Photocatalytic Reduction of Selenate and Selenite: Water/Wastewater Treatment and The Formation of Nano-Selenium Compounds
by Thatt Yang Timothy TAN, B.E. (Hons Class 1)
A thesis submitted in fulfilment of the requirements for the degree of Doctor of Philosophy
The School of Chemical Engineering and Industrial Chemistry The University of New South Wales Sydney Australia
July 2003
A CERTIFICATE OF ORIGINALITY
I hereby declare that this submission is my own work and to the best of my knowledge it contains no materials previously published or written by another person, nor material which to a substantial extent has been accepted for the award of any other degree or diploma at UNSW or any other educational institution, except where due acknowledgment is made in the thesis. Any contribution made to the research by others, with whom I have worked at UNSW or elsewhere, is explicitly acknowledged in the thesis.
I also declare that the intellectual content of this thesis is the product of my own work, except to the extent that assistance from others in the project’s design and conception or in style, presentation and linguistic expression is acknowledged.
………………………………………………………..
B
This thesis is especially dedicated to my parents, Boon Chew Tan & Ah Sioh Lee, And my sister, Foong Yee Tan.
C Abstract
The current work investigates the photocatalytic reduction of selenium (Se) ions, selenate Se(VI) and selenite Se(IV), from two perspectives: Se ion removal from water and wastewater and the formation of nano-Se compounds. Se ion pollution has become an environmental issue in recent years, and hence there is an urgent need for an efficient removal technique. In addition, there is increasing interest in the formation of nano-size semiconductors for niche applications. Since Se is a semiconductor, its formation onto the semiconductor TiO2 could lead to the discovery of new composite materials.
The current study has successfully elucidated the mechanism of Se ions reduction by photocatalysis. Factors such as the simultaneous adsorption of the Se ions (the electron scavenger in this case) and a suitable organic compound (the hole scavenger), and the chemical properties of the hole scavenger were crucial for effective and efficient Se ions photoreduction. Optimum conditions in relation to pH, concentrations and types of hole scavenger were reported and discussed. It was also found that stoichiometric adsorption ratio of formate and selenate resulted to optimum photoreduction rate. A modified Langmuir-Hinshelwood kinetic model that considered the simultaneous adsorption of both solutes was derived.
The current investigation has also seen the successful formation Se deposits of different morphologies onto the TiO2 particles. Discrete Se particles of various sizes in the nano-size range as well as a Se film were deposited onto the TiO2 particles under different initial experimental conditions. The Se-TiO2 composite semiconductor was explored for the 2+ removal of cadmium Cd ions, which resulted in the formation of CdSe-TiO2 systems. The photoreduction of Se ions using silver-modified TiO2 showed the enhanced reduction of Se 2- ions to Se in the form of H2Se gas. It is suggested that the H2Se gas generated from the current photoreduction process could be used as a safer and cheaper technique in the formation of Se-compounds such copper selenide, cadmium selenide and zinc selenide. All these compounds were widely used in optical and semiconducting devices.
D Acknowledgements
I would like to extend my most sincere acknowledgement and gratitude to a number of people and organisations. Without them, the quest towards my Ph.D. degree would be less rewarding, interesting and challenging.
My greatest appreciation goes to my university supervisors, A/Prof Rose Amal and Dr Donia Beydoun, for their omnipresence in providing help, support, encouragement, guidance and never being tired of reading my works. I am particularly grateful to them for imparting their valuable knowledge and time. I would also like to thank Dr Myint Zaw for his help and guidance when I was conducting experiments in the Australian Nuclear Science and Technology Organisations (ANSTO).
I would also like to thank Prof Paul Munroe and Miss Katie Levick from the Electron Microscopy unit in the University of New South Wales (UNSW) for their interpretation and assistance with electron microscopy analysis. Special thanks to Dr Kath Smith and Dr David Mitchell for their assistance in electron microscopy analysis in ANSTO. I also thank Dr Grainne Moran for her assistance in Uv-Vis spectroscopy analysis and Inna Bolkovsky for her help in XRD analysis, and of course not forgetting Mr John Starling for his ever- timely ordering of materials.
The awarding of the Australian Institute of Nuclear Science and Engineering (AINSE) Postgraduate Scholarship, the International Postgraduate Research Scholarship from The University of New South Wales (UNSW) and the Department of Education, Training and Youth Affairs (DETYA), the Supplementary Engineering Award from the Faculty of Engineering, and the scholarship from the Centre of Particle and Catalysts Technologies are all gratefully acknowledged.
My sincere thanks also extend to everyone in the Centre of Particle and Catalyst Technologies; for their continual support, friendship and ceaseless humour in making a drilling day more tolerable, and especially for their generosity in lending laboratory apparatus during crucial moments.
Most importantly, I would like to thank my parents, sister and all my family and friends in Sydney for their love and support. Last but not least, the continued support, love and care from Chun Poe Yuen and Pui Yik Chan are greatly appreciated.
E Lists of Publications
Journal Publications
1. Timothy T. Y. Tan, Donia Beydoun, Rose Amal. “Photocatalytic Reduction of Se(VI) in Aqueous Solutions in UV/TiO2 System : Importance of Stoichiometric Ratio of Reactants on TiO2 Surface”. Journal of Molecular Catalysis A: Chemical, 2003 202(1-2 ) Page. 73-85. (Chapter 3A) 2. Timothy T. Y. Tan, Donia Beydoun, Rose Amal. “Effects of Organic Hole Scavengers on the Photocatalytic Reduction of Se ions”. Journal of Photochemistry and Photobiology A: Chemistry, 2003 159(3), Pages 273-280. (Chapter 3B) 3. Timothy T. Y. Tan, Donia Beydoun, Rose Amal. “Photocatalytic Reduction of Se(VI) in Aqueous Solutions in UV/TiO2 System: Kinetic Modeling and Reaction Mechanism”. Journal of Physical Chemistry B, 2003 107(18) Pages 4296-4303. (Chapter 3C) 4. Timothy T. Y. Tan, Myint Zaw, Donia Beydoun, Rose Amal. “The Formation of Nano-Sized Selenium-Titanium Dioxide Composite Semiconductors by Photocatalysis”. Journal of Nanoparticle Research, 2002 4(6), Pages 541-552. (Chapter 4A) 5. Timothy T. Y. Tan, Chee Kin Yip, Donia Beydoun, Rose Amal. “Effects of Nano- Ag Particles Loading on TiO2 Photocatalytic Reduction of Selenate Ions”. Corrected proof available online in The Chemical Engineering Journal: Environmental. (Chapter 4B) 6. Timothy T. Y. Tan, Pierre Pichat, Donia Beydoun, Rose Amal. “A Study on the Photoconductance and Oxygen Isotopic Exchange of Se-loaded TiO2 Particles and their Photocatalytic Efficiency in Gas-Phase Methanol Photooxidation”. In preparation.
Refereed Conference Proceedings
1. Timothy T. Y. Tan, Myint Zaw, Donia Beydoun, Rose Amal. “The Formation of Nano-sized Selenium-Titanium Dioxide Composite Semiconductors by Photocatalysis” in the World Congress of Particle Technology 4 (21 – 25 July, 2002), Sydney, Australia. 2. Timothy T. Y. Tan, Chee Kin Yip, Myint Zaw, Donia Beydoun, Rose Amal. “Effects of Nano-Ag Particles Loading on TiO2 Photocatalytic Reduction of Selenate Ions” in the 9th APCChE Congress and CHEMECA 2002, 29 September – 3 October 2002, Christchurch, New Zealand.
Other Conference Proceedings
1. Timothy T.Y. Tan, Jason Scott, Myint Zaw, Donia Beydoun, Rose Amal. “Photocatalytic Reducitons of Se(IV) in Aqueous Solutions” in the The Fifth
F International Conference on TiO2 Photocatalytic Purification and Treatment of Water and Air, 2000, Ontario, Canada. 2. Timothy T. Y. Tan, Myint Zaw, Donia Beydoun, Rose Amal. “The Formation of Nano-sized Selenium-Titanium Dioxide Composite Semiconductors by Photocatalysis” in the 1st International Student Congress of Particle Technology (19 – 21 July, 2002), Terrigal, Australia. 3. Timothy T. Y. Tan, Donia Beydoun, Myint Zaw, Rose Amal. “Deposition of Se Nano-particles onto TiO2 Photocatalyst using Se(IV) and Se(VI) as Se precursors” in the 14th International Conference On Photochemical Conversion And Storage Of Solar Energy (IPS-14) Hokkaido University, Sapporo, Japan August 4-9, 2002. 4. Timothy T. Y. Tan, Myint Zaw, Donia Beydoun, Rose Amal. “The Synthesis Of Nano Se-Tio2 P-N Composite Semiconductors By Photocatalysis” in the The Seventh International Conference on TiO2 Photocatalysis, November 17-21, 2002, Ontario, Canada. 5. Timothy T. Y. Tan, Myint Zaw, Donia Beydoun, Rose Amal. “Application Of The Photocatalytic Reduction of Se(VI) And Se(IV) by Ag-Modified And Unmodified TiO2 In The Synthesis of Nano Metallic Selenide Semiconductors” in the The Seventh International Conference on TiO2 Photocatalysis, November 17-21, 2002, Ontario, Canada. 6. Timothy T. Y. Tan, Donia Beydoun, Rose Amal. “The Synthesis of Se-TiO2 p-n Composite Semiconductor by Photocatalysis and its Implicationas” in the 7th International Conference on Solar Energy and Applied Photochemistry, 23-28 February, 2003, Luxor, Egypt. 7. Timothy T. Y. Tan, Donia Beydoun, Rose Amal. “Effects of Organic Hole Scavengers on the Photocatalytic Reduction of Se ions” in The Eighth International Conference on TiO2 Photocatalytic Purification and Treatment of Water and Air th (TiO2-8), October 26-30 2003, Sheraton Centre Hotel of Montreal, Canada. 8. R. Enriquez, T. Tan, and P. Pichat, “Probing By Oxygen Isotope Exchange The Effects Of Various TiO2 Treatments On Surface Oxygen Lability And TiO2 Accessibility. Comparison With Variations In Methanol Photocatalytic Removal In Air,” in The Eighth International Conference on TiO2 Photocatalytic Purification th and Treatment of Water and Air (TiO2-8), October 26-30 2003, Sheraton Centre Hotel of Montreal, Canada.
G TABLE OF CONTENTS
CHAPTER 1: INTRODUCTION...... 1
CHAPTER 2: LITERATURE REVIEW...... 6
2.1 HETEROGENEOUS PHOTOCATALYSIS...... 6 2.1.1 Introduction...... 6 2.1.2 Semiconductor Photocatalysis...... 8 Electronic Properties of Semiconductor...... 8 Mechanistic of TiO2 Semiconductor Photocatalysis...... 10 Photocatalytic Oxidation...... 15 Photocatalytic Reduction...... 16 2.1.3 Quantum Yield...... 17 2.1.4 Fundamental Parameters in Photocatalysis...... 19 Effect of Oxygen...... 19 Effect of pH...... 21 Effect of Temperature...... 23 Effect of Catalyst Loading...... 23 Effect of Initial Solute Concentration and Kinetic Modelling...... 24 2.1.5 Strategies for the Photocatalytic Enhancement: Quantum Efficiency and Visible- light Absorption...... 26 Addition of Electrons and Holes Scavenger...... 27 Noble Metal Deposition on TiO2...... 27 Sensitization of TiO2: Metal Ion Implantation, Dye Sensitization, Composite Semiconductors...... 30 Quantum-size (Q-size) Effect on Photocatalysis...... 33 2.2 SELENIUM ...... 35 2.2.1 Introduction...... 35 2.2.2 Properties of Selenium and Selenium Anions...... 35 Physical Properties...... 35 Chemical Properties...... 36 2.2.4 Applications...... 38 Industrial Applications...... 38 Selenium as Essential Nutrient...... 39 2.2.5 Selenium Distribution in the Ecosystem...... 40
I Selenium Speciation in Natural Water...... 40 2.2.6 Toxicity of Selenium Compound...... 42 2.2.7 Selenium Sources...... 43 2.2.8 Removal of Selenium Compounds...... 44 Comparison of Se Anion Removal Techniques...... 47 CHAPTER 3. PHOTOCATALYTIC REDUCTION OF SELENATE AND SELENITE...... 49
3.1 INTRODUCTION...... 49
3A. INVESTIGATIONS OF SE IONS PHOTOREDUCTION AT VARIOUS PARAMETERS ...... 51 3A.1 Introduction...... 51 3A.2 Equipment and Procedure...... 54 Catalyst and Reagent...... 54 Apparatus and Procedure...... 54 Adsorption Data and Isotherm Determination of Se Ions on TiO2 ...... 57 Analysis...... 57 3A.3 Results and Discussions...... 58
Determination of TiO2 Surface Charge...... 58 Se(VI) Adsorption Studies...... 59 Preliminary Se(VI) Reduction Studies...... 63 The Redox Reactions on the TiO2 Surface...... 65 Effect of pH on Se(VI) Photoreduction...... 67 Effects of Initial Solute Concentration on Se(VI) Photoreduction...... 70 Validation of the Significance of Optimum Adsorption for Se(VI) Photoreduction71 Se(IV) Adsorption Studies...... 74 Preliminary Se(IV) Reduction Studies...... 78 The Effect of pH on Se(IV) Photoreduction...... 79 Effects of Initial Solute Concentration on Se(IV) Photoreduction...... 80 Validation of the Significance of Optimum Adsorption for Se(IV) Photoreduction81 Comparison of the Photocatalytic Reductions for Se(VI) and Se(IV)...... 83 3A.4 Conclusions...... 84
3B. EFFECTS OF ORGANIC HOLE SCAVENGERS ON THE PHOTOCATALYTIC REDUCTION OF
SELENATE AND SELENITE OVER UV-ILLUMINATED TI02...... 86 3B.1 Introduction...... 86 3B.2 Equipment and Procedure...... 88 Catalyst and Reagent...... 88 Photoreactor and Experimental Procedure...... 88 Analysis...... 89
II 3B.3 Results and Discussions...... 89 3B.4 Conclusions...... 99
3C. KINETIC MODELLING AND REACTION MECHANISM OF SE(VI) PHOTOREDUCTION...101 3C.1. Introduction...... 101 3C.2 Equipment and Procedure...... 102 3C.3 Results and Discussions...... 103 Kinetic Modelling of Se(VI) Photocatalytic Reduction...... 103 Proposed Mechanism...... 110 3C.4 Conclusions...... 111
CHAPTER 4. SYNTHESIS OF (NANO) SE-COMPOUNDS BY PHOTOCATALYSIS...... 113
4.1 INTRODUCTION...... 113
4A. THE FORMATION OF NANO-SIZED SELENIUM-TITANIUM DIOXIDE COMPOSITE
SEMICONDUCTORS BY PHOTOCATALYSIS...... 114 4A.1 Introduction...... 114 4A.2 Equipment and Procedure...... 116 Catalyst and Reagent...... 116 Procedures for Selenium Photodeposition...... 117 Procedures for Photocatalytic Removal of Cadmium Ions...... 117 Procedures for Se ion Adsorption...... 117 Analysis...... 117 4A.3 Results and Discussions...... 118 Se Particle Characterisation...... 118 Se Formation Mechanism from Se(VI) Photoreduction...... 120 Se Formation Mechanism from Se(IV) Photoreduction...... 124 Effects of pH on Se Formation...... 128 Implications of Current Work...... 130 4A.4 Conclusions...... 133
4B. EFFECTS OF NANO-AG PARTICLES LOADING ON TIO2 PHOTOCATALYTIC REDUCTION
OF SELENATE AND SELENITE IONS TO H2SE GAS ...... 134 4B.1 Introduction...... 134 Metal-n-type-semiconductor Junction...... 135 4B.2 Equipment and Procedures...... 137 Catalyst and Reagent...... 137
III Preparation of Ag-TiO2...... 137 Experimental Procedures for Se Ion Adsorption and Photoreduction...... 138 Particle Characterisation...... 139 4B.3 Results and Discussions...... 139
Characterisations of Ag Deposits on TiO2...... 139 Photocatalytic Reduction of Se(VI) by Bare TiO2 and Ag-deposited TiO2 ...... 141 Effects of Ag loading and pH on the Photocatalytic Reduction of Se(VI)...... 144 Photocatalytic Reduction of Se(IV) by Bare TiO2 and Ag-deposited TiO2 ...... 147 Effects of Ag Loading, pH and Formic Acid Concentration on the Photocatalytic Reduction of Se(IV)...... 148 Comparison of Se(VI) and Se(IV) Photoreduction with Ag-TiO2 ...... 150 Implications of Current Work...... 151 4B.4 Conclusions...... 153
CHAPTER 5. CONCLUSIONS...... 155
CHAPTER 6. RECOMMENDATIONS...... 158
REFERENCES...... 161
APPENDIX...... 177
A. DERIVATION OF SE(VI) AND FORMATE ADSORPTION MODEL ...... 177
B. ACTINOMETRY: PHOTON FLUX AND QUANTUM EFFICIENCY DETERMINATION ...... 182
IV LIST OF FIGURES
Figure 2. 1. The effect of the increase in the number N of monomeric units from unity to cluster of more than 2000 on the electronic structure of a semiconductor compound (Hoffmann et al, 1995)...... 9 Figure 2. 2. Illustrations of electrochemistry of TiO2 particles in contact with an electrolyte and after irradiation. (a) Fermi level near the conduction band in TiO2 prior to contact with an electrolyte. (b) Band-bending and the presence of an electric field after contact with electrolyte. (c) Transfer of the photogenerated electrons and holes upon irradiation. EVB: Valence band, ECB: Conduction band...... 11 Figure 2. 3. The possible pathways of the main charge carriers occurring in a TiO2 particle following photoexcitation. EVB: Valence band, ECB: Conduction band...... 12 Figure 2. 4. Electron mediation by Ag in contact with TiO2 surface...... 28 Figure 2. 5. Electron transfer during the photoexcitation of a dye in the presence of TiO2 and an electron acceptor (A+)...... 31 Figure 2. 6. Charge transfer in a CdS-TiO2 coupled system...... 32 Figure 2. 7. Charge transfer in a SnO2-TiO2 shell-and-core semiconductor system...... 33 Figure 2. 8. Structure of Se8 monoclonic structure (Langner et al, 1996)...... 37 Figure 2. 9. Average composition of Se species in 3 different samples of groundwater and soil water (Reddy, 1998)...... 42 Figure 3. 1. (a) Schematic diagram and (b) photo of the experimental setup. P: power supply, L: lamp housing with Hg lamp, R: glass reactor, M: magnetic stirrer, Q: quartz window, S: sampling port, Pr: pH probe, S1: CuSO4 scrubber for H2Se, S2: NaOH scrubber for H2Se...... 56 Figure 3. 2. Surface charge of TiO2 at various pH in the absence and presence of Se ions (20 ppm)...... 58 Figure 3. 3. Adsorption Isotherm of Se(VI). Experimental conditions: pH 3.5, 1 L test solution, 1.1 gTiO2/L, N2 purging, 293 K. Insert: Linearised BET model plot.59 Figure 3. 4. Adsorption Isotherm of HCOOH. Experimental conditions: pH 3.5, 1 L test solution, 1.1 gTiO2/L, N2 purging, 293 K. Insert: Linearised LH Adsorption Model plot...... 60 Figure 3. 5. Effect of pH on Se(VI) adsorption. Experimental conditions: 1 L test solution, [Se(VI)]0=20ppm (0.256 mM), [HCOOH]0=100ppmC (8.3 mMC), 1.1 gTiO2/L, N2 purging, 293K...... 61 Figure 3. 6. The simultaneous adsorption of Se(VI) and HCOOH at various pH. Experimental conditions: 1 L test solution, [Se(VI)]0=20 ppm, [HCOOH]0=100 ppmC, 1.1 gTiO2/L, N2 purging, 293 K...... 62 Figure 3. 7. Adsorption of Se(VI) and HCOOH at various initial HCOOH Concentration. Experimental conditions: pH 3.5, 1 L test solution, 1.1 gTiO2/L, N2 purging, 293 K. ¨ Se(VI) adsorbed, ■ formate adsorbed...... 63 Figure 3. 8. Degussa TiO2 powder collected on filter paper (a) before and (b) after Se ions photoreduction...... 66 Figure 3. 9. Effect of pH on the photoreduction of Se(VI). □ pH 2.2 (R2=0.996), × pH 3.0 (R2=0.996), ○ pH 6.4 (R2=0.854). Experimental conditions: 1 L test solution,
V [Se(VI)]0=20 ppm, [HCOOH]0=100 ppmC, 1.1 gTiO2/L, N2 purging, 293 K, photon intensity 3.15 mmol/sec...... 68 Figure 3. 10. The effect of pH on Se(VI) photoreduction rate. Experimental conditions: 1 L test solution, [Se(VI)]0=20 ppm, [HCOOH]0=100 ppmC, 1.1 gTiO2/L, N2 purging, 293 K, photon intensity 3.15 mmol/sec...... 69 Figure 3. 11. The effect of initial HCOOH concentrations on Se(VI) photoreduction rates. Experimental conditions: pH 3.5, 1 L test solution, 1.1 gTiO2/L, N2 purging, 293 K. □ 20 ppm and ○ 40 ppm initial Se(VI) concentration...... 71 Figure 3. 12. Adsorption Isotherm of Se(IV) at pH 3. Experimental conditions: 1 L test solution, 1 gTiO2/L, N2 purging, 293 K...... 75 Figure 3. 13 The ionic structure of (a) Se(VI) and (b) Se(IV)...... 76 Figure 3. 14. The effects of pH on Se(IV) adsorption in the presence and absence of HCOOH. Experimental conditions: 1 L test solution, [Se(IV)]0 =20 ppm (0.256 mM), [HCOOH]0 =100ppmC (8.3 mMC), 1.1 gTiO2/L, N2 purging, 293K.....77 Figure 3. 15. Equilibrium dark adsorption of Se(IV) and HCOOH on TiO2 surface at 20 ppm initial Se(IV) concentrations, 50-2000 ppmC HCOOH, pH 3.0, 293 K, 1L test solution, N2 purging...... 78 Figure 3. 16. The effect of pH on Se(IV) photoreduction rate (■) and the corresponding formate:selenite molar adsorption ratio (▲). Experimental conditions: 1 L test solution, [Se(IV)]0=20 ppm, [HCOOH]0=300 ppmC, 1.1 gTiO2/L, N2 purging, 293 K, photon intensity 3.17 mmol/sec...... 80 Figure 3. 17. The effect of initial HCOOH concentrations on Se(IV) photoreduction rates and the corresponding molar adsorption ratio. Experimental conditions: 20 ppm initial Se(IV) concentrations, pH 3.0, 1 L test solution, 1.1 gTiO2/L, N2 purging, 293 K, photon intensity 3.17 mmol/sec...... 81 Figure 3. 18. Comparison of the disappearance of Se(VI) and Se(IV) ions and the generation of H2Se from Se ions precursors. Experimental condition: 10 ppm initial Se ions concentrations, 300 ppmC HCOOH concentrations, pH 3.5, 1 L test solution, 1.1 gTiO2/L, N2 purging, 293 K, photon intensity 3.17 mmol/sec...... 84 Figure 3. 19. The extent of mineralisation of various organic compounds in the presence of nitrogen. Experimental conditions: 15 mg carbon of organic used, 0.15 L test solution, 2.0 gTiO2/L, N2 purging, pH 3.0...... 92 Figure 3. 20. Relative position of the conduction band and valance band edges for TiO2 in a pH 3 aqueous medium and selected redox levels (Chenthamarakshan et al, 2000a; Seby et al, 2001)...... 94 Figure 3. 21. Photocatalytic reduction of Se(IV) in various concentration of formic acid, methanol and ethanol at pH 3.0 in 120 minutes. Experimental conditions: 20 ppm initial Se(IV) concentration, 1 L test solution, 1.1 gTiO2/L, N2 purging, 293 K...... 97 Figure 3. 22. Comparison of Se ions dark adsorption and photoreduction at various pH and 300 ppmC methanol and ethanol. 20 ppm initial Se(IV) concentrations, 1.1 g/LTiO2, 293 K, 1L test solution, N2 purging...... 98 Figure 3. 23. Comparison of experimental and modeled data for Se(VI) adsorption 2 -1 -1 according to Equation (3-7). R =0.996, KSe=298 mmol , KF=0.079 mmol and Ct=0.071 mmol/gTiO2...... 104
VI Figure 3. 24. Comparison of experimental and modeled data for formate adsorption 2 5 -1 -1 according to Equation (3-8). R =0.97, KSe=8.5x10 mmol , KF=3.9 mmol and Ct=1.2 mmol/gTiO2...... 105 Figure 3. 25. Comparison of experimental and modeled data for Se(VI) photoreduction rate according to Equation (3-9). □: Experimental rate and —: Modeled rate for 0.256 mM initial Se(VI) concentration, ◊: Experimental rate and ---: Modeled 2 -9 rate for 0.512 mM. initial Se(VI) concentration. R =0.96, krxn=8.16x10 -1 -1 gTiO2/mol.min, KSe=1320.5 mmol , KF=42.5 mmol and Ct=1320.5 mmol/gTiO2...... 106 Figure 3. 26. Optimisation of Equation (3-8) with respect to molar adsorption ratio of 3:1 at HCOO-:Se(VI). —: Modeled rate for 0.256 mM initial Se(VI) concentration, --- : Modeled rate for 0.512 mM. initial Se(VI) concentration...... 108 Figure 3. 27. Influence of catalyst loading on Se(VI) photoreduction rate. Experimental conditions: pH=3.5, 1 Litre test solution, [Se(VI)]0=0.256 mM, [HCOOH]0=16.7 mMC, N2 purging, 293K...... 108 Figure 4. 1. Band diagram showing the electron energy as a function of distance of a p-n junction (Diagram redrawn from Dalven, 1990a)...... 115 Figure 4. 2. Effect of applying a forward bias on a p-n junction. Electric field Ea resulting from the applied forward bias, E is the built-in electric field at the junction and W is the width of the space charge region (Dalven, 1990a)...... 116 Figure 4. 3. TEM image of Se particles on TiO2 particles...... 119 Figure 4. 4. Electron diffraction pattern by EDX of a Se particle deposited during the photoreduction process...... 120 Figure 4. 5. Experimental data showing Se(VI) reduction, H2Se generation and formic acid oxidation. Experimental conditions: Initial Se(VI) concentration: 20 ppm, initial formic acid concentration: 300 ppmC, irradiation time: 420 minutes, pH 3.5, N2 purging, 0.5 gTiO2/L...... 121 Figure 4. 6. Energy diagram of Se-TiO2 system (pH=4) showing the position of the flatband potentials and the bandgap energy of TiO2 (n-type) and Se (p-type) upon UV irradiation. Bandgap energy (Eg): Eg-TiO2=3.2eV (Hagfeldt & Gratzel, 1995; Tada et al, 1998), Eg-Se=1.95eV (Streltsov et al, 2002). Conduction band potential (CB): CBTiO2=-0.3eV (Tada et al, 1998; Chenthamarakshan et al, 2000b), CBSe=-1.65eV (Streltsov et al, 2002)...... 122 Figure 4. 7. a and b show the TEM images of Se-TiO2 particles with Se(IV) and Se(VI) as precursors respectively. c and d show the corresponding Se mapping by TEM with Se(IV) and Se(VI) as precursors respectively. Experimental conditions: Initial Se(IV)/Se(VI) concentration: 20 ppm, initial formic acid concentration: 300 ppmC, irradiation time: 90 minutes, pH 3.5, N2 purging...... 126 Figure 4. 8. Adsorption of Se(IV) and Se(VI) onto TiO2 at various pH. Experimental conditions: Initial Se(IV)/Se(VI) concentration: 20 ppm, initial formic acid concentration: 300 ppmC, N2 purging...... 126 Figure 4. 9. High magnification (970K) of TiO2 particles before (a) and after (b) Se(IV) photoreduction. Initial Se(IV) concentration = 20 ppm...... 127 Figure 4. 10 Film of Se deposits on TiO2 particles. Initial Se(IV) concentration = 80 ppm...... 127
VII Figure 4. 11. Effect of pH on the size of Se particles deposited and the corresponding amount of Se(VI) reduced. Experimental conditions: Initial Se(VI) concentration: 20 ppm, initial formic acid concentration: 300 ppmC, N2 purging, 0.5 L test solution and 0.5 gTiO2...... 129 Figure 4. 12. Se mapping by TEM with Se(IV) as precursor. Experimental conditions: Initial Se(IV): 20 ppm, initial formic acid concentration: 300 ppmC, irradiation time: 90 minutes, N2 purging, 0.5 gTiO2. RSe = Amount of Se(IV) reduced (mg)...... 130 Figure 4. 13. Comparison of Cd ion removal using (b) UV/Degussa P25, (c) UV/Se-TiO2 and (a) UV only...... 131 Figure 4. 14. TEM image of CdSe on TiO2 particles...... 132 Figure 4. 15. UV-Vis absorbance of TiO2, Se-TiO2 and CdSe-TiO2...... 132 Figure 4. 16. Simplified version of the metal-n-type semiconductor junction for the case in which Fm is greater than Fs (Diagram modified from Dalven, 1990a)...... 136 Figure 4. 17. X-ray Diffractogram of Ag Deposited Degussa P25 particles...... 140 Figure 4. 18. Figure a, b and c show the mapping of Ti, Ag and Se respectively by EDX. Figure d shows the corresponding TEM image of Se on TiO2. Se(VI) was used as the Se precursor...... 140 Figure 4. 19. Se(VI) photoreduction experiments performed using unmodified TiO2 and 2 atomic % Ag-TiO2. ¨ Se(VI) concentration, ■ H2Se generation, ▲ formic acid oxidation. Experimental conditions: 1 L test solution, [Se(IV)]0=20 ppm, [HCOOH]0=300 ppmC, pH 3.5, 1.1 gTiO2/L, N2 purging, 293 K, photon intensity 3.06 mmol/sec...... 142 Figure 4. 20. Energy diagram of the TiO2-Ag-Se system at pH 3.5. Bandgap energy (Eg): Eg-TiO2= 3.2eV (Hagfeldt & Gratzel, 1995), Eg-Se=1.95eV (Streltsov et al, 2002). Conduction band potential (CB): CBTiO2=-0.3eV (Tada et al, 1998), CBSe=- 1.65eV (Streltsov et al, 2002).. Work function (F): FTiO2=4.2eV, FAg=4.6eV, FSe=4.8eV (assuming Fermi levels are near the conduction and valence band for TiO2 and Se respectively and the work functions are calculated by the equations given in Henglein ( 1997)...... 143 Figure 4. 22. Surface charge of TiO2 and Ag-TiO2 particles at various pH...... 146 Figure 4. 23.Comparison on of Se(IV) reduction using TiO2 and Ag-TiO2 (1 atomic %). ■ Se(IV) concentration using TiO2, × Se(IV) concentration using Ag-TiO2, ¨ H2Se generation using TiO2, ▲ H2Se generation using Ag-TiO2. Experimental conditions: 1 L test solution, [Se(IV)]0=10 ppm, [HCOOH]0=400 ppmC, pH 3.0, 1.1 gTiO2/L, N2 purging, 293 K, photon intensity 3.17 mmol/sec...... 148 Figure 4. 24. Comparison of Se(IV) and Se(VI) photoreduction using 1 atomic % Ag-TiO2. × Se(IV) concentration, ■ Se(VI) concentration, ▲H2Se generation with Se(IV) as precursor, ¨ H2Se generation with Se(VI) as precursor. Experimental conditions: 1 L test solution, [Se(IV)/Se(VI)]0=11 ppm, 1.1 gAg-TiO2/L, N2 purging, 293 K, photon intensity 3.17 mmol/sec...... 151
VIII LIST OF TABLES
Table 2. 1. Summary of Technologies for Selenium Compounds Removal in Water and Wastewater...... 45 Table 3. 1. Preliminary Experiments of Se(VI) Reduction. Experimental conditions: [Se(VI)]0=20 ppm (0.256 mM), [HCOOH]0=100 ppmC (8.3 mMC), 1L test solution, pH= 2.6±0.1, 1.1 g TiO2/L, 293 K. Irradiation time: Expt 1 & 2: 30 mins, Expt 3-5: 60 mins...... 65 Table 3. 2. Mass balance of selenium species before and after photoreduction. TiO2 loading = 1.1g/L, reaction volume = 1L, residence time = 65min, pH=3.5...... 67 Table 3. 3. Effect of formate:Se(VI) molar adsorption ratio on Se(VI) photoreduction rate. Experimental conditions for Set A and B: 1 L test solution, 1.1 gTiO2/L, 293 K, N2 purging, 60 min residence time. Set A: [Se(VI)]0=10 ppm, [HCOOH]0=50 ppm, Set B: [Se(VI)]0=40 ppm, [HCOOH]0=200 ppm...... 72 Table 3. 4. Effect of varying the order of Se(VI) and formic acid addition into the TiO2 suspension. Experimental conditions for Set C and D: pH 3.5, 0.5 L test solution, 0.5 g TiO2/L, 293 K, N2 purging, 60 min residence time. Set C : [Se(VI)]0=20 ppm, [HCOOH]0=20 ppmC, Set D: [Se(VI)]0=20 ppm, [HCOOH]0=100 ppmC...... 74 Table 3. 5. Preliminary experiments of Se(IV) photoreduction. Experimental conditions: [Se(VI)/ Se(IV)]0=20 ppm (0.256 mM), [HCOOH]0=100 ppmC (8.3 mMC), pH= 2.6±0.1, 1.1 g TiO2/L, 293 K...... 79 Table 3. 6. Effect of formate:Se(IV) molar adsorption ratio on Se(IV) photoreduction rate. Experimental conditions: 1 L test solution, 1.1 gTiO2/L, 293 K, N2 purging, 60 min residence time. [Se(IV)]0= 10 ppm, Set E: [HCOOH]0=235 ppmC, Set F: [HCOOH]0=400 ppmC...... 82 Table 3. 7. Results of the adsorption and photocatalytic reduction of Se anions in the presence of various organic additives. Experimental conditions: 20 ppm initial Se anions concentration, 300 ppmC initial organic concentration, pH 3.0, 1 L test solution, 1.1 gTiO2/L, N2 purging, 293 K...... 90 Table 3. 8. The comparison between the individual adsorption of the organic additives and the simultaneous adsorption of Se(IV) ions and the organic additives. Experimental conditions: 20 ppm initial Se(IV) concentration, 300 ppmC initial organic concentration, pH 3.0, 1 L test solution, 1.1 gTiO2/L, N2 purging, 293 K...... 91 Table 4. 1. Results for Se(VI) photoreduction. Experimental conditions: 1 L test solution, [Se(IV)]0 =20 ppm, [HCOOH]0=300 ppmC, 1.1 gTiO2/L, N2 purging, 293 K, photon intensity 3.17 mmol/sec...... 145 Table 4. 2. Summary of results for Se(IV) photoreduction. Experimental conditions: 1 L -2 test solution, [Se(IV)]0=10 ppm (12.7×10 mM), 1.1 gTiO2/L, N2 purging, 293 K, photon intensity 3.06 mmol/sec...... 149 Table 4. 3. Metal Selenide formed by passing H2Se into metal ions solution and the corresponding yield. pKsp values obtained from Seby et al (2001).1)...... 153
IX NOMENCLATURE
u frequency of light q fraction of surface coverage by the substrate A+ electron acceptor B constant expressive of energy of adsorption C substrate concentration at any time t C0 initial substrate concentration C0 catalyst concentration Ce residual solute concentration at equilibrium CF amount of solute adsorbed per unit weight of adsorbent CF formic acid concentrations at equilibrium CF-m monolayer saturation concentration CFs formate ion adsorbed onto one site on the TiO2 surface Cs saturation concentration CSe residual Se(VI) concentrations at equilibrium CSess Se(VI) ion adsorbed onto the two sites in the TiO2 surface Ct total available sites on the catalyst surface D electron donor E electric field e- Electrons Ea apparent activation energy ECB conduction band EF(m) Fermi energy in the metal EF(s) Fermi energy in the semiconductor EF. Fermi level Eg optical band-gap of the semiconductor Et true activation energy eV energy barrier for electron transfer Evac vacuum level/energy EVB valence band h Planck’s constant hu photon energy h+ Holes K adsorption constant k’ apparent rate constant kC rate constant for the catalyst KF adsorption constants for HCOO kr reaction rate constant krxn apparent rate constant krxn′ new reduction rate constant KSe adsorption constants for Se(VI) N monomeric units in a particle
X • O2 superoxide radicals OH• hydroxyl radicals pHzpc point of zero charge Q0 solid-phase concentration corresponding to complete coverage of available sites qe amount of solute adsorbed per unit weight of adsorbent r reaction rate r’ initial Se(VI) photoreduction rates Sc Semiconductor W depletion zone or space charge region ζr relative photonic efficiency DEms energy barrier between the metal and the semiconductor DEsm energy barrier for an electron going from the bulk semiconductor into the metal F work function Fm work function of the metal Fs work function of the semiconductor c, electron affinity f Quantum yield w light attenuation factor
XI Chapter 1: Introduction
Remediation of polluted surface and groundwater as well as the treatment of wastewater are long-term goals to improve the quality of water for human consumption and to maintain proper functioning of the ecosystem. This is especially so in arid areas where water is a very valuable commodity and where reuse and recycling of treated water is of high priority. Purification of water, such as municipal wastewaters, is conventionally achieved by combined processes usually including both physical and biological processes, such as flocculation, filtration and biological treatment. These processes are relatively low in cost and well-established. However, such removal techniques are not able to remove low level of toxic inorganic and organic contaminants, which are commonly found in industrial effluent discharge. For such wastewaters, chemical treatment methods have been employed to remove the low levels of toxic compounds to sufficiently safe concentrations for human use.
Chemical treatment methods can involve the addition of powerful oxidant such as ozone or hydrogen peroxide, which are capable of oxidising most of the toxic organic compounds. In recent decades, a stream of chemical treatment process known as Advanced Oxidation Processes (abbreviated as AOPs) has come under intense research. All AOPs are characterised by the generation of highly oxidising hydroxyl radicals (OH•), which are capable of achieving complete abatement of the pollutants through mineralisation to carbon dioxide, water, and inorganic ions such as nitrate, chloride, sulfate and phosphate. Some of these processes involve the use of oxidising chemicals alone, such as ozone and hydrogen peroxide, while others combine such chemicals with UV-visible light irradiation. The combined use of oxidants and irradiation has been found to significantly accelerate the degradation of the pollutants. Common photoassisted AOPs are UV/Fenton processes,
Ozone/UV, H2O2/UV and TiO2/UV (Andreozzi et al, 1999).
1 The AOPs could be further categorized into homogeneous and heterogeneous photocatalysis, the difference being the involvement of single phase and dual phase systems respectively. Among the above-mentioned AOPs, titanium dioxide (TiO2) photocatalysis, a heterogeneous process, is the fastest growing in terms of research as reflected in the numerous publications in the last three decades (Blake, 1994; 1995; 1997). This material is highly versatile, being able to operate in a wide range of pH and degrade a huge variety of organic compounds. It is activated by near UV-irradiation, and hence could be activated by • sunlight. When TiO2 is irradiated, oxidising species such as the OH could be generated, rendering the addition of expensive and toxic oxidising chemicals unnecessary. The generation of oxidising species (the “holes”) in TiO2 photocatalysis involves the simultaneous generation of reducing species (the electrons). Hence, TiO2 photocatalysis is also capable of reducing toxic pollutants, such as heavy metal ions and some organic pollutants. Due to these synergistic processes, photocatalysis should indeed be considered as a combined advanced oxidation and reduction technology.
Many investigations involving TiO2 photocatalysis have focused on the oxidation of organic pollutants. This is because most pollutants are organic in nature. On the US EPA national primary drinking water guidelines, there are 56 listed organic pollutants and only 16 listed inorganic pollutants. Nevertheless, many inorganic pollutants are extremely toxic and carcinogenic at low concentrations. It is hence of great urgency to find a low cost, as well as an effective and efficient removal method. The inorganic pollutants, which mainly exist in soluble forms and higher oxidation states, could be reduced to insoluble forms for subsequent removal. The reduced form of the pollutants could then be recovered and reused. TiO2 photocatalysis shows great promise (Fujishima et al, 2000b).
2- The pollutants of interest in this work are the selenium (Se) oxyanions: selenate (SeO4 , 2- Se(VI)) and selenite (SeO3 , Se(IV)). These inorganic anions are highly mobile due to their high solubility. In recent years, mining, industrial and agricultural activities have caused significant elevation and widespread Se(VI)/Se(IV) in the natural waterways, presenting toxicological hazards to plants, marine and human lives. This warrants an efficient and economical technique to remove the Se ions from the waterways. The present maximum
2 contaminant level for Se for US and Australia drinking water standard is 10 mg/L or ppb (Jones & French, 1999).
Even though TiO2 photocatalysis shows tremendous potential in water and wastewater treatment processes, there are several issues which have to be addressed before it can be considered as a mainstream technology. One of the major challenges is to overcome its low quantum efficiency in order to favourably compete with other AOPs. The rapid recombination of the photogenerated holes and electrons, which are responsible for the activities in photocatalysis, results in their short-lived presence on the catalyst surface and hence low quantum efficiency. The surface modification of the TiO2 catalysts by deposition of noble metals, such as silver or platinum, has been widely investigated as one of the methods to enhance the photocatalytic efficiency. The presence of the noble metal is believed to mediate the electrons from the TiO2 surface, preventing their recombination with the holes.
