Open Collections

UBC Theses and Dissertations

UBC Theses Logo

UBC Theses and Dissertations

Development and evaluation of a composite photocatalyst for water treatmment processes Vega, Adrian A. 2009

Your browser doesn't seem to have a PDF viewer, please download the PDF to view this item.

Item Metadata

Download

Media
24-ubc_2009_fall_vega_adrian.pdf [ 8.92MB ]
Metadata
JSON: 24-1.0058900.json
JSON-LD: 24-1.0058900-ld.json
RDF/XML (Pretty): 24-1.0058900-rdf.xml
RDF/JSON: 24-1.0058900-rdf.json
Turtle: 24-1.0058900-turtle.txt
N-Triples: 24-1.0058900-rdf-ntriples.txt
Original Record: 24-1.0058900-source.json
Full Text
24-1.0058900-fulltext.txt
Citation
24-1.0058900.ris

Full Text

DEVELOPMENT AND EVALUATION OF A COMPOSITE PHOTOCATALYST FOR WATER TREATMENT PROCESSES  by Adriân A. Vega  B.Sc, University of Costa Rica, 2005 ATHESIS SUBMITTED IN PARTIAL FULFILMENT OF THE REQUIREMENTS FOR THE DEGREE OF MASTER OF APPLIED SCIENCE in THE FACULTY OF GRADUATE STUDIES (Chemical and Biological Engineering) THE UNIVERSITY OF BRITISH COLUMBIA (Vancouver) August 2009 ©Adriân A. Vega, 2009  Abstract Titanium dioxide (Ti0 2 ) has been extensively studied for photocatalytic oxidation of organic pollutants, in water and gas effluents, into smaller and harmless substances such as C0 2 and H 2 0. However, there are still a number of drawbacks that hinder its large scale applications, for example expensive downstream filtration of the photocatalyst powder, mass transfer limitations, loss of activity, attrition, among others. Fluidized  bed photoreactor (FBPR) with its advantages for potential  commercial applications, suffers from the problem of attrition and elutriation of particles. This research focused on addressing this challenge and worked on the development of a template-free composite photocatalyst to be used in a FBPR. It involved the production of Ti0 2 nanoparticles, using sol-gel process, and mixing them with Ti0 2 pre-calcined powder (Degussa P-25), used as filler material bonded together with the gel-derived Ti0 2 . This solution was then mixed with a polymeric structure to produce the sphères. A complète characterization of the Ti0 2 sphères was performed to détermine their morphology (XRD - X Ray Diffraction, SEM Scanning électron microscopy), physicochemical properties (BET surface area, pore volume and pore diameter) and thermal behaviour (TGA - Thermo gravimétrie analysis). The major achievement of this work was to modify the formulation and synthesis procédure of the composite Ti0 2 sphères, producing a photocatalyst with good attrition résistance and high photoactivity. The improvements were made by working on the préparation conditions (modifying the hydrolysis and condensation rates during sol-gel préparation, aging time, sphères formation, drying process, and calcination conditions) and evaluating their impacts on the attrition résistance (amount of Ti0 2 particles released to the solution during the normal opération of the FBPR). Photocatalyst activity was measured based on its efficacy at decomposing organic compounds such as formic acid (FA) and 2,4-Dichlorophenoacetic acid (2,4D), as well as natural organic matter (NOM). ii  The attrition résistance of the TÏO2 sphères was improved by 70%, corresponding to a réduction in the amount of TiC>2 particles released from the catalyst from 22 mg L"1 to 7 mg L"1. This was achieved via the following modifications in the formulation and synthesis procédure:  • Accelerating the rate of hydrolysis and condensation reactions that occur during the formation of the sol-gel matrix, by increasing the amount of water from 0.040 g H2O per mL"1 of Ti precursor to 0.053 g H2O per ml_"1 Ti precursor and reducing the amount of catalyst (HCI) from 0.20 mL HCI per mL'1 Ti precursor to 0.13 mL HCI per mL"1 Ti precursor. .  Increasing pH of the solution, where the sphères are formed, from 11.75 to 12.  • Accelerating the drying process from 15 days at 23°C to 20 h at 80°C. • Increasing the calcination time from 1 h to 3 h.  In terms of the photocatalytic activity, the composite TiÛ2 sphères demonstrated high activity to dégrade either FA or 2,4-D. The dégradation of both model pollutants followed first order kinetics with rates constants of 0.317 min"1 (with fluence rate of 5.32 mW cm"2) and 0.736 min"1 (with a fluence rate of 4.16 mW cm"2) for FA and 2,4-D, respectively. A comparison with Degussa P-25 was made showing that the activity of the Ti02 sphères is higher than that of the commercial Ti0 2 powder for both model organic compounds. For NOM dégradation, water from Trepanier Creek in Central British Columbia (initial TOC of approximately 5 mg L"1) was treated. There was about 50% réduction of NOM after 1 h, but this was primarily due to the adsorption of NOM on the sphères.  iii  Table of Contents Abstract  ii  Table of Contents  iv  List of Tables  viii  List of Figures  xi  Nomenclature  xviii  Acknowledgements  xxvi  Dedication  Quote 1  2  xxviii  xxix  Introduction  1  1.1 Background  1  1.2 General Objective and scope of the présent work  4  1.3 Structure of this thesis  5  Literature Review  7  2.1 Background  7  2.2 Photocatalysis and environmental remediation  8  2.3 Heterogeneous photocatalysis  9  2.3.1 Electronic excitation process in semiconductor materials  10  2.3.2 Titanium Dioxide (Ti0 2 ) as a photocatalyst  16  2.4 Photocatalytic efficiency and kinetic overview  19  2.4.1 Photocatalytic efficiency  19  2.4.2 Kinetics of photocatalyzed reactions  20  2.5 Overview of photocatalytic reactors  24  2.6 Sol-gel process  27  2.6.1 Background and fundamentals  27  2.6.2 Sol-gel process steps  28  2.6.3 Reactions and chemical characteristics  33 iv  3  4  2.6.4 Sol-gel advantages and disadvantages  38  2.6.5 Sol-gel applications  39  2.6.6 Composite materials  40  Scope and Objectives  44  3.1 Scope  44  3.2 Objectives  45  3.3 Significance of this work  46  3.4Layout  46  Expérimental Methodology  49  4.1 Introduction  49  4.2 Composite TIO2 sphères préparation  49  4.2.1 Chemicals and reagents  46  4.2.2 Composite sol-gel material (CSG)  50  4.2.3 Polymeric solution  50  4.2.4 Composite TÏO2 sphères formation  51  4.3 Composite Ti02 sphères characterization  52  4.3.1 Surface area, pore volume and pore size détermination  53  4.3.2 X-Ray diffraction spectroscopy (XRD)  56  4.3.3 Scanning électron microscopy (SEM)  58  4.3.4 Thermo gravimétrie analysis (TGA)  59  4.4 Fluidized bed photoreactor setup  60  4.5 Photocatalytic activity test  62  4.5.1 High performance liquid chromatography analysis (HPLC)  64  4.5.2 Total organic carbon analysis (TOC)  65  4.5.3 Organic pollutants dégradation  65  4.5.3.1 Formic acid (FA) dégradation  65  4.5.3.2 Micropollutant dégradation (2,4-Dichlorophenoacetic acid)  66  4.5.4 Natural organic matter (NOM) dégradation  67  4.6 Attrition résistance détermination  67  4.7 Photocatalyst deactivation and attrition évolution  68 v  4.8 Fluence rate détermination 5  6  Preliminary Evaluation of Composite T1O2 Photocatalyst  69 72  5.1 Introduction  72  5.2 Composite TÏO2 sphères activity  72  5.2.1 Bed expansion estimation  73  5.2.2 Photocatalyst activity at différent calcination températures  74  5.2.3 Mass transfer résistance calculations  79  5.3 Attrition détermination  81  5.4 Composite photocatalyst characterization  82  5.5 Final remarks  87  Photocatalyst Improvement and Characterization  88  6.1 Introduction  88  6.2 Compositional study of the TÏO2 sphères formulation  88  6.2.1 Preliminary tests and effects of various parameters  90  6.2.1.1 Type of catalyst used during the sol-gel préparation  91  6.2.1.2 Water and catalyst (HCI) concentration  93  6.2.1.3 Amountoffillermaterial (Degussa P-25)  95  6.2.1.4 pH of the chitosan solution  96  6.2.1.5 Concentration of chitosan in the polymeric solution  97  6.2.1.6 Drying processof the sphères  99  6.2.1.7 Heat treatment conditions  100  6.2.2 Expérimental design  102  6.2.3 Modified formulation to produce the Ti02 composite sphères  104  6.3 Catalyst characterization  107  6.3.1 XRD analysis  107  6.3.2 BET analysis  110  6.3.3 TGA analysis  113  6.3.4 SEM analysis  114  6.4 Ti0 2 sphères durability and attrition détermination over time  116 vi  6.5 Final remarks 7  Activity of Composite Ti02 Sphères  119  7.1 Introduction  120  7.2 Fluence rate détermination  117  7.3 Photocatalytic dégradation of Formic Acid (FA)  125  7.3.1 Effect of the radiation flux on FA dégradation  125  7.3.2 Deactivation of the Ti0 2 sphères  126  7.3.3 Comparison with Degussa P-25  128  7.4 Photocatalytic dégradation of 2,4-D  133  7.4.1 Adsorption studyof 2,4-D on the composite TÏO2 sphères  134  7.4.2 Effect of the différent levels of irradiation in the 2,4-D dégradation  138  7.4.3 Effect of the initial concentration of 2,4-D in its dégradation  141  7.4.4 Deactivation of the Ti0 2 sphères  147  7.4.5 Comparison with Degussa P-25  149  7.5 Photocatalytic dégradation of natural organic matter (NOM) 7.5.1 Effect of NOM On the dégradation of 2,4-D 7.6 Final remarks 8  118  Conclusions and Recommendations  152 ..154 156 158  8.1 Conclusions  158  8.2 Recommendations for future work  162  Bibliography  165  Appendices  183  AppendixA  183  AppendixB  192  vii  List of Tables Table 2.1  Band gap énergies for some semiconductor materials at OK (Bhatkhande et al., 2001 ; Thiruvenkatachari et al., 2008) 11  Table 2.2  Photocatalytic reaction scheme for the Ti0 2 (Turchi and Ollis, 1989; Turchi and Ollis, 1990; Kabir, 2006 14  Table 2.3 Comparison between suspended and immobilized photocatalytic Systems (De Lasa et al., 2005; Kabir, 2006) 25 Table 4.1 Spécifications of Chemicals used during the préparation of the composite Ti0 2 photocatalyst 49 Table 4.2  Spécifications of chemicals used during the activity test of the CSG photocatalyst 62  Table 4,3 Spécifications of chemicals used for the attrition résistant of the composite Ti0 2 sphères 68 Table 4.4  Spécifications of chemicals used for the fluence rate détermination  69  Table 5.1  Data to calculate the height of the expanded bed in the fluidized bed photocatalytic reactor (FBPR) of sphères at 600°C 4 -3 (rsPh = 5.75 x 10" m; psph = 3109 kg m ) 76  Table 5.2 Data to détermine the activity of sphères at différent calcination températures (msph = 0.025 kg) 78 Table 5.3  FA photodegradation rate constants (kr) of composite Ti0 2 sphères produced at différent calcination températures (mass of catalyst loaded to the reactor equal to 0.025 kg) 79  Table 5.4 Surface area and anatase composition (using BET and XRD respectively) of the composite Ti0 2 sphères calcined at différent températures 84  viii  Table 6.1  Variables involved during the production process of the composite Ti02 sphères 89  Table 6.2  Range of the variables to be study during the preliminary tests with the composite TiÛ2 91  Table 6.3  Fractional factorial design (25"1) to study the effects of and interactions among variables involved in the production of the composite TiÛ2 sphères 103  Table 6.4 Response variables of the fractional factorial design (25"1) for the composite Ti0 2 sphères (mass of catalyst equal to 0.025 Kg) 104 Table 6.5  Variables chosen from the expérimental design analysis  104  Table 6.6 Surface area and anatase composition of pre-calcined powder (Degussa P-25 - as received) and composite TiÛ2 sphères prepared using the new formulation and calcined at différent températures 108 Table 6.7  Evolution of the attrition particles from the composite Ti0 2 sphères over time 116  Table 7.1  Computed and expérimental fluence rates obtained for the reaction System shown in Figure 4.7 123  Table 7.2  FA photodegradation rate constants (kr) of composite TiÛ2 sphères at différent fluence rates using 100 mg L"1 as initial concentration of FA and 0.025 kg of composite Ti0 2 sphères 126  Table 7.3  Initial reaction rate constant for the FA dégradation in the présence of 131 composite Ti0 2 sphères and Degussa P-25  Table 7.4 2,4-D photodegradation rate constants (k2,4-D) of composite TiÛ2 sphères at différent fluence rates using 10 mg L"1 of initial concentration and 25 g of composite TiÛ2 sphères 139 Table 7.5 Pseudo-first-order kinetic rate constants (/f2,4-oobs) in photocatalytic dégradation of 2,4-D with différent initial concentration (fluence rate=4.16 mW cm"2) using 25 g of composite TiC>2 sphères 143 ix  Table 7.6 Apparent reaction rate constant for the 2,4-D dégradation in the présence of composite Ti0 2 sphères and Degussa P-25 151  x  List of Figures Figure 2.1  Schematic photochemical process over photon activated semiconductor showing the photogeneration of electron/hole pair: (a) oxidation of donor on the surface of the semiconductor particle; (b) diffusion of acceptor and réduction on the surface of the semiconductor; (c) recombination in the bulk, and (d) surface recombination 15  Figure 2.2 Structures of Rutile and Anatase Ti0 2 (Adapted from Linsebigler et al., 1995) 18 Figure 2.3 Sol-Gel chemistry séquence stages (Adapted from Optoweb, 2009)....29 Figure 2.4 Schematic of routes for structural évolution of métal organic in solution (Adapted from Ring, 1996) 32 Figure 2.5 Overview of the sol-gel process and various sol-gel derived products (Adapted from Brinker and Scherer, 1990) 33 Figure 2.6 Hydrolysis and condensation reactions of titanium isopropoxidé in the présence of water (Adapted from Laine, 1990) 37 Figure 2.7 Proceeding routes for sol-gel-derived composites (Adapted from Nazeri étal., 1993) 42 Figure 4.1 Préparation of composite "IÏO2 sphères  52  Figure 4.2 Multipoint BET equipment - Micrometrics ASAP 2020  54  Figure 4.3 X-ray diffractometer - Siemens D5000  57  Figure 4.4 Scanning Electron Microscope (SEM) - Phillips XL-30  59  Figure 4.5 Thermo gravimétrie analyzer (TGA) - TA Instruments  60  Figure 4.6 Schematic diagram of the quartz tube photocatalytic reactor  61 xi  Figure 4.7 Expérimental setup: (1) wood box, (2) UV lamps, (3) fluidized bed photocatalytic reactor (FBPR), (4) différent lamp positions 63 Figure 4.8 Expérimental setup: (1) fluidized bed photocatalytic reactor (FBPR), (2) tank, (3) pump, (4) Flowmeter 64 Figure 5.1 Apparent reaction rate constant of Formic acid (FA) versus cataiyst loadings in the FBPR using sphères calcined at 600°C 77 Figure 5.2 Formic acid dégradation versus reaction time with sphères calcined at: (0) 900°C, (o) 800°C, (A) 700°C, (o) 600°C; error bars represent 95% Cl of the triplicate runs) 78 Figure 5.3 Attrition résistance for TÏO2 sphères calcined at différent températures (Error bars represent 95% Cl of triplicate runs) 82 Figure 5.4 SEM micrograph of a photocatalytic composite sphère, calcined at 600°C (a and b) and calcined at 900°C (c and d) 83 Figure 5.5 X-ray diffraction patterns of composite TÏO2 sphères at différent températures 85 Figure 5.6 Surface area for samples calcined at différent conditions, (0) 600°C, (o) 700°C, (a) 800°C 86 Figure 5.7 Pore size distribution for samples calcined at différent températures, (0) 600°C, (o) 700°C, (o) 800°C 86 Figure 6.1 Attrition résistance of the photocatalyst produced with a) original formulation (OF); b) using HNO3 as a cataiyst instead of HCI (CHNO3). Error bars represent 95% Cl of triplicate runs 92 Figure 6.2 Attrition résistance for three différent processes conditions; a) original formulation (OF); b) increased amount of water (CH2O) in the sol-gel matrix; c) reduced the amount of HCI (CHci) in the sol-gel matrix. Error bars represent 95% Cl of triplicate runs 94 Figure 6.3 Attrition résistance for three différent process conditions; a) original formulation (OF); b) using 0.30 g P-25 ml_"1 TTIP (LP-25); c) using xii  0.50 g P-25 mL"1 TTIP (MP-25). Error bars represent 95% Cl of triplicate runs 95 Figure 6.4 Attrition résistance for two différent process conditions; a) original formulation (OF); b) when the pH of the chitosan solution was 4.58 (MpH); c) when the pH of the chitosan solution was 4.01 (LpH). Error bars represent 95% Cl of triplicate runs 97 Figure 6.5 Attrition résistance for two différent process conditions; a) original formulation (OF); b) when the chitosan solution had a concentration of 15 g chitosan L"1 (ChCOnc)- Error bars represent 95% Cl of triplicate runs 98 Figure 6.6 Attrition résistance for three différent process conditions; a) original formulation (OF); b) catalyst dried for 20 h at 80°C, cooled down, and then placed in the furnance at 600°C for 1 h (DC00i); b) catalyst dried in the furnace at 80°C for 20 h, followed by immédiate température increase to 600°C for 1 h (DCOnt.). Error bars represent 95% Cl for triplicate runs 100  Figure 6.7  Attrition résistance of TÏO2 sphères for two différent process conditions; a) original formulation (OF); b) catalyst fired for 3 h at 600°C instead of 1 h at 600°C (Ctime). Error bars represent 95% Cl for triplicate runs 101  Figure 6.8 Attrition résistance comparisons between the original formulation of the catalyst (OF) and the modified formulation (MF) considering variables listed in Table 6.5. Error bars represent 95% Cl of triplicate runs 105  Figure 6.9 Attrition résistance for five différent process conditions; a) modified formulation (MF); b) reducing the ammonia concentration to 10% (CAmm); c) reducing the amount of water to 0.040 g H 2 0 mL"1 TTIP (CH2O); d) the catalyst was cooled down to RT after the drying process and before the heat treatment at 600°C for 3 h (CCooi)- Error bars represent 95% Cl for triplicate runs 106  xiii  Figure 6.10 X-ray diffraction patterns of the commercial Ti0 2 powder (Degussa P-25 as received) and the composite Ti0 2 sphères produced with the modified formulation (MF) calcined at 600°C 108  Figure 6.11 X-ray diffraction patterns of composite Ti0 2 sphères calcined at différent températures (600, 700, 800 and 900°C) 109  Figure 6.12 Surface area and pore size distribution for samples from différent conditions, (a) surface area of commercial Ti0 2 powder (Degussa P-25) and the composite Ti0 2 sphères calcined at 600°C, (b) pore size distribution of Degussa P-25 and composite Ti0 2 sphères calcined at 600°C 112  Figure 6.13 TGA curves of the composite Ti0 2 sphères dried at 80°C for 20 h and calcined at 600°C for 3 h, (a) first 25 min of the drying process, (b) isotherm drying process, (c) heat treatment for 3 h 113  Figure 6.14 SEM micrograph of a photocatalytic composite sphère calcined at 600°C for 3 h 115  Figure 6.15 Cumulative attrition of the composite Ti0 2 sphères versus time  117  Figure 6.16 SEM micrograph of a photocatalytic composite sphère calcined at 600°C after 25 h of continuous use 117  Figure 7.1 Fluence rate dependence on the distance of the UV Lamps from the centre of the quartz tube using two différent actinometers: (o) Potassium Ferrioxalate; (0) lodite-iodate actinometer. Error bars represent 95% Cl of the triplicate runs 121  Figure 7.2 Schematic représentation of plausible photon paths for the Monte Carlo (MC)Method 122  Figure 7.3  Expérimental fluence rate variations with respect to the position of the UV lamps using Potassium Ferrioxalate actinometer and the predicted xiv  behaviour using Monte Carlo (MC) model. Error bars represent 95% Cl for the triplicate runs 124  Figure 7.4 Effect of différent lamp positions on FA photocatalytic dégradation, (o) 5.32 mW cm"2, (0) 4.16 mW cm-2, (•) 3.04 mW cm' 2 , (A) 2.14 mW cm-2. Error bars represents 95% Cl for the triplicate runs 126  Figure 7.5 Photoactivity of the composite Tï0 2 sphères with respect to time using 100mgL' 1 ofFA 128  Figure 7.6 Effect of catalyst loading on the reaction rate constant using commercial Ti0 2 powder (Degussa P-25) as a photocatalyst: CFAO = 100 mg L"1, Vtot = 1L 129  Figure 7.7 Photocatalytic dégradation of Formic acid, (a) commercial Ti0 2 (Degussa P-25): (a) with UV lamps off (0) with UV irradiation; (b) composite Ti0 2 sphères 130  Figure 7.8 Mechanism of 2,4-D dégradation - major route (Adapted from Djebbar and Sehili, 1998) 135  Figure 7.9 Adsorption isotherm of 2,4-D on the composite Ti02 sphères  136  Figure 7.10 Adsorption isotherm of 2,4-D on the composite TÏO2 sphères, (o) isotherm study, (A) measurements at the FBPR with différent batches of sphères 137  Figure 7.11 2,4-D (0) dégradation and 2,4-DCP formation (A) using the composite Ti0 2 sphères 138  Figure 7.12 Dégradation of 2,4-D using the composite Ti0 2 sphères under différence fluence rates, (0) 4.16 mW cm"2, (D) 3.04 mW cm"2, (o) 2.14 mW cm"2. The initial concentration of 2,4-D was 10 mg L"1 139  xv  Figure 7.13 Comparison between the 2,4-D photodegradation pathways at différent irradiation levels with fluences of: (0) 4.16 mW cm"2, (o) 3.04 mW cm"2, (o) 2.14 mW cm"2. Error bars represent 95% Cl of the triplicate runs.. 140  Figure 7.14 Formation of 2,4-DCP as an intermediate of the 2,4-D dégradation using the composite TÏO2 sphères at différent fluence rates: (•) 4.16 mW cm"2, (0) 3.04 mW cm"2, (0) 2.14 mW cm"2. The initial concentration of 2,4-D was 10 mg L"1 141  Figure 7.15 2,4-D photocatalytic dégradation at différent initial concentrations, (0) 10 mg L"1, (•) 5 mg L"1, (A) 1 mg L"1. Fluence rate of 4.16 mW cm"2 (lamps at 5 cm from the centre of the reactor) 142 Figure 7.16 Effect of the différent initial 2,4-D concentration, (A) 1 mg L"1, (•) 5 mg L"1, (0) 10 mg L"1. Fluence rate equal to 4.16 mW cm' 2 (lamps at 5 cm from the centre of the reactor) 143  Figure 7.17 Formation of 2,4-DCP as an intermediate of the 2,4-D dégradation using the composite TIO2 sphères at différent initial concentration of 2,4-D: (0) 10 mg L"1, (o) 5 mg L'1, (A) 1 mg L'1. The lamp position in ail the cases was 5 cm from the centre of the reactor (fluence rate equal to 4.16 mW 146 cm"2) Figure 7.18 pH évolution in the reaction System with 5 mg L"1 of initial 2,4-D concentration 147 Figure 7.19 Dégradation of 2,4-D using 10 mg L"1 as initial concentration reusing the Ti0 2 sphères, (0) 3 h, (•) 8 h, (A) 15 h 148 Figure 7.20 Dégradation of 2,4-D using 10 mg L"1 as initial concentration, (0) 2,4-D for the first run; (•) 2,4-DCP génération during the first run; (•) 2,4-D for the second run; (•) 2,4-DCP génération during the second run; (0) 2,4-D for the third run; (•) 2,4-DCP génération during the third run 149  xvi  Figure 7.21 Photocatalytic dégradation of 10 mg L"1 of initial concentration of 2,4-D (fluence rate equal to 4.16 mW cm"2) 150  Figure 7.22 Photocatalytic dégradation of 2,4-D using Degussa (A) and the composite Ti0 2 sphères (a) with an initial concentration of 10 mg L"1. Lamps position at 5 cm from the centre of the reactor (fluence rate equal to4.16mWcm" 2 ) 151 Figure 7.23 Dégradation of "Peachiand water" (initial TOC - 5 mg L"1) using the composite Ti0 2 sphères, (o) using UV (fluence rate of 4.16 mW cm"2) without the Ti0 2 sphères, (•) with the Ti0 2 sphères without UV light, (A) with Ti0 2 sphères and UV (4.16 mW cm"2). Error bars represent 95% Cl of triplicate runs 153 Figure 7.24 Adsorption of "Peachiand water" (initial TOC ~ 5 mg L"1) on the composite Ti0 2 sphères. Error bars represent 95% Cl of duplicate runs 154  Figure 7.25 Effect of the "Peachiand water" on the 2,4-D dégradation using the composite Ti0 2 sphères, (o) 2,4-D mix with Peahland water, (•) 2,4-D mixed with Milli-Q water 155  Figure 7.26 2,4-D dégradation using Peachiand water as the matrix, (•) fresh batch of Ti0 2 sphères, (•) reusing the same batch of sphères the first time, (A) reusing the batch of sphères the second time 156  xvii  Nomenclature a - External surface to volume ratio of the catalyst particles acat - Activity of the catalyst d - Interplanar spacing between crystals dsph - diameter of the composite "TÏO2 sphères d* - Dimensionless particle diameter e - Expansion index e" - Electrons f - Total number of tested TiÛ2 sphères formulations g - Gravity force h+ - Holes f w - Energy i - Integer exponent in power law rate expression m - Individual observations nrisph - Mass of composite TiC-2 sphères n - Integer representing order of diffraction nPh, tôt - Total number of photons nPh,trans - Number of photons entering into the reactor volume nt - number of observations k" - Surface second order rate constant kc - Fractional site coverage by hydroxyl radicals kCT - Charge transfer rate kLH -Apparent Langmuir-Hinshelwood rate constant kMT - Mass transfer coefficient kobs - Observed reaction rate constant kr - Rate constant for the FA dégradation  kV - Rate constant for the FA dégradation in m s"1 kR - Electron/ hole recombination rate k2,4-D - 2,4-D photodegradation rate constant pKa - Acid dissociation constant qx - Spectral radiation flux r - Rate r« - Kelvin radius of the pore of the catalyst rp - Real pore radius of the catalyst rsph - Radius of the composite Ti0 2 sphères ST2 - Between-formulations mean square SR2 - Within-formulations mean square t - Thickness of the adsorbed layer in the catalyst t| - Time of irradiation tv - T distribution v-Volume ofgas adsorbed (m3) ve - Terminal velocity of isolated single particles vm - Molar volume of liquid N2 (34.7 cm3 mol ~1) vt - Terminal velocity of the composite TiÛ2 sphères y - Grand average of ail the data yt - Average of three expérimental values (reaction rate constants) ytm - mth observation in the tth formulation z - Oxidation state  A - Electron acceptor / Titania polymorphic phase (Anatase) AN - Nucleophilic addition Ar - Archimedes number AOP - Advance oxidation processes xix  BeXp - Expansion of the bed BET - Brunauer-Emmet-Teller method to measure surface area BDDT -  Brunauer, Demming, Deming and Teller classification for isotherms generated suring surface area détermination  BJH - Barrett, Joyce and Halenda method for pore size distribution détermination BSE - Backscattered électrons C - Molar concentration of the iodide solution Cads - Amount of 2,4-D adsorbed in the catalyst CAmm - Ammonia concentration in water (10% v/v) Ceiem - Covariance élément Ceq - Concentration of 2,4-D in equilibrium CFA - Concentration of FA in solution CHCI -  Concentration of HCI during the sol-gel préparation  CH2O -  Concentration of water during the sol-gel préparation  Ctime - Time that takes the heat treatment of the composite TÏO2 sphères C2,4-D C2,4-DO -  Concentration of 2,4-D Initial concentration of 2,4-D  CB - Conduction band Chconc - Concentration of Chitosan (15 g L"1) Cl - Confidence interval (95%) COD - Chemical oxygen demand CSG - Composite sol-gel D - Electron donor Dcooi - Drying process where the composite Ti0 2 sphères were dried at 80°C for 20 h, then cool down to room température and then the heat treatment was applied Dcont - Continuous drying and heat treatment process (drying process at 80°C for 20 h and immediately after that, the température was risen to 600°C for 1 h) DFA-W  - Diffusive coefficient for FA in water xx  Dg - Monosaccharide residue (D-glucosamine) DBP - Desinfection by-products DSC - Differential scanning calorimetry Eest - Standard error of the estimation F - F value for statistical analysis for vT/vR degrees of freedom Fexp - F value for statistical analysis obtained from expérimental results FA - Formic acid FBPR - Fluidized bed photoreactor Ga - Galileo number Hn - Fluence rate using lodide-lodate Actinometry HPF - Fluence rate using Potassium Ferrioxalate Actinometry H2 - Characteristic hystérésis loop H3 - Characteristic hystérésis loop HBed - Height of the bed HBed.o- Height of the static bed HPLC - High performance liquid chromatography I - Photon or irradiation flux li - Intermediate product IEP - Isoelectric point K - 2,4-D adsorption constant Kads - Apparent adsorption equilibrium constant Kappads - Apparent adsorption constant Kj - Equilibrium adsorption constant for the intermediate i KL - Langmuir adsorption constant K02 - Apparent rate constant for oxygen Ko2a - Equilibrium adsorption constant for oxygen K2,4-D -  Equilibrium adsorption constant for 2,4-D  L-H - Langmuir Hinshelwood adsortion/kinetic model LpH - Chitosan solution at pH = 4.01 LP-25 - Low amount of P25 in the sol-gel (0.30 g mL"1 TTIP) M - Network forming élément during the sol-gel process Mv - Density number Mx - Mean MA - Minerai acids MC - Monte Carlo method MF - Modified formulation to produce the composite TiÛ2 sphères MpH - Chitosan solution at pH = 4.58 MP-25 - High amount of P25 in the sol-gel (0.50 g mL"1 TTIP) N - Maximum coordination number NA - Monosaccharide residue (N-acetyl-D-glucosamine) Nt - Total number of observations NOM - Natural organic matter OF - Original formulation P - Pollutant or Pressure in the System Pads - Pollutant adsorbed on the catalyst Piamp,?i - Power of the lamp at 254nm Piiq - Pollutant dissolved in a liquid phase P02 - Oxygen pressure in the System PPCP - Pharmaceutical and personal care products Q - Volumetric flow rate R - Titania polymorphic phase (Rutile) Rreactor - Radius of the quartz tube ROH - Alcohol RT - Room température  Se - Schmidt number SE - Standard error Serrer - Standard error SN - Nucleophilic substitution mechanism Sparam - Standard error of the parameter SR - Within-formulation sum of squares ST - Sum of squares Sh - Sherwood number SD - Standard déviation SE - Secondary Electrons SEM - Scanning électron microscopy T - Température TEOS - Tetraethyl orthosilicate TGA - Thermo gravimétrie analysis THM - Trihalomethanes TOC - Total organic carbon TTIP - Titanium IV isopropoxide U - Fluidization velocity UV - Ultraviolet light UV/H 2 0 2 - Advance oxidation process utilizing ultraviolet and hydrogen peroxide UV/O3 - Advance oxidation process utilizing ultraviolet and ozone UV/O3/H2O2 - Advance oxidation process utilizing ultraviolet, ozone and hydrogen peroxide VB - Valence band V - Solution volume VdH - volume of dilution for the samples taken during the potassium ferrioxalate actinometry Vrrad - Irradiated volume during the potassium ferrioxalate actinometry xxiii  Vsamp - Volume of the samples taken during potassium ferrioxalate actinometry W-Watts WHO - World health organization X - Individual observation XRD - X-ray diffraction  a - Proportional constant to calculate the reaction rate constant based on the photon flux (3 - Exponential constant to calculate the reaction rate constant based on the photon flux s - Voidage of the bed or molar extinction coefficient (1.11 x 104 L gmol"1 cm"1) a - Variance se - Molar extinction coefficient used for the calculations of the irradiation flux using potassium ferrioxalate actinometry S450 - Molar absorption coefficient of triodide at 450 nm. y - Surface tension of N2 at its boiling point (8.85 ergs cm 2 at 77K) X - Wavelength Ç - Relative efficiency PH20 - Density of water Psph - Density of water I^H2o - Viscosity of the water § - Quantum yield <> |s - Sphericity 0 - Bragg angle AA450 - Change in the absorbance at 450 nm AG - Change of Gibbs free energy 0P - Equilibrium coverage of the pollutant over the surface of the catalyst QOH - Fractional site coverage by hydroxyl radicals xxiv  02,4-D -  Fractional site coverage by 2,3-D  VR - Degree of freedom within formulations VT - Degree of freedom between formulations  xxv  Acknowledgements My sincère gratitude is due toward my supervisor, Dr. Madjid Mohseni, who provided me with this incredible opportunity to corne to UBC and be part of his research group. His support, generosity and constant guidance gave me the chance to overcome many of the challenges that I faced during this process. I would like also to acknowledge Dr. Mehrdad Keshmiri for his help, trust and support during this process. I am also gratefui to Dr. Naoko Ellis and Dr. Dusko Posarac for reviewing my thesis and being part of my research committee. I am very thankful for their valuable advice and comments. During the last five months of my masters, I received an incredible help from Maxime Thevenin during his exchange program at UBC. He gave an enormous contribution helping me in the conduction and data collection of many experiments. Thanks a lot for your work, your désire to learn, for your diligent work and friendship. I am especially gratefui to Gustavo Imoberdorf, Siva Sarathy, Esteban Durân, Sandra Robaire for their help, advise, constant support and friendship during this process. Certainly, every discussion and brainstorming generated valuable ideas that helped me to achieve my goals during my masters. I want also to thanks to Dr. Kevin Smith and Dr. Fariborz Taguipour from Chemical Engineering Department for letting me used their laboratories and equipments. I would like also to thank Dr. Tom Troczynski from Materials Department for providing access to the Ceramic Laboratories and Dr. Matti Raudsepp, from the Department of Earth and Océan Sciences, for teaching and helping me with many important tools and techniques use in materials characterization. My thanks are also extended to my friends Jana Schmidtova, Jidon Janau, Masakazu Sakaguchi, Anne-Marie Kietzig, Bojan Petkovic, Ryan Anderson, Mohammad Alquaad and so many other graduate students for their friendship and support. I want to thank them ail for sharing their expertise on spécifie topics and for providing their time and minds to solve issues. xxvi  Spécial thanks also to the faculty and staff of the Department of Chemical and Biological Engineering for their support and for the friendly environment that I found hère. I would like to acknowledge also the generous funding contributed by Bl Pure Water (Canada) Inc. and NSERC. Lastly, I would like to express my deepest appréciation and respect to my family, especially my parents and my wife for giving me hope, courage and an amazing support throughout this process.  xxvii  Dedication First of ail, I would like to dedicate this work to God, for His amazing grâce, everlasting love and guidance through my académie and personal life. Secondly, I would like to dévote ail this work to my family, especially to my wife and parents, for their constant support and advice. They hâve been always there in my most difficult times of struggles and sorrows. Additionally, I would like to mention my godfather Luis Emilio (R.I.P) who always supported me during my engineering career and encouraged me to continue with my graduate studies abroad. Unfortunately, he is not longer with us to enjoy this moment, but I am sure he would be proud of this new accomplishment. I also want to dedicate this work to my godson David. He was born almost at the same time I was finishing my work at UBC and, as every new born, he has become in a source of inspiration for everyone.  xxviii  ffie victory qfsuccess is haCfwon when one gains the habit qf setting goaCs and achieving them. %ven the most tedious chore wiCC ôecome enduraôfe as you parade through each day convinced that every tasf^ no matter how tneniafor ôoring, ôrings you doser to fuCfitting your dreams. Og !Mandino  xxix  Chapter One Introduction 1.1 Background Contaminated water, along with its conséquent effects on the safety of drinking water resources, has been a major concern in récent décades. In particular, there hâve been increasing concerns over the contamination of drinking water supplies with trace levels of organic (solvents) and inorganic (metallic compounds) contaminants, as well as bioactive materials like pharmaceuticals and agricultural products. Beyond the impact of population growth, the contamination of water has been rising as a resuit of industrial development and agricultural activities. For example, pesticides from agricultural field run-offs and other groundwater contaminants are now frequently found in water resources. Also, pharmaceuticals and personal care products (PPCPs) hâve been identified in outflows from sewage treatment plants and surface waters Worldwide (Boyd et al., 2003). Conventional treatment technologies, such as filtration, chlorination, ozonation and boiling hâve been applied for many years to remove or inactive pathogens and reduce the levels of contaminants in drinking water. However, it is known that the advantages of thèse processes are limited to certain levels and forms of microbial/chemical contaminants. In addition, thèse methods require a lot of chemicals, fuels and electricity (Belapurkar et al., 2006). Further, some of thèse traditional methods (e.g., chlorination) may be responsible for more drinking water issues, such as génération of disinfection by-products (DBP's) like trihalomethanes (THMs). Additionally, some of the pathogenic viruses and bacteria {Campylobacter, Yersina, Mycobacteria or Legionella) and protozoa (Cryptosporidium or Giardia lamblia cysts) are known to be résistant to chlorine disinfection (Kabir, 2006). Therefore, alternative treatment methods to overcome thèse and other limitations hâve been increasingly studied in récent years.  1  1. Introduction Besides the abovementioned issues, récent environmental régulations hâve led to the development of novel and more effective treatment technologies that are capable of removing the pollutants rather than transferring them from one phase to another (such as the case for activated carbon adsorption). One such technology that has received significant attention over the past two décades is photochemical oxidation, which is also related to advance oxidation processes (AOPs). Thèse methods are based on the irradiation of the contaminated water with ultraviolet light (UV) under différent conditions, such as H2O2/UV, Fenton's reagent, O3/UV, O3/H2O2/UV and photocatalysis (Parson, 2004; Kabir, 2006; Catalkaya and Kargi, 2008). Photocatalysis is a process where a semiconductor catalyst is activated by UV irradiation promoting électrons from the valence band (VB) to the conduction band (CB) and leaving holes in the VB. Thèse electron-hole pairs migrate to the surface of the photocatalyst where they initiate redox reactions with absorbed molécules. Thus, organic pollutants are degraded to smaller non-toxic species, such as carbon dioxide (CO2) and water (H2O) (Turchi and Ollis, 1990; Herrmann, 1999). Photocatalytic oxidation has gained high levels of attention Worldwide, not only because of the attractive multi-science aspects of the field, but also the associated challenging parameters when it cornes to the practical large scale applications such as the improvement of the photon utilization, réduction of the overall cost of the process, among others. Nonetheless, there are many advantages associated with thèse processes (Mills et al., 1993; Kabir, 2006): 1. almost ail organic pollutants can be mineralized. 2. the process is generally considered as a green technology because many of the dégradation products are environmentally harmless. 3. the most commonly used photocatalysts are cheap, non-toxic, stable, biologically and chemically inert, and insoluble under most conditions and reusable. 4. in most cases a low energy UV source is needed for catalyst activation.  2  1. Introduction The sélection and synthesis of proper semiconductor material are key to having an efficient photocatalytic process. Among many différent semiconductor photocatalysts, there is a gênerai consensus that Titanium Dioxide (TÏO2) is more superior because it is cheap, easy to produce, has high chemical stability, and its photogenerated holes are highly oxidizing. The redox potential for titania's photogenerated holes is +2.53V versus the standard hydrogen électrode in pH 7 solution and the redox potential for its conduction band électrons is -0.52V, which is négative enough to reduce dyoxygen to superoxygen or to hydrogen peroxide (Fujishima et al., 2000; Hashimoto et al., 2005; Fujishima and Zhang, 2006; Thiruvenkatachari et al., 2008). TiÛ2 exists in three main crystallographic forms of anatase, rutile and brookite. The anatase type has been selectively used for photocatalytic applications because it offers many advantages over the other two forms, e.g., higher surface area and surface density of active sites (Herrmann, 1999; Ambrus et al., 2008; Wetchakun and Phanichphant, 2008). Sol-gel (solution - gelation) process is one of the methods that has been increasingly used to produce many heterogeneous photocatalysts, including nanosize Ti0 2 particles with high photocatalytic activities (Campanati et al., 2003; Su et al., 2004). This technique involves the formation of homogeneous solution of raw materials and the subséquent gelation of the solution to form a porous oxide. Therefore, it can be defined as a séquence of chemical synthesis methods, mainly inorganic polymeralization reactions that need to be carefully controlled (Wen and Wilkes, 1996). In gênerai, the process starts when a métal or semimetal alkoxide precursor, M(OR)n, where M represents a network-forming élément such as Si, Al, Ti, etc. and R is typically an alkyl group (CxH2x+i), is dissolved in alcohol to give a homogenous solution. Then, afterthe addition of water, hydrolysis and condensation occur simultaneously producing a gel (Hench and West, 1990; Wen and Wilkes, 1996). The process is followed by aging and then drying which are necessary to extract the solvents and reaction liquids trapped inside the gel. Finally, heat treatment and sintering are required to densify the gel. Among the many advantages of the sol-gel process, the low température requirement is one of the most important  3  1. Introduction ones. This allows solids with large spécifie surface areas and high porosity in the meso and macropore ranges (Livage, 1998). The produced photocatalyst can be applied for the treatment of contaminated water in différent photocatalytic reactors, e.g., slurry reactors, supported catalytic reactors, among others. In slurry reactors, the catalyst is dispersed in an aqueous médium as suspension, with a problem associated with the pénétration of the UV irradiation due to the strong absorption by photocatalyst and organic pollutants. In addition, the séparation and recycling of the photocatalyst from treated liquid are required. Supported photocatalytic reactors are the ones that hâve the catalysts immobilized on an inert support. In thèse reactors, the overall rate of dégradation could be affected due to mass transfer limitation and low surface area as well as catalyst attrition (Werther and Reppenhagen, 1999; Kabir, 2006). Fluidized bed photocatalytic reactors (FBPR) are among the promising alternatives mainly because they présent many advantages such as efficient contact between the catalyst and the pollutants, low mass transfer résistance, and high TÏO2 surface exposure to UV radiation (Pozzo et al., 1999, 2000; Bouchy and Zahraa, 2003). However, attrition of particles and elutriation in FBPR are a problem and could induce a decay in the catalyst activity and an increase in the cost of the overall process due to downstream filtration (Werther and Reppenhagen, 1999; Nelson et al., 2007).  