets often exceed the size limit of the colloidal system (1 (im) (Schramm 1992).

Emulsions may be classified depending on the organization of the oil and aqueous phase

as: 1) a system consisting of oil droplets dispersed in an aqueous phase which is called

"oil-in-water (O/W) emulsions", and 2) a system consisting of water droplets dispersed in

an oil phase called "a water-in-oil (W/O) emulsion" (Williams 2007).

As a result of hydrophobicity, the contact between oil and water in the emulsions

is thermodynamically unfavorable. The stability of emulsions may be improved with the

addition of surfactants or emulsifiers (Becher 1977). Emulsifiers are surface-active

molecules that adsorb droplets in the emulsions at the surface (Williams 2007). The type

of oil, its concentration, and the hardness and the temperature of water affect the stability

of an oil-in-water emulsion. The stability of emulsions is determined by the surface

11

tension, the double electrical layers over the surfaces and the size of particles. Emulsifiers

are added to decrease surface tension and change the potential at the double layer.

Adsorption of emulsifier molecules that contain ionized or ionizable groups on the

surface of droplets changes the electrical charge of emulsion droplets. The double layer

makes the emulsion stable, preventing coagulation of the particles (Pushkarev 1983).

2.1.2 Physical Characteristics of Emulsions

Most oil-in-water emulsions may appear to have a watery, creamy, viscous and

oily texture. The majority of emulsions exhibit a milky opaqueness in their appearance.

When a light beam is incident upon the surface of an emulsion, a portion of the incident

light beam is transmitted through the emulsion while another portion is reflected.

Transmission and absorption of light by an emulsion depends upon size, concentration,

refractive index, and the presence of any chromophoric material in the emulsion

(Williams 2007). An oil-in-water emulsion is colored by dyes that are soluble in water.

Hence, the overall appearance of an emulsion is influenced by its structure and

composition. The color intensity of an emulsion may decrease with an increasing droplet

concentration. Generally, an oil-in-water emulsion has a high specific conductance

(Schramm 1992). Phase inversion may occur when mixed with additional oil.

12

tension, the double electrical layers over the surfaces and the size of particles. Emulsifiers

are added to decrease surface tension and change the potential at the double layer.

Adsorption of emulsifier molecules that contain ionized or ionizable groups on the

surface of droplets changes the electrical charge of emulsion droplets. The double layer

makes the emulsion stable, preventing coagulation of the particles (Pushkarev 1983).

2.1.2 Physical Characteristics of Emulsions

Most oil-in-water emulsions may appear to have a watery, creamy, viscous and

oily texture. The majority of emulsions exhibit a milky opaqueness in their appearance.

When a light beam is incident upon the surface of an emulsion, a portion of the incident

light beam is transmitted through the emulsion while another portion is reflected.

Transmission and absorption of light by an emulsion depends upon size, concentration,

refractive index, and the presence of any chromophoric material in the emulsion

(Williams 2007). An oil-in-water emulsion is colored by dyes that are soluble in water.

Hence, the overall appearance of an emulsion is influenced by its structure and

composition. The color intensity of an emulsion may decrease with an increasing droplet

concentration. Generally, an oil-in-water emulsion has a high specific conductance

(Schramm 1992). Phase inversion may occur when mixed with additional oil.

12

2.1.3 Emulsion Droplet Characteristics

Important physicochemical and functional properties of emulsions are governed

by the presence and characteristics of the droplets. The emulsifier type, its concentration

and the homogenization conditions used in the preparation of an emulsion largely

determine the size of the droplets in an emulsion (Pushkarev 1983). An emulsion that

contains droplets of the same size is denoted as monodisperse, and droplets having a

range of different sizes are referred to as "polydisperse" (Williams 2007). Droplets in a

polydisperse emulsion are represented either by full particle size distribution, or in most

cases, by the mean size of the emulsion droplets. Manipulating their electrical charge can

control the physicochemical properties of emulsions. The electrical charge of emulsion

droplets depends upon the type and concentration of surface-active molecules present at

the interface and the pH and ionic concentration of the aqueous phase. The droplet's

electrical properties are characterized by the surface charge density (a) and the electrical

surface potential (410) (Williams 2007). When the electrical charge of the emulsion

droplets is sufficiently large, the droplets are prevented from aggregating due to

electrostatic repulsion between them. By changing the pH of the aqueous phase, the

emulsifier can be made to lose its charge resulting in weak repulsive forces that may lead

to droplet aggregation. Thus, choosing emulsifiers with desirable charge properties and

controlling the aqueous phase properties can manipulate the electrical charge on an

emulsion droplet.

The droplet interfacial region plays a major role in determining the properties of

the emulsion. The droplet interface is comprised of a narrow region (2 — 20 urn thick) that

13

2.1.3 Emulsion Droplet Characteristics

Important physicochemical and functional properties of emulsions are governed

by the presence and characteristics of the droplets. The emulsifier type, its concentration

and the homogenization conditions used in the preparation of an emulsion largely

determine the size of the droplets in an emulsion (Pushkarev 1983). An emulsion that

contains droplets of the same size is denoted as monodisperse, and droplets having a

range of different sizes are referred to as "polydisperse" (Williams 2007). Droplets in a

polydisperse emulsion are represented either by full particle size distribution, or in most

cases, by the mean size of the emulsion droplets. Manipulating their electrical charge can

control the physicochemical properties of emulsions. The electrical charge of emulsion

droplets depends upon the type and concentration of surface-active molecules present at

the interface and the pH and ionic concentration of the aqueous phase. The droplet's

electrical properties are characterized by the surface charge density (a) and the electrical

surface potential (\|/o) (Williams 2007). When the electrical charge of the emulsion

droplets is sufficiently large, the droplets are prevented from aggregating due to

electrostatic repulsion between them. By changing the pH of the aqueous phase, the

emulsifier can be made to lose its charge resulting in weak repulsive forces that may lead

to droplet aggregation. Thus, choosing emulsifiers with desirable charge properties and

controlling the aqueous phase properties can manipulate the electrical charge on an

emulsion droplet.

The droplet ihterfacial region plays a major role in determining the properties of

the emulsion. The droplet interface is comprised of a narrow region (2 - 20 nm thick) that

surrounds each emulsion droplet (Williams 2007). The composition and structure of the

interfacial region are determined by the type and concentration of surface-active species

present. The physicochemical properties and stability of emulsions depend upon the

extent of droplet aggregation and the characteristics of any aggregates formed. The

interactions between two emulsion droplets can be described in terms of an inter-droplet

pair potential. The overall inter-droplet pair potential acting between two droplets is the

result of different interactions, such as van der Waals, steric, electrostatic, depletion,

hydrophobic and hydration interactions (Williams 2007). Generally, droplets tend to

aggregate when attractive interactions dominate but droplets would remain as individual

entities when repulsive forces dominate.

2.1.4 Emulsion Stability

The term "emulsion stability" is broadly used to describe the ability of an

emulsion to resist changes in its properties, over time. The stability of an emulsion can be

regarded as a situation within the system wherein the suspended droplets do not settle out

or float rapidly, and the droplets do not coalesce quickly (Schramm 1992). Most

importantly, physical mechanisms that may work against the stability of an emulsion are:

a) creaming; b) sedimentation; c) aggregation; and d) coalescence.

Creaming is the upward movement of droplets resulting from a density difference

between the two liquid phases where droplets have less of a density than in the aqueous

phase. Sedimentation is the opposite of creaming, resulting in a downward movement of

droplets. Droplets in emulsions are in constant motion due to the effects of thermal 14

surrounds each emulsion droplet (Williams 2007). The composition and structure of the

interfacial region are determined by the type and concentration of surface-active species

present. The physicochemical properties and stability of emulsions depend upon the

extent of droplet aggregation and the characteristics of any aggregates formed. The

interactions between two emulsion droplets can be described in terms of an inter-droplet

pair potential. The overall inter-droplet pair potential acting between two droplets is the

result of different interactions, such as van der Waals, steric, electrostatic, depletion,

hydrophobic and hydration interactions (Williams 2007). Generally, droplets tend to

aggregate when attractive interactions dominate but droplets would remain as individual

entities when repulsive forces dominate.

2.1.4 Emulsion Stability

The term "emulsion stability" is broadly used to describe the ability of an

emulsion to resist changes in its properties, over time. The stability of an emulsion can be

regarded as a situation within the system wherein the suspended droplets do not settle out

or float rapidly, and the droplets do not coalesce quickly (Schramm 1992). Most

importantly, physical mechanisms that may work against the stability of an emulsion are:

a) creaming; b) sedimentation; c) aggregation; and d) coalescence.

Creaming is the upward movement of droplets resulting from a density difference

between the two liquid phases where droplets have less of a density than in the aqueous

phase. Sedimentation is the opposite of creaming, resulting in a downward movement of

droplets. Droplets in emulsions are in constant motion due to the effects of thermal

14

energy, gravity or applied mechanical force, as a result of which two or more droplets

frequently bump together. After the collisions, droplets may either move apart or remain

aggregated depending upon the dominance of attractive and repulsive forces between

them. Aggregation is sometimes referred to as "flocculation".

Coalescence is a process whereby two or more liquid droplets merge together to

form a single layer droplet with a reduced surface area. Coalescence causes emulsion

droplets to cream or sediment due to an increased size of the droplets (Williams 2007).

The rate at which coalescence proceed and the physical mechanism by which it occurs is

dependent upon the nature of the emulsifier used to stabilize the system.

The stability of an emulsion can be described in making reference to the drop size

distribution of the dispersed phase. Based on the drop diameter of oil, oil-in-water

emulsions can be classified as follows (Tabakin et al. 1978a,b):

1) Free oil: the oil droplet size is greater than 150 gm. In free oils, oil droplets rise

rapidly to the surface under quiescent conditions;

2) Mechanical dispersion: fine droplets range in size from 150 to 20 gm. Mechanical

dispersions are stabilized by forces such as electrical charges and not through the

influence of surface active agents;

3) Chemically stabilized emulsions: droplets range in size from 20 to 5 gm. Oil

droplets in chemically stabilized emulsions are similar to mechanical dispersions

but have enhanced stability resulting from surface active agents at the oil-water

interface;

15

energy, gravity or applied mechanical force, as a result of which two or more droplets

frequently bump together. After the collisions, droplets may either move apart or remain

aggregated depending upon the dominance of attractive and repulsive forces between

them. Aggregation is sometimes referred to as "flocculation".

Coalescence is a process whereby two or more liquid droplets merge together to

form a single layer droplet with a reduced surface area. Coalescence causes emulsion

droplets to cream or sediment due to an increased size of the droplets (Williams 2007).

The rate at which coalescence proceed and the physical mechanism by which it occurs is

dependent upon the nature of the emulsifier used to stabilize the system.

The stability of an emulsion can be described in making reference to the drop size

distribution of the dispersed phase. Based on the drop diameter of oil, oil-in-water

emulsions can be classified as follows (Tabakin et al. 1978a,b):

1) Free oil: the oil droplet size is greater than 150 (im. In free oils, oil droplets rise

rapidly to the surface under quiescent conditions;

2) Mechanical dispersion: fine droplets range in size from 150 to 20 nm. Mechanical

dispersions are stabilized by forces such as electrical charges and not through the

influence of surface active agents;

3) Chemically stabilized emulsions: droplets range in size from 20 to 5 nm. Oil

droplets in chemically stabilized emulsions are similar to mechanical dispersions

but have enhanced stability resulting from surface active agents at the oil-water

interface;

15

4) Dissolved oil: the droplet size is typically less than 5 pm and are truly soluble

species in the chemical sense;

5) Micro emulsions: very small droplet sizes of < 0.01 1.tm. These are systems with

low interfacial tension that allow spontaneous or nearly spontaneous

emulsification and are thought to be thermodynamically stable.

2.2 Removal of Oil from Water

2.2.1 Dissolved Air Flotation

Dissolved Air Flotation (DAF) is a process for removing suspended particles from

liquid by bringing the particles to the surface of the liquid. Air is dissolved, at high

pressure in a saturator, and micro bubbles are formed when water is released into the

flotation cell at atmospheric pressure. The micro bubbles become attached to the

particles, increasing their buoyancy and allowing them to rise to the surface. Volesky and

Agathos (1974) critically reviewed air floatation processes for oil refinery effluent

treatment and concluded that removal efficiencies of 65% to more than 90% can be

achieved for oil and suspended solids. DAF, in combination with flocculation, can reduce

oil content in refinery wastewater to levels approaching oil solubility (API 1969). DAF

equipment, generally used for the treatment of refinery waste, would include packaged

units of steel construction with capacities of up to 2000 gpm (Wang et al. 2004). A

relatively new design of a high rate DAF unit uses a shallow bed system with only a 3

minute retention time and operates at an overflow rate of 3.5 gpm/sq. ft. This unit has

16

4) Dissolved oil: the droplet size is typically less than 5 nm and are truly soluble

species in the chemical sense;

5) Micro emulsions: very small droplet sizes of < 0.01 (im. These are systems with

low interfacial tension that allow spontaneous or nearly spontaneous

emulsification and are thought to be thermodynamically stable.

2.2 Removal of Oil from Water

2.2.1 Dissolved Air Flotation

Dissolved Air Flotation (DAF) is a process for removing suspended particles from

liquid by bringing the particles to the surface of the liquid. Air is dissolved, at high

pressure in a saturator, and micro bubbles are formed when water is released into the

flotation cell at atmospheric pressure. The micro bubbles become attached to the

particles, increasing their buoyancy and allowing them to rise to the surface. Volesky and

Agathos (1974) critically reviewed air floatation processes for oil refinery effluent

treatment and concluded that removal efficiencies of 65% to more than 90% can be

achieved for oil and suspended solids. DAF, in combination with flocculation, can reduce

oil content in refinery wastewater to levels approaching oil solubility (API 1969). DAF

equipment, generally used for the treatment of refinery waste, would include packaged

units of steel construction with capacities of up to 2000 gpm (Wang et al. 2004). A

relatively new design of a high rate DAF unit uses a shallow bed system with only a 3

minute retention time and operates at an overflow rate of 3.5 gpm/sq. ft. This unit has

16

been used in petrochemical complexes as a secondary clarifier to improve the operation

and capacity of an existing activated sludge system (Wang et al. 2004). Moursi and Abo-

Elela (1982) evaluated the use of chemical coagulation, followed by dissolved air

flotation as a treatment process for the primary effluent of refinery wastewater.

Considerable reductions of 94.4%, 95%, 95%, 85%, and 96.5%, respectively in turbidity,

COD, total oil and grease, phenols, and total oils were obtained. The data showed that

when wastewater was chemically pretreated to break the oil emulsion, dissolved air

flotation units were capable of removing most of the emulsified oil in addition to the

original free oil content. Removal rate of hydrocarbon compounds was generally found to

increase as the solubility of compounds decreased. The results obtained following the

treatment of refinery wastewater showed some of the hydrocarbon compounds were

completely removed.

Galil and Wolf (2001) conducted laboratory batch experiments to evaluate the

removal efficiencies of suspended solids, general organic matter, hydrocarbons and

phenols by DAF using chemical flocculation, followed by DAF. The flocculant dose

(aluminum sulfate (alum) or a cationic polyelectrolyte) and the air to solids ratio were

controlled. The DAF process reduced the general hydrocarbon content by 50 to 90%. The

results indicated that chemical flocculation, followed by DAF, efficiently removed the

emulsified phase, which could be aggregated and separated at the surface. However, it

was found that the process could also remove substantial amounts of dissolved organic

matter. Tansel and Pascual (2004) conducted studies to evaluate the individual and

interactive effects of the operational variables of the DAF process regarding the removal

17

been used in petrochemical complexes as a secondary clarifier to improve the operation

and capacity of an existing activated sludge system (Wang et al. 2004). Moursi and Abo-

Elela (1982) evaluated the use of chemical coagulation, followed by dissolved air

flotation as a treatment process for the primary effluent of refinery wastewater.

Considerable reductions of 94.4%, 95%, 95%, 85%, and 96.5%, respectively in turbidity,

COD, total oil and grease, phenols, and total oils were obtained. The data showed that

when wastewater was chemically pretreated to break the oil emulsion, dissolved air

flotation units were capable of removing most of the emulsified oil in addition to the

original free oil content. Removal rate of hydrocarbon compounds was generally found to

increase as the solubility of compounds decreased. The results obtained following the

treatment of refinery wastewater showed some of the hydrocarbon compounds were

completely removed.

Galil and Wolf (2001) conducted laboratory batch experiments to evaluate the

removal efficiencies of suspended solids, general organic matter, hydrocarbons and

phenols by DAP using chemical flocculation, followed by DAF. The flocculant dose

(aluminum sulfate (alum) or a cationic polyelectrolyte) and the air to solids ratio were

controlled. The DAF process reduced the general hydrocarbon content by 50 to 90%. The

results indicated that chemical flocculation, followed by DAF, efficiently removed the

emulsified phase, which could be aggregated and separated at the surface. However, it

was found that the process could also remove substantial amounts of dissolved organic

matter. Tansel and Pascual (2004) conducted studies to evaluate the individual and

interactive effects of the operational variables of the DAF process regarding the removal

17

efficiency of petroleum hydrocarbons from water sources contaminated with fuel oils. A

series of batch and continuous experiments (utilizing full pressurization and effluent

recirculation) were conducted using a 60 L DAF system, which could be operated either

in batch or continuous modes. The factorial analysis showed that for the batch mode of

operation, oil concentration, detention time, coagulant use, and water type had a

significant effect on petroleum hydrocarbon removal. However, with respect to the

continuous DAF runs, the only variable that was significant at the 95% confidence level

was detention time. The average petroleum hydrocarbon removal efficiency for batch

runs was 77 ± 2 and for continuous runs, with full pressurization, it was 86 ± 2%.

Azbar and Yonar (2004) studied various schemes of treating Vegetable Oil

Refinery Wastewater (VORW) and concluded it can be successfully treated using a

combination of physicochemical and biological methods. A pretreatment scheme of the

DAF unit, after coagulation and flocculation, was observed to be effective in minimizing

operating costs. The physicochemical treatment processes were found to significantly

influence the relative biodegradability of the VORW. Al-Shamrani et al. (2002)

conducted studies on synthetic industrial effluent, prepared by stabilizing low

concentrations of oil (Catenex 11, Shell, UK) in aqueous dispersion with a non-ionic

surfactant (Span 20), and DAF was used to clarify this wastewater. Measurements

indicated a saturator efficiency of approximately 90% was achieved and it was found that

increasing the working pressure of the saturator had less of an effect on the separation of

oil droplets than increasing the recycle ratio. Optimum conditions for separation are

obtained with an air to oil ratio of 0.0075, corresponding to a recycle ratio of 10%. The

18

efficiency of petroleum hydrocarbons from water sources contaminated with fuel oils. A

series of batch and continuous experiments (utilizing full pressurization and effluent

recirculation) were conducted using a 60 L DAF system, which could be operated either

in batch or continuous modes. The factorial analysis showed that for the batch mode of

operation, oil concentration, detention time, coagulant use, and water type had a

significant effect on petroleum hydrocarbon removal. However, with respect to the

continuous DAF runs, the only variable that was significant at the 95% confidence level

was detention time. The average petroleum hydrocarbon removal efficiency for batch

runs was 11 ±2 and for continuous runs, with full pressurization, it was 86 ± 2%.

Azbar and Yonar (2004) studied various schemes of treating Vegetable Oil

Refinery Wastewater (VORW) and concluded it can be successfully treated using a

combination of physicochemical and biological methods. A pretreatment scheme of the

DAF unit, after coagulation and flocculation, was observed to be effective in minimizing

operating costs. The physicochemical treatment processes were found to significantly

influence the relative biodegradability of the VORW. Al-Shamrani et al. (2002)

conducted studies on synthetic industrial effluent, prepared by stabilizing low

concentrations of oil (Catenex 11, Shell, UK) in aqueous dispersion with a non-ionic

surfactant (Span 20), and DAF was used to clarify this wastewater. Measurements

indicated a saturator efficiency of approximately 90% was achieved and it was found that

increasing the working pressure of the saturator had less of an effect on the separation of

oil droplets than increasing the recycle ratio. Optimum conditions for separation are

obtained with an air to oil ratio of 0.0075, corresponding to a recycle ratio of 10%. The

18

polyelectrolytes were found to be ineffective in enhancing separation. When aluminum

sulphate was used, it was found that it was important to decrease the magnitude of the

zeta potential in order to decrease electrostatic repulsion so that the emulsion was

destabilized prior to flotation. Under these conditions, DAF yielded near complete oil

separation even at a moderate working pressure and recycle ratio when the oil droplets

were destabilized. The treatment of oil-in-water emulsions containing n-octane (used as

simulated wastewater) was investigated by means of dissolved-air flotation jar-tests

(Zouboulis and Avranas 2000). The use of polyelectrolytes was observed to be

ineffective, while the addition of ferric chloride and a subsequent application of

dissolved-air flotation were found to be very efficient. At optimum defined experimental

conditions (recycle ratio: 30%, pH: 6, [Fe31: 100 mg/L and [sodium oleate]: 50 mg/L) of

more than 95% of the emulsified oil was effectively separated from an initial

concentration of 500 mg/L.

2.2.2 Aerobic Treatment of Oily Wastewater

In aerobic wastewater treatment systems, oils are generally believed to be

biodegradable and therefore, considered part of the treated organic load (Young, 1979).

Young (1979) reported that oxygen demand of influent dispersed polar oil should be

considered part of the normal BOD load headed to the treatment plant so that effluent

BOD measurements would include the oxygen demand of biodegradable oil in the

effluent samples. Moreover, the amount of oil present in an activated sludge system was

19

polyelectrolytes were found to be ineffective in enhancing separation. When aluminum

sulphate was used, it was found that it was important to decrease the magnitude of the

zeta potential in order to decrease electrostatic repulsion so that the emulsion was

destabilized prior to flotation. Under these conditions, DAF yielded near complete oil

separation even at a moderate working pressure and recycle ratio when the oil droplets

were destabilized. The treatment of oil-in-water emulsions containing n-octane (used as

simulated wastewater) was investigated by means of dissolved-air flotation jar-tests

(Zouboulis and Avranas 2000). The use of polyelectrolytes was observed to be

ineffective, while the addition of ferric chloride and a subsequent application of

dissolved-air flotation were found to be very efficient. At optimum defined experimental

conditions (recycle ratio: 30%, pH: 6, [Fe3+]: 100 mg/L and [sodium oleate]: 50 mg/L) of

more than 95% of the emulsified oil was effectively separated from an initial

concentration of 500 mg/L.

2.2.2 Aerobic Treatment of Oily Wastewater

In aerobic wastewater treatment systems, oils are generally believed to be

biodegradable and therefore, considered part of the treated organic load (Young, 1979).

Young (1979) reported that oxygen demand of influent dispersed polar oil should be

considered part of the normal BOD load headed to the treatment plant so that effluent

BOD measurements would include the oxygen demand of biodegradable oil in the

effluent samples. Moreover, the amount of oil present in an activated sludge system was

19

to be related to the occurrence of filamentous actinomycete Narcodia amarae, known to

be involved in the formation of scum and stable foams (Becker et al. 1999).

2.2.2.1 Treatment of Oily Wastewater in an Activated Sludge Process

Young (1979) carried out studies by mixing biological solids with vegetable oil

and correlated the effluent BOD characteristics with the amount of oil added. It was

observed that the removal of oil by mixed microbial population was equal to, or better

than BOD removal, suggesting not only biodegradation occurred, but also adsorption

took place. Hsu et al. (1983) conducted a batch study using acclimated activated sludge

for olive oil degradation. The results showed the maximum removal rate of olive oil was

in the order of 0.1 g/L/h. Adsorption of oil was found to also contribute to the removal

from wastewater. The results suggest that oil adsorption may influence the performance

of aerobic processes such as the Activated Sludge Process (ASP). In a study carried out

by Jung et al. (2002), an activated sludge process which treated oil at concentrations of

400 and 600 mg/L, achieved COD removal efficiencies of 86 and 75%, respectively.

However, at 800 mg/L of oil in the feed, the COD removal efficiency of the bioreactor

dropped markedly. Liu et al. (2004) achieved a 90% COD removal efficiency at a high

oil concentration of 660 mg/L on a batch activated sludge system under mesophilic

conditions, with a temperature as low as 21°C. Also, at 13,500 mg/L of oil concentration,

activated sludge still achieved removals in the range of 64 — 76%. The results

contradicted with those of Jung et al. (2002) and it was postulated that the nature of oil in

the wastewater, studied by Liu et al. (2004), was predominantly more of animal origin

20

to be related to the occurrence of filamentous actinomycete Narcodia amarae, known to

be involved in the formation of scum and stable foams (Becker et al. 1999).

2.2.2.1 Treatment of Oily Wastewater in an Activated Sludge Process

Young (1979) carried out studies by mixing biological solids with vegetable oil

and correlated the effluent BOD characteristics with the amount of oil added. It was

observed that the removal of oil by mixed microbial population was equal to, or better

than BOD removal, suggesting not only biodegradation occurred, but also adsorption

took place. Hsu et al. (1983) conducted a batch study using acclimated activated sludge

for olive oil degradation. The results showed the maximum removal rate of olive oil was

in the order of 0.1 g/L/h. Adsorption of oil was found to also contribute to the removal

from wastewater. The results suggest that oil adsorption may influence the performance

of aerobic processes such as the Activated Sludge Process (ASP). In a study carried out

by Jung et al. (2002), an activated sludge process which treated oil at concentrations of

400 and 600 mg/L, achieved COD removal efficiencies of 86 and 75%, respectively.

However, at 800 mg/L of oil in the feed, the COD removal efficiency of the bioreactor

dropped markedly. Liu et al. (2004) achieved a 90% COD removal efficiency at a high

oil concentration of 660 mg/L on a batch activated sludge system under mesophilic

conditions, with a temperature as low as 21°C. Also, at 13,500 mg/L of oil concentration,

activated sludge still achieved removals in the range of 64 - 76%. The results

contradicted with those of Jung et al. (2002) and it was postulated that the nature of oil in

the wastewater, studied by Liu et al. (2004), was predominantly more of animal origin

than mineral origin, which was not only non-inhibitory but also biodegradable. Also, it

was suggested that mechanisms such as coating the biological floc and hindering oxygen

transfer by oil had adverse impacts upon the biological system.

A study by Wakelin and Forster (1997) showed that acclimatized activated sludge

exhibited a higher performance than a non-acclimatized activated sludge even though the

microbial growth pattern and removal of oil were similar. Since activated sludge is a

mixture of different microorganisms, which can be dominated by different species, their

respective domination can be dictated by the type and concentration of the substrate

(Kovarova-Kovar and Egli 1998). Therefore, the results, as reported by Wakelin and

Forster (1997, suggest the differences in overall performancei of the various microbial

cultures could be due to differences in enzyme systems. This further suggests the use of

mixed microbial cultures such as activated sludge, particularly when it has been

acclimatized to oil, can offer a best option for the treatment of oily wastewaters. Mkhize

et al. (2000) evaluated the efficiency of a laboratory scale activated sludge process to

treat edible oil refining industry effluent. Treatability studies were conducted using an

anaerobic/aerobic Sequencing Batch Reactor (SBR). The results showed a 75% influent

COD reduction and more than 90% removal of oils and suspended solids. Aggelis et al.

(2001) evaluated the performance of separate aerobic, anaerobic and a combined

anaerobic - aerobic process for the biological treatment of green olive debittering

wastewater. Aerobic treatment resulted in a degradation efficiency of 72 — 76% but

required a pH correction and thus, hardly affected the polyphenolic compounds.

21

than mineral origin, which was not only non-inhibitory but also biodegradable. Also, it

was suggested that mechanisms such as coating the biological floe and hindering oxygen

transfer by oil had adverse impacts upon the biological system.

A study by Wakelin and Forster (1997) showed that acclimatized activated sludge

exhibited a higher performance than a non-acclimatized activated sludge even though the

microbial growth pattern and removal of oil were similar. Since activated sludge is a

mixture of different microorganisms, which can be dominated by different species, their

respective domination can be dictated by the type and concentration of the substrate

(Kovarova-Kovar and Egli 1998). Therefore, the results, as reported by Wakelin and

Forster (1997, suggest the differences in overall performances of the various microbial

cultures could be due to differences in enzyme systems. This further suggests the use of

mixed microbial cultures such as activated sludge, particularly when it has been

acclimatized to oil, can offer a best option for the treatment of oily wastewaters. Mkhize

et al. (2000) evaluated the efficiency of a laboratory scale activated sludge process to

treat edible oil refining industry effluent. Treatability studies were conducted using an

anaerobic/aerobic Sequencing Batch Reactor (SBR). The results showed a 75% influent

COD reduction and more than 90% removal of oils and suspended solids. Aggelis et al.

(2001) evaluated the performance of separate aerobic, anaerobic and a combined

anaerobic - aerobic process for the biological treatment of green olive debittering

wastewater. Aerobic treatment resulted in a degradation efficiency of 72 - 76% but

required a pH correction and thus, hardly affected the polyphenolic compounds.

21

However, the combined anaerobic — aerobic treatment achieved a COD and polyphenol

reduction by 74% and 20%, respectively.

2.2.2.2 Treatment of Oily Wastewater in an Attached Growth Process

A continuous bench scale Rotating Biological Contactor (RBC) was studied by

Najafpour et al. (2005) to treat Palm Oil Mill Effluents (POME) using S. cerevisiae as the

biomass. In the study, about 88% COD was removed with the lowest flow rate being 1.1

L/h of POME. When the flow rate was increased to 3.6 L/h, COD removal decreased to

57%. It was found that the initial biomass loading and fixing biofilm of S. cerevisiae on

the surface of RBC significantly improved the treatability of POME. In a study carried

out by El-Masry et al. (2004), a sand biofilm filter system, with bacterial strains of

Pseudomonas sp. and P. diminuta, was used for the removal of vegetable oil. Even in

cases of a high degree of pollution, an oil removal of 100% efficiency, a BOD of 96%

and a COD of 96%, at a 50 mL/h flow rate using one unit of biofilm system, was

achieved. When using two units in sequence, a complete removal of oil, BOD and COD,

with an efficiency of 100%, at a flow rate of 100 mL/h, was achieved.

2.2.2.3 Treatment of Oily Wastewater in a Combined Suspended and Attached

Growth System

In order to enhance the biodegradation of oils, Keenan and Sabelnikov (2000)

proposed using a combination of suspended and attached growth treatment systems, using

22

However, the combined anaerobic - aerobic treatment achieved a COD and polyphenol

reduction by 74% and 20%, respectively.

2.2.2.2 Treatment of Oily Wastewater in an Attached Growth Process

A continuous bench scale Rotating Biological Contactor (RBC) was studied by

Najafpour et al. (2005) to treat Palm Oil Mill Effluents (POME) using S. cerevisiae as the

biomass. In the study, about 88% COD was removed with the lowest flow rate being 1.1

L/h of POME. When the flow rate was increased to 3.6 L/h, COD removal decreased to

57%. It was found that the initial biomass loading and fixing biofilm of S. cerevisiae on

the surface of RBC significantly improved the treatability of POME. In a study carried

out by El-Masiy et al. (2004), a sand biofilm filter system, with bacterial strains of

Pseudomonas sp. and P. diminuta, was used for the removal of vegetable oil. Even in

cases of a high degree of pollution, an oil removal of 100% efficiency, a BOD of 96%

and a COD of 96%, at a 50 mL/h flow rate using one unit of biofilm system, was

achieved. When using two units in sequence, a complete removal of oil, BOD and COD,

with an efficiency of 100%, at a flow rate of 100 mL/h, was achieved.

2.2.2.3 Treatment of Oily Wastewater in a Combined Suspended and Attached

Growth System

In order to enhance the biodegradation of oils, Keenan and Sabelnikov (2000)

proposed using a combination of suspended and attached growth treatment systems, using

selected bacterial strains that were capable of degrading oils. The authors found that the

oil content in the effluent could not be reduced to values below 0.3 g/L from 1.5 g/L

using a suspended growth treatment system. However, adding a biofilter to the suspended

growth system substantially reduced the oil content in the wastewater to 0.03 g/L. An

increase in the efficiency of the system was the result of an increased concentration of

bacterial cells, which was accompanied by increased microbial activity, growth and

maintenance of microbial populations associated with attached growth systems. However

the treatment system reported by Keenan and Sabelnikov (2000) sporadically failed and

the content of oil in wastewater increased to 0.4 g/L. The authors attributed this failure to

a malfunction of the pH adjustment system.

2.2.3 Anaerobic Treatment of Oily Wastewaters

During the anaerobic treatment of oily wastewater, biotransformation of oil takes

place according to the following steps (Novak and Carlson 1970):

• Hydrolysis to unsaturated long chain fatty acids (LCFA);

• Saturation of unsaturated LCFAs; and

• 13- oxidation of saturated LCFA to volatile fatty acids.

During anaerobic treatment processes, the characteristics of oily wastewater lead to many

challenges. Sludges with diverse characteristics of poor activity can develop and form

foam on the surface of water. This may result in a loss of biomass with effluent, a

decrease in biomass quantity within the reactor and an efficiency of the treatment system.

In addition, oil can be adsorbed on the surface of the anaerobic sludge, which may limit

23

selected bacterial strains that were capable of degrading oils. The authors found that the

oil content in the effluent could not be reduced to values below 0.3 g/L from 1.5 g/L

using a suspended growth treatment system. However, adding a biofilter to the suspended

growth system substantially reduced the oil content in the wastewater to 0.03 g/L. An

increase in the efficiency of the system was the result of an increased concentration of

bacterial cells, which was accompanied by increased microbial activity, growth and

maintenance of microbial populations associated with attached growth systems. However

the treatment system reported by Keenan and Sabelnikov (2000) sporadically failed and

the content of oil in wastewater increased to 0.4 g/L. The authors attributed this failure to

a malfunction of the pH adjustment system.

2.2.3 Anaerobic Treatment of Oily Wastewaters

During the anaerobic treatment of oily wastewater, biotransformation of oil takes

place according to the following steps (Novak and Carlson 1970):

• Hydrolysis to unsaturated long chain fatty acids (LCFA);

• Saturation of unsaturated LCFAs; and

• P- oxidation of saturated LCFA to volatile fatty acids.

During anaerobic treatment processes, the characteristics of oily wastewater lead to many

challenges. Sludges with diverse characteristics of poor activity can develop and form

foam on the surface of water. This may result in a loss of biomass with effluent, a

decrease in biomass quantity within the reactor and an efficiency of the treatment system.

In addition, oil can be adsorbed on the surface of the anaerobic sludge, which may limit

23

the transport of soluble substrate to the biomass. At a lower temperature, oils may

solidify and create operational problems such as clogging and the production of

unpleasant odours (Cammarota and Freire 2006).

Problems with respect to anaerobic treatment of wastewater containing oil are the

result of two phenomena: (1) adsorption of a light oil layer around biomass particles

causing biomass flotation and washout and (2) acute toxicity of LCFA, especially

unsaturated LCFA, to both methanogens and acetogens, the two main trophic groups

involved in LCFA degradation (Hanalei et al. 1981). The conversion of LCFA to acetate

in an anaerobic process limits gas production and the removal of COD (Saacti et al.

2003). The design of anaerobic systems with which to treat wastewaters containing

vegetable oils account for the slow degradation of LCFA and the potential inhibition by

LCFA. The breakdown of LCFA is often the rate-limiting step in the degradation of a

complex substrate (Salminen et al. 2000).

Experimental work, carried out by Beccari et al. (1998), on the anaerobic

treatment of Olive Oil Mill Wastewater (OMW) showed the inhibition of methane

production was mainly caused by the presence of lipids in the OMW. Experimental work

carried out in a semi- continuous two-reactor system, fed with diluted OMW, identified

the saturation of unsaturated LCFA as the key factor preventing inhibition of

methanogenesis. It was concluded that the presence of modest hydrogenotrophic activity

in the first reactor was sufficient to obtain an almost quantitative conversion of

unsaturated LCFAs to palmitic acid, thus, drastically lowering inhibition on

methanogenesis in the second reactor. However, degradation of saturated LCFAs did not

24

the transport of soluble substrate to the biomass. At a lower temperature, oils may

solidify and create operational problems such as clogging and the production of

unpleasant odours (Cammarota and Freire 2006).

Problems with respect to anaerobic treatment of wastewater containing oil are the

result of two phenomena: (1) adsorption of a light oil layer around biomass particles

causing biomass flotation and washout and (2) acute toxicity of LCFA, especially

unsaturated LCFA, to both methanogens and acetogens, the two main trophic groups

involved in LCFA degradation (Hanaki et al. 1981). The conversion of LCFA to acetate

in an anaerobic process limits gas production and the removal of COD (Saacti et al.

2003). The design of anaerobic systems with which to treat wastewaters containing

vegetable oils account for the slow degradation of LCFA and the potential inhibition by

LCFA. The breakdown of LCFA is often the rate-limiting step in the degradation of a

complex substrate (Salminen et al. 2000).

Experimental work, carried out by Beccari et al. (1998), on the anaerobic

treatment of Olive Oil Mill Wastewater (OMW) showed the inhibition of methane

production was mainly caused by the presence of lipids in the OMW. Experimental work

carried out in a semi- continuous two-reactor system, fed with diluted OMW, identified

the saturation of unsaturated LCFA as the key factor preventing inhibition of

methanogenesis. It was concluded that the presence of modest hydrogenotrophic activity

in the first reactor was sufficient to obtain an almost quantitative conversion of

unsaturated LCFAs to palmitic acid, thus, drastically lowering inhibition on

methanogenesis in the second reactor. However, degradation of saturated LCFAs did not

24

continue beyond the formation of palmitic acid. Concentration of hydrogenotrophic

bacteria within the acidogenic reactor was not high enough to allow the progression of

the 0- oxidation down to the VFAs. Another study was carried out by Beccari et al.

(1996) regarding OMW in order to investigate interaction between the acidogenesis and

methanogenesis occurring in anaerobic digestion. It was observed that most of the lipids

in OMW were degraded in both stages. A low methanogenic activity, established in

acidogenic conditions due to the partial degradation of the chemical inhibitor, was

observed to be the key factor in determining lipid degradation, even in acidogenesis. The

results of the study suggested two-phase anaerobic digestion might be adopted as a

suitable process to optimize OMW degradation.

Saacti et al. (2003) investigated the treatment of wastewater from a sunflower oil

industry in a pilot-scale mesophilic Upflow Anaerobic Sludge Blanket digestion (UASB)

reactor. Removal efficiencies of total lipids and fatty acids were observed to be above

70% at organic loading rates between 1.6 and 7.8 kg COD/ m3 d and optimum retention

times of 2.0 and 2.8 days. The conversion rate of removed COD to methane was between

0.16 and 0.35 m3 CH4 / kg COD. Hanaki et al. (1990) anaerobically treated cafeteria

wastewater containing approximately 30% oil on a COD basis, in single-phase and two-

phase systems. The two-phase system achieved approximately 85% of methane

conversion from the removed COD, while the single-phase system achieved about 70%

on average. The oil content in the wastewaters on a COD basis reflected the differences

between the two systems. Therefore, the results showed that oil was degraded well in the

two-phase system, while little degradation took place in a single-phase system. POME

25

continue beyond the formation of palmitic acid. Concentration of hydrogenotrophic

bacteria within the acidogenic reactor was not high enough to allow the progression of

the P- oxidation down to the VFAs. Another study was carried out by Beccari et al.

(1996) regarding OMW in order to investigate interaction between the acidogenesis and

methanogenesis occurring in anaerobic digestion. It was observed that most of the lipids

in OMW were degraded in both stages. A low methanogenic activity, established in

acidogenic conditions due to the partial degradation of the chemical inhibitor, was

observed to be the key factor in determining lipid degradation, even in acidogenesis. The

results of the study suggested two-phase anaerobic digestion might be adopted as a

suitable process to optimize OMW degradation.

Saacti et al. (2003) investigated the treatment of wastewater from a sunflower oil

industry in a pilot-scale mesophilic Upflow Anaerobic Sludge Blanket digestion (UASB)

reactor. Removal efficiencies of total lipids and fatty acids were observed to be above

70% at organic loading rates between 1.6 and 7.8 kg COD/ m3 d and optimum retention

times of 2.0 and 2.8 days. The conversion rate of removed COD to methane was between

0.16 and 0.35 m3 CH4 / kg COD. Hanaki et al. (1990) anaerobically treated cafeteria

wastewater containing approximately 30% oil on a COD basis, in single-phase and two-

phase systems. The two-phase system achieved approximately 85% of methane

conversion from the removed COD, while the single-phase system achieved about 70%

on average. The oil content in the wastewaters on a COD basis reflected the differences

between the two systems. Therefore, the results showed that oil was degraded well in the

two-phase system, while little degradation took place in a single-phase system. POME

25

treatability in an anaerobic hybrid digester, studied by Borja et al. (1996), showed that at

an Organic Loading Rate (OLR) of 16.2 g/L day and an HRT of 3.5 days, COD removal

efficiency and a methane yield of 92% and 0.34 m3/kg were achieved, respectively.

Najafpour et al. (2006) studied anaerobic digestion of POME in an Upflow Anaerobic

Sludge-fixed Film (UASFF) bioreactor. The study revealed the use of the UASFF reactor

was a good strategy in order to accelerate anaerobic granulation and to achieve high COD

removal efficiency in a short period of time. The reactor was found to be very efficient in

the treatment of diluted and high strength POME at a high OLR and a short HRT. High

COD removals of 89 and 97%, at an HRT of 1.5 and 3 days, were achieved, respectively.

Also, it showed that the use of packing media in the middle portion reduced channeling

and loss of biomass due to flotation associated with poorly performing UASB reactors.

Faisal and Unno (2001) studied the treatability of POME using a Modified Anaerobic

Bioreactor (MABR). It was observed the removal of COD and oil varied from 87 to 95%

and from 44 to 91%, respectively.

Filidei et al. (2003) studied anaerobic digestion of OMW on a laboratory scale

involving chemical-physical processes, followed by biological treatment. Anaerobic

digestion was observed to reduce organic load by 78 - 89% and phytotoxocity tests

carried out on Lepidium sativum seeds showed that the anaerobic treatment considerably

reduced the phytotoxic character of OMW. Anaerobic digestion of chemically pretreated

OMW was observed by Banitez et al. (2001) to reduce COD in the range of 0.77 - 0.65 g

of COD degraded per g of COD fed to the reactor. Wahaab and El-Awady (1999) studied

the treatment of meat processing wastewater containing oil and grease in a UASB reactor,

26

treatability in an anaerobic hybrid digester, studied by Borja et al. (1996), showed that at

an Organic Loading Rate (OLR) of 16.2 g/L day and an HRT of 3.5 days, COD removal

efficiency and a methane yield of 92% and 0.34 m3/kg were achieved, respectively.

Najafpour et al. (2006) studied anaerobic digestion of POME in an Upflow Anaerobic

Sludge-fixed Film (UASFF) bioreactor. The study revealed the use of the UASFF reactor

was a good strategy in order to accelerate anaerobic granulation and to achieve high COD

removal efficiency in a short period of time. The reactor was found to be very efficient in

the treatment of diluted and high strength POME at a high OLR and a short HRT. High

COD removals of 89 and 97%, at an HRT of 1.5 and 3 days, were achieved, respectively.

Also, it showed that the use of packing media in the middle portion reduced channeling

and loss of biomass due to flotation associated with poorly performing UASB reactors.

Faisal and Unno (2001) studied the treatability of POME using a Modified Anaerobic

Bioreactor (MABR). It was observed the removal of COD and oil varied from 87 to 95%

and from 44 to 91%, respectively.

Filidei et al. (2003) studied anaerobic digestion of OMW on a laboratory scale

involving chemical-physical processes, followed by biological treatment. Anaerobic

digestion was observed to reduce organic load by 78 - 89% and phytotoxocity tests

carried out on Lepidium sativum seeds showed that the anaerobic treatment considerably

reduced the phytotoxic character of OMW. Anaerobic digestion of chemically pretreated

OMW was observed by Banitez et al. (2001) to reduce COD in the range of 0.77 - 0.65 g

of COD degraded per g of COD fed to the reactor. Wahaab and El-Awady (1999) studied

the treatment of meat processing wastewater containing oil and grease in a UASB reactor,

26

followed by a RBC. A removal efficiency of only 58% was achieved for oil by the UASB

and when a RBC followed, an overall efficiency of 91% was achieved with an oil

concentration of 10 mg/L in the final effluent. Nakhla et al. (2003) evaluated the use of a

new biosurfactant, derived from cactus, for the treatment of oily wastewater

anaerobically. The addition of biosurfactant to the oily wastewater produced a 164 —

238% increase in the total COD biodegradation rate coefficient. Results from a full-scale

mesophilic anaerobic digestion system indicated that the addition of biosurfactant at

doses of 130 - 200 mg/L decreased oil concentrations from 66,300 to 10,200 mg/L over a

two-month period. Aerobic and anaerobic processes for the removal of oil from

wastewater are given in Table 2.1.

2.2.1 Enzymes for the Treatment of Oily Wastewaters

Lipases are enzymes or biocatalysts, which have the ability to catalyze the

cleavage of carboxyl ester bonds in tri-, di-, and monoacylglycerols (Balcao et al. 1996).

Lipases cleave ester bonds of triacylglycerols with the consumption of water molecules

(hydrolysis). There are several research studies available regarding the treatment of oily

wastes using lipase (Wakelin and Forster 1997; Cammarota et al. 2001). However, the

majority of the studies were focused on the pre-treatment of wastewater with

synthetically added oils. There is a wide range of scientific studies investigating

enzymatic hydrolysis processes to precede traditional biological treatment. The enzymes

27

followed by a RBC. A removal efficiency of only 58% was achieved for oil by the UASB

and when a RBC followed, an overall efficiency of 91% was achieved with an oil

concentration of 10 mg/L in the final effluent. Nakhla et al. (2003) evaluated the use of a

new biosurfactant, derived from cactus, for the treatment of oily wastewater

anaerobically. The addition of biosurfactant to the oily wastewater produced a 164 -

238% increase in the total COD biodegradation rate coefficient. Results from a full-scale

mesophilic anaerobic digestion system indicated that the addition of biosurfactant at

doses of 130 - 200 mg/L decreased oil concentrations from 66,300 to 10,200 mg/L over a

two-month period. Aerobic and anaerobic processes for the removal of oil from

wastewater are given in Table 2.1.

2.2.1 Enzymes for the Treatment of Oily Wastewaters

Lipases are enzymes or biocatalysts, which have the ability to catalyze the

cleavage of carboxyl ester bonds in tri-, di-, and monoacylglycerols (Balcao et al. 1996).

Lipases cleave ester bonds of triacylglycerols with the consumption of water molecules

(hydrolysis). There are several research studies available regarding the treatment of oily

wastes using lipase (Wakelin and Forster 1997; Cammarota et al. 2001). However, the

majority of the studies were focused on the pre-treatment of wastewater with

synthetically added oils. There is a wide range of scientific studies investigating

enzymatic hydrolysis processes to precede traditional biological treatment. The enzymes

27

Table 2.1: Biological treatment

Biological treatment Source of oil

Efficiency! Remarks Reference

Acclimated ASP OMW Maximum removal rate Hsu et al. (1983) 0.1 g/L/h

ASP Dairy wastewater

86% COD removal Jung et al. (2002)

Batch ASP Pet food wastewater

90% COD removal Liu et al. (2004)

Anaerobic/aerobic Edible oil 90% oil removal Mkhize et al. SBR industrial

effluent 75% COD reduction (2000)

Aerobic treatment OMW 72 -76% COD reduction Aggelis et al. Combined anaerobic-aerobic treatment

74% COD reduction (2001)

RBC POME 88% COD removal Najafpour et al. (2005)

Sand biofilm filter Vegetable 100% oil removal El-Masary et al. oil 96% COD reduction (2004)

Combined suspended Bakery Reduction in oil content Keenan and and attached growth wastewater from 0.3 g/L to 0.028 Sabelnikov (2000)

Two-reactorTwo-reactor anaerobic system

OMW -

82% total lipids removal Baccari et al. (1998)

UASB Sunflower oil industry

70% removal efficiency of total lipids

Saacti et al. (2003)

Anaerobic treatment Cafeteria 85% of CH4 converted Hanaki et al. (2- phase system) wastewater from removed COD (1990) 2-stage UASB POME Borja et al. (1996) Anaerobic hybrid digester

POME 92% COD removal Najafpour et al. (2006)

UASFF POME 89 — 97% COD removal Faisal and Unno (2001)

Modified anaerobic POME 87 - 95% COD removal Falidei et al. bioreactor 44 -91% oil removal (2003) Anaerobic digestion OMW 78 - 89% reduction in

organic load Banitez et al. (2001)

UASB followed by Meat 91% oil removal Nakhla et al. RBC processing (2003)

28

Table 2.1: Biological treatment

Biological treatment Source of oil

Efficiency/ Remarks Reference

Acclimated ASP OMW Maximum removal rate Hsu et al. (1983) 0.1 g/L/h

ASP Dairy 86% COD removal Jung et al. (2002) wastewater

Batch ASP Pet food 90% COD removal Liu et al. (2004) wastewater

Anaerobic/aerobic Edible oil 90% oil removal Mkhize et al. SBR industrial 75% COD reduction (2000)

effluent Aerobic treatment OMW 72 -76% COD reduction Aggelis et al. Combined 74% COD reduction (2001) anaerobic-aerobic treatment RBC POME 88% COD removal Najafpour et al.

(2005) Sand biofilm filter Vegetable 100% oil removal El-Masary et al.

oil 96% COD reduction (2004) Combined suspended Bakery Reduction in oil content Keenan and and attached growth wastewater from 0.3 g/L to 0.028 Sabelnikov (2000)

g/L Two-reactor OMW 82% total lipids removal Baccari et al. anaerobic system (1998) UASB Sunflower 70% removal efficiency Saacti et al. (2003)

oil industry of total lipids Anaerobic treatment Cafeteria 85% of CH> converted Hanaki et al. (2- phase system) wastewater from removed COD (1990) 2-stage UASB POME Boija et al. (1996) Anaerobic hybrid POME 92% COD removal Najafpour et al. digester (2006) UASFF POME 89 - 97% COD removal Faisal and Unno

(2001) Modified anaerobic POME 87 - 95% COD removal Falidei et al. bioreactor 44-91% oil removal (2003) Anaerobic digestion OMW 78 - 89% reduction in Banitez et al.

organic load (2001) UASB followed by Meat 91% oil removal Nakhla et al. RBC processing (2003)

28

catalyze the hydrolysis of complex organic compounds, transforming them into

substances which can be readily biodegraded by the microbial consortium present in

subsequent biological treatments (Cammarota and Freire 2006). However, the use of

enzyme preparations is not attractive due to the fact it is only used for the hydrolysis of

oils to fatty acids and glycerol (Malcata et al. 1992; Paiva et al. 2000). The liberated fatty

acids can form colloidal particles that aggregate and precipitate from solutions during

changes in environmental conditions during the treatment system, causing clogging and

process failure (Chao and Yang 1981). Therefore, this approach provides only a partial

solution to the problem. Process stability may also depend upon the state of the added

enzyme preparations. If added to a solution, enzymes would disappear from the system. It

is also difficult to recover enzymes from reactor effluent at the end of the catalytic

process, which is an even more expensive exercise. In contrast, immobilized enzymes

would be retained within the system and most likely would improve stability in relation

to environmental conditions (Malcata et al. 1992). Immobilized enzymes have the

advantages of multiple usage, controlled reactions, and mechanical stability (Murray et

al. 1997). Despite the advantages, the use of immobilized enzymes in wastewater

treatment has been limited by several factors, mainly due to the high cost of the enzymes

connected with the immobilization procedure. Also, loss of enzyme activity during the

immobilization procedure and during the reaction is the main hurdle in the way of

widespread commercialization (Murray et al. 1997).

In a study conducted by Jeganathan et al. (2006), hydrolysis of oil originating

from pet food industrial wastewater was evaluated. Lipase from Candida rugosa was

29

catalyze the hydrolysis of complex organic compounds, transforming them into

substances which can be readily biodegraded by the microbial consortium present in

subsequent biological treatments (Cammarota and Freire 2006). However, the use of

enzyme preparations is not attractive due to the fact it is only used for the hydrolysis of

oils to fatty acids and glycerol (Malcata et al. 1992; Paiva et al. 2000). The liberated fatty

acids can form colloidal particles that aggregate and precipitate from solutions during

changes in environmental conditions during the treatment system, causing clogging and

process failure (Chao and Yang 1981). Therefore, this approach provides only a partial

solution to the problem. Process stability may also depend upon the state of the added

enzyme preparations. If added to a solution, enzymes would disappear from the system. It

is also difficult to recover enzymes from reactor effluent at the end of the catalytic

process, which is an even more expensive exercise. In contrast, immobilized enzymes

would be retained within the system and most likely would improve stability in relation

to environmental conditions (Malcata et al. 1992). Immobilized enzymes have the

advantages of multiple usage, controlled reactions, and mechanical stability (Murray et

al. 1997). Despite the advantages, the use of immobilized enzymes in wastewater

treatment has been limited by several factors, mainly due to the high cost of the enzymes

connected with the immobilization procedure. Also, loss of enzyme activity during the

immobilization procedure and during the reaction is the main hurdle in the way of

widespread commercialization (Murray et al. 1997).

In a study conducted by Jeganathan et al. (2006), hydrolysis of oil originating

from pet food industrial wastewater was evaluated. Lipase from Candida rugosa was

29

immobilized and applied into the hydrolysis experiment. Lipase from Candida rugosa is

a microbial enzyme produced by the fermentation of yeasts. It was extensively used for

oil hydrolysis (Murray et al. 1997; Knezeric et al. 1998; Murthy et al. 2004) because it is

one of the commercially available lipases that have the ability to liberate all types of acyl

chains. The study showed approximately 50% of oil was hydrolyzed due to enzyme

activity. A significant increment in COD and VFA production was also observed. During

the 3-day experiment, approximately 65% of the beads were recovered and

approximately 70% of the enzyme activity remained in the beads after hydrolysis. The

study showed the immobilized lipase could be used up to 4 cycles with a 55% activity

recovery. However, after 4 cycles, the activity loss was critical, probably due to lipase

leakage from the beads and/or blockage of substrate/product. The study showed the

potential of immobilized lipase as a pre-treatment step in the biological treatment of pet

food manufacturing wastewater.

Jung et al. (2002) treated dairy wastewater in a batch activated sludge system with

and without an enzymatic pre-hydrolysis stage. A fermented babassu cake containing

Penicillium restrictum lipase was used in the pre-hydrolysis stage. The efficiency of pre-

hydrolysis was monitored via the formation of free acids. During the study, a marked

difference between the concentration of free acids, before and after hydrolysis, was

observed, indicating a modification in composition of wastewater even when a small

quantity of fermented cake (0.2) was utilized. Following enzymatic hydrolysis, COD

values were observed to rise due to the presence of organic compounds other than

microorganisms and proteins in the enzymatic cake. Although fed with high COD

30

immobilized and applied into the hydrolysis experiment. Lipase from Candida rugosa is

a microbial enzyme produced by the fermentation of yeasts. It was extensively used for

oil hydrolysis (Murray et al. 1997; Knezeric et al. 1998; Murthy et al. 2004) because it is

one of the commercially available lipases that have the ability to liberate all types of acyl

chains. The study showed approximately 50% of oil was hydrolyzed due to enzyme

activity. A significant increment in COD and VFA production was also observed. During

the 3-day experiment, approximately 65% of the beads were recovered and

approximately 70% of the enzyme activity remained in the beads after hydrolysis. The

study showed the immobilized lipase could be used up to 4 cycles with a 55% activity

recovery. However, after 4 cycles, the activity loss was critical, probably due to lipase

leakage from the beads and/or blockage of substrate/product. The study showed the

potential of immobilized lipase as a pre-treatment step in the biological treatment of pet

food manufacturing wastewater.

Jung et al. (2002) treated dairy wastewater in a batch activated sludge system with

and without an enzymatic pre-hydrolysis stage. A fermented babassu cake containing

Penicillium restrictum lipase was used in the pre-hydrolysis stage. The efficiency of pre-

hydrolysis was monitored via the formation of free acids. During the study, a marked

difference between the concentration of free acids, before and after hydrolysis, was

observed, indicating a modification in composition of wastewater even when a small

quantity of fermented cake (0.2) was utilized. Following enzymatic hydrolysis, COD

values were observed to rise due to the presence of organic compounds other than

microorganisms and proteins in the enzymatic cake. Although fed with high COD

30

concentrations, the hydrolyzed bioreactor maintained COD removal efficiencies in the

range of 82% to 93%, indicating that products of enzymatic hydrolysis were easily

metabolized by the microbial consortium present in the activated sludge. Also, biomass

growth was observed to be more intense in the hydrolyzed bioreactor.

Leal et al. (2006) conducted a study regarding the biological treatment of dairy

wastewater containing high levels of oil. The study used two identical Upflow Anaerobic

Sludge Blanket (UASB) reactors, one fed with wastewater from an upstream enzymatic

hydrolysis step and the other with raw wastewater. Hydrolysis was carried out at 35°C for

14 hours using an enzyme prepared via solid-state fermentation, showing pronounced

lipase activity. The UASB reactor fed with the hydrolyzed wastewater assimilated the

increase of oil content in the influent and showed stable COD removal even when the oil

concentration was raised to 1000 mg/L. Dharmshiti and Kuhasuntisuk (1998)

investigated the effect of the addition of lipase to a biological system treating restaurant

wastewater. The oil content was totally removed after 48 hours of incubating the

wastewater with the enzyme. However, the amount of enzyme preparation used in the

study was appreciably high. Cammarota and Freire (2006) first proposed hydrolysis of oil

by solid preparation of lipase using dairy wastewater. The results of the work showed that

pre-treatment with enzymatic hydrolysis improved the treatment of dairy wastewater in a

bench scale UASB reactor. Leal et al. (2002) carried out a similar work using a liquid

enzyme and batch operated anaerobic reactors. The highest level of oil investigated by

the authors was 1200 mg/L and the removal attained in reactors fed with the hydrolyzed

stream was significantly higher than that reached in the control reactor without pre-

31

concentrations, the hydrolyzed bioreactor maintained COD removal efficiencies in the

range of 82% to 93%, indicating that products of enzymatic hydrolysis were easily

metabolized by the microbial consortium present in the activated sludge. Also, biomass

growth was observed to be more intense in the hydrolyzed bioreactor.

Leal et al. (2006) conducted a study regarding the biological treatment of dairy

wastewater containing high levels of oil. The study used two identical Upflow Anaerobic

Sludge Blanket (UASB) reactors, one fed with wastewater from an upstream enzymatic

hydrolysis step and the other with raw wastewater. Hydrolysis was carried out at 35°C for

14 hours using an enzyme prepared via solid-state fermentation, showing pronounced

lipase activity. The UASB reactor fed with the hydrolyzed wastewater assimilated the

increase of oil content in the influent and showed stable COD removal even when the oil

concentration was raised to 1000 mg/L. Dharmshiti and Kuhasuntisuk (1998)

investigated the effect of the addition of lipase to a biological system treating restaurant

wastewater. The oil content was totally removed after 48 hours of incubating the

wastewater with the enzyme. However, the amount of enzyme preparation used in the

study was appreciably high. Cammarota and Freire (2006) first proposed hydrolysis of oil

by solid preparation of lipase using dairy wastewater. The results of the work showed that

pre-treatment with enzymatic hydrolysis improved the treatment of dairy wastewater in a

bench scale UASB reactor. Leal et al. (2002) carried out a similar work using a liquid

enzyme and batch operated anaerobic reactors. The highest level of oil investigated by

the authors was 1200 mg/L and the removal attained in reactors fed with the hydrolyzed

stream was significantly higher than that reached in the control reactor without pre-

31

treatment. Therefore, enzymatic pre-treatment was found to improve the performance of

anaerobic reactors, which treat industrial wastewaters containing high oil concentrations.

Gombert et al. (1999) studied lipase production by Penicillium restrictum in solid-state

fermentation using a babassu oil cake as a substrate. The highest lipase activity was

observed after 24 hours of cultivation with 2% olive oil enrichment.

2.2.2 Microbial Degradation of Oils

Generally, microbial oil degradation is considered to occur as a result of

hydrolysis of oil by secretion of lipase, an oil-degrading enzyme, which degrades the oil

to organic acids and Volatile Fatty Acids (VFA) via beta-oxidation, a fatty acid

degradation pathway. Finally, oil is decomposed to CO2 and H2O (Taken et al. 2005).

Keenan and Sabelnikov (2000) studied the biodegradation of corn, olive, sunflower and

waste oils employing a variety of bacterial strains such as Acinetobacter sp.,

Rhodococcus sp. and Caseobacter sp. that were isolated from different environments

based on their ability to grow on vegetable and waste oils and by commercial bacterial

preparations specifically designed for oil degradation. As for all bacterial strains and

preparations, only corn oil and waste oils supported microbial growth more efficiently

than either olive or sunflower oil. Moreover, the Caseobacter strain and one commercial

preparation could not grow on olive oil at all. Wakelin and Forster (1997) reported

similar results regarding the Acinetobacter strain, which was found to grow efficiently on

oils including Rhodococcus rubra, Nocardia amarae, and Microthrix parvicella.

However, even Acinetobacter sp. could not reduce the content of lipids in wastewater to

32

treatment. Therefore, enzymatic pre-treatment was found to improve the performance of

anaerobic reactors, which treat industrial wastewaters containing high oil concentrations.

Gombert et al. (1999) studied lipase production by Penicillium restrictum in solid-state

fermentation using a babassu oil cake as a substrate. The highest lipase activity was

observed after 24 hours of cultivation with 2% olive oil enrichment.

2.2.2 Microbial Degradation of Oils

Generally, microbial oil degradation is considered to occur as a result of

hydrolysis of oil by secretion of lipase, an oil-degrading enzyme, which degrades the oil

to organic acids and Volatile Fatty Acids (VFA) via beta-oxidation, a fatty acid

degradation pathway. Finally, oil is decomposed to CO2 and H2O (Takeno et al. 2005).

Keenan and Sabelnikov (2000) studied the biodegradation of com, olive, sunflower and

waste oils employing a variety of bacterial strains such as Acinetobacter sp.,

Rhodococcus sp. and Caseobacter sp. that were isolated from different environments

based on their ability to grow on vegetable and waste oils and by commercial bacterial

preparations specifically designed for oil degradation. As for all bacterial strains and

preparations, only corn oil and waste oils supported microbial growth more efficiently

than either olive or sunflower oil. Moreover, the Caseobacter strain and one commercial

preparation could not grow on olive oil at all. Wakelin and Forster (1997) reported

similar results regarding the Acinetobacter strain, which was found to grow efficiently on

oils including Rhodococcus rubra, Nocardia amarae, and Microthrix parvicella.

However, even Acinetobacter sp. could not reduce the content of lipids in wastewater to

values lower than 0.1 g/L. From an initial oil content of 1.5 g/L, the lowest values

achieved using Acinetobacter sp. were 0.35 g/L for corn oil and 0.267 g/L for waste oil.

Keenan and Sabelnikov (2000) also reported that microbial growth was more so on

unrefined oils than on refined oils; this was likely because the former contained nutrients,

which were removed while making refined oil.

In a study carried out by El-Bestawy et al. (2005), eight bacterial species were

isolated from vegetable oil contaminated industrial wastewater, four of which were found

to have the ability to degrade oil in the contaminated wastewater. The isolates were

identified as Pseudomonas sp., P. diminuta, P. Pseudoalcaligenes and Escherichia sp.

Results showed that all tested bacteria were able to degrade palm oil completely and

utilized free fatty acids as a carbon source. The combination, Pseudomonas sp. and P.

diminuta, produced the highest degradative activity followed by combination,

Pseudomonas sp., P. diminuta and P. Pseudoalcaligenes. Also, the combination of

Pseudomonas sp. and P. diminuta produced the highest activity in reducing COD by 93%

and BOD by 100%.

An inoculum containing a mixed culture of fifteen bacterial isolates from various

fatty wastewater samples was developed by Tano-Debrah et al. (1999) to degrade oils.

The inoculum was effective in wastewaters with pH ranging from 4.5 to 9.5 and the

optimum treatment temperature was 20 — 25°C. It was observed that the growth of

organisms was better in an alkaline medium due to better emulsification. Microorganisms

generally live at the oil-water interfaces and the attack on oil molecules and microbial

oxidation is rapid when the hydrocarbon molecule is in intimate contact with water at

33

values lower than 0.1 g/L. From an initial oil content of 1.5 g/L, the lowest values

achieved using Acinetobacter sp. were 0.35 g/L for corn oil and 0.267 g/L for waste oil.

Keenan and Sabelnikov (2000) also reported that microbial growth was more so on

unrefined oils than on refined oils; this was likely because the former contained nutrients,

which were removed while making refined oil.

In a study carried out by El-Bestawy et al. (2005), eight bacterial species were

isolated from vegetable oil contaminated industrial wastewater, four of which were found

to have the ability to degrade oil in the contaminated wastewater. The isolates were

identified as Pseudomonas sp., P. diminuta, P. Pseudoalcaligenes and Escherichia sp.

Results showed that all tested bacteria were able to degrade palm oil completely and

utilized free fatty acids as a carbon source. The combination, Pseudomonas sp. and P.

diminuta, produced the highest degradative activity followed by combination,

Pseudomonas sp., P. diminuta and P. Pseudoalcaligenes. Also, the combination of

Pseudomonas sp. and P. diminuta produced the highest activity in reducing COD by 93%

and BOD by 100%.

An inoculum containing a mixed culture of fifteen bacterial isolates from various

fatty wastewater samples was developed by Tano-Debrah et al. (1999) to degrade oils.

The inoculum was effective in wastewaters with pH ranging from 4.5 to 9.5 and the

optimum treatment temperature was 20 - 25°C. It was observed that the growth of

organisms was better in an alkaline medium due to better emulsification. Microorganisms

generally live at the oil-water interfaces and the attack on oil molecules and microbial

oxidation is rapid when the hydrocarbon molecule is in intimate contact with water at

33

temperatures ranging from 15 to 35*C, which means adequate mixing or dispersion of the

water and oil is necessary for effective degradation. Adequate mixing occurs in emulsion

formation. Biodegradative tests with the inoculum on samples of eight different oils,

developed by Tano-Debrah et al. (1999), showed that the organisms in the culture could

degrade several oils of plant and animal origin.

Takeno et al. (2005) studied the removal of cooking oil from domestic

wastewater. A continuous treatment of oil containing wastewater was carried out with

agar-immobilized Rhodobacter Shaeroide S, a photosynthetic bacterium, at a fixed

dilution rate of 0.4 /day. The results indicated that 96% of the oil was removed from the

wastewater, and the maximum removal rate of oil reached approximately 3.83 kg

oil/m3/day. The study showed that immobilized photosynthetic bacteria directly

decomposed one part of the oil contained in the wastewater, while oil-decomposing

bacteria (naturally growing Pseudomonas and Bacillus) decomposed another part of the

oil, degrading it to organic fatty acids such as propionic and acetic acid. Photosynthetic

bacteria consumed these acids. As a result, oil degradation was promoted by the addition

of photosynthetic bacteria and the combination showed significant decomposition of oil

from wastewater.

Prasertasan et al. (2004) investigated the bioseparation of suspended solids and oil

from POME. Two thermotolerant lipase producing fungal isolates, ST4 and ST29, were

observed to remove 98.9% and 88.9% of oil, respectively. The high treatment efficiency

was found to be the consequence of the entrapment of the majority of suspended solids

and residual oil of POME in the polymer excreted from these strains. Lipases are widely

34

temperatures ranging from 15 to 35°C, which means adequate mixing or dispersion of the

water and oil is necessary for effective degradation. Adequate mixing occurs in emulsion

formation. Biodegradative tests with the inoculum on samples of eight different oils,

developed by Tano-Debrah et al. (1999), showed that the organisms in the culture could

degrade several oils of plant and animal origin.

Takeno et al. (2005) studied the removal of cooking oil from domestic

wastewater. A continuous treatment of oil containing wastewater was carried out with

agar-immobilized Rhodobacter Shaeroide S, a photosynthetic bacterium, at a fixed

dilution rate of 0.4 /day. The results indicated that 96% of the oil was removed from the

wastewater, and the maximum removal rate of oil reached approximately 3.83 kg

oil/m3/day. The study showed that immobilized photosynthetic bacteria directly

decomposed one part of the oil contained in the wastewater, while oil-decomposing

bacteria (naturally growing Pseudomonas and Bacillus) decomposed another part of the

oil, degrading it to organic fatty acids such as propionic and acetic acid. Photosynthetic

bacteria consumed these acids. As a result, oil degradation was promoted by the addition

of photosynthetic bacteria and the combination showed significant decomposition of oil

from wastewater.

Prasertasan et al. (2004) investigated the bioseparation of suspended solids and oil

from POME. Two thermotolerant lipase producing fungal isolates, ST4 and ST29, were

observed to remove 98.9% and 88.9% of oil, respectively. The high treatment efficiency

was found to be the consequence of the entrapment of the majority of suspended solids

and residual oil of POME in the polymer excreted from these strains. Lipases are widely

34

produced in fungi and their production and activity in organisms such as Rhizopus,

Mucor and Aspergillus, have been well studied (Jensen 1971; Akhtar et al. 1980; Chopra

and Khuller 1983). In a study conducted by Gopinath et al. (2005) on thirty-four wild

fungal species associated with edible oil mill waste, the species Absidia colymbifera,

Aspergillus fumigatus, Aspergillus japonicus, Aspergillus nidulans, Aspergillus terreus,

Cunninghamella verticillata, Curvularia pallescens, Fusarium oxysporum, Geotrichum

candidum, Mucor racemosus, Penicillium citrinum, Penicillium frequentans, Rhizopus

stolonifer, and Trichoderma viride were found to exhibit maximum lipase activity. The

literature advises that lipase-producing fungi can utilize oil as a main carbon source,

which has approximately twice the energy value of glucose, and are capable of reducing

the COD of oil effluent (Koritala et al. 1987; Ratledge 1992). When fungi are confronted

with fats and oils in a growth medium, triacylglycerol is hydrolyzed by fungal lipases to

yield diacylglycerol, monoacylglycerols, free fatty acids and glycerol.

Mucor circinelloides was found to effectively degrade a laboratory-based

sunflower oil waste (Joseph et al. 2005). The fungus was found to exhibit a preference

towards the utilization of unsaturated fatty acids in the presence of acetate. The study

suggested that fungi could be used in oil waste removal while, at the same time, biomass,

rich in essential fatty acids such as linoleic acid and GLA, can also be produced. Fabritius

et al. (1998) studied bioconversions of sunflower oil and rapeseed oil in fed-batch

cultures fermented with Candida tropicalis. Co-fermentations with palmitic acid resulted

in successful transformations to different 3- hydroxydioic acids.

35

produced in fungi and their production and activity in organisms such as Rhizopus,

Mucor and Aspergillus, have been well studied (Jensen 1971; Akhtar et al. 1980; Chopra

and Khuller 1983). In a study conducted by Gopinath et al. (2005) on thirty-four wild

fungal species associated with edible oil mill waste, the species Absidia corymbifera,

Aspergillus fumigatus, Aspergillus japonicus, Aspergillus nidulans, Aspergillus terreus,

Cunninghamella verticillata, Curvularia pallescens, Fusarium oxysporum, Geotrichum

candidum, Mucor racemosus, Penicillium citrinum, Penicillium frequentans, Rhizopus

stolonifer, and Trichoderma viride were found to exhibit maximum lipase activity. The

literature advises that lipase-producing fungi can utilize oil as a main carbon source,

which has approximately twice the energy value of glucose, and are capable of reducing

the COD of oil effluent (Koritala et al. 1987; Ratledge 1992). When fungi are confronted

with fats and oils in a growth medium, triacylglycerol is hydrolyzed by fungal lipases to

yield diacylglycerol, monoacylglycerols, free fatty acids and glycerol.

Mucor circinelloides was found to effectively degrade a laboratory-based

sunflower oil waste (Joseph et al. 2005). The fungus was found to exhibit a preference

towards the utilization of unsaturated fatty acids in the presence of acetate. The study

suggested that fungi could be used in oil waste removal while, at the same time, biomass,

rich in essential fatty acids such as linoleic acid and GLA, can also be produced. Fabritius

et al. (1998) studied bioconversions of sunflower oil and rapeseed oil in fed-batch

cultures fermented with Candida tropicalis. Co-fermentations with palmitic acid resulted

in successful transformations to different 3- hydroxydioic acids.

35

The ability of Aspergillus fumigatus Fres. and Aspergillus nidulans (Eidam) wint,

from Nigerian palm products, to degrade vegetable oils was examined (Ogundero 1982).

Both species were found to readily hydrolyze palm oil and palm kernel oil and good

growth was recorded with respect to triglycerides, which were used as a source of carbon.

Extracellular lipases were detected and found to be active at a pH level of 5.6. Barker and

Worgan (1981) investigated the growth of Aspergillus oryzae on palm oil effluent. The

fungus grew in a batch culture indicating the utilization of principal substrate

components, followed by a period of protein assimilation and then, lipid and

carbohydrate utilization. Supplementation with an inorganic nitrogen source was found to

be necessary and BOD reductions of 75% and COD reductions of 75% to 80% were

observed in batch cultures.

