xml version 1.0 encoding UTF-8 standalone no
record xmlns http:www.loc.govMARC21slim xmlns:xsi http:www.w3.org2001XMLSchema-instance xsi:schemaLocation http:www.loc.govstandardsmarcxmlschemaMARC21slim.xsd
leader nam Ka
controlfield tag 001 001935122
007 cr mnu|||uuuuu
008 080421s2007 flua sbm 000 0 eng d
datafield ind1 8 ind2 024
subfield code a E14-SFE0002295
Thermochemical treatment of TiO2 nanoparticles for photocatalytic applications
h [electronic resource] /
by Mark Schmidt.
[Tampa, Fla.] :
b University of South Florida,
ABSTRACT: Titanium Dioxide (TiO2) has been considered an ideal photocatalyst due to factors such as its photocatalytic properties, chemical stability, impact on the environment and cost. However, its application has been primarily limited to ultraviolet (UV) environments due to its high band gap (3.2 eV). This high band gap limits the harvesting of photons to approximately 4% of sunlight radiation. Research today is focused on lowering this gap by doping or coupling TiO2 with other semiconductors, transition metals and non-metal anions, thereby expanding its effectiveness well into the visible range. This thesis explores the effects of thermal and thermochemical ammonia treatment of nano-particulated TiO2. The objective is to synthesize a photocatalyst with a lower band gap energy that demonstrates photocatalytic activity in the visible range while at the same time retaining its photocatalytic properties in the UV range.Specifically, this study utilizes pure commercial nano-particulated TiO2 powder (Degussa P-25), and uses this untreated TiO2 as a baseline to investigate the effects of thermal and thermochemical treatments. Nitrogen-doping is carried out by gas phase impregnation using anhydrous ammonia as the nitrogen source and a tube furnace reactor. The effects of temperature, time duration and gas flow rate on the effectiveness of thermally and thermochemically treated TiO2 are examined. Thermally treated TiO2 was calcinated in a dry inert nitrogen (N2) atmosphere and the effects of temperature and treatment duration are investigated. The band gap of the thermally treated and thermochemically ammonia treated TiO2 have been measured and calculated using an optical spectrometer.The photocatalytic properties of all materials have been investigated by the degradation of methyl orange (MO) in an aqueous solution using both visible simulated solar spectrum (VSSS) and simulated solar spectrum (SSS) halogen light sources. Methyl orange degradation has been measured and calculated using an optical spectrometer. The phase structure and particle size of the materials is determined using x-ray diffraction (XRD). The BET surface area of the samples has been obtained using an Autosorb. Surface or microstructure characterization has also been obtained by scanning electron microscopy (SEM) and transmission electron microscopy (TEM)
Thesis (M.S.)--University of South Florida, 2007.
Includes bibliographical references.
Text (Electronic thesis) in PDF format.
System requirements: World Wide Web browser and PDF reader.
Mode of access: World Wide Web.
Title from PDF of title page.
Document formatted into pages; contains 147 pages.
Advisor: Elias Stefanakos, Ph.D.
x Electrical Engineering
t USF Electronic Theses and Dissertations.
Thermochemical Treatment of TiO2 Nanoparticles for Photocatalytic Applications by Mark Schmidt A thesis submitted in partial fulfillment of the requirement s for the degree of Master of Science in Electrical Engineering Department of Electrical Engineering College of Engineering University of South Florida Major Professor: Elias Stefanakos, Ph.D. D. Yogi Goswami, Ph.D. Nikolai Kislov, Ph.D. Date of Approval October 31, 2007 Keywords: photocatalysis, methyl orange, doping, calcination, thermal treatment Copyright 2007, Mark Schmidt
DEDICATION I dedicate this thesis to my wife. Without her, this journey would not have been possible. I will be forever gratef ul for her support, encouragement and trust. I hope to honor her selfle ssness in the work that lies ahead.