1.2 General objective and scope of the présent work Commercial Tï0 2 powder (Degussa P-25) is often selected as a référence photocatalyst. It shows high photocatalytic activity mainly because it has a comparatively high surface area and substantial amount of anatase. Nonetheless, complications such as the need for filtration and further séparation of the suspended particles after the treatment make the large scale application of P-25 TIO2 photocatalyst difficult and economically unfeasible. The main objective of the présent study was to develop and standardize a novel composite photocatalyst to be used in a FBPR. The approach focused on a  4  1. Introduction novel processing method to produce Ti0 2 nanoparticles, using sol-gel process, and mixing them with Ti0 2 pre-calcined powder (Degussa P-25). TÏO2 powder was used as filler bonded together with the gel-derived Ti0 2 . The resulting Composite Sol-Gel (CSG) titania was then mixed with a polymeric structure to produce the sphères. The evaluation-standardization aspects of the work included better production process conditions, such as reactant quantities, température conditions, among others. The photo-efficiency of the composite photocatalyst was evaluated in terms of its ability to décompose formic acid (FA) from a water solution. FA was selected as a model pollutant because it is oxidised directly to C0 2 without the formation of any stable intermediate products, and has been used extensively in other photocatalytic studies (McMurray et al., 2004). In addition, 2,4-D (2,4 Dichlorophenoxyacetic acid), a well known micro-pollutant, was selected to test the catalyst with a more complex molécule. Finally, preliminary tests with natural organic matter (NOM) were also conducted.  1.3 Structure of this thesis The présent thesis is organized into seven chapters. In Chapter Two, an extensive review of literature on relevant issues is presented. Chapter Three describes the scope and the objectives of this investigation. Chapter Four présents in détail the expérimental methodology that has been followed to reach the goals of this project. Chapter Five explains the preliminary expérimental results obtained with the CSG Ti0 2 sphères before any modification was made. Photocatalyst activity, attrition résistance détermination, and photocatalyst characterization are some of the aspects included in this chapter. Chapter Six discusses in détail the catalyst préparation and modification and how those changes in the catalyst synthesis affected the attrition résistance and the photoactivity of the catalyst. Chapter Seven, présents data on Formic Acid (FA) dégradation (under différent conditions) and a study of the 2,4-D dégradation using this CSG catalyst in a fluidized bed photoreactor. In addition, preliminary results of the NOM dégradation are also included along with an analysis of the fluence rate within the photoreactor. 5  1. Introduction Finally, Chapter Eight highiights the main conclusions obtained after this investigation and présent some recommendations for future work.  6  Chapter Two Literature Review 2.1 Background Water is a natural resource that has direct impact on human health. It is considered vital to humanity regardless of time (Kamble et al., 2006). According to the World Health Organization (WHO), 70% of the Earth's surface is water, but only 2.5% to 3% is considered fresh water. Also, less than 1 % is accessible as surface water located in biomass, rivers, lakes, soi! moisture and as water vapour distributed in the atmosphère (Kabir, 2006; Kamble et al., 2006; WHO, 2006). One of the major environmental problems nowadays is water pollution. Many locations with high quality water supplies hâve been developed intensively, producing a significant industrialization and extensive urbanization resulting in a progressive détérioration of water quality (Matthews, 1992). Particularly, the industrialized world has many industries that require large amounts of water for processing. In fact, it is predicted that water demand by industry, including energy and agriculture sectors, will grow rapidly to keep up with growing population. Nonetheless, the real problem is that the water discharged from thèse industries is contaminated with toxic organic compounds (Balapurkar et al., 2006; Kamble et al., 2006). In addition, the use of pesticides and herbicides in the agriculture sector causes scarcity of clean drinking water due to the toxicity of thèse compounds (Krysa et al., 1999; Kabir, 2006). A balance between continuous improvement and sustainability should be reached in order to minimize water pollution and préserve this vital resource. In addition, developments and improvements of wastewater treatment technologies should be achieved (Kabir, 2006). Conventional treatment technologies, such as filtration and chlorination, hâve been applied for many years to remove or inactive pathogens and reduce the levels of contaminants in drinking water. However, alternative treatment methods hâve been increasingly studied in récent years to  7  2. Literature Review overcome some of the drawbacks of thèse technologies (e.g., génération of disinfection by-products). One group of such technologies are referred to as advance oxidation processes (AOPs), which are based on the irradiation of the contaminated water with ultraviolet light (UV). AOPs hâve gained high level of attention in récent years because they offer an interesting approach to remove contaminants from water. In particular, heterogeneous photocatalysis has been the subject of intensive studies by researchers around the world, because of the appeal of the technology in which low energy UV is coupled with a semiconductor inducing mineralization of pollutants into harmless compounds without producing any other waste streams (Ray, 1998). In addition, photocatalysis does not require the use of strong oxidizing chemicals of potentially hazardous nature, e.g., H2O2, O3 (Mills et al., 1993)  2.2 Photocatalysis and environmental remediation The first mention of photocatalysis was by Plotnikov in the 1930's in his book entitled Allgemeine Photochemie (Tremblay, 2001). The word "photocatalysis" is composed of two parts; the prefix "photo" meaning light and the suffix "catalysis" referring to the process in which a substance participâtes in modifying the rate of transformation of reactants without being altered in the end (Kabir, 2006). From a thermodynamic point of view, the term "photocatalysis" can be fully applied only to reactions occurring with a réduction in the free energy (AG < 0) where the rate of thèse reactions is increased thanks to a particular reaction pathway involving photocreated species. This pathway then leads to différent reaction product selectivity than those for the thermal reactions. In the case of thermodynamic unfavourable reactions (AG > 0), the energy of UV irradiation is converted into chemical energy and thus, the term "photosynthesis" is applied (Teichner, 2008). Thèse kinds of reactions can be qualified as photocatalytic only when one catalyst (homogenous or heterogeneous) is clearly highlighted and the energy of irradiation compensâtes for the potential barrier, which is usually called the activation energy (Ollis et al., 1991).  8  2. Literature Review There has been an increased interest among the scientific community to the development and improvement in photocatalytic processes, especially heterogenous photocatalysis. Much of the interest in photocatalysis is due to the many advantages it has over traditional water and air treatment technologies, particularly the fact that organic pollutants can be degraded to carbon dioxide, water and minerai acids (Ollis et al., 1991). Additionally, other advantages of this technology are: mild reaction conditions and modest reaction times, less Chemical input requirement, minimal secondary waste génération, and the capacity for using renewable solar energy (Kabraefal., 2004) In the last three décades, photocatalysis has been studied more and more for environmental remediation. For example, some of the major applications of photocatalysis recently reported include the removal of colour and the destruction of dye; réduction of COD (chemical oxygen demand); mineralization of hazardous organics; destruction of hazardous inorganic compôunds; treatment of heavy metals; dégradation of harmfui fungicides, herbicides and pesticides; decontamination of soil; destruction of cancer cells and viruses; purification and disinfection of water; among other (Bhatkhande et al., 2001; Kabra et al., 2004; Syoufian et al., 2007; Gaya and Abdullah, 2008; Higgins étal., 2009).  2.3 Heterogeneous photocatalysis Heterogeneous photocatalysis was first introduced and developed in 1970 to describe the partial oxidation of alkanes and olefinic hydrocarbons at ambient température and under UV irradiation (Teichner, 2008). The "heterogeneous" adjective spécifies the nature of the reaction médium which is not homogeneous but comprises at least two phases: the solid (catalyst) and the fluid reagent that could be in gas phase, pure organic liquid phases or aqueous solutions (Herrmann, 1999; Teichner, 2008). This is a process in which illumination of an oxide semiconductor (heterogeneous catalyst), usually Titanium Dioxide (Ti0 2 ), produces photoexcited électrons (e") and holes (h+). Then, thèse électrons and holes can migrate to the oxide surface and participate in half-cell reactions that are part of a closed catalytic cycle (Ollis et al., 1991). 9  2. Literature Review A heterogeneous catalyst is frequently defined as a solid or mixture of solids which accelerate chemical reactions without themselves undergoing changes. However, this définition is limited in scope considering that the properties of catalysts can change significantly with service lives and with certain applications (Campanati et al., 2003). As a composite material, a heterogeneous photocatalyst can be characterized by: i) the relative amount of différent components (active species, physical and/or chemical promoters and supports); ii) shape; iii) size; iv) pore volume and distribution; and v) surface area (Campanati et al., 2003). Ail those characteristics would define operational characteristics of the catalyst, such as photocatalytic activity and mechanical strength of the catalyst.  2.3.1 Electronic excitation process in semiconductor materials The initiation of the electronic excitation process is defined upon irradiation of a semiconductor. By définition, a semiconductor has a band structure which is usually characterized by a séries of energetically closed spaced energy levels (valence band) and spatially diffused levels at higher energy (conduction band). The magnitude of the energy gap between the electronically populated valence band and the vacant conduction band, which is called the Band Gap, governs the extent of thermal population of the conduction band and its intrinsic state (Fox and Dulay, 1993; Linsebigler et al., 1995). This band gap also defines the wavelength sensitivity of the semiconductor to irradiation. In gênerai, photocatalytic process is described by Equation 2.1, when the semiconductor is irradiated with ultraviolet (UV): hv + semiconductor —» h+ + e~  (2.1 )  When a semiconductor catalyst of the chalcogenide type (oxides -TiC>2, ZnO, Zr0 2 , Ce0 2 , etc. - or sulfides - CdS, ZnS, etc.) is illuminated by light with a photon energy (hv) equal to or larger than the band gap energy, it excites the électrons in the valence band to the conduction band, resulting in the formation of a positive hole (h+) in the valence band and an électron (e") in the conduction band. Thèse will hâve 10  2. Literature Review sufficient life-time, in the nanosecond regimen, to undergo charge transfer to adsorbed species on the semiconductor surface (Linsebigler et al., 1995; Herrmann, 1999; Bhatkhande et al., 2001; Kabra et al., 2004; Thiruvenkatachari et al., 2008). Alternatively, thèse photogenerated électrons and holes can recombine with the subséquent release of phonons (a phonon is a quantum mechanical version of a spécial type of vibrational motion), inducing one of the major drawbacks of this process because it reduces the efficiency of the photocatalytic process (Krishna et al., 2006). The band gap énergies of some semiconductors are given in Table 2.1.  Table 2.1: Band gap énergies for some semiconductor materials at 0K (Bhatkhande et al., 2001 ; Thiruvenkatachari et al., 2008). Band gap energy (eV) Semiconductor Band gap energy (eV) Semiconductor 5.4 Fe 2 0 3 Diamond 2.3 2.42 PbS CdS 0.286 3.87 ZnS 3.6 Zr0 2 3.436 Cu 2 0 2.172 ZnO CdSe 1.7 3.03 Ti0 2 2.582 PbSe 0.165 CdS 3.54 Si 1.17 Sn0 2 0.744 Ge 2.76 W03  During heterogeneous photocatalysis in water treatment, the electronic excitation of a polycrystalline semiconductor, caused by light absorption, drastically alters its ability to lose or gain électrons, promoting décomposition of pollutants to harmless products (Ray, 1998; Lehr et al., 2005).  Therefore, the photoreaction  accélération, by the action of the solid catalyst, will dépend on the reaction mechanism, that itself dépends on the interaction between the species to be degraded and/or the intermediates (Lehr et al., 2005). The classical heterogeneous photocatalytic process can be divided to seven independent steps, as is detailed below (Herrmann, 1999): 1. Transfer of the reactants from the fluid phase to the catalyst surface (external diffusion)  11  2. Literature Review 2. Mass transfer of reactants from the catalyst surface into its pore structure (internai diffusion) 3. Adsorption of at least one of the reactants 4. Reaction on the surface 5. Desorption of the product(s) 6. Mass transfer of the products out of the pore structures of the catalyst 7. Removal of the products from the interface région  The photocatalytic reaction occurs mainly on the adsorbed surface (step 4) and the only différence with conventional catalysis is the mode of activation of the catalyst. That is, the thermal activation is replaced by photonic activation. This activation is not concerned with steps 1, 2, 3, 5, 6 and 7, although photoadsorption and photodesorption of reactants (mainly oxygen) do exist (Herrmann, 1999). The séquence of chain reactions that can take place during the adsorption step hâve been described (Gaya and Abdullah, 2008):  •  Photoexitation: hv + semiconductor -» h+ + e  •  Oxygen ionosorption: (02)ads +e~ ->02"  (2.2)  •  lonizationofwater:  H20-*OH~ + H+  (2.3)  •  Protonation of superoxides: 02' +H+-+HOO'  (2.1)  (2.4)  The hydroperoxyl radical formed in Equation 2.4 also has scavenging properties that help prolong the lifetime of photon hole (preventing the recombination between h+ and e) because it can react with électrons to produce HOi and subsequently H202:  HOO'+e^>H02~ HOO~ + H+ -» H202  (2.5)  (2.6) 12  2. Literature Review Considering that hydroxyl radical (•OH) is the primary oxidant in a photocatalytic System, Turchi and Ollis (1989) proposed a route where four possible mechanisms take place. This séquence is listed in Table 2.2, where P? represents an organic molécule and P1tads represents the absorbed organic molécule. As was already pointed out, the initiating step in photocatalysis is the excitation of the semiconductor by irradiation to produce electron-hole pairs (Equations 2.1 and 2.7). The typically low quantum yield of photocatalytic reactions is due to high rate of recombination between the electron-hole pairs (Equation 2.11). Recombination can be avoided if thèse species are separated and "trapped" by surface absorbents. For instance, the photogenerated électrons can be trapped by oxygen to form superoxides (02~) and hydroperoxyl radical (H02-) and subsequently H 2 0 2 , as shown in Equations 2.2 and 2.5, respectively (Okamoto et al., 1985; Gaya and Abdullah, 2008). Besides, one of the principal hole traps is the adsorbed hydroxide ions or water molécules (Equations 2.8) that proceed to form hydroxyl radicals as shown in Equation 2.12 (Turchi and Ollis, 1989). Another possibility to avoid the recombination between e and h+ is to mix noble metals such as silver, gold, platinum, with the semiconductor photocatalyst (Chao et al., 2003; Falaras, et al., 2003; AprWe et al., 2008). The direct hole-organic reaction (Equation 2.14), while thermodynamically possible, is not believed to be significant because of the lack of reactivity shown in water-free organic solutions. For example, it has been demonstrated that aromatic molécules preferably adsorb to surface hydroxyls rather than directly to the Ti0 2 lattice. For this case, the oxidative attack of thèse molécules would be analogous to Equation 2.19 (Turchi and Ollis, 1989). Photo-excited électrons are trapped by Ti l v centres to form Ti'" as is shown in Equation 2.15. In Systems without a reducible adsorbate, conduction band électrons remain on the semiconductor resulting in the formation of a blue colour, which is characteristic of Ti'". When oxygen is in the System, the surface Ti'" is readily oxidized by molecular oxygen to form the superoxide ion radical (Equation 2.16). Thèse ions may be further reduced to H 2 0 2 (Equation 2.21). Under acidic conditions, H+ may protonate the superoxide to form perhydroxyl radical (Equation 2.22). 13  2. Literature Review Equations 2.17 to 2.20 represent four différent possible pathways for »OH attack on organics, depending on the photocatalyst surface or the fluid phase (Turchi and Ollis, 1990; Mattews, 1992; Hoffman, et al., 1995).  Table 2.2: Photocatalytic reaction scheme for the TiC>2 (Turchi and Ollis, 1989; Turchi and Ollis, 1990; Kabir, 2006). Séquence Excitation of the catalyst by photon energy greater than the band gap Adsorption on the catalyst surface and lattice oxygen (OC2)  Reaction + h+  Ti02—^e-  (2.7)  OC1 + Tiw + H20 *> 0LH~ + Tiw • • • OH~ (2.8) Tiw +H20<r^TiIV  •••H20  (2.9)  «te + J J o P ^  (2.10)  e~ + h+ —> heat  (2.11)  +  Recombination of the e" - h pair, releasing heat  Tiw •••OH' +h+ <r>Tiw Tiw •••H20 + h+ <^Tiw  -OH' •••OH'+H+  Trapping of the hole and the électron  (2.12) (2.13) (2.14)  Ti^+e-^Ti1" TÏ" + 02±>Ti  lv  (2.15) •02'-  (2.16)  Case 1:  Ti^-OH'+P^^Ti^+P^  (2.17)  Case II: Attack by hydroxyl radical (adsorbed or free) under  (2.18)  OH'+Puds^P2tads Case III: Tiw •••OH'+Pl<^TiIV  +P2  (2.19)  Case IV: OH'+P1<r>P2  e + Tiw • • • 02'~ + 2H+ <-> Tiw{H202) Reaction of other radicals  Tiw • • • 02" +H+*> Tiw [HO;) H202 + OH' <-> H02" + H20  (2.20)  (2.21 ) (2.22) (2.23)  Both oxidation and réduction can take place at the surface of the photoexcited semiconductor photocatalyst, as is shown in Figure 2.1. The génération 14  2. Literature Review of electron-hole pairs takes place in the semiconductor particle due to the excitation of an électron from the valence band to the conduction band initiated by light absorption. Upon excitation, the fate of separated électron and hole can follow several pathways; while at the surface, the semiconductor can donate électrons to reduce an électron acceptor (usually oxygen) (pathway b). Also, a hole can migrate to the surface where an électron from a donor species can combine with the surface hole oxidizing the donor species (pathway a). In compétition with the charge transfer to adsorbed species is the électron and hole recombination; a separate electron-hole recombination can occur in the volume of the semiconductor particle (pathway c) or at the surface (pathway d) with the release of heat (Linsebigler et al., 1995).  Figure 2.1: Schematic photochemical process over photon activated semiconductor showing the photogeneration of electron/hole pair: (a) oxidation of donor on the surface of the semiconductor particle; (b) diffusion of acceptor and réduction on the surface of the semiconductor; (c) recombination in the bulk, and (d) surface recombination.  15  2. Literature Review As depicted in Figure 2.1, if an électron donor, D, adsorbs on the surface of the semiconductor particles, it can react with the photogenerated holes to generate an oxidized product, D+. Similarly, an électron acceptor présent at the surface, A, can react with the photogenerated conduction band électrons to generate product A". In the application of semiconductor photocatalysis to the purification of water, the électron acceptor, A, is invariably dissolved oxygen, and the électron donor, D, is the pollutant (Parsons, 2004). The activity of the photocatalyst to dégrade pollutants can be defined as the relative or absolute rate of photocatalytic reaction (Kaneko and Okura, 2002). Based on the mechanism involved in a photocatalytic process, the photocatalytic activity of semiconductor materials must be controlled by the following parameters: i) light absorption properties, which are mostly governed by the bulk structure of the semiconductor and difficult to modify; ii) rate of réduction and oxidation of the substrate by e" and h+, respectively, which is dépendent on the catalyst surface; and iii) rate of e" and h+ recombination (Kabir, 2006). During the illumination of the photocatalyst, only a portion of the photocatalyst particles can absorb incident photons and the remainder of the photocatalyst does not take part in the reaction. The total number of the absorbed photons should be constant if a sufficient amount of catalyst absorbs ail the incident photons. Large catalyst surface area leads to faster reaction of e" - h+ with substrate because more substrate molécules are available to participate in the reaction (Ohtani et al., 1987; Su et al., 2004; Kabir, 2006). However, literature reports suggested that e" and h+ recombination is dépendent on the crystallinity of the catalyst, which increases with the crystal defects in the semiconductor surface. The crystal defects are higher with larger surface area; hence, the appropriate surface area should be determined based on its influence on the catalyst activity (Kabir, 2006).  2.3.2 Titanium Dioxide (Ti0 2 ) as a photocatalyst Photocatalytic activity of titanium dioxide CTÏO2) was first reported by Fujishima and Honda (1972) for light induced water splitting on Ti0 2 surface. A  16  2. Literature Review complète electrochemical circuit with a rutile form of Ti0 2 and platinum black, as the two électrodes, allowed water to be decomposed using UV-visible light, into oxygen and hydrogen, without the application of an external voltage (Macak et al., 2007; Fujishima and Zhang, 2006). This resuit, which is referred to as the "HondaFujishima effect", has opened novel approaches in heterogeneous catalysis and Ti0 2  applications  (Gaya  and Abdullah, 2008). The Worldwide  interest  in  photocatalysis has increased significantly ever since and much research has been done in this area. Titanium dioxide exists in three main crystallographic forms: rutile, anatase and brookite. The anatase type has been selectively used for photocatalytic applications because it is the most active form of TiÛ2 (Wetchakun and Phanichphant, 2008). The octahedron (Ti06) is the fundamental structure units in the TiÛ2 crystals. In the rutile structure, each octahedron is in contact with ten neighbour octahedrons (two sharing edge oxygen pairs and eight sharing corner oxygen atoms), while in the anatase structure each octahedron is in contact with eight neighbours (four sharing an edge and four sharing a corner). As can be seen in Figure 2.2, each Ti 4+ ion is surrounded by an octahedron of six O2" ions. This octahedron in rutile is not regular, showing a slight orthorhombic distortion. On the other hand, the octahedron in anatase is significantly distorted so that its symmetry is lower than orthorhombic. The 77 - 77 distances in anatase are greater than those in rutile (3.79 and 3.04 versus 3.57 and 2.96 Â) whereas the 77 - O distances are shorter than those in rutile (1.934 and 1.980  in anatase versus 1.949 and 1.980  in rutile). Thèse différences in lattice structures cause différent mass densities and electronic band structures between anatase and rutile, possible reason why the former is more active than the latter (Linsebigler et al., 1995; Gao et al., 2006; Bojinovaefa/.,2007). In gênerai, TiÛ2 is a material of great importance due to its useful electrochemical, dielectric, electroconductive and optical properties. It is widely used in cosmetics, pigments, photocatalysis, adsorbents and catalytic supports (Gao et al., 2006). As a photocataiyst, Ti0 2 is well known because of several reasons: low cost, easy to produce, high chemical stability, and highly oxidizing photogenerated  17  2. Literature Review holes. The redox potential for photogenerated holes is +2.53V versus the standard hydrogen électrode in pH 7 solution and the redox potential for conduction band électrons is -0.52V, which is négative enough to reduce dyoxygen to superoxygen or to hydrogen peroxide (Fujishima and Zhang, 2006). Besides, Ti0 2 has biological and environmental safety advantages that give it particular characteristics (Fujishima étal., 2000; Hashimoto et al., 2005; Fujishima and Zhang, 2006).  Figure 2.2: Structures of Rutile and Anatase Ti0 2 (Adapted from Linsebigier et al., 1995).  The mineralization process by the Ti0 2 photocatalytic oxidation of organic compounds is shown in Equation 2.24 where P represents the organic compounds (pollutant) and MA the minerai acids produced in the process:  P + 02+hv  Ti 2  ° >C02 + H20 + MA  (2.24)  18  2. Uterature Review Since 1977, when Frank and Bard examined the possibilities of using Ti0 2 to décompose cyanide in water, there has been increasing interest in environmental applications of photocatalysis (Frank and Bard, 1977; Fujishima et al., 2000). In the last few décades, suspensions of Ti0 2 powder hâve been used for photocatalytic oxidation of organic and inorganic compounds in aqueous solutions to remove the impurities from water. Many organic compounds, such as aromatics, haloaromatics, aliphatics, dyes, dioxins, etc., hâve been demonstrated to be oxidized through this process to carbon dioxide (C0 2 ) and water, which are categorized as harmless compounds (Matthews, 1987; Kosanic, 1998; Kabra et al., 2004; Mrowetz and Selli, 2006; Belapounkar et al., 2006; Bojinova et al., 2007; Osajima étal., 2008).  2.4. Photocatalytic efficiency and kinetic overview 2.4.1 Photocatalytic efficiency The efficiency of the photocatalytic process is measured by quantum yield, (/>, which is defined as the number of events occurring in the System per photon absorbed (Herrmann, 1999; Rahn et al., 2003). In heterogeneous catalysis, the quantum yield is difficult to measure since the semiconductor particles absorb, scatter and transmit radiation. It may vary on a wide range of conditions including: a) the nature of the catalyst; b) the expérimental conditions such as température, concentrations, etc.; and c) the nature of the reaction itself. Nonetheless, the knowledge of this parameter is essential because it allows one to compare the activity of différent catalysts for the same reaction, to estimate the relative feasibility of différent reactions, and to calculate the energetic yield of the process and the associated cost (Hermann, 1999). Given that it is difficult to measure the absorbed light in heterogeneous photocatalysis, it is assumed that ail the light is absorbed and the efficiency is quoted as an apparent quantum yield (Linsebigler et al., 1995). Hence, to détermine the efficiency of the process, a combination of ail the pathway probabilities for the électron and hole must be considered.  19  2. Literature Review Ideally, <j> is proportional to the rate of the charge transfer process (kcr) and inversely proportional to the sum of the charge transfer rate (kcr) and the electronhole recombination rate (kR) (bulk and surface). Mathematically, this relation is given by (Linsebigler et al., 1995):  ^ 7 ^ 7 -  (2-25)  KCT + KR  Equation 2.25, assumes that diffusion of the products into the solution occurs quickly without the reverse reaction of électrons recombining with donors and holes recombining with acceptors. Without recombination, the quantum yield would take on the idéal value of 1. However, in a real System recombination does occur and the concentration of électrons and holes at the surface is not equal (Linsebigler et al., 1995). One way to improve the yield in a photocatalytic process is to sélect a semiconducting material that should be capable of reversibly changing its valence state to accommodate a hole without decomposing the semiconductor. In addition, the semiconductor should also hâve suitable band-gap énergies, stability toward photocorrosion, nontoxic nature, low cost, and physical characteristics that enable them to act as catalyst (Kabra et al., 2004).  2.4.2 Kinetic of photocatalyzed reactions In heterogeneous Systems, the reaction kinetics relied for many years on the Langmuir-Hinshelwood  (L-H)  model  to  interpret  expérimental  observations.  Considering Equation 2.24, the observed variation in the initial dégradation rate of the organic compound (-r) with organic concentration (P) can be described by Equation 2.26 (L-H kinetics) where kLH is the apparent LH rate constant and KL is the L adsorption constant.  r  _ kLHKLP l + KLP  (2.26) 20  2. Literature Review This kinetic model, in its simplest form, assumes a relatively rapid reaction achieving adsorption equilibrium, followed by a single, slow surface reaction step (rate determining step) (Emeline et al., 2005; Ollis, 2005). The following équations describe the mechanism where P represents the organic pollutant either in bulk solution (liq) or adsorbed on the catalyst surface (ads) (Mills et al., 2006): Piui<r±U^Pads  Pads —^-> Products  (2 .27)  (2.28)  The L-H adsorption/kinetic model assumes the following statements: i) at equilibrium, the number of surface adsorption sites is fixed; ii) only one substrate may bind at each surface site; iii) the heat of adsorption by the substrate is identical for each site and is independent of surface coverage; iv) there is no interaction between adjacent adsorbed molécules; v) the rate of surface adsorption of the substrate is greater than the rate of any subséquent chemical reactions; and vi) no irréversible blocking of active sites by binding to product occurs (Fox and Dulay, 1993). Based on thèse assumptions, the reactant equilibrium coverage, 9, is related to the concentration of the pollutant P, and to the apparent adsorption equilibrium constant  {Kads=->-)by.  K e  adsP  (2.29)  Hence, the reaction rate could be:  -r = kLH6P=^-f^ i + V  (2.30)  21  2. Literature Review where kw = k2 is the reaction rate constant, which is expected to be dépendent on the photon flux (/) incident over the catalyst surface (TIO2 for example). However, during the last décade, some key experiments hâve brought évidence that this dependency is not consistent with the obtained results (Ollis, 2005). Some of the flaws of the L-H kinetic model discussed by Ollis (2005) are the followings: i) while adsorption isotherm and reaction rate may each follow the same analytic form, i.e., a saturation function of the Langmuir adsorption isotherm form, the dark adsorption equilibrium constant, Kacis, is not found to be the same as the apparent adsorption constant, K%%, in the rate équation; ii) some reaction and adsorption Systems hâve différent analytic forms, iii) the L-H parameters kLn and K%% collected from différent experiments using phénol and 4-chlorophenol hâve positive corrélations, iv) a significant numbers of papers hâve explored the influence of the reactant concentration (in this case P) and light intensity, /, in the same photocatalyzed reaction, v) Emeline et al., (2000) found that the L-H rate parameters for phénol oxidation vary as follows: ki_H proportional to intensity and K"% inversely with intensity. Therefore, a différent approach was proposed to define a new simple model that can be described with the same Equations 2.26 and 2.27. Instead of assuming equilibrium adsorption of reactants and, correspondingly, a slow ratecontrolling surface step, the revised model assumes a pseudo-steady-state hypothesis to the surface coverage (d0/dt = 0). Under thèse assumptions, the reaction rate expression is (Ollis, 2005):  r.  -r=  m  f^appp ads  l + K?JP  (2.31)  where  KZ=T\~  (2.32)  22  2. Literature Review As it can be easily identified, Equation 2.31 has exactly the same mathematical form as the L-H model; however, it can explain the dependence of KaiadS on the incident light intensity (/). One important contribution is that k is related to / based on the following expression (Mills et al., 2006):  kLH=odp  (2.33)  where a is a proportional constant and (3 has been traditionally reported with a value of 0.5 under high-photon-flux conditions when the surface reaction dominâtes. Nonetheless, Emeline et al. (2005) stated that the dependence of the reaction rate to the photon flux, depending on the concentration of the reactant molécules, varies from zero-order kinetics at low concentration to first order kinetics at high concentration. Therefore, at intermediate concentrations, the order of the reaction on photon flux is between 0 and 1, and then, the square root dependence is a particular case lying within this range. Another important aspect in photocatalytic reactions is the availability of oxygen which plays an important rôle in the photooxidation process. Oxygen acts as an électron sink for photogenerated carriers, mainly électrons (Kabra et al., 2004). Therefore, the dependence of the photocatalytic dégradation rate on oxygen concentration can be also modeled by non-competitive Langmuir adsorption relation (Turchiefa/., 1989):  roc  K  °*C°> l + K0C0i  (2.34)  where K02 is the apparent constant for oxygen and C02 is the dissolved oxygen concentration. Hence, a complète model of the photocatalytic System should consider not only Equation 2.31, but also the dependence on oxygen concentration given by Equation 2.34.  23  2. Literature Review  2.5 Overview of photocatalytic reactors Photocatalytic reactors constitute an important part of every photocatalytic process. The reactor will not only affect the overall rate of photocatalytic dégradation of the organic/inorganic compounds, but also define aspects such as irradiation distribution, mass transfer phenomena, among others (Bouchy and Zahraa, 2003). While some of the physic-chemical principles of photocatalysis (e.g. production process, photoactivity, etc.) are relatively well understood, reactor design and reactor engineering of photocatalytic units still require a lot of considérations (Fox and Dulay, 1993; Linsebigler et al., 1995). In particular, this lack of knowledge is true for scaled reactors processing large volumes of water and using high levels of irradiation (De Lasa et al., 2005). There are several important factors affecting the efficiency of photocatalytic reactors with heterogeneous photocatalyst. Thèse include mode of opération, flow characteristics, reactor geometry, sélection of irradiation source, contact between the catalyst and the reactants, the durability of the photocatalyst, etc. (Vaisman et al., 2005; Kabir, 2006). There is a gênerai agreement that controlling and improving ail thèse factors will help establish intermediate and large scale reactors, as demanded by industrial and commercial applications (De Lasa et al., 2005). Photocatalytic reactors can be classified in terms of the type of illumination (artificial or natural) and in terms of the amount of the irradiation applied (number and kind of lamps to be used). However, a classification based on the state of the photocatalyst is more frequently used. In gênerai, two main groups of reactors can be defined: when the solid phase (the catalyst) is stationary within the reactor (immobilized photocatalyst), and when the catalyst is dispersed within the reactor (suspension or slurry reactors). In the latter case, fine powdered photocatalyst (e.g., Ti0 2 ) is freely dispersed in water providing high efficiency due to the large surface area available for reaction and efficient mass transfer within such Systems (Bouchy and Zahraa, 2003). However, due to the small size of the Ti0 2 particles (30-200 nm), it is difficult to separate them from water after used, making the post-treatment catalyst recovery stage necessary. This post-treatment catalyst recovery would be undesirable at larger scales as it would add to the overall cost of the treatment 24  2. Literature Review process (De Lasa et al., 2005). Thus, there is a practical necessity to immobilize the photocatalyst to overcome this issue. Immobilization, on the other hand, leads to mass transfer résistance which results in the réduction of reactor efficiency and in the accuracy of measured catalyst efficiency and kinetics (Kulas et al., 1998; Lee and Lee, 2004; McMurray et al., 2004). Table 2.3 summarizes the advantages and disadvantages of slurry and immobilized photocatalytic reactors.  Table 2.3: Comparison between suspended and immobilized photocatalytic Systems (De Lasa et al., 2005; Kabir, 2006). Reactor Advantages Disadvantages  Slurry reactors  Uniform catalyst distribution  Requires filtration  High surface area  Light scattering and adsorption in the suspended médium  No mass transfer effects  Operational costs are high  Well mixed particles suspension Low pressure drop through the reactor Initial cost is low If the catalyst has good attrition résistance, there is no need of an additional séparation opération Immobilized reactors  Opération cost are generally cheaper  Low light utilization efficiencies due to light scattering Mass transfer limitations Possible catalyst deactivation and catalyst wash out Initial cost is high due to the catalyst immobilization  25  2. Literature Review One alternative to the above mentioned photocatalytic reactors is the fluidized bed reactor (FBPR). The catalyst, which is usually immobilized on a support material, is highly agitated while the fluid is passed up through a bed of catalyst. Some of the support materials frequently used in FBPR are ceramic particles (Kanki et al., 2005), glass beads and silica gel particles (Bideau et al., 1995; Zhang ef al., 2006; Choi et al., 2007; Qiu and Zheng, 2007), zeolites (Haque ef al., 2005), activated carbon (Lee et al., 2004; Dong et al., 2008), etc. In gênerai, it is reported that fluidized bed photoreactors provide an efficient contact between the catalyst and the pollutants, low mass transfer résistance, and better surface exposure to UV radiation (Satterfield, 1980; Haarstrick étal., 1996; Pozzo étal., 1999, 2000; Bouchy and Zahraa, 2003; Imoberdorf et al., 2008). It is also possible to control and to improve the pénétration of light into the fluidized bed by varying its expansion (Haarstrick et al., 1996). Among différent configurations of FBPR, an annular photoreactor (two coaxial cylinders that define the reaction zone with the lamp usually placed on the axis) has been suggested as a very convenient one because the light is absorbed entirely by the reactive System in the annular région (Vaisman et al., 2005). In addition, this configuration meets the requirements of higher surface area-to-volume ratio, which is typically much lower in fix-bed configurations. Attrition of particles is a serious concern in fluidized bed reactors. The particle motion inside the reactor causes the catalyst attrition due to interparticle collisions and bed-to-wall impacts. It is one of the major obstacles in the development of large scale applications and/or for existing processes whenever the catalyst is changed (Werther and Reppenhagen, 1999). The main conséquence of the attrition is the génération of fines or smaller particles due to the severe impacts which occurs in a fluidized bed and the elutriation of thèse fine particles from the bed results in high catalyst losses, with the subséquent activity decays. Also, potential downstream line blockage with the generated fines will happen and additional filtration stage would be necessary (Werther and Reppenhagen, 1999; Qiu and Zheng, 2007). Catalysts containing major quantities of Ti0 2 hâve reported to hâve the disadvantage of poor  26  2. Literature Review physical integrity and attrition résistance when used in fluidized bed photocatalytic reactors (Cullo et al., 1986).  2.6 Sol-gel Process 2.6.1 Background and fundamentals The interest in the sol-gel (solution - gelation) processing of différent inorganic, ceramic and glasses materials, began as early as the mid-1800s with the study on silica gels by Ebelman (1846 and 1847) and Graham (1864) who reported that the hydrolysis of tetraethyl orthosilicate (TEOS) yielded Si02 under acidic conditions (Hench and West, 1989). Since early 1970's, this technology has gained much attention, especially in glass and ceramics fields, after a borosilicate glass lens was produced by heating a compact mass formed from sol-gel derivate oxide powders (Sakka, 2006). The most important reason for the significant growth and popularity of the solgel technology may be attributed to its low température processing. In comparison to other methods that require much higher températures, the sol-gel process allows better homogeneity of multi-components materials. Also, the rheological properties of sols or gels allow the formation of fibres, films or composites materials (Livage et al., 1988; Livage et al., 1989; Wen and Wilkes, 1996; Campanati et al., 2003). The sol-gel process, which is based mainly on inorganic polymeralization reactions, is defined as a séquence of Chemical synthesis method that involves the low température synthesis of an inorganic network (Bischoff and Anderson, 1995; Wen and Wilkes, 1996). A sol is defined as a colloïdal (suspension in which the dispersed phase is so small that gravitational forces are negligible and interactions are dominated by short-range forces) of solid particles in a liquid. On the other hand, a gel is defined as a semirigid mass formed when the colloïdal particles are iinked to form a network or when the polymer molécules are cross-linked or interlinked. Therefore, in the sol - gel process, the precursors or starting compounds for the préparation of a colloid consist of a métal or metalloid élément surrounded by 27  2. Literature Review various ligands (Brinker and Scherer, 1990; Rahaman, 2007). During the sol-gel process, hydrolysis and polycondensation of the metal-organic compounds, such as métal alkoxide M(OR)n where M represents a network-forming élément such as Si, Ti, Zr, Al, etc., and R is typically an alkyl group (CxH2x+i), occur to form oxopolymers, which are then transformed into an oxide network (Su et al., 2004; Gao et al., 2006). Métal alkoxides are the most common precursors used in sol-gel process because they react readily with water (Brinker and Scherer, 1990; Rahaman, 2007). Sol-gel method is one of the most important techniques for preparing nanosized metallic oxide materials with high photocatalytic activities (Su et al., 2004). During the photocatalyst production, the morphology, average particle size and particle size distribution, porosity, phase composition and crystallinity are important factors to be controlled (Gao et al., 2006). Therefore, by tailoring the chemical structure of the primary precursor and carefully controlling the différent process steps and variables during the sol-gel préparation, ail those parameters can be controlled, giving a final product with good homogeneity, purity and desired microstructure (Su et al., 2004; Gao et al., 2006; Keshmiri et al., 2006).  