Aspergillus niger was observed to posses extracellular enzymes that hydrolyze

pectins, polyphenols and tannins. It degrades many phenolic compounds and can be used

in the pretreatment of olive black water (Hamdi et al. 1991). The organic fraction of olive

oil mill waste includes sugars, tannins, polyphenols, polyalcohols, pectins and lipids

(Hamdi 1993). Hamdi et al. (1991) observed a reduction in COD of 52.5% using A. niger

in aerobic conditions. Average COD removals of 55%, 52.5% and 62.8% in olive mill

wastewater, fermented with Geotrichium sp, Aspergillus sp, Candida tropicalis,

respectively, were observed by Fadil et al. (2003). Under optimum conditions, C.

versicolor and F. trogii removed 63% and 70% of COD, respectively (Yesilada et al.

1998). The addition of sulphate, glucose and nitrogen had no effect upon biodegradation

and immobilized fungi in calcium alginate gels, obtained at high yields. In a study

36

The ability of Aspergillus fumigatus Fres. and Aspergillus nidulans (Eidam) wint,

from Nigerian palm products, to degrade vegetable oils was examined (Ogundero 1982).

Both species were found to readily hydrolyze palm oil and palm kernel oil and good

growth was recorded with respect to triglycerides, which were used as a source of carbon.

Extracellular lipases were detected and found to be active at a pH level of 5.6. Barker and

Worgan (1981) investigated the growth of Aspergillus oryzae on palm oil effluent. The

fungus grew in a batch culture indicating the utilization of principal substrate

components, followed by a period of protein assimilation and then, lipid and

carbohydrate utilization. Supplementation with an inorganic nitrogen source was found to

be necessary and BOD reductions of 75% and COD reductions of 75% to 80% were

observed in batch cultures.

Aspergillus niger was observed to posses extracellular enzymes that hydrolyze

pectins, polyphenols and tannins. It degrades many phenolic compounds and can be used

in the pretreatment of olive black water (Hamdi et al. 1991). The organic fraction of olive

oil mill waste includes sugars, tannins, polyphenols, polyalcohols, pectins and lipids

(Hamdi 1993). Hamdi et al. (1991) observed a reduction in COD of 52.5% using A. niger

in aerobic conditions. Average COD removals of 55%, 52.5% and 62.8% in olive mill

wastewater, fermented with Geotrichium sp, Aspergillus sp, Candida tropicalis,

respectively, were observed by Fadil et al. (2003). Under optimum conditions, C.

versicolor and F. trogii removed 63% and 70% of COD, respectively (Yesilada et al.

1998). The addition of sulphate, glucose and nitrogen had no effect upon biodegradation

and immobilized fungi in calcium alginate gels, obtained at high yields. In a study

36

conducted by Yesilada et al. (1999), while treating OMW aerobically with fungi, it was

found that Coriolus versicolor, Funalia trogii and Pleurotus sarorcaju showed high

reductions in COD. The ability of Phanerochaete chrysosporium to degrade the phenolic

compounds of olive oil mill wastewater, using cells immobilized on loofah, was

examined by Ahmadi et al. (2006). The fungus did not grow on the concentrated

wastewater and a reduction in COD of 5% was observed. Pleurotus ostreatus grown in

bioreactor batch cultures in a model OMW (diluted and sterilized) was observed to cause

significant phenolic removal (Aggelis et al. 2003). Laccase, the sole ligninolytic enzyme

detected in the growth environment, was found during primary metabolic growth. The

toxicity of OMW against the seeds of Lepidium sativum and the marine Branchiopoda

Artemia sp. was observed to decrease after biotreatment. Pleurotus ostreatus was able to

reduce phenolic content and the toxicity of sterilized OMW, in bioreactor cultures.

During a study carried out by Zheng et al. (2003) regarding the treatment of salad oil

manufacturing wastewater, using yeast isolates, a large loss of biomass was observed

which subsequently reduced the treatment efficiency. Of the five yeast species used in the

study, only Candida tropicalis remained in the aeration tank, possibly because it had

better settleability. An addition of nitrogen to the final effluent stimulated the activity of

this yeast, which resulted in treatment efficiency similar to that of the mixed yeast

system. Lanciotti et al. (2005) examined the ability of different Yarrowia lipolytica yeast

strains to grow in OMW and to reduce its COD level. Results showed that several

Yarrowia lipolytica strains are good candidates for the reduction of the pollution potential

of OMW and for the production of enzymes and metabolites such as lipase and citric

37

conducted by Yesilada et al. (1999), while treating OMW aerobically with fungi, it was

found that Coriolus versicolor, Funalia trogii and Pleurotus sarorcaju showed high

reductions in COD. The ability of Phanerochaete chrysosporium to degrade the phenolic

compounds of olive oil mill wastewater, using cells immobilized on loofah, was

examined by Ahmadi et al. (2006). The fungus did not grow on the concentrated

wastewater and a reduction in COD of 5% was observed. Pleurotus ostreatus grown in

bioreactor batch cultures in a model OMW (diluted and sterilized) was observed to cause

significant phenolic removal (Aggelis et al. 2003). Laccase, the sole ligninolytic enzyme

detected in the growth environment, was found during primary metabolic growth. The

toxicity of OMW against the seeds of Lepidium sativum and the marine Branchiopoda

Artemia sp. was observed to decrease after biotreatment. Pleurotus ostreatus was able to

reduce phenolic content and the toxicity of sterilized OMW, in bioreactor cultures.

During a study carried out by Zheng et al. (2003) regarding the treatment of salad oil

manufacturing wastewater, using yeast isolates, a large loss of biomass was observed

which subsequently reduced the treatment efficiency. Of the five yeast species used in the

study, only Candida tropicalis remained in the aeration tank, possibly because it had

better settleability. An addition of nitrogen to the final effluent stimulated the activity of

this yeast, which resulted in treatment efficiency similar to that of the mixed yeast

system. Lanciotti et al. (2005) examined the ability of different Yarrowia lipolytica yeast

strains to grow in OMW and to reduce its COD level. Results showed that several

Yarrowia lipolytica strains are good candidates for the reduction of the pollution potential

of OMW and for the production of enzymes and metabolites such as lipase and citric

37

acid. In another study conducted on Yarrowia lipolytica, it was observed that the yeast

reduced the COD level of OMW by 80% and produced a useful biomass and enzyme

lipase (Sciolli and Vollaro 1997).

2.2.2.1 Biodegradation of Oil under Thermophilic Conditions

At temperatures above 50°C, the oil aggregates melt, and stable emulsions of

substrates with large surface areas are formed during agitation. Under such conditions,

the bioavailability of such hydrophobic substrates for enzymes and microorganisms is

significantly increased. From a technological point of view, high temperatures are

desirable, because the viscosity of the streams decrease, and thus, diffusion and mass

transfer are accelerated (Becker et al. 1999). The biological treatment of oils under

thermophilic conditions, i.e., above 60 °C, is expected to be advantageous due to

favorable changes in physical properties of the hydrophobic compounds with increasing

temperature.

Becker et al. (1999) studied the aerobic thermophilic degradation of olive oil

using a pure culture of bacterial strain, Bacillus thermoleovorans 1111-91, in a

continuously operated stirred tank reactor. It was observed that an oil removal of more

than 90% of the initial oil concentration of 2 g/L was obtained at a residence time of 2

hours. A relatively high maximum biomass yield of 1.05 g dry cell per g of olive oil

consumed was measured in the study. A severe growth inhibition was observed when the

feed olive oil concentration was increased to more than 4 g/L. Compared with data from

mesophilic processes, the oil degradation rates obtained under thermophilic conditions,

38

acid. In another study conducted on Yarrowia lipolytica, it was observed that the yeast

reduced the COD level of OMW by 80% and produced a useful biomass and enzyme

lipase (Sciolli and Vollaro 1997).

2.2.2.1 Biodegradation of Oil under Thermophilic Conditions

At temperatures above 50°C, the oil aggregates melt, and stable emulsions of

substrates with large surface areas are formed during agitation. Under such conditions,

the bioavailability of such hydrophobic substrates for enzymes and microorganisms is

significantly increased. From a technological point of view, high temperatures are

desirable, because the viscosity of the streams decrease, and thus, diffusion and mass

transfer are accelerated (Becker et al. 1999). The biological treatment of oils under

thermophilic conditions, i.e., above 60 °C, is expected to be advantageous due to

favorable changes in physical properties of the hydrophobic compounds with increasing

temperature.

Becker et al. (1999) studied the aerobic thermophilic degradation of olive oil

using a pure culture of bacterial strain, Bacillus thermoleovorans IHI-91, in a

continuously operated stirred tank reactor. It was observed that an oil removal of more

than 90% of the initial oil concentration of 2 g/L was obtained at a residence time of 2

hours. A relatively high maximum biomass yield of 1.05 g dry cell per g of olive oil

consumed was measured in the study. A severe growth inhibition was observed when the

feed olive oil concentration was increased to more than 4 g/L. Compared with data from

mesophilic processes, the oil degradation rates obtained under thermophilic conditions,

were extremely high. Lapara and Alleman (1999) attributed this difference to the fact that

nitrifying bacteria, floc-forming organisms and higher organisms that aid flocculations

were not present, which affected sludge sedimentation. Markossian et al. (2000) carried

out a study regarding the same strain, Bacillus thermoleovorans IHI-91, and observed

that the aerobic bacterium grew optimally at 65°C with a pH of 6.0 and secreted a high

level of lipase. The strain was observed to utilize several oils such as olive oil, sunflower

oil, soya oil, and fish oil as sole carbon and energy sources without an additional supply

of growth factors. A degradation of about 93% of triolein, which is present in olive oil,

was observed after only 7 hours of fermentation, at a maximal growth rate of 1.0 /h. Due

to its production of high concentrations of thermoactive lipases and esterases, the

bacterium was found to be capable of degrading a wide range of oils at high

temperatures. Results related to the microbial degradation of oil from wastewater are

summarized in Table 2.2.

2.2.3 Removal of Oil by Various Sorbents

In the past, many sorbents such as activated carbon, peat, bentonit, and

organoclay have been used to remove oil from wastewater. Activated carbon is an •

adsorbent that is commonly used in the removal of a wide variety of organic compounds

including oil from water (Ahmad et al. 2005b; Alther 1995; Alther 1996; Mathavan and

Viraraghavan 1989; Viraraghavan and Moazed 2003; Zunan et al. 1995). The

disadvantage of using activated carbon in removing oil from water is that emulsified oil

can bind its pore spaces during operation (Mysore et al. 2006).

39

were extremely high. Lapara and Alleraan (1999) attributed this difference to the fact that

nitrifying bacteria, floc-forming organisms and higher organisms that aid flocculations

were not present, which affected sludge sedimentation. Markossian et al. (2000) carried

out a study regarding the same strain, Bacillus thermoleovorans IHI-91, and observed

that the aerobic bacterium grew optimally at 65°C with a pH of 6.0 and secreted a high

level of lipase. The strain was observed to utilize several oils such as olive oil, sunflower

oil, soya oil, and fish oil as sole carbon and energy sources without an additional supply

of growth factors. A degradation of about 93% of triolein, which is present in olive oil,

was observed after only 7 hours of fermentation, at a maximal growth rate of 1.0 /h. Due

to its production of high concentrations of thermoactive lipases and esterases, the

bacterium was found to be capable of degrading a wide range of oils at high

temperatures. Results related to the microbial degradation of oil from wastewater are

summarized in Table 2.2.

2.2.3 Removal of Oil by Various Sorbents

In the past, many sorbents such as activated carbon, peat, bentonit, and

organoclay have been used to remove oil from wastewater. Activated carbon is an

adsorbent that is commonly used in the removal of a wide variety of organic compounds

including oil from water (Ahmad et al. 2005b; Alther 1995; Alther 1996; Mathavan and

Viraraghavan 1989; Viraraghavan and Moazed 2003; Zunan et al. 1995). The

disadvantage of using activated carbon in removing oil from water is that emulsified oil

can bind its pore spaces during operation (Mysore et al. 2006).

Table 2.2: Microbial degradation

Microorganism Source Efficiency/ Remarks Reference

Acinetobacter sp. Rhodococcus sp.

Corn, olive, sunflower and

Corn and waste oil supported all bacterial

Keenan and Sabelnikov

Caseobacter sp. waste oils strains than olive and

(2000) sunflower oil

Acinetobacter P. Corn and waste oil

Reduce lipid content from 1.5 g/L to 0.35 g/L for corn and to 0.267 g/L for waste oil.

Wakelin and Forster (1997)

Pseudomonas sp. Pseudomonas sp. And P. diminuta P. pseudoalcaligens Vegetable oil P. diminuta combined,

reduced COD by 93% and El-Bestawy et al. (2005)

Escherichia sp. BOD by 100% Rhodobacter Domestic Takeno et al. shaeroide S wastewater

96% oil removal (2005)

Mucor circinelloides Sunflower oil waste

Presence of acetate enhanced sunflower oil utilization

Joseph et al. (2005)

Aspergillus fumigatus Aspergillus nidulans

Vegetable oils Hydrolyze palm and palm kernel oil

Ogundero (1982)

Barker and Aspergillus oryzae POME 75-80% COD reduction Worgan

(1981)

Aspergillus niger OMW 52.5% COD reduction Hamdi et al. (1991)

C. versicolor OMW

63% COD reduction Yesilada et F. trogii 70% COD reduction al. (1998) C. versicolor F. trogii P. sarorcaju

OMW High COD reduction Yesilada et al. (1999)

Phanerochaete chlysosporium OMW 50% COD reduction Ahmadi et al.

(2006) Scioli and Vollaro

Yarrowia lipolytica OMW 80% COD reduction (1997); Lanciotti et al. (2005)

40

Table 2.2: Microbial degradation

Microorganism Source Efficiency/ Remarks Reference

Acinetobacter sp. Rhodococcus sp. Caseobacter sp.

Acinetobacter sp.

Pseudomonas sp. P. diminuta P. pseudoalcaligens Escherichia sp. Rhodobacter shaeroide S

Mucor circinelloides

Aspergillus fumigatus Aspergillus nidulans

Aspergillus oryzae

Aspergillus niger

C. versicolor F. trogii C. versicolor F. trogii P. sarorcaju Phanerochaete chrysosporium

Yarrowia lipolytica

Corn, olive, sunflower and waste oils

Corn and waste oil

Vegetable oil

Domestic wastewater

Sunflower oil waste

Vegetable oils

POME

OMW

OMW

OMW

OMW

OMW

Corn and waste oil supported all bacterial strains than olive and sunflower oil Reduce lipid content from 1.5 g/L to 0.35 g/L for corn and to 0.267 g/L for waste oil. Pseudomonas sp. And P. diminuta combined, reduced COD by 93% and BOD by 100%

96% oil removal

Presence of acetate enhanced sunflower oil utilization

Hydrolyze palm and palm kernel oil

75-80% COD reduction

52.5% COD reduction

63% COD reduction 70% COD reduction

High COD reduction

50% COD reduction

80% COD reduction

Keenan and Sabelnikov (2000)

Wakelin and Forster (1997)

El-Bestawy et al. (2005)

Takeno et al. (2005)

Joseph et al. (2005)

Ogundero (1982)

Barker and Worgan (1981) Hamdi et al. (1991) Yesilada et al. (1998)

Yesilada et al. (1999)

Ahmadi et al. (2006) Scioli and Vollaro (1997); Lanciotti et al. (2005)

40

Crushed and/or processed plant materials have been used to adsorb oil from surfaces for

some time. For example, cotton linters, crushed corncobs and modified pulp have been

investigated (Johnson et al. 1973; Silva-Tilak 2002; Sun et al. 2002). Deschamps et al.

(2003) used cotton as a bed material to recover oil from oily water and found the oil—

water separation could be improved at a lower temperature, lower flow, a deeper bed

material and larger oil drop. Mathavan and Viraraghavan (1992) used horticultural peat

as a bed material to examine the capability of the filter bed in oily water clean-up.

Moazed and Viraraghavan (2002) examined the ability of an anthracite/organoclay

mixture in removing oil from water. Varghese and Cleveland (1998) studied the

feasibility of using a kenaf-filled filter in removing oil from a surfactant-stabilized oil-in

water emulsion. The possibilities of reed canary grass, flax and hemp fiber as oil

absorbing filter materials were investigated by Pasila (2004). It was found that these

materials were able to separate oil from water during the filtration of oily water.

Oil removal efficiencies of 98.2%, 46.3%, 86.5%, 98.2 %, 99.2%, 92.4%, and

94.2% were obtained for peat, coal, sand, amberlite, activated carbon, polypropylene, and

fiberglass, respectively (Viraraghavan and Mathavan 1990). Sodium bentonite was found

to be effective in removing oil and efficiencies of 88-94% for Standard Mineral Oil

(SMO), 85-96% and 87-96% for two cutting oils, 84-86% for produced water from

production wells, and 54-87% for refinery effluent were obtained (Viraraghavan and

Moazed 2003). Moazed and Viraraghavan (2001) showed that bentonite organo-

clay/anthracite were quite effective in breaking down a number of oil-in-water emulsions.

In a recent study, Ahmad et al. (2005a) investigated adsorption of residue oil from Palm

41

Crushed and/or processed plant materials have been used to adsorb oil from surfaces for

some time. For example, cotton linters, crushed corncobs and modified pulp have been

investigated (Johnson et al. 1973; Silva-Tilak 2002; Sun et al. 2002). Deschamps et al.

(2003) used cotton as a bed material to recover oil from oily water and found the oil-

water separation could be improved at a lower temperature, lower flow, a deeper bed

material and larger oil drop. Mathavan and Viraraghavan (1992) used horticultural peat

as a bed material to examine the capability of the filter bed in oily water clean-up.

Moazed and Viraraghavan (2002) examined the ability of an anthracite/organoclay

mixture in removing oil from water. Varghese and Cleveland (1998) studied the

feasibility of using a kenaf-filled filter in removing oil from a surfactant-stabilized oil-in

water emulsion. The possibilities of reed canary grass, flax and hemp fiber as oil

absorbing filter materials were investigated by Pasila (2004). It was found that these

materials were able to separate oil from water during the filtration of oily water.

Oil removal efficiencies of 98.2%, 46.3%, 86.5%, 98.2 %, 99.2%, 92.4%, and

94.2% were obtained for peat, coal, sand, amberlite, activated carbon, polypropylene, and

fiberglass, respectively (Viraraghavan and Mathavan 1990). Sodium bentonite was found

to be effective in removing oil and efficiencies of 88-94% for Standard Mineral Oil

(SMO), 85-96% and 87-96% for two cutting oils, 84—86% for produced water from

production wells, and 54-87% for refinery effluent were obtained (Viraraghavan and

Moazed 2003). Moazed and Viraraghavan (2001) showed that bentonite organo-

clay/anthracite were quite effective in breaking down a number of oil-in-water emulsions.

In a recent study, Ahmad et al. (2005a) investigated adsorption of residue oil from Palm

41

Oil Mill Effluent (POME) using chitosan from crab shells. Chitosan was found to remove

almost 99% of residue oil from POME. Chitosan, in powder form, had a higher capacity

for residue oil compared to the flake form. Several of the physicochemical parameters of

adsorption were evaluated at dynamic and equilibrium conditions. The results of this

study can be useful with respect to further applications in treatments of oily wastewaters

(Ahmad et al. 2005a). Studies have been conducted to evaluate the use of walnut shell

filters as an alternative to conventional sand filters to remove free oils and suspended

solids (Blumenschein et al. 2001). USFilter Zimpro products, USA have developed an

Auto-Shell Walnut shell filter to remove oily contaminants from water (USFilter

Corporation 2005). The walnut shell media was found to have good oil adsorbent ability

and low media gravity (Yang et al. 2002). The media have been observed to be tough

enough to maintain their size during the back washing operation, thereby reducing

possible loss during the filtration process (Rahman 1992). Srinivasan and Viraraghavan

(2008) conducted batch kinetic studies to evaluate the equilibrium time required by

walnut shell media to sorb oil. As for pure oil medium, sorption capacities of 0.30 g/g,

0.51 g/g and 0.58 g/g were obtained for standard mineral oil, canola oil and Bright- Edge

oil, respectively. Walnut shells remove oil by coalescing oil droplets but do not bind with

them. Similar sorption capacity tests have been conducted by Mysore et al. (2004) on

vermiculite for both pure oil and oil on aqueous medium. The external surface of the

hydrophobic vermiculite had a wax coating, which resulted in adsorption followed by

absorption of oil (Mysore et al. 2004). Oil sorption capacity values of different materials

are provided in Table 2.3.

42

Oil Mill Effluent (POME) using chitosan from crab shells. Chitosan was found to remove

almost 99% of residue oil from POME. Chitosan, in powder form, had a higher capacity

for residue oil compared to the flake form. Several of the physicochemical parameters of

adsorption were evaluated at dynamic and equilibrium conditions. The results of this

\

study can be useful with respect to further applications in treatments of oily wastewaters

(Ahmad et al. 2005a). Studies have been conducted to evaluate the use of walnut shell

filters as an alternative to conventional sand filters to remove free oils and suspended

solids (Blumenschein et al. 2001). USFilter Zimpro products, USA have developed an

Auto-Shell Walnut shell filter to remove oily contaminants from water (USFilter

Corporation 2005). The walnut shell media was found to have good oil adsorbent ability

and low media gravity (Yang et al. 2002). The media have been observed to be tough

enough to maintain their size during the back washing operation, thereby reducing

possible loss during the filtration process (Rahman 1992). Srinivasan and Viraraghavan

(2008) conducted batch kinetic studies to evaluate the equilibrium time required by

walnut shell media to sorb oil. As for pure oil medium, sorption capacities of 0.30 g/g,

0.51 g/g and 0.58 g/g were obtained for standard mineral oil, canola oil and Bright- Edge

oil, respectively. Walnut shells remove oil by coalescing oil droplets but do not bind with

them. Similar sorption capacity tests have been conducted by Mysore et al. (2004) on

vermiculite for both pure oil and oil on aqueous medium. The external surface of the

hydrophobic vermiculite had a wax coating, which resulted in adsorption followed by

absorption of oil (Mysore et al. 2004). Oil sorption capacity values of different materials

are provided in Table 2.3.

42

Table 2.3: Oil sorption capacity of different media

Media Oil studied Sorption capacity (g/g)

Reference

Penalia Peat Finland SMO 3.42 Thun et al (1983) Saskatchewan peat SMO 7.85 Mathavan and Canada Viraraghavan (1990) Expanded vermiculite SMO 2.53 Hydrophobic vermiculite

SMO 2.45

Expanded vermiculite Canola 2.57 Hydrophobic vermiculite

Canola 2.48 Mysore et al. (2004)

Expanded vermiculite Kutwell 45 2.62 Hydrophobic vermiculite

Kutwell 45 2.53

Oclansorb (treated natural peat)

Crude 2.23

Zugol (modified pine bark)

Crude 1.56 Environment Canada

Verdyol (treated vegetable fibre)

Crude 0.52 (1985)

Seaclean (natural feathers)

Crude 0.65

Walnut shell media SMO 0.30 Walnut shell media Canola 0.51 Srinivasan and

Walnut shell media Bright-Edge 80 0.58 Viraraghavan (2008)

Hemp fiber and reed canary grass

Fuel oil No. 2 2.0 - 4.0 Pasila (2004)

Hollow cellulosic B - heavy oil 11.1 Kobayashi et al. kapok fiber Machine oil 7.80 (1977) Bagasse mesh Light crude 3.38

Gas oil 4.07 Bayab et al. (2005) Heavy crude 5.30

Polypropylene Light crude 8.26 nonwoven mesh Gas oil 8.46 Bayab et al. (2005)

Heavy crude 9.12 Rice hull mesh Light crude 3.80 Bayab et al. (2005)

Gas oil 3.81 Heavy crude 5.15

Salvinia sp. Vaseline Marlin oil

7.30 11.60 Ribeiro et al. (2003)

43

Table 2.3: Oil sorption capacity of different media


Reference

Penalia Peat Finland SMO 3.42 Thun et al (1983) Saskatchewan peat SMO 7.85 Mathavan and Canada Viraraghavan (1990) Expanded vermiculite SMO 2.53 Hydrophobic SMO 2.45 vermiculite Expanded vermiculite Canola 2.57 Hydrophobic Canola 2.48 Mysore et al. (2004) vermiculite Expanded vermiculite Kutwell 45 2.62 Hydrophobic Kutwell 45 2.53 vermiculite Oclansorb (treated Crude 2.23 natural peat) Zugol (modified pine Crude 1.56 bark) Environment Canada Verdyol (treated Crude 0.52 (1985) vegetable fibre) Seaclean (natural Crude 0.65 feathers) Walnut shell media SMO 0.30

Srinivasan and Walnut shell media Canola 0.51

Srinivasan and

Walnut shell media Bright-Edge 80 0.58 Viraraghavan (2008)

Hemp fiber and reed Fuel oil No. 2 2.0-4.0 Pasila (2004) canary grass Hollow cellulosic B - heavy oil 11.1 Kobayashi et al. kapok fiber Machine oil 7.80 (1977) Bagasse mesh Light crude 3.38

Gas oil 4.07 Bayab et al. (2005) Heavy crude 5.30

Polypropylene Light crude 8.26 nonwoven mesh Gas oil 8.46 Bayab et al. (2005)

Heavy crude 9.12 Rice hull mesh Light crude 3.80 Bayab et al. (2005)

Gas oil 3.81 Heavy crude 5.15

Salvinia sp. Vaseline 7.30 Ribeiro et al. (2003)

Salvinia sp. Marlin oil 11.60

Ribeiro et al. (2003)

43

Table 23: Oil sorption capacity of different media (Contd.)


Reference

Chitosan powder POME 0.60 Ahmad et al. (2005a) Chitosan flake POME 0.50 Kapok fiber Gas station runoff 827 Cattail fiber Gas station runoff 1107 Salvinia sp. Gas station runoff 944 Polyester fiber Wood chip

Gas station runoff Gas station runoff

1008 343

Khan et al. (2004)

Rice husk Gas station runoff 298 Coconut husk Gas station runoff 58 Bagasse Gas station runoff 19 Felt A — grade heavy oil 17 Activated carbon fibers Cotton fiber fabric

A — grade heavy oil

A — grade heavy oil

20

7.40 Inagaki et al. (2002)

Exfoliated graphite A — grade heavy oil 83 B — grade heavy oil 67

Ground chrome Diesel motor oil 7.0 shavings Premium motor oil 6.5

Used oil 7.6 Gammoun et al. Raw chrome Diesel motor oil 3.7 (2007) shavings Premium motor oil 3.8

Used oil 3.8 Loose natural wool Oily wastewater 5.56 Wool based nonwoven materials

Oily wastewater 5.48 Rajakovic et al. (2007)

Sepiolite Oily wastewater 0.19 SDS modified chitosan

Cutting fluid effluent 2.50 Piyamongkala et al. (2008)

Polyacrylonitrile Motor oil SAE -30 10 fiber Multigrade engine oil 12

Modified Motor oil SAE — 50 Motor oil SAE -30

14 17 Ji et al. (2009)

polyacrylonitrile Multigrade engine oil 16 fiber Motor oil SAE — 50 18 Recycled wool based Diesel fuel 9.62 nonwoven material Crude oil

Base oil 11.06 12.98

Radetic et al. (2008)

Vegetable oil 13.16

44

Table 2.3: Oil sorption capacity of different media (Contd.)

Media Oil studied Sorption Reference capacity (R/R)

Chitosan powder POME 0.60 Ahmad et al. (2005a) Chitosan flake POME 0.50 Kapok fiber Gas station runoff 827 Cattail fiber Gas station runoff 1107 Salvinia sp. Gas station runoff 944 Polyester fiber Gas station runoff 1008 IfHnn *»t nl Wood chip Gas station runoff 343

jsjidn ei di.

Rice husk Gas station runoff 298 Coconut husk Gas station runoff 58 Bagasse Gas station runoff 19 Felt A - grade heavy oil 17 Activated carbon A - grade heavy oil 20 fibers Cotton fiber fabric A - grade heavy oil 7.40

Inagaki et al. (2002)

Exfoliated graphite A - grade heavy oil 83 B - grade heavy oil 67

Ground chrome Diesel motor oil 7.0 shavings Premium motor oil 6.5

Used oil 7.6 Gammoun et al. Raw chrome Diesel motor oil 3.7 (2007) shavings Premium motor oil 3.8

Used oil 3.8 Loose natural wool Oily wastewater 5.56 Wool based Oily wastewater 5.48 Rajakovic et al. nonwoven materials (2007) Sepiolite Oily wastewater 0.19 SDS modified Cutting fluid effluent 2.50 Piyamongkala et al. chitosan (2008) Polyacrylonitrile Motor oil SAE -30 10 fiber Multigrade engine oil 12

Motor oil SAE - 50 14 Ji et al. (2009) Modified Motor oil SAE -30 17 Ji et al. (2009)

polyacrylonitrile Multigrade engine oil 16 fiber Motor oil SAE - 50 18 Recycled wool based Diesel fuel 9.62 nonwoven material Crude oil 11.06

Radetic et al. (2008) Base oil 12.98

Radetic et al. (2008)

Vegetable oil 13.16

44

2.2.4 Summary

A review of the literature showed conventional treatment methods such as

dissolved air flotation and biological treatment are generally effective in the removal of

oil. Adsorption methods are equally effective, especially for smaller systems. Activated

carbon is not only costly but also creates binding problems. Therefore, a search for less-

costly alternative adsorbents has opened a new area of investigation where biosorbents

have been studied. However, no research with respect to the use of microbial sorbents for

oil removal has been undertaken. Although live biomass is part of the biological

treatment for oil removal, no studies regarding the use of non-viable microbial biomass

for oil removal exist. Such a study would fill a void in this field.

2.3 Adsorption Processes in Environmental Engineering

2.3.1 Theoretical Background

Adsorption is a surface phenomenon that is defined as the accumulation of

dissolved substances at interfaces or between phases. The dissolved substance being

adsorbed is called "adsorbate" and the solid phase on which the accumulation of

adsorbate takes place is called "adsorbent". Adsorption of dissolved substances on a

microbial cell surface has been defined as "biosorption". Adsorption takes place as a

result of either a solvent motivated force, which relates to surface tension, and/or an

adsorbent motivated force that combines chemical, electrostatic, and physical interactions

between an adsorbate and an adsorbent (Weber and DiGiano 1996). Chemisorption

involves the formation of a chemical bond between the adsorbate and the surface while

45

2.2.4 Summary

A review of the literature showed conventional treatment methods such as

dissolved air flotation and biological treatment are generally effective in the removal of

oil. Adsorption methods are equally effective, especially for smaller systems. Activated

carbon is not only costly but also creates binding problems. Therefore, a search for less-

costly alternative adsorbents has opened a new area of investigation where biosorbents

have been studied. However, no research with respect to the use of microbial sorbents for

oil removal has been undertaken. Although live biomass is part of the biological

treatment for oil removal, no studies regarding the use of non-viable microbial biomass

for oil removal exist. Such a study would fill a void in this field.

2.3 Adsorption Processes in Environmental Engineering

2.3.1 Theoretical Background

Adsorption is a surface phenomenon that is defined as the accumulation of

dissolved substances at interfaces or between phases. The dissolved substance being

adsorbed is called "adsorbate" and the solid phase on which the accumulation of

adsorbate takes place is called "adsorbent". Adsorption of dissolved substances on a

microbial cell surface has been defined as "biosorption". Adsorption takes place as a

result of either a solvent motivated force, which relates to surface tension, and/or an

adsorbent motivated force that combines chemical, electrostatic, and physical interactions

between an adsorbate and an adsorbent (Weber and DiGiano 1996). Chemisorption

involves the formation of a chemical bond between the adsorbate and the surface while

adsorption, due to electrostatic interaction, is the result of coulombic forces of attraction

between the oppositely charged ionic species of the adsorbent and the adsorbate.

Physisorption involves weaker interactions that occur as a result of van der Waals forces

of attraction between the adsorbate and adsorbent rather than electron transfer between

the two.

2.3.2 Adsorption Kinetics

The adsorption process results in the removal of an adsorbate from a solution,

which simultaneously increases the concentration of an adsorbate at the surface of an

adsorbent. This process continues until the concentration of the adsorbate remaining in

the solution achieves a dynamic equilibrium with the concentration of adsorbent on the

adsorbent surface. The time required for this stage to be completed is called "equilibrium

time for adsorption". When the equilibrium stage is reached, adsorption process is

assumed to be complete and no further removal of adsorbate takes place.

2.3.2.1 The Legergren Model

Lagergren (1898) showed that the rate of adsorption of pollutants on the adsorbent

followed a pseudo-first order equation. The non-linear equation is expressed as follows:

q, = qe [1— exp (—kit)]

46

2.1

adsorption, due to electrostatic interaction, is the result of coulombic forces of attraction

between the oppositely charged ionic species of the adsorbent and the adsorbate.

Physisorption involves weaker interactions that occur as a result of van der Waals forces

of attraction between the adsorbate and adsorbent rather than electron transfer between

the two.

2.3.2 Adsorption Kinetics

The adsorption process results in the removal of an adsorbate from a solution,

which simultaneously increases the concentration of an adsorbate at the surface of an

adsorbent. This process continues until the concentration of the adsorbate remaining in

the solution achieves a dynamic equilibrium with the concentration of adsorbent on the

adsorbent surface. The time required for this stage to be completed is called "equilibrium

time for adsorption". When the equilibrium stage is reached, adsorption process is

assumed to be complete and no further removal of adsorbate takes place.

2.3.2.1 The Legergren Model

Lagergren (1898) showed that the rate of adsorption of pollutants on the adsorbent

followed a pseudo-first order equation. The non-linear equation is expressed as follows:

2.1

46

where k1 is the Lagergren rate constant for adsorption (10, qe is the amount of adsorbate

adsorbed at equilibrium (mg/g) and q, is the amount of adsorbate adsorbed (mg/g) at any

given time t (h).

2.3.2.2 The Ho Model

A pseudo-second order reaction rate equation was proposed by Ho and McKay

(1998) to study the kinetics of adsorption. The model involves the basic assumption that

an adsorption reaction on the surface of adsorbent is the rate-controlling step. The

equation is expressed as follows:

- k2tqe2 qt

1+(k2tqe2) 2.2

where k2 is the pseudo-second order adsorption rate constant (g/(mg h)) and qe and qr

were defined earlier.

2.3.2.3 The Weber and Morris Model

The intra-particle diffusion model proposed by Weber and Morris (1963) is

expressed by the following equation:

q,- k,t Y2 +C 2.3

where lc, is the intra-particle diffusion rate constant (mg/g h"2) and C is the intercept

(mg/g). A plot of q, versus t"2 will be a straight line with a slope k, and intercept C when

adsorption mechanism follows the intra-particle diffusion process.

47

where k\ is the Lagergren rate constant for adsorption (h-1), qe is the amount of adsorbate

adsorbed at equilibrium (mg/g) and qt is the amount of adsorbate adsorbed (mg/g) at any

given time t (h).

2.3.2.2 The Ho Model

A pseudo-second order reaction rate equation was proposed by Ho and McKay

(1998) to study the kinetics of adsorption. The model involves the basic assumption that

an adsorption reaction on the surface of adsorbent is the rate-controlling step. The

equation is expressed as follows:

where k2 is the pseudo-second order adsorption rate constant (g/(mg h)) and qe and q,

were defined earlier.

2.3.2.3 The Weber and Morris Model

The intra-particle diffusion model proposed by Weber and Morris (1963) is

expressed by the following equation:

where &,• is the intra-particle diffusion rate constant (mg/g h1/2) and C is the intercept

(mg/g). A plot of qt versus tm will be a straight line with a slope kt and intercept C when

adsorption mechanism follows the intra-particle diffusion process.

k2tq) 2.2

, * ^ q, -k t t +C 2.3

47

2.3.3 Adsorption Isotherm

An adsorption isotherm gives the relationship between the solid phase

concentration of an adsorbate or the amount of adsorbate adsorbed per unit weight of

adsorbent and the solution phase concentration of an adsorbate at equilibrium condition at

a particular temperature. The most common forms of adsorption isotherms used in

environmental engineering are the Langmuir and Freundlich models (Benefield et al.

1982).

2.3.3.1 The Langmuir Model

The Langmuir model can be described as follows (Langmuir, 1918; Weber,

1972):

QobCe (1+bCe )

2.4

where q is the amount of adsorbate adsorbed per unit mass of adsorbent (mg/g); Qo the

constant, indicating the mass of adsorbed solute completely required to saturate a unit

mass of adsorbent (mg/g); b the constant, related to the net enthalpy of adsorption

(L/mg); and Ce is the equilibrium solute concentration (mg/L). The Langmuir model was

developed based on assumptions: 1) the maximum adsorption corresponds to a saturated

monolayer of solute molecules on the adsorbent surface; 2) the energy of adsorption is

constant; 3) there is no transmigration of adsorbate in the plane of the surface; and 4) the

adsorption process is reversible. The Langmuir model applies strictly to homogenous

48

2.3.3 Adsorption Isotherm

An adsorption isotherm gives the relationship between the solid phase

concentration of an adsorbate or the amount of adsorbate adsorbed per unit weight of

adsorbent and the solution phase concentration of an adsorbate at equilibrium condition at

a particular temperature. The most common forms of adsorption isotherms used in

environmental engineering are the Langmuir and Freundlich models (Benefield et al.

1982).

2.3.3.1 The Langmuir Model

The Langmuir model can be described as follows (Langmuir, 1918; Weber,

1972):

q. Q°bC< 2.4 (1 + BC.)

where q is the amount of adsorbate adsorbed per unit mass of adsorbent (mg/g); Qo the

constant, indicating the mass of adsorbed solute completely required to saturate a unit

mass of adsorbent (mg/g); b the constant, related to the net enthalpy of adsorption

(L/mg); and Ce is the equilibrium solute concentration (mg/L). The Langmuir model was

developed based on assumptions: 1) the maximum adsorption corresponds to a saturated

monolayer of solute molecules on the adsorbent surface; 2) the energy of adsorption is

constant; 3) there is no transmigration of adsorbate in the plane of the surface; and 4) the

adsorption process is reversible. The Langmuir model applies strictly to homogenous

48

surfaces. The characteristics of the Langmuir isotherm could be expressed in terms of

separation factor RL that is defined as follows (Hall et al. 1966):

RL (1+ bC0 )

1 2.5

where Co is the initial concentration (mg/L) and b the Langmuir constant. According to

the value of RL, the isotherm is considered to be of the following types: RL > 1,

unfavorable; RL = 1, linear; 0 < RL < 1, favorable; and RL = 0, irreversible.

2.3.3.2 The Freundlich Isotherm

The Freundlich isotherm has a general form (Freundlich, 1906; Weber and

DiGiano, 1996):

q- K FCr 2.6

where q is the amount of adsorbate adsorbed per unit mass of adsorbent (mg/g); KF the

equilibrium constant indicative of adsorption capacity (mg/g); Ce the concentration of

adsorbate in solution at equilibrium; and n is the adsorption equilibrium constant

indicative of adsorption intensity.

2.3.4 Thermodynamic and Activation Parameters

2.3.4.1 Gibb's Free Energy Change

Gibbs Free Energy Change (G°) is the fundamental criterion of spontaneity.

Reaction occurs spontaneously at a given temperature if the value of AG° is negative. The

49

surfaces. The characteristics of the Langmuir isotherm could be expressed in terms of

separation factor Rl that is defined as follows (Hall et al. 1966):

where Co is the initial concentration (mg/L) and b the Langmuir constant. According to

the value of RL, the isotherm is considered to be of the following types: RL > 1,

unfavorable; RL = 1, linear; 0 <RL< 1, favorable; and RL = 0, irreversible.

2.3.3.2 The Freundlich Isotherm

The Freundlich isotherm has a general form (Freundlich, 1906; Weber and

DiGiano, 1996):

2.6

where q is the amount of adsorbate adsorbed per unit mass of adsorbent (mg/g); KF the

equilibrium constant indicative of adsorption capacity (mg/g); Ce the concentration of

adsorbate in solution at equilibrium; and n is the adsorption equilibrium constant

indicative of adsorption intensity.

2.3.4 Thermodynamic and Activation Parameters

2.3.4.1 Gibb's Free Energy Change

Gibbs Free Energy Change (G°) is the fundamental criterion of spontaneity.

Reaction occurs spontaneously at a given temperature if the value of AG0 is negative. The

49

value of AG° can be determined according to the following equation (Weber and

DiGiano, 1996):

AG° -RT: In b 2.7

where AG° is the Gibbs free energy change (kJ/mol), R the universal gas constant

(8.314x 10-3 kJ/mol-K, Te the absolute temperature (K) and b is the constant from the

Langmuir isotherm model (L/mol). Gibbs free energy change is related to enthalpy

change (AH°) and entropy change (AS°) and expressed by the following equation (Weber,

1972):

AG° AH° - TAS° 2.8

-RT, In b MI° -TAS° 2.9

where AH° is the heat of adsorption or enthalpy change (kJ/mol) and AS° is the entropy

change (kJ/mol-K).

2.3.4.2 Activation Parameters

The rate constant k for pseudo-second-order reaction shows an Arrhenius

dependence upon reciprocal temperature. The relationship can be expressed by:

Ink-InA+Ea/ RT 2.10

where k is the rate constant of adsorption, A is the Arrhenius factor, and E ad is the

activation energy (J/mol) for adsorption. The values of E ad and A are calculated from the

slope and intercept of a plot of Ink vs 1/T.

50

value of AG0 can be determined according to the following equation (Weber and

DiGiano, 1996):

AG0 "-RT e \nb 2.7

where AG0 is the Gibbs free energy change (kJ/mol), R the universal gas constant

(8.314xl0-3 kJ/mol-K, Te the absolute temperature (K) and b is the constant from the

Langmuir isotherm model (L/mol). Gibbs free energy change is related to enthalpy

change (Aff) and entropy change (AS0) and expressed by the following equation (Weber,

1972):

&G°-AH°-T eM° 2.8

-RTe In = Aif0 - TeAS° 2.9

where A//0 is the heat of adsorption or enthalpy change (kJ/mol) and A5° is the entropy

change (kJ/mol-K).

2.3.4.2 Activation Parameters

The rate constant k for pseudo-second-order reaction shows an Arrhenius

dependence upon reciprocal temperature. The relationship can be expressed by:

lnfc = ln A + E°y^j 2.10

where k is the rate constant of adsorption, A is the Arrhenius factor, and Eaj is the

activation energy (J/mol) for adsorption. The values of Ead and A are calculated from the

slope and intercept of a plot of In k vs 1 IT.

50

2.3.5 Fixed Bed Column in Adsorption Studies

The successful design of a column adsorption process requires the prediction of the

concentration-time profile or breakthrough curve for the effluent. A number of

mathematical models have been developed to describe the concentration time profile in

order to assist in the design. In general, mechanisms operating in a biosorption column

involve axial dispersion in the direction of the liquid flow, film diffusion resistance, intra-

particle diffusion resistance which may include both pore and surface diffusion, and

sorption kinetics at the adsorbent surface (Chu 2004; Pagnanelli 2011). Non-linearity,

associated with equilibrium expressions, leads to sets of partial differential equations that

may require a complicated numerical solution. In addition, independent experiments may

be required to estimate the numerous equilibrium, transport, and sorption kinetic

parameters involved (Chu, 2004). Otherwise, multi-parameter fitting of breakthrough

curves may reduce the physical significance of the mechanistic parameters (Pagnanelli

2011). Obtaining a numerical solution with current computing facilities may no longer be

rigorous but semi-emperical or approximate methods are still extensively used to

simulate breakthrough curves of adsorption columns (Cooney 1999). Different kinetic

models, useful in process design, can be fitted to the column data to determine the

characteristic parameters of the column. Several widely used models for column design

are: Thomas, Yan, Belter, Chu, Yoon—Nelson, Oulman, and Wolborska models. The

Thomas model (Thomas 1948) is a second order reversible reaction model while Bohart—

Adams is a quasi-chemical kinetic model (Wang et al. 2010). The Yan model (Yan et al.

2001) is a dose—response model while the Belter and Cussler model (Belter and Cussler

51

2.3.5 Fixed Bed Column in Adsorption Studies

The successful design of a column adsorption process requires the prediction of the

concentration-time profile or breakthrough curve for the effluent. A number of

mathematical models have been developed to describe the concentration time profile in

order to assist in the design. In general, mechanisms operating in a biosorption column

involve axial dispersion in the direction of the liquid flow, film diffusion resistance, intra-

particle diffusion resistance which may include both pore and surface diffusion, and

sorption kinetics at the adsorbent surface (Chu 2004; Pagnanelli 2011). Non-linearity,

associated with equilibrium expressions, leads to sets of partial differential equations that

may require a complicated numerical solution. In addition, independent experiments may

be required to estimate the numerous equilibrium, transport, and sorption kinetic

parameters involved (Chu, 2004). Otherwise, multi-parameter fitting of breakthrough

curves may reduce the physical significance of the mechanistic parameters (Pagnanelli

2011). Obtaining a numerical solution with current computing facilities may no longer be

rigorous but semi-emperical or approximate methods are still extensively used to

simulate breakthrough curves of adsorption columns (Cooney 1999). Different kinetic

models, useful in process design, can be fitted to the column data to determine the

characteristic parameters of the column. Several widely used models for column design

are: Thomas, Yan, Belter, Chu, Yoon-Nelson, Oulman, and Wolborska models. The

Thomas model (Thomas 1948) is a second order reversible reaction model while Bohart-

Adams is a quasi-chemical kinetic model (Wang et al. 2010). The Yan model (Yan et al.

2001) is a dose-response model while the Belter and Cussler model (Belter and Cussler

51

1988) is a simple two-parameter model. The design criteria for lab scale packed bed

columns, as suggested in the literature, is given in Table 2.4 and the design details

regarding the lab scale packed columns, used in various biosorption studies, are provided

in Table 2.5. The design criteria, available in the literature, differ considerably.

Crittenden and Sowerby (1990) stated that conflicting criteria for packed bed

reactors arose when attempting to minimize the effects of radial temperature gradients

and channeling and wall effects. Carberry (1976) recommends a column to particle

diameter of less than 5 to 6 to avoid excessive radial temperature gradients. Whereas

Rase (1977) recommends the ratio should be greater than about 10 to avoid channeling

and wall effects (Crittenden and Sowerby 1990).

2.3.5.1 The Thomas Model

The Thomas model is widely used to evaluate column performance. It can be used

to predict a breakthrough curve and the maximum solute uptake by an adsorbent. The

Thomas model (Thomas 1948) has the following form (Reynolds and Richards 1996):

C0 1+ exp (K7/ Qfqom — Coll)

2.11

where Ce is the effluent adsorbate concentration (mg/L); Co the influent adsorbate

concentration (mg/L); KT the Thomas rate constant (L/min mg); qo the maximum solid

phase concentration of the solute (mg/g); m the mass of the adsorbent (g); V the

throughput volume (mL); and Q is the volumetric flow rate (mL/min).

52

1988) is a simple two-parameter model. The design criteria for lab scale packed bed

columns, as suggested in the literature, is given in Table 2.4 and the design details

regarding the lab scale packed columns, used in various biosorption studies, are provided

in Table 2.5. The design criteria, available in the literature, differ considerably.

Crittenden and Sowerby (1990) stated that conflicting criteria for packed bed

reactors arose when attempting to minimize the effects of radial temperature gradients

and channeling and wall effects. Carberry (1976) recommends a column to particle

diameter of less than 5 to 6 to avoid excessive radial temperature gradients. Whereas

Rase (1977) recommends the ratio should be greater than about 10 to avoid channeling

and wall effects (Crittenden and Sowerby 1990).

2.3.5.1 The Thomas Model

The Thomas model is widely used to evaluate column performance. It can be used

to predict a breakthrough curve and the maximum solute uptake by an adsorbent. The

Thomas model (Thomas 1948) has the following form (Reynolds and Richards 1996):

where Ce is the effluent adsorbate concentration (mg/L); C0 the influent adsorbate

concentration (mg/L); Kj the Thomas rate constant (L/min mg); q0 the maximum solid

phase concentration of the solute (mg/g); m the mass of the adsorbent (g); V the

throughput volume (mL); and Q is the volumetric flow rate (mL/min).

Q 2.11

52

Table 2.4: Design criteria for packed bed columns

Remarks Design criteria Reference

Lab columns, to Diameter of column — At least 4 in; Height Benefield et al. reduce wall effects of column -25-35 ft; Flow rate — Minimum (1982)

1 gpm/ft2; Diameter of column: particle diameter - 25:1; Empty bed contact time -20 to 50 min

Lab or pilot Diameter of column - Minimum 2.54 cm columns (1 in); Height of column - Minimum 61cm

(24 in)

Diameter of column - 3.5 cm; Flow rate - Reynolds and

Lab or pilot Maximum 0.5 gpm/ ft2 Richards (1995)

columns - To avoid channeling

Flow rate - Maximum 1 gpm/ft2

Diameter of column - 9.5 cm; Flow rate -Maximum 1.5 gpm/ft2

To investigate Flow rate - 1-3 gpm/ ft2; Recommended effects of flow rates

midrange flow rate - 0.5-4 gpm/ ft2

Perrich (1981) Typical values feasible for most applications

Empty bed contact time - 15, 30, 60, 90, 120 min

For GAC Empty bed contact time - 5-30 min Asano et al. contactors (2006)

Rapid small scale Diameter of column - 2-4 cm; height of Crittenden et al. column test column — 30 cm; length of bed - 5-15 cm; (1991)

Height of column: diameter of particle -20:1 (or greater)

For isothermal Turbulent conditions (Re>10); Height of operation column: diameter of column more than 20

Axial peclet number important

Height of column: particle diameter - Less than 50 Carberry (1976)

Avoid channeling Diameter of column: particle diameter -& wall effects Less than 5 to 6

Avoid channeling Diameter of column: particle diameter - Rase (1977) & wall effects Greater than about 10

Account column height: particle dia 100 or less Hill (1977) Dispersion

53

Table 2.4: Design criteria for packed bed columns

Remarks Design criteria Reference

Lab columns, to reduce wall effects

Lab or columns

pilot

Lab or pilot columns - To avoid channeling

To investigate effects of flow rates

Typical values feasible for most applications

For contactors

GAC

Rapid small scale column test

For isothermal operation

Axial peclet number important

Avoid channeling & wall effects

Avoid channeling & wall effects

Account Dispersion

Diameter of column - At least 4 in; Height of column -25-35 ft; Flow rate - Minimum 1 gpm/ft2; Diameter of column: particle diameter - 25:1; Empty bed contact time -20 to 50 min

Diameter of column - Minimum 2.54 cm (1 in); Height of column - Minimum 61cm (24 in)

Diameter of column - 3.5 cm; Flow rate -Maximum 0.5 gpm/ ft2

Flow rate - Maximum 1 gpm/ft2

Diameter of column - 9.5 cm; Flow rate -Maximum 1.5 gpm/ft2

Flow rate - 1-3 gpm/ ft2; Recommended midrange flow rate - 0.5-4 gpm/ ft2

Empty bed contact time - 15, 30, 60, 90, 120 min

Empty bed contact time - 5-30 min

Diameter of column - 2-4 cm; height of column - 30 cm; length of bed - 5-15 cm; Height of column: diameter of particle -20:1 (or greater)

Turbulent conditions (Re>10); Height of column: diameter of column more than 20

Height of column: particle diameter - Less than 50

Diameter of column: particle diameter -Less than 5 to 6

Diameter of column: particle diameter -Greater than about 10

column height: particle dia 100 or less

Benefield et al. (1982)

Reynolds and Richards (1995)

Perrich (1981)

Asano et al. (2006)

Crittenden et al. (1991)

Carberry (1976)

Rase (1977)

Hill (1977)

53

Table 2.5: Design details of lab scale columns used in biosorption studies

Dia of column (cm)

Length of bed (cm)

Flow rate (mL/ min)

Empty bed contact time (min)

Hydraulic loading rate (gpm/ft2)

Sorbent used/ material removed from water

Reference

3 10.3 0.58 A. niger/ Cd, Kapoor and

1.27 24.5 Cu, Pb Viraraghavan

1.8 17.2 0.35 A. niger/Ni (1998)

Fu and 1.27 23.5 6 4.9 1.16 A. niger! dyes Viraraghavan

(2003) 2.28 16.1 0.44 M rouxii/ Pb Yan and

1.27 29 2.59 2.66

14.2 13.8

0.50 0.52

M rouxii/ Cd M rouxii/Ni

Viraraghavan (2001)

2.22 16.5 0.43 M rouxii/ Zn

1.6 55 2.4 46.1 0.29 A. niger/ Cr Mungasavalliet al. (2007) Mathialagan

3.18 38 8 37.7 0.25 Perlite/ Cd andViraraghavan (2002)

1.25 20 12 2.0 2.40 Vermicul-ite! oil

Mysore et al. (2006)

Organoc-lay Moazed and 1.9 100 12 23.6 1.04 anthracite! oil Viraraghavan

(Breakthrough) (2001) 12 23.6 1.04

1.9 100

16 20 24 28

17.7 14.2 11.8 10.1

1.39 1.73 2.08 2.43

Organoclay- anthracite! oil (Breakdown)

Moazed and Viraraghavan (2002)

32 8.9 2.77

10 30 50 47.1 0.16 Peat/ oil (B reakthrough)

Mathavan and Viraraghavan (1989)

4.3 120.0 6.0 290.3 0.1 S. cerevisiae/ Ramirez et al. 12.0 145.1 0.2 Cr (III) and Cr (2007) 15.0 116.1 0.3 (VI)

54

Table 2.5: Design details of lab scale columns used in biosorption studies

Dia of column (cm)

Length of bed (cm)

Flow rate (mL/ min)


1.27 24.5 3

1.8

10.3

17.2

1.27 23.5 6 4.9

1.27 29

2.28 2.59 2.66 2.22

16.1 14.2 13.8 16.5

1.6 55 2.4 46.1

3.18 38 8 37.7

1.25 20 12 2.0

1.9 100 12 23.6

1.9 100

12 16 20 24 28 32

23.6 17.7 14.2 11.8 10.1 8.9

10 30 50 47.1

4.3 120.0 6.0 12.0 15.0

290.3 145.1 116.1

Hydraulic loading rate (gpm/ft2)


Reference

0.58

0.35

1.16

0.44 0.50 0.52 0.43

0.29

0.25

2.40

1.04

1.04 1.39 1.73 2.08 2.43 2.77

0.16

0.1 0.2 0.3

A. niger/ Cd, Cu, Pb

A. niger/ Ni

A. niger/ dyes

M. rouxii/ Pb M. rouxii/ Cd M. rouxii/ Ni M. rouxii/ Zn

A. niger/ Cr

Perlite/ Cd

Vermicul-ite/ oil Organoc-lay anthracite/ oil (Breakthrough)

Organoclay-anthracite/ oil (Breakdown)

Peat/ oil (Breakthrough)

S. cerevisiae/ Cr (III) and Cr (VI)

Kapoor and Viraraghavan (1998)

Fu and Viraraghavan (2003) Yan and Viraraghavan (2001)

Mungasavalli etal. (2007) Mathialagan and Viraraghavan (2002) Mysore et al. (2006) Moazed and Viraraghavan (2001)

Moazed and Viraraghavan (2002)

Mathavan and Viraraghavan (1989) Ramirez et al. (2007)

54

Table 2.5: Design details of lab scale columns used in biosorption studies (contd.)

Dia of column (cm)

Length of bed (cm)

Flow rate (mL/ min)


Hydraulic loading rate (gpm/ft)


Reference

1.0 30.0 0.5 47.1 0.2 Bottom ash and de-oiled soya/

Mittal et al. (2008)

Metanil Yellow

0.8 8.0 0.4 10.5 0.2 Cedar saw Hamdaoui (2006) 0.9 4.6 0.4 dust/ 2.0 2.0 1.0 Methylene

0.8 16.0 0.4 21.0 0.2 Blue 0.9 9.3 0.4 2.0 4.0 1.0

1.0 6.0 0.8 5.4 0.3 Immobilize Aksu and Gonen 1.6 2.7 0.5 d activated (2004) 3.2 1.4 1.1 sludge/

phenol 1.2 15.0 8.2 2.1 1.8 Natural Han et al. (2007a)

5.2 3.3 1.1 zeolite/ 2.2 7.7 0.5 Methylene

Blue 1.5 8.4 8.2 1.8 1.1 Rice husk/ Han et al. (2007b)

13.0 8.2 2.8 Methylene 25.4 8.2 5.5 Blue 39.0 8.2 8.4

2.8 30.6 6.0 31.4 0.2 Sargassum sp. / copper

Borba et al. (2008)

(II)

55

Table 2.5: Design details of lab scale columns used in biosorption studies (contd.)

Dia of Length Flow Empty Hydraulic Sorbent Reference column of bed rate bed loading used/ (cm) (cm) (mL/ contact rate material

min) time (gpm/ft2) removed (min) from water

1.0 30.0 0.5 47.1 0.2 Bottom ash Mittal et al. (2008) and de-oiled soya/ Metanil Yellow

0.8 8.0 0.4 10.5 0.2 Cedar saw Hamdaoui (2006) 0.9 4.6 0.4 dust/ 2.0 2.0 1.0 Methylene

0.8 16.0 0.4 21.0 0.2 Blue 0.9 9.3 0.4 2.0 4.0 1.0

1.0 6.0 0.8 5.4 0.3 Immobilize Aksu and Gonen 1.6 2.7 0.5 d activated (2004) 3.2 1.4 1.1 sludge/

phenol 1.2 15.0 8.2 2.1 1.8 Natural Han et al. (2007a)

5.2 3.3 1.1 zeolite/ 2.2 7.7 0.5 Methylene

Blue 1.5 8.4 8.2 1.8 1.1 Rice husk/ Han et al. (2007b)

13.0 8.2 2.8 Methylene 25.4 8.2 5.5 Blue 39.0 8.2 8.4

2.8 30.6 6.0 31.4 0.2 Sargassum Borba et al. (2008) sp. / copper (II)

55

The rate constant, KT, is a lumped parameter containing the effects of both intrinsic

kinetics and mass transfer. Its value is dependent upon the relative magnitudes of the

processes, which may be affected by the operating conditions and system variables of the

fixed bed column (Wang et al. 2010). The difference between breakthrough curves

calculated from the Thomas model and from other realistic diffusion equation models is

small (Ruthven, 1984).

2.3.5.2 The Yan Model

Column kinetics in a biosorption column was described by a modified dose—

response model proposed by Yan et al. (2001) and the Thomas model. The model was

developed for heavy metal removal in a biosorption column and is used to describe the

binary response regression problem. The logistic equation is expressed as follows:

Ce = 1

Co

1

1+{/} a

2.12

where d is the throughput volume that produces the same Ce/C0 value at 50% removal

and a is the constant, denoting the slope of the function.

2.3.5.3 The Belter and Chu Models

The models used in this study were the models developed by Belter et al. (1988)

(Equation 2.13) and modified by Chu (2004), as shown in Equations 2.14 and 2.15.

56

The rate constant, Kj, is a lumped parameter containing the effects of both intrinsic

kinetics and mass transfer. Its value is dependent upon the relative magnitudes of the

processes, which may be affected by the operating conditions and system variables of the

fixed bed column (Wang et al. 2010). The difference between breakthrough curves

calculated from the Thomas model and from other realistic diffusion equation models is

small (Ruthven, 1984).

2.3.5.2 The Yan Model

Column kinetics in a biosorption column was described by a modified dose-

response model proposed by Yan et al. (2001) and the Thomas model. The model was

developed for heavy metal removal in a biosorption column and is used to describe the

binary response regression problem. The logistic equation is expressed as follows:

where d is the throughput volume that produces the same CJCQ value at 50% removal

and a is the constant, denoting the slope of the function.

2.3.5.3 The Belter and Chu Models

The models used in this study were the models developed by Belter et al. (1988)

(Equation 2.13) and modified by Chu (2004), as shown in Equations 2.14 and 2.15.

2.12

56

t —t0_1i(i+erfCo y2ato

Ce 1 Co 2

Ce 1

Co 2

1+ erf

1+ erf

t—t0)exp(a(X0))

(t— t0)exp(—a(yto ))

Via;

2.13

2.14

2.15

where erf(x) is the error function of x, t the residence time inside the column, to the

temporal parameter which indicates the time required for the outlet contaminant

concentration to be half of the one inlet contaminant concentration and a is the standard

deviation which is a measure of the slope of the breakthrough curve. Model parameters to

and a can be estimated by fitting Equations 2.14 and 2.15 into the experimental

breakthrough data.

2.3.5.4 The Yoon and Nelson Model

Yoon and Nelson (1984) have developed a relatively simple model addressing

adsorption and breakthrough of adsorbate vapors or gases with respect to activated

charcoal. The Yoon and Nelson equation, related to a single- component system, is

expressed as:

57

£l Co

2.13

('-'•)rap(CT(X)) 2.14

(,-,„)exp (-a(^J) 2.15

where erf(x) is the error function of x, t the residence time inside the column, to the

temporal parameter which indicates the time required for the outlet contaminant

concentration to be half of the one inlet contaminant concentration and a is the standard

deviation which is a measure of the slope of the breakthrough curve. Model parameters to

and a can be estimated by fitting Equations 2.14 and 2.15 into the experimental

breakthrough data.

2.3.5.4 The Voori and Nelson Model

Yoon and Nelson (1984) have developed a relatively simple model addressing

adsorption and breakthrough of adsorbate vapors or gases with respect to activated

charcoal. The Yoon and Nelson equation, related to a single- component system, is

expressed as:

57

In t -t K Co - C

YN y 2 YN 2.16

where KYN is the rate constant (min-I); tut, the time required for 50% adsorbate

breakthrough (min) and t is the breakthrough time (min). Values of KYN and ti /2 may be

determined from a plot of in C/(Co-C) versus sampling time (t) according to Equation

2.16. If the theoretical model accurately characterizes the experimental data, this plot will

result in a straight line with slope of Km and an intercept KYN t1/2.

2.3.5.5 The Oulman Model

Oulman proposed the use of a bed depth service model to simulate Granular

Activated Carbon (GAC) adsorption beds (Oulman 1980). The model was first developed

by Bohart and Adams (1920) and based upon surface reaction theory. The Bohart-Adams

equation is as follows:

ln ...1).KNx

KC °t t 2.17

in which C is the effluent concentration (mg/L); Co is the influent concentration (mg/L);

K is the adsorption rate coefficient (L /mg min); N is the adsorption capacity coefficient

(mg/L); x is the bed depth (cm); u is the linear velocity (cm/min); and t is the time (min).

The Bohart-Adams equation can be rewritten as Equation 2.18 where a = KNxIu and b =

KC0.

C 1

C o 1+ e °-b`

58

2.18

2.16

where KYN is the rate constant (min-1); ti/2, the time required for 50% adsorbate

breakthrough (min) and t is the breakthrough time (min). Values of KYN and ti/2 may be

determined from a plot of In C/(Co-C) versus sampling time (t) according to Equation

2.16. If the theoretical model accurately characterizes the experimental data, this plot will

result in a straight line with slope of KYN and an intercept KYN ti/2.

2.3.5.5 The Oulman Model

Oulman proposed the use of a bed depth service model to simulate Granular

Activated Carbon (GAC) adsorption beds (Oulman 1980). The model was first developed

by Bohart and Adams (1920) and based upon surface reaction theory. The Bohart-Adams

equation is as follows:

in which C is the effluent concentration (mg/L); Co is the influent concentration (mg/L);

K is the adsorption rate coefficient (L /mg min); N is the adsorption capacity coefficient

(mg/L); x is the bed depth (cm); u is the linear velocity (cm/min); and t is the time (min).

The Bohart-Adams equation can be rewritten as Equation 2.18 where a = KNx/u and b =

2.17

KC0.

C 1

C0 " l + ea~b' 2.18

58

The Oulman model, as given by Equation 2.18 is equivalent to the 'logistic curve', an S-

shaped curve that is symmetrical around its midpoint at 1= a/b, C = C0/2.

2.3.5.6 The Wolborska Model

The Wolborska (1989) model generally describes the concentration distribution in

the bed for the low concentration region (low C/C0) of the breakthrough curve:

C Cot I 3 In C: a a

C 0 N 0 U 0

2.19

where pa is the kinetic coefficient of the external mass transfer (min-1), Na is the

saturation concentration in the Wolborska model (mg/L), Uo is the superficial velocity

(mm/min), and Z is the height of the column (mm). The expression of the Wolborska

model is equivalent to the Adams-Bohart model if coefficient k is equal to 13./N0. So, the

parameters in the two models can be determined from a plot of In (C/C0) against t at a

given bed height and flow rate.

2.4 Breakdown Mechanisms

2.4.1 Filtration Mechanisms

Particle removal mechanisms of granular media filtration may be divided into two

sequential steps: transport and attachment. Mechanisms involved in the transport of a

particle, from the bulk fluid to the surface of the filter medium, may include straining,

sedimentation, diffusion, interception, hydrodynamics, and flocculation. The mechanisms

59

The Oulman model, as given by Equation 2.18 is equivalent to the 'logistic curve', an S-

shaped curve that is symmetrical around its midpoint at t ~ alb, C = Co/2.

2.3.5.6 The Wolborska Model

The Wolborska (1989) model generally describes the concentration distribution in

the bed for the low concentration region (low C/Co) of the breakthrough curve:

I„£.ASL, M 219 c0 N, u„

where pa is the kinetic coefficient of the external mass transfer (min-1), N0 is the

saturation concentration in the Wolborska model (mg/L), U0 is the superficial velocity

(mm/min), and Z is the height of the column (mm). The expression of the Wolborska

model is equivalent to the Adams-Bohart model if coefficient k is equal to pa/iVo- So, the

parameters in the two models can be determined from a plot of In (C/Co) against t at a

given bed height and flow rate.

2.4 Breakdown Mechanisms

2.4.1 Filtration Mechanisms

Particle removal mechanisms of granular media filtration may be divided into two

sequential steps: transport and attachment. Mechanisms involved in the transport of a

particle, from the bulk fluid to the surface of the filter medium, may include straining,

sedimentation, diffusion, interception, hydrodynamics, and flocculation. The mechanisms

59

of particle attachment to filter media may occur via physical adsorption (electrostatic,

electrokinetic, and vander Waals forces), chemical adsorption (bonding and chemical

interaction), and biological growth. Sherony and Kintner (1971a,b) presented the

following filtration models to predict the head-loss across a filter bed involving single and

two-phase flows, respectively:

36LUyck, (1— 0 2API - (For single-phase flow)

cqgce32.20

36LUyck2 (1— e, )2 2.21 (For two-phase flow)

(.1;

2 where, AP1= pressure drop across the bed for single-phase flow [M/L ]; AP2 = pressure

2 drop across the bed for two-phase flow [MA, yc= dynamic viscosity of the continuous

phase [M/LT]; k1= Carman-Kozeny constant for single-phase flow (dimensionless); k2=

Carman-Kozeny constant for two-phase flow (dimensionless); gc = acceleration due to

gravity, 9.8 m/s2 [M/T2]; E = porosity of immobilized biomass bed in single-phase flow

(dimensionless); et = porosity of immobilized biomass bed in two-phase flow

(dimensionless); df = average diameter of immobilized biomass beads [L]; L = bed length

[L]; and U = superficial velocity [LIT]. The Carman-Kozeny constant (k1) is related to the

shape factor (k0) by the following equation (Carman 1956):

k,- koT2 2.22

where T is tortuosity. The value of T depends upon the pore structure of the medium.

With respect to pores of a medium having a straight path, the value of T is 1. As for a

medium having a non-straight path of pores, the value of T can be estimated as 1/e

60

of particle attachment to filter media may occur via physical adsorption (electrostatic,

electrokinetic, and vander Waals forces), chemical adsorption (bonding and chemical

interaction), and biological growth. Sherony and Kintner (1971a,b) presented the

following filtration models to predict the head-loss across a filter bed involving single and

two-phase flows, respectively:

. _ 36LUy ck.(l-£)2 2 20 AP, -—^—— (For single-phase flow)

d fg c£

A n 36LUy ck2{l-£,)2 ^ s 2.21 AP, —^ — (For two-phase flow)

df8A

2 where, APi= pressure drop across the bed for single-phase flow [M/L ]; AP2 = pressure

2 drop across the bed for two-phase flow [M/L ]; yc= dynamic viscosity of the continuous

phase [M/LT]; ki= Carman-Kozeny constant for single-phase flow (dimensionless); k2=

Carman-Kozeny constant for two-phase flow (dimensionless); gc = acceleration due to

2 , gravity, 9.8 m/s [M/T ]; e = porosity of immobilized biomass bed in single-phase flow

(dimensionless); et = porosity of immobilized biomass bed in two-phase flow

(dimensionless); df = average diameter of immobilized biomass beads [L]; L = bed length

[L]; and U = superficial velocity [L/T], The Carman-Kozeny constant (ki) is related to the

shape factor (ko) by the following equation (Carman 1956):

*1 - KT1 2.22

where T is tortuosity. The value of T depends upon the pore structure of the medium.

With respect to pores of a medium having a straight path, the value of T is 1. As for a

medium having a non-straight path of pores, the value of T can be estimated as 1/e

60

(Johnston 1983).

One very important concept in applying filtration equations to two-phase flow is

that of relative permeability. Bear (1972) outlined factors affecting relative permeability

and stated saturation had a considerable effect upon the other parameters. The specific

permeability coefficient (Bo) of materials having high porosities can be calculated using

the following relationship, as proposed by Kozeny (Carman 1956):

d2E32.23 Bo -

16k1 (1 —

The Kozeny equation for a specific permeability coefficient is applicable to laminar

flows. According to Sherony et al. (1978), the condition for laminar flow is:

NRe <10 (1—e)

where NRe is the Reynolds number.

2.24

2.4.2 Coalescence Mechanisms

The coalescence process can be described as a phenomenon in which droplets

from the discontinuous phase of an emulsion tend to produce extensively large droplets

until eventually a separate phase may be created (Schramm 1992). The process is usually

initiated by a collision between the droplets that is sufficient enough to break the

interfacial film. Once the droplets are in physical contact, the process is completed by

surface forces (Chieu et al. 1975). The coalescence mechanism takes place according to

three major steps (Hazlett 1969): approach, attachment, and release. The approach

61

(Johnston 1983).

One very important concept in applying filtration equations to two-phase flow is

that of relative permeability. Bear (1972) outlined factors affecting relative permeability

and stated saturation had a considerable effect upon the other parameters. The specific

permeability coefficient (Bo) of materials having high porosities can be calculated using

the following relationship, as proposed by Kozeny (Carman 1956):

B - ^ 2.23 ° 16^(1-e)2

The Kozeny equation for a specific permeability coefficient is applicable to laminar

flows. According to Sherony et al. (1978), the condition for laminar flow is:

7^ <10 2.24 (l-e)

where Afa is the Reynolds number.

2.4.2 Coalescence Mechanisms

The coalescence process can be described as a phenomenon in which droplets

from the discontinuous phase of an emulsion tend to produce extensively large droplets

until eventually a separate phase may be created (Schramm 1992). The process is usually

initiated by a collision between the droplets that is sufficient enough to break the

interfacial film. Once the droplets are in physical contact, the process is completed by

surface forces (Chieu et al. 1975). The coalescence mechanism takes place according to

three major steps (Hazlett 1969): approach, attachment, and release. The approach

61

mechanism involves bringing a dispersed-phase droplet to the surface of the filter

medium or to the surface of a droplet attached to the surface medium. The main forces

involved in the approach step include interception, diffusion, inertial impaction,

electrostatic attraction, van der Waals force and sedimentation (Hazlett 1969). In the

attachment process, the oil droplets displace water film from the surface of the material

and preferentially wet its surface. If the media is wetted by the dispersed phase, drops

will coalescence onto the media surface to form a film upon which subsequent drops will

coalesce. However, if the media is not wetted by the dispersed phase, coalescence

between drops will occur in the pores of the media. In the release step, the hydrodynamic

force acting upon a drop overcomes the adhesive force between the droplet and the

material (Hazlett 1969). Several mathematical models have been developed to describe

the process of coalescence. Sheroney and Kintner (1971 a,b) developed a collision model,

which relates the overall collision frequency between droplets and particles of the bed to

the overall coalescence efficiency. The theory of the model was based largely upon the

impaction contact mechanism: droplets moving with the continuous phase collide with

other droplets previously attached to the medium. Although other mechanisms were

identified, it was shown that the order of magnitude of collision frequencies due to

impaction was high enough to neglect the other mechanisms (Sheroney and Kintner,

1971a). This model predicts the ratio of outlet to inlet particle number densities (Y) in

terms of other parameters as follows:

62

mechanism involves bringing a dispersed-phase droplet to the surface of the filter

medium or to the surface of a droplet attached to the surface medium. The main forces

involved in the approach step include interception, diffusion, inertial impaction,

electrostatic attraction, van der Waals force and sedimentation (Hazlett 1969). In the

attachment process, the oil droplets displace water film from the surface of the material

and preferentially wet its surface. If the media is wetted by the dispersed phase, drops

will coalescence onto the media surface to form a film upon which subsequent drops will

coalesce. However, if the media is not wetted by the dispersed phase, coalescence

between drops will occur in the pores of the media. In the release step, the hydrodynamic

force acting upon a drop overcomes the adhesive force between the droplet and the

material (Hazlett 1969). Several mathematical models have been developed to describe

the process of coalescence. Sheroney and Kintner (1971a,b) developed a collision model,

which relates the overall collision frequency between droplets and particles of the bed to

the overall coalescence efficiency. The theory of the model was based largely upon the

impaction contact mechanism: droplets moving with the continuous phase collide with

other droplets previously attached to the medium. Although other mechanisms were

identified, it was shown that the order of magnitude of collision frequencies due to

impaction was high enough to neglect the other mechanisms (Sheroney and Kintner,

1971a). This model predicts the ratio of outlet to inlet particle number densities (Y) in

terms of other parameters as follows:

62

2.25

Y — exp 4 (1 —S,) d

where: c = porosity of the immobilized biomass beads (dimensionless); Sd = average

saturation (the fraction of the void volume occupied by the dispersed phase adhering to

the solid, dimensionless); di = average particle size of the distribution [L]; df =

immobilized biomass bead diameter; ic = the overall coalescence efficiency (fraction);

and L = bed length [L]. The Sherony and Kintner model was recently modified by Li and

Gu (2005), taking into consideration both the collisions between a droplet in the stream

and a droplet deposited on the solid surface and the collisions between two droplets in the

stream. However, this modified model has undergone only limited-testing and the

specific mechanisms affecting the overall coalescence efficiency of the granular beds

could not be adequately assessed.