ACKNOWLEDGEMENTS I would like to thank my advisor, Dr Elias Stefanakos, for the opportunity to work with him and the members of t he Clean Energy Research Center. I greatly appreciate the trus t and confidence he placed in me. I would like to especially thank Dr. Nikolai Kislov. His wisdom, guidance, patience and enthusiasm will never be forgotten. I woul d also like to thank Dr. Yogi Goswami for his guidance during our time together. His insights and leadership are greatly appreciated. I would also like to thank the sta ff and faculty of the CERC. I would especially like to recognize the guidance and assistance given to me by Dr. Nikhil Kothurkar, Mr. Chuck Garretson, Dr. Se sha Srinivasan and Dr. Burton Krakow. Finally, I would like to thank my fellow students of the CERC for their help and support, especially Ms. Paula Algarin.
NOTE TO READER Note to Reader: The original of th is document contains color that is necessary for understanding the data. The or iginal thesis is on file with the USF library in Tampa, Florida.
i TABLE OF CONTENTS LIST OF TABLES.................................................................................................v LIST OF FI GURES...............................................................................................vi ABSTRACT..........................................................................................................xi CHAPTER 1: THE NEED FOR A VI SIBLE-LIGHT PH OTOCATALYST...............1 CHAPTER 2: PHOT OCATALYSI S.......................................................................3 2.1 Photocatal ysis Defi ned................................................................................3 2.2 Photogenerated Catalysis Vers us Catalyzed Photolysis.............................4 2.3 Principles of Photocatal ysis.........................................................................5 2.4 Electronic Excitation..................................................................................10 2.5 Band Gap Ex citation..................................................................................11 2.6 Band Edge Position...................................................................................15 2.7 Charge Separation, Tr apping and Recomb ination....................................16 2.8 Quantum Size Effects................................................................................18 CHAPTER 3: TITANIUM DIOXIDE AS A MODEL PHOT OCATALYST..............20 3.1 The Importance of Titanium Dioxide..........................................................20 3.2 The Lattice Structure of Anatase and Rutile TiO2......................................21 3.3 Electronic Structure of TiO2.......................................................................24 3.4 Photoadsorption and Phot odesorption of Oxygen on Ti02.........................24 3.5 Photooxidation at the Liqu id-Solid Interface of TiO2..................................25 3.6 Trapping, Recombination, and Interfacial Elec tron Transfer.....................26
ii 3.7 Photomineralization R eactions in Aqueous Photocatalytic Suspensions..28 3.8 Modifications of TiO2 for Improved Photocatal ytic Performance................28 3.8.1 Surface Modifica tion Using Metals ......................................................28 3.8.2 Transition Metal D oping ......................................................................30 3.8.3 Coupled Phot ocatalyst s......................................................................30 3.9 Degussa P-25 TiO2...................................................................................31 CHAPTER 4: METHYL ORANGE AS A MODEL PO LLUTANT..........................35 4.1 The Real World Probl em of Meth yl Or ange...............................................35 4.2 Chemical Compositi on of Methyl Orange..................................................36 4.3 Chemical Stability of Methyl Orange .........................................................38 4.4 Calculation of Rate Cons tants...................................................................40 4.5 Catalyst Loading a nd Discolora tion...........................................................45 4.6 pH and Disco lorati on.................................................................................51 4.7 Initial Concentration and Discolorati on Rates............................................55 4.8 Photocatalytic In termediates of MO...........................................................58 4.9 Adsorption of Methyl Orange on Titanium Dioxi de....................................60 4.10 Photocatalytic Degradat ion of Meth yl Or ange.........................................62 CHAPTER 5: FUNDAMENTALS OF NITROGEN -DOPING...............................69 5.1 Decreasing the Band Gap of TiO2.............................................................69 5.2 Mechanism of Visible-Light Absorp tion and Photocatalytic Activity...........70 5.2.1 Decrease in Band Gap Due to t he Overlap of N2p and O2p Orbitals.71 5.2.2 Creation of New El ectronic States Above the Valence Band .........73 5.2.3 Creation of New Electronic St ates Below the Conduction Band.........74 5.3 Creation of Oxygen Vacancie s Due to Therma l Effects............................75