Another drawback of TiO2 photocatalysis is the fact that the semiconductor TiO2 can only be activated by near UV irradiation (wavelength less than 380 nm), which constitutes less than 5% of the total solar spectrum. Many works have investigated the enhancement of the photo-response of TiO2 in the visible region. Methods used include the coupling with a semiconductor which is sensitive to visible light, such as CdS, to form a composite semiconductor of which efficient activation can be achieved by visible light irradiation alone (Spanhel et al, 1987). Other methods investigated include dye sensitization of TiO2
(Kamat, 1989) and ion implantation into the TiO2 crystal matrix (Yamashita el al, 2002).
The current work presented in this Ph.D. thesis aims to investigate the photocatalytic reduction of Se(VI) and Se(IV). This work is divided into two parts. The first part of this work demonstrates the feasibility of using photocatalysis as a technique for Se(VI) and Se(IV) photoreduction in the context of water and wastewater remediation. More specifically, the aims for the first part of the study are (i) to understand the effect of various process parameters, such as pH, presence of oxygen and nitrogen, initial solute
3 concentrations, catalyst dosage, types of organic compounds (as hole scavengers) on the kinetics of the Se(VI) and Se(IV) photoreduction process, and (ii) to model the kinetics of the photoreduction process.
The second part of this work investigates the formation of nano-size Se on the TiO2 particles as a result of photoreduction of the Se ions to Se on the TiO2 surface. Elemental selenium has been used in the 1950s as the photoconductor in xerography due to its semiconducting property. Hence, it is also envisaged that the deposition of Se particles onto
TiO2 can initiate a new research area in the formation of nano-size Se-TiO2 composite semiconductors, or possibly other types of composite semiconductors. Nano-size material synthesis has attracted tremendous research interest due to their abundant potential benefits (Loughran, 2001; Lu, 2001). In the context of this work, the formation of nano-size Se-
TiO2 and the subsequent demonstration of this technique in the preparation of nano-size metal selenide semiconductors (such as CuSe, CdSe and ZnSe) and composite semiconductors have been investigated. The electronic and optical properties of these metal selenide semiconductors or composite semiconductors (such as CdSe/ CuSe/ ZnSe-TiO2) may be of interest for applications in optical instruments and the semiconductor industry.
With the above aims in mind, this thesis has been subdivided into the following sections. A review on TiO2 photocatalysis and selenium will be given in Chapter 2. Chapter 3 describes the experimental results for the photocatalytic reduction of Se(IV) and Se(VI), and Chapter 4 focuses on the formation of nano-Se compounds preceding the findings from Chapter 3. Chapter 3 is subdivided into 3 sections, each presenting (1) the preliminary investigations of Se(VI) and Se(IV) photoreduction under different experimental conditions and the optimisation of Se(VI) and Se(IV) photoreduction, (2) the effects of different types of organic additives on the Se ions photoreduction and (3) the kinetic modelling of Se(VI) photoreduction. Chapter 4 is subdivided into 2 sections. The first section discusses the formation of different morphologies of nano-Se particles on TiO2 and presents the possibility of removing Cd ions using Se-TiO2. The second section demonstrates the role of
Ag-modified TiO2 particles in the photoreduction of Se(VI) and Se(IV) and the feasibilities
4 of using TiO2 photocatalysis in the preparation of metal selenide semiconductors such as CuSe, CdSe and ZnSe.
Chapter 5 presents the conclusions of this study, followed by the recommendations for further work in Chapter 6.
5 Chapter 2: Literature Review
2.1 Heterogeneous Photocatalysis
2.1.1 Introduction
Since the start of the Industrial Revolution in the early nineteenth century, industries have created tremendous wealth for many nations. However, the creation of wealth also entailed the increased production of waste materials from industrial processes, which were often conveniently discharged into the ecosystem. Lessons learnt from many incidents, an example of which is the Minamata disease from mercury poisoning at Minamata Bay, Japan, have alerted us to the graveness of the problems and their repercussions if waste materials are not properly disposed of. Physical removal of the waste products by methods such as filtration and reverse osmosis merely concentrate the wastes and offer short-term solutions. The best approach is to covert waste products into innocuous compounds, which could then be safely released to the environment.
However, many waste treatment technologies utilise expensive chemicals which are usually toxic in high concentrations. For example, one type of AOPs utilises ozone and hydrogen peroxide for oxidising toxic organic compounds. The use of these chemicals for drinking water treatment could be highly controversial as there are concerns of residues in the treated water. On the other hand, technologies such as electrolysis and membrane processes utilise large amounts of electrical energy. These technologies are hence not “green” or environmentally friendly. In addition, there are also concerns about the ability of existing technologies to effectively remove the new type of pollutants. Hence, alternative treatment technologies conforming to the notion of green technology are urgently required in water treatment. Heterogeneous photocatalysis could be a suitable and promising candidate.
6 Photocatalysis could utilise sunlight as the energy source, which is available in abundance to initiate the photodecomposition process of pollutants. The end products of this treatment process are usually harmless compounds such as carbon dioxide, water and inorganic ions such as chloride and nitrate. The most intensely researched photocatalyst, titanium dioxide
(TiO2), is stable, non-toxic, relatively inexpensive to produce and can be reused. The fact that TiO2 could be reused makes this treatment process more economically attractive.
Since Fujishima and Honda (1972) demonstrated water photolysis into hydrogen and oxygen gas by the irradiation of a TiO2 electrode (anode) connected to a platinum electrode (cathode), photocatalytic processes have evolved into a possible technique in water remediation. The potential of this technology has been demonstrated by the intense research in the recent three decades. Various investigations have been performed to optimise and commercialise titanium dioxide photocatalysis, not only in waste water treatment but also in air purification. The outlook so far has been exciting and promising.
Many review papers on photocatalytic remediation processes have been presented in recent years (Chen et al, 2000; Mills & Le Hunte, 1997; Robertson, 1996; Hagfelt & Gratzel, 1995; Surender et al, 1998; Litter, 1999; Herrmann, 1999). These papers provide extensive details on the various aspects of photocatalytic processes. The prime intention of the following section is to briefly review the principles and chemistry of photocatalysis relevant to the current studies. The use of quantum yield to evaluate and compare different experimental results will be discussed first before the review on fundamental parameters affecting the photocatalytic rates. The discussion on kinetic modelling of photocatalytic reactions will also be incorporated into this section. Finally, an overview of different strategies to achieve enhancement in photocatalytic efficiency and photosensitization will be presented.
7 2.1.2 Semiconductor Photocatalysis
Electronic Properties of Semiconductor
TiO2 is able to degrade a wide range of recalcitrant organics and inorganic pollutants due to its ability to generate highly oxidising and reducing species. This is because as a semiconductor (Sc), it could be “photoexcited” by absorbing light of suitable wavelength to generate two types of electronic carriers, the electrons (e-) (the reducing species) and holes (h+) (the oxidising species) according to equation 2-1
n >Eh +- Sc ¾®¾ g +¾ he (2-1) where h is the Planck’s constant, u is the frequency of light and hu is the photon energy. To achieve photoexcitation, hu must be greater than Eg, which is known as the optical band- gap of the semiconductor (Rajeshwar & Ibanez, 1998). Eg is defined as the region devoid of energy levels as illustrated in Figure 2. 1 (Mills & Le Hunte, 1997). For many compounds, as the number N of monomeric units in a particle increases, the energy necessary to photoexcite the particle decreases. When N is much greater than 2000, a particle which exhibits the band electronic structure of a semiconductor could form. Such a particle would consist of a highest occupied energy band, the valence band (EVB), and the lowest unoccupied energy band, the conduction band (ECB), separated by Eg. Photoexcitation - involves the promotion of an e from the valence band (EVB) to the conduction band (ECB) upon adsorption of a photon of energy exceeding the band-gap (that is hu > Eg) by the semiconductor, simultaneously generating a hole, h+, in the valence band. A hole can be rationalised as a vacancy devoid of an electron.
The band-gap for two crystalline phases of the TiO2 semiconductor, rutile and anatase, are 3.02 eV and 3.23 eV, corresponding to photons of wavelength 413 nm and 388 nm
8 respectively (Chen et al, 2000). Hence, near-UV irradiation is required to generate the e- + and h in anatase TiO2.
Energy Atomic Molecule Cluster Q-size Semi- Orbitals Particle conductor N=1 N=2 N=10 N=2000 N>>2000
LUMO
Eg Eg Eg Eg
HOMO
Figure 2. 1. The effect of the increase in the number N of monomeric units from unity to cluster of more than 2000 on the electronic structure of a semiconductor compound (Hoffmann et al, 1995).
Semiconductors can be categorized into two main types: intrinsic and extrinsic semiconductors. An intrinsic semiconductor consists of negligible concentrations of defects and impurities. Thermal excitation is only possible in small band-gap semiconductors, which leads to e--h+ generation of identical concentration, giving rise to electrical conductivity. Extrinsic semiconductors have impurities added to their lattice and contain defects. If donor impurities are present in the extrinsic semiconductor, the semiconductor is termed an n-type semiconductor. Donor impurities contribute extra electrons to the intrinsic semiconductor. The resulting semiconductor has electrons as its major charge carriers and its Fermi level is close to its conduction band. The Fermi level (EF) shall be defined as the chemical potential of electrons in a solid (metal, semiconductors or insulators) or in an electrolyte solution. For a p-type semiconductor, the major charge carriers are holes due to
9 the presence of acceptor impurities. Acceptor impurities have a deficit of electrons and hence result in an excess of holes in the semiconductor. Hence, a p-type semiconductor has a Fermi level close to its valence band (Parmon et al, 2002).
Mechanistic of TiO2 Semiconductor Photocatalysis
TiO2 is an n-type semiconductor. The photoelectrochemistry of the n-type TiO2 semiconductor in contact with an electrolyte is depicted in Figure 2. 2. When the TiO2 particles are brought into contact with an electrolyte, the excess charge carriers (e-) are transferred from the TiO2 surface to the electrolyte in order to equilibrate the Fermi levels of the TiO2 and the electrolyte (Rajeshwar & Ibanez, 1998). This leads to band-bending and the formation of a thin-layer of space charge region, normally a few Angstroms in width on the surface of the TiO2. For TiO2 particles of sizes greater than 10 nm, the space charge layer formed on the planar electrodes must be formed in the surface region of the particles (Yoneyama, 1993). This layer is called the depletion layer as the majority charge carriers are depleted into the electrolyte (Hagfeldt & Gratzel, 1995). This region is positively charged as the excess electrons on the surface are being transferred to the electrolyte, hence resulting in an electric field directed from the TiO2 surface (positively-charged) to the electrolyte.
Upon near-UV irradiation, an electron is excited from the valance band (EVB) to the conduction band (ECB), and the subsequent separation of the photogenerated charge carriers may occur as follows. Under the influence of the electric field, it has been suggested that the majority of the holes generated are transferred to the surface while the majority of the electrons are transferred to the interior of the TiO2 particle. It was suggested that prolonged irradiation could result in the accumulation of electrons in the interior. This could cause the flattening of the band near the surface and the narrowing of the space charge layer, easing the transfer of electrons from the interior to the surface (Memming, 1990).
10
Direction of the Electric Field
TiO2 Electrolyte TiO2 Electrolyte TiO2 e- E E CB CB Fermi level of TiO2
Fermi level of the electrolyte EVB E VB + h
(a) Before contact with (b) After contact with (c) Irradiation of TiO2- electro lyte. electrolyte. electrolyte suspension.
Figure 2. 2. Illustrations of electrochemistry of TiO2 particles in contact with an electrolyte and after irradiation. (a) Fermi level near the conduction band in TiO2 prior to contact with an electrolyte. (b) Band-bending and the presence of an electric field after contact with electrolyte. (c) Transfer of the photogenerated electrons and holes upon irradiation. E VB: Valence band, ECB: Conduction band.
When the e- and h+ are generated following irradiation, they can take part in processes as illustrated in Figure 2. 3. e--h+ recombination can occur in the bulk (reaction 1a) and/or on + - the surface (reaction 1b). When the h and e are successfully transferred to the TiO2 surface, the h+ can oxidise an electron donor D (reaction 2a) while the e- can reduce an electron acceptor A+ (reaction 2b) (Litter, 1999). The redox reactions 2a and 2b are desirable as they lead to the destruction of organics and inorganic pollutants via oxidation and reduction processes respectively. These processes are believed to occur mainly on the + TiO2 surface. The presence of the species D and A on the TiO2 surface is hence imperative to the capture of h+ and e- respectively. Species D could be a surface hydroxyl group or an adsorbed organic molecule.
11
hv - e TiO2 ECB
+ EVB 1a. recombination h V
+ e- h+ A - + 2b. reduction e h D
2a. oxidation
A + D V
1b. recombination
Figure 2. 3. The possible pathways of the main charge carriers occurring in a TiO 2 particle following photoexcitation. EVB: Valence band, ECB: Conduction band.
To date, evidence found has supported the hydroxyl group or water on the TiO2 surface behaves as the electron donors D. Upon capturing the photogenerated holes, they form the highly oxidising hydroxyl radicals (OH•). OH• have been suggested to be the main species
initiating oxidation processes on the TiO2 surface due to their high oxidising potential (1.77 eV vs SHE) (Rajeshwar & Ibanez, 1998).
To summarise the reactions depicted in Figure 2. 3, the major processes in TiO2 photocatalysis can be generalised by the following set of simplified equations (Equations 2- 1 to 2-6) (Litter, 1999):
TiO2 photoexcitation leading to the generation of electrons and holes, similar to equation (2-1):
n >Eh g +- TiO2 ¾®¾ +¾ he (2-2)
12
Reaction of holes with the surface adsorbed water molecules (Equation 2-3a) and hydroxyl groups (Equation 2-3b) to form the hydroxyl radicals (OH•):
+ · + 2 ads ads +®+ HOHOHh (2-3a)
+ - · OHh ads ®+ OH ads (2-3b)
Direct oxidation of an adsorbed electron donor (Dads) by photogenerated holes (Equation 2- 4a) (Carraway et al, 1994), or by OH• attack (Equation 2-4b) (Tunesi et al, 1991; Turchi & Ollis, 1990):
+ + ads ®+ DDh ads ®® oxidised products (2-4a)
· + ads ®+ DDOH ads )( ®® oxidised products (2-4b)
+ Direct reduction of surface-adsorbed electron acceptor (A ads) by the photogenerated electrons (Equation 2-5):
+- ads ®+ AAe ads (2-5) e--h+ recombination with the release of thermal energy (Equation 2-6):
+- ®+ heathe (2-6)
The e--h+ recombination (Equation 2-6) drastically lowers the redox reactions (Equation 2-3 and 2-4). The estimated duration for these processes is compared as follows: charge carrier generation (in femto seconds), electron trapping (10 nanoseconds), hole trapping (100 picoseconds), charge carrier recombination (10-100 nanoseconds) and interfacial charge transfer (100 nanoseconds to milliseconds) (Surender et al, 1998; Martin et al, 1994). Hence, in order to improve the efficiency of the photocatalytic processes, the redox
13 reactions, which involve the interfacial charge transfer, must be made to compete more effectively with the recombination process.
For systems in which oxidation of a compound is desired, the presence of an electron scavenger is important. The most commonly used electron scavenger is oxygen. It has been • suggested that oxygen could form superoxide radicals (O2 ) upon scavenging an electron • • (Litter, 1999). O2 is suggested to be further reduced to HO2 , which could lead to the formation of OH• as summarised by the sequence of reactions in Equation (2-7) to (2-11) • (Matthews, 1984; Al-Ekabi & Serpone, 1988). The redox couple of O2/ HO2 has been reported to be 0.2 eV vs SHE at pH 0 (Mills & Le Hunte, 1997).
- + · +-· 2 ads 2 2 +«®++ HOHOHOe (2-7)
+-· 2 ®++ 2OHHeHO (2-8)
· 2 2 +® OOHHO 222 (2-9)
-· · - 222 2 ++®+ OHOOHOOH (2-10)
hn · OH 22 ¾®¾ 2OH (2-11)
- · - 22 +®+ OHOHeOH (2-12)
One major factor determining whether a pollutant could be degraded by photocatalysis is its redox potential. The reduction potential of the pollutant has to fall between the valence and conduction band positions of the semiconductor, also known as the flat band potentials, so that the pollutant can be either oxidised or reduced. Many studies have presented the flat band potentials of TiO2 at various pH values (Chen et al, 2000; Mills & Le Hunte, 1997;
Robertson, 1996; Hagfelt & Gratzel, 1995) for example, anatase TiO2 has VB and CB potentials at 3.0 eV and -0.2 eV respectively at pH 0 (Mills& Le Hunte, 1997). The position of the flat band of a semiconductor in a solution follows a Nernstian pH dependence, which decreases 59 mV per pH unit (Ward et al, 1983). Many organic pollutants have reduction potentials falling between these values. This explains why TiO2 photocatalysis has been successfully applied for the degradation of numerous organic
14 pollutants. Three exhaustive reviews by Blake (1994, 1996, & 1997) describe almost 1800 studies on organic photooxidation studies carried out before 1996.
Photocatalytic Oxidation
In the context of water purification, the main objective of the photocatalysis is to convert - the harmful pollutants into innocuous compounds, such as CO2, H2O and ions such as NO3 , 2- - 3- SO4 , Cl , PO4 . Many reviews have shown the successes of photocatalytic mineralisation of many organic compounds (Robertson, 1996; Penuela & Barcelo, 1998) and bacteria such as E.Coli (Butterfield et al, 1997). However, the disappearance of the initial compound is not sufficient to demonstrate the effectiveness of photocatalytic process. This is because the intermediate compounds formed could be even more toxic and resilient towards degradation than the parent compound. It is hence of interest not only to monitor the degradation of the parent compound but also to identify the intermediate compounds formed. Gas and liquid chromatography combined with mass spectrometry (GC-MS and LC-MS) have been used to identify intermediate compounds (Chiron et al, 1997) while the rate of organic mineralisation has been monitored by the Total Organic Carbon (TOC) analysis, or by monitoring the evolution of carbon dioxide.
In general, the oxidation of straight-chained hydrocarbon is relatively easy. However, dearomatisation of aromatic compounds has been found to be harder, longer and involves the formation of many intermediate compounds before mineralisation is achieved (Hermann et al, 1999). In some instances, total mineralisation is not achieved. This is exemplified by the case of photocatalytic degradation of s-triazine herbicides, for which the final product was mainly cyanuric acid (Pelizzetti et al, 1990). Fortunately, this product is non-toxic. However, in the work of Maillard-Dupuy et al (1994) in which the photodegradation of nitrobenzene was investigated, undesirable end products such as nitrite and ammonia ions were formed. This reinforces the need to elucidate all intermediate products formed.
15 Photocatalytic oxidation studies of inorganic ions have also been carried out. An example is the highly toxic cyanide ions (CN-), which has been shown to oxidise to the less toxic - cyanate ion (CNO ) in the presence of TiO2 by surface hydroxyl radicals or holes (Chiang et al, 2002):
- ·/ hOH + - - +22 OHCN ¾¾®¾¾ ++ 2OHCNCNO (2-13)
CNO- ion has been found to be further oxidised to nitrate ions (Pollema et al, 1992).
Photocatalytic Reduction
Photocatalytic reduction is mainly used for the removal of dissolved metal ions by reducing them to their insoluble forms. A widely investigated inorganic ion is the carcinogenic Cr(VI) anions (Khalil et al, 1998; Chenthamarakshan & Rajeshwar, 2000; Ku & Jung, 2001). Other inorganic compounds investigated in photoreduction include Hg(II) (Khalil et al, 2002), Cd(II) (Chenthamarakshan et al, 2000a) and Ag(I) (Huang et al, 1996). The reduction of organic compounds, such as benzoquinone (Richard, 1994), 4-nitrophenol (Brezova et al, 1997) and hydrazine (Chatterjee, 2000) has also been investigated.
However, many of metal ions have reduction potentials more negative than or close to the
CB of TiO2. Examples are zinc, cadmium, cobalt and nickel for which the standard reduction potentials are -0.76, -0.40, -0.28 and -0.23 eV respectively (Chang, 1994). The photocatalytic reduction of such metal ions is deemed thermodynamically unfavourable.
The photoreduction of water to hydrogen gas using TiO2 is also highly challenging since + the reduction potential of H /H2 redox couple falls near the CB of TiO2. Some work has also been done on the photocatalytic reduction of carbon dioxide in an effort to reduce the greenhouse gas (Kohno et al, 1999; Yoneyama, 1997). Other photocatalysts, such as ZnO,
CdS and CdSe, of which the CB is more negative than that of TiO2, could be better photocatalysts for the reduction of water and the abovementioned metal ions due to greater
16 thermodynamic driving force. However, their stability is questionable as they have been reported to photocorrode during the process (Herrmann et al, 1983). A more detailed review on the photocatalytic reduction of metal ions is given in Section 3A.1.
Finally, with respect to photocatalytic reduction process, the presence of a hole scavenger (electron donor) is important. Organic compounds and water are two potential reductants. The mere presence of water is insufficient but the addition of a organic hole scavenger has been found to enhance the reduction process significantly (Litter, 1999). Besides being hole scavengers, some organic compounds could form reducing intermediates upon reacting with the hydroxyl radicals which could enhance the reduction reactions (Kaise et al, 1994). This is further described in section 3B.1. It must be emphasised that oxidation and reduction kinetics are inter-related and either process can be rate-limiting.
2.1.3 Quantum Yield
A common parameter employed in photochemistry to evaluate and compare photocatalytic results obtained under different experimental conditions is the quantum yield. Quantum yield (f) in photocatalysis may be defined as the number of moles of products formed or destroyed per mole of photons introduced into the system:
reactionofrate f = (2-14) fluxphoton
The photon flux into a photochemical or photocatalytic process can be determined by a chemical actinometer. An actinometer is a solution or liquid which is capable of measuring small doses of UV light of both short and long wavelength. The resultant photon flux is reported as Einstein (moles of photon) per unit time (Bunce, 1989).
17 The three commonly used actinometers are potassium ferrioxalate, uranyl oxalate and benzophenone – benzhydrol. Potassium ferrioxalate actinometry is the most popular since it is easy to handle, suitable for the range of wavelength between 200 nm to 480 nm, can detect low radiation flux and has a stable photolysis product. Its quantum yields are relatively insensitive to wavelength, concentration, temperature and light intensity (Murov et al, 1993).
When the potassium ferrioxalate solution is exposed to light of wavelength less than about 490 nm, the ferrioxalate ions undergo decomposition to ferrous ions according to the equation:
3- hn 2- OCFe 342 ¾®¾ 42 + OCOCFe 42 + 2][3)(2])([2 CO2 (2-15)
The concentration of ferrous ions increases with irradiation time according to Equation 2- 15. Ferrous ions form highly stable and soluble magenta complex when reacted with ferrozine. This magenta complex has maximum absorbance at 562 nm (Stookey, 1970). Hence, ferrous ions concentration can be determined by UV-Vis spectrophotometry, allowing for the determination of the rate of ferrous ions formation. Using information obtained from a tabulated quantum yields for ferrous ions at various wavelengths and the rate of ferrous ions formed as determined above, the photon flux into the system can hence be calculated.
In photocatalysis, the reaction rate is usually expressed as a function of grams of catalyst. Therefore, in order to have a more meaningful interpretation of the quantum yield, the incident photons should be calculated based on the photon absorption by the photocatalyst but not the incident photon into the system. This could be a very difficult task to undertake as the photocatalytic system is very complex. The incident photons into the system might be interacting with a solution of ever changing composition owing to the formation of intermediates and products and the depletion of the initial substrate. The geometry and the
18 materials of the reactor will also affect the absorption and scattering photons (Alfano et al, 2000).
To overcome this problem, a simple expression known as the relative photonic efficiency
(ζr) is used (Serpone et al, 1996). This method assesses process efficiencies in heterogeneous photocatalytic oxidation using the initial photoconversion of phenol as the standard process and Degussa P-25 as the standard photocatalyst. The use of ζr renders the comparison of experiments conducted with different photocatalyst materials, and also at different laboratories and pilot plants possible because ζr is basically independent of some of the fundamental photocatalysis parameters, such as light intensity, reactor geometry and
TiO2 concentration for a given type of catalyst. However, it is dependent on initial substrate concentration and temperature.
2.1.4 Fundamental Parameters in Photocatalysis
Several chemical and physical parameters have been shown to affect the photocatalytic process. The main chemical parameters include oxygen, pH, initial solute concentrations and catalyst dosage while physical parameters include agitation rate, temperature and light intensity. Many works have been done to investigate the effects of such parameters on the photocatalytic kinetics (Chen & Ray, 1999; Lea & Adesina, 1998; Augugliaro et al, 1995). Matthews (1993) and Herrmann (1999) have provided a summary on the effects of these parameters. The next section will briefly discuss the influence of a few selected parameters investigated in this thesis.
Effect of Oxygen
The effects of oxygen have been widely studied in photocatalytic oxidation systems. Some studies have shown that the presence of oxygen in the system enhanced the degradation of the organic pollutants. Wang et al (1993) and Sclafani et al (1991) have found that the
19 photooxidation of methanol and propanol was nearly inhibited in the absence of oxygen while the changes of oxygen concentration in the system affected the rate of decomposition of the organic compounds. The overall mineralisation process for an organic compound in the presence of oxygen can be summarised by the following equation (Chen et al, 2000):
ykzsz ,u³EhTiO - kzn (mXOHC --++ )O ¾¾®2 ¾¾¾ bg mCO + + ++ zXOkzHOH -k zynm 2 4 2 2 22 2 s
(2-16)
Oxygen is also suggested to function as an electron acceptor. This is because it is strongly -• electrophillic (Lea & Adesina, 1998), hence forming the superoxide radicals (O2 ) upon scavenging of the electrons (Equation 2-7). This helps to prevent the recombination of electrons and holes and enables the system to maintain a favourable charge balance necessary for the photocatalytic redox process. EPR studies by Howe et al (1987) found that when the hydrated anatase TiO2 was UV illuminated in the presence of oxygen, the trapped electrons can be removed, stabilising the trapped holes. Oxygen is the most commonly used electron acceptor in photocatalytic oxidation processes as it is readily available, soluble under most conditions and is non-toxic (Chen & Ray, 1999). Gerischer & Heller (1991) and Wang et el (1992) have suggested that the scavenging of the electrons could be the rate-determining step in the photocatalytic oxidation.
The concentration of oxygen in the solution is affected by the partial pressure of oxygen in the gaseous phase in contact with the aqueous phase. Variation of oxygen partial pressure in the photocatalytic systems has been found to affect the reaction rate (Bangun & Adesina, 1998). Increasing the oxygen partial pressure was shown to increase the organic photodegradation rate until an optimum oxygen partial pressure was reached. The optimum is symptomatic of competitive inhibition between the reactants for adsorption of the same site. In the presence of excess oxygen, the semiconductor surface was suggested to become
20 highly hydroxylated to the point of inhibiting the adsorption of the other reactants (Pozzo et al, 1997).
With respect to photocatalytic reduction process, the presence of oxygen was found to retard the photoreduction rate. This is simply because in photoreduction processes, the scavenging of electrons by the substrate of interest is desired. However, oxygen competes with the substrate for electrons, reducing the photoreduction efficiency. This was shown by the studies of Ku & Jung (2001), who found that the rate of chromate photoreduction was significantly retarded in the presence of oxygen.
Effect of pH
The pH of the solution affects the TiO2 properties in two ways. Firstly, the pH alters the redox potential of the system and secondly, it changes the surface properties of TiO2.
A change in the solution pH causes a change in the conduction and valence band potentials of TiO2. This is because the flat band potential of colloidal TiO2 particles exhibit Nernstian pH dependence behaviour. A unit increase in pH results in the cathodic shift of 0.059 V of the flatband potentials (Ward et al, 1983). The cathodic shift in conduction band at higher pH improves the overpotential for the reduction process, possibly rendering the reduction of some heavy metal ions such as Cd2+ and Ni2+ thermodynamically feasible (Chen & Ray, 2001).
The surface charge of the TiO2 is also affected by the pH of the solution. This is because
TiO2 is amphoteric in nature. When TiO2 is in contact with water, its surface becomes hydroxylated. Upon changing the pH, the surface hydroxyl group could undergo protonation and deprotonation according to the following equations (Kormann et al, 1991):
+ pK1 + TiOH 2 ¾®¬- ¾ TiOH + H (2-17)
21
-+ TiOH ¾®¬- pK¾2 TiO +H (2-18)
The equilibrium constants of these reactions for Degussa P25 were determined to be pK1=2.4 and pK2=8.0 for Equation 2-17 and 2-18 respectively (Kormann et al, 1991). The above equations show that the surface of the TiO2 has a net positive charge at low pH and net negative charge at high pH. The pH at which the surface has an equivalence of positive and negatives charges is defined as the point of zero charge (pHzpc). For Degussa P25 TiO2, the pHzpc can be estimated by (pK1+pK2)/2=5.2.
Owing to the change of net surface charge with pH, the photocatalytic degradation of ionisable compounds by TiO2 is hence affected. Firstly, a change in the solution pH affects the extent of ionisation and hence the adsorption of the ionisable compounds on the TiO2 surface. It is generally agreed that photocatalytic reactions take place on the TiO2 surface. Therefore the amount of the ionisable compound adsorbed can determine the extent of the photocatalytic degradation. For example, the photocatalytic degradation of strongly adsorbed species such as benzoic acid was greatly affected by pH while the rate of photocatalytic degradation of weakly adsorbed species such as phenol was affected to a lesser extent (Al-Ekabi et al, 1988). Secondly, the intermediate products formed during the photocatalytic degradation of organic compounds may behave differently depending on the pH of the solution. If they are strongly adsorbed, they could possibly inhibit the reactions (Chan et al, 2001). Thirdly, the solution pH was found to affect the degree of aggregation of the photocatalyst particles while in suspension. The TiO2 particles were found to aggregate much faster near the pHzpc due to the decrease in electrostatic repulsion among the particles (O’Shea et al, 1999). Particle aggregation could be undesirable as it decreases the available active surface area.
To study the pH effects, the reagents used to change and maintain the pH must contain counterions that have minimal effect on the rate of degradation. This means that the reagents must not contain counterions that competitively adsorb onto the TiO2 surface and
22 also must not take part in the reaction. Abdullah et al (1990) have performed investigations on the effect of various anions on the rate of mineralisation of some selected model organic compounds. They found that perchlorate and nitrate ions have minimal effect whereas chloride, sulfate and phosphate reduce the rate by up to 70% at a low concentration of 1 mM. In the light of these findings, sodium hydroxide and perchloric acid should be used to produce basic and acidic pH respectively.
Effect of Temperature
Herrmann (1999) observed that the effect of temperature on the photocatalytic activity at the temperature range of 20°C to 80°C was not significant. This is because a photocatalyst is activated by the absorption of photon, and hence its true activation energy (Et) is nil whereas its apparent activation energy (Ea) is often very low in this temperature range. At very low temperature such as < 0°C, activities decrease and Ea becomes positive. This is attributed to the low temperature favouring adsorption, which is a spontaneous exothermic phenomenon. Hence, the intermediates and the final products formed could be strongly adsorbed, impeding the adsorption of the substrates to be degraded and hence decreasing the photocatalytic activity. At temperature greater than 80°C, the exothermic adsorption process is suppressed and the adsorption of substrates becomes difficult. Under these conditions, the activity decreases and Ea tends to negative.
Effect of Catalyst Loading
For non-photosensitized heterogeneous catalysis, the reaction rate should increase with increasing the mass of the catalyst (assuming the amount of substrate is not limited). However, in photocatalysis, above a certain value of catalyst loading, the reaction rates plateau-off or decrease. The occurrence of optimum catalyst loading has been attributed to the screening effect of excess particles (Herrmann, 1999). In addition to this, Bangun & Adesina (1998) suggested that increased agglomeration of the particles, due to greater
23 particle-particle interactions at high particle concentration, could contribute to an optimum catalyst loading. However, since agglomeration was observed to be most pronounced at pHzpc, it is suggested that the optimum catalyst loading would also be a function of pH. The determination of the optimum catalyst loading is hence important from a process economics perspective.
Effect of Initial Solute Concentration and Kinetic Modelling
In heterogeneous catalysis, the substrates have to be adsorbed on the catalyst surface sites for bond breaking or formation. The adsorption of substrates and the availability of sites are hence important parameters in photocatalytic reactions (Litter, 1999; Parmon et al, 2002). The rate of substrate conversion is proportional to the available active sites. As the reaction proceeds, the amount of substrate adsorbed on the catalysts surface will decrease until the substrate is completely converted.
Kinetic models are often formulated to describe photocatalytic reactions with respect to the initial substrate concentrations. The kinetic models for photocatalytic reactions are derived based on the classical heterogeneous catalysis model, which is the Langmuir-Hinshelwood (LH) kinetic model. This model assumes that the reaction occurs on the surface and the reaction rate (r) is proportional to the fraction of surface coverage by the substrate (q):
dC KC r q ==-= kk (2-19) dt rr 1+ KC
where kr is the reaction rate constant, K is the adsorption constant and C is the substrate concentration at any time t. Integrating the above expression yields:
C 0 )(ln =-+ KtkCCK (2-20) C 0 r
24 where C0 is the initial substrate concentration. At high C0, equation 2-20 can be simplified to a zero order rate law:
' 0 )( r ==- tkKtkCC (2-21)
At low C0, a first order rate equation can be obtained from equation 2-20:
C ln 0 == 'tkKtk (2-22) C r where k’ is the apparent rate constant. It is generally agreed that the rate constants are apparent and only serves to describe the rate of degradation but have no physical meaning (Galvez & Rodrigues, 2002). They may not be used to identify surface processes but may be used for reactor optimisation. Chen et al (2000) have shown that the photocatalytic degradation profiles during the first few minutes of the reaction followed the zero order rate law when the substrate concentration was high, but became first order at lower substrate concentration.
Attempts have been made to modify the basic LH equation in order to account for the presence of other possible adsorbed species on the catalyst surface. The competitive adsorption of solvent (Al-Ekabi & Serpone, 1988), oxygen (Terzian et al, 1991) and intermediate products (Turchi & Ollis, 1989; Chan et al, 2001) have been incorporated into the LH model. EPR studies have suggested that electrons produced by band-gap irradiation are trapped as Ti3+ species (Howe & Gratzel, 1985, 1987), and holes are trapped by surface hydroxyl group forming the oxidising Ti4-O• species (Micic et al, 1993a &1993b). A modified LH model accounting for the presence of the two different reactive sites has also been proposed to better describe the photodegradation rate (Rideh et al, 1999; Matthews, 1988).
25 A review on kinetic modelling of photocatalytic reduction reactions will be presented in section 3C.1.
2.1.5 Strategies for the Photocatalytic Enhancement: Quantum Efficiency and Visible-light Absorption
TiO2 photocatalytic processes generally suffer from low quantum efficiency due to the rapid recombination rate of the photogenerated electrons and holes. In addition, TiO2 can only be excited by near-UV irradiation, which only constitutes about 5% of the total solar spectrum. Hence, TiO2 photocatalysis would lose its economic credibility if extra energy is required from near-UV light for its photoexcitation. Hence, the ongoing aims of photocatalytic water and wastewater remediation are to improve the quantum efficiency and to shift its absorption to the visible light region so that sunlight could be employed as a more efficient irradiation source.
The first part in the next section will discuss the various techniques used to improve quantum efficiency by minimising the recombination of photogenerated electrons and holes. One way to achieve this is through the addition of hole or electron scavengers.
Another way is through noble metals deposition onto the TiO2 surface and metal ions doping to further minimise electron-hole recombination.
The discussion on the second part will focus on the techniques used to shift the photoresponse of TiO2, a wide band-gap semiconductor, to the visible region. Such techniques include metal ion implantation, dye photosensitization and coupling TiO2 with another semiconductor to form composite semiconductors. The latter technique could also act as a mean of charge separation and was suggested to prevent electron-hole recombination (Spanhel et al, 1987).
26 Other methods used in the improvement of photocatalytic efficiency such as the use of quantum-sized semiconductor in photoreduction will also be briefly discussed.
Addition of Electrons and Holes Scavenger
As discussed earlier, for the case of the photooxidation of organic compounds, the presence of oxygen was found to enhance the photooxidation rate due to its ability to scavenge the electrons and hence preventing the recombination of the photogenerated electrons and holes. Similarly, for the case of photoreduction, it was found that the presence of hole scavengers were important. Chenthamarakshan & Rajeshwar (2000) reported that the presence of formic acid, added as the hole scavenger, enhanced the reduction of Cr(VI) to Cr(III). In a separate study in the photoreduction of zinc and cadmium ions (Chenthamarakshan et al, 2001a), they also suggested that formate ion did not only serve as the hole scavenger but also as a specie that promoted the adsorption of the zinc and cadmium on the TiO2 surface. The organic additives have also been suggested to form reducing radicals upon reacting with the photogenerated holes. The radicals could then participate in the reduction reactions. Sanuki at el (1999, 2000) found that in the absence of formic acid, the photoreduction of selenate and selenite ions did not proceed. More discussions on the importance of organic additives are presented in section 3A.1 and 3B.1.
Noble Metal Deposition on TiO2
Metalisation of TiO2 by noble metal has been extensively investigated as a technique to increase the quantum efficiency of TiO2. The noble metals used for TiO2 deposition include platinum (Xi et al, 1995, Yang et al, 1997), palladium (Wang et al, 1992), silver (Sahyun & Serpone, 1997; Sclafani et al, 1991; Vamathevan et al, 2001; Tada et al, 1998, 2000) and gold (Kamat, 2001; Li & Li, 2001). These studies reported an enhancement in quantum efficiency with metal-loaded TiO2. However, studies by Hufschmidt et al (2002) showed that the efficiency of the noble-metal modified photocatalytic processes were strongly
27 dependent on the nature of the substrates. Mu et al (1989) reported that smaller oxidation rates of cyclohexane were observed when TiO2 was loaded with 0.5 to 10 wt% Pt.
Therefore, the use of metal loaded TiO2 might not always enhance the photodegradation of some substrates. For technical application, it is important to investigate whether the modified catalyst is suitable for the photodegradation of the investigated substrate.
The presence of noble metals on semiconductor alters the semiconductor surface properties. Figure 2. 4 illustrates the ability of Ag, a noble metal in contact with a semiconductor surface, to capture electrons. Upon excitation, electrons can migrate to the metal deposit where they can be trapped or captured by an oxidant (A+). The hole is then free to migrate to the surface where oxidation can occur (D→D+).
hv
Ag
e- h+ + A e- h+ D
+ A Ag D
TiO2
Figure 2. 4. Electron mediation by Ag in contact with TiO 2 surface.
28 The ability of a metal to mediate the photogenerated electrons away from the TiO2 has been explained by the presence of a Schottky barrier (Dalven, 1990a). A Schottky barrier is created at a metal-semiconductor interface. The potential barrier formed at the interface depends on the work functions of the two materials. The work function (F) is defined as the energy difference between the vacuum level Evac and the Fermi level EF. The vacuum level
Evac is defined as the energy of an electron at rest (and hence with zero kinetic energy) just outside the crystal surface and not interacting with the crystal (Dalven, 1990a). Electrons flow from a region of lower work function to one of higher work function.
The transfer of electrons could be rationalised as follows: the energy barrier for the electron transfer from a region of higher work function (for example, Ag) to one of lower work function (TiO2) is greater than that from a lower work function to one of higher work function. The work function of TiO2 (FTiO2=4.2eV, Henglein, 1997) is smaller than that of a noble metal, such as Ag (FAg=4.6eV, Henglein, 1997). Hence electrons could be transferred from TiO2 to Ag. A more detail discussions on the electron transfer mechanism across the metal-semiconductor junction will be presented in section 4B.1.
A common method for preparing noble metal deposits on TiO2 surface is by photoreduction of the metal ions from solutions to form metal deposits (Li & Li, 2001; Kudo et al, 1987; Xi et al, 1995, Yang et al, 1997; Sahyun & Serpone, 1997; Sclafani et al, 1991). An optimum metal loading usually exists for which the quantum efficiency is maximum. Beyond the optimum loading, the efficiency decreases possibly because the excess negative charge on the metal could become attractive for the holes, resulting in their recombination and generating non-productive thermal energy (Equation 2-6) (Li & Li, 2001).
Metal-loaded photocatalysts have also been used in the investigation of the photoreduction of carbon dioxide to a variety of organic compounds. Some examples of metal-loaded photocatalysts used include Rh-TiO2 (Ranjit et al, 1995), Pd-TiO2 (Wang et al, 1992) and
Pt-TiO2 (Rophael & Malati, 1987; Kudo et al, 1987). The photocatalytic reduction of nitrite
29 and nitrate ions on metal-loaded TiO2 has also been investigated (Ranjit et al, 1995; Ranjit & Viswanathan, 1997). The efficiency for such photoreduction reactions has been demonstrated to be higher compared to unmodified photocatalysts. This was attributed to the greater availability of the photogenerated electrons to be utilised in photoreduction reactions. Hence, this could suggest that photoreduction reactions, in general, could be enhanced using noble metal-modified photocatalyst.