If, Sd, di, df, E, L and Y are known, Tic can be determined using the above equation.

The values of di, df, c, L and Y can be measured experimentally. The assumptions based

upon which the above model was developed are (Sherony and Kintner 1971a,b): 1) the

model is confined to emulsion concentrations of 1000 mg/L or less flowing into the bed;

2) the bed is randomly packed fiber with a resident population of adhering drops and size

distribution of the held drops is constant with time; 3) the particles from the liquid

interact with the hold-up by collision; 4) the velocity is independent of particle mass; 5)

in the velocity region of interest, the impaction mechanism is predominant; 6) the number

density of particles to the fiber is dependent upon the degree of saturation of the bed; 7)

63

where: e = porosity of the immobilized biomass beads (dimensionless); Sa = average

saturation (the fraction of the void volume occupied by the dispersed phase adhering to

the solid, dimensionless); di = average particle size of the distribution [L]; df =

immobilized biomass bead diameter; r|c = the overall coalescence efficiency (fraction);

and L = bed length [L]. The Sherony and Kintner model was recently modified by Li and

Gu (2005), taking into consideration both the collisions between a droplet in the stream

and a droplet deposited on the solid surface and the collisions between two droplets in the

stream. However, this modified model has undergone only limited-testing and the

specific mechanisms affecting the overall coalescence efficiency of the granular beds

could not be adequately assessed.

If, Sd, di, df, E, L and Y are known, r|c can be determined using the above equation.

The values of dj, df, e, L and Y can be measured experimentally. The assumptions based

upon which the above model was developed are (Sherony and Kintner 1971a,b): 1) the

model is confined to emulsion concentrations of 1000 mg/L or less flowing into the bed;

2) the bed is randomly packed fiber with a resident population of adhering drops and size

distribution of the held drops is constant with time; 3) the particles from the liquid

interact with the hold-up by collision; 4) the velocity is independent of particle mass; 5)

in the velocity region of interest, the impaction mechanism is predominant; 6) the number

density of particles to the fiber is dependent upon the degree of saturation of the bed; 7)

emulsion flowing through the bed does not form a continuum; and 8) the possibility of a

collision between two drops is independent of drop diameter.

The average saturation (Sd) of the dispersed phase is related to the change in

porosity (et) by:

sd = (1 — ft) 2.26

The porosity of the bed in two-phase flow will be calculated using the following

relationship developed by dividing Equation 2.20 by Equation 2.21 and re-arranging:

dPik2e3 )3

'3—

AP2k, (1— e)2 (1 — e ,)22.27

where f3 is a dimensionless parameter.

The average saturation (Sd) in Equation 2.26 can be calculated using Equation

2.27. The average holdup is defined as the volume of dispersed phase per volume of solid

retained. The average holdup O H) of a bed can be calculated by the following equation,

as presented by Sherony et al. (1978):

Sd (1— e)

£

2.28

Breaking of naphtha/water and water/naphtha emulsions with a packed bed coalescence,

using Ottawa sand, was examined by Crickmore et al. (1989). Process variables

influencing the packed bed in coalescing emulsions were screened and the qualitative

relationship between the influential process variables and emulsion coalescence were

determined. The flow rate was identified as affecting the emulsion coalescence and other

variables such as the type of packing, packing size range, and temperature and was found

64

emulsion flowing through the bed does not form a continuum; and 8) the possibility of a

collision between two drops is independent of drop diameter.

The average saturation (S<j) of the dispersed phase is related to the change in

porosity (et) by:

The porosity of the bed in two-phase flow will be calculated using the following

relationship developed by dividing Equation 2.20 by Equation 2.21 and re-arranging:

where P is a dimensionless parameter.

The average saturation (Sd) in Equation 2.26 can be calculated using Equation

2.27. The average holdup is defined as the volume of dispersed phase per volume of solid

retained. The average holdup (0h) of a bed can be calculated by the following equation,

as presented by Sherony et al. (1978):

Breaking of naphtha/water and water/naphtha emulsions with a packed bed coalescence,

using Ottawa sand, was examined by Crickmore et al. (1989). Process variables

influencing the packed bed in coalescing emulsions were screened and the qualitative

relationship between the influential process variables and emulsion coalescence were

determined. The flow rate was identified as affecting the emulsion coalescence and other

variables such as the type of packing, packing size range, and temperature and was found

2.26

WAP"*)' "(l-e,)! "P

APfe3 E> 2.27

64

to have no significant effect. Preliminary examination revealed that coalescence of an

emulsion was solely a function of the residence time in the bed and a kinetic expression

should be able to correlate the coalescence data. The relationship between emulsion

coalescence and residence time was described by a first-order rate equation:

CA ws CA0 exp(-01 2.29

where, CA = fraction of fluid emulsified at t', CA0 = fraction of fluid emulsified at

entrance to the packed bed, IQ = rate constant for coalescence, t'= modified residence

time.

r

AE 2.30

where, t= residence time, A = cross-sectional area of the empty tube, and E = void

fraction of the packed bed.

65

to have no significant effect. Preliminary examination revealed that coalescence of an

emulsion was solely a function of the residence time in the bed and a kinetic expression

should be able to correlate the coalescence data. The relationship between emulsion

coalescence and residence time was described by a first-order rate equation:

C>CAoexp(-*cT') 2 29

where, CA = fraction of fluid emulsified at x', CA0 = fraction of fluid emulsified at

entrance to the packed bed, kc = rate constant for coalescence, x - modified residence

time.

T'-— 2.30 Ae

where, x= residence time, A = cross-sectional area of the empty tube, and e = void

fraction of the packed bed.

65

Chapter 3

Materials and Methods

3.1 Glassware Preparation

All glassware for use in the experiments was washed with a laboratory detergent,

rinsed with tap water, further washed with hydrochloric acid, and finally rinsed twice

with deionized water and dried prior to use.

3.2 Experimental Materials

Two fungal strains, Mucor rouxii and Absidia coerulea, were purchased from

American Type Culture Collection (ATCC), Rockville, Maryland, USA (ATCC #24905,

ATCC #10738a). Chitosan from crab shells (Aldrich 417963) was purchased from

Sigma-Aldrich Corporation, Ontario, Canada. Walnut shell media was supplied by

USFilter, part of the Water Technology division of Siemen's Industrial Solutions and

Services (I&S) Group, USA. Types of oil found in contaminated water include fats;

lubricants; cutting fluids; heavy hydrocarbons such as tars, grease, crude oil, diesel oils;

and light hydrocarbons such as kerosene, jet fuel, and gasoline (Braden, 1994). The

following oils were used in the study typically representing a refined mineral oil, a

vegetable oil and a cutting oil:

• Standard (light) Mineral Oil (SMO) marketed by Fisher Scientific Company,

USA, emulsified with oleic acid and triethanolamine using Regina tap water

according to the procedure used by Biswas (1973);

66

Chapter 3

Materials and Methods

3.1 Glassware Preparation

All glassware for use in the experiments was washed with a laboratory detergent,

rinsed with tap water, further washed with hydrochloric acid, and finally rinsed twice

with deionized water and dried prior to use.

3.2 Experimental Materials

Two fungal strains, Mucor rouxii and Absidia coerulea, were purchased from

American Type Culture Collection (ATCC), Rockville, Maryland, USA (ATCC #24905,

ATCC #10738a). Chitosan from crab shells (Aldrich 417963) was purchased from

Sigma-Aldrich Corporation, Ontario, Canada. Walnut shell media was supplied by

USFilter, part of the Water Technology division of Siemen's Industrial Solutions and

Services (I&S) Group, USA. Types of oil found in contaminated water include fats;

lubricants; cutting fluids; heavy hydrocarbons such as tars, grease, crude oil, diesel oils;

and light hydrocarbons such as kerosene, jet fuel, and gasoline (Braden, 1994). The

following oils were used in the study typically representing a refined mineral oil, a

vegetable oil and a cutting oil:

• Standard (light) Mineral Oil (SMO) marketed by Fisher Scientific Company,

USA, emulsified with oleic acid and triethanolamine using Regina tap water

according to the procedure used by Biswas (1973);

66

• Vegetable oil, Canola Oil (CO) marketed in Canada, emulsified in the same

manner as SMO; and

• DoALL Bright-Edge 80, a cutting oil manufactured by DoALL Company, IL,

USA, emulsified in the same manner as SMO;

3.3 Preparation of Oil-in-water Emulsions

Oil (21.25 g) was poured into a beaker to which 18.80 g of oleic acid was added.

This mixture was stirred continuously using a magnetic stirrer for 15 minutes, producing

a clear mixture followed by an addition of 9.95 g of triethanolamine. The resulting

mixture was stirred at a high speed to produce a clear mixture with bubbles inside the

solution. The oil mixture was diluted with tap water to two litres. This was then mixed

thoroughly in a blender for 15 minutes. This oil emulsion was used as a 25,000 mg/L

concentrate. The working stock solution was prepared by diluting with tap water. This oil

emulsion was prepared based on the procedure adopted by Biswas (1973).

3.4 Characterization of Oil-in-water Emulsions

The density of oils was determined by weighing a known volume (25 mL) of oil

in a pycnometer. Viscosity was measured according to ASTM (1963). The viscosity of

the oils was determined by comparing the efflux time of a liquid having a known

viscosity to that of oil with unknown viscosity, as given below:

67

• Vegetable oil, Canola Oil (CO) marketed in Canada, emulsified in the same

manner as SMO; and

• DoALL Bright-Edge 80, a cutting oil manufactured by DoALL Company, IL,

USA, emulsified in the same manner as SMO;

3.3 Preparation of Oil-in-water Emulsions

Oil (21.25 g) was poured into a beaker to which 18.80 g of oleic acid was added.

This mixture was stirred continuously using a magnetic stirrer for 15 minutes, producing

a clear mixture followed by an addition of 9.95 g of triethanolamine. The resulting

mixture was stirred at a high speed to produce a clear mixture with bubbles inside the

solution. The oil mixture was diluted with tap water to two litres. This was then mixed

thoroughly in a blender for 15 minutes. This oil emulsion was used as a 25,000 mg/L

concentrate. The working stock solution was prepared by diluting with tap water. This oil

emulsion was prepared based on the procedure adopted by Biswas (1973).

3.4 Characterization of Oil-in-water Emulsions

The density of oils was determined by weighing a known volume (25 mL) of oil

in a pycnometer. Viscosity was measured according to ASTM (1963). The viscosity of

the oils was determined by comparing the efflux time of a liquid having a known

viscosity to that of oil with unknown viscosity, as given below:

67

12= WI/ P2t2 3.1

Where, t = Efflux time; p = Density of the liquid; rl 1 = Viscosity of water; and rl 2 =

Viscosity of oil. Interfacial tension of the oil was measured with a Spinning Drop

Interfacial Tensiometer Model 510 marketed by Temco Inc., Tulsa, Oklahoma, USA. The

surface charges of SMO and CO were taken from measurements done by Mysore et al.

(2005). Surface charge of Bright-Edge 80 cutting oil was measured using Zetasizer,

model HSA 3000 (Malvern, Worcestershire, England).

3.5 Preparation of Fungal Seed

Freeze dried cultures received from ATCC were activated in sterile water for

seeding. Deionized water (100 mL) was autoclaved for 15 minutes at 121 °C and 124 kPa

and allowed to cool to room temperature. Freeze dried cultures were soaked in 10 mL of

sterile water for 45 minutes before transferring to a potato-dextrose agar surface. Bacto

potato dextrose agar (PDA) 39 g/L was used to prepare the agar surfaces. Approximately

2 mL of this suspension was transferred to each PDA plate. The plates were incubated for

7-10 days at 22±2°C to fully activate the fungal strain. This culture was used to grow the

biomass for experimental purposes.

3.6 Preparation of Non-viable Fungal Biomass

M rouxii strain was routinely maintained on potato dextrose agar plates. It was

grown aerobically by the shake flask method. It was cultivated using a growth medium

68

til/ 112= pit]/ p2t2 3 i

Where, t = Efflux time; p = Density of the liquid; TI i = Viscosity of water; and t] 2 =

Viscosity of oil. Interfacial tension of the oil was measured with a Spinning Drop

Interfacial Tensiometer Model 510 marketed by Temco Inc., Tulsa, Oklahoma, USA. The

surface charges of SMO and CO were taken from measurements done by Mysore et al.

(2005). Surface charge of Bright-Edge 80 cutting oil was measured using Zetasizer,

model HSA 3000 (Malvern, Worcestershire, England).

3.5 Preparation of Fungal Seed

Freeze dried cultures received from ATCC were activated in sterile water for

seeding. Deionized water (100 mL) was autoclaved for 15 minutes at 121 °C and 124 kPa

and allowed to cool to room temperature. Freeze dried cultures were soaked in 10 mL of

sterile water for 45 minutes before transferring to a potato-dextrose agar surface. Bacto®

potato dextrose agar (PDA) 39 g/L was used to prepare the agar surfaces. Approximately

2 mL of this suspension was transferred to each PDA plate. The plates were incubated for

7-10 days at 22±2°C to fully activate the fungal strain. This culture was used to grow the

biomass for experimental purposes.

3.6 Preparation of Non-viabie Fungal Biomass

M. rouxii strain was routinely maintained on potato dextrose agar plates. It was

grown aerobically by the shake flask method. It was cultivated using a growth medium

68

comprised of yeast extract (3 g/L), peptone (10 g/L) and glucose (replaced by dextrose)

(20 g/L) (Bartnickni-Garcia and Nickerson 1962; Muzzarelli et al. 1994). The pH of the

growth medium was maintained at 4.5 by 1.0 N HCI. The culture was grown in an

aerobic condition at room temperature (22 ± 2°C) with 100 mL of the liquid medium in

250 mL conical flasks on a rotary shaker agitated at 125 rpm. M rouxii was harvested

after three days of growth by filtering the growth medium through a 150 gm sieve. The

harvested fungal biomass was washed with generous amounts of deionized water and

autoclaved for 30 minutes at 121°C and 103 kPa. The autoclaved biomass was allowed to

cool and dried in an oven at 60 °C for 24 hours. The dried biomass was powdered into a

fine size using a grinder. The biomass passing through a 400 mesh sieve was used for the

experiment. A. coerulea strains were also prepared in the same manner as those of M

rouxii, except that the pH of the growth medium for A. coerulea was maintained at 5.0 by

1.0 N HCI and was harvested after 4 days of growth.

3.7 Characterization of Fungal Biomass

Porosity of the adsorbent was determined by the graduated cylinder technique

(AEEP 1988) by placing a known volume of adsorbent into a measuring cylinder and

filling it with water. The amount of water required to fill in the voids was measured and

the porosity determined. The moisture content of the adsorbent was measured by

electronic balance according to ASTM (1992). The samples were mixed thoroughly and a

known amount was placed in the pre-weighed porcelain dish. The sample was then oven-

dried at 105°C for 16 hours, cooled and weighed. The loss in weight was used to

69

comprised of yeast extract (3 g/L), peptone (10 g/L) and glucose (replaced by dextrose)

(20 g/L) (Bartnickni-Garcia and Nickerson 1962; Muzzarelli et al. 1994). The pH of the

growth medium was maintained at 4.5 by 1.0 N HC1. The culture was grown in an

aerobic condition at room temperature (22 ± 2°C) with 100 mL of the liquid medium in

250 mL conical flasks on a rotary shaker agitated at 125 rpm. M. rouxii was harvested

after three days of growth by filtering the growth medium through a 150 |im sieve. The

harvested fungal biomass was washed with generous amounts of deionized water and

autoclaved for 30 minutes at 121 °C and 103 kPa. The autoclaved biomass was allowed to

cool and dried in an oven at 60 °C for 24 hours. The dried biomass was powdered into a

fine size using a grinder. The biomass passing through a 400 mesh sieve was used for the

experiment. A. coerulea strains were also prepared in the same manner as those of M.

rouxii, except that the pH of the growth medium for A coerulea was maintained at 5.0 by

1.0 N HCI and was harvested after 4 days of growth.

3.7 Characterization of Fungal Biomass

Porosity of the adsorbent was determined by the graduated cylinder technique

(AEEP 1988) by placing a known volume of adsorbent into a measuring cylinder and

filling it with water. The amount of water required to fill in the voids was measured and

the porosity determined. The moisture content of the adsorbent was measured by

electronic balance according to ASTM (1992). The samples were mixed thoroughly and a

known amount was placed in the pre-weighed porcelain dish. The sample was then oven-

dried at 105°C for 16 hours, cooled and weighed. The loss in weight was used to

calculate the moisture content. The pH of the adsorbent was determined by electronic

ASTM measurements (ASTM 1989). A known weight of the adsorbent was placed in 350

mL deionized water and allowed to soak for 30 minutes with occasional stirring. The pH

was then determined using an Accumet pH meter (Fisher Scientific Ltd, Model 600).

3.7.1 Surface Area Analysis and Surface Charge Measurement

Surface area and pore size measurements of the powdered autoclaved biomass

were carried out using Micromeritics® ASAP 2020 accelerated surface area and

porosimetry analyzer. The surface charge of the autoclaved M rouxii biomass was

measured using Zetasizer, model HSA 3000 (Malvern, Worcestershire, England) at the

University of Alberta, Canada. An average particle size of the biomass was also

measured using the Zetasizer.

3.7.2 Scanning Electron Microscope (SEM) Studies

Scanning Electron Microscope (SEM) images of biomass samples (before and after

sorption of oil) were obtained using a LEO FESEM 1530 unit with a field emission

Gemini Column. The samples were gold coated for 120 s prior to SEM scanning.

70

calculate the moisture content. The pH of the adsorbent was determined by electronic

ASTM measurements (ASTM 1989). A known weight of the adsorbent was placed in 350

mL deionized water and allowed to soak for 30 minutes with occasional stirring. The pH

was then determined using an Accumet pH meter (Fisher Scientific Ltd, Model 600).

3.7.1 Surface Area Analysis and Surface Charge Measurement

Surface area and pore size measurements of the powdered autoclaved biomass

were carried out using Micromeritics® ASAP 2020 accelerated surface area and

porosimetry analyzer. The surface charge of the autoclaved M. rouxii biomass was

measured using Zetasizer, model HSA 3000 (Malvern, Worcestershire, England) at the

University of Alberta, Canada. An average particle size of the biomass was also

measured using the Zetasizer.

3.7.2 Scanning Electron Microscope (SEM) Studies

Scanning Electron Microscope (SEM) images of biomass samples (before and after

sorption of oil) were obtained using a LEO FESEM 1530 unit with a field emission

Gemini Column. The samples were gold coated for 120 s prior to SEM scanning.

70

3.7.3 Fourier Transform Infrared (FTIR) Analysis

Autoclaved biomass and oil-loaded biomass were analyzed qualitatively on a Bio-

Rad FTS-60 infrared system using the potassium bromide (KBr) pellet technique (6 mg

sample/400 mg KBr).

3.8 Oil Concentration Measurement

Oil and grease are defined as any material recovered as a substance soluble in the

solvent (Franson and Eaton 2005). In the determination of oil and grease, an absolute

quantity of a specific substance is not measured; instead oil and grease are defined by the

method used for their determination (Franson and Eaton 2005). Although a variety of

non-hydrocarbon solvents are commercially available for the determination of oil and

grease using the Infrared (IR) method, for certain regulatory purposes, USEPA currently

recommends only n-hexane (Pushkarev et al. 1983; Franson and Eaton 2005; Tyrie and

Caudle 2007). As for liquid samples, the 21s` edition of Standard Methods prescribes four

methods for oil and grease determination: 1) the liquid/liquid partition-gravimetric

(5520B); 2) the partition-infrared (5520C); 3) the soxhelet method (5520D); and 4) the

solid-phase, partition-gravimetric (Franson and Eaton 2005). Several oil measurement

techniques such as gravimetric, IR, UltraViolet (UV) and colorimetric methods are listed

in Table 3.1. In the partition-gravimetric method (5520B), dissolved or emulsified oil is

extracted from water by intimate contact with an extracting solvent. However, this

method is not suitable for samples that contain volatile hydrocarbons, because they would

be lost in the solvent-removal operations of this procedure. Soxhlet method (5520D) is

71

3.7.3 Fourier Transform Infrared (FTIR) Analysis

Autoclaved biomass and oil-loaded biomass were analyzed qualitatively on a Bio-

Rad FTS-60 infrared system using the potassium bromide (KBr) pellet technique (6 mg

sample/400 mg KBr).

3.8 Oil Concentration Measurement

Oil and grease are defined as any material recovered as a substance soluble in the

solvent (Franson and Eaton 2005). In the determination of oil and grease, an absolute

quantity of a specific substance is not measured; instead oil and grease are defined by the

method used for their determination (Franson and Eaton 2005). Although a variety of

non-hydrocarbon solvents are commercially available for the determination of oil and

grease using the Infrared (IR) method, for certain regulatory purposes, USEPA currently

recommends only n-hexane (Pushkarev et al. 1983; Franson and Eaton 2005; Tyrie and

Caudle 2007). As for liquid samples, the 21st edition of Standard Methods prescribes four

methods for oil and grease determination: 1) the liquid/liquid partition-gravimetric

(5520B); 2) the partition-infrared (5520C); 3) the soxhelet method (5520D); and 4) the

solid-phase, partition-gravimetric (Franson and Eaton 2005). Several oil measurement

techniques such as gravimetric, IR, Ultraviolet (UV) and colorimetric methods are listed

in Table 3.1. In the partition-gravimetric method (5520B), dissolved or emulsified oil is

extracted from water by intimate contact with an extracting solvent. However, this

method is not suitable for samples that contain volatile hydrocarbons, because they would

be lost in the solvent-removal operations of this procedure. Soxhlet method (5520D) is

71

Table 3.1: Oil in water measurement methods

Method Extraction solvent Remarks Reference Gravimetric n-hexane Legally defines oil in the US. Franson and (USEPA Method 1664;

Measures only organic compounds that are extractable from water in n-hexane at pH

Eaton (2005)

Standard Methods

2.0 and that remain after hexane is evaporated.

5520B)

Infrared Trichlorotrifluro- C-H bond present in the oil Franson and (Standard ethane, non- absorbs IR energy at a Eaton (2005); Methods hydrocarbon wavelength of 3.41 µ (2900 cm" Tyrie and 5520C) solvents such as 1). Since water absorbs IR Caudle

tetrachlorothylene, carbon

energy at this wavelength, measurements must be made on

(2007); Ibrahim et al.

tetrachloride. a sample using a solvent that does not absorb IR radiation. Calibrated with a standard sample of known concentration.

(2009); Sokolovic et al. (2010)

Ultra violet Extraction solvents Aromatic compounds absorb UV Bastow et al. method such as isopropanol radiation and fluoresce. UV is (1997); Tyrie

have been reported. Measurements can also be made without an extraction solvent.

not absorbed by water. However, other components in the wastewater such as iron may fluoresce. So instruments using extraction have an advantage.

and Caudle (2007)

Colorimetric Generally no The absorption of energy in the Biswas methods extraction is visible light range is used as the (1973);

required. A solvent detection process. The Tellez et al. like methylene chloride has been used to impart color to the oil containing wastewater.

instrument needs to be calibrated with a known concentration.

(2005)

Particle No extraction is No calibration is required. Tyrie and counting required Particle counting methods

cannot normally see below 2 µ.. Caudle (2007)

72

Table 3.1: Oil in water measurement methods

Method Extraction solvent Remarks Reference Gravimetric (USEPA Method 1664; Standard Methods 5520B)

Infrared (Standard Methods 5520C)

n-hexane

Ultra violet method

Colorimetric methods

Particle counting

Trichlorotrifluro-ethane, non-hydrocarbon solvents such as tetrachlorothylene, carbon tetrachloride.

Extraction solvents such as isopropanol have been reported. Measurements can also be made without an extraction solvent.

Generally no extraction is required. A solvent like methylene chloride has been used to impart color to the oil containing wastewater.

No extraction is required

Legally defines oil in the US. Measures only organic compounds that are extractable from water in n-hexane at pH 2.0 and that remain after hexane is evaporated.

C-H bond present in the oil absorbs IR energy at a wavelength of 3.41 n (2900 cm" '). Since water absorbs IR energy at this wavelength, measurements must be made on a sample using a solvent that does not absorb IR radiation. Calibrated with a standard sample of known concentration.

Aromatic compounds absorb UV radiation and fluoresce. UV is not absorbed by water. However, other components in the wastewater such as iron may fluoresce. So instruments using extraction have an advantage.

The absorption of energy in the visible light range is used as the detection process. The instrument needs to be calibrated with a known concentration.

No calibration is required. Particle counting methods cannot normally see below 2 it.

Franson and Eaton (2005)

Franson and Eaton (2005); Tyrie and Caudle (2007); Ibrahim et al. (2009); Sokolovic et al. (2010)

Bastow et al. (1997); Tyrie and Caudle (2007)

Biswas (1973); Tellez et al. (2005)

Tyrie and Caudle (2007)

72

the method of choice when relatively polar, heavy petroleum fractions are present, or

when the levels of the nonvolatile greases may challenge the solubility limit of the

solvent. The infrared spectroscopic method may have an advantage over other methods

when determining the sum of extractables of unknown origin for which it is generally

difficult to plot a calibration curve. This is possible due to the fact that light absorption

bands of the C-H bond belong to the infrared region at a wavelength of 2900 aril . The C-

H bonds are mostly characteristic of any organic compound and the number of bonds per

unit of mass varies less for substances of variable composition than does the number of

any other functional group. Therefore, the calibration curves plotted for various types of

oils diverge by not more than 5 — 10% (Pushkarev et al. 1983). However, the infrared

absorption bands of water (0-H) bonds are also localized in the region around 3000 cm I.

Hence, the method requires complete elimination of water from the sample for which an

extracting solvent containing no C-H bonds should be used.

Due to its advantages and precision, the partition-infrared method was used in this

study. The use of IR techniques to measure oil in water has been widely reported in the

literature (Mathavan and Viraraghavan 1992; Moazed and Viraraghavan 2005; Mysore et

al. 2005). Other than conventional IR spectroscopy (Sokolovic 2010), equipment such as

Horiba OCMA 220, and InfraCal CVH have been reported to provide an analysis of oil

content of water samples (Mathavan and Viraraghavan 1992; Moazed and Viraraghavan

2005; Mysore et al. 2005; Ibrahim et al. 2009). Among IR techniques, oil determination

by Horiba OCMA 220 using different solvents has been extensively studied (Mathavan

and Viraraghavan 1990). Different solvents such as carbon tetrachloride and

73

the method of choice when relatively polar, heavy petroleum fractions are present, or

when the levels of the nonvolatile greases may challenge the solubility limit of the

solvent. The infrared spectroscopic method may have an advantage over other methods

when determining the sum of extractables of unknown origin for which it is generally

difficult to plot a calibration curve. This is possible due to the fact that light absorption

bands of the C-H bond belong to the infrared region at a wavelength of 2900 cm*1. The C-

H bonds are mostly characteristic of any organic compound and the number of bonds per

unit of mass varies less for substances of variable composition than does the number of

any other functional group. Therefore, the calibration curves plotted for various types of

oils diverge by not more than 5 - 10% (Pushkarev et al. 1983). However, the infrared

absorption bands of water (O-H) bonds are also localized in the region around 3000 cm"1.

Hence, the method requires complete elimination of water from the sample for which an

extracting solvent containing no C-H bonds should be used.

Due to its advantages and precision, the partition-infrared method was used in this

study. The use of IR techniques to measure oil in water has been widely reported in the

literature (Mathavan and Viraraghavan 1992; Moazed and Viraraghavan 2005; Mysore et

al. 2005). Other than conventional IR spectroscopy (Sokolovic 2010), equipment such as

Horiba OCMA 220, and InfraCal CVH have been reported to provide an analysis of oil

content of water samples (Mathavan and Viraraghavan 1992; Moazed and Viraraghavan

2005; Mysore et al. 2005; Ibrahim et al. 2009). Among IR techniques, oil determination

by Horiba OCMA 220 using different solvents has been extensively studied (Mathavan

and Viraraghavan 1990). Different solvents such as carbon tetrachloride and

73

tetrachloroethylene have been used for extraction prior to oil determination by IR

techniques (Mathavan and Viraraghavan 1992; Moazed and Viraraghavan 2005; Mysore

et al. 2005). A Horiba OCMA-350 Oil Content Analyzer will be used in the measurement

of oil. Horiba OCMA-350 has an inbuilt NDIR spectrophotometer and displays oil

concentration directly in mg/L on a digital panel. Oil will be extracted with

tetrachloroethylene (ultra-resi analyzed) before being analyzed by OCMA-350. The

measurement range of Horiba OCMA-350 is from 0 to 200 mg/L and 0 to 1000 mg/g.

3.9 Batch Biosorption Experiments

Oil-in-water emulsions of 100 mL volume were contacted with 0.2 g of the fungal

biomass at a speed of 175 rpm in a platform shaker (Model: Classic C2), manufactured

by New Brunswick Scientific, New Jersey, USA. The study was carried out with an

initial oil concentration of 200 mg/L for 6 hours. The pH was adjusted using 0.1 M HCI

or 0.1 M NaOH solution. The experiments were conducted under controlled pH

conditions using buffer solutions (Lange, 1973). 0.2 M of sodium phosphate and 0.1 M of

citric acid were used in different ratios to adjust the pH. The oil-in-water emulsions were

vacuum-filtered through a 1.2 lam glass micro-filter after biosorption experiments were

taken. A control with no biomass was also set up for each run. All experiments were

conducted in duplicate and the mean values were used in the analysis of the data. The

filtrate was analyzed for oil concentration using Horiba OCMA-350 oil content analyzer

(Horiba Instruments Inc., CA).

74

tetrachloroethylene have been used for extraction prior to oil determination by IR

techniques (Mathavan and Viraraghavan 1992; Moazed and Viraraghavan 2005; Mysore

et al. 2005). A Horiba OCMA-350 Oil Content Analyzer will be used in the measurement

of oil. Horiba OCMA-350 has an inbuilt NDIR spectrophotometer and displays oil

concentration directly in mg/L on a digital panel. Oil will be extracted with

tetrachloroethylene (ultra-resi analyzed) before being analyzed by OCMA-350. The

measurement range of Horiba OCMA-350 is from 0 to 200 mg/L and 0 to 1000 mg/g.

3.9 Batch Biosorption Experiments

Oil-in-water emulsions of 100 mL volume were contacted with 0.2 g of the fungal

biomass at a speed of 175 rpm in a platform shaker (Model: Classic C2), manufactured

by New Brunswick Scientific, New Jersey, USA. The study was carried out with an

initial oil concentration of 200 mg/L for 6 hours. The pH was adjusted using 0.1 M HC1

or 0.1 M NaOH solution. The experiments were conducted under controlled pH

conditions using buffer solutions (Lange, 1973). 0.2 M of sodium phosphate and 0.1 M of

citric acid were used in different ratios to adjust the pH. The oil-in-water emulsions were

vacuum-filtered through a 1.2 jim glass micro-filter after biosorption experiments were

taken. A control with no biomass was also set up for each run. All experiments were

conducted in duplicate and the mean values were used in the analysis of the data. The

filtrate was analyzed for oil concentration using Horiba OCMA-350 oil content analyzer

(Horiba Instruments Inc., CA).

74

3.9.1 Preliminary Studies

Two fungal biomasses, Mucor rouxii and Absidia coerulea, along with chitosan

and walnut shell media were the biomaterials used. Oil-in-water emulsions of 100 mL

volume were contacted with 0.2 g of each adsorbent at two pH values of 5.0 and 7.6 for 6

hours. The objective of this study was to examine the removal of three oils from water

using non-viable fungal biomasses M rouxii and A. coerulea and compare their

adsorption capacities with those of chitosan and walnut shell media.

3.9.2 Factorial Design of Experiments

In order to evaluate factors that influence the percent removal of oil by M rouxii

biomass, a two-level-five factors fractional factorial experiment was designed. Five

factors that are generally known to influence batch adsorption were chosen to study the

response as percentage removal of oil by sorption on M rouxii biomass. The five factores

chosen were pH of the solution, temperature, adsorbent dose, concentration of oil and

rotational speed of the shaker. Each factor was studied at two levels - low level and high

level. To analyze the factorial design, the original measurement units for the experimental

factors (uncoded units) were transformed into coded units (Minitab 2007). Five factors

were analyzed at 2 levels and the factor levels were coded as -1 (low) and +1 (high)

(Table 3.2). The response was expressed as the percent removal of oil by M rouxii

biomass. Two pH conditions of 3.0 and 9.0 were chosen to represent highly acidic and

highly alkaline conditions. Temperatures of 5 and 30°C were chosen to represent a range

of cold and warm conditions. Typical mean influent oil concentration into wastewater

75

3.9.1 Preliminary Studies

Two fungal biomasses, Mucor rouxii and Absidia coerulea, along with chitosan

and walnut shell media were the biomaterials used. Oil-in-water emulsions of 100 mL

volume were contacted with 0.2 g of each adsorbent at two pH values of 5.0 and 7.6 for 6

hours. The objective of this study was to examine the removal of three oils from water

using non-viable fungal biomasses M. rouxii and A. coerulea and compare their

adsorption capacities with those of chitosan and walnut shell media.

3.9.2 Factorial Design of Experiments

In order to evaluate factors that influence the percent removal of oil by M. rouxii

biomass, a two-level-five factors fractional factorial experiment was designed. Five

factors that are generally known to influence batch adsorption were chosen to study the

response as percentage removal of oil by sorption on M. rouxii biomass. The five factores

chosen were pH of the solution, temperature, adsorbent dose, concentration of oil and

rotational speed of the shaker. Each factor was studied at two levels - low level and high

level. To analyze the factorial design, the original measurement units for the experimental

factors (uncoded units) were transformed into coded units (Minitab 2007). Five factors

were analyzed at 2 levels and the factor levels were coded as -1 (low) and +1 (high)

(Table 3.2). The response was expressed as the percent removal of oil by M. rouxii

biomass. Two pH conditions of 3.0 and 9.0 were chosen to represent highly acidic and

highly alkaline conditions. Temperatures of 5 and 30°C were chosen to represent a range

of cold and warm conditions. Typical mean influent oil concentration into wastewater

Table 3.2: Coded and uncoded values of the factors

Factor Coded low level

Corresponding uncoded low values

Coded high level

Corresponding uncoded high values

pH -1 3 +1 9 Temperature -1 5°C +1 30°C Dose of biomass -1 0.05 g +1 0.5 g Concentration of oil

-1 50 mg/L +1 350 mg/L

Speed -1 100 rpm +1 200 rpm

76

Table 3.2: Coded and uncoded values of the factors

Factor Coded low Corresponding Coded Corresponding level uncoded low high level uncoded high

values values pH -1 3 +1 9 Temperature -1 5°C +1 30°C Dose of biomass -1 0.05 g +1 0.5 g Concentration of oil Speed

-1 50 mg/L +1 350 mg/L Concentration of oil Speed -1 100 rpm +1 200 rpm

76

treatment plant in refineries such as Consumer's Cooperative Refineries Limited

(CCRL), Regina is 350 mg/L. Hence a high oil concentration of 350 mg/L was chosen for

the study. The minimum number of experimental runs which must be carried out for a

two level - five factor design is 25 = 32 runs. This is called a 25 full factorial design. With

two replicates, the number of test runs increases to 64, which is large. When the number

of factors is more than four, fractional factorial designs can be used. The information on

the main effects and two-order interactions can be obtained by running only a fraction of

the full-factorial design (Antony 2003). A fractional factorial design is represented by 2(k-

P), where k is the number of factors and 1/2P represents the fraction of the full factorial 2k.

A 2(5-1) fractional factorial design is 1/2th fraction of a 25 full factorial experiment.

In this way, five factors at two levels may be studied in just 16 (i.e. 2(54)) experimental

trials instead of 32 trials (25). MINITABTM release 15 (Minitab 2007) statistical software

was used to create and analyze the experimental data, in order to measure the effect of

various factors regarding the removal of oil from water.

Five factors were analyzed at 2 levels using a 1/2 fraction factorial 25-1 Resolution

V design resulting in 16 runs. The resolution is a description of the design that gives the

extent to which interactions will be confounded with other factors and interactions

(Minitab 2007). In Resolution V design, no main effect or two-factor interaction is

confounded with any other main effect or two-factor interaction (Minitab 2007). All

experimental trials were replicated and thus, 32 experiments were conducted in random

order as generated by MINITAB.

Oil-in-water emulsions of 100 mL volume were contacted for 6 hours with the M

77

treatment plant in refineries such as Consumer's Cooperative Refineries Limited

(CCRL), Regina is 350 mg/L. Hence a high oil concentration of 350 mg/L was chosen for

the study. The minimum number of experimental runs which must be carried out for a

two level - five factor design is 2s = 32 runs. This is called a 25 full factorial design. With

two replicates, the number of test runs increases to 64, which is large. When the number

of factors is more than four, fractional factorial designs can be used. The information on

the main effects and two-order interactions can be obtained by running only a fraction of

the full-factorial design (Antony 2003). A fractional factorial design is represented by 2(k"

p), where k is the number of factors and 1/2P represents the fraction of the full factorial 2k.

A 2(s_1) fractional factorial design is 172th fraction of a 25 full factorial experiment.

In this way, five factors at two levels may be studied in just 16 (i.e. 2(5"1)) experimental

trials instead of 32 trials (25). MINITAB™ release 15 (Minitab 2007) statistical software

was used to create and analyze the experimental data, in order to measure the effect of

various factors regarding the removal of oil from water.

Five factors were analyzed at 2 levels using a V4 fraction factorial 25"1 Resolution

V design resulting in 16 runs. The resolution is a description of the design that gives the

extent to which interactions will be confounded with other factors and interactions

(Minitab 2007). In Resolution V design, no main effect or two-factor interaction is

confounded with any other main effect or two-factor interaction (Minitab 2007). All

experimental trials were replicated and thus, 32 experiments were conducted in random

order as generated by MINITAB.

Oil-in-water emulsions of 100 mL volume were contacted for 6 hours with the M.

11

rouxii biomass at a desired rotational speed. Experiments were conducted under

controlled pH conditions using buffer solutions. Experiments at temperatures other than

room temperature were conducted in a temperature—controlled chamber. A control (oil in

water with no biomass) was also set up for each run. Based on the fractional factorial

design analysis, the influential parameters that affect the removal of different types of oils

from water by M rouxii biomass were found. The effect of parameters such as pH, dose,

oil concentration, temperature, and time on biosorption of oil using M rouxii biomass

was studied in further detail.

3.9.3 Effect of pH

pH studies were conducted by shaking 100 mL of 200 mg/L of oil-in-water

emulsion separately, with 0.2 g of autoclaved M rouxii biomass, over a range of pH

values from 3 to 9 inch increments of 1. The pH was kept constant during the study.

Samples were shaken for 6 hours at 175 rpm at room temperature (22 ± 2 °C). An

optimum pH was selected for further study.

3.9.4 Effect of Concentration

Concentration studies were conducted by shaking 100 mL of oil-in-water

emulsion with 0.2 g of autoclaved M rouxii biomass, over a range of concentration

values from 50 to 350 mg/L in increments of 50 mg/L at pH 3.0. Samples were shaken

for 6 hours at 175 rpm at room temperature (22 ± 2 °C).

78

rouxii biomass at a desired rotational speed. Experiments were conducted under

controlled pH conditions using buffer solutions. Experiments at temperatures other than

room temperature were conducted in a temperature-controlled chamber. A control (oil in

water with no biomass) was also set up for each run. Based on the fractional factorial

design analysis, the influential parameters that affect the removal of different types of oils

from water by M, rouxii biomass were found. The effect of parameters such as pH, dose,

oil concentration, temperature, and time on biosorption of oil using M. rouxii biomass

was studied in further detail.

3.9.3 Effect of pH

pH studies were conducted by shaking 100 mL of 200 mg/L of oil-in-water

emulsion separately, with 0.2 g of autoclaved M. rouxii biomass, over a range of pH

values from 3 to 9 inch increments of 1. The pH was kept constant during the study.

Samples were shaken for 6 hours at 175 rpm at room temperature (22 ± 2 °C). An

optimum pH was selected for further study.

3.9.4 Effect of Concentration

Concentration studies were conducted by shaking 100 mL of oil-in-water

emulsion with 0.2 g of autoclaved M. rouxii biomass, over a range of concentration

values from 50 to 350 mg/L in increments of 50 mg/L at pH 3.0. Samples were shaken

for 6 hours at 175 rpm at room temperature (22 ± 2 °C).

3.9.5 Batch Kinetic Studies

Kinetic studies were conducted at an optimum pH of 3.0 with 100 mL of 200

mg/L of oil and 0.2 g of autoclaved biomass. The samples were collected at the following

intervals of time: 0.25, 0.5, 0.75, 1, 1.25, 1.5, 1.75, 2.0, 2.25, 2.5, 2.75, 3.0, 3.5, 4.0, 5.0,

and 6.0 h. Batch kinetic experiments were conducted at 5±2, 15±2, 22±2 (room

temperature), and 30±2°C. Experiments at temperatures other than room temperature

were conducted in a temperature—controlled chamber. Data from kinetic studies were

fitted to the models developed by Lagregren (1898), Ho and McKay (1998) and Weber

and Morris (1963) to examine biosorption kinetics. Non-linear regression was performed

with statistical software STATISTICA, version 5.0 for Windows (Statistica, 1997).

3.9.6 Batch Isotherm Studies

Isotherm studies were conducted at pH 3 with 100 mL of 200 mg/L of oil and of

varying doses of autoclaved biomass (0.03, 0.04, 0.06, 0.08, 0.1, 0.2, and 0.4 g) at 5, 15,

22 ± 2 and 30 °C. Samples were shaken for 6 hours to ensure that equilibrium was

reached, and then filtered and analyzed for oil concentration. Data from isotherm studies

were fitted to the Langmuir and Freundlich isotherm models. Non-linear regression was

performed with the statistical software STATISTICA, version 5.0 for Windows

(Statistica, 1997).

79

3.9.5 Batch Kinetic Studies

Kinetic studies were conducted at an optimum pH of 3.0 with 100 mL of 200

mg/L of oil and 0.2 g of autoclaved biomass. The samples were collected at the following

intervals of time: 0.25, 0.5,0.75, 1, 1.25, 1.5, 1.75, 2.0, 2.25, 2.5, 2.75, 3.0, 3.5, 4.0, 5.0,

and 6.0 h. Batch kinetic experiments were conducted at 5±2, 15±2, 22±2 (room

temperature), and 30±2°C. Experiments at temperatures other than room temperature

were conducted in a temperature-controlled chamber. Data from kinetic studies were

fitted to the models developed by Lagregren (1898), Ho and McKay (1998) and Weber

and Morris (1963) to examine biosorption kinetics. Non-linear regression was performed

with statistical software STATISTICA, version 5.0 for Windows (Statistica, 1997).

3.9.6 Batch Isotherm Studies

Isotherm studies were conducted at pH 3 with 100 mL of 200 mg/L of oil and of

varying doses of autoclaved biomass (0.03, 0.04, 0.06, 0.08, 0.1, 0.2, and 0.4 g) at 5, 15,

22 ± 2 and 30 °C. Samples were shaken for 6 hours to ensure that equilibrium was

reached, and then filtered and analyzed for oil concentration. Data from isotherm studies

were fitted to the Langmuir and Freundlich isotherm models. Non-linear regression was

performed with the statistical software STATISTICA, version 5.0 for Windows

(Statistica, 1997).

79

3.9.7 Batch Desorption Studies

Desorption studies were preceded by isotherm studies. Isotherm studies were

conducted at room temperature (22 ± 2 °C), as described earlier. For each oil-in-water

emulsion, after the equilibrium shaking time, samples were filtered and collected for the

analysis of oil concentration. The biomass collected on the filter paper, which was loaded

with oil, was thoroughly scraped off and put back into clean 125 mL conical flasks.

Conical flasks were filled with oil-free deionized water. The resulting oil-loaded

autoclaved biomass and oil-free solutions were shaken for a time interval equal to

equilibrium time (obtained previously in the batch kinetic studies). Solution mixtures

were filtered and analyzed for oil. When using deionized water, buffer salts were not used

and the pH of the solutions during the desorption experiments was monitored.

3.9.8 Modification of Functional Groups and Lipid Extraction

Portions of the autoclaved biomass were chemically treated in different ways to

modify specific functional groups that may be present on the biomass. Kapoor and

Viraraghavan (1997) conducted a similar procedure for metal biosorption using A. niger

biomass and they confirmed the modification of different surface functional groups using

FTIR analysis. Chemical treatments applied to the biomass were as follows: the

modification of carboxyl functional groups was carried out by shaking 2 g of the

autoclaved biomass in 130 mL of anhydrous methanol (CH3OH) and 1.2 mL of

concentrated hydrochloric acid for 6 hours at 125 rpm. This treatment causes

80

3.9.7 Batch Desorption Studies

Desorption studies were preceded by isotherm studies. Isotherm studies were

conducted at room temperature (22 ± 2 °C), as described earlier. For each oil-in-water

emulsion, after the equilibrium shaking time, samples were filtered and collected for the

analysis of oil concentration. The biomass collected on the filter paper, which was loaded

with oil, was thoroughly scraped off and put back into clean 125 mL conical flasks.

Conical flasks were filled with oil-free deionized water. The resulting oil-loaded

autoclaved biomass and oil-free solutions were shaken for a time interval equal to

equilibrium time (obtained previously in the batch kinetic studies). Solution mixtures

were filtered and analyzed for oil. When using deionized water, buffer salts were not used

and the pH of the solutions during the desorption experiments was monitored.

3.9.8 Modification of Functional Groups and Lipid Extraction

Portions of the autoclaved biomass were chemically treated in different ways to

modify specific functional groups that may be present on the biomass. Kapoor and

Viraraghavan (1997) conducted a similar procedure for metal biosorption using A. niger

biomass and they confirmed the modification of different surface functional groups using

FTIR analysis. Chemical treatments applied to the biomass were as follows: the

modification of carboxyl functional groups was carried out by shaking 2 g of the

autoclaved biomass in 130 mL of anhydrous methanol (CH3OH) and 1.2 mL of

concentrated hydrochloric acid for 6 hours at 125 rpm. This treatment causes

80

esterification of the carboxylic acids that may be present on the biomass (Gardea-

Torresdey et al., 1990; Drake et al., 1996). The general reaction scheme is:

RCOOH + CH 30H —9. RCOOCH3 + H2O 3.2

The biomass residue obtained was referred to as B 1. The modification of amino

functional groups was carried out by shaking 1 g of autoclaved biomass with 20 mL of

formaldehyde (HCHO) and 40 mL of formic acid (HCOOH) for 6 hours at 125 rpm. This

treatment causes methylation of amines that may be present on the biomass. The general

reaction scheme is (Loudon, 1984):

RCH2NH2 --, RCH 2N (CH 3)2 + CO2 + H2O 3.3

The obtained biomass residue was referred to as B2. The modification of phosphate

functional groups was carried out by heating 1 g of autoclaved biomass under reflux

conditions with 40 mL of triethylphosphite and 30 mL of nitromethane for 6 hours. This

treatment causes esterification of the phosphate groups that may be present on the

biomass (Tobin et al. 1990). The obtained biomass residue was referred to as B3. The

lipid fractions of the biomass were extracted by heating 1 g of autoclaved biomass under

reflux conditions with 75 mL of acetone and 75 mL of benzene, separately, for 6 hours

(Tobin et al. 1990). The obtained biomass residues were referred to as B4 and B5. All

biomass residues were gently washed with deionized water, dried at 40°C and finely

powdered before use. Biosorption experiments were conducted at room temperature by

shaking 0.2 g of biomass B1—B5, separately, in 100 mL of solution containing 200 mg/L

oil at pH 3 for 6 hours. A control was also run under the same conditions with 0.2 g of the

autoclaved biomass.

81

esterification of the carboxylic acids that may be present on the biomass (Gardea-

Torresdey et al., 1990; Drake et al., 1996). The general reaction scheme is:

RCOOH + CH3OH -» RCOOCH3 + H20 3 2

The biomass residue obtained was referred to as Bl. The modification of amino

functional groups was carried out by shaking 1 g of autoclaved biomass with 20 mL of

formaldehyde (HCHO) and 40 mL of formic acid (HCOOH) for 6 hours at 125 rpm. This

treatment causes methylation of amines that may be present on the biomass. The general

reaction scheme is (Loudon, 1984):

RCH2NH2 — RCH2N(CH3\ + C02 + H20 33

The obtained biomass residue was referred to as B2. The modification of phosphate

functional groups was carried out by heating 1 g of autoclaved biomass under reflux

conditions with 40 mL of triethylphosphite and 30 mL of nitromethane for 6 hours. This

treatment causes esterification of the phosphate groups that may be present on the

biomass (Tobin et al. 1990). The obtained biomass residue was referred to as B3. The

lipid fractions of the biomass were extracted by heating 1 g of autoclaved biomass under

reflux conditions with 75 mL of acetone and 75 mL of benzene, separately, for 6 hours

(Tobin et al. 1990). The obtained biomass residues were referred to as B4 and B5. All

biomass residues were gently washed with deionized water, dried at 40°C and finely

powdered before use. Biosorption experiments were conducted at room temperature by

shaking 0.2 g of biomass B1-B5, separately, in 100 mL of solution containing 200 mg/L

oil at pH 3 for 6 hours. A control was also run under the same conditions with 0.2 g of the

autoclaved biomass.

81

3.10 Use of Immobilized Biomass in Oil Removal

3.10.1 Procedure for immobilization of Biomass

Powdered M rouxii biomass was immobilized in a polysulfone matrix following the

procedure described by Kapoor and Viraraghavan (1998), but with certain changes.

Seven grams of powdered pretreated biomass and 7 g of polysulfone were mixed in 100

mL of N, N-dimethyl Formamide (DMF). The beaker was immediately sealed to avoid

volatilization of DMF and then shaken for approximately 16 hours on a magnetic shaker

for polysulfone to completely dissolve in DMF and form a uniformally consistent slurry.

The slurry was sucked into a 10 mL syringe and allowed to drop slowly into a tub of

deionized water. Due to the consistency of the slurry and the contact with air for a brief

moment, spherical droplets formed beads upon contact with the deionized water in the

tub. The biomass was thus immobilized within the solidified polysulfone matrix. The

beads were cured in a moderately agitated deionized water bath for 24 hours to diffuse

the DMF. After curing, the beads were air dried for 3 days at room temperature (22 ± 2

°C).

3.10.2 Characterization of Immobilized M. rouxii Beads

Porosity of the beads was determined by graduated cylinder technique (AEEP 1988)

placing a known volume of adsorbent in a measuring cylinder and filling it with water.

The amount of water required to fill in the voids was measured and the porosity

determined. A sieve analysis of the immobilized biomass beads was conducted with the

82

3.10 Use of Immobilized Biomass in Oil Removal

3.10.1 Procedure for Immobilization of Biomass

Powdered M. rouxii biomass was immobilized in a polysulfone matrix following the

procedure described by Kapoor and Viraraghavan (1998), but with certain changes.

Seven grams of powdered pretreated biomass and 7 g of polysulfone were mixed in 100

mL of N, TV-dimethyl Formamide (DMF). The beaker was immediately sealed to avoid

volatilization of DMF and then shaken for approximately 16 hours on a magnetic shaker

for polysulfone to completely dissolve in DMF and form a uniformally consistent slurry.

The slurry was sucked into a 10 mL syringe and allowed to drop slowly into a tub of

deionized water. Due to the consistency of the slurry and the contact with air for a brief

moment, spherical droplets formed beads upon contact with the deionized water in the

tub. The biomass was thus immobilized within the solidified polysulfone matrix. The

beads were cured in a moderately agitated deionized water bath for 24 hours to diffuse

the DMF. After curing, the beads were air dried for 3 days at room temperature (22 ± 2

*C).

3.10.2 Characterization of Immobilized M. rouxii Beads

Porosity of the beads was determined by graduated cylinder technique (AEEP 1988)

placing a known volume of adsorbent in a measuring cylinder and filling it with water.

The amount of water required to fill in the voids was measured and the porosity

determined. A sieve analysis of the immobilized biomass beads was conducted with the

82

use of sieves of sizes 2.36, 2.0, 1.18, 0.85, 0.6 and 0.42 mm. Beads with an irregular

shape and beads smaller than 0.42 mm or bigger than 2.36 mm were discarded.

3.11 Column Studies

3.11.1 Continuous Breakthrough Studies

Immobilized biomass beads (4.5 g) were packed into a plastic column 1.27 cm in

diameter and 45 cm in height, with a bed depth of 30 cm (Figure 3.1). Glass beads were

placed on both ends of the column (height of 1 cm), to allow for even distribution of the

influent and also to prevent the biomass beads from floating. Glass wool was placed

between the glass beads and the biomass beads at the bottom of the column. The column

was sealed at the bottom using a rubber stopper with a single bore. Tygon tubing was

used for the connections. The ratio of column diameter to particle diameter was 6.5,

which agrees with the criteria suggested by Carberry (1976). The empty bed contact time

was 15 minutes, which is the suggested typical value for feasible applications by Perrich

(1981). The oil-in-water emulsion was fed by a peristaltic pump through the column at a

flow rate of 2.6 mL/min (0.5 gpm/ft2). Reynolds and Richards (1995) suggested a flow

rate of 0.5 to 1.5 gpm/ft2 to avoid channeling. The initial concentration of the oil was 50

mg/L. Effluent samples were collected at regular time intervals. Once the ratio of effluent

to influent concentration reached a value of approximately 0.95 or higher, the column

study was terminated. At this point, the column was considered to have reached

exhaustion. Six kinetic models were fitted to the column data to predict breakthrough

curves. They are: Thomas, Yan, Belter, Chu, Yoon—Nelson, Oulman, and Wolborska 83

use of sieves of sizes 2.36, 2.0, 1.18, 0.85, 0.6 and 0.42 mm. Beads with an irregular

shape and beads smaller than 0.42 mm or bigger than 2.36 mm were discarded.

3.11 Column Studies

3.11.1 Continuous Breakthrough Studies

Immobilized biomass beads (4.5 g) were packed into a plastic column 1.27 cm in

diameter and 45 cm in height, with a bed depth of 30 cm (Figure 3.1). Glass beads were

placed on both ends of the column (height of 1 cm), to allow for even distribution of the

influent and also to prevent the biomass beads from floating. Glass wool was placed

between the glass beads and the biomass beads at the bottom of the column. The column

was sealed at the bottom using a rubber stopper with a single bore. Tygon tubing was

used for the connections. The ratio of column diameter to particle diameter was 6.5,

which agrees with the criteria suggested by Carberry (1976). The empty bed contact time

was 15 minutes, which is the suggested typical value for feasible applications by Perrich

(1981). The oil-in-water emulsion was fed by a peristaltic pump through the column at a

flow rate of 2.6 mL/min (0.5 gpm/fit2). Reynolds and Richards (1995) suggested a flow

rate of 0.5 to 1.5 gpm/ft2 to avoid channeling. The initial concentration of the oil was 50

mg/L. Effluent samples were collected at regular time intervals. Once the ratio of effluent

to influent concentration reached a value of approximately 0.95 or higher, the column

study was terminated. At this point, the column was considered to have reached

exhaustion. Six kinetic models were fitted to the column data to predict breakthrough

curves. They are: Thomas, Yan, Belter, Chu, Yoon-Nelson, Oulman, and Wolborska

83

Influent line

Siosorbent beads

111111111111111•11 Emulsion reservoir

Effluent to drain

Figure 3.1: Schematics of the experimental set up used for breakthrough studies

84

Influent line

# *12.7

E E

-7 mn

Siosorbent beads

•S2L—i—— 4u> y

\

Emulsion reservoir

Effluent to drain

Figure 3.1: Schematics of the experimental set up used for breakthrough studies

84

models. Non-linear regression will be performed with the statistical software

STATISTICA for Windows (Statistica 1997).

3.11.1 Column Regeneration and Reuse

After the column reached exhaustion, the column saturated with oil was eluted using

DI water to desorb oil from beads. Samples were collected at regular time intervals to

measure the concentration of oil in the effluent. Then, the column was fed again with the

same oil-in-water emulsion with a concentration of 50 mg/L oil under similar

experimental conditions as that of the initial run for the second cycle of operation to

investigate the potential of reusing the beads for oil removal. Effluent oil samples were

collected at regular time intervals. The breakthrough curve for the second run of

operation was fitted to the Thomas model. Non-linear regression was performed with the

statistical software STATISTICA for Windows (Statistica 1997).

3.11.2 Coalescence/ Filtration Experiments

An experimental column of 1600 mm length and 19 mm inner diameter was used in

the study (Figure 3.2). An immobilized biomass bed of 1000 mm depth was used and

manometers were installed at 200, 400, 600, 800, and 1000 mm depths to monitor the

pressure drop across the bed. Correspondingly, five sample ports were used to collect

samples to measure the oil concentration, diameter of the oil droplets, and drop densities.

Six different flow rates of 12, 16, 20, 24, 28, and 32 mL/min (1.0, 1.3, 1.7, 2.0, 2.3, and

85

models. Non-linear regression will be performed with the statistical software

STATISTICA for Windows (Statistica 1997).

3.11.1 Column Regeneration and Reuse

After the column reached exhaustion, the column saturated with oil was eluted using

DI water to desorb oil from beads. Samples were collected at regular time intervals to

measure the concentration of oil in the effluent. Then, the column was fed again with the

same oil-in-water emulsion with a concentration of 50 mg/L oil under similar

experimental conditions as that of the initial run for the second cycle of operation to

investigate the potential of reusing the beads for oil removal. Effluent oil samples were

collected at regular time intervals. The breakthrough curve for the second run of

operation was fitted to the Thomas model. Non-linear regression was performed with the

statistical software STATISTICA for Windows (Statistica 1997).

3.11.2 Coalescence/ Filtration Experiments

An experimental column of 1600 mm length and 19 mm inner diameter was used in

the study (Figure 3.2). An immobilized biomass bed of 1000 mm depth was used and

manometers were installed at 200, 400, 600, 800, and 1000 mm depths to monitor the

pressure drop across the bed. Correspondingly, five sample ports were used to collect

samples to measure the oil concentration, diameter of the oil droplets, and drop densities.

Six different flow rates of 12, 16,20,24,28, and 32 mL/min(1.0, 1.3, 1.7, 2.0,2.3, and

Influent line

1000 mm EMuent to drain

Emulsion reservoir

Figure 3.2: Schematics of the experimental set up used for breakdown studies

86

Influent line

J »13 mm

JL—H

200 mm

400 mm

600 mm

800 mm

1000 mm

Manometers

Emulsion reservoir

Effluent to drain

Figure 3.2: Schematics of the experimental set up used for breakdown studies

86

2 2.7 gpm/ft , respectively) were used in the experiments. As for lab scale columns,

Benefield et al. (1982) suggested a flow rate of 1 gpm/ft2 to reduce wall effects and

Perrich (1981) suggested 1 — 3 gpm/ft2 to investigate the effects of flow rates. The initial

concentration of the oil was 50 mg/L. Experiments were terminated when the ratio of

effluent oil concentration to influent oil concentration reached a value of approximately

0.95. The diameter of the SMO-in-water emulsion droplets and drop density of the

effluent samples was measured by a Spectrex Laser Particle Counter (Spectrex

Corporation, U.S.A., Model SPC-510). The experimental procedure involves filling the

system with single-phase (Regina tap water without oil) first and measuring the pressure

drop across the bed for various flow rates. Experiments were monitored for equilibrium

pressure drop. The equilibrium pressure drop was visually monitored by observing min-

imum fluctuations of water levels in the manometers. With respect to single-phase flow,

the pressure drop reached equilibrium within 45 minutes from the start of the run. After

reaching a steady-state condition, oil-in-water emulsion was allowed in the immobilized

M rouxii biomass bed and the pressure drop was monitored for the two-phase flow. With

respect to the two-phase flow, the equilibrium pressure drop was measured at the end of

the runs, for all the flow rates studied.

Data collected from the experimental column were analyzed to predict the head-loss

across the immobilized biomass bed for single-phase flow using Equation 2.20. Then, the

Carman-Kozeny constant (k1) was calculated from the pressure drop data. The pressure

drop data from the two-phase (oil-in-water emulsion) flow through the immobilized M

rouxii biomass bed was analyzed using Equation 2.21 to evaluate the Carman-Kozeny

87

2 2.7 gpm/ft, respectively) were used in the experiments. As for lab scale columns,

Benefield et al. (1982) suggested a flow rate of 1 gpm/ft2 to reduce wall effects and

Perrich (1981) suggested 1 - 3 gpm/ft2 to investigate the effects of flow rates. The initial

concentration of the oil was 50 mg/L. Experiments were terminated when the ratio of

effluent oil concentration to influent oil concentration reached a value of approximately

0.95. The diameter of the SMO-in-water emulsion droplets and drop density of the

effluent samples was measured by a Spectrex Laser Particle Counter (Spectrex

Corporation, U.S.A., Model SPC-510). The experimental procedure involves filling the

system with single-phase (Regina tap water without oil) first and measuring the pressure

drop across the bed for various flow rates. Experiments were monitored for equilibrium

pressure drop. The equilibrium pressure drop was visually monitored by observing min

imum fluctuations of water levels in the manometers. With respect to single-phase flow,

the pressure drop reached equilibrium within 45 minutes from the start of the run. After

reaching a steady-state condition, oil-in-water emulsion was allowed in the immobilized

M. rouxii biomass bed and the pressure drop was monitored for the two-phase flow. With

respect to the two-phase flow, the equilibrium pressure drop was measured at the end of

the runs, for all the flow rates studied.

Data collected from the experimental column were analyzed to predict the head-loss

across the immobilized biomass bed for single-phase flow using Equation 2.20. Then, the

Carman-Kozeny constant (ki) was calculated from the pressure drop data. The pressure

drop data from the two-phase (oil-in-water emulsion) flow through the immobilized M.

rouxii biomass bed was analyzed using Equation 2.21 to evaluate the Carman-Kozeny

87

constant, k2. The average saturation (Sd) was calculated using Equation 2.26. To

determine the average saturation, initially the value of p was calculated using Equation

2.27. In order to calculate 13, experimental values of a i (single-phase pressure drop), AP2

(two-phase pressure drop), and 8 (porosity of the immobilized biomass bed) were used

assuming k2/k1 =1. With this f3 value obtained, c was determined using the relationship

between p and et given in Equation 2.27. The iteration was carried out using Matlab to

obtain 8, This et value was immediately used again in Equation 2.27 to estimate k2. With

the new value of k2, a new value of p was determined. For each set of p, et, k2, the average

saturation was calculated. The overall coalescence efficiency (Tic) was calculated using

Equation 2.25 for each flow rate. Equations 2.29 and 2.30 were applied to the

coalescence data to evaluate the model proposed by Crickmore et al. (1989).

88

constant, k2. The average saturation (S<j) was calculated using Equation 2.26. To

determine the average saturation, initially the value of P was calculated using Equation

2.27. In order to calculate P, experimental values of APi (single-phase pressure drop), AP2

(two-phase pressure drop), and e (porosity of the immobilized biomass bed) were used

assuming k2/ki =1. With this P value obtained, £t was determined using the relationship

between P and et given in Equation 2.27. The iteration was carried out using Matlab to

obtain £t. This et value was immediately used again in Equation 2.27 to estimate k2. With

the new value of k2, a new value of p was determined. For each set of P, £t, k2, the average

saturation was calculated. The overall coalescence efficiency (rjc) was calculated using

Equation 2.25 for each flow rate. Equations 2.29 and 2.30 were applied to the

coalescence data to evaluate the model proposed by Crickmore et al. (1989).

88

Chapter 4

Results and Discussion

4.1 Data and Supplemental Figures

All raw data used in the analysis are provided in Appendix A. Figures A.1 to A.36

in Appendix A supplement the figures in the main body of the thesis.

4.2 Characterization of Oil

The characteristics of oils used in the study are presented in Table 4.1. Surface

charges of the three oil-in-water emulsions at different pH conditions are presented in

Figure 4.1. The characteristics of the solvent, tetrachloroethylene (Ultra Resi - Analysed)

used in the measurement of oil by Horiba OCMA 350 oil content analyzer are shown in

Table 4.2. The characteristics of the three oil-in-water emulsions used in the study based

on the diameter of oil droplets are summarized in Table 4.3. The droplet size analysis

showed that SMO had droplet sizes in the range of 18.2 to 25.1 gm and can be classified

as dispersed and emulsified oils. Both CO and Bright-Edge 80 had droplet sizes in the

range of 13.3 to 15.1 gm and 12 to 7.9 gm, respectively and therefore, are classified as

emulsified and soluble oils.

89

Chapter 4

Results and Discussion

4.1 Data and Supplemental Figures

All raw data used in the analysis are provided in Appendix A. Figures A. 1 to A.36

in Appendix A supplement the figures in the main body of the thesis.

4.2 Characterization of Oil

The characteristics of oils used in the study are presented in Table 4.1. Surface

charges of the three oil-in-water emulsions at different pH conditions are presented in

Figure 4.1. The characteristics of the solvent, tetrachloroethylene (Ultra Resi - Analysed)

used in the measurement of oil by Horiba OCMA 350 oil content analyzer are shown in

Table 4.2. The characteristics of the three oil-in-water emulsions used in the study based

on the diameter of oil droplets are summarized in Table 4.3. The droplet size analysis

showed that SMO had droplet sizes in the range of 18.2 to 25.1 nm and can be classified

as dispersed and emulsified oils. Both CO and Bright-Edge 80 had droplet sizes in the

range of 13.3 to 15.1 jim and 12 to 7.9 (im, respectively and therefore, are classified as

emulsified and soluble oils.

89

Table 4.1: Characteristics of oils used for the study at 20°C

Type of oil Density (kg/m3) Viscosity (Pa.$) Interfacial tension (dynes/cm)

SMO

Canola oil

Bright-Edge 80

841.9

913.2

821.5

0.143

0.070

0.023

5.3

3.1

2.7

90

Table 4.1: Characteristics of oils used for the study at 20°C

Type of oil Density (kg/m3) Viscosity (Pa.s) Interfacial tension (dynes/cm)

SMO 841.9 0.143 5.3

Canola oil 913.2 0.070 3.1

Bright-Edge 80 821.5 0.023 2.7

90

40

20

E ri -20

O' a -40

N

-60

-80

-100

—'--Zeta potential of SMO —0-- Zeta potential of CO

Zeta potential of Bright-Edge 80 zeta potential of Mucor rouxii

Figure 4.1: Zeta potential of autoclaved M rouxii biomass and oil-in-water emulsions

91

> E

I "20

c ® o Q. • 8 N

-60 -

-80 -

100

• Zeta potential of SMO —o— zeta potential of CO

—a— Zeta potential of Bright-Edge 80 —*— zeta potential of Mucor rouxii

Figure 4.1: Zeta potential of autoclaved M. rouxii biomass and oil-in-water emulsions

91

Table 4.2: Characteristics of the solvent used in the measurement of oil

Characteristics Description

Chemical name Tetrachloroethylene

Commercial name Ultra Resi-Analyzed

Formula C12C:CC12

Molecular weight 165.83 g/mol

Boiling Point 121.1 °C

Melting Point -19 °C

Flash point 38 °C

Vapor Pressure 18 mm Hg at 25 °C

Solubility 0.015 gin 100 g of water

Appearance Clear, colorless liquid

Odor Ethereal odor

92

Table 4.2: Characteristics of the solvent used in the measurement of oil

Characteristics Description

Chemical name T etrachloroethylene

Commercial name Ultra Resi-Analyzed

Formula C12C:CC12

Molecular weight 165.83 g/mol

Boiling Point 121.1 °C

Melting Point -19 °C

Flash point 38 °C

Vapor Pressure 18 mm Hg at 25 °C

Solubility 0.015 g in 100 g of water

Appearance Clear, colorless liquid

Odor Ethereal odor

92

Table 4.3: Emulsion classification based on the diameter of oil droplets

Type of emulsion Range of diameter (Am)

Mean diameter (gm) Classification

SMO 18.2 — 25.1 21 Dispersed and emulsified

Canola oil 13.3 — 15.1 14 Emulsified and soluble

Bright-Edge 80 7.9 — 12 10 Emulsified and soluble

93

Table 4.3: Emulsion classification based on the diameter of oil droplets

Type of emulsion Range of diameter (Hm)

Mean diameter (nm) Classification

SMO 18.2-25.1 21 Dispersed emulsified

and

Canola oil 13.3-15.1 14 Emulsified soluble

and

Bright-Edge 80 7.9-12 10 Emulsified soluble

and

93

4.3 Characterization of M. rouxii Biomass and Other Adsorbents

The characteristics of autoclaved powdered M rouxii biomass are presented in

Table 4.4. The surface area of M rouxii biomass was 20.55 m2/g and the average particle

size was 5.47 x 104 cm. Surface charges of autoclaved M rouxii biomass at different pH

conditions are presented in Figure 4.1. SEM images of the raw M rouxii biomass are

shown in Figure 4.2a. Pores of the raw M rouxii biomass are clearly observed from

Figure 4.2a. The surface morphology of the biomass after adsorption of SMO, CO, and

Bright-Edge 80 are shown in Figure 4.2b, c and d, respectively. The SEM images with

best resolutions are presented in the figure. A definite change in the surface morphology

of the biomass following adsorption of oil was observed in all three cases as the SEM

images showed adsorbed oil in the form of a muddy-like substance on the biomass

surface. The characteristics of other biomaterials are presented in Table 4.5. A plot of the

diameter versus the percent (weight %) of walnut shell media passing the related sieve is

shown in Figure 4.3 and the particle mean size was 56.28 mm.

4.4 Biosorption of Oil using Different Blomaterials

The pH of all three oil-in-water emulsions was in the range of 7.5-7.6. The study

examined the effect of slightly acidic and slightly basic pH. The adsorption capacities of

the four biomaterials media at pH 7.6 and 5.0 are shown in Figures 4.4 to 4.6. Residual

oil concentrations obtained by using four different biomaterials are presented in Table

4.6.

94

4.3 Characterization of M. rouxii Biomass and Other Adsorbents

The characteristics of autoclaved powdered M. rouxii biomass are presented in

Table 4.4. The surface area of M. rouxii biomass was 20.55 m2/g and the average particle

size was 5.47 x 10"4 cm. Surface charges of autoclaved M. rouxii biomass at different pH

conditions are presented in Figure 4.1. SEM images of the raw M. rouxii biomass are

shown in Figure 4.2a. Pores of the raw M. rouxii biomass are clearly observed from

Figure 4.2a. The surface morphology of the biomass after adsorption of SMO, CO, and

Bright-Edge 80 are shown in Figure 4.2b, c and d, respectively. The SEM images with

best resolutions are presented in the figure. A definite change in the surface morphology

of the biomass following adsorption of oil was observed in all three cases as the SEM

images showed adsorbed oil in the form of a muddy-like substance on the biomass

surface. The characteristics of other biomaterials are presented in Table 4.5. A plot of the

diameter versus the percent (weight %) of walnut shell media passing the related sieve is

shown in Figure 4.3 and the particle mean size was 56.28 mm.

4.4 Biosorption of Oil using Different Biomaterials

The pH of all three oil-in-water emulsions was in the range of 7.5-7.6. The study

examined the effect of slightly acidic and slightly basic pH. The adsorption capacities of

the four biomaterials media at pH 7.6 and 5.0 are shown in Figures 4.4 to 4.6. Residual

oil concentrations obtained by using four different biomaterials are presented in Table

4.6.