2.6.2 Sol-gel process steps The sol-gel process has a number of stages that will define the properties of the final product. Figure 2.3 shows différent stages of the process and the séquence in which they are carried out.  •  28  2. Literature Review  Figure 2.3: Sol-Gel chemistry séquence stages (Adapted from Optoweb, 2009).  The properties of the final product dépend on how well controlled are the seven steps involved during the sol-gel process (Figure 2.3). The main characteristics of some of thèse steps are described below (Hench and West, 1989; Brinkerand Scherer, 1990; Scherer, 1990; Richerson, 2006; Rahaman, 2007): •  Mixing: a suspension of colloïdal powders is formed by mechanical mixing of colloïdal particles in water at a pH that prevents précipitation. In gênerai, the liquid alkoxide precursor is hydrolyzed by mixing with water. The hydrated product interacts in a condensation reaction producing an oxide network. The water and the alcohol expelled from the reaction remain in the pores of the network. The size of the sol particles and the cross-linking within the particles dépend upon the pH and the ratio between water and the precursor added to form the sol, among other variables.  29  2. Literature Review •  Gelation: During the gelation, the colloïdal particles and condensed species link together to become a three-dimensional network. As the sol particles grow and collide, condensation occurs and macroparticles are formed; the sol becomes a gel when it can support a stress elastically, which is typically defined as the gelation point or gelation time. This change is graduai as more and more particles become interconnected. The physical characteristics of the gel network dépend greatly upon the size of particles and cross-linking before gelation. It is believed that the sharp increase in viscosity that accompanies gelation essentially freezes in a particular polymer structure at the gel point.  •  Aging: During the aging process, the gel is maintained in contact with the liquid and its structure and properties continue to change long after the gel point. Four différent processes can occur, separately or simultaneously, during aging. Thèse include polycondensation, syneresis, coarsening, and phase transformation. Polycondensation reactions continue to occur along with localized solution and reprecipitation of the gel network, increasing the thickness of interparticle necks and decreasing the porosity. Syneresis is defined as the spontaneous shrinkage of the gel and the resulting expulsion of liquid from the pores is expected. Coarsening is the irréversible decrease in surface area through dissolution and reprecipitation processes. In gênerai, it is expected that the strength of the gel increases with aging.  •  Drying: During drying, the liquid is removed from the interconnected pore network. Large stresses can be developed during this process causing the gels to crack uniess the drying process is controlled via decreasing the liquid surface energy, hypercritical evaporation, or obtaining monodisperse pore sizes by controlling the rates of hydrolysis and condensation. There are three main stages during the drying process. Stage 1 is characterized by the decrease in volume of the gel and the gel network is deformed by the large capillary forces which cause shrinkage of the object. This stage ends when shrinkage ceases. Stage 2 begins when the critical point is reached; this critical point occurs when 30  2. Literature Review the strength of the network has increased, due to the greater packing density of the solid phase, sufficient to resist further shrinkage. Stage 3 is reached when the pores hâve substantially emptied and the remaining liquid can be released only by evaporation from within the pores and diffusion to the surface.  •  Densification:  Heating the porous gel at high température induces its  densification. This stage is also called sintering and it is essentially the removal of the pores between the starting particles, combined with the growth and strong bonding between adjacent particles.  Based on the process conditions, such as the starting solution, pH, and température, différent products, in terms of gel physical characteristics, are formed. Thèse products can be classified as particle networks, aggregate networks, and polymer networks. The polymer network results in a three dimensional gel structure which can be transformed to a monolithic spécimen through the control of drying and heat treatments. When the aggregate clusters formed in the sol exhibit low and nonuniform density distribution, a précipitation of the solid phase occurs (aggregate network). High and nonuniform density distribution of particle aggregates resuit in a particle network which is considerably uniform and monosized (Ring, 1996). Figure 2.4 represents thèse three products.  31  2. Literature Review  Monomeric solution  Polymer synthesis.  >*~ /  Particle synthesis  //  Linear Polymer  Crosslink  Polymer network  Random Aggregates  Aggregate  Aggregate network  Particulates  Aggregate  Particulate network  Figure 2.4: Schematic of routes for structural évolution of métal organic in solution (Adapted from Ring, 1996).  As was already pointed out, the concept of sol-gel processing covers a large variety of materials, specially glasses and ceramics. Nonetheless, modifying the séquence or even the conditions of one or more than one of the stages shown in Figure 2.3, leads to différent products in différent forms. Some of thèse are xerogel, which is just a dried gel (xero means dry), aerogel which is a gel where the liquid component has been replaced with a gas, uniform particles (powders), films and self-supported products (Figure 2.5).  32  2. Literature Review  Heat  Métal alkoxide solution  Heat  Evaporatior Xerogel  Dense Ceramic  Hydrolysis Polymerization Extraction of solvent Aerogel Sol  Precipitating  OOOOOOOO  0000000e 000000*0 O000OO@O 00000000  Uniform particles Figure 2.5: Overview of the sol-gel process and various sol-gel derived products (Adapted from Brinker and Scherer, 1990).  2.6.3 Reactions and chemical characteristics Many chemical reactions hâve been found along the seven stages described in Section 2.6.2. The sol-gel chemistry is based on inorganic polymerization reactions and two routes are usually described depending on whether the precursor is an aqueous solution of an inorganic sait or an alkoxide in an organic solvent (Livage et al., 1989). In either case, the solid network is formed, from the solution, via the hydrolysis and condensation of the molecular precursors in solution (Livage étal., 1998). The sol-gel process generally starts with alcoholic or other low molecular weight organic solutions of monomeric, métal or semimetal alkoxide precursors. Once thèse solutions are mixed with water, hydrolysis and condensation reactions  33  2. Literature Review occur simultaneously (Wen and Wilkes, 1996). Hydrolysis and condensation occur by nucleophilic substitution (SN) mechanism which involves three steps: 1. Nucleophilic addition (AN) 2. Proton transfer within the transition states 3. Removal of the protonated species as either alcohol or water. Thèse two reactions (hydrolysis and condensation) usually proceed with either an acid or base as catalyst; therefore, the structure and morphology of the resulting network strongly dépend on the nature of the catalyst. Other important factors are the pH of the reaction, the solvent used, the amount of water added, the séquence of mixing, and the size of the aikoxy group because of its steric effect (Nazeri étal., 1993; Wen and Wilkes, 1996). The structure of a gel is established at the time of gelation. Subséquent processes such as aging, drying, stabilization and densification dépend upon the gel structure. Since it is the rate of hydrolysis and condensation that détermine the structure of the gel, it is important to understand the kinetics of the hydrolysis and condensation reactions:  Hydrolysis: Hydrolysis of the alkoxide occurs upon adding water or a water/alcohol solution, and a reactive M-OH hydroxo group is generated. The gênerai hydrolysis reactions can be detailed as follows:  M(OR)n + nH20 -» M(OH\ + nROH  (2.35)  However, this reaction is more complex and can be represented with Equation 2.36 (Livageefa/., 1989):  34  2. Literature Review  H \ -> 0:->M-OR  H-O + M-OR I l H  -> HO-M<-0  H  (a)  (b)  I  R  \ H  -> M-OH+ROH  (c)  (2.36)  (d)  The first step (a) is a nucleophilic addition of a water molécule to the positively charged métal atom M. This leads to a transition state (b) where the coordination number of M has increased by one. The second step involves a proton transfer within (b) leading to the intermediate (c). Then, a proton from the entering water molécule is transferred to the negatively charged oxygen of an adjacent OR group. The last step is the departure of the living group which should be the most positively charged species within the transition state (c). AN thèse processes follow a nucleophilic substitution mechanism and the charge distribution governs the thermodynamics of thèse reactions (Livage et al., 1988).  Condensation: Condensation occurs as soon as hydroxo groups, M(OH)n, are formed. Depending on the expérimental conditions, three compétitive mechanisms are considered (Livage et al., 1988; Livage étal., 1989):  > Alcoxolation: This is a reaction by which the bridging oxo group is formed through the élimination of an alcohol molécule. The mechanism is basically the same as that for hydrolysis with M replacing H in the entering group.  R  MM \ O + M-OR  \ <+  /  H  0:->M-OR  I <-> M-0-M<-:0  /  H  <-> M-O -M + R -OH (2.37)  \  H  > Oxolation: Follows the same mechanism as alcoxolation, but the R group of the leaving species is a proton (water molécule).  35  2. Literature Review  M M \ 0 + M-OH ^ H  H \ I 0:->M-OH +* M-0-M<-:0 <-> M-O-M + H20 H  (2.38)  H  > Olation: It can occur when the full coordination of the métal atom is not satisfied in the alkoxide (N-z *• 0). In this case, bridging hydroxo groups can be formed through the élimination of the solvent molécule which will be water or alcohol, depending on the water concentration.  M-OH+  M<-:0  I  H  R H I M-OH + M<-:0 H  M \ <-> 0:->H+R -OH  (2.39)  M M  \ <-> 0:->H+H20 M  (2.40)  Titanium alkoxide, particularly titanium isopropoxide (Ti(OC3H7)4), is one of the typical starting precursors for Ti0 2 photocatalysts. Therefore, the overall chemical reaction is as follows:  T^OC.H, \ + 2H20 -> Ti02 + 4(C3//7 )OH  (2.41 )  Figure 2.6 shows the hydrolysis and condensation of this alkoxide in the présence of water. In gênerai, Ti (IV) aikoxides form pentacoordinate and hexacoordinate complexes in the présence of water. Then, condensation happens once the M-OH bonds are formed. The resulting tetrahedral Ti(OR)(4-n)(OH)n groups can form linear polymers as shown in Figure 2.6 (Laine, 1990).  36  2. Literature Review  H  X  > -  O  xL O •••- Ti! -•••• no^ = + H20 o  y  H  H >0,  X1A\ R  R7X';rR  / Ti \ \ Ô'  / / Ti \ \ Ô \  R  Ô \R  R  X  > ~  O  O H  x^o  Ti  •", ••-.o  0  0  O ^ /  H...  ...- Ti .... O' ! O-H  Ao'  h  -IPrOH •  ,  0  o  y-  / /° \  H  A .-- -•.. /  O  y  H  y  r 0>-  Ao.  rS° :•;;,. /  ^  "  O••••"'/  •  i  ""O-H  y  Higher Oligomers Figure 2.6: Hydrolysis and condensation reactions of titanium isopropoxide in the présence of water (Adapted from Laine, 1990).  Because of the high reactivity of transition métal aikoxides with water, précipitation is usually observed instead of gelation. However, this can be modified by introducing additives such as solvents, acidic or basic catalysts, etc. In many cases, nucleophilic molécules, e.g. XOH, react with aikoxides resulting in a new 37  2. Literature Review molecular precursor. The new molecular precursor reacts differently with the nucleophilic agent. The more electronegative ligands are removed during the condensation reactions, while the less electronegative ones are quickly removed at the initial stages of hydrolysis. Then, instead of having a fast précipitation, a controlled formation of polymeric gels is expected (Danuwila et al., 1994) In some cases, like in this investigation, the hydrolysis of the alkoxide is performed in the présence of an acid catalyst (e.g., HCI, HNO3) that allows for controlling the rate and the extent of the hydrolysis reaction. At the same time, polycondensation reactions can be delayed either by the use of titanium organic compounds with high molecular weight alkyl groups or through the reactions which are acid catalyzed increasing the rate of hydrolysis.  2.6.4 Sol- gel advantages and disadvantages The advantages of the sol-gel processing over other conventional processes can be listed as follows (Brinker et al., 1984; Brinker et al., 1986; Brinker et al., 1988; Brinker and Scherer, 1990; Wen and Wilkes, 1996; Livage, 1998; Sakka, 2006):  •  High purity materials can be obtained if the necessary care is exercised during the sol-gel process.  •  Processed materials can be very reactive and hence, the sintering températures are much lower than those employed for conventional by prepared powders. This is mainly because of the high surface energy of the sol-gel-derivated colloidal particles.  •  Lower température process leads to products with large spécifie surface area and high porosity and macro pore ranges.  •  Low calcination températures, needed for the sol-gel-derived materials, help to better control the size and crystallinity of the resulting particles. At the same time, lower températures mean lower energy costs.  •  Spécial materials or configurations hâve been developed altering the sol conditions, e.g., thin films, controlled-size spherical powders, fibres, among other. 38  2. Literature Review •  Mixing of multiple components to produce a variety of organic/inorganic hybrid materials.  •  Gelation permits the moulding of near-net shapes in applications where machining of those shapes is expensive. The shape and surface configuration will be retained in the body, despite the large dimensional change due to shrinkage.  •  Improved homogeneity of multicomponent species can be obtained by blending a variety of métal alkoxides, colloidal dispersions or easily diffused soluble salts.  On the other hand, the disadvantages of the sol-gel processing are the followings (Brinker et al., 1984; Yang, 1999; Keshmiri, 2004): •  The high-purity alkoxides are relatively expensive raw materials.  •  The process is time consuming and multi-step.  •  Génération of a significant amount of waste of some Chemicals (e.g. alcohol).  •  The shrinkage of the resulting gel during drying and heat treatment is high (as high as 90% for materials with high solvent content), potentially causing micro and macro cracks within the dried and calcined body.  •  If the thermal processing is not properly completed, the remaining solvent and carbonaceous residues, if not properly removed, can lead to defects such as bloating, residual bubbles or crystal formation.  2.6.5 Sol-gel applications Because of its multiple advantages, the sol-gel process is one of the most important techniques for preparing nanosized metallic oxide materials with high photocatalytic activities (Su et al., 2004). In récent years, différent materials hâve been coated with photocatalyst using sol-gel principles. Examples of thèse include activated carbon (Lee et al., 2004; Dong et al., 2008), ceramic particles (Kanki et al., 2005), glass beads and silica gel particles (Bideau et al., 1995; Choi et al., 2007), zeolites (Hisanaga and Tanaka, 2002; Haque et al., 2005), glass slides (Kumara et  39  2. Literature Review al., 1999; Guillard et al., 2004), stainless steel (Fernândez et al., 1995) among others. Also, many other materials hâve been produced for photocatalytic purposes, such as membranes, powders, nanoparticles, microspheres, etc. (Bessekhouad et al., 2003; Su et al., 2004; Choi et al., 2006; Choi et al., 2007; Junin et al., 2008; Wetchakum and Phanichphant, 2008; Xu et al., 2008; Shi et al., 2009). There are also a significant number of other materials produced using sol-gel principles for différent applications. Some of those materials include différent kinds of composite materials, aerogels, plastic ophthalmic lenses, water repellent coatings etc., (Que et al., 2001; Mackenzie, 2003; Chen et al., 2004). Among ail those materials, composite materials hâve gained a lot of attention because of their advantages and properties.  2.6.6 Composite materials One of the main applications of sol - gel process is related to the production of the sol-gel-derived composites. Generally speaking, materials formed by the solgel process hâve a relatively high rate of volume shrinkage, which causes considérable cracking and subséquent mechanical failure of the final products. At the same time, it is well known that the dried gel structure is usually weak due to continuous  pores with unreacted and trapped organics, water, and other  compounds. Thèse problems can be overcome by the introduction of a secondary phase (e.g., pre-calcined powder) as a filler to reinforce the material structure, producing a material with a composite structure. Composite materials, by définition, usually contain two (or more) distinct materials as a unified combination. The principal material of a composite that envelopes the reinforcement is called matrix and the reinforcement is usually called a filler or inclusion (Keshmiri, 2004). Through this process, not only is the shrinkage of the gel expected to decrease because of the présence of a significant amount of inert ceramic powder, but also the sol-gel-derived material acts as a binder to bond the entire structure. The sol-gel-derived material, présent in between the neighbouring particles, also acts as a sintering aid to decrease the sintering température (Keshmiri, 2004). 40  2. Literature Review There are various pathways for processing sol-gel-derivated composites; Figure 2.7 shows thèse pathways (Nazeri et al., 1993): 1.  Mixing of two or more sols to form a homogeneous solution (Figure 2.7 (a)). The différent components may be tailored so that they do not react with each other to form a new component. This method allows uniformity of the composites.  2.  Dispersion of a solid phase, such as fine powders or fibres, in a sol before gelation, which leads to a composite with good homogeneity and good contact between particles and matrix (Figure 2.7 (b)).  3.  Imprégnation of the fine interconnecting pores by organic or inorganic phases (Figure 2.7 (c)).  4.  Infiltration of the fine interconnecting pores by organic or inorganic phases (Figure 2.7 (d)).  5.  Combination of # 2 and # 3 in to give a "triphasic" composite.  While preparing a sol-gel composite material, there are some limitations to take into account, like the ones listed below (Lutz and Swain, 1991 ; Yang, 1999):  1. Agglomération of ceramic particles in sol precursor bécomes critical to further improve the technology of composite-gels. 2. Différent thermal expansion coefficients between the inclusions and sol-gel matrices inducing additional stresses in the composite material. Hence, cracks may be formed from small pre-existing defects at or near the filler/matrix interfaces. 3. Composite sol-gel coatings cannot be densified enough to gain high strength and hardness at curing température below 1000°C  41  2. Literature Review  a) o  o  a®  l  ©  o°  +  0  Sol 1  Mixing •  : o . o *  % ° © 0 ° • 0 O e 0 °0  •  Sol 2  Gelation Drying » Firing  LmJ  Composite  Sol 3  b) •®  ©  • • •• •  0  : «  +  js&ISâli&lte.  Mixing •  Gelation Drying Firing  ns ssi  ©  0  Composite  Second phase  son  c)  l  Imprégnation  Gelation or polymerization Composite  d)  Gelation Drying 1 Firing Composite Figure 2.7: Proceeding routes for sol-gel-derived composites (Adapted from Nazeri et ai, 1993).  2. Literature Review The novel catalyst that is evaluated in this investigation is based in the principles described above. Basically a sol-gel matrix is produced to bond together pre-calcined particles together (TÏO2 powder). Then, the formation of the catalyst is done by the combination of the sol-gel-composite material with a polymeric solution leading to a fast and uniform formation of Ti0 2 catalyst.  43  Chapter Three Scope and Objectives 3.1 Scope Over the past two décades, significant efforts hâve been made towards the applications of photocatalytic TÏO2 in various forms such as coatings on différent surfaces and supports, fibres and powders with a wide range of particle sizes and surface areas. However, there are still important number of unsolved problems or drawbacks with respect to the photocatalyst itself and its large scale applications. Some of thèse challenges are: •  the séparation and filtration of the catalyst from the processed fluid (e.g., treated water) when a powder form of photocatalyst is used,  •  the diffusive mass transfer résistance and low surface area being limiting factors  for  the  effectiveness  of  the  processes  involving  supported  photocatalyst, •  relatively low photoactivity and significant attrition issue, producing low photoefficiency of the System and process complications such as downstream filtration and séparation of the suspended particles, for the fluidized bed photocatalytic process. This research focuses on the development and standardization of template  free composite titania photocatalysts, in the form of sphères, to be used in a fluidized bed photoreactor. The photocatalyst is synthesized using sol-gel principles to bond T1O2 precalcined particles (Degussa P-25) together, which produced a composite Ti0 2 sphères with high photoefficiency and high attrition résistance. Thèse sphères are evaluated in a laboratory scale fluidized bed photocatalytic reactor treating two model compounds.  44  3. Scope and Objectives  3.2 Objectives The spécifie objectives of this research were to: 1. Examine the effect of parameters such as: a. sol composition b. amount of catalyst to be added to the sol c. calcination température and duration of the heat treatment d. drying process conditions (température and time) in order to détermine their effects on the photocatalytic activity and attrition résistance of the TÏO2 sphères. 2. Evaluate the attrition résistance of the TÏO2 sphères, developing a method to détermine the amount of TÏO2 particles released in water after the normal opération of the fluidized bed photoreactor. 3. Evaluate the ability of the composite "TÏO2 sphères to dégrade water contaminated with a. Formic Acid (FA) as a model organic compound b. 2,4-Dichlorophenoacetic acid (2,4-D) as a model micropollutant c. Natural Organic Matter (preliminary tests) 4. Improve the formulation and synthesis process of the photocatalytic sphères considering the results obtained in objectives 1-3. 5. Characterize the Ti02 sphères in terms of a. spécifie surface area of the TIO2 sphères (BET) b. change in weight with respect to the drying process and heat treatment using thermo gravimétrie analysis (TGA) c. phase évolution of the photocatalyst using X-ray diffraction (XRD) d. microstructure analysis of the catalyst using Scanning Electron Microscope (SEM)  45  3. Scope and Objectives  3.3 Significance of this work This research with its objectives provides the first steps towards the development and standardization of the novel template-free composite Ti0 2 photocatalyst. The photocatalyst, with its high attrition résistance and high photoactivity, will significantly contribute to the research and development activities performed towards the implementation of large-scale commercial photocatalytic treatment Systems. Although the treatment of contaminated water by means of heterogeneous photocatalysis has been demonstrated to be a promising technology, the development of commercial  Systems has not been successfully achieved.  Depending on the type of application, a number of limitations and challenges, contributing to this issue, include low photo-efficiency and quantum yield of the catalyst, expensive séparation of the catalyst from treated water, and weak adhésion of the catalyst on the support materials. Nonetheless, the development of such commercial Systems would be a great improvement to the water treatment industry, since i) many organic compounds can be mineralized during photocatalysis, ii) a significant amount of the dégradation by-products are environmentally benign, iii) in most cases low energy UV would be needed for the catalyst activation. The development of composite TiÛ2 sphères, that possess good attrition résistance and provide high photo-efficiency, address some of the outstanding challenges and will contribute  significantly  to  the  implementation  of  large-scale  commercial  photocatalytic Systems. The whole photocatalyst is made of pure Ti0 2 (no support required), has good photoactivity to dégrade pollutants in water, and has high mechanical stability (attrition résistance) that reduces the amount of Ti0 2 particles that need to be filtrated downstream.  3.4 Layout The abovementioned objectives were achieved at différent stages along the project. A detailed description of how each objective has been met with respect to their présentation in the following chapters is shown below:  46  3. Scope and Objectives Chapter Four: describes the expérimental methodology that was followed to meet ail the objectives described above. A detailed description of the procédure to produce the catalyst and the variables involved during the process is presented. Additionally, a detailed explanation of the characterization of the catalyst in terms of the surface area (BET), microstructure (SEM), phase évolution (XRD) and TGA analyses are provided giving the necessary background to ail thèse techniques. Finally, the procédures to détermine the activity of the catalysts (using FA, 2,4-D and natural organic matter), the attrition résistance of the catalyst and the fluence rate in the System are explained. Chapter Fîve: describes primarily the preliminary expérimental results that were obtained with the sphères using the formulation before any modification. Aspects such as catalyst loadings in the reactor, bed expansion, mass transfer within the System, and the impact of the calcination température on the FA dégradation are some of the points developed and presented along this chapter. The original level of attrition of the sphères is presented as référence and benchmark for additional improvement. Preliminary characterization of the sphères was done in terms of microstructure (SEM), spécifie surface area (BET) and polymorphic phases détermination (XRD). Chapter Six: covers the analysis of the effect of différent variables on the attrition résistance and activity of the catalyst (objective 1 ) and the détermination of the final formulation to produce the composite Ti0 2 sphères (objective 5). Detailed characterization of the catalyst produced with the new formulation was performed and the results are presented. Chapter Seven: présents the results of the experiments performed with the composite Ti0 2 sphères to dégrade FA under différent UV irradiations. Also, 2,4-D was used as a contaminant to evaluate the ability of the photocatalyst to dégrade more complex organic molécules in water. Preliminary results with NOM are presented and discussed as well. In addition, a detailed study of the fluence rate détermination in the reaction System used in this research is included in this chapter.  47  3. Scope and Objectives Chapter Eight: provides the main conclusions of this research as well as recommendations for future work.  48  Chapter Four Expérimental Methodology 4.1 Introduction This chapter provides a detailed description of the expérimental work performed during the course of this investigation. Detailed explanations of the production process of the photocatalyst as well as différent techniques used to characterize the photocatalyst are discussed. Moreover, analytical techniques used during the expérimentation, e.g., the protocol to monitor Ti0 2 in suspension are presented. Finally, a detailed description of the fluidized bed photoreactor (FBPR) and its operational procédure are presented.  4.2 Composite Ti0 2 sphères préparation 4.2.1 Chemicals and reagents The list of chemicals that used to prépare the photocatalyst is provided in Table 4.1. Most of the chemicals were obtained from Sigma-AIdrich (Ontario, Canada) and Fisher Scientific (Ontario and Edmonton, Canada). The Ti0 2 Degussa P-25 was obtained from Degussa, Germany.  Table 4.1: Spécifications of the chemicals used during the préparation of the composite Ti02 photocatalyst. Chemicals Purity Chemical formula Ethyl Alcohol Denatured 85%/15% C2H6OI CH3OH Hydrochloric Acid Titanium (IV) Isopropoxide  HCl Ti(OCH(CH3\\  Titanium Dioxide Chitosan  Ti02 t-'n"  24™ 2*^9  38% 97% —  Médium Molecular Weight  Glacial Acetic Acid  C2H402  99.7%  Ammonium Hydroxide  NH4OH  30% 49  4. Expérimental Methodology 4.2.2 Composite sol-gel material (CSG) To prépare the composite sol-gel (CSG) material, anhydrous ethyl alcohol denatured was used as a solvent to prevent fast hydrolysis of the titanium alkoxide. Since water content of the sol has a critical rôle in the hydrolysis and polycondensation reactions during the sol-gel process, the amount of water needs to be well controlled. Milli-Q water was added to the alcohol followed by the addition of the catalyst, in this case HCI, while the solution was magnetically stirred. The catalyst is used to prevent fast gelation of the sol, control the rate of condensation, and eliminate the possibility of precipitate formation due to uncontrolled hydrolysis reaction. Hydrolysis was carried out by the addition of the precursor. Titanium (IV) isopropoxide was used as a precursor without fùrther purification, and it was added to the prepared solution (alcohol with water and HCI) while stirring. Différent molar ratios of the Chemicals were investigated. The mixture was stirred for 2 h and a clear and stable alkoxide sol was obtained using this process. Pre-calcined commercial Ti0 2 powder, Degussa P-25, was used as a filler material. This material was used because it exhibits high photoefficiency, high surface area (54 m2 g"1) and low charge recombination rate. Characteristics and properties of Degussa P-25 (e.g., spécifie surface area and phase composition) hâve been evaluated prior to incorporating it to the composite préparation process (Section 6.3). Good agitation was required to dissolve the powder into the solution and create a homogeneous solution called Composite SolGel (CSG). A similar procédure was reported before by Keshmiri et al. (2004).  4.2.3 Polymeric solution The polymeric solution was prepared by dissolving chitosan in water. Chitosan is a natural, non-toxic, biodégradable, high molecular weight polymer that was first described in 1811 and named by Odier in 1823. From the commercial point of view, chitosan is obtained by deacetylation of chitin, which is the structural élément in the exoskeleton of crustaceans (e.g., crabs, shrimps, etc.) and the cell walls of fungi and certain yeasts, among other sources (Ayers and Hunt, 2001).  50  4. Expérimental Methodology Chitosan is a linear cationic polysaccharide made out of two kinds p(1->4)linked monosaccharide residues, namely N-acetyl-D-glucosamine  (A) and D-  glucosamine (D). It is established that D-residues at acidic pH provide protonated NH3 groups which render chitosan soluble, whereas A-residues hâve an important rôle on the stiffness of the chains (Arguelles-Monal ef al., 1998) determined by hydrogen bonding between two adjacent hexopyranose rings along the chain, hindering the free rotation around glycosidic bonds and reinforced by an incrément on the hydrophobic macromolecular interactions (Skjak-Braek  ef al., 1989;  Arguelles-Monal et al., 1998). Therefore, chitosan was dissolved in Milli-Q water that was acidified with glacial acetic acid. To ensure a homogeneous solution, the solution was stirred for at least 24 h. Différent molar ratios of chitosan-water-glacial acetic acid were investigated. The pH of the solution was measured using a pH meter from Denver Instruments - UB5. The viscosity was measured using a viscometer from Cambridge Viscosity - VISCOlab 3000.  4.2.4 Composite Ti0 2 sphères formation Once the CSG material was ready to be used (after approximately 14 hours of agitation), it was mixed with the chitosan solution to create a solution with the desired viscosity needed for the sphères formation. Chitosan also acted as a binder of the CSG material. The CSG material was added to the chitosan solution while the solution was magnetically stirred. This new solution was agitated for 2 h before it was added drop wise to the ammonia solution to form the TiÛ2 sphères. The formation of the sphères was based on the principle that the CSG - chitosan solution would undergo a fast gelation when in contact with basic pH. After the sphère formation, drying was required to remove the liquids contained in the body of the sphères. Différent drying conditions were studied during this investigation; specifically, TiÛ2 sphères were placed in a furnace for 20 h at 80°C or they were left at room température (23°C) for 15 days. After the completion of the drying process, heat treatment was applied to densify the composite TiÛ2 sphères. The heat treatment was carried out at the Materials Engineering Ceramic Laboratory using high température furnaces. The composite TiÛ2 sphères were 51  4. Expérimental Methodology placed in ceramic crucibles to be able to heat them up to high température (600°C to 900°C). The heating rate for ail the samples were approximately 9.5°C min"1 up to 470 °C and 4.4°C min"1 up to the final températures. Among différent calcination températures and heat treatment conditions, the best conditions were selected to give a product with the désirable characteristics. Figure 4.1 shows the schematic and séquence of the abovementioned procédure for the production of the composite TiÛ2 sphères.  H20  C2H50  •  \  HCI  /„  fc w  Ti(OC3H7)  Stirring  Mixing with Ti0 2 precalcined powder (Degussa P-25) Stirring _J f Mixing with Chitosan  Drop-wise Heat treatment  • *  ^  Drying process  4  w  Sphères formation in NH4OH  M  ^  Figure 4.1 : Préparation of composite Ti0 2 sphères.  4.3 Composite Ti0 2 sphères characterization The objective of catalyst characterization was to examine the catalyst surface and its bulk properties that offer substantial information about the catalyst attributes. The information obtained from ail the différent characterization tools will significantly improve the understanding of how the physicochemical properties and catalytic performance are related.  52  4. Expérimental Methodology In this section, brief reviews of the basic concepts of various catalyst characterization techniques are provided.  In particular, BET surface area, pore  volume, pore size distribution, X-ray diffraction (XRD), scanning électron microscope (SEM) and thermogravimetric analysis (TGA) are described.  4.3.1 Surface area, pore volume and pore size détermination The détermination of total surface area, not only external but also internai, is generally  considered  an  important  catalyst characterization  (Haber,  1991).  Nowadays, adsorption is one of the most widely used methods to détermine the surface area and pore size distribution (Gregg and Sing, 1982). When a gas is in contact with a solid surface, physical adsorption will occur. Of course, the amount of gas adsorbed dépends on many variables such as equilibrium pressure, température, and the spécifie characteristics of the gas-solid System. Brunauer-Emmet-Teller (BET) method is a well known technique for the détermination of physical adsorption of gas molécules on a solid surface (Brunauer et al., 1938). The theory is an extension of the Langmuirtheory, developed based on the monolayer molecular adsorption on a catalyst surface. BET équation was derived for multilayer adsorption and its fundaments are on the relation between the volume of gas physically adsorbed and the total area of adsorbent, which is given in Equation 4.1:  c-\ f u\ fp\  _J  V  j  +—  (4.1)  mC V ' o /  v^y where P is the gas pressure; P0 is the saturation pressure of the adsorbate gas, v is the volume of gas adsorbed, v m is the volume of gas adsorbed corresponding to monolayer coverage, and c is a characteristic constant of the adsorbate. The total surface area was determined using a multipoint BET method using Micromeritics ASAP 2020 equipment (Figure 4.2) located at the Department of 53  4. Expérimental Methodology Chemical and Biological Engineering. The pore volume and the pore size distribution were calculated as well using the same equipment. To détermine the porosity of most solid materials, nitrogen (N2) at 77K is usually utilized as a suitable adsorbate. Thus, by constructing an adsorption isotherm, a better understanding of the porosity of the materials can be obtained. This isotherm is made by measuring the quantity of adsorbate on the surface of the solid over a wide range of relative pressures at constant température (Gregg and Sing, 1982). Similarly, desorption isotherm can be obtained by measuring the amounts of gas removed from the sample as the relative pressure is lowered. The isotherm shapes, which is recognized as BDDT (Brunauer, Deming, Deming and Teller) classification, dépends on the kind of pores that the solid has (e.g. micropores, mesopores and macropores) (Gregg and Sing, 1982; Sing et al., 1985; Hosokawa, 2007).  Figure 4.2: Multipoint BET equipment - Micrometrics ASAP 2020  The total pore volume is derived from the amount df vapour adsorbed at a relative pressure close to one, assuming that the pores are filled with liquid nitrogen. Pores which would not be filled below P/Po of one are considered negligible to the total pore volume; then, the average pore size can be estimated from the pore volume.  54  4. Expérimental Methodology The distribution of pore volume with respect to pore size is known as a pore size distribution. The pore size distribution was calculated using the method proposed by Barrett, Joyner and Halenda (BJH). This method is based on the emptying of the pores by a step-wise réduction of P/Po (Barrett et al., 1951). In gênerai, the desorption isotherm is more appropriate than the adsorption isotherm for the évaluation of the pore size distribution of a solid material because it exhibits a lower relative pressure resulting in a lower free energy state; therefore, the desorption isotherm is doser to the true thermodynamic stability (Gregg and Sing, 1982). Pore size calculations are made assuming cylindrical pore geometry using the Kelvin Equation (Sing et al., 1985):  r  «=-^TJS  («)  RT\n \P0J where y represents the surface tension of N2 at its boiling point (8.85 mN m"1 at 77K); vm is the molar volume of liquid N2 (34.7 cm3 mol"1), T is the N2 boiling point, P/P0 is the relative pressure of N 2 and /* is the Kelvin radius of the pore. The Kelvin radius is the radius of the pore in which condensation happens at P/Po. Nonetheless, the r* does not represent the actual pore radius because some adsorption has taken place prior to condensation on the walls of the pores. Besides, during desorption an adsorbed layer remains on the walls when evaporation occurs. Therefore, the real pore radius (rp) is given by Equation 4.3:  rp=rk+t  (4.3)  where t is the thickness of the adsorbed layer that can be estimated using the Equation 4.4 (DeBoer et al., 1966):  55  4. Expérimental Methodology  13.99  (4.4)  + 0.034  log [Pô)  The measurements with BET equipment were conducted by placing few grams of the catalyst (usually around 0.4 g) in the degasser section for at least 3 h at 100°C. The degassing was done prior to the analysis in order to eliminate any physisorbed species from the surface of the adsorbent. Then, the BET analysis was done using N2 as an adsorbate and liquid N2 to keep the température constant.  