iii 5.4 Nitrogen Concentration and Phases Changes Due to Thermal Effects.....76 5.5 Effect on Photocat alytic Ac tivity.................................................................79 5.5.1 Positive Effect on Photocatalytic Activity.............................................79 5.5.2 No Effect on Phot ocatalytic Activi ty.....................................................80 5.5.3 Negative Effect on P hotocatalytic Activi ty...........................................81 5.6 Reducing the Recombination Ra te of ElectronHole Pairs........................82 5.7 Current Methods for Producing Nitrogen-Doped TiO2...............................83 5.8 Formation of Titanium Nitride as a Result of Nitrogen-D oping..................84 5.9 Decreasing t he Band Ga p.........................................................................85 CHAPTER 6: EXPERIMEN TAL SYST EMS........................................................91 6.1 Thermal and Thermochemic al Treatment System.....................................91 6.2 Photocatalyt ic Reac tor..............................................................................93 6.3 Characterization of Light Source...............................................................97 CHAPTER 7: EXPERIMENTA L PROCEDUR ES..............................................100 7.1 Preparation of Thermally Treated Photocatalytic Materials.....................100 7.2 Preparation of Thermochemically Tr eated Photocatalytic Materials........101 7.3 Experimental Procedures for P hotocatalytic Ex periments.......................103 7.4 Control Experiments Using Degussa P-25 TiO2......................................106 CHAPTER 8: EXPERIMEN TAL RESULT S......................................................108 8.1 Untreated Degussa P-25 TiO2.................................................................108 8.2 Effects of Oxygen Concentrati on on Photocatalyt ic Rate........................112 8.3 Effect of Thermal Treatment on the Photocatalytic Activity of TiO2.........115 8.4 Characterization of Thermally Treated Degussa P-25 TiO2.....................119 8.5 Effect of Thermochemical Amm onia Treatment on t he Photocatalytic Activity of Degussa P-25 TiO2.......................................................................120
iv 8.6 Optimization of Thermochemical Ammonia Treatment Flow Rate...........124 8.7 Effects of Nitrification of The rmochemically Ammonia Treated TiO2.......127 8.8 Characterization of Thermoch emically Ammonia Treated TiO2...............132 CHAPTER 9: CONCLUSIONS A ND RECOMMENDAT IONS...........................137 LIST OF REFE RENCES..................................................................................141
v LIST OF TABLES Table 1 SalomonÂ’s Classificati on of Photocatalysis Â…Â…Â…Â…Â…Â…Â…Â…Â…Â…Â…5
vi LIST OF FIGURES Figure 1 Formation of Energy Bands Involving Hybrid ization ........................12 Figure 2 Formation of an El ectron-Hole Pair Due to the Absorption of a Photon ................................................................................................................13 Figure 3 Photoexcitation and De -Excitation Pathwa ys.......................................14 Figure 4 Band Edge Positions of Comm on Semiconductor Photocatalyst.........16 Figure 5 Surface and Bu lk Traps ...................................................................18 Figure 6 Stick and Ball Model of Anatase Titani um Dioxide...............................22 Figure 7 Stick and Ball Model of Rutile Titani um Dioxi de...................................23 Figure 8 Oxygen Vacancies in TiO2 Lattice Stru cture [6 ]....................................24 Figure 9 Metal as an Electron Trap....................................................................29 Figure 10 Coupled TiO2-WO3.............................................................................31 Figure 11 TEM Images of Degussa TiO2 P-25 Powder, (a) Anatase Particles (b) Rut ile Partic les................................................................................34 Figure 12 Chemical Composit ion of Meth yl Or ange...........................................37 Figure 13 Molecular Struct ure of Meth yl Or ange................................................37 Figure 14 Spectra of a 20 ppm Methyl Orange Solution.....................................38 Figure 15 Stability Spectra of Meth yl Orange Without Irradiat ion.......................39 Figure 16 Concentration as a Function of Time for Methyl Orange in the Presence of Degussa P-25 TiO2 But Without an Irr adiation S ource...................40 Figure 17 Absorbance Spectra as a Function of Time for Methyl Orange in the Presence of Calcinated TiO2 and Simulated Solar Spectrum (SSS) Irradiat ion...........................................................................................................41