Sensitization of TiO2: Metal Ion Implantation, Dye Sensitization, Composite Semiconductors
As mentioned previously, TiO2 is only activated by near-UV irradiation. Hence its quantum efficiency is very low when irradiated by sunlight, of which the near-UV radiation only constitutes less than 5%. In order to extend its absorption into the visible region, metal ion implantation has been performed on TiO2 powders (Yamashita et al, 2002). This is achieved by accelerating the vanadium, manganese and iron ions under the influence of an electric field and injecting them into the TiO2 matrix. The results indicate a definite shift in the absorption spectra towards the visible light regions. This method is similar to the concept of metal ion doping which also involves the incorporation of metal ions into the matrix of the TiO2 particles. However, mixed results were encountered as the dopant ions can either function as the sites for the charge mediation or recombination centre (Choi et al, 1994).
The addition of a second photoactive component to shift the irradiation wavelength into the visible region has also been utilised. One of the techniques known as dye sensitization involves the addition of dyes into the TiO2 photocatalytic system. Dye sensitizers have been used without TiO2 and they are found to be effective in the photolysis of some pesticides
(Tsao & Eto, 1994). The reaction pathway of the combined role of the dye/TiO2 photosensitized reaction has been described by Kamat (1989) and is presented in Figure 2. 5. In order to achieve sensitization, the dye must possess oxidative energy level more
30 negative than the conduction band of TiO2 when excited by visible light. The excited electrons from the dye can then be transferred to the TiO2 conduction band and subsequently used to reduce an electron acceptor. The same study also found that only 10% of the photoexcited electrons could be used for reduction while the others recombine with the dye itself. This system could be used in the oxidative degradation of the dye itself, provided that the photocatalytic decomposition rate of the dye is faster than that of the regeneration rate (Zhang et al, 1997).
+ Dye* A
e- hv CB A
Dye VB
TiO2
Figure 2. 5. Electron transfer during the photoexcitation of a dye in the presence of TiO 2 and an electron acceptor (A+).
Another approach taken to sensitize TiO2 has been to couple the wider band gap semiconductor, such as TiO2 or ZnO (both Ebg=3.2 eV), with a semicondiuctor of a smaller bangap of which the conduction band is higher than that of the wider band gap semiconductor, such as CdS (Ebg=2.5 eV). Figure 2. 6 shows such a CdS-TiO2 coupled system which has been investigated by Spanhel et al (1987). As CdS has a narrower band gap, a visible light photon (l<497 nm) is sufficient to excite the electrons in CdS. Due to the relative positions of the TiO2 and CdS conduction bands, the electrons from CdS could be transferred to that of attached TiO2, while the holes remain in the CdS particles. Although this charge separation mechanism resulted in enhanced charge transfer, the CdS particles were unstable and underwent photocorrosion.
31 In the same study by Spanhel et al (1987), it was found that charge-carrier separation was less pronounced for CdS-ZnO couples. This was because the conduction bands of these particles are at similar levels. Electron injection into ZnO only occurred in the case of Q- size (quantum-size) CdS as quantisation of semiconductor particles resulted in a negative shift in the conduction (Q-size semiconductors will be discussed in greater details in the next section). The effect of coupling Q-size composite semiconductors was further demonstrated by Vogel et al (1994), who reported up to 80% increase in photocurrent quantum yield in their investigations.
V vs SHE -2 CdS + -1 e- A e- CB 0 CB hv A 1 + h+ VB D 2 VB 3 D
TiO2
Figure 2. 6. Charge transfer in a CdS-TiO2 coupled system.
Another type of composite semiconductor system is the core-and-shell configuration as shown in Figure 2. 7. This configuration has a semiconductor at the core, of which the conduction band is lower than that of the external semiconductor surrounding the core. Upon excitation, electrons from the shell are injected into the core, providing efficient charge separation. The holes could then be used for oxidation reactions. This configuration is unfavourable for reduction reactions as the electrons are trapped by the core and hence not readily available for reduction. Bedja et al (1995) have synthesized TiO2-capped SnO2 composite semiconductors and observed a 2-3 times enhancement of photoactivity with this
32 coupled semiconductor system. However, the stability of the composite semiconductor was not mentioned.
-2
-1
- TiO2 0 CB e
1 CB hv V vs SHE vs V 2 VB 3 SnO2 + VB + h D D
Figure 2. 7. Charge transfer in a SnO2-TiO2 shell-and-core semiconductor system.
Quantum-size (Q-size) Effect on Photocatalysis
Quantum-size effects occur when the size of the semiconductor particles become smaller than the Bohr radius of the first excitation state (Hagfeldt et al, 1995). One of the most distinct effect of reducing the size of a particle to that of quantum size is the increase in band gap energy as can be seen in Figure 2. 1. This is because as the particle size approaches that of nano-scale, the photogenerated electron and hole cannot fit into such a particle unless they assume a state of higher kinetic energy (Weller & Eychmuller, 1995).
There seems to be many discrepancies in the reported quantum size of TiO2 particle, possibly due to the different crystalline phases of the particle. However, it can be taken that crystalline TiO2 starts to show Q-size effect at about 10 nm (100 Å) (Beydoun et al, 1999). Q-size effects could be manifested by the absorption shift to shorter wavelength, also known as blue-shift. The levels of valence band are moderately shifted to lower energies, while those of the conduction band are strongly shifted to higher energies (Henglein, 1997).
33 The resultant increase of band-gap energy owing to the Q-size effect has prompted many studies on the photocatalytic activity by Q-size semiconductors. This is due to the anticipated increase in thermodynamic driving forces for interfacial charge transfer. Znaidi et al (2001) have used Q-size TiO2 particles in the degradation of phenol but the result indicated no enhancement compared to the commercially available Degussa P25 particles. However, Q-size semiconductors were shown to have enhanced activities in the photoreduction of carbon dioxide (Yoneyama, 1997). These findings present possibilities of using Q-size semiconductor in the phoreduction of heavy metal ions to their elemental forms, for example Cd2+ and Fe2+, which redox potentials are slightly less negative than that of bulk TiO2 particle, rendering their photoreduction thermodynamically unfavourable.
On the other hand, charge-carrier dynamics investigations for Q-size and P25 TiO2 particles conducted by Martin et al (1994) have suggested that indirect band gap semiconductors
(such as TiO2) in the Q-size region could be less photoefficient owing to faster rates of charge-carrier recombination, off-setting the increased thermodynamic driving force for interfacial charge transfer.
34 2.2 Selenium
2.2.1 Introduction
Selenium (Se) is a rare element that is almost never found in its native state. It is associated with a few rare minerals such as eucairite (CuAgSe), crooksite (CuThSe) and clausthalite (PbSe), and coexists with sulfide minerals. It was officially isolated and characterised in 1817 by a Swedish chemist, Jons Jakob Berzelius. It was named after the word, Selene, which means the moon in Greek (Reilly, 1996). It is about 66th in abundance among the 88 elements that naturally occur in the earth crust (Greenwood & Earnshaw, 1997).
The following section will present a brief review on the physical and chemical properties of selenium and selenium anions which are relevant to the current investigation. Selenium is a semiconductor which has found many industrial applications. The use of selenium as a semiconductor and other applications will also be discussed. Owing to its industrial application and natural occurrence, elevated amount of selenium compounds has been found in the ecosystems in recent decades. Excess selenium in water is detrimental to the health of animals, especially water fowls, and humans. Its distribution in the environment and its toxicity will be discussed. Finally, the existing techniques for Se compounds removal will be compared and discussed.
2.2.2 Properties of Selenium and Selenium Anions
Physical Properties
Selenium has an atomic weight of 78.96 and its atomic number is 34. Elemental selenium boils at 684°C. Its outer electronic configuration is 3d104s24p4, and its three inner shells are completely full (Chang, 1994). It lies between sulfur and tellurium in Group 6A and hence
35 its compounds have chemical properties intermediate between those of sulfur and tellurium, bearing closer resemblance to sulfur. This is the reason for the association of selenide with sulfide minerals. There are no ores from which Se can be directly mined as a primary product. It is commonly found in conjunction with sulfide minerals of copper (berzelianite, klockmannite, umangite) and lead (clausthalite) (Mineral Galleries, 2003).
Like sulfur, several allotropic forms of Se are known to exist. The three red monoclinic (a, b and g) forms consist of Se8 rings and differ only in the intermolecular packing of the rings in the crystals. Other ring sizes have also been synthesised in the form of cyclo-Se6, cyclo-
Se7 and heterocyclic analogues cyclo-Se5S and cyclo-Se5S2 red allotropes. The red crystalline selenium, like sulfur, is soluble in carbon disulfide (Reilly, 1996). The grey “metallic” Se, which exists as hexagonal crystalline featuring helical polymeric chain, is the most thermodynamically stable and is the only allotropic form that conducts electricity in the dark (Greenwood & Earnshaw, 1997). Somewhat related to the metallic Se is the amorphous Se (a-Se), analogous to the yellow sulfur. It has a deformed chain structure but does not conduct electricity in the dark. However, a-Se is a well-known photoconductor which led to the invention of xerography. More applications of a-Se will be discussed in Section 2.2.4. It has been reported that heating amorphous Se to 180°C transforms it into grey metallic Se. Finally, the commercially available black Se comprises of extremely complex and irregular structure of large polymeric rings having up to 1000 atoms per ring (Pejova & Grozdanov, 2001).
Chemical Properties
The transitional position of selenium between non-metallic sulfur and the metallic tellurium renders it characteristics of both metals and non-metals (Hoffman & King, 1998).
Chemically, Se bears a closer resemblance to sulfur. The structure of the Se8 monoclinic form (Figure 2. 8) bears a close resemblance to that of the S8 orthorhombic sulfur.
36
233.5 nm (mean)
105.7° (mean)
Figure 2. 8. Structure of Se8 monoclonic structure (Langner et al, 1996).
The chemistry of selenium is complicated because it exists in four oxidation states (+6, +4, 0, -2) and a variety of compounds (oxides/hydroxide, sulfides, organoselenium compounds and selenides). This review will focus on the redox chemistry of selenate (+6), selenite (+4) and their protonated ions since they are the most common selenium species in water and wastewater. The chemistry of elemental selenium and hydrogen selenide (H2Se) will also be briefly discussed.
Selenate ion (Se(VI)) is isostructural with sulfate. Both selenate and sulfate ions are tetrahedral in structure, with the central Se or S atom tetrahedrally bonded to four oxygen atoms. The two species are of comparable size with the mean tetrahedral bond lengths of 1.49Å and 1.65Å for S-O and Se-O respectively. The redox chemistry of selenate in the presence of H+ can be expressed by (Seby et al, 2001):
2- -+ - 4 23 +«++ 23 OHHSeOeHSeO (2-23)
2- - E -= pH + SeO4 HSeO3 ][][log295.00886.0075.1 (2-23a)
Selenite ions (Se(IV)) in its protonated state can be reduced according to the reaction:
- -+ 0 HSeO3 45 +«++ 3 2OHSeeH (2-24)
37
- E -= pH + 148.00739.0778.0 log[HSeO3 ] (2-24a)
Sanuki et al (1999) have reported the one-step reduction of Se(VI) to elemental selenium using photocatalysis. This could be represented by the following equation:
2- -+ 0 4 68 +«++ 4 2OHSeeHSeO (2-25)
2- E -= pH + 0099.00788.0876.0 g[SeO4 ]lo (2-25a)
At low pH, elemental Se could be further reduced to hydrogen selenide (H2Se):
0 -+ 22 «++ 2SeHeHSe (2-26)
E --= pH - log0295.00591.0369.0 r (2-26a) 2SeH
Hydrogen selenide normally exists in gaseous form. It is extremely toxic, flammable and has a strong oxidising ability, and should be handled with extreme care.
2.2.3 Applications
Industrial Applications
Bearing both non-metallic and metallic properties, Se is a well-known semiconductor. The popularity of selenium (Se) as a semiconductor was heralded even before the use of silicon and germanium (Greenwood & Earnshaw, 1997). a-Se exhibits photoconducting properties: its electrical conductivity is low in the dark, but increases several hundredfold upon
38 exposure to light (Dalven, 1990b). This photoconducting property of a-Se has seen its patented use in xerography by Xerox (Kasap et al, 1991) and in TV camera tubes (Neuhauser, 1987). More recently, a-Se was regarded as the most suitable photoconducting material in X-ray image detector (Rowlands & Kasap, 1997, Kasap et al, 2000). The ability of a-Se to function as semiconductor was attributed to the presence of shallow and deep traps in its structure. Shallow traps reduce drift mobility while deep traps prevent carrier recombination (Rowlands & Kasap, 1997). In addition, Se possesses what is known as asymmetrical conductivity, which allows it to conduct an electrical current more easily in one direction. This has seen its application in rectifiers (devices used for the conversion of an alternating current into direct current)
Other than its semiconducting properties, Se is also used in metallurgy. Addition of small amounts of selenium to alloys improves the machinability of wrought products and steel castings. Selenium enhances the corrosion resistance of chromium, platinum and magnesium alloys. Selenium is used as a red-orange tint for glass manufacturing (Oldfield, 1995).
Selenium as Essential Nutrient
Selenium in minute quantity and in its organic form is an essential nutrient in the human diet. A dietary intake of about 50 mg/day for women and 70 mg/day for men is required to meet the daily nutritional need (ATSDR, 2003). Se binds with tRNA to form seleno-tRNA which is required in the normal functioning of enzymatic activities. It also forms a selenoprotein called selenocysteine (Burk, 1994).
A small amount of selenium is especially important in the dietary requirement in livestock. Some disease associated with selenium deficiency for animals include white muscle (skeletal and heart muscle) disease, pancreatic degeneration and liver necrosis (Oldfield, 1995). Se is also an important nutrient for the growth of plant. The lack of Se in the
39 environment prompted the Finish government to deliberately add Se in artificial fertilizers (Wang et al, 1995).
Selenium might also play a protective role against the toxic effects of heavy metals. Feroci et al. (1997) have examined the interaction between selenium ions and heavy metal ions, namely Cu2+ and Pb2+. He found out that the Se(VI) ion does not interact with either ions; 2+ 4 while selenite forms a 1:1 soluble complex with Pb (Kf = 4.9x10 ) and a poorly soluble 2+ -9 salt with Cu (Kps = 3.8x10 ). This information can represent a defence mechanism of living organism against toxic substances.
2.2.4 Selenium Distribution in the Ecosystem
Selenium Speciation in Natural Water
Attempts have been made to investigate the speciation of Se in natural waterways. However, generalisation in Se speciation is difficult due to the different conditions experienced in different environments. Nevertheless, three major factors may affect the redistribution and remobilisation of Se in the environment: (1) changes in the state of redox in the aquatic environment, (2) bioturbation and (3) bioaccumulation (Peter et al, 1999).
Zhang and Moore (1997) have also found that Se concentration in the natural waterways is mostly affected by microbial or chemical reduction and the formation of organic Se. Microbial and chemical reduction are the results of changing pH and dissolved oxygen in the water system. Also, various forms of sediments in the water system can act as adsorption surfaces and may reduce Se concentration in water. Wang et al (1995) have suggested that the sediments act as a sink for Se. Se can be immobilised when sedimentation occurs in the bottom of the lake. However, this form of removal has been found to be not as effective as compared to the microbial or chemical reduction. Undercurrent and bioturbation might cause remobilisation of Se.
40 In natural water, Se can exist as free ionic species (Se(VI) and Se(IV)), inorganic 0 0 complexes (CaSeO 4, MgSeO 3), and dissolved organic carbon complexes. Changes in selenium’s oxidation state and the differences in chemical properties of various forms of selenium species strongly affect the movement and toxicity of selenium. In acidic solutions, 2- 2- the state of selenium changes according to the sequence SeO4 , HSeO3 , and Se to H2Se, by the decrease of pH (Pourbaix, 1966). However, extremely low pH would be required for the reduction of selenium ions to elemental Se and further to H2Se. This is highly uncommon in the natural environment.
The solubility of selenium is of great importance as it determines its mobility in natural waters. Pollutants with higher mobility are considered to be more hazardous due to their greater ease to be circulated around the ecosystem. Masscheleyn et al. (1990) have determined the influence of Eh and pH on the mobility of selenium species. The higher oxidation state selenium species such as Se(VI) and Se(IV) are more mobile compared to selenium and selenide. The adsorption properties of selenium ions are also important factors controlling its mobility in the environment. It is reported that Se(IV) adsorption on soil particulates decreases with increasing pH in the range 4-9 and that Se(VI) adsorption is minimal under most pH conditions (Masscheleyn et al. 1990). Hence, Se(VI) could be considered to be the more hazardous due to its higher mobility.
Figure 2. 9 shows the average composition of Se species in three different samples of soil water and groundwater. The results show that the dominating Se species in the total dissolved Se concentrations was Se(VI), which was about 58 and 52% in soil water and ground water respectively. As expected, Se(IV) in soil water (5%) is lower than that of groundwater (16%). This is because Se(IV) is more strongly adsorbed onto particulates and hence its composition is lower in soil water. The Se-organic and Se-metal complex compounds have similar compositions in groundwater and soil water (Reddy, 1998).
41 60
50
40
30
20
Composition in % 10
0 selenateseleniteorganicothers
Soil water Ground water Se species
Figure 2. 9. Average composition of Se species in 3 different samples of groundwater and soil water (Reddy, 1998)
2.2.5 Toxicity of Selenium Compound
Selenium and its ions are toxic at high concentration. Bioaccumulation of selenium in living organisms at the top of the food chain has created serious health hazards (Wang et al, 1995). More stringent regulations have been introduced to control the discharge of Se(VI) and Se(IV) into the waterways. As an example, the present maximum Se contaminant level for US and Australia drinking water standard is 10 mg/L (Jones & French, 1999). In Japan, the effluent standard for Se is 100 mg/L (Sanuki, et al., 1999, 2000).
Several occasions of selenium toxicity, known as selenosis, were reported in North America in 1930s, Latin America in early 1970s and the most endemic intoxication in China in the early 1980s. Body tissues most affected include hair, nails, teeth, skin and the nervous system. In most cases, hair became dry, nails became brittle and the teeth became mottled and eroded easily. In extremities, permanent damage to the nervous system was observed (Yang et al, 1983).
Selenium can build up in the food chain when total selenium concentrations are as low as 0.5-3.0 mg/l. Lemly (1996) considered that dissolved selenium concentration of 2 mg/l or
42 greater are highly “hazardous to the health and long survival of fish and wildlife”. In addition, dissolved organic selenium has been shown to bioconcentrate at much higher levels than either the anionic Se(VI) or Se(IV) (Besser et al. 1989, 1993). Animals at the bottom of the food chain such as Marphysa Sanguinea and Spisula Trigonella in Lake Macquarie NSW Australia accumulated significant amounts of selenium when exposed to highly Se-contaminated water. The selenium in such animals and mollusc was found to be associated with the protein fraction as selenomethionine and an unidentified compound instead of inorganic Se(IV) and Se(VI) (Peter et al, 1999). Se is also rapidly accumulated in phytoplankton (Turner and Rudd, 1983), and then further enriched in herbivorous fish and finally in predatory fish (Wang et al, 1995).
A study by Lindstrom (1983) suggested that Se is a nutrient for algae-bloom in fresh water. The study demonstrated that the growth rate of one specie of green algae was rapid even at low Se(IV) level (5ng /l) and the highest growth rate was observed at a range of 20-50 ng/l of Se(IV). The algae depleted oxygen and nutrients and produced toxic chemical.
2.2.6 Selenium Sources
Selenium exists naturally in seleniferous soil and rocks. The leaching of these soils and rocks accounts for the Se species that occur naturally in waterways (Zhang and Sparks, 1990). However, human activities such as agriculture, mining and industries have increased the concentration of various Se species in the waterways to levels that might be harmful to the living organisms (Zhang & Moore, 1996).
Agricultural Se output arises from livestock feed containing selenium as a nutrient. In addition, soil selenium deficiency is sometimes corrected by selenium addition. For example, Finland is the first country in the world to add Se(VI) (6-16 mg Se per kg fertiliser, producing 20 tons Se per year) to artificial fertilizer nationwide in 1985 (Wang et al, 1995). Another minor use of selenium compounds is as an insecticide in the horticultural
43 industry. Potassium ammonium sulfoselenide has insecticidal properties and was discouraged by US authorities due to its likelihood of food contamination (Reily, 1996).
In mining, oxidative leaching of sulfide ores results in the release of Se(IV) or Se(VI) (Dutrizac, 1981). The leachate is usually left in the leachate pond, increasing the risk of Se(IV) and Se(VI) escape into the environment.
Industrial output includes effluent from glass manufacturing and metallurgy. Needless to say, Xerox produces a significant amount of Se waste. In Lake Macquarie of the state of New South Wales Australia, the presence of a smelter and a power station results in a high Se concentration of about 69 times in concentration to that of unpolluted water. The smelter processed sulfide ore and other lead-zinc mineral in which selenide is present while the power plant uses Se-bearing coal as the fuel for the boiler (Peters et al, 1999). In the Nile Delta lagoon, Egypt, contamination of the lagoon water was attributed to both industrial activities in the area surrounding the lagoon and agricultural drainage from upstream. Predominant Se species were Se(IV) (38-47%) and organic Se (37-40%). About an estimated of 1500 kg per year of total Se reaches the lagoon (Abdel-Moati, 1998).
2.2.7 Removal of Selenium Compounds
The increased Se concentrations in natural waterways to potentially toxic levels has justified the need for effective Se removal. The three main methods investigated for Se species removal of involve physical, biological and chemical means. When selenium species become insoluble, they can be removed from the water by conventional physical methods such as sedimentation and filtration. For the soluble Se species, the charge density will be an important parameter to determine its adsorption extent onto the adsorbate. In order to render the soluble and highly mobile Se anion species insoluble, reduction could be carried out to achieve immobilization. This could be achieved by biological and chemical means.
44
Table 2. 1 summarises the various methods for the removal of Se compound from wastewater. These removal methods are then compared and discussed in the next paragraphs.
Table 2. 1. Summary of Technologies for Selenium Compounds Removal in Water and Wastewater. Processes Description References
Physical treatment
Coagulation Optimal Se(IV) removal was 80% when ferric sulfate Sorg & Logsdon, was used as the coagulants in the pH 6-7 range. 1978. However, this technique is ineffective in the removal of Se(VI).
Adsorption Se(IV) adsorption by ferrohydrite and alumina is Merrill et al, effective. Se(VI) adsorption by the above methods is 1996, Trussell et ineffective. Se(IV) and Se(VI) adsorption by activated al, 1991, Rozelle, carbon is ineffective. 1987
Ion Exchange Se(IV) is removed more effectively than Se(VI) when a Maneval et al, strong base anion ion exchange resin was used. 1985. However, other anion species, such as sulfate and sulfite which always coexist with Se(VI) and Se(IV), may load in preference to the Se ions. Efforts to identify selective resins were unsuccessful.
Membrane Reverse osmosis and nanofiltration hold promise for Se Pontius, 1995, Separation ion removal. However, they might be more suited as a Kapoor et al, final polishing steps as pretreatment is required to 1995, Kharaka et remove suspended solids to prevent fouling. They are al, 1996. also more suited to treat drinking water instead of acid mine water.
Chemical treatment
Ferrous A patented process discovered by Murphy (1988) could Murphy, 1988, Hydroxide reduce both Se(IV) and, to certain extent, Se(VI) ions Nishimura & (green rust to elemental Se by using ferrous hydroxide as the Umetsu, 2000, method) reducing agent at pH 9. The process forms precipitate Lien et al, 1990, containing selenium, possibly containing unreacted Zingaro et al, Se(VI) and Se(IV), and magnetite and maghemite. The 1997, Refait et
45 reduction is very sensitive to pH, with an optimum at al, 2000 pH 9. The rate decreases drastically below pH 8 and above pH 10. Nitrate, dissolved oxygen and bicarbonate interfered with the process. When this technology was applied to mine water, about 25 g/L ferrous sulfate was required to attain Se effluent concentration below 0.1 mg/L, generating a large quantity of iron bearing sludge and costing about US$10/1000 gallon of water treated (US$2.27 per m3).
Zero-valent Se(VI) reduction was effective when iron was used as McGrew et al, Iron the reductant, catalysed in the presence of copper. 1996, Nishimura Elemental Se, iron selenide and hydrogen selenide are et al, 2000, Qiu et formed in the process. The process is not influenced by al, 2000. the presence of sulfate. When an optimum condition of 10 mg/L copper and 10 g/L iron was used to treat mine leach effluent bearing 0.257 mg/L Se, a residence time of 20 minutes was required to lower the Se content below the detection limit (< 5 mg/L).
Se-sol The reduction kinetics of Se(IV) to Se-sol have been Shaker, 1996. formation investigated by using L-ascorbic acid in buffered aqueous media of varying pH (2.5-3.5) and ionic strength. Se-sol manifests as a pink substance. The reaction rate was found to increase with acidity but decreased with ionic strength. The reaction was suggested to proceed via the formation of an intermediate dimer species [Se(IV)2].
Photocatalysis The use of semiconductor photocatalysis in the Sanuki et al, reduction of Se ions on TiO2 was investigated. 1999, 2000. Complete removal of Se ions was suggested via the reduction of the Se ions to elemental selenium, indicated by the colour change of the initial white TiO2 suspension to a orange/red colour. The photoreduction rate of Se(IV) was faster than that of Se(VI), indicating the importance of the Se ion adsorption onto the photocatalyst surfaces. Upon exhaustion of the Se ions, the elemental Se was further reduced to H2Se, a toxic gas, which was removed as copper (II) selenide in a copper (II) scrubber.
46 Biological reduction
Pseudomonous The use of P. stutzeri bacteria shows great potential in Adams et al, Stutzeri reducing both Se(IV) and Se(VI). Two different 1996. Bacteria approaches can be adopted: the growth of the actual bacteria as biofilm or the extraction of enzymes from the bacteria. Continuous selenium removal from about 600 to <10 mg/L has been demonstrated for up to 4 months period.
Combined treatment
Bicarbonate This method involves the use of ferrous hydroxide but U.S. Department and oxygen include pretreatment steps. Dissolved bicarbonate and of the Interior, removal oxygen are found to inhibit ferrous hydroxide-Se(VI) 1991. followed by reduction and hence are treated prior to removal. Ferrous Bicarbonate can be removed by lime precipitation or Hydroxide acidification followed by aeration. Oxygen can be removed by introducing sulfur dioxide. The estimated cost for combined pretreatment and selenium removal is US$150/acre-foot (US$0.12 per m3).
Ferrous This technique combines ferrous oxide treatment, Lykins and Clark, Coagulation- followed by coagulation and filtration. This is listed by 1994. Filtration US-EPA as one of the Best Available Technologies for treating selenium-bearing waters but its success in treating mine waters remains doubtful due to the interference of sulfate and sulfite ions.
Comparison of Se Anion Removal Techniques
The removal of Se(VI) and Se(IV) by physical methods has claimed little success. Most of these techniques demonstrated reasonable removal of Se(IV) ions but were ineffective against Se(VI) due to its minimal adsorption.
Considerable attention has been given to the biological reduction of Se ions to elemental Se. Biological removal has claimed more success by using bacteria isolated from selenium rich soil. Reported problems have included difficulty in promoting growth of the bacteria
47 due to the inhibitory effect of the elemental Se formed (Maiers et al, 1988, Fujita et al, 1997). Ike et al (2000) have found that there was a tendency for Se(IV) reduction to precede more rapidly and easily then the Se(VI) reduction. In the work of Guo et al (1999), they observed that the rate of Se(VI) reduction is greatly favoured in the presence of organics and in low oxygen content.
Both physical and biological removal techniques seem to claim little success in Se ion removal. Chemical treatment such as the ferrous hydroxide method developed by Murphy (1988) seems to be the most promising since it is the Best Available Technology enlisted by the US- EPA due to its effectiveness. However, this technology is not sustainable as it generates a large amount of iron-selenium-bearing sludge which requires secondary disposal. The chemical cost is also very high.
The photocatalytic reduction of Se(VI) and Se(IV) to elemental Se seems highly promising
(Sanuki et al, 1999, 2000). The formation of elemental Se on the TiO2 particles shows that immobilization of the selenium is achieved by removing the highly mobile Se ions from water by reducing them into the zero valent form (Se0). As demonstrated in the same study, the TiO2 could be regenerated by further photoreducing the Se formed to H2Se, which was then reacted with copper(II) ions to form copper selenide precipitates. Copper selenide is a useful product and has been used as a semiconductor (Garcia et al, 1999).
Comparing photocatalysis with the other removal techniques, this novel approach can remove both Se(VI) and Se(IV) ions from water and does not generate waste products that require expensive secondary treatment and disposal. Upon further reduction of the Se anions to H2Se gas, useful products such as copper selenide semiconductor could be produced. This step could also be regarded as the catalyst regeneration steps. Therefore, even though this technology has some drawbacks such as low quantum efficiency and an additional cost due to the need for irradiation, its many advantages warrant the further investigation of using TiO2 photocatalysis in the removal of Se ions from water.
48 Chapter 3. Photocatalytic Reduction of Selenate and Selenite
3.1 Introduction
Many research groups have demonstrated the viability of heterogenous photocatalysis to treat contaminated water containing inorganic ions by reducing them to their insoluble states. The most widely investigated inorganic ions include the chromate (Alam & Montalvo, 1998; Ku & Jung, 2001), nitrate ions (Ranjit et al, 1995; Ranjit &Viswanathan, 1997; Kudo et al, 1987) and other metal ions namely Hg2+, Ag+, Pt+, Pb2+, Cd2+, Ni2+, Cu2+ and Fe3+ (Prairie et al, 1993). The aims for their reduction were to remove them from contaminated waters and also for their recovery due to economic reasons. The photodeposition of noble metal by the reduction of the noble metal ions has also been investigated with an aim to modify the photocatalyst’s surface properties to enhance electron-hole separation. More details on noble-metal modified photocatalyst will be presented in Section 4B.1.
As discussed in Chapter 2, Se ions (selenate, Se(VI) and selenite, Se(IV)) have been found to be the most common form of Se compound pollutants. This chapter will focus on the fundamental and mechanistic studies of Se ion removal by photocatalytic reduction. These studies will place more emphasis on Se(VI) photoreduction since it has been found that the removal Se(VI) is more difficult than that of Se(IV).
This chapter will be subdivided into 3 different sections to elucidate the effects of various parameters on the Se ions photoreduction process and to report the kinetic modelling of Se(VI) photoreduction. The first section, Section 3A, reports on the preliminary investigation of Se ion photoreduction at various pH and formic acid concentrations. Section 3B investigates the use of different organic additives and their effects on the rate of
49 Se ions reduction. Section 3C models the Se(VI) ion photoreduction rates in the presence of formic acid.
50 3A. Investigations of Se Ions Photoreduction at Various Parameters
3A.1 Introduction
The standard reduction potential of the inorganic ions is an important factor in determining whether the ions could be reduced by TiO2. Depending on the pH of the system, and ruling out any effects due to quantum size and additives, it has been reported that only ions with half reactions more positive than the conduction band of the TiO2 (using unmodified- Degussa P-25 as an example) can be reduced (Prairie et al, 1993). Half reactions of redox potentials very close to the conduction band, such as those of Cd2+ and Pb2+, show very slow reduction rates, which could be ascribed to either being thermodynamically unfeasible or kinetically sluggish (Chenthamarakshan et al, 2000a). Many research investigating photocatalytic reductions have found that the major parameters affecting the process are the presence of an organic hole scavenger, O2 and pH. The following paragraphs will discuss each of these effects separately.
In photocatalytic systems involving the oxidation of organic pollutants, oxygen is normally used as the natural and cheap source oxidant which can effectively scavenge electrons, hence preventing electron-hole recombination. In systems for which reduction reactions are emphasised, it is desired that the photogenerated electrons be captured by the species to be reduced rather than dissolved oxygen. Hence, the presence of oxygen will inevitably compete with such species, lowering the efficiency of the desired photoreduction reaction. To minimise this competitive effect, the oxygen can be removed by saturating the system with inert gases such as argon and nitrogen. Khalil et al (2002) demonstrated that the photoreduction of Hg2+ is more efficient in the absence of oxygen. Chen & Ray (2001) observed that the photocatalytic reduction of Ag(I) and Hg(II) ions was significantly inhibited in the presence of oxygen. Alam & Montalvo (1998) reported that in the presence
51 of dissolved oxygen in the system, the majority of the photogenerated electrons reduced the dissolved oxygen instead of reducing the chromate ions as desired.
Photocatalysis is a synergistic process involving the redox reactions initiated by the simultaneously photogenerated holes and electrons. Hence, to achieve photoreduction, the presence of a hole scavenger is essential. The presence of hole scavengers enhances the electron-hole separation, consequently enhancing the quantum efficiency. The importance of organic hole scavengers present in a photocatalytic reduction system has been affirmed in many studies (Prairie et al, 1993; Alam & Montalvo, 1998; Khalil et al, 2002; Ku & Jung, 2001). Chenthamarakshan & Rajasehwar (2000) focused their study on the role of the hole reaction pathway in the photocatalytic reduction of chromate ions. They compared the effect of using ammonium ions to formate ions as the hole scavenger and found the former gives better chromate photoreduction rate. This was explained as, besides being the hole scavenger, the ammonium ions produced more protons upon photooxidation, hence stabilising the pH of the system and maintaining favourable solute adsorption. However, it has also been reported that formate ions and some simple alcohols are capable of forming reducing radicals upon reaction with hydroxyl radicals. These reducing radicals were suggested to enhance the photoreduction process, for example during the photoreduction of zinc and cadmium (Chenthamarakshan et al, 2000a).
As discussed in the literature review, it is generally agreed that photocatalytic reactions take place on the surface of the photocatalyst. The adsorption of the solute onto the photocatalyst surface is hence important in determining the extent of reduction. Besides having an influence on the flatband potential of the TiO2 photocatalyst, the pH of the system has a direct effect on the adsorption of the species since pH determines the surface charge of TiO2. The pH of zero point charge (pHzpc) for anatase TiO2 has been reported to be between 4.7-5.5 (Sanuki et al, 1999). Degussa P-25 TiO2 has been reported to have pHzpc at about 5.6 (Terzian et al, 1991). The photocatalytic reduction of chromate was found to be more effective at pH less than 6. This was attributed to the affinity of the chromate ions Cr(VI) to the TiO2 surface at this pH (Ku & Jung, 2001). The pH could also influence the chemistry of the solutes in the system. Gimenez et al (1996) investigated the
52 photocatalytic reduction of chromate and observed good reduction rates at pH lower than 4.
At pH greater than 4, the TiO2 was suggested to be deactivated due to the formation of chromium (III) hydroxide, preventing the adsorption of chromate adsorption onto the surface.
The research on photocatalytic reduction of Se(VI) or Se(IV) species by titanium dioxide is limited. Sanuki et al (1999, 2000) have investigated such photocatalytic reduction and found that the adsorption of Se(VI) or Se(IV) species onto titanium dioxide and the presence of suitable hole scavengers played a significant role in the reduction process. The adsorption of Se(IV) onto the TiO2 surface and the removal of Se(IV) upon UV irradiation were both found to be higher than that of Se(VI). The presence of sulfate ions in the solution was found to depress Se(VI) or Se(IV) reduction rate, indicating competitive adsorption between the sulfate ions and the Se ions on the TiO2 surface. Kikuchi & Sakamoto (2000) have proposed that when all Se(VI) or Se(IV) ions were reduced to elemental Se, no more species in the solution could capture the electrons from the TiO2 surface. They suggested that the accumulation of electrons in the TiO2 particles shifted the conduction band more negative and further reduced the elemental Se to hydrogen selenide gas. It has been demonstrated that the generated H2Se gas from reactions could be removed as copper selenide precipitate when the H2Se gas is scrubbed in a Cu(II) solution. Following the Cu(II) scrubber, an alkaline scrubber, for example one containing sodium hydroxide, could be introduced to provide an alkaline condition so as to prevent the protonation of the selenide ions, which could lead to the formation of H2Se gas (Sanuki et al, 1999 & 2000)
The studies presented in the next section reports the effects of various experimental parameters, such as pH, initial concentrations of holes scavengers (formic acid) on the photocatalytic reduction of Se(VI) and Se(IV) ions. These studies focus on the synergistic and competitive effects of formate and Se ions on the TiO2 surface have on the photocatalytic reduction of the Se ions.
53 Experimental results using Se(VI) ions as the Se precursors will be discussed first followed by that of Se(IV) ions. A comparison of the photoreduction results of Se(VI) and Se(IV) ions is also included.
3A.2 Equipment and Procedure
Catalyst and Reagent
Degussa P25 titanium dioxide was selected as the photocatalyst. Its polycrystalline structure is composed of approximately 70% anatase and 30% rutile, with specific surface area of about 58.6 m2/g obtained from BET analysis. The size of the primary particles is about 20-30 nm in diameter. Sodium selenate (Na2Se04) and sodium selenite pentahydrate
(Na2Se03·H2O) were used as the Se(VI) and Se(IV) sources respectively and formic acid (77%) as the organic reductant (hole scavenger). 5×10-4 M copper(II) sulfate and 0.1 M sodium hydroxide solutions were used as two scrubbers in series to trap the hydrogen selenide gas from the system. In the CuSO4 scrubber, the H2Se gas would be removed from the system by reacting with Cu2+ to form copper selenide precipitate (CuSe). The NaOH scrubber would provide an alkaline condition to dissolve any remaining H2Se that might have escaped the CuSO4 scrubber. Sodium hydroxide and perchloric acid were used to adjust the pH. The perchlorate ions were found to have minimal adsorption onto the TiO2 particles (Abdullah et al, 1990). For colorimetric analysis of Se(IV), 2,3- diaminonaphthalene (DAN), cyclohexane, hydroxylamine chlorohydrate and EDTA were used. The chemicals were used without further purifications. Deionized pure water was used for the preparation of all solutions.
Apparatus and Procedure
The photoreactor used in this study was designed for the purpose of fundamental study and hence the irradiation efficiency was not the main priority but the ease of handling, sampling
54 and cleaning were important design factors. Figure 3. 1 shows the schematic diagram (Figure 3. 1a) and a photograph of the experimental set-up (Figure 3. 1b). The photoreactor is made of glass and is cylindrical in shape with a 1.2 L capacity. A side quartz-window enables UV irradiation from a 200 W Mercury lamp (Oriel 66001-373). The lamp mainly provides irradiation of wavelength below 380 nm. The reactor was placed on top of a magnetic stirrer for agitation (250 rpm), while the purging of gas (nitrogen or air) (1.5 Lmin-1), introduced from the top of the reactor, provided additional agitation. Gas exhausted from the reactor was introduced into the two CuSO4 and NaOH scrubbers in series. The required Se ions (20 and 40 ppm Se) and formic acid (20 parts per million in terms of carbon (ppmC) to 600 ppmC) concentrations were then made up into a test solution of volume 0.5 or 1 L. This was followed by pH adjustment and the addition of titanium dioxide powder (1.1 gTiO2/L). The reported pH of the experiments was the one obtained after the addition of TiO2. The suspension was allowed to mix for 30 minutes before irradiation to allow for equilibrium adsorption of the solutes. After irradiation, samples were taken from the suspension at regular time intervals.
Two different mercury lamps of the same type were used in the photocatalytic studies. The photon fluxes into the reactor for the different lamps were determined to be 3.15 and 3.17 mmol photon/sec (refer to Appendix B) by chemical actinometry (Hatchard et al, 1956; Bunce, 1989).
55
Direction of N2 gas flow
Pr
Q
S
M
S2 S1 R L P
(a)
(b)
Figure 3. 1. (a) Schematic diagram and (b) photo of the experimental setup. P: power supply, L: lamp housing with Hg lamp, R: glass reactor, M: magnetic stirrer, Q: quartz window, S: sampling port, Pr: pH probe, S1: CuSO4 scrubber for H2Se, S2: NaOH scrubber for H2Se.
56 Adsorption Data and Isotherm Determination of Se Ions on TiO2
Solutions containing sodium Se(VI) and Se(IV) concentration ranging from 2 to 80 ppm (0.025 – 1 mM) and formic acid concentration ranging from 20 to 500 ppm (1.5 - 42 mMC, mMC meaning milli-molar in terms of carbon) were subjected to adsorption isotherm analysis. 1 L of these solutions was added to the reactor with 1.1 gTiO2 powder and stirred continuously with nitrogen bubbling for 30 minutes at atmospheric conditions. From the kinetics of adsorption, it was found that 15 minutes was sufficient to reach adsorption equilibrium. After the suspension was filtered, the concentration of Se and total organic carbon in the filtrate was analysed. The amount of Se or formic acid adsorbed onto the TiO2 particles was calculated from the difference between the initial and residual Se anions or formic acid concentrations. It was also found that the amount Se anions or formic acid adsorbed onto the reactor walls was insignificant. All adsorption experiments were conducted at room temperature at 293±2 K.
Analysis
All samples were collected using a syringe and then immediately filtered using 0.22mm Millipore membrane. Total Se concentration (Se(VI) or Se(IV)) in the filtrate was determined using Induced Coupled Plasma-Atomic Emission Spectroscopy (Varian ICP- AES). For some experiments, Se(IV) was also discriminated from Se(VI) by a colorimetry analytical method developed by Holtzclaw et al (1987). The amount of H2Se generated was determined by analysing the amount of Cu(II) remaining in the trap using ICP-AES. Formic acid concentration was determined by analysing the total organic carbon (TOC) in the solution using a Shimadzu TOC-5000A analyzer or Anatoc II TOC analyser. The amount of elemental Se deposited on titanium dioxide was determined by digestion using concentrated nitric acid and analyzing the digested Se ions by Varian ICP-AES. Zeta potential and particle size of the TiO2 suspensions were determined using the electrophoresis and Photon Correlation Spectroscopy (PCS) techniques (using Brookhaven 3-in-1 system). The surface area of Degussa P25 was obtained by Micromeritics ASAP
57 2000 BET surface analyser. X-ray diffraction was performed using Siemens D5000 Diffractometer. Energy Dispersive X-ray analyser (EDX) (in-built in the TEM-Phillips CM200) was also used to identify the Se element.