4.3.2 X-Ray diffraction spectroscopy (XRD) X-ray diffraction (XRD) spectroscopy has been commonly used for crystalline phase identification and for the average crystalline size détermination in a catalyst sample. It dépends on the fact that crystalline matter is composed of periodic arrays of atoms in three dimensions. As the wavelength of X-rays is of the same order of magnitude as the spacing between the atoms in crystals (~1 Â), crystals behave as a diffraction grating for X-rays. Then, by passing a monochromatic bean of X-rays through a crystal and by measuring the intensifies and angles of the diffracted beams at spécifie orientation of the crystal, the symmetry of the atomic arrangement of the crystal may be deduced and the nature of the atomic structure of the crystal can be solved and quantified (Raudsepp and Pani, 2003). In catalyst characterization, the catalyst sample is irradiated with X-ray of a known wavelength (A). This condition results in the reflection of X-rays by atomic layers with interplanar spacing d at a certain angle of incidence and reflection known as the Bragg angle (0). The relationship between thèse variables is given by Bragg's Law (Reed, 2005):  n = 2dsin0  (4.5)  56  4. Expérimental Methodology This équation follows directly from the différence in path lengths between successive planes where the integer n is the order of diffraction. The XRD measurements were performed on a Siemens D5000 X-ray diffractometer (Figure 4.3) at 40kV, 40mA. X-ray diffraction patterns were recorded using CoKa radiation (X= 1.7889Â) over the range 26 from 3 to 80°, with a scanning speed of 3°/min. The peaks obtained from the analysis were identified using the Diffracplus XRD identification method to détermine the crystalline phases présent in the sample. This equipment is located at the Department of Earth and Océan Science at UBC.  Figure 4.3: X-ray diffractometer- Siemens D5000  To prépare a sample, approximately 4 g of sphères were crushed to a fine powder (less than 10|im) and then placed on a spécimen holder and pressed using a glass slide. If the sample was not smaller than 10>m inaccurate intensities of the diffraction peaks could resuit because of extinction, micro-absorption contrast among the phases, prefer orientation of particles, etc. (Raudsepp and Pani, 2003). Rietveld refinement was carried out as well to détermine the quantitative composition of the phases formed during the calcination processes. In Rietveld method, a model for the shapes and widths of the diffraction peaks, a model for any aberrations in the shapes and positions of the peaks and a model for the background are defined for the crystal-structure of each phase. Thèse models were used to 57  4. Expérimental Methodology calculate a simulated powder diffraction pattern for each phase. The sum of the individual calculated patterns was then fitted to the digital expérimental diffraction pattern. The fitting was done by least-squares refinement of structural parameters of each phase (Raudsepp et al., 1990; Raudsepp and Pani, 2003).  4.3.3 Scanning Electron microscopy (SEM) Scanning Electron Microscope (SEM) was used to analyze the morphology (size and shape), topography (surface features) and crystallographic (atomic arrangements) of the photocatalyst. In the SEM unit, an électron gun is used to generate an électron beam with high intensity. Once the beam bombard the metal/carbon coated spécimen, secondary (SE) and backscattered électrons (BSE) are emitted from highlighted area. Local variation in surface topography, e.g. orientation of surface with respect to the électron beam, gives rise to contrast and then, images are produced by scanning the beam while displaying the signal from an électron detector (Haber, 1991; Reed, 2005). SE are produced by inelastic interactions of high energy électrons with valence électrons of atoms in the spécimen which cause the éjection of the électrons from the atoms; on the other hand, BSE consist of high-energy électrons originated in the électron beam, that are reflected or back-scattered out of the spécimen interaction volume by elastic scaterring interaction with spécimen atoms (Bozzola and Russell, 1998). A Philips XL-30 SEM unit, located at the Materials Engineering Analytical Laboratory, was used (Figure 4.4). SEM samples were prepared by placing few sphères, randomly picked, on the surface of the spécimen, and then either gold or carbon was used as coating to enhance the conductivity of the sample. The majority of the images were taken using SE with an approximate distance of 14 mm from the sample.  58  4. Expérimental Methodology  Figure 4.4: Scanning Electron Microscope (SEM) - Phillips XL-30  4.3.4 Thermogravimetric analysis (TGA) Thermogravimetric analysis (TGA) is an analytical technique used to détermine either the material's thermal stability or the change of sample mass with change of température. This loss of mass is due to the amount of volatile components or solvents that are released from the sample when the spécimen is heated up (Hill, 2005). TGA analysis relies on a high degree of précision in three différent measurements: weight, température and température change. In gênerai, the analyzer consists of a high-precision balance with a pan, generally made of platinum, where the sample is placed. This pan is placed inside the electrically heated oven with a thermocouple to accurately measure the température. The measurement is normally carried out in air or in an inert atmosphère, such as Hélium or Argon, to prevent oxidation or other undesirable reactions and the weight is recorded as a function of increasing température. In addition to weight changes, some instruments record the température différence between the spécimen and one or more référence pans (differential thermal analysis, or DTA) or the heat flow into the spécimen pan compared to that of the référence pan (differential scanning calorimetry, or DSC). The latter can be used  59  4. Expérimental Methodology to monitor the energy released or absorbed via chemical reactions during the heating process (Hill, 2005). A TA Instruments - Q600 thermo gravimétrie analyzer, located at the Department of Chemical and Biological Engineering (Figure 4.5), with N 2 flow of 100 cm3 min"1 and température range of 20°C to 600°C was used. Heating rates of 12°C min"1 up to 80°C (an isothermal study was done for 1200 min), 9.6°C min"1 up to 473°C and 4.4°C min"1 to 600°C (an isothermal study was done for 180 min) were utilized. Right after the production process, a single composite TÏO2 sphère was placed in the pan for 23 hours. The change in the weight of the Ti0 2 with System température was used to détermine the adsorbed moisture and solvent content of the catalyst.  Figure 4.5: Thermo gravimétrie analyzer (TGA) - TA Instruments  4.4 Fluidized Bed Photoreactor Setup The fluidized bed photoreactor was built using an 80 ml_ quartz tube (15 cm height and 2.6 cm of internai diameter) as is shown in Figure 4.6. Two acrylic pièces, held together with three threaded rods, were part of the main body of the reactor. The conic shape in the bottom pièce (calming section) helps develop the 60  4. Expérimental Methodology fluid to hâve a better catalyst expansion. The catalyst inside the quartz tube was irradiated by three UV-Hg lamps (5.7 W at 254 nm output, GPH357T5L/4P, Light Sources Inc.) that were longitudinally placed surrounding the reactor. The position of the lamps could be changed in order to adjust the level of irradiation. A cylindrical reflector was also included in some experiments surrounding the lamps to improve the radiation to the reactor.  Water outlet (D = 1.20 cm)  S3  .  ifuy' T u * Mesh  Quartz tube (ID = 2.60 cm)  E o Threaded rod (D = 0.80 cm)  r— CM  Mesh  O-rings  i irn  Water inlet (D = 1.20 cm)  7.62 cm  Figure 4.6: Schematic diagram of the quartz tube photocatalytic reactor.  61  4. Expérimental Methodology  4.5 Photocatalyst activity test In order to measure the activity of the composite TÏO2 photocatalyst sphères, two différent model organic pollutants were put in contact with the sphères in the présence of UV lamps. In both cases, the reaction proceeded at 27°C and at ambient pressure and the changes in concentration were monitored using High Performance Liquid Chromatography (HPLC). The list of Chemicals use for thèse experiments is provided in Table 4.2.  Table 4.2: Spécifications of chemicals used during the activity test of the CSG photocatalyst Chemical Chemicals Purity Supplier 3 Rôle formula Formic acid  HCOOH  99%  AO  Model organic contaminant  Acetonitrile  CH3CN  HPLC - 99%  S-A  HPLC eluent  Methanol  CH.OH  HPLC- 99%  FS  HPLC eluent  99.7%  FS  HPLC eluent  Glacial Acetic Acid  CH.COOH  2,4-Dichlorophenoacetic acid  CSH6C1203  98%  S-A  Model micropollutant  2,4-Dichlorophenol  C6H4C120  99%  S-A  2,4-D by-product  4-Chlorophenol  C6H5C10  99%  S-A  2,4-D by-product  2-Chloro-1,4benzoquinone  C6H3C102  95%  S-A  2,4-D by-product  4-Chlorocatechol  C6H5C102  97%  , S-A  2,4-D by-product  Chlorohydroquinone  C6H5C102  85%  S-A  2,4-D by-product  AO: Acros Organics; S-A: Sigma Aldrich; FS: Fisher Scientific  In addition to the experiments with FA and 2,4-D, preliminary tests were conducted with raw drinking water containing natural organic matter (NOM) to détermine the ability of the composite sphères to dégrade NOM. Total organic carbon analyzer (Shimatzu, TOC - VCPH) were used to monitored the change in  62  4. Expérimental Methodology concentration with respect to time. Same as above, thèse experiments were run at 27°C and ambient pressure. The optimum amount of catalyst to be loaded on the photocatalytic reactor was experimentally determined monitoring the apparent rate constant of FA with respect to time for a given bed expansion. The optimum mass of catalyst was chosen based on the highest dégradation rate constant. The photocatalytic reactor was placed inside a wooded box to avoid any direct contact with UV light and to be able to change the level of irradiation applied to the reactor by changing the position of the UV lamps (Figure 4.7). The photoreactor was fed from a storage tank (Figure 4.8) that was equipped with a porous diffuser, sparging air through the solution in order to maintain relatively constant dissolved oxygen needed for the reaction. The recycling flow rate was relatively high (3.4 L min"1), and the bed expansion of the Ti0 2 photocatalyst was approximately 500%.  Figure 4.7: Expérimental setup: (1) wooden box, (2) UV lamps, (3) fluidized bed photocatalytic reactor (FBPR), (4) différent lamp positions.  63  4. Expérimental Methodology  (1)  (4)  (3)  D  Figure 4.8: Expérimental setup: (1) fluidized bed photocatalytic reactor (FBPR), (2) tank, (3) pump, (4) Flowmeter.  As mentioned before, in ail of the experiments carried out in this investigation, the High Performance Liquid Chromatography (HPLC) and the Total Organic Carbon analyzer (TOC) were used. Therefore, a brief explanation of thèse two equipments is necessary to hâve a brief background about them.  4.5.1 High Performance Liquid Chromatography analysis (HPLC) HPLC is a widely used analysis to separate, identify and quantify corhpounds because of the multiple advantages such as high précision and low volume of samples. In this investigation, a Water HPLC Instrument (WATERS 2695), equipped with a Symmetry C-18 column (4-um particle diameter) was used for the purpose of analyzing FA and 2,4-D. This equipment has degasser chamber to degas the mobile phase in the System, reducing the amount of dissolved air from the mobile phase and preventing air bubbles and conséquent drift or other baseline irregularities. The mobile phase changed depending on the organic pollutant that was analyzed. Other  64  4. Expérimental Methodology variables such as column température, flow rate, among others change also depending of the nature of the pollutant as is going to be explained later.  4.5.2 Total Organic Carbon Analysis (TOC) TOC analysis has been recognized as non-specific analytical technique to measure the water quality during the water purification process, because it represents the amount of carbon bound in an organic compound. Typically, when the TOC is measured, the total carbon and the inorganic carbon are measured at the same time. Then, subtracting the inorganic carbon from the total carbon yields the TOC. For this particular work, samples of 25 ml_ were analyzed in a Shimadzu TOC-VCPH  combustion  oxidation TOC  analyzer.  This  equipment  employs  combustion oxidation at 680°C to improve the détection regardiess the type of organic matter. Calibration curve of two points using Milli-Q water and a standard solution of Benzoic Acid was used.  4.5.3 Organic pollutants dégradation Two différent organic pollutants were used to détermine the photoactivity of the composite Ti02 sphères. Formic Acid (FA) and 2,4-Dichlorophenoacetic acid (2,4-D) were selected. As a gênerai procédure, catalyst washing was needed to eliminate any small particles of Ti02 adhered to the surface of the catalyst sphères and to remove any contamination on the surface of the catalyst. At the same time washing helped clean the system itself. One L of milli-Q water was used to wash the catalyst for 60 min. This water was changed in the first 30 min.  4.5.3.1 Formic acid (FA) dégradation Formic acid was selected as a model pollutant because it is oxidized directly to CO2 without the formation of any stable intermediate products and because it has 65  4. Expérimental Methodology been used extensively in other photocatalytic studies (Ha and Anderson, 1996; Kim and Anderson, 1996; Muggli and Backes, 2002; McMurray et al., 2004; Krysa, 2006). After the washing process, 1 L of water was re-circulated through the reactor for 15 min, with the UV lamps. This ensured that the UV lamps were warm and reached their maximum output prior to the start of the experiment. After this time, a concentrated FA solution (10 g L"1) was added to the water, the amount of this solution determined the initial FA concentration in the System. The concentrations of FA were quantified using the HPLC with Acetonitrile / water (40:60% v/v) as the mobile phase. The flow rate of mobile phase for the analysis was kept at 1ml_ min"1 using A=214 nm for UV détention. The température of the column was kept at 25°C during the détection. Reagent grade standard was used to calibrate the HPLC for the model organic compound.  4.5.3.2 Micro-pollutant dégradation (2,4 Dichlorophenoxyacetic acid) 2,4-Dichlorophenoxyacetic acid (2,4-D) is the third most commonly used herbicide in the United States and Canada. Worldwide, 2,4-D and différent salts derivatives and ester forms are used in a variety of places including home lawns, cereal and grain crops, commercial areas, commercial turf, rights-of-way, and forest (Trillas et al., 1995). This widespread use of 2,4-D leads to certain environmental impact due to the fact that during its application it could be easily spread within the environment. Once on the ground, it can be incorporated in the natural aqueous stream showing différent half-lifes, depending on factors such as oxygen concentration, acidity, solar light, among others (EPA, 2008). Therefore, evaluating the photocatalytic activity of composite Ti0 2 sphères using 2,4-D as a micro-pollutant is important for real large scale applications. Prior to thèse experiments, an adsorption study was done to détermine the amount of 2,4-D that is being adsorbed in the Ti0 2 sphères. After the initial washing of the photocatalyst (as discussed above), 500 ml_ of 2,4-D solution, at différent initial concentrations (1 - 10 mg L"1), was re-circulated through the reactor with the UV Lamps off. This process was allowed for approximately 1 h, ensuring that the adsorption of the 2,4-D on the catalyst reached 66  4. Expérimental Methodoiogy equilibrium. Immediately after that, the UV lamps were turned on for about 1 h. Samples were collected at différent intervais and the concentrations of 2,4-D and its photodegradation by-products were quantified using HPLC with methanol / water / glacial acetic acid (58:40:2% v/v) as a mobile phase. The flow rate of mobile phase for analysis was kept at 1.23 mL min"1 using A=280 nm for UV détention. The température of the column was kept at 35°C during the détection. Reagent grade standards were used to calibrate the HPLC for the parent contaminant and its oxidation by-products.  4.5.4 Natural Organic Matter (NOM) dégradation Natural organic matter (NOM) is defined as a complex mixture of organic compounds présent in surface water, with a composition that changes with respect to season and location (Bursill et al., 2002). Therefore, testing Ti02 sphères with NOM will give us a good understanding about the photocatalytic effect of CSG titania in complex mixture of compounds. Water from Trepanier Creek in the Peachland area in central British Columbia (with the initial TOC of approximately 5 mg L"1) was treated in the fluidized bed reactor using 25 g of titania photocatalyst sphères. The photodegradation study was performed running 1 L of raw water through the FBPR for 1 h. Prior to thèse experiments, an adsorption study was done to détermine the amount of NOM that is being adsorbed in the Ti02 sphères At each intervais, 25 mL of water sample was collected and analyzed by the TOC  4.6 Attrition résistance détermination Photocatalyst attrition was determined by measuring and monitoring the amount Ti0 2 in the solution. A spectrophotometric technique was adopted and used for this purpose (Jackson, 1883; Jackson et al., 1991). The list of chemicals required to perform this analysis is provided in Table 4.3.  67  4. Expérimental Methodology Table 4.3: Spécifications of Chemicals used for the attrition résistant of the composite TÏO2 sphères. Chemical formula Purity Chemicals Supplier {NHA\SOA 100% Ammonium Sulfate Fisher Scientific S u Ifu rie acid  H2S04  98%  Fisher Scientific  Hydrogen Peroxide  H202  30%  Fisher Scientific  In this technique, 10 mL of the solution after the photocatalytic treatment (containing Ti02 in suspension) and 10 mL of solution of ammonium sulfate, (NH4)2S04 in sulfuric acid, H2SO4, (3 M) was heated up to approximately 100°C in a porcelain dish for approximately 10 min to produce an homogeneous solution. This solution was then transferred to a 25 mL volumetric flask which was filled to exactiy 25 mL with 6 mL of H2O and with a H2SO4 solution in water (1.1 M). Three drops of hydrogen peroxide, H2O2, were added to the sample to produce a yellow-orange colored complex ions (Ti02-H202)  (Eisenberg, 1943; Chariot, 1964; Marczenko,  1986): Ti4+ + H202 + 2H20 « Ti02 • H202 + 4H+  /4>gv  The amount of Ti 4+ présent in the solution was determined using a spectrophotometer (Shimadzu 1240) at A = 410 nm.  4.7 Photocatalyst deactivation and attrition évolution For the FA, 2,4-D and NOM, some tests were performed to détermine and quantify photocatalyst deactivation. In each case, the same photocatalyst samples were used for several consécutive runs to détermine if there was a change in the activity of the catalyst with respect 0 the dégradation of each contaminant. A change in the photocatalytic dégradation for any of those compounds was considered as and indicator for photocatalyst deactivation. 68  4. Expérimental Methodology  4.8 Fluence rate détermination As discussed before, a photocatalyst is activated by light that is provided to generate the électron and holes (e" - h+) pairs. Therefore, identifing the amount of light or irradiation supplied to the reactor is required to understand thoroughly the phenomena behind the photocatalysis process. The fluence rate is defined as the energy provided by the UV lamps per unit time per illuminated area of the reactor (J s"1 m"2). For the purpose of this research, the irradiation flux to the photoreactor was measured using two différent methods: potassium ferrioxalate and iodite-iodate actinometry techniques. The Chemicals required for carrying out thèse measurements, with both actinometric solutions, are listed in Table 4.4. In the case of potassium ferrioxalate, the photochemical reaction that describes the chemical change in the actinometer is given by Equation 4.6. As can be seen, the Fe3+ is converted to Fe2+ in the présence of UV irradiation (Zalazar et al., 2005):  2Fe3++C20/  hv  >2Fe 2+ +2Ca  (4.6)  Table 4.4: Spécifications of chemicals used for the fluence rate détermination. Chemicals Chemical formula Purity Supplier Sodium Acétate anhydrous  C2H302Na  S u Ifu rie acid  H2SOA  Iran (III) Sulfate pentahydrate Potassium oxalate monohydrate 1,10 Phenantroline Potassium lodate Potassium lodide  —  Sigma Aldrich  98%  Fisher Scientific  Fe2OuS,-5H20  97%  Acros Organics  K2C204H20  —  Sigma Aldrich  Cjj/ZgiVj  99%  Acros Organics  KI03  100%  Kl  99%  Fisher Scientific Fisher Scientific  69  4. Expérimental Methodology The potassium ferrioxalate method involved a solution of (Fe2(S04)3) (98.02 g Fe2012S3-5H20 + 55 mL of H 2 S0 4 made up to 1 L) that was prepared from the hydrate compound. Then, the actinometry solution was prepared mixing 50 mL of iron sulfate solution with 50 mL of potassium oxalate (K2C2C>4) (1.2 M) solution. This new solution was diluted with water up to 1 L. Prior to each experiment, UV lamps were warmed up for 15 min to reach stable radiation. Then, the actinometry solution was pumped through the System shown in Figure 4.8. Différent lamp positions were used to measure the irradiation reaching the reactor. In every case, 2 mL of sample was taken. Each sample was mixed with 8 mL of phenantroline (0.2 % w/w), 1 mL of buffer solution (49.43 g C2H302Na + 10 mL of H2SO4 made up to 1 L) and 9 mL of water. This final solution was left under dark conditions for 30 min. Finally, the amount of ferrous ions produced was measured via spectrophotometric measurement of the concentration of complex formed with 1,10 phenantroline ([(C-i2H8N2)3Fe]2+) at À = 510 nm. The fluence rate was proportional to the absorbance measured at 510 nm (Parker, 1953) and was calculated as follows:  e-<l>-t-Vsamp-Area  where HPF is in mW cm"2, Vin-ad is the irradiated volume in mL, Vdli is the volume of the dilution in mL, e is the molar extinction coefficient (1.11 x 104 L gmol"1 cm"1), ^is the quantum yield (-1.48 according with Golsdtein and Rabani (2008)), t is the irradiation time in min and Vsamp is the volume of the sample taken for analysis in mL. The Area is the irradiated area of the reactor (127 cm2) In the case of the lodite-lodate actinometry, the reaction that describes the process is the following (Rahn et al., 2003):  SKI + KIOz + 3H20 + hv-> 3/ 3 " + 60H~ + 9K+  (4.8)  70  4. Expérimental Methodology According  to  this  reaction,  the  triiodide  could  be  determined  spectrophotometrically at A = 450 nm. A 0.01 M solution of sodium borate buffer was required to prépare 0.1 M solution of potassium iodate in the borate buffer. Finally, 0.6 M of potassium iodide was needed to be prepared in borate-iodate solution. Additional détails that was followed could be obtained elsewhere (Rahn, 1997; Rahn et al., 2003; Rahn et al., 2006). Finally, the quantum yield was determined using Equation 4.8: </> = 0.75[l + 0.02(7" - 20.7)5 + ° - 2 3 ( c " 0.577)]  (4.9)  where the quantum yield (<$ is given in mol einstein"1 and T is the température of the solution in °C and C is the molar concentration of the iodide solution. Then, the fluence rate was calculated as follows:  4.72ri0'-^,-y-1000 £•450 • <j) • Area  •t  where Hu is the fluence in mW cm"2, AA450 is the change in the absorbance at X = 450 nm, V is the solution volume in L, E450 is the molar absorption coefficient of triodide at X = 450 nm (M"1 cm"1), ^ is quantum yield calculated with Equation 4.9, Area is the irradiated area (127 cm2) and t is the time of sampling in s.  71  Chapter Five Preliminary Evaluation of Composite TïCh Photocatalyst 5.1 Introduction The photocatalytic activity of a photocatalyst is one of the most important parameters in photocatalyst development. High activity is always désirable because it translates into better photodegradation of pollutants in contaminated water. Another important parameter to evaluate, in the case of FBPR, is the mechanical strength of the particles because it will ensure good durability and stability of the catalyst. Therefore, this section describes the preliminary évaluation of the photocatalyst activity, using Formic acid (FA) as a model organic compound, and the attrition résistance of the CSG TÏO2 sphères as a référence and benchmark for additional improvement. Preliminary characterization of the original composite photocatalyst was performed using BET method to measure the surface area of the catalyst, X-ray diffraction (XRD) to identify the polymorphic phases présent in the TiC>2 sphères, and scanning électron microscopy (SEM) to identify the microstructural characteristics of the catalyst.  5.2 Composite Ti0 2 photocatalyst activity In order to evaluate the activity of the catalyst, photocatalytic dégradation of Formic Acid (FA) was investigated. FA was selected as a model pollutant because it is oxidized directly to CO2 and H 2 0 without the formation of any stable intermediate products (McMurray et al., 2004). As was described in Section 4.5, HPLC was used to analyze the change in FA concentration with respect to the reaction time.  72  5. Preliminary Evaluation of Composite TIQ2 Photocatalyst As part of this preliminary évaluation, sphères calcined at différent températures were used, mainly because it is well known that the calcination température plays a critical rôle defining the crystalline phases, the surface area and most importantly, the activity of the photocatalyst (So et al. 1997; Pozzo et al., 2000; Yin et al., 2001; Su et al., 2004; Qiu and Zheng, 2007). In gênerai, it is expected that the activity of the photocatalyst will decrease with an incrément in the calcination température because the surface area will decrease, and the anatase will be converted to rutile. Therefore, calcination températures between 600°C and 1300°C were studied to ensure that a wide range of températures were covered and its impact was studied thoroughly. It is also known that physical characteristics of the sphères, such as size and density, changed with the calcination température. Therefore, to hâve a constant height of the expanded bed in the FBPR (to hâve a similar light distribution in the System), certain parameters such as flow rate and mass of catalyst should be adjusted for sphères produced under différent températures. The height of the bed is an important characteristic of the FBPR that needs to be maintained constant from one test to another, in order to allow comparison between the experiments. Inconsistent bed's heights may lead to différent photon distributions, flow behaviour, and mass transfer within the reactor. Hence, prior to any photocatalytic experiments a thorough assessment of the bed expansion was performed as discussed in the following section.  5.2.1 Bed expansion estimation The bed expansion of the composite TiC^ sphères within the reactor was calculated based on the terminal velocity (i/t) and bed voidage (E) of the Ti0 2 sphères.  In this approach, the fluidization velocity (U) can be calculated with  Equation 5.1 where Q is the volumetric flow rate and Rmactor is the radius of the quartztube(0.012m).  73  5. Preliminary Evaluation of Composite T1O2 Photocatalyst (5.1)  U = ^V^reactor)  The Vt is calculated using a corrélation proposed by Haider and Levenspield (1989) (Equation 5.2) were d* is the dimensionless particle diameter, pH2o and psph are the density of the water (1000 kg m"3) and the sphères respectively; 1JH20 is the viscosity of the water (0.001 kg m"1 s"1), rsph is the radius of the sphères, and g is the gravitational force.  18 0.5909 —r + —==-  PH2O  v, = sph  ~PH7 0 )MH}O  d*=2r. sph\  g\Psph  ,  -PH2O)PH2O  (5.2)  (5.3)  t*H20  The terminal velocity of isolated single particles, ve, is calculated by Equation 5.4 (Khan and Richardson, 1987). The bed voidage, s, is predicted using Equation 5.5 which is the corrélation given by Richardson and Zaki (1954). The exponent e is known as the expansion index and can be calculated by Equation 5.6.  ,0.6  (  -s- = l - 1 . 1 5  'sph  V  U  4.8 -e e-2.4  (5.4)  p reactor J  = £  (5.5)  0.043(^r) 0 5 7  (5.6)  where Ar is known as the Archimedes number, defined by  74  5. Preliminary Evaluation of Composite Ti02 Photocatalyst 3\  8  Ar  8(PsPh-PH1o)pH2orsl'sph 2  (5.7)  MH2O  With the bed voidage for the catalyst, it is possible to predict the expansion of the bed in the reactor using Equation 5.8. The value of e0 is calculated based on the physical characteristics of the catalyst and it is estimated to be equal to 0.39 for the sphères calcined at 600°C and 0.37 for the sphères calcined at 900°C (Benenati and Brosilow, 1962; Mueller, 1997):  1-f  The height of the static bed in the reactor, Hbed.o, and the bed height (Hbed) corresponding to a given bed expansion can be evaluated applying Equations 5.9 and 5.10 (Imoberdorf et al., 2008):  n  Psph V  S  0 /^reactor  Hbed=HbedfiB^  (5.10)  As shown in Equation 5.9, the mass of catalyst is an important parameter for defining the expansion of the bed. In addition, the amount of catalyst influences the efficiency of the System to dégrade contaminants, mainly because the light distribution in the reactor might change with the amount of catalyst loaded in the reactor. Therefore, the optimum mass of catalyst that should be loaded in the reactor (for a given catalyst characteristics such as size and density) needs to be determined. Sphères at 600°C were selected as a baseline because it was expected that they would show higher activity. Table 5.1 shows différent masses of catalyst calcined at 600°C and the flow rates required to obtain a fairly constant height of the 75  5. Prelirninary Evaluation of Composite TiC>2 Photocatalyst expanded bed (the bed was expanded up to approximately 93% of the 15 cm height of the quartz tube). As can be seen, the greater the amount of catalyst, the lower the volumetric flow rate for a constant height of the bed.  Table 5.1: Data to calculate the height of the expanded bed in the fluidized bed photocatalytic reactor (FBPR) of sphères at 600°C (rsph = 5.75 x 10"4 m; psph = 3109 kg m"3) Mass of Catalyst loaded in the FBPR (kg) Volumetric flow rate (ma s"1) 0.015 6.5x10- b 0.025 5.7 x10' 5 0.035 4.8 x10" 5 0.040 4.5 x10' 5  Figure 5.1 shows the dependence of the FA dégradation on the photocatalyst loadings in the reactor using sphères calcined at 600°C. The voidage of the bed for ail the cases is also specified. At catalyst loading of above 25 g, the reaction rate constant was approximately constant for FA dégradation considering a relatively constant height of the expanded bed (H^d œ 14cm). In addition, the change in the rate constant between 25 g and 35 or 40 g was almost negligible, but the différence in the amount of sphères loaded in the reactor was very significant. Therefore, 25 g was used as the loading used for ail the experiments in this study, giving a bed expansion of 500% with respect to the height of the static bed (~2.8 cm) It is important to mention that increasing the amount of catalyst in the reactor, the light pénétration might change within the reactor. Therefore, the light distribution might be différent for différent catalyst loadings. Even though this could affect the FA dégradation, this effect was not considered because this analysis was done as a prelirninary test of the System with composite TiÛ2 sphères.  76  5. Prelirninary Evaluation of Composite Ti02 Photocatalyst 0.035 -  c (A C  o u  8 = 0.83  0.030 # e = 0.88  0.025  8 = 0.80  0)  *•>  2 _ §:= co Si *•• c  2 m  Q.  a <  9 e = 0.93  0.020 0.015 0.010 0.005 0.000 1 0  10  20  30  40  Catalyst loadings (g)  Figure 5.1: Apparent reaction rate constant of Formic acid (FA) versus catalyst loadings in the FBPR using sphères calcined at 600°C (s = voidage of the bed)  5.2.2 Photocatalyst activîty at différent calcination températures To compare the activity of sphères calcined at différent températures, the same mass of sphères determined with sphères at 600°C (25 g) was used for the other sphères (calcined at différent températures) and the flow rate was adjusted as is detailed in Table 5.2 to hâve a similar height of the expanded bed (Hbed7* 14cm) in ail cases. This criterion was followed just for the prelirninary comparative studies, even though it was recognized that changing either the mass of the catalyst or the flow rate in the system would change the dynamic of the System. As expected, sphères at 600°C showed the highest activity to dégrade FA. Figure 5.2 shows the effect of the calcination température on the FA dégradation using the Ti0 2 sphères. First order dégradation reaction was determined for the cases at 600°C, 700°C and 800°C. As is shown, there is a strong relation between the heat treatment and the activity of the catalyst. For the catalyst calcined at 900°C or higher températures, there was no or little FA dégradation after three hours of opération. On the other hand, for the catalyst calcined at 600°C the FA dégradation was completed after 3 h. Table 5.3 shows the rate constants for the experiments 77  5. Preliminary Evaluation of Composite "TIO2 Photocatalyst presented in Figure 5.2. To calculate this rate constant, the apparent rate constant was normalized multiplying it by the ratio of the total volume (1000 mL) to the active volume (80 mL) according with the approach followed by Turchi and Wolfrum (1992). The rate constant was the highest at 0.315 min"1 for the catalyst calcined at 600°C, but it decreased with increasing température. At 800°C, for example, the rate constant decreased by 92%. Thèse results can be explained from the point of view that higher calcination températures induce higher amount of rutile in the catalyst and lower surface area.  Table 5.2: Data to détermine the activity of sphères at différent calcination températures {mSPh = 0.025 kg) Value at différent températures Parameter Nomenclature 600°C 700°C 800°C 900°C Volumetric flow rate 5.7 x10"5 7.3x10* 7.8x10* 8.0x10-5 Q (m3 s"1) Density of sphères 4527 3109 4020 4399 Psph (kg m"3) Radius of sphères -4 5.75x10"4 5.59x1 (T 5.25x10"4 4.65x10 l"sph (m)  1.00 I  <> Ê3  0.80 -  <>  ï  r3 []  * O  <  0.60 -  ()  •*-  0 <> E]  1  1  i  tJ  r1  —m*.  O 0.40 -  ïl ()  {i  0.20  (  n nc\ u.uu 0  30  60  f  90  A  <D 120  0 150  180  Reaction time (min) Figure 5.2: Formic acid dégradation versus reaction time with sphères calcined at: (0) 900°C, (•) 800°C, (A) 700°C, (o) 600°C; error bars represent 95% Cl of the triplicate runs) 78  5. Preliminary Evaluation of Composite TIO2 Photocatalyst The calcination température had an impact on the sphères ability to dégrade organic compound (in this case FA) because it defined characteristics such as the surface area, pore size, pore volume and the amount of anatase in the catalyst (Section 5.4).  Table 5.3: FA photodegradation rate constants (kr) of composite TiC>2 sphères produced at différent calcination températures (mass of catalyst loaded to the reactor equal to 0.025 kg) Calcination température Dégradation rate constant Standard error of 1 (min" ) the parameter (°C) 0.315 600 0.061 0.164 700 0.008 0.025 800 0.013 N/A 900 N/A  5.2.3 Mass transfer résistance calculations In a heterogeneous reaction séquence, mass transfer of reactants first takes place from the bulk to the extemal surface of the catalyst. Occurrence and completion of the photocatalytic reaction is then followed by the mass transfer of the products from the catalyst surface to the solution. Hence, mass transfers of the reactants and products significantly affect and sometime hinder the overall reaction rates. In fluidized bed reactors, it is well known that the mass transfer résistance is negligible. Nonetheless, calculations were done to confirm that the reaction rate was not limited by mass transfer. In that regards, the catalyst-liquid mass transfer coefficient was calculated using mass transfer corrélations proposed in the literature (Tournié et al., 1979; Arters and Fan, 1984; Prakash et al., 1987) and the Sherwood number (Sh):  Sh = ^ * D  (5.11)  79  5. Preliminary Evaluation of Composite T1Q2 Photocatalyst  Sh = O.2280h35GaO323MvO3ScOA  (5.12)  where Ga is the Galileo number, Mv is the density number, Se is the Schmidt number, <j>s is the sphericity (in this case considered equal to 1 ),  I<MT  is the mass  transfer coefficient and DFA-W is the diffusive coefficient for FA in water. Thèse three dimensionless numbers can be calculated by:  Ga = d-fslfMl  MV  =  PSP»-PH2O  (5.13)  (  5  M  )  PH2O  Sc=  *"*  (5.15)  PH2O^FA-W  Using some of the data listed in Table 5.2 for the sphères at 600°C and the diffusivity of FA in water equal to 1.45 x 10"9 m2 s"1 (Perry and Green, 1997), Equations 5.13 to 5.15 could be solved. Thus, the mass transfer coefficient was calculated, using Equation 5.11, to 1.09 xlO"4 m s"1. To compare the mass transfer coefficient with the reaction rate constant (kr) at 600°C (0.315 min'1), Equation 5.16 was applied:  *,'=A  (5.16)  a Where kr' is the reaction rate constant expressed in m s"1 and a is the external surface to volume ratio of the catalyst particle and it can be calculated by:  80  5. Preliminary Evaluation of Composite "TIQ2 Photocatalyst  a =  _*i_ 4 3 —Ttr u sph  3  =  ^_  (5.17)  Sr Hu  P  Knowing a, the value of kr' was calculated to be 1.01 x 10"6 m s"1, which can be compared with the mass transfer coefficient using a criterion similar to that developed by Mears (1971). In this approach, the reaction rate constant needs to be:  M^V<oi5 k  (518)  r  "•MT^FA  where /' is the integer exponent in power law rate expression. In the case of this work / was equal to 1 because the dégradation followed first order kinetics. Simplifying Equation 5.18, gives:  h  " <0.05  h-  (5.19)  H-MT  According to Equation 5.19, the reaction rate constant should be smallerthan 5% of the mass transfer coefficient in order to neglect mass transfer résistance. With the values of /c/and km being 1.01 x 10"6 m s"1 and 1.09 x10"4 m s"1, respectively, it is concluded that mass transfer résistance was negligible in the System used in this investigation.  5.3 Attrition détermination The motion of particles in a fluidized bed reactor usually causes catalyst attrition due to the collision between particles and bed-to-wall impacts (Vaisman et al., 2005; Qiu and Zheng, 2007). One of the main conséquences of attrition in any System, and particularly in water treatment processes, is the loss of catalyst and génération of fines that are difficult to separate downstream.  81  5. Preliminary Evaluation of Composite TiQ2 Photocatalyst Figure 5.3 shows the amount of TÏO2 detached into water after 3 h of opération for catalyst produced at différent calcination températures. As expected, the résistance of the catalyst improves when the calcination température increases because the structure of the sphères becomes harder and more compact. At 600°C, the amount of TÏO2 in suspension after 3 h of continuous opération was approximately 22 mg L"1; while at 800°C this value dropped to nearly 8 mg L"1.  n oo^  5" "c  O W  T 0.020 -  1 600°C  0)  a j»  0.015 -  (A  700°C  c (M  Çj 1-  0.010 -  "5 C  800°C  •4->  g E <  T  T JL  1 900°C  0.005 -  n nnn u.uuu  Calcination température (°C)  Figure 5.3: Attrition résistance for TÏO2 sphères calcined at différent températures (Error bars represent 95% Cl of triplicate runs)  5.4 Composite photocatalyst characterization Microstructure of the produced TiÛ2 sphères was analyzed by using SEM. Images of the composite TÏO2 sphères calcined at 600°C and 900°C are shown in Figure 5.4. As it can be seen, there are no significant différences between thèse two batches. The sphères at 700°C and 800°C showed similar characteristics as well. In ail the cases, the Ti0 2 sphères had a relatively smooth surface despite the présence of some roughness along the surface of the catalyst. No cracks were found 82  5. Preliminary Evaluation of Composite TiQ2 Photocatalyst after heat treatment confirming that there were not extrême stresses during the catalyst production process. Grooves around the equator of the sphère as well as the crater shown in Figure 5.4a, 5.4c & 5.4d are the resuit of the production process. Sphères calcined at 600°C had an average diameter of 1.17 mm. At higher calcination température, e.g. 900°C, the average diameter was reduced to 0.930 mm. In either case, the catalyst particles showed consistency in size and shape.  m Figure 5.4: SEM micrograph of a photocatalytic composite sphère, calcined at 600°C (a and b) and calcined at 900°C (c and d)  As briefly discussed before, the calcination température played a significant rôle in the activity of the catalyst. It defined many of the photocatalyst characteristics, e.g., the surface area, pore volume, pore size, percentage of anatase, among others. From Table 5.4 it is clear that the higher the température, the smaller the surface area and the pore volume, but the higher the pore size. The percentage of anatase in the catalyst also changed with calcination température; the amount of 83  5. Preliminary Evaluation of Composite "TIQ2 Photocatalyst anatase was reduced significantly when the calcination température increased from 600°C to 700°C (from approximately 72% to 1.9%). At higher températures there was no anatase in the catalyst.  Table 5.4: Surface area and anatase composition (using BET and XRD respectively) of the composite TiC>2 sphères calcined at différent températures. Calcination Pore size Pore volume Surface area 2 1 Anatase (%) (cm3 g-1) (nm) (m g" ) température (°C) 600 0.2577 72.23 36.93 22.03 700 1.86 6.58 30.43 0.0995 0.0425 800 0 3.79 42.73 N/A N/A 0 N/A 900  Figure 5.5 shows the X-ray diffraction patterns of TiÛ2 sphères, at différent calcination températures. When the sphères were calcined at 600°C, there was a broad peak at 20 = 30°, which is identified as the most intensive peak ( 1 0 1) for anatase. Another characteristic peak was found at 20 = 32° which represents the rutile (1 1 0) phase. The peak intensity of anatase ( 1 0 1) decreased dramatically and the peak intensity of rutile (1 1 0) increased significantly when the calcination température increased from 600°C to 700°C. This demonstrates that increasing the température  of  heat  treatment  accelerated  phase  transformation  from  thermodynamically metastable anatase to more stable and more condense rutile phase (So et al. 1997). BET was used for the surface area, pore size distribution and pore volume analyses of the composite Ti0 2 sphères at différent calcination températures. As shown in Table 5.4, a strong dependence was observed for the surface area, pore size and pore volume to the calcination température of the composite Ti0 2 sphères. The spécifie surface area and pore volume shifted towards smaller values at higher calcination températures, but the average pore diameters increased upon heat treatment, which is consistent with the literature (Bischoff and Anderson, 1995; Xu et al., 2008). The réduction in surface area with the calcination température was  84  5. Preliminary Evaluation of Composite TiQ2 Photocatalyst significant. At 700°C the spécifie surface area decreased to 6.58 m2 g"1 and the pore volume reduced almost by 60% with respect to the value reported when the catalyst was calcined at 600°C (36.93 m2 g"1). Although a large portion of the mesoporous structure was maintained at 700°C (Figure 5.7), the pore volume was reduced significantly, likely because higher calcination températures induce crystallization of Ti0 2 , followed by crystal growth, and collapse of the pores (Peng et al., 2005). The isotherms and pore size distributions of the composite TÏO2 sphères, calcined at différent températures, are shown in Figures 5.6 and 5.7. It is clear that sphères calcined at 600°C hâve higher surface area than those calcined at higher températures. According to the BDDT (Brunauer, Deming, Deming and Teller) classification, the sphères showed isotherm type IV, exhibiting H2 hystérésis loops. This resuit indicates that the composite photocatalyst porosity is in the meso - range (2~50nm) (Sink et al., 1985).  