94

Table 4.4: Characteristics of M rouxii biomass

Characteristics Values M rouxii

pH 6.4

Moisture content (%) 4.6

Porosity (%) 85

Surface area (m2/g) 20.55

Color Light brown

Chemical analysis (% by weight)' Chitosan — 32.7

Chitin — 9.4

• Lipids — 7.8

Fucose — 3.8

Mannose — 1.6

Galactose — 1.6

Protein — 6.3

Phosphate 23.3

Magnesium — 1.0

Calcium — 1.0

Glucuronic acid — 11.8

1. Composition of M rouxii cell wall was obtained from Bartnicki-Garcia and Nickerson, 1962.

95

Table 4.4: Characteristics of M. rouxii biomass

Characteristics Values M. rouxii

pH 6.4

Moisture content (%) 4.6

Porosity (%) 85

Surface area (m2/g) 20.55

Color Light brown

Chemical analysis (% by weight)1 Chitosan - 32.7

Chitin - 9.4

Lipids - 7.8

Fucose - 3.8

Mannose - 1.6

Galactose - 1.6

Protein - 6.3

Phosphate 23.3

Magnesium -1.0

Calcium - 1.0

Glucuronic acid -11.8

1. Composition of M. rouxii cell wall was obtained from Bartnicki-Garcia and Nickerson, 1962.

95

(a)

Figure 4.2: Scanning electron micrographs (a) raw M rouxii biomass, (b) SMO-adsorbed M rouxii biomass, (c) CO-adsorbed M

rouxii biomass and (d) Bright-Edge 80-adsorbed M rouxii biomass

96

Figure 4.2: Scanning electron micrographs (a) raw M. rouxii biomass, (b) SMO-adsorbed M. rouxii biomass, (c) CO-adsorbed M.

rouxii biomass and (d) Bright-Edge 80-adsorbed M. rouxii biomass

96

Table 4.5: Characteristics of other adsorbents used in the preliminary study

Characteristics Values A. coeruleal Chitosan 2 Walnut shell media3

pH 4.5 6.8 7.5

Moisture content (%) 4.7 2.8 2.0

Density (kg/L) 0.6

Specific gravity 1.2-1.4

Porosity (%) 60 62 52

Surface area (m2/g) 0.68 0.65 for 0.17 Sigma C3646

Color Light brown Faintly beige Light Brown

Chemical analysis (% by weight)

Chitosan — 10.4 > 85% deacetylation

Nitrogen - 0.10

Sp. Cellulose -40.60

Lignin - 20.30

Toluene solubility -0.5-1.0

Methoxyl - 6.5

Chlorine - 0.10

Ash - 1.5

Cutin - 1.0

Note: 1. Percentage of chitosan on A. coerulea cell wall was obtained from Miyoshi et al. 1992. 2. The degree of deacetylation is > 85%; surface area of chitosan was obtained from Uzun 2006. 3. Density, specific gravity, and chemical analysis of walnut shell media were provided by USFilter, USA.

97

Table 4.5: Characteristics of other adsorbents used in the preliminary study

Characteristics Values A. coerulea1 Chitosan2 Walnut shell media3

PH 4.5 6.8 7.5

Moisture content (%) 4.7 2.8 2.0

Density (kg/L) 0.6

Specific gravity 1.2-1.4

Porosity (%) 60 62 52

Surface area (m2/g) 0.68 0.65 for 0.17 Sigma C3646

Color Light brown Faintly beige Light Brown

Chemical analysis (% by Chitosan -10.4 > 85% Nitrogen - 0.10 weight) deacetylation Sp. Cellulose

40.60

Lignin - 20.30

Toluene solubility -0.5-1.0

Methoxyl - 6.5

Chlorine - 0.10

Ash -1.5

Cutin -1.0

Note: 1. Percentage of chitosan on A. coerulea cell wall was obtained from Miyoshi et al. 1992. 2. The degree of deacetylation is > 85%; surface area of chitosan was obtained from Uzun 2006. 3. Density, specific gravity, and chemical analysis of walnut shell media were provided by USFilter, USA.

97

70 -

60

50

0 40

To

a. 30

20

10

• • 0 , • ,

0 0.2 0.4 0.6 0.8

Diamter (mm)

1 1.2 1.4

Figure 4.3: Plot of diameter versus % walnut shell media passing the related sieve

98

O) c 10 M (0 Q.

0.4 0.6 1.4 0 0.2 0.8 1 1.2

Diamter (mm)

Figure 4.3: Plot of diameter versus % walnut shell media passing the related sieve

98

100

90

_80

-5470

7460

1.50

40

x,30

20

<10

0 2 4

Time (h) 6

M. rouxii pH 7.6

A. Coerulea pH 7.6

Chitosan pH 7.6

Walnut shell media pH 7.6 M. rouxii pH 5

- A. coerulea pH 5

-- Chitosan pH 5

Walnut shell media

8

Figure 4.4: Adsorption capacity versus time for SMO

99

Time (h)

M. rouxii pH 7.6

—a— A. Coeailea pH 7.6

—A— Chitosan pH 7.6

—X— Walnut shell media pH 7.6

.0.- M. rouxii pH 5

- -Q-- A. coeailea pH 5

& - Chitosan pH 5

-X-- Walnut shell media

Figure 4.4: Adsorption capacity versus time for SMO

99

100

90

80

I s: 70 E

60

. 50 Uc 40 0 T.: IS 30

i < 20

10

0

--0-- M. rouxii pH 7.6

--e-- A. coerulea pH 7.6

--A-- Chitosan pH 7.6

x-- Walnut shell media pH 7.6

- - -0- - - M. rouxii pH 5.0

0 A. coerulea pH 5.0

---0--- Chitosan pH 5.0

- - -x- - - Walnut shell media pH 5.0

3 4 5 6 7 Time (h)

Figure 4.5: Adsorption capacity versus time for CO

100

• • • • » M. rouxii pH 7.6

—B— - A. coerulea pH 7.6

A— Chitosan pH 7.6

X— - Walnut shell media pH 7.6

- M. rouxii pH 5.0

• A. coerulea pH 5.0

- - A - - - Chitosan pH 5.0

- -X- - - Walnut shell media pH 5.0

2 3 4 Time (h)

Figure 4.5: Adsorption capacity versus time for CO

100

100

590

1.80

.970 o. co uc 60 o 150 0 to Q40

30

20

10

0 1 2 3 4

Time (h) 5

--4)-- M. rouxii pH 7.6

—a-- A. coerulea pH 7.6

--a-- Chitosan pH 7.6

-x-- Walnut shell media pH 7.6

- - -4- - - M. rouxii pH 5.0

---0-- - A. Coerulea pH 5.0

- --a- - - Chitosan pH 5.0

---x--- Walnut shell media pH 50

7

Figure 4.6: Adsorption capacity versus time for Bright-Edge 80

101

100

»90 ::

£

A a

-A

Q - •

X X -——X—

0 ..0 0 —-a— —a —B— —•

2 3 4 Time (h)

—0 M. rouxii pH 7.6

• A. coerulea pH 7.6

—a— - Chitosan pH 7.6

- Walnut shell media pH 7.6

- M. rouxii pH 5.0

—0- -- A. Coerulea pH 5.0

—A- - - Chitosan pH 5.0

—X- -- Walnut shell media DH 5.0

Figure 4.6: Adsorption capacity versus time for Bright-Edge 80

101

Table 4.6: Residual oil concentration obtained using different biomaterials

M rouxii A. coerulea Chitosan Walnut shell Oil Time, pH pH pH pH pH pH pH pH

h 7.6 5.0 7.6 5.0 7.6 5.0 7.6 5.0 SMO

1 75.6 65.7 88.0 72.6 26.6 17.3 150.7 67.7 2 72.0 57.3 82.0 64.5 23.7 12.2 142.0 49.3 3 69.1 46.8 79.1 55.8 13.3 5.0 135.1 35.6 4 66.9 45.8 78.1 55.4 9.1 2.1 128.0 33.8 5 66.3 45.8 77.6 55.6 9.6 1.2 118.1 34.1 6 66.0 45.6 78.0 55.8 9.2 0.8 123.9 35.0

CO 1 35.4 28.9 37.1 32.1 8.0 5.3 116.0 27.1 2 28.9 22.6 33.1 26.7 5.3 2.8 94.3 19.8 3 24.6 18.4 29.9 19.2 1.8 1.2 85.6 12.4 4 23.2 15.3 28.6 18.9 1.7 0.6 84.4 8.8 5 23.9 15.6 27.8 18.7 1.8 0.5 82.8 9.1 6 22.2 15.0 27.2 18.9 1.7 0.3 80.4 8.0

Bright-Edge 80 1 81.1 47.8 84.7 49.8 32.8 14.6 61.7 17.1 2 78.7 38.1 81.1 44.2 27.1 12.2 58.9 14.6 3 73.7 34.7 79.8 41.7 24.7 8.1 56.4 . 10.5 4 72.2 32.1 79.1 39.9 23.7 6.8 54.3 9.4 5 71.8 31.9 78.8 40.1 24.9 6.1 54.5 9.8 6 71.1 32.0 79.0 39.8 22.2 6.3 53.5 9.3

Note: All values 'n mg/L; initial oil concentration = 200 mg/L

102

Table 4.6: Residual oil concentration obtained using different biomaterials

M. rouxii A. coerulea Chitosan Walnut shell Oil Time, pH pH PH pH PH PH PH PH

h 7.6 5.0 7.6 5.0 7.6 5.0 7.6 5.0 SMO

1 75.6 65.7 88.0 72.6 26.6 17.3 150.7 67.7 2 72.0 57.3 82.0 64.5 23.7 12.2 142.0 49.3 3 69.1 46.8 79.1 55.8 13.3 5.0 135.1 35.6 4 66.9 45.8 78.1 55.4 9.1 2.1 128.0 33.8 5 66.3 45.8 77.6 55.6 9.6 1.2 118.1 34.1 6 66.0 45.6 78.0 55.8 9.2 0.8 123.9 35.0

CO 1 35.4 28.9 37.1 32.1 8.0 5.3 116.0 27.1 2 28.9 22.6 33.1 26.7 5.3 2.8 94.3 19.8 3 24.6 18.4 29.9 19.2 1.8 1.2 85.6 12.4 4 23.2 15.3 28.6 18.9 1.7 0.6 84.4 8.8 5 23.9 15.6 27.8 18.7 1.8 0.5 82.8 9.1 6 22.2 15.0 27.2 18.9 1.7 0.3 80.4 8.0

Bright-Edge 8( 1 81.1 47.8 84.7 49.8 32.8 14.6 61.7 17.1 2 78.7 38.1 81.1 44.2 27.1 12.2 58.9 14.6 3 73.7 34.7 79.8 41.7 24.7 8.1 56.4 10.5 4 72.2 32.1 79.1 39.9 23.7 6.8 54.3 9.4 5 71.8 31.9 78.8 40.1 24.9 6.1 54.5 9.8 6 71.1 32.0 79.0 39.8 22.2 6.3 53.5 9.3

Note: All values in mg/L; initial oil concentration = 200 mg/L

102

Each batch test was conducted in triplicate and the mean is represented by one data point

in the figures. The deviations of the three points from the mean were found to be within

±1%; the standard deviation for each set of data was found to be relatively small

compared to the mean. Among the three oils studied, maximum oil removals were

observed in the case of canola oil for all four adsorbents (Table 4.7). The adsorption

capacities for canola oil at pH 5.0 were 92.5, 90.5, 99.9, and 96 mg/g by M rouxii, A.

coerulea, chitosan and walnut shell media, respectively. This could be due to the fact that

in comparison with the other two emulsions, vegetable oil emulsions are known for their

limited stability (Vesala et al. 1985). Vegetable oils predominantly consist of

triglycerides (Pushkarev et al. 1983) and emulsions based on oils from the triglyceride

group showed poor stability in comparison with emulsions based on hydrocarbon oils

(Ozgen et al. 2006). At pH 5.0, minimum oil removal was obtained in the case of Bright-

Edge 80 using the four adsorbents studied. Bright-Edge 80 cutting oil is the one with the

least interfacial tension among the oils studied. Low interfacial tension would result in

reduced free surface energy associated with the formation of droplets. It enhances the

formation of smaller emulsion droplets with greater kinetic stability (Aserin 2007). Better

oil removals for all three oils were obtained at pH 5.0 than at pH 7.6 for all four

adsorbents. According to Figure 4.1, in the case of all three oils, the zeta-potential value

was less electronegative at pH 5.0 when compared to that at pH 7.0. Zeta-potential

indicates the stability of oil-in-water emulsions. A higher magnitude of zeta-potential

would refer to greater forces of repulsion between particles in the emulsion, which would

lead to better stability of the emulsions (Reynolds and Richards 1995). With a decrease in

103

Each batch test was conducted in triplicate and the mean is represented by one data point

in the figures. The deviations of the three points from the mean were found to be within

±1%; the standard deviation for each set of data was found to be relatively small

compared to the mean. Among the three oils studied, maximum oil removals were

observed in the case of canola oil for all four adsorbents (Table 4.7). The adsorption

capacities for canola oil at pH 5.0 were 92.5, 90.5, 99.9, and 96 mg/g by M. rouxii, A,

coerulea, chitosan and walnut shell media, respectively. This could be due to the fact that

in comparison with the other two emulsions, vegetable oil emulsions are known for their

limited stability (Vesala et al. 1985). Vegetable oils predominantly consist of

triglycerides (Pushkarev et al. 1983) and emulsions based on oils from the triglyceride

group showed poor stability in comparison with emulsions based on hydrocarbon oils

(Ozgen et al. 2006). At pH 5.0, minimum oil removal was obtained in the case of Bright-

Edge 80 using the four adsorbents studied. Bright-Edge 80 cutting oil is the one with the

least interfacial tension among the oils studied. Low interfacial tension would result in

reduced free surface energy associated with the formation of droplets. It enhances the

formation of smaller emulsion droplets with greater kinetic stability (Aserin 2007). Better

oil removals for all three oils were obtained at pH 5.0 than at pH 7.6 for all four

adsorbents. According to Figure 4.1, in the case of all three oils, the zeta-potential value

was less electronegative at pH 5.0 when compared to that at pH 7.0. Zeta-potential

indicates the stability of oil-in-water emulsions. A higher magnitude of zeta-potential

would refer to greater forces of repulsion between particles in the emulsion, which would

lead to better stability of the emulsions (Reynolds and Richards 1995). With a decrease in

103

Table 4.7: Oil removals by biomaterials

Oil removals, % Oil pH M rouxii A. coerulea Chitosan Walnut shell SMO 7.6 67 61 96 38

5.0 77 72 99 83 CO 7.6 89 87 99 60

5.0 93 91 99 96 Bright-Edge 80 7.6 65 61 89 73

5.0 84 80 97 96 Note: Initial concentration = 200 mg/L; Time of contact = 6 h; Adsorbent dose = 0.2 g.

104

Table 4.7: Oil removals by biomaterials

Oil removals, % Oil PH M. rouxii A. coerulea Chitosan Walnut shell SMO 7.6 67 61 96 38

5.0 77 72 99 83 CO 7.6 89 87 99 60

5.0 93 91 99 96 Bright-Edge 80 7.6 65 61 89 73

5.0 84 80 97 96 Note: Initial concentration = 200 mg/L; Time of contact = 6 h; Adsorbent dose = 0.2 g.

104

the magnitude of zeta-potential, the possibility of coagulation of dispersed particles

increases and the emulsion becomes less stable (Riddick 1968; Pushkarev et al. 1983).

Hence, a lower electro-negativity value of pH 5.0 for the three oils studied resulted in

higher removal efficiency. Similar behavior was found by Ahmad et al. (2005a) with

chitosan, which was used to remove residue oil from POME. They observed that when

the pH was lower than 5.0, the percentage of residue oil adsorption could be increased to

99%. Also, at a pH of more than 5.0, the percentage of adsorption was found to decrease

and when the pH was 7.0, the percentage of residue oil adsorption was the lowest. Similar

trends were observed with bentonite and activated carbon which, when used for

adsorption of residue oil from POME, showed a higher oil removal at a pH of less than

5.0 (Ahmad et al. 2005b). It was suggested that the acidic conditions acted as a catalyst to

catalyze the reaction between the residue oil molecules and the adsorption site of chitosan

(-NH2 group) (Ahmad et al. 2005a). In the case of fungal biomass, their surfaces have

been generally observed to be negatively charged because of the ionization of functional

groups present (Yan and Viraraghavan 2003). At acidic pH, some of the functional

groups present in the fungal cell wall will be positively charged and the negative charge

intensity on the sites will be reduced, which might have an effect on the sorption

characteristics of the biomass (Yan and Viraraghavan 2003). Uptake by dead fungal cells

takes place as a result of the functional groups of the cell and, in particular, the cell wall.

The mechanism of uptake by the cell wall has been broadly categorized as: (1) uptake

directed by functional groups such as phosphate, carboxyl, amine and phosphate diester

species of the compounds; and (2) physio-chemical interactions directed by adsorption

105

the magnitude of zeta-potential, the possibility of coagulation of dispersed particles

increases and the emulsion becomes less stable (Riddick 1968; Pushkarev et al. 1983).

Hence, a lower electro-negativity value of pH 5.0 for the three oils studied resulted in

higher removal efficiency. Similar behavior was found by Ahmad et al. (2005a) with

chitosan, which was used to remove residue oil from POME. They observed that when

the pH was lower than 5.0, the percentage of residue oil adsorption could be increased to

99%. Also, at a pH of more than 5.0, the percentage of adsorption was found to decrease

and when the pH was 7.0, the percentage of residue oil adsorption was the lowest. Similar

trends were observed with bentonite and activated carbon which, when used for

adsorption of residue oil from POME, showed a higher oil removal at a pH of less than

5.0 (Ahmad et al. 2005b). It was suggested that the acidic conditions acted as a catalyst to

catalyze the reaction between the residue oil molecules and the adsorption site of chitosan

(-NH2 group) (Ahmad et al. 2005a). In the case of fungal biomass, their surfaces have

been generally observed to be negatively charged because of the ionization of functional

groups present (Yan and Viraraghavan 2003). At acidic pH, some of the functional

groups present in the fungal cell wall will be positively charged and the negative charge

intensity on the sites will be reduced, which might have an effect on the sorption

characteristics of the biomass (Yan and Viraraghavan 2003). Uptake by dead fungal cells

takes place as a result of the functional groups of the cell and, in particular, the cell wall.

The mechanism of uptake by the cell wall has been broadly categorized as: (1) uptake

directed by functional groups such as phosphate, carboxyl, amine and phosphate diester

species of the compounds; and (2) physio-chemical interactions directed by adsorption

105

phenomena (Kapoor and Viraraghavan 1995). Biosorption mechanisms have been found

to include ion-exchange, co-ordination, complexation, chelation, adsorption and

microprecipitation (Yan and Viraraghavan 2000). At this stage, it is necessary to carry

out more detailed studies to understand the specific mechanism involved in the

biosorption of oil by fungal adsorbents. The mechanism of oil removal in the case of

walnut shell media has been observed to be sorption (Rahman 1992). Oil is found at the

interstices of the finely divided granules of the walnut shell media, which was observed

to be an excellent coalescing material.

Among the two fungal biomasses studied, M rouxii was found to be better for oil

sorption than A. coerulea. The adsorption capacity of M rouxii at pH 5.0 for SMO,

canola oil and Bright-Edge 80 were 77.2, 92.5, and 84 mg/g, respectively. The higher

adsorption capacity for oil by M rouxii, when compared to A. coerulea, could be due to

its higher BET surface area. Most adsorbents with a higher surface area have been found

to facilitate adsorption of residue oil (Ahmad et al. 2005a). Mycelium of M rouxii grows

in the form of suspended' growth leading to a larger surface area of the biomass for

adsorption (Yan and Viraraghavan 2000). In the case of A. coerulea, it grows in the form

of pellets with a lower surface area. Most Absidia species are found to have a strong

inclination to form pellets during their growth in a fermenter (Davoust and Persson

1992). Yan and Viraraghavan (2000) have observed that the biosorption capacity of

heavy metals were higher for M rouxii than for A. niger and the difference was ascribed

to the larger surface area of M rouxii. The effectiveness of M rouxii in removing the oils

follows this order: CO, SMO, and Bright-Edge 80. The oil adsorption capacity by M

106

phenomena (Kapoor and Viraraghavan 1995). Biosorption mechanisms have been found

to include ion-exchange, co-ordination, complexation, chelation, adsorption and

microprecipitation (Yan and Viraraghavan 2000). At this stage, it is necessary to carry

out more detailed studies to understand the specific mechanism involved in the

biosorption of oil by fungal adsorbents. The mechanism of oil removal in the case of

walnut shell media has been observed to be sorption (Rahman 1992). Oil is found at the

interstices of the finely divided granules of the walnut shell media, which was observed

to be an excellent coalescing material.

Among the two fungal biomasses studied, M. rouxii was found to be better for oil

sorption than A. coerulea. The adsorption capacity of M. rouxii at pH 5.0 for SMO,

canola oil and Bright-Edge 80 were 77.2, 92.5, and 84 mg/g, respectively. The higher

adsorption capacity for oil by M. rouxii, when compared to A. coerulea, could be due to

its higher BET surface area. Most adsorbents with a higher surface area have been found

to facilitate adsorption of residue oil (Ahmad et al. 2005a). Mycelium of M. rouxii grows

in the form of suspended growth leading to a larger surface area of the biomass for

adsorption (Yan and Viraraghavan 2000). In the case of A. coerulea, it grows in the form

of pellets with a lower surface area. Most Absidia species are found to have a strong

inclination to form pellets during their growth in a fermenter (Davoust and Persson

1992). Yan and Viraraghavan (2000) have observed that the biosorption capacity of

heavy metals were higher for M. rouxii than for A. niger and the difference was ascribed

to the larger surface area of M. rouxii. The effectiveness of M. rouxii in removing the oils

follows this order: CO, SMO, and Bright-Edge 80. The oil adsorption capacity by M.

106

rouxii was lower when compared with that of crab shell chitosan. However, M rouxii is

found to contain approximately 32% of chitosan in its cell wall (Bartnicki-Garcia and

Nickerson 1962, exhibiting an adsorption capacity of 92.5 mg/g for canola oil at pH 5.0

while pure chitosan showed an adsorption capacity of 99.9 mg/g. Oil uptake by M rouxii

is also influenced by other functional groups present in the cell wall.

4.5 Factorial Design of Experiments

The design matrix of uncoded values for the factors and the response in terms of

the percent removal of SMO, CO and Bright-Edge 80 for all experimental runs including

replicates, are shown in Table 4.8. A linear regression model was fitted for the

experimental data using the least square technique using MINITAB. The model

coefficients for the removal, the effects and standardized effects of the factors and

interactions, and p-values of the effects in the model are shown in Tables 4.9 - 4.11 for

SMO, CO, and Bright-Edge 80, respectively. The net effect is a difference between the

responses of two levels (high and low level) of factors; the regression model coefficients

are obtained by dividing the net effects by two. The standardized effects are obtained by

dividing the regression coefficients by standard error coefficient (Antony, 2003). The p-

value is the probability value used to determine the effects in the model that are

statistically significant. P-value is termed as the probability that measures the strength of

the evidence against a null hypothesis (Moore, 2009). As for a 95% confidence level, the

p-value should be less than or equal to 0.05 (0 < 0.05) for the effect to be statistically

significant.

107

rouxii was lower when compared with that of crab shell chitosan. However, M, rouxii is

found to contain approximately 32% of chitosan in its cell wall (Bartnicki-Garcia and

Nickerson 1962, exhibiting an adsorption capacity of 92.5 mg/g for canola oil at pH 5.0

while pure chitosan showed an adsorption capacity of 99.9 mg/g. Oil uptake by M. rouxii

is also influenced by other functional groups present in the cell wall.

4.5 Factorial Design of Experiments

The design matrix of uncoded values for the factors and the response in terms of

the percent removal of SMO, CO and Bright-Edge 80 for all experimental runs including

replicates, are shown in Table 4.8. A linear regression model was fitted for the

experimental data using the least square technique using MINITAB. The model

coefficients for the removal, the effects and standardized effects of the factors and

interactions, and /^-values of the effects in the model are shown in Tables 4.9 - 4.11 for

SMO, CO, and Bright-Edge 80, respectively. The net effect is a difference between the

responses of two levels (high and low level) of factors; the regression model coefficients

are obtained by dividing the net effects by two. The standardized effects are obtained by

dividing the regression coefficients by standard error coefficient (Antony, 2003). The p-

value is the probability value used to determine the effects in the model that are

statistically significant. P-value is termed as the probability that measures the strength of

the evidence against a null hypothesis (Moore, 2009). As for a 95% confidence level, the

p-value should be less than or equal to 0.05 (0 < 0.05) for the effect to be statistically

significant.

107

Table 4.8: Uncoded design table for the factors and response

Run pH Temp- erature

(°C)

Dose (g)

Conc- entration (mg/L)

Speed (rpm)

% removal of SMO

% removal of CO

% removal

of Bright-Edge 80

1 3 5 0.05 350 100 99.90 99.97 99.90 2 3 5 0.50 50 100 80.20 99.40 89.14 3 9 30 0.05 50 200 38.00 54.00 87.60 4 3 5 0.05 50 200 94.80 99.60 91.80 5 9 5 0.50 50 200 27.00 59.60 68.80 6 9 5 0.50 350 100 51.70 46.80 13.40 7 9 30 0.50 50 100 60.00 61.60 57.40 8 3 30 0.50 50 200 96.40 98.00 88.20 9 3 5 0.50 350 200 99.77 99.77 98.80 10 9 5 0.05 350 200 9.10 23.10 8.00 11 9 30 0.50 350 200 58.00 79.70 69.40 12 9 30 0.05 50 200 39.10 56.00 88.00 13 9 5 0.05 50 100 23.40 90.40 44.60 14 3 30 0.50 50 200 96.70 97.70 88.00 15 3 5 0.05 350 100 99.70 99.80 99.10 16 3 30 0.05 350 200 98.40 99.80 99.60 17 3 30 0.05 50 100 99.60 99.80 90.60 18 9 5 0.50 50 200 28.00 60.10 68.90 19 9 30 0.50 350 200 58.40 80.10 70.00 20 3 5 0.50 50 100 80.60 99.60 89.90 21 3 30 0.05 50 100 99.50 99.30 91.00 22 3 30 0.50 350 100 97.40 97.90 97.00 23 9 5 0.50 350 100 52.00 47.10 14.20 24 3 5 0.50 350 200 99.40 98.60 98.60 25 3 30 0.05 350 200 98.80 99.70 99.30 26 9 30 0.05 350 100 37.10 55.40 36.80 27 9 5 0.05 50 100 25.00 91.00 44.40 28 9 30 0.05 350 100 37.40 55.10 37.20 29 3 30 0.50 350 100 97.70 97.20 97.20 30 9 30 0.50 50 100 61.20 62.00 58.00 31 9 5 0.05 350 200 10.00 22.80 9.00 32 3 5 0.05 50 200 95.10 99.40 92.00

108

Table 4.8: Uncoded design table for the factors and response

Run PH Temp Dose Conc Speed % % % erature (g) entration (rpm) removal removal removal

(°C) (mg/L) ofSMO of CO of Bright-Edge 80

1 3 5 0.05 350 100 99.90 99.97 99.90 2 3 5 0.50 50 100 80.20 99.40 89.14 3 9 30 0.05 50 200 38.00 54.00 87.60 4 3 5 0.05 50 200 94.80 99.60 91.80 5 9 5 0.50 50 200 27.00 59.60 68.80 6 9 5 0.50 350 100 51.70 46.80 13.40 7 9 30 0.50 50 100 60.00 61.60 57.40 8 3 30 0.50 50 200 96.40 98.00 88.20 9 3 5 0.50 350 200 99.77 99.77 98.80 10 9 5 0.05 350 200 9.10 23.10 8.00 11 9 30 0.50 350 200 58.00 79.70 69.40 12 9 30 0.05 50 200 39.10 56.00 88.00 13 9 5 0.05 50 100 23.40 90.40 44.60 14 3 30 0.50 50 200 96.70 97.70 88.00 15 3 5 0.05 350 100 99.70 99.80 99.10 16 3 30 0.05 350 200 98.40 99.80 99.60 17 3 30 0.05 50 100 99.60 99.80 90.60 18 9 5 0.50 50 200 28.00 60.10 68.90 19 9 30 0.50 350 200 58.40 80.10 70.00 20 3 5 0.50 50 100 80.60 99.60 89.90 21 3 30 0.05 50 100 99.50 99.30 91.00 22 3 30 0.50 350 100 97.40 97.90 97.00 23 9 5 0.50 350 100 52.00 47.10 14.20 24 3 5 0.50 350 200 99.40 98.60 98.60 25 3 30 0.05 350 200 98.80 99.70 99.30 26 9 30 0.05 350 100 37.10 55.40 36.80 27 9 5 0.05 50 100 25.00 91.00 44.40 28 9 30 0.05 350 100 37.40 55.10 37.20 29 3 30 0.50 350 100 97.70 97.20 97.20 30 9 30 0.50 50 100 61.20 62.00 58.00 31 9 5 0.05 350 200 10.00 22.80 9.00 32 3 5 0.05 50 200 95.10 99.40 92.00

108

Table 4.9: Estimated effects and coefficients for the removal of SMO (% coded units)

Term Net effect

Regression coefficient

Standardized effect (T)

p-value

Constant 67.17 753.50 0.000

pH -57.41 -28.71 -322.02 0.000

Temperature 12.38 6.19 69.42 0.000

Dose 8.72 4.36 48.93 0.000

Concentration 3.76 1.88 21.09 0.000

Speed -3.46 -1.73 -19.43 0.000

pH*Temperature 8.00 4.00 44.86 0.000

pH*Dose 13.43 6.71 75.31 0.000

pH*Concentration -2.26 -1.13 -12.68 0.000

pH*speed -6.56 -3.28 -36.80 0.000

Temperature*Dose 1.01 0.51 5.69 0.000

Temperature*Concentration -4.67 -2.34 -26.21 0.000

Temperature*speed 2.70 1.35 15.15 0.000

Dose*Concentration 6.77 3.39 37.99 0.000

Dose*speed 1.32 0.66 7.42 0.000

Concentration*speed -1.66 -0.83 -9.34 0.000

Note: Standard error coefficient for all cases = 0.08914

109

Table 4.9: Estimated effects and coefficients for the removal of SMO (% coded units)

Term Net effect



p-v alue

Constant 67.17 753.50 0.000

PH -57.41 -28.71 -322.02 0.000

Temperature 12.38 6.19 69.42 0.000

Dose 8.72 4.36 48.93 0.000

Concentration 3.76 1.88 21.09 0.000

Speed -3.46 -1.73 -19.43 0.000

pH*Temperature 8.00 4.00 44.86 0.000

pH*Dose 13.43 6.71 75.31 0.000


pH*speed -6.56 -3.28 -36.80 0.000

Temperature ""Dose 1.01 0.51 5.69 0.000

Temperature*Concentration -4.67 -2.34 -26.21 0.000

T emperature* speed 2.70 1.35 15.15 0.000

Dose* Concentration 6.77 3.39 37.99 0.000

Dose*speed 1.32 0.66 7.42 0.000

Concentration* speed -1.66 -0.83 -9.34 0.000


109

Table 4.10: Estimated effects and coefficients for the removal of CO (% coded units)

Term Net effect



p-value

Constant 79.07 922.86 0.000

pH -40.05 -20.02 -233.69 0.000

Temperature 3.52 1.76 20.52 0.000

Dose 2.50 1.25 14.59 0.000

Concentration -7.79 -3.90 -45.47 0.000

Speed -4.65 -2.32 -27.13 0.000

pH*Temperature 4.36 2.18 25.44 0.000

pH*Dose 3.65 1.83 21.30 0.000


pH*Speed -4.60 -2.30 -26.84 0.000

Temperature*Dose 4.39 2.19 25.60 0.000

Temperature*Concentration 12.35 6.18 72.09 0.000

Temperature*Speed 9.24 4.62 53.90 0.000


Dose*Speed 12.40 6.20 72.34 0.000

Concentration*Speed 5.19 2.59 30.27 0.000


110

Table 4.10: Estimated effects and coefficients for the removal of CO (% coded units)

Term Net Regression Standardized p-value effect coefficient effect (T)

Constant 79.07 922.86 0.000

pH -40.05 -20.02 -233.69 0.000

Temperature 3.52 1.76 20.52 0.000

Dose 2.50 1.25 14.59 0.000

Concentration -7.79 -3.90 -45.47 0.000

Speed -4.65 -2.32 -27.13 0.000

pH*Temperature 4.36 2.18 25.44 0.000

pH*Dose 3.65 1.83 21.30 0.000


pH* Speed -4.60 -2.30 -26.84 0.000

Temperature * Dose 4.39 2.19 25.60 0.000

T emperature * Concentration 12.35 6.18 72.09 0.000


Dose* Concentration 8.94 4.47 52.15 0.000

Dose* Speed 12.40 6.20 72.34 0.000

Concentration* Speed 5.19 2.59 30.27 0.000


110

Table 4.11: Estimated effects and coefficients for the removal of Bright-Edge 80 (% coded units)

Term Net effect



p-value

Constant 71.43 1095.02 0.000

pH -45.90 -22.95 -351.83 0.000

Temperature 14.05 7.02 107.67 0.000

Dose 3.00 1.50 23.01 0.000

Concentration -11.93 -5.96 -91.42 0.000

Speed 10.39 5.19 79.60 0.000

pH*Temperature 15.09 7.55 115.66 0.000

pH*Dose 5.06 2.53 38.78 0.000


pH*Speed 10.08 5.04 77.24 0.000

Temperature*Dose -3.62 -1.81 -27.71 0.000

Temperature*Concentration 6.64 3.32 50.89 0.000



Dose*Speed 6.42 3.21 49.23 0.000

Concentration*Speed -3.15 -1.57 -24.12 0.000


111

Table 4.11: Estimated effects and coefficients for the removal of Bright-Edge 80 (% coded units)

Term Net Regression Standardized p-value effect coefficient effect (T)

Constant 71.43 1095.02 0.000

pH -45.90 -22.95 -351.83 0.000

Temperature 14;05 7.02 107.67 0.000

Dose 3.00 1.50 23.01 0.000

Concentration -11.93 -5.96 -91.42 0.000

Speed 10.39 5.19 79.60 0.000

pH*Temperature 15.09 7.55 115.66 0.000

pH*Dose 5.06 2.53 38.78 0.000


pH* Speed 10.08 5.04 77.24 0.000

Temperature *Dose -3.62 -1.81 -27.71 0.000

Temperature * Concentration 6.64 3.32 50.89 0.000

Temperature * Speed 5.23 2.61 40.07 0.000


Dose*Speed 6.42 3.21 49.23 0.000

Concentration* Speed -3.15 -1.57 -24.12 0.000


111

A statistical analysis (normal probability plot) of the data in terms of the standardized

residual was also conducted to verify the normality of the data. The absolute value of the

estimated effect determines its relative strength over the response. The higher the value of

the effect, the higher the influence over the response. The significance level for this

model was chosen to be 0.05 (95% confidence level). Solution pH had the highest effect

on the removal of all three oils. A negative sign of the effect indicates that a low factor

setting (-1) would result in a higher removal (Minitab 2007). A decrease in solution pH to

3.0 from 9.0 increased the percent removal of SMO, CO and Bright-Edge 80 to as high as

99%. It was found that pH 3.0 is the point at which the zeta-potential of autoclaved M

rouxii was zero. Acidic pH has been found to increase the percentage of residue oil

adsorption to 99% according to Ahmad et al (2005a). Similar trends were observed with

bentonite and activated carbon which when used for adsorption of residue oil from

POME, showed higher oil removal at a pH of less than 5.0 (Ahmad et al 2005b). As for

the removal of SMO, the magnitude of effects of each factor and their interactions were

found to increase in the following order: pH > pH*Dose > Temperature. With respect to

the removal of CO, the increasing order of the effects was given by: pH > Dose*Speed >

Temperature*Concentration. Bright-Edge 80, was given by: pH > pH*Concentration >

pH*Temperature > Temperature. The p-values in the estimated effects and coefficients

were used to determine statistically significant individual and interaction effects. All

factors or combination of factors were found to be statistically significant (p-values <

0.05) for the removal of SMO, CO and Bright-Edge 80.

112

A statistical analysis (normal probability plot) of the data in terms of the standardized

residual was also conducted to verify the normality of the data. The absolute value of the

estimated effect determines its relative strength over the response. The higher the value of

the effect, the higher the influence over the response. The significance level for this

model was chosen to be 0.05 (95% confidence level). Solution pH had the highest effect

on the removal of all three oils. A negative sign of the effect indicates that a low factor

setting (-1) would result in a higher removal (Minitab 2007). A decrease in solution pH to

3.0 from 9.0 increased the percent removal of SMO, CO and Bright-Edge 80 to as high as

99%. It was found that pH 3.0 is the point at which the zeta-potential of autoclaved M.

rouxii was zero. Acidic pH has been found to increase the percentage of residue oil

adsorption to 99% according to Ahmad et al (2005a). Similar trends were observed with

bentonite and activated carbon which when used for adsorption of residue oil from

POME, showed higher oil removal at a pH of less than 5.0 (Ahmad et al 2005b). As for

the removal of SMO, the magnitude of effects of each factor and their interactions were

found to increase in the following order: pH > pH*Dose > Temperature. With respect to

the removal of CO, the increasing order of the effects was given by: pH > Dose*Speed >

Temperature*Concentration. Bright-Edge 80, was given by: pH > pH*Concentration >

pH * Temperature > Temperature. The /^-values in the estimated effects and coefficients

were used to determine statistically significant individual and interaction effects. All

factors or combination of factors were found to be statistically significant (p-values <

0.05) for the removal of SMO, CO and Bright-Edge 80.

112

4.5.1 Pareto Plot of Effect

The Pareto plot visually represents the absolute values of the effects of the main

factors and the effects of the interaction of factors. A reference line is drawn to indicate

the factors, which extend past this line, are potentially important (Antony 2003). pH had

the greatest effect on the removal for all three oils (Figures 4.7 - 4.9). In the figures, the

reference line can be found near the vertical axis. The effects above the reference line are

statistically significant at a 95% confidence level and all effects were found to be

statistically significant. The relative importance of each factor and the combination of

factors can be observed according to the Pareto plots for all three oils.

4.5.2 Main Effects Plot

The main effects plot is shown in Figures 4.10 — 4.12 with respect to the removal

of SMO, CO, and Bright-Edge 80, respectively. It indicates the relative strength of the

effects of various factors. A main effect is present when the mean response changes

across the level of a factor. The sign of the main effect indicates the directions of the

effect. It can be observed from the main effects plot that for all three oils, the effect of pH

was characterized by a greater degree of departure from the overall mean. The pH had a

negative effect upon the removal of all three oils. For SMO, temperature and dose

showed a slight positive effect upon removal. As for CO, the concentration had a slight

negative effect upon removal. With respect to the Bright-Edge 80, temperature and speed

had a slight positive effect while concentration was found to have a slight negative effect.

All other factors showed a smaller change. The patterns were previously identified by

113

4.5.1 Pareto Plot of Effect

The Pareto plot visually represents the absolute values of the effects of the main

factors and the effects of the interaction of factors. A reference line is drawn to indicate

the factors, which extend past this line, are potentially important (Antony 2003). pH had

the greatest effect on the removal for all three oils (Figures 4.7 - 4.9). In the figures, the

reference line can be found near the vertical axis. The effects above the reference line are

statistically significant at a 95% confidence level and all effects were found to be

statistically significant. The relative importance of each factor and the combination of

factors can be observed according to the Pareto plots for all three oils.

4.5.2 Main Effects Plot

The main effects plot is shown in Figures 4.10-4.12 with respect to the removal

of SMO, CO, and Bright-Edge 80, respectively. It indicates the relative strength of the

effects of various factors. A main effect is present when the mean response changes

across the level of a factor. The sign of the main effect indicates the directions of the

effect. It can be observed from the main effects plot that for all three oils, the effect of pH

was characterized by a greater degree of departure from the overall mean. The pH had a

negative effect upon the removal of all three oils. For SMO, temperature and dose

showed a slight positive effect upon removal. As for CO, the concentration had a slight

negative effect upon removal. With respect to the Bright-Edge 80, temperature and speed

had a slight positive effect while concentration was found to have a slight negative effect.

All other factors showed a smaller change. The patterns were previously identified by

Pareto Chart of the Standardized Effects (response is % Removal, Alpha = 0.05)

2..1 A

AC

B

C AB

CD AE BD D

E

BE AD DE

CE BC

0 50 100 150 200 250 Standardized Effect

300 350

Factor Name A pH B Temperature C Dose D Concentration E speed

Figure 4.7: Pareto chart for standardized effects for the removal of SMO

114


2.1 Factor Name A PH B Temperature C Dose

D Concentration E speed

CD-

0 50 100 150 200 250 300 350 Standardized Effect

Figure 4.7: Pareto chart for standardized effects for the removal of SMO

114


2.1 A

CE

BD BE CD

D AD DE

AE

BC AB AC

B

0 50 100 150 Standardized Effect

200 250

Factor Name A pH B Temperature C Dose D Concentration E Speed

Figure 4.8: Pareto chart for standardized effects for the removal of CO

115

[2

2.1

A

CE

BD

BE

CD

D

AD

DE

E

AE

BC

AB

AC

B

C


~50 100 150 Standardized Effect

Factor Name A PH B Temperature C Dose D Concentration E Speed

200 *250

Figure 4.8: Pareto chart for standardized effects for the removal of CO

115


1

2.1 A

AD

AB B D

E AE BD CE

CD BE

AC sc DE

C

0 . I

100 200 300 Standardized Effect

400

Factor Name A pH B Temperature C Dose D Concentration E Speed

Figure 4.9: Pareto chart for standardized effects for the removal of Bright-Edge 80

116

2.1

A AO

AB

B D E

AE

BD

CE CD BE

AC

BC

DE C

w


1

loo"

Factor Name A PH B Temperature C Dose D Concentration E Speed

200 300 Standardized Effect

"400

Figure 4.9: Pareto chart for standardized effects for the removal of Bright-Edge 80

116

Main Effects Plot for % Removal Data Means

100

80

60

C 40

i 100

80

60

40

pH Temperature Dose

_.-----.

\

•- e-----

3 5 Concentration speed

•

50 350 100 200

0.

Figure 4.10: Main effects plot for the removal of SMO

117

o.


Temperature Dose 100-

80-

60-

40-c

i 0.05 0.50

speed Concentration 100-

80-

60-

50 350 100 200

Figure 4.10: Main effects plot for the removal of SMO

117


100

90

80

70

60

pH Temperature Dose

3 9

100-

90-

80-

70-

60-

Concentration Speed

50 350 100 200

0.05

Figure 4.11: Main effects plot for the removal of CO

118

0.50


Concentration

Figure 4.11: Main effects plot for the removal of CO

118


C

i

90

80

70

60

50

90

80

70

60

50

pH Temperature Dose

—,=,",-----'°---.

\

•--' .. . .'a

3 Concentration Speed

.----'--.

50 350 100 200

0.05 0.50

Figure 4.12: Main effects plot for the removal of Bright-Edge 80

119


Temperature Dose

90-

80-

70-

60-

0.05 0.50

Speed Concentration

90-

80-

70-

60-

50-

50 350 100 200

Figure 4.12: Main effects plot for the removal of Bright-Edge 80

119

statistical significance. Statistical analysis of experimental data showed the effects of the

factors upon removal percentage of the three oils were not similar. This could be due to

the fact that the composition of the three oil-in-water emulsions used in the study was

different.

4.5.3 Interaction Effects Plot

The interaction effects plots are shown in Figures 4.13 — 4.15 for the removal of

SMO, CO, and Bright-Edge 80, respectively. The plots provide the mean response of two

factors at all possible combinations of their settings. If the lines are not parallel, it is an

indication of interaction between the two factors (Antony 2003). The interaction plots for

all three oils showed the pH interacted strongly with all other factors indicating pH to be

a predominant influencing factor in removal. A decrease in the solution pH to 3.0

increased the percent removal of the oils. The concentration of SMO and the dose of the

adsorbent showed a minor interaction with each other (Figure 4.13). When an adsorbent

dose of 0.5 g was used, the percent removal of SMO decreased at a SMO concentration

of 50 mg/L and the percent removal of SMO increased at a SMO concentration of 350

mg/L. Percent removal of SMO appeared not to be affected at a low dose of 0.05 g,

irrespective of the SMO concentration. A higher removal of SMO was observed at a dose

of 0.5 g for both concentrations than at a dose of 0.05 g. Adsorbent dose was previously

found to be statistically significant in the removal of SMO. As for CO, combinations of

the adsorbent dose and the rotational speed, solution temperature and rotational speed,

120

statistical significance. Statistical analysis of experimental data showed the effects of the

factors upon removal percentage of the three oils were not similar. This could be due to

the fact that the composition of the three oil-in-water emulsions used in the study was

different.

4.5.3 Interaction Effects Plot

The interaction effects plots are shown in Figures 4.13-4.15 for the removal of

SMO, CO, and Bright-Edge 80, respectively. The plots provide the mean response of two

factors at all possible combinations of their settings. If the lines are not parallel, it is an

indication of interaction between the two factors (Antony 2003). The interaction plots for

all three oils showed the pH interacted strongly with all other factors indicating pH to be

a predominant influencing factor in removal. A decrease in the solution pH to 3.0

increased the percent removal of the oils. The concentration of SMO and the dose of the

adsorbent showed a minor interaction with each other (Figure 4.13). When an adsorbent

dose of 0.5 g was used, the percent removal of SMO decreased at a SMO concentration

of 50 mg/L and the percent removal of SMO increased at a SMO concentration of 350

mg/L. Percent removal of SMO appeared not to be affected at a low dose of 0.05 g,

irrespective of the SMO concentration. A higher removal of SMO was observed at a dose

of 0.5 g for both concentrations than at a dose of 0.05 g. Adsorbent dose was previously

found to be statistically significant in the removal of SMO. As for CO, combinations of

the adsorbent dose and the rotational speed, solution temperature and rotational speed,

120

Interaction Plot for % Removal Data Means

Figure 4.13: Interaction effects plot for the removal of SMO

121


30 0,05 0,^0 SO 3y 100 2g0

w .s*

W

•- *

*- » ft-

"" *•«

*- • •— — Hi i I

"

Dole

-III

Concentration —

PH

• 3

-m- 9

Temperature

m 5

-m- 30

Dose

0.05

—*- 0.50

Concentration

50

350

Figure 4.13: Interaction effects plot for the removal of SMO

121


Figure 4.14: Interaction effects plot for the removal of CO

122


J 0.(15 Q.ft) 2 2L 200

wr

Tmpanhn ^

f :

DOM

Concvtattn

PH

3

9

Temperature

5

30

Dose

m 0.05

0.50

Concentration

50

-9- 350

Figure 4.14: Interaction effects plot for the removal of CO

122

Figure 4.15: Interaction effects plot for the removal of Bright-Edge 80

123


PH

J 2L

Tupwhw

0.05 OfO

DMI

100 29a -100

pH

• 75 • 3

-50 9

-100 -100 Temperature

•75 • 5

-50 30

-100 Dose

-75 —•— 0.05

-50 —Wr~ 0.50

- 100 - 100 Concentration

k 75 m 50

-so 350

Figure 4.15: Interaction effects plot for the removal of Bright-Edge 80

123

solution temperature and CO concentration and adsorbent dose and CO concentration

showed antagonistic effects (Figure 4.14). The concentration of CO and solution

temperature showed a slight interaction with each other. The percent removal of CO was

not affected by the rotational speed at a CO concentration of 350 mg/L while the percent

removal of CO increased at a CO concentration of 50 mg/L and a rotational speed of 100

rpm. The interaction effect between temperature and adsorbent dose showed that percent

removal of CO was higher at a solution temperature of 30°C and an adsorbent dose of 0.5

g. The percent removal of CO remained the same at a solution temperature of 5°C,

irrespective of the adsorbent dose. In the case of Bright-Edge 80, the solution temperature

and Bright-Edge 80 concentration, solution temperature and adsorbent dose, and solution

temperature and rotational speed had an interaction with each other (Figure 4.15). In all

cases, the percent removal of Bright-Edge 80 was found to be higher at 30 °C. The

adsorbent dose and Bright-Edge 80 concentration and adsorbent dose and rotational

speed were also found to have an interaction with each other. In both cases, the percent

removal of Bright-Edge 80 was found to increase at a higher adsorbent dose. Other

interactions showed no prominent features as far as this discussion.

4.5.4 Normal Probability Plot of Residuals

One of the key assumptions for the statistical analysis of data from the

experiments is that the data came from a normal distribution (Antony 2003). Plotting a

normal probability plot of the residuals can verify the normality of the data. If the points

on the plot fall fairly close to a straight line, then the data are normally distributed

124

solution temperature and CO concentration and adsorbent dose and CO concentration

showed antagonistic effects (Figure 4.14). The concentration of CO and solution

temperature showed a slight interaction with each other. The percent removal of CO was

not affected by the rotational speed at a CO concentration of 350 mg/L while the percent

removal of CO increased at a CO concentration of 50 mg/L and a rotational speed of 100

rpm. The interaction effect between temperature and adsorbent dose showed that percent

removal of CO was higher at a solution temperature of 30°C and an adsorbent dose of 0.5

g. The percent removal of CO remained the same at a solution temperature of 5°C,

irrespective of the adsorbent dose. In the case of Bright-Edge 80, the solution temperature

and Bright-Edge 80 concentration, solution temperature and adsorbent dose, and solution

temperature and rotational speed had an interaction with each other (Figure 4.15). In all

cases, the percent removal of Bright-Edge 80 was found to be higher at 30 °C. The

adsorbent dose and Bright-Edge 80 concentration and adsorbent dose and rotational

speed were also found to have an interaction with each other. In both cases, the percent

removal of Bright-Edge 80 was found to increase at a higher adsorbent dose. Other

interactions showed no prominent features as far as this discussion.

4.5.4 Normal Probability Plot of Residuals

One of the key assumptions for the statistical analysis of data from the

experiments is that the data came from a normal distribution (Antony 2003). Plotting a

normal probability plot of the residuals can verify the normality of the data. If the points

on the plot fall fairly close to a straight line, then the data are normally distributed

124

(Antony 2003). The normal probability plot of the residuals for SMO, CO and Bright-

Edge 80 are shown in Figures 4.16 — 4.18, respectively. It can be observed that for all

three oils, the points are fairly close to the straight line. Therefore, the experimental data

is from a normally distributed population.

4.6 Effect of pH on Biosorption

Maximum oil removals for all three oils were obtained at pH 3.0. The results of

pH studies and zeta potentials of the three oils and the biomass are shown in Figure 4.19.

The percent removal of oil decreased with an increase in pH for all three oils. At pH 3, an

almost 98% removal of oil was achieved for all three oils and at pH 9, the oil removal

decreased to < 35% for SMO and Bright-Edge 80 and to 87% approximately for CO. The

initial pH of all three oil-in-water emulsions was in the range of 7.5-7.6. The percentage

of oil removal for SMO, CO and Bright-Edge 80 at this pH was approximately 66%, 87%

and 63 %, respectively. The isoelectric point of the biomass was found to be at pH 3.0.

The surface charge of the M rouxii biomass above the isoelectric pH was negative.

Surface charges of the three oils were, for the most part, negative above pH 3.0. As the

pH is lowered, the overall surface charge on the biomass will become positive or less

negative which will promote a stronger attraction towards negatively charged groups

present in the oil. As the pH of the system increases, the number of negatively charged

sites increases and the number of positively charged sites decreases. This causes a

hindrance to the sorption of the negatively charged functional groups in the oil resulting

125

(Antony 2003). The normal probability plot of the residuals for SMO, CO and Bright-

Edge 80 are shown in Figures 4.16 - 4.18, respectively. It can be observed that for all

three oils, the points are fairly close to the straight line. Therefore, the experimental data

is from a normally distributed population.

4.6 Effect of pH on Biosorption

Maximum oil removals for all three oils were obtained at pH 3.0. The results of

pH studies and zeta potentials of the three oils and the biomass are shown in Figure 4.19.

The percent removal of oil decreased with an increase in pH for all three oils. At pH 3, an

almost 98% removal of oil was achieved for all three oils and at pH 9, the oil removal

decreased to < 35% for SMO and Bright-Edge 80 and to 87% approximately for CO. The

initial pH of all three oil-in-water emulsions was in the range of 7.5-7.6. The percentage

of oil removal for SMO, CO and Bright-Edge 80 at this pH was approximately 66%, 87%

and 63 %, respectively. The isoelectric point of the biomass was found to be at pH 3.0.

The surface charge of the M. rouxii biomass above the isoelectric pH was negative.

Surface charges of the three oils were, for the most part, negative above pH 3.0. As the

pH is lowered, the overall surface charge on the biomass will become positive or less

negative which will promote a stronger attraction towards negatively charged groups

present in the oil. As the pH of the system increases, the number of negatively charged

sites increases and the number of positively charged sites decreases. This causes a

hindrance to the sorption of the negatively charged functional groups in the oil resulting

125

Normal Probability Plot (response is % Removal)

Figure 4.16: Normal probability plot of the residuals for removal of SMO

126


95-

90-

80-

70-

t 2 £

30-

20-

10-

-1.0 -0.5 0.0 0.5- 1.0 Residual

Figure 4.16: Normal probability plot of the residuals for removal of SMO

126

Figure 4.17: Normal probability plot of the residuals for removal of CO

127


95-

90-

80-

70-

60-

50-

40-

30-

20-

8 £

10-

-1.0 -0.5 0.5 0.0 1.0 Residual

Figure 4.17: Normal probability plot of the residuals for removal of CO

127

Figure 4.18: Normal probability plot of the residuals for removal of Bright-Edge 80

128


95-

90-

80-

70-

60-

so

lo-

-0.50 -0.75 -0.25 0.00 0.25 0.50 Residual

Figure 4.18: Normal probability plot of the residuals for removal of Bright-Edge 80

128

r 150

- % removal of SMO % removal of Bright-Edge 80

—0—Zeta potential of CO —a—zeta potential of Mucor rouxii

—2-- % removal of CO --e—Zeta potential of SMO - -a—Zeta potential of Bright-Edge 80

125 1

100

75

50

25

-25

-50

-75

-100

Zet

a po

tent

ial,

mV

Figure 4.19: Effect of pH on biosorption of oils and zeta potential of autoclaved M rouxii biomass and three oil-in-water emulsions

129 -

120 150

125

100 - 100

5 60

-25

-50

-75

-100 9 2 3 5 6 8 4 7

' % removal of SMO —•— % removal of CO —*— % removal of Bright-Edge 80 —•— Zeta potential of SMO —o—Zeta potential of CO —*—Zeta potential of Bright-Edge 80 —*— zeta ootentlal of Mucor rouxii

Figure 4.19: Effect of pH on biosorption of oils and zeta potential of autoclaved M. rouxii biomass and three oil-in-water emulsions

129

in a decrease in sorption of oil at higher pH levels. A negatively charged surface site on

the adsorbent does not favor adsorption of oil, due to the electrostatic repulsion. Also,

lower adsorption of oil at alkaline pH is due to the presence of excess hydroxyl ions

competing with the negatively charged functional groups in oil for the adsorption sites.

However, Bright-Edge 80 showed a percent removal of 100, indicating other types of

interactions, other than electrostatic interactions, could be involved. At pH 3.0, the

magnitude of zeta potential of Bright-Edge 80 emulsion was found to be low. The

magnitude of zeta potential refers to forces of repulsion between oil particles in the

emulsion that lead to the stability of emulsions; the lower the magnitude of zeta potential,

the lesser the stability of emulsion.

4.7 Effect of Concentration on Biosorption

Results showing the effect of the initial oil concentration between 50 mg/L and

350 mg/L at room temperature (22 ± 2 °C), and at a pH of 3.0 are given in Figure 4.20.

The percent removal rate of oil via M rouxii increased upon increasing the initial

concentration of up to 200 mg/L. The percent removal of oil dropped with an initial oil

concentration higher than 200 mg/L. The adsorption capacity of the biomass was found to

increase with an initial concentration of oil for the entire concentration range. This could

be due to the fact that the initial concentration provides an important driving force with

which to overcome all mass transfer resistances of the oil between the aqueous and solid

phases.

130

in a decrease in sorption of oil at higher pH levels. A negatively charged surface site on

the adsorbent does not favor adsorption of oil, due to the electrostatic repulsion. Also,

lower adsorption of oil at alkaline pH is due to the presence of excess hydroxyl ions

competing with the negatively charged functional groups in oil for the adsorption sites.

However, Bright-Edge 80 showed a percent removal of 100, indicating other types of

interactions, other than electrostatic interactions, could be involved. At pH 3.0, the

magnitude of zeta potential of Bright-Edge 80 emulsion was found to be low. The

magnitude of zeta potential refers to forces of repulsion between oil particles in the

emulsion that lead to the stability of emulsions; the lower the magnitude of zeta potential,

the lesser the stability of emulsion.

4.7 Effect of Concentration on Biosorption

Results showing the effect of the initial oil concentration between 50 mg/L and

350 mg/L at room temperature (22 ± 2 °C), and at a pH of 3.0 are given in Figure 4.20.

The percent removal rate of oil via M. rouxii increased upon increasing the initial

concentration of up to 200 mg/L. The percent removal of oil dropped with an initial oil

concentration higher than 200 mg/L. The adsorption capacity of the biomass was found to

increase with an initial concentration of oil for the entire concentration range. This could

be due to the fact that the initial concentration provides an important driving force with

which to overcome all mass transfer resistances of the oil between the aqueous and solid

phases.

130

100

_ 98

2 96

0 E 94

92

90

88

86

SMO

-0- CO

0 50 100 150 200 250 300

Concentration mg/L

350 400

Bright-Edge 80

Figure 4.20: Percentage removal of oil at various initial concentrations

131

Concentration mg/L

Figure 4.20: Percentage removal of oil at various initial concentrations

131

4.8 Batch Kinetic Studies

Batch kinetic studies showed that equilibrium was reached within 3 h, 2 h and 1 h

of contact for SMO, CO, and Bright-Edge 80, respectively using M rouxii biomass.

Residual oil concentration versus time profile for SMO, CO, and Bright-Edge 80 are

given in Figures 4.21 — 4.23, respectively. As for SMO, an increase in the temperature led

to an increase in the initial adsorption rate, but adsorption capacities at 6 hours for the

four temperatures studied, were close to each other. With respect to SMO, the time taken

to reach equilibrium shortened from 2.75 hours to 1 hour when the temperature was

increased from 5°C to 3 °C while it took approximately 0.75 hours at all four

temperatures for Bright-Edge 80. In the case of CO, equilibrium occurred at 0.75 h at 5°C

while it took 1.75 hours before the temperature was increased to 22°C. However, at 30°C,

equilibrium was reached at 0.5 h. The reason for the attainment of rapid equilibrium at

30°C could be due to the fact that stability of emulsions decreases when temperatures are

increased above room temperature; this decreases the external phase viscosity and the

relative solubilities of the emulsifying agents (Manning and Thompson 1995). Studies

regarding the rate of adsorption by the M rouxii biomass showed that sorption of the oils

to the M rouxii biomass consisted of two distinct phases: a primary phase characterized

by rapid sorption, and a secondary phase characterized by slow sorption (Moazed and

Viraraghavan 2005). The rapid phase accounted for a major portion of the total oil

sorption, while the secondary phase contributed to a relatively small portion. The rapid

phase lasted approximately 15 to 30 minutes in all cases. Residual oil concentrations of

132

4.8 Batch Kinetic Studies

Batch kinetic studies showed that equilibrium was reached within 3 h, 2 h and 1 h

of contact for SMO, CO, and Bright-Edge 80, respectively using M. rouxii biomass.

Residual oil concentration versus time profile for SMO, CO, and Bright-Edge 80 are

given in Figures 4.21 - 4.23, respectively. As for SMO, an increase in the temperature led

to an increase in the initial adsorption rate, but adsorption capacities at 6 hours for the

four temperatures studied, were close to each other. With respect to SMO, the time taken

to reach equilibrium shortened from 2.75 hours to 1 hour when the temperature was

increased from 5°C to 3 °C while it took approximately 0.75 hours at all four

temperatures for Bright-Edge 80. In the case of CO, equilibrium occurred at 0.75 h at 5°C

while it took 1.75 hours before the temperature was increased to 22°C. However, at 30°C,

equilibrium was reached at 0.5 h. The reason for the attainment of rapid equilibrium at

30°C could be due to the fact that stability of emulsions decreases when temperatures are

increased above room temperature; this decreases the external phase viscosity and the

relative solubilities of the emulsifying agents (Manning and Thompson 1995). Studies

regarding the rate of adsorption by the M. rouxii biomass showed that sorption of the oils

to the M. rouxii biomass consisted of two distinct phases: a primary phase characterized

by rapid sorption, and a secondary phase characterized by slow sorption (Moazed and

Viraraghavan 2005). The rapid phase accounted for a major portion of the total oil

sorption, while the secondary phase contributed to a relatively small portion. The rapid

phase lasted approximately 15 to 30 minutes in all cases. Residual oil concentrations of

132

250

E 200

0 0

150 05C

rzt 015 C

A22 C 100

X30 C

50

0 0.5 1 1.5 2 2.5

Time, h

4-3 3.5

Figure 4.21: SMO concentration versus time for different temperatures

133

4

250

£ 200 B

0 CM

§ 1 5 0

1 g gioo o o

0 5 C

• 15 C

A22C

X 3 0 C

^ A & £ —j^ |—n—^—

1.5 2 2.5 3

Time, h

# 3.5

Figure 4.21: SMO concentration versus time for different temperatures

133

250

1). 200

0 0

•

❑ 150 0

0

`4 100 0 U

so

05 C

015 C

A22 C

X30 C

0 - -1-144-4-1141-411-11-11-111- --EU---11 0 0.5 1 1.5 2 2.5 3 3.5 4

Time, h

Figure 4.22: CO concentration versus time for different temperatures

134

250

200 s o c 150

<a H c u u c o o C3 ri

e$

100

50

05 c • 15 C

A22C

X 3 0 C

0.5 1.5 2 2.5

Time, h

3.5

Figure 4.22: CO concentration versus time for different temperatures

134

Res

idu

al c

once

ntr

atio

n o

f o

il, i

ng/L

250

200

150 <> 5C

015C 100 -

50 -

22C

x 30C

a I 4 0 4 0 0 0 0.5 1 1.5 2 2.5 ' 3 3.5 4

Time, h

Figure 4.23: Bright-Edge 80 concentration versus time for different temperatures

135

• 15C

A 22C

X30C

1 1.5 2

Time.h

Figure 4.23: Bright-Edge 80 concentration versus time for different temperatures

135

approximately 52 mg/L, 18 mg/L and 12 mg/L were observed at 22°C after the rapid

phase for SMO, CO and Bright-Edge 80, respectively. This was because active

adsorption sites of the adsorbent could become involved in adsorption as soon as the

adsorbent was introduced into the system. The slow secondary phase of oil uptake may

be due to the reduced availability of the active sites. During the mixing period, breakage

of the oil droplets is enhanced, encouraging adsorption of oil by the adsorbent (Ahmad et

al. 2005a). Oil adsorption capacities of the biomass after 6 hours of contact were 97.7

mg/g, 98 mg/g and 98 mg/g for SMO, CO and Bright-Edge 80, respectively at 22°C. A

comparison of oil removal efficiencies of M rouxii biomass and other sorbents is given

in Table 4.12. Batch studies using similar oil-in-water emulsions were conducted by

Moazed (2000), Moazed and Viraraghavan (2005) and Mysore et al. (2005). M rouxii

biomass showed a higher oil adsorption capacity than vermiculite, powdered organo clay

or organo clay/ anthracite. The Adsorption capacity of SMO by M rouxii biomass and

bentonite were comparable. Kinetic data were fitted to pseudo first-order, pseudo second-

order and infra-particle diffusion models. The rate constants and the correlation

coefficients (r) for all three models at four temperatures are provided in Table 4.13. The

correlation coefficient (r) is the Pearson product-moment correlation coefficient, which is

a dimensionless measure of the linear association between two variables. Based on the

correlation coefficient values (r) and test of statistical significance at a 95% confidence

level, it was found that the Ho pseudo second-order model was able to describe the

adsorption kinetics of SMO, CO and Bright-Edge 80 onto the M rouxii biomass.

136

approximately 52 mg/L, 18 mg/L and 12 mg/L were observed at 22°C after the rapid

phase for SMO, CO and Bright-Edge 80, respectively. This was because active

adsorption sites of the adsorbent could become involved in adsorption as soon as the

adsorbent was introduced into the system. The slow secondary phase of oil uptake may

t be due to the reduced availability of the active sites. During the mixing period, breakage

of the oil droplets is enhanced, encouraging adsorption of oil by the adsorbent (Ahmad et

al. 2005a). Oil adsorption capacities of the biomass after 6 hours of contact were 97.7

mg/g, 98 mg/g and 98 mg/g for SMO, CO and Bright-Edge 80, respectively at 22°C. A

comparison of oil removal efficiencies of M. rouxii biomass and other sorbents is given

in Table 4.12. Batch studies using similar oil-in-water emulsions were conducted by

Moazed (2000), Moazed and Viraraghavan (2005) and Mysore et al. (2005). M. rouxii

biomass showed a higher oil adsorption capacity than vermiculite, powdered organo clay

or organo clay/ anthracite. The Adsorption capacity of SMO by M. rouxii biomass and

bentonite were comparable. Kinetic data were fitted to pseudo first-order, pseudo second-

order and intra-particle diffusion models. The rate constants and the correlation

coefficients (r) for all three models at four temperatures are provided in Table 4.13. The

correlation coefficient (r) is the Pearson product-moment correlation coefficient, which is

a dimensionless measure of the linear association between two variables. Based on the

correlation coefficient values (r) and test of statistical significance at a 95% confidence

level, it was found that the Ho pseudo second-order model was able to describe the

adsorption kinetics of SMO, CO and Bright-Edge 80 onto the M, rouxii biomass.

136

Table 4.12: Comparison of oil removal efficiencies of M rouxii biomass and other sorbents obtained in batch studies

Sorbents Oil Adsorption capacity (mg/g)

Initial concentration (1110-)

Final concentration (mgt)

% removal

M rouxii biomass SMO 97.5 200 5 97.5 Bentonite SMO 98.2 502 11 97.8 Powdered organo clay SMO 44.8 235.5 11.2 95.3 Organo clay/anthracite SMO 2.3 235.2 120 49.0 Bentonite organoclay SMO 29.0 236 10 94.0 Expanded vermiculite SMO 11.5 218 45 79.0 Hydrophobized vermiculite

SMO 10.2 218 65 56.0

M rouxii biomass CO 98.0 200 4 98.0 Expanded vermiculite CO 7.3 205 95 53.6 Hydrophobized vermiculite

CO 7.9 205 86 58.0

M rouxii biomass Bright- 98.0 200 4 98.0 Edge 80

Bentonite KUT45 58.5 305 12.4 95.9 Powdered organo clay KUT45 75.4 381 3.8 99.0 Organo clay/anthracite KUT45 5.2 330 70.8 78.6 Bentonite organoclay KUT45 358 381 10 97.0 Expanded vermiculite KUT45 10.1 170 18 89.4 Hydrophobized vermiculite

KUT45 5.3 170 90 49.0

Bentonite RE 0.9 5.2 0.7 86.5 Powdered organo clay RE 2.2 25.7 14.5 43.6 Organo clay/anthracite RE 0.1 5.2 1.0 80.8 Bentonite organoclay RE 0.7 26 10 61.0 Expanded vermiculite RE 0.4 11.5 4.9 57.0 Hydrophobized vermiculite

RE 0.3 11.5 6.5 43.0

Note: Values for bentonite, powdered organo clay, organo clay/anthracite and bentonite organoclay were obtained from Moazed (2000) and Moazed and Viraraghavan (2005). Values for expanded and hydrophobized vermiculite were obtained from Mysore et al. (2005). KUT45 is Kutwell oil; RE is refinery effluent. All studies were conducted at 22 ± 2 °C

137

Table 4.12: Comparison of oil removal efficiencies of M. rouxii biomass and other sorbents obtained in batch studies

Sorbents Oil Adsorption Initial Final % capacity conc conc removal (mg/g) entration

(mg/L) entration (mg/L)

M. rouxii biomass SMO 97.5 200 5 97.5 Bentonite SMO 98.2 502 11 97.8 Powdered organo clay SMO 44.8 235.5 11.2 95.3 Organo clay/anthracite SMO 2.3 235.2 120 49.0 Bentonite organoclay SMO 29.0 236 10 94.0 Expanded vermiculite SMO 11.5 218 45 79.0 Hydrophobized SMO 10.2 218 65 56.0 vermiculite M. rouxii biomass CO 98.0 200 4 98.0 Expanded vermiculite CO 7.3 205 95 53.6 Hydrophobized CO 7.9 205 86 58.0 vermiculite M. rouxii biomass Bright-

Edge 80

98.0 200 4 98.0

Bentonite KUT45 58.5 305 12.4 95.9 Powdered organo clay KUT45 75.4 381 3.8 99.0 Organo clay/anthracite KUT45 5.2 330 70.8 78.6 Bentonite organoclay KUT45 358 381 10 97.0 Expanded vermiculite KUT45 10.1 170 18 89.4 Hydrophobized KUT45 5.3 170 90 49.0 vermiculite Bentonite RE 0.9 5.2 0.7 86.5 Powdered organo clay RE 2.2 25.7 14.5 43.6 Organo clay/anthracite RE 0.1 5.2 1.0 80.8 Bentonite organoclay RE 0.7 26 10 61.0 Expanded vermiculite RE 0.4 11.5 4.9 57.0 Hydrophobized RE 0.3 11.5 6.5 43.0 vermiculite Note: Values for bentonite, powdered organo clay, organo clay/anthracite and bentonite organoclay were obtained from Moazed (2000) and Moazed and Viraraghavan (2005). Values for expanded and hydrophobized vermiculite were obtained from Mysore et al. (2005). KUT45 is Kutwell oil; RE is refinery effluent. All studies were conducted at 22 ± 2 °C

137


Oil-in- water

emulsions

Lagergren model Ho model Intra-particle diffusion model

k1 q, r k2 q, r ki C r (1/h) (mg/g) (g/m (mg/g) (na/g (mg/

g-h) h") g) SMO 5 °C 5.85 97.21 0.93 0.13 100.75 0.98 13.73 77.4 0.90 15 °C 5.32 96.37 0.84 0.11 100.75 0.98 20.03 68.9 0.97 22 °C 3.75 94.91 0.85 0.06 101.21 0.96 21.54 61.9 0.91 30 °C 5.71 98.48 0.95 0.14 101.73 0.96 29.97 65.7 0.94 CO 5 °C 16.87 98.22 0.85 2.44 98.46 0.95 2.11 95.9 0.90 15 °C 15.54 97.97 0.89 1.82 98.28 0.95 2.83 94.9 0.90 22 °C 9.79 97.09 0.79 0.39 98.36 0.95 7.69 86.8 0.98 30 °C 15.10 97.94 0.90 1.62 98.29 0.96 8.64 91.0 0.90 Bright-Edge 80 5 °C 15.17 97.94 0.97 1.79 98.24 0.95 8.42 94.5 0.90 15 °C 14.01 97.86 0.89 1.21 98.33 0.97 4.99 92.8 0.96 22 °C 12.99 97.59 0.74 0.79 98.33 0.96 6.19 90.8 0.97 30 °C 17.51 97.92 0.88 2.84 98.12 0.98 5.43 91.4 0.96 Note: r - correlation coefficient; all model parameters are statistically significant (t-test) at 95% confidence level.

138


Oil-in-water

Lagergren model Ho model Intra-particle diffusion model

emulsions ki

(1/h) <le

(mg/g) r k2

(g/m g-h)

<h (mg/g)

r C (mg/ 8)

r

SMO 5 °C 5.85 97.21 0.93 0.13 100.75 0.98 13.73 77.4 0.90 15 °C 5.32 96.37 0.84 0.11 100.75 0.98 20.03 68.9 0.97 22 °C 3.75 94.91 0.85 0.06 101.21 0.96 21.54 61.9 0.91 30 °C 5.71 98.48 0.95 0.14 101.73 0.96 29.97 65.7 0.94 CO 5 °C 16.87 98.22 0.85 2.44 98.46 0.95 2.11 95.9 0.90 15 °C 15.54 97.97 0.89 1.82 98.28 0.95 2.83 94.9 0.90 22 °C 9.79 97.09 0.79 0.39 98.36 0.95 7.69 86.8 0.98 30 °C 15.10 97.94 0.90 1.62 98.29 0.96 8.64 91.0 0.90 Bright-Edge 80 5 °C 15.17 97.94 0.97 1.79 98.24 0.95 8.42 94.5 0.90 15 °C 14.01 97.86 0.89 1.21 98.33 0.97 4.99 92.8 0.96 22 °C 12.99 97.59 0.74 0.79 98.33 0.96 6.19 90.8 0.97 30 °C 17.51 97.92 0.88 2.84 98.12 0.98 5.43 91.4 0.96 Note: r - correlation coefficient; all model parameters are statistically significant (/-test) at 95% confidence level.