3A.3 Results and Discussions
Determination of TiO2 Surface Charge
Prior to the adsorption studies, the surface charge of TiO2 photocatalyst was examined. The pHzpc of Degussa P-25 TiO2 was determined to be 5.6 in the absence of the Se anions. The values for the pHzpc in the presence of 20 ppm Se(VI) and Se(IV) ions were found to be 5.4 and 5.15 respectively. This is shown in Figure 3. 2. In the presence of the Se anions, adsorption of the negatively-charged Se ions onto the positively-charged TiO2 surface resulted in the decrease in the magnitude of the charge. The lower pHzpc value in the presence of Se(IV) ions compared to that of Se(VI) ions is the result of greater adsorption of the Se(IV) ions on the TiO2 surface. The comparison for Se(VI) and Se(IV) adsorption will be discussed later.
50
40
30
20
10
0
-10 012345678910 pH
Zeta-potential (mV) -20
-30
-40 Se Absent Se(VI) Se(IV)
Figure 3. 2. Surface charge of TiO2 at various pH in the absence and presence of Se ions (20 ppm).
58 Se(VI) Adsorption Studies
The individual adsorption isotherms of Se(VI) and formate ions onto the TiO2 surface were then investigated. The adsorption isotherm of Se(VI) conforms to the trend of multi-layer adsorption for a non-porous particle, which can be described by the Brunauer-Emmett- Teller (BET) multi-layer adsorption isotherm model (equation 1). This is shown in Figure 3. 3. This model assumes that the adsoprtion layers beyond the first have equal energies of adsorption (Gregg & Sing, 1982):
BCeQ0 qe= (3-1) éCeù Ce-Csê+B-)1(1)( ú ëCsû
where Cs is the saturation concentration, qe is the amount of solute adsorbed per unit weight of adsorbent, Ce is the residual solute concentration at equilibrium, B is the constant expressive of energy of adsorption and Q0 is the solid-phase concentration corresponding to complete coverage of the available sites.
12.0
e 0.6 10.0 )*q
e 0.4 ) 2 -C 8.0 s 0.2 y = 0.653x R2 = 0.988 /(C 0.0 e
6.0 C 0.00.51.0 Ce/Cs 4.0
2.0 Se(VI) Ads (mg/gTiO
0.0 01020304050607080
Equilibrium Se(VI) Conc (ppm)
Figure 3. 3. Adsorption Isotherm of Se(VI). Experimental conditions: pH 3.5, 1 L test solution, 1.1 gTiO2/L, N2 purging, 293 K. Insert: Linearised BET model plot.
59 The adsorption isotherm of formic acid is shown in Figure 3. 4. It follows the Langmuir adsorption isotherm which can be expressed as:
C CK F = eF (3-2) C -mF 1+ CK eF
where CF-m is the monolayer saturation concentration, CF is the amount of solute adsorbed per unit weight of adsorbent, Ce is the residual solute concentration at equilibrium and KF is the adsorption constant.
4.0 8 3.5 6 ) 2
3.0 adsorbed 4 2 y = 107.342x + 0.027 R2 = 0.995 2.5 0 00.050.1 2.0 1/HCOOH 1/HCOOHequilibrium 1.5
1.0
0.5 HCOOH Ads (mgC/gTiO Ads HCOOH
0.0 050100150200250300 Equilibrium HCOOH Conc (ppmC)
Figure 3. 4. Adsorption Isotherm of HCOOH. Experimental conditions: pH 3.5, 1 L test solution, 1.1 gTiO2/L, N2 purging, 293 K. Insert: Linearised LH Adsorption Model plot.
Comparing the adsorption isotherms in Figure 3. 3 and Figure 3. 4, the adsorption of formic acid onto TiO2 was not as significant as Se(VI) adsorption at pH 3.5. Only 3.0 mgC/gTiO2 of formate ions were adsorbed when the amount of formic acid added into 1.1 g/l TiO2 was 300 mgC. However, when the amount of Se(VI) added in solution was 70 mg, about 11.3 mg/gTiO2 of Se(VI) ions were adsorbed, forming multi-layers of Se(VI) ions. This
60 indicated a higher affinity of Se(VI) ions to the TiO2 surface, which could be attributed to the higher negative charge of Se(VI) ions.
The simultaneous adsorption of Se(VI) and formic acid on TiO2 at various pH values and initial solute concentrations were also investigated. Figure 3. 5 shows the adsorption of
Se(VI) ions onto TiO2 particles in the absence and presence of formic acid at various pH values. The adsorption of Se(VI) ions was found to reach equilibrium within 1-2 minutes.
In the investigated pH range of 1.5 to 6.5, sodium selenate (pKd=0.02) (Seby et al, 2001) was completely dissociated while formic acid (pKa=3.77) (Chang, 1994) was increasingly ionized to the negatively-charged formate ions as the pH increased. Se(VI) is mostly 2- present as the negatively charged SeO4 ions at pH 2 and above (H2SeO4, pKa= -2.01±0.06 - and HSeO4 , pKa=1.8±0.1) (Seby et al, 2001). The negatively-charged nature of the Se(VI) and formate ions enabled the adsorption of these ions onto the positively charged TiO2 below its pHzpc.
7.0
6.0 HCOOH
) Presence 2 5.0 HCOOH Absence 4.0
3.0
2.0
1.0 Se(VI) Ads (mg/gTiO 0.0 0.01.02.03.04.05.06.07.0 pH
Figure 3. 5. Effect of pH on Se(VI) adsorption. Experimental conditions: 1 L test solution, [Se(VI)]0=20ppm (0.256 mM), [HCOOH]0=100ppmC (8.3 mMC), 1.1 gTiO2/L, N2 purging, 293K.
Also shown in Figure 3. 5, at pH 1.5 and 2.5, the amount of Se(VI) ions adsorbed was not affected by the presence of formic acid. This was due to the low formate ionisation at high H+ concentration (low pH). As the pH increased, the amount of Se(VI) ions adsorbed
61 decreased since the TiO2 surface became less positively charged. At pH higher than 2.5 and in the presence of formic acid, less Se(VI) ions were adsorbed onto TiO2 at pH higher than 2.5, indicating a competition between the two negatively-charged Se(VI) and formate ions for the positively-charged adsorption sites on the TiO2 particles. As the pH was raised, formate ions concentration increased due to a greater extent of formic acid ionisation, resulting in the increase in formate ion adsorption which further depressed the adsorption of
Se(VI) ions. From pH 5.5 to 6.5, once the pHzpc of TiO2 was reached, very little Se(VI) ions were adsorbed.
The competitive adsorption of the Se(VI) and formate ions at various pH is further depicted in Figure 3. 6. Besides further consolidating the finding that increasing pH resulted in the decreased adsorption of Se(VI) and the increased adsorption of formate ions, the data from Figure 3. 6 also enables the determination of the formate:selenate ratio adsorbed onto the
TiO2 surface. The molar adsorption ratios were calculated to be 0.012, 0.152, 2.75 and 16.8 at the pH of 1.5, 2.5, 3.5 and 4.5, respectively. The relationship between these ratios and the photocatalytic reduction of Se(VI) will be discussed in the later sections.
7.0 4.5
4 6.0 ) ) 2 3.5 5.0
3 ) 2
4.0 2.5
3.0 2
Se(VI) 1.5 (mgC/gTiO 2.0 HCOOH HCOOH Adsorbed HCOOH 1 1.0 0.5 Se(VI) (mg/gTiOAdsorbed 0.0 0 0.01.02.03.04.05.0 pH
Figure 3. 6. The simultaneous adsorption of Se(VI) and HCOOH at various pH. Experimental conditions: 1 L test solution, [Se(VI)]0=20 ppm, [HCOOH]0=100 ppmC, 1.1 gTiO2/L, N2 purging, 293 K.
62 The simultaneous adsorption of Se(VI) and formic acid at different initial solute concentrations was also investigated. Figure 3. 7 (a) and Figure 3. 7(b) show the adsorption of Se(VI) and formic acid for an initial Se(VI) concentration of 20 and 40 ppm respectively at varying formic acid concentrations. For both initial Se(VI) concentrations, as the formic acid concentration increased, the amount of Se(VI) adsorbed decreased, again indicating the competitive adsorption of these two solutes.
5.0 3.0 5.0 3.0 ) 2
) 2.5 2 4.0 2.5 4.0 ) ) 2
2 2.0 2.0 3.0 3.0 1.5 1.5 2.0 2.0 1.0 1.0 HCOOH Ads HCOOH (mgC/gTiO HCOOH Ads HCOOH 1.0 (mgC/gTiO 1.0 0.5 0.5 Se(VI)Ads (mg/gTiO 0.0 0.0 Se(VI)Ads (mg/gTiO 0.0 0.0 0.0100.0200.0300.0400.0 0.0100.0200.0300.0400.0 HCOOH Concentration (ppmC) HCOOH Concentration (ppmC)
Figure 3. 7. Adsorption of Se(VI) and HCOOH at various initial HCOOH Concentration. Experimental conditions: pH 3.5, 1 L test solution, 1.1 gTiO2/L, N2 purging, 293 K. ¨ Se(VI) adsorbed, ■ formate adsorbed.
Preliminary Se(VI) Reduction Studies
Preliminary studies of Se(VI) reduction were conducted in the presence and absence of UV irradiation and formic acid. Formic acid was chosen as the organic hole scavenger due to its single-carbon linear structure which results in simple oxidation products, that is carbon dioxide and water (Aguado & Anderson, 1993). In the absence of TiO2, UV and formic acid, the reduction of Se(VI) and Se(IV) was only found to occur at a highly acidic condition when concentrated sulfuric acid of greater than 2 M was used at a reaction temperature of 80°C. The reduction was indicated by the formation of a pink precipitate in the solution. The pink particles were determined to be elemental Se after filtering and digesting the particles using concentrated nitric acid. In a study involving stirring Se ions in solutions of pH 2.6 for an hour at room temperature, a negligible amount of Se ions was
63 removed and no colour change was observed. This indicates the chemical reduction of the Se ions only occurred at very low pH and at elevated temperature. The pH 2.6 was used since in some of the preliminary experiments to be discussed later, 100 ppmC of formic acid was added which resulted in a final pH of 2.6. For experiments without formic acid, perchloric acid was used to adjust the pH to the various desired values.
The results from the preliminary studies are summarised in Table 3. 1. Experiments (Expt) 1 to 4 were carried out in anoxic environment (under nitrogen gas sparging) while Expt 5 was carried out using air as the sparging gas. In the absence of UV irradiation and formic acid (Expt. 1), about 28% of Se(VI) ions were removed from the solution and this was attributed to dark adsorption of the Se(VI) ions onto the TiO2 surfaces. A similar observation was encountered when formic acid was added to the solution in the absence of UV irradiation (Expt. 2) as well as in the presence of UV irradiation but without formic acid addition (Expt. 3).
When formic acid was introduced into the solution in the presence of UV irradiation (Expt. 4), about 92% of Se(VI) was removed from the solution after 60 minutes of irradiation. The colour of the test suspension changed from white to an intense orange-pink. This is attributed to the deposition of elemental Se onto the TiO2 surface. Upon further irradiation, the colour of the suspension turned white again due to the further photoreduction of Se to
H2Se. This shows that the presence of formic acid was essential in achieving Se(VI) photoreduction. In the presence of formic acid, the photogenerated holes were effectively scavenged from the TiO2 particles, preventing recombination with the photogenerated electrons and hence allowing electrons to reduce Se(VI) ions. This explained why Se(VI) reduction was not observed when formic acid was not added to the UV/TiO2 system (in Expt. 3).
In addition to having the role as the hole scavengers, formate ions can form reducing radicals upon reaction with the hydroxyl radicals formed. These reducing radicals can also play a role in the reduction of the Se(VI) ions. More discussions on the role of reducing formate radicals will be presented in Section 3B.
64 When the reaction suspension was sparged with air (see Table 3. 1 Expt. 5), Se(VI) removal after 60 minutes was about 62%, as compared to 92% for the continuous nitrogen sparging experiment. This observation of reduced photoreduction rate was similar to those encountered in other investigations on the photoreduction of chromate (Alam & Montalvo, 1998) and Hg+ (Khalil et al, 2002). In those studies, the reduced photoreduction observed was attributed to the dissolved oxygen in the suspension competing with the chromate or Hg+ for the electrons. Similarly in the present case, the dissolved oxygen is believed to compete with the Se(VI) ions for electrons, hence retarding the Se(VI) photoreduction rate.
Table 3. 1. Preliminary Experiments of Se(VI) Reduction. Experimental conditions: [Se(VI)] 0=20 ppm (0.256 mM), [HCOOH]0=100 ppmC (8.3 mMC), 1L test solution, pH= 2.6±0.1, 1.1 g TiO 2/L, 293 K. Irradiation time: Expt 1 & 2: 30 mins, Expt 3-5: 60 mins.
Expt. UV Formic Sparging Se(VI) Dark Total Se(VI) Colour H2Se Irradiation Acid Gas Adsorption (%) Removed after change Formed irradiation (%)
1 No No N2 28.0 28.0 No No
2 No Yes N2 27.0 27.0 No No
3 Yes No N2 26.9 26.9 No No
4 Yes Yes N2 26.9 91.6 Yes Yes 5 Yes Yes Air 26.7 62.5 Yes No
The Redox Reactions on the TiO2 Surface
The simultaneous oxidation and reduction reactions on the TiO2 surface may be explained as follows. The formation of elemental Se (indicated by the orange-pink colour) upon the photocatalytic reduction of Se(VI) could be described by equations (2-24). Photos showing the colour of the TiO2 particles before and after the photoreduction of Se ions are presented in Figure 3. 8 (a and b respectively). The orange-pink particles are subjected EDX analysis
65 and were confirmed to be elemental Se. A detailed discussion on the Se particle formation mechanism will be presented in the Chapter 4A.
When Se(VI) was used as the Se precursor, from the colorimetric analysis, no Se(IV) ions could be detected in the solution during the reduction. Since Se(IV) is more strongly adsorbed onto the TiO2 surface (the adsorption of Se(IV) is presented in the later section), it is suggested that the Se(IV) ions were adsorbed on the TiO2 surface after they were formed from Se(VI) reduction, and were subsequently reduced to elemental Se. The overall photoreduction of the Se(VI) ions could be taken to proceed directly to elemental Se as described by equation (2-24).
(a) (b)
Figure 3. 8. Degussa TiO2 powder collected on filter paper (a) before and (b) after Se ions photoreduction.
The oxidation reaction could involve either the direct attack of formic acid by holes or by hydroxyl radicals. As mentioned previously, this could form reducing formate radicals which could play a role in the Se ions reduction. Complete oxidation or mineralization of formic acid may be described by equations 3-3a and 3-3b:
-++ (3-3a) HCOO +2 ®COh 2 +H
66 -·- (3-3b) HCOO +2 ®+22 +OHOHCOOH
When the near complete reduction of Se(VI) was encountered, the formation of a black precipitate was observed in the clear-blue copper sulfate solution (used as the H2Se trap). This also corresponded to the most intense orange colour of the suspension observed in the photoreactor. The black precipitate formed has been confirmed to be copper (II) selenide
(CuSe) by XRD (Sanuki et al, 1999, 2000). The generation of H2Se was the result of the further reduction of elemental Se to H2Se as given in equation (2-25). The oxidation of formic acid continued during the reduction of Se to H2Se.
A mass balance on Se (as shown in Table 3. 2) further confirmed the reduction of Se(VI) to elemental Se and the further reduction of elemental Se to H2Se gas. The mechanism of
H2Se generation will be presented and discussed in Chapter 4A.
Table 3. 2. Mass balance of selenium species before and after photoreduction. TiO 2 loading = 1.1g/L, reaction volume = 1L, residence time = 65min, pH=3.5. Mass of Se added (mg) Mass of Se after reduction (mg) % Error
Se(VI) in Se(0) on TiO2 Se(VI) in Se(0) on Se as
suspension suspension TiO2 H2Se
20.5 0 0.5 14.3 4.5 5.8
Effect of pH on Se(VI) Photoreduction
Following the adsorption and the preliminary Se(VI) reduction experiments, the Se(VI) photoreduction reaction was studied at various pH values. The results presented in Figure 3. 9 shows the concentration profiles of Se(VI) as the photoreduction proceeded at pH 2.2, 3.0 and 6.4. The decrease in the initial Se(VI) concentration at time 0 minute was reflective of the Se(VI) dark adsorption. Se(VI) dark adsorptions at pH 2.2, 3.0 and 6.4 were 5.42, 3.68 and 0.12 mg/gTiO2 respectively. These results consolidated the previously presented results
67 that an increase in pH led to a decrease in Se(VI) adsorption. Once the solution was exposed to UV irradiation, the decrease in Se(VI) concentration was attributed to the reduction of Se(VI) to Se. From Figure 3. 9, it seemed that a zero-order rate could represent the Se(VI) photoreduction rates well and hence was used for the rate estimation. The high adsorption of the Se(VI) species onto the TiO2 surface may have resulted in the well- representation of the photoreduction rates by the zero-order rate law. This agreed with the simplification of the general kinetic equation 2-20 to equation 2-12, which is the result of high initial substrate concentration (section 2.1.4). Comparing the Se(VI) photoreduction rates at the different pH values, it can be seen that the rate was fastest at pH 3.0 and slowest at pH 6.4.
24
20
16
12
8
4
0 Se(VI) Concentration (ppm) solution in 020406080100120140 time (min)
Figure 3. 9. Effect of pH on the photoreduction of Se(VI). □ pH 2.2 (R2=0.996), × pH 3.0 (R2=0.996), ○ 2 pH 6.4 (R =0.854). Experimental conditions: 1 L test solution, [Se(VI)]0=20 ppm, [HCOOH]0=100 ppmC, 1.1 gTiO2/L, N2 purging, 293 K, photon intensity 3.15 mmol/sec.
In order to further evaluate the effect of pH, Se(VI) photoreduction experiments were carried out at other pH values. These results, presented as the Se(VI) photoreduction rates as a function of pH are shown in Figure 3. 10. The Se(VI) photoreduction rate increased
68 and reached a maximum value at pH 3.5, followed by a decrease until the photoreduction ceased at pH 6.5. This observed trend can again be correlated with the effects of pH on Se(VI) and formate adsorption shown in Figure 3. 6. It is interesting to note that the maximum rate was achieved at pH 3.5 which corresponded to a formate-to-selenate molar ratio of 2.75. This is very close to the stoichiometry of 3:1 of the overall reaction as determined from balancing the number of e- and h+ generated in equations (2-24) and (3- 3a):
- 2- + 0 (3-4) 3HCOO SeO4542++®++ 3COOHSeH 2
0.35
0.30 )
2 0.25
0.20
0.15
(mg/min.gTiO 0.10
0.05 Se(VI) Photoreduction Rate 0.00 0.02.04.06.08.0 pH
Figure 3. 10. The effect of pH on Se(VI) photoreduction rate. Experimental conditions: 1 L test solution, [Se(VI)]0=20 ppm, [HCOOH]0=100 ppmC, 1.1 gTiO2/L, N2 purging, 293 K, photon intensity 3.15 mmol/sec.
This finding may be suggestive of the importance of a near stoichiometric adsorption of formate and Se(VI) ions on the TiO2 surface in order to achieve the highest efficiency in capturing the photogenerated electrons and holes by Se(VI) and formate ions respectively. The importance of stoichiometric adsorption is demonstrative that the redox reaction took place on the TiO2 surface, of which the surface sites for the adsorption of both solutes could
69 be limited. The limitation of surface sites was demonstrated earlier by the competitive adsorptions between the Se(VI) and formate ions.
The previously suggested role of the reducing formate radicals was still valid under the above postulate. This is because the effect of formate radicals could only be significant after formate ions were adsorbed onto the TiO2 surface since formate ions are most likely to scavenge the hydroxyl radicals on the surface. The adsorption of the “right” amount of Se(VI) and formate on the surface would hence entail an efficient photoreduction rate, be it (formate ions) as a hole scavenging or the subsequent reduction reactions by the reducing species.
Effects of Initial Solute Concentration on Se(VI) Photoreduction
In the next series of experiments, the effect of initial Se(VI) and formic acid concentrations on the Se(VI) photoreduction rates were studied. The experimental conditions were similar to those described earlier for the adsorption experiments (Figure 3. 7). The photoreduction results, represented by the zero-order rates, are shown in Figure 3.9. It can be observed from Figure 3. 11 that increasing the formic acid concentrations from 10 to 300 ppmC resulted in the increase in reduction rates until a maximum value was reached. Beyond this maximum value, a higher formic acid concentration was found to depress the reduction rate. The optimum formate ion concentrations corresponding with an initial Se(VI) concentration of 20 and 40 ppm were 100 and 200 pppmC, respectively. The adsorption tests (Figure 3. 7 (a) and (b)) showed that as the formic acid concentration increased from
15 to 300 ppmC, Se(VI) adsorption onto TiO2 surface decreased. From the adsorption data presented in Figure 3. 7 (a) and (b), the formate-Se(VI) molar adsorption ratios for the two optimum conditions discussed above were 2.6 and 2.7 for the initial Se(VI) concentrations of 20 and 40 ppm respectively. These two values are again comparable to the stoichiometric ratio of 3:1 in the reaction described in Equation (3-8). Hence, the existence of an optimum formate ions concentration (as observed in Figure 3. 11) may also be attributed to the near 3:1 stoichiometric adsorption ratio of formate:Se(VI) ions. The
70 similar value for the maximum photoreduction rates of about 0.3 mg/min.gTiO2 obtained at 20 amd 40 ppm Se(VI) concentration is also highly supportive of the occurrence of optimum conditions which closely corresponded to a stoichimetric adsorption ratio.
0.35
0.30 )
2 0.25
0.20
0.15
0.10 (mg/min.gTiO
0.05
Se(VI) Photoreduction Rates 0.00 0100200300400 Formic Acid (ppmC)
Figure 3. 11. The effect of initial HCOOH concentrations on Se(VI) photoreduction rates. Experimental conditions: pH 3.5, 1 L test solution, 1.1 gTiO2/L, N2 purging, 293 K. □ 20 ppm and ○ 40 ppm initial Se(VI) concentration.
The maximum photoreduction rates were also manifested in the results of the investigations by Chenthamarakshan and Rajeshwar (2000) during the photocatalytic reduction of chromate using TiO2. In their study, it was found that by increasing the concentration of organic additives, even though the adsorption of chromate was decreased, the reduction rate was significantly enhanced, indicating the existence of an optimum organic concentration. However, no explanation was provided as to why an optimum concentration existed.
Validation of the Significance of Optimum Adsorption for Se(VI) Photoreduction
To further validate the importance of stoichiometric adsorption, four different sets of adsorption and photoreduction experiments (Sets A to D) of different initial Se(VI) and
71 formic acid concentrations. The experiments of Set A had initial Se(VI) and formic acid concentrations of 10 ppm and 50 ppmC respectively, while those of Set B were 40 ppm and 200 ppmC, respectively. Experiments Set A and B were carried out at various pH values. The results are summarised in Table 3. 3.
Table 3. 3. Effect of formate:Se(VI) molar adsorption ratio on Se(VI) photoreduction rate. Experimental conditions for Set A and B: 1 L test solution, 1.1 gTiO2/L, 293 K, N2 purging, 60 min residence time. Set A: [Se(VI)]0=10 ppm, [HCOOH]0=50 ppm, Set B: [Se(VI)]0=40 ppm, [HCOOH]0=200 ppm. Adsorption Experiments Photoreduction Experiments
pH Se(VI) Formate Formate:Se(VI) Se(VI) photoreduction (mg/gTiO2) (mgC/gTiO2) Molar Ratio rate (mg/min.gTiO2) Set A
3.5 1.91 0.102 0.351 0.229
3.9 1.62 0.211 0.856 0.253
4.4 0.851 0.432 3.32 0.303
4.8 0.301 0.502 11.0 0.263
Set B 3.5 2.86 1.35 3.10 0.309
4.0 2.58 1.61 4.27 0.263
4.4 2.21 2.48 7.34 0.187
By adjusting the pH of the solution, the amount of Se(VI) and the formate ions adsorbed onto the TiO2 surface was varied. The two different starting initial reactant concentrations (i.e. the difference between Sets A and B) were used since this allowed for a wider range of adsorbed molar ratios of formate to Se(VI) ions. As can be seen from the presented results, when the adsorption of formate and Se(VI) ions was near the stoichiometric molar ratio, the optimum Se(VI) photoreduction rate was encountered. The optimum pH for Set A was found to be 4.4 while that of Set B was 3.5. Again, the maximum rates were also found to
72 be close to the value of 0.3 mg/min.gTiO2, strongly supporting the existence of optimum conditions at a stoichiometric adsorption ratio.
Another sets of experiment (Sets C and D) were performed by varying the order of Se(VI) and formic acid addition into the TiO2 suspension. The results are summarised in Table 3. 4. These experiments were carried out at a different catalyst loading and reaction volume to test the validity of the abovementioned postulate under a different condition. The order of addition was as follows: Method 1: Addition of formic acid and stirred for 30 minutes, followed by the addition of Se(VI) ions and stirred for 30 minutes, Method 2: Addition of Se(VI) ions followed by the formic acid, each stirred for 30 minutes and Method 3: Simultaneous addition of both Se(VI) and formic acid and stirred for 30 minutes, followed by irradiation of the suspension. These experiments were performed based on the knowledge that Se(VI) ions have a stronger affinity to TiO2 than formate ions as observed from the adsorption isotherms obtained earlier (refer to Figure 3. 3 and Figure 3. 4 for
Se(VI) and formate adsorption isotherm respectively). By subjecting TiO2 to formic acid adsorption first, formate ions would have a greater chance of adsorption onto TiO2 while Se(VI) could still probably be adsorbed when introduced later.
Both experiments of Sets C and D indicated that adsorbing formic acid first resulted in a higher Se(VI) photoreduction rate than by adsorbing Se(VI) first (compare method 1 with 2 in Sets C and D). However, the photoreduction rate was comparable to that of simultaneous adsorption (compare method 1 with 3 in Sets C and D). This again showed that Se(VI) ions have greater affinity to TiO2. The results from Set D also indicated that the fastest rate also correlated well to stoichiometric adsorption ratio of 3:1. Comparing Sets C and D, increasing the initial formic acid concentration from 20 to 100 ppmC would increase the reduction rates. This shows that a high formic acid concentration was necessary to maintain a favourable formate adsorption. The dissimilarity in the optimum rates as compared to Sets A and B was due to the different initial conditions used.
73 Table 3. 4. Effect of varying the order of Se(VI) and formic acid addition into the TiO 2 suspension. Experimental conditions for Set C and D: pH 3.5, 0.5 L test solution, 0.5 g TiO 2/L, 293 K, N2 purging, 60 min residence time. Set C : [Se(VI)]0=20 ppm, [HCOOH]0=20 ppmC, Set D: [Se(VI)]0=20 ppm, [HCOOH]0=100 ppmC. Adsorption Experiments Photoreduction Experiments
Method Se(VI) Formate Formate:Se(VI) Se(VI) photoreduction (mg/gTiO2) (mgC/gTiO2) Molar Ratio rate (mg/min.gTiO2) Set C
1 3.40 0.352 0.682 0.470
2 3.62 0.101 0.183 0.372
3 3.55 0.301 0.557 0.446
Set D 1 2.52 1.21 3.16 0.588
2 3.18 0.542 1.12 0.492
3 2.61 1.11 2.79 0.593
Se(IV) Adsorption Studies
In this section, the adsorption and photoreduction of Se(IV) ions are examined by investigating the effects experimental parameters as was carried out for the Se(VI) photoreduction studies. The adsorption isotherm for Se(IV) at pH 3.0 is shown in Figure 3. 12. It conforms to the Langmuir adsorption model similar to equation (3-2). By fitting the experimental data according to equation (3-2), values of 8.65 mg/gTiO2 and 0.609 L/mg were obtained for the parameters CS-m and KS respectively.
74
9.0
)
2 8.0 7.0 6.0 5.0 4.0 3.0 2.0
Se(IV) Adsorbed (mg/gTiO Adsorbed Se(IV) 1.0
0.0 0.05.010.015.020.025.030.0 Se(IV) equilibrium concentration (ppm)
Adsorption Data LAI Model
Figure 3. 12. Adsorption Isotherm of Se(IV) at pH 3. Experimental conditions: 1 L test solution, 1 gTiO2/L, N2 purging, 293 K.
Comparing Figure 3. 3 and Figure 3. 12, the adsorption of Se(IV) was more significant than that of Se(VI). This could be attributed to the difference in ionic structure of the two ions: Se(VI) has four oxygen atoms centrally bonded to the Se atom, forming a tetrahedral structure, while Se(IV) has three oxygen atoms centrally bonded to the Se atom, forming a triagonal-planar structure. Ionic repulsion of Se(VI) ions on the TiO2 surface would be more significant than that of Se(IV) owing to the higher electron density of the four oxygen atoms. In addition, the greater spatial volume of the tetrahedral Se(VI) ion could result in less number of Se(VI) ions to be adsorbed. These combined factors were postulated to cause lower Se(VI) adsorption onto the TiO2 surface. The geometry of the Se ions is illustrated in Figure 3. 13.
75 2- 2- (a) Se(VI): SeO4 (b) Se(IV): SeO3 Tetrahedral Triagonal Planar
O O Legend: Se O Se O O O Selenium O Atom
Oxygen Atom
Figure 3. 13 The ionic structure of (a) Se(VI) and (b) Se(IV).
The difference in the adsorption trend for Se(VI) and Se(IV) (multilayer versus monolayer adsorption) was postulated as follows. The greater extent of Se(IV) adsorption, as mentioned previously due to their smaller size and less repulsive force exerted on each other, could reduce the magnitude of the effective surface charge of the TiO2 due to the large number of the Se(IV) ions adsorbed on the surface. Consequently, only monolayer adsorption was attained. On the contrary, since less Se(VI) ions were adsorbed, the magnitudude of the effective surface charge was not significantly shielded and hence was still strong enough to cause the adsorption of subsequent layers of Se(VI) ions. This postulate could be supported by the TiO2 surface charge investigation in the presence of the Se ions as presented in Figure 3. 2. From Figure 3. 2, it can been seen that at a selected pH, for example pH 2, and with the same concentration of Se anions (20 ppm), the magnitude of the surface charge of TiO2, as reflected by zeta potential, in the presence of Se(VI) was 30 mV as compared to that of Se(IV) of 20 mV. As discussed, the higher magnitude could possible result in the adsorption of Se(VI) beyond the first layer, and hence manifesting multilayer adsorption.
Figure 3. 14 shows the effects of pH on Se(IV) adsorption in the absence and presence of formic acid. It was found that and in the range of pH approximately 1.5 to 2.5, the amount of Se(IV) ions adsorbed was similar. As the pH increased, the amount of Se(IV) adsorbed
76 decreased since the TiO2 surface became less positively charged. In the presence of formic acid, less Se anions were adsorbed onto TiO2 at pH higher than 2.0, indicating a competition between the two negatively-charged Se(VI) and formate ions for the positively-charged adsorption sites on the TiO2 particles. The explanations given for Se(VI) adsorption in the previous section could be applied for Se(IV) as well.
10 9 ) 2 8 7 6 5
4 HCOOH 3 Absent 2 HCOOH Se(IV) Ads (mg/gTiO Ads Se(IV) 1 Present 0 01234567 pH
Figure 3. 14. The effects of pH on Se(IV) adsorption in the presence and absence of HCOOH. Experimental conditions: 1 L test solution, [Se(IV)]0 =20 ppm (0.256 mM), [HCOOH]0 =100ppmC (8.3 mMC), 1.1 gTiO2/L, N2 purging, 293K.
Figure 3. 15 illustrates the effect of formic acid concentration on Se(IV) adsorption. The competitve adsorption of Se(IV) and formate ions was encountered only until a formic concentration of about 500 ppmC. Beyond this concentration, competitive adorption of the two ions became less significance. As explained previously, the competitve adsorption of the ions could indicate the existence of limited adsorption sites on the TiO2 surface. It is hence postulated that the Se(IV) ions could have saturated the limited adsorption sites on the TiO2 surface. This is supported by data in Figure 3. 12, showing the complete monolayer surface coverage encountered at 20 ppm initial Se(IV) concentration. This prevented the further adsorption of the formate ions at high formic acid concentration (beyond 500 ppmC), resulting in little change in the formate adsorption.
77 8.0 1.4 ) 2 7.5 1.2
1.0 )
7.0 2 0.8 6.5 0.6
6.0 (mgC/gTiO 0.4 HCOOH Adsorbed Adsorbed HCOOH
5.5 0.2 Se(IV) Formate Se(IV) (mg/gTiOAdsorbed 5.0 0.0 05001000150020002500
Formic Acid Concentration (ppmC)
Figure 3. 15. Equilibrium dark adsorption of Se(IV) and HCOOH on TiO2 surface at 20 ppm initial Se(IV) concentrations, 50-2000 ppmC HCOOH, pH 3.0, 293 K, 1L test solution, N 2 purging.
Preliminary Se(IV) Reduction Studies
The preliminary experiments for Se(IV) ions photoreduction were performed at similar experimental conditions to those of Se(VI) for the purpose of comparison. The results for the preliminary experiments for Se(IV) are summarised in Table 3. 5. Similar to the results for Se(VI) photoreduction (see Table 3. 1), it could be seen that in the absence of UV irradiation, photoreduction did not take place and the Se(IV) removal was due to dark adsorption alone (Expt 6 & 7). Again, similar to Se(VI), photoreduction was also not observed in the absence of formic acid (Expt 8) and the rate of photoreduction was retarded in the presence of oxygen (Expt 10). Sanuki et al (1999) also reported insignificant photoreduction of Se(IV) and Se(VI) ions in the absence of formic ions. However, this was contradictory to that observed by Chenhamarakshan et al (2000), who reported photoreduction of Se(IV) in the absence of organic additive. The reason for this difference was not elucidated in this work and could be followed up in future studies. The highest rate of Se(IV) reduction was found when nitrogen was used as the sparging gas and formic acid was used as the organic additive (Expt 9). The dark adsorption of Se(IV) was more significant than that of Se(VI) (comparing Expt 1 to 5 in Table 3. 1 with that of 6 to10 in
78 Table 3. 5). As explained previously, this is due to the different ionic structures of the two ions. It was also observed that the combined removal of Se(IV) by dark adsorption and photoreduction was faster compared to that of Se(VI) (comparing Expt 9 in Table 3. 5 with Expt 4 in Table 3. 1). The higher rate of Se(IV) removal was attributed to the greater adsorption of Se(IV) onto TiO2 and the greater ease in bond-breaking of the Se(IV) ions due to its ionic structure. More discussions and comparative studies on the Se(VI) and Se(IV) ion removals will be presented in the later section.
Table 3. 5. Preliminary experiments of Se(IV) photoreduction. Experimental conditions: [Se(VI)/ Se(IV)]0=20 ppm (0.256 mM), [HCOOH]0=100 ppmC (8.3 mMC), pH= 2.6±0.1, 1.1 g TiO2/L, 293 K.
Expt. UV Formic Sparging Irradiation Se(IV) Total Se(IV) Colour H2Se Irradiation Acid Gas time Adsorption Removed change Formed (min) (%) (%)
6 No No N2 30 39.9 39.9 No No
7 No Yes N2 30 39.2 39.2 No No
8 Yes No N2 60 39.5 39.5 No No
9 Yes Yes N2 60 38.4 100 Yes Yes
10 Yes Yes O2 60 38.9 76.9 Yes No
The Effect of pH on Se(IV) Photoreduction
The rate of Se(IV) photoreduction at various pH is shown in Figure 3. 16. Similar to the photoreduction of Se(VI), the Se(IV) photoreduction follows a zero order rate. The corresponding formate:selenite molar adsorption ratios were also plotted on the same figure. It could be seen that the optimum at pH 4 corresponds to a formate:selenite molar adsorption ratio at 1.6. This is relatively close to the molar stoichiometric ratio of 2 (2 moles of formate required to react with one mole of selenite) in the overall equation involving formic oxidation (Equation 3-3a) and Se(IV) reduction (Equation 2-24) as shown
79 in equation (3-5). Hence, the previous postulate on the importance of stoichiometric adsorption to achieve optimum rate might also be applicable to Se(IV) photoreduction.
- - + 0 (3-5) 2HCOO HSeO3332++®++ 2COOHSeH 2
0.15 3
) 2 0.10 2
0.05 1
Adsorption Ratio Rate (mg/min.gTiO HCOOH:Se(IV) Molar Se(IV) Photoreduction
0.00 0 2.03.04.05.06.0 pH
Figure 3. 16. The effect of pH on Se(IV) photoreduction rate ( ■) and the corresponding formate:selenite molar adsorption ratio (▲). Experimental conditions: 1 L test solution, [Se(IV)]0=20 ppm, [HCOOH]0=300 ppmC, 1.1 gTiO2/L, N2 purging, 293 K, photon intensity 3.17 mmol/sec.
Effects of Initial Solute Concentration on Se(IV) Photoreduction
The rates of Se(IV) photoreduction at various formic acid concentration, together with the molar formate-to-selenite adsorption ratio, were shown in Figure 3. 17. Contrary to that observed in Se(VI) photoreduction, no optimum formic acid concentration was encountered for the case of Se(IV) photoreduction in the range of the investigated formic acid concentration. The Se(IV) photoreduction rate increased rapidly in the range of 20 to about 500 ppmC. This is attributed to the increase in formate ions, and hence the greater amount of hole scavenger adsorption. This is manifested by the increase in molar adsorption ratio from 0.2 to 1.0. Beyond the 500 ppmC formic acid concentration, it can be seen that the
80 rate of Se(IV) photoreduction reached a plateau. As previously explained and as was evident from Figure 3. 15, competitive adsorption of the Se(IV) and formate ions became less significant above 500 ppmC. The little change in adsorption of the two ions could explain the little variation in Se(IV) photoreduction rate and the molar adsorption ratio above 500 ppmC of formic acid. As a result of this, it is suggested that, for in the current study, and under the experimental conditions studied, , the molar adsorption ratio near the predicted 2:1 stoichiometry could not be achieved.
0.12 1.6 ) 2 1.2 0.08
0.8
0.04 0.4 Adsorption Ratio rate (mg/min.gTiO Formate:Se(IV) Molar
Se(IV) Photoreduction Se(IV) Rate Ratio 0.00 0 0500100015002000 HCOOH Concentration (ppmC)
Figure 3. 17. The effect of initial HCOOH concentrations on Se(IV) photoreduction rates and the corresponding molar adsorption ratio. Experimental conditions: 20 ppm initial Se(IV) concentrations, pH 3.0, 1 L test solution, 1.1 gTiO2/L, N2 purging, 293 K, photon intensity 3.17 mmol/sec.
Validation of the Significance of Optimum Adsorption for Se(IV) Photoreduction
Even though the predicted 2:1 stoichiometry could not be achieved in the previous section, the importance of a stoichiometric molar adsorption ratio for achieving the optimum Se(IV) photoreduction rate at different conditions was still investigated. In the following experiments, the initial Se(IV) concentrations used was 10 ppm at two different initial formic concentrations of 235 (Set E) and 400 (Set F) ppmC. Within each set, the
81 experiments were performed at three different pH values. The results are summarised in Table 3. 6.
Table 3. 6. Effect of formate:Se(IV) molar adsorption ratio on Se(IV) photoreduction rate. Experimental conditions: 1 L test solution, 1.1 gTiO2/L, 293 K, N2 purging, 60 min residence time. [Se(IV)]0= 10 ppm, Set E: [HCOOH]0=235 ppmC, Set F: [HCOOH]0=400 ppmC. Set E Adsorption Experiments Photoreduction Experiments
pH Se(IV) Formate Formate:Se(IV) Se(IV) photoreduction (mg/gTiO2) (mgC/gTiO2) Molar Ratio rate (mg/min.gTiO2) 2.6 6.54 0.100 0.101 0.141
3.5 4.33 0.782 1.19 0.402
5.0 2.20 1.05 3.14 0.233
Set F 2.6 6.03 0.2 0.218 0.272
3.5 4.03 0.92 1.50 0.554
5.0 1.71 1.22 4.65 0.261
It was found that an optimum molar adsorption ratio existed for the experiments with the highest photoreduction rate. For the case of Set E, the optimum ratio of 1.19 was not close to the suggested molar stoichiometric adsorption ratio of 2:1. It is possible that a maximum photoreduction rate corresponding to a stoichiometric ratio of 2:1 may be obtained if a similar experiment was conducted in the pH range of 3.5 to 5.0. For Set F, the maxiumum rate at a ratio of 1.5 was relatively close to the suggested molar stoichiometric adsorption ratio of 2:1. This may consolidate the existence of a stoichiometric adsorption ratio to achieve maximum photoreduction rates, which for set F is believed to be within pH 3.5 to 5. It could also be seen that the Se(IV) photoreduction rate was higher at the higher formic acid concentration (comparing results from Set E with that from Set F). Similar to previous observation for Se(VI) (see Table 3. 4), the high formic concentration could act to maintain
82 the favourable formate adsorption during the photoreduction reactions to provide efficient hole scavengers.