Figure 5.5: X-ray diffraction patterns of composite Ti0 2 sphères at différent températures.  85  5. Preliminary Evaluation of Composite Ti02 Photocatalyst  Figure 5.6: Surface area for samples calcined at différent conditions, (0) 600°C, (o) 700°C, (n) 800°C.  0.120 0.100 U)  E  0.080  E  0.060  o o  3 O  >  0.040  o Q.  0.020 -Q  0.000 20  40  60  80  100  120  Pore diameter (nm) Figure 5.7: Pore size distribution for samples calcined at différent températures, (0) 600°C, (o) 700°C, (D) 800°C  86  5. Preliminary Evaluation of Composite TiQ2 Photocatalyst  5.5 Final remarks Based on the results presented in this section, 25 g of TÏO2 sphères calcined at 600°C was selected as a practical loading for the fluidized bed reactor experiments carried out in the rest of the research. This calcination température was selected for further study because the activity of the catalyst prepared at this température was superior. At 600°C, the percentage of anatase (72%) and the surface area (-37 m2 g) of the Ti0 2 sphères were higher than those of the other samples, which can explain the différence in activity. High percentages of anatase and high surface area hâve been reported as désirable characteristics that lead to more efficient photocatalyst with  greater  dégradation  of  contaminants  (Berry  and  Mueller,  1994;  Thiruvenkatachari et al., 2008). In terms of attrition résistance of the catalyst, it has been demonstrated that the higher the calcination température, the higher the résistance of the catalyst. However, when the température increases, the activity drops significantly. At 600°C, the attrition in the System was determined to be 22 mg L"1 of Ti0 2 particles in suspension while at 800°C this value dropped to nearly 8 mg L"1. However, the dégradation rate constant dropped from 0.315 ± 0.061 min"1 at 600°C to 0.025 ± 0.013 min"1 at 800°C. Hence, it was decided to use 600°C as the appropriate calcination température, but to perform further optimization on différent variables that lead to improve attrition résistance.  87  Chapter Six Photocatalyst Improvement and Characterization 6.1 Introduction Attrition of the catalyst particles within a photoreactor is a key challenge requiring attention. Catalyst attrition in a System will induce low photoefficiency of the System and process complications such as downstream filtration and séparation of the suspended particles, with the subséquent incrément in cost. This chapter describes the results of the work performed to improve the attrition résistance of the composite Ti0 2 sphères. Variables such as the rate of hydrolysis and condensation during the sol-gel processing, the sphères formation process, and drying and heat treatment conditions were studied. At the same time, the activity of the sphères to dégrade formic acid (FA) was determined. Photocatalyst characterization was performed using BET method to measure the surface area of the sphères, X-ray diffraction (XRD) to identify the polymorphic phases présent in the Ti02 sphères, and scanning électron microscopy (SEM) to identify the microstructural characteristics of the catalyst.  6.2 Compositional study of the Ti02 sphères formulation The préparation of the composite TiÛ2 sphères involves a séries of variables that affect not only the chemical composition of the sphères, but also their physical characteristics, such as the size and shape of the catalyst. Table 6.1 shows some of the variables involved during the préparation process of the composite Ti0 2 sphères. Thèse are variables that may influence the chemical composition of the sol-gel, the shape of the sphères and their production process.  88  6. Photocatalyst Improvement and Characterization Table 6.1: Variables involved during the production process of the composite TÏO2 sphères. Variable - Composite Ti0 2 production Effect / Purpose of the variable Amount of Alcohol  Solvent  Amount of water  Induce hydrolysis reactions  Amount and type of catalyst  Control hydrolysis and polycondensation reactions  Amount of TTIP  Precursor of Titanium  Stirring time  Homogenize the solution  Mass Degussa P-25  Filler material  Chitosan solution characteristics  Helps with the sphères formation and aging  pH of the Ammonia solution  Helps with the sphères formation  Size of the droppers used to produce the sphères  Size of the TIO2 sphères  Height of the droppers  Shape of the sphères  Volume of solution in the droppers  Production of the sphères (define shape of the sphères)  Drying conditions (température and time)  Define the characteristics of the xerogel  Heat treatment conditions (température and time)  Densification of the sphères  Despite many variables involved in the process, only some hâve direct and significant impact on the attrition résistance of the catalyst. For example, the sol-gel matrix properties, as well as the drying and heat treatment conditions could hâve important contributions to the properties of the sphères. Particularly, the hydrolysis and condensation reactions that take place during the sol-gel process may influence the attrition résistance of the catalyst since the condensation reactions are responsible of the network construction (bonds between particles) along the gel. In gênerai, hydrolysis and condensation occur simultaneously once the hydrolysis reaction has been initiated. Many factors that can influence thèse two processes include the amount of water added, the nature and concentration of the catalyst, the 89  6. Photocatalyst Improvement and Characterization solvent used and the séquence of mixing (Hench and West, 1990; Nazeri et al., 1993; Wen and Wilkes, 1996). Thèse condensation reactions continue through the drying stage. Therefore, the drying conditions of the sphères should be adéquate to promote thèse reactions to continue. Finally, the heat treatment is necessary to densify the gel and produce a body dense enough to give the sphères strength. Preliminary tests with thèse and other variables were done to détermine their effects on the attrition résistance and the activity of the catalyst. The activity was monitored using FA as a model organic compound.  6.2.1 Preliminary tests and effects of various parameters The photocatalyst had originally been developed in the Advance Oxidation Research Laboratory at the Department of Chemical and Biological Engineering at UBC. Nonetheless, the original development did not focus on the attrition résistance of the catalyst. Therefore, experiments were performed modifying the original formulation of the sphères to détermine the impact of some of the variables involved during the production process of the catalyst. The response variables for this study were the level of attrition after 3 h of continuous use in the reaction System, shown in Figure 4.8, and the photocatalytic activity of the sphères to dégrade FA. In ail the cases, 25 g of catalyst were used. Among ail the variables listed in Table 6.1, the following were selected for this stage of the investigation: i) amount of water used to prépare the sol; ii) type and concentration of the catalyst for sol préparation; iii) amount of filler material (Degussa P-25) used to produce the CSG; iv) pH of the chitosan solution and its concentration; and v) drying and heat treatment conditions. As pointed out before, it was believed that the characteristics of the sol (and eventually of the gel) hâve an impact on the attrition résistance of the catalyst. This is because the hydrolysis and condensation reactions that take place during the whole process are responsible for the network formed during gelation. Thus, the proper amount of water and catalyst are key factors to investigate. At the same time, the chitosan solution could play an important rôle in the sphères characteristics since  90  6. Photocatalyst Improvement and Characterization it is used as a binder during the sphères formation. In addition, chitosan solution is responsible for the hardening process of the sphères once they are produced. The drying process and the heat treatment are needed to remove the liquid that is trapped inside the sphères and to densify the photocatalyst. Table 6.2 lists the range of variables explored in this stage of the research. For each experiment, only the indicated variable was changed while ail the other variables were kept constant, according to the original formulation of the catalyst. The main focus was to détermine if the variable has an impact on the response variables.  Table 6.2: composite Experiment 1 2 3 4 5 fi  7 8  Range of the variables to be studied during the preliminary tests with the TiQ2. Variable Original Formulation Range of study 0.040 ~ 0.053-0.070 Amount of water (g H20 ml/ 1 TTIP) Type of the catalyst HCI Amount of catalyst (mLmL 1 TTIP) 0.20 0.10-0.13 Amount offiller (g P-25 mL'1 TTIP) 0.40 0.30-0.50 pH of the Chitosan solution (adim) 4.94 4.01 - 4.58 Concentration of Chitosan in the 1Q 15 polymeric solution (g L"1) Drying conditions 15days-23°C 20 h - 80°C Calcination time (h) 1 3  6.2.1.1 Type of catalyst used during the sol-gel préparation As mentioned above, the type of catalyst could affect the properties of the gel. Hence, différent acid catalysts were tested to détermine the effect on hydrolysis and condensation reactions. Phosphoric acid (H3PO4), acetic acid (CH3COOH) and nitric acid (HNO3) were used in the same ratio as HCI (0.20 ml_ ml_"1 TTIP), which was used as the catalyst in the original formulation of the sphères, as shown in Table 6.2. Detailed studies on the impact of the catalyst, performed by Brinker and Scherer (1990), Livage (1998) and Hench and West (1990) for différent sols, e.g. silica sols, showed that the catalyst used during the process has a clear impact on the gelation time, syneresis and the change of the isoelectric point (IEP).  91  6. Photocatalyst Improvement and Characterization In this research, the use of H3PO4 significantly increased the viscosity of the solution producing some clusters when the solution was in contact with titanium (IV) isopropoxide (TTIP). With CH3COOH, the clusters were formed when the CSG was mixed with the chitosan solution. In both cases, the sphères were not produced because of the rapid increase in the viscosity of the solution. Using HN0 3 , on the other hand, no significant changes were observed with respect to the original formulation because the characteristics of HCI and HN0 3 are similar, e.g., pKa. In this case, the sphères were formed and the attrition was measured showing a value of 20 mg L"1 which represents no change with respect to the original formulation (Figure 6.1).  n C\IK -, u.uzo  --  - --  -  -  -  -  T  J  e 0 0.020 < 'ô » 0.015 -  1 OF CHN03  3 0)  ç 0 0.010 H N-  0  +rf  1 0.005 E < n nnr\ u.uuu Composite Ti0 2 sphères  Figure 6.1: Attrition résistance of the photocatalyst produced with a) original formulation (OF); b) using HN0 3 as a catalyst instead of HCI (CHNO3)- Error bars represent 95% Cl of triplicate runs.  The results with HN0 3 are consistent with theory, as it is well known that more acidic conditions strongly delay condensation processes. Thus, it could be expected that a weaker acid would lead to faster condensation reactions, resulting in  92  6. Photocatalyst Improvement and Characterization a différent gelation time. In other words, it is expected that the higher the pH, the smaller the gelation time (Livage et al., 1988; Brinker and Scherer, 1990). When no acid is added, the resuit is not a gel but a precipitate. Based on the results obtained hère, and because HN0 3 did not give significant improvements, HCI was kept as the catalyst for this application.  6.2.1.2 Water and catalyst (HCI) concentration Water is one of the key reagents during sol-gel processing. It is partially responsible for the hydrolysis reactions that take place during the sol-gel formation. At the same time, the structure of the final gel dépends on the contribution of the hydrolysis and condensation reactions, which define the structure and morphology of the resulting oxide. Hence, the characteristics of the final product dépend on the relative contribution of thèse two reactions. With water content and acid concentration being very important, their impact on the attrition résistance of the catalyst was investigated. First, the effect of increasing the amount of water in the system was investigated by changing the water content from 0.040 g H 2 0 per mL TTIP to 0.053 g H 2 0 per mL TTIP and 0.070 g H 2 0 per mL TTIP. When 0.070 g H 2 0 per mL TTIP was used, the viscosity of the solution increased and a lot of clusters were formed in the reaction média. Therefore the sphères could not be produced at this concentration of water. On the other hand, the effect of acid was studied by reducing its concentration from 0.20 mL HCI to 0.13 mL HCI and 0.10 mL HCI per mL TTIP. With 0.10 mL HCI per mL TTIP, immediately after the addition of TTIP, clusters were formed and the sphères could not be produced. Figure 6.2 shows the effect of using 0.053 g H 2 0 per mL TTIP in one formulation and 0.13 mL HCI per mL TTIP in another formulation. As shown, there was around 32% réduction in attrition (Le., the amount of Ti0 2 particles released in the water) for each of the abovementioned conditions. Even though there could be an interaction between thèse two variables, only the individual contribution to the  93  6. Photocatalyst Improvement and Characterization attrition résistance is shown in Figure 6.2. The interaction between thèse two variables will be discussed later.  n DOR T ? 0.020 o «5  1 OF  c  |  0.015 -  3 V> C  2 0.010 -  c,HCI  CH20  H 4-  o  | 0.005 E < n nnn u.uuu Composite Ti0 2 sphères  Figure 6.2: Attrition résistance for three différent processes conditions; a) original formulation (OF); b) increased amount of water (CH2O) in the sol-gel matrix; c) reduced the amount of HCI (CHci) in the sol-gel matrix. Error bars represent 95% Cl of triplicate runs.  The viscosity of the sol-gel changed with respect to the amount of water and acid added. With 0.053 g H 2 0 per mL TTIP the viscosity at 25°C was 5.55 cP (compared with the 4.83 cP of the sol-gel used for the OF). On the other hand, when the amount of acid was 0.13 mL HCI per mL TTIP the viscosity was 5.39 cP. As expected, the viscosity in both cases increased compared to that in the OF, an indication that the hydrolysis and condensation reactions were promoted by adding more water and less acid.  94  6. Photocataiyst Improvement and Characterization 6.2.1.3 Amount of filler material (Degussa P-25) Commercial Ti0 2 (Degussa P-25) was dispersed in the sol as filler to fabricate composite high performance sol-gel materials. P-25 was chosen because of its demonstrated high photoefficiency to dégrade organic pollutants (Porter et al., 1999; Arana et al., 2009). Figure 6.3 shows the effect of the amount of Degussa P-25. It is shown that reducing the amount of P-25 from 0.40 g P-25 per mL TTIP to 0.30 g P25 per mL TTIP led to slightly lower attrition. The amount of TiÛ2 particles released to the solution was reduced by about 9%. On the other hand, increasing the amount of P-25 added (Le., 0.40 g P-25 to 0.50 g P-25 per mL TTIP) showed similar level of attrition compared to the original formulation.  \J.\J£.yJ  -  T  _l B  c 0.020 o V) c o  g- 0.015 -  1  T  OF  1  MP-25  LP-25  3 (0 C  OI  2 0.010 iM--  o  +J  c  g 0.005 E  <  n nnn Composite Ti0 2 sphères  Figure 6.3: Attrition résistance for three différent process conditions; a) original formulation (OF); b) using 0.30 g P-25 mL"1 TTIP (LP-25); c) using 0.50 g P-25 mL"1 TTIP (MP-25). Error bars represent 95% Cl of triplicate runs.  Thèse results showed that the amount of filler material did not hâve a significant impact on the attrition of the composite Ti02 sphères. However, some  95  6. Photocatalyst Improvement and Characterization impacts would be observed on the production process of the sphères. With the amount of Degussa P-25 reduced significantly, the viscosity of the solution changes as well. Therefore, lower amounts of P-25 could resuit in a solution that has low viscosity, making it difficult to produce the sphères. On the contrary, higher amounts of P-25 would increase the viscosity of the CSG and the process to produce the sphères would take longer than that of the OF.  6.2.1.4 pH of the chitosan solution The polymeric matrix (chitosan solution) was used in this application because it has a certain viscosity needed for the formation of sphères and also for the fast hardening property that chitosan has upon contact with basic solution. As discussed earlier, chitosan is soluble in dilute acids. Therefore, for this application glacial acetic acid was used to acidify the water and propitiate the solution of the polymer. Then, by modifying the amount of acid added to the water, the pH of the final solution could be adjusted. Figure 6.4 shows the effect of the pH of the chitosan solution on the attrition résistance of the sphères. Reducing the pH to 4.58 (the original pH was 4.94) induced a réduction in the level of attrition in the System. After 3 h of continuous opération, approximately 14 mg L"1 of TiÛ2 was measured in the solution, representing a 32% réduction compared to that in the original formulation (OF). With further réduction of the pH to 4.01, no additional improvement was observed, but the viscosity of the Ti02-chitosan solution increased significantly. Thèse results suggest that decreasing the pH of the chitosan solution resulted in a lower pH for the polymeric matrix-Ti02 CSG solution inducing a harder structure when it came in contact with the ammonia solution. The différence between the pH of the chitosan-Ti02 CSG solution and the ammonia solution helped to increase the strength of TiÛ2 sphères, producing a harder three-dimensional structure. The reports from literature suggested that chitosan shows chemical high cross-linked gels when it is in contact with Glutaraldehyde and that the gelation of this chitosan chains dépends on parameters such as pH, chitosan concentration,  96  6. Photocatalyst Improvement and Characterization ionic strength, etc. (Arguelles-Monal et al., 1998). It is believed that the degree of gelation of chitosan in ammonia solution dépends on the pH différence between thèse two solutions, making stronger structure when the différence increases. It is important to note that more acidic condition in the chitosan solution could also retard the condensation reactions of the TÏO2 CSG material.  n no1^  T f  0.020 -  0  1  OF  "35 c 0  g- 0.015 3 (0 e  j -  2 0.010 -  MpH  LpH  O C  4-1  g 0.005 £ < n nnn u.uuu Composite Ti0 2 sphères  Figure 6.4: Attrition résistance for two différent process conditions; a) original formulation (OF); b) when the pH of the chitosan solution was 4.58 (MpH); c) when the pH of the chitosan solution was 4.01 (LpH). Error bars represent 95% Cl of triplicate runs.  6.2.1.5 Concentration of chitosan in the polymeric solution The main purpose of chitosan in this application was to propitiate the necessary conditions to form the sphères. The effect of chitosan concentration on the attrition résistance of the catalyst is shown in Figure 6.5. In gênerai, increasing the concentration of chitosan to 15 g L"1 (in the original formulation the concentration of chitosan was 10 g L"1) induced a réduction in the attrition by 38% compared to that in the original formulation. With this incrément in the chitosan concentration, the  97  6. Photocatalyst Improvement and Characterization viscosity of the solution increased from 135 cP to 206 cP (measured at 25°C) and the pH of the solution increased to 5.52. As mentioned before, condensation reactions continued to happen during the aging, drying and heat treatment stages. It is expected that less acidic conditions in the chitosan-Ti02 CSG solution (based on the incrément in pH of the chitosan solution) might promote those reactions to continue and to do so even faster. As mentioned in Chapter 2, condensation reactions are responsible to build bridges between chains; hence, this change in acidic conditions of the chitosan-Ti02 CSG solution might be the cause of a strong structure in the sphères showing a réduction in the amount of Ti0 2 particles in suspension.  n no*; u.uzo  T f  0.020 -  o » c  1  OF  1 0.015 3 (0 C  T  1  Q 0.010 -  ^•"•conc.  H H-  O  1 0.005 E < n nr\r\ u.uuu Composite Ti0 2 sphères  Figure 6.5: Attrition résistance for two différent process conditions; a) original formulation (OF); b) when the chitosan solution had a concentration of 15 g chitosan L"1 (Chconc)- Error bars represent 95% Cl of triplicate runs.  As mentioned before, chitosan is used in this application due to the fast hardening property that it has upon contact with basic solution. Therefore, increasing the concentration of chitosan in solution, might lead to a harder structure in the 98  6. Photocatalyst Improvement and Characterization sphères. Nonetheless, it is important to mention that increasing the concentration of chitosan also increases the viscosity of the polymer-CSG solution. If the viscosity of this solution increases by a significant degree, the production process of the sphères will be more difficult and take longer time because the solution flows slowly.  6.2.1.6 Drying process of the sphères The fresh sphères contain significant amount of liquid (alcohol, water, ammonia solution, etc.) that must be removed from the structure. That is, the gel must be dried before the appropriate heat treatment is applied. As the drying process proceeds, the network formed during the gelation process becomes stiffer because of the formation of new bonds due to condensation reactions (Rahaman, 2007). Thus, an accelerated drying process increases the solid network matrix which has an impact on the mechanical strength of TiÛ2 sphères. However, if the drying process is too aggressive, cracks can be induced in the body of the gel. The relation between the drying process and the attrition résistance can be observed on Figure 6.6. As can be seen, when the drying process was done at 80°C for 20 h, instead of 15 days at room température (RT - 23°C), the amount of TiÛ2 particles in suspension was reduced. When the température of the furnace was increased to 600°C for 1h immediately after the drying process, the amount of TiÛ2 in suspension dropped further to 11 mg L"1 from around 22 mg L"1 in the OF (DCOnt.)However, when the catalyst was first cooled down to RT and then placed in the furnace for 1h at 600°C, the amount of TiÛ2 in suspension was 15 mg L"1 (DCOoi). This effect could be explained in terms of the polycondensation reactions that take place during the drying and heat treatment. If the température of the catalyst is decreased after the drying process, many of those reactions would slow or stop and the solid network will not be as compact as the case where such reactions continue with immédiate heat treatment after drying.  99  6. Photocatalyst Improvement and Characterization  T e 0.020 o w c  1 OF  g>  g- 0.015 3  m c  DCool  Q* 0.010 i*o  Dcont.  4-i  | 0.005 E  <  Composite Ti02 sphères Figure 6.6: Attrition résistance for three différent process conditions; a) original formulation (OF); b) catalyst dried for 20 h at 80°C, cooled down, and then placed in the furnance at 600°C for 1 h (Dcooi); b) catalyst dried in the furnace at 80°C for 20 h, followed by immédiate température increase to 600°C for 1 h (DCOnt.)- Error bars represent 95% Cl for triplicate runs.  6.2.1.7 Heat treatment conditions The main purpose of heat treatment is to densify the gel and produce a strong body. The relation between heat treatment and the attrition résistance of Ti0 2 sphères has already been demonstrated in section 5.3 (Figure 5.3). An increase in the calcination température of the sphères reduced the attrition, i.e., the amount of particles released from the catalyst to the water. Nonetheless, the activity is reduced by increasing the heat treatment température (Figure 5.2 in Section 5.2.2). Another important factor affecting photocatalyst attrition is the length (or time) of heat treatment. Figure 6.7 shows the effect of increasing the calcination time from 1 h to 3 h at 600°C. As shown, the amount of Ti0 2 in suspension drops to 13 mg L"1 from around 22 mg L"1 in the OF. The effect of thèse two conditions (time and température) can be explained based on the impact that higher températures and/or longer exposure to a given température induce tighter solid network that will resist  100  6. Photocatalyst Improvement and Characterization better the attrition effect. The higher the température and/or the longer the calcination time, the denser the sphères would be, which means the solid network is tighter and more compact. Therefore, the strength of the composite TÏO2 sphères increases, improving the attrition résistance of the photocatalyst.  0.025 T c 0.020 w c  »3 0.015 (A  o 0.010 -  1  OF  T  J_  Ctime  Io 0.005 E < 0.000 Composite Ti02 sphères Figure 6.7: Attrition résistance of TÏO2 sphères for two différent process conditions; a) original formulation (OF); b) catalyst fired for 3 h at 600°C instead of 1 h at 600°C (Ctime)- Error bars represent 95% Cl for triplicate runs.  In ail the above mentioned cases, the activity of the catalyst was not drastically impacted by the changes in the sphères formulation or synthesis procédure. Indeed, there was no statistical différence between the apparent rates constants obtained in ail the expérimental runs involving différent batches of sphères. This inference was based on the comparison between F values obtained for the apparent rate constants obtained from every expérimental run (F = 2.47) and the F9,2o of 4.57 (at 99% confidence level). Appendix A provides more détails on thèse analyses.  101  6. Photocatalyst Improvement and Characterization 6.2.2 Expérimental design Based on the results obtained in Section 6.2.1, an expérimental design was developed to study the effects and interactions of some of the variables involved in the production process of the sphères. Thèse tests were performed to complément the earlier results on the effect of each individual parameter (determined independently). As discussed earlier, the amount of water and catalyst, the drying and heat treatment conditions, and the pH and concentration of chitosan had a positive impact on the sphères résistance to attrition. However, the pH of the chitosan solution and its concentration had a strong impact on the production process of the catalyst because of the changes in the viscosity of the solution, which represent a change in the production process. Thus, in order to hâve a similar effect, the pH of the ammonia solution can be increased from 11.75 to 12.0 (increasing the percentage of ammonium hydroxide from 10% to 20%) to increase the pH différence between the chitosan-Ti02 CSG solution and the ammonia solution. Higher concentrations of ammonia made the production of the sphères more difficult because the présence of vapours induced a fast hardening of the chitosan-Ti02 CSG solution in the droppers. With respect to the amount of filler (Degussa P-25) added to the sol-gel, 0.30 g P-25 per mL TTIP was used to produce the catalyst. A fractional factorial expérimental design (25"1) was set-up to détermine the effect of each variable and the interactions among variables. This approach is useful in an investigation of a multivariable System since it can indicate major trends with promising directions using relatively fewer runs (16 runs instead of the 32 runs for the complète 2 5 factorial design) in the factor space (Box et al., 1978). At the same time, the effects and interactions among variables can be identified. Table 6.3 shows the expérimental design distribution and the levels of the variables investigated. The expérimental runs were carried out using the FBPR and the System shown in Figure 4.8. In ail the experiment, attrition was determined using the procédure detailed in Chapter 4 (Section 4.7) and the activity was measured using FA as a model organic compound.  102  6. Photocatalyst Improvement and Characterization Table 6.3: Fractional factorial design (25"1) to study the effects of and interactions among variables involved in the production of the composite Ti02 sphères. .. _ . .. Amount of Amount of Concentration of ~ . 0 , . P Runs Water(g HCI (mL NH4OH solution ^ °nZT oZ pr0CeSS t i m e (h) 0rder mL-1TTIP) mL-1TTIP) (%) 10 5 RT-15days 3 1 0.040 0.133 6 0.053 2 0.133 10 RT-15days 1 7 0.040 0.200 10 RT-15days 1 3 4 0.053 0.200 4 10 RT-15days 3 3 0.040 0.133 5 20 RT-15days 1 8 0.053 0.133 6 20 RT-15days 3 2 0.040 0.200 3 7 20 RT-15days 1 0.053 0.200 8 20 RT-15days 1 12 0.040 0.133 9 10 80°C - 20 h 1 16 80°C - 20 h 3 0.133 10 0.053 10 9 80°C - 20 h 3 11 0.040 0.200 10 15 1 10 80°C - 20 h 12 0.053 0.200 14 80°C - 20 h 3 13 0.133 20 0.040 11 80°C - 20 h 1 14 0.053 20 0.133 13 1 20 80°C - 20 h 15 0.040 0.200 10 80°C - 20 h 3 16 0.200 0.053 20  The results obtained from the factorial experiments are shown in Table 6.4. The amount of particles in suspension reduced from 22 mg L"1 (run # 3 which représenta the original formulation) to a minimum 6.73 mg L"1, representing a 70% réduction in attrition. In terms of photocatalytic activity to dégrade FA (with the initial concentration of 100 mg L"1), the results indicated that the dégradation rate constants were not statistically différent in ail cases. The effects and interactions among variables were calculated using Yates algorithm (Box et al., 1978; Riedwyl, 1998). The analysis was done to détermine the combination of variables that led to the lowest attrition and highest activity. Detailed calculations are presented in Appendix A.  103  6. Photocatalyst Improvement and Characterization Table 6.4: Response variables of the fractional factorial design (25"1) for the composite Ti0 2 sphères (mass of catalyst equal to 0.025 kg) Reaction rate constant Amount of TiC"2 particles in Runs suspension (mg L"1) (min"1) 8.63 1 0.286 2 9.65 0.283 3 21.92 0.281 4 9.78 0.297 5 8.51 0.314 6.73 6 0.325 7 14.19 0.306 13.53 8 0.323 9 7.68 0.338 10 15.71 0.336 11 16.55 0.345 0.344 12 12.68 8.81 0.316 13 14 7.39 0.344 11.97 0.330 15 13.23 0.305 16  6.2.3 Modified formulation to produce the Ti0 2 composite sphères Considering the results obtained in the previous section, the level of the variables giving the best resuit in terms of mechanical strength and photoefficiency of the photocatalyst are listed in Table 6.5. As can be seen, higher amount of water, lower amount of acid, higher concentration of ammonia in solution, faster drying process and longer time at 600°C gave the best combination of variables. Those variables were then used for the production of photocatalytic sphères in ail the remaining experiments in this research.  Table 6.5: Variables chosen from the expérimental design analysis. Variables studied Level Amount of water "~ " 0.053 g H 2 0 mL"1 TTIP Amount of HCI 0.13 mL HCI mL 1 TTIP Concentration of NH4OH solution 20% v/v Drying process conditions 80°C - 20 h Calcination time 3h  104  6. Photocatalyst Improvement and Characterization Considering the variables listed in Table 6.5, a batch of catalyst was prepared and tested in the FBPR. Figure 6.8 shows the effect of ail thèse variables collectively in one formulation (modified formulation - MF) versus the original formulation (OF). As shown in the figure, the combination of ail thèse variables gave an attrition of 7 mg L"1 (with a particle size distribution around 15 |xm, as can be seen in Appendix B) which represents approximately 70% réduction in the amount of Ti0 2 parades released in the water.  n n*?^  T !  i  0.020 -  1 OF  0.015 -  w £  £  0.010 -  s  l  T  0.005 -  1 MF  <  n nnn Composite Ti0 2 sphères  Figure 6.8: Attrition résistance comparison between the original formulation of the catalyst (OF) and the modified formulation (MF) considering variables listed in Table 6.5. Error bars represent 95% Cl of triplicate runs.  It is important to mention that the viscosity of the sol-gel used in the MF was 7.94 cP, which is higher than the viscosity of the sol-gel solution in the OF (4.83 cP). Hence, increasing the amount of water and reducing the amount of acid in one formulation induce a 64% incrément in the viscosity of the sol-gel. As discussed before, this increase in viscosity is a reflection of the condensation reactions that take place in the solution.  105  6. Photocatalyst improvement and Characterization To further confirm the results obtained in Figure 6.8, some additional tests were performed by varying some variables in the modified formulation (MF). As shown in Figure 6.9, reducing the concentration of NH4OH (from 20% to 10%) increased the attrition by 80% toi 3 mg L"1. Reducing the amount of water during the préparation of sol-gel (from 0.053 g H 2 0 to 0.040 g H 2 0 per ml_ TTIP) gave 43% more attrition in the System (10 mg L"1). Finally, drying the catalyst at 80°C for 20 h followed by cooling down to room température, before heat treatment, led to higher attrition (14 mg of TÏO2 L"1 in a litre of solution).  0.025 1 f  o  0.020 -  c  g0.015 3 M  | 0.005 E < 0.000 -  T  1 Ccool  ^Amm  1-  o  I  T  c O 0.010 -  CH20  MF  Composite Ti0 2 Sphères  Figure 6.9: Attrition résistance for five différent process conditions; a) modified formulation (MF); b) reducing the ammonia concentration to 10% (CAmm); c) reducing the amount of water to 0.040 g H 2 0 mL"1 TTIP (CH2o); d) the catalyst was cooled down to RT after the drying process and before the heat treatment at 600°C for 3 h (Ccooi). Error bars represent 95% Cl for triplicate runs.  There are few literature reports for the quantitative détermination of Ti0 2 particles in suspension in a fluidized bed System. Nonetheless, it has been reported that using silica-gel as a support, the amount of Ti0 2 in suspension varies from 6 mg L"1 to 312 mg L"1, mainly due to the préparation of photocatalyst and its  106  6. Photocatalyst Improvement and Characterization interaction with the support itself (Bideau et al., 1995). Also, using soda-lime glass beads, the Ti0 2 in suspension was identified between 1 mg L"1 to 130 mg L"\ depending on the catalyst préparation procédure (Qiu and Zheng, 2007). Comparing the literature results with those obtained in this study for the composite Ti0 2 sphères, it is clear that the level of attrition is certainly low with the advantage that the activity is not hindered due to the catalyst detachment.  6.3 Catalyst characterization As mentioned before, one of the most important parameters in the synthesis of the TIO2 sphères is the calcination température, mainly because it plays a critical rôle defining the crystalline phases, the surface area and most importantly, the activity of the catalyst. Even though the calcination température used in this application is 600°C, the following discusses four différent températures (600°C 900°C) in order to analyze différent characteristics of the catalyst (prepared using modified formulation). As shown in Table 6.6, commercial TÏO2 (Degussa P-25) had the highest surface area (54.03 m2 g"1) and the highest percentage of anatase (87.76%). Even though the composite Ti0 2 sphères showed a significantly high percentage of anatase (84.20%) and a very high surface area (187.3 m2 g"1) before the heat treatment, once the heat treatment was applied, the surface area, percentage of anatase, pore volume dropped noticeably. The same tendency was shown when the température during the heat treatment increased from 600°C to 900°C  6.3.1 XRD analysis Figures 6.10 shows the X-ray diffraction pattems of the pre-calcined TiÛ2 powder (Degussa P-25) and the composite sphères calcined at 600°C. There are broad peaks at 20 = 30° (identify as the most intensive peak (1 0 1) for the anatase) and at 26 = 32° which represents the rutile (1 1 0) phase. As shown, in both cases the peaks were located in exactly the same position, but Degussa P-25 has a  107  6. Photocatalyst Improvement and Characterization smaller rutile peak than the composite sphères which explains the high percentage of anatase in the commercial titania.  Table 6.6: Surface area and anatase composition (using BET and XRD respectively) of pre-calcined powder (Degussa P-25 - as received) and composite TÏO2 sphères prepared using the new formulation and calcined at différent températures. Anatase Surface area Pore volume Pore size Material Heat treatment (cm3 g"1) (nm) (m2 g"1) (%) (!ÇJ Degussa — 87.76 54.03 0.2275 16.56 P-25 — 84.20 187.32 0.2707 7.17 Ti0 2 Sphères  600  64.69  29.37  0.1680  18.16  700  0.0  6.44  0.0671  31.31  800  0.0  2.13  0.0315  38.68  900  0.0  N/A  N/A  N/A  A: Anatase R: Rutile  Ti02 sphères Degussa P-25  v  0  I I 1 I I I I I  2-Theta - Scale  Figure 6.10: X-ray diffraction patterns of the commercial Ti0 2 powder (Degussa P-25 as received) and the composite TÏO2 sphères produced with the modified formulation (MF) calcined at 600°C.  108  6. Photocatalyst Improvement and Characterization Figure 6.11 shows the effect of calcination températures on the X-ray diffraction patterns of the photocatalyst sphères calcined at différent températures (600 900°C). The peak intensity of anatase (10 1) dramatically disappeared and the peak intensity of rutile (1 1 0) significantly increased when higher températures were applied, demonstrating that increasing heat treatment température accélérâtes phase transformation from thermodynamically metastable anatase to the most stable and more condense rutile phase (So et al. 1997). In gênerai, the primary formed TÏO2 particles usually contain large portions of defect sites and because at high température some bond breaking and atom arrangements are induced, the transformation of anatase to rutile is facilitated. Because anatase has a structural similarity to rutile this transformation is initiated by forming the rutile nuclei along the anatase interface. Thus, the rutile structure can be developed at the expense of anatase crystals (Yin et al., 2001; Su et al., 2004).  A: Anatase R: Rutile  R 900°C .Ç 3000  800°C 700°C 600°C i '  ' i ' 1 1 1 r  2-Theta - Scale  Figure 6.11: X-ray diffraction patterns of Composite Ti0 2 sphères calcined at différent températures (600, 700, 800 and 900°C).  109  6. Photocatalyst Improvement and Characterization Using Rietveld refinement method, the amount of anatase and rutile in the composite photocatalyst was determined. The weight fractions of anatase and rutile for sphères calcined at 600°C were 64.69% and 35.31%, respectively. In comparison, the commercial Ti0 2 (Degussa P-25) has a weight fraction of 87.76% for anatase and 12.24% for rutile. A 100% phase transformation from anatase to rutile took place at a température between 600°C and 700°C. Increasing the calcination température by 100°C (from 600°C to 700°C) led to the complète disappearance of anatase as shown in Figure 6.11 (no peak is présent for 700°C).  6.3.2 BET analysis BET was used for the surface area, pore size distribution and pore volume analysis of the composite TÏO2 sphères. Table 6.6 shows the dependence of the surface area, pore size and pore volume on the calcination température of the composite Ti0 2 sphères. The spécifie surface area and pore volume shifted towards smaller values at higher calcination températures. At 700°C the spécifie surface area decreased to 6.44 m 2 g"1 and the pore volume reduced by 60% with respect to the value reported when the catalyst was calcined at 600°C. Although a large portion of the mesoporous structure was maintained at 700°C, the pore volume was reduced significantly, likely because higher calcination températures induce crystallization of Ti0 2 , subséquent crystal growth, and shrinkage of the pores (Peng et al., 2005). At 800°C, the mesoporous structure has significantly disappeared, mainly due to the particle size of the catalyst. In gênerai, the spécifie surface area and the pore volume decreased with the increase of calcination température; at the same time, the average pore diameter increased. On the other hand, the average pore diameters increased upon heat treatment, which is consistent with the literature reports (Xu et al., 2008; Bischoff and Anderson, 1995). The isotherms and pore size distributions of the composite Ti0 2 sphères calcined at 600°C, and commercial Ti0 2 powder (Degussa P-25) are shown in Figure 6.12. It is clear that Degussa P-25 has a higher surface area than the composite photocatalyst, with a value of 54.03 m2 g"1 for the powder and  110  6. Photocatalyst Improvement and Characterization 29.37 m2 g"1 for the composite TÏO2 sphères calcined at 600°C. According to the BDDT (Brunauer, Deming, Deming and Teller) classification, the powder and the sphères show isotherm type IV, exhibiting hystérésis loops mostly of H2 and H3. This resuit indicates that the composite photocatalyst porosity in the meso-range (2~50nm) (Sink et al., 1985). In the sol-gel process, nanosize particles are formed by hydrolysis and condensation processes. If gelation occurs, thèse particles are connected together to form three dimensional aggregates. During heat treatment (sintering), the primary particles are grown into larger ones and bond together to form a solid network, which exhibits a narrow pore size distribution and smaller surface area at higher température (Brinker and Scherer, 1990) (see Figure 6.12). On the other hand, Degussa P-25 powder is prepared at elevated température (>1000°C) by flame synthesis, containing larger crystalline anatase and rutile particles. As a resuit, Degussa P-25 has the widest pore diameter région. Comparing the BET results and XRD analysis with the ones obtained for the sphères produced using the original formulation (Table 5.3), it is clear that the accelerated drying process, the calcination time and/or the sol-gel composition had an impact not only on the percentage of anatase, but also on surface area, pore size and pore volume. In gênerai, higher percentages of anatase and higher surface area hâve been determined for the sphères produced with the original formulation.  