138

The predicted fit for the pseudo first-order and pseudo second-order models for the

observed data for at 22°C SMO, CO and Bright-Edge 80 are shown in Figures 4.24 —

4.26, respectively. Plots at other temperatures are shown in Figures A.1 to A.9. It was

found that the correlation coefficient (r) for the first-order reaction was lower than 0.9 in

most cases. It was likely that adsorption of oil onto the M rouxii biomass was not a first-

order reaction. The Ho pseudo-second order model predicted higher values for qe than

the Lagergren model, and the qe values obtained from the Ho model were closer to the

observed values. The 'ye values obtained from both models for SMO, CO and Bright-Edge

80 were all above 95.0 mg/g, which showed the ability of the M rouxii biomass to adsorb

such oils. The basic assumption of the pseudo-second order model is that adsorption

reaction on the surface of the adsorbent is the rate-controlling step. In general, the

mechanism of adsorption may be assumed to involve the following four steps: (i) bulk

diffusion; (ii) film diffusion; (iii) intra-particle diffusion or pore diffusion; and (iv)

chemical reaction (Mathews and Weber, 1976). Since adsorption of oil by M rouxii

biomass is a rapid process, the rate of adsorption is possibly governed either by an

external mass transfer rate or an intra-particle diffusion rate. Therefore, the intra-particle

diffusion model was analyzed. A plot of q, versus t 112 should be a straight line if intra-

particle diffusion is involved in adsorption process and if this line passes through the

origin, the intra-particle diffusion is the rate-controlling step. As for all three oils, the plot

was identified as an initial curve, followed by a flat portion (the linear portion of the plots

are given in Figures A.10 to A.21). Previous work showed that the plot of sql versus t1/2

might present as multi-linearity having two or more stages (Ahmad et al. 2005a).

139

The predicted fit for the pseudo first-order and pseudo second-order models for the

observed data for at 22°C SMO, CO and Bright-Edge 80 are shown in Figures 4.24 -

4.26, respectively. Plots at other temperatures are shown in Figures A.l to A.9. It was

found that the correlation coefficient (r) for the first-order reaction was lower than 0.9 in

most cases. It was likely that adsorption of oil onto the M. rouxii biomass was not a first-

order reaction. The Ho pseudo-second order model predicted higher values for qe than

the Lagergren model, and the qe values obtained from the Ho model were closer to the

observed values. The qe values obtained from both models for SMO, CO and Bright-Edge

80 were all above 95.0 mg/g, which showed the ability of the M. rouxii biomass to adsorb

such oils. The basic assumption of the pseudo-second order model is that adsorption

reaction on the surface of the adsorbent is the rate-controlling step. In general, the

mechanism of adsorption may be assumed to involve the following four steps: (i) bulk

diffusion; (ii) film diffusion; (iii) intra-particle diffusion or pore diffusion; and (iv)

chemical reaction (Mathews and Weber, 1976). Since adsorption of oil by M. rouxii

biomass is a rapid process, the rate of adsorption is possibly governed either by an

external mass transfer rate or an intra-particle diffusion rate. Therefore, the intra-particle

diffusion model was analyzed. A plot of qt versus tm should be a straight line if intra-

particle diffusion is involved in adsorption process and if this line passes through the

origin, the intra-particle diffusion is the rate-controlling step. As for all three oils, the plot

was identified as an initial curve, followed by a flat portion (the linear portion of the plots

are given in Figures A. 10 to A.21). Previous work showed that the plot of q, versus tin

might present as multi-linearity having two or more stages (Ahmad et al. 2005a).

139

100 -

95 -

90

ao @) 85 E

80

Z3 75 0

70 0

65

60

55

50

0 0 0 ' . --- - - - 0

0 Adsorption capacity - observed

Adsorption capacity predicted -Lagergren (r = 0.85)

Adsorption capacity predicted - Ho (r = 0.96)

3 Time, h 4 5 6

Figure 4.24: Rate of SMO biosorption predicted by Lagergren and Ho models at 22°C

140

100

95

90

80

75 O Adsorption capacity - observed

70

Adsorption capacity predicted -Lagergren (r = 0.85) 65

60 Adsorption capacity predicted - Ho (r = 0.96)

55

2 3 Time, h 4 5 6 7

Figure 4.24: Rate of SMO biosorption predicted by Lagergren and Ho models at 22°C

140

100 -

98

to

E96-

Cd

94 o 0 Adsorption capacity observed

,,§92 Adsorption capacity predicted -

Lagergren (r = 0.79)

90 Adsorption capacity predicted - Ho (r = 0.95)

AQQQQQ .. - -Q- -Q __

88 0 2 3 4

Time, h 5 6 7

Figure 4.25: Rate of CO biosorption predicted by Lagergren and Ho models at 22°C

141

° Adsorption capacity observed


Adsorption capacity predicted - Ho (r = 0.95)

88 3 4

Time, h

Figure 4.25: Rate of CO biosorption predicted by Lagergren and Ho models at 22°C

141

100 •i

99 1

98 Q _Q- 0- - -0- ----- --------

96

95

94 -

93 4 -r

0 1

0 Adsorption capacity - Observed

7


Adsorption capacity predicted -Ho (r = 0.96)

2 3 4 5

Time, h

Figure 4.26: Rate of Bright-Edge 80 biosorption predicted by Lagergren and Ho models at 22°C

142

100

99 -

Ztr „ 8 0 .0 -Q-0 .—0-—Q-—0—Q-

O Adsorption capacity - Observed

• Adsorption capacity predicted -Lagergren (r = 0.74)

Adsorption capacity predicted -Ho (r = 0.96)

B 4

Time, h

Figure 4.26: Rate of Bright-Edge 80 biosorption predicted by Lagergren and Ho models at 22°C

142

The first stage refers to the instantaneous adsorption stage, the second stage is the

gradual adsorption where the intra-particle diffusion is controlled and the third stage is

the final equilibrium stage where the intra-particle starts to slow down due to extremely

low residue oil in the solution (Ahmad et al. 2005a). In the case of the three oils studied,

during the initial curve portion, adsorption process followed boundary layer diffusion and

the later flat portion was an indication of the intra-particle diffusion. This showed the

presence of boundary layer diffusion along with intra-particle diffusion, which controls

oil adsorption by M rouxii biomass. According to the majority of systems reported in the

literature, there is some evidence of boundary layer resistance in the initial stages of

adsorption process. A good correlation of rate data (r > 0.9) for all three oils justifies the

intra-particle diffusion mechanism. The values of the intercept, C, from Table 4.13, gives

an idea about the boundary layer thickness: the larger the intercept, the greater the

boundary layer effect. According to Table 4.13, it can be observed that the thickness of

the boundary layer was higher for CO and Bright-Edge 80 than for SMO.

4.9 Batch Isotherm Studies

Isotherm studies yield important data regarding the equilibrium distribution of a

solute between the adsorbent and liquid phases at a constant temperature. It is critical in

optimizing the use of adsorbents. With respect to all three oils, an increase in biomass

dose from 0.03 g to 0.4 g resulted in a decrease in its adsorption capacity. The number of

adsorption sites increased with an increase in the adsorbent dosage. Oil removal

increased from 82% to 95% for Bright-Edge 80 with an increase in biomass dose from

143

The first stage refers to the instantaneous adsorption stage, the second stage is the

gradual adsorption where the intra-particle diffusion is controlled and the third stage is

the final equilibrium stage where the intra-particle starts to slow down due to extremely

low residue oil in the solution (Ahmad et al, 2005a). In the case of the three oils studied,

during the initial curve portion, adsorption process followed boundary layer diffusion and

the later flat portion was an indication of the intra-particle diffusion. This showed the

presence of boundary layer diffusion along with intra-particle diffusion, which controls

oil adsorption by M. rouxii biomass. According to the majority of systems reported in the

literature, there is some evidence of boundary layer resistance in the initial stages of

adsorption process. A good correlation of rate data (r > 0.9) for all three oils justifies the

intra-particle diffusion mechanism. The values of the intercept, C, from Table 4.13, gives

an idea about the boundary layer thickness: the larger the intercept, the greater the

boundary layer effect. According to Table 4.13, it can be observed that the thickness of

the boundary layer was higher for CO and Bright-Edge 80 than for SMO.

4.9 Batch Isotherm Studies

Isotherm studies yield important data regarding the equilibrium distribution of a

solute between the adsorbent and liquid phases at a constant temperature. It is critical in

optimizing the use of adsorbents. With respect to all three oils, an increase in biomass

dose from 0.03 g to 0.4 g resulted in a decrease in its adsorption capacity. The number of

adsorption sites increased with an increase in the adsorbent dosage. Oil removal

increased from 82% to 95% for Bright-Edge 80 with an increase in biomass dose from

143

0.03 g to 0.2 g. This result was expected because as the biomass dose increased, the

number of adsorbent particles surrounding the oil droplets increased and therefore, the

particles sorbed more oil to their surfaces. An increase in biomass dose over 0.2 g did not

result in any increase in percent removal of oil in all cases. This may be due to the

binding of almost all of the oil to the adsorbent. The decrease in adsorption capacity with

the increase in the adsorbent dosage was mainly attributed to the fact that some of the

adsorption sites remained unsaturated during adsorption process (Achak et al. 2009). The

Langmuir and Freundlich isotherm models were evaluated to examine biosorption. The

constants of the adsorption parameters and the correlation coefficient (r) at all

temperatures for all three oils are summarized in Table 4.14. Isotherm plots are shown in

Figures A.22 to A.30. Model parameters regarding the biosorption of oil on the M rouxii

biomass were statistically significant at a 95% confidence level for both the Langmuir

and Freundlich isotherm models. Based on the correlation coefficient values (r), the

Freundlich isotherm model was able to fit the experimental data better. The Langmuir

isotherm assumes homogenous surface energy distribution (Langmuir 1916). The value

of the separation factor RL, calculated from the Langmuir isotherm constant, was between

0.2 and 0.8 for all three oils, indicating that adsorption was favorable (Table 4.15). The

Freundlich isotherm assumes a heterogeneous surface with a non-uniform distribution of

heat of adsorption over the surface (Freundlich, 1906). The high value of Freundlich

constant KF for all three oils showed the oil adsorption capacity of the biomass was very

high. The values of n were greater than 1, reflecting favorable adsorption.

144

0.03 g to 0.2 g. This result was expected because as the biomass dose increased, the

number of adsorbent particles surrounding the oil droplets increased and therefore, the

particles sorbed more oil to their surfaces. An increase in biomass dose over 0.2 g did not

result in any increase in percent removal of oil in all cases. This may be due to the

binding of almost all of the oil to the adsorbent. The decrease in adsorption capacity with

the increase in the adsorbent dosage was mainly attributed to the fact that some of the

adsorption sites remained unsaturated during adsorption process (Achak et al. 2009). The

Langmuir and Freundlich isotherm models were evaluated to examine biosorption. The

constants of the adsorption parameters and the correlation coefficient (r) at all

temperatures for all three oils are summarized in Table 4.14. Isotherm plots are shown in

Figures A.22 to A.30. Model parameters regarding the biosorption of oil on the M. rouxii

biomass were statistically significant at a 95% confidence level for both the Langmuir

and Freundlich isotherm models. Based on the correlation coefficient values (r), the

Freundlich isotherm model was able to fit the experimental data better. The Langmuir

isotherm assumes homogenous surface energy distribution (Langmuir 1916). The value

of the separation factor RL, calculated from the Langmuir isotherm constant, was between

0.2 and 0.8 for all three oils, indicating that adsorption was favorable (Table 4.15). The

Freundlich isotherm assumes a heterogeneous surface with a non-uniform distribution of

heat of adsorption over the surface (Freundlich, 1906). The high value of Freundlich

constant KF for all three oils showed the oil adsorption capacity of the biomass was very

high. The values of n were greater than 1, reflecting favorable adsorption.

144


Oil-in- water

emulsions

Langmuir Freundlich

b (L/mg) Q° (mg/g) r KF (L/mg) N r

SMO 5 °C 0.014 2564.82 0.95 41.11 1.24 0.98

15 °C 0.008 4174.15 0.95 33.36 1.05 0.97

22 °C 0.02 1660.08 0.96 40.16 1.12 0.97

30 °C 0.004 8901.18 0.95 30.12 1.00 0.97

CO 5 °C 0.018 2107.43 0.95 42.84 1.17 0.98

15 °C 0.005 5827.36 0.96 30.65 1.03 0.97

22 °C 0.015 2727.53 0.95 30.6 1.01 0.97

30 °C 0.019 1793.44 0.95 41.29 1.21 0.97

Bright-Edge 80 5 °C 0.005 6533.35 0.95 25.95 1.01 0.98

15 °C 0.004 6175.04 0.96 26.85 1.02 0.98

22 °C 0.011 2019.25 0.96 26.56 1.16 0.97

30 °C 0.0011 33982.13 0.95 19.4 1.01 0.98

Note: r - correlation coefficient; model parameters are statistically significant (t-test) at 95% confidence level

145


Oil-in-water

emulsions

Langmuir Freundlich

b (L/mg) e"(mg/g) r KF (L/mg) N r

SMO 5 °C 0.014 2564.82 0.95 41.11 1.24 0.98

15 °C 0.008 4174.15 0.95 33.36 1.05 0.97

22 °C 0.02 1660.08 0.96 40.16 1.12 0.97

30 °C 0.004 8901.18 0.95 30.12 1.00 0.97

CO 5 °C 0.018 2107.43 0.95 42.84 1.17 0.98

15 °C 0.005 5827.36 0.96 30.65 1.03 0.97

22 °C 0.015 2727.53 0.95 30.6 1.01 0.97

30 °C 0.019 1793.44 0.95 41.29 1.21 0.97

Bright-Edge 80 5 °C 0.005 6533.35 0.95 25.95 1.01 0.98

15 °C 0.004 6175.04 0.96 26.85 1.02 0.98

22 °C 0.011 2019.25 0.96 26.56 1.16 0.97

30 °C 0.0011 33982.13 0.95 19.4 1.01 0.98

Note: r - correlation coefficient; model parameters are statistically significant (t-test) at 95% confidence level

145

Table 4.15: Separation factor, RL based on the Langmuir equation

Temperature (`) C) RL values

SMO CO Bright-Edge 80

5 0.26 0.22 0.50

15 0.38 0.50 0.55

22 0.20 0.25 0.31

30 0.55 0.21 0.82

146

Table 4.15: Separation factor, Ri based on the Langmuir equation

Temperature (° C) Ri values

SMO CO Bright-Edge 80

5 0.26 0.22 0.50

15 0.38 0.50 0.55

22 0.20 0.25 0.31

30 0.55 0.21 0.82

146

4.10 Thermodynamics and Activation Parameters

The values of 4H° and AS° were determined from the slope and intercept at the

plot of In b versus 1/T. Thermodynamic and activation parameters obtained in the study

are shown in Table 4.16. The values of AG° were found to be negative for the three oils at

all four temperatures, indicating that adsorption was a spontaneous process. The AG°

values obtained for the three oils in the study were between -9.0 to —17.0 kJ/mol

approximately. The AG° values up to -20 kJ/mol involve physisorption while AG° values

greater than -40 kJ/mol were consistent with chemisorption (Pokhrel and Viraraghavan,

2008). The AH° value associated with chemisorption may be 100 to 400 kJ/mol while that

of physisorption may be as high as 20 kJ/mol (Weber and DiGiano, 1996). The

magnitude of the AH° values for the three oils, given in Table 4.16, showed the process

was not chemisorption but physisorption. The positive values of AH° for SMO, CO and

Bright-Edge 80 showed the endothermic nature of the process. The values of AS° were

positive for SMO and Bright-Edge 80, reflecting the affinity of oil towards the biomass.

The negative value of AS° for CO indicates the randomness of the decreases at the solid—

solution interface during adsorption of CO on the biomass. The rates of most chemical

reactions increase as the temperature is increased. In the rate law, temperature

dependence appears in the rate constant. The activation energy was determined from the

slope of linear plot of In k versus 1/T. The activation energies obtained were 27, 11.3 and

33.2 kJ/mol for adsorption of SMO, CO and Bright-Edge 80, respectively. Activation

energies lower than 42 kJ/mol indicate a diffusion-controlled process while activation

energies higher than 42 kJ/mol indicate a chemically controlled process (Sparks, 1989).

147

4.10 Thermodynamics and Activation Parameters

The values of AH0 and AS? were determined from the slope and intercept at the

plot of In b versus 1/T. Thermodynamic and activation parameters obtained in the study

are shown in Table 4.16. The values of AG0 were found to be negative for the three oils at

all four temperatures, indicating that adsorption was a spontaneous process. The AG0

values obtained for the three oils in the study were between -9.0 to -17.0 kJ/mol

approximately. The AG0 values up to -20 kJ/mol involve physisorption while AG0 values

greater than -40 kJ/mol were consistent with chemisorption (Pokhrel and Viraraghavan,

2008). The AH0 value associated with chemisorption may be 100 to 400 kJ/mol while that

of physisorption may be as high as 20 kJ/mol (Weber and DiGiano, 1996). The

magnitude of the AH0 values for the three oils, given in Table 4.16, showed the process

was not chemisorption but physisorption. The positive values of AH0 for SMO, CO and

Bright-Edge 80 showed the endothermic nature of the process. The values of AS0 were

positive for SMO and Bright-Edge 80, reflecting the affinity of oil towards the biomass.

The negative value of AS0 for CO indicates the randomness of the decreases at the solid-

solution interface during adsorption of CO on the biomass. The rates of most chemical

reactions increase as the temperature is increased. In the rate law, temperature

dependence appears in the rate constant. The activation energy was determined from the

slope of linear plot of In k versus 1/T. The activation energies obtained were 27, 11.3 and

33.2 kJ/mol for adsorption of SMO, CO and Bright-Edge 80, respectively. Activation

energies lower than 42 kJ/mol indicate a diffusion-controlled process while activation

energies higher than 42 kJ/mol indicate a chemically controlled process (Sparks, 1989).

147


Temperature (`) C) AG° (kJ/mo1) AH° (kJ/mol) AS° (J/mol) Ead

(kJ/mol) SMO 5 15 22 30

-9.86 -11.56 -9.59 -13.90

25.23 125.29 27.0

CO 5 15 22 30

-9.28 -12.68 -10.30 -9.98

11.89 -78.91 11.3

Bright-Edge 80 5 15 22 30

-12.24 -13.22 -11.06 -17.10

30.17 149.75 33.2

148


Temperature (° C) AG0 (kJ/mol) AHU (kJ/mol) AS0 (J/mol) Ead (kJ/mol)

SMO 5 -9.86 15 22

-11.56 -9.59

25.23 125.29 27.0

30 -13.90 CO 5 -9.28 15 22

-12.68 -10.30

11.89 -78.91 11.3

30 -9.98 Bright-Edge 80 5 -12.24 15 22

-13.22 -11.06

30.17 149.75 33.2

30 -17.10

148

Hence, it can be concluded that oil adsorption by M rouxii biomass could be controlled

by diffusion.

4.11 Batch Desorption Studies

The results of desorption tests using DI water are presented in Figure 4.27.

Similar results were observed when using DI water as eluent to desorb oil from oil-

adsorbed vermiculite (Mysore et al., 2005). Oil recovered by desorbing M rouxii biomass

that had adsorbed CO was 18% while for SMO and Bright-Edge 80 it was approximately

14 and 12%, respectively. The percent of desorption was found to increase with the mass

of oil-adsorbed biomass. The use of warm water at 40°C or higher may be an effective

method to regenerate an oil adsorbent (Moazed 2000).

4.12 Surface Functional Groups on M. rouxii Biomass

Functional groups present on the cell surface can be identified by Fourier

Transform Infrared (FTIR) spectroscopy as each group has a unique energy absorption

band. The FTIR spectra obtained for the autoclaved biomass is shown in Figure 4.28. The

infrared absorption wavelengths of each peak and the corresponding functional groups of

the autoclaved and oil adsorbed biomass are presented in Table 4.17. The spectra of raw

biomass showed a trough (marked A) observed at 3417 cm-I resulting from the NH2

asymmetric stretch mode of amines. The troughs resulting from bonded OH groups were

also observed in the range of 3300 - 3450 crril .

149

Hence, it can be concluded that oil adsorption by M. rouxii biomass could be controlled

by diffusion.

4.11 Batch Desorption Studies

The results of desorption tests using DI water are presented in Figure 4.27.

Similar results were observed when using DI water as eluent to desorb oil from oil-

adsorbed vermiculite (Mysore et al., 2005). Oil recovered by desorbing M. rouxii biomass

that had adsorbed CO was 18% while for SMO and Bright-Edge 80 it was approximately

14 and 12%, respectively. The percent of desorption was found to increase with the mass

of oil-adsorbed biomass. The use of warm water at 40°C or higher may be an effective

method to regenerate an oil adsorbent (Moazed 2000).

4.12 Surface Functional Groups on M. rouxii Biomass

Functional groups present on the cell surface can be identified by Fourier

Transform Infrared (FTIR) spectroscopy as each group has a unique energy absorption

band. The FTIR spectra obtained for the autoclaved biomass is shown in Figure 4.28. The

infrared absorption wavelengths of each peak and the corresponding functional groups of

the autoclaved and oil adsorbed biomass are presented in Table 4.17. The spectra of raw

biomass showed a trough (marked A) observed at 3417 cm"1 resulting from the NH2

asymmetric stretch mode of amines. The troughs resulting from bonded OH groups were

also observed in the range of 3300 - 3450 cm"1.

149

% re

cove

ry o

f oil

20

18

16

14

12

1

6

4

--11— CO

Bright-Edge 80

,

0 0.05 0.1 0.15 0.2 0.25 0.3 0.35 0.4 0.45 Mass of biomass, g

Figure 4.27: Desorption plot for M rouxii biomass using water as an eluent

150

20 -

18

16

14

"3 12 o

« 10 o o u i-

0s

8

6

4

2

SMO

CO

Bright-Edge 80

0.05 0.1 0.15 0.2 0.25 0.3 0.35 0.4 0.45 Mass of biomass, g

Figure 4.27: Desorption plot for M. rouxii biomass using water as an eluent

150

100

90 •

40 •

Before adsorption SMO adsortvd

— Bright-Edge SO adiatied - - a&orbed

...ore. • •

• • . ••e".. •

I, ' k ••

't • ...... "' 1 ••••e• , /

' 1

•

A

a

F

BC

••

30 4000 3500 3000 2500 2000

Wave number, cm'

D

1500 1000 500

Figure 4.28: FTIR spectra of M rouxii biomass before and after oil adsorption

151

0

too ~ Before adsorption

~ Bnght-Fdgc SO adsorkd 90

80

.r*

70

50

40

4000 3500 3000 2500 2000 1500 500 1000 0

Wave number, cm1

Figure 4.28: FTIR spectra of M. rouxii biomass before and after oil adsorption

151

Table 4.17: Functional groups of autoclaved biomass of M rouxii biomass, oil adsorbed biomass and the corresponding infrared absorption wavelengths (Lin-Vien 1991)

Observed wavenumber (cm-i) Wavenumber (cm -1)

Assignment

Raw SMO- adsorbed

CO- adsorbed

Bright-Edge 80-adsorbed

3300- 3446 3433 3300- 3200-3500 0—H stretching 3450 3430 3100-3500 N—H stretching

(Primary and secondary amines and amides stretching)

2935 2935 2932 2932 2700-2950 C—H stretching 2889 2912 2889 2891 1658 1658 1653 1655 1640-1670 C=0 amide 1558 1558 1558 1551 1400-1660 N—H bending

(Primary and secondary amines and amides)

1402 1385 1385 1384 1300-1420 COO- of carboxylate group

1230 1319 1269 1255 1000-1300 C-O stretching of COOH

1157 1158 1159 1157 1150 Phosphate groups 1086 1086 1082 1087 1030-1100 1053 1053 1056 1036 910-1040

152

Table 4.17: Functional groups of autoclaved biomass of M. rouxii biomass, oil adsorbed biomass and the corresponding infrared absorption wavelengths (Lin-Vien 1991)

Observed wavenumber (cm ') Wavenumber (cm-1)

Assignment

Raw SMO- CO- Bright-adsorbed adsorbed Edge 80-

adsorbed 3300- 3446 3433 3300- 3200-3500 O-H stretching 3450 3430 3100-3500 N-H stretching

(Primary and secondary amines and amides stretching)

2935 2935 2932 2932 2700-2950 C-H stretching 2889 2912 2889 2891 1658 1658 1653 1655 1640-1670 C=0 amide 1558 1558 1558 1551 1400-1660 N-H bending

(Primary and secondary amines and amides)

1402 1385 1385 1384 1300-1420 COO- of carboxylate group

1230 1319 1269 1255 1000-1300 C-0 stretching of COOH

1157 1158 1159 1157 1150 Phosphate groups 1086 1086 1082 1087 1030-1100 1053 1053 1056 1036 910-1040

152

Thus, the broad trough observed at 3,418 cm-1 might be an indication of both amine and

bonded OH groups. The C-H group was indicated by the trough (marked B and C)

observed at 2935 and 2889 cm'. A trough (marked D) was observed at 1659 cm-1 which

resulted from the C=0 stretching mode of primary and secondary amides and was

indicative of an amide 1 band. The trough (marked E) observed at 1,558 cm' showed the

presence of amide-2 which resulted from a combination of N-H bending and C=N

stretching. The trough (marked F) present at 1402 cm-1 showed the presence of COO" of

carboxylate group. The C-O stretch of carboxylic acid was expected in the range of 1320

- 1210 cm' and the trough (marked G) at 1230 cm' can be indicative of the same. The

troughs (marked H, I and J) at 1157, 1086 and 1053 cm-1 were due to the presence of a

phosphate functional group in the biomass. The functional groups assigned to the M.

rouxii biomass, used in this study, were in good agreement with that reported in the

literature (Majumdar et al. 2008; Yan and Viraraghavan, 2008). The FTIR spectra of the

three different oil adsorbed biomasses showed changes of band over the raw autoclaved

biomass. The trough observed at 3417 cm 1 corresponding to N-H stretching was shifted

to 3446, 3433 and 3431 cm 1 for the SMO-adsorbed, CO-adsorbed and Bright-Edge 80-

adsorbed biomass, respectively. Further, troughs at 1402 and 1230 cm' corresponding to

COO- and C-O stretch of —COOH groups, respectively, as observed for raw biomass, had

been shifted to a lower energy in all three cases after adsorption. A possible interaction of

oil molecules with amine and carboxyl groups of the biomass was inferred.

153

Thus, the broad trough observed at 3,418 cm-1 might be an indication of both amine and

bonded OH groups. The C-H group was indicated by the trough (marked B and C)

observed at 2935 and 2889 cm-1. A trough (marked D) was observed at 1659 cm-1 which

resulted from the C=0 stretching mode of primary and secondary amides and was

indicative of an amide 1 band. The trough (marked E) observed at 1,558 cm-1 showed the

presence of amide-2 which resulted from a combination of N-H bending and C=N

stretching. The trough (marked F) present at 1402 cm"1 showed the presence of COO" of

carboxylate group. The C-0 stretch of carboxylic acid was expected in the range of 1320

-1210 cm"1 and the trough (marked G) at 1230 cm'1 can be indicative of the same. The

troughs (marked H, I and J) at 1157, 1086 and 1053 cm"1 were due to the presence of a

phosphate functional group in the biomass. The functional groups assigned to the M.

rouxii biomass, used in this study, were in good agreement with that reported in the

literature (Majumdar et al. 2008; Yan and Viraraghavan, 2008). The FTIR spectra of the

three different oil adsorbed biomasses showed changes of band over the raw autoclaved

biomass. The trough observed at 3417 cm"1 corresponding to N-H stretching was shifted

to 3446, 3433 and 3431 cm"1 for the SMO-adsorbed, CO-adsorbed and Bright-Edge 80-

adsorbed biomass, respectively. Further, troughs at 1402 and 1230 cm'1 corresponding to

COO- and C-0 stretch of -COOH groups, respectively, as observed for raw biomass, had

been shifted to a lower energy in all three cases after adsorption. A possible interaction of

oil molecules with amine and carboxyl groups of the biomass was inferred.

153

4.13 Role of Surface Functional Groups, Lipids and Surface Charge on Oil

Biosorption

The modification of the functional groups and lipid extraction from the cell wall

of M rouxii biomass resulted in blocking those functional groups and the lipids from

taking part in oil biosorption. The different chemical treatments resulted in five types of

biomass: B1-biomass with carboxyl groups modified; B2-biomass with amino groups

modified; B3- biomass with phosphate groups modified; and B4 and B5-biomasses in

which lipid fractions of the cell walls were extracted. The IR spectra obtained for

biomass treated with methanol and hydrochloric acid (B1) indicated the trough observed

at 1402 cm-I was reduced in intensity, showing a partial removal of ester from the

biomass after chemical modification and as a result of esterification (Figure 4.29). The IR

spectrum of the biomass treated with formic acid and formaldehyde (B2), as presented in

Figure 4.30, showed the trough as observed at wave number 3417 cm-I in the raw

biomass, was reduced in intensity and had broadened, indicating changes in the amino

group or a reduction of the N-H bond of amines in the biomass. In the IR spectrum of

biomass treated with nitromethane and triethyl phosphite (B3), the shoulder near 1053

--1 cm was not apparent, indicating an esterification of the phosphate groups (Figure 4.31).

The IR spectrum of biomass treated with acetone and benzene (B4 and B5), as presented

in Figure 4.32 and Figure 4.33, respectively showed reduced intensity in troughs with

wave numbers 1402 and 1230 cm-I. This may be attributed to the extraction of

carboxylates from the biomass. The percent oil removal by raw and chemically modified

M rouxii biomass is given in Table 4.18.

154

4.13 Role of Surface Functional Groups, Lipids and Surface Charge on Oil

Biosorption

The modification of the functional groups and lipid extraction from the cell wall

of M. rouxii biomass resulted in blocking those functional groups and the lipids from

taking part in oil biosorption. The different chemical treatments resulted in five types of

biomass: Bl-biomass with carboxyl groups modified; B2-biomass with amino groups

modified; B3- biomass with phosphate groups modified; and B4 and B5-biomasses in

which lipid fractions of the cell walls were extracted. The IR spectra obtained for

biomass treated with methanol and hydrochloric acid (Bl) indicated the trough observed

at 1402 cm-1 was reduced in intensity, showing a partial removal of ester from the

biomass after chemical modification and as a result of esterification (Figure 4.29). The IR

spectrum of the biomass treated with formic acid and formaldehyde (B2), as presented in

Figure 4.30, showed the trough as observed at wave number 3417 cm"1 in the raw

biomass, was reduced in intensity and had broadened, indicating changes in the amino

group or a reduction of the N-H bond of amines in the biomass. In the IR spectrum of

biomass treated with nitromethane and triethyl phosphite (B3), the shoulder near 1053

cm"1 was not apparent, indicating an esterification of the phosphate groups (Figure 4.31).

The IR spectrum of biomass treated with acetone and benzene (B4 and B5), as presented

in Figure 4.32 and Figure 4.33, respectively showed reduced intensity in troughs with

wave numbers 1402 and 1230 cm-1. This may be attributed to the extraction of

carboxylates from the biomass. The percent oil removal by raw and chemically modified

M. rouxii biomass is given in Table 4.18.

154

50

45

40

35

30 .

25

20

15

10

5'

0

l'rudsorbed biomass

""— SMO adsorbed blooms,.

'CO adsorbed biomass

" 8'1011-Edge SO actuated "-ass.%

0.00 4...%

mo w'•

-I'

1%

..... $ •

0 • 't 0

•

n 5 11t I 1.1.

I.I

: \

. , 4000 3500 3000 2500 2000 1500 1000 500 0

Win nunthcr, cm 1

Figure 4.29: FTIR spectra of biomass residue (B1) after methanol and hydrochloric acid treatment before and after oil adsorption

155

so

SMO o&orted bmnuss. 45

~ ' ' B n g h t - E d g c M a d K H f c t d KiTat 40

35

30

25

20

15

10

5

0 4000 3500 3000 2500 2000 1500 1000 500 0

Wave number, cm!

Figure 4.29: FTIR spectra of biomass residue (Bl) after methanol and hydrochloric acid treatment before and after oil adsorption

155

Tra

nsr

nit

tan

ec

10

0

—'-uoadsorbed biomass

"''' "" SMO adsorbed biomass

CO adsorbed biomass

' • Bright-Edge 80 adsorbed biomass

...••••••••

• • ,•'• /1 • • "I

• • •

• "ft • •

/ ,• ,

„ Ot • '••

A / ts` I

e` '. - / %% I 4 ...,./' ,./

.,., '14 I ,1 % $ • • : $ %., f i

• .. ''''' , .II • t • '

1 1 I , 0

I ' o 11 I I V

1

0.' I ; . 0 e. It

I ‘t ,Iii. j ' t'i I I ' Fq • 1 1

' i : l ot % 4

11;

4000 3500 3000 2500 2000 1500 Ws% e number. cm

1000 500

Figure 4.30: FTIR spectra of biomass residue (B2) after formic acid and formaldehyde treatment before and after oil adsorption

156

60

50

30

20

10

' • ¥

»•» <»

""•unadsoibed biomass SMO adsaiwd hcmms

' "CO adsotwd biomass '' Bnghl-Edy 80 adtortwi biomMS

0 •L-> •— 4000 3500 2500 2000 1500

Wave number, em 1000 500

Figure 4.30: FTIR spectra of biomass residue (B2) after formic acid and formaldehyde treatment before and after oil adsorption

156

70

40

10

Unadsorbed biomass

SMO adsorbed bornass

- CO adsorbed biomass

" Blight-Edge BO adsorbed biomass

h001 * PIWON ir .0 0 NI

I. -5 v 5 ,

\

r"0..

• 1 I ,t;

,

. , 51 t 1 I •

3500 3000 2500 2000 Wave number. cm'

1500

****.-111

1

1000 500 0

Figure 4.31: FTIR spectra of biomass residue (B3) after nitromethane and triethylphosphite treatment before and after oil adsorption

157

70

50-

40

30

20

10

0 4000 3500 3000 0 2500 2000 1500 1000 500

Wave number, era-'

Figure 4.31: FTIR spectra of biomass residue (B3) after nitromethane and triethylphosphite treatment before and after oil adsorption

157

40

t.. 30

4000

--- Unadsorbed biomass

SMO adsorbed biomass

—CO adsorbed biomass

" Bri9ht-Edge 80 adsorbed biomass

3500 3000 2500 2000 Wave number. cm

1500

ay. ,•'

a 1

1000 500 0

Figure 4.32: FTIR spectra of biomass residue (B4) after acetone treatment before and after oil adsorption.

158

60 —Unadswbsd biomass -• SMO Khub«d two mast ' "CO adsorbed biomass '' Bright-Edge 80 adsortwd btomaa

50

40

30

20

10

0 4000 3500 3000 2500 2000 1500 1000 500 0

Wave number, cm

Figure 4.32: FTIR spectra of biomass residue (B4) after acetone treatment before and after oil adsorption.

158

rransm

ettan..z

70 Unadimbirl Moms

''''' SMO adsorbed biomass

- -00 adsothed Nom.-

- " Bright.tdre, 80 adsorbed biomass

1%.. ./ k 1,1

r • I

40

30

20

10

............

3500 3000 2500

Was munIvr.

2000

••• 1 -•,•••••

1,0, •

1500

•

1000 500 0

Figure 4.33: FTIR spectra of biomass residue (B5) after benzene treatment before and after oil adsorption

159

70

SMOadsiiitaibwass

• "CO adsorM tanas

•' Bright-Edge St) adsorbed bioniass

60

50

40

30

20

10

0 4000 3500 3000 2500 2000 1500 1000 0

Wave number, cm'1

Figure 4.33: FTIR spectra of biomass residue (B5) after benzene treatment before and after oil adsorption

159

Table 4.18: Oil removal by raw and chemically modified M rouxii biomass

Biomass Oil removal (%) SMO CO Bright-Edge 80

Raw 98 99 98 B1 98 80 43 B2 98 99 98 B3 99 98 98 B4 98 98 97 B5 90 78 54

160

Table 4.18: Oil removal by raw and chemically modified M. rouxii biomass

Biomass Oil removal (%) SMO CO Bright-Edge 80

Raw 98 99 98 B1 98 80 43 B2 98 99 98 B3 99 98 98 B4 98 98 97 B5 90 78 54

160

Results of oil biosorption studies for the five chemically treated biomasses showed that

percent removal of SMO was lower for B5 by 10% over that obtained by raw biomass.

As for CO and Bright-Edge 80, percent removal of oil was lower than raw biomass when

B1 and B5 were involved. Percent removal of Bright-Edge 80 was reduced to 42.5% and

53.5% when BI and B5 were used, respectively. Both B1 and B5 have been modified to

extract the carboxyl functional groups and hence, the results of the oil biosorption study

suggest carboxyl groups might have taken part in oil biosorption. The trough observed at

the raw M rouxii at wavenumber 1402 cm' is indicative of deprotonated COO- of

carboxylate groups. At pH 6.4, which is the pH of the raw M rouxii biomass, the surface

charge of the material remains negative (Figure 4.1). When the oil biosorption study was

conducted at pH 3.0, protonation of such oxygen containing functional groups (COO-)

could have reduced the repulsion between the oil and other dissociated groups in the

biomass resulting in better removals. The FTIR spectra of oil adsorbed chemically

modified biomasses showed peak shift and an evident reduction in the intensity of the

3417 cm -1 N-H stretching trough, for most cases. However, oil removals using biomass

with methylation of amino groups (B2) were not significant. It should be noted that FTIR

spectra of chemically modified B2 biomass had indicated the presence of N-H stretching,

after chemical modification, with a reduced intensity in comparison with that of the raw

M rouxii biomass. The presence of the N-H stretching peak after chemical modification

of the amine groups (B2) was observed by Yan and Viraraghavan (2008). There was a

possibility that the strong band observed in the region of 3300 - 3450 cm-1 for raw M

rouxii biomass was a result of N-H and 0-H stretching vibrations. Since the M rouxii

161

Results of oil biosoiption studies for the five chemically treated biomasses showed that

percent removal of SMO was lower for B5 by 10% over that obtained by raw biomass.

As for CO and Bright-Edge 80, percent removal of oil was lower than raw biomass when

B1 and B5 were involved. Percent removal of Bright-Edge 80 was reduced to 42.5% and

53.5% when B1 and B5 were used, respectively. Both B1 and B5 have been modified to

extract the carboxyl functional groups and hence, the results of the oil biosorption study

suggest carboxyl groups might have taken part in oil biosorption. The trough observed at

the raw M. rouxii at wavenumber 1402 cm"1 is indicative of deprotonated COO- of

carboxylate groups. At pH 6.4, which is the pH of the raw M. roiaii biomass, the surface

charge of the material remains negative (Figure 4.1). When the oil biosorption study was

conducted at pH 3.0, protonation of such oxygen containing functional groups (COO-)

could have reduced the repulsion between the oil and other dissociated groups in the

biomass resulting in better removals. The FTIR spectra of oil adsorbed chemically

modified biomasses showed peak shift and an evident reduction in the intensity of the

3417 cm"1 N-H stretching trough, for most cases. However, oil removals using biomass

with methylation of amino groups (B2) were not significant. It should be noted that FTIR

spectra of chemically modified B2 biomass had indicated the presence of N-H stretching,

after chemical modification, with a reduced intensity in comparison with that of the raw

M. rouxii biomass. The presence of the N-H stretching peak after chemical modification

of the amine groups (B2) was observed by Yan and Viraraghavan (2008). There was a

possibility that the strong band observed in the region of 3300 - 3450 cm"1 for raw M.

rouxii biomass was a result of N-H and O-H stretching vibrations. Since the M. rouxii

161

cell wall contains different polymers such as chitosan, glucan, proteins and lipids, it may

not be possible that there is overlapping of N-H and 0-H stretching vibrations (Majumdar

et al., 2008). The FTIR analysis of the M rouxii biomass showed the biomass surface

consists mainly of amino, amide, carboxyl, hydroxyl, and phosphate functional groups.

The study showed the involvement of —COON and to some extent -NH2 in oil uptake. Oil

removal was not affected in the cases of B2, B3 and B4 biomasses.

4.14 Immobilization of Fungal Biomass and its use in Biosorption

The use of dead biomass in powdered form in a fixed bed column has limitations

such as difficulty in the separation of biomass after adsorption, mass loss after

regeneration (Kapoor and Viraraghavan 1998; Yan and Viraraghavan 2000), and low

strength and small particle size (Tsezos 1990). The dead biomass must be immobilized in

a granular or polymeric matrix to improve the mechanical strength of the biosorbent.

Polysulfone was used as the solid matrix in the study as it has been proven to be an

amorphous, rigid, heat resistant and chemically stable thermoplastic material (Kapoor and

Viraraghavan 1998). The sieve analysis of the beads showed that the size of the majority

of the immobilized M rouxii biomass beads ranged from 0.42 to 2.36 mm. The

coefficient of uniformity Cu, of the beads could be calculated using the equation:

C = 6°u Dw

162

4.1

cell wall contains different polymers such as chitosan, glucan, proteins and lipids, it may

not be possible that there is overlapping of N-H and O-H stretching vibrations (Majumdar

et al., 2008). The FTIR analysis of the M. rouxii biomass showed the biomass surface

consists mainly of amino, amide, carboxyl, hydroxyl, and phosphate functional groups.

The study showed the involvement of-COOH and to some extent -NH2 in oil uptake. Oil

removal was not affected in the cases of B2, B3 and B4 biomasses.

4.14 Immobilization of Fungal Biomass and its use in Biosorption

The use of dead biomass in powdered form in a fixed bed column has limitations

such as difficulty in the separation of biomass after adsorption, mass loss after

regeneration (Kapoor and Viraraghavan 1998; Yan and Viraraghavan 2000), and low

strength and small particle size (Tsezos 1990). The dead biomass must be immobilized in

a granular or polymeric matrix to improve the mechanical strength of the biosorbent.

Polysulfone was used as the solid matrix in the study as it has been proven to be an

amorphous, rigid, heat resistant and chemically stable thermoplastic material (Kapoor and

Viraraghavan 1998). The sieve analysis of the beads showed that the size of the majority

of the immobilized M. rouxii biomass beads ranged from 0.42 to 2.36 mm. The

coefficient of uniformity Cu, of the beads could be calculated using the equation:

C 4.1 " D ^10

162

where D is the diameter of the particles and the subscript (10, 60) refers to the percent

that is smaller. Results showed that the value of Cu was 1.71, the effective size (D10) of

the beads was 1.28 mm and the average particle size was 2.11 mm. The plot of diameter

versus weight percent of immobilized biomass beads passing the related sieve is shown in

Figure 4.34. The characteristics of the immobilized beads are given in Table 4.19.

4.14.1 Batch Studies using Immobilized Biomass Beads

Batch kinetic studies similar to that of powdered biomass were conducted using

immobilized biomass to investigate changes in equilibrium time. Polysulfone beads

without biomass were used as a control and it was found to remove less than 2% of oil, in

all three cases, which was negligible. Also, the beads did not disintegrate due to agitation

at 175 rpm for 6 hours. Batch studies with immobilized beads were conducted at two

different pH conditions of 3.0 and 7.6 for 6 hours. Following 6 hours, approximately 86

to 89% and 25 to 76% of oil was removed at a pH of 3.0 and 7.6, respectively. The values

were lower than those obtained using powdered biomass (Tables A.10 and A.11). Batch

kinetic studies conducted using immobilized beads at a pH of 6.0 consisted of a primary

rapid phase followed by a secondary slow phase. Equilibrium times were 5, 3 and 4 hours

for SMO, CO and Bright-Edge 80, respectively. Immobilization of the biomass may have

either affected the free movement of solute to the sites or masked the sites (Prakashan et

al., 1999). A change in the characteristics of the biomass after immobilization may be

responsible for the longer equilibrium time and lower performance.

163

where D is the diameter of the particles and the subscript (10, 60) refers to the percent

that is smaller. Results showed that the value of Cu was 1.71, the effective size (Dio) of

the beads was 1.28 mm and the average particle size was 2.11 mm. The plot of diameter

versus weight percent of immobilized biomass beads passing the related sieve is shown in

Figure 4.34. The characteristics of the immobilized beads are given in Table 4.19.

4.14.1 Batch Studies using Immobilized Biomass Beads

Batch kinetic studies similar to that of powdered biomass were conducted using

immobilized biomass to investigate changes in equilibrium time. Polysulfone beads

without biomass were used as a control and it was found to remove less than 2% of oil, in

all three cases, which was negligible. Also, the beads did not disintegrate due to agitation

at 175 rpm for 6 hours. Batch studies with immobilized beads were conducted at two

different pH conditions of 3.0 and 7.6 for 6 hours. Following 6 hours, approximately 86

to 89% and 25 to 76% of oil was removed at a pH of 3.0 and 7.6, respectively. The values

were lower than those obtained using powdered biomass (Tables A. 10 and A.l 1). Batch

kinetic studies conducted using immobilized beads at a pH of 6.0 consisted of a primary

rapid phase followed by a secondary slow phase. Equilibrium times were 5, 3 and 4 hours

for SMO, CO and Bright-Edge 80, respectively. Immobilization of the biomass may have

either affected the free movement of solute to the sites or masked the sites (Prakashan et

al., 1999). A change in the characteristics of the biomass after immobilization may be

responsible for the longer equilibrium time and lower performance.

163

90

80

70

r 60

cs. 50

1 40

10

0 0 0.5 1 1.5

Diameter (mm)

1 2.5

Figure 4.34: Diameter size versus percent of immobilized beads passing the sieve

164

90

80

70

60

50

40

30

20

10

0 0 0.5 1.5 2.5 1 2

Diameter (mm)

Figure 4.34: Diameter size versus percent of immobilized beads passing the sieve

164

Table 4.19: Characteristics of immobilized M rouxii biomass

Characteristics Values

Uniformity coefficient 1.71

Effective size (mm) 1.28

Average size (mm) 2.11

Bulk density (g/cm3) 0.119

Porosity (%) 47.5

165

Table 4.19: Characteristics of immobilized M. rouxii biomass

Characteristics Values

Uniformity coefficient 1.71

Effective size (mm) 1.28

Average size (mm) 2.11

Bulk density (g/cm3) 0.119

Porosity (%) 47.5

165

4.15 Column Breakthrough Studies

4.15.1 The Thomas Model

Breakthrough experiments lasted 42, 24, and 30 hours for SMO, CO and Bright-Edge 80,

respectively. The initial pH of the three emulsions was in the range of 7.5 to 7.6 and in

the effluent it was in the range of 7.0 to 7.3. Data were fitted to the Thomas model

through non-linear regression analysis to determine the maximum solid-phase

concentration (qo) using STATISTICA (Release 5.0). Breakthrough curves, predicted

using the Thomas model for the biosorption of SMO, CO and Bright-Edge 80 on the

immobilized beads, are shown in Figures 4.35 — 4.37, respectively. According to the high

correlation coefficient (r > 0.95) and the t statistical significance at the 95% confidence

level, it can be said that the Thomas equation could reasonably describe the breakthrough

data. However, the Thomas model does not give a good fit to the experimental data at the

beginning. Initially adsorption was very rapid probably due to the availability of reaction

sites to remove oil. When these sites are gradually occupied, the removal becomes less

effective. A possible reason could be that when the residence time of the solute in the

column is not long enough for adsorption equilibrium to be reached, the oil in the

emulsion leaves the column before equilibrium occurs. The Thomas Equation parameters

such as Kr and qo values are shown in Table 4.20. The Thomas model was derived from

the equation of mass conservation in a flow system (Thomas 1948). The model is

developed based on the assumption that the rate driving force obeys second-order

reversible reaction kinetics, and adsorption equilibrium follows the Langmuir model with

no axial dispersion (Thomas 1948).

166

4.15 Column Breakthrough Studies

4.15.1 The Thomas Model

Breakthrough experiments lasted 42, 24, and 30 hours for SMO, CO and Bright-Edge 80,

respectively. The initial pH of the three emulsions was in the range of 7.5 to 7.6 and in

the effluent it was in the range of 7.0 to 7.3. Data were fitted to the Thomas model

through non-linear regression analysis to determine the maximum solid-phase

concentration (qo) using STATISTICA (Release 5.0). Breakthrough curves, predicted

using the Thomas model for the biosorption of SMO, CO and Bright-Edge 80 on the

immobilized beads, are shown in Figures 4.35 - 4.37, respectively. According to the high

correlation coefficient (r > 0.95) and the t statistical significance at the 95% confidence

level, it can be said that the Thomas equation could reasonably describe the breakthrough

data. However, the Thomas model does not give a good fit to the experimental data at the

beginning. Initially adsorption was very rapid probably due to the availability of reaction

sites to remove oil. When these sites are gradually occupied, the removal becomes less

effective. A possible reason could be that when the residence time of the solute in the

column is not long enough for adsorption equilibrium to be reached, the oil in the

emulsion leaves the column before equilibrium occurs. The Thomas Equation parameters

such as Kt and qo values are shown in Table 4.20. The Thomas model was derived from

the equation of mass conservation in a flow system (Thomas 1948). The model is

developed based on the assumption that the rate driving force obeys second-order

reversible reaction kinetics, and adsorption equilibrium follows the Langmuir model with

no axial dispersion (Thomas 1948).

166

1

0.9

0.8

0.7

0.6

0 0.5

U

0.4

0.3

0.2

0.1

0 0.5 1 1.5

4

lK 4

3: Experimental

iThomas Predicted

2.5 3 3.5 4 4.5 5 5.5

Volume (L)

Figure 4.35: Breakthrough curves for SMO predicted using Thomas model

167

0.9

0.8

0.7

0.6

0.5

0.4

0.3

* Experimental 0.2

Thomas Predicted 0.1

0 0 0.5 1 1.5 2 2.5 3 3.5 4 4.5 5 5.5 6 6.5 7

Volume (L)

Figure 4.35: Breakthrough curves for SMO predicted using Thomas model

167

0.8 •

0.7

0.6

c) Sz..) 0.5

0.4

0.3

0.2

0.1

0

3: Experimental

- r

Thomas Predicted

0.5 1 1.5 2 2.5 3 3.5 4

Volume (L)

Figure 4.36: Breakthrough curves for CO predicted using Thomas model

168

0.9

0.8

0.7

0.6

0.4

0.3

0.2 * Experimental

Thomas Predicted 0.1

0.5 0 1.5 1 2 2.5 3.5 3 4

Volume (L)

Figure 4.36: Breakthrough curves for CO predicted using Thomas model

168

0

0.9

0.8

0.7

0.6

0.5

0.4

1 1.5 2 2.5

Volume (L)

3: Experimental

3.5

Thomas Predicted

4 4.5

Figure 4.37: Breakthrough curves for Bright-Edge 80 predicted using Thomas model

169

0.9

0.8

0.7

0.6

o ^ 0.5 U

0.4

0.3

0.2

* Experimental

—Thomas Predicted 0.1

0 0.5 1.5 2.5 1 2 3 3.5 4 4.5 5

Volume (L)

Figure 4.37: Breakthrough curves for Bright-Edge 80 predicted using Thomas model

169


SMO CO Bright-Edge 80 Model Constants r Constants r Constants r Thomas KT = 7x10-5

L/min mg qo = 14.67 mg/g

0.95 KT = 8.3x10-5L/min mg qo = 9.87 mg/g

0.98 KT = 6x10-5L/min mg qo = 10.67 mg/g

0.97

Yan a = 1.29 0.99 a = 1.29 0.95 a = 0.98 0.95 Belter to = 520 min

a = 0.98 0.92 to = 347.1

min a= 1.17

0.96 to = 376.7 min a= 1.39

0.95

Chu form 1 (Eq. 2.14)

to = 410.5 min a = 0.84

0.81 to = 381.8 min a= 1.19

0.82 to = 405.2 min a= 1.36

0.74


to = 420 min a = 0.46

0.93 to = 311.6 min a = 0.49

0.86 to = 338 min a = 0.48

0.76

Yoon Nelson KyN = 0.002 min-' t112 = 775 min

0.91 KyN = 0.003 min-' tin = 406 min

0.96 KyN = 0.003 min i• -

t112 = 424 min

0.96

Oulman K = 6.5x10-5L/min mg N = 25430.6 mg/L

0.96 K = 8.3x10-5L/min mg N = 17081.6 mg/L

0.98 K = 6.4x10-5 L/min mg N = 18508.8 mg/L

0.98

Wolbroska 13a = 0.16 min-' No = 2058.8 mg/L

0.87 a = 0.11 min-1No = 2143 mg/L

0.9 pa = 0.1 min-' No = 2292 mg/L

0.95

Second run (after regeneration) Thomas

KT = 1.4x10-4L/min mg qo = 2.54 mg/g

0.91 KT = 1.1x10-4L/min mg qo = 0.03 mg/g

0.9 KT = 1.05x10-4L/min mg qo = 0.49 mg/g

0.9

Note: r - correlation coefficient. All model parameters statistically significant (t-test) at 95% confidence

level

170


SMO CO Bright-Edge 80 Model Constants r Constants r Constants r Thomas Kj = 7xl0'5

L/min mg q0 = 14.67 mg/g

0.95 £T = 8.3xl0"5


0.98 Kt = 6xl0"5


0.97

Yan a= 1.29 0.99 a = 1.29 0.95 a = 0.98 0.95 Belter to = 520 min

a = 0.98 0.92 to = 347.1

min o= 1.17

0.96 to = 376.7 min a = 1.39

0.95


to = 410.5 min a = 0.84

0.81 to = 381.8 min

1.19

0.82 to = 405.2 min a = 1.36

0.74


to = 420 min a = 0.46

0.93 to = 311.6 min o = 0.49

0.86 to = 338 min a = 0.48

0.76

Yoon Nelson KYN = 0.002 min"1

tl/2 = 775 min

0.91 KYN = 0.003 min"1

tj/2 = 406 min

0.96 KYN =0.003 min"1

ti/2 = 424 min

0.96

Oulman K = 6.5x10 s

L/min mg N = 25430.6 mg/L

0.96 K = 8.3xl0"5

L/min mg N = 17081.6 mg/L

0.98 K = 6.4x10' 5 L/min mg N 18508.8 mg/L

0.98

Wolbroska Pa = 0.16 min'1

N0 = 2058.8 mg/L

0.87 Pa = 0.11 min"1

No = 2143 mg/L

0.9 Pa = 0.1 min"1

No = 2292 mg/L

0.95

Second run (after regeneration) Thomas

Kr= 1.4xl0"4


0.91 Kt= l.lxl 0"4


0.9 Kr

1.05xl0"4


0.9

Note: r - correlation coefficient. All model parameters statistically significant (/-test) at 95% confidence

level

170

Based on the qo values obtained for the Thomas model, the adsorption capacity of oil on

the immobilized M rouxii beads were 14.67, 9.87 and 10.67 mg/g for SMO, CO, and

Bright-Edge 80, respectively. The maximum solid phase concentrations were based on

the mass of beads. If the concentrations were based on the biomass mass, the maximum

solid phase concentration would be twice as high as the values shown above, as the bead

composition consisted of 50% biomass and 50% polysulfone. Thus, qo values based on

the mass of biomass would be 29.34, 19.74 and 21.34 mg/g for SMO, CO, and Bright-

Edge 80, respectively. The values of the Thomas constants from column studies using

other sorbents for oil removal are given in Table 4.21. The adsorption capacity (q0)

values, calculated from the Thomas model, imply that the immobilized M rouxii biomass

showed a higher oil adsorption capacity than vermiculite for both SMO and CO.

However, the adsorption capacity of SMO by horticultural peat was higher than that of

the immobilized M rouxii biomass, which may be attributed to the fibrous nature of peat.

4.15.2 The Yan Model

The value of maximum solid phase concentration of solute (q0) obtained from the

Yan's model, based on the immobilized M rouxii bead mass were 11.6, 9.4, and 10.4

mg/g for SMO, CO, and Bright-Edge 80, respectively. The go value, based on the mass of

biomass, would be 23.2, 18.8 and 20.8 mg/g for SMO, CO, and Bright-Edge 80,

respectively. The qo values predicted by the Thomas model were slightly higher than the

values predicted by the Yan model for all three oils. Yan's empirical model is developed

171

Based on the qg values obtained for the Thomas model, the adsorption capacity of oil on

the immobilized M. rouxii beads were 14.67, 9.87 and 10.67 mg/g for SMO, CO, and

Bright-Edge 80, respectively. The maximum solid phase concentrations were based on

the mass of beads. If the concentrations were based on the biomass mass, the maximum

solid phase concentration would be twice as high as the values shown above, as the bead

composition consisted of 50% biomass and 50% polysulfone. Thus, qo values based on

the mass of biomass would be 29.34, 19.74 and 21.34 mg/g for SMO, CO, and Bright-

Edge 80, respectively. The values of the Thomas constants from column studies using

other sorbents for oil removal are given in Table 4.21. The adsorption capacity (q0)

values, calculated from the Thomas model, imply that the immobilized M. rouxii biomass

showed a higher oil adsorption capacity than vermiculite for both SMO and CO.

However, the adsorption capacity of SMO by horticultural peat was higher than that of

the immobilized M. rouxii biomass, which may be attributed to the fibrous nature of peat.

4.15.2 The Yan Model

The value of maximum solid phase concentration of solute (q0) obtained from the

Yan's model, based on the immobilized M. rouxii bead mass were 11.6, 9.4, and 10.4

mg/g for SMO, CO, and Bright-Edge 80, respectively. The q0 value, based on the mass of

biomass, would be 23.2, 18.8 and 20.8 mg/g for SMO, CO, and Bright-Edge 80,

respectively. The qQ values predicted by the Thomas model were slightly higher than the

values predicted by the Yan model for all three oils. Yan's empirical model is developed

171

Table 4.21: Comparison of Thomas constants for other oil sorbents

Oil Media Comg/ L

C mg/ L

Bed mass/ depth kg/m

Flow mL/ min

Breakt- hrough volume L

K L/h- kg

co kg/kg

SMO Peat 218 84 0.175/ 50 782 27.6 1.58 0.3

MCO Peat 210 187 0.175/ 50 1080 30.2 1.33 0.3

CUT Peat 278 205 0.175/ 25 120 32.4 0.14 0.3

RE Peat 8.9 8.3 0.175/ 25 480 21 0.03 0.3

PW Peat 38 37 0.175/ 50 810 85.9 0.19 0.3

SMO Organolcay/ anthracite

46 44 0.225/1 12 47.5 113 0.0025

KUT Organolcay/ anthracite

49 44 0.225/1 12 34.6 1748 0.0014

VAL Organolcay/ anthracite

51 48 0.225/1 12 30.2 9328 0.0008

RE Organoclay/ anthracite

8.3 6 0.225/1 12 95.0 3892 0.0019

SMO Vermiculite 47 43 0.016/ 12 8.64 0.34 0.0012 0.2

CO Vermiculite 45 41 0.016/ 12 8.64 0.08 0.0077 0.2

KUT Vermiculite 30 27.3 0.016/ 12 14.4 0.17 0.0061 0.2

RE Vermiculite 11 10 0.016/ 12 14.4 0.08 0.0027 0.2

SMO Immobilized 50 48 0.0045/ 2.6 6.55 4200 0.0098 M rouxii 0.3

CO Immobilized 50 48 0.0045/ 2.6 3.74 4980 0.0146 M rouxii 0.3

Bright Immobilized 50 48 0.0045/ 2.6 4.68 3600 0.0106 -Edge M rouxii 0.3 80

Note: Co - Influent concentration; C - Effluent concentration MCO - Midale crude oil; CUT- Cutting RE - Refinery effluent; PW - Produced water; KUT- Kutwell 45 oil; VAL - Valcool oil. The values of horticultural peat were obtained from Viraraghavan and Mathavan (1989); the values of organo clay/anthracite were obtained from Moazed (2001); the values of vermiculite were obtained from Mysore et al. (2006). All model parameters statistically significant (t-test) at 95% confidence level.

172

Table 4.21: Comparison of Thomas constants for other oil sorbents

Oil Media Co C Bed Flow Breakt K qo mg/ mg/ mass/ mL/ hrough L/h- kg/kg L L depth min volume kg

kg/m L SMO Peat 218 84 0.175/ 50 782 27.6 1.58

0.3 MCO Peat 210 187 0.175/ 50 1080 30.2 1.33

0.3 CUT Peat 278 205 0.175/ 25 120 32.4 0.14

0.3 RE Peat 8.9 8.3 0.175/

n i 25 480 21 0.03

PW Peat 38 37 V/.J

0.175/ 50 810 85.9 0.19 0.3

SMO Organolcay/ 46 44 0.225/1 12 47.5 113 0.0025 anthracite

KUT Organolcay/ 49 44 0.225/1 12 34.6 1748 0.0014 anthracite

VAL Organolcay/ 51 48 0.225/1 12 30.2 9328 0.0008 anthracite

RE Organoclay/ 8.3 6 0.225/1 12 95.0 3892 0.0019 anthracite

SMO Vermiculite 47 43 0.016/ 12 8.64 0.34 0.0012 0.2

CO Vermiculite 45 41 0.016/ 12 8.64 0.08 0.0077 0.2

KUT Vermiculite 30 27.3 0.016/ 12 14.4 0.17 0.0061 0.2

RE Vermiculite 11 10 0.016/ 12 14.4 0.08 0.0027 0.2

SMO Immobilized 50 48 0.0045/ 2.6 6.55 4200 0.0098 M. rouxii 0.3

CO Immobilized 50 48 0.0045/ 2.6 3.74 4980 0.0146 M. rouxii 0.3

Bright Immobilized 50 48 0.0045/ 2.6 4.68 3600 0.0106 -Edge M. rouxii 0.3 80

Note: C0 - Influent concentration; C - Effluent concentration; MCO - Midale crude oil; CUT- Cutting oil, RE - Refinery effluent; PW - Produced water; KUT- Kutwell 45 oil; VAL - Valcool oil. The values of horticultural peat were obtained from Viraraghavan and Mathavan (1989); the values of organo clay/anthracite were obtained from Moazed (2001); the values of vermiculite were obtained from Mysore et al. (2006). All model parameters statistically significant (/-test) at 95% confidence level.

172

to minimize the error resulting from the use of the Thomas model, especially at lower or

higher time periods of the breakthrough curve.Breakthrough curves, predicted using the

Yan model for biosorption of SMO, CO and Bright-Edge 80 on the immobilized beads,

are shown in Figures 4.38 — 4.40, respectively. High correlation coefficient (r > 0.95)

values and t statistical significance at a 95% confidence level were obtained from the Yan

model. This showed the Yan model described the breakthrough curves with great

accuracy. The values of the empirical constant 'a' and correlation coefficient for the oils

are given in Table 4.20. The limitation of the Yan model is that it may be difficult to

relate the empirical parameter 'a' with the experimental conditions, so the scale up of the

system may be difficult (Lodeiro et al. 2006).

4.15.3 The Belter and Chu Models

Experimental data were fitted to the mathematical models proposed by Belter et

al. (1988) and Chu (2004) in Figures 4.41 — 4.43 for SMO, CO, and Bright-Edge 80,

respectively. Equations 2.14 and 2.15 were modified from Equation 2.13 by Chu (2004)

so that the experimental data would fit better even if the breakthrough curve were

symmetrical or asymmetrical. The Belter model is capable of describing only symmetric

breakthrough curve behavior. The process variables such as the flow rate, bed height and

average size of the sorbent were not used to develop Equation 2.13 and it is necessary to

empirically correlate the two model parameters with these variables (Chu 2004). The

model parameters a and to and the correlation coefficient (r) values obtained for the oils

are shown in Table 4.20. With respect to all three oils, the breakthrough curve showed a

173

to minimize the error resulting from the use of the Thomas model, especially at lower or

higher time periods of the breakthrough curve.Breakthrough curves, predicted using the

Yan model for biosorption of SMO, CO and Bright-Edge 80 on the immobilized beads,

are shown in Figures 4.38 - 4.40, respectively. High correlation coefficient (r > 0.95)

values and t statistical significance at a 95% confidence level were obtained from the Yan

model. This showed the Yan model described the breakthrough curves with great

accuracy. The values of the empirical constant'a' and correlation coefficient for the oils

are given in Table 4.20. The limitation of the Yan model is that it may be difficult to

relate the empirical parameter 'o' with the experimental conditions, so the scale up of the

system may be difficult (Lodeiro et al. 2006).

4.15.3 The Belter and Chu Models

Experimental data were fitted to the mathematical models proposed by Belter et

al. (1988) and Chu (2004) in Figures 4.41 - 4.43 for SMO, CO, and Bright-Edge 80,

respectively. Equations 2.14 and 2.15 were modified from Equation 2.13 by Chu (2004)

so that the experimental data would fit better even if the breakthrough curve were

symmetrical or asymmetrical. The Belter model is capable of describing only symmetric

breakthrough curve behavior. The process variables such as the flow rate, bed height and

average size of the sorbent were not used to develop Equation 2.13 and it is necessary to

empirically correlate the two model parameters with these variables (Chu 2004). The

model parameters a and to and the correlation coefficient (r) values obtained for the oils

are shown in Table 4.20. With respect to all three oils, the breakthrough curve showed a

173

0 0.5 1 1.5

r Experimental

Yan Predicted

2.5 3 3.5 4 4.5 5 5.5

Volume (L)

6.5

Figure 4.38: Breakthrough curves for SMO predicted using Yan model

174

1

0.9

0.8

0.7

0.6

0.5

0.4

0.3

0.2 * Experimental

•Yan Predicted 0.1

0 0 0.5 1 1.5 2 2.5 3 3.5 4 4.5 5 5.5 6 6.5 7

Volume (L)

Figure 4.38: Breakthrough curves for SMO predicted using Yan model

174

0

R. 0.5 .

0.4

4

0 0.5 1 1.5 2 2.5

Volume (L)

3: Experimental

••• •Yan Predicted

3.5

Figure 4.39: Breakthrough curves for CO predicted using Yan model

175

0.9

0.8

0.7

0.6

^0.5

0.4

0.3

0.2 * Experimental

0.1 Yan Predicted

0 0.5 1 1.5 2.5 2 3.5 3 4

Volume (L)

Figure 4.39: Breakthrough curves for CO predicted using Yan model

175

4

4

x Experimental

Yan Predicted

0 0.5 1 1.5 2 2.5 Volume (L)

Figure 4.40: Breakthrough curves for Bright-Edge 80 predicted using Yan model

176

3.5

0.9

0.8

0.7

0.6

0.4

0.3

0.2 * Experimental

0.1 Yan Predicted

0 0.5 4 1 1.5 2 2.5 3 3.5 Volume (L)

Figure 4.40: Breakthrough curves for Bright-Edge 80 predicted using Yan model

176

0.8

0 0.6

C.)

0

4 ............... ........

.....

-SMO C/CO Belter and Cussler Predicted

•-• SMO C/CO Chu 1 Predicted

SMO C/C0 Chu 2 Predicted

SMO C/CO Experimental

500 1000 1500

Time, min

2000 2500 3000

Figure 4.41: Breakthrough curves for SMO predicted using Belter and Chu models

177

0.8

SMO C/CO Belter and Cussler Predicted

0.4

SMO C/CO Chu 1 Predicted

SMO C/CO Chu 2 Predicted

0.2

* SMO C/CO Experimental

1000 3000 0 500 1500 2000 2500

Time, min

Figure 4.41: Breakthrough curves for SMO predicted using Belter and Chu models

177

X CO MO Experimental

CO MO Belter and Cussler Predicted

—• — CO MO Chu 1 Predicted

CO C/C0 Chu 2 Predicted

0 200 400 600 800 1000

Time, min

1200 1400 1600

Figure 4.42: Breakthrough curves for CO predicted using Better and Chu models

178

1

0.8

* CO C/CO Experimental 0.4

CO C/CO Belter and Cussler Predicted

CO C/CO Chu 1 Predicted 0.2

CO C/CO Chu 2 Predicted

0 200 600 0 400 1000 1200 1400 1600 800

Time, min

Figure 4.42: Breakthrough curves for CO predicted using Belter and Chu models

178

c0.6 frig

.1.1It WNW

4

.....

....................................

U Bright-Edge 80 C/CO Experimental

0.4 / """""•••"Bright-Edge 80 C/CO Belter and Cussler : Predicted

• ••• Bright-Edge 80 C/CO Chu 1 Predicted

0.2 Bright-Edge 80 C/CO Chu 2 Predicted

*•

1••••0 +

0 200 400 600 800 1000 1200 1400 1600 1800 2000

Time, min

Figure 4.43: Breakthrough curves for Bright-Edge 80 predicted using Belter and Chu models

179

0.8

* Bright-Edge 80 C/CO Experimental

0.4 Bright-Edge 80 C/CO Belter and Cussler

Predicted

Bright-Edge 80 C/CO Chu 1 Predicted

0.2 Bright-Edge 80 C/CO Chu 2 Predicted

0 200 400 600 800 1000 1200 1400 1600 1800 2000

Time, min

Figure 4.43: Breakthrough curves for Bright-Edge 80 predicted using Belter and Chu models

179

good fit to the model described by Equation 2.13 with r2 values higher than 0.92 at a 95%

confidence level. The model equation employed by Chu in Equation 2.15 (with negative

sign in exponential terms) fitted better than Equation 2.14 (with positive sign in

exponential terms) for all oils. In the case of SMO, the breakthrough curve was well

described by both Equations 2.13 and 2.15 showing a high r2 value (> 0.92) at a 95%

confidence interval. In the case of all oils, Equation 2.15 underestimates the effluent oil

concentration, while the model in Equation 2.14 overestimates the experimental values.

In addition, Equations 2.13 and 2.14 predict a nonzero effluent concentration at t = 0

which contradicts real conditions for the three oils. The modeling results of Chu (2004)

showed that Equation 2.14 was capable of simulating breakthrough curves with a broad

leading edge whileEquation 2.15 could be used to describe breakthrough curves with a

broad trailing edge. However, using these simple models to design or optimize fixed bed

biosorption columns requires knowledge of the effect of the process variables on the two

model parameters, to and a (Chu 2004).

4.15.4 The Yoon and Nelson Model

A simple theoretical model developed by Yoon—Nelson was applied to investigate the

breakthrough behavior of the oils on immobilized biomass beads. This model is based

upon the assumption that the rate of decrease in the probability of adsorption for each

adsorbate molecule is proportional to the probability of adsorbate adsorption and the

probability of adsorbate breakthrough on the adsorbent (Hamdaoui 2006). The values of

KyN and t12 are listed in Table 4.20. Breakthrough curves predicted using the Yoon and

180

good fit to the model described by Equation 2.13 with r2 values higher than 0.92 at a 95%

confidence level. The model equation employed by Chu in Equation 2.15 (with negative

sign in exponential terms) fitted better than Equation 2.14 (with positive sign in

exponential terms) for all oils. In the case of SMO, the breakthrough curve was well

described by both Equations 2.13 and 2.15 showing a high r2 value (> 0.92) at a 95%

confidence interval. In the case of all oils, Equation 2.15 underestimates the effluent oil

concentration, while the model in Equation 2.14 overestimates the experimental values.

In addition, Equations 2.13 and 2.14 predict a nonzero effluent concentration at t = 0

which contradicts real conditions for the three oils. The modeling results of Chu (2004)

showed that Equation 2.14 was capable of simulating breakthrough curves with a broad

leading edge whileEquation 2.15 could be used to describe breakthrough curves with a

broad trailing edge. However, using these simple models to design or optimize fixed bed

biosorption columns requires knowledge of the effect of the process variables on the two

model parameters, to and a (Chu 2004).

4.15.4 The Yoon and Nelson Model

A simple theoretical model developed by Yoon-Nelson was applied to investigate the

breakthrough behavior of the oils on immobilized biomass beads. This model is based

upon the assumption that the rate of decrease in the probability of adsorption for each

adsorbate molecule is proportional to the probability of adsorbate adsorption and the

probability of adsorbate breakthrough on the adsorbent (Hamdaoui 2006). The values of

Kyn and ti/2 are listed in Table 4.20. Breakthrough curves predicted using the Yoon and

180

Nelson model are given in Figures 4.44 — 4.46 for SMO, CO, and Bright-Edge 80,

respectively. Both CO and Bright-Edge 80 showed high correlation values (r > 0.95) and

the correlation was statistically significant (from t-test) at the 95% confidence level. The

model equations obtained with the non-linear regression analysis of the Yoon Nelson

model could describe the breakthrough data better for CO and Bright-Edge 80.

4.15.5 The Oulman Model

Experimental data were fitted to the Oulman model given in Equation 2.17

(Figures A.31 to A.33). Equation parameters such as K and No values are given in Table

4.20. The high correlation coefficient (r > 0.96) and the t statistical significance at the

95% confidence level indicated that the Oulman equation could describe the

breakthrough data reasonably. The Oulman equation, a re-written form of the Bohart—

Adams model can be considered a simple form of the Thomas model. When the isotherm

is highly favorable, the Thomas model is reduced to the Bohart—Adams. Bohart—Adams

model and is derived based on the assumption that adsorption is an irreversible second

order reaction. The rate constant K obtained from the Oulman model is approximately the

same as Kr , as predicted by the Thomas model.

4.15.6 The Wolborska Model

The Wolborska sorption model was applied to experimental data for the

description of the initial part of the breakthrough curve (Figures A.34 to A.36). This

approach was focused upon the estimation of characteristic parameters such as maximum

181

Nelson model are given in Figures 4.44 - 4.46 for SMO, CO, and Bright-Edge 80,

respectively. Both CO and Bright-Edge 80 showed high correlation values (r > 0.95) and

the correlation was statistically significant (from t-test) at the 95% confidence level. The

model equations obtained with the non-linear regression analysis of the Yoon Nelson

model could describe the breakthrough data better for CO and Bright-Edge 80.