Comparison of the Photocatalytic Reductions for Se(VI) and Se(IV)
In this study, the removal of Se ions was reported separately as dark adsorption and photoreduction. This is because separate analyses were applied to the dark adsorption and photoreduction steps in order to clearly elucidate the effects of the above processes on the removal of the Se ions. For example, separate dark adsorption experiments performed using different Se precursors enabled the elucidation of greater Se(IV) adsorbed compared to Se(VI) at the investigated pH. The subsequent Se ion removal after irradiation was then attributed to the photoreduction of Se ions.
Figure 3. 18 shows the dark adsorption, rate of removal after irradiation and the rate of
H2Se generation for Se(VI) and Se(IV) ions. It can be seen that the overall rate of Se(IV) removal was faster than that of Se(VI). This could be attributed to the greater extent of adsorption of Se(IV) ions, followed by the faster exhaustion of Se(IV) ions in the solution and then the quicker revolution of H2Se gas compared to that of Se(VI). In addition, it is postulated that the triagonal-planar structure of the Se(IV) would allow easier access for the electrons to achieve bond-breaking compared to the tetrahedral structure of Se(VI).
In order to compare the photoreduction rates, the quantum yields, taking into consideration the electrons requirement for both the photoreduction reactions for Se(VI) and Se(IV), were evaluated to be 1.98% and 1.6% respectively (see calculation in Appendix B). This shows that Se(VI) photoreduction was in fact more efficient than that of Se(IV). The rate of Se(VI) removal after irradiation was faster (gradient = -0.043 ppm/min) than that of Se(IV) (gradient = -0.031 ppm/min). The reason behind this observation could not be elucidated at this stage. Finally, it is also important to note that the different size and morphology of the
Se particles formed from the different Se ions precursors could also affect the rate of H2Se generation. This will be discussed in greater details in Chapter 4.
83
10 10 Se(IV) Se(VI) H2Se-Se(IV) H2Se-Se(VI) 8 8
6 6
(ppm) 4 4 Se Generated (mg) 2 2 2 H Se ions Concentration
0 0 050100150200250300 Time (min)
Figure 3. 18. Comparison of the disappearance of Se(VI) and Se(IV) ions and the generation of H2Se from Se ions precursors. Experimental condition: 10 ppm initial Se ions concentrations, 300 ppmC HCOOH concentrations, pH 3.5, 1 L test solution, 1.1 gTiO 2/L, N2 purging, 293 K, photon intensity 3.17 mmol/sec.
3A.4 Conclusions
The effects of pH and initial formic acid concentration on the UV/TiO2 reduction process of
Se ions were investigated. It was found that the adsorption of Se ions onto the TiO2 surface, and the presence of formate ions as the hole scavengers were essential for Se ions photoreduction to elemental Se. This suggests that the hole scavenging process could be rate-limiting to the Se ions photoreduction. Formate ions could also function as reducing radicals upon reacting with the hydroxyl radicals. The photoreduction rate was depressed in the presence of oxygen. The elemental Se was further reduced to hydrogen selenide, H2Se, once the Se ions were exhausted from the solution. The optimum Se(VI) photoreduction rates were found to be closely correlated to the stoichiometric molar adsorption ratio of 3:1 of formate-to-selenate on the TiO2 surface. For Se(IV) photoreduction, optimum molar formate-to-selenate adsorption ratio was also encountered but its correlation with the stoichiometry of 2:1 was not conclusive. The importance of optimum molar adsorption of
84 the precursors demonstrated the synergism between photocatalytic oxidation and reduction. It showed that efficient photocatalytic systems could be realised if both oxidant and reductant were present in the ‘right’ amount on the photocatalyst surface. It was observed that the overall rate of Se(IV) removal was faster than that of Se(VI). However, the rate of Se(VI) photoreduction, comparing the quantum yield of Se(VI) and Se(IV) photoreduction, was found to be higher.
85 3B. Effects of Organic Hole Scavengers on the Photocatalytic Reduction of Selenate and Selenite over UV-Illuminated Ti02
3B.1 Introduction
The importance of a hole scavenger, usually an organic compound, in the photocatalytic reduction process has been highlighted in the previous section. For efficient photoreduction, it is desired that the degradation of the organic hole scavenger is relatively straightforward so that the reduction process will not be complicated. Hence, the choice of the organic hole scavenger plays an important role photoreduction processes.
Prairie et al (1993) have investigated the effects of various organic additives on the photoreduction of chromate. The organics investigated were EDTA, citric acid, salicylic acid, acetic acid, ethanol and methanol. They found that the type of organic was important in determining the rate of chromate reduction, with EDTA and citric acid resulting in the fastest reduction rates. They suggested chelating and enhanced adsorption may contribute to the observed results. They have also suggested that conduction band reduction reactions were faster for organics that were more easily oxidised.
The photoreduction of other compounds using different organics have also been investigated. The photocatalytic reduction rates of nitro-aromatic compounds (Brezova et al, 1997) and Cd2+ (Chenthamarakshan et al, 2000a) were found to follow the order of formate > methanol > ethanol > n-propanol. The trend was explained in terms of the different rates of reducing radicals formation when these organic compounds reacted with hydroxyl radicals. The reactions involved the abstraction of a hydrogen atom from the a-position carbon to the hydroxyl group of the organic compounds (Kaise et al, 1994). The process, using formate and alcohols as examples, occurs as described in reactions 3-6a and 3-6b respectively,
86
- -·· HCOO OH COO +®+ H2O (3-6a)
· · 2 +--®+-- 2OHOHCHROHOHCHR (3-6b)
The relative yields of hydrogen abstraction have been reported to be in the order of formate > methanol > ethanol > n-propanol, with primary alcohols having greater relative yields than the secondary alcohols (Asmus et al, 1973). Attempts have also been made to correlate solvent properties (such as viscosity, solvent polarity and polarisability) of the alcohols to the photocatalytic reduction rate (Brezova et al, 1997). The solvent properties were shown to affect the electron transfer kinetics and hence affecting the photocatalytic activities (Hecht & Fawcett, 1995).
The current section aims to investigate the role of different organic compounds on the photoreduction of Se(VI) and Se(IV) ions. The organic compounds investigated in this study are sucrose, acetic acid, salicylic acid, formic acid, methanol and ethanol. To the best of our knowledge, all previous investigation on Se ions photoreduction utilised formic acid as the organic hole scavenger [Kikuchi & Sakamoto, 2000; Sanuki et al, 1999; 2000]. This could be due to formic acid having a simple one-carbon molecular structure and hence its oxidation to carbon dioxide is straightforward and involves a minimal number of intermediate products (Aguado & Anderson, 1993). In addition, as mentioned previously, formic acid is capable of forming reducing radicals, which could help in the reduction reaction (Kaise et al, 1994).
87 3B.2 Equipment and Procedure
Catalyst and Reagent
Catalysts and chemicals relevant to this section have been described in section 3A.2. In addition, the organic compounds used were acetic acid, salicylic acid, methanol, ethanol, and sucrose.
Photoreactor and Experimental Procedure
Two different reactors were used in the current study. The first was for the Se ions photoreduction study and was illustrated and described in section 3A.2. The second was for the study of organic mineralisation. This reactor was capable of measuring CO2 generation in situ under an anoxic environment. The reactor consisted of a glass coil (borosilicate tubing, 5.0mm inside diameter, 6.0 mm outside diameter) of 300 ml. A blue-blacklight fluorescent lamp (NEC, 15W, maximum emission at ~350nm, emission range 300nm to
400nm) fitted through the centre of the glass coil was used for the irradiation of the TiO2 suspensions. ThepH of the TiO2 suspension (2.0 gTiO2/L, 150 mL) was adjusted to 3.0 prior to the reaction using perchloric acid. Nitrogen was then purged through the system for 1 hour, followed by the introduction of the organic compound. The photocatalyst suspension containing the organic was circulated through the reactor for 10 minutes to allow for uniform mixing, after which the lamp was switched on. The amount of carbon dioxide generated during the reaction entered a gas loop which was then detected by an online conductivity meter (Alpha 800). Conductivity readings were converted to the mass of carbon dioxide generated (in terms of mass of carbon) by means of a calibration curve. Each experiment was duplicated to ensure the reproducibility of the results.
88 Analysis
The total Se concentration in the filtrate was determined by Varian Induced Coupled Plasma-Atomic Emission Spectroscopy (ICP-AES). Total organic concentration was determined by analysing the total organic carbon (TOC) in the solution using Anatoc II TOC analyser.
3B.3 Results and Discussions
The photoreduction and dark adsorption of Se ions in the presence of the various organic additives were performed in an anoxic environment and the experimental results are summarised in Table 3. 7. The photocatalytic reduction experiments involving Se(IV) are discussed first. It can be seen in Table 3. 7 that after 120 minutes of irradiation, the greatest amount of Se(IV) was removed from the system in which formic acid was used as the organic additive (Set 1). The second highest amount of Se(IV) removed was observed when using methanol (Set 2) and the lowest with ethanol (Set 3). The colour change from white to orange pink of the system was also observed for these three experiment sets, indicating the reduction of Se(IV) ions to elemental Se. However, in the cases of acetic acid (Set 4), salicylic acid (Set 5) and sucrose (Set 6), and when no no organic additive was present (Set 7), no colour change and no further reduction of Se(IV) ions from the solution was observed during and after irradiation. Hence, it can be deduced that the Se(IV) removal in Sets 4-7 was solely due to adsorption.
In addition, as can be seen from Table 3. 7, the amount of organics mineralized was related to the amount of Se ions reduced; a greater reduction of Se ions corresponded to a greater extent of organic mineralisation, showing the synergism of the photocatalytic redox reactions. A negligible amount of organic was mineralized in the experiments described in Sets 4-7. This prompted the suggestion that the mere presence of an organic compound was not sufficient for the effective photocatalytic reduction of Se ions.
89 Table 3. 7. Results of the adsorption and photocatalytic reduction of Se anions in the presence of various organic additives. Experimental conditions: 20 ppm initial Se anions concentration, 300 ppmC initial organic concentration, pH 3.0, 1 L test solution, 1.1 gTiO 2/L, N2 purging, 293 K. Adsorption Experiments Photocatalytic reactions (120 minutes)
Set Organic Se anions Organic adsorbed Se ions TOC remaining in
Additives adsorbed (mgC/gTiO2) in Concentration solution (ppmC)
(mg/gTiO2) presence of Se remaining in anions solution (ppm) Se(IV) Se(VI) Se(IV) Se(VI) Se(IV) Se(VI) Se(IV) Se(VI)
1 Formic Acid 6.87 3.63 0.73 2.31 1.98 2.76 287 291
2 methanol 7.71 4.52 ng ng 7.72 10.28 291 296
3 ethanol 7.78 4.45 ng ng 9.02 12.45 292 298 4 Acetic Acid 7.69 4.36 ng ng 12.26 15.03 298 301 5 Salicylic Acid 7.85 4.47 ng ng 12.23 15.78 303 301
6 Sucrose 7.79 4.56 ng ng 12.42 15.45 305 303 7 NIL 7.84 5.52 NIL NIL 12.44 15.02 302 302 Error ± 2.0% ± 5.0% ± 2.0% ± 5.0% ng=negligible
The fastest rate of Se(IV) photoreduction observed when formic acid was used as the hole scavenger could be attributed to a number of factors. Firstly, formic acid adsorbs well onto the TiO2 surface. Unlike the other studied organic compounds, formate ions in fact competed with Se(IV) ions for the TiO2 surface. When the other organic compounds were used (Sets 2-6), their extent of adsorption was found to be negligible and the adsorption of Se(IV) in the presence of those organic compounds was similar to that in the absence of the organic compounds, at an average value of 7.75 mg/gTiO2.
Table 3. 8 presents the results of the extent of adsoprtion for the different organics in the absence and presence of Se(IV). The adsorption results in the absence of Se(IV) showed that all of the studied organic compounds adsorbed significantly onto TiO2 surface, with formic acid adsorbed the most followed by acetic acid and salicylic acid, and a similar amount of methanol, ethanol and sucrose. However, only formic acid was adsorbed
90 significantly in the presence of Se(IV). This could be rationalised by the fact that formic acid is a small molecule and could be ionized into the negatively charged formate ions near the pH 3.0, hence it is able to adsorb onto the positively charged TiO2 surface (as discussed previously in section 3A.3). The other organic compounds were either less likely to form negatively charged ions at the studied pH (Chang, 1994) or/and too big to compete with
Se(IV) and Se(VI) for the TiO2 surface.
Table 3. 8. The comparison between the individual adsorption of the organic additives and the simultaneous adsorption of Se(IV) ions and the organic additives. Experimental conditions: 20 ppm initial Se(IV) concentration, 300 ppmC initial organic concentration, pH 3.0, 1 L test solution, 1.1 gTiO2/L, N2 purging, 293 K. Organic additives Adsorption of the organic additives Adsorption of the organic additive in the
(mgC/gTiO2) presence of Se(IV) (mgC/gTiO2) Sucrose 0.74 - Acetic Acid 1.38 - Salicylic Acid 1.12 - Formic Acid 2.84 0.73 Methanol 0.56 - Ethanol 0.68 -
From the results in Table 3. 7, while the adsorption of the organic was shown to be important for effective photoreduction, the adsorption results observed for methanol and ethanol indicated the importance of other factors. Both methanol and ethanol did not show significant adsorption on the TiO2 surface, however, Se(IV) photoreduction was still observed. Prairie at al (1993) have suggested that the rate of photoreduction was faster when a more easily photooxidised hole scavenger was used. Hence, a possibly higher rate of mineralisation of the organic compounds such as methanol and ethanol (in an anoxic environment), might explain their effectiveness as hole scavengers despite their minimal adsorption.
In the next set of experiments, the rate of photomineralisation of the six selected organic compounds in an anoxic environment and in the absence of Se(IV) ions was studied. The results are summarised in Figure 3. 19. In the anoxic environment, formic acid had the
91 fastest mineralisation rate. Acetic acid also had a fast mineralisation rate but slower than that of formic acid. Only a small amount of methanol, ethanol, and negligible amount of sucrose and salicylic acid were mineralized in the anoxic environment.
18
16 Formic Acid
14
12
10 Acetic Acid 8
6 Carbon OxidisedCarbon (mg) 4 Methanol
2 Ethanol Sucrose 0 Salicylic 051015202530354A0cid 45 Time (min)
Formic Acid Methanol Acetic Acid Sucrose Ethanol Salicylic Acid
Figure 3. 19. The extent of mineralisation of various organic compounds in the presence of nitrogen. Experimental conditions: 15 mg carbon of organic used, 0.15 L test solution, 2.0 gTiO 2/L, N2 purging, pH 3.0.
The greater mineralisation of formic and acetic acids in an anoxic environment can be attributed to their abilities to react with the TiO2 lattice oxygen (Muggli & Falconer, 1999; Muggli & Backes, 2002). In comparison with the mineralisation of formic acid, the lower mineralisation of acetic acid is believed to be due to the formation of intermediate compounds (such as methane and ethane) instead of just carbon dioxide and water (as was the case of formic acid). The importance of oxygen for the mineralisation for methanol, ethanol, sucrose and salicylic acid might explain their limited mineralisation in an anoxic environment. Studies involving the photooxidation of the alcohol propanol found that the presence of gaseous oxygen was crucial for its mineralisation (Larson et al, 1995).
92 Penpolcharoen et al (2001) also observed minimal mineralisation of sucrose in an anoxic environment. They suggested that the low sucrose mineralisation in anoxic environment was due to the formation of intermediates that could block the active surface sites. These intermediates might require oxygen to be further degraded. The importance of oxygen for the mineralisation for methanol, ethanol, sucrose and salicylic acid might explain their limited mineralisation in an anoxic environment. Hence, the photomineralisation results (in the absence of Se ions) presented another positive characteristic of formate ions as an effective hole scavenger. In addition to its ability to compete with Se ions to adsorb on
TiO2 surface, formate ions can be easily mineralized in an anoxic environment.
The results obtained for acetic acid mineralisation were interesting. Despite its high photocatalytic mineralisation in anoxic environment, the photoreduction of Se(IV) or Se(VI) ions was not observed in its presence. As shown previously in Table 3. 7 and Table 3. 8, acetic acid adsorption was negligible in the presence of Se ions. This indicates that to be an effective hole scavenger, the organic compound has to be able to adsorb on the TiO2 surface as well as can be mineralized easily in that environment.
Returning to the results obtained for metahnol and ethanol as hole scavengers, the insignificance adsorption (dark adsorption) of methanol and ethanol in the presence of Se ions and their negiligible mineralisation in an anoxic environment suggested that these organic compounds have another role in assisting the photocatalytic reduction of Se ions. It is suggested that their ability to form reducing radicals by reacting with hydroxyl radicals could play an important role in the photoreduction of Se ions.
The formation of reducing radicals from formate and alcohols by the reactions with hydroxyl radicals has been well discussed in literature (Asmus et al, 1973; Perissinotti, 2001) and has been shown to contribute to the photoreduction of Cd ions (Chenthamarakshan et al, 2000a). Formate ions could form reducing radicals more effectively than the alcohols (Asmus et al, 1973). The redox potentials of the radicals formed from formate, methanol and ethanol with respect to the Se ions are depicted in Figure 3. 20. These reducing radicals have potentials greater than that of the Se(IV)/Se(0)
93 and Se(VI)/Se(0) redox couples and hence are thermodynamically capable of reducing the Se ions.
Potential V (vs. SHE)
·- CO2 /CO2
-1.6
-1.2 · CH2CH3OH · -0.8 CH2OH
E CB -0.4
0.0
0.4
4+ 0 0.8 Se /Se 6+ 0 Se /Se
1.2
1.6
»
EVB 3.2
Figure 3. 20. Relative position of the conduction band and valance band edges for TiO 2 in a pH 3 aqueous medium and selected redox levels (Chenthamarakshan et al, 2000a; Seby et al, 2001).
It is well established that photooxidation of methanol produces intermediates following an oxidation sequence: formaldehyde, formic acid and carbon dioxide (Chen et al, 1999, Muggli et al, 2001). It could hence be argued that given a sufficiently long irradiation time, the formate ions formed, which were found to be strongly adsorbed onto the TiO2 surface in the current study, could lead to the formation of strongly reducing radicals and hence may lead to delayed enhanced reduction rate of the Se ions. This is however not found in our current study. It is important to note that the reactions were being performed in an anoxic environment and hence the oxidation of methanol would be impeded. This has already been
94 shown in the above discussion pertaining to Figure 3. 19. If the oxidation of methanol were successful, the formation of the volatile formaldehyde may halt the oxidation sequence before formic acid was produced, hence limiting the effect of the reducing radicals in the reaction system. Given these reasons, the reaction irradiation time of 120 minutes would be likely to be insufficient for the photooxidation of methanol to formate.
It is interesting to see that even though acetic acid had a greater extent of mineralisation than methanol or ethanol under N2 environment and in the absence of Se(VI) and Se(IV) (refer to Figure 3. 19), no Se ions reduction was observed when it was used as the hole scavenger. This could be attributed to the fact that acetate ions could not compete with
Se(VI) and Se(IV) for adsorption on TiO2 surface and also that it was not capable of forming reducing radicals due to the absence of hydrogen at the a-carbon position.
Similarly, sucrose and salicylic acid did not adsorb onto TiO2 and could not form reducing radicals. From Table 3. 7, the increase in the rate of Se ions reduction in the presence of the different organics follows the order of formate> methanol> ethanol. This is in agreement with another study (Chenthamarakshan et al, 2000a) involving the use of these three organic compounds as the hole scavengers. They explained their observation by relating the photocatalytic activity to an increase in the extent of hydrogen abstraction to form reducing radicals from formic acid, ethanol and methanol.
The photoreduction and adsorption of Se(VI) were found to follow similar trends to those of Se(IV), although less Se(VI) ions were removed from solution compared to the Se(IV) system for the same organic additive. This can be explained by the greater adsorption of Se(IV) ions and the difference in the ionic structure of the two ions as discussed section 3A.3. The extent of Se(VI) photoreduction was found to be higher in the presence of formate compared to methanol and ethanol. It was found that formate ions could competitively adsorb onto the TiO2 surface while the other organic additives could not significantly adsorb in the presence of Se(VI) and Se(IV) ions.
95 To summarise the above findings, it can be deduced that the factors for efficient Se ions photoreduction were adsorption of the organic compounds, ease of mineralisation of the organics and/or the organics’ ability to form reducing radicals. Formic acid was found to be the best organic additive in achieving efficient Se ions reduction. This could be attributed to its ability to simultaneously adsorb with Se(VI) and Se(IV) on the TiO2 surface in the form of formate ion, and its ease of mineralisation, rendering it an effective hole scavenger. In addition, it is also suggested that formate could efficiently form highly reducing radicals by scavenging the hydroxyl radicals formed (a hole scavenging process). For the case of methanol and ethanol, even though they were not significantly adsorbed (Table 3. 8), their abilities to form reducing radicals with hydroxyl radicals resulted in some Se ions photoreduction with efficiencies lower than that of formic acid.
Further studies were carried out for the above systems using formic acid, methanol and ethanol in the photoreduction of Se ions. These involved varying the organic concentrations as well as the reaction system pH. Figure 3. 21 shows the extent of Se(IV) photoreduction in 120 minutes with respect to the different concentrations and types of organic additives used. It was again observed that the increasing order of formic acid> methanol> ethanol was encountered. Increasing the organic concentration resulted in an increase in the amount of Se(IV) reduced. The latter observation could be attributed to the increase in mass transfer of the organic compounds to the TiO2 surface at higher organic concentration, hence resulting in an enhanced hole scavenging and possibly an enhanced formation of the reducing radicals.
96
14.0
12.0
10.0 8.0 6.0 4.0 (mg) irradiation 2.0 300ppmC
min Reduced in Se(IV) 120 150ppmC 0.0 50ppmC Formic Methanol Acid Ethanol Organic Concentration (ppm) Types of Organic Additives
Figure 3. 21. Photocatalytic reduction of Se(IV) in various concentration of formic acid, methanol and ethanol at pH 3.0 in 120 minutes. Experimental conditions: 20 ppm initial Se(IV) concentration, 1 L test solution, 1.1 gTiO2/L, N2 purging, 293 K.
The investigation of Se(IV) and Se(VI) photoreduction at different pH values have been previously discussed in section 3A.3. To enable comparison and discussion in the latter section, it is wished to highlight the occurrence of an optimum pH for which maximum Se(IV) and Se(VI) photoreduction was encountered when formic acid was used as the hole scavenger.
The effect of pH on the extent of Se ions photoreduction in the presence of methanol and ethanol was also carried out. Figure 3. 22a and Figure 3. 22b show the results for adsorption and photoreduction of Se(IV) and Se(VI) ions respectively in the presence of methanol and ethanol in the pH range of 2.2 to 4.0. The adsorption of Se ions in the presence of methanol and ethanol was found to decrease with increasing pH owing to an increase in the net surface negative charge of TiO2. It can also be seen that, for the same pH value, the amount of Se(IV) adsorbed was similar in the presence of methanol and ethanol (Figure 3. 22a). This observation also applied to the Se(VI) ions (Figure 3. 22b) and suggested that both methanol and ethanol did not affect the Se ion adsorption. It has also
97 been found that in the presence of Se ions, negligible methanol and ethanol was adsorbed in the investigated pH range.
15.0 8.0
12.0 6.0 ) 2 9.0 4.0 6.0 (mg/gTiO 2.0 3.0 presenceof organics min irradiation(mg) Se(IV) reductionafter 120 Se(IV) darkadsorption thein 0.0 0.0 2.23.0pH 4.0 Se(IV) Ads (M) Se(IV) Ads (E) Se(IV) Red (M) Se(IV) Red (E)
a. Se(IV) experiments.
15.0 6.0
10.0 4.0 ) 2
5.0 2.0 (mg/gTiO min irradiation (mg) presence of organics Se(IV) reduction after 120
Se(VI) dark adsorptionthe in 0.0 0.0 2.23.0pH 4.0 Se(VI) Ads (M) Se(VI) Ads (E) Se(VI) Red (M) Se(VI) Red (E)
b. Se(VI) experiments.
Figure 3. 22. Comparison of Se ions dark adsorption and photoreduction at various pH and 300 ppmC methanol and ethanol. 20 ppm initial Se(IV) concentrations, 1.1 g/LTiO 2, 293 K, 1L test solution, N2 purging.
98 Upon UV irradiation, the extent of Se ions photoreduction in the presence of methanol was greater than that of ethanol at the range of pH values investigated. This can be explained in terms of the faster rate of hydrogen abstraction to from reducing radicals for methanol compared to ethanol as discussed earlier. The extent of Se(IV) photoreduction was also faster than that of Se(VI). It was also found that the highest amount of Se ions reduced after 120 mins irradiation occurred at the lowest pH investigated (a value of 2.2), while the least Se ions reduced was encountered at pH 5.0. These corresponded to the highest and the least amount of Se ions adsorbed respectively.
The above observation suggests that the role of formic acid in the Se ions photoreduction was different from that of methanol and ethanol. Since formic acid could be adsorbed in the form of formate ions, the amount of formate ions present on the TiO2 surface had to be optimised relative to the Se ions in order to maximise its ability as an effective hole scavenger. Since methanol and ethanol could not adsorb on the TiO2 surface in the presence of Se ions, the role of these organic compounds in the photoreduction of Se ions was mainly through their ability to form reducing radicals. Hence, in the current study, a pH value of 2.2, which resulted in the best Se ion adsorption, would produce the most efficient Se ions photoreduction.
3B.4 Conclusions
The photocatalytic reduction of Se(VI) and Se(IV) to elemental Se was performed using sucrose, acetic acid, salicylic acid, formic acid, methanol and ethanol. Photoreduction was only observed in the presence of formic acid, methanol or ethanol and the rate of Se ions photoreduction was found to be in the order formic acid> methanol> ethanol. The reason for formic acid being the most efficient hole scavenger is postulated to be due to its ability to compete with Se ions for TiO2 surface, its fast mineralisation rate and its ability to form reducing radicals quickly. Even though methanol and ethanol did not adsorb significantly in the presence of Se ions and were not easily mineralized, their presence enabled the reduction of Se ions due to their ability to form reducing radicals. Increasing the
99 concentration of the organic additives resulted in greater extent of Se(IV) reduction for all the three organic compounds used. When formic acid was used, optimum pH values were encountered for the Se(VI) and Se(IV) photoreduction. When methanol and ethanol were used as the hole scavenger in the pH range of 2.2 to 4.0, the greatest extent of Se ions photoreduction was encountered at pH 2.2 which corresponded to the highest amount of Se ions adsorbed. This shows that methanol and ethanol have a different role compared to formic acid in the photoreduction of Se ions.
100 3C. Kinetic Modelling and Reaction Mechanism of Se(VI) Photoreduction
3C.1. Introduction
Numerous works have been done to model heterogeneous photocatalytic reactions, based on the classical heterogeneous catalysis mechanism of Langmuir-Hinshelwood (LH) reaction mechanism. However, photocatalytic reactions are very complex and are affected by a number of operating parameters, such as catalyst dosage, light intensity and temperature, of which the LH model cannot fully predict. Some studies have proposed separate models for each operational parameter to elucidate their effects on the photocatalytic oxidation rate (Lea & Adesina, 1998; Chen & Ray, 1999).
Most of the published kinetic modeling studies have focused on the photooxidation of organic and inorganic pollutants with dissolved oxygen gas as the electron acceptor. Very little work on kinetic modeling has been done on the photoreduction of anionic toxic pollutants, such as chromate, nitrate and selenate, which involve the simultaneous adsorption of the pollutant and organic scavenger. Gimenez et al (1996) proposed a model for catalyst deactivation in the photocatalytic reduction of chromate in the presence of oxygen. Alam and Montalvo (1998) developed a kinetic model based on various electron and hole consuming reactions involving chromate as the anionic pollutant and salicylic acid and oxygen as the electron scavengers. Kikuchi and Sakamoto (2000) proposed that in the photocatalytic reduction of Se(VI) ions, when all the Se(VI) ions were reduced to elemental
Se, no species in the solution could capture electrons from the conduction band of the TiO2 while photogenerated holes continue to be captured by the organic reductant. This resulted in accumulated electrons in the conduction band, further reducing the elemental Se on the
TiO2 surface to hydrogen selenide (H2Se). A model based on electron accumulation was
101 proposed to explain their results. No work has been previously carried out on the kinetic modeling of Se(VI) photoreduction in the presence of formate ions as the hole scavenger.
From the investigations on Se(VI) photoreduction in Section 3A, it was found that an optimum molar adsorption ratio of formate:selenate, closely correlated to the stoichiometric ratio of 3:1, played a significant role in optimising the photocatalytic reduction of Se(VI). Following on from these experimental findings, it was aimed to formulate a kinetic model and propose a reaction mechanism of Se(VI) photoreduction rates in the presence of various formic acid concentrations based on the adsorption of both Se(VI) and formate ions. The proposed model will be used as the basis for optimisation studies which will be carried out to verify the aforementioned ratio necessary to achieve optimum Se(VI) photoreduction rate.
3C.2 Equipment and Procedure
In the current study, the catalysts and the reagents, the reactor, the analytical equipments and procedure for adsorption and photoreduction have been described in section 3A.1. All concentrations are expressed as mM (not ppm as in the previous section). The Se(VI) photoreduction rates were estimated by zero(not sure about this- I guess it could be either)- order rate law. The experimental data was then fitted in accordance to the proposed models by the mathematical software MicroMath Scientists by the least-squares methods.
102 3C.3 Results and Discussions
Kinetic Modelling of Se(VI) Photocatalytic Reduction
In the current investigation, the pH was maintained at 3.5. The competitive adsorptions of - both the negatively charged Se(VI) and HCOO ions onto the positively charged TiO2 surface at pH 3.5, and at the initial Se(VI) concentrations of 0.256 mM (20 ppm) and 0.512 mM (40 ppm), have been shown in Figure 3. 7. Se(VI) adsorption in the presence of formate ions was modeled under the consideration that each Se(VI) ion possesses two negative charges and hence one Se(VI) ion is capable of being adsorbed onto two positively charged sites on the TiO2 surface. The derivation of the Se(VI) Langmuir adsorption isotherm based on the above reasons gives the following equation 3-7 (details of the derivation are included in the Appendix 1):
éæ+CK )1( 2 öæ+CK 2 öæ+CK )1()1( 2 öù C =´êç25.0 Ct +FF ÷±ç4Ct +FF ÷çFF ÷ú(3-7) Sess ç÷ç÷ êçCK ÷CK CK ú ëèSeSe øèSeSe øèSeSe øû
where CSess represents the Se(VI) ion adsorbed onto the two sites in the TiO2 surface, KSe - and KF are the adsorption constants for Se(VI) and HCOO , Ct represents the total available sites on the catalyst surface, and CSe and CF are the residual Se(VI) and formic acid concentrations at equilibrium respectively.
Equation (3-7) indicates the possibility of 2 sets of solutions as shown by the “±” sign in the equation. The experimental data was successfully fitted when the equation with the “–” sign was used. A graphical representation of fitting the Se(VI) adsorption to Equation (3-7) is presented in Figure 3. 23. The values for the fitted parameters are given in the figure captions. The fitting results show that equation (3-7) was able to describe the adsorption of Se(VI) in the presence of formate ions successfully. The fitted parameters were determined
103 to have uncertainties of less than 5% of the regressed values. However, these parameters have no physical meaning and hence are not used to describe the surface processes.
0.08 ) 2
0.06
0.512 mM Se(VI) 0.04
0.02 0.256 mM Se(VI)
Se(VI) adsorbed0 (mmol/gTiO.00 0102030 HCOOH Concentration (mM)
Figure 3. 23. Comparison of experimental and modeled data for Se(VI) adsorption according to 2 -1 -1 Equation (3-7). R =0.996, KSe=298 mmol , KF=0.079 mmol and Ct=0.071 mmol/gTiO2.
Subsequently, formate ion adsorption in the presence of Se(VI) was derived. The derivation is given in the Appendix 1. This derivation assumes the adsorption of one formate ion onto one positively charged site on the TiO2 surface in the presence of Se(VI) ions, assuming one Se(VI) ion is adsorbed onto two active sites:
CK 2 C =FF é(-CK -)±()11 ++4 CKCCK ù (3-8) FS ëêFF FF SeSet ûú 2 CK SeSe
where CFs represents the formate ion adsorbed onto one site on the TiO2 surface.
Figure 3. 24 shows the comparison of the experimental data with the modeled data from Equation (3-8) when the equation with the “+” sign was used. The model provides a reasonable representation of the data with the average percentage difference between the experimental and calculated data being about 13.3 %. The fitted parameters were determined to have uncertainties of about 10 % of the regressed values. Similar to the
104 earlier discussion on Se(VI) adsorption, these parameters are only apparent and have no physical meaning.
0.25 ) 2 0.20 0.256 mM Se(VI)
0.15
0.10
0.05 0.512 mM Se(VI)
0.00 Formate adsorbed (mmol/gTiO 0102030
HCOOH Concentration (mM)
Figure 3. 24. Comparison of experimental and modeled data for formate adsorption according to 2 5 -1 -1 Equation (3-8). R =0.97, KSe=8.5x10 mmol , KF=3.9 mmol and Ct=1.2 mmol/gTiO2.
Assuming that surface reaction is the rate-limiting step and that the adsorption models represented by equations (3-7) and (3-8) are reasonably accurate, the following expression (equation 3-9) was derived to describe the Se(VI) photoreduction rate using the Langmuir- Hinshelwood reaction mechanism (Fogler, 1992):
'=CCkr FSerxn sss (3-9) ì 2 2 2 ü 1 ïé()1+CK ùé()1+CK ùé()1+CK ùï =íê2Ck +FF ú-ê4C +FF úêFF úý rxn 2 t CK t CK CK îïë SeSe ûë SeSe ûëSeSe ûþï
ìCK 2 ü íFF é(-CK -)+()11 ++4 CKCCK ùý ëêFF FF SeSet ûú î2 CK SeSe þ
where r’ is defined as the initial Se(VI) photoreduction rates and krxn is the apparent rate constant. Other parameters have been previously described.
105 Figure 3. 25 shows the comparison of the Se(VI) photoreduction rates at various formic acid concentrations (as shown previously in Figure 3. 11) with the modeled data for equation (3-9). The model could fit the data reasonably well. The average percentage error between the experimental and calculated data from the model was determined to be about 15%. Again, it is suggested that these fitted parameters only represent apparent constants with no physical meaning. The optimum formic acid concentration reflected in the experimental data was also shown in the model.
4.0 3
3.0 ) 2
2.0
mmol/min.gTiO1.0 Se(VI) Photoreduction rate (x10
0.0 0102030 HCOOH Concentration (mM)
Figure 3. 25. Comparison of experimental and modeled data for Se(VI) photoreduction rate according to Equation (3-9). □: Experimental rate and —: Modeled rate for 0.256 mM initial Se(VI) concentration, ◊: Experimental rate and ---: Modeled rate for 0.512 mM. initial Se(VI) concentration. 2 -9 -1 -1 R =0.96, krxn=8.16x10 gTiO2/mol.min, KSe=1320.5 mmol , KF=42.5 mmol and Ct=1320.5 mmol/gTiO2.
The deviation of the newly derived rate model from the experimental rate may be explained as follows. The LH reaction mechanism assumes that the reactive sites are uniform and that the surface reaction is the rate-determining step. The former assumption means that all active surface sites have the same attraction for the solute. However, the formation of Se particles as the photoreduction reaction proceeded is believed to have resulted in changes in the surface properties. This would hence affect the dynamic adsorption of Se(VI) and formate ions onto the TiO2 surface and result in the nonuniformity of the surface. When
106 there were more Se deposits on the TiO2 surface as the Se(VI) photoreduction proceeds, the desorption of Se might become rate-limiting. As observed from the experiments, Se did not desorb from the surface until the Se(VI) was exhausted from the solution. Hence, the presence of Se deposits on the TiO2 surface would mostly likely block the reactive sites and subsequently affect the photoreduction rate. Another phenomenon unaccounted for in the above model was the change in the electrophoretic properties for illuminated TiO2 (Boxall & Kelsall, 1991a; 1991b). This would again affect the dynamic adsorption of the substrate onto the TiO2 surface during irradiation. The error in the adsorption models would inevitably contribute to the deviation as well.
- The importance of the molar ratio of Se(VI) and HCOO adsorbed onto the TiO2 surface to gave the optimum Se(VI) photoreduction rate was highlighted in the section 3A. An optimum Se(VI) photoreduction rate was encountered when a molar adsorption of HCOO- and Se(VI) ions in the ratio of approximately 3:1 on the TiO2 surface was achieved. This experimental observation is also reflected by the model (Equation 3-9). Figure 3. 26 shows the Se(VI) photoreduction rate as predicted by the new model as a function of the molar - ratio of HCOO :Se(VI) ions adsorbed onto the TiO2 surface. From Figure 3. 26, it can be seen that the optimum photoreduction rate corresponded to the ratio of 3.3+0.2 and 2.6+0.1 for 0.256 and 0.512 mM initial Se(VI) concentration respectively. These ratios are in close agreement to the stoichiometric ratio of 3:1 HCOO-:Se(VI) as suggested earlier, indicating the reasonable representation of the model to the experimental data.
As part of the kinetic studies, the effect of catalyst loading on the Se(VI) photoreduction rates was also investigated. The experiments were conducted with a catalyst loading -1 ranging from 0.25 to 2.2 gTiO2L . The results are shown in Figure 3. 27. Increasing the catalyst loading increased the reduction rate constant until an optimum catalyst loading was attained at about 1.1 g/L. This observation is typical in photocatalysis and its reasoning has been discussed in section 2.1.4 under “Effect of Catalyst Loading”.
107 4.5
4.0
3.5 ) ) 2 3.0
2.5
2.0
1.5 (mmol/min.gTiO
1.0
Modeled Se(VI) photoreduction rate rate photoreduction Se(VI) Modeled 0.5
0.0 04812 Formate:Selenate ratio
Figure 3. 26. Optimisation of Equation (3-8) with respect to molar adsorption ratio of 3:1 at HCOO - :Se(VI). —: Modeled rate for 0.256 mM initial Se(VI) concentration, ---: Modeled rate for 0.512 mM. initial Se(VI) concentration.
3.5
3.0 3 3
2.5
2.0
1.5 mmol/min ) 1.0
0.5 Se(VI) Photoreduction Rate (x10 Rate Photoreduction Se(VI) 0.0 0.00.51.01.52.02.5
TiO2 Loading (g/L)
Figure 3. 27. Influence of catalyst loading on Se(VI) photoreduction rate. Experimental conditions: pH=3.5, 1 Litre test solution, [Se(VI)]0=0.256 mM, [HCOOH]0=16.7 mMC, N2 purging, 293K.
108 Although the catalyst may not be explicitly regarded as a reactant, the non-linear trend of the experimental data shown in Figure 3. 27 may be fitted to the model shown in the work of Lea & Adesina (1998):
kC (3-10) r=C0 w+2 ()C0
where kC is the rate constant for the catalyst, C0 is the catalyst concentration, w is a light attenuation factor and the squared denominator indicates interparticle interactions.
The regression based on the Equation (3-10) yielded an excellent agreement with a -1 -2 correlation coefficient of 0.99 while w was evaluated as 0.93 gTiO2L and kc was 1.3x10 -2 -1 mol gTiO2L min .
From the use of all the information from the foregoing discussion, an empirical composite rate equation, which describes the competitive adsorption of the reactants and incorporating the effect of catalyst loading, was derived as follows:
ìé 2 ùé 2 ùé2 ùü 1 ï()1+CK FF ()1+CK FF ()1+CK FF ï =kr '' rxn íê2Ct + ú-ê4Ct + úêúý 2 ïëCK SeSe ûëCK SeSe ûëCK SeSe ûï î þ(3-11) ì ü CK FF é 2 ù í(-CK -)+()11 ++4 CKCCK ý 2 CK ëêFF FF SeSet ûú îSeSe þ where
kkC k'= rxn C0 rxn 2 ()w + C0
krxn′ is the new reduction rate constant and it is a function of catalyst loading.
109 Proposed Mechanism
From the preceding investigation, it was found that the photocatalytic reduction of Se(VI) in the presence of HCOOH as the organic scavenger was strongly dependent on the adsorption of both Se(VI) and HCOO- ions. These ions adsorbed competitively on the same site. The following sequence of elementary reactions is proposed to explain the mechanism of photoreduction of Se(VI):
1. Complete dissociation of sodium selenate:
+2- 2SeONa 4®2+SeONa 4 (3-12)
2. Ionisation of Se(VI) and formic acid (This is a pH and concentration dependent process). 2- At pH 3.5, selenate ions exist as the unprotonated (SeO4 ), while some of the formic acid is ionized to formate ions. (3-13) HSeO - + +® SeOH 2- 4 4
HCOOH H + +Û HCOO- (3-14)
3. Competitive adsorption of ions onto hydrated TiO2 sites (sites denoted by *):
2- - *)2(2 (3-15) SeO4 *2 Û+ SeO4
- - HCOO * Û+ HCOO * (3-16)
- + 4. Photogeneration of electrons (e cb) and holes (h vb):
(3-17) TiO ¾®¾hv + he +- 2 vbcb
110 5a. Reduction of adsorbed Se(VI) ions by electrons:
- *)2(2 -+ *0 (3-18) SeO4 68 cb +®++ 4 2OHSeeH
5b. Oxidation of adsorbed HCOO- ions by holes (or hydroxyl radicals):
HCOO 2 +*- HCOh + *++®+ vb 2 (3-19)
6. Recombination of electrons and holes: (3-20) +- +®+ heatTiOhe vbcb 2
By balancing the electrons and holes generated in steps 5a and 5b (Equation 3-18 and 3- 19), it could be seen that when 3 moles of HCOO- ions are oxidised, 1 mole of Se(VI) ions is reduced, giving the stoichiometry of the redox reactions as 3:1. Hence, it is anticipated that this corresponds to the optimum molar adsorption ratio.