vii Figure 18 C/C0 Versus Time for Methyl Orange Discoloration in the Presence of Calcinated TiO2 and SSS Irradi ation..............................................42 Figure 19 Calculation of Apparent Ra te Constant for a First Order Chemical R eaction .............................................................................................43 Figure 20 Creation of Intermediate Products During Photocatalytic Reaction of Ammonia Thermochemically Treated TiO2 Under SSS Irradiat ion...........................................................................................................45 Figure 21 Effects of Catalyst Loading (g /L) on the Rate of Discoloration for Untreated Degussa P-25 TiO2 Under SSS Irr adiatio n........................................46 Figure 22 Calculation of Rate Const ants for Catalyst Loading (g/L) for Untreated Degussa P-25 TiO2 Under SSS Irr adiatio n........................................47 Figure 23 Apparent Rate Constant for Catalyst Loading (g/L) for Untreated Degussa P-25 TiO2 Under SSS Irr adiatio n.........................................................48 Figure 24 Apparent Rate Constant for Catalyst Loading (g/L) for Untreated Degussa P-25 TiO2 Under Visible Simulated Sola r Spectrum Irradiation...........49 Figure 25 SSS and VSSS Light Intensity as a Function of Distance from the Source for a 20 ppm Methyl Or ange Solution and 1 g/L of Untreated Degussa P-25 TiO2.............................................................................................50 Figure 26 SSS and VSSS Light Intensity as a Function of Distance from the Source for a 20 ppm Methyl Or ange Solution and 4 g/L of Untreated Degussa P-25 TiO2.............................................................................................51 Figure 27 Spectra of 20 ppm Soluti ons of Methyl Orange and 1 g/L of Degussa P-25 TiO2 for Varying pH Level s..........................................................55 Figure 28 Change in Concentration as a Function of Time Due only to the Adsorption of Methyl Orange to Untreated Degussa P-25 TiO2 (3 g/L)..............57 Figure 29 Absorbance Spectra as a Function of Time for Methyl Orange in the Presence of Untreated Degussa P-25 TiO2 and Simulated Solar Spectrum Irr adiatio n...........................................................................................60 Figure 30 Effects of Catalyst Loading (g /L) on the Rate of Discoloration for Untreated Degussa P-25 TiO2 Under SSS Irr adiatio n........................................63
viii Figure 31 Calculation of Rate Const ants for Catalyst Loading (g/L) for Untreated Degussa P-25 TiO2 Under SSS Irr adiatio n........................................64 Figure 32 Shift in Spectra Due to In termediates for 20 ppm Methyl Orange Solution in the Presence of 3 g/L of Degussa P-25 TiO2 With No Oxygen Source and Under VS SS Irradiat ion...................................................................66 Figure 33 Hypothetical Band Gap of Nitrogen-Doped TiO2 Proposed by Madhsudan et al. ...........................................................................................73 Figure 34 BET Surface Area Measur ements for Pure, Calcinated and Thermochemically Ammonia Treated Degussa P-25 TiO2.................................78 Figure 35 Diffuse Reflectance of Pure, Thermochemically Ammonia Treated Degussa P-25 TiO2...............................................................................87 Figure 36 Optical Absorption (Kubelka-Munk) of Thermochemically Ammonia Treated Degussa P-25 TiO2...............................................................89 Figure 37 Block Diagram of T ube Furnace Reac tor Syst em..............................91 Figure 38 Temperature Profile for Tube Furnac e Reacto r..................................93 Figure 39 Batch Reactor for P hotocatalytic Experimen ts...................................95 Figure 40 One Liter Photocatalytic Batch Reactor with Stainless Steel Cooling Coils ......................................................................................................96 Figure 41 Spectra for Solar and Simulat ed Solar Spectrum Irradiation..............98 Figure 42 Spectra for Halogen Lamps With and Without Ultraviolet NaNO2 Solution Fi ltering.................................................................................................99 Figure 