4.15.5 The Oulman Model

Experimental data were fitted to the Oulman model given in Equation 2.17

(Figures A.31 to A.33). Equation parameters such as K and No values are given in Table

4.20. The high correlation coefficient (r > 0.96) and the t statistical significance at the

95% confidence level indicated that the Oulman equation could describe the

breakthrough data reasonably. The Oulman equation, a re-written form of the Bohart-

Adams model can be considered a simple form of the Thomas model. When the isotherm

is highly favorable, the Thomas model is reduced to the Bohart-Adams. Bohart-Adams

model and is derived based on the assumption that adsorption is an irreversible second

order reaction. The rate constant K obtained from the Oulman model is approximately the

same as Kt , as predicted by the Thomas model.

4.15.6 The Wolborska Model

The Wolborska sorption model was applied to experimental data for the

description of the initial part of the breakthrough curve (Figures A.34 to A.36). This

approach was focused upon the estimation of characteristic parameters such as maximum

181

0.9

0.8

0.7

0.6

9 .`-te. 0.5 • U

0.4

0.3

0.2 j 40.1 0.0

0 500 1000 1500 2000 2500 3000

Time (min)

Yoon Nelson Predicted

Experimental

Figure 4.44: Breakthrough curves for SMO predicted using Yoon and Nelson model

182

1.0

0.9

0.8

0.7

0.6

0.5

0.4

0.3

0.2 Yoon Nelson Predicted

* Experimental 0.1

0.0 500 1500 0 2500 1000 2000 3000

Time (min)

Figure 4.44: Breakthrough curves for SMO predicted using Yoon and Nelson model

182

1.0

0.9

0.8

0.7

0.3 4

0.2 4

0.1 -$

0.0 --0

4

200 400 600 800 1000

time (min)

Yoon Nelson Predicted

Experimental

1200 1400 1600

Figure 4.45: Breakthrough curves for CO predicted using Yoon and Nelson model

183

1.0

0.9

0.8

0.7

0.6

o ^ 0.5 O

0.4

0.3

Yoon Nelson Predicted 0.2

* Experimental 0.1

0.0 200 0 400 600 800 1000 1200 1400 1600

time (min)

Figure 4.45: Breakthrough curves for CO predicted using Yoon and Nelson model

183

1.0

0.9

0.8

0.7

0.6

0 Cz? 0.5

0.4

0.3

IS 0.2 i x ••••••••Yoon Nelson Predicted

0.1 x Experimental

0.0 4-- —/ , 0 200 400 600 800 1000 1200 1400 1600 1800 2000

Time (min)

Figure 4.46: Breakthrough curves for Bright-Edge 80 using Yoon and Nelson model

184

1.0

0.9

0.8

0.7

0.6

0.5

0.4

0.3

0.2 Yoon Nelson Predicted

0.1 * Experimental

0.0

0 200 400 600 800 1000 1200 1400 1600 1800 2000

Time (min) ^

Figure 4.46: Breakthrough curves for Bright-Edge 80 using Yoon and Nelson model

184

sorption capacity (No) and kinetic coefficient of the external mass transfer (pa). After

applying Equation 2.19 to the experimental data, a linear relationship between In (C/Co)

and t was obtained for In (C/Co) < —0.3, for all three breakthrough curves (r > 0.85).

Breakthrough curves for SMO and CO showed a high correlation coefficient (r > 0.9) at a

95% confidence interval. Respective values of No and pa calculated from the ln(C/Co)

versus t plots for the oils are presented in Table 4.20. This showed that the overall system

kinetics was dominated by external mass transfer in the initial part of sorption in the

column. The kinetic coefficient pa reflects the effect of both mass transfer in the liquid

phase and axial dispersion. Wolborska observed that in short beds or at high flow rates of

solution through the bed, the axial diffusion is negligible and 13a = Po, the external mass

transfer coefficient. Although the Wolborska model provides a simple and

comprehensive approach to running and evaluating sorption- column tests, its validity

was limited to the range of the used conditions. The expression of the Wolborska model

is equivalent to the Adams—Bohart relation if the coefficient KAB is equal to Pa /N0.

4.16 Column Regeneration and Reuse

Concentrations of oil in elutant for columns that had been fed SMO, CO, and

Bright-Edge 80, respectively are shown in Figures 4.47 — 4.49, respectively. The results

showed the adsorbed oils can be desorbed from the beads in the column. The desorption

produced no visible effects upon the physical properties of the beads. The re-generated

column was then fed with the same oil-in-water emulsion. The Thomas model equations

185

sorption capacity (No) and kinetic coefficient of the external mass transfer (pa). After

applying Equation 2.19 to the experimental data, a linear relationship between In (C/Co)

and t was obtained for In (C/Co) < -0.3, for all three breakthrough curves (r > 0.85).

Breakthrough curves for SMO and CO showed a high correlation coefficient (r > 0.9) at a

95% confidence interval. Respective values of No and pa calculated from the ln(C/Co)

versus t plots for the oils are presented in Table 4.20. This showed that the overall system

kinetics was dominated by external mass transfer in the initial part of sorption in the

column. The kinetic coefficient pa reflects the effect of both mass transfer in the liquid

phase and axial dispersion. Wolborska observed that in short beds or at high flow rates of

solution through the bed, the axial diffusion is negligible and Pa = p0) the external mass

transfer coefficient. Although the Wolborska model provides a simple and

comprehensive approach to running and evaluating sorption- column tests, its validity

was limited to the range of the used conditions. The expression of the Wolborska model

is equivalent to the Adams-Bohart relation if the coefficient KAB is equal to Pa /N0.

4.16 Column Regeneration and Reuse

Concentrations of oil in elutant for columns that had been fed SMO, CO, and

Bright-Edge 80, respectively are shown in Figures 4.47 - 4.49, respectively. The results

showed the adsorbed oils can be desorbed from the beads in the column. The desorption

produced no visible effects upon the physical properties of the beads. The re-generated

column was then fed with the same oil-in-water emulsion. The Thomas model equations

185

400 -

350

t 300

5 e 250 O • 200

6 • 150

(.5) 100 - 50

0

**

0 50 100 150 200 250 300 350 400 450 500

Time, min

Figure 4.47: Desorption profile for SMO using de-ionized water

186

400

350

300

„ 250 a o 13 200 -b a g 150 fi o u 100

50

0 0 50 100 150 200 250 300 350 400 450 500

Time, min

Figure 4.47: Desorption profile for SMO using de-ionized water

186

180

160

140

e 120

o 100

• 80

c.) O 60 0

40

20

0

•

•

• • # *

•

0 50 100 150 200 250 300 350

Time, min

Figure 4.48: Desorption profile for CO using de-ionized water

187

180 "

160 • , 140

"a. a 120

•

• o 100 t

1 80

a 60 -o U

40 •

20 •

N

0 0 50 100 150 200 250 300 350

Time, min

Figure 4.48: Desorption profile for CO using de-ionized water

187

Con

cent

rati

on,

160

140

120

100

80

60

40

20

0

•

50

* # #

100 150 200 250 300 350 400

Time, min

Figure 4.49: Desorption profile for Bright-Edge 80 using de-ionized water

188

160

140

120

100

80

60

40

20

0

•

50 100 150 200 250

Time, min

t * #

300 350 400

Figure 4.49: Desorption profile for Bright-Edge 80 using de-ionized water

188

and parameters kT and qo for the second run are shown in Table 4.20. The value of qo

obtained was 2.5, 0.03, and 0.5 mg/g for SMO, CO and Bright-Edge 80, respectively.

However, it may be noted that the correlation coefficient was much lower than that

obtained from the first cycle. Under the effluent pH and other similar operation

conditions, the qo value in the second cycle should not generally exceed the value in the

first cycle (Yan and Viraraghavan 2001). Nevertheless, it could be observed that the

regenerated beads retained the capability to adsorb oil. Figures 4.50 — 4.52 show the plot

of the ratio of effluent to influent oil concentration versus the bed volume for SMO, CO,

and Bright-Edge 80, respectively. As in the first cycle, the non-linear regression analysis

provided a higher correlation coefficient (r > 0.9) for all oils. Thus, based on the results, it

can be expected that the regenerated immobilized biomass can also be reused in a

biosorption column.

4.17 Coalescence/Filtration Mechanism

The head-loss across the immobilized biomass bed for single-phase flow was

predicted using Equation 2.20 and the linearized plot is presented in Figure 4.53. The plot

shows the pressure data fitted the equation very well, with an R2 value of 0.96. The

average value of the Carman-Kozeny constant (k1), determined from the slope of the

regression line, was 5.03. The value of k1, obtained for the immobilized M rouxii

biomass bed, appeared to be reasonable as similar k1 values were obtained by Carman

(1956) (4.5-5.1 for medium spheres) and Wiggins et al (1939) (4 to 6.5 for glass fibers,

189

and parameters kj and qo for the second run are shown in Table 4.20. The value of qo

obtained was 2.5, 0.03, and 0.5 mg/g for SMO, CO and Bright-Edge 80, respectively.

However, it may be noted that the correlation coefficient was much lower than that

obtained from the first cycle. Under the effluent pH and other similar operation

conditions, the qo value in the second cycle should not generally exceed the value in the

first cycle (Yan and Viraraghavan 2001). Nevertheless, it could be observed that the

regenerated beads retained the capability to adsorb oil. Figures 4.50 - 4.52 show the plot

of the ratio of effluent to influent oil concentration versus the bed volume for SMO, CO,

and Bright-Edge 80, respectively. As in the first cycle, the non-linear regression analysis

provided a higher correlation coefficient (r > 0.9) for all oils. Thus, based on the results, it

can be expected that the regenerated immobilized biomass can also be reused in a

biosorption column.

4.17 Coalescence/Filtration Mechanism

The head-loss across the immobilized biomass bed for single-phase flow was

predicted using Equation 2.20 and the linearized plot is presented in Figure 4.53. The plot

shows the pressure data fitted the equation very well, with an R2 value of 0.96. The

average value of the Carman-Kozeny constant (ki), determined from the slope of the

regression line, was 5.03. The value of ki, obtained for the immobilized M. rouxii

biomass bed, appeared to be reasonable as similar ki values were obtained by Carman

(1956) (4.5-5.1 for medium spheres) and Wiggins et al (1939) (4 to 6.5 for glass fibers,

189

1.2

4

0.8

U 0.

0.4 SMO C/Co - Experimental

0.2 —SMO C/Co - Thomas Predicted 4

0 0.5 1 1.5 2 2.5 3 3.5 4

Volume, L

Figure 4.50: Breakthrough curve for SMO for the second run

190

1.2

0.8

o

0.6 U

* SMO C/Co - Experimental 0.4

0.2 SMO C/Co - Thomas Predicted

0 0.5 1 1.5 2 2.5 3 3.5 4

Volume, L

Figure 4.50: Breakthrough curve for SMO for the second run

190

0 0.5

)1'. CO C/Co - Experimental

CO C/Co - Thomas Predicted

1 1.5

Volume, L

2

Figure 4.51: Breakthrough curve for CO for the second run

191

2.5

1.2

0.8

* CO C/Co - Experimental

0.4

CO C/Co - Thomas Predicted

0.2

0 0.5 1 1.5 2.5 2

Volume, L

Figure 4.51: Breakthrough curve for CO for the second run

191

1.2

0.8

0 U 0.6

0.4

0.2

0

Bright-Edge 80 C/Co - Experimental

Bright-Edge 80 C/Co - Thomas Predicted

0.5 1 1.5 2 2.5 3

Volume, L

Figure 4.52: Breakthrough curve for Bright-Edge 80 for the second run

192

1.2

1

0.8

* Bright-Edge 80 C/Co - Experimental 0.6

0.4

Bright-Edge 80 C/Co - Thomas

Predicted 0.2

0 0.5 1.5 2.5 0 1 2 3

Volume, L

Figure 4.52: Breakthrough curve for Bright-Edge 80 for the second run

192

0.35

0.3

0.25

0.1 1

0.05

0 4-

0 0.01 0.02 0.03 0.04 0.05 0.06 0.07

y = 5.03x - 0.02 R2 =0.96

36uy,(1-E)2/(df2gcE3)

Figure 4.53: Linearized plot of single-phase flow pressure drop

193

0.05

0

y = 5.03x- 0.02

R2 = 0.96

0 0.01 0.02 0.03 0.04 0.05 0.06 0.07

36UYc(l-£)V(df2gcE3)

Figure 4.53: Linearized plot of single-phase flow pressure drop

193

glass wool, fiber glass and yarn). The condition of laminar flow defined by Equation 2.23

with Reynolds numbers of 1.6 to 4.3 and a porosity value of 0.475 was valid. A summary

of coalescence/filtration performance and the values of k1, as reported in literature, are

shown in Table 4.22. The specific permeability coefficient (Bo) of the immobilized

biomass beads, calculated using Equation 2.23 with a k1 value of 5.03, was found to be

-8 2 2.135 x 10 m . The value of Tortuosity (T) for immobilized biomass medium with a

porosity of 0.475 was found to be 2.105. Using the value of T in Equation 2.22, the value

of ko (shape factor) was calculated as 1.13. Carman (1956) reported the ko value to be in

the range of 1.2 to 3.0 depending upon the media shape. Moazed and Viraraghavan

(2002) observed a low T value of 1.0 for an organo-clay/anthracite mixture used in their

study and attributed that it may be due to the characteristics of the mixed media. The

filtration model of Carman-Kozeny for single-phase flow, as given by Equation 2.20, was

used to predict the head-loss across the immobilized M rouxii biomass bed at different

depths. The results of the predicted and experimental head-losses across the bed for

single-phase flow (Regina tap water) are shown in Figure 4.54. The high correlation

coefficient with a R2 value of 0.99 indicated that the Carman-Kozeny filtration model can

be applied to predict the head-loss from the single-phase flow through the immobilized

M rouxii biomass bed. The data of pressure drop from the two-phase (oil-in-water

emulsion) flow through the immobilized M rouxii biomass bed was analyzed using

Equation 2.21 to evaluate the Carman-Kozeny constant, k2. The linear plot of Equation

2.21 for two-phase flow is shown in Figure 4.55. The results of predicted and

experimental head-losses across the bed for two-phase flow are shown in Figure 4.56. A

194

glass wool, fiber glass and yarn). The condition of laminar flow defined by Equation 2.23

with Reynolds numbers of 1.6 to 4.3 and a porosity value of 0.475 was valid. A summary

of coalescence/filtration performance and the values of ki, as reported in literature, are

shown in Table 4.22. The specific permeability coefficient (Bo) of the immobilized

biomass beads, calculated using Equation 2.23 with a kj value of 5.03, was found to be

-8 2

2.135 x 10 m . The value of Tortuosity (T) for immobilized biomass medium with a

porosity of 0.475 was found to be 2.105. Using the value of T in Equation 2.22, the value

of ko (shape factor) was calculated as 1.13. Carman (1956) reported the ko value to be in

the range of 1.2 to 3.0 depending upon the media shape. Moazed and Viraraghavan

(2002) observed a low T value of 1.0 for an organo-clay/anthracite mixture used in their

study and attributed that it may be due to the characteristics of the mixed media. The

filtration model of Carman-Kozeny for single-phase flow, as given by Equation 2.20, was

used to predict the head-loss across the immobilized M. rouxii biomass bed at different

depths. The results of the predicted and experimental head-losses across the bed for

single-phase flow (Regina tap water) are shown in Figure 4.54. The high correlation

coefficient with a R2 value of 0.99 indicated that the Carman-Kozeny filtration model can

be applied to predict the head-loss from the single-phase flow through the immobilized

M. rouxii biomass bed. The data of pressure drop from the two-phase (oil-in-water

emulsion) flow through the immobilized M. rouxii biomass bed was analyzed using

Equation 2.21 to evaluate the Carman-Kozeny constant, k.2. The linear plot of Equation

2.21 for two-phase flow is shown in Figure 4.55. The results of predicted and

experimental head-losses across the bed for two-phase flow are shown in Figure 4.56. A

194

Table 4.22: Summary of coalescence and filtration data

Influent characteristics

Media characteristics

Design factors

Final oil concen-tration

K1 Reference

•

Lab-prepared Fibrous glass Filter area 30 ppm 4.4 Chieu et al. emulsified coker #2 oil, 100 ppm

material, 10 mm fiber size, 75% porosity

0.018 m2, loading rate 5 m3

m2/day

(1975)

Lab-prepared Horticultural Filter area 79 126 — 3.4 Mathavan and emulsified peat, 1.6 mm cm2

, loading 130 Viraraghavan mineral oil in water, 131 mg/L

particle dia, 79% porosity

rate 0.26 — 0.51 m3

m2/day

mg/L (1992)

Lab-prepared Organo- Filter area 2.8 45 — 46 4.29 Moazed and emulsified clay/anthracite cm2, loading mg/L Viraraghavan mineral oil in water, 48 mg/L

mixture, 2.1 mm particle dia, 48.5% porosity

rate 1 — 2.7 gpm/ft2

(2002)

Lab-prepared Polypropylene 254 — 581 12.8 Li and Gu emulsified instow crude oil/ mineral oil in water, 2000, 5000 and 10000 ppm

fibres, 170 gm, 48% porosity

Nylon fibres, 80

'1111, 45%

porosity

mL/min

.367 — 813 mL/min

(2005)

Granular polypropylene, 684 iirn, porosity 38.4%

177 — 542 mL/min

9.67

Produced water 1 - 264

Granular media, 0.569 mm

Filter area 5.5 cm2, 2 — 10

13.3 — 23.2

1.5 Multon and Viraraghavan

mg/L

Produced water 2 — 212 mg/L

particle size, 56% porosity

gpm/ft2 mg/L

7.1 — 56.8 mg/L

(2006)

Lab-prepared Immobilized M Filter area 2.8 47 — 48 5.03 Present study emulsified mineral oil in water, 50 mg/L

rouxii, 2.11 mm particle dia, 47.5% porosity

cm2, loading rate 1 — 2.7 gpm/ft2

mg/L

195

Table 4.22: Sximmary of coalescence and filtration data

Influent characteristics

Media characteristics

Design factors

Final oil concentration

K, Reference

Lab-prepared emulsified coker #2 oil, 100 ppm

Fibrous glass material, 10 mm fiber size, 75% porosity

Filter area 0.018 m2, loading rate 5 m3 m2/day

30 ppm 4.4 Chieu et al. (1975)

Lab-prepared emulsified mineral oil in water, 131 mg/L

Horticultural peat, 1.6 mm particle dia, 79% porosity

Filter area 79 cm2, loading rate 0.26 -0.51 m3

m2/day

126 -130 mg/L

3.4 Mathavan and Viraraghavan (1992)


Organo-clay/anthracite mixture, 2.1 mm particle dia, 48.5% porosity

Filter area 2.8 cm2, loading rate 1 - 2.7 gpm/ft2

45 -46 mg/L

4.29 Moazed and Viraraghavan (2002)

Lab-prepared emulsified instow crude oil/ mineral oil in water, 2000, 5000 and 10000 ppm

Polypropylene fibres, 170 [im, 48% porosity

Nylon fibres, 80 Hm, 45% porosity

Granular polypropylene, 684 nm, porosity 38.4%

254 - 581 mL/min

•367 - 813 mL/min

177 - 542 mL/min

12.8

9.67

Li and Gu (2005)

Produced water 1 - 264 mg/L

Produced water 2 - 212 mg/L

Granular media, 0.569 mm particle size, 56% porosity

Filter area 5.5 c m 2 , 2 - 1 0 gpm/ft2

13.3 -23.2 mg/L

7.1 -56.8 mg/L

1.5 Multon and Viraraghavan (2006)


Immobilized M. rouxii, 2.11 mm particle dia, 47.5% porosity

Filter area 2.8 •y

cm, load ing rate 1 - 2.7 gpm/ft2

47-48 mg/L

5.03 Present study

195

0.35

0.3 -

a..3

0.25

0.2

0.15 -

6' 0 1a. •

0.05

0 0 0.05 0.1 0.15 0.2

Experimental APJL

y 1.09x + 0.0004 R2 =0.99

- - • - 45 degree line

Linear (Trendline)

0.25


196

0.3

•

0.35

0.3

>0.25 Qm | 0.2

2 0.15 13 t 01

0.05

0

y = 1.09x + 0.0004 R2 = 0.99

0.05 0.1 0.15 0.2

Experimental AP/L

45 degree line

Linear (Trendline)

0.25 0.3


196

0.9 -

0.8 -

0.7 -

0.6

0.5 -a <3 0.4 -

0.3 -

0.2

0.1 -

0

y = 2.98x + 0.04 R2 =0.86 vAk

•

0 0.02 0.04 0.06 0.08 0.1 0.12 0.14 0.16 0.18

3611,4:(1-02/(d?g,f t3)


197

0.9

0.8

0.7

0.6

< 0.5 h <1 0.4

0.3

0.2

0.1

0

y = 2.98x + 0.04

R2 = 0.86

0.02 0.04 0.06 0.08 0.1 0.12

36Uyc(l-hy/(d(%^)

0.14 0.16 0.18


197

0.6

0.5

0.4

,E•1 0.3 Q

"CJ 2

41" 0.2

0.1

0

♦ *

y = 1.05x - 0.03 R2 = 0.97 •

- • - • - 45 degree line

0.05 0.1 0.15 0.2 0.25 0.3

Experimental AWL

0.35

Linear (Trendline)

r

0.4 0.45 0.5


198

0.6

y = 1.05x - 0.03

R2 = 0.97

45 degree line 0.2

Linear

(Trendline)

—r—

0.1

—r

0.2

~f~~ """ '"T 0.25 0.3

-p. ..

0.35

""" T "'"T 0.4 0.45 0 0.05 0.15 0.5

Experimental AP/L


198

high correlation coefficient with a R2 value of 0.97 was obtained. The average value of k2,

as determined from the slope of the regression line was 2.98. In general, as far as the two-

phase flow, the head-loss predictions were closer to the measured values at lower flow

rates. As for higher flow rates, especially in the case of two-phase flow, the predicted

head-loss values were higher than the measured values at all depths of the bed. A similar

trend has been found both by Mathavan and Viraraghavan (1992) and Moazed and

Viraraghavan (2002) and the difference was attributed to be between predicted and

measured head-loss values to differences in average saturation for various flow rates.

The specific permeability coefficient (Bo) obtained for two-phase flow, using Equation

-8 2 2.23, was found to be 1.58 x 10 m . The permeability coefficient for two-phase flow did

not differ appreciably compared with the single-phase flow permeability. Both Mathavan

and Viraraghavan (1992) and Moazed and Viraraghavan (2002) observed a slight

decrease in permeability for two-phase flow due to the negligible void space occupied by

the oil droplets in the bed. The average saturation (SO was calculated using Equation

2.26. Table 4.23 provides average saturation values for the immobilized M rouxii

biomass bead bed used in this study. The degree of saturation was found to increase with

an increasing flow rate. Average saturations of 10.3% and 20.8% were obtained for flow

rates of 12 mL/min and 32 mL/min, respectively. The overall coalescence efficiency (Tic)

was calculated using Equation 2.25 for each flow rate. Results are given in Table 4.24. A

plot of coalescence efficiency versus Reynolds number for different bed depths is given

in Figure 4.57. The coalescence efficiency is plotted against Reynolds number to ensure

199

high correlation coefficient with a R2 value of 0.97 was obtained. The average value of k2,

as determined from the slope of the regression line was 2.98. In general, as far as the two-

phase flow, the head-loss predictions were closer to the measured values at lower flow

rates. As for higher flow rates, especially in the case of two-phase flow, the predicted

head-loss values were higher than the measured values at all depths of the bed. A similar

trend has been found both by Mathavan and Viraraghavan (1992) and Moazed and

Viraraghavan (2002) and the difference was attributed to be between predicted and

measured head-loss values to differences in average saturation for various flow rates.

The specific permeability coefficient (Bo) obtained for two-phase flow, using Equation

-8 2

2.23, was found to be 1.58 * 10 m . The permeability coefficient for two-phase flow did

not differ appreciably compared with the single-phase flow permeability. Both Mathavan

and Viraraghavan (1992) and Moazed and Viraraghavan (2002) observed a slight

decrease in permeability for two-phase flow due to the negligible void space occupied by

the oil droplets in the bed. The average saturation (Sd) was calculated using Equation

2.26. Table 4.23 provides average saturation values for the immobilized M. rouxii

biomass bead bed used in this study. The degree of saturation was found to increase with

an increasing flow rate. Average saturations of 10.3% and 20.8% were obtained for flow

rates of 12 mL/min and 32 mL/min, respectively. The overall coalescence efficiency (r]c)

was calculated using Equation 2.25 for each flow rate. Results are given in Table 4.24. A

plot of coalescence efficiency versus Reynolds number for different bed depths is given

in Figure 4.57. The coalescence efficiency is plotted against Reynolds number to ensure

199

Table 4.23: Average saturation values of the immobilized M rouxii biomass bed

Flow (mL/min) A P1 A P2 A P1/is P2 B Et Sd

12 0.098 0.162 0.605 0.235 0.426 0.103

16 0.125 0.219 0.568 0.198 0.41 0.137

20 0.156 0.281 0.557 0.167 0.395 0.169

24 0.201 0.332 0.609 0.157 0.388 0.182

28 0.230 0.392 0.589 0.144 0.381 0.198

32 0.276 0.454 0.613 0.136 0.376 0.208

200

Table 4.23: Average saturation values of the immobilized M. rouxii biomass bed

Flow (mL/min) API A P2 AP1/AP2 B sd

12 0.098 0.162 0.605 0.235 0.426 0.103

16 0.125 0.219 0.568 0.198 0.41 0.137

20 0.156 0.281 0.557 0.167 0.395 0.169

24 0.201 0.332 0.609 0.157 0.388 0.182

28 0.230 0.392 0.589 0.144 0.381 0.198

32 0.276 0.454 0.613 0.136 0.376 0.208

200

Table 4.24: Coalescence efficiency

Coalescence efficiency (fraction)

Bed height (mm) Flow rate (mL/min)

12 16 20 24 28 32

20 0.159 0.12 0.10 0.081 0.080 0.080

40 0.086 0.067 0.054 0.053 0.047 0.047

60 0.060 0.048 0.039 0.039 0.035 0.032

80 0.045 0.039 0.028 0.032 0.027 0.026

100 0.041 0.034 0.027 0.025 0.023 0.020

201

Table 4.24: Coalescence efficiency

Coalescence efficiency (fraction)

Bed height (mm) Flow rate (mL/min)

12 16 20 24 28 32

20 0.159 0.12 0.10 0.081 0.080 0.080

40 0.086 0.067 0.054 0.053 0.047 0.047

60 0.060 0.048 0.039 0.039 0.035 0.032

80 0.045 0.039 0.028 0.032 0.027 0.026

100 0.041 0.034 0.027 0.025 0.023 0.020

201

0200 mm 0.18 —

❑400 mm ›, 0.16 -

A600 mm -6 0.14

T)X800 mm

0.12

8 ce) a)

0.1

0.08 0 0 0

X1000 mm

U 0.06

0.04 U

0.02

0 0 0.1 0.2 0.3

log NRe

0.4 0.5 0.6


202

0.18

>> 0.16 o c 0) 0.14

0.12

o B o u ca 8 o.i CO <D K 0-08 o

0.06

0.04

0.02

0

0

•

6 0

A 1

•

0200 mm

• 400 mm

A 600 mm

X800 mm

*1000 mm

0 0 0

• •

0.1 0.2 0.3 0.4 0.5 0.6

log NRe


202

laminar flow in the system. In the laminar flow regime, the oil droplets will be considered

as moving in streamlines and wall effects will become negligible. The coalescence

efficiency was a maximum for the flow rate of 12 mL/min at all depths. The overall

coalescence efficiency ranged from 15.9 to 4% at the lowest flow rate of 12 mL/min and

from 8 to 2% at the highest flow rate of 32 mL/min. It is possible that a flow rate of 12

mL/min might be the optimum velocity to attain maximum coalescence efficiency in the

immobilized M rouxii biomass bed. However, Sherony et al. (1978) observed an increase

in coalescence efficiency with increasing velocity. Yet, it was pointed out that Sherony's

filtration theories were not very successful at predicting the capture of particulates from

water by sand beds, likely due to the underestimation of surface chemical forces in the

filtration theory (Speilman and Goren 1970). The coalescence efficiency was observed to

decrease with an increase in bed depth. Maximum and minimum coalescence efficiencies

of 15.9% and 4.0% were obtained at 200 mm and 1000 mm depth, respectively for a flow

rate of 12 mL/min. Both Mathavan and Viraraghavan (1992) and Moazed and

Viraraghavan (2002) who conducted similar studies on SMO observed decreasing

efficiency with an increasing bed depth. However, this contradicts the findings of

Sherony et al. (1971b) where the efficiency was found to increase with depth. Chieu et al.

(1975) predicted that a minimum oil saturation of 10 to 15% was necessary for

coalescence to occur in a synthetic media. In the present study, the saturation values

ranged from 10 to 20%. However, the low coalescence efficiency obtained could be

attributed to high stability of the SMO emulsion that was stabilized using emulsifiers in

the study. Mathavan and Viraraghavan (1992) concluded the peat bed used in their study

203

laminar flow in the system. In the laminar flow regime, the oil droplets will be considered

as moving in streamlines and wall effects will become negligible. The coalescence

efficiency was a maximum for the flow rate of 12 mL/min at all depths. The overall

coalescence efficiency ranged from 15.9 to 4% at the lowest flow rate of 12 mL/min and

from 8 to 2% at the highest flow rate of 32 mL/min. It is possible that a flow rate of 12

mL/min might be the optimum velocity to attain maximum coalescence efficiency in the

immobilized M. rouxii biomass bed. However, Sherony et al. (1978) observed an increase

in coalescence efficiency with increasing velocity. Yet, it was pointed out that Sherony's

filtration theories were not very successful at predicting the capture of particulates from

water by sand beds, likely due to the underestimation of surface chemical forces in the

filtration theory (Speilman and Goren 1970). The coalescence efficiency was observed to

decrease with an increase in bed depth. Maximum and minimum coalescence efficiencies

of 15.9% and 4.0% were obtained at 200 mm and 1000 mm depth, respectively for a flow

rate of 12 mL/min. Both Mathavan and Viraraghavan (1992) and Moazed and

Viraraghavan (2002) who conducted similar studies on SMO observed decreasing

efficiency with an increasing bed depth. However, this contradicts the findings of

Sherony et al. (1971b) where the efficiency was found to increase with depth. Chieu et al.

(1975) predicted that a minimum oil saturation of 10 to 15% was necessary for

coalescence to occur in a synthetic media. In the present study, the saturation values

ranged from 10 to 20%. However, the low coalescence efficiency obtained could be

attributed to high stability of the SMO emulsion that was stabilized using emulsifiers in

the study. Mathavan and Viraraghavan (1992) concluded the peat bed used in their study

203

could not filter or coalesce highly stable emulsions having droplet diameters of less than

10 gm. In the present study, droplet diameters of less than 6.9 gm were not found and the

results may agree with the findings of Mathavan and Viraraghavan (1992). The plot of

the ratio of drop diameter to immobilized M rouxii biomass bead diameter for various

flow rates at various depths is given in Figure 4.58. In the study, the initial average drop

diameter of the oil droplet was measured to be 21 gm and as the experiments proceeded,

the diameter of the oil droplets was lower than approximately 12.5 gm (range 6.9 — 12. 5

gm) at the end of the runs. The ratio of the drop diameter to biomass bead diameter

decreased as the bed depth increased from 200 mm to 1000 mm for all flow rates. The

ratio followed a decreasing trend with the bed depth for all flow rates. This decrease in

the ratio can be attributed to the filtration of drops with the biomass bed depth. The

influent drop diameter was 21 gm while the drop diameter of the effluent was less than

12.5 p.M. The decrease in the drop diameter in the effluent could be due to larger droplets

being filtered out and smaller droplets coalescing to form particles that are large enough

to be filtered out. The average holdup of the immobilized M rouxii biomass bed was

calculated using Equation 2.28. The plot of average holdup versus Reynolds number is

given in Figure 4.59. The holdup equation showed a fair agreement with experimental

data. The average holdup values of the bed increased as the Reynolds number increased.

However, Sherony et al. (1978) observed a decreasing bed holdup with an increasing

Reynolds number. It should be noted that the Reynolds numbers in their study were in the

range of 0.1 —1.0 in comparison to Reynolds numbers found in the present study (1.49 to

204

could not filter or coalesce highly stable emulsions having droplet diameters of less than

10 Jim. In the present study, droplet diameters of less than 6.9 jim were not found and the

results may agree with the findings of Mathavan and Viraraghavan (1992). The plot of

the ratio of drop diameter to immobilized M. rouxii biomass bead diameter for various

flow rates at various depths is given in Figure 4.58. In the study, the initial average drop

diameter of the oil droplet was measured to be 21 urn and as the experiments proceeded,

the diameter of the oil droplets was lower than approximately 12.5 (im (range 6.9 - 12. 5

(Am) at the end of the runs. The ratio of the drop diameter to biomass bead diameter

decreased as the bed depth increased from 200 mm to 1000 mm for all flow rates. The

ratio followed a decreasing trend with the bed depth for all flow rates. This decrease in

the ratio can be attributed to the filtration of drops with the biomass bed depth. The

influent drop diameter was 21 nm while the drop diameter of the effluent was less than

12.5 (Am. The decrease in the drop diameter in the effluent could be due to larger droplets

being filtered out and smaller droplets coalescing to form particles that are large enough

to be filtered out. The average holdup of the immobilized M. rouxii biomass bed was

calculated using Equation 2.28. The plot of average holdup versus Reynolds number is

given in Figure 4.59. The holdup equation showed a fair agreement with experimental

data. The average holdup values of the bed increased as the Reynolds number increased.

However, Sherony et al. (1978) observed a decreasing bed holdup with an increasing

Reynolds number. It should be noted that the Reynolds numbers in their study were in the

range of 0.1 - 1.0 in comparison to Reynolds numbers found in the present study (1.49 to

204

0.065 012 mL/min ❑16 mL/min

20 mL/min X24 mL/min

0.06 X28 mL/min 032 mL/min

0.055 - 0 b

7:1

0.05

0 0 A

0.045

0.04 0 20 40 60 80 100

Bed height, mm

Figure 4.58: Plot of the ratio of the drop diameter to the immobilized M rouxii biomass bead diameter versus bed depth for various flow rates

205

0.065

0.06

0.055

T3 ^5 TJ

0.05

0.045

0.04

• 0 6

20

& • 0

012 mL/min • 16 mL/min

A20 mL/min X24 mL/min

*28 mL/min 032 mL/min

0

B y

0

&

40 60

Bed height, mm

0

A

80

6

100

Figure 4.58: Plot of the ratio of the drop diameter to the immobilized M. rouxii biomass bead diameter versus bed depth for various flow rates

205

0.14

0.12

♦ 0.1 - y = 0.13x + 0.058

R2 =0.97

0,, 0.08 ♦

0 0.06

0.04 -

0.02 -

0 0 0.1 0.2 0.3 0.4 0.5 0.6

log NRe

Figure 4.59: Average holdup versus Reynolds number

206

y = 0.13x +0.058

0 0.1 0.2 0.3 0.4 0.5 0.6

log NRe

Figure 4.59: Average holdup versus Reynolds number

206

3.65). A decrease in coalescence efficiency, effluent oil concentration and effluent drop

diameter can all be attributed to the action of filtration and coalescence. Equations 2.29

and 2.30 were applied to the coalescence data to evaluate the model proposed by

Crickmore et al. (1989). The simplistic empirical model describes coalescence as a first

order rate equation. Influent and effluent oil concentrations were used in the equation as

the fraction of the fluid emulsified at the entrance and exit to the packed bed (Multon and

Viraraghavan 2006). ln(CA/CAO) versus t' was plotted to determine the rate constants for

coalescence (Figure 4.60). The coalescence data for SMO fitted the Crickmore model

reasonably well (R2 = 0.93). Rate constant (1c) from the plot was determined to be

0.0011 cm2/h. The value of the rate constant in this study appeared to be low compared to

the range of values obtained by Crickmore et al. (1989) (0.021 to 0.056 cm2/s) and

Multon and Viraraghavan (2006) (0.0032 and 0.0143 cm2/min). Additional experimental

runs with different bed depths, as carried out by Crickmore et al. (1989), may be required

to evaluate the suitability of this model in order to predict coalescence kinetics. Due to

the simplicity of this equation, it was difficult to determine whether it models

coalescence, filtration, or a combination of both (Multon and Viraraghavan 2006).

207

3.65). A decrease in coalescence efficiency, effluent oil concentration and effluent drop

diameter can all be attributed to the action of filtration and coalescence. Equations 2.29

and 2.30 were applied to the coalescence data to evaluate the model proposed by

Crickmore et al. (1989). The simplistic empirical model describes coalescence as a first

order rate equation. Influent and effluent oil concentrations were used in the equation as

the fraction of the fluid emulsified at the entrance and exit to the packed bed (Multon and

Viraraghavan 2006). ln(CA/CAO) versus T' was plotted to determine the rate constants for

coalescence (Figure 4.60). The coalescence data for SMO fitted the Crickmore model

reasonably well (R2 = 0.93). Rate constant (Kc) from the plot was determined to be

0.0011 cm2/h. The value of the rate constant in this study appeared to be low compared to

the range of values obtained by Crickmore et al. (1989) (0.021 to 0.056 cm2/s) and

Multon and Viraraghavan (2006) (0.0032 and 0.0143 cm2/min). Additional experimental

runs with different bed depths, as carried out by Crickmore et al. (1989), may be required

to evaluate the suitability of this model in order to predict coalescence kinetics. Due to

the simplicity of this equation, it was difficult to determine whether it models

coalescence, filtration, or a combination of both (Multon and Viraraghavan 2006).

207

1 rr- 1

10 20 30 40 50 60 70 80 90 100

-0.02

-0.04

-0.06 u

-0.0s

-0.1

-0.12

-0.14 hfcm2

y = -0.0011x - 0.022 Rz = 0.93

Figure 4.60: Coalescence kinetics for SMO predicted by the Crickmore model

208

0 0 I ( I f I J 1 I 1

1 10 20 30 40 50 60 70 80 90 100

-0.02 - y =-0.001 lx-0.022 Rz = 0.93

-0.04 "

§-0.06 " y • u 5 - 0 0 8 -

-0.1

-0.12 • -0.14 ^ -0.14 ^

t', h/cm2

Figure 4.60: Coalescence kinetics for SMO predicted by the Crickmore model

208

Chapter 5

Conclusion and Recommendations

5.1 Conclusion

The potential of fungus M rouxii to remove three different oils from water was

investigated via a series of batch and dynamic column studies. The following summations

were drawn based on the studies:

1. The factorial design of the experiments showed the pH was the most influential

parameter with respect to the removal of SMO, CO and Bright-Edge 80 by M

rouxii biomass. Adsorbent dose and temperature were found to have an effect on

the removal of SMO, oil concentration had an effect on the removal of CO and

temperature was found to have an effect on the removal of Bright-Edge 80. It is

important to note that these statements are valid within the lower and upper limits

of the factors: adsorbent dose (0.05 g to 0.5 g), temperature (5 to 30°C),

concentration (50 to 350 mg/L) and speed (100 to 200 rpm).

2. The results of batch adsorption studies showed M rouxii biomass could remove

up to 98% of oil from water. A decrease in the pH of solution led to a significant

increase in the oil-sorption capacity of the biomass. Adsorption of the three oils

was best at pH 3.0, the zero point charge of the biomass. Oil sorption on M rouxii

biomass was rapid with an equilibrium time of less than 3 hours. The rate of

adsorption of the three oils onto the M rouxii biomass was initially fast, followed

by a second phase characterized by slow sorption. Temperature had a relatively

minor effect as adsorption capacities of the biomass.

209

Chapter 5

Conclusion and Recommendations

5.1 Conclusion

The potential of fungus M. rouxii to remove three different oils from water was

investigated via a series of batch and dynamic column studies. The following summations

were drawn based on the studies:

1. The factorial design of the experiments showed the pH was the most influential

parameter with respect to the removal of SMO, CO and Bright-Edge 80 by M.

rouxii biomass. Adsorbent dose and temperature were found to have an effect on

the removal of SMO, oil concentration had an effect on the removal of CO and

temperature was found to have an effect on the removal of Bright-Edge 80. It is

important to note that these statements are valid within the lower and upper limits

of the factors: adsorbent dose (0.05 g to 0.5 g), temperature (5 to 30°C),

concentration (50 to 350 mg/L) and speed (100 to 200 rpm).

2. The results of batch adsorption studies showed M. rouxii biomass could remove

up to 98% of oil from water. A decrease in the pH of solution led to a significant

increase in the oil-sorption capacity of the biomass. Adsorption of the three oils

was best at pH 3.0, the zero point charge of the biomass. Oil sorption on M. rouxii

biomass was rapid with an equilibrium time of less than 3 hours. The rate of

adsorption of the three oils onto the M. rouxii biomass was initially fast, followed

by a second phase characterized by slow sorption. Temperature had a relatively

minor effect as adsorption capacities of the biomass.

209

3. Adsorption process was found to follow the pseudo second-order model. An

analysis of data showed that an infra-particle diffusion mechanism could play a

significant role in adsorption of oil by the biomass, and it was likely that

adsorption rate was controlled by a diffusion process.

4. The equilibrium data for the tested three oils were adequately described by the

Langmuir and Freundlich isotherm models. The oil sorption process was found to

be spontaneous and endothermic. Carboxyl and amide functional groups in the

biomass were likely involved in oil biosorption.

5. The Wolborska model better described the initial part of the breakthrough curve.

The Belter model reasonably predicted the experimental data. The breakthrough

curves were not described well by the modified Chu model equations. The

simulation of the whole breakthrough curve was better reflected by the Thomas,

Yan and Yoon Nelson models.

6. Both single-phase and two-phase flow data were shown to fit the Carman-Kozeny

equation. Average saturation values of the immobilized M rouxii biomass,

obtained from the Carman-Kozeny equation, increased with an increasing flow

rate. Maximum and minimum coalescence efficiencies of 15.9% and 4.0% were

obtained at 200 mm and 1000 mm depth, respectively, for a flow rate of 12

mL/min. The Crickmore model was adequate to describe the

coalescence/filtration of an oil-in-water emulsion as a first order rate equation.

210

3. Adsorption process was found to follow the pseudo second-order model. An

analysis of data showed that an intra-particle diffusion mechanism could play a

significant role in adsorption of oil by the biomass, and it was likely that

adsorption rate was controlled by a diffusion process.

4. The equilibrium data for the tested three oils were adequately described by the

Langmuir and Freundlich isotherm models. The oil sorption process was found to

be spontaneous and endothermic. Carboxyl and amide functional groups in the

biomass were likely involved in oil biosorption.

5. The Wolborska model better described the initial part of the breakthrough curve.

The Belter model reasonably predicted the experimental data. The breakthrough

curves were not described well by the modified Chu model equations. The

simulation of the whole breakthrough curve was better reflected by the Thomas,

Yan and Yoon Nelson models.

6. Both single-phase and two-phase flow data were shown to fit the Carman-Kozeny

equation. Average saturation values of the immobilized M. rouxii biomass,

obtained from the Carman-Kozeny equation, increased with an increasing flow

rate. Maximum and minimum coalescence efficiencies of 15.9% and 4.0% were

obtained at 200 mm and 1000 mm depth, respectively, for a flow rate of 12

mL/min. The Crickmore model was adequate to describe the

coalescence/filtration of an oil-in-water emulsion as a first order rate equation.

210

5.2 Practical Research Applications

By using the maximum adsorption capacity data of an adsorbent for the particular

oil-in-water emulsion, it is possible to calculate the amount of adsorbent required to

reduce oil in oily water to a desired concentration. The amount and cost of different

biomaterials used in the preliminary study, as required to obtain an effluent concentration

of 10 mg/L for SMO for a daily flow of 100 m3/d, are shown in Table 5.1. The amount of

powdered M rouxii biomass required to reduce oil from water to a desired concentration

can be determined by applying the relevant adsorption isotherm (Table 5.2). Batch

adsorption systems, which are basically operated on a fill-and-draw basis, are suitable to

treat small wastewater volumes. The CCRL, Regina, Saskatchewan, Canada treats

approximately 4700 m3/d of effluent with an average initial oil concentration of 350

mg/L and the treated effluent has an average oil concentration of 10-15 mg/L. The

wastewater treatment plant at the CCRL includes an American Petroleum Institute (API)

separator followed by an equalization pond, two aeration ponds and finally a stabilization

pond, whose effluent is finally discharged to Regina municipal sewer system. Total

Petroleum Hydrocarbon (TPH) concentration of the influent to API separator range from

24 — 558 mg/L (mean 350 mg/L), TPH concentration of the effluent from the API

separator that is fed to the stabilization range from 14 — 438 mg/L (mean 173 mg/L) and

in the final effluent from the stabilization pond, it ranges from 5 — 13 mg/L (mean 9

mg/L) (Aruldoss and Viraraghavan, 1998). A continuous adsorption system with an

immobilized fungal biomass medium (beads) may be required to treat such a large

quantity of wastewater.

211

5.2 Practical Research Applications

By using the maximum adsorption capacity data of an adsorbent for the particular

oil-in-water emulsion, it is possible to calculate the amount of adsorbent required to

reduce oil in oily water to a desired concentration. The amount and cost of different

biomaterials used in the preliminary study, as required to obtain an effluent concentration

of 10 mg/L for SMO for a daily flow of 100 m3/d, are shown in Table 5.1. The amount of

powdered M. rouxii biomass required to reduce oil from water to a desired concentration

can be determined by applying the relevant adsorption isotherm (Table 5.2). Batch

adsorption systems, which are basically operated on a fill-and-draw basis, are suitable to

treat small wastewater volumes. The CCRL, Regina, Saskatchewan, Canada treats

approximately 4700 m3/d of effluent with an average initial oil concentration of 350

mg/L and the treated effluent has an average oil concentration of 10-15 mg/L. The

wastewater treatment plant at the CCRL includes an American Petroleum Institute (API)

separator followed by an equalization pond, two aeration ponds and finally a stabilization

pond, whose effluent is finally discharged to Regina municipal sewer system. Total

Petroleum Hydrocarbon (TPH) concentration of the influent to API separator range from

24 - 558 mg/L (mean 350 mg/L), TPH concentration of the effluent from the API

separator that is fed to the stabilization range from 14 - 438 mg/L (mean 173 mg/L) and

in the final effluent from the stabilization pond, it ranges from 5-13 mg/L (mean 9

mg/L) (Aruldoss and Viraraghavan, 1998). A continuous adsorption system with an

immobilized fungal biomass medium (beads) may be required to treat such a large

quantity of wastewater.

211

Table 5.1: Cost of adsorbents required to treat 100 m3 per day of oily water based on SMO data

M rouxii A. coerulea Chitosan Walnut shell

Adsorption capacity, mg/g

77.2 72.1 99.6 82.5

Mass of adsorbent required, kg/d

246 264 191 230

Cost of adsorbent, $/kg

1 - 51 1 - 51 132 3.843

Cost of adsorbent required, $/day

246 —1231 264 —1318 2480 884

Note: Initial concentration = 200 mg/L; Final concentration = 10 mg/L Cost of specifically cultured fungi was obtained from Kuyucak 1990.

2 Cost of chitosan obtained from Chi and Cheng 2006. 3 Cost of walnut shell media provided by USFilter, USA in 2006.

212

Table 5.1: Cost of adsorbents required to treat 100 m3 per day of oily water based on SMO data

M. rouxii A. coerulea Chitosan Walnut shell

Adsorption capacity, mg/g

77.2 72.1 99.6 82.5

Mass of adsorbent required, kg/d

246 264 191 230

Cost of adsorbent, $/kg

1 - 51 1 -51 132 3.843

Cost of adsorbent required, $/day

246- 1231 264- 1318 2480 884

Note: Initial concentration = 200 mg/L; Final concentration = 10 mg/L 1 Cost of specifically cultured fiingi was obtained from Kuyucak 1990. 2 Cost of chitosan obtained from Chi and Cheng 2006. 3 Cost of walnut shell media provided by USFilter, USA in 2006.

212

Table 5.2: Design of batch adsorber system for M rouxii biomass and SMO with a flow rate of 100 m3 per day

Descriptions Calculations Isotherm model used Freundlich Isotherm equation x/m = kCA(1/n) Isotherm constants

K 40.16 1/n (n = 1.12) 0.893

Initial oil concentration, mg/L 200 Final oil concentration, mg/L 5 Percentage of oil removal 97.5 x/m value mg/g 168.99 Mass of oil removed kg/d 19.5 Mass of biomass required using x/m value (kg/d) 115.39

213

Table 5.2: Design of batch adsorber system for M. rouxii biomass and SMO with a flow rate of 100 m3 per day

Descriptions Calculations Isotherm model used Freundlich Isotherm equation x/m = kCA(l/n) Isotherm constants

K 40.16 l/n(n= 1.12) 0.893

Initial oil concentration, mg/L 200 Final oil concentration, mg/L 5 Percentage of oil removal 97.5 x/m value mg/g 168.99 Mass of oil removed kg/d 19.5 Mass of biomass required using x/m value (kg/d) 115.39

213

The design of an immobilized M rouxii biomass filter can be carried out using

kinetic rate constants and a scale-up approach. In the present study, the Thomas kinetic

model was used in the design. Details of such a design for SMO for a daily flow of 100

m3/d, using both approaches, are presented in Table 5.3. Details of design of batch and

column adsorber systems for three oils are given in Appendix B.

5.3 Contribution to the Field

So far, no research has been conducted regarding the use of microbial biomass

sorbents to remove oil from water. Until now, purely studies on the use of live biomass,

as part of biological treatment to remove oil, are available. The current research regarding

the use of non-viable fungi provides a new basis upon which to promote an understanding

of the processes involved in the biosorption of oil from water. Biomass of fungus Mucor

rouxii was chosen as a candidate for use in the study because of the presence of a high

content of chitosan in its cell wall. Chitosan was shown earlier shown to remove oil from

water. The hypothesis was that if pure chitosan effectively removed oil, then a fungal

biomass having a high chitosan content in its cell wall could be effective in removing oil

from water. Batch adsorption and column studies were conducted to test this hypothesis.

M rouxii biomass was shown to efficiently remove oil from water. Batch

adsorption studies exposed the optimum conditions required for biosorption of oil,

determined the rate and thermodynamics of adsorption and mechanisms involved in the

214

The design of an immobilized M. ronxii biomass filter can be carried out using

kinetic rate constants and a scale-up approach. In the present study, the Thomas kinetic

model was used in the design. Details of such a design for SMO for a daily flow of 100

m3/d, using both approaches, are presented in Table 5.3. Details of design of batch and

column adsorber systems for three oils are given in Appendix B.

5.3 Contribution to the Field

So far, no research has been conducted regarding the use of microbial biomass

sorbents to remove oil from water. Until now, purely studies on the use of live biomass,

as part of biological treatment to remove oil, are available. The current research regarding

the use of non-viable fungi provides a new basis upon which to promote an understanding

of the processes involved in the biosorption of oil from water. Biomass of fungus Mucor

rouxii was chosen as a candidate for use in the study because of the presence of a high

content of chitosan in its cell wall. Chitosan was shown earlier shown to remove oil from

water. The hypothesis was that if pure chitosan effectively removed oil, then a fungal

biomass having a high chitosan content in its cell wall could be effective in removing oil

from water. Batch adsorption and column studies were conducted to test this hypothesis.

M. rouxii biomass was shown to efficiently remove oil from water. Batch

adsorption studies exposed the optimum conditions required for biosorption of oil,

determined the rate and thermodynamics of adsorption and mechanisms involved in the

214

Table 5.3: Design of immobilized M rouxii filter for SMO with a flow rate of 100 m3 per day

Inside diameter of column, mm Length of bed, mm Mass of biomass in the column, kg Packed density of biomass in the column, kg/m3

By scale-up approach: Unit liquid flow rate, L/h Bed volume, L Bed volume per unit time, BV/h Design bed volume, m3Mass of immobilized biomass required, kg From breakthrough curve for allowable 10 mg/L oil concentration, the corresponding volume, L Emulsion treated per kg of biomass, L/kg Biomass exhausted per hour, kg/h Breakthrough time d Breakthrough volume m3

By kinetic approach: KT, L/h-kg

go, kg/kg Co, mg/L C, mg/L Flow, m3/d Applying Thomas equation and solving for M, kg/h

12.70 300.00 0.0045

118.50

0.16 0.13 1.26 3.31 391.91

0.45

100 41.67 0.39 39.19

4200

0.0098 50 10 100

422.4

215

Table 5.3: Design of immobilized M. rouxii filter for SMO with a flow rate of 100 m3 per day

Inside diameter of column, mm Length of bed, mm Mass of biomass in the column, kg Packed density of biomass in the column, kg/m3

By scale-up approach: Unit liquid flow rate, L/h Bed volume, L Bed volume per unit time, BV/h Design bed volume, m3

Mass of immobilized biomass required, kg From breakthrough curve for allowable 10 mg/L oil concentration, the corresponding volume, L Emulsion treated per kg of biomass, L/kg Biomass exhausted per hour, kg/h Breakthrough time d Breakthrough volume m3

By kinetic approach: Kt, L/h-kg

q0, kg/kg Co, mg/L C, mg/L Flow, m3/d Applying Thomas equation and solving for M, kg/h

300.00 0.0045

118.50

391.91

41.67

0.0098

422.4

215

process. The research was further supported by a detailed investigation regarding the role

of various functional groups present on the M rouxii biomass towards oil biosorption.

The aforementioned analyses produced a very unique study and contributed to the

fundamental understanding of oil biosorption by microbial sorbents.

Column breakthrough studies predicted the concentration-time profile for the

effluent, which will ultimately assist in the successful design of a column adsorption

process. The powdered biomass was immobilized to ensure its use as a packing medium

in a fixed bed column. Investigations with respect to the breakdown mechanisms and

flow characteristics involved in an immobilized biomass bed evaluated the applicability

of the well-known Carman-Kozeny filtration equation and coalescence efficiency. The

column studies, involving immobilized biomass, lend themselves to be used as a medium

in a filtration mode.

This work is very original and unique as it provides an important foundation for

studies on biosorption of oil by fungal sorbents, focuses on understanding adsorption

process mechanisms and highlights the use of biomass in fixed bed adsorption and

filtration applications.

5.4 Recommendations for Further Study

The following topics are recommended for further study:

1. Examine the removal of other pollutants such as heavy metals and odor producing

compounds that are typically found in oil containing wastewaters by M rouxii

biomass;

216

process. The research was further supported by a detailed investigation regarding the role

of various functional groups present on the M. rouxii biomass towards oil biosorption.

The aforementioned analyses produced a very unique study and contributed to the

fundamental understanding of oil biosorption by microbial sorbents.

Column breakthrough studies predicted the concentration-time profile for the

effluent, which will ultimately assist in the successful design of a column adsorption

process. The powdered biomass was immobilized to ensure its use as a packing medium

in a fixed bed column. Investigations with respect to the breakdown mechanisms and

flow characteristics involved in an immobilized biomass bed evaluated the applicability

of the well-known Carman-Kozeny filtration equation and coalescence efficiency. The

column studies, involving immobilized biomass, lend themselves to be used as a medium

in a filtration mode.

This work is very original and unique as it provides an important foundation for

studies on biosorption of oil by fungal sorbents, focuses on understanding adsorption

process mechanisms and highlights the use of biomass in fixed bed adsorption and

filtration applications.

5.4 Recommendations for Further Study

The following topics are recommended for further study:

1. Examine the removal of other pollutants such as heavy metals and odor producing

compounds that are typically found in oil containing wastewaters by M. rouxii

biomass;

216

2. Conduct experiments to evaluate the potential of immobilized M rouxii biomass

in removing oil from various industrial effluents such as refineries, food

processing, and metalworking units as only synthetically prepared oil-in-water

emulsions were considered in the present study;

3. Conduct experiments to study the possibility of a combined treatment scheme for

the effective removal of oil by powdered biosorbents with subsequent flotation.

4. Conduct experiments to evaluate the effect of surfactants on oil removal

efficiency;

5. Conduct experiments to evaluate the effect of bed length and particle size on oil

removal efficiency in a packed bed column containing immobilized M rouxii

biomass;

6. Establish an association between functional groups present in the fungal cell wall

and different chemical structures and compositions of oil molecules on

biosorption of oil by fungal biomass;

7. Investigate the effects of DMF on chemical modifications of fungal biomass

during the bead production process; and

8. Investigate the disposal options available for spent biomass such as landfilling in

order to understand the long-term environmental effects due to possible

desorption and migration.

217

2. Conduct experiments to evaluate the potential of immobilized M. rouxii biomass

in removing oil from various industrial effluents such as refineries, food

processing, and metalworking units as only synthetically prepared oil-in-water

emulsions were considered in the present study;

3. Conduct experiments to study the possibility of a combined treatment scheme for

the effective removal of oil by powdered biosorbents with subsequent flotation.

4. Conduct experiments to evaluate the effect of surfactants on oil removal

efficiency;

5. Conduct experiments to evaluate the effect of bed length and particle size on oil

removal efficiency in a packed bed column containing immobilized M. rouxii

biomass;

6. Establish an association between functional groups present in the fungal cell wall

and different chemical structures and compositions of oil molecules on

biosorption of oil by fungal biomass;

7. Investigate the effects of DMF on chemical modifications of fungal biomass

during the bead production process; and

8. Investigate the disposal options available for spent biomass such as landfilling in

order to understand the long-term environmental effects due to possible

desorption and migration.

217

References

Achak, M., Hafid, A., Ouazzani, N., Sayad, S. and Mandi, L. (2009). Low cost biosorbent

`banana peel' for the removal of phenolic compounds from olive oil mill

wastewater: kinetic and equilibrium studies. Journal of Hazardous Materials, 166,

117-125.

AEEP (Association of Environmental Engineering Professors) (1988). Environmental

engineering unit operations and unit processes laboratory manual. M. T. Suidan,

Editor, Department of Civil Engineering, University of Illinois at Urbana-

Champaign, USA.

Aggelis, G., Gavala, H. N. and Lyberatos, G. (2001). Combined and separate arobic and

anaerobic biotreatment of green olive debittering wastewater. Journal of

Agricultural Engineering Research, 80(3), 283-292.

Aggelis, G., Iconomou, D., Christou, M., Bokas, 0., Kotzailias, S., Christou, G., Tsagou,

V. and Papanikolaou, S. (2003). Phenolic removal in a model olive oil mill

wastewater using Pleurotus ostreatus in bioreactor cultures and biological

evaluation of the process. Water Research, 37(16), 3897-3904.

Ahmad, A.L., Sumathi, S. and Hameed, B.H. (2005a). Adsorption of residue oil from

palm oil mill effluent using powder and flake chitosan: equilibrium and kinetic

studies. Water Research, 39, 2483-2494.

Ahmad, A.L., Sumathi, S. and Hameed, B.H. (2005b). Residual oil and suspended solid

removal using natural adsorbents chitosan, bentonite and activated carbon: a

comparative study. Chemical Engineering Journal, 108, 179-185.

218

References

Achak, M., Hafid, A., Ouazzani, N., Sayad, S. and Mandi, L. (2009). Low cost biosorbent

'banana peel' for the removal of phenolic compounds from olive oil mill

wastewater: kinetic and equilibrium studies. Journal of Hazardous Materials, 166,

117-125.

AEEP (Association of Environmental Engineering Professors) (1988). Environmental

engineering unit operations and unit processes laboratory manual. M. T. Suidan,

Editor, Department of Civil Engineering, University of Illinois at Urbana-

Champaign, USA.

Aggelis, G., Gavala, H. N. and Lyberatos, G. (2001). Combined and separate arobic and

anaerobic biotreatment of green olive debittering wastewater. Journal of

Agricultural Engineering Research, 80(3), 283-292.

Aggelis, G., Iconomou, D., Christou, M., Bokas, O., Kotzailias, S., Christou, G., Tsagou,

V. and Papanikolaou, S. (2003). Phenolic removal in a model olive oil mill

wastewater using Pleurotus ostreatus in bioreactor cultures and biological

evaluation of the process. Water Research, 37(16), 3897-3904.

Ahmad, A.L., Sumathi, S. and Hameed, B.H. (2005a). Adsorption of residue oil from

palm oil mill effluent using powder and flake chitosan: equilibrium and kinetic

studies. Water Research, 39, 2483-2494.

Ahmad, A.L., Sumathi, S. and Hameed, B.H. (2005b). Residual oil and suspended solid

removal using natural adsorbents chitosan, bentonite and activated carbon: a

comparative study. Chemical Engineering Journal, 108, 179-185.

218

Ahmadi, M., Vahabzadeh, F., Bonakdarpour, B.Mehranian, M. and Mofarrah, E. (2006).

Phenolic removal in olive oil mill wastewater using loofah-immobilized

Phanerochaete chrysosporium. World Journal of Microbiology and Biotechnology,

22(2), 119-127.

Akhtar, M. W., Mirza, A. Q. and Chuhati, M. I. D. (1980). Lipase production in Mucor

hiemalis. Applied and Environmental Microbiology, 40(2), 257-263.

Aksu, Z. and Gonen, F. (2004). Biosorption of phenol by immobilized activated sludge in

a continuous packed bed: prediction of breakthrough curves. Process Biochemistry,

39, 599-613.

Al-Shamrani, A.A., James, A. and Xiao, H. (2002) Separation of oil from water by

dissolved air flotation. Colloids and Surfaces A: Physicochemical and Engineering

Aspects, 209, 15-26.

Alther, G.R. (1995). Organically modified clay removes oil from water. Waste

Management, 15(8), 623-628.

Alther, G.R. (1996). Organoclays lead the way to zero discharge. Environmental

Solutions, 9(8), 22 — 25.

Amorim' R., de Souza, W., Fukushima, K. and de Campos-Takaki, G. (2001). Faster

chitosan production by mucoralean strains in submerged culture. Brazilian Journal

of Microbiology, 32 (1), 20-23.

Antony, J. (2003). Design of Experiments for Engineers and Scientists. Butterworth-

Heinemann, New York, USA.

219

Ahmadi, M., Vahabzadeh, F., Bonakdarpour, B.Mehranian, M. and Mofarrah, E. (2006).

Phenolic removal in olive oil mill wastewater using loofah-immobilized

Phanerochaete chrysosporium. World Journal of Microbiology and Biotechnology,

22(2), 119-127.

Akhtar, M. W., Mirza, A. Q. and Chuhati, M. I. D. (1980). Lipase production in Mucor

hiemalis. Applied and Environmental Microbiology, 40(2), 257-263.

Aksu, Z. and Gonen, F. (2004). Biosorption of phenol by immobilized activated sludge in

a continuous packed bed: prediction of breakthrough curves. Process Biochemistry,

39, 599-613.

Al-Shamrani, A.A., James, A. and Xiao, H. (2002) Separation of oil from water by

dissolved air flotation. Colloids and Surfaces A: Physicochemical and Engineering

Aspects, 209, 15-26.

Alther, G.R. (1995). Organically modified clay removes oil from water. Waste

Management, 15(8), 623-628.

Alther, G.R. (1996). Organoclays lead the way to zero discharge. Environmental

Solutions, 9(8), 22 - 25.

Amorim' R., de Souza, W., Fukushima, K. and de Campos-Takaki, G. (2001). Faster

chitosan production by mucoralean strains in submerged culture. Brazilian Journal

of Microbiology, 32 (1), 20-23.

Antony, J. (2003). Design of Experiments for Engineers and Scientists. Butterworth-

Heinemann, New York, USA.

219

API (American Petroleum Institute) (1969). Manual on disposal of refinery wastes.

American Petroleum Institute, Washington, DC.

Arcidiacono, S. and Kaplan, L.D. (2004). Molecular weight distribution of chitosan

isolated from Mrouxii under different culture and processing conditions.

Biotechnology and Bioengineering, 39(3), 281-286.

Arcidiacono, S., Lombardi, S.J. and Kaplan, L.D. (1988). Fermentation, processing and

enzyme charecterization for chitosan biosynthesis — by Mucor rouxii. Proceedings

from the 4th International Conference on Chitin and Chitosan, Norway, August 22-

24.

Asano, T., Tsuchihashi, R. and Metcalf and Eddy Inc. (2006). Water reuse: issue,

technologies and applications. Pt edition, McGraw Hill, USA.

Aserin, A. (2007). Multiple emulsions: technology and applications. Wiley-Interscience,

Hoboken, New Jersey, USA.

ASTM (American Society for Testing and Materials) (1963). Standard test method for

finding viscosity of fluids, ASTM-D445. Philadelphia, USA.


pH of soils, ASTM-D4972-89. Philadelphia, USA.


laboratory determination of water (moisture) content of soil and rock, ASTM

D2216-92. Philadelphia, USA.

Ayorinde, F.O., Garvin, K. and saeed, K. (2000). Determination of the fatty acid

composition of saponified vegetable oils using matrix-assisted laser

220

API (American Petroleum Institute) (1969). Manual on disposal of refinery wastes.

American Petroleum Institute, Washington, DC.

Arcidiacono, S. and Kaplan, L.D. (2004). Molecular weight distribution of chitosan

isolated from M.rouxii under different culture and processing conditions.

Biotechnology andBioengineering, 39(3), 281-286.

Arcidiacono, S., Lombardi, S.J. and Kaplan, L.D. (1988). Fermentation, processing and

enzyme charecterization for chitosan biosynthesis - by Mucor rouxii. Proceedings

from the 4th International Conference on Chitin and Chitosan, Norway, August 22-

24.

Asano, T., Tsuchihashi, R. and Metcalf and Eddy Inc. (2006). Water reuse: issue,

technologies and applications. 1st edition, McGraw Hill, USA.

Aserin, A. (2007). Multiple emulsions: technology and applications. Wiley-Interscience,

Hoboken, New Jersey, USA.


finding viscosity of fluids, ASTM-D445. Philadelphia, USA.


pH of soils, ASTM-D4972-89. Philadelphia, USA.


laboratory determination of water (moisture) content of soil and rock, ASTM

D2216-92. Philadelphia, USA.

Ayorinde, F.O., Garvin, K. and saeed, K. (2000). Determination of the fatty acid

composition of saponified vegetable oils using matrix-assisted laser

220

desorption/ionization time-of-flight mass spectroscopy. Rapid Communications in

Mass Spectroscopy, 14(7), 600-615.

Azbar, N. and Yonar, T. (2004). Comparative evaluation of a laboratory and full-scale

treatment alternatives for the vegetable oil refining industry wastewater (VORW).

Process Biochemistry, 39, 869-875.

Azizian, S. (2004). Kinetic models of sorption: a theoretical analysis. Journal of Colloid

and Interface Science, 276, 47-52.

Balcao, V.M., Paiva, A.L. and Malcata, F.X. (1996). Bioreactors with immobilized

lipases - State of the art. Enzyme and Microbial Technology, 18(6), 392-416.

Banitez F.J., Acero, J.L., Gonzalez, T. and Garcia, J. (2001). Organic matter removal

from wastewaters of the black olive industry by chemical and biological

procedures. Process Biochemistry, 37(3), 257-265.

Barker, T.W. and Worgan J.T. (1981). The utilization of palm oil processing effluents as

substrates for microbial protein production by the fungus Aspergillus oryzae.

Applied Microbiology and Biotechnology, 11(4), 234-240.

Bartnicki-Garcia, S. and Nickerson, W.Y. (1962). Isolation, composition and structure of

cell walls of filamentous and yeast-like forms of Mucor rouxii. Biochemistry and

Biophysics Acta, 58, 102-119.

Bastow, T., Durnie, W.H., Jefferson, A. and Pang, J. (1997). Ultraviolet spectroscopy for

the analysis of oil-in-water effluent using isopropanol as co-solvent. Applied

Spectroscopy, 51, 318-322.

Bayab, A., Aghamiri, S.F., Moheb, A. and Vakili-Nezhrad, R. (2005). Oil spill cleanup

221

desorption/ionization time-of-flight mass spectroscopy. Rapid Communications in

Mass Spectroscopy, 14(7), 600-615.

Azbar, N. and Yonar, T. (2004). Comparative evaluation of a laboratory and full-scale

treatment alternatives for the vegetable oil refining industry wastewater (VORW).

Process Biochemistry, 39, 869-875.

Azizian, S. (2004). Kinetic models of sorption: a theoretical analysis. Journal of Colloid

and Interface Science, 276,47-52.

Balcao, V.M., Paiva, A.L. and Malcata, F.X. (1996). Bioreactors with immobilized

lipases - State of the art. Enzyme and Microbial Technology, 18(6), 392-416.

Banitez F.J., Acero, J.L., Gonzalez, T. and Garcia, J. (2001). Organic matter removal

from wastewaters of the black olive industry by chemical and biological

procedures. Process Biochemistry, 37(3), 257-265.

Barker, T.W. and Worgan J.T. (1981). The utilization of palm oil processing effluents as

substrates for microbial protein production by the fungus Aspergillus oryzae.

Applied Microbiology and Biotechnology, 11(4), 234-240.

Bartnicki-Garcia, S. and Nickerson, W.Y. (1962). Isolation, composition and structure of

cell walls of filamentous and yeast-like forms of Mucor rouxii. Biochemistry and

Biophysics Acta, 58, 102-119.

Bastow, T., Dumie, W.H., Jefferson, A. and Pang, J. (1997). Ultraviolet spectroscopy for

the analysis of oil-in-water effluent using isopropanol as co-solvent. Applied

Spectroscopy, 51, 318-322.

Bayab, A., Aghamiri, S.F., Moheb, A. and Vakili-Nezhrad, R. (2005). Oil spill cleanup

221

from seawater by sorbent materials. Chemical Engineering Technology, 28 (12),

1525 —1528.

Bear, J. (1972). Dynamics of Fluids in Porous Media. American Elsevier Publishing

Company, Inc., New York, USA.

Beccari, M., Bonemazzi, F., Majone, M. and Ricardi, C. (1996). Interaction between

acidogenesis and methanogenesis in the anaerobic treatment of olive oil mill

effluents. Water Research, 30(1), 183-189.

Beccari, M., Majone, M. and Torrisi, L. (1998). Two-reactor system with partial phase

separation for anaerobic treatment of oil mill effluents. Water Science and

Technology, 38(4-5), 53-60.

Becher, P. (1977). Emulsions theory and practice. R.E. Krieger Publishing Company,

New York, USA.

Becker, P., Koster, D., Popov, M. N., Markossian, S., Antranikian, G. and Markl H.

(1999). The biodegradation of olive oil and the treatment of lipid-rich wool

scouring wastewater under aerobic thermophilic conditions. Water Research, 33(3),

653-660.

Belter, P.A., Cussler, E.L. and Hu, W-S. (1998). Bioseparations: Downstream

Processing for Biotechnology. Wiley, New York, USA.

Benefield, L.D., Judkins, J.F. and Weand, B.L. (1982). Process biochemistry for water

and wastewater treatment. Prentice-Hall Inc.: New Jersey, USA.

Biswas, N. (1973). Electrochemical Treatment of Oil Emulsions. MASc Thesis.

Department of Civil Engineering, University of Ottawa, Ottawa, Canada.

222

from seawater by sorbent materials. Chemical Engineering Technology, 28 (12),

1525- 1528.

Bear, J. (1972). Dynamics of Fluids in Porous Media. American Elsevier Publishing

Company, Inc., New York, USA.

Beccari, M., Bonemazzi, F., Majone, M. and Ricardi, C. (1996). Interaction between

acidogenesis and methanogenesis in the anaerobic treatment of olive oil mill

effluents. Water Research, 30(1), 183-189.

Beccari, M., Majone, M. and Torrisi, L. (1998). Two-reactor system with partial phase

separation for anaerobic treatment of oil mill effluents. Water Science and

Technology, 38(4-5), 53-60.

Becher, P. (1977). Emulsions theory and practice. R.E. Krieger Publishing Company,

New York, USA.

Becker, P., Koster, D., Popov, M. N., Markossian, S., Antranikian, G. and Markl H.

(1999). The biodegradation of olive oil and the treatment of lipid-rich wool

scouring wastewater under aerobic thermophilic conditions. Water Research, 33(3),

653-660.

Belter, P.A., Cussler, E.L. and Hu, W-S. (1998). Bioseparations: Downstream

Processing for Biotechnology. Wiley, New York, USA.

Benefield, L.D., Judkins, J.F. and Weand, B.L. (1982). Process biochemistry for water

and wastewater treatment. Prentice-Hall Inc.: New Jersey, USA.

Biswas, N. (1973). Electrochemical Treatment of Oil Emulsions. MASc Thesis.

Department of Civil Engineering, University of Ottawa, Ottawa, Canada.

222

Blumenschein, C.D., Severing, K.W. and Boyle, E. (2001). Walnut shell filtration for oil

and solids. AISE Steel Technology (USA), 78 (4), 33-37.

Bohart, G. and Adams, E.N. (1920). Some aspects of the behavior of charcoal with

respect to chlorine. Journal of the American Chemical Society, 42, 523-544.

Borba, E. A., da Silva, E. A., Fagundes-Klen, M. R., Kroumov, A. D. and Guirardello, R.

(2008). Prediction of the copper (II) ions dynamic removal from a medium by using

mathematical models with analytical solution. Journal of Hazardous Materials,

152(1), 366 — 372.