111  6. Photocatalyst Improvement and Characterization  160  û. 140  A: Commercial Ti0 2 powder 0: Composite Ti0 2 Sphères  u> 1?0  J?-*  E  o^ 100 •o fl> .Q i_  o»  •o  ra >>  HO fi()  40  C (0 3  a  20  0.00  0.20  0.40  0.60  1.00  0.80  Relative pressure (p/p°)  0.120  A: Commercial Ti0 2 powder 0: Composite Ti0 2 Sphères  0.100  3 •0  F  0.080  o. <D  b 0.060 o >  ?  0.040  o  Q.  0.020 0.000 20  40  60 100 80 Pore Diameter (nm)  120  140  160  Figure 6.12: Surface area and pore size distribution for samples from différent conditions, (a) surface area of commercial Ti0 2 powder (Degussa P-25) and the composite Ti0 2 sphères calcined at 600°C, (b) pore size distribution of Degussa P25 and composite TiÛ2 sphères calcined at 600°C.  112  6. Photocatalyst Improvement and Characterization 6.3.3 TGA analysis In order to better understand the thermal behaviour of the composite TÏO2 sphères during différent drying and heat treatment conditions the catalyst was analyzed using TGA at différent heating rates to emulate the actual process. The TGA analysis was performed in three stages. The first stage was from room température to 80°C as shown in Figure 6.13a. The sample weight was reduced by 87% within the 25 min analysis. The second stage involved maintaining the température constant (80°C) for 20 h, showing an additional réduction of 7.6% with a final value of 1.59 mg (Figure 6.13b). Finally, the third stage constituted heat treatment, for which the température increased from 80°C to 600°C, showing a final catalyst weight of 1.38 mg (Figure 6.13c). The loss of mass for the TiÛ2 sphères was mainly due to the evaporation of the residual water, alcohol, chitosan décomposition and ammonia that had been trapped in composite structure. Weight réduction of the sample was as high as 9 1 % resulted by the sintering and densification of the composite.  y 90  - 70 •  bb  ght (mg)  c  ID.U  "S 8.0 -  -60 /  /  -- 50  h.  w  H (D  3  T3 <0 •1  fi)  - 40 C -1  ^  4  ^  E (0  - 30  CD 0  - 20  4 0-  O  -- 10  2.0 0.0 -  10  1  1  15  20  -- 0  Time (min) Figure 6.13: TGA curves of the composite TiÛ2 sphères dried at 80°C for 20 h and calcined at 600°C for 3 h, (a) first 25 min of the drying process, (b) isotherm drying process, (c) heat treatment for 3 h.  113  6. Photocatalyst Improvement and Characterization  1.68  80.30  1.66  80.25  D)  E  • — •  1.64  <-•  •c  •£  1.62  5  a> D.  E  1 60  m  V)  1.58  80.05  1.56  80.00 25  150  275  400  525  650  775  900  1025 1150  Time (min)  1 R^  - 700 -600  1.60 /  \  ^  X  JL  500 5» 3 - 400©  155 -  O)  S)  • | 1.50 ©  |  - 300 5 ©  1.45 -  (0 (0  1.40 i .ou 1 ^  /  ^  1206  C  - 200 O - 100  -  1256  1306  1356  1406  -'0 1456  Time (min)  Figure 6.13 (cont): TGA curves of the composite Ti0 2 sphères dried at 80°C for 20 h and calcined at 600°C for 3 h, (a) first 25 min of the drying process, (b) isothermal drying process, (c) heat treatment for 3 h.  6.3.4 SEM analysis Microstructure of the produced TiC>2 sphères was analyzed using SEM (Figure 6.14). Composite TIO2 sphères using the new formulation had an average  114  6. Photocatalyst Improvement and Characterization diameter of 1.15 mm. At higher calcination température, i.e. 900°C, the average diameter was reduced to 0.90 mm. In either case, the catalyst particles showed consistency in size and shape. Figure 6.14 shows that the composite Ti0 2 sphères had a relatively smooth surface despite the présence of some roughness along the surface of the catalyst. No cracks were found after heat treatment, confirming that there were no extrême stresses during the catalyst production process. Grooves around the sphère (Figure 6.13b & 6.13d) as well as the crater shown in Figures 6.13b & 6.13c are the resuit of the production process; the same for ail the sphères.  Figure 6.14: SEM micrograph of a photocatalytic composite sphère calcined at 600°Cfor3h.  115  6. Photocatalyst Improvement and Characterization  6.4 Ti0 2 sphères durability and attrition détermination over time A batch of composite Ti0 2 sphères (25 g) was used for 25 h in a fluidized bed photocatalytic reactor to détermine how the attrition was changing over time. Every 3 h a sample was collected to détermine the amount of Ti0 2 in water as a measure of attrition, as shown in Table 6.7. The attrition incrément was reduced with respect to the initial value (7 mg L"1) showing a maximum value of 6 mg L"1 after 18 h of use.  Table 6.7: Evolution of the Ti0 2 particles released from the composite Ti0 2 sphères over time. Duration (h) Cumulative attrition (mg L"1) Attrition incrément (mg L"1) 3 7.392 6 11.073 3.681 9 14.285 3.212 12 19.076 4.791 15 23.137 4.061 18 29.197 6.060 21 34.845 5.648 25 39.859 5.014  Figure 6.15 shows the cumulative attrition versus time. As can be seen, there is a linear incrément of the attrition with approximately 40 mg L"1 of Ti0 2 particles after 25 h of continuous opération. As listed in Table 6.7, the incrément every three hours was between 3.2 and 6.1 mg L'1. SEM micrographs of the catalyst after 25 h of continuous use are shown in Figure 6.16 showing that the surface of the catalyst is smoother than that of the catalyst right after the heat treatment. Comparing thèse images with the micrograph shown in Figure 6.14, the catalyst after 25 h of continuous use shows practically no roughness.  116  6. Photocataîyst Improvement and Characterization  09  E  "*•"  c o "c 25 0) (0 3  45 i 40 -  O  35 -  O  30  O  25  « ,r.S 1J 20 -  O H •^ O  E <  O  15 -  O  o  10  c 3 O  O  5 -  oeY J  ()  o I  3  I  1  1  '  I  1  6  9  12  15  18  21  24  27  Time (min)  Figure 6.15: Cumulative attrition of the composite TÏO2 sphères versus time.  Figure 6.16: SEM micrograph of a photocatalytic composite sphère caicined at 600°C after 25 h of continuous use. 117  6. Photocatalyst Improvement and Characterization  6.5 Final remarks This section demonstrated the significant improvement obtained on the attrition résistance of the TÏO2 sphères through modifications made in the formulation and synthesis procédure. The amount of "IÏO2 particles in suspension were reduced from 22 mg L"1 to 7 mg L"1 based on the following changes in the procédure/formulation: a) inducing more hydrolysis and condensation reactions by increasing the amount of water from 0.040 g H2O to 0.053 g H20 per ml_ TTIP and reducing the amount of HCI from 0.20 ml_ HCI to 0.13 ml_ HCI per ml_ TTIP during the sol-gel production; b) increasing the pH of the ammonia solution from 11.75 to 12 (increasing the percentage of ammonia from 10% to 20% v/v); c) accelerating the drying process conditions to 80°C for 20 h; and d) increasing the calcination time from 1 h to 3 h at 600°C. The activity of the catalyst prepared with MF, was similar to that of sphères prepared with the OF (i.e., with none of the abovementioned modifications). Using the MF, the FA dégradation rate constant was between 0.281 min"1 and 0.344 min"1 and using the OF, the dégradation rate constant was 0.315 ± 0.061 min"1. In terms of the characteristics of the sphères, the modifications in the sphères at 600°C resulted in a surface area of 29.37 m2 g"1 (compared with 37 m2 g"1 for that of OF) and a percentage anatase of 65% (compared with 72% for that of OF). The shape and size of the sphères were similar to those in the original formulation. Besides the better attrition résistance and the good activity of the photocatalyst, this new formulation to produce the composite Ti0 2 sphères are more efficient in terms of the sphères formation because less time would be needed to prépare the final product.  118  Chapter Seven Activity of Composite TïCh Sphères 7.1 Introduction Titanium Dioxide is a photocatalyst studied extensively in the last three décades because of its ability to destroy organic pollutants in liquid and gas effluents. It has been shown that under UV irradiation, TÏO2 photocatalyst can dégrade many organic pollutants into harmless substances such as CO2, H20 and minerai acids as a resuit of oxidative redox reactions on the surface of the photocatalyst (Bhatkhande et al., 2002; Kabra et al., 2004; Syoufian et al., 2007; Gaya and Abdullah, 2008; Higgins et al., 2009). This investigation has focused on the development of a composite photocatalyst with high attrition résistance and good photoactivity. As shown in Chapter Six, the catalyst formulation and synthesis were improved, resulting in better attrition résistance. The improvement came through the modification of the hydrolysis and condensation rates during sol-gel préparation, aging time, sphères formation, drying process, and calcination conditions. This chapter focuses specifically on the photocatalytic activity of the catalyst sphères in a fluidized bed photoreactor. Formic acid (FA) and 2,4-Dichlorophenoacetic acid (2,4-D) were selected as model compounds, representing simple and more complex organic contaminants, respectively. In addition, preliminary tests were carried out for the dégradation of natural organic matter (NOM). Expérimental conditions involved différent initial concentrations and fluence rates within the reactor. The performance of the composite TiÛ2 sphères at degrading the organic molécules was also compared with that of Degussa P-25, known as a standard photocatalyst with many literature data (Porter et al., 1999; Arana et al., 2009). Experiments with P-25 were conducted with the amount of powder P-25 (catalyst loading) that gave the maximum activity.  119  7. Activity of Composite TiQ2 Sphères  7.2 Fluence rate détermination The fluence rate applied to the photocatalytic reactor was determined for the expérimental set up showed in Figure 4.7. The fluence rate is defined as the energy provided by the UV lamps per unit time per irradiated area of the reactor (J s"1 m"2). For this expérimental setup, there were three possible lamp positions around the reactor, located at 5, 7.5 and 10 cm from the centre of the reactor. Therefore, the fluence rate was determined for each of those positions, so the effect of distance on the amount of irradiation in the reactor could be determined. It is important to clarify that for the initial fluence rate déterminations, the cylindrical reflector was not used. Potassium ferrioxalate and iodide-iodate actinometers were used to détermine the fluence rate in the System. Potassium ferrioxalate actinometer was first introduced by Hatchard and Parker (1956) and it is probably the most commonly used. Iodide-iodate actinometer was introduced by Rahn (1997) and it has the advantange that it is optically opaque to light below 290 nm and it is optically transparent to wavelengths greater than 330 nm. In other words, it absorbs practically ail of the light below 290 nm but little of the ambient light normally présent in the laboratory, which represents a différence from the potassium ferrioxalate actinometer that absorbs slightly in visible région (Goldstein and Rabani, 2008). Nonetheless, the experiments with potassium ferrioxalate were performed under dark conditions to avoid any interférence in the results. Figure 7.1 shows the différence between the two actinometers. As is shown, this différence was significant especially when the irradiation was high. Potassium ferrioxalate actinometer showed higher fluence rate values at two différent lamp positions (5 cm and 7.5 cm). However, there was no statistical différence between the two actinometers when the lamps were located 10 cm from the reactor. In addition, the iodide-iodate actinometer seemed to show a "saturation effect", meaning that the fluence rate only increased by 16% when the lamps were moved doser to the reactor (from 10 cm to 5 cm). This "saturation effect", as it suggests, was likely due to the saturation of the actinometric solution at very high photon flux that prevents proper differentiation between différent levels of irradiation.  120  7. Activity of Composite TIO2 Sphères To détermine which actinometer represented better the actual irradiation and fluence rate in the System, the fluence rate was converted into the lamp output. This effective output was then compared with the nominal output provided by the manufacturer.  Figure 7.1: Fluence rate dependence on the distance of the UV Lamps from the centre of the quartz tube using two différent actinometers: (0) Potassium Ferrioxalate; (0) lodide-iodate actinometer. Error bars represent 95% Cl of the triplicate runs.  To calculate the effective output of the UV lamps, a model was developed in collaboration with Dr. Gustavo Imoberdorf (of the Advance Oxidation Research Laboratories) based on the data obtained from the actinometrical measurements and the photon absorption distribution in the System. Monte Carlo (MC) algorithm was used in this model considering the trajectories of a large number of photons, as well as possible locations where they are absorbed in the reactor (Imoberdorf et al., 2008). The model considers two main parts: i) the photon émission and ii) the photon propagation in the System. The photon émission in the lamp is modeled according to 121  7. Activity of Composite Ti02 Sphères two stochastic rules, i) the probability for a photon to be emitted is uniform in ail the lamp volume, and (ii) the photon émission is considered isotropic. Once a plausible émission point inside the lamp and a plausible propagation direction of a given photon are evaluated, the photon trajectory is tracked until it leaves the lamp. In addition, the optical properties of the reactor quartz tube are considered to discount the reflection of some of the photons. This procédure is repeated until ail the photons are tracked. Finally, once the photon trajectories are known, the local net spectral radiation flux (qu) is evaluated as is shown in Equation 7.1:  «„  Q& Y/  , «  (7.1)  ~ "lamp,X 2L  nph,tot  R^Rreactor  where Piamp,z is the power of the lamp at 254 nm, nphitrans is the number of photons entering into the reactor volume, and nPh,tot is the total number of photons considered in the MC model. Figure 7.2 shows a représentation of the possible photon paths for the MC method.  *- J  o  © o  o  28o  ©  ©  G ©©  o  @  ©  o  o  © 0  O  0 ©  0 O  UV lamp Quartz tube  UV lamp  Figure 7.2: Schematic représentation of plausible photon paths for the Monte Carlo (MC) Method in the reaction System shown in Figure 4.7.  122  7. Activity of Composite TiQ2 Sphères Table 7.1 shows the expérimental results and the predicted values for the fluence rate using the above mentioned model. At the same time, the effective output of each UV lamp is shown. It is important to mention that the nominal output of the lamp, according to the manufacturer, is 5.7 W (LightSources & LightTech technical sheet, 2005).  Table 7.1: Computed and expérimental fluence rates obtained for the reaction System shown in Figure 4.7. Lamp positions from the Fluence rate (mW cm-2) Predicted UV Actinometer centre of the reactor (cm) Experiment lamps output (W)a Model 4.44 5 4.16 Potassium 3.04 2.83 7.5 4.73 Ferrioxalate 2.14 10 2.04 2.61 5 2.09 lodide 1.50 7.5 2.06 2.80 lodate 1.20 10 1.79 a  According with the manufacturer the output at 254 nm is 5.7 W  Based on thèse results, it is clear that potassium ferrioxalate provided a more realistic prédiction of the nominal output of the lamps. The différence between the effective and the nominal output of the lamps is 17% using potassium ferrioxalate actinometer, while using iodide-iodate actinometer the différence is 51%, the later being high and likely not representing the real output of the lamps. A radiometer (International Light - IL 1700) was used to verify the émission of the lamps at 1 m distance  from  the  lamp,  as  recommended  by  the  manufacturer.  Thèse  measurements showed that the output of each of thèse lamps were similar to the value reported by the manufacturer, with a maximum déviation of 7%. The différences between the effective and the nominal output of the lamps can be explained in terms of the température variations inside the wooden enclosure. When the UV lamps were on, the température inside the box increased due to the lack of fresh air circulating through the System. This led to the increase in the température of the UV lamps, thereby affecting their real output. A change in the ê  UV lamp température would change the vapour pressure of the mercury inside the lamp, changing the equilibrium between the gas and liquid phases that affect the  123  7. Activity of Composite TiQ2 Sphères émission of the lamp (this phenomenon is commonly called "wind chill" effect) (Lau et al., 2009; UVTA technical sheet, 1999). Potassium ferrioxalate was proven to better represent the actual fluence rate within the FBPR. Figure 7.3 shows the corrélation between the expérimental results obtained using potassium ferrioxalate actinometer and the results obtained using the proposed model with the effective output of each UV lamp being 4.73 W.  5.0 -r  „~  4.0-  E u |  S  3.0  a, u c  2.0  u-  1.0 -  V 3  0.0 4  5  6  7  8  9  10  11  Lamp distance from the centre of the reactor (cm) Figure 7.3: Expérimental fluence rate variations with respect to the position of the UV lamps using Potassium Ferrioxalate actinometer and the predicted behaviour using Monte Carlo (MC) model. Error bars represent 95% Cl for the triplicate runs.  Potassium ferrioxalate actinometer was then used to détermine the fluence rate when the cylindrical reflector was placed around the lamps. In this case, ail the lamps were located at 5 cm from the centre of the reactor. A value of 5.32 mW cm"2 was determined using the reflector, which is 26% higher than the value obtained without the reflector (4.16 mW cm"2). As expected, the fluence rate was higher using the reflector, mainly because more photons reached the reactor.  124  7. Activity of Composite TiQ2 Sphères  7.3 Photocatalytic dégradation of Formic Acid (FA) The dégradation of FA was studied under différent irradiation levels (Le., lamps positioned at différent distances from the reactor) as is shown below.  7.3.1 Effect of the radiation flux on FA dégradation The effect of radiation flux was investigated by changing the position of the UV lamps or taking out the reflector (cylinder covered with aluminum foil). The expérimental setup was exactly the same as the one described in Chapter 4. FA concentration was monitored with respect to time to détermine the effect of différent levels of irradiation on the dégradation pathway of the pollutant. Figure 7.4 shows the impact of the différent irradiation levels in the FA dégradation showing that the higher the irradiation, the faster the FA dégradation. Therefore, when the reflector was used, the dégradation was the fastest. Table 7.2 shows the rate constants (kr) of the FA photodegradation under différent fluence rates. In ail the cases pseudo-first order kinetics was assumed. The normalized rate constants were calculated multiplying the apparent rate constant by the ratio of the total volume (1000 mL) to the active volume (80 mL) according to the approach followed by Turchi and Wolfrum (1992). The value of kr was the highest for the highest fluence rate, Le., when the lamps were doser to the reactor and the cylindrical reflector was used (0.317 min"1). At 5 cm from the reactor, but without the reflector, the rate constant dropped by 36% (0.205 min"1). As expected, for the other positions of the lamps, the rate constants were smaller, because rate constant dépends on the amount of irradiation available in the System as was described in Chapter 2. On the other hand, the lower the irradiation, the smaller the amount of electron-hole (e" - h+) pairs that are generated on the surface of the catalyst and, as a conséquence, the smaller the rate of dégradation of the organic pollutants (in this case FA).  125  7. Activity of Composite J\02 Sphères  Figure 7.4: Effect of différent lamp positions on FA photocatalytic dégradation, (o) 5.32 mW cm-2, (0) 4.16 mW cm-2, (•) 3.04 mW cm-2, (A) 2.14 mW cm-2. Error bars represents 95% Cl for the triplicate runs.  Table 7.2: FA photodegradation rate constants (kr) of composite TiÛ2 sphères at différent fluence rates using 100 mg L-1 as initial concentration of FA and 0.025 kg of composite Ti02 sphères Standard error Dégradation rate Lamp position Fluence rate (mW cm-2) of the parameter (cm) constant (min-1) 0.317 0.038 5 (reflector) 5.32 0.048 4.16 0.205 5 (no reflector) 0.014 3.04 0.114 7.5 (no reflector) 0.025 2.14 0.094 10 (no reflector)  7.3.2 Deactivation of the Ti0 2 sphères Deactivation  of the  photocatalyst  is one of the main concerns in  heterogeneous photocatalytic processes. As was discussed before, photocatalytic reactions mainly take place on the surface of the photocatalyst; therefore, adsorption of pollutants is one of the most important steps in photocatalytic reactions. If the  126  7. Activity of Composite TiQ2 Sphères products of the photocatalytic reactions adsorb firmly on the surface of the photocatalyst, the surface charge carrier transportation and pollutant adsorption can be altered, resulting in the loss of photocatalyst activity (Liqiang et al., 2004). One way to détermine this effect is to run the same catalyst more than once and to calculate changes in the dégradation rate as a function of time and use. The dégradation rates of the reuse catalyst can be compared with that of the fresh catalyst to détermine if there was any change in the activity. The deactivation of the composite Ti0 2 sphères, a(t), was calculated using the same approach as shown in Equation 7.2.  •-«•^§5  (72  >  A similar approach has been used elsewhere (Szépe and Levenspiel, 1968; Fogler, 2006). Nonetheless, it is important to clarify that this approach was used hère just to show the effect of FA, or any other possible contaminant, on the deactivation of the catalyst. Figure 7.5 shows the results after 27 h of continuous use, using FA (100 mg L"1) as a pollutant. The level of FA was maintained by adding fresh quantities every 3 h. As can be seen, the activity of the catalyst remained practically constant after 27 h of use. That is, FA dégradation was not affected with repeated use of the catalyst. This implies that no contamination or fouling happened on the CSG photocatalyst and hence there was no inhibition or photocatalytic activity loss.  127  7. Activity of Composite TiQ2 Sphères  O 1.00 -  O  O  O  o  o O  §  0.80  3? £ u «  0.60  O s,  0.40 -  O  O  1  1  24  27  Q.  0.20 0.00 ()  I  3  6  9  12  15  18  21  30  Time (h)  Figure 7.5: Photoactivity of the composite TÏO2 sphères with respect to time using 100mgl_'1ofFA.  7.3.3 Comparison with Degussa P-25 The performance of the composite TiÛ2 sphères at degrading FA was compared with that of particulate Ti0 2 (Degussa P-25). In both cases, the experiments were performed under exactly identical conditions, e.g., initial concentration of FA, irradiation flux, and flow rate. Prior to the comparative investigation, the set ups shown in Figure 4.7 and 4.8 were operated with différent amounts of P-25 to détermine the impact of P-25 loading on FA dégradation. As shown in Figure 7.6, the reaction rate constant changed with P-25 loadings. The rate constant increased with increasing the Ti0 2 loading until a plateau was reached after a concentration of 0.4 g of P-25 in 1 L of water. This indicates that photon absorption starts to be the limiting factor as the amount of P-25 powder goes beyond 0.4 g L'1. With further increase in catalyst concentration (Le., beyond 1 g L"1), FA dégradation rate decreases since suspended particles absorb photons and block the illumination path. At the same time, diffusion and mass transfer limitations could also dominate since the radiation is higher at the 128  7. Activity of Composite T1O2 Sphères wall and most reactions take place in this région. For the purpose of this research and given that P-25 loadings of between 0.4 g L"1 and 1 g L"1 gave similar FA dégradation rates, a loading of 0.75 g L"1 was selected.  T-  C  14.0 -  E F o>  12.0  c ra <n c  8.0  0 0)  6.0  0  >*-»  2  •  10.0  •  4.0 -  c  0 0  re 0) tt  2.0 0.0 E; 0.00  0.20  0.40  0.60  0.80  1.00  1.20  1.40  1.60  Amount of Catalyst (g L'1)  Figure 7.6: Effect of catalyst loading on the reaction rate constant using commercial TÏO2 powder (Degussa P-25) as a photocatalyst: CFAO = 100 mg L"1, Vtot = 1L.  It should be noted that thèse results dépend on the photoreactor geometry, characteristics of UV radiation (power and wavelength), pollutant concentration, and also the kind of photocatalyst used (Minero and Vione, 2006; Muruganandham and Swaminathan, 2006). Chen et al. (2008) found that 3 g L'1 was the most efficient for the dégradation of FA. Dijkstra et al. (2001) reported that a catalyst loading of 2 g L"1 was found to be optimum for FA dégradation. AH those literature studies were performed in set-ups différent from that used in this investigation. A significant réduction in the concentration of FA was observed when the reactor was loaded with Degussa P-25 and treated with UV irradiation (Figure 7.7a). Langmuir-Hinshelwood kinetic model was used to describe the FA dégradation, giving an average reaction rate constant of 11.87 gFA m"3 min"1. Similar to the case with composite Ti0 2 sphères (Figure 7.7b), there was no significant photolytic 129  7. Activity of Composite Ti02 Sphères dégradation when the reactor was operated with the UV lamps on, but without the P25. At the same time, when there was no UV-irradiation, no réduction of FA was observed, showing that there was neither adsorption nor catalytic oxidation in the reaction System. Note that Figure 7.10 présents one set of data from three replicate experiments with the same P-25 loading of 0.75 g L"1.  100 T""  _l D)  80 E ^^ e o 60 (8  l_ +J  C  0)  u c o  40  o < 20 u.  0 0  30  60  90  120  150  180  150  180  Reaction time (min)  a 100 ,*s T-  -1  E 80 c o 60 m !_  "••-*  4-i  C  u 40 c o  o <  70 -  u. 00  b  30  60  90  120  Reaction time (min)  Figure 7.7: Photocatalytic dégradation of Formic acid, (a) commercial Ti0 2 (Degussa P-25): (•) with UV lamps off (0) with UV irradiation; (b) composite Ti0 2 sphères.  130  7. Activity of Composite TiC?2 Sphères For a better comparison between the composite Ti0 2 sphères and the commercial Ti0 2 (Degussa P-25), the relative efficiency (Ç) was determined for the experiments performed with both photocatalysts. This criterion is usefui for comparing quantitatively the efficiency of différent photocatalysts under the same expérimental conditions. Even though the mass of photocatalyst was not the same, the amount of irradiation that was applied to the catalyst and other expérimental conditions were the same. The initial reaction rates were used to compare the two photocatalysts (composite Ti0 2 sphères and Degussa P-25). The relative efficiency was calculated following a similar approach described by Serpone (1996 and 1997) and Shankar et al. (2006). However, FA dégradation using Degussa P-25 was used as the benchmark because P-25 has been known as a standard photocatalyst for many years:  l-r „  \  ) FA,0 /sphères  I-J Q \  #—7—-j  (7.3)  where -rFA,o is the initial reaction rate for the sphères and for the P-25 (shown in Table 7.3) calculated in the first 30 min of reaction.  Table 7.3: Initial reaction rate for the FA dégradation in the présence of composite Ti02 sphères and Degussa P-25. Photocatalyst  Initial reaction rate, -rFA,o (9,FA m"3 min"1)  Composite Ti0 2 sphères Degussa P-25  1.78 OJ54  Using the results shown in Table 7.3, and through Equation 7.3, the relative efficiency (Ç) came to be equal to 2.12. This resuit indicates that the initial oxidative dégradation process for FA is more effective with the composite Ti0 2 sphères than with Degussa P-25. Analyzing the dégradation pattern using the sphères and  131  7. Activity of Composite TiQ2 Sphères Degussa P-25 (Figure 7.10), it is clear that the composite Ti0 2 sphères provides faster dégradation of the FA. Another way to compare the Ti0 2 sphères with P-25, using FA as a model pollutant, is using the apparent quantum yield (Qapp)- One assumption for the calculation of quantum yield was that ail the light that is emitted from the lamp is absorbed by the catalyst. Therefore, using Equation 7.4, the quantum yield was calculated:  Kp^  rFA  >«Yes,p-25-V  (7.4)  o  where -rFA.o is the initial dégradation rate (mol m"3 s"1) per unit of reaction volume V, and l0 the amount of photons entering the reactor (Einsteins s"1). The value of l0 was determined using Potassium Ferrioxalate Actinometry as 1 x 10"4 Einsteins s"1 (with the lamps at 5 cm from the centre of the reactor using the cylindrical reflector). The quantum yield for the sphères was determined to be 0.60 and for the Degussa P-25 was 0.18, which is comparable to the one reported in other studies under similar conditions (McMurray et al., 2004). Thèse results indicated that more molécules of FA were degraded per photon of light absorbed using the composite TiÛ2 than using the commercial TiÛ2. In terms of the characteristics of thèse two photocatalysts, the composite TiÛ2 sphères had smaller surface area (29.37 m 2 g"1) than the commercial TiÛ2 powder (54 m2 g"1), but the pore volume and pore size were not that différent (Table 6.6 - Section 6.3). In terms of the percentage of anatase, P-25 had 87% in comparison with 65% for the TiÛ2 sphères. Based on thèse data, P-25 was expected to provide higher dégradation than the CSG titania. However, it is postulated that light distribution in the reactor might hâve played a factor in bridging higher performance for the CSG catalyst, as is confirmed with the apparent quantum yield value.  132  7. Activity of Composite TiQ2 Sphères  7.4 Photocatalytic dégradation of 2,4-D A pesticide and related products are defined as any substance or mixture of substances, e.g. insecticide, herbicide, fungicide, etc., and any product related to any of thèse including any growth regulators, and their relevant metabolites, dégradation and reaction products (Sinclair et al., 2006). Phenoxyalkanoic acid herbicides are extensively used around the world mainly for the control of annual and perennial weeds. They are also used in lakes, ponds and ditches for aquatic weed control. One of the most frequently used herbicides in this group is 2,4Dichlorophenoxyacetic acid (2,4-D) (Djebbar et al., 2006). This widespread use of 2,4-D leads to certain environmental impacts due to the fact that during its application it could be easily spread within the environment. Once on the ground, it can be incorporated in the natural aqueous stream showing différent half-lifes, depending on factors such as oxygen concentration, acidity, solar light, among others (EPA, 2008). 2,4-Dichlorophenoxyacetic acid dégradation has been studied in aquatic environments by différent methods such as microbial, chemical and photochemical processes with varying success (Bell, 1956; Gréer et al., 1990; Veeh et al., 1996; Ghertner étal., 2001; MacAdam and Parsons, 2009). Photocatalytic dégradation has also been used as a potential technology to dégrade 2,4-D. Différent photocatalysts, such as Zr0 2 , WO x , Mn0 2 , Ti0 2 , Iran (II) Sulfate, among other, hâve been investigated and showed différent results (Kwan and Chu, 2003; Alvarez et al., 2007; Alvarez et al., 2008; Zhang et al., 2008). Among the métal oxide semiconductors suitable for photocatalytic processes, titanium dioxide (Ti0 2 ) has received more attention because of ail the advantages that this photocatalyst has such as its high photocatalytic activity, its low price, etc. Photocatalytic décomposition of 2,4-D has been investigated using différent reactor configurations such as solar reactors (Kamble et al., 2006; Bandala et al., 2007) and with différent catalyst configurations like slurry Systems (Trillas et al., 1995; Djebbar and Sehili, 1998; Kamble étal., 2006; Galindo étal., 2008; Rodriguez - Gonzalez et al., 2008) and supported catalyst on différent surfaces (Trillas et al., 1996; Modestov and Lev, 1998; Shankar et al., 2006; Giri et al., 2008) showing a  133  7. Activity of Composite TiQ2 Sphères variety of results. At the same time, différent mechanisms hâve been proposed to describe the photocatalytic dégradation of 2,4-D; however, it has been reported that the action of O H radicals contribute in about 70% to the conversion of 2,4-D and the other 30% can be attributed to the capture of positive holes by the adsorbed 2,4-D on the surface of the photocatalyst (Djebbar and Sehili, 1998). Figure 7.8 shows the mechanism and pathway of 2,4-D dégradation with its intermediates, including the more prédominant 2,4-Dicholorophenol (2,4-DCP). For the purpose of this work and prior to any photocatalytic oxidation tests, it was necessary to détermine the adsorption of this pollutant on the composite photocatalyst. Hence, a detailed adsorption study was performed first. The photocatalytic dégradation was then performed with three différent irradiation levels (i.e., fluence rates) and three différent initial concentrations of 2,4-D. In addition, a comparison between the composite TIO2 sphères and Degussa P-25 was done to détermine the performance of "ÏÏO2 sphères with respect to the standard photocatalyst.  7.4.1 Adsorption study of 2,4-D on the composite Ti0 2 sphères Adsorption equilibrium studies were conducted under dark conditions using 30 ml_ aliquots of 2,4-D solution (concentrations up to 20 mg L"1) in contact with 1.5 g of TIO2 sphères for about 2 h to allow equilibrium. The ratio of 20 ml_ of 2,4-D solution g"1 catalyst was the same as that used in the photocatalytic reactor. The equilibrium was established when the amount of soluté being adsorbed onto the adsorbent was equal to the amount being desorbed, i.e., there was no further change in the concentration of 2,4-D in the solution.  134  7. Activity of Composite TiQ2 Sphères  o  2,4-D  OH Cl  ^  Cl hv  Ti0 2  OH 2,4-DCP  Cl  Cl OH  OH  1 OH Cl  Cl  Cl OH  lî Cl  Figure 7.8: Mechanism of 2,4-D dégradation - major route (Adapted from Djebbar andSehili, 1998).  135  7. Activity of Composite T1Q2 Sphères Figure 7.9 shows the amount of contaminant adsorbed per mass of catalyst plotted against the concentration of the contaminant in the bulk solution upon the attainment of the equilibrium.  According to the results and in the range of  concentrations studied, the adsorption isotherm follows a linear relation, as shown in Equation 7.5, which is in agreement with literature reports (Allen et al., 2004, Goetz et al., 2009):  Cads  (7.6)  =K C  ' eq  ^,3 g „-1 . which Kbeing equal to 2.1 x 10 6 m  .E+oo O.E+00 5.E-03  1.E-02 2.E-02  2.E-02  3.E-02  3.E-02  4.E-02  4.E-02  Concentration of 2,4 D in the bulk of the solution (mol/m3 of solution)  Figure 7.9: Adsorption isotherm of 2,4-D on the composite Ti0 2 sphères.  Figure 7.10 shows the relation between the above mentioned isotherm (obtained in 30 mL aliquots) and some of the measurements obtained in the fluidized bed photoreactor operating under dark. AH thèse measurements were conducted using différent batches of the composite Ti0 2 sphères. As can be seen, there is a  136  7. Activity of Composite Ti02 Sphères close agreement between the data obtained from the isotherm tests and those obtained during real experiments in the FBR. Nonetheless, it is worth indicating that there were some variations between différent batches of TÏO2 sphères, in terms of the adsorption of 2,4-D on the catalyst, but the overall trend was very similar and consistent as shown in Figure 7.9.  7.E-08 -i  .S ô •0 0) •O tZ 0 (0 T3  £ w *i V) >> re -e  « re  5.E-08 -  2.E-08 1 .E-08  0  «5  £ « O 0  JXf  3.E-08  .2 S  "S o-  t^L  4.E-08 -  entr; lyst  n of 2,4 masso  o£  6.E-08 -  A  Vf'  ••• i O.E+00 i 3"^ O.E+00 5.E-03  <Xs^  Qs^  i  *  1.E-02 2.E-02  i  2.E-02  3.E-02  3.E-02  4.E-02  4.E-02  Concentration of 2,4 D in the bulk of the solution (mol/m 3 of solution)  Figure 7.10: Adsorption isotherm of 2,4-D on the composite TÏO2 sphères, (o) isotherm study, (À) measurements at the FBPR with différent batches of sphères.  The adsorption phenomenon and its impact on the removal of 2,4-D from the contaminated water are further demonstrated in Figure 7.11. The concentration of 2,4-D decreases rapidly upon the start of the experiment, reaching a plateau after about 30 min. No further change in the concentration of 2,4-D demonstrates the establishment of adsorption equilibrium. Hence, ail the photodegradation studies were carried out after 75 min, ensuring that there was no adsorption taking place. When the UV lamps were turned on, the photocatalytic dégradation of 2,4-D and the formation of the main intermediate (2,4-DCP) started.  137  7. Activity of Composite TiQ2 Sphères (0  c o c 0) o c o  10< > Without UV 8;>  ^  O ? 9 •=-  •4  With UV  6 - O  o o  o o  £  4 -  «M •o C  «s o  0 2 -  <* n J  0  A  m  A  m  30  A  M  A  m  60  /  *  ^  . ^  90  A M  MI  120  150  Reaction time (min)  Figure 7.11: 2,4-D (0) dégradation and 2,4-DCP formation (A) using the composite Ti0 2 sphères.  7.4.2 Effect of différent levels of irradiation on 2,4-D dégradation The effect of three différent lamp positions (Le., fluence rates) on the dégradation of 2,4-D was studied. As shown in Figure 7.12, différent levels of irradiation induced a différence in the dégradation pathway. When the UV lamps were placed farthest from the centre of the reactor (i.e., the fluence was the lowest), the dégradation was slower than when the lamps were in doser positions. Therefore, after 30 min of irradiation the conversion of 2,4-D (with respect to the equilibrium concentration at 75 min) was 50% when the fluence rate was 2.14 mW cm"2 and 90% when the fluence rate was 4.16 mW cm"2.  138  7. Activity of Composite Ti0 2 Sphères 12  10 H Without UV  With UV  D)  £ 8 c o  O  1c 6 0 o o c o O  <§> o  O  9  8  d>  o •  Q  4  D  CM"  O -030  60 90 Reaction time (min)  120  150  Figure 7.12: Dégradation of 2,4-D using the composite TiÛ2 sphères under différence fluence rates, (0) 4.16 mW cm"2, (D) 3.04 mW cm"2, (o) 2.14 mW cm"2. The initial concentration of 2,4-D was 10 mg L"1.  The dégradation of 2,4-D followed pseudo-first-order kinetics at low 2,4-D concentration (Figure 7.13), which is in agreement with the literature (Shankar et al., 2006; Terashima et al., 2006; Giri et al., 2007; Giri et al., 2008). Therefore, the influence of différent levels of irradiation on the dégradation rate constant was determined as is shown in Table 7.4. It is observed that the dégradation rate decreased with decreasing the fluence rate or increasing the distance of the lamps from the centre of the reactor.  Table 7.4: 2,4-D photodegradation rate constants (/C2,4-D) of composite Ti0 2 sphères at différent fluence rates using 10 mg L"1 of initial concentration and 25 g of composite Ti02 sphères Standard error of the Fluence rate (mW cm"2) k2,4-D (min"1) parameter 0.736 0.022 4.16 0.555 0.014 3.04 2.14 0.330 0.089  139  7. Activity of Composite TiQ2 Sphères  1.0 H 0.8 o  §  $  <>U  CD  •4  O  0.4  0.2  i  0.0  10  20  30 40 50 Reaction time (min)  60  70  Figure 7.13: Comparison between the 2,4-D photodegradation pathways at différent irradiation levels with fluences of: (0) 4.16 mW cm"2, (a) 3.04 mW cm"2, (o) 2.14 mW cm"2. Error bars represent 95% Cl of the triplicate runs.  The reaction rate constants listed in Table 7.4 were normalized multiplying the apparent rate constant by the ratio of the total volume (500 ml_) to the active volume (80 ml_) according to the approach followed by Turchi and Wolfrum (1992). As shown in Table 7.4, the dégradation rate constant changed with respect to the fluence rate. For high fluence rates (e.g., 4.16 mW cm"2), the reaction rate constant (0.736 min"1) was more than 2 times higher than the reaction rate constant (0.330 min"1) at lower fluence rate (2.14 mW cm"2). This resuit is expected since the reaction rate constant is related to the incident light intensity as was previously discussed in Chapter 2. Once the dégradation of 2,4-D started with the UV lamps turned on, 2,4-DCP (an aromatic intermediate of the 2,4-D dégradation) was detected. As can be seen in Figure 7.14, the maximum amount of this intermediate was found when the fluence rate was 2.14 mW cm"2 (lamps at 10 cm from the centre of the reactor) mainly because with less irradiation, longer time was required to dégrade the generated  140  7. Activity of Composite TiQ2 Sphères 2,4-DCP into other compounds. After 20 min of irradiation, with lamps at 10 cm distance, there was more than 60% more of 2,4-DCP in the System than when the lamps were in the closest position. Other four intermediates (Chlorohydroquinone, 4-Chloropyrocatechol, 2,4-Dichloropyrocatechol and Chlorobenzoquinone) were présent in the system, but in very small quantifies and were not tracked in this investigation. Ail the intermediates, including 2,4-DCP, were removed with further irradiation (e.g., over 80% removal of 2,4-DCP after 90 min).  1 ./LVI  1.00  E. c o  o 0.80 -  '.5  S c Q) O C  o o  o  0.60 -  o O D O  O  â  D O  0.40 -  a  Q.  O  a 4 ci  o o  0.20 -  8  D 0 00 F  —,  — i  20  40  60  80  a 100  Irradiation time (min) Figure 7.14: Formation of 2,4-DCP as an intermediate of the 2,4-D dégradation using the composite TÏO2 sphères at différent fluence rates: (•) 4.16 mW cm"2, (0) 3.04 mW cm"2, (o) 2.14 mW cm"2. The initial concentration of 2,4-D was 10 mg L .  