43 Original Calc ination Sy stem.............................................................101 Figure 44 Tube Furnace Reactor S ystem.........................................................102 Figure 45 Photocatalytic Batch R eactors for SSS and VSSS Irradiation Experiment s......................................................................................................105 Figure 46 Change in Concentration as a Function of Time for Methyl Orange in the Presence of 1 g/L of Untreated Degussa P-25 TiO2 Under No, SSS and VSSS Ir radiatio n.........................................................................110 Figure 47 Integrated Rate Law Plot for Untreated Degussa P-25 TiO2 Under SSS Irr adiation .......................................................................................111
ix Figure 48 Apparent Rate Constants for Untreated Degussa P-25 TiO2 Under No and SSS I rradiatio n..........................................................................112 Figure 49 Apparent Rate Constant for Untreated Degussa P-25 TiO2 Under SSS Irradiating Using O ne and Two Aerati ng Stones ............................113 Figure 50 Comparison of Air and Oxygen on the Apparent Rate Constant......114 Figure 51 HRTEM Showing the Grain Boundaries of Pure Degussa P-25 TiO2..................................................................................................................115 Figure 52 Change in Concentration as a Function of Time for Methyl Orange in the Presence of 1 g/L of Calcinated Degussa P-25 TiO2 Under SSS Irradiat ion.................................................................................................117 Figure 53 Integrated Rate Law Plot for Calcinated Degussa P-25 TiO2 Under SSS Irr adiation .......................................................................................118 Figure 54 Apparent Rate Constants for Calcinated Degussa P-25 TiO2 Under SSS Irr adiation .......................................................................................119 Figure 55 Characterization of Therma lly Treated (Calcinated) Degussa P25 TiO2 Under SSS Irr adiation ..........................................................................120 Figure 56 Change in Concentration as a Function of Time for Methyl Orange in the Presence of 1 g/L of Thermochemically Ammonia Treated Degussa P-25 TiO2 Under VSSS Irr adiation .....................................................122 Figure 57 De-coloration Decay Rate as a Function of Treatment Temperature for Thermochemically Ammonia Treated Degussa P-25 TiO2 Under VSSS Irr adiation ....................................................................................123 Figure 58 Dye Concentration as a Func tion of Treatment Temperature for Nitrogen-Doped TiO2 for Reactive Red and Phenol by Kosowska et al. ....124 Figure 59 De-coloration Decay Rate as a Function of Ammonia Flow Rate at 675C for Thermochemically Ammonia Treated Degussa P-25 TiO2 Under VSSS Irr adiation ....................................................................................125 Figure 60 Thermochemically Ammonia Treated Degussa P-25 TiO2 at 675C for 3 Hours at 12.7 mL/m in....................................................................125 Figure 61 Thermochemically Ammonia Treated Degussa P-25 TiO2 at 675C for 3 Hours at 24.8 mL/m in....................................................................126
x Figure 62 Depiction of Color Gradient After Thermochemically Ammonia Treated At or Above 24.8 mL/min .....................................................................127 Figure 63 Thermochemically Ammonia Treated Degussa P-25 TiO2 at 675C at 87.3 mL/min. for 3 Hour s...................................................................127 Figure 64 Effects of Nitride on t he Photocatalytic Effects of Thermochemically Ammonia Treated Degussa P-25 TiO2...............................128 Figure 65 Apparent Rate Constant for Thermochemically Ammonia Treated (24.8 mL/min.) Degussa P-25 TiO2 Under SSS Irr adiation..................129 Figure 66 Formation of Titanium Nitri de During Thermochemical Ammonia Treatments at 825C of Degussa P-25 TiO2.....................................................131 Figure 67 XRD Comparison for Thermochemically Ammonia Treated Degussa P-25 TiO2...........................................................................................132 Figure 68 Characterization of Thermochemically Ammonia Treated Degussa P-25 TiO2 Under SSS Irr adiation .......................................................133 Figure 69 HRTEM of Thermochemically Ammonia Treated Degussa P-25 TiO2..................................................................................................................134 Figure 70 Optical Absorbance (Kubel ka-Munk) of Thermochemically Ammonia Treated Degussa P-25 TiO2.............................................................135 Figure 71 (A) TiO2-XNX and (B) Pure TiO2 Calcinated at 400C  .................136