Borja, R., Banks, C.J., Khalaoui, B. and Martin, A. (1996). Performance evaluation of an

anaerobic hybrid digester treating palm oil mill effluent. Journal of Environmental

Science and Health, A31(6), 1379-1393.

Braden, M.L. (1994). Demulsification of oily waste waters. In: WPCF 64th Annual

Conference & Exposition Toronto, AC91-061-005, WPCF, 601 Wythe St.,

Alexandria, VA, USA.

Cammarota, M.C. and Freire, D.M.G. (2006). A review on hydrolytic enzymes in the

treatment of wastewater with high oil and grease content. Bioresource Technology,

97(17), 2195-2210.

Cammarota, M.C., Teixeira, G.A. and Freire, D.M.G. (2001). Enzymatic pre-hydrolysis

and anaerobic degradation of wastewaters with high fat content. Biotechnology

Letters, 23(19), 1591-1595.

Carberry, J.J. (1976). Chemical and catalytic reaction engineering. McGraw Hill, New

York, USA.

223

Blumenschein, C.D., Severing, K.W. and Boyle, E. (2001). Walnut shell filtration for oil

and solids. AISE Steel Technology (USA), 78 (4), 33-37.

Bohart, G. and Adams, E.N. (1920). Some aspects of the behavior of charcoal with

respect to chlorine. Journal of the American Chemical Society, 42, 523-544.

Borba, E. A., da Silva, E. A., Fagundes-Klen, M. R., Kroumov, A. D. and Guirardello, R.

(2008). Prediction of the copper (II) ions dynamic removal from a medium by using

mathematical models with analytical solution. Journal of Hazardous Materials,

152(1), 366-372.

Boija, R., Banks, C.J., Khalaoui, B. and Martin, A. (1996). Performance evaluation of an

anaerobic hybrid digester treating palm oil mill effluent. Journal of Environmental

Science and Health, A31(6), 1379-1393.

Braden, M.L. (1994). Demulsification of oily waste waters. In: WPCF 64th Annual

Conference & Exposition Toronto, AC91-061-005, WPCF, 601 Wythe St.,

Alexandria, VA, USA.

Cammarota, M.C. and Freire, D.M.G. (2006). A review on hydrolytic enzymes in the

treatment of wastewater with high oil and grease content. Bioresource Technology,

97(17),2195-2210.

Cammarota, M.C., Teixeira, G.A. and Freire, D.M.G. (2001). Enzymatic pre-hydrolysis

and anaerobic degradation of wastewaters with high fat content. Biotechnology

Letters, 23(19), 1591-1595.

Carberry, J.J. (1976). Chemical and catalytic reaction engineering. McGraw Hill, New

York, USA.

223

Carman, P. C. (1956). Flow of gases through porous media, Academic Press, Inc.: New

York, USA.

Chang, I.S., Chung, C.M. and Han, S.H. (2001). Treatment of oily wastewater by

ultrafiltration and ozone. Desalination, 133(3), 139-144.

Chao, A.C. and Yang, W. (1981). Treatment of wool scouring wastewater. Journal Water

Pollution Control Federation, 53(3), 311-317.

Chi, F.H. and Cheng, W.P. (2006). Use of chitosan as coagulant to treat wastewater from

milk processing plant. Journal of Polymers and the Environment, 14,411-417.

Chieu, J.N., Schechter, R.S., Humenick, M.J. and Gloyna, E.F. (1975). Coalescence of

emulsified wastes by fibrous bed. Technical report EHE-75-05, CRWR-126,

Department of Civil Engineering, University of Texas at Austin, USA.

Chopra, A. and Khuller, G.K. (1983). Lipids of pathogenic fungi. Progress in Lipid

Research, 22(3), 189-220.

Chu, K. (2004). Improved fixed bed models for metal biosorption. Chemical Engineering

Journal, 97,233-239.

Cooney, D.O. (1999). Adsorption Design for Wastewater Treatment. CRC Press, Boca

Raton.

Crickmore, P. J., Veljkovic, M. and Cooke, E.D. (1989). Breaking naphtha/water and

water/naphtha emulsions with a packed bed coalesce. AOSTRA Journal of Research,

5,391-396.

224

Carman, P. C. (1956). Flow of gases through porous media, Academic Press, Inc.: New

York, USA.

Chang, I.S., Chung, C.M. and Han, S.H. (2001). Treatment of oily wastewater by

ultrafiltration and ozone. Desalination, 133(3), 139-144.

Chao, A.C. and Yang, W. (1981). Treatment of wool scouring wastewater. Journal Water

Pollution Control Federation, 53(3), 311-317.

Chi, F.H. and Cheng, W.P. (2006). Use of chitosan as coagulant to treat wastewater from

milk processing plant. Journal of Polymers and the Environment, 14,411-417.

Chieu, J.N., Schechter, R.S., Humenick, M.J. and Gloyna, E.F. (1975). Coalescence of

emulsified wastes by fibrous bed. Technical report EHE-75-05, CRWR-126,

Department of Civil Engineering, University of Texas at Austin, USA.

Chopra, A. and Khuller, G.K. (1983). Lipids of pathogenic fungi. Progress in Lipid

Research, 22(3), 189-220.

Chu, K. (2004). Improved fixed bed models for metal biosorption. Chemical Engineering

Journal, 97, 233-239.

Cooney, D.O. (1999). Adsorption Design for Wastewater Treatment. CRC Press, Boca

Raton.

Crickmore, P. J., Veljkovic, M. and Cooke, E.D. (1989). Breaking naphtha/water and

water/naphtha emulsions with a packed bed coalesce. AOSTRA Journal of Research,

5, 391-396.

224

Crittenden, B.D. and Sowerby, B. (1990). Scale-up of vapor phase adsorption columns

for breaking the ethanol-water azeotrope. I.Chem.E Symposium Series No. 118,

South Wales Branch, Great Britian.

Crittenden, J.C., Reddy, P.S., Arora, H., Trynoski, J., Hand, D.W., Perram, D.L. and

Summers, R.S. (1991). Prediction of GAC performance with rapid small-scale

column tests. Journal of American Water Works Association, 83.

Davoust, N. and Persson, A. (1992). Effect of growth morphology and time of harvesting

on the chitosan yield of Absidia repens. Applied Microbiology and Biotechnology, 37,

572-575.

Deschamps, G., Camel, H., Borredon, M.E., Bonnin, C. and Vignoles, C. (2003). Oil

removal from water by selective sorption on hydrophobic cotton fibres. 1. Study of

sorption properties and comparison with other cotton fibre-based sorbents.

Environmental Science and Technology, 37,1013-1015.

Dharmsthiti, S. and Kuhasntisuk, B. (1998). Lipase from Pseudomonas aeruginosa

LP602: Biochemical properties and application for wastewater treatment. Journal

of Industrial Microbiology and Biotechnology, 21(1-2), 75-80.

Drake, L.R., Lin, S., Rayson, G.D. and Jackson, P.J. (1996). Chemical modifications and

metal binding sites of Datura innoxia. Environmental Science and Technology, 30,

110-116.

El-Bestawy, E., El-Masry, M.H. and El-Adl, N.E. (2005). The potentiality of free Gram-

negative bacteria for removing oil and grease from contaminated industrial

effluents. World Journal of Microbiology and Biotechnology, 21(6-7), 815-822.

225

Crittenden, B.D. and Sowerby, B. (1990). Scale-up of vapor phase adsorption columns

for breaking the ethanol-water azeotrope. LChem.E Symposium Series No. 118,

South Wales Branch, Great Britian.

Crittenden, J.C., Reddy, P.S., Arora, H., Trynoski, J., Hand, D.W., Perram, D.L. and

Summers, R.S. (1991). Prediction of GAC performance with rapid small-scale

column tests. Journal of American Water Works Association, 83.

Davoust, N. and Persson, A. (1992). Effect of growth morphology and time of harvesting

on the chitosan yield of Absidia repens. Applied Microbiology and Biotechnology, 37,

572-575.

Deschamps, G., Caruel, H., Borredon, M.E., Bonnin, C. and Vignoles, C. (2003). Oil

removal from water by selective sorption on hydrophobic cotton fibres. 1. Study of

sorption properties and comparison with other cotton fibre-based sorbents.

Environmental Science and Technology, 37, 1013-1015.

Dharmsthiti, S. and Kuhasntisuk, B. (1998). Lipase from Pseudomonas aeruginosa

LP602: Biochemical properties and application for wastewater treatment. Journal

of Industrial Microbiology and Biotechnology, 21(1-2), 75-80.

Drake, L.R., Lin, S., Rayson, G.D. and Jackson, P.J. (1996). Chemical modifications and

metal binding sites of Datura innoxia. Environmental Science and Technology, 30,

110-116.

El-Bestawy, E., El-Masry, M.H. and El-Adl, N.E. (2005). The potentiality of free Gram-

negative bacteria for removing oil and grease from contaminated industrial

effluents. World Journal of Microbiology and Biotechnology, 21(6-7), 815-822.

225

El-Masry, M.H., El-Bestawy, E. and El-Adl, N. (2004). Bioremediation of vegetable oil

and grease from polluted wastewater using a sand biofilm system. World Journal of

Microbiology and Biotechnology, 20(6), 551-557.

Environment Canada (1976). Guidelines for Effluent Quality and Wastewater Treatment

at Federal Establishments. EPS 1-EC-76-1, Canada.

Environment Canada (1985). Selection criteria and laboratory evaluation of oil spill

sorbent, update II, EPS report — EPS4-EP-83-4, Canada.

European Committee for Standardization (2002). Separator systems for light liquids (e.g.

oil and petrol) - Part 1: Principles of product design, performance and testing,

marking and quality control. 93/38/EEC (No) 89/106/EEC (C 319, 2005-12-14).

Fabritius, D., Schafer, H.J. and Steinbuchel, A. (1998). Bioconversion of sunflower oil,

rapeseed oil and ricinoleic acid by Candida tropicalis M25. Applied Microbiology

Biotechnology, 50(5), 573-578.

Fadil, K., Chahlaoui, A., Ouahbi, A., Zaid, A. and Borja, R. (2003). Aerobic

biodegradation and detoxification of wastewaters from the olive oil industry.

International Biodeterioration & Biodegradation, 51 (1), 37-41.

Faisal, M. and Unno, H. (2001). Kinetic analysis of palm oil mill wastewater treatment

by a modified anaerobic baffled reactor. Biochemical Engineering Journal, 9(1),

25-31.

Filidei, S., Masuandaro, G. and Ceccanti, B. (2003). Anaerobic digestion of olive oil mill

effluents: Evaluation of wastewater organic load and phytotoxicity reduction.

Water, Air and Soil Pollution, 145(1-4), 79-94.

226

El-Masry, M.H., El-Bestawy, E. and El-Adl, N. (2004). Bioremediation of vegetable oil

and grease from polluted wastewater using a sand biofilm system. World Journal of

Microbiology and Biotechnology, 20(6), 551-557.

Environment Canada (1976). Guidelines for Effluent Quality and Wastewater Treatment

at Federal Establishments. EPS l-EC-76-1, Canada.

Environment Canada (1985). Selection criteria and laboratory evaluation of oil spill

sorbent, update II, EPS report - EPS4-EP-83-4, Canada.

European Committee for Standardization (2002). Separator systems for light liquids (e.g.

oil and petrol) - Part 1: Principles of product design, performance and testing,

marking and quality control. 93/38/EEC (No) 89/106/EEC (C 319, 2005-12-14).

Fabritius, D., Schafer, H.J. and Steinbuchel, A. (1998). Bioconversion of sunflower oil,

rapeseed oil and ricinoleic acid by Candida tropicalis M25. Applied Microbiology

Biotechnology, 50(5), 573-578.

Fadil, K., Chahlaoui, A., Ouahbi, A., Zaid, A. and Boija, R. (2003). Aerobic

biodegradation and detoxification of wastewaters from the olive oil industry.

International Biodeterioration & Biodegradation, 51 (1), 37-41.

Faisal, M. and Unno, H. (2001). Kinetic analysis of palm oil mill wastewater treatment

by a modified anaerobic baffled reactor. Biochemical Engineering Journal, 9(1),

25-31.

Filidei, S., Masuandaro, G. and Ceccanti, B. (2003). Anaerobic digestion of olive oil mill

effluents: Evaluation of wastewater organic load and phytotoxicity reduction.

Water, Air and Soil Pollution, 145(1-4), 79-94.

226

Franson, M.A.H. and Eaton, A.D. (2005). Standard methods for the examination of water

and wastewater. 21st Ed., APHA, AWWA, WPCF, Washington DC.

Freundlich, H. (1906). Over the adsorption in solution. Journal of Physical Chemistry,

57A, 385-470.

Fu, Y. and Viraraghavan, T. (2003). Column studies for biosorption of dyes from

aqueous solutions on immobilized Aspergillus niger fungal biomass. Water SA, 29

(4), 465-472.

Galil, N.I. and Wolf, D. (2001). Removal of hydrocarbons from petrochemical

wastewater by dissolved air flotation. Water Science and Technology, 43(8), 107-

113.

Gammoun, A., Tahiri, S., Albizane, A., Azzi, M., Moros, J., Garrigues, S. and de la

Guardic, H. (2007). Separation of motor oils, oily wastes and hydrocarbons from

contaminated water by sorption

Materials, 145, 148 — 153.

Gardea-Torresdey, J., Becker-Hapak, M.K., Hosea, J.M. and Darnall, D.W. (1990).

Effect of chemical modification of algal carboxyl groups on metal ion binding.


Gombert, A.K., Lopes, A., Castilho, L.R. and Freire, D.M.G. (1999). Lipase, production

by Penicillium restrictum in a solid-state fermentation using babassu oil cake.

Process Biochemistry, 35(1), 85-90.

Gopinath, S.C.B., Anbu, P. and Hilda, A. (2005). Extracellular enzymatic activity

profiles in fungi isolated from oil-rich environments. Mycoscience, 46(2), 119-126.

on chrome shavings. Journal of Hazardous

227

Franson, M.A.H. and Eaton, A.D. (2005). Standard methods for the examination of water

and wastewater. 21st Ed., APHA, AWWA, WPCF, Washington DC.

Freundlich, H. (1906). Over the adsorption in solution. Journal of Physical Chemistry,

57A, 385-470.

Fu, Y. and Viraraghavan, T. (2003). Column studies for biosorption of dyes from

aqueous solutions on immobilized Aspergillus niger fungal biomass. Water SA, 29

(4), 465-472.

Galil, N.I. and Wolf, D. (2001). Removal of hydrocarbons from petrochemical

wastewater by dissolved air flotation. Water Science and Technology, 43(8), 107—

113.

Gammoun, A., Tahiri, S., Albizane, A., Azzi, M., Moros, J., Garrigues, S. and de la

Guardic, H. (2007). Separation of motor oils, oily wastes and hydrocarbons from

contaminated water by sorption on chrome shavings. Journal of Hazardous

Materials, 145, 148 - 153.

Gardea-Torresdey, J., Becker-Hapak, M.K., Hosea, J.M. and Darnall, D.W. (1990).

Effect of chemical modification of algal carboxyl groups on metal ion binding.


Gombert, A.K., Lopes, A., Castilho, L.R. and Freire, D.M.G. (1999). Lipase, production

by Penicillium restrictum in a solid-state fermentation using babassu oil cake.

Process Biochemistry, 35(1), 85-90.

Gopinath, S.C.B., Anbu, P. and Hilda, A. (2005). Extracellular enzymatic activity

profiles in fungi isolated from oil-rich environments. Mycoscience, 46(2), 119-126.

227

Hall, K.R., Eagleton, L.C., Acrivos, A. and Vermeulen, T. (1966). Pore and solid-

diffusion kinetics in fixed-bed adsorption under constant-pattern conditions.

Industrial and Engineering Chemistry Fundamentals, 5 (2), 212-223.

Hamdaoui, 0. (2006). Dynamic sorption of methylene blue by cedar sawdust and crushed

brick in fixed bed columns. Journal of Hazardous Materials B138,293-303.

Hamdi, M. (1993). Future prospects and constraints of olive oil mill wastewater use and

treatment: A review. Bioprocess Engineering, 8(5-6), 155-159.

Hamdi, M., Khadir, A. and Garcia, J. (1991). The use of Aspergillus niger for

bioconversion of olive oil mill waste-waters. Applied Microbiology Biotechnology,

34(6), 828-831.

Han, R., Wang, Y., Zou, W., Wang, Y. and Shi, J. (2007a). Comparison of linear and

nonlinear analysis in estimating the Thomas model parameters for methylene blue

adsorption onto natural zeolite in fixed-bed column. Journal of Hazardous

Materials, 145,331-335.

Han, R., Wang, Y., Yu, W., Zou, W., Shi, J. and Liu, H. (2007b). Biosorption of

methylene blue from aqueous solution by rice husk in a fixed-bed column. Journal

of Hazardous Materials, 141,713-718.

Hanaki, K., Matsuo, T. and Kumazaki, K. (1990). Treatment of oily cafeteria wastewater

by singl-phase and two-phase anaerobic filter. Water Science and Technology, 22(3-4),

299-306.

228

Hall, K.R., Eagleton, L.C., Acrivos, A. and Vermeulen, T. (1966). Pore and solid-

diffusion kinetics in fixed-bed adsorption under constant-pattern conditions.

Industrial and Engineering Chemistry Fundamentals, 5 (2), 212-223.

Hamdaoui, O. (2006). Dynamic sorption of methylene blue by cedar sawdust and crushed

brick in fixed bed columns. Journal of Hazardous Materials B138, 293-303.

Hamdi, M. (1993). Future prospects and constraints of olive oil mill wastewater use and

treatment: A review. Bioprocess Engineering, 8(5-6), 155-159.

Hamdi, M., Khadir, A. and Garcia, J. (1991). The use of Aspergillus niger for

bioconversion of olive oil mill waste-waters. Applied Microbiology Biotechnology,

34(6), 828-831.

Han, R., Wang, Y., Zou, W., Wang, Y. and Shi, J. (2007a). Comparison of linear and

nonlinear analysis in estimating the Thomas model parameters for methylene blue

adsorption onto natural zeolite in fixed-bed column. Journal of Hazardous

Materials, 145,331-335.

Han, R., Wang, Y., Yu, W., Zou, W., Shi, J. and Liu, H. (2007b). Biosorption of

methylene blue from aqueous solution by rice husk in a fixed-bed column. Journal

of Hazardous Materials, 141, 713-718.

Hanaki, K., Matsuo, T. and Kumazaki, K. (1990). Treatment of oily cafeteria wastewater

by singl-phase and two-phase anaerobic filter. Water Science and Technology, 22(3-4),

299-306.

228

Hanalei, K., Matsuo, T. and Nagase, M. (1981). Mechanism of inhibition caused by long-

chain fatty acids in anaerobic digestion process. Biotechnology and Bioengineering,

23(7), 1591-1610.

Hazlett, R.N. (1969). Fibrous bed coalescence of water. Industrial and Engineering

Chemistry Fundamentals, 8 (4), 625-632.

Hill, C.G.J. (1977). An introduction to chemical engineering kinetics and reactor design.

Wiley, New York.

Ho, Y.S. and McKay, G. (1998). Kinetic model for lead(II) sorption on to peat.

Adsorption Science and Technology, 16 (4), 243-255.

Hsu, T.C., Hanaki, K. and Matsumoto, J. (1983). Kinetics of hydrolysis, oxidation and

adsorption during olive oil degradation by activated sludge. Biotechnology and

Bioengineering, 25(7), 1829-1839.

Ibrahim, S., Ang, H-M. and Wang, S. (2009). Removal of emulsified food and mineral

oils from wastewater using surfactant modified barley straw. Bioresource

Technology, 100, 5744-5749.

Inagaki, M., Kawahara, A., Nishi, Y. and Iwashita, N. (2002). Heavy oil sorption and

recovery by using carbon fiber felts. Carbon, 40, 1487 — 1492.

Jawarska, M.M. (2003). Fungi: a source of chitosan. Recent Research Developments in

Applied Microbiology and Biotechnology, 1, 219-231.

Jeganathan, J., Bassi, A. and Nakhla, G. (2006). Pre-treatment of high oil and grease pet

food industrial wastewater using immobilized lipase hydrolization. Journal of

Hazardous Materials, 137(1),121-128.

229

Hanaki, K., Matsuo, T. and Nagase, M. (1981). Mechanism of inhibition caused by long-

chain fatty acids in anaerobic digestion process. Biotechnology andBioengineering,

23(7), 1591-1610.

Hazlett, R.N. (1969). Fibrous bed coalescence of water. Industrial and Engineering

Chemistry Fundamentals, 8 (4), 625-632.

Hill, C.G.J. (1977). An introduction to chemical engineering kinetics and reactor design.

Wiley, New York.

Ho, Y.S. and McKay, G. (1998). Kinetic model for lead(II) sorption on to peat.

Adsorption Science and Technology, 16 (4), 243-255.

Hsu, T.C., Hanaki, K. and Matsumoto, J. (1983). Kinetics of hydrolysis, oxidation and

adsorption during olive oil degradation by activated sludge. Biotechnology and

Bioengineering, 25(7), 1829-1839.

Ibrahim, S., Ang, H-M. and Wang, S. (2009). Removal of emulsified food and mineral

oils from wastewater using surfactant modified barley straw. Bioresource

Technology, 100, 5744-5749.

Inagaki, M., Kawahara, A., Nishi, Y. and Iwashita, N. (2002). Heavy oil sorption and

recovery by using carbon fiber felts. Carbon, 40,1487 - 1492.

Jawarska, M.M. (2003). Fungi: a source of chitosan. Recent Research Developments in

Applied Microbiology and Biotechnology, 1, 219-231.

Jeganathan, J., Bassi, A. and Nakhla, G. (2006). Pre-treatment of high oil and grease pet

food industrial wastewater using immobilized lipase hydrolization. Journal of

Hazardous Materials, 137(1),121-128.

229

Jensen, R.G. (1971). Lipolytic enzymes. Progress in the Chemistry of Fats and other

Lipids, 11(3), 347-394.

Ji, F., Li, C., Dong, X., Li, Y. and Wang, D. (2009). Separation of oil from oily

wastewater by sorption and coalescence technique using ethanol grafted

polyacrylonitrile. Journal of Hazardous Materials, 164, 1346 — 1351.

Johnson, R.F., Manjreker, T.G. and Halligan, J.E. (1973). Removal of oil from water

surfaces by sorption on unstructured fibers. Environmental Science and

Technology, 7 (5), 439-443.

Johnston, P.R. (1983). The most probable pore-size distribution in fluid filter media.

Journal of Testing and Evaluation, 11 (2): 117-125.

Joseph, M., Kock, J.L.F., Pohl, C.H., Botes, P.J., van Heerdan, E. and Hugo, A. (2005).

Actate-enhanced polymerized triacylglycerol utilization by Mucor circinelloides.

World Journal of Microbiology and Biotechnology, 21(1), 97-99.

Jung, F., Cammarota, M.C. and Freire, D.M.G. (2002). Impact of enzymatic pre-

hydrolysis on batch activated sludge systems dealing with oily wastewaters.

Biotechnology Letters, 24(21), 1797-1802.

Kapoor, A. and Viraraghavan, T. (1997). Heavy metal biosorption sites in Aspergillus

niger. Bioresource Technology, 61 (3), 221-227.

Kapoor, A. and Viraraghavan, T. (1998). Removal of heavy metals from aqueous

solutions using immobilized fungal biomass in continuous mode. Water Research, 32

(6), 1968-1977.

230

Jensen, R.G. (1971). Lipolytic enzymes. Progress in the Chemistry of Fats and other

Lipids, 11(3), 347-394.

Ji, F., Li, C., Dong, X., Li, Y. and Wang, D. (2009). Separation of oil from oily

wastewater by sorption and coalescence technique using ethanol grafted

polyacrylonitrile. Journal of Hazardous Materials, 164, 1346- 1351.

Johnson, R.F., Manjreker, T.G. and Halligan, J.E. (1973). Removal of oil from water

surfaces by sorption on unstructured fibers. Environmental Science and

Technology, 7 (5), 439-443.

Johnston, P.R. (1983). The most probable pore-size distribution in fluid filter media.

Journal of Testing and Evaluation, 11 (2): 117-125.

Joseph, M., Kock, J.L.F., Pohl, C.H., Botes, P.J., van Heerdan, E. and Hugo, A. (2005).

Actate-enhanced polymerized triacylglycerol utilization by Mucor circinelloides.

World Journal of Microbiology and Biotechnology, 21(1), 97-99.

Jung, F., Cammarota, M.C. and Freire, D.M.G. (2002). Impact of enzymatic pre-

hydrolysis on batch activated sludge systems dealing with oily wastewaters.

Biotechnology Letters, 24(21), 1797-1802.

Kapoor, A. and Viraraghavan, T. (1997). Heavy metal biosorption sites in Aspergillus

niger. Bioresource Technology, 61 (3), 221-227.

Kapoor, A. and Viraraghavan, T. (1998). Removal of heavy metals from aqueous

solutions using immobilized fungal biomass in continuous mode. Water Research, 32

(6), 1968-1977.

230

Kapoor, A. and Virarghavan, T. (1995). Fungal biosorption - an alternative treatment

option for heavy metal bearing wastewaters: a review. Bioresource Technology, 53,

195-206.

Keenan, D. and Sabelnikov, A. (2000). Biological augmentation eliminates grease and oil

in bakery wastewater. Water Environment Research, 72(2), 141-146.

Khan, E., Virojnagud, W. and Ratpukdi, T. (2004). Use of biomass sorbents for oil

removal from gas station runoff. Chemosphere 57, 681 — 689.

Knezeric, Z., Mojovic, L. and Adnadjevic, B. (1998). Palm oil hydrolysis by lipase from

Candida cylindracea immobilized on zeolite type. Enzyme and Microbial

Technology, 22 (4), 275-280.

Knorr, D. and Klein, J. (1986). Production and conversion of chitosan with cultures of

Mucor rouxii or Phycomycus blakesleeanus. Biotechnology Letters, 8(10), 691-694.

Kobayashi,Y., Matsuo, R., and Nishiyama, M. (1977). Method for adsorption of oils.

Japanese Patent 52,138,081, Japan.

Koritala, S., Hesseltine, C.W., Pryde, E.H. and Mounts, T.L. (1987). Biochemical

modification of fats by microorganisms: A preliminary survey. Journal of the

American Oil Chemists' Society, 64(4), 509-513.

Kovarova-Kovar, K. and Egli, T. (1998). Growth kinetics of suspended microbial cells:

From single-substrate-controlled growth to mixed-substrate kinetics. Microbiology

and Molecular Biology Reviews, 62(3), 646-666.

Kuyucak, N. (1990). Feasibility of biosorbents application. In Volesky, B. (Ed.)

Biosorption of Heavy Metals, CRC Press, Boca Raton, Florida, 371-380.

231

Kapoor, A. and Virarghavan, T. (1995). Fungal biosorption - an alternative treatment

option for heavy metal bearing wastewaters: a review. Bioresource Technology, 53,

195-206.

Keenan, D. and Sabelnikov, A. (2000). Biological augmentation eliminates grease and oil

in bakery wastewater. Water Environment Research, 72(2), 141-146.

Khan, E., Virojnagud, W. and Ratpukdi, T. (2004). Use of biomass sorbents for oil

removal from gas station runoff. Chemosphere 57, 681 - 689.

Knezeric, Z,, Mojovic, L. and Adnadjevic, B. (1998). Palm oil hydrolysis by lipase from

Candida cylindracea immobilized on zeolite type. Enzyme and Microbial

Technology, 22 (4), 275-280.

Knorr, D. and Klein, J. (1986). Production and conversion of chitosan with cultures of

Mucorrouxii or Phycomycus blakesleeanus. Biotechnology Letters, 8(10), 691-694.

Kobayashi.Y., Matsuo, R., and Nishiyama, M. (1977). Method for adsorption of oils.

Japanese Patent 52,138,081, Japan.

Koritala, S., Hesseltine, C.W., Pryde, E.H. and Mounts, T.L. (1987). Biochemical

modification of fats by microorganisms: A preliminary survey. Journal of the

American Oil Chemists' Society, 64(4), 509-513.

Kovarova-Kovar, K. and Egli, T. (1998). Growth kinetics of suspended microbial cells:

From single-substrate-controlled growth to mixed-substrate kinetics. Microbiology

and Molecular Biology Reviews, 62(3), 646-666.

Kuyucak, N. (1990). Feasibility of biosorbents application. In Volesky, B. (Ed.)

Biosorption of Heavy Metals, CRC Press, Boca Raton, Florida, 371-380.

231

Lagergren, S. (1898). About the theory of so-called adsorption of soluble substances.

Kungliga Svenska Vetenskapsakademiens. Handlingar, Band 24 (4), 1-39.

Lanciotti, R., Gianotti, A., Baldi, D., Angrisani, R., Suzzi, G., Mastrocola, D. and

Guerzoni, M.E. (2005). Use of Yarrowia lipolytica strains for the treatment of olive

mill wastewater. Bioresource Technology, 96(3), 317— 322.

Lange, A. (1973). Lange's handbook of chemistry (12th edition). In: John, A.D. (Ed.).

McGraw-Hill Book Company, New York, USA.

Langmuir, I. (1918). The adsorption of gases on plane surface of glass, mica and

platinum. Journal of the American Chemical Society, 40,1361-1403.

Lapara, T.M. and Alleman, J.E. (1999). Thermophilic aerobic biological wastewater

treatment. Water Research, 33(4), 895-908.

Leal, M.C.M.R., Cammarota, M.C., Freire D.M.G. and Sant'Anna Jr G.L. (2002).

Hydrolytic enzymes as co-adjuvant in the anaerobic treatment of dairy wastewaters.

Brazilian Journal of Chemical Engineering, 19(2), 175-80.

Leal, M.C.M.R., Freire, D.M.G., Cammarota, M.C., Sant'Anna G.L. Jr. (2006). Effect of

enzymatic hydrolysis on anaerobic treatment of dairy wastewater. Process

Biochemistry, 41(5), 1173-1178.

Li, J. and Gu, Y. (2005). Coalescence of oil-in-water emulsions in fibrous and granular

beds. Separation and Purification Technology, 42,1-13.

Lin-Vien, D. (1991). The Handbook of infrared and Raman characteristic frequencies of

organic molecules. Academic Press Inc., Boston.

232

Lagergren, S. (1898). About the theory of so-called adsorption of soluble substances.

Kungliga Svenska Vetenskapsakademiens. Handlingar, Band 24 (4), 1-39.

Lanciotti, R., Gianotti, A., Baldi, D., Angrisani, R., Suzzi, G., Mastrocola, D. and

Guerzoni, M.E. (2005). Use of Yarrowia lipolytica strains for the treatment of olive

mill wastewater. Bioresource Technology, 96(3), 317- 322.

Lange, A. (1973). Lange's handbook of chemistry (12th edition). In: John, A.D. (Ed.).

McGraw-Hill Book Company, New York, USA.

Langmuir, I. (1918). The adsorption of gases on plane surface of glass, mica and

platinum. Journal of the American Chemical Society, 40, 1361-1403.

Lapara, T.M. and Alleman, J.E. (1999). Thermophilic aerobic biological wastewater

treatment. Water Research, 33(4), 895-908.

Leal, M.C.M.R., Cammarota, M.C., Freire D.M.G. and Sant'Anna Jr G.L. (2002).

Hydrolytic enzymes as co-adjuvant in the anaerobic treatment of dairy wastewaters.

Brazilian Journal of Chemical Engineering, 19(2), 175-80.

Leal, M.C.M.R., Freire, D.M.G., Cammarota, M.C., Sant'Anna G.L. Jr. (2006). Effect of

enzymatic hydrolysis on anaerobic treatment of dairy wastewater. Process

Biochemistry, 41(5), 1173-1178.

Li, J. and Gu, Y. (2005). Coalescence of oil-in-water emulsions in fibrous and granular

beds. Separation and Purification Technology, 42, 1-13.

Lin-Vien, D. (1991). The Handbook of infrared and Raman characteristic frequencies of

organic molecules. Academic Press Inc., Boston.

232

Lipp, P., Lee, C. H., Fane, A. G. and Fell, C. J. D. (1998). A fundamental study of the

ultrafiltration of oil-water emulsions. Journal of Membrane Science, 36, 161-177.

Liu, V.L., Nakhla, G. and Bassi A. (2004). Treatability and kinetics studies of mesophilic

aerobic biodegradation of high oil and grease pet food wastewater. Journal of

Hazardous Materials, 112(1-2), 87-94.

Lodeiro, P. Herrero, R. and Sastre de Vicente, M.E. (2006). The use of protonated

Sargassum muticum as biosorbent for cadmium removal in a fixed-bed column.

Journal of Hazardous Materials B137, 244-253.

Loudon, G.M. (1984). Organic Chemistry. Addison-Wesley Publishing Company,

Reading, Massachusetts, USA.

Majumdar, S.S., Das, S.K., Saha, T., Panda, G.C., Bandyopadhyoy, T. and Guha, A.K.

(2008). Adsorption behavior of copper ions on Mucor rouxii biomass through

microscopic and FTIR analysis. Colloids and Surfaces B: Biointerfaces, 63, 138-

145.

Malcata, F.X., Reyes, H.R., Garcia, H.S., Hill, C.G. and Amundson, C.H. (1992).

Kinetics and mechanisms of reactions catalyzed by immobilized lipases. Enzyme

and Microbial Technology, 14(6), 426-446.

Manning, F.S. and Thompson, R.E. (1995). Oilfield processing of petroleum: crude oil.

Penn Well Publishing Co., Oklahoma, USA.

Markossian, S., Becker, P., Mark, H. and Antranikian, G. (2000). Isolation and

characterization of lipid-degrading Bacillus thermoleovorans IHI-91 from an

icelandic hot spring. Extremophiles, 4(6), 365-371.

233

Lipp, P., Lee, C. H., Fane, A. G. and Fell, C. J. D. (1998). A fundamental study of the

ultrafiltration of oil-water emulsions. Journal of Membrane Science, 36, 161-177.

Liu, V.L., Nakhla, G. and Bassi A. (2004). Treatability and kinetics studies of mesophilic

aerobic biodegradation of high oil and grease pet food wastewater. Journal of

Hazardous Materials, 112(1-2), 87-94.

Lodeiro, P. Herrero, R. and Sastre de Vicente, M.E. (2006). The use of protonated

Sargassum muticum as biosorbent for cadmium removal in a fixed-bed column.

Journal of Hazardous Materials B137, 244—253.

Loudon, G.M. (1984). Organic Chemistry. Addison-Wesley Publishing Company,

Reading, Massachusetts, USA.

Majumdar, S.S., Das, S.K., Saha, T., Panda, G.C., Bandyopadhyoy, T. and Guha, A.K.

(2008). Adsorption behavior of copper ions on Mucor rouxii biomass through

microscopic and FTIR analysis. Colloids and Surfaces B: Biointerfaces, 63, 138-

145.

Malcata, F.X., Reyes, H.R., Garcia, H.S., Hill, C.G. and Amundson, C.H. (1992).

Kinetics and mechanisms of reactions catalyzed by immobilized lipases. Enzyme

and Microbial Technology, 14(6), 426-446.

Manning, F.S. and Thompson, R.E. (1995). Oilfield processing of petroleum: crude oil.

Penn Well Publishing Co., Oklahoma, USA.

Markossian, S., Becker, P., Mark, H. and Antranikian, G. (2000). Isolation and

characterization of lipid-degrading Bacillus thermoleovorans IHI-91 from an

icelandic hot spring. Extremophiles, 4(6), 365-371.

233

Mathavan, G. N. and Viraraghavan, T. (1992). Coalescence/filtration of an oil-in-water

emulsion in a peat bed. Water Research, 26 (1), 91-98.

Mathavan, G.N. and Viraraghavan, T. (1989). Use of peat in the treatment of oily waters.

Water, Air and Soil Pollution, 45 (1-2), 17-26.

Mathavan, G.N. and Viraraghavan, T. (1990). Comparisons of IR determinations of oil

and grease in water. Environmental Technology, 11, 455-462.

Mathews A.P. and Weber W.J. (1976). Effects of external mass transfer and inter-particle

diffusion on adsorption. AIChE Symposium Series, 73, 91-98.

Mathialagan, T. and Viraraghavan, T. (2002). Adsorption of cadmium from aqueous

solutions by perlite. Journal of Hazardous Materials, 94 (3), 291-303.

Minitab (2007). Meet Minitab 15 for Windows, Minitab Inc., USA.

Mittal, A., Gupta, V.K., Malviya, A. and Mittal, J. (2008). Process development for the

batch and bulk removal and recovery of a hazardous, water-soluble azo dye

(Metanil Yellow) by adsorption over waste materials (Bottom Ash and De-Oiled

Soya). Journal of Hazardous Materials, 151(2-3), 821-32.

Miyoshi, H., Shimura, K., Watanabe, K. and Onodera, K. (1992). Characterization of

some fungal chitosans. Bioscience, Biotechnology and Biochemistry, 56, 1901-1905.

Mkhize, S.P., Atkinson, B.W. and Bux F. (2000). Evaluation of a laboratory-scale

biological process for the treatment of edible oil effluent. Water South Africa,

26(4), 555-558.

Moazed, H. (2000). Removal of oil from water by organo-clay and other sorbents. Ph.D.

Thesis, University of Regina, Regina, Canada.

234

Mathavan, G. N. and Viraraghavan, T. (1992). Coalescence/filtration of an oil-in-water

emulsion in a peat bed. Water Research, 26 (1), 91-98.

Mathavan, G.N. and Viraraghavan, T. (1989). Use of peat in the treatment of oily waters.

Water, Air and Soil Pollution, 45 (1-2), 17-26.

Mathavan, G.N. and Viraraghavan, T. (1990). Comparisons of IR determinations of oil

and grease in water. Environmental Technology, 11, 455-462.

Mathews A.P. and Weber W.J. (1976). Effects of external mass transfer and inter-particle

diffusion on adsorption. AIChE Symposium Series, 73, 91-98.

Mathialagan, T. and Viraraghavan, T. (2002). Adsorption of cadmium from aqueous

solutions by perlite. Journal of Hazardous Materials, 94 (3), 291-303.

Minitab (2007). Meet Minitab 15 for Windows, Minitab Inc., USA.

Mittal, A., Gupta, V.K., Malviya, A. and Mittal, J. (2008). Process development for the

batch and bulk removal and recovery of a hazardous, water-soluble azo dye

(Metanil Yellow) by adsorption over waste materials (Bottom Ash and De-Oiled

Soya). Journal of Hazardous Materials, 151(2-3), 821-32.

Miyoshi, H., Shimura, K., Watanabe, K. and Onodera, K. (1992). Characterization of

some fungal chitosans. Bioscience, Biotechnology and Biochemistry, 56, 1901-1905.

Mkhize, S.P., Atkinson, B.W. and Bux F. (2000). Evaluation of a laboratory-scale

biological process for the treatment of edible oil effluent. Water South Africa,

26(4), 555-558.

Moazed, H. (2000). Removal of oil from water by organo-clay and other sorbents. Ph.D.

Thesis, University of Regina, Regina, Canada.

234

Moazed, H. and Viraraghavan, T. (2001). Organo-clay anthracite filtration for oil

removal. Journal of Canadian Petroleum Technology, 49 (9), 37-42.

Moazed, H. and Viraraghavan, T. (2002). Coalescence/filtration of an oil-in-water

emulsion in a granular organo-clay/anthracite mixture bed. Water, Air and soil

Pollution 138, 253-270.

Moazed, H. and Viraraghavan, T. (2005). Removal of oil from water by bentonite

organoclay. Practice Periodical of Hazardous Toxic and Radioactive Waste

Management, 9, 130-134.

Moursy, A.S. and Abo EI-Ela, S.E. (1982). Treatment of oily refinery wastes using a

dissolved air flotation process. Environment International, 7, 267-270.

Moore, D.S. (2009). The basic practice of statistics. Fifth Eds., W.H. Freeman and

Company, New York, USA.

Multon, L.M. and Viraraghavan, T. (2006). Removal of oil from produced water by

coalescence/filtration in a granular bed. Environmental Technology, 27, 529-544.

Mungasavalli, D.P., Viraraghavan, T. and Jin, Y-C. (2007). Biosorption of chromium

from aqueous solutions by pretreated Aspergillus niger: Batch and column studies.

Colloids and Surfaces A: Physicochemical Engineering Aspects, 301, 214-223.

Murray, M., Rooney, D., Neikerk, V. M., Montenegro, A. and Weatherley, L.B. (1997).

Immobilization of lipase onto lipophilic polymer particles and application to oil

hydrolysis. Process Biochemistry, 32(6), 479-486.

235

Moazed, H. and Viraraghavan, T. (2001). Organo-clay anthracite filtration for oil

removal. Journal of Canadian Petroleum Technology, 49 (9), 37-42.

Moazed, H. and Viraraghavan, T. (2002). Coalescence/filtration of an oil-in-water

emulsion in a granular organo-clay/anthracite mixture bed. Water, Air and soil

Pollution 138, 253-270.

Moazed, H. and Viraraghavan, T. (2005). Removal of oil from water by bentonite

organoclay. Practice Periodical of Hazardous Toxic and Radioactive Waste

Management, 9, 130-134.

Moursy, A.S. and Abo EI-Ela, S.E. (1982). Treatment of oily refinery wastes using a

dissolved air flotation process. Environment International, 7, 267-270.

Moore, D.S. (2009). The basic practice of statistics. Fifth Eds., W.H. Freeman and

Company, New York, USA.

Multon, L.M. and Viraraghavan, T. (2006). Removal of oil from produced water by

coalescence/filtration in a granular bed. Environmental Technology, 27, 529-544.

Mungasavalli, D.P., Viraraghavan, T. and Jin, Y-C. (2007). Biosorption of chromium

from aqueous solutions by pretreated Aspergillus niger: Batch and column studies.

Colloids and Surfaces A: Physicochemical Engineering Aspects, 301,214-223.

Murray, M., Rooney, D., Neikerk, V. M., Montenegro, A. and Weatherley, L.B. (1997).

Immobilization of lipase onto lipophilic polymer particles and application to oil

hydrolysis. Process Biochemistry, 32(6), 479-486.

235

Murthy, V.R., Bhat, J. and Muniswaran, P.K.A. (2004). Hydrolysis of rice bran oil using

an immobilized lipase from Candida rugosa in isooctane. Biotechnology Letters,

26(7), 563-567.

Muzzarelli, R.A.A. (1977). Chitin. Pergamon Press Ltd., Oxford, England, UK.

Muzzarelli, R.A.A., lari, P., Tarsi, R., Dubini, B. and Xia, W. (1994). Chitosan from

Absidia coerulea. Carbohydrate Polymers, 25,45-50.

Mysore, D., Viraraghavan, T. and Jin, Y-C. (2006). Vermiculite filtration for removal of

oil from water. Practice Periodical of Hazardous Toxic and Radioactive Waste

Management, 10(3), 156-161.

Mysore, D., Viraraghavan, T. and Jin, Y.C. (2004). Removal of oil by vermiculite.

Fresenius Environmental Bulletin, 13(6), 560-567.

Mysore, D., Viraraghavan, T. and Jin, Y.C. (2005). Treatment of oily waters using

vermiculite. Water Research, 39,2643-2653.

Najafpour, G., Yieng, H.A., Younesi, H. and Zinatizadeh A. (2005). Effect of organic

loading on performance of rotating biological contactors using palm oil mill

effluents. Process Biochemistry, 40(8), 2879-2884.

Najafpour, G., Zinatizadeh, A.A.L., Mohamed, A.R., Hasnain Isa M. and. Nasrollahzadeh

H. (2006). High-rate anaerobic digestion of palm oil mill effluent in an upflow

anaerobic sludge-fixed film bioreactor. Process Biochemistry, 41(2), 370-379.

Nakhla, G., Al-Sabawi, M., Bassi, A. and Liu V. (2003). Anaerobic treatability of high

oil and grease rendering wastewater. Journal of Hazardous Materials, 102(2-3),

243-255.

236

Murthy, V.R., Bhat, J. and Muniswaran, P.K.A. (2004). Hydrolysis of rice bran oil using

an immobilized lipase from Candida rugosa in isooctane. Biotechnology Letters,

26(7), 563-567.

Muzzarelli, R.A.A. (1977). Chitin. Pergamon Press Ltd., Oxford, England, UK.

Muzzarelli, R.A.A., Illari, P., Tarsi, R., Dubini, B. and Xia, W. (1994). Chitosan from

Absidia coerulea. Carbohydrate Polymers, 25,45-50.

Mysore, D., Viraraghavan, T. and Jin, Y-C. (2006). Vermiculite filtration for removal of

oil from water. Practice Periodical of Hazardous Toxic and Radioactive Waste

Management, 10(3), 156-161.

Mysore, D., Viraraghavan, T. and Jin, Y.C. (2004). Removal of oil by vermiculite.


Mysore, D., Viraraghavan, T. and Jin, Y.C. (2005). Treatment of oily waters using

vermiculite. Water Research, 39, 2643-2653.

Najafpour, G., Yieng, H.A., Younesi, H. and Zinatizadeh A. (2005). Effect of organic

loading on performance of rotating biological contactors using palm oil mill

effluents. Process Biochemistry, 40(8), 2879-2884.

Najafpour, G., Zinatizadeh, A.A.L., Mohamed, A.R., Hasnain Isa M. and. Nasrollahzadeh

H. (2006). High-rate anaerobic digestion of palm oil mill effluent in an upflow

anaerobic sludge-fixed film bioreactor. Process Biochemistry, 41(2), 370-379.

Nakhla, G., Al-Sabawi, M., Bassi, A. and Liu V. (2003). Anaerobic treatability of high

oil and grease rendering wastewater. Journal of Hazardous Materials, 102(2-3),

243-255.

236

Novak, J.N. and Carlson, D.A. (1970). The kinetics of anaerobic long chain fatty acid

degradation. Journal Water Pollution Control Federation, 42(11), 1932-1948.

Ogundero, V.W. (1982). Hydrolysis of vegetable oils and triglycerides by thermotolerant

and zoopathogenic species of Aspergillus from Nigerian palm produce.

Mycopatholgia, 77(1), 43-46.

Outman, C.S. (1980). The logistic curve as a model for carbon bed design. Journal

American Water Works Association, 72(1), 50-53.

Owens, N. and Lee, D. (2007). The use of micro bubble flotation technology in

secondary and tertiary produced water treatment — a technical comparison with other

separation technologies. In: TUV NEL, 5th Produced Water Workshop, 30th-31st

May, Aberdeen, Scotland.

Ozgen, 0., Burcu, A. and Yasemin, Y. (2006). Effect of oil type on stability of W/O/W

emulsions. Cosmetics and toiletries, 121(7), 57-64.

Pagnanelli, F. (2011). Equilibrium, kinetic and dynamic modeling of biosorption

processes. In Microbial biosorption of metals, (Eds.) Kotrba, P., MAckova, M. and

Macek, T., Springer, Heidelberg, Germany.

Paiva, A.L., Balcao, V.M. and Malcata, F.X. (2000). Kinetics and mechanisms of

reactions catalyzed by immobilized lipases. Enzyme and Microbial Technology,

27(3-5), 187-204.

Pasila, A. (2004). A biological oil adsorption filter. Marine Pollution Bulletin, 49 (11-

12), 1006-1012.

Pasila, A. (2004). A biological oil adsorption filter. Marine Pollution Bulletin, 49, 1006 —

237

Novak, J.N. and Carlson, D.A. (1970). The kinetics of anaerobic long chain fatty acid

degradation. Journal Water Pollution Control Federation, 42(11), 1932-1948.

Ogundero, V.W. (1982). Hydrolysis of vegetable oils and triglycerides by thermotolerant

and zoopathogenic species of Aspergillus from Nigerian palm produce.

Mycopatholgia, 77(1), 43-46.

Oulman, C.S. (1980). The logistic curve as a model for carbon bed design. Journal

American Water Works Association, 72(1), 50-53.

Owens, N. and Lee, D. (2007). The use of micro bubble flotation technology in

secondary and tertiary produced water treatment - a technical comparison with other

separation technologies. In: TUV NEL, 5th Produced Water Workshop, 30th—31st

May, Aberdeen, Scotland.

Ozgen, O., Burcu, A. and Yasemin, Y. (2006). Effect of oil type on stability of W/O/W

emulsions. Cosmetics and toiletries, 121(7), 57-64.

Pagnanelli, F. (2011). Equilibrium, kinetic and dynamic modeling of biosorption

processes. In Microbial biosorption of metals, (Eds.) Kotrba, P., MAckova, M. and

Macek, T., Springer, Heidelberg, Germany.

Paiva, A.L., Balcao, V.M. and Malcata, F.X. (2000). Kinetics and mechanisms of

reactions catalyzed by immobilized lipases. Enzyme and Microbial Technology,

27(3-5), 187-204.

Pasila, A. (2004). A biological oil adsorption filter. Marine Pollution Bulletin, 49 (11-

12), 1006-1012.

Pasila, A. (2004). A biological oil adsorption filter. Marine Pollution Bulletin, 49, 1006 -

237

1012.

Perrich, J.R. (1981). Activated carbon adsorption for wastewater treatment, CRC Press:

Florida, USA.

Piyamongkala, K., Mekasut, L. and Pongstabodea, S. (2008). Cutting fluid effluent

removal by adsorption on chitosan and SDS — modified chitosan. Macromolecular

Research, 16 (6), 492 — 502.

Prakasham, R.S., Merrie, J.S., Sheela, R., Saraswathi, N. and Ramakrishna, S.V. (1999).

Biosorption of chromium VI by free and immobilized Rhizopus arrhizus.

Environmental Pollution, 104, 421-427.

Pokhrel D. and Viraraghavan, T. (2008). Arsenic removal from an aqueous solution by

modified A. niger biomass: batch kinetic and isotherm studies. Journal of

Hazardous Materials, 150, 818-825.

Prasertsan, P. Kithkun, A.H. and Mueesri, P. (2004). Bioseparation of suspended solid

and oil from palm oil mill effluent and secondary treatment by photosynthetic

bacteria. European Symposium on Environmental Biotechnology, April 25-28,

Oostende, Belgium, 771-773.

Pushkarev, V.V., Yuzhaninov, A.G. and Men, S.K. (1983). Treatment of oil-containing

wastewater. Allerton Press, Inc. NY, USA.

Quemeneur, M. and Marty, Y. (1994). Fatty acids and sterols in domestic wastewater.

Water Research, 28 (5), 1217-1226.

Radetic, M., Ilic, V., Radojevic, D., Miladinovic, R., Jocic, D. and Jovacic, P. (2008).

Efficiency of recycled wool — based nonwoven material for the removal of oils

238

1012.

Perrich, J.R. (1981). Activated carbon adsorption for wastewater treatment, CRC Press:

Florida, USA.

Piyamongkala, K., Mekasut, L. and Pongstabodea, S. (2008). Cutting fluid effluent

removal by adsorption on chitosan and SDS - modified chitosan. Macromolecular

Research, 16 (6), 492-502.

Prakasham, R.S., Merrie, J.S., Sheela, R., Saraswathi, N. and Ramakrishna, S.V. (1999).

Biosorption of chromium VI by free and immobilized Rhizopus arrhizus.

Environmental Pollution, 104,421-427.

Pokhrel D. and Viraraghavan, T. (2008). Arsenic removal from an aqueous solution by

modified A. niger biomass: batch kinetic and isotherm studies. Journal of

Hazardous Materials, 150,818-825.

Prasertsan, P. Kithkun, A.H. and Mueesri, P. (2004). Bioseparation of suspended solid

and oil from palm oil mill effluent and secondary treatment by photosynthetic

bacteria. European Symposium on Environmental Biotechnology, April 25-28,

Oostende, Belgium, 771-773.

Pushkarev, V.V., Yuzhaninov, A.G. and Men, S.K. (1983). Treatment of oil-containing

wastewater. Allerton Press, Inc. NY, USA.

Quemeneur, M. and Marty, Y. (1994). Fatty acids and sterols in domestic wastewater.

Water Research, 28 (5), 1217-1226.

Radetic, M., Ilic, V., Radojevic, D., Miladinovic, R., Jocic, D. and Jovacic, P. (2008).

Efficiency of recycled wool - based nonwoven material for the removal of oils

238

from water. Chemosphere, 70, 525 — 530.

Rahman, S.S. (1992). Evaluation of filtering efficiency of walnut granules as deep-bed

filter media. Journal of Petroleum Science and Engineering, 7 (3-4), 239-246.

Ramirez C.M., da Silva, M.P., Ferreira L.S.G. and Vasco E.O. (2007). Mathematical

models applied to the Cr(III) and Cr(VI) breakthrough curves. Journal of


Rane, K. D. and Hoover, D.G. (1993). Production of chitosan by fungi. Food

Biotechnology, 7, 11-33.

Rase, H.F. (1977). Chemical reactor design for process plants, Volume 1: Principles and

techniques. Wiley and Sons, New York, USA.

Ratledge, C. (1992). Microbial oxidation of fatty alcohols and fatty acids. Journal of

Chemical Technology and Biotechnology, 55(4), 399-400.

Raunkjaer, K., Hvitved-Jacobsen, T. and Nielsen, P.H. (1994). Measurement of pools of

protein, carbohydrate and lipid in domestic wastewater. Water Research, 28(2),

251-262.

Ravi Kumar, N.V.M. (2000). A review of chitin and chitosan applications. Reactive and

Functional Polymers, 46, 1-27.

Reynolds, T. D. and Richards, P. (1995). Unit operations and processes in environmental

engineering. Second ed. PWS Publishing Company, Boston, MA.

Ribeiro, T., Rubio, J. and Smith, R.W. (2003). A dried hydrophobic aquaphyte as an oil

filter for oil/water emulsions. Spill Science and Technology Bulletin, 8 (5), 483 —

489.

239

from water. Chemosphere, 70, 525 - 530.

Rahman, S.S. (1992). Evaluation of filtering efficiency of walnut granules as deep-bed

filter media. Journal of Petroleum Science and Engineering, 7 (3-4), 239-246.

Ramirez C.M., da Silva, M.P., Ferreira L.S.G. and Vasco E.O. (2007). Mathematical

models applied to the Cr(III) and Cr(VI) breakthrough curves. Journal of


Rane, K. D. and Hoover, D.G. (1993). Production of chitosan by fungi. Food

Biotechnology, 7, 11-33.

Rase, H.F. (1977). Chemical reactor design for process plants, Volume 1: Principles and

techniques. Wiley and Sons, New York, USA.

Ratledge, C. (1992). Microbial oxidation of fatty alcohols and fatty acids. Journal of

Chemical Technology and Biotechnology, 55(4), 399-400.

Raunkjaer, K., Hvitved-Jacobsen, T. and Nielsen, P.H. (1994). Measurement of pools of

protein, carbohydrate and lipid in domestic wastewater. Water Research, 28(2),

251-262.

Ravi Kumar, N.V.M. (2000). A review of chitin and chitosan applications. Reactive and

Functional Polymers, 46, 1-27.

Reynolds, T. D. and Richards, P. (1995). Unit operations and processes in environmental

engineering. Second ed. PWS Publishing Company, Boston, MA.

Ribeiro, T., Rubio, J. and Smith, R.W. (2003). A dried hydrophobic aquaphyte as an oil

filter for oil/water emulsions. Spill Science and Technology Bulletin, 8 (5), 483 -

489.

239

Riddick, T. M. (1968). Control of colloid stability through Zeta Potential. Zeta-Meter,

Inc., New York, N.Y., USA.

Roux-Van der Merwe, M. P., Badenhorst, J. and Britz, T. J. (2005). Fungal treatment of

edible oil-containing industrial effluent. World Journal of Microbiology and

Biotechnology, 21(6-7), 947-953.

Ruthven, D.M. (1984). Principles of Adsorption and Adsorption Processes. Wiley, New

York, USA.

Saatci, Y., Arslan, E.I. and Konar, V. (2003). Removal of total lipids and fatty acids from

sunflower oil factory effluent by UASB reactor. Bioresource Technology, 87(3),

269-272.

Salminen, E., Rintala, J., Lokshina, L. and Vavilin, V.A. (2000). Anaerobic batch

degradation of solid poultry slaughterhouse waste. Water Science and Technology,

41(3), 33-41.

Schramm, L.L. (1992). Petroleum emulsions. In Schramm. L. L. (ed.) Emulsions,

fundamentals and applications in the petroleum industry. The American Chemical

Society, Washington DC, USA.

Scioli, C. and Vollaro, L. (1997). The use of Yarrowia lipolytica to reduce pollution in

olive mill wastewaters. Water Research, 31(10), 2520-2524.

Sherony, D.F. and Kintner, R.C. (1971a). Coalescence of an emulsion in a fibrous bed:

Part I, Theory. Canadian Journal of Chemical Engineering, 49(6), 314-320.

Sherony, D.F. and Kintner, R.C. (1971b). Coalescence of an emulsion in a fibrous bed:

Part II, Experiment. Canadian Journal of Chemical Engineering, 49(6), 321-325.

240

Riddick, T. M. (1968). Control of colloid stability through Zeta Potential. Zeta-Meter,

Inc., New York, N.Y., USA.

Roux-Van der Merwe, M. P., Badenhorst, J. and Britz, T. J. (2005). Fungal treatment of

edible oil-containing industrial effluent. World Journal of Microbiology and

Biotechnology, 21(6-7), 947-953.

Ruthven, D.M. (1984). Principles of Adsorption and Adsorption Processes. Wiley, New

York, USA.

Saatci, Y., Arslan, E.I. and Konar, V. (2003). Removal of total lipids and fatty acids from

sunflower oil factory effluent by UASB reactor. Bioresource Technology, 87(3),

269-272.

Salminen, E., Rintala, J., Lokshina, L. and Vavilin, V.A. (2000). Anaerobic batch

degradation of solid poultry slaughterhouse waste. Water Science and Technology,

41(3), 33-41.

Schramm, L.L. (1992). Petroleum emulsions. In Schramm. L. L. (ed.) Emulsions,

fundamentals and applications in the petroleum industry. The American Chemical

Society, Washington DC, USA.

Scioli, C. and Vollaro, L. (1997). The use of Yarrowia lipolytica to reduce pollution in

olive mill wastewaters. Water Research, 31(10), 2520-2524.

Sherony, D.F. and Kintner, R.C. (1971a). Coalescence of an emulsion in a fibrous bed:

Part I, Theory. Canadian Journal of Chemical Engineering, 49(6), 314—320.

Sherony, D.F. and Kintner, R.C. (1971b). Coalescence of an emulsion in a fibrous bed:

Part II, Experiment. Canadian Journal of Chemical Engineering, 49(6), 321-325.

240

Sherony, D.F., Kintner, R.C. and Wasan, D.T. (1978). Coalescence of secondary

emulsions in fibrous bed, in Matijevic. E. (Ed.), Surface and Colloid Science. Vol.

10, Plenum Press, New York, USA.

Silva-Tilak, V. (2002). Method of oil cleanup using coconut coir pith. US Patent and

Trademark office. 1258 (3), USA.

Sokolovic, R.M.S., Govedarica, D.D. and Sokolovic, D.S. (2010). Separation of oil-in-

water emulsions using two coalescers of different geometry. Journal of Hazardous

Materials, 175(1-3), 1001-1006.

Sparks, D.L., (1989). Kinetics of Soil Chemical Processes. Academic Press, Inc.,

California, USA.

Spielman, L.A. and Goren, S.L. (1970). Progress in induced coalescence and a new

theoretical framework for coalescence by porous media. Industrial and Engineering

Chemistry, 62, 10-24.

Srinivasan, A. and Viraraghavan, T. (2007). Biological processes for removal of oil from

wastewater- A review. Fresenius Environmental Bulletin, 16(12a), 1532-1543.

Srinivasan, A. and Viraraghavan, T. (2008). Removal of oil by walnut shell media.

Bioresource Technology, 99, 8217-8220.

Stams, A.G. and Oude, E.S.J. (1997). Understanding and advancing wastewater

treatment. Current Opinion in Biotechnology, 8(3), 328-334.

Statistica, (1997). Statistica for Windows, Release 5.1, Statsoft Inc., Tulsa, USA.

Sun, X-F., Sun, R-C. and Sun, J-X. (2002). Acetylation of rice straw with or without

catalysts and its characterization as a natural sorbent in oil spill cleanup. Journal of

241

Sherony, D.F., Kintner, R.C. and Wasan, D.T. (1978). Coalescence of secondary

emulsions in fibrous bed, in Matijevic. E. (Ed.), Surface and Colloid Science. Vol.

10, Plenum Press, New York, USA.

Silva-Tilak, V. (2002). Method of oil cleanup using coconut coir pith. US Patent and

Trademark office. 1258 (3), USA.

Sokolovic, R.M.S., Govedarica, D.D. and Sokolovic, D.S. (2010). Separation of oil-in-

water emulsions using two coalescers of different geometry. Journal of Hazardous

Materials, 175(1-3), 1001-1006.

Sparks, D.L., (1989). Kinetics of Soil Chemical Processes. Academic Press, Inc.,

California, USA.

Spielman, L.A. and Goren, S.L. (1970). Progress in induced coalescence and a new

theoretical framework for coalescence by porous media. Industrial and Engineering

Chemistry, 62, 10-24.

Srinivasan, A. and Viraraghavan, T. (2007). Biological processes for removal of oil from

wastewater- A review. Fresenius Environmental Bulletin, 16(12a), 1532-1543.

Srinivasan, A. and Viraraghavan, T. (2008). Removal of oil by walnut shell media.

Bioresource Technology, 99, 8217-8220.

Stams, A.G. and Oude, E.S.J. (1997). Understanding and advancing wastewater

treatment. Current Opinion in Biotechnology, 8(3), 328-334.

Statistica, (1997). Statistica for Windows, Release 5.1, Statsoft Inc., Tulsa, USA.

Sun, X-F., Sun, R-C. and Sim, J-X. (2002). Acetylation of rice straw with or without

catalysts and its characterization as a natural sorbent in oil spill cleanup. Journal of

241

Agricultural and Food Chemistry, 50 (22), 6428-6433.

Synoweicki, J. and Al-Khateeb, A.A.A.Q. (1997). Mycelia of Mucor rouxii as a source of

chitosan. Food Chemistry, 60(4), 605-610.

Tabakin, R.B., Trattner, R. and Cheremisinoff, P.N. (1978a). Oil/water separation

technology: The options available Part I. Water and Sewage Works, 125(7), 74 — 77.

Tabakin, R.B., Trattner, R. and Cheremisinoff, P.N. (1978b). Oil/water separation

technology: The options available Part II. Water and Sewage Works, 125(8), 72 — 75.

Takeno, K., Yamaoka, Y. and Sasaki, K. (2005). Treatment of oil-containing sewage

wastewater using immobilized photosynthetic bacteria. World Journal of

Microbiology and Biotechnology, 21(8-9), 1385-1391.

Tan, S., Tan, T., Wong, S. and Khor, E. (1996). The chitosan yield of zygomycetes at

their optimum harvesting time. Carbohydrate Polymers, 30 (4), 239-242.

Tano-Debrah, K., Fukuyama, S., Otonari, N., Taniguchi, F. and Ogura M. (1999). An

inoculum for the aerobic treatment of wastewaters with high concentrations of fats

and oils. Bioresource Technology, 69(2),133-139.

Tansel, B. and Pascual, B. (2004). Factorial evaluation of operational variables of a DAF

process to improve PHC removal efficiency. Desalination, 169, 1-10.

Tellez, G.T., Nirmalakhandan, N. and Gardea-Torresdey, J.L. (2005). Comparison of

purge and trap GC/MS and spectrophotometry for monitoring petroleum hydrocarbon

degradation in oilfield produced waters. Microchemical Journal, 81, 12-18.

Thomas, H.C. (1948). Chromatography: a problem in kinetics, Annals of the New York

Academy of Sciences, 49, 161-182.

242

Agricultural and Food Chemistry, 50 (22), 6428-6433.

Synoweicki, J. and Al-Khateeb, A.A.A.Q. (1997). Mycelia of Mucor rouxii as a source of

chitosan. Food Chemistry, 60(4), 605-610.

Tabakin, R.B., Trattner, R. and Cheremisinoff, P.N. (1978a). Oil/water separation

technology: The options available Part I. Water and Sewage Works, 125(7), 74 - 77.

Tabakin, R.B., Trattner, R. and Cheremisinoff, P.N. (1978b). Oil/water separation

technology: The options available Part II. Water and Sewage Works, 125(8), 72-75.

Takeno, K., Yamaoka, Y. and Sasaki, K. (2005). Treatment of oil-containing sewage

wastewater using immobilized photosynthetic bacteria. World Journal of

Microbiology and Biotechnology, 21(8-9), 1385-1391.

Tan, S., Tan, T., Wong, S. and Khor, E. (1996). The chitosan yield of zygomycetes at

their optimum harvesting time. Carbohydrate Polymers, 30 (4), 239-242.

Tano-Debrah, K., Fukuyama, S., Otonari, N., Taniguchi, F. and Ogura M. (1999). An

inoculum for the aerobic treatment of wastewaters with high concentrations of fats

and oils. Bioresource Technology, 69(2),133-139.

Tansel, B. and Pascual, B. (2004). Factorial evaluation of operational variables of a DAF

process to improve PHC removal efficiency. Desalination, 169,1-10.

Tellez, G.T., Nirmalakhandan, N. and Gardea-Torresdey, J.L. (2005). Comparison of

purge and trap GC/MS and spectrophotometry for monitoring petroleum hydrocarbon

degradation in oilfield produced waters. Microchemical Journal, 81,12-18.

Thomas, H.C. (1948). Chromatography: a problem in kinetics, Annals of the New York

Academy of Sciences, 49, 161-182.

242

Thun, R., Fagernas, L. and Brandt, J. (1983). Use of thermally treated peat for water

purification. Proceedings of the International Symposium on peat utilization, C.H.

Fuchsman and S.A. Spigarelli, (ed.), Bemidji State University, Minnesota, USA,

365-379.

Tobin, J.M., Cooper, D.G. and Neufeld, R.J. (1990). Investigation of the mechanism of

metal uptake by denatured Rhizopus arrhizus biomass. Enzyme and Microbial

Technology, 12, 591-595.

Tsezos, M. (1990). Engineering aspects of metal binding by biomass. In: Ehrlich, H.L.,

Brierley, C.L. (Eds.), Microbial Minerial Recovery. McGraw-Hill, New York,

USA.

Tyrie, C.C. and Caudle, D.D. (2007). Comparing oil in water measurement methods.

Exploration and Production — Oil and Gas Review, November (2), 31-35.

USFilter Corporation, 2005. Auto-ShellTM filter: walnut shell filtration.

<www.Zimpro.usfilter.com> (accessed 05.03.07).

Uzun, I., 2006. Kinetics of the adsorption of reactive dyes by chitosan. Dyes and

Pigments, 70(2), 76-83.

Varghese, B.K. and Cleveland, T.G. (1998). Kenaf as a deep-bed filter medium to

remove oil from oil-in-water emulsions. Separation Science and Technology, 33

(14), 2197-2220.

Vesala, A., Rosenholm, J.B., Laiho, S. (1985). Increasing the stability of vegetable oil

solutions with the aid of monoglycerides and a cosurfactant. Journal of the American

Oil Chemists' Society, 62(9), 1379-1385.

243

Thun, R., Fagernas, L. and Brandt, J. (1983). Use of thermally treated peat for water

purification. Proceedings of the International Symposium on peat utilization, C.H.

Fuchsman and S.A. Spigarelli, (ed.), Bemidji State University, Minnesota, USA,

365-379.

Tobin, J.M., Cooper, D.G. and Neufeld, R.J. (1990). Investigation of the mechanism of

metal uptake by denatured Rhizopus arrhizus biomass. Enzyme and Microbial

Technology, 12, 591-595.

Tsezos, M. (1990). Engineering aspects of metal binding by biomass. In: Ehrlich, H.L.,

Brierley, C.L. (Eds.), Microbial Minerial Recovery. McGraw-Hill, New York,

USA.

Tyrie, C.C. and Caudle, D.D. (2007). Comparing oil in water measurement methods.

Exploration and Production - Oil and Gas Review, November (2), 31-35.

USFilter Corporation, 2005. Auto-ShellTM filter: walnut shell filtration.

<www.Zimpro.usfilter.com> (accessed 05.03.07).

Uzun, I., 2006. Kinetics of the adsorption of reactive dyes by chitosan. Dyes and

Pigments, 70(2), 76-83.

Varghese, B.K. and Cleveland, T.G. (1998). Kenaf as a deep-bed filter medium to

remove oil from oil-in-water emulsions. Separation Science and Technology, 33

(14), 2197-2220.

Vesala, A., Rosenholm, J.B., Laiho, S. (1985). Increasing the stability of vegetable oil

solutions with the aid of monoglycerides and a cosurfactant. Journal of the American

Oil Chemists' Society, 62(9), 1379-1385.

243

http://www.Zimpro.usfilter.com

Viraraghavan, T. and Mathavan, G.N. (1990). Treatment of oily waters using peat. Water

Pollution Research Journal of Canada, 25 (1), 73-90.

Viraraghavan, T. and Mathavan, G.N. (1989). Peat filtration for oil removal. Proceedings

of the 1989 specialty conference, Environmental Engineering Division, ASCE,

Ausin, Texas, July 10-12, 635-642.

Viraraghavan, T. and Moazed, H. (2003). Removal of oil from water by bentonite.

Fresenius Environmental Bulletin, 12 (9), 1092-1097.

Volesky, B. (1993). Biosorption of heavy metals. CRC, Boca Raton, FL, USA.

Volesky, B. (2003). Sorption and Biosorption. BV Sorbex Inc., Quebec, Canada.

Volesky, B. and Agathos, S. (1974). Oil removal from refinery wastes by air flotation.

Water Pollution Research Journal of Canada, 25(1), 73-90.

Volesky, B. and Holan, Z.R. (1995). Biosorption of heavy metals. Biotechnology

Progress, 11, 235 — 250.

Wahaab R.A. and El-Awady. M.H. (1999). Anaerobic/aerobic treatment of meat

processing wastewater. The Environmentalist, 19(1), 61-65.

Wakelin, N.G. and Forster, C.F. (1997). An investigation into microbial removal of fats,

oils and greases. Bioresource Technology, 59(1), 37-43.

Wang, L., Hung, Y., Lo, H., and Yapijakis, C. (2004). Handbook of industrial treatment

and hazardous wastes treatment. Second edition, Marcel Dekker Inc., New York,

USA.

Wang, L.K., Tay, J-H., Tay, S. and Hung, Y-T. (2010). Environmental Bioengineering.

Humana Press, New York, USA.

244

Viraraghavan, T. and Mathavan, G.N. (1990). Treatment of oily waters using peat. Water

Pollution Research Journal of Canada, 25 (1), 73-90.

Viraraghavan, T. and Mathavan, G.N. (1989). Peat filtration for oil removal. Proceedings

of the 1989 specialty conference, Environmental Engineering Division, ASCE,

Ausin, Texas, July 10-12, 635-642.

Viraraghavan, T. and Moazed, H. (2003). Removal of oil from water by bentonite.


Volesky, B. (1993). Biosorption of heavy metals. CRC, Boca Raton, FL, USA.

Volesky, B. (2003). Sorption and Biosorption. BV Sorbex Inc., Quebec, Canada.

Volesky, B. and Agathos, S. (1974). Oil removal from refinery wastes by air flotation.

Water Pollution Research Journal of Canada, 25(1), 73-90.

Volesky, B. and Holan, Z.R. (1995). Biosorption of heavy metals. Biotechnology

Progress, 11, 235 - 250.

Wahaab R.A. and El-Awady. M.H. (1999). Anaerobic/aerobic treatment of meat

processing wastewater. The Environmentalist, 19(1), 61-65.

Wakelin, N.G. and Forster, C.F. (1997). An investigation into microbial removal of fats,

oils and greases. Bioresource Technology, 59(1), 37-43.

Wang, L., Hung, Y., Lo, H., and Yapijakis, C. (2004). Handbook of industrial treatment

and hazardous wastes treatment. Second edition, Marcel Dekker Inc., New York,

USA.

Wang, L.K., Tay, J-H., Tay, S. and Hung, Y-T. (2010). Environmental Bioengineering.

Humana Press, New York, USA.

244

Weber Jr, W.J. and DiGiano, F.A. (1996). Adsorption Processes. In: Process Dynamics in

Environmental Systems. John Wiley & Sons Inc., New York, USA.

Weber Jr., W. J. (1972). Physicochemical Processes for Water Quality Control, John

Wiley & Sons Inc., New York, USA.

Weber, W.J. and Morris, J.C. (1963). Kinetics of adsorption on carbon from solution, in:

Proceedings of the American Society of Civil Engineers, Journal of Sanitary

Engineering Division, 89 (SA2), 31-59.

White, S.A., Farina, P.R. and Fulton, I. (1979). Production and isolation of chitosan from

Mucor rouxii. Applied and Envionmental Microbiology, 38(2), 323-328.

Wiggins, E. J., Campbell, W. B. and Maass, 0. (1939). Determination of the specific

surface of fibrous materials. Canadian Journal of Research, 17B, 318-324.

Williams, P.A. (2007). Handbook of industrial water soluble polymers. Blackwell

Publishing, Oxford, UK.