7.4.3 Effect of initial concentration of 2,4-D in its dégradation The influence of initial concentration of 2,4-D on its photodegradation rate was studied for 1, 5 and 10 mg L"1 using 25 g of Ti0 2 sphères (Figure 7.15). The adsorption equilibrium was achieved after 30 min but the system was run under dark  141  7. Activity of Composite TiQ2 Sphères up to 75 min. After this point, the UV irradiation was applied for ail the concentrations. Figure 7.16 shows the effect of the initial concentration on the photocatalytic dégradation of 2,4-D. As shown, the lower the initial concentration of 2,4-D, the faster its dégradation (Table 7.5). In ail the cases, pseudo-first order reaction was found for the différent initial concentrations, with a kinetic constant that increased as the initial reactant concentration decreases.  11 I I  i  > 10 -  i  € 9" O)  fi  c O « *• g  76-  o o  U Q  t  1  Without UV  !  With UV  i  >  i  .i ! i  5E]<> 4  <  o3 (3 -  2  n  CM  1i n ^ u 0  o  o  n  n  A  A i  30  o a A  o  1 i o; i  <> n n ' P o A A À A i  A*  60  90  â  ÎH  A  \/  120  O  D  i  150  r» Li  180  Reaction time (min)  Figure 7.15: 2,4-D photocatalytic dégradation at différent initial concentrations, (0) 10 mg L"1, (n) 5 mg L"1, (A) 1 mg L"1. Fluence rate of 4.16 mW cm"2 (lamps at 5 cm from the centre of the reactor).  142  7. Activity of Composite Ti0 2 Sphères  Figure 7.16: Effect of the différent initial 2,4-D concentration, (À) 1 mg L"\ (a) 5 mg L"1, (0) 10 mg L"1. Fluence rate equal to 4.16 mW cm"2 (lamps at 5 cm from the centre of the reactor).  Table 7.5: Pseudo-first-order kinetic rate constants (/f2,4-Dobs) in photocatalytic dégradation of 2,4-D with différent initial concentration (fluence rate = 4.16 mW cm"2) using 25 g of composite Ti02 sphères Standard error of the 1 2,4-D initial concentration <2,4-Dobs (min" ) parameter (mg L"1) 1 1.383 0.195 5 0.984 0.084 10 0.736 0.022  In pure first order kinetics, it is expected to see a linear relation between the concentration and the reaction rate. In otherwords, the higherthe concentration, the faster the reaction rate. At the same time, the reaction rate constant should be relatively constant no matter which concentration of pollutant is used. However, the expérimental results showed that the reaction rate constant changed with the initial concentration of 2,4-D (the lower the initial concentration, the higher the rate  143  7. Activity of Composite Ti02 Sphères constant). Beltran-Heredia et al. (2001) proposed a model to explain similar results considering that the rate determining step of the catalyzed reaction is the reaction between «OH radicals and organic molécules over the catalyst surface. The proposed hypothesis was based on the theory that TÏO2 surface possesses both acidic and basic sites. The acidic sites are associated with coordinatively unsaturated surface métal ions while the basic sites are associated with surface anions or anion vacancies. Thèse two types of sites can be involved in the adsorption process of the reacting species. One will adsorb the reacting specie (2,4D) and its dégradation products; the other one will be able to adsorb oxygen (Beltran-Heredia et al., 2001). Therefore, the reaction rate for second order surface décomposition might be written in terms of Langmuir-Hinshelwood kinetics as is shown in Equation 7.6:  r = k"eoHe2,D  (7.6)  in which k" is the surface second order rate constant, QOH is the fractional site coverage by hydroxyl radicals and B2A-0 ' s the fraction of sites covered by 2,4-D. Thèse last two variables can be written as follows:  0OH= K ° 2 a P ° 2 ° l + K0P02  -^2,4-J>^2  e2A-D =1 : - T  ^  + ^2,4-zA,4-0 +  (7-7)  „r  zliKJi  (7-8)  where K02, K2A-D and K-, are equilibrium adsorption constants and / refers to the various intermediate products of 2,4-D. Turchi and Ollis (1989) proposed a kinetic simplification that helps to explain results from several researchers investigating dégradation of aromatics compounds. In this simplification it is assumed that the dégradation is a step-wise process (more than one intermediate is involved in the  144  7. Activity of Composite TiQ2 Sphères dégradation pathway of an aromatic compound) where the K\ values of the generated intermediates are fairly similar to the reactant K2J-D value (K-, «  K^-D)-  Thus, simplifying Equation 7.8, the following relation can be obtained:  i +  where  C2,4-DO  - ft -2,4-Z) t -'2,4-£lo  is the initial concentration of 2,4-D, that is equal to the concentration of  2,4-D plus the intermediates présent in the System  (C2,4-DO  «  C2,4-D +  !£)• In addition,  considering that the partial pressure of oxygen is constant during the photocatalytic run, the fractional site coverage by hydroxyl radicals was constant and equal to kc. Then, Equation 7.6 can be written as:  1 +A 2 4 _ £ ) (_- 2 j 4 _ £ ) o  which is a pseudo-first order kinetic équation with respect to the pollutant concentration. In this way, the results listed in Table 7.6 can be explained in terms of the Langmuir-Hinshelwood kinetic model in which the dégradation rate of 2,4-D would be faster for the more dilute solution. In dilute solutions there will be more active sites available for the adsorption of 2,4-D and, at the same time, less amount of dégradation intermediates will compete with the 2,4-D for the active sites of the photocatalyst. The formation of 2,4-DCP at différent initial concentrations is shown in Figure 7.17. As can be seen, the maximum amount of this intermediate was found when the initial concentration was 10 mg L"1. This resuit was expected because the higher the concentration of 2,4-D, the higher the concentration of the intermediates that are formed (for a given fluence rate) along the photocatalytic dégradation of the parent compound.  145  7. Activity of Composite "TIO2 Sphères n 7 -,  ^r  0.6 -  £ c  0.5  g c  0.4  0  °  0  0  n 0  0)  0 0 O 0. O Q  "*  0.3  D  D  0.2 -  A 0.1 - 1  o*  )  (  O A  A  0  1  1  ta  1  20  40  60  80  8  1—1  100  Irradiation time (min)  re 7.17: Formation of 2,4-DCP as an intermediate of the 2,4-D de<arad using the composite Ti0 2 sphères at différent initial concentration of 2,4-D: (0) 10 mg L"1, (D) 5 mg L'1, (A) 1 mg L"1. The lamp position in ail the cases was 5 cm from the centre of the reactor (fluence rate equal to 4.16 mW cm"2).  The pH évolution was monitored during the 2,4-D dégradation with the initial concentration of 5 mg L"1. As shown in Figure 7.18, the pH increased rapidly in a short period of time from 4.70 to 5.10, and then remained stable up to 80 min (which is consistent with the start of the irradiation). Upon the start of irradiation, pH increased again to a value of 5.40 (possibly due to the formation of the intermediates, specially 2,4-DCP) and then remained constant until the end of the experiment. The initial pH and its évolution along the photocatalytic process need to be carefully determined because pH might affect the surface charge on the photocatalyst and the state of the ionization of the substrate, and hence its adsorption (Marczewska and Marczewski, 1997; Kamble et al., 2004). This change in pH might explain some of the variability that was found during the adsorption study as was discussed by Marczewska and Marczewski (1997) and Pettibone et al.  146  7. Activity of Composite TiQ2 Sphères (2008) where différent organic compounds showed différent adsorption behaviours on différent adsorbents, such as TÏO2 nanoparticles, at différent pH.  5.50  E '•v  ro  E~ S  5.40 5.30  </)  5.20  >» c  5.10  0  0  0  0  0  0  0  *-> 0 « 0 ».  5.00 4.90  <D  4.80  c X  4.70  a 4.60  0  20  40  60  80  100  120  140  160  Reaction time (min)  Figure 7.18: pH évolution in the reaction System with 5 mg L"1 of initial 2,4-D concentration.  7.4.4 Deactivation of the Ti0 2 sphères Photocatalytic deactivation was investigated through repeated use of the Ti02 sphères for the dégradation of 2,4-D. Figure 7.19 shows the removal of 2,4-D over three runs. After each run was completed, the reaction System was charged with 2,4D to bring the concentration to the original level. From the results, it is évident that the activity of the catalyst remained the same and the pollutant was removed completely within 60 min of irradiation. The same was observed for the adsorption capacity of the photocatalytic sphères.  147  7. Activity of Composite TiQ2 Sphères  ~ 10 I With UV  Without UV c g  &  +3  «S c a> o  6  *J  Û»  â  â> 6 ft  i ^ O  a •*  â  2  CM  â a 30  60  90  -H120  150  Reaction time (min) Figure 7.19: Dégradation of 2,4-D using 10 mg L"1 as initial concentration reusing the Ti0 2 sphères, (0) 3 h, (o) 8 h, (A) 15 h.  The experiments shown in Figure 7.19 were performed in batch modes, that is, each run was done independently (after the first run, the UV lamps were turned off and the concentration of 2,4-D was re-established). Nonetheless, it was important to perform the same experiment with continuous irradiation, meaning that immediately after the first run, the concentration of 2,4-D was re-established (without turning off the UV lamps). Figure 7.20 shows the results of such experiments with 10 mg L"1 of 2,4-D added to the solution each time when the concentration depleted in the solution. As shown, 2,4-D dégradation did not change with subséquent additions of 2,4-D. At the same time, the 2,4-DCP, as the main intermediate, was formed and degraded in ail of the above mentioned runs.  148  7. Activity of Composite T1O2 Sphères  u< 9 - Without ' UV ; 8-  S. c o +3 2 c d> o c o O •«->  ]—  cj>  0  7,>  e5-O 0  c:  : O  <r*  3  9 f>  4 3-  |o ïo  D  2 -  a  10< »_#.—4—£—.1  30  60  $&—JL 90  1O  B B—£ 120  150  Q B—0 210 180  Reaction time (min) Figure 7.20: Dégradation of 2,4-D using 10 mg L"1 as initial concentration: (0) 2,4-D for the first run; (•) 2,4-DCP génération during the first run; (n)2,4-D for the second run; (•) 2,4-DCP génération during the second run; (o) 2,4-D for the third run; (•)2,4DCP génération during the third run.  7.4.5 Comparison with Degussa P-25 The performance of the composite Ti0 2 sphères at degrading 2,4-D was compared with that of particulate Ti0 2 (Degussa P-25). In both cases, the experiments were conducted  under exactly  identical conditions, e.g. initial  concentration of 2,4-D and same irradiation to the reactor. Similar to the case for FA (described in Section 7.3.4), the P-25 catalyst loading was 0.75 g L"1. Also, this analysis was done just for the lamps at 5 cm from the centre of the reactor and without the cylindrical reflector (fluence rate equal to 4.16 mW cm"2). Other lamp positions were not investigated. Figure 7.21 shows the dégradation of 2,4-D with 0.75 g L"1 using 10 mg L"1 initial concentration of the pollutant. The concentration of 2,4-D decreased upon the start of the experiment, reaching a plateau after 30 min. No further change in the concentration of 2,4-D demonstrate the establishment of adsorption equilibrium. No  149  7. Activity of Composite TiQ2 Sphères significant photolytic dégradation was observed when the reactor was operated with the UV lamps on, but without the P-25.  10<> _l OJ  8  C  o  •-s  6  o c o u ci  15  30  45  60  75  90  105  120  135  150  Reaction time (min) Figure 7.21: Photocatalytic dégradation of 10 mg L~?r of initial concentration of 2 ,4-D (fluence rate equal to 4.16 mW cm"2).  As can be seen in Figure 7.21, the dégradation follows a pseudo-first-order kinetics with an average reaction rate constant of 0.09 min"1. Figure 7.22 shows the photocatalytic dégradation of 2,4-D using the commercial Ti0 2 and the composite Ti0 2 sphères. As shown, the sphères degraded the 2,4-D faster than the other catalyst, i.e., after 15 min of irradiation there was 50% less 2,4-D using the sphères than with Degussa P-25. For a better comparison between the composite Ti0 2 sphères and the commercial Tï0 2 powder (Degussa P-25), the relative efficiency (Ç) was determined for the experiments performed with both catalysts. Even though the mass of the catalyst was not the same, the amount of irradiation that was applied to the catalysts and other expérimental conditions were the same.  150  7. Activity of Composite TiQ2 Sphères  Figure 7.22: Photocatalytic dégradation of 2,4-D using Degussa (A) and the composite TÏO2 sphères (n) with an initial concentration of 10 mg L"1. Lamps position at 5 cm from the centre of the reactor (fluence rate equal to 4.16 mW cm"2).  The relative efficiency is defined as the ratio of the initial rates of the pollutants for the two photocatalysts as shown in Equation 7.11. In this équation, -r2,4-D,o is the initial reaction rate for the sphères and for P-25 (shown in Table 7.6), calculated in the first 15 min of reaction.  \  r  2,4-D,0  ) sphères ,  (7.11)  P-25  Table 7.6: Apparent reaction rate constant for the 2,4-D dégradation in the présence of composite Ti02 sphères and Degussa P-25 Initial reaction rate, -fy-DX 10" Photocatalyst (mg2,4-D L"1 min"1) Composite TiÛ2 sphères 5.58 Degussa P-25 3.91  151  7. Activity of Composite TiQ2 Sphères Using the results shown in Table 7.6, and evaluating Equation 7.11, the relative efficiency (Ç) was equal to 1.26. This indicates that the oxidative dégradation process for 2,4-D is more effective with the composite TiC>2 sphères than with Degussa P-25. Analyzing the dégradation pattern using the sphères and Degussa P25 (Figure 7.22), it is also clear that the composite Ti0 2 sphères degraded the pollutant (2,4-D) faster.  7.5 Photocatalytic dégradation of natural organic matter (NOM) Natural organic matter (NOM) is defined as a complex mixture of organic compounds présent in surface water. The structure and chemical composition of NOM is not well understood due to its complexity and variability with respect to season and location (Bursill et al., 2002). However, it is known that from the chemical point of view, NOM is grouped into non-humic substances and humic substances. The humic substances are non-polar organic acids derived from soil humus and terrestrial aquatic plants and are subdivided into humic acids (that precipitate out of solution at pH<2) and fulvic acids (thèse do not precipitate) (Thurman, 1985). NOM brings several challenges to drinking water treatment opérations mainly because of its contribution to the formation of disinfection by-products (DBPs) (Nikolaou and Lekkas, 2001). It may also induce subséquent détérioration of water quality due to bacterial re-growth in distribution Systems (Zhang and DiGiano, 2002). Therefore, différent efforts hâve been made to eliminate or minimize the level of NOM in source water before treatment. Advanced oxidation processes hâve been extensively studied for the removal of NOM from natural water, finding that NOM could be mineralized leading to a réduction of DBPs formation (Speitel étal., 2000). This section présents the results of some preliminary tests with the composite Ti0 2 sphères to détermine the feasibility of using photocatalysis to reduce the amount of NOM in natural water. Water from Trepanier Creek in the Peachland area in Central British Columbia (initial TOC of approximately 5 mg L"1) was treated in the  152  7. Activity of Composite TiQ2 Sphères fluidized bed reactor (the same setup used for the dégradation of FA and 2,4-D). As shown in Figure 7.23, 40% of the initial TOC was eliminated when the FBPR was run just with the composite photocatalyst (no UV). When the reactor was running with the catalyst and UV (photocatalytic process) there was also approximately 40% of réduction in the TOC. Therefore, ail the TOC réduction that was detected during the photocatalytic process was probably due to the adsorption of NOM in the composite Ti02 sphères (Figure 7.23). No significant photocatalytic dégradation was observed when the reactor was operated with the UV lamps on, but without the composite Ti0 2 .  Figure 7.23: Dégradation of "Peachland water" (initial TOC ~ 5 mg L"1) using the composite Ti0 2 sphères, (o) using only UV (fluence rate of 4.16 mW cm"2) without the Ti0 2 sphères, (n) with the Ti0 2 sphères without UV light, (A) with Ti0 2 sphères and UV (4.16 mW cm"2). Error bars represent 95% Cl of triplicate runs.  An adsorption study was done using the composite Ti0 2 sphères and natural organic matter. As can be seen in Figure 7.24, after 60 min approximately 40% of  153  7. Activity of Composite TIO2 Sphères the NOM was adsorbed on the sphères. After that point, the UV lamps were turned on to détermine if there was some photocatalytic dégradation. As shown, the concentration of NOM (or TOC) did not decrease any further after the UV lamps were turned on. Based on thèse preliminary results, can be concluded that under the expérimental conditions used in this investigation, there is not significant photocatalytic dégradation of Peachland water using CSG titania sphères. Further studies should be done to better understand the mechanism and variables involved during NOM photocatalytic dégradation.  1.0(3  Without UV  With UV  0.8  *î *  o O  0.6 -  1  -T  O  o H  0.4 0.2 -  0.0  20  40  60  80  100  120  140  Reaction time (min)  Figure 7.24: Adsorption of "Peachland water" (initial TOC ~ 5 mg L"1) on the composite Ti02 sphères. Error bars represent 95% Cl of duplicate runs.  7.5.1 Effect of NOM on the dégradation of 2,4-D In order to better emulate real applications of photocatalytic processes, 2,4-D dégradation was studied in the présence of NOM. The effect of NOM présence on  154  7. Activity of Composite TiQ2 Sphères the dégradation of 2,4-D is shown in Figure 7.25. For this study, 2,4-D was added to 500 ml_ of Peachiand water, bringing the concentration of the pollutant to 10 mg L"1. As can be seen, the dégradation of 2,4-D was affected significantly and was slower than the dégradation of the same concentration of 2,4-D without NOM. After 30 min of irradiation, 40% of 2,4-D was still in the solution when NOM was présent as the matrix, while without NOM ail the 2,4-D was degraded within the same time. The présence of NOM, not only hindered the dégradation of 2,4-D, but also reduced its adsorption on the CSG catalyst. As was shown in Figures 7.23 and 7.24 approximately 40% of NOM was adsorbed on the TÏO2 sphères; hence, this réduction in 2,4-D adsorption could be due to higher affinity of the NOM to be adsorbed on the catalyst. This clearly demonstrate the compétitive effect of NOM for occupying the adsorption sites and also scavenging the «OH radicals formed during the irradiation process.  1 1 1 1  e  ~ 10 T  (  b  E c  8  o  0  0  0  0  c^ ! 0 : 0 ; 0  n  n  n  •  li  13  0  •*->  <0  _.  6 -  c  0) 0  §  n O  4-  O Q "*-  «M  With UV  D D 0  2-  Without UV  O O  D D  n  n 100  M  0  20  40  60  80  n 120  C) rb  Reaction time (min) Figure 7.25: Effect of the "Peachiand water" on the 2,4-D dégradation using the composite Ti0 2 sphères, (0) 2,4-D mix with Peahland water, (D) 2,4-D mixed with Milli-Q water.  155  7. Activity of Composite Ti02 Sphères In terms of the deactivation of the catalyst, the same batch of catalyst was run for 6 h as shown in Figure 7.26. It is shown that the activity of the catalyst was decreased after every run, indicating the conséquence and impact of the NOM on the activity of the catalyst. After thèse three runs, the colour of the catalyst changed to light brown (originally the sphères are white); this could indicate contamination on the surface of the catalyst diminishing its activity.  1.0  ià  *  *  *  é \  0.8  S N  o"  D 06  •  5 o  A  0.4  A  • 0.2  0.0 20  40  60  80  100  120  Reaction time (min)  Figure 7.26: 2,4-D dégradation using Peachland water as the matrix, (•) fresh batch of Ti0 2 sphères, (D) reusing the same batch of sphères the first time, (A) reusing the batch of sphères the second time.  7.6 Final remarks The fluence rate within the reactor was determined to be 5.32 mW cm"2 when the UV lamps were located at 5 cm from the centre of the reactor and when the cylindrical reflector was used. This fluence rate decreased when the cylindrical  156  7. Activity of Composite TiQ2 Sphères reflector was removed (4.16 mW cm"2) and when the lamps were moved away from the reactor. FA was used as a model organic compound to test the activity of the composite TIO2 sphères. Using a 100 mg L"1 as initial concentration of FA, the dégradation (showed a first order kinetics) was complète after 105 min of irradiation applying a fluence rate of 5.32 mW cm"2. The effect of différent irradiation levels was studied  showing  that  the  dégradation  rate  constant  changed  from  0.317 ± 0.038 min"1 (at 5.32 mW cm"2) to 0.094 ± 0.025 min"1 (at 2.14 mW cm"2). No photocatalyst deactivation was found after 27 h of use. The herbicide 2,4-D showed a significant adsorption on the composite TiÛ2 sphères, reaching an adsorption equilibrium after 75 min. The effect of différent irradiation levels and différent initial concentrations were studied. At différent fluence rates, the dégradation rate constant (following pseudo-first order kinetics) changed from 0.736 ± 0.022 min"1 (at 4.16 mW cm"2) to 0.330 ± 0.089 min"1 (at 2.14 mWcm"2). At différent initial concentrations, pseudo-first order reaction was found with a kinetic constant that increased as the initial reactant concentration decreased, which can be explained using Langmuir-Hinshelwood kinetic model. No photocatalyst deactivation was found after 6 h of continuous use. Degussa P-25 (0.75 g L"1) was used to compare the photoactivity of the composite TiÛ2 sphères using the above mentioned organic compounds. In both cases the T1O2 sphères showed higher activity than the commercial TiÛ2. Preliminary tests with NOM ("Peachland water" with an initial TOC of ~5 mg L"1) suggested that using the sphères there was no photocatalytic dégradation. On the other hand, ~40% of adsorption was found on the CSG titania. In addition, 2,4-D was mixed with NOM to détermine the effect of NOM in the pesticide dégradation, showing that the NOM hindered the dégradation of 2,4-D and diminished the CSG sphères activity after consécutive runs.  157  Chapter Eight Conclusions and Recommendations 8.1 Conclusions This research investigated the attrition of the composite TÏO2 sphères in FBPR and how attrition can be controlled and reduced by adjusting some of the conditions in the production process of the TÏO2 sphères. In addition, the activity of the sphères to dégrade contaminants in water was tested using two model contaminants, formic acid (FA) and 2,4-Dichlorophenoacetic acid (2,4-D). Also, preliminary tests with natural organic matter (in water from Trapanier Creek in Central British Columbia, Canada) were performed. Based on the expérimental work carried out in this study, the following conclusions hâve been made: 1. The template-free composite Ti02 sphères were produced by mixing a composite sol-gel (CSG) titania with a viscous acidic solution of chitosan. The CSG material was a sol-derived TiÛ2 with fine particles (commercial TiÛ2 powder) dispersed in the liquid by high shear mixing. The commercial powder (Degussa P-25) acted as filler material providing higher efficiency of the photocatalyst. The chitosan acidic solution was used to produce the TiÛ2 sphères because of its desired viscosity (needed for the sphères formation), and because of its fast hardening when in contact with basic solutions. This composite TiÛ2 sphères possessed consistency in their shape and size (with a narrow size distribution around 1.15 mm at 600°C) and relatively smooth surface. 2. Modifying the composite TiÛ2 sphères formulation and some of the synthesis procédures, it was possible to reduce the attrition effect of the photocatalyst. Hydrolysis and condensation reactions during the sol-gel production, the pH différence between the chitosan-Ti02 CSG solution and the ammonia (NH4OH) 158  8. Conclusions and Recommandations solution, and the drying and heat treatment conditions contributed greatly to a stronger structure of photocatalyst with higher attrition résistance. The amount of TiC>2 particles in suspension were reduced from 22 mg L'1 to 7 mg L"1 (70% réduction) after 3 h of continuous opération in a FBPR based on the following changes in the procédure/formulation: i) inducing more hydrolysis and condensation reactions by increasing the amount of water from 0.040 g H20 to 0.053 g H20 per ml_ TTIP and reducing the amount of HCI from 0.20 mL HCI to 0.13 mL HCI per mL TTIP during the sol-gel production, ii) increasing the pH of the ammonia solution from 11.75 to 12.0 (increasing the percentage of ammonia from 10% to 20% v/v), iii) accelerating the drying process conditions to 80°C for 20 h, and iv) increasing the calcination time from 1 h to 3 h at 600°C. 3. During the sol-gel préparation of the Ti0 2 sphères, the amount of water and catalyst (HCI) used during the gelation stage defined the rates of hydrolysis and condensation reactions in the System. With an incrément in the water content and with a réduction in the amount of HCI, the viscosity of the sol increased from 4.83 cP in the original formulation to 7.94 cP. This incrément in the viscosity showed that the condensation reactions during the sol-gel préparation were accelerated. Thèse condensation reactions continued occurring even during the drying process of the sphères inducing a stiffer network because of the formation of more Ti-O-Ti bonds. Therefore, an accelerated drying process induced stronger sphères structure. At the same time, an increase in the calcination température of the sphères reduced the attrition effect of the photocatalyst, i.e., the amount of Ti0 2 particles released from the catalyst to the water. 4. The calcination température had an enormous impact on catalyst characteristics such as the percentage of anatase and surface area. At 600°C, the surface area of the sphères was 29.37 m2 g"1 and at 700°C, the surface area dropped by 78% (6.44 m2 g"1). The percentage of anatase changed from 64.7% at 600°C to 0% at 700°C, demonstrating that phase transformation from anatase to rutile was accelerated by increasing the température of heat treatment. Even though high anatase/rutile ratio had proved to be bénéficiai for the activity of the 159  8. Conclusions and Recommandations photocatalyst, it was found that surface area had also a significant rôle in photocatalyst activity, i.e., at 800°C and with 100% of rutile, the composite TIO2 sphères showed little dégradation of FA (0.025 ± 0.013 min"1) and possessed a surface area of 3.79 m 2 g"1. On the other hand, at 900°C (100% rutile) and no surface area, no FA was degraded. In photocatalytic reactions, larger surface area leads to faster reaction of e" - h + with substrate because more substrate molécules are available to participate in the reaction. 5. The production process of the TiÛ2 sphères using the new/modified formulation (MF) was more efficient (i.e., time devoted to obtain the final product was less) than that with the old formulation (OF). Using the original formulation (OF), approximately 20 days were required to produce one batch of sphères, whereas with the MF the same batch could be produced in three days. This réduction in processing time gives an added advantage to the composite Ti0 2 sphères for large-scale applications. 6. Sphères at 600°C showed high activities for the dégradation of FA (100 mg L"1). Within 105 min of irradiation, FA was degraded in the System (with a fluence rate of 5.32 mW cm"2) showing first-order reaction kinetics with a dégradation rate constant of 0.315 ± 0.061 min"1. At higher calcination températures, the dégradation rate constant significantly dropped, i.e., at 800°C the dégradation rate constant was 0.025 ± 0.013 min"1, mainly because there was no anatase in the catalyst and the surface area was 90% smaller than that at 600°C. On the other hand, at lower fluence rates, the dégradation rate constant for the FA also decreased, i.e., at 2.14 mW cm"2 the dégradation rate constant was 0.094 ± 0.025 min"1, confirming that the reaction rate constant dépend on the irradiation flux applied to the System. There was no statistical différence between the photocatalytic activity of the sphères produced by the original formulation and by the new formulation. 7. 2,4-D (10 mg L"1) was degraded with composite Ti0 2 sphères showing pseudo first-order kinetics (reaction rate constant equal to 0.736 ± 0.022 min"1) when the  160  8. Conclusions and Recommendations fluence rate was 4.16 mW cm"2. At lower fluence rates, the dégradation rate constant significantly decreased, Le., at 2.14 mW cm"2 the dégradation rate constant was reduced by 45% (0.330 ± 0.089 min"1). At différent initial concentrations of the pesticide, pseudo-first order reaction was found with a kinetic rate constant that increased as the initial reactant concentration decreased. With an initial concentration of 1 mg L"1 the dégradation rate constant was 1.383 ±0.195 min"1 and at 5 mg L"1 the dégradation rate constant was 0.984 ± 0.084 min"1. This effect can be explained based on the hypothesis that in dilute solutions there will be more active sites available for the adsorption of 2,4-D and, at the same time, less amount of dégradation intermediates will compete with the 2,4-D for the active sites of the photocatalyst. 8. Reusing the composite TÏO2 sphères more than once, it was found that using FA (100 mg L"1) and 2,4-D (at 1, 5 and 10 mg L"1), the sphères showed the same photocatalytic activity. There was no fouling or contamination of the Ti0 2 sphères that led to inhibition or photocatalytic activity loss. On the other hand, when using NOM as matrix for the 2,4-D, it was found that the activity changed and reduced over time. After the third run (6 h of use) with the same batch of sphères, the dégradation of 2,4-D was diminished significantly. While 83% of the 2,4-D was degraded using the fresh catalyst and after 45 min of irradiation, only 60% of 2,4-D was eliminated after once the same catalyst was used for the third consécutive run. 9. The relative efficiency parameter (Ç) was used to compare the activity of the composite Ti0 2 sphères with the commercial Ti02 photocatalyst (Degussa P-25 with a concentration of 0.75 g L"1). Ti0 2 sphères showed higher activity than the P-25 (Ç = 2.12) when 100 mg L"1 of FA was used as a model organic compound. In addition, the apparent quantum yield (<S>app) was calculated for the composite Ti0 2 sphères and for the commercial Ti0 2 powder (Degussa P-25). For the sphères, a $>app = 0.60 was determined in contrast with the Qapp = 0.21 for the P25, confirming that the sphères were more efficient to dégrade FA in the FBPR. For the dégradation of 2,4-D, the relative efficiency was Ç = 1.26 suggesting that 161  8. Conclusions and Recommandations with more complex molécules (Le., pesticides) the composite photocatalyst showed slightly higher activity. Even though the composite TÏO2 sphères had a comparable high percentage of anatase and surface area (64% and 29 m2 g"1, respectively) with commercial Ti0 2 (88% and 54 m 2 g"1), the différence in activity could be due to the différences in terms of the light distribution in the System. lO.Using composite TÏO2 sphères to dégrade NOM in raw drinking water (Peachland water), it was found that the sphères had a strong adsorption capacity. However, no significant photocatalytic degradation/mineralization of NOM was observed when the reactor was operated. In the présence of NOM, the dégradation of 2,4-D was hindered and its adsorption on the CSG catalyst was reduced. 11. Potassium ferrioxalate actinometer better represented the fluence rate within the FBPR. lodide-lodate actinometer, on the other hand, showed a "saturation effect" at différent lamp positions, meaning that the fluence rate slightly changed at différent lamp positions around the reactor. This "saturation effect", as it suggests, was likely due to the saturation of the actinometric solution at very high photon flux that prevented proper differentiation among différent levels of irradiation. The effective output of the UV lamps was calculated based on the fluence rate values of each actinometry showing a value of 4.73 W using potassium ferrioxalate and 2.80 W with iodide-iodate. Comparing thèse values with the nominal output of the lamps (5.7 W - given by the manufacturer) it was concluded  that  potassium  ferrioxalate  better  represented  the  radiation  characteristics within the FBPR.  8.2 Recommendations for future work Although this investigation has shown that the modifications performed in the formulation and synthesis of the composite Ti02 sphères provided better attrition résistance to the photocatalyst, further studies would be needed to further improve the characteristics  and  performance  of this  particular  photocatalyst.  It is 162  8. Conclusions and Recommendations recommended that further investigations be carried out to quantify any additional changes the formulation and/or préparation process may offer to the mechanical stability and photoefficiency of the photocatalyst. The following represents some of the spécifie recommendations for future research on the composite template-free TIO2 sphères:  1. Study in more détail the effect of hydrolysis and condensation reactions in the attrition résistance of the composite TÏO2 sphères. As was proven in this investigation, the hydrolysis and condensation reactions helped produce a tighter network, improving the attrition résistance of the photocatalyst. Therefore, further studies in the procédure to produce the CSG could lead to a stronger Ti0 2 sphères structure. 2. Explore différent drying process conditions (température and time) to understand the real effect of thèse conditions on the attrition résistance of the photocatalyst. 3. Investigate the effect of lower calcination températures (i.e., 500°C) on the activity and attrition résistance of the composite photocatalyst produced with the new formulation (MF). It is expected that the activity of the photocatalyst would increase at lower calcination températures because the surface area and the anatase/rutile ratio would be higher than those at 600°C; nonetheless, the attrition résistance of the sphères might be reduced. 4. Incorporate traces of noble metals (or other materials) to the photocatalyst formulation in order to further improve the activity of the Ti0 2 sphères. It is well known that adding materials such as silver, platinum or manganèse help to prevent electron-hole recombination, increasing the activity of the photocatalyst. However, the production process of the sphères needs to be carefully assessed to identify the best stage to add thèse materials. At the same time, the characteristics of the CSG photocatalyst need to be evaluated to détermine the impact of thèse materials on the properties of the sphères (e.g., attrition résistance)  163  8. Conclusions and Recommandations 5. Optimize the production process of the composite T1O2 sphères in order to produce larger amounts of sphères in a more effective way, through the design of equipment with greater throughput. 6. Perform further studies with the composite sphères in the fluidization process to détermine the best reactor design and operational conditions. 7. Continue the study with NOM and NOM with 2,4-D to understand better the true photocatalytic mechanism and the performance of composite TÏO2 sphères in setting that include natural water Systems. Thèse kinds of assessments are important because they will give a better understanding of the feasibility of photocatalytic process using the CSG photocatalyst under more realistic conditions. 8. Study the performance of the composite Ti02 sphères (in terms of photoactivity and attrition résistance) in larger scale fluidized bed reactors (e.g., pilot scale reactors). 9. Investigate the deactivation of the photocatalyst (e.g., in the case of NOM and NOM with 2,4-D) and détermine possible mechanism to regenerate the photocatalyst particles.  164  BIBLIOGRAPHY Allen, S.J., G. Mckay and J.F. Porter. (2004). Adsorption isotherm models for basic dye adsorption by peat in single and binary component Systems. J. Colloid Interf. Sci., 280, 322-333. Alvarez, M., T. Lôpez, J.A. Odriozola, M.A. Centeno, M.l. Domînguez, M. Montes, P. Quintana, D.H. Aguilar and R.D. Gonzalez. (2007). 2,4-Dichlorophenoacetic acid (2,4-D) photodegradation using an Mn+/Zr02 photocatalyst: XPS, UV-vis, XRD characterization. Appl. Catal. B - Environ., 73 (1-2), 34-41. Alvarez, M., T. Lôpez, S. Recillas, D.M. Frîas, M. Montes, J.J. Delgado, M.A. Centeno and J.A. Odriozola. (2008). Photocatalytic dégradation of 2,4dichlorophenoxyacetic acid using nanocrystalline cryptomelane composite catalyst. J. Mol. Catal. A-Chem., 281,107-112. Ambrus, Z., K. Mogyorôsi, A. Szalai, T. Alapi, K. Demeter, A. Dombi and P. Sipos. (2008). Low température synthesis, characterization and substrate-dependent photocatalytic activity of nanocrystalline TÏO2 with tailor-made rutile to anatase ratio. Appl. Catal. A - G e n . , 340,153-161. Aprile, C , A. Corma and H. Garcia. (2008). Enhancement of the photocatalytic activity of Ti0 2 through spatial structuring and particle size control: from subnanometric to submillimetric length scale. Phys. Chem. Chem. Phys., 10, 769-783. Araha, J., A. Pena Alonso, J.M. Dona Rodrîguez, G. Colon, J.A. Navio and J. Pérez Peha. (2009). FTIR study of photocatalytic dégradation of 2-propanol in gas phase with différent Ti0 2 catalyst. Appl. Catal. B - Environ., 89, 204-213. Arguelles-Monal, W., F.M. Goycoolea, C. Péniche and I. Higuera-Ciapara. (1998). Rheological study of the chitosan/glutaraldehyde chemical gel system. Polym Gels Netw., 6, 429-440. Arters, D.C. and L. Fan. (1986). Solid-liquid mass transfer in a gas-liquid-solid fluidized Bed. Chem. Eng. Sci., 41(1), 107-115. Ayers, M. and A. J. Hunt. (2001). Synthesis and properties of chitosan-silica hybrid aerogels. J. Non-cryst solids, 285,123-127. Bandala, E.R., M.A. Pelâez, D.D. Dionysiou, S. Gelover, J. Garcia and D. Macias. (2007). Dégradation of 2,4-dichlorophenoacetic acid (2,4-D) using cobalt-  165  peroxymonosulfate in Fenton-like process. J. Photoch. Photobio. A., 186 (2-3), 357-363. Barrett, E., L.G. Joyner and P.P. Halenda. (1951). The détermination of pore volume and area distributions in porous substances: computations from nitrogen isotherms. J. Am. Chem. Soc, 73(1), 373-380. Belapurkar, A.D., P. Sherkhane and S.P. Kale. (2006). Disinfection of drinking water using photocatalytic technique. Curr. Sci. India, 91 (1), 73-76. Bell, G.R. (1956). On the photochemical dégradation of 2,4-Dichlorophenoacetic acid and structurally related compounds in the présence and absence of riboflavin. Botanical Gazzette, 118(2), 132-136. Beltran-Heredia, J., J. Torregrosa, J.R. Dominguez and J.A. Pères. (2001). Oxidation of p-hydroxybenzoic acid by UV radiation and by Ti02/UV radiation: comparison and modeling of reaction kinetics. J. Hazard. Mater., B83, 255-264. Benenati, R.F. and C.B. Brosilow. (1962). Void fraction distribution in beds of sphères. A.I.CH.E. Journal, 8(3), 359-361. Bessenkhouad, Y., D. Robert and J.V. Weber. (2003). Synthesis of photocatalytic TiC>2 nanoparticles: optimization of the préparation conditions. J. Photoch. Photobio. A., 157,47-53. Bhatkhande, D.S., V.G. Pangarkar and A.A.C.M. Beenackers. (2002). Photocatalytic dégradation for environmental applications - a review. J. Chem. Technol. Biot, 77(1), 102-116. Bideau, M., B. Claudel, C. Dubien, L. Faure, and H. Kozouan. On the "immobilization" of titanium dioxide in the photocatalytic oxidation of spent waters. J. Photoch. Photobio. A., 91,137-144. Bischoff, B. and Anderson, M.A. (1995). Peptization process in the sol-gel préparation of porous anatase (Ti02). Chem. Mater., 7 (10), 1772-1778. Bojinova, A., R. Kralchevska, I. Poulios and C. Dushkin. (2007). Anatase/rutile TIO2 composites: Influence of the mixing ratio on the photocatalytic dégradation of malachite green and orange II in slurry. Mater. Chem. Phys., 106,187-192. Bouchy, M., and O. Zahraa. (2003). Photocatalytic reactors. Int. J. Photoenergy, 05, 191-197. Box, G., W.G. Hunter and J.S. Hunter. Statistics for experimenters: An introduction to design, data analysis and model building. John Wiley & Sons, Inc., USA.  166  Boyd, G.R., H. Reemtsma, D.A. Grimm and S. Mitra. (2003). Pharmaceutical and Personal care products (PPCPs) in surface and treated waters of Louisiana, USA and Ontario, Canada. Sci. Total. Environ., 311,135-149. Bozzola, J.J. and L.D. Russell. (1998). Electron microscopy. Jones and Bartlett Publishers, Second Edition, Canada. Brinker, C.J., and G. Scherer. (1990). Sol-Gel Science: The physics and chemistry of Sol-Gel processing. Académie Press, USA. Brinker, C.J., D.E. Clark and D.R. Ulrich. (1984). Better ceramics through chemistry. North-Holland. Brinker, C.J., D.E. Clark and D.R. Ulrich. (1986). Better ceramics through chemistry II. Materials Research Society. Brinker, C.J., D.E. Clark and D.R. Ulrich. (1988). Better ceramics through chemistry III. Materials Research Society. Brunauer, S., P.H. Emmet and E. Teller. (1938). Adsorption of Gases in Multimolecular Layers. J. Am. Chem. Soc, 60, 309-319. Bursill, D., J. Van Leeuwen and M. Drikas. (2002). Problems related to particulate and dissolved components in water: the importance of organic matter. Water Sci. Technol.,2(1), 155-162. Campanati, M., G. Fornasari and A. Vaccari. (2003). Fundamentals in the préparation of heterogeneous catalysts. Catal. Today, 77, 299-314. Catalkaya, E.C. and F. Kargi. (2008). Advanced oxidation treatment of pulp mill effluent for TOC and toxicity removals. J. Environ. Manage., 87, 236-404. Chao H.E., Y.U. Yun, H.U. Xingfang and A. Larbot. (2003). Effect of silver doping on the phase transformation and grain growth of the sol-gel titania powder. Journal of European Ceramic Socierty, 23,1457-1464. Chariot, G. (1964). Colorimetric Détermination of Eléments: Principles and Methods. Elsevier Pub. Co., N Y - USA. Chen, Q., J. Chang, L. Li and J.Y. Yuan. (2008). A new kinetic model of photocatalytic dégradation of formic acid in UV/Ti02 suspension System with insitu monitoring. React. Kinet. Catal. L., 93 (1), 157-164. Chen, W., J. Zhang, Q. Fang, S. Li, J. Wu, F. Li and K. Jiang. (2004). Sol-gel préparation of thick titania coatings aided by organic binder materials. Sensor Actuator, B 100, 195-199.  