As mentioned previously, reducing formate radicals could form and take part in the reduction of the Se(VI) ions. However, the detailed reaction mechanism pertaining to this step could not be elucidated at this stage.
3C.4 Conclusions
New Se(VI) and formic acid adsorption models have been proposed and were derived assuming that one Se(VI) ion is adsorbed onto two active sites on the TiO2 surface and the competitive adsorption of both Se(VI) and formate ions. The models fitted the experimental adsorption data with correlation coefficients of 0.97 and 0.99 respectively. The use of these
111 adsorption models enabled the derivation of a rate equation on the basis of the LH reaction mechanism. The model confirmed that the optimum photoreduction rate of Se(VI) was achieved at formate-to-selenate molar adsorption ratios of 3.3+0.2 and 2.6+0.1 for 0.256 and 0.512 mM initial Se(VI) concentration respectively, supporting the postulate for an optimum formate to selenate molar adsorption ratio of 3:1. A composite rate law incorporating the effect of catalyst loading was also derived. On the basis of the above model, a reaction mechanism was proposed. This reaction mechanism described the redox reaction based on the competitive adsorption of both Se(VI) and HCOO- ions and the subsequent scavenging of the photogenerated electrons and holes on the catalyst surface.
112 Chapter 4. Synthesis of (nano) Se- compounds by Photocatalysis
4.1 Introduction
Before the last decade, photocatalysis has been demonstrated to be mainly useful in gas and water remediation. In the recent years, alternate applications, including self-cleaning building tiles and photoinduced superhydrophilicity (anti-fogging) show that the unique properties of TiO2 photocatalysis can bring about many interesting applications to our daily lives (Fujishima et al, 2000a and 2000b). In the field of materials science, photocatalysis has also been used in the preparation of surface-modified TiO2 by noble metal photodeposition. In many cases, the modified photocatalysts have demonstrated enhanced rate of pollutant degradation.
This chapter attempts to demonstrate the alternate use of TiO2 photocatalysis as a particle synthesis technique. The chapter is divided into 2 sections. Section 4A first gives a brief theory on n-p composite conductors, which result when elemental Se is deposited on TiO2 particles. It then goes on to examine the Se particles formed on the TiO2 surface under different experimental conditions. A brief feasibility study on the potential uses of the Se- 2+ TiO2 particles for toxic Cd ion removal is then carried out. Section 4B investigates the effect of Ag-metal modified TiO2 on the photocatalytic reduction of Se ions and also demonstrates the possibility of forming metal selenide compounds from the Se ions photoreduction by Ag-TiO2.
113 4A. The Formation of Nano-Sized Selenium- Titanium Dioxide Composite Semiconductors by Photocatalysis
4A.1 Introduction
The abundant uses of elemental selenium, especially due to its unique property as a photoconductor, coupled with the emerging trend of “tailoring” new materials at atomic and molecular levels, could provide considerable opportunity for the use of nanoparticle Se in photoelectric devices with novel characteristics. The numerous benefits in nano-sized materials have been extensively described in many literatures (Beydoun et al, 1999; Choe et al, 2000; Hagfeldt & Gratzel, 1995; Loughran, 2001; Lu, 2001; Umehara et al, 1997).
Se is a p-type semiconductor (Streltsov et al, 2002) while TiO2 is an n-type semiconductor
(Yoneyama et al, 1979). The formation of Se on TiO2 results in the contact of a p-type with an n-type semiconductor, forming a p-n junction at the interface. The difference in the concentration of electrons and holes at the interface induces an initial diffusion of holes towards the n-region and electrons towards the p region. The diffusing electrons recombine with holes in the p-region while the diffusing holes capture electrons on the n-side. Since the p-region is injected with extra electrons, it becomes negative while the n-region becomes positive. This sets up an electric field (E) directed from n to the p region, creating an energy barrier (eV) for electron transfer from n to p region (Dalven, 1990a). The region across which the electron energy (related to the potential) varies is known as the depletion zone or space charge region (W). The width of the depletion zone depends on the concentration of n and p impurities. This accumulation of charges at the junction and the corresponding electron energy profile are shown schematically in Figure 4. 1.
114
p-n n-type junction p-type
W
E E CB eV
EVB
Electron Energy
Distance
Figure 4. 1. Band diagram showing the electron energy as a function of distance of a p-n junction (Diagram redrawn from Dalven, 1990a).
The application of a forward bias on the p-n junction will decrease the energy barrier and reduce the width of the space charge layer, facilitating electron transfer from n to p (Dalven, 1990a). The irradiation of the n-p junction is akin to the action of applying a forward bias. Upon irradiation, the strong internal field of the space charge region will rapidly sweep the carriers across the region, thus resulting in a net current. The net effect of the photogeneration of charge carriers is that a constant current (a negative, or electron current) is created that flows from the n to p regions of the composite semiconductors (Thiel, 1999). Figure 4. 2 depicts the application of a forward bias on a p-n junction.
115 W
n-type p-type
E Ea
Figure 4. 2. Effect of applying a forward bias on a p-n junction. Electric field Ea resulting from the applied forward bias, E is the built-in electric field at the junction and W is the width of the space charge region (Dalven, 1990a).
The formation of nano-sized Se particles on TiO2 could present a new type of application for the p-n dual-semiconductor system, such as nano-sized photodiodes. Such material could also provide a further depth in the investigation of enhancing the optical and electronic properties of TiO2 akin to the composite semiconductor systems as mentioned in the literature review.
In the previous chapters, the removal of Se(VI) and Se(VI) ions from solution was studied.
In this section, emphasis will be given to the formation of elemental Se onto TiO2. The morphology of the Se deposited under different conditions, such as when using different Se ions precursors (namely Se(VI) and Se(IV)) and pH will be investigated.
4A.2 Equipment and Procedure
Catalyst and Reagent
Catalysts and chemicals relevant to this section have been described in section 3A.1. In addition, cadmium perchlorate is used as the cadmium ion source.
116 Procedures for Selenium Photodeposition
Procedures described in section 3A.2 are adhered to with minor modifications: 0.5 L of deionized water was used and 0.5 gTiO2/L was added. The photon flux into the reactor for the experiments performed was determined to be 3.17 mmol photon/sec by chemical actinometry procedures described in Parker (1953a &1953b).
Procedures for Photocatalytic Removal of Cadmium Ions
After the formation of Se-TiO2, UV irradiation was halted temporarily by closing the shutter of the lamp. The pH was adjusted to 7 and then the required concentration of Cd ions (30 ppm in the present study) was added to the suspension. The suspension was stirred in the dark for 30 minutes to allow for adsorption of Cd ions to take place. A sample was taken after the dark adsorption and then the suspension was re-illuminated. Samples were taken at 10, 20 and 30 minutes interval. The procedure for the photocatalytic Cd ion removal by Degussa P25 was similar. For blank experiment, Cd ions solution at pH 7 was irradiated in the absence of a photocatalyst.
Procedures for Se ion Adsorption
The adsorption procedures as described in section 3A.2 were followed.
Analysis
Analysis of the total Se concentration (Se(VI) or Se(IV)) in the filtrate, the Cu(II) ions remaining in the Cu(II) trap and the formic acid concentration were the same as described in section 3A.2. Quantitative analysis of elemental Se was performed by digesting the Se particles using concentrated HNO3 followed by determining the Se concentration by ICP-
117 AES. The Se-TiO2 particles were viewed under a Transmission Electron Microscope (TEM-Phillips CM200). The crystallinity of the deposited selenium particles was determined by X-ray Diffraction (Siemens D5000 Diffractometer) and Energy Dispersive X-ray analyser (EDX) in-built in the TEM. Se elemental mapping was also performed by EDX. Possible signals from precursors were minimised by washing the particles by deionized water and 0.1 mM NaOH to alter the surface charge of TiO2 to a negative value such that the negatively charge Se ions will be desorbed. The reported size of the largest Se particles was estimated by selecting the largest Se particles seen under 3 different TEM images and measuring their diameters. UV-Vis absorbance was performed a using Cary 500 UV-Vis spectrophotometer.
4A.3 Results and Discussions
The formation of elemental Se particles on the TiO2 was the result of reduction of Se(VI) or
Se(IV) ions to Se following equations 2-24 and 2-25. The Se particles formed on the TiO2 particles from the photoreduction of Se(VI) were different to those formed using Se(IV) as precursor. The formation of the Se particles from Se(VI) photoreduction will be discussed first.
Se Particle Characterisation
Characterisation of the Se formed on TiO2 (using Se(VI) as the Se precursor) was performed first. X-ray diffraction was performed to investigate the crystallinity of the Se formed. The results indicated the absence of crystalline Se. The Se-TiO2 particles were then viewed under Transmission Electron Microscope and the image is shown in Figure 4. 3. Two different types of particles can be distinguished from Figure 4. 3; the smaller, irregular shaped particles and the bigger, round particles. The latter was identified as Se by EDX while the former were the TiO2 crystals. When the Se particles were subjected to electron diffraction analysis by EDX, the diffraction pattern as shown in Figure 4. 4 were observed.
118 The diffraction pattern resembles the Kikuchi line, which is the result of inelastic scattering due to thick crystals (Loretto, 1994), indicating that the Se particles formed are comparatively thicker than the TiO2 particles. Due to the colour observed and the crystallinity nature of the particles, the deposited Se nanoparticles are suggested to be the red monoclinic form. This was contradictory to the suggestion by Sanuki et al (1999), which did not observe any Se crystals using XRD analyis XRD. The absence of crystalline Se from the X-ray Diffraction analysis could be the result of too little selenium particles presents in the sample relative the amount of TiO2 particles.
Se
TiO2
100
Figure 4. 3. TEM image of Se particles on TiO 2 particles
119
Figure 4. 4. Electron diffraction pattern by EDX of a Se particle deposited during the photoreduction process.
Se Formation Mechanism from Se(VI) Photoreduction
TiO2 particles normally exist in solutions as aggregates with sizes ranging from 300 to 500 nm. Due to its aggregate form, it is most likely that the Se particles are deposited and not coated onto the TiO2 particles. From Figure 4. 3, it is also seen that the Se particles are of different sizes and formed discretely as individual particle onto the TiO2 particles.
The formation of larger and round Se onto the TiO2 is postulated as follows. When elemental Se particles are deposited onto TiO2, they become the reduction sites for Se(VI) ions, making the Se particles grow spherically. This requires the initial Se deposits onto
TiO2 to be rich in electrons in order to act as the reduction sites for the Se(VI) ions. The accumulation of electrons in the Se particles can be explained by the phenomenon of p-n junction under irradiation. When the n-type TiO2 particles are brought into contact with an electrolyte and upon prolonged irradiation, the accumulation of electrons in the interior could be resulted as described in the literature review. Likewise, in the case of a p-type semiconductor (Se in this case) brought into contact with an electrolyte followed by irradiation, electrons migrate to the surface while the holes are transferred to the bulk of the semiconductor, leading to accumulation of the positively charged holes within the Se interior.
120 It is postulated that the accumulation of electrons in the interior of TiO2 and the accumulation of holes in that of Se as mentioned above could set up an electric field directed from the bulk of Se towards the bulk of TiO2. This is akin to applying a forward bias to a p-n junction. Under such influence, electrons migrate from TiO2 to Se, hence making the Se particles rich in electrons, resulting in the reduction of Se(VI) ions on the Se particles and consequently the growth of Se particles. Being a photoconductor, Se’s conductivity increases with irradiation (Dalven, 1990b). The increased conductivity resulted in the even distribution of charge in the Se particles, causing a uniform reduction of Se(VI) onto the Se particles. This could explain the sphericity of the Se particles formed.
After the Se(VI)/TiO2 slurry was irradiated for about 270 minutes, black precipitates began to form in the Cu(II) scrubber due to the formation of CuSe, indicating the generation of
H2Se gas. Figure 4. 5 shows the disappearance of Se(VI) ions, the H2Se generation and the
HCOOH oxidation monitored at progressing irradiation time. The onset of H2Se generation was observed at about 270 minutes when the Se(VI) ions were almost exhausted.
10 90
75 8 60 6
45 4 30
Se Generation (mg) 2 2 15 H (mg) Oxidised HCOOH orSe(VI) Disappearance 0 0 0100200300400 Time (min) Se(VI) H2Se HCOOH
Figure 4. 5. Experimental data showing Se(VI) reduction, H2Se generation and formic acid oxidation. Experimental conditions: Initial Se(VI) concentration: 20 ppm, initial formic acid concentration: 300 6+ 0 ppmC, irradiation time: 420 minutes, pH 3.5, N2 purging, 0.5 gTiO2/L. Quantum yield for Se /Se : 0.18%.
121 Figure 4. 6 shows the flatband potentials of the TiO2 and Se semiconductors after UV irradiation and the reduction potential of the Se/Se2- and Se6+/Se0 couples. Since the Se/Se2- couple is near or above the TiO2 conduction band, the reduction of Se to H2Se is thermodynamically unfavourable. However, from our observation as discussed previously,
H2Se was still generated. This observation hence warrants further explanation.
eV vs SHE Se p-type
-2.0 Eg=1.95eV e-
-1.0 2- - Se/Se e 0 0 E =-0.4eV
- 6+ 0 e + Se /Se h 0 1.0 E =0.9eV
2.0
3.0 Oxidation of HCOOH by h+ holes TiO2 n-type Eg=3.2eV
Figure 4. 6. Energy diagram of Se-TiO2 system (pH=4) showing the position of the flatband potentials and the bandgap energy of TiO2 (n-type) and Se (p-type) upon UV irradiation. Bandgap energy (E g): Eg-TiO2=3.2eV (Hagfeldt & Gratzel, 1995; Tada et al, 1998), Eg-Se=1.95eV (Streltsov et al, 2002). Conduction band potential (CB): CBTiO2=-0.3eV (Tada et al, 1998; Chenthamarakshan et al, 2000b), CBSe=-1.65eV (Streltsov et al, 2002).
Kikuchi and Sagamoto (2000) have proposed that the accumulation of electrons in the bulk of the TiO2 raised the conduction band of TiO2, making it reducing enough to cause the further reduction of Se to H2Se. Other studies have reported experiments which showed that an accumulation of electrons causes the flattening of the conduction band, hence reducing the width of the space charge layer and facilitating the transfer of electrons from the interior
122 to the surface (Memming, 1990). Our proposed mode of electron transfer from TiO2 to Se (as a result of the presence of a forward bias on a p-n junction) might also be used to explain the further reduction of Se to H2Se. As seen in Figure 4. 6, the reduction potential of Se to H2Se lies at -0.4 eV, which is within the bandgap of Se semiconductor but not within that of TiO2. The Se reduction to H2Se by the Se-photogenerated electrons is highly favourable in terms of thermodynamics. From Figure 4. 5, it could also be seen that HCOOH was continuously oxidised even though Se(VI) ions have been exhausted from the suspension. It is hence suggested that even when all the Se(VI) ions in the system were completely reduced, electrons-hole pairs were still being generated. It is proposed that while the holes were being captured for the oxidation of HCOOH, the electrons were being transferred from the TiO2 to the Se particles under the previously proposed mode of electron transfer. This results in the accumulation of electrons in the Se particles, facilitating its self-reduction to H2Se gas by the Se-generated electrons.
It should be noted that the photogenerated electrons transferred from the TiO2 to Se particles do not have sufficient reducing power for the further reduction of Se since the 2- conduction band of TiO2 is near or below the reduction potential of the Se/Se couple. If the TiO2-transferred electrons have sufficient reducing power to reduce Se to H2Se, the generation of H2Se would have been observed at the early stage of the reduction, not when the Se(VI) ions were exhausted. It is hence postulated that the Se-photogenerated electrons are responsible for the further reduction of Se to H2Se due to the position of the Se conduction band (-1.65 eV) with respect to the Se/Se2- couples (reduction potential = -0.3 eV). However, it is also of interest to know which of the TiO2 or Se-generated electrons are responsible for the reduction of Se(VI) to Se. It can be seen from Figure 4. 6 that the redox potential for the Se(VI)/Se(0) couple is not within the bandgap of the Se semiconductor and hence the Se(VI) reduction by the Se-photogenerated electrons is not feasible. It is hence postulated that the TiO2-photogenerated electrons are responsible for the Se(VI)/Se(0) photoreduction. The dynamics of these electrons are yet to be elucidated at this stage.
Another possibility proposed for the formation of Se particles and the generation of H2Se after the exhaustion of Se(VI) is given as follow. From equations 2-23 and 2-24, it was
123 possible that Se(VI) was reduced to the intermediate Se(IV) and then to elemental Se. Therefore, the intermediate Se(IV) could react Se2- according to the Equation 4-1 to form Se particles (Bouroushian et al, 2000):
0 eOH 32 2 SeSeHS +®+ 332 2OH (4-1)
It is hence suggested that the Se particles could be formed from the direct photoreduction of the Se anions to elemental Se and also by reaction 4-1. The generation of H2Se only upon exhaustion of the Se anions could be explained even if the Se particles were formed via the second suggested pathway. When there was no more Se anions in the solution, reaction 4-1 could no longer proceed, hence the Se2- formed would proceed to form H2Se gas via the electrons photogenerated in the elemental Se discussed above.
Upon inspection of Figure 4. 3 again, the size of the Se particles formed on the TiO2 particles is observed to be non-uniform. Some particles seem to have “outgrown” others. The Se particles are also formed discretely on certain sites instead of uniformly dispersed onto the TiO2. The non-uniformity of UV irradiation falling onto the TiO2 particle within the photoreactor may explain the above observation. The irradiation employed was a collimated beam from the side of the reactor and hence was not uniformly irradiated throughout the suspension. It was hence possible that some Se particles were formed quicker than others when the TiO2 particles were first exposed to UV radiation. Once a Se particle was formed, it became the growing centre due to its higher electron density, resulting in a faster growth rate and hence a bigger particle.
Se Formation Mechanism from Se(IV) Photoreduction
The formation of the Se from Se(IV) photoreduction was also studied. When Se(IV) was used as the precursor, the colour of the suspension changed from white to orange-pink, indicating the formation of elemental Se. Upon exhaustion of Se(IV) ions, H2Se was also
124 generated. This shows that the reduction mechanism for Se(VI) and Se(IV) to elemental Se is similar.
However, when the particles from the Se(IV) photoreduction experiments were viewed under TEM, a significant difference in the morphology of Se formed was observed compared to that when Se(VI) was used as the precursor. Figure 4. 7 shows the Se mapping results (Figure 4. 7c and Figure 4. 7d) with the corresponding TEM images (Figure 4. 7a and Figure 4. 7b) with Se(IV) and Se(VI) ions used as the precursors, respectively. As seen from the TEM images, no Se particles were evident when Se(IV) ions were used as the precursors. Comparing Figure 4. 7c and Figure 4. 7d, Figure 4. 7c shows signals of Se dispersed evenly among the TiO2 particles while Figure 4. 7d shows signals of Se concentrating at certain regions, indicating the presence of bigger, round Se particles of varying size (with some overlapping of particles shown) forming discretely on some sites.
It was postulated that the difference in the morphology of the Se particles formed using the different Se precursors was due to their different adsorption mechanism on TiO2. The adsorption of Se(IV) and Se(VI) onto TiO2 at various pH values is shown in Figure 4. 8. From this figure, it could be seen that both the Se(IV) and Se(VI) adsorptions increased with decreasing pH, attributed to the increasing net positive charge as pH increases. However, the adsorption of Se(IV) ions was found to be higher than that of Se(VI) ions due to the difference in ionic structure of the two ions as explained in Section 3A.3. With less Se(VI) ions adsorbed initially, more Se(VI) ions were available in the solution to grow onto the existing Se particles, resulting in the formation of larger and spherical particles.
125
100 nm 100 nm
(a) (b)
100 nm 100 nm
(c) (d)
Figure 4. 7. a and b show the TEM images of Se-TiO2 particles with Se(IV) and Se(VI) as precursors respectively. c and d show the corresponding Se mapping by TEM with Se(IV) and Se(VI) as precursors respectively. Experimental conditions: Initial Se(IV)/Se(VI) concentration: 20 ppm, initial formic acid concentration: 300 ppmC, irradiation time: 90 minutes, pH 3.5, N 2 purging.
9 8 7
) 2 6 5 4
(mg/gTiO 3 2
Se Species AdsorbedSeSpecies 1
0 2.23.04.05.0 pH Se(IV) Se(VI)
Figure 4. 8. Adsorption of Se(IV) and Se(VI) onto TiO2 at various pH. Experimental conditions: Initial Se(IV)/Se(VI) concentration: 20 ppm, initial formic acid concentration: 300 ppmC, N 2 purging.
126 Figure 4. 9 shows high magnification TEM images of a TiO2 particle before (Figure 4. 9a) and after (Figure 4. 9b) Se(IV) photoreduction (using 20 ppm initial Se(IV) concentration).
It could be seen that a layer was present beyond the crystalline edge of the TiO2 particle.
This layer was suggested to be the Se deposited on the TiO2 surface.
Presence of Se Layer
TiO2 crystal 5 nm
a b
Figure 4. 9. High magnification (970K) of TiO 2 particles before (a) and after (b) Se(IV) photoreduction. Initial Se(IV) concentration = 20 ppm.
Evidence of Se Layer TiO2 crystal
TiO2 crystal
Figure 4. 10 Film of Se deposits on TiO2 particles. Initial Se(IV) concentration = 80 ppm.
127 Figure 4.10 further shows the evidence of the presence of Se layer around the TiO2 particles when a higher initial Se(IV) concentration (80 ppm) was used. The presence of Se was confirmed by EDX. The formation of Se layer when Se(IV) was used as the precursor was postulated to be due to the high amount of the Se(IV) ions adsorbed onto the TiO2 surface. This could result in the initial formation many tiny Se particles from Se(IV) photoreduction, possibly fusing together subsequently to form the Se layer. As seen from Figure 4.10, the Se layer formed was about 10 to 20 nm thick, much smaller in dimension compared to the round Se particles formed (maximum about 145 nm) from Se(VI) reduction, even though the initial Se(IV) concentration used was 80 ppm compared to 20 ppm of initial Se(VI) concentration. This further supported the occurrence of thin Se layer formed throughout the TiO2 particles when Se(IV) was used instead of forming discrete but bigger Se particles with Se(VI).
Effects of pH on Se Formation
Earlier investigations showed that the rate of Se(VI) photoreduction was influenced by pH and initial solute concentrations (Section 3A.3). In the current study, the effect of pH on the size of the Se particles formed and the amount of Se(VI) reduced after 90 minutes of irradiation were investigated. The results are presented in Figure 4. 11. It can be seen that the size of the Se particles formed was related to the amount of Se(VI) photoreduced. The largest Se particle size (145 nm) was formed when the most Se(VI) was reduced (11.43 mg) at pH 4. Since the amount of Se(VI) reduced in a given time indicates the rate of the reduction, it can be said that the largest Se particles were formed when the Se(VI) reduction rate was the fastest, which was influenced by pH. The explanation for the pH effect on the Se(VI) photoreduction rate has been given in Section 3A.3. Based on our previous postulate, the Se(VI) reduction on the existing Se particles contributes to the growth of the particles. Hence a higher amount of Se(VI) reduced would result in the formation of bigger Se particles.
128 12 200
10 150 8
6 100 (mg)
4 50 Se particle size (nm)
Amount Se(VI) Reduced 2
0 0 2.54.05.0 pH Se(VI) Reduced Size
Figure 4. 11. Effect of pH on the size of Se particles deposited and the corresponding amount of Se(VI) reduced. Experimental conditions: Initial Se(VI) concentration: 20 ppm, initial formic acid concentration: 300 ppmC, N2 purging, 0.5 L test solution and 0.5 gTiO2.
It was also of interest to elucidate the effect of pH on the amount of Se particles deposited when Se(IV) was used as the precursors. Consequently, Se-mapping was performed on two
Se-TiO2 samples collected from Se(IV) photoreduction experiments at pH 3.5 and 5.0 after 90 minutes of irradiation. This is shown in Figure 4. 12. Se-mapping of the samples collected from Se(IV) photoreduction performed at pH 3.5 (Figure 4. 12a) gave stronger mapping signals and more specks compared to that of pH 5.0 (Figure 4. 12b). Also, the amount of Se(IV) reduced at pH 3.5 was found to be higher. This indicated that Se formation at pH 3.5 was more than that at pH 5.0. Hence, changing the pH of the system could control the rate of Se(IV) photoreduction and hence the amount of Se formed on
TiO2. The Se layer formed in Figure 4. 12a was found to be less than 3 nm thick while that of Figure 4. 12b could not be detected under the Transmission Electron Microscope. Hence it is believed that the layer is less than that of Figure 4. 12a in thickness.
129
100 nm 100 nm
a. pH 3.5, RSe = 3.58 mg. b. pH 5.0, RSe = 2.87 mg. . Figure 4. 12. Se mapping by TEM with Se(IV) as precursor. Experimental conditions: Initial Se(IV): 20 ppm, initial formic acid concentration: 300 ppmC, irradiation time: 90 minutes, N2 purging, 0.5 gTiO2. RSe = Amount of Se(IV) reduced (mg).
Implications of Current Work
The current investigations have successfully demonstrated the ability to form Se particles of different morphologies onto TiO2 particles by manipulating the initial conditions such as the Se precursors and the pH of the system. The formation of nano Se-TiO2 might be useful for applications in which the unique property of a dual n-p composite semiconductor is required. One possible use of Se-TiO2 is as nano-sized n-p photodiodes. The photodeposition of a Se film on TiO2 could also imply an alternative technique for forming Se films for application in xerography.
Chenthamarakshan et al (2000b) have reported the enhanced Cd ion removal by Se-TiO2, forming CdSe-TiO2 composite semiconductors. Several mechanisms for CdSe formation on
TiO2 has also been proposed in their studies. To validate their findings and also to demonstrate the implications of Se-modified TiO2, Se-TiO2 was used to remove Cd ions from simulated wastewater. The photocatalytic reduction of Cd ions to metallic Cd has been demonstrated to be thermodynamically difficult due to the redox potential of this reduction being close to or even more negative than the conduction band of unmodified
130 TiO2 (Chenthamarakshan et al, 2000a). When Se-TiO2, prepared by using Se(IV) as the precursor, was used as the photocatalyst, it was found that the Cd ion removal was enhanced when compared to the experiments using Degussa P25. When the experiment was carried out in the absence of a photocatalyst, no Cd ion removal was observed. The results are given in Figure 4. 13. It is postulated that the enhancement is due to the formation of CdSe particles. The mechanism for CdSe formation was not elucidated in this work. The presence of CdSe was confirmed by TEM as shown in Figure 4. 14, with the Cd and Se identified in the particle by EDX. The presence of Cd could not be detected by EDX in the case of using unmodified TiO2 (Degussa P25).
35
30
25
20
15 Conc(ppm) 2+ 10 Cd 5
0 01020304050 time (min) Se-TiO2 TiO2 Blank
Figure 4. 13. Comparison of Cd ion removal using (b) UV/Degussa P25, (c) UV/Se-TiO 2 and (a) UV only.
131
TiO2
CdSe
Figure 4. 14. TEM image of CdSe on TiO2 particles.
UV-Vis absorbance of the prepared Se-TiO2 and CdSe-TiO2 particles was performed and compared with Degussa P25. The adsorption profile is shown in Figure 4. 15. Three distinct peaks in the figure shows the adsorption of TiO2, Se-TiO2 and Se/CdSe/TiO2 at 388, 636,
717 nm respectively. The absorbance of the CdSe modified TiO2 particles were encountered at two regions in the visible light range. One of the absorbance is evidently due to that of Se. This shows that these samples also contain Se particles, forming a system of composite semiconductors consisting of three different materials. The electrons and holes dynamics of such system upon irradiation would be an interesting research topic. Such system might find its niche in composite semiconductor applications.
6.0 5.0 4.0
3.0 Abs(Se)=636nm 2.0 Absorbance
1.0 Abs(CdSe)=717nm Abs(TiO2)=388nm 0.0 350450550650750 wavelength (nm)
Figure 4. 15. UV-Vis absorbance of TiO2, Se-TiO2 and CdSe-TiO2.
132
4A.4 Conclusions
The current investigation has presented possibilities of tailoring the growth of Se particles of different sizes and distribution onto TiO2 particles by changing the pH and using different Se precursors. The pH was found to determine the amount of Se photoreduced at a given time, hence affecting the size of Se particles formed as well as the amount of Se particles deposited onto TiO2. Using Se(VI) as the precursor, large and spherical Se particles, which were discretely formed on different TiO2 sites, could be obtained. The largest Se particle was measured at 145 nm, corresponded to the highest amount of Se(VI) reduced at pH 4. When Se(IV) ions were used as the precursor, Se was seen as a thin film and formed more uniformly distributed onto the TiO2 particles. At a pH of 3.5, it was found that the highest amount of Se(IV) ions were reduced and hence resulting in the largest amount of Se particles being deposited on the TiO2 particles in the pH investigated.
Possible implications of the current work include the formation of Se on TiO2 support in electronic devices such as X-ray image detector or xerography machines. Another application demonstrated in the current work include the enhanced photocatalytic removal of Cd ions using Se modified TiO2 and also the enhanced absorbance into the visible regions of the Se- and CdSe-TiO2 composite semiconductors.
133 4B. Effects of Nano-Ag Particles Loading on
TiO2 Photocatalytic Reduction of Selenate and Selenite Ions to H2Se Gas
4B.1 Introduction
The use of metal (Ag, Pt, Pd) modified TiO2 for the photocatalytic oxidation of organic compounds has been reported with mixed results. Sahyun & Serpone (1997) and Alberici & Jardim (1994) have reported enhancement in chlorophenol and phenol photooxidation rate respectively using Ag-TiO2. However, the presence of metal on TiO2 does not necessarily contribute to rate enhancement. In the investigation of cyclohexane photooxidation by Pt-modified TiO2 (Mu et al, 1989), of which the usual main product was cyclohexanone, it was reported that increasing the Pt content decreased the formation of cyclohexanone due to the formation of by-product cyclohexanol. In terms of photoreduction, it has been reported that the presence of metals (Pd, Rh, Pt) on TiO2 has resulted in ammonia formation from the photoreduction of nitrite and nitrate ions, which was not observed when unmodified TiO2 was used (Ranjit & Viswanathan, 1997). As mentioned in the literasture review, many studies have also observed enhanced photoreduction using metal modified TiO2, such as the photoreduction of carbon dioxide with Rh-modified TiO2 (Xie et al, 2001; Kohno et al, 1999). However, to the best of the author’s knowledge, the use of metal modified photocatalyst in the photoreduction of toxic anionic species, such as selenate and slenite, and even the most commonly investigated chromate anions, has not been encountered.
In the following section, a discussion on the theory of the electron transfer mechanism at the metal-semiconductor junction will be presented. The discussion, based on the detail description of the physics of metal-semiconductor junctions in the reference by Dalven (1990), will be simplified for the sufficient understanding of this work.
134 Metal-n-type-semiconductor Junction
As mentioned in the literature review, TiO2 is an n-type semiconductor in which the electrons are the majority charge carriers. The presence of metal deposits on the TiO2 surface forms a metal-n-type semiconductor junction, at which the energy barrier is referred to the Schottky barrier. Figure 4. 16 shows the band diagram of a metal in contact with an n-type semiconductor with their work functions (F) indicated. The definition of a work function has been given in section 2.1.5. The bold solid line in the figure indicates the electron energy at the various regions of the metal and n-type semiconductor. The work function of the metal Fm, which can be rationalised as the energy necessary to move an electron from the Fermi energy in the metal EF(m) to the vacuum energy Evac (also defined previously), is given as:
m vac -=F EE mF )( (4-2)
Similarly the work function of the semiconductor Fs is defined by:
S -=F EE sFvac )( (4-3)
where EF(s) is the Fermi energy in the semiconductor. It is also necessary to define electron affinity, c, the energy required to move an electron from the conduction band minimum, Ec, to the vacuum energy:
c -= EE cvac (4-4)
For the case of a metal, for example Ag, and a n-type semiconductor, the Fermi level of the semiconductor (EF(s)) is higher than that of the metal (EF(m)). As mentioned previously, when an n-type semiconductor is brought into contact with a metal, electrons are transferred from the semiconductor to the metal until the Fermi levels (chemical potentials) equilibrate. This results in a region of excess charges, W, near the junction, and
135 consequently creates a potential energy gradient which is higher at the junction. As shown in Figure 4. 16, the energy of an electron at the junction EC ́ is larger than the energy ECB of an electron in the bulk of the semiconductor and is equal to c at the junction. The energy barrier DEms between the metal and the semiconductor is given by:
Ems m -F=D c (4-5a)
metal n-type semiconductor
Evac W Evac Fm c Fs
EC ́ DEsm DEms ECB F -c s EF(m) EF(s)
Energy Electron
EVB
W
Distance
Figure 4. 16. Simplified version of the metal-n-type semiconductor junction for the case in which Fm is greater than Fs (Diagram modified from Dalven, 1990a).
It can also be seen that the energy difference between the conduction band edge ECB and
EF(s) is equal to (Fs –c), therefore the energy barrier DEsm for an electron going from the bulk semiconductor into the metal is given by:
Esm m c s c )()( F-F=-F--F=D sm (4-5b)
136 From equations 4-3a and 4-3b, it can be seen that the energy barrier for electron transfer
DEms is greater than DEsm for a metal and n-type semiconductor system of which Fm is greater than Fs, and hence the transfer of electrons from the semiconductor to the metal is favoured.
Many studies using metal-modified TiO2 have attributed the photocatalytic rate improvement to the enhanced electron transfer from the TiO2 to the metal, based on the above-described mechanism (Hufschmidt et al, 2002; Li & Li, 2001). Also, the enhanced photocatalytic reduction rates of carbon dioxide and nitrate have been reported with metal- modified TiO2 as discussed in the previous section. Based on these studies, the work presented in this section aims to elucidate the effect of Ag-modified TiO2 on the photocatalytic reduction of Se(VI) and Se(IV) ions. Experimental parameters such as Ag- loading, pH and formic acid concentration will be studied.
4B.2 Equipment and Procedures
Catalyst and Reagent
Catalysts and chemicals relevant to this section are described in section 3A.2. In addition, silver nitrate (purity 99.99 wt%, Aldrich Chemical Co.) was used as the silver source. Sucrose (supplied by Fisons Scientific Equipment) was used as the hole scavenger in the preparation of the Ag-loaded TiO2.
Preparation of Ag-TiO2
+ The Ag-TiO2 particles were prepared by photoreducing Ag ions (from AgNO3) to Ag metal on the TiO2 surface. The TiO2 used was Degussa P25. 4 grams of TiO2 were added into 1 litre of ultra-pure water in the photoreactor and then irradiated for 10 minutes to remove surface impurities that might be present on the TiO2 surface. Aliquots of various
137 amounts of Ag+ ions, prepared by dissolving silver nitrate salt in deionized water, were + added into the suspension of TiO2 such that the Ag concentration was of 0.5, 1.0, 2.0 or
5.0 atom % in relation to TiO2. The various Ag/TiO2 samples prepared are identified by the Ag+ loading (atom %) added to the suspension prior to the irradiation of the suspension.
During the formating of the Ag-TiO2 particles, sucrose was added to act as the hole scavenger. After the suspension pH was adjusted to 3.5 using perchloric acid, the solution was then irradiated for 40 minutes with continuous nitrogen purging. The suspension was then filtered and the collected Ag-TiO2 particles were washed, dried and ground.
Experimental Procedures for Se Ion Adsorption and Photoreduction
The reaction suspensions were prepared by adding a pre-sonicated solution containing 0.5g of Ag-TiO2 catalyst into a solution consisting 0.254 mM Se(VI) or Se(IV) (from sodium selenate or sodium selenite pentahydrate) and 25 mMC formic acid to make up a 1 L suspension. The pH of the suspension was adjusted and controlled by addition of either perchloric acid or sodium hydroxide. Prior to UV irradiation, the suspension was stirred for 0.5 hours to establish adsorption equilibrium conditions. The suspension was then irradiated with continuous stirring and nitrogen purging for 2 hours. At given time intervals, samples were collected from the suspension and immediately filtered through a 0.45 mm Millipore filter. The filtrate was analysed as required. The total Se concentration (Se(VI) or Se(IV)) in the filtrate was determined by Varian Induced Coupled Plasma-
Atomic Emission Spectroscopy (ICP-AES). The amount of H2Se generated was determined by analysing the amount of Cu(II) remaining in the trap using Varian ICP-AES. The 2+ generation of H2Se was indicated by the presence of black precipitates of CuSe in the Cu scrubber. Formic acid concentration was determined by analysing the total organic carbon (TOC) in the solution using Shimadzu TOC-5000A analyser.
138 Particle Characterisation
Electrophoresis experiments, carried out using the Brookhaven 3-in-1 system, were carried out to measure the zeta potential of the particles. The procedure involved the following.
Suspensions containing approximately 20 mg/L particles (TiO2 or Ag-TiO2) and 0.1 mM NaCl (to control the ionic strength) were prepared. The pH of the suspension was adjusted using dilute HCl and NaOH solution. The suspensions were then subjected to electrophoresis analysis. Powder X-ray Diffraction (Siemens D5000 Diffractometer) was used to study the crystalline nature of the deposited Ag particles. The morphology of the Se particles and the Ag-deposited TiO2 particles was examined using Transmission Electron Microscope (TEM, Phillips CM200). An Energy Dispersive X-ray Spectometer (EDX) in the TEM enabled elemental identification.
4B.3 Results and Discussions
Characterisations of Ag Deposits on TiO2
The Ag-deposited TiO2 particles were brownish in colour. The X-ray diffraction analysis confirmed the presence of Ag-metal on the TiO2 surface. The X-ray diffractogram is shown in Figure 4. 17 where the main peaks of silver metal (marked Ag) and the crystal phases of
TiO2 (marked A for anatase and R for rutile) were identified. Elemental X-ray mapping using EDX connected to the TEM also showed the presence of elemental Ag on the TiO2 particles. This is shown in Figure 4. 18. Based on the XRD and elemental mapping results, it can be said that metallic Ag had been photodeposited on the TiO2 particles. Since the metallic deposits were indistinguishable under TEM, it is estimated that they were nanosized and less than 10 nm diameter.
139
X-Ray Diffraction Pattern of 2.0 atomic% Ag Deposited Degussa P25 particles 800 A 700 600 500
400 Ag
Counts 300 Ag A R 200 R Ag A A Ag A R 100 R A A A A R A 0 0102030405060708090100 Degrees - 2 Theta
Figure 4. 17. X-ray Diffractogram of Ag Deposited Degussa P25 particles.
X
a. Ti signal. b. Ag signal.
c. Se signal. d. Corresponding TEM
image. Figure 4. 18. Figure a, b and c show the mapping of Ti, Ag and Se respectively by EDX. Figure d shows the corresponding TEM image of Se on TiO2. Se(VI) was used as the Se precursor.
140 Photocatalytic Reduction of Se(VI) by Bare TiO2 and Ag-deposited TiO2
When the Ag-TiO2 particles were used as the photocatalysts, the following was observed. With Se(VI) ions as the Se precursor, the colour of the suspension changed from brown
(Ag-TiO2) to orange, indicating the photoreduction of Se(VI) to elemental Se onto the surface of the TiO2. The position marked X on Figure 4. 18 a showed an absence of Ti signal from the mapping. This is due to the Se formed on the TiO2 as shown in Figure 4. 18c and Figure 4. 18d. The presence of elemental Se was confirmed by EDX mapping. From Figure 4. 18d, it can also be seen that the Se particles formed were round and bigger than TiO2. The mechanism of Se formation on the TiO2 particles has been previously discussed in section 4A.
A significant difference in terms of H2Se generation was observed between the systems in which the Ag-TiO2 particles were used as compared to that in which unmodified TiO2 was used. As previously discussed in Section 4a, when unmodified TiO2 was used, H2Se generation was observed when the Se ions had been completely exhausted from the solution. On the other hand, when Ag-TiO2 was used as the photocatalyst, simultaneous
Se(VI) photoreduction and H2Se gas generation was observed. Figure 4. 19 shows the photoreduction of Se(VI) with the corresponding H2Se generation for when Ag-TiO2 was used. The photoreduction of Se(VI) using unmodified TiO2 has been shown in Figure 4. 5.
The mechanism for Se(VI) photoreduction and H2Se generation using unmodified TiO2 as a photocatalyst was previously discussed in Section 4a. It is believed that a different charge transfer mechanism for Ag-TiO2 is leading to the simultaneous Se(VI) photoreduction and
H2Se generation. The following charge transfer mechanism using Ag-TiO2 is postulated and will be explained with the aid of the illustrations in Figure 4. 20. In this figure, the relative bandgaps and work functions of these materials and the direction of electron transfer are shown. When two materials of different work functions are in contact, an energy barrier, also known as the Schottky barrier, for electron transfer is created at the junction (Dalven, 1991a). The transfer of electrons is facilitated from a material of lower
141 work function to one that has a higher work function. For the TiO2-Ag-Se system, the work function of Ag is greater than that of TiO2 while that of Se is greater than that of Ag.