Wolborska, A. (1989). Adsorption on activated carbon of p-nitrophenol from aqueous

solution. Water Research, 23,85-91.

Wu, T. Zivanovic, S., Draughon, F., Conway, W. and Sams, C. (2005). Physicochemical

Properties and Bioactivity of Fungal Chitin and Chitosan. Agriculture Food

Chemistry, 53 (10), 3888-3894.

Yan, G. and Viraraghavan, T. (2000). Effect of pretreatment on the bioadsorption of

heavy metals on Mucor rouxii. Water SA, 26(1) 119-124.

Yan, G. and Viraraghavan, T. (2001). Heavy metal removal in a biosorption column by

immobilized M rouxii biomass. Bioresource Technology, 78,243-249.

245

Weber Jr, WJ. and DiGiano, F.A. (1996). Adsorption Processes. In: Process Dynamics in

Environmental Systems. John Wiley & Sons Inc., New York, USA.

Weber Jr., W. J. (1972). Physicochemical Processes for Water Quality Control, John

Wiley & Sons Inc., New York, USA.

Weber, WJ. and Morris, J.C. (1963). Kinetics of adsorption on carbon from solution, in:

Proceedings of the American Society of Civil Engineers, Journal of Sanitary

Engineering Division, 89 (SA2), 31-59.

White, S.A., Farina, P.R. and Fulton, I. (1979). Production and isolation of chitosan from

Mucor rouxii. Applied and Envionmental Microbiology, 38(2), 323-328.

Wiggins, E. J., Campbell, W. B. and Maass, O. (1939). Determination of the specific

surface of fibrous materials. Canadian Journal of Research, 17B, 318-324.

Williams, P.A. (2007). Handbook of industrial water soluble polymers. Blackwell

Publishing, Oxford, UK.

Wolborska, A. (1989). Adsorption on activated carbon of /J-nitrophenol from aqueous

solution. Water Research, 23, 85-91.

Wu, T. Zivanovic, S., Draughon,F., Conway, W. and Sams, C. (2005). Physicochemical

Properties and Bioactivity of Fungal Chitin and Chitosan. Agriculture Food

Chemistry, 53 (10), 3888-3894.

Yan, G. and Viraraghavan, T. (2000). Effect of pretreatment on the bioadsorption of

heavy metals on Mucor rouxii. Water SA, 26(1) 119-124.

Yan, G. and Viraraghavan, T. (2001). Heavy metal removal in a biosorption column by

immobilized M. rouxii biomass. Bioresource Technology, 78, 243-249.

245

Yan, G. and Viraraghavan, T. (2003). Heavy-metal removal from aqueous solution by

fungus Mucor rouxii. Water Research, 37(18), 4486-4496.

Yan, G. and Viraraghavan, T. (2008). Mechanism of biosorption of heavy metals by

Mucor rouxii. Engineering in Life Sciences, 8(4), 363-371.

Yan, G., Viraraghavan, T. and Chen, M. (2001). A new model for heavy metal removal

in a biosorption column. Adsorption Science and Technology, 19 (1), 25-43.

Yang, Y., Zhang, X. and Wang, Z. (2002). Oilfield produced water treatment with

surface-modified fiber ball media filtration. Water Science and Technology, 46 (11-

12): 165-170.

Yesilada, D., Sik, S. and Sam, M. (1998). Biodegradation of olive oil mill wastewater by

Coriolus versicolor and Funalia trogii: effects of agitation, initial COD

concentration, inoculum size and immobilization. World Journal of Microbiology

and Biotechnology, 14(1), 37-42.

Yesilada, D., Sik, S. and Sam, M. (1999). Treatment of olive oil mill wastewater with

fungi. Turkish Journal of Biology, 23,231-240.

Yoon, Y.H. and Nelson, J.H. (1984). Application of gas adsorption kinetics. I. A

theoretical model for respirator cartridge service time. American Industrial Hygiene

Association Journal, 45,509-516.

Yoshihara, K., Yoshihiro, S., Hirotsu, T. and Izumoi, K. (2003). Chitosan productivity

enhancement in Rhizopus oryzae YPF-61A by D-Psicose. Journal of Bioscience

and Bioengineering, 95(3), 293-297.

246

Yan, G. and Viraraghavan, T. (2003). Heavy-metal removal from aqueous solution by

fungus Mucor rouxii. Water Research, 37(18), 4486-4496.

Yan, G. and Viraraghavan, T. (2008). Mechanism of biosorption of heavy metals by

Mucor rouxii. Engineering in Life Sciences, 8(4), 363-371.

Yan, G., Viraraghavan, T. and Chen, M. (2001). A new model for heavy metal removal

in a biosorption column. Adsorption Science and Technology, 19(1), 25-43.

Yang, Y., Zhang, X. and Wang, Z. (2002). Oilfield produced water treatment with

surface-modified fiber ball media filtration. Water Science and Technology, 46 (11-

12): 165-170.

Yesilada, D., Sik, S. and Sam, M. (1998). Biodegradation of olive oil mill wastewater by

Coriolus versicolor and Funalia trogii: effects of agitation, initial COD

concentration, inoculum size and immobilization. World Journal of Microbiology

and Biotechnology, 14(1), 37-42.

Yesilada, D., Sik, S. and Sam, M. (1999). Treatment of olive oil mill wastewater with

fungi. Turkish Journal of Biology, 23, 231-240.

Yoon, Y.H. and Nelson, J.H. (1984). Application of gas adsorption kinetics. I. A

theoretical model for respirator cartridge service time. American Industrial Hygiene

Association Journal, 45, 509-516.

Yoshihara, K., Yoshihiro, S., Hirotsu, T. and Izumoi, K. (2003). Chitosan productivity

enhancement in Rhizopus oryzae YPF-61A by D-Psicose. Journal of Bioscience

and Bioengineering, 95(3), 293-297.

246

Young, J.C. (1979). Removal of grease and oil by biological treatment processes. Journal

Water Pollution Control Federation, 51(8), 2071— 2087.

Zheng, S., Yang, M., Park, Y.H. and Liu, F. (2003). Washout of a yeast population

during continuous treatment of salad-oil-manufacturing wastewater. Bioresource

Technology, 86(3), 235-237.

Zouboulis, A.I. and Avranas, A. (2000). Treatment of oil-in-water emulsions by

coagulation and dissolved-air flotation. Colloids and Surfaces A: Physicochemical

and Engineering Aspects, 172 (1-3), 153-161.

Zunan, Q., Yiz, Z. and Yucliao, F. (1995). Removal of oil from concentrated wastewater

by attapulgite and coagulant. Water Quality Research Journal of Canada, 30(1),

89-99.

247

Young, J.C. (1979). Removal of grease and oil by biological treatment processes. Journal

Water Pollution Control Federation, 51(8), 2071- 2087.

Zheng, S., Yang, M., Park, Y.H. and Liu, F. (2003). Washout of a yeast population

during continuous treatment of salad-oil-manufacturing wastewater. Bioresource

Technology, 86(3), 235-237.

Zouboulis, A.I. and Avranas, A. (2000). Treatment of oil-in-water emulsions by

coagulation and dissolved-air flotation. Colloids and Surfaces A: Physicochemical

and Engineering Aspects, 172(1-3), 153-161.

Zunan, Q., Yiz, Z. and Yuqiao, F. (1995). Removal of oil from concentrated wastewater

by attapulgite and coagulant. Water Quality Research Journal of Canada, 30(1),

89-99.

247

Appendix A

Data and Supplementary Figures

Table A.1: Batch kinetic data for SMO

Time (h) Adsorption capacity, Q mg/g

5 °C 15 °C 22 °C 30 °C

0 0 0 0 0 0.25 76.15 75.5 74 78 0.5 90.75 85 79.1 87 0.75 92.1 86.5 80.75 95 1 93.3 91 82.5 98.7 1.25 94.85 92.5 83.6 98.8 1.5 95.25 94 83.9 98.85 1.75 96.8 95.5 85.8 98.8 2 97.15 97 97.95 98.7 2.25 97.55 98.6 97.65 98.85 2.5 97.5 98.65 97.85 98.95 2.75 98.9 98.55 97.7 98.9 3 98.8 98.6 97.7 98.8 3.5 98.85 98.6 97.7 98.85 4 98.9 98.55 97.75 98.75 4.5 98.85 98.65 97.7 98.8 5 98.95 98.6 97.65 98.85 6 98.9 98.55 97.7 98.95

248

Appendix A

Data and Supplementary Figures

Table A. 1: Batch kinetic data for SMO

Time (h) Adsorption capacity, Q mg/g

5 °C 15 °C 22 °C 30 °C

0 0 0 0 0

0.25 76.15 75.5 74 78

0.5 90.75 85 79.1 87

0.75 92.1 86.5 80.75 95

1 93.3 91 82.5 98.7

1.25 94.85 92.5 83.6 98.8

1.5 95.25 94 83.9 98.85

1.75 96.8 95.5 85.8 98.8

2 97.15 97 97.95 98.7

2.25 97.55 98.6 97.65 98.85

2.5 97.5 98.65 97.85 98.95

2.75 98.9 98.55 97.7 98.9

3 98.8 98.6 97.7 98.8

3.5 98.85 98.6 97.7 98.85

4 98.9 98.55 97.75 98.75

4.5 98.85 98.65 97.7 98.8

5 98.95 98.6 97.65 98.85

6 98.9 98.55 97.7 98.95

248

Table A.2: Batch kinetic data for CO Time (h) Adsorption capacity, Q mg/g

5 °C 15 °C 22 °C 30 °C

0 0 0 0 0 0.25 96.8 96 91 95 0.5 97.4 97 92.5 98 0.75 98.2 98 93 98.05 1 98.25 97.95 94 98.1 1.25 98.3 98.05 95 97.95 1.5 98.25 98.05 96 98 1.75 98.3 98.1 98 98.05 2 98.3 98 98.05 98.1 2.25 98.25 98.05 97.95 98.05 2.5 98.3 98.1 98.05 98 2.75 98.3 98.05 98.1 98.05 3 98.3 98 98 98 3.5 98.2 97.95 98.05 97.95 4 98.25 98 98.1 98.05 4.5 98.3 98.05 98.1 98 5 98.3 98.1 98.15 98.05 6 98.25 98.05 98 98.05

249

Table A.2: Batch kinetic data for CO

Time (h) Adsorption capacity, 0 mg/g

5 °C 15 °C 22 °C 30 °C

0 0 0 0 0 0.25 96.8 96 91 95 0.5 97.4 97 92.5 98 0.75 98.2 98 93 98.05 1 98.25 97.95 94 98.1 1.25 98.3 98.05 95 97.95 1.5 98.25 98.05 96 98 1.75 98.3 98.1 98 98.05 2 98.3 98 98.05 98.1 2.25 98.25 98.05 97.95 98.05 2.5 98.3 98.1 98.05 98 2.75 98.3 98.05 98.1 98.05 3 98.3 98 98 98 3.5 98.2 97.95 98.05 97.95 4 98.25 98 98.1 98.05 4.5 98.3 98.05 98.1 98 5 98.3 98.1 98.15 98.05 6 98.25 98.05 98 98.05

249

Table A.3: Batch kinetic data for Bright-Edge 80 Time (h) Adsorption capacity, Q (mg/g)

5 °C 15 °C 22 °C 30 °C 0 0 0 0 0 0.25 95.75 95 94 96.5 0.5 97.5 96.5 95 97.75 0.75 97.65 97.5 96 97.9 1 97.95 98 97.65 97.95 1.25 98 97.95 97.4 98 1.5 97.95. 98 97.55 97.95 1.75 98 98 97.8 98 2 97.95 97.95 97.95 98 2.25 98.05 97.95 98 97.95 2.5 98 98 98 98 2.75 98 97.95 97.95 97.95 3 97.95 98 98 98 3.5 98 98 97.9 97.95 4 97.95 97.95 97.95 98 4.5 98 97.95 98 98.05 5 98 98 97.95 98 6 98 98 98 98

Table A.4: Batch isotherm data for SMO Dose (g) Adsorption capacity, Q (mg/g)

5 °C 15 °C 22 °C 30 °C 0.03 593.33 593.33 573.33 596.67 0.04 465 465 457.5 465 0.06 318.33 318.33 315 316.67 0.08 241.5 241.25 240.63 241.25 0.1 194.5 194 194 194 0.2 98 97.6 97.75 97.75 0.4 . 49.13 48.88 48.93 49

250

Table A.3: Batch kinetic data for Bright-Edge 80

Time (h) Adsorption capacity, Q (mg/g) 5 °C 15 °C 22 °C 30 °C

0 0 0 0 0 0.25 95.75 95 94 96.5

0.5 97.5 96.5 95 97.75 0.75 97.65 97.5 96 97.9

1 97.95 98 97.65 97.95

1.25 98 97.95 97.4 98 1.5 97.95 98 97.55 97.95 1.75 98 98 97.8 98

2 97.95 97.95 97.95 98 2.25 98.05 97.95 98 97.95 2.5 98 98 98 98 2.75 98 97.95 97.95 97.95 3 97.95 98 98 98 3.5 98 98 97.9 97.95

4 97.95 97.95 97.95 98 4.5 98 97.95 98 98.05

5 98 98 97.95 98 6 98 98 98 98

Table A.4: Batch isotherm data for SMO Dose (g) Adsorption capacity, Q (mg/g)

5 °C 15 °C 22 °C 30 °C

0.03 593.33 593.33 573.33 596.67 0.04 465 465 457.5 465 0.06 318.33 318.33 315 316.67 0.08 241.5 241.25 240.63 241.25 0.1 194.5 194 194 194 0.2 98 97.6 97.75 97.75 0.4 . 49.13 48.88 48.93 49

250

Table A.5: Batch isotherm data for CO Dose (g) Adsorption capacity, Q (mg/g)

5 °C 15 °C 22 °C 30 °C 0.03 590 593.33 600 580 0.04 465 462.5 467.5 457.5 0.06 318.33 316.67 318.33 316.67 0.08 241.25 240 241.25 241.25 0.1 194 193 194 194 0.2 98 97.5 97.75 97.9 0.4 49.25 49 49 49

Table A.6: Batch isotherm data for Bright-Edge 80 Dose (g) I Adsorption capacity, Q mg/g

5 °C 15 °C 22 °C 30 °C 0.03 590 593.33 600 580 0.04 465 462.5 467.5 457.5 0.06 318.33 316.67 318.33 316.67 0.08 241.25 240 241.25 241.25 0.1 194 193 194 194 0.2 98 97.5 97.75 97.9 0.4 49.25 49 49 49

Table A.7: Desorption data for SMO

Mass of Oil Oil retained Oil Oil Percent biomass, g removed,

mg/L in the biomass

removed after

recovered in solution

recovery of oil after

after adsorption, mg/L

desorption, mg/L.

after desorption, mg/L

desorption

0.03 28 172 6 0.035 3.489 0.04 17 183 7 0.038 3.825 0.06 11 189 8 0.042 4.232 0.08 7.5 192.5 10.5 0.055 5.455 0.1 6 194 12.5 0.064 6.443 0.2 4.5 195.5 22 0.112 11.253 0.4 4.3 195.7 28 0.143 14.308

251

Table A.5: Batch isotherm data for CO Dose (g) Adsorption capacity, 0 (mg/g)

5 °C 15 °C 22 °C 30 °C

0.03 590 593.33 600 580 0.04 465 462.5 467.5 457.5 0.06 318.33 316.67 318.33 316.67 0.08 241.25 240 241.25 241.25 0.1 194 193 194 194

0.2 98 97.5 97.75 97.9 0.4 49.25 49 49 49

Table A.6: Batch isotherm data for Bright-Edge 80

Dose (g) Adsorption capacity, Q mg/g 5 °C 15 °C 22 °C 30 °C

0.03 590 593.33 600 580 0.04 465 462.5 467.5 457.5 0.06 318.33 316.67 318.33 316.67 0.08 241.25 240 241.25 241.25 0.1 194 193 194 194 0.2 98 97.5 97.75 97.9 0.4 49.25 49 49 49

Table A.7: Desorption data for SMO

Mass of Oil Oil retained Oil Oil Percent biomass, g removed, in the removed recovered in recoveiy

mg/L biomass after solution of oil after after desorption, after desorption adsorption, mg/L. desorption, mg/L mg/L

0.03 28 172 6 0.035 3.489

0.04 17 183 7 0.038 3.825

0.06 11 189 8 0.042 4.232

0.08 7.5 192.5 10.5 0.055 5.455

0.1 6 194 12.5 0.064 6.443

0.2 4.5 195.5 22 0.112 11.253

0.4 4.3 195.7 28 0.143 14.308

251

Table A.8: Desorption data for CO

Mass of Oil Oil retained Oil Oil recovered Percent biomass, g removed,

mg/L in the biomass after

removed after

in solution after


adsorption, mg/L

desorption, mg/L

desorption, m/L

desorption

0.03 20 180 6 0.033 3.333 0.04 13 187 7 0.037 3.743 0.06 9 191 9 0.047 4.712 0.08 7 193 11 0.057 5.699 0.1 6 194 12.5 0.064 6.443 0.2 4.5 195.5 24 0.123 12.276 0.4 4 196 36 0.184 18.367

Table A.9: Desorption data for Bright-Edge 80

Mass of Oil Oil retained Oil Oil recovered Percent biomass, g removed,

mg/L in the biomass after

removed after

in solution after


adsorption, mg/L

desorption, mg/L

desorption, mg/L

desorption

0.03 38 162 6 0.037 3.704 0.04 24 176 7 0.039 3.977 0.06 16 184 9 0.049 4.891 0.08 12 188 11 0.059 5.851 0.1 9 191 12 0.063 6.283 0.2 5.8 194.2 18 0.093 9.269 0.4 5 195 23 0.118 11.795

252

Table A. 8: Desorption data for CO

Mass of Oil Oil retained Oil Oil recovered Percent biomass, g removed, in the removed in solution recovery of

mg/L biomass after after after oil after adsorption, desorption, desorption, desorption mg/L mg/L mg/L

0.03 20 180 6 0.033 3.333

0.04 13 187 7 0.037 3.743

0.06 9 191 9 0.047 4.712

0.08 7 193 11 0.057 5.699

0.1 6 194 12.5 0.064 6.443

0.2 4.5 195.5 24 0.123 12.276

0.4 4 196 36 0.184 18.367

Table A.9: Desorption data for Bright-Edge 80

Mass of Oil Oil retained Oil Oil recovered Percent biomass, g removed, in the removed in solution recovery of

mg/L biomass after after after oil after adsorption, desorption, desorption, desorption mg/L mg/L mg/L

0.03 38 162 6 0.037 3.704

0.04 24 176 7 0.039 3.977

0.06 16 184 9 0.049 4.891

0.08 12 188 11 0.059 5.851

0.1 9 191 12 0.063 6.283

0.2 5.8 194.2 18 0.093 9.269

0.4 5 195 23 0.118 11.795

252

Table A.10: Batch study with immobilized biomass beads for 6 h

Oil pH Mass of immobilized beads, g

Initial concentration of oil, mg/L

Final concentration, mg/L

Percent oil removal

SMO 3.0 0.2 200 32 89 SMO 7.6 0.2 200 150 25 CO 3.0 0.2 200 25 87.5 CO 7.6 0.2 200 48 75 Bright-Edge 80 3.0 0.2 200 28 86 Bright-Edge 80 7.6 0.2 200 138 39

Table A.11: Batch kinetic studies with immobilized biomass beads at pH 3.0

Oil Oil concentration, mg/L 0.25h 0.5h lh 2h 3h 4h 5h 6h

SMO 105 82 64 45 37 35 32 32 CO 76 61 48 39 26 24 25 25 Bright- Edge 80 92 83 74 51 36

29 28 ...

28

253

Table A. 10: Batch study with immobilized biomass beads for 6 h

Oil PH Mass of Initial Final Percent immobilized concentration concentration, oil beads, g of oil, mg/L mg/L removal

SMO 3.0 0.2 200 32 89

SMO 7.6 0.2 200 150 25

CO 3.0 0.2 200 25 87.5

CO 7.6 0.2 200 48 75

Bright-Edge 80 3.0 0.2 200 28 86

Bright-Edge 80 7.6 0.2 200 138 39

Table A.l 1: Batch kinetic studies with immobilized biomass beads at pH 3.0

Oil Oil concentration, mg/L

0.25 h 0.5 h 1 h 2 h 3 h 4 h 5 h 6 h

SMO 105 82 64 45 37 35 32 32

CO 76 61 48 39 26 24 25 25

Bright-Edge 80

92 83 74 51 36 29 28 28

253

Table A.12: Column breakthrough data

Volume (L) SMO concentration (mgt-)

CO concentration (mg/-)

Bright-Edge 80concentration (mg/L)

0.013 1.3 4.8 4.2 0.039 3.8 7.8 8.4 0.078 4.8 10.2 11.6 0.156 5.5 13.4 16.9 0.312 7.8 15.1 18.2 0.468 10.7 17 19.5 0.624 16.9 20.2 20.4 0.78 20.7 23.1 22.9 0.936 23.1 26.7 25.2 1.092 26 31.8 28.5 1.248 30.6 35.2 32.1 1.872 33.4 38.5 36 2.496 36.1 41 40.5 3.12 39.5 44.9 44.2 3.744 42.1 48.2 47.8 4.68 43.7 48.1 5.616 45.4 6.552 48

254

Table A. 12: Column breakthrough data

Volume (L) SMO concentration (mg/L)

CO concentration (mg/L)

Bright-Edge 80 concentration (mg/L)

0.013 1.3 4.8 4.2

0.039 3.8 7.8 8.4

0.078 4.8 10.2 11.6

0.156 5.5 13.4 16.9

0.312 7.8 15.1 18.2

0.468 10.7 17 19.5

0.624 16.9 20.2 20.4

0.78 20.7 23.1 22.9

0.936 23.1 26.7 25.2

1.092 26 31.8 28.5

1.248 30.6 35.2 32.1

1.872 33.4 38.5 36

2.496 36.1 41 40.5

3.12 39.5 44.9 44.2

3.744 42.1 48.2 47.8

4.68 43.7 48.1

5.616 45.4

6.552 48

254

Table A.13: Column desorption data for SMO Volume (L) SMO Concentration (mg/L)

0.013 310 0.026 380 0.039 321 0.052 202 0.078 174 0.104 116 0.13 74.5

0.156 45.2 0.195 20.7 0.312 18.9 0.468 16.3 0.624 14.5 0.78 11.8 0.936 8.3 1.092 6.1 1.248 4.9

Table A.14: Column desorption data for CO Volume (L) CO concentration (mg/L)

0.013 154.5 0.026 132.7 0.039 110.1 0.052 98.7 0.078 51.4 0.104 28.6 0.13 20.5

0.156 16.8 0.195 14.1 0.312 12.3 0.468 11.7 0.624 10.9 0.702 10.3 0.78 9.8

255

Table A. 13: Column desorption data for SMO Volume (L) SMO Concentration (mg/L)

0.013 310 0.026 380 0.039 321 0.052 202 0.078 174 0.104 116 0.13 74.5

0.156 45.2 0.195 20.7 0.312 18.9 0.468 16.3 0.624 14.5 0.78 11.8 0.936 8.3 1.092 6.1 1.248 4.9

Table A. 14: Column desorption data for CO

Volume (L) CO concentration (mg/L)

0.013 154.5 0.026 132.7 0.039 110.1 0.052 98.7 0.078 51.4 0.104 28.6 0.13 20.5

0.156 16.8 0.195 14.1 0.312 12.3 0.468 11.7 0.624 10.9 0.702 10.3 0.78 9.8

255

Table A.15: Column desorption data for Bright-Edge 80

Volume (L) Bright-Edge 80 concentration (mg/L)

0.013 142.6 0.026 129.4 0.039 107.4 0.052 101 0.078 64.3 0.104 31.2 0.13 19.8

0.156 17.7 0.195 15.8 0.312 15.1 0.468 13.4 0.624 12.7 0.78 11.3

0.858 10.1 0.936 9.6

Table A.16: Column breakthrough second run data for SMO Volume (L) SMO concentration (mg/L)

0.013 5 0.039 19 0.078 22 0.156 26.7 0.312 34.1 0.468 36.3 0.624 37.8 0.78 39.2 0.936 40.9 1.092 41.7 1.248 42.9 1.872 44.1 2.496 45 3.12 47.1

3.744 48.2

256

Table A. 15: Column desorption data for Bright-Edge 80


0.013 142.6

0.026 129.4

0.039 107.4

0.052 101

0.078 64.3

0.104 31.2

0.13 19.8

0.156 17.7

0.195 15.8

0.312 15.1

0.468 13.4

0.624 12.7

0.78 11.3

0.858 10.1

0.936 9.6

Table A. 16: Column breakthrough second run data for SMO

Volume (L) SMO concentration (mg/L)

0.013 5

0.039 19

0.078 22

0.156 26.7

0.312 34.1

0.468 36.3

0.624 37.8

0.78 39.2

0.936 40.9

1.092 41.7

1.248 42.9

1.872 44.1

2.496 45

3.12 47.1

3.744 48.2

256

Table A.17: Column breakthrough second run data for CO Volume (L) CO concentration (mg/L)

0.013 12 0.039 28.4 0.078 31.8 0.156 34.2 0.312 35.6 0.468 37.7 0.624 39.1 0.78 41.5 0.936 42.9 1.092 44.1 1.248 45.5 1.872 46.8 2.184 48

Table A.18: Column breakthrough second run data for Bright-Edge 80 Volume (L) Bright-Edge 80 concentration (mg/L)

0.013 10 0.039 27.1 0.078 29.8 0.156 33.2 0.312 34.9 0.468 36.7 0.624 39.1 0.78 40.9

0.936 41.7 1.092 42.9 1.248 43.7 1.872 45.3 2.496 46.4 2.81 47.6

257

Table A. 17: Column breakthrough second run data for CO

Volume (L) CO concentration (mg/L)

0.013 12

0.039 28.4

0.078 31.8

0.156 34.2

0.312 35.6

0.468 37.7

0.624 39.1

0.78 41.5

0.936 42.9

1.092 44.1

1.248 45.5

1.872 46.8

2.184 48

Table A. 18: Column breakthrough second run data for Bright-Edge 80


0.013 10

0.039 27.1

0.078 29.8

0.156 33.2

0.312 34.9

0.468 36.7

0.624 39.1

0.78 40.9

0.936 41.7

1.092 42.9

1.248 43.7

1.872 45.3 2.496 46.4

2.81 47.6

257

Table A.19: Experimental and predicted head loss for single-phase flow Flow rate (mL/min) Depth of bed (m)

Experimental head loss (m of water)

Predicted head loss (m of water)

12 0.2 0.4 0.6 0.8 1

0.020 0.041 0.063 0.084 0.098

0.022 0.044 0.066 0.088 0.110

16 0.2 0.4 0.6 0.8 1

0.028 0.052 0.076 0.102 0.125

0.029 0.058 0.088 0.117 0.146

20 0.2 0.4 0.6 0.8 1

0.034 0.062 0.097 0.129 0.156

0.037 0.073 0.101 0.146 0.183

24 0.2 0.4 0.6 0.8 1

0.042 0.079 0.112 0.174 0.201

0.044 0.088 0.131 0.175 0.219

28 0.2 0.4 0.6 0.8 1

0.049 0.093 0.145 0.175 0.23

0.051 0.102 0.153 0.204 0.256

32 0.2 0.4 0.6 0.8 1

0.061 0.126 0.168 0.209 0.276

0.058 0.117 0.175 0.234 0.292

258

Table A. 19: Experimental and predicted head loss for single-phase flow

Flow rate (mL/min)

Depth of bed (m) Experimental head loss (m of water)


12 0.2 0.020 0.022

0.4 0.041 0.044

0.6 0.063 0.066

0.8 0.084 0.088

1 0.098 0.110

16 0.2 0.028 0.029

0.4 0.052 0.058

0.6 0.076 0.088

0.8 0.102 0.117

1 0.125 0.146

20 0.2 0.034 0.037

0.4 0.062 0.073

0.6 0.097 0.101

0.8 0.129 0.146

1 0.156 0.183

24 0.2 0.042 0.044

0.4 0.079 0.088

0.6 0.112 0.131

0.8 0.174 0.175

1 0.201 0.219

28 0.2 0.049 0.051

0.4 0.093 0.102

0.6 0.145 0.153

0.8 0.175 0.204

1 0.23 0.256

32 0.2 0.061 0.058

0.4 0.126 0.117

0.6 0.168 0.175

0.8 0.209 0.234

1 0.276 0.292

258

Table A.20: Experimental and predicted head loss for two-phase flow Flow rate (mL/min) Depth of bed (m)

Experimental head loss (m of water)


12 0.2 0.4 0.6 0.8 1

0.036 0.068 0.096 0.128 0.162

0.022 0.043 0.065 0.087 0.108

16 0.2 0.4 0.6 0.8 1

0.049 0.097 0.129 0.177 0.219

0.034 0.069 0.103 0.137 0.171

20 0.2 0.4 0.6 0.8 1

0.061 0.111 0.166 0.224 0.281

0.051 0.101 0.152 0.202 0.253

24 0.2 0.4 0.6 0.8 1

0.078 0.139 0.201 0.281 0.332

0.065 0.129 0.194 0.259 0.324

28 0.2 0.4 0.6 0.8 1

0.089 0.162 0.241 0.32

0.392

0.082 0.164 0.247 0.329 0.411

32 0.2 0.4 0.6 0.8 1

0.156 0.253 0.336 0.421 0.454

0.099 0.199 0.298 0.397 0.497

259

Table A.20: Experimental and predicted head loss for two-phase flow

Flow rate (mL/min)

Depth of bed (m) Experimental head loss (m of water)


12 0.2 0.036 0.022

0.4 0.068 0.043

0.6 0.096 0.065

0.8 0.128 0.087

1 0.162 0.108 16 0.2 0.049 0.034

0.4 0.097 0.069 0.6 0.129 0.103

0.8 0.177 0.137

1 0.219 0.171 20 0.2 0.061 0.051

0.4 0 .111 0.101

0.6 0.166 0.152

0.8 0.224 0.202 1 0.281 0.253

24 0.2 0.078 0.065 0.4 0.139 0.129

0.6 0.201 0.194 0.8 0.281 0.259

1 0.332 0.324 28 0.2 0.089 0.082

0.4 0.162 0.164 0.6 0.241 0.247 0.8 0.32 0.329

1 0.392 0.411 32 0.2 0.156 0.099

0.4 0.253 0.199 0.6 0.336 0.298 0.8 0.421 0.397

1 0.454 0.497

259

Table A.21: DroD diameters m for 12 mL/min Time (h) 200 mm 400 mm 600 mm 800 mm 1000 mm

2 9.4 9.3 11.2 9.8 9.7 4 9.7 9.6 9.1 9.7 9.9 6 10.5 9.8 8.7 9.1 10 8 13.6 12.2 10.5 10.9 11.2

10 11.3 11 9.6 9.9 10.2 12 10.9 10.1 8.9 9.7 9.8 16 10.6 10 8.7 9.4 9.9 20 13.5 12.8 10.6 11.3 11.6 24 11.1 10.8 9.8 10.3 10.9 28 14.1 13 10.6 11.5 11.8 32 11.8 11.1 10.2 10.8 11.1 38 10.9 10.1 9.1 9.8 10.1 44 13.1 12.3 10.5 11.1 11.7 50 11.3 10.9 9.6 10.1 10.7

Table A.22: Drop density no./ cm3) for 12 mL/min Time (h) 200 mm 400 mm 600 mm 800 mm 1000 mm

2 151 147 143 139 135 4 135 . 131 128 124 121 6 149 142 136 131 124 8 161 151 142 133 123 10 158 149 139 128 120 12 164 153 144 136 125 16 142 132 130 136 118 20 149 138 137 141 125 24 152 142 139 146 126 28 146 137 134 140 122 32 143 134 131 138 119 38 138 131 128 134 118 44 134 127 124 129 115 50 132 124 122 127 112

260

Table A.21: Drop diameters (urn) for 12 mL/min

200 mm 400 mm 600 mm 800 mm 1000 mm

2 9.4 9.3 11.2 9.8 9.7

4 9.7 9.6 9.1 9.7 9.9

6 10.5 9.8 8.7 9.1 10

8 13.6 12.2 10.5 10.9 11.2

10 11.3 11 9.6 9.9 10.2

12 10.9 10.1 8.9 9.7 9.8

16 10.6 10 8.7 9.4 9.9

20 13.5 12.8 10.6 11.3 11.6

24 11.1 10.8 9.8 10.3 10.9

28 14.1 13 10.6 11.5 11.8

32 11.8 11.1 10.2 10.8 11.1

38 10.9 10.1 9.1 9.8 10.1

44 13.1 12.3 10.5 11.1 11.7

50 11.3 10.9 9.6 10.1 10.7

Table A.22: Drop density (no./ cm3) for 12 mL/min

Time (h) 200 mm 400 mm 600 mm 800 mm 1000 mm

2 151 147 143 139 135

4 135 131 128 124 121

6 149 142 136 131 124

8 161 151 142 133 123

10 158 149 139 128 120

12 164 153 144 136 125

16 142 132 130 136 118

20 149 138 137 141 125

24 152 142 139 146 126

28 146 137 134 140 122

32 143 134 131 138 119

38 138 131 128 134 118

44 134 127 124 129 115

50 132 124 122 127 112

260

Table A.23: Coalescence efficiency for 12 mL/min Bed length (mm)

Average drop diameter (gm)

Average drop density (no./cm3) Coalescence efficiency

200 11.56 146.71 0.159 400 10.93 138.43 0.086 600 9.79 134.07 0.060 800 10.24 134.43 0.045

1000 10.61 121.64 0.041

Table A.24: Dron diameters m for 16 mL/min Time (h) 200 mm 400 mm 600 mm 800 mm 1000 mm

2 11.3 10 10.8 10.2 9.9 4 11.8 11.1 10.6 10.9 10.4 6 13.9 12.9 12.3 12.7 12.2 8 11.1 10.9 10.5 10.8 10.2

10 10.8 10.4 10.1 10.3 9.9 12 10.4 10.1 9.8 9.9 9.4 16 11.6 10.8 10.5 10.7 10.4 20 13.1 12.3 11.5 11.9 11.2 24 12.2 12 11.2 11.6 11 28 11.5 10.9 10.5 10.8 10.3 32 11.8 11.1 10.7 10.9 10.5 38 12.7 12 11.4 11.7 11.1

Table A.25 Drop density (no./ cm3) for 16 mL/min Time (h) 200 mm 400 mm 600 mm 800 mm 1000 mm

2 145 138 130 123 116 4 141 135 127 119 113 6 139 132 125 117 112 8 142 136 128 121 114 10 146 138 131 124 117 12 132 121 112 101 92 16 135 123 115 102 96 20 138 129 121 113 105 24 140 132 125 117 110 28 136 127 119 112 103 32 133 124 116 107 99 38 131 123 112 103 91

261

Table A.23: Coalescence efficiency 'or 12 mL/min

Bed length (mm)

Average drop diameter (urn)


200 11.56 146.71 0.159

400 10.93 138.43 0.086

600 9.79 134.07 0.060

800 10.24 134.43 0.045

1000 10.61 121.64 0.041



2 11.3 10 10.8 10.2 9.9

4 11.8 11.1 10.6 10.9 10.4

6 13.9 12.9 12.3 12.7 12.2

8 11.1 10.9 10.5 10.8 10.2

10 10.8 10.4 10.1 10.3 9.9

12 10.4 10.1 9.8 9.9 9.4

16 11.6 10.8 10.5 10.7 10.4

20 13.1 12.3 11.5 11.9 11.2

24 12.2 12 11.2 11.6 11

28 11.5 10.9 10.5 10.8 10.3

32 11.8 11.1 10.7 10.9 10.5

38 12.7 12 11.4 11.7 11.1



2 145 138 130 123 116 4 141 135 127 119 113

6 139 132 125 117 112 8 142 136 128 121 114 10 146 138 131 124 117 12 132 121 112 101 92 16 135 123 115 102 96 20 138 129 121 113 105 24 140 132 125 117 110 28 136 127 119 112 103 32 133 124 116 107 99 38 131 123 112 103 91

261

Table A.26: Coalescence efficiency for 16 mL/min

Bed length (mm) Average drop diameter (pm)


200 11.85 138.17 0.12 400 11.21 129.83 0.067 600 10.83 121.75 0.048 800 11.03 113.25 0.039

1000 10.54 105.67 0.034

Table A.27: Dron diameters(um) for 20 mL/min Time (h) 200 mm 400 mm 600 mm 800 mm 1000 mm

2 13.2 9.7 9.2 9.5 9 4 12.7 9.5 9 9.2 8.8 6 11 9.6 9 9.3 8.9 8 10.7 11.3 11 11.2 10.8

10 11.3 11.3 11.1 11.2 11 12 13.6 12.8 12.2 12.5 11.8 16 10.5 10.4 10.1 10.3 10 20 10.9 10.8 10.7 10.8 10.5 24 10.2 10.5 10.2 10.4 10 28 9.7 10.1 9.7 10 9.9 32 9.4 9.8 9.5 9.8 9.7

Table A.28 Drop density (no./ cm3) for 20 mL/min Time (h) 200 mm 400 mm 600 mm 800 mm 1000 mm

2 144 136 129 132 115 4 142 135 127 130 112 6 145 138 131 134 118 8 139 133 124 129 109 10 138 131 122 125 106 12 146 137 132 131 118 16 127 120 110 117 94 20 132 127 117 125 103 24 131 125 114 122 98 28 129 120 112 119 95 32 125 119 108 115 91

262


Bed length (mm) Average drop diameter (nm)


200 11.85 138.17 0.12

400 11.21 129.83 0.067

600 10.83 121.75 0.048

800 11.03 113.25 0.039

1000 10.54 105.67 0.034

Table A.27: Drop diameters ("nm) for 20 mL/min


2 13.2 9.7 9.2 9.5 9

4 12.7 9.5 9 9.2 8.8

6 11 9.6 9 9.3 8.9

8 10.7 11.3 11 11.2 10.8

10 11.3 11.3 11.1 11.2 11

12 13.6 12.8 12.2 12.5 11.8

16 10.5 10.4 10.1 10.3 10

20 10.9 10.8 10.7 10.8 10.5

24 10.2 10.5 10.2 10.4 10 28 9.7 10.1 9.7 10 9.9

32 9.4 9.8 9.5 9.8 9.7



2 144 136 129 132 115

4 142 135 127 130 112 6 145 138 131 134 118 8 139 133 124 129 109 10 138 131 122 125 106 12 146 137 132 131 118 16 127 120 110 117 94 20 132 127 117 125 103

24 131 125 114 122 98 28 129 120 112 119 95 32 125 119 108 115 91

262


Bed length (mm) Average drop diameter (p.m)

Average drop density (no./cm3)

Coalescence efficiency

200 11.3 134.5 0.100 400 9.75 127.5 0.0536 600 9.35 118.5 0.0389 800 9.65 123.5 0.0279

1000 9.35 103.0 0.0270

Table A.30: Dron diameters In for 24 mL/min Time (h) 200 mm 400 mrn 600 nun 800 mm 1000 mm

2 13.9 13.4 12.8 13.1 12.3 4 13.2 12.9 12.4 12.7 12.1 6 10.4 12.6 12.1 12.4 11.8 8 11.3 10.7 10.1 10.4 9.7

10 10.9 10 9.4 9.7 9 12 14.4 13.5 12.9 13.2 12.6 16 11.2 10.9 10.3 10.7 9.8 20 12.1 11.2 10.5 10.9 10.2 24 10.9 9.8 9.2 9.5 8.9 28 11.3 10.8 10.1 10.5 9.6

Table A.31 Drop density (no / cm3) for 24 mL/min Time (h) 200 mm 400 mm 600 mm 800 mm 1000 mm

2 165 135 128 120 111 4 161 133 125 118 110 6 155 145 139 115 108 8 138 126 119 110 98 10 132 123 115 106 92 12 166 146 129 119 112 16 137 125 118 107 96 20 142 129 121 112 100 24 135 124 116 104 93 28 131 121 113 101 90

263


Bed length (mm) Average drop diameter (|im)



200 11.3 134.5 0.100

400 9.75 127.5 0.0536

600 9.35 118.5 0.0389

800 9.65 123.5 0.0279

1000 9.35 103.0 0.0270



2 13.9 13.4 12.8 13.1 12.3

4 13.2 12.9 12.4 12.7 12.1

6 10.4 12.6 12.1 12.4 11.8

8 11.3 10.7 10.1 10.4 9.7

10 10.9 10 9.4 9.7 9

12 14.4 13.5 12.9 13.2 12.6

16 11.2 10.9 10.3 10.7 9.8

20 12.1 11.2 10.5 10.9 10.2

24 10.9 9.8 9.2 9.5 8.9

28 11.3 10.8 10.1 10.5 9.6



2 165 135 128 120 111

4 161 133 125 118 110 6 155 145 139 115 108 8 138 126 119 110 98 10 132 123 115 106 92 12 166 146 129 119 112 16 137 125 118 107 96 20 142 129 121 112 100 24 135 124 116 104 93 28 131 121 113 101 90

263


Bed length (mm) Average drop diameter (gm)



200 12.6 148 0.081 400 12.1 128 0.053 600 11.45 120.5 0.039 800 11.8 110.5 0.032

1000 10.95 100.5 0.025

Table A.33: Drop diameters(um) for 28 mL/min Time (h) 200 mm 400 mm 600 mm 800 mm 1000 mm

2 14.7 13.4 12.7 13.1 12.4 4 13.6 12.2 11.5 11.8 11.2 6 12.4 11.7 11.1 11.4 10.7 8 11.1 10.8 10.2 10.5 9.8

10 14.3 13.1 12.5 12.9 12.1 12 10.8 10.1 9.5 9.7 9.1 16 11 10.5 9.9 10.2 9.3 20 10.3 9.8 9 9.3 8.5

Table A.34: Drop density (no.! cm3) for 28 rnL/min Time (h) 200 mm 400 mm 600 mm 800 mm 1000 mm

2 149 137 131 121 109 4 142 132 126 116 103 6 139 128 121 112 101 8 136 125 118 103 99 10 140 130 123 113 102 12 129 121 112 99 94 16 131 122 115 100 97 20 126 119 111 98 91


Bed length (mm) Average drop diameter (Am)



200 12.5 137.5 0.080 400 11.6 128 0.047 600 10.85 121 0.035 800 11.2 109.5 0.029

1000 10.45 100 0.023

264


Bed length (mm) Average drop diameter (um)



200 12.6 148 0.081

400 12.1 128 0.053

600 11.45 120.5 0.039

800 11.8 110.5 0.032

1000 10.95 100.5 0.025

Table A.33: Drop diameters (nm) for 28 mL/min


2 14.7 13.4 12.7 13.1 12.4

4 13.6 12.2 11.5 11.8 11.2

6 12.4 11.7 11.1 11.4 10.7

8 11.1 10.8 10.2 10.5 9.8

10 14.3 13.1 12.5 12.9 12.1

12 10.8 10.1 9.5 9.7 9.1

16 11 10.5 9.9 10.2 9.3

20 10.3 9.8 9 9.3 8.5



2 149 137 131 121 109

4 142 132 126 116 103

6 139 128 121 112 101 8 136 125 118 103 99

10 140 130 123 113 102 12 129 121 112 99 94 16 131 122 115 100 97

20 126 119 111 98 91


Bed length (mm) Average drop diameter (um)



200 12.5 137.5 0.080 400 11.6 128 0.047 600 10.85 121 0.035

800 11.2 109.5 0.029

1000 10.45 100 0.023

264

Table A.36: Dron diameters In for 32 mL/min Time (h) 200 mm 400 mm 600 mm 800 mm 1000 mm 2 14.1 13.3 12.5 12.8 12.1 4 13.7 12.6 11.8 12.3 11.6 6 11.7 11.2 10.7 10.9 10.3 8 12.6 12.1 11.5 11.9 11.2 10 11.9 11 10.2 10.5 10 12 10.5 9.8 9.1 9.5 8.7 16 11.1 10.5 9.6 10.1 9.2

Table A.37: Drop density no./ cm3) for 32 mL/min Time (h) 200 mm 400 mm 600 mm 800 mm 1000 mm

2 136 124 119 115 110 4 133 120 116 112 107 6 129 117 115 109 103 8 135 122 117 113 109 10 131 119 112 109 104 12 121 115 107 100 92 16 125 117 109 103 96


Bed length (mm) Average drop diameter (gm)



200 12.60 130.5 0.081 400 11.90 120.5 0.047 600 11.05 114.0 0.033 800 11.45 109.0 0.026

1000 10.65 103.0 0.021

Table A.39: Data used to fit Crickmore model

Flow rate, rnL/min

Initial oil concentration (Co), mg/L

Effluent oil concentration (C), mg/L Time, h t', h/cm2 In (C/Co)

12 50 44.2 50 88.407 -0.123 16 50 45.4 38 69.845 -0.096 20 50 46.1 32 61.113 -0.0812 24 50 46.2 28 54.272 -0.0790 28 50 47.1 20 39.549 -0.0596 32 50 46.8 16 32.068 -0.066

265

Table A.36: Drop diameters (urn) for 32 mL/min Time (h) 200 mm 400 mm 600 mm 800 mm 1000 mm 2 14.1 13.3 12.5 12.8 12.1 4 13.7 12.6 11.8 12.3 11.6 6 11.7 11.2 10.7 10.9 10.3 8 12.6 12.1 11.5 11.9 11.2 10 11.9 11 10.2 10.5 10 12 10.5 9.8 9.1 9.5 8.7 16 11.1 10.5 9.6 10.1 9.2

Table A.37: Drop density (no./ cm3) for 32 mL/min Time (h) 200 mm 400 mm 600 mm 800 mm 1000 mm

2 136 124 119 115 110 4 133 120 116 112 107 6 129 117 115 109 103 8 135 122 117 113 109 10 131 119 112 109 104 12 121 115 107 100 92 16 125 117 109 103 96


Bed length (mm) Average drop diameter (|im)



200 12.60 130.5 0.081 400 11.90 120.5 0.047 600 11.05 114.0 0.033 800 11.45 109.0 0.026

1000 10.65 103.0 0.021

Table A.39: Data used to fit Cric cmore model Initial oil Effluent oil

Flow rate, concentration concentration mL/min (Co), mg/L (C), mg/L Time, h t', h/cm2 In (C/Co)

12 50 44.2 50 88.407 -0.123 16 50 45.4 38 69.845 -0.096 20 50 46.1 32 61.113 -0.0812 24 50 46.2 28 54.272 -0.0790 28 50 47.1 20 39.549 -0.0596 32 50 46.8 16 32.068 -0.066

265

1 380

5, 70

60

8 so ..52 40

oo 30

d••t

20

10

0

0 Adsorption capacity observed

Adsorption capacity predicted -Lagergren Adsorption capacity predicted - Ho

3 4 Time, h

Figure A.1: Rate of SMO biosorption by Lagergren and Ho kinetic models at 5°C

100

90

ho 80

5 70

c.) 60 cs, 5 so

40

30 -66

20

10

0 0

"0" -0- - - - 4-


Adsorption capacity predicted -Lagergren

Adsorption capacity predicted - Ho

3 Time, h 4


266

100 - - - - o

.2 40 -o Adsorption capacity observed

o 30 -

3 < 20 -Adsorption capacity predicted -Lagergren Adsorption capacity predicted - Ho

Time, h

Figure A. 1: Rate of SMO biosorption by Lagergren and Ho kinetic models at 5°C

100 i

/q'Q

9-

O Adsorption capacity observed



0 1 2 3 Time, h 4 5 6 7


266

Ads

orpt

ion

capa

city

, mg/

g

100

80

60

40

20

0

o Adsorption capacity observed

Adsorption capacity predicted - Lagergren


0 3 Time, h


98.4

98.2 -

to 98.0 - --a) E 97.8 - -6 .5 97.6vs cd • 97.4

. 0 97.2

O 97.0 - •t1 < 96.8

96.6

Q -




3 4 Time, h

5

Figure A.4: Rate of CO biosorption by Lagergren and Ho kinetic models at 5°C

267

100

80

Ph 60

40




20

0 3 Time, h 4 0 2 1 5 6 7


I5

& 8* o a o ••p & o •8 <

98.4

98.2

98.0

97.8

97.6

97.4

97.2

97.0

96.8

96.6

o 9 Q.o.ft-Q-fr-9'Q"" 0 o ® ? 0




3 Time, h 4


267

98.5

ao 98.0

5 X97.5

"c"3 co

6' 97.0

0

46 , 96.5 I

96.0

95.5 0




3 Time, h 4


98.5

98.0

11137.5

,s5 97.0

896.5 • 1.1

95.5




3 Time, h 7


268

98.5

bo 98.0

I £97.5 G

S 97.0 a o

1*96.5 o •8 ^ 96.0

95.5

9 qq g 'o""Q




3 Time, h 4


98.5

98.0

f»7.5

897.0 & o §96.5

*•£3

&

J96.0 <

95.5

0 o 0 o..0-o-q-Q"Q""o""6""o" 0 ?

p o Adsorption capacity observed

• Adsorption capacity predicted -Lagergren


3 Time, h 4


268

98.5

bo 98.0

5 .6 97.5 Z)

COQ

97.0 0

4-6 o 96.5

96.0

95.5




0 1 2 3 Ti

4 Time, h

Figure A.7: Rate of Bright-Edge 80 biosorption by Lagergren and Ho models at 5°C

98.5

98.0

ao -a) 97.5 5

.6 97.0

ct ▪ 96.5 U

.9. • 96.0 - 46' - • 95.5

95.0

94.5

L

-- -------------------------




3 Time, h 4 5


269

98.5

to 98.0

I

£ 97-5

8 & u 97.0 C o v» fr o 96.5 •8 <

96.0

95.5

q Q g q"—2 9 Q. SL


• Adsorption capacity predicted - Lagergren


Time, h


98.5

98.0

bo ~Sb 97.5 £

& 97.0

U 03-96.5 Q

•2 96.0

&

J 95.5 <

95.0

94.5

o q oq.-q-o-q~o~q~ 0 q q_ 9




3 Time, h 4


269

Ads

orpt

ion

capa

city

, mg/

g

98.5 -

98.0 -

97.5

97.0

96.5 -

96.0

95.5

95.0 -

94.5

94.0

93.5

----------------------------------




3 Time, h 4 7


Ads

orbt

ion

capa

city

, Q m

g/g

105

100

95

90

85

80

75

0 0

0 0

o Adsorption capacity -Observed

—Adsorption capacity - Predicted

o

05 0.7 0.9 1.3 1.5 1.7

Figure A.10: Rate of SMO biosorption by intra-particle diffusion model at 5°C

270

1.9

1 c3 & o d o

• m +->

& o •3 <

98.5

98.0

97.5

97.0

96.5

96.0

95.5

95.0

94.5

94.0

93.5

0 0 gL.Q 0-Q"P"o""0 * Q 0 2 0-


• Adsorption capacity predicted -Lagergren


3 Time, h 4


105 -i


Adsorption capacity - Predicted

Vt, h0-5

Figure A. 10: Rate of SMO biosorption by intra-particle diffusion model at 5°C

270

105

100

11)E 95 0'

.. ,..' 90 O RS cl. 8 85 = 0 2 80 0 ...

-6' ..:t 75

05 0.7 0.9


— Adsorption capacity - Predicted

1.1 'Nit, 110.5

1.3 1.5


97

to - 92 E 0' p: 87 0 is to. RS O • 82 2

:r:i 4.4

., • 77 d

72

O

0 O

0 0



1.7

0 5 0.7 0.9 1.1

lit, 11°.5

1.3 1.5 1.7


271

105

100 oo

S 95

8«

° Adsorption capacity -Observed

Adsorption capacity - Predicted j3 80 o M TJ < 75

0.5 0.7 0.9 1.1 1.5 1.7 1.3

Vt, h0-5

Figure A. 11: Rate of SMO biosorption by intra-particle diffiision model at 15°C

97

©«

82

-£ o -3 77



72 0.5 0.7 1.5 1.3 1.7

Vt, h0-5


271

103

98

0 E 93

CY

88 art

0 83

0 -1:1 78 44t 0.5 0.6 0.7 0.8 0.9 1 1.1 1.2 1.3

h"

0 • 0

0 Adsorption capacity -Observed



1 )1)E

.6

0

A

99

98.5

0

98

97.5 •

0 Adsorption capacity -Observed

97 —Adsorption capacity - Predicted

96.5 0 5 0.6 0.7 0.8 0.9 1 1.1 1.2 1.3

h"

Figure A.14: Rate of CO biosorption by intra-particle diffusion model at 5°C

272

103

6 93

u 88

O 83 o Adsorption capacity -Observed


1.2 0.6 0.7 0.8


"5b 98.5


Adsorption capacity - Predicted •3 97

Figure A. 14: Rate of CO biosorption by intra-particle diffusion model at 5°C

272

99

98.5

04 98

E 0 '97.5

97

uo 96.5 0

-° 96

< 95.5 05

0



0

0.6 0.7 0.8 0.9 Alt, 110.5

1 1.1 1.2


99

98

to 97

E 96

F. 95

94

e) 93 0

,14 92

.g 91 90

05 0.7 0.9 1.1

0



h"

1.3 1.5


273

1.3

1.7

98.5

° Adsorption capacity -Observed 96.5


<95.5

1.2 1.3 0.5 0.6 0.7 0.8 0.9 1.1 1 Vt, h0 5


° Adsorption capacity -Observed


1.1

Vt, h0-5


273

Ads

orbt

ion

capa

city

, Q m

g/g

99

98.5

98 0 0

97.5

97

96.5

96

95.5

95

94.5 0 5 0.55 0.6 0.65 0.7 0.75 0.8 0.85 0.9

119.5




98.5

98

143E 97.5

0 1

-41.; 97

c) 96.5 0

6- 96

d 95.5

05

O

0



0.6 0.7 0.8

'It, 11.0.5

0.9 1.1 1.2


274

99

98.5

97.5

97

g, 96.5

96


—Adsorption capacity - Predicted 95

94.5

0.8 0.85 0.9 0.55 0.75 0.5 0.65


98.5

6 97.5

& ° 96.5



95.5

0.5 0.7 0.6 0.8 0.9

Vt, h0-5


274

99

98.5

0 98 a) E 97.5 a i, 97

a 96.5cs cc.)

96 0

€ :a 95.5 0 434 95 <

94.5 05

0

0



0.6 0.7 0.8 0.9

'It, V.51 1.1 1.2


Ads

orbt

ion

capa

city

, Q m

g/g

98.5

98

97.5

97

96.5

96

95.5

95

94.5

94

93.5 05 0.6 0.7

0 0

o Adsorption capacity -Observed — Adsorption capacity - Predicted

0.8 0.9

Alt, ha5

1 1.1 1.2 1.3


275

98.5

E 97.5

8 96.5

95.5 o Adsorption capacity -Observed


94.5

1.2 1.1 0.7 0.8 0.9 1 0.5 0.6

Vt, h°-5


98.5 i



93.5 ~i > 1 1 1 1 1 1 1

0.5 0.6 0.7 0.8 0.9 1 1.1 1.2 1.3

Vt, h0-5


275

Ads

orbt

ion

capa

city

, Q m

g/g

98.4

98.2

98

97.8

97.6

97.4

97.2

97

96.8

96.6

96.4 05

0

0

o Adsorption capacity -Observed —Adsorption capacity - Predicted

0.6 0.7 0.8

h"

0.9 1 1.1


L

700

600 ao

E 500 e,os 400

g 300

en9, 200

•cs

100

0

0 Experimental

Langmuir predicted

Freundlich predicted

10 15 Ce, mg/L

20

Figure A.22: The Langmuir and Freundlich isotherm for biosorption of SMO at 5°C

276

25

98.4 i



< )

96.4 4 . . 1 . . .

0.5 0.6 0.7 0.8 0.9 1 1.1

Vt, h05


700

600

500

« 400

300

O Experimental

Langmuir predicted


g 200

100

0 0 5 10 15 20 25

Ce, mg/L


276

700

600 to be g 500

co 400

gi 300

200

100

0 0

o Experimental

Langmuir predicted


10 15 Ce, mg/L

20 25


700

600 eo too

500

eo 400

300

200

100

0 10 15

Ce, mg/L

o Experimental

Langmuir predicted


20 25 30


277

700

BO S 500

g 200

O Experimental

Langmuir predicted


10 ^ /. 15 Ce, mg/L

20 25


700

600

M g 500

S 400

O Experimental

Langmuir predicted


15 Ce, mg/L

20 25 30


277

700

ba 600

soo a 400 a g 300

I :0 P.

200

4< 100

0 0

o Experimental

Langmuir predicted


10 15 20 25 Ce, mg/L

Figure A.25: The Langmuir and Freundlich isotherm for biosorption of CO at 5°C

600 OD

itik 500

.0 es 400

300

100

10 15 Ce, mg/L

Experimental

Langmuir predicted


20

Figure A.26: The Langmuir and Freundlich isotherm for biosorption of CO at 22 °C

278

700

600

500

400

O Experimental

Langmuir predicted


300

200

100

0 0 5 10 15 20 25

Ce, mg/L


700

600

500

400

300 O Experimental

Langmuir predicted


200

100

0 0 5 10 15 20 25 Ce, mg/L

Figure A.26: The Langmuir and Freundlich isotherm for biosorption of CO at 22 °C

278

700

600

iha 500

x. 400

I 8 g 300 0BA.

0 200

100 Q

0 5

--r

10

o Experimental

Langmuir predicted


1

15 20 25 Ce, mg/L


700

600 ba

r 500 jr ,400 a g 300

200

100

0

o Experimental

Langmuir predicted


3

10 15 Ce, mg/L

20 25

Figure A.28: The Langmuir and Freundlich isotherm model plots for biosorption of Bright-Edge 80 at 5°C

279

700

„600 M S 500

S 400

£ 300 O Experimental

Langmuir predicted


o 200

100

0 0 5 10 15 ,

Ce, mg/L 20 25 30


700 1

600 as

if 500

300 -o Experimental

Langmuir predicted


8 200 -

100

0 5 10 15 20 25 Ce, mg/L


279

700

ea 600 -1

g 500

1 . 400 -I 8 .1 300

• 200

[ ▪ 100 .1

0

o Experimental

Langmuir predicted


10 15 20 25 30 35 40 1 Ce, mg/L


700

600 J [ IDA

soo

eta 400

O • 300

51 200

100

o Experimental

Langmuir predicted


8 10 12 14 16 18 Ce, mg/L


280

700

600 H

S 500

2 400 -

O Experimental

Langmuir predicted


300 -<

o 200

100

0 10 15 5 20 25 30 35 40 Ce, mg/L


700

600

« 400

o 300

Experimental

Langmuir predicted


3 200

100

0 0 2 6 4 8 10 12 14 16 18

Ce, mg/L


280

1.0

0.9

0.8

0.7

0.6

$40.5

0.4

0.3

0.2

0.1

0.0

—Outman Predicted

• Experimental

0 500 1000 2000 2500 3000 Time

Figure A.31: Breakthrough curve predicted by Oulman model for SMO

1.0 1

0.9 -

0.8

0.7

0.6 -

0.5u

0.4

0.3

0.2 •

0.1 •

0.0

•

•

—Oulman Predicted

• Experimental

0 200 400 600 800 1000 1200 1400 1600 Time (min)

Figure A.32: Breakthrough curve predicted by Oulman model for CO

281

1.0

0.9

0.8

0.7

0.6 o ii.0.5 o

0.4

Oulman Predicted

• Experimental

0.3

0.2

0.1

2000 2500 0 500 1000 Timeim

3000 i

Figure A.31: Breakthrough curve predicted by Oulman model for SMO

l.o

0.9

0.8

0.7

0.6

Oulman Predicted

• Experimental 0.4

0.3

0.2

0.1

0.0 200 0 600 400 800 1000 1200 1400 1600

Time (min)

Figure A.32: Breakthrough curve predicted by Oulman model for CO

281

1.0

0.9

0.8

0.7

0.6 0Q0.5

0.4

0.3 J

0.2

0.1

0.0 0 200 400 600 800 1000 1200 1400 1600 1800 2000

Time (min)

— Oulman Predicted

• Experimental

Figure A.33: Breakthrough curve predicted by Oulman model for Bright-Edge 80

1.4

1.2

1 0.8

v

0.6

0.4

0.2

0

• Experimental

— Wolbroska Predicted

500 t(min)

Figure A.34: Breakthrough curve predicted by Wolbroska model for SMO

282

1000 I

1.0

0.9

0.8

0.7

0.6

Oulman Predicted 0.5

0.4 • Experimental

0.3

0.2

0.1

0.0 200 400 600 800 1000 1200 1400 1600 1800 2000 0

Time (mill)

Figure A.33: Breakthrough curve predicted by Oulman model for Bright-Edge 80

1.6

• Experimental

Wolbroska Predicted 1.4

1.2

ii 0.8

0.6

0.4

0.2

0 500 t(min)

1000

Figure A.34: Breakthrough curve predicted by Wolbroska model for SMO

282

• Experimental — Wolbroska Predicted

0 t.) (4 0.6

0 500 Time (min)

Figure A.35: Breakthrough curve predicted by Wolbroska model for CO

1.2

1

0.8

0.4

0.2

0 0

4

• Experimental — Wolbroska Predicted

4 4

TimiCOriin)

4

1000

Figure A.36: Breakthrough curve predicted by Wolbroska model for Bright-Edge 80

283

1.4

• Experimental

Wolbroska Predicted 1.2

1

0.8

0.6

0.4

0.2

0 0 500

Time (min) 1000

Figure A.35: Breakthrough curve predicted by Wolbroska model for CO

1.2

A Experimental

—Wolbroska Predicted 1

0.8

< 0.6

0.4

0.2

0 0 Timl^in) 1000

Figure A.36: Breakthrough curve predicted by Wolbroska model for Bright-Edge 80

283

Appendix B

Design of Batch and Column Adsorber System

Table B 1: Design of batch adsorber system for M rouxii biomass and SMO with a flow rate of 100 m3/d Descriptions Calculations Isotherm model used Freundlich Isotherm equation x/m = kCA(1/n) Isotherm constants

K 40.16 1/n (n = 1.12) 0.893

Initial oil concentration, mg/L 200 Final oil concentration, mg/L 5 Percentage of oil removal 97.5

=(40.16)(5)0.893)x/rn value mg/g

=168.99 =(100)(200-5)/1000

Mass of oil removed kg/d =19.5

Mass of biomass required using x/m value (kg/d) =(19.5)(1000)/(168.99) =115.39

Table B2: Design of batch adsorber system for M rouxii biomass and CO with a flow rate of 100 m3/d Descriptions Calculations Isotherm model used Freundlich Isotherm equation x/m = kCA(1/n) Isotherm constants

K 30.6 l/n (n = 1.01) 0.99


x/m value mg/g =(30.6)(5)(0.99)=150.58

Mass of oil removed kg/d =(100)(200-5)/1000 =19.5


284

Appendix B

Design of Batch and Column Adsorber System

Table Bl: Design of batch adsorber system for M. rouxii biomass and SMO with a flow rate of 100 m3/d Descriptions Calculations Isotherm model used Isotherm equation Isotherm constants

K l/n(n = 1.12)

Initial oil concentration, mg/L Final oil concentration, mg/L Percentage of oil removal

x/m value mg/g

Mass of oil removed kg/d

Mass of biomass required using x/m value (kg/d)

Freundlich x/m = kCA(l/n)

40.16 0.893 200 5 97.5 =(40.16)(5)(0893)

=168.99 =(100)(200-5)/1000 =19.5 =(19.5)(1000)/(168.99) =115.39

Table B2: Design of batch adsorber system for M. rouxii biomass and CO with a flow rate of 100 m3/d Descriptions Calculations Isotherm model used Isotherm equation Isotherm constants

K l/n(n= 1.01)

Initial oil concentration, mg/L Final oil concentration, mg/L Percentage of oil removal

x/m value mg/g

Mass of oil removed kg/d

Mass of biomass required using x/m value (kg/d)

Freundlich x/m = kCA(l/n)

30.6 0.99 200 5 97.5 =(3O.6)(5)(0 99)

=150.58 =(100)(200-5)/1000 =19.5 =(19.5)(1000)/( 150.58) =129.49

284

Table B3: Design of batch adsorber system for M rouxii biomass and Bright-Edge 80 with a flow rate of 100 m3/d Descriptions Calculations Isotherm model used Freundlich Isotherm equation x/m = kCA(1/n) Isotherm constants

K 26.56 1/n (n = 1.16) 0.862


x/m value mg/g =(26.56)(5)(0.862)

=106.36

Mass of oil removed kg/d =(100)(200-5)/1000 =19.5


285

Table B3: Design of batch adsorber system for M. rouxii biomass and Bright-Edge 80 with a flow rate of 100 m3/d Descriptions Calculations Isotherm model used Freundlich Isotherm equation x/m = kCA(l/n) Isotherm constants

K 26.56 l/n(n= 1.16) 0.862

Initial oil concentration, mg/L 200 Final oil concentration, mg/L 5 Percentage of oil removal

x/m value mg/g

97.5 =(26.56)(5)(0'862)

=106.36

Mass of oil removed kg/d =( 100)(200-5)/1000 =19.5


285

Table B4: Design of immobilized M rouxii biomass column filter for SMO with a flow rate of 100 m3/d By scale-up approach Flow rate, mL/min Unit liquid flow rate, L/h

Bed volume, cm3

Bed volume, L

Bed volume per unit time, BV/h

Design bed volume, m3

Inside diameter of column, mm Length of bed, mm Mass of biomass in the column, kg Packed density of biomass in the column, kg/m3Mass of immobilized biomass required, kg From breakthrough curve for allowable 10 mg/L oil concentration, the corresponding volume, L

Emulsion treated per kg of biomass, L/kg

Biomass exhausted per hour, kg/h

Breakthrough time, h

Breakthrough time, d

Breakthrough volume, m3

By kinetic approach

=2.6 =(2.6)(60)/1000 =0.156 =(3.14)(1.27)2(100)/4 =126.61 =126.61/1000 =0.127 =0.156/0.127 =1.23 =100/(1.23)(24) =3.38 =12.70 =300.00 =0.004500 =118.50 =400.74

=0.45

=0.45/0.0045 =100 =(100)(1000)/(100)(24) =41.67 =(400.74)/(41.67) =9.62 =9.62/24 =0.40074 =(0.40074)(100) =40.074

KT, L/h-kg (From Table 4.21) =4200 go, kg/kg (From Table 4.21) =0.0098 Co, mg/L =50 C, mg/L =5 Flow, m3/d =100 Flow, Q, L3/h =(100)(1000)/24 =4166.67

Volume, V, L3 =(40.074)(1000) =40074

Applying Thomas model solving for M, kg/h

(Equation 2.12) and 422.37

286

Table B4: Design of immobilized M. rouxii biomass column filter for SMO with a flow rate of 100 m3/d By scale-up approach Flow rate, mL/min =2.6

Unit liquid flow rate, L/h =(2.6)(60)/1000 =0.156

Bed volume, cm3 =(3.14)( 1.27)2( 100)/4 Bed volume, cm3

=126.61

Bed volume, L =126.61/1000 =0.127

Bed volume per unit time, BV/h =0.156/0.127 =1.23

Design bed volume, m3 =100/(1.23)(24) =3.38

Inside diameter of column, mm =12.70 Length of bed, mm =300.00 Mass of biomass in the column, kg =0.004500 Packed density of biomass in the column, kg/m3 =118.50 Mass of immobilized biomass required, kg =400.74 From breakthrough curve for allowable 10 mg/L oil =0.45 concentration, the corresponding volume, L

=0.45

Emulsion treated per kg of biomass, L/kg =0.45/0.0045 =100

Biomass exhausted per hour, kg/h =( 100)( 1000)/(100)(24) =41.67

Breakthrough time, h =(400.74)/(41.67) =9.62

Breakthrough time, d =9.62/24 =0.40074

Breakthrough volume, m3 =(0.40074)(100) =40.074

By kinetic approach Kt, L/h-kg (From Table 4.21) =4200 q0, kg/kg (From Table 4.21) =0.0098 Co, mg/L =50 C, mg/L =5

Flow, m3/d =100 Flow, Q, L3/h =(100)(1000)/24 =4166.67

Volume, V, L3 =(40.074)(1000) Volume, V, L3

-40074 Applying Thomas model (Equation 2.12) and 422.37 solving for M, kg/h

286

Table B5: Design of immobilized M rouxii biomass column filter for CO with a flow rate of 100 m3/d By scale-up approach Flow rate, mL/min Unit liquid flow rate, L/h

Bed volume, cm3

Bed volume, L









By kinetic approach

=2.6 =(2.6)(60)/1000 =0.156 =(3.14)(1.27)2(100)/4 =126.61 =126.61/1000 =0.127 =0.156/0.127 =1.23 =100/(1.23)(24) =3.38 =12.70 =300.00 =0.0045 =118.50 =400.74

=0.08

=0.08/0.0045 =17.33 =(100)(1000)/(17.33)(24) =240.38 =(400.74)/(240.38) =1.67 =1.67/24 =0.069 =(0.069)(100) =6.95

KT, L/h-kg (From Table 4.21) =4980 qo, kg/kg (From Table 4.21) =0.0146 Co, mg/L =50 C, mg/L =5 Flow, m3/d =100 Flow, Q, L3/h =(100)(1000)/24 =4166.67

Volume, V, L3 =(6.95)(1000) =6950

Applying Thomas model (Equation solving for M, kg/h

2.12) and 149.72

287

Table B5: Design of immobilized M. rouxii biomass column filter for CO with a flow rate of 100 m3/d By scale-up approach Flow rate, mL/min =2.6


Bed volume, cm3 =(3.14)(1.27)2(100)/4 Bed volume, cm3

=126.61

Bed volume, L =126.61/1000 =0.127



Inside diameter of column, mm =12.70 Length of bed, mm =300.00 Mass of biomass in the column, kg =0.0045 Packed density of biomass in the column, kg/m3 =118.50 Mass of immobilized biomass required, kg =400.74 From breakthrough curve for allowable 10 mg/L oil =0.08

=0 08/0 0045 concentration, the corresponding volume, L

=0.08

=0 08/0 0045 Emulsion treated per kg of biomass, L/kg V*V W' V/• V/\y TmJ

=17.33

Biomass exhausted per hour, kg/h =(100)(1000)/(17.33)(24) =240.38




By kinetic approach KT, L/h-kg (From Table 4.21) =4980 qo, kg/kg (From Table 4.21) =0.0146 Co, mg/L =50 C, mg/L =5

Flow, m3/d =100 Flow, Q, L3/h =(100)(1000)/24 =4166.67

Volume, V, L3 =(6.95)(1000) Volume, V, L3 =6950

Applying Thomas model (Equation 2.12) and 149.72 solving for M, kg/h

287

Table B6: Design of immobilized M rouxii biomass column filter for Bright-Edge 80 with a flow rate of 100 m3/d By scale-up approach Flow rate, mL/min Unit liquid flow rate, L/h

Bed volume, cm3

Bed volume, L









By kinetic approach

=2.6 =(2.6)(60)/1000 =0.156 =(3.14)(1.27)2(100)14 =126.61 =126.61/1000 =0.127 =0.156/0.127 =1.23 =100/(1.23)(24) =3.38 =12.70 =300.00 =0.0045 =118.50 =400.74

=0.05

=0.05/0.0045 =11.11 =(100)(1000)/(11.11)(24) =375.0 =(400.74)/(375.0) =1.01 =1.01/24 =0.045 =(0.045)(100) =4.45

KT, L/h-kg (From Table 4.21) =3600 go, kg/kg (From Table 4.21) =0.0106 Co, mg/L =50 C, mg/L =5 Flow, m3/d =100 Flow, Q, L3/h =(100)(1000)/24 =4166.67

Volume, V, L3 =(4.45)(1000) =4450

Applying Thomas model (Equation solving for M, kg/h

2.12) and 260.904

288

Table B6: Design of immobilized M. rouxii biomass column filter for Bright-Edge 80 with a flow rate of 100 m3/d By scale-up approach Flow rate, mL/min =2.6


Bed volume, cm3 =(3.14)(1.27)2(100)/4 Bed volume, cm3

=126.61

Bed volume, L =126.61/1000

Bed volume, L =0.127



Inside diameter of column, mm =12.70 Length of bed, mm =300.00 Mass of biomass in the column, kg =0.0045 Packed density of biomass in the column, kg/m3 =118.50 Mass of immobilized biomass required, kg =400.74 From breakthrough curve for allowable 10 mg/L oil =0.05

=0 05/0 0045 concentration, the corresponding volume, L

=0.05

=0 05/0 0045 Emulsion treated per kg of biomass, L/kg ViVk/r \/< W i

= 11.11

Biomass exhausted per hour, kg/h =(100)(1000)/(11.11 )(24) =375.0




By kinetic approach KT, L/h-kg (From Table 4.21) =3600 qo, kg/kg (From Table 4.21) =0.0106 Co, mg/L =50 C, mg/L =5

Flow, m3/d =100 Flow, Q, L3/h =(100)(1000)/24 =4166.67

Volume, V, L3 =(4.45)(1000) Volume, V, L3 =4450

Applying Thomas model (Equation 2.12) and 260.904 solving for M, kg/h

288

Documents

REMOVAL OF OIL FROM WATER USING FUNGAL BIOMASS A …