167  Choi, J., J. Ban, S. Choung, J. Kim, H. Abimanyu and K. Yoo. (2007). Sol-gel synthesis, characterization and photocatalytic activity of mesoporous Ti02/yAl 2 0 3 granules. J. Sol-Gel Sci. Techn., 44, 21-28. Choi, H., E. Stathatos and D. Dionysiou. (2006). Sol-gel préparation of mesoporous photocatalytic Ti0 2 films and Ti0 2 /Al 2 0 3 composite membranes for environmental applications. Appl. Catal. B - Environ., 63, 60-67. Cullo, L., Jr. Nagle, V. Elliot, Jr. Restelli, F. Edward and T. Yogan. (1987). Method to making anatase-titania attrition résistant catalyst composition. Aristech Chemical Corporation. US Patent 4 705 770. Cumming, G., F. Fidler and D. L. Vaux. (2007). Error bars in expérimental biology. J. Cell.Biol., 177(1), 7-11. De Lasa, H., B. Serrano and M. Salaius. (2005). Photocatalytic reaction engineering. Springer, USA. DeBoer, J.H., B.C. Lippens, B.G. Lippens, J.C.P. Broekhoff, A. Van den Heuvel and Th. V. Osinga. (1966). T-curve of multimolecular N2-Adsorption. J. Colloid. Interf. Sci., 21(4), 405-414. Dijkstra, M.F.J., A. Michorius, H. Buwalda, H.J. Panneman, J.G.M. Winkelman and A.A.C.M. Beenackers. (2001). Comparison of the efficiency of immobilized and suspended Systems in photocatalytic dégradation. Catal. Today, 66, 487-494. Djebbar, K., and T. Sehili. (1998). Kinetics of heterogeneous photocatalytic décomposition of 2,4-Dichlorophenoxyacetic acid over titanium dioxide and zinc oxide in aqueous solution. Pestic. Sci., 54, 269-276. Djebbar, K., and T. Sehili. (2006). Photocatalytic dégradation of 2,4Dichlorophenoxyacetic acid and 4-Chloro-2-Methylphenoxyacetic acid in water byusingTi0 2 . Environ. Technol., 27,1191-1197. Dong, S., D. Zhou and X. Bi. (2008). Effect of liquid and gas flow rates on the performance of a fluidized bed photocatalytic reactor. Int. J. Chem. React. Eng., 6.A100. Dunuwila, D.D., C D . Gagliardi and K.A. Berglud. (1994). Application of controlled hydrolysis of titanium (IV) Isopropoxide to produce sol-gel-derivate thin films. Chem. Mater., 6, 1556-1562. Eisenberg, G. (1943). Colorimetric détermination of hydrogen peroxide. Ind. Eng. Chem., 15(5), 327-328.  168  Emeline, A., V. Ryabachuk and N. Serpone. (2000). Factors affecting the efficiency of photocatalyzed process in aqueous metal-oxide dispersion: Prospect of distinguising between two kinetic models. J. Photoch. Photobio. A., 133, 89-98. Emeline, A.V., V.K. Ryabchuk and N. Serpone. (2005). Dogmas and misconceptions in heterogeneous photocatalysis. Some Enlightened Reflections. J. Phys. Chem. B., 109, 18515-18521. EPA (2008). Pesticides: 2,4-D Red Facts. Retrieved on 03/14/2009 from http://www.epa.gov/oppsrrd1/REDs/factsheets/24d_fs.htm Falaras, P., I.M. Arabatzis, T. Stergiopoulos and M.C. Bernard. (2003). Enhanced activity of silver modified thin-film Ti0 2 photocatalysts. Int. J. Photoenergy, 05, 123-130. Fernândez, A., G. Lassaletta, V.M. Jiménez, A. Justo, A.R. Gonzâlez-Elipe, J.M. Herrmann, H. Tahiri and Y. Ait-lchou. (1995). Préparation and characterization of Ti0 2 photocatalyst supported on various rigid supports (glass, quartz and stainless steel). Comparative studies of photocatalytic activity in water purification. Appl. Catal. B - Environ.,7, 49-63. Fogler, H.S. (2006). Eléments of chemical reaction engineering. Fourth Edition. Pearson Education Inc. USA. Fox, M.A. and M.T. Dulay. (1993). Heterogeneous photocatalysis. Chem. Rev., 83, 341-357. Frank, S. and A.J. Barn. (1977). Heterogeneous photocatlytic oxidation of cyanide and sulfite in aqueous solutions at semiconductor powders. J. Phys. Chem-US, 81(15), 1484-1488 Fujishima, A., T. Rao, and D. Tryk. (2000). Titanium dioxide photocatalysis. J. Photoch. Photobio. C , 1, 1-21. Fujishima, A. and X. Zhang. (2006). Titanium dioxide photocatalysis: présent situation and future approaches. C.R. Chimie, 9, 750-760. Galindo, F., R. Gômez and M. Aguilar. (2008). Photodegradation of the herbicide 2,4-dichlorophenoxyacetic acid on nanocrystalline Ti02-Ce0 2 sol-gel catalysts. J. Mol. Catal. A-Chem., 281, 119-125. Gao, B., Y. Ma, Y. Cao, J. Zhao and J. Yao. (2006). Effect of ultraviolet irradiation on crystallization behaviour and surface microstructure of titania in the sol-gel process. J. Solid State Chem., 179, 41-48.  169  Gaya, U.I., and A.H. Abdullah. (2008). Heterogeneous photocatalytic dégradation of organic contaminants over titanium dioxide: A review of fundamentals, progress and problems. J. Photoch. Photobio. C , 9,1-12. Ghertner, D.A., T. Bertram and C. Langelier. (2001). Photochemical dégradation of 2,4-D and atrazine in well-defined média. Abstr. Pap. Am. Chem. S., 221, U176. Giri, R.R., H. Ozaki, T. Ishida, R. Takanami and S. Taniguchi. (2007). Synergy of ozonation and photocatalysis to mineralize low concentration 2,4dicholorophenoacetic acid in aqueous solutions. CHEMOSPHERE, 66(9), 1610-1617. Giri, R.R., H. Ozaki, R. Takanami and S. Taniguchi. (2008). Heterogenous photocatalytic ozonation in dilute aqueous solution with TiÛ2 fiber. Water Sci. Technol., 58(1), 207-216. Goetz, V., J.P. Cambon, D. Sacco and G. Plantard. (2009). Modeling aqueous heterogeneous photocatalytic dégradation of organic pollutants with immobilized Ti0 2 . Chem. Eng. Process, 48, 532-537. Goldstein, S. and J. Rabani. (2008). The ferrioxalate and iodide-iodate actinometers in the UV région. J. Photoch. Photobio. A., 193, 50-55. Gréer, C.W., J. Hawari and R. Samson. (1990). Influence of environmental factors on 2,4-dichlorophenoxyacetic acid dégradation by Pseudomonas cepacia isolated from peat. Arch. Microbiol., 154, 317-322. Gregg, S.J. and K.S.W. Sing. (1982). Adsorption, surface area and porosity. Second Edition, Acemic Press Inc, NY - USA. Guillard, O , D. Debayle, A. Gagnaire, H. Jaffrezic and J. Herrmann. (2004). Physical properties and photocatalytic efficiencies of TiÛ2 films prepared by PECVD and sol-gel methods. Mater. Res. Bull., 39, 1445-1458. Haarstrick, A., O. Kut, and E. Heinzie. (1996). Ti02-asisted dégradation of environmentally relevant organic compounds in wastewater using a novel fluidized bed photoreactor. Environ. Sci. Technol., 30, 817-824. Haber, J. (1991). Manual on catalyst characterization. Pure Appl. Chem., 63 (9), 1227-1246. Haider, A. and O. Levenspield. (1989). Drag coefficient and terminal velocity of spherical and non spherical particles. Powder Technol., 58, 63-70.  170  Haque, F., E. Vaisman, C. Langford and A. Kantzas. (2005). Préparation and performance of integrated photocatalyst adsorbent (IPCA) employed to dégrade model organic compounds in synthetic wastewater. J. Photoch. Photobio. A., 169, 21-27. Harris, G., M. Asce, V. D. Adams, W.M. Moore and D.L. Sorensen. (1987). Potassium ferrioxalate as chemical actinometer in ultraviolet reactors. J. Environ. Eng.-ASCE., 113 (3), 612-627. Hashimoto, K., H. Irie, and A. Fujishima. (2005). Ti0 2 photocatalyst: A historical overview and future prospects. Jpn. J. Appl. Phys., 44 (12), 8269-8285. Hatchard, C.G. and C.A Parker. (1956). A new sensitive chemical actinometer. II. Potassium ferrioxalate as a standard chemical actinometer. Proc. R. Soc. A., 235, 518-536. Hench, L. and J. West. (1990). The Sol-Gel process. Chem. Rev., 90 (1), 33-68. Herrmann, J.M. (1999). Heterogeneous photocatalysis: fundamentals and applications to the removal of various types of aqueous pollutants. Catal. Today, 53, 115-129. Higgins, R., B. A. Bishop and R. Goldsmith. A photocatalytic membrane reactor for enhanced destruction of chloro-organics in aqueous média. CeraMem Corporation Technical document. Retrieved 06-01-2009 from http://www.netl.doe.gov/publications/proceedings/00/ind_part00/em7-4.pdf Hill, J.O. (2005). Encyclopedia of chemical processing. Taylor & Francis Group. Hisanaga, T. and K. Tanaka. (2002). Photocatalytic dégradation of benzène on zeolite-incorporated Ti0 2 film. J. Hazard. Mater. B93, 331-337. Hoffmann, M., S T . Martin, W. Choi and D. Bahnemann. (1995). Environmental applications of semiconductor photocatalysis. Chem. Rev., 95, 69-96. Hosokawa, M., K. Nogi, M. Naito and T. Yokoyama. (2007). Nanoparticle technology handbook. Elsevier, USA. Imoberdorf, G., F. Taghipour, M. Keshmiri and M. Mohseni. (2008). Prédictive radiation field modeling for fluidized bed photocatalytic reactors. Chem. Eng. Sci., 63, 4228-4238. Jackson, E. (1883). A new test for titanium and the formation of a new oxide of the métal. Chem. News., 47,157.  171  Jackson, N.B., C M . Wang, Z. Luo, J, Schwitzgebel, J. G. Ekerdt, J.R. Brock and A. Heller. (1991). Attachment of Ti0 2 powders to hollow glass microbeans: activity of the Ti0 2 -coated beads in the photoassisted oxidation of ethanol to acetaldehyde. J. Electrochem. Soc, 138 (12), 1991 Junin, C , C. Thanachayanont, C. Euvananont, K. Inpor and P. Limthongkul. (2008). Effects of précipitation, sol-gel synthesis conditions, and drying methods on the properties of nano-Ti0 2 for photocatalysis applications. Eur. J. Inorg. Chem., 974-979. Kabir, M. (2006). Fluidized bed photocatalytic reactor System for disinfection and wastewater treatment. PhD Thesis. Department of Chemical and Petroleum Engineering, University of Calgary, AB - Canada. Kabra, K., R. Chaudhary and R. Sawhney. (2004). Treatment of hazardous organic and inorganic compounds through aqueous-phase photocatalysis: A Review. Ind. Eng. Chem. Res., 43 (24), 7683-7696. Kamble, S., S.P. Deosarkar, S.B. Sawant, J.A. Moulijn and V.G. Pangarkar. (2004). Photocatalytic dégradation of 2,4-Dichlorophenoxyacetic acid using concentrated solar radiation: batch and continuous opération. Ind. Eng. Chem. Res., 43, 8178-8187. Kamble, S.P., S.B. Sawat and V.G. Pangarkar. (2006). Photocatalytic mineralization of phenoxyacetic acid using concentrated solar radiation and titanium dioxide in slurry photoreactor. Chem. Eng. Res. Des., 84(A5), 355-362. Kaneko, M. and I. Okura. (2002). Photocatalysis science and technology. Kodousha, Springer. Japan. Kanki, T., S. Hamasaki, N. Sano, A. Toyoda and K. Hirano. (2005). Water purification in a fluidized bed photocatalytic reactor using Ti02-coated ceramic particles. Chem. Eng. J., 108,155-160. Khan, R.A. and J.F. Richardson. (1987). The résistance to motion of a solid sphère in a fluid. Chem. Eng. Commun., 62,135-150. Keshmiri, M., T. Troczynsky and M. Mohseni. (2006). Oxidation of gas trichloroethylene and toluène using composite sol-gel Ti0 2 photocatalytic coatings. J. Hazard. Mater., B128, 130-137. Keshmiri, M. (2004). Composite sol-gel process for photocatalytic titanium dioxide. PhD thesis. Department of Metals and Materials Engineering - University of British Columbia, BC - Canada.  172  Keshmiri, M., M. Mohseni and T. Troczynsky. (2004). Development of novel TÏO2 sol-gel derived composite and its photocatalytic activities for trichloroethylene oxidation. Appl. Catal. B - Environ., 53, 209-219. Kim, D.H. and M.A. Anderson. (1996). Solution factors affecting the photocatalytic and photoelectrocatalytic dégradation of formic acid using supprted Ti0 2 thin films. J. Photoch. Photobio. A., 94, 221-229. Kosanic, M. (1998). Photocatalytic dégradation of oxalic acid over TiC>2 powder. J. Photoch. Photobio. A., 119, 119-122. Krishna, V., N. Noguchi, B. Koopman and B. Moudgil. (2006). Enhancement of titanium dioxide photocatalysis by water-soluble fullerenes. J. Colloid Interf. Sci., 304, 166-171. Krysa, J., G. Waldner, H. Mest'ânkovâ, J. Jirkovsky and G. Grabner. (2006). Photocatalytic dégradation of model organic pollutants on an immobilized particulate TÏO2 layer: Rôles of adsorption processes and mechanistic complexity. Appl. Catal. B - Environ., 64, 290-301. Kuhn, H.J., S.E. Braslavsky and R. Schmidt. (2004). Chemical actinometry. Pure Appl. Chem., 76(2), 2105-2146. Kulas, J., I. Rousar and J. Krysa. (1998). Photocatalytic dégradation rate of oxalic acid on the semiconductive layer of n-Ti02 particles in the batch mode plate reactor: Mass transfer limits. J. Appl. Electrochem., 28, 843-853. Kumara, G.R.R.A., F.M. Sultanbawa, V.P.S. Perera, I.R.M. Kottegoda and K. Tennakone. (1999). Continuous flow photochemical reactor for solar decontamination of water using immobilized Ti02. Sol. Energ. Mat. Sol. C , 58, 167-171. Kwan, C.Y. and W. Chu. (2003). Photodegradation of 2,4-Dichlorophenoxyacetic acid in various iron-mediated oxidation Systems. Water Res., 36(9), 2043-2051. Laine, R.M. (1990). Chemical processing of glasses. SPIE, vol. 1328, Sol-Gel Optics, 16-28. Lau, J., W. Bahnfleth and J. Freihaut. (2009). Estimating the effects of ambient conditions on the performance of UVGI air cleaners. Build. Environ., 44, 13621370. Lee, D., S.C. Kim, I.C. Cho, S.J. Kim and S.W. Kim. (2004). Photocatalytic oxidation of microcystin-LR in a fluidized bed reactor having Ti02-coated activated carbon. Sep. Purif. Technol., 34, 59-66.  173  Lee, S. and K. Lee. (2004). Mass transfer effect on the photocatalytic activity of Uv7Tï02 packed-bed Systems. J. Ind. Eng. Chem., 10 (3), 492-498. Lehr, J., J. Keelay and J. Lehr. (2005). Water Encyclopedy. John Wiley and Sons, 755-823. LightSources Inc., and LightTech Lamp Technology. (2005). Technical sheet on germicidal lamps. USA. Linsebigler, A.L., G. Lu and J.T. Yates. (1995). Photocatalysis on Ti0 2 surfaces: principles, mechanisms and selected results. Chem. Rev., 95 (3), 735-758. Liqiang, J., X. Baifu, Y. Fulong, W. Baiqi, S. Keying, C. Weimin and F. Honggang. (2004). Deactivation and régénération of ZnO and TÏO2 nanoparticles in the gas phase photocatalytic oxidation of n-C7Hi6 or S0 2 . Appl. Catal. A-Gen., 275, 4954. Livage, J. (1998). Sol-gel synthesis of heterogeneous catalyst from aqueous solutions. Catal. Today, 41, 3-19. Livage, J., C. Sanchez, M. Henry and S. Doeuff. (1989). The chemistry of the sol-gel process. Solid State lonics, 32(33), 633-638. Livage, J., M. Henry and C. Sanchez (1988). Sol-gel chemistry of transition métal oxides. Prog. Solid State Ch., 18, 259-341. Lutz, E.H. and M.V. Swain. (1991). Mechanical and thermal shock properties of duplex-ceramics. A review. Materials Forum, 15, 307-323. MacAdam, J. and S.A. Parsons. (2009). An investigation into advance oxidation of three chlorophenoxy pesticides in surface water. Water Sci. Technol., 59 (8), 1665-1671. Macak, J., M. Zlamal, J. Krysa and P. Schmuki. (2007). Self-organized Ti0 2 nanotubes layers as highly efficiency photocatalysts. Small, 3(2), 300-304. Mackenzie, J. (2003). Sol-Gel Research - Achievements since 1981 and Prospects for the Future. J. Sol-Gel Sci. Techn., 26, 23-27. Marczenko, Z. (1986). Séparation and spectrophotometric éléments. Halsted Press, NY - USA.  détermination of  Marczewska, A.D. and Marczewski, A.W. (1997). A gênerai model for adsorption of organic solutés from dilute aqueous solutions on heterogeneous solids: application for prédiction of multisolute adsorption. Langmuir, 13,1245-1250.  174  McMurray, T.A., J.A. Byrne, P.S.M. Dunlop, J.G.M. Winkelman, B.R. Eggins and ET. McAdams. (2004). Intrinsec kinetics of photocatalytic oxidation of formic and oxalic acid on immobilised "ÏÏO2 films. Appl. Catal. A - Gen., 262,105-110. Matthews, R. (1992). Photocatalytic oxidation of organic contaminants in water: an aid to environmental préservation. Pure & Appl. Chem., 64(9), 1285-1290. Matthews, R. (1987). Photo-oxidation of organic impurities in water using thin films of titanium dioxide. J. Phys. Chem., 91, 3328-3333. McMurray, T A , J.A. Byrne, P.S.M. Dunlop, J.G.M. Winkelman, B.R. Eggins and ET. McAdams. (2004). Intrinsec kinetics of photocatalytic oxidation of formic and oxalic acid on immobilised Ti0 2 films. Appl. Catal. A - Gen., 262, 105-110. Mears, D.E. (1971). Test for transport limitations in expérimental catalytic reactors. Ind. Eng. Chem. Process Des. Dev., 10(4), 541-547. Muggi, D. and M. J. Backes. (2002). Two active sites for photocatalytic oxidation of formic acid on Ti0 2 : Effect of H 2 0 and température. J. Catal., 209, 105-113. Mills, A., R.H. Davies and D. Worsiey. (1993). Water purification by semiconductor photocatalysis. Chem. Soc. Rev., 22 (6), 417-425. Mills, A., J. Wang and D. Ollis. (2006). Dependence of the kinetics of liquid-phase Photocatalyzed reactions on oxygen concentration and light intensity. J. Catal., 243, 1-6. Minera, C. and D. Vione. (2006). A quantitative évolution of the photocatalytic performance of Ti0 2 slurries. Appl. Catal. B - Environ., 67, 257-269. Modestov, A.D. and O. Lev. Photocatalytic oxidation of 2,4-dichlorophenoxyacetic acid with titania photocatalyst. Comparison of supported and suspended Ti0 2 . J. Photoch.Photobio. A., 112, 261-270. Mrowetz, M. and E. Selli. (2006). Photocatalytic dégradation of formic and benzoic acids and hydrogen peroxide évolution in Ti0 2 and ZnO water suspensions. J. Photoch. Photobio. A., 180,15-22. Muller, G.E. (1997). Numerical simulation of packed beds with monosized sphères in cylindrical containers. PowderTechnol., 92,179-183. Muraganandham, M. and M. Swaminathan. (2006). Ti02 - UV photocatalytic oxidation of reactive Yellow 14: Effect of operational parameters. J. Hazard. Mater. B135, 78-86.  175  Nazeri, A., Bescher, E. and Mackenzie, J. (1993). Ceramic composites by the sol-gel method: A review. Ceram. Eng. Sci. Proc, 14(11-12), 1-19. Nelson R.J., C L . Flakker and D.S. Muggli (2007). Photocatalytic oxidation of methanol using titania-based fluidized beds. Appl. Catal. B - Environ. 69(3-4), 189-195. Nikolaou, A. and T.D. Lekkas. (2001). The rôle of natural organic matter during formation of chlorination by-products: A Review. Acta Hydroch. Hydrob., 29, 6377. Ohtami, M., M. Kakimoto, S. Nishimoto and T. Kagia. (1993). Photocatalytic reaction of neat alcohols by metal-loaded Titanium (IV) oxide particles. J. Photoch. Photobio. A., 70(3), 265-272. Okamoto, K., Y. Yamamoto, H. Tanaka and A. Itaya. (1985). Kinetics of heterogenous photocatalytic décomposition of phénol over anatase TiC>2 powder. Bull. Chem. Soc. Jpn., 58, 2023-2028. Ollis, D.F. and C. Turchi. (1990). Heterogeneous photocatalysis for water purification: contaminant mineralization kinetics and elementary reactor analysis. Environ. Prog. , 9 (4), 229-234. Ollis, D.F., E. Pelizzetti and N. Serpone. (1991). Photocatalyzed destruction of water contaminants. Environ. Sci. Technol., 25 (9), 1522-1529. Ollis, D. (2005). Kinetics of liquid phase photocatalyzed reactions: An llluminating approach. J. Phys. Chem. B. 109, 2439-2444. Optoweb (2009). Sol-gel process. http://optoweb.fis.uniforma2.it/optp/solgel/  Retrieved  02-16-2009  from  Osajima, J.A., H.M. Ishiki and K. Takashima. (2008). The Photocatalytic dégradation of Imazapyr. Monatsh. Chem., 139, 7-11. Parker, C.A. (1953). A new sensitive chemical actinometer. I. Some trials with potassium ferrioxalate. Proc. R. Soc. A., 220,104-116. Parson, S. (2004). Advance oxidation processes for water and wastewater treatment. IWA Publising, USA. Peng, T., D. Zhao, K. Dai, W. Shi and K. Hirao. Synthesis of titanium dioxide nanoparticles with mesoporous anatase wall and high photocatalytic activity. J. Phys. Chem. B., 109 (11), 4947-4952.  176  Perry, R.H. and D.W. Green. (1997). Perry's Chemical engineering handbook. Seventh Edition, McGraw Hill, USA. Pettibone, J.M., D.M. Cwiertny, M. Scherer and V.H. Grassian. (2008). Adsorption or Organic acids on Ti02 Nanoparticles: Effect of pH, Nanoparticle Size, and Nanoparticle Aggregation. Langmuir, 24, 6659-6667. Porter, J.F., Y. Li and C. Chan. (1999). The effect of calcination on the microstructural characteristics and photoreactivity of Degussa P-25 Ti0 2 . J. Mater. Sci., 34, 1523-1531. Pozzo, R., M. Baltanas and A. Cassano. (1999). Towards a précise assessment of the performance of supported photocatalysts for water detoxification processes. Catal. Today, 143-157. Pozzo, R., J. Giombi, M. Baltanas and A. Cassano. (2000). The performance in fluidized bed reactor of photocatalysts immobilized onto inert supports. Catal. Today, 62, 175-187. Prakash, A., C L . Briens and M.A. Bergougnou. (1987). Mass transfer between Solid particles and liquid in a three phase fluidized bed. Can. J. Chem. Eng., 65, 228236. Qiu, W. and Y. Zheng. (2007). A comprehensive assessment of supported titania photocatalyst in a fluidized bed photoreactor: Photocatalytic activity and adhérence stability. Appl. Catal. B - Environ, 71,151-162. Que, W., Y. Zhou, Y.L. Lam, Y.C. Chan and C.H. Kam. (2001). Préparation and characterizations of the Ti02/organically modified silane composite materials Produced by the Sol-Gel Method. J. Sol-Gel Sci. Techn., 20,187-195. Rahaman, M. (2007). Ceramic processing. Taylor & Francis Group, USA. Rahn, R.O. (1997). Potassium iodide as a chemical actinometer for 254 nm radiation: use of iodate as an électron scavenger. Photochem. Photobiol., 66(4), 450-455. Rahn, R.O., J. Bolton and M.l. Stefan. (2006). The iodide/iodate actinometer in UV disinfection: détermination of the fluence rate distribution in UV reactors. Photochem. Photobiol., 82, 611-615. Rahn, R.O., M.l. Stefan, J.R. Bolton, E. Goren, P. Shaw and K.R. Lykke. (2003). Quantum yield of the iodide-iodate chemical actinometer: dependence on wavelength and concentration. Photochem. Photobiol., 78(2), 146-152.  177  Raudsepp, M., F.C. Hawthorne and A.C. Turnock. (1990). Evaluation of the Rietveld method for the characterization of fine-grained products of minerai synthesis: the diopside-hedenbergite Join. Can. Minerai., 28, 93-109. Raudsepp, M. and E. Pani. (2003). Application of Rietveld analysis to environmental mineralogy. In: Mineralogical Association of Canada, Short Courses Séries, Volume 31, Chapter8, 165-180. Ray, A. (1998). A new photocatalytic reactor for destruction of toxic water pollutants by advanced oxidation process. Catalysis Today, 44, 357-368. Reed, S.J.B. (2005). Electron microprobe analysis and scanning électron microscopy in geology. Second Edition, Cambridge University Press, UK. Richardson, J.F. and W.N. Zaki. (1954). Sédimentation and fluidization. Part I. Trans. Inst. Chem. Eng., 32, 35-53. Richerson, D.W. (2006). Modem ceramic engineering: properties, processing and use in design. Taylor& Francis Group. Third Edition. USA. Riedwyl, H. (1998). Modifying and using Yates' algorithm. Stat. Pap., 39, 41-60. Ring, T.A. (1996). Fundamentals of ceramic powder processing and synthesis. Académie Press. Rodriguez-Gonzâalez, V., A. Moreno-Rodrîguez, M. May., F. Tzompantzi and R. Gômez. (2008). Slurry photodegradation of 2,4-dichlorophenoacetic acid: A comparative study of impregnated and sol-gel In203-Ti02 mixed oxide catalysts. J. Photoch. Photobio. A., 193, 266-270. Sakka, S. (2006). Current sol-gel activities in Japan. Sol-Gel Sci. Techn., 37, 135140. Satterfield, C.N. (1980). Heterogenous catalysis in practice. McGraw-Hill, Inc. USA. Scherer, G.W. (1990). Theory of drying. J. Am. Ceram. Soc, 73 (1), 3-14. Serpone, N., G. Sauvé, R. Koch, H. Tahiri, P. Pichat, P. Piccinini, E. Pelizzeti and H. Hidaka. (1996). Standardization protocol of process efficiencies and activation parameters in heterogenous photocatalysis: relative photonic efficiencies Çr- JPhotoch. Photobio. A., 94, 191-203. Serpone, N. (1997). Relative photonic efficiencies and quantum yields in heterogeneous photocatalysis. J. Photoch. Photobio.A, 104,1-12.  178  Shankar, M.V., S. Anandan, N. Venkatachalam, B. Arabindoo, and V. Murugesan. (2006). Fine route for an efficient removal od 2,4-dichlorophenoxyacetic acid (2,4-D) by zeolite-supported Ti0 2 . Chemosphere, 63, 1014-1021. Shi, J., J. Zheng and P. Wu. (2009). Préparation, characterization and photocatalytic activities of holmium-doped titanium dioxode nanoparticles. J. Hazard. Mater., 161,416-422. Sinclair, C.J., A.B.A. Boxall, S.A. Parsons and M.R. Thomas. (2006). Prioritization of pesticide environmental transformation products in drinking water supplies. Environ. Sci. Technol., 40, 7283-7289. Sing, K.S.W., R.A.W. Haul, L. Moscou, R.A. Pierotti, J. Rouquerol and T. Siemieniewska. (1985). Reporting physisorption data for gas/solid Systems with spécial référence to the détermination of surface area and porosity. Pure Appl. Chem., 57(4), 603-619. Sink, K.S.W., R.A.W. Haul, R.A. Pierotti and T. Siemieniewska (1985). Reporting physisorption data for gas/solid Systems with spécial référence to the détermination of surface area and porosity. Pure Appl. Chem., 57, 603-619. Skjak-Braek, G., T. Anthonsen and P. Sandford. (1989). Chitin and chitosan. Elsevier Applied Science, USA. Stangroom, S.J., C D . Collins and J.N. Lester. (1998). Sources of organic micropollutants to lowland rivers. Environ. Technol., 19, 643-666. So, W.W., S.B. Park, K.J. Kim and S.J. Moon. (1997). Phase transformation behavior at low température in hydrothermal treatment of stable and unstable titania sol. J. Colloid Interf. Sci., 191, 398-406. Speitel, G.E., J.M. Symons, J.M. Mialaret and M.M.E. Wanielista. (2000). AOP/biomilm process for DOX precursors. Journal-American Water works Association, 92, 59-73. Su, C , B.-Y. Hong and C.-M. Tseng (2004). Sol-gel préparation and photocatalysis of titanium dioxide. Catal. Today, 96,119-126. Syoufian, A., O. H. Satriya and K. Nakashima. (2007). Photocatalytic activity of titania hollow sphères: Photodecomposition of méthylène blue as a target molécule. Catal. Commun., 8, 755-759. Szépe, S., and O. Levenspiel. (1968). Optimal température policies for reactors subject to catalyst deactivation - I Batch reactor. Chem. Eng. Sci., 23, 881-894.  179  Teichner, S.J. (2008). The origins of photocatalysis. J. Porous Mater., 15, 311-314. Terashima, Y., H. Ozaki, R.R. Giri, T. Tano, N. Nakatsuji, R. Takanami and S. Taniguchi. (2006). Photocatalytic oxidation of low concentration of 2,4-D solution with new Ti0 2 fiber catalyst in a continuous flow reactor. WA. Sci. Technol., 54 (8), 55-63. Thiruvenkatachari, R., S. Vigneswaran and I.S. Moon. (2008). A review on UV/TiC>2 photocatalytic oxidation process. Korean J. Chem. Eng., 25 (1), 64-72. Thurman, E.M. (1985). Organic geochemistry of natural waters. Nijhoff & Junk Publ. USA. Tournié, P., C. Laguerie and J.P. Couderc. (1979). Corrélations for mass transfer between fluidized sphères and a liquid. Chem. Eng. Sci., 34,1247-1255. Tremblay, K.M.A. (2001). Photocatalytic treatment of wastewater: a model for a fast process monitoring analysis method. M.Se Thesis. Department of Chemistry, University of Calgary, AB. Trillas, M., J. Peral and X. Domenech. (1995). Redox photodegradation of 2,4dichlorophenoxyacetic acid over Ti0 2 . Appl. Catal. B - Environ.,5, 377-387. Trillas, M., J. Peral and X. Domenech. (1996). Photocatalyzed dégradation of phénol, 2,4-Dichlorophenol, phenoxyacetic acid and 2,4-Dichlorophenoxyacetic acid over supported TiÛ2 in a flow system. J. Chem. Tech. Biotechnol., 67, 237242. Turchi, C , and D. Ollis. (1989). Mixed reactant photocatalysis: intermediates and mutual rate inhibition. J. Catal., 119, 483-496. Turchi, C , and D. Ollis. (1990). Photocatalytic dégradation of Organic water contaminants: mechanism involving hydroxyl radiael attack. J. Catal., 122, 178-192. Turchi, C. and E. Wolfrum. (1992). Letters to the Editors: comments on "Reactor dynamics in the évaluation of photocatalytic oxidation kinetics". J. Catal., 136, 626-628. Ultraviolet Technology of Australasia PTY LTDA (UVTA). (1999). Technical sheet on critical factors for ultraviolet desinfection Systems. Vaisman, E., M.F. Kabir, A. Kantzas and C.H. Langford. (2005). A fluidized bed photoreactor exploiting a supported photocatalyst with adsorption preconcentration capacity. J. Appl. Electrochem., 35, 675-681.  180  Veeh, R.H., W.P. Inskeep and A.K. Camper. (1996). Soil depth and température effects on microbial dégradation of 2,4-D. J. Environ. Quai., 25, 5-12. Wen, J. and G. Wilkes (1996). Organic/inorganic hybrid network materials by the solgel approach. Chem. Mater., 8,1667-1681. Werther, J. and J. Reppenhagen (1999). Catalyst attrition in fluidized-bed Systems. AlChE Journal, 45 (9), 2001 -2010. Wetchakun, N. and S. Phanichphant. (2008). Effect of température on the degree of anatase-rutile transformation in titanium dioxide nanoparticles synthesized by the modified sol-gel method. Curr. Appl. Phys., 8, 343-346. WHO (2006). Water. Training Package for the Health Sector. Retrieved 08-27-2008 from http://www.who.int/ceh/capacity/Water.pdf. WHO (2008). Water and sanitation. Retrieved http://www.euro.who.int/watsan/issues/20030903_1  08-27-2008  from  Xu, J., W. Dai, J. Li, Y. Cao, H. Li and K. Fan. (2008). Novel core-shell structure mesoporous titania microspheres: préparation, characterization and excellent photocatalytic activity in phénol abatement. J. Photoch. Photobio. A., 195, 284294. Yang, Q. (1999). Composite sol-gel ceramics. PhD thesis. Department of Metals and Materials Engineering - University of British Columbia, BC - Canada. Yin, H., Y. Wada, T. Kitamura, S. Kambe, S. Murasawa, H. Mori, T. Sakata and S. Yanagida. (2001). Hydrothermal synthesis of nanosized anatase and rutile Ti0 2 using amorphous phase TÏO2. J. Mater. Chem., 11,1694-1703. Zalazar, C , M. Labas, C. Martin., R. Brandi, O. Alfano and A. Cassano. (2005). The extended use of actinometry in the interprétation of photochemical reaction engineering data. Chem. Eng. J., 109, 67-81. Zhang, X., H. Liu, W. Li, G. Cui, H. Xu, K. Han and Q. Long. (2008). Visible-light photocatalytic dégradation of aromatic contaminants with simultaneous H2 génération: comparison of 2,4-Dichlorophenoxyacetic acid and 4-Chlorophenol. Catal. Lett., 125, 371-375. Zhang, M., T. An, J. Fu, G. Sheng, X. Wang, X. Hu and X. Ding. (2006). Photocatalytic dégradation of mixed gaseous carbonyl compounds at low level on adsorptive Ti0 2 /Si0 2 photocatalyst using a fluidized bed reactor. Chemosphere, 64, 423-431.  181  Zhang, W. and F. A. DiGiano. (2002). Comparison of bacterial regrowth in distribution Systems using free chlorine and chloramine: statistical study of causative factors. Water Res., 36,1469-1482.  182  Appendix A Statistical analysis: Yates eff ects, F values and CI The calculation of the effects in the response variables (catalyst attrition and photocatalyst activity) of the différent formulation used to prépare the composite TIO2 sphères will be described in this section.  • Comparison between différent formulation to produce TiOg sphères To compare the photocatalytic activity of the différent catalyst formulations (Section 6.2.1 - Preliminary tests and effects of various parameters), an analysis between différent formulations was done by comparing the mean of the reaction rate constant of each of those formulations following the same approach described by Box et al. (1978). For each of the experiments described in Section 6.2.1, three replicates were done per expérimental run; therefore, three independent values of the rate constant were determined during the FA dégradation in a FBPR. An average (yt) was calculated for each of those sets of data as well as the grand average for ail the data (y) obtained during in thèse preliminary experiments, which is the sum of ail the reaction rate constants experimentally obtained divided by the total number of observations. For the variations within formulations, the sum of squares, St (déviations from the average), was calculated by:  st=tiyti-yt)2  (A.1)  183  Appendix A - Statistical analysis: Yates effects, F values and Cl where the ytm is the mth observation (each rate constant) in the f1 formulation and yt is the average of the three expérimental values. Then, the within-formulation sum of squares for ail the formulations (fin total) was given by:  s^tt^-yJ  (A.2)  t=\ m=\  and the within-formulation mean square was given by:  2  sR=-^—  SR  (A.3)  Nt-f where f is the total number of formulations tested and Nt is the total number of observations (total number of reaction rate constants). For the variations between formulations, the between-formulations sum of squares can be calculated by:  sT=t«t(yt-y)2  (A.4)  where nt is the number of observations (reaction rate constants) per formulation (f). Then, the between-formulations mean square was calculated as follows:  sr2=^-  (A.5)  Once the residues analysis was done to probe that the errors are independently and identically distributed in a normal distribution with mean zéro and variance <x2 (usually abbreviated as IIDN (0, a2)), the hypothesis of the analysis was probed. This hypothesis was that the formulations means were ail equal and that there was no différence between formulations. The expérimental F value was calculated by:  184  Appendix A - Statistical analysis: Yates effects, F values and Cl  ^=£4  (A.6)  which is the F value for VTIVR degrees of freedom, where vT= f-1 and vR = Nt-f. If the Fexp is within the F distribution for a given confidence level (Le., 99%), the null hypothesis is true; on the other hand, if the Fexp is outside of the F distribution, the null hypothesis is discrédit and some différences exist between formulations. In this particular case, doing the above mentioned calculation for the data listed in Table A.1, the Fexp was 2.47 and the Fg,2o was 4.57 (at 99% confidence level). Therefore, from the statistical point of view, there was no différence between formulations.  Table A.1: Reaction rate constants for the différent formulation utilized during the preliminary tests in Section 6.2.1 PT1 PT2 PT3 PT4 PT5 PT6 PT7 PT8 PT9 PTi 0 0.02154 0.02290 0.02415  0.02985 0.02154 0.02451  0.02506 0.02389 0.02489  0.02290 0.01990 0.02430  0.02204 0.02440 0.01990  0.02676 0.02501 0.02315  0.02967 0.02814 0.02468  0.02911 0.02714 0.02830  0.02413 0.02569 0.02147  0.02968 0.02353 0.02687  where: PTi: is referring to the formulation that used Nitric acid (HNO3) instead of HCI PT2: is referring to the formulation that used 0.053 g H 2 0 ml_"1 TTIP PT3: is referring to the formulation that used 0.13 ml_ HCI ml_"1 TTIP PT4: is referring to the formulation that used 0.30 g P-25 ml_"1 TTIP PT5: is referring to the formulation that used 0.50 g P-25 ml_"1 TTIP PT6: is referring to the formulation that used a chitosan solution with a pH of 4.58 PT7: is referring to the formulation that used a chitosan solution with a pH of 4.01 PT8: is referring to the formulation that used a chitosan concentration of 15 g chitosan L"1 PTg: is referring to the formulation that had a drying process at 80°C for 20h PT-|0: is referring to the formulation that was calcined for 3 h at 600°C  185  Appendix A - Statistical analysis: Yates effects, F values and Cl • Yates algorithm used for the expérimental design Yates algorithm was used to calculate the effects and interactions among the différent variables utilized during the expérimental design described in Section 6.2.2 (Expérimental design) for the catalyst improvement. In this expérimental design, the activity and the catalyst attrition were the response variables. As shown in Table A.2, five variables were analyzed in two différent levels (différent values of the variable). Using this information, the signs arrangement shown in Table A.3 was done, showing the différent combination between variables. In this particular case, the levels of variable E were generated based on the signs of ABCD interaction (E = ABCD) Table A.2: Variables included in the 25"1 expérimental design Levels  Variable 1  Amount of Water, A (g mL" TTIP) Amount of HCI, B (mL mL'1 TTIP) Concentration of NH4OH solution, C (%) Drying conditions, D Calcination time, E (h) Table A.3: Sic ns arrangement for the 2 Run Number B A 5 + 6 + 7 + + 4 3 + 8 + 2 + + 1 12 + 16 + 9 + + 15 14 + 11 + 13 + + 10  -  +  0.040 0.133 10 RT-15days 1  0.053 0.200 20 80°C - 20 h 3  factorial design Variables C -  + + + + -  + + + +  D -  + + + + + + + +  E + -  + -  + + -  + + -  + -  + 186  Appendix A - Statistical analysis: Yates effects, F values and Cl Table A.4 shows the results obtained after applying Yates algorithm. One important considération at this point was that ail the effects of third and fourth order (represented by three-factor and four factor interactions) were ignored, assuming that détectable effects could be identify only with first and second order interactions.  nding pattern and estima tes from 25-1 design of Table A.3 Confounding Pattern Estimate -1.1939 /A—A /B-»B 5.0939 -2.2757 /C->C /D->D 0.1371 /E-»E 0.0379 /AB-^AB -2.6568 /AC-+AC 0.5455 /AD^AD 2.1939 /AE^AE 0.5106 0.2789 /BC^BC /BD-*BD -1.3829 /BE->BE -1.6264 /CD-+CD -0.5227 /CE->CE 0.3520 /DE->DE 3.6077 11.6841 Average  With the data listed in Table A.4, a normal plot (Figure A.1) can be done to détermine the effects and interactions that are significant for the analysis. As can be seen, there are some points (close to zéro) that can be adjusted to a linear régression which means that ail those values can be explained by noise. The other values are considered significant for the analysis. Therefore, using just the significant effects, the estimate conversion for the process development data was given at the vertices of the design as is shown in Equation A.7.  (lAB^ X  AB+\  (IDE}  (IB^  xB +  (f  (IA  xc +  (IBD  IBE '•BE  + -  l 2  (IAD X  BD  +  \  *ÂD  (A.7)  x„  \^J  187  Appendix A - Statistical analysis: Yates effects, F values and Cl  2.50 -t 2.00 N  1.50  §  1.00 -  «S  jj  0.50  ixî  0.00  •c  -0.50  ro  fc -1.00  o z  -1.50 -2.00 -2.50 -  4  -  2  0  2  4  6  5 1  Effects - 2 " design Figure A.1 Normal plot of effects for the expérimental design described in Section 6.2.2  The différence between the estimated value (Equation A.7) and the values obtained experimentally (reaction rate constant and attrition values - Table 6.4) was defined as the residuals. Therefore, in order to check the model, a normal plot between of thèse residuals can be done (Figure A.2). In this case, ail the points from this residual plot should lie close to the line, confirming the conjecture that effects other than AB, C, BE, BD, AD, DE, B and A are readily explained by random noise. As can be seen in Figure A.2, the points are fairly close to the line, which was enough to confirm the conjecture. With thèse results, the values of the variables A,B,C,D and E (Table A.2) can be selected based on the combination that gives lower attrition and higher activity. The effect of the amount of water (A), the amount of HCI (B) and the concentration of NH4OH solution (C) was selected directiy from the results obtained based on the signs arrangements shown in Table A.3. But the effects of BE, BD, AD and DE cannot be interpreted separately because of the interaction between variables and is  188  Appendix A - Statistical analysis: Yates effects, F values and Cl better to consider the two-way table were the combine effect of the variables was analyzed.  2.50 -, 2.00  N  1.50 -  c o  1.00 -  '•C 3 X2  (A •D  ko z  0.50 0.00  -0.50 -1.00 -1.50 -2.00 -2.50 -1.50  -1.00  -0.50  0.00  0.50  1.00  1.50  2.00  Residual Figure A.2 Normal plot of residuals for the expérimental design described in Section 6.2.2  • Confidence interval calculations In order to calculate the error bars in ail the graphs described in this document, the confidence interval was calculated following the approach described by Cumming et al. (2007). In this approach, the standard déviation (Equation A.8) was calculated to quantify the average différence between the data points and their mean. Then, the standard error was calculated (Equation A.9) to détermine how variable the mean would be. Finally the confidence interval was calculated with Equation A. 10.  189  Appendix A - Statistical analysis: Yates effects, F vaiues and Cl  SD = y\±^  iien  (A.8)  SE = ^= 4nt  (A.9)  CI = Mx±tÀni_x)SE  (A.10)  where X is each individual data, M is the mean, n is the number of independent samples, tV(n-i) is a critical value of tv for nr1 degree of freedom at 95%. The confidence interval can be defined like the range of values where you can be 95% confident that contains the true mean.  • Standard Error of the parameter The standard error of the parameter (rate constant dégradation) was estimated using the standard error of the estimation and the covariance matrix (2x2) given for the exponential fit. Thèse two parameters were obtained by the program CurveExpert 1.3. The standard error (Equation A.11) was calculated multiplying the standard error of the estimation (Eest) by the covariance élément of the parameter (Ce/em)  Serror=(Eest-CelemT5  (A.11)  With this standard error and using the t distribution, the standard error of the parameter can be estimated as follows:  "param  =  ^ error ' *v(n,-2)  (A. 11 )  190  Appendix A - Statistical analysis: Yates effects, F values and Cl where tv is the T distribution at nt - 2 degrees of freedom (nt is the number of points considered to do the exponential adjustment)  191  Appendix B Particle Size Distribution of the TiCh Particles in Suspension The particle size distribution analysis was done for the treated water after three hours of normal opération of the FBPR with sphères calcined at 600°C. The equipment used for the analysis (Mastersizer 2000 - Malvern Instruments) was located at UBC Environmental Group at the Department of Civil Engineering.  D -, y  mil  j  8 -  I lillll d  7 -  ihyr h  5 -  i  4  1  3-  ni iT I !• \w  j  \  i 1 |t  =  ;| Ï  .  !1 Iléil llfil  ...  j (il  i;  m  1  2 -  I  1 u * 0. 31  !  Il ni* T  6 -  £  i  À 0.1  "IN*  1  •  11  11 |  ^t\  A f  |  \  T  \ ill  II!  •^  100 10 Particle size ( nm)  1000  10C)00  Figure B.1 Particle size distribution for the attrition in the FBPR after 3 h of normal opération using sphères calcined at 600°C As can be seen in Figure B.1, a significant amount of particles had an average size around 15 (j.m.  192  

Cite

Citation Scheme:

        

Citations by CSL (citeproc-js)

Usage Statistics

Share

Embed

Customize your widget with the following options, then copy and paste the code below into the HTML of your page to embed this item in your website.
                        
                            <div id="ubcOpenCollectionsWidgetDisplay">
                            <script id="ubcOpenCollectionsWidget"
                            src="{[{embed.src}]}"
                            data-item="{[{embed.item}]}"
                            data-collection="{[{embed.collection}]}"
                            data-metadata="{[{embed.showMetadata}]}"
                            data-width="{[{embed.width}]}"
                            async >
                            </script>
                            </div>
                        
                    
IIIF logo Our image viewer uses the IIIF 2.0 standard. To load this item in other compatible viewers, use this url:
http://iiif.library.ubc.ca/presentation/dsp.24.1-0058900/manifest

Comment

Related Items