xi Thermochemical Treatment of TiO2 Nanoparticles for Photocatalytic Applications Mark Schmidt ABSTRACT Titanium Dioxide (TiO2) has been considered an ideal photocatalyst due to factors such as its photocatalytic proper ties, chemical stability, impact on the environment and cost. However, its app lication has been pr imarily limited to ultraviolet (UV) environments due to its high band gap (3.2 eV). This high band gap limits the harvesting of photons to approx imately 4% of sunlight radiation. Research today is focused on loweri ng this gap by doping or coupling TiO2 with other semiconductors, transition metals and non-metal anions, thereby expanding its effectiveness we ll into the visible range. This thesis explores the effects of thermal and thermochemical ammonia treatment of nano-particulated TiO2. The objective is to synthesize a photocatalyst with a lower band gap energy that demonstrates photocatalytic activity in the visible range while at t he same time retaining its photocatalytic properties in the UV range. Specifica lly, this study utilizes pure commercial nano-particulated TiO2 powder (Degussa P-25), and uses this untreated TiO2 as a baseline to investigate the effects of thermal and thermoc hemical treatments.
xii Nitrogen-doping is carried out by gas phase impregnation using anhydrous ammonia as the nitrogen sour ce and a tube furnace reactor. The effects of temperature, time duration and gas flow rate on the effectiveness of thermally and thermochemically treated TiO2 are examined. Thermally treated TiO2 was calcinated in a dry inert nitrogen (N2) atmosphere and the effects of temperature and treatment dur ation are investigated. The band gap of the thermally trea ted and thermochemically ammonia treated TiO2 have been measured and calculated us ing an optical spectrometer. The photocatalytic properties of all ma terials have been investigated by the degradation of methyl orange (MO) in an aqueous solution using both visible simulated solar spectrum (VSSS) and simu lated solar spectrum (SSS) halogen light sources. Methyl orange degradat ion has been measured and calculated using an optical spectrometer. The phas e structure and particle size of the materials is determined using x-ray diffrac tion (XRD). The BET surface area of the samples has been obtained using an Autosorb. Surface or microstructure characterization has also been obtained by scanning electron microscopy (SEM) and transmission electron microscopy (TEM).
1 CHAPTER 1: THE NEED FOR A VISIBLE-LIGHT PHOTOCATALYST Today, one of the greatest health th reats to our planet is the lack of potable water. According to United Na tions ChildrenÂ’s Fund (UNICEF), more than one billion people lack safe drinking water. Further, nearly 6,000 people die from water related illness each day. Of these two million deaths annually, the majority are children . A substantial pr oblem exists in remote areas that lack clean water and other basic essentials such as proper sanitation. These places also tend to lack the electrical power necessary for water treatment. Today, there is hope that polluted water source s can be disinfected by utilizing the energy of the sun and using a class of materials referred to as photocatalysts. Titanium dioxide (TiO2) has been considered an ideal photocatalyst due to factors such as its photocatalytic proper ties, chemical stability, impact on the environment and cost. Decomposing organic pollutants using TiO2 is currently considered a possible decont amination process that could relieve much of the worldÂ’s problems with potable water . However, TiO2 is a wide band gap (3.2 eV) semiconductor and as such its applic ation has been primarily limited to ultraviolet (UV) environments. This high band gap limits the harvesting of photons to approximately 4% of the sunÂ’s available radiation, which is far too small for practical use. Research t oday is focused on lowering the band gap of TiO2 by doping or coupling TiO2 with other semiconductors, transition metals and
2 non-metal anions, and thereby expanding it s effectiveness well into the visible range. This thesis explores the effects of thermal and thermochemical ammonia treatments on nano-pa rticulated TiO2. The objective is to synthesize a photocatalyst with a lower band gap energy that demonstrates photocatalytic activity in the visible range while at t he same time retaining its photocatalytic properties in the UV range. Specifica lly, this study will utilize pure commercial nano-particulated TiO2 powder (Degussa P-25), using this untreated TiO2 as a baseline to investigate the effects of thermal and thermochemical ammonia treatments.