Electrons can hence be transferred from TiO2 to Ag and then to Se. Many studies involving surface modification of TiO2 by metal loading such as Ag (Tada et al, 1998), Au (Li & LI, 2001; Tada et al, 2001) and Pt (Hufschmidt, 2002) have explained the transfer of electron with this theory.
0.12 2.5
0.10 2
0.08 1.5 0.06 1 0.04 Se Generation (mmol) 2 0.5 Se(VI) Disappearance or H 0.02 (mmol) Oxidised HCOOH
0.00 0 0100200300400 time (min)
Figure 4. 19. Se(VI) photoreduction experiments performed using unmodified TiO2 and 2 atomic % Ag-TiO2. ¨ Se(VI) concentration, ■ H2Se generation, ▲ formic acid oxidation. Experimental conditions: 1 L test solution, [Se(IV)]0=20 ppm, [HCOOH]0=300 ppmC, pH 3.5, 1.1 gTiO2/L, N2 purging, 293 K, photon intensity 3.06 mmol/sec.
In the presence of Ag metal on TiO2, it is postulated that the photogenerated electrons transferred from TiO2 to Se via Ag results in the accumulation of electrons in the Se particles even before the exhaustion of Se(VI) in the solution, resulting in the reduction of 2- Se to Se . It is also important to note that the electrons transferred from the TiO2 particles via the Ag metal would not have enough reducing power to initiate the reduction of Se to 2- 2- Se as the conduction band of TiO2 is lower than that of the reduction potential of Se/Se .
Otherwise, H2Se would have been generated by the unmodified TiO2 before the exhaustion of Se(VI) ions. When unmodified TiO2 was used, it is postulated that the accumulation of electrons in the Se semiconductor becomes more significant only upon exhaustion of
142 Se(VI) ions, enabling the further reduction of Se to H2Se (refer to the discussions in section
4a for more details). Hence, the presence of Ag on TiO2 is believed to have greatly increased the rate of electron transfer from the TiO2 to Se, making the self reduction of Se possible even before the exhaustion of Se(VI) ions.
Vac Level eV vs SHE
-3 -2 -2 CB -3 -1 e- 2- -4 CB Se/Se - 0 0 e E =-0.3eV Ag -5 VB Se 1 p-type
-6 Eg=1.95eV 2 -7 VB 3 TiO2 n-type E =3.2eV g
Figure 4. 20. Energy diagram of the TiO2-Ag-Se system at pH 3.5. Bandgap energy (Eg): Eg-TiO2= 3.2eV (Hagfeldt & Gratzel, 1995), Eg-Se=1.95eV (Streltsov et al, 2002). Conduction band potential (CB): CBTiO2=-0.3eV (Tada et al, 1998), CBSe=-1.65eV (Streltsov et al, 2002).. Work function (F): FTiO2=4.2eV, FAg=4.6eV, FSe=4.8eV (assuming Fermi levels are near the conduction and valence band for TiO2 and Se respectively and the work functions are calculated by the equations given in Henglein ( 1997).
When further analysing the photoreduction process using the Ag-TiO2 photocatalyst, it was observed that the rate of Se(VI) reduction was slower compared to that when unmodified
TiO2 was used. From Figure 4. 19, the disappearance rate of Se(VI) with Ag-TiO2 was slower at 6.2×10-6 mmol/sec compared to that of 6.8×10-6 mmol/sec when unmodified
TiO2 was used (Figure 4. 5). The rate of H2Se generation when Ag-TiO2 was used was calculated to be 6.0×10-6 mmol/sec, which was slower compared to 1.9×10-5 mmol/sec
143 when unmodified TiO2 was used. The slower reduction rates could be rationalised as follows. The reduction process could be visualised as two separate steps: the Se(VI) ions were reduced to Se first, and upon Se(VI) exhaustion, elemental Se was further reduced to 2- Se , which resulted in the generation of H2Se gas. When unmodified TiO2 was used, the photogenerated electrons were utilised in the Se(VI) reduction to Se and then from Se to 2- Se in two distinctive reduction steps. However, when Ag-TiO2 was used, the two reduction steps occurred simultaneously, resulting in the competition of the available electrons between the separate steps described above. Consequently, the apparent Se(VI) reduction and H2Se generation rates were slower.
In order to quantify the difference between the observed overall rates in the TiO2 and the
Ag-TiO2 systems, the quantum yields (f) were evaluated. The calculations are shown in
Appendix 2. From these calculation, it was found that when Ag-TiO2 was used, the quantum yield was calculated to be 0.35%, higher then that to the case when unmodified
TiO2 was used, for which was calculated to be 0.23%. The enhanced overall quantum yield indicated an overall better utilisation of the photogenerated electrons owing to the presence of Ag.
Effects of Ag loading and pH on the Photocatalytic Reduction of Se(VI)
To further understand the photoreduction mechanism using the Ag-TiO2 photocatalyst, samples with different Ag loading were prepared. The pH of the system was also varied. The results are summarised in Table 4.1. The effect of the Ag loading will be discussed first. From the results it can be seen that the simultaneous H2Se generation was encountered in all cases when Ag-TiO2 was used. When comparing the different atomic % of Ag- deposited on TiO2 particles, it was found that increasing the Ag deposited amount up to 2 atomic% resulted in an increase in the production of H2Se gas. From TEM mapping, increasing the Ag loading was found to increase the Ag amount distributed on the TiO2 surface, and hence increasing the chance of Ag in contact with Se, resulting in more H2Se production. Nevertheless, when the highest Ag loading of 5.0 atomic% was used, both
144 H2Se generation and Se(VI) reduction seemed to be suppressed. This indicates an optimum
Ag loading existed between 2.0 and 5.0 Ag atomic% on the TiO2 photocatalyst. It has been suggested that a higher amount of Ag deposits could act as the recombination sites and hence resulting in lower quantum efficiency (Li & Li, 2001).
Table 4. 1. Results for Se(VI) photoreduction. Experimental conditions: 1 L test solution, [Se(VI)] 0 =20 ppm, [HCOOH]0=300 ppmC, 1.1 gTiO2/L, N2 purging, 293 K, photon intensity 3.17 mmol/sec.
Ag atomic% on H2Se generated in Se(VI) reduced in Formic Oxidation in -2 -2 -1 TiO2 120min (×10 mmol, 120min (×10 mmol, 120min (×10 mmol, Expt No +/- 2% error) +/- 2% error) +/- 4% error) pH = 2.5 1.1 0.0 0.00 5.26 3.33 1.2 0.5 0.249 5.81 3.40 1.3 1.0 0.800 5.71 3.29 1.4 2.0 1.10 5.19 3.58 1.5 5.0 0.844 4.62 3.42 pH = 3.5 2.1 0.0 0.00 5.61 3.93 2.2 0.5 1.17 6.30 3.81 2.3 1.0 2.44 5.95 4.08 2.4 2.0 4.09 5.38 4.26 2.5 5.0 1.17 4.92 3.97 pH = 5.0 3.1 0.0 0.00 3.76 2.52 3.2 0.5 0.231 4.06 2.27 3.3 1.0 0.487 4.15 2.45 3.4 2.0 0.950 3.54 2.55 3.5 5.0 0.667 2.80 2.13
Also presented in Table 4. 1 are the results of varying the pH during the photocatalytic reduction of Se(VI) using the Ag-TiO2 photocatalyst. As can be seen from these results, the rates of H2Se generation, Se(VI) photoreduction and formic acid oxidation increased as pH was increased from 2.5 to 3.5, but decreased as pH was further increased to pH 5. This
145 observation was similar to the occurrence of optimum pH discussed in section 3A.3, which was attributed to the optimum adsorption ratio of formate:selenate. In order to correlate these trends, the surface charge of the Ag-TiO2 particles was examined using electrophoresis analysis. The electrophoresis analysis was carried out on unmodified, 0.5
Ag atomic% and 5.0 Ag atomic% TiO2. The results are presented in Figure 4. 22. The Ag-
TiO2 particles, both 0.5 and 5.0 Ag atomic%, did not show significant difference in electrophoresis in comparison with unmodified TiO2. It could hence be assumed that the adsorption of Se(VI) and formate ions by unmodified and modified TiO2 was relatively similar. The observed optimum pH at 3.5 with Ag-TiO2 was hence attributed to the same reasoning discussed in section 3A.3. The slowest rate of Se(VI) photoreduction at pH 5.0 was due to the very little Se(VI) adsorption at this pH.
40 30
20
10
0 012345678910 -10 Zeta-potential (mv) pH -20
-30
0.0 Ag atomic% 0.5 Ag atomic% 5.0 Ag atomic%
Figure 4. 22. Surface charge of TiO2 and Ag-TiO2 particles at various pH.
It was also found that the maximum Se(VI) reduction rate was encountered at pH 3.5 and at 0.5 Ag atomic%. This is an improvement of about 10% in terms of the total Se(VI) reduced over a duration of 120 minutes when compared to unmodified TiO2. The highest amount of
H2Se generation was also encountered at pH 3.5 and was lowest at pH 5.0. It could hence be said that the rate of H2Se generation was dependent on the rate of Se(VI) reduced. The
146 + lower amount of H2Se formed at pH 5.0 could also be due to the lack of H ions in the suspension at that pH. According to equation (2-26), H+ ions are necessary for the formation of H2Se.
By referring to Table 4. 1 again, it was also found that the oxidation rate of formic acid was the highest at pH 3.5 and lowest at pH 5.0. This was expected since improved oxidation was an indication of more effective hole scavenging and hence effective electron scavenging. It was not surprising that the optimum conditions for formic acid oxidation corresponded with that of optimum Se(VI) reduction. As discussed in Section 3B.3, the formate ions could be attacked by the photogenerated holes, yielding reducing radicals, and that could contribute to the reduction of Se(VI). An increase in formic acid oxidation rate could suggest an increase in reducing radicals formation and hence an enhanced Se(VI) reduction.
Photocatalytic Reduction of Se(IV) by Bare TiO2 and Ag-deposited TiO2
The Se(IV) photoreduction using bare TiO2 and Ag-TiO2 is shown in Figure 4. 23. When the photocatalytic reduction of Se(IV) was carried out using Ag-TiO2, similar trends to those of Se(VI) were encountered. The colour of the suspension changed from white (TiO2) or brown (Ag-TiO2) to orange, indicating the photoreduction of Se(IV) to elemental Se onto the surface of the TiO2. When unmodified TiO2 was used, H2Se was only generated when
Se(IV) was exhausted from the solution. When Ag-TiO2 was used, H2Se gas was simultaneously generated as Se(IV) was being reduced. It is proposed that the same electron transfer mechanism for the case of Se(VI) discussed earlier applied to the reduction of Se(IV) as well. Similarly, it was also observed that the rates of Se(IV) reduction and H2Se generation using Ag-TiO2 were slower than that of unmodified TiO2. Again, this is attributed to the competition of the available electrons for the two separate reduction steps, which were now occurring at simultaneously in the presence of Ag-TiO2.
147
12 12
10 10
8 8
6 6
Se Generated (mg) 2 4 4
2 2
Mass of H Concentration of Se(IV) (ppm) Se(IV) of Concentration 0 0 050100150 Time (min)
Figure 4. 23.Comparison on of Se(IV) reduction using TiO2 and Ag-TiO2 (1 atomic %). ■ Se(IV) concentration using TiO2, × Se(IV) concentration using Ag-TiO2, ¨ H2Se generation using TiO2, ▲ H2Se generation using Ag-TiO2. Experimental conditions: 1 L test solution, [Se(IV)]0=10 ppm, [HCOOH]0=400 ppmC, pH 3.0, 1.1 gTiO2/L, N2 purging, 293 K, photon intensity 3.17 mmol/sec.
Effects of Ag Loading, pH and Formic Acid Concentration on the Photocatalytic Reduction of Se(IV)
The effect of Ag loading on TiO2, pH and formic acid concentration on the Se(IV) photoreduction were also investigated. The results are summarised in Table 4. 2. For all the experiments performed with unmodified TiO2, no H2Se generation was observed after 60 minutes of irradiation. When Ag-TiO2 was used, H2Se gas was generated simultaneously with the reduction of Se(IV). The quantum yields were also evaluated. Similar to the case of Se(VI) photoreduction, the quantum yield was improved when Ag-TiO2 was used. Two different Ag loadings of 1 and 2 atomic% were investigated. It was found that increasing
Ag loading on the TiO2 from 1 to 2 atomic% increased the rate of H2Se generation but decreased the rate of Se(IV) photoreduction in the solution. This indicated an optimum loading of 1 atomic% for Se(IV) photoreduction for the current Ag loading range investigated, corresponding to the optimum quantum yield. An efficient quantum yield is an indication of efficient utilisation of the photogenerated electrons. It is not surprising that the quantum yield corresponded closely with the Se(IV) photoreduction step since this
148 - process required more electrons (4 e, see equation 2-24) compared to the H2Se generation - step (2 e, see equation 2-26). The optimum loading for H2Se generation has not been encountered in this case, probably due to the lesser number of Ag-loaded TiO2 investigated. However, similar trends to that of Se(VI) were observed for the current case for both
Se(IV) reduction and H2Se generation patterns.
Table 4. 2. Summary of results for Se(IV) photoreduction. Experimental conditions: 1 L test solution, -2 [Se(IV)]0=10 ppm (12.7×10 mM), 1.1 gTiO2/L, N2 purging, 293 K, photon intensity 3.06 mmol/sec. Se(IV) reduced in Ag H2Se generated in Expt Conditions and 60 min (×10-2 Quantum Yield atomic% 60 min (×10-2 mmol, Number: mmol, +/- 2% (f) on TiO2 +/- 2% error) error) pH = 2.5 @ 235 ppmC 4.1 0 0 10.4 0.94 4.2 1 0.7 12.7 1.2 4.3 2 1.32 9.62 1.0 pH = 2.5 @ 400 ppmC 5.1 0 0 10.1 0.91 5.2 1 2.31 12.7 1.4 5.3 2 3.48 9.84 1.2 pH = 3.5 @ 235 ppmC 6.1 0 0 10.1 0.91 6.2 1 0.94 12.1 1.2 6.3 2 1.59 9.25 1.0 pH = 3.5 @ 400 ppmC 7.1 0 0 10.1 0.91 7.2 1 2.10 12.7 1.3 7.3 2 3.35 9.42 1.1
As the formic acid concentration was increased from 235 to 400 ppmC, the H2Se generation rate was found to increase too. This is suggested to be the result of the presence of a greater amount of hole scavenger, resulting in enhanced photoreduction. The greater concentration of H+ due to the higher amount of formic acid added could have also contributed to the increase in H2Se generation rate. The increase in formic acid
149 concentration also brought about an increase in Se(IV) photoreduction rate too. This was explained in section 3A.3, that the Se(IV) photoreduction was enhanced at higher formic acid concentration, possibly due to the increase in formic radicals.
Comparison of Se(VI) and Se(IV) Photoreduction with Ag-TiO2
Figure 4. 24 compares the rates of Se(VI) and Se(IV) removal in the solution and H2Se generation at different pH values of 2.5 and 3.5 and formic acid concentrations of 235 and
400 ppmC using 1 atomic % Ag-TiO2. It could be seen that the removal of Se(IV) from the solutions in all the experiments was greater than that of Se(VI). It was also found that at the same pH of 3.5, the rate of H2Se generation was comparable for experiments using Se(IV) and Se(VI) at 235 ppmC, but was higher for Se(IV) at a higher formic acid concentration of 400 ppmC (compare Figure 4. 24a and Figure 4. 24b). As discussed earlier, the rate of Se ions removal could play an important role on the rate H2Se generation and the Se(IV) photoreduction was enhanced at higher formic acid concentration. Therefore, the increase in formic acid concentration from 235 to 400 ppmC resulted in higher rate of Se(IV) photoreduction (gradient of -0.0541 and -0.0783 ppmSe(IV)/min for 235 and 400 ppmC respectively) and hence a higher rate of H2Se of generation. In addition, another factor that + might contribute to enhanced H2Se generation rate was the greater concentration of H at higher formic acid concentration. The rate of Se(VI) photoreduction was depressed at high formic acid concentration (as found in section 3A.3), resulting in slower rate of H2Se generation compared to the case when Se(IV) was used.
At the lower pH of 2.5 and higher formic acid concentration (see Figure 4. 24c and Figure
4. 24d), the rate of H2Se generation was significantly higher when Se(IV) was used as the Se precursor. This could be attributed to the combined effects of lower pH, which resulted + in greater amount of Se(IV) adsorption and provided more H for H2Se formation, and the higher formic acid concentration, which favoured Se(IV) photoreduction. For the case of
Se(VI) photoreduction at pH 2.5, H2Se generation rates at both formic concentrations were lowered compared to that at pH 3.5. This consolidated the previous finding that at the lower
150 pH of about 2.5, Se(VI) photoreduction was not optimised, which in this case resulted in the lower rate of H2Se generation.
12 12 12 12 10 10 10 10 8 8 8 8
6 6 6 6 4 4 4 4
GenerationSe Se Generation 2 2 H H Concentration of of Concentration Concentration of Concentration 2 2 2 2 Se(VI)/Se(IV) (ppm) Se(VI)/Se(IV) Se(VI)/Se(IV) (ppm) Se(VI)/Se(IV) 0 0 0 0 050100150200 050100150200 Time (min) Time (min)
a. pH 3.5, 235 ppmC. b. pH 3.5, 400 ppmC.
12 12 12 12 10 10 10 10 8 8 8 8 6 6 6 6
4 4 4 4 Se Generation Se GenerationSe 2 2 H 2 2 H 2 2 Concentration of Concentration Concentration of of Concentration Se(VI)/Se(IV) (ppm) Se(VI)/Se(IV)
(ppm) Se(VI)/Se(IV) 0 0 0 0 050100150200 050100150200 Time (min) Time (min)
c. pH 2.5, 235 ppmC. d. pH 2.5, 400 ppmC.
Figure 4. 24. Comparison of Se(IV) and Se(VI) photoreduction using 1 atomic % Ag-TiO 2. × Se(IV) concentration, ■ Se(VI) concentration, ▲H2Se generation with Se(IV) as precursor, ¨ H2Se generation with Se(VI) as precursor. Experimental conditions: 1 L test solution, [Se(IV)/Se(VI)] 0=11 ppm, 1.1 gAg-TiO2/L, N2 purging, 293 K, photon intensity 3.17 mmol/sec.
From the results obtained from both Se(VI) and Se(IV) photoreduction using Ag-TiO2, the relative rates of H2Se generation and Se ions photoreduction could be manipulated by manipulating the experimental parameters, which were the Ag loading, the pH of the system, the formic acid concentration or the choice of Se precursor.
Implications of Current Work
From a waste water treatment point of view, the generation of toxic H2Se gas is undesirable. The need for its entrapment also constitutes an additional cost to the process.
151 However, if the H2Se gas generated could be trapped and transformed into products with commercial value, this technology would be regarded as more appealing. In addition, the further reduction of the photodeposited Se to H2Se is necessary for the regeneration of the
TiO2 photocatalyst.
As mentioned previously, a copper (II) solution was used to remove the H2Se generated in the form of copper selenide precipitate. Copper selenide is a useful semiconductor and a promising material for solar cell application (Garcia et al, 1999). This approach of scrubbing the H2Se gas generated inspired the precipitation of metal selenides from metal ions solutions from the photocatalytic reduction of Se ions. Previously, a similar approach utilised the toxic H2Se gas as the precursors directly and subsequently passed the H2Se gas into Cu(II) solution to precipitate out the CuSe (Li et al, 1999). This method was deemed hazardous due to the use of large quantity of H2Se gas. The more accepted preparatory method such as chemical vapour deposition also involves complex chemicals reactions
(Khanna et al, 2003; Garcia et al, 1999). TiO2 photocatalysis could provide a relatively safe, cheap and simple technique in the preparation of these compounds. The precursors used (Se ions) are safer to handle and could be obtained from industrial and mining effluent.
Besides copper selenide, other metal selenides such as cadmium selenide (Khanna et al, 2003) and zinc selenide (Gard et al, 2003) have also been extensively used as light emitting diodes, in solar cells and catalysis. In this study, the simultaneous production of metal selenides as part of the Se ions photoreduction process was also investigated. The following describes the preliminary results from the investigations of metal selenide precipitation from Se ions photoreduction.
In three separate experiments, the generated H2Se gas was passed into scrubbing solutions containing 100 ppm of copper(II) , cadmium(II) or zinc(II) ions solutions, forming copper, cadmium and zinc selenide respectively as precipitates. The colours of the precipitates were black, orange and yellow respectively. The yield of the product was shown in Table 4. 3 and was found to be related to the solubility products of the precipitates. Being less soluble,
152 CuSe precipitated out of the solution at a higher quantity and hence resulted in a greater yield. Other metal selenide compounds (such as silver and lead selenide) may also be prepared by this method.
Table 4. 3. Metal Selenide formed by passing H2Se into metal ions solution and the corresponding yield. pKsp values obtained from Seby et al (2001).1).
Semiconductor Formed Yield pKsp CuSe 68% 48.1 CdSe 42% 35.2 ZnSe 24% 29.4
When viewed under TEM, these metal selenide particles were found to be amorphous. Further processing would be required (such as heating) to convert them to useful crystalline materials. Optimisation of the process is still needed to increase the yield. Other experimental parameters, such as the rate of H2Se generation, which could be altered by using different initial conditions, must be further investigated to see its effect on the metal selenide particles formed. However, these were not done in the current work as it only intended to demonstrate the feasability of forming such particles utilising photocatalytic technology.
4B.4 Conclusions
Both unmodified and Ag-modified TiO2 particles were able to photoreduce the Se ions
(Se(VI) and Se(IV)) to elemental Se. In the case in which unmodified TiO2 was used, H2Se generation occurred only when the Se ions in the suspension was exhausted. However,
H2Se was generated simultaneously during the reduction of Se ions when Ag-TiO2 was used. This was explained in terms of enhanced electron mediation from TiO2 to Se via the Ag metals, greatly increasing the electron density in the Se particles and hence leading to the formation of H2Se via the self-reduction of Se. The quantum efficiency was found to improve when Ag-TiO2 was used. When Se(VI) was used as the precursor, a different
153 optimum Ag-loading for H2Se generation and Se(VI) photoreduction was encountered when compared to the system in which Se(IV) ions were used as precursor. When Se(IV) was used as the precursors, the optimum quantum yield was found to coincide with the maximum Se(IV) photoreduction rate. It was also found that the optimum conditions for
H2Se generation was pH 2.5, 400 ppmC of formic acid, 2 atomic % Ag-TiO2 and using
Se(IV) as the Se precursor. The rate of Se ions reduction and H2Se generation was affected by a combination of factors of Ag loading, pH and formic acid concentration. Depending on the requirement, these parameters could be manipulated to achieve either high H2Se generation but low Se ions reduction or the reverse. It was successfully demonstrated that the generation of H2Se by the photocatalytic reduction of Se(VI) and Se(IV) ions could be extended as a relatively simple and safe technique in the preparation of copper, cadmium and zinc selenide.
154 Chapter 5. Conclusions
The first section of this work aimed to optimise the photoreduction of the Se ions under different experimental parameters and to formulate a kinetic model representing the reaction mechanisms. The second part investigated the formation of nano-size Se compounds and the effects of using Ag-modified TiO2 for Se ions photoreduction with subsequent implications for Se-compound synthesis.
In the first section, preliminary experiments found that the adsorption of Se ions onto the
TiO2 surface, and the presence of formate ions as the hole scavenger were essential for Se ions photoreduction to elemental Se. The latter could imply that the hole-scavenging step was the rate-limiting step. The photoreduction rate was depressed in the presence of oxygen. The elemental Se was further reduced to hydrogen selenide, H2Se, once the Se ions were exhausted from the solution. It was observed that the overall rate of Se(IV) removal was faster than that of Se(VI). However, the rate of Se(VI) photoreduction, comparing the quantum yield of Se(VI) and Se(IV) photoreduction, was found to be higher. When the effects of pH and initial formic acid concentration on the Se ions photoreduction process were investigated, the optimum Se(VI) photoreduction rates were found to be closely correlated to the stoichiometric molar adsorption ratio of 3:1 of formate-to-selenate on the
TiO2 surface. For Se(IV) photoreduction, optimum molar adsorption ratio was also encountered but its correlation with the stoichiometry of 2:1 was not conclusive. The importance of optimum molar adsorption of the precursors demonstrated the synergism between photocatalytic oxidation and reduction. It showed that efficient photocatalytic systems could be realised if both oxidant and reductant were present in the ‘right’ amount on the photocatalyst surface.
In the investigation of using different organic hole scavengers in the photoreduction of Se ions, photoreduction was only observed in the presence of formic acid, methanol or ethanol and the rate of Se ions photoreduction was found to be in the order formic acid> methanol> ethanol. The reason for formic acid being the most efficient hole scavenger was suggested
155 to be due to its ability to effectively compete with Se ions for adsorption on the TiO2 surface, its fast mineralisation rate and its ability to form reducing radicals quickly. Even though methanol and ethanol did not adsorb significantly in the presence of Se ions and were not easily mineralized, their presence enabled the reduction of Se ions due to their ability to form reducing radicals. Increasing the concentration of the above organic additives resulted in a greater extent of Se(IV) reduction for all three organic compounds used. An optimum pH was encountered for the Se(VI) and Se(IV) photoreduction when formic acid was used as the hole scavenger. When methanol and ethanol were invetigated as hole scavengers in the pH range of 2.2 to 4.0, the greatest extent of Se ions photoreduction was encountered at pH 2.2 which corresponded to the highest amount of Se ions adsorbed. This showed that methanol and ethanol had a different role compared to formic acid in the photoreduction process.
New Se(VI) and formic acid adsorption models were derived assuming the adsorption of one Se(VI) ion onto two active sites on the TiO2 surface and the competitive adsorption between Se(VI) and formate ions. The new models fitted the experimental adsorption data reasonably well. The use of these adsorption models enabled the derivation of a rate equation on the basis of the LH reaction mechanism. The model confirmed that the optimum photoreduction rate of Se(VI) was achieved at formate-to-selenate molar adsorption ratios of 3.3+0.2 and 2.6+0.1 for 0.256 and 0.512 mM initial Se(VI) concentration respectively, supporting the postulate for an optimum formate to selenate molar adsorption ratio of 3:1. A composite rate law incorporating the effect of catalyst loading was also derived. On the basis of the above model, a reaction mechanism describing the redox reaction based on the competitive adsorption of both Se(VI) and formate ions and the subsequent scavenging of the photogenerated electrons and holes on the catalyst surface was proposed.
In the second part of this work, the possibility of forming Se particles of different morphologies onto TiO2 particles was presented. Using Se(VI) as the precursor, large and round Se particles, which were discretely formed on different TiO2 sites, were obtained.
When Se(IV) ions were used as the precursor, Se was deposited as a film on the TiO2
156 particles. The pH was found to determine the amount of Se ions photoreduced at a given time, hence affecting the size of Se particles formed as well as the amount of Se particles deposited onto TiO2. Possible implications of the current work include the depostion of Se on TiO2 support for application in electronic devices such as X-ray image detector or xerography machines. The enhanced photocatalytic removal of Cd ions using Se modified
TiO2 was also demonstrated in the current work.
When Ag-modified TiO2 photocatalysts were employed for the Se ions photoreduction,
H2Se was generated simultaneously during the reduction of Se ions. This was different to the observations made for the systems which utilised unmodified TiO2 as photocatalysts, in which the generation of H2Se did not occur until the Se ions were exhausted from solution.
This was explained in terms of enhanced electron mediation from TiO2 to Se via the Ag metals, greatly increasing the electron density in the Se particles and hence leading to the formation of H2Se via the self-reduction of Se. It was found that the quantum efficiency was increased when Ag-TiO2 was used. When Se(VI) was used as the precursor, an optimum H2Se generation rate was encountered at Ag-loading of 2.0 atomic %, pH 3.5 and
300 ppmC of formic acid. The optimum conditions for H2Se generation when using Se(IV) as the Se precursor was 2 atomic % Ag-TiO2, pH 2.5 and 400 ppmC. It was suggested that the generation of H2Se by the photocatalytic reduction of Se(VI) and Se(IV) ions could be used as a relatively simpler and safer technique in the preparation of copper, cadmium, zinc selenide.
157 Chapter 6. Recommendations
The author proposed the following recommendations for future improvement and development of the current work:
1. As discussed in section 3A.3, it was rather difficult to compare the photoreduction rate between experiments using different Se precursors due to the difference in their extent of adsorption. In the current work, the photoreduction rate was estimated from the disappearance of the Se ions after irradiation by analysisng the remaining Se ions in the solutoin. This method of analysis was common for similar investigations, such as the photoreduction of chromate in Chenthamarakshan & Rajeshwar (2001). However, in order to obtain a direct measure of photoreduction, it was recommended that the direct quantification of the photoreduced product on
the TiO2 surface, which is selenium in this case, should be done. This could be
achieved by digesting the Se particles on the TiO2 and then analyzing the dissolved Se ions after washing. It would be ideal if another analytical method could be
devised or employed to measure the quantity of Se formed on TiO2 in-situ. As a suggestion, this could be done by exploiting the unique property of Se, such as its photoconductance and its absorbance at 616 nm.
2. It would also be beneficial in terms of mechanistic study to analyse the different Se
species present on the TiO2 particles during photoreduction, especially for the case when Se(VI) was used as the precursor. For Se(VI) photoreduction, an overall reduction step was assumed from Se(+6) to Se(0), by-passing the Se(+6) to Se(+4) reduction step. The ability to elucidate the different Se species on the surface would enable the validation of the assumption of overall reduction step and also enable us to propose a more comprehensive reduction mechanism. This may be done by X-ray Photon Spectroscopy (XPS).
158 3. A further study investigating the role of formate ions as the organic hole scavenger, and the reducing radicals’ role (formed from the reactions between formate/methanol/ethanol and hydroxyl radicals) in the photoreduction reactions would be contributive to the mechanistic study of the present work. Techniques such as Electron Paramagnetic Resonance (EPR) might be helpful in determining
the electron transfer mechanism between these species and the Se ions or TiO2 surface.
4. The current work has successfully demonstrated the ability of TiO2 photocatalysis to reduce both Se ions effectively and efficiently under optimum experimental conditions. This is especially important for Se(VI) removal since exisitng technologies have found difficulty in its removal, and if the technique is efficient (such as the ferrous hydroxide method) it is relatively expensive. However, to further investigate photocatalysis as a viable removal technique, actual effluent samples from industries or mines should be used. This is because such effluents consist of many dissolved ions, which could be inhibitory to the photocatalytic reduction. Existing studies by Sanuki et al (1999) have already shown the inhibitory effects of sulfate and sulfite on the adsorption and hence the photoreduction of Se anoins.
5. The formation of metal selenide semiconductor compounds by passing the H2Se gas into metal ions scrubbing solution could be further studied by varying experimental
parameters such as the concentrations of the scrubbing solution, the rate of H2Se generation rate and the addition of surfactant to explore the possibility of forming particles of various size and structure. It was suggested that this technique of metal selenide formation could be simpler compared to conventional methods as the photoreduction of the Se anions was relatively straightforward involving simple reduction procedure. It could also be less expensive since Se anoins were used as the precursors, which might be obtained from industrial effluents.
159 6. The mechanism postulated in Equation 4.1 could lead to the precipitation of Se in the solution. However, from the observation in the current work, if Se precipitation were to occur in the solution, it would be rather difficult to be distinguished from
the deposition of Se onto the TiO2 surface as the TiO2 particles eventually assumed the colour pink like the Se element. It is hence recommended to carry out future work to investigate the above as it may help to further elucidate the mechanism of Se ions photoreduction.
7. The addition of Se(VI) solution to TiO2 formed from the reduction of Se(IV) or vice versa could also be an interesting technique for elucidating the mechanism of Se
anions photoreduction and might lead to the formation of composite Se-TiO2 particles of intriguing morphology. This may also be performed in future studies.
160 References
Abdel-Moati, M. A. R., Speciation of Selenium in a Nile Delta Lagoon and SE Mediterranean Sea Mixing Zone, Estuarine, Coastal and Shelf Science 46 (1998) 621-628.
Abdullah, M., Low, G. K. C., Matthews, R.W., Effects of Common Inorganic Anions on Rates of Photocatalytic Oxidation of Organic Carbon over Illuminated Titanium Dioxide, Journal of Physical Chemistry 94 (1990) 6820-6825.
Adams, J., Pickett, T., Montgomery, J. Biotechnologies for Metal and Toxic Inorganic Removal from Mining Process and Waste Solutions; 1996; Denver CO., 143-146. ATSDR - Agency for Toxic Substances and Disease Registry – USA website, “Public Health Statement for Selenium”, CAS# 7782-49-2, retrieved 17th April 2003. (http:// www.atsdr.cdc.gov/toxprofiles/phs92.html.) Aguado, M. A., Anderson, M.A., Degradation of Formic Acid over Semiconducting Membranes Supported on Glass: Effects of Structure and Electronic Doping, Solar Energy Materilas and Solar Cells 28 (1993) 345-361.
Al-Ekabi, H., Serpone, N., Kinetic Studies in Heterogeneous Photocatalysis. 1. Photocatalytic Degradation of Chlorinated Phenols in Aerated Aqueous Solutions over TiO2 Supported on a Glass Matrix, Journal of Physical Chemistry 92 (1988) 5726-5731.
Alam, M., Montalvo, R.A., Titania-Assisted Photoreduction of Cr(VI) to Cr(III) in Aqueous Media: Kinetics and Mechanisms, Metallurgical and Materials Transaction B 29B (1998) 95-140.
Alberici, R. M., Jardim, W. F., Photocatalytic degradation of phenol and chlorinated phenols using Ag-TiO2 in a slurry reactor, Water Research 28 (8) (1994) 1845-1849.
Alfano, O. M., Bahnemann, D., Cassano, A. E., Dillert, R., Goslich, R., Photocatalysis in Water Environments using Artificial and Solar Light, Catalysis today 58 (2000) 199-230.
Andreozzi, R., Caprio V., Insola, A., Marotta, R., Advanced Oxidation Processses (AOP) for Water Purification and Recovery, Catalysis Today 53 (1999) 51-59.
Asmus, K. D., Mockel, H., Henglein, A., Pulse Radiolytic Studies of the site of OH Radical Attack on Aliphatic Alcohols in Aqueous Solution, Journal of Physical Chemistry 77 (10) (1973) 1218-1221.
Augugliaro, V., Loddo, V., Palmisano, L., Schiavello, M., Performance of Heterogeneous Photocatalytic Systems: Influence of Operational Variables on Photoactivity of Aqueous Suspension of TiO2, Journal of Catalysis 153 (1995) 32-40.
161
Bangun, J., Adesina, A. A., The Photodegradation Kinetics of Aqueous Sodium Oxalate Solution using TiO2 Catalyst, Applied Catalysis A: General 175 (1998) 221-235.
Bedja, I., Kamat., P. V., Capped Semiconductor Colloids. Synthesis and Photoelectrochemical Behavior of TiO2-Capped SnO2 Nanocrystallites, Journal of Physical Chemistry 99 (1995) 9182-9188.
Besser, J. M., Huckins, J. N., Little, E. E., La Point, T. W., Distribution and bioaccumulation of selenium in aquatic microcosms, Environmental Pollution 62 (1) (1989) 1-12.
Besser, J. M. P., JCanfield, T., Point, T. W. L., Bioaccumulation of organic and inorganic selenium in a laboratory food chain, Environmental Toxicology and Chemistry 12 (1) (1993) 57-72.
Beydoun, D., Amal. R., Low, G., McEvoy, S., Role of Nanoparticles in Photocatalysis, Journal of Nanoparticle Research 1 (1999) 439-458.
Blake, D. M.; 1994. Biblography of Work on the Photocatalytic Removal of Hazardous of Hazardous Compounds from Water and Air. Springfield: National Technical Information Service, US Department of Commerce May. Report nr VA22161.
Blake, D. M.; 1996. Biblography of Work on the Photocatalytic Removal of Hazardous of Hazardous Compounds from Water and Air: Update Number 1. Springfield: National Technical Information Service, US Department of Commerce October. Report nr VA22161.
Blake, D. M.; 1997. Biblography of Work on the Photocatalytic Removal of Hazardous of Hazardous Compounds from Water and Air. Update Number 2. Springfield: National Technical Information Service, US Department of Commerce January. Report nr VA22161.
Bouroushian, M., Kosanovic, T., Loizos, Z., Spyrellis, N., On a Thermodynamic Description of Se(IV) Electroreduction and Cdse Electrolytic Formation on Ni, Ti and Pt Cathodes in Acidic Aqueous Solution, Electrochemistry Communications 2 (2000) 281- 285.
Boxall, C., Kelsall, G.H., Photoelectrophoresis of Colloidal Semiconductors: Part 1. The Techniques and its Application, Journal of Chemical Society Faraday Transaction 87 (21) (1991a) 3357-3545.
Boxall, C., Kelsall, G.H., Photoelectrophoresis of Colloidal Semiconductors: Part 2. Transient Experiments on TiO2 Particle, Journal of Chemical Society Faraday Transaction 87 (21) (1991b) 3347-3556.
Brezova, V., Blazkova, A., Surina, I., Havlinova, B., Solvent Effect on the Photocatalytic
162 Reduction of 4-Nitrophenol in Titanium Dioxide Suspensions, Journal of Photochemistry and Photobiology A: Chemistry 107 (1997) 233-237.
Bunce, N. J., Chp 9: Actinometry, In: Scarano JC, editor. CRC Handbook of Organic Photochemistry. Volume 1. Florida: CRC Press, Inc; (1989). p 241-259.
Burk, R. F. Selenium in biology and human health. Burk RF, editor. New York: Springer- Verlag; 1994.
Butterfield, I. M., Christensen, P. A., Curtis, T. P., Gunlazuardi, J., Water disinfection using an immobilised titanium dioxide film in a photochemical reactor with electric field enhancement, Water Research 31 (3) (1997) 675-677.
Carraway, E. R., Hoffman, A. J., Hoffmann, M. R., Photocatalytic Oxidation of Organic Acids on Quantum-Sized Semiconductor Colloids, Environmental Science & Technology 28 (5) (1994) 786-793.
Chan, Y. C., Chen, J.N., Lu, M.C., Intermediate Inhibition in the Heterogeneous UV- Catalysis using a TiO2 Suspension System, Chemosphere 45 (2001) 29-35.
Chang, R., Chapter 16 Acid Base Equilibria, In: Chang, R., editor. Chemistry. Fifth Edition, McGraw Hill; (1994) 631-678.
Chatterjee, D., Photocatalytic Reduction of Hydrazine to Ammonia Catalysed by [Ru(III)(edta)(H2O)-] Complex in a Pt/TiO2 Semiconductor Particulate System, Journal of Molecular Catalysis A: Chemical 154 (2000) 1-3.
Chen, D., Ray, A. K., Photocatalytic Kinetics of Phenol and its Derivatives over UV irradiated TiO2, Applied Catalysis B: Environmental 23 (1999) 143-157.
Chen, J., Ollis, D.F., Rulkens, W.H., Bruning, H., Photocatalyzed oxidation of alcohols and organochlorides in the presence of native TiO2 and metallized TiO2 suspensions. Part (I): photocatalytic activity and pH influence, Water Research 33 (3) (1999) 661-668.
Chen, D., Sivakumar, M., Ray, A. K., Heterogeneous Photocatalysis in Environmental Remediation, Developments in Chem. Eng. and Mineral Process 8 (5/6) (2000) 505-550.
Chen, D., Ray, A. K., Removal of Toxic Metal Ions from Wastewater by Semiconductor Photocatalysis, Chemical Engineering Science 56 (2001) 1561-1570.
Chenthamarakshan, C. R., Rajeshwar, K., Heterogeneous Photocatalytic Reduction of Cr(VI) in UV-Irradiated titania Suspension: Effects of Protons, Ammonium Ions, and Other Interfacial Aspects, Langmuir 16 (6) (2000) 2715-2721.
Chenthamarakshan, C. R., Yang, H., Ming, Y., Rajeshwar, K., Photocatalytic Reactivity of Zinc and Cadmium Ions in UV-Irradiated Titania Suspensions, Journal of Electroanalytical
163 Chemistry 494 (2000a) 79-86.
Chenthamarakshan, C. R., Ming, Y., Rajasehwar, K., Underpotential Photocatalytic Deposition: A New Preparative Route to Composite Semiconductors, Chemical Materials 12 (2000b) 3538-3540.
Chiang, K., Amal, R., Tran, T., Photocatalytic degradation of cyanide using titanium dioxide modified with copper oxide, Advances in Environmental Research 6 (4) (2002) 471-485.
Chiron, S., Fernandez-Alba, A. R., Rodriguez, A., Pesticide chemical oxidation processes: an analytical approach, Trends in Analytical Chemistry 16 (9) (1997) 518-527.
Choe, S., Chang, Y., Hwang, K., Khim, J., Kinetics of Reductive Denitrification by Nanoscale Zero-Valent Iron, Chemosphere 41 (2000) 1307-1311.
Choi, W., Termin, A., Hoffmann, M. R., The Role of Metal Ion Dopants in Quantum-sized TiO2: Correlation between Photoreactivity and Charge Carrier Recombination Dynamics, journal of Physical Chemistry 98 (1994) 13669-13679.
Cocivera, M., Darkowski, A., Love, B., Thin Film CdSe Electrodeposited from Selenosulfite Solution, Journal of Electrochemical Society 131 (11) (1984) 2514-2517.
Dalven, R., Chapter 7: Detectors and Generators of EM Radiation. Introduction to Applied Solid State Physics. second ed: Plenum Press. New York and London; (1990a) 177-216.
Dalven, R., Chapter 4: Physics of Metal-Semiconductor and Metal-Insulator- Semiconductor Junctions. Introduction to Applied Solid State Physics. Second ed: Plenum Press. New York and London; (1990b) 111-128.
Dutrizac, J. E., Dinardo, O., Kaiman, S., Selenite Analogues of Jarosite-Type Compounds, Hydrometallurgy 6 (1981) 327-337.
Feroci, G., Fini, A., Badiello, R., Breccia, A., Interaction between Selenium Derivatives and Heavy Metal Ions: Cu2+ and Pb2+, Microchemical Journal 57 (1997) 379-388.
Fogler, H. S. Elements of Chemical Rection Engineering. New Jersey: Prentice Hall; 1992.
Fujishima, A., Honda, K., Electrochemical Photolysis of Water at a Semiconductor Electrode, Nature 238 (1972) 37-38.
Fujishima, A., Rao, T. N., Tryk, D. A., TiO2 Photocatalysts and Diamond Electrodes, Electrochimica Acta 45 (28) (2000a) 4683-4690.
Fujishima, A., Rao, T. N., Tryk, D. A., Titanium dioxide photocatalysis, Journal of Photochemistry and Photobiology C: Photochemistry Reviews 1 (1) (2000b) 1-21.
164
Fujita, M., Ike, M., Nishimoto, S., Takahashi, K., Kashiwa, M., Isolation and Characterisation of a Novel Selenate -Reducing Bacterium, Bacillus sp. SF-1, Journal of Fermentation Bioengineering 83 (1997) 517-522.
Garcia, V. M., Nair, P. K., Nair, M. T. S., Copper Selenide Thin Films by Chemical Bath Deposition, Journal of Crystal Growth 203 (1999) 113-124.
Gard, F. S., Riley, J. D., Dowsett, M. G., Prince, K., The effect of Se and Zn pre-deposition on thermal diffusion of elements across the ZnSe/GaAs interface studied by SIMS, Applied Surface Science 203-204 (2003) 490-494.
Gerischer, H., Heller, A., The Role of Oxygen in Photooxidation of Organic Molecules on Semiconductor Particles, Journal of Physical Chemistry 95 (1991) 5261-5267.
Gimenez, J., Aguado, M.A., Cervera-March, S., Photocatalytic Reduction of Chromium(VI) with Titania Powders in a Flow System. Kinetics and Catalyst Activity, Journal of Molecular Catalysis A: Chemical 105 (1996) 67-78.
Greenwood, N. N., Earnshaw, A., Chapter 16: Selenium, Tellurium and Polonium, In: Greenwood NN, Earnshaw, A., editor. Chemistry of the Elements. Second ed: Butterworth- Heinemann; (1997) 747-788.
Gregg, S. J., Sing, K.S.W., The Physical Adsorption of Gases by Nonporous Solids: The type II Isotherm, In: Gregg SJ, Sing, K.S.W., editor. Adsorption, Surface Area and Porosity. United Kingdom: Academic Press; (1982).
Guo, L., Frankenberger Jr., W. T., Jury, W.A., Evaluation of Simultaneous Reduction and Transport of Selenium in Saturated Soil Columns, Water Resources Research 35 (3) (1999) 663-669.
Hagfeldt, A., Gratzel, M., Light-induced redox Reactions in Nanocrystalline Systems, Chemical Review 95 (1) (1995) 49-68.
Hatchard, C. G., Parker, C. A., A New Sensitive Chemical Actinometer II. Potassium Ferrioxalate as a Standard Chemical Actinometer, Proc. Roy. Soc. (1956) 518-536.
Hecht, M., Fawcett, W. R., Solvent Effects in the Electroreduction of Ethylenediaminetetraacetatochromium(III) at a Mercury Electrode, Journal of Physical Chemistry 99 (1995) 1311-1316.
Henglein, A., Nanoclusters of Semiconductor and Metals: Colloidal Nano-particles of Semiconductors and Metals: Electronic Structure and Processes, Ber. Bunsenges. Physical Chemistry 11 (1997) 1562-1572.
Herrmann, J.-M., Mozzanega, M.-N., Pichat, P., Oxidation of Oxalic Acid in Aqueous
165 Suspensions of Semiconductors illuminated with UV or Visible Light, Journal of Photochemistry: Chemistry 22 (1983) 333-343.
Herrmann, J. M., Heterogeneous Photocatalysis: Fundamentals and Applications to the removal of Various types of Aqueous Pollutants, Catalysis Today 53 (1999) 115-129.
Hoffmann, M. R., Martin, S. T., Choi, W., Bahnemann, D. W., Environmental Applications of Semiconductor Photocatalysis, Chemistry Review 95 (1995) 69-96.
Hoffmann, J. E., King, M. G., Selenium and Selenium Compounds, In: Kroschwitz JIH-G, M., editor. Encyclopedia of Chemical Technology. 4th Edition ed. Volume 21, Kirk- Othmer encyclopedia of chemical technology. New York: Wiley; (1998) 686-719.
Holtzclaw, K. M., Neal, R.H., Sposito, G., Traina, S.J., A Sensitive Colorimetry Method for the Quantification of Selenite in Soil Solutions and Natural Waters, Soil Science Society 51 (1987) 75-78.
Howe, R. F., Gratzel, M., EPR Observation of Trapped Electrons in Colloidal TiO2, Journal of Physical Chemistry 89 (1985) 4495-4499.
Howe, R. F., Gratzel, M, EPR Study of Hydrated Anatase under UV Irradiation, Journal of Physical Chemistry 91 (1987) 3906-3909.
Huang, M., Tso, E., Datye, A.K., Prairie, M.R., Stange, B.M., Removal of Silver in Photographic Processing Waste by TiO2-Based Photocatalysis, Environmental Science Technology 30 (1996) 3084-3088.
Hufschmidt, D., Bahnemann, D., Testa, J. J., Emilio, C. A., Litter, M. I., Enhancement of the photocatalytic activity of various TiO2 materials by Platinisation, Journal of Photochemistry and Photobiology A: Chemistry 148 (1-3) (2002) 223-231.
Ike, M., Takahashi, K., Fujita, T., Kashiwa, M., Fujita, M., Selenate Reduction by Bacteria Isolated from Aquatic Environment Free from Selenium Contamination, Water Research 34 (11) (2000) 3019-3025.
Jones, M. J., French, R. Local Government Engineering in Australia: The Federation Press; 1999 140-141.
Kaise, M., Nagai, H., Tokuhashi, K., Kondo, S., Nimura, S., Kikuchi, O., Electron Spin Resonance Studies of Photocatalytic Interface Reactions of Suspended M/TiO2 (M = Pt, Pd, Ir, Rh, Os, or Ru) with Alcohol and Acetic Acid in Aqueous Media, Langmuir 10 (5) (1994) 1345 - 1347.
Kamat, P. V., Photoelectrochemistry in Particulate Systems. 9. Photosensitized Reduction in a Colloidal TiO2 System using Anthracene-9-carboxylic Acid as the Sensitizer, journal of Physical Chemistry 1989 (93) (1989) 859-864.
166
Kamat, A. D. a. P. V., Semiconductor-Metal Nanocomposites. Photoinduced Fusion and Photocatalysis of Gold-Capped TiO2 (TiO2/Gold) Nanoparticles, Journal of Physical Chemistry B 105 (2001) 960-966.
Kapoor, A., Tanjore, T., Viraraghavan, T., Removal of Se from Water and Wastewater, Environmental Science Technology 49 (2) (1995) 137-147.
Kasap, S. O., Baxendale, M., Juhasz, C., Xerographic Properties of a-Se:Te Photoconductors, IEEE Transactions on Industry applications 27 (4) (1991) 620-626.
Kasap, S. O., Haugen, C., Nesdoly, M., Rowlands, J. A., Properties of a-Se for use in flat panel X-ray image detectors, Journal of Non-Crystalline Solids 266-269 (2000) 1163- 1167.
Khalil, L. B., Mourad, W.E., Ropheal, M.W., Photocatalytic Reduction of Environmental Pollutant Cr(VI) over some Semiconductor under UV/visible light Illumination, Applied Catalysis B: Environmental 17 (1998) 267-273.
Khalil, L. B., Ropheal, M.W., Mourad, W.E., The Removal of the Toxic Hg(II) salts from Water by Photocatalysis, Applied Catalysis B: Environmental 36 (2002) 125-130.
Khanna, P. K., Gorte, R. M., Morley, C. P., Synthesis of Cadmium Selenide by use of Selenadiazol in Ethylene Glycol - A Greener Source of Selenium, Materials Letters 57 (2003) 1464-1466.
Kharaka, Y. K., Removal of Selenium from Contaminated Agricultural Drainage Water by Nanofiltration Membranes, Applied Geochemistry 11 (1996) 797-802.
Kikuchi, E., Sakamoto, H., Kinetics of the Reduction Reaction of Selenate Ions by TiO2 Photocatalyst, Journal of the Electrochemical Society 147 (12) (2000) 4589-4593.
Kohno, Y., Hayashi, H., Takenaka, S., Tanaka, Tsunehiro, Funabiki, T., Yoshida, S., Photoenhanced Reduction of Carbon Dioxide with Hydrogen over Rh/TiO2, Journal of Photochemistry and Photobiology A: Chemistry 126 (1999) 117-123.
Kormann, C., Bahnemann, D. W., Hoffmann, M. R., Photolysis of Chloroform and other Organic Molecules in Aqueous TiO2 Suspensions, Environmental Science & Technology 25 (1991) 494-500.
Ku, Y., Jung, I.L., Photocatalytic Reduction of Cr(VI) in Aqueous solutions by UV Irradiation with the Presence of Titanium Dioxide, Water Research 35 (1) (2001) 135-142.
- Kudo, A., Domen, K., Maruya, K., Onishi, T., Photocatalytic Reduction of NO3 to form NH3 over Pt-TiO2, Chemistry Letters (1987) 1019-1022.
167 Langner, B. E., Selenium and Selenium Compounds, In: Wolfgang Gerhartz ; editor. Ullmann's encyclopedia of industrial chemistry. Volume A23. Weinheim, Federal Republic of Germany ; Deerfield Beach, FL, USA; (1996) 525-536.
Larson, S. A., Widegren, J. A., Falconer, J. L., Transient Studies of 2-Propanol Photocatalytic Oxidation on Titania, Journal of Catalysis 157 (1995) 611-625.
Lea, J., Adesina, A.A., The Photo-oxidative Degradation of Sodium Dodecyl Sulphate in Aerated Aqueous TiO2 Suspension, Journal of Photochemistry and Photobiology A: Chemistry 118 (1998) 111-122.
Li, B., Xie, Y., Huang, J., Qian, Y., Sonochemical synthesis of silver, copper and lead selenides, Ultrasonics Sonochemistry 6 (4) (1999) 217-220.
3+ Li, X. Z., Li, F.B., Study of Au/Au -TiO2 Photocatalysts toward Visible Photooxidation for Water and Wastewater Treatment, Environmental Science Technology 35 (11) (2001) 2381-2387.
Lien, R. Chemical and Biological Cyanide Destruction and Se Removal from Precious Metals Tailings and Pond Water. In: Gaskell DR, editor; 1990; Warrendale, PA. TMS. p 323-329.
Lindstrom, K., Selenium as a growth factor for plankton algae in laboratory experiments and in some Swedish lakes, Hydrobiologia 101 (1983) 35-48.
Litter, M. I., Review: Heterogeneous Photocatalysis Transition Metals Ions in Photocatalytic Systems, Applied Catalysis B: Environmental 23 (1999) 89-114.
Loretto, M. H., Chapter 4. Interpretation of Diffraction Information, In: Loretto, M. H., editor. Electron Beam Analysis of Materials. Second Edition ed. London: Chapman & Hall; (1994).
Loughran, O. Nanotechnology: Molecular Self-assembly - the Chemical Engineer's Final Frontier. The Chemical Engineer 2001 March: 38-40.
Lu, G. Q., Editorial: Preface-Nanomaterials in Catalysis, Catalysis Today 68 (1) (2001) 1.
Lykins, B., Clark, R., U.S. Drinking Water Regulations: Treatment Technologies and Cost, Journal of Environmental Engineering 120 (4) (1994) 783-802.
Maiers, D. T., Wichlacz, P. L., Thompson, D. L., Bruhn, D. F., Selenate Reduction by Bacteria from A Selenium-Rich Environment, Applied Environmental Microbiology 54 (1988) 2591-2593.
Maillard-Dupuy, C., Guillard, C., Pichat, P., The Degradation of Nitrobenzene in Water by Photocatalysis over TiO2: Kinetics and Products; Simultaneous Elimination of Benzamide
168 or Phenol or Pb2+ Cations, New Journal of Chemistry 18 (8-9) (1994) 941-948.
Maneval, J.; 1985. Se Removal from Drinking Water by Ion Exchange. Washington DC: US-EPA. Report nr EPA/600/2-85/074.
Martin, S. T., Herrmann, H., Hoffmann, M. R., Time-resolved Microwave Conductivity. Part 2.-Quantum-sized TiO2 and the effect of Adsorbates and Light Intensity on Charge- carrier Dynamic, Journal of Chemical Society Faraday Transactions 90 (21) (1994) 3323- 3330.
Masscheleyn, P. H., Delaune, R.D, Patrick Jr, H.P., Transformation of Selenium as Affected by Sediment Oxidation-Reduction Potential and pH, Environmental Science & Technology 24 (1990) 91-96.
Matthews, R. W., Hydroxylation Reactions Induced by Near-Ultraviolet Photolysis of Aqueous Titanium Dioxide Suspensions, Journal of Chemical Society Faraday Transactions 1 80 (1984) 457-71.
Matthews, R. W., Kinetics of Photocatalytic Oxidation of Organic Solutes over Titanium Dioxide, Journal of Catalysis 111 (1988) 264-272.
Matthews, R. W., Photocatalysis in Water Purification: Possibilities, Problems and Prospects, In: D F Ollis HA-E, editor. Photocatalytic Purification and Treatment of Water and Air: Elsevier Science Publisher (1993).
McGrew, D. T., Murphy, J., Williams, D. Se Reduction via Conventional Water Treatment, 1996; Denver CO., 129-141.
Memming, R., Photochemical Utilization of Solar Energy. Photochemistry and Photophysics. Volume 3. Florida: CRC Press; (1990).
Merrill, D. T., Manzione, M., Parker, D., Petersen, J., Crow, W., Hobbs, A., Field Evaluation of As and Se Removal by Iron Coprecipitation, Journal of Water Pollution Control Federation 58 (1) (1996) 82-89.
Micic, O. I., Zhang, Y., Cromack, K.R., Trifunac, A.D., Thurnauer, M.C., Photoinduced Hole Transfer from TiO2 to Methanol Molecules in Aqueous Solutions Studied by Electron Paramagnetic Resonance, Journal of Physical Chemistry 97 (1993a) 13284-13288.
Micic, O. I., Zhang, Y., Cromack, K.R., Trifunac, A.D., Thurnauer, M.C., Trapped Holes on TiO2 Colloids Studied by Electron Paramagnetic Resonance, Journal of Physical Chemistry 97 (1993b) 7277-7283.
Mills, A., Le Hunte, S., An Overview of Semiconductor Photocatalysis, Journal of Photochemistry and Photobiology A: Chemistry 108 (1997) 1-35.
169 Mineral Galleries website, “The Mineral Native Selenium”, retrieved 17th April 2003. (http://mineral.galleries.com/minerals/elements/selenium/selenium.htm.)
Mu, W., Herrmann, J. M., Pichat, P., Room Temperature Photocatalytic Oxidation of Liquid Cyclohexane into Cyclohexanone over Neat and Modified TiO2, Catalysis Letters 3 (1989) 73-84.
Muggli, D. S., Falconer, J. L., UV-Enhanced Exchange of O2 with H2O Adsorbed on TiO2, Journal of Catalysis 181 (1999) 155-159.
Muggli, D. S., Odland, M. J., Schmidt, L. R., Effect of Trichloroethylene on the Photocatalytic Oxidation of Methanol on TiO2, Journal of Catalysis 203 (2001) 51-63.
Muggli, D. S., Backes, M. J., Two Active Sites for Photocatalytic Oxidation of Formic Acid on TiO2: Effects of H2O and Temperature, Journal of Catalysis 209 (2002) 105-113.
Murov, S. L., Carmichael, I., Hug, G.L., Chp 13: Chemical Actonimetry. Handbook of Photochemistry. New York: Marcel Dekker, Inc; (1993). p 298-305.
Murphy, A., Removal of Selenate from Water by Chemical Reduction, Industrial Engineering and Chemistry Research 27 (1988) 187-191.
Neuhauser, R. G., Photoconductors Utilized in TV Camera Tubes, Journal of The Electrochemical Society 134 (1) (1987) 5C-13C.
Nishimura, T., Umetsu, Y., Chemistry on Elimination of Arsenic, Antimony and Selenium from Aqueous Solution with Iron(III) Species, In: Young C, editor. Minor Elements 2000: Processing and Environmental Aspects of As, Sb, Se, Te and Bi: Society for MIning, Metallurgy and Exploration Inc; (Feb 2000) 105-111.
Nishimura, T., Hata, R., Umetsu, Y., Removal of Selenium from Industrial Waste Water, In: Young C, editor. Minor Elements 2000 Processing and Environmental Aspects of As, Sb, Se, Te and Bi: Society for Mining, Metallurgy and Exploration, Inc.; (Feb 2000). 355- 361.
O'Shea, K. E., Pernas, E., Saiers, J., The Influence of Mineralization Products on the Coagulation of TiO2 Photocatalyst, langmuir 15 (1999) 2071-2076.
Oldfield, J. E., SeRendipity, Chemtech March (1995) 52-55.
Parker, C. A., A New Sensitive Chemical Actinometer I. Some Trials with Potassium Ferrioxalate, Proc. Roy. Soc. 220 (1953) 104-116.
Parmon, V., Emeline, A.V., Serpone, N., Glossary of Terms in Photocatalysis and Radiocatalysis, International Journal of Photoenergy 4 (2002) 91-131.
170 Pejova, B., Grozdanov, I., Solution growth and characterization of amorphous selenium thin films. Heat transformation to nanocrystalline gray selenium thin films, Applied Surface Science 117 (2001) 152-157.
Pelizzetti, E. M., Valter; Minero, Claudio; Carlin, Vilma; Pramauro, Edmondo; Zerbinati, Orfeo; Tosato, Maria L., Photocatalytic degradation of atrazine and other s-triazine herbicides, Environmental Science and Technology 24 (10) (1990) 1559-1565.
Penpolcharoen, M., Amal, R., Brungs, M., Degradation of sucrose and nitrate over titania coated nano-hematite photocatalysts, Journal of Nanoparticle Research (2001) 289-302.
Peñuela, G. A., Barceló, D., Photosensitized degradation of organic pollutants in water: processes and analytical applications, Trends in Analytical Chemistry 17 (10) (1998) 605- 612.
Perissinotti, L. L., Brusa, M. A., Grela, M. A., Yield of Carboxyl Anion Radicals in the Photocatalytic Degradation of Formate over TiO2 Particles, Langmuir 17 (2001) 8422- 8427.
Peters, G. M., Maher, W.A., Jolley, D., Carroll, B.I., Gomes, V.G., Jenkinson, A.V., McOrist, G.D., Selenium Contamination, Redistribution and Remobilisation in Sediments of Lake Macquarie, NSW., Organic Geochemistry 30 (1999) 1287-1300.
Pollema, C. H., HendrixEmil, J. L., Milosavljevi, B., Soluji, L., Nelson, J. H., Photocatalytic oxidation of cyanide to nitrate at TiO2 particles, Journal of Photochemistry and Photobiology A: Chemistry 66 (2) (1992) 235-244.
Pontius, F., An Update of the Federal Drinking Water Regulations, Journal of the American Water Works Association 87 (2) (1995) 48-58.
Pourbaix, M., Atlas of electrochemical equilibria in aqueous solutions. Oxford, New York: Pergamon Press; (1966).
Pozzo, R. L., Baltanás, M. A., Cassano, A. E., Supported titanium oxide as photocatalyst in water decontamination: State of the art, Catalysis Today 39 (3) (1997) 219-231.
Prairie, M. R., Evans, L., R., Stange, B. M., Martinez, S. L., An Investigation of TiO2 Photocatalysis for the Treatment of Water Contaminated with Metals and Organic Chemicals, Environment Science and Technology 27 (9) (1993) 1776-1782.
Qiu, S. R., Lai, H.F., Roberson, J., Hunt, M.L., Amrhein, C., Giancarlo, L.C., Flynn, G.W., Yarmoff, J.A., Removal of Contaminants from Aqueous Solution by Reaction with Iron Surfaces, Langmuir 16 (5) (2000) 2230-2236.
Rajeshwar, K., Ibanez, J.G., Chapter 2: Semiconductor Electrochemistry. Environmental Electrochemistry: Fundamentals and Applications in Pollution Abatement: Academic
171 Press; (1998) 116-124.
Ranjit, K. T., Varadarajan, T. K., Viswanathan, B., Photocatalytic Reduction of Nitrite and Nitrate ions to Ammonia on Ru/TiO2 Catalysts, Journal of Photochemistry and Photobiology A: Chemistry 89 (1995) 67-68.
Ranjit, K. T., Viswanathan, B., Photocatalytic Reduction of Nitrite and Nitrate ions to Ammonia on M/TiO2 Catalysts, Journal of Photochemistry and Photobiology A: Chemistry 108 (1997) 73-78.
Reddy, K. J., Selenium Speciation in Natural Waters, In: Myers RA, editor. Encyclopaedia of environmental analysis and remediation. Volume 7. New York: Wiley; (1998) 4291- 4300.
2- Refait, P., Simon, L., Genin, J.R., Reduction of SeO4 Anions and Anoxic Formation of Iron(II)-Iron(III) Hydroxy-Selenate Green Rust, Environmental Science Technology 34 (5) (2000) 819-825.
Reilly, C., Chapter 1: Introduction, In: Reilly C, editor. Selenium in Food and Health. London: Blackie Academic & Professional (An Imprint of Chapman & Hall) (1996).
Richard, C., Photocatalytic Reduction of Benzoquinone in Aqueous ZnO or TiO2 Suspensions, New Journal of Chemistry 18 (1994) 443-445.
Rideh, L., Wehrer, A., Ronze, D., Zoulalian, A., Modelling of Kinetic of 2-Chlorophenol Catalytic Photooxidation, Catalysis Today 48 (1999) 357-362.
Robertson, P. K. J., Semiconductor photocatalysis: an environmentally acceptable alternative production technique and effluent treatment process, Journal of cleaner production 4 (3-4) (1996) 203-212.
Rophael, M. W., Malati, M.A., The Photocatalysed Reduction of Aqueous Sodium Carbonate to Carbon Using Platinised Titania, Journal of Chemical Society Communication (1987) 1418-1420.
Rowlands, J., Kasap, S., Amorphous Semiconductors Usher in Digital X-Ray Imaging, Physics Today (1997) 24-30.
Rozelle, L. T., Point-of-use/point-of-entry Drinking Water treatmenrt: General State of the Art. Our Water Makes the Difference. Kansas City MO: American Water Works Association; (1987) 779-1802.
Sahyun, M. R. V., Serpone, N., Primary events in the photocatalytic deposition of silver on nanoparticulate TiO2, Langmuir 13 (19) (1997) 5082-5088.
Sanuki, S., Kojima, T., Arai, K., Nagaoka, S., Majima, H., Photocatalytic Reduction of
172 Selenate and Selenite Solutions using TiO2 Powders, Metallurgical and Materials Transactions B 30B (1999) 15-20.
Sanuki, S., Shako, K., Nagaoka, S., Majima, H., Photocatalytic Reduction of Se ions Using Suspended Anatase Powders, Materials Transaction, JIM 41 (7) (2000) 799-805.
Sclafani, A., Mozzanega, M.-N., Pichat, P., Effect of silver deposits on the photocatalytic activity of titanium dioxide samples for the dehydrogenation or oxidation of 2-propanol, Journal of Photochemistry and Photobiology. A. Chemistry 59 (2) (1991) 181-189.
Seby, F., Potin-Gautier, M., Giffaut, E., Borge, G., Donard, O.F.X., A Critical Review of Thermodynamics data for Selenium Species at 25C, Chemical Geology 171 (2001) 173- 191.
Serpone, N., Sauvé, G., Koch, R., Tahiri, H., Pichat, P., Piccinini, P., Pelizzetti, E., Hidaka, H., Standardization protocol of process efficiencies and activation parameters in heterogeneous photocatalysis: relative photonic efficiencies ζ, Journal of Photochemistry and Photobiology A: Chemistry 94 (2-3) (1996) 191-203.
Shaker, A. M., Kinetics of the Reduction of Se(IV) to Se-Sol, Journal of Colloid and Interface Science 180 (1996) 225-231.
Singh, M., Singh, N., Relan, P. S., Adsorption and desorption of selenite and selenate selenium on different soils, Soil Science 132 (2) (1981) 134-141.
Sorg, T. T., Logsdon, G. S., Treatment Technology to Meet the Interim Primary Drinking Water Regulations for Inorganics: Part 2, Journal of American Water Works (1978) 379- 393.
Spanhel, L., Weller, H., Henglein, A., Photochemistry of Semiconductor Colloids. 22. Electron Injection from Illuminated CdS into Attached TiO2 and ZnO particles, Journal of American Chemical Society 109 (1987) 6632-6635.
Stookey, L., Ferrozine - A new Spectrophotometric reagent for Iron, Analytical Chemistry 42 (7) (1970) 779-781.
Streltsov, E. A., Poznyak, S.K., Osipovich, N.P., Photoinduced and Dark Underpotential Deposition of Lead on Selenium, Journal of Electroanalytical Chemistry 518 (2002) 103- 114.
Surender, G. D., Fotou, G.P., Pratsinis, S.E., Titania-based Aqueous Phase Photocatalytic Remediation Processes: Strategies for Enhancement of Quantum Efficiency, Trends in Chemical Engineering 4 (1998) 145-159.
Tada, H., Teranishi, K., Inubushi, Y., Ito, S., TiO2 photocatalytic Reduction of bis(2- dipyridyl)disulfide to 2-mercaptopyridine by H2O: Incorporation effect of Nanometer-sized
173 Ag Particles, Chemical Communication (1998) 2345-2346.
Tada, H., Teranishi, K., Inubushi, Y., Ito, S., Ag Nanocluster Loading Effect on TiO2 Photocatalytic Reduction of Bis(2-dipyridyl)disulfide to 2-mercaptopyridine by H2O, Langmuir 16 (2000) 3304-3309.
Tada, H., Suzuki, F., Yoneda, S., Ito, S., Kobayashi, H., The effect of nanometre-sized Au particle loading on TiO2 photocatalysed reduction of bis(2-dipyridyl)disulfide to 2- mercaptopyridine by H2O, Phys. Chem. Chem. Phys. 3 (7) (2001) 1376-1382.
Terzian, R., Serpone, N., Minero, C., Pelizzeyyi, E., Photocatalyzed Mineralization of Cresols in Aqueous Media with Irradiated Titania, Journal of Catalysis 128 (1991) 352- 365.
Thiel, C. W, An Introduction to Semiconductor Radiation Detectors. Bozeman: Montana State University (1999) 1-20.
Trussell, R. A., Trussell, A., Kraft, P. Se Removal with Activated Alumina. AWWA Research FOundation Water Quality Research News (1991) 4-5.
Tsao, R., Eto, M., Effect of some Natural Photosensitizers on Photolysis of some Pesticides, In: G.R. Helz, R.G. Zepp, editor. Aquatic and Surface Photochemistry: Lewis Publishers; (1994) 163-171.
Tunesi, S., Anderson, M., Influence of Chemisorption on the Photodecomposition of Salicylic Acid and Related Compounds using Suspended TiO2 Ceramic Membranes, 1991 95 (1991) 3399-3405.
Turchi, C. S., Ollis, D.F., Photocatalytic Degradation of Organic Water Contaminants: Mechanisms Involving Hydroxyl Radical Attack, Journal of Catalysis 122 (1990) 178-192.
Turner, M. A., Rudd, J. W. M., The English- Wabigoon River system: III. Selenium in lake enclosures: its geochemistry, bioaccumulation, and ability to reduce mercury bioaccumulation, Canadian Journal of Fisheries & Aquatic Sciences 40 (12) (1983) 2228- 2240.
Umehara, A., Nitta, S., Furukawa, H., Nonomura, S., Preparation and Properties of Nanocrystalline Semiconductor Selenium Films, Applied Surface Science 119 (1997) 176- 180.
Vamathevan, V., Tse, H., Amal, R., Low, G., McEvoy, S., Effects of Fe3+ and Ag+ ions on the Photocatalytic Degradation of Sucrose in Water, Catalysis Today 68 (2001) 201-208.
Vogel, R., Hoyer, P., Weller, H., Quantum-sized PbS, CdS, Ag2S, Sb2S3, and BiS3 particles as Sensitizers for Various Nanoporous Wide-Bandgap Semiconductors, journal of Physical Chemistry 98 (1994) 3183-3188.
174
Wang, C.M., Heller, A., Gerischer, H., Palladium Catalysis of O2 Reduction by Electrons Accumulated on TiO2 Particles During Photoassisted Oxidation of Organic Compounds, Journal of American Chemical Society 114 (13) (1992) 5230-5234.
Wang, Z. H., Zhuang, Q.X., Photocatalytic reduction of pollutant Hg(II) on doped WO3 dispersion, Journal of Photochemistry and Photobiology A. 75 (1993) 105-111.
Wang, D., Alfthan, G., Aro, A., Makela, A., Knuuttila, S., Hammar, T., The Impact of Selenium supplemented Fertilization on Selenium in Lake Ecosystems in Finland, Agriculture Ecosystems & Environment 54 (1995) 137-148.
Ward, M. D., White, J. R., Bard, A. J., Electrochemical Investigation of the Energetics of particulate Titanium Dioxide Photocatalysts. The Methyl Viologen-Acetate System, Journal of the American Chemical Society 105 (1983) 27-31.
Weller, H., Eychmuller, A., Photochemistry and Photoelectrochemistry of Quantized Matter: Properties of Semiconductor Nanoparticles in Solution and Thin-film Electrodes, In: Neckers DC, Volman, D.H., von Bunau, G., editor. Advances in Photochemistry. Volume 20: John Wiley and Sons; (1995) 165-216.
+ - Xi, C., Chen, Z., Li, Q., and Jin, Z., Effects of H, Cl and CH3COOH on the 2- photocatalytic conversion of PtCl6 in aqueous TiO2 dispersion, Journal of Photochemistry and Photobiology A: Chemistry 87 (3) (1995) 249-255.
Xie, T. F., Wang, D.J., Zhu, L.J., Li, T.J., Xu, Y.J., Application of Surface Photovoltage Technique in Photocatalysis Studies on Modified TiO2 Photocatalysts for Photoreduction of CO2, Materials Chemistry and Physics 70 (2001) 103-106.
Yamashita, H., Harada, M., Misaka, J., Takeuchi, M., Ikeue, K., Anpo, M., Degradation of Propanol Diluted in water under Visible Light Irradiation using Metal Ion-implanted Titanium Dioxide Photocatalysts, Journal of Photochemistry and Photobiology A; Chemistry 6003 (2002) 1-5.
Yang, J. C., Kim, Y.C., Shul, Y.G., Shin, C.H., Lee, T.K., Characterization of Photoreduced Pt/TiO2 and Decomposition of Dichloroacetic Acid over Photoreduced Pt/TiO2 Catalysts, Applied Surface Science 121/122 (1997) 525-529.
Yoneyama, H., Yamashita, Y., Tamura, H., Heterogeneous Phtocatalytic Reduction of Dichromate on n-type Semiconductor Catalysts, Nature 282 (20/27) (1979) 817-818.
Yoneyama, H., Electrochemical Aspects of Light-induced Heterogeneous Reactions on Semiconductors, Critical Reviews in Solid State and Material Sciences 18 (1) (1993) 69- 111.
Yoneyama, H., Photoreduction of carbon dioxide on quantized semiconductor nanoparticles
175 in solution, Catalysis Today 39 (3) (1997) 169-175.
Zhang, P., Sparks, D. L., Kinetics of Selenate and Selenite Adsorption/Desorption at the Goethite/Water Interface, Environmental Science & Technology 24 (1990) 1848-1856.
Zhang, Y., Moore, J. N., Selenium Fractionation and Speciation in a Wetland System, Environmental Science & Technology 30 (1996) 2613-2619.
Zhang, F., J. Zhao, L. Z., Shen, T., Hidaka, H., Pelizzetti, E., Serpone, N., Photoassisted degradation of dye pollutants in aqueous TiO2 dispersions under irradiation by visible light, Journal of Molecular Catalysis A: Chemical 120 (1-3) (1997) 173-178.
Zhang, Y., Moore, J.N., Interaction of Selenate with a Wetland Sediment, Applied Geochemistry 12 (1997) 685 - 691.
Zingaro, R. A., Dufner, D.C., Murphy, A.P, Moody, C.D., Reduction of Oxoselenium Anions by Iron(II) Hydroxide, Environmental International 23 (3) (1997) 299 - 304.
Znaidi, L., Séraphimova, R., Bocquet, J. F., Colbeau-Justin, C., Pommier, C., A semi- continuous process for the synthesis of nanosize TiO2 powders and their use as photocatalysts, Materials research bulletin 36 (2001) 811-825.
176 Appendix
A. Derivation of Se(VI) and formate adsorption model
The adsorption of Se(VI) (represented by Se) onto 2 adsorption sites (represented by S) to yield adsorbed Se(VI) ion (represented by SeSS) is given as:
2 «+ SeSSe SS
Similarly, the adsorption of formate ion (represented by F) onto 1 adsorption site to yield adsorbed formate ion (represented by FS) is given as:
«+ FSF S
Assuming equilibrium Se(VI) adsorption,
2 - = - CkCCkr =- 0 SeSevSeSeSeAD SS
where rAD-Se represents the rate Se(VI) ions adsorbed onto the TiO2 surface, kSe and k-Se are the forward and reversed adsorption rates respectively, CV represents the vacant sites available on the catalyst surface, and CSe and CF are the residual Se(VI) and formic acid concentrations at equilibrium respectively.
For Se(VI) adsorption, simplifying the above expression gives,
=\ 2CCKC SeSS SevSe (I)
Similarly, for formate adsorption,
=\ CCKC FS FvF (II)
177
Performing a site balance,
++= CCCC Sevt SS FS
--=\ CCCC Setv SS FS (III)
where Ct represents the total sites available on the catalyst surface, CV represents the vacant sites available on the catalyst surface, and CSess and CFs represent the site occupied by Se(VI) and formate ions respectively.
A. To derive Se(VI) adsorption onto 2 sites in the presence of formate ions:
Subst (III) into (II) and making CFs the subject:
( - CCCK ) C =\ SetFF SS FS 1+ CK FF (IV)
Subst (IV) and (III) into (I):
2 2 C é ( - CCCK )ù æ - CC ö SeSS SetFF SS ç Set SS ÷ ê CC Set --= ú = CK SS 1+ CK ç1+ CK ÷ SeSe ë FF û è FF ø
(1+ CK )2 \ FF 2 2 +-= CCCCC 2 CK SeSS Sett SS SeSS SeSe
2 2 é (1+ CK ) ù ê2CC +-\ FF ú CC 2 =+ 0 SeSS t CK SeSS t ë SeSe û
2 For cbxax =++ 0
2 -±- 4acbb x = 2a
178 2 é 2ù é 2ù ()1+ CK FF ()1+ CK FF 2 ê2Ct + ú ê2Ct +± ú - 4Ct ë CK SeSe û ë CK SeSe û C =\ SeSS 2
Using a2–b2 =(a+b)(a-b) for the expressions in the square root, the adsorption of Se(VI) onto two sites in the presence of formate ions is derived as follows:
ì 2 2 2 ü 1 ïé()1+CK ùé()1+CK ùé()1+CK ùï C =íê2C +FF ú±ê4C +FF úêFF úý SeSS 2 t CK t CK CK îïë SeSe ûë SeSe ûëSeSe ûþï
B. To derive formate adsorption onto 1 site in the presence of Se(VI) ions:
Substitute (III) into (I)
= ( -- CCCCKC )2 SeSS setSese SS FS using (a-b-c)2= a2+b2+c2-2ab-2ac+2bc,
C SeSS 2 2 2 +--++= 222 CCCCCCCCC CK set SS FS set SS Ft S Fse SSS Sese
2æ1ö2 \+ç2-2CCC -÷+()2-2+CCCCC =0 SeSS çFS t CK ÷SeSS SFFtt S èSeSe ø
2æ 1ö 2 \+ç2-2CCC -÷+()-CCC =0 SeSS çFS t CK ÷SeSS Ft S è SeSe ø
2 æ 1ö æ 1ö 2 ç22CC --- ÷ ç22CC --± ÷ 4()-- CC ç FSt ÷ ç FSt ÷ Ft S è CK SeSe ø è CK SeSe ø CSe =\ SS 2
Again, using a2–b2 =(a+b)(a-b) for the expressions in the square root:
179 æ1öæ1öæ1ö -ç2-2CC -÷±ç-÷ç4-4CC -÷ çFS t ÷ç÷çFS t ÷ èCK SeSe øèCK SeSe øèCK SeSe ø \CSe = SS 2 (V)
Substitute (V) & (III) into (II)
æ æ 1 ö æ 1 öæ 1 ö ö ç ç 22 CC -- ÷ ç-± ÷ç 44 CC -- ÷ ÷ ç ç FS t ÷ ç ÷ç FS t ÷ ÷ è CK SeSe ø è CK SeSe øè CK SeSe ø = ç CCCKC +- ÷ FS ç FtFF S 2 ÷ ç ÷ è ø
2 é æ CF ö 1 ù æ 1 öæ 1 ö ê2çC S -+ ÷ 22 CCC ++- ú ç-= ÷ç 44 CC -- ÷ ç FS t ÷ FS t ç ÷ç FS t ÷ ê CK CK ú CK CK ë è FF ø SeSe û è SeSe øè SeSe ø
2 æ 2CF 1ö 4CF 4C 1 ç S + ÷ S t ++-= ç ÷ 2 CKCK CK CK ()CK è FF SeSe ø SeSe SeSe SeSe
4C 4C 42 FS FS 4Ct CF + +-=0 ()CK 2S CKCK CK CK FF SeSeFF SeSe SeSe
42(4+4CK FF )4Ct CF +CF -=0 ()CK 2S CKCK S CK FF SeSeFF SeSe
222 2(+CKCK )()CCK C +FFFF C -tFF =0 FS CK FS CK SeSe SeSe
Similarly, for ax2+bx+c=0
-±2-4acbb x = 2a
180 2 é 2ù é 2ù 2 FF + ()CKCK FF FF + ()CKCK FF 4()CKC FFt - ê ú ± ê ú + 2CK SeSe ë CK SeSe û ë CK SeSe û ()CK SeSe CF = S 2
(CKCK )[()11 2 ++±-- 4 ]()CKCKCCK 2 C = FFFF FF FFSeSet FS 2 CK SeSe
The adsorption of formate onto one site in the presence of Se(VI) ions is derived as follows:
CK 2 C =FF é(-CK -)±()11 ++4 CKCCK ù FS 2 CK ëêFF FF SeSet ûú SeSe
181 B. Actinometry: Photon Flux and Quantum efficiency determination
Actinometry on mercury lamp
250 y = 12.97x + 99.82
200 2 R = 0.95 150 y = 13.26x + 84.08 R2 = 1.00 100 mg/L Fe(II) 50 A B 0 024681012 Time (min)
Photon flux for ‘A’ = 3.15 mE/sec Photon flux for ‘B’ = 3.17 mE/sec
Calculation of quantum efficiency for unmodified and Ag modified TiO2.
From equation 2-14,
Quantum efficiency for reduction reactions using unmodified TiO2 =
Total mols of Se(VI) reduced and H2Se formed/time= Photon flux
(0.093 + 0.084 mmol)/(420 X 60 sec)= 0.00317 mE/sec =0.22%
Quantum efficiency for reduction reactions using Ag-TiO2 = (0.10 + 0.093 mmol)/(300 X 60 sec)= 0.00317 mE/sec =0.34%
182 Quantum yield (QY) calculations due to rate of electrons utilisation calculations.
Using Photon flux = 3.15 mE/sec =0.19 mmol/min,
For Se(IV):
8.48 mg reduced in 140 minutes. Hence 0.76 X 10-3 mmol/min reduced.
Since 1 mol Se(IV) utilized 4 electrons, 3.04 X 10-3 mmol/min electrons utilized.
Hence, QY= 3.04 X 10-3 mmol/min electrons 0.19 mmol/min =1.6%
For Se(VI):
8.23 mg reduced in 165 minutes. Hence 0.63 X 10-3 mmol/min reduced.
Since 1 mol Se(VI) utilized 6 electrons, 3.79 X 10-3 mmol/min electrons utilized.
Hence, QY= 3.79 X 10-3 mmol/min electrons 0.19 mmol/min =1.9%
183