3 CHAPTER 2: PHOTOCATALYSIS 2.1 Photocatalysis Defined The term catalysis can be thought of as a chemical reaction that results from the action of a catal yst. Simply defined, a catalyst is a substance that increases the rate of reaction without being consumed during the reaction. Photocatalysis is a specialized catalytic process. Over the years, the exact definition of photocatalysis has grown and evolved along with this area of research. One of the ear ly definitions of what is now known today as photocatalysis was Â“a catalytic reaction invo lving light absorption by a catalyst or a substrateÂ” . Another definition for photocatalysis, which was described as a complimentary definition, wa s a Â“catalytic reaction involving the production of a catalyst by absorption of lightÂ”. This was commonly referred to as photo-assisted catalysis . Today, catalysis is defined by Serpone et al. as Â“a process in which a substance (the catalyst) accelerates, th rough intimate interaction(s) with the reactant(s) and concomitantly providi ng a lower energy pathway, an otherwise thermodynamically favored but kinetically slow reaction with the catalyst fully regenerated quantitatively at t he conclusion of the catalytic cycleÂ” . More simply put, the catalytic process allows us to increase the rate of a reaction by
4 the addition of a substance, the catalyst, which itself remains unaltered after the reaction has completed. From its early beginnings, photocat alysis has now grown to become an extension of catalysis. Photocatalysis can simply be thought of as a catalytic chemical reaction where photons are invo lved. This photoreaction only takes place in the presence of both the cata lyst and photons of sufficient energy. Ultimately, simpler more inclusive def initions were sought. The term photocatalysis now can be thought of simp ly as a reaction where Â“light and a substance (the catalyst or initiator) are necessary entities to influence a reactionÂ” . 2.2 Photogenerated Catalysis Ve rsus Catalyzed Photolysis Under this very broad definition of photocatalysis is a number of differing reaction mechanisms. As a result, phot ocatalysis has been further refined into two main groups by Salomon et al., photogenerated catalysis, and catalyzed photolysis, as is shown in Table 1. In photogenerated catalysis, the ground states of the catalyst and substrate are involved in a thermodynamically spontaneous catalytic step . By contra st, in catalyzed photolysis the catalyst, the substrate or both are in an excit ed state during the catalytic step .
5 Table 1 SalomonÂ’s Classifica tion of Photocatalysis  Catalytic in Photons Non-Catalytic in Photons Photogenerated catalysis Catalyzed photolysis photoinduced catalytic reactions (stoichiometric photogenerated catalysis) catalyzed photochemistry catalyzed photoreaction sensitized photoreaction photosensitized reactions photo-assisted catalysis (stoichiometric photogenerated catalysis) substance-assisted photoreactions substance-catalyzed photoreactions 2.3 Principles of Photocatalysis Serpone et al. , Carp et al. [5 ], and Linsebigler et al.  provide excellent overviews of photocatalysi s with specific reference to TiO2. Many of the ideas expressed in this section come from these and other authors that have written on this thoroughly examined subject. The reader is encouraged to see the complete writings for detailed analysis. The concept of photocatalysis can be thought of as an extension of a catalytic reaction where some compound A is converted to some compound(s) B: B A The use of a catalyst in this reacti on changes the rate of the reaction, but the catalyst itself, by definition, must be preserved in both quantity and state, and
6 is completely separable from the other reactants at the completion of the catalytic cycle. This reaction which is now a catalytic process is given by: Catalyst B Catalyst A Although the reaction mechanics for t hese two processes are different, the end result is the same with a resultant species (compound B). However, in the catalytic process the catalyst in its origi nal state is also one of the final products. Regarding the actual rate of reaction of these two processes, the reaction rate will be increased using a catalyst only if the total activation energy is lower than that of the reactants alone. If this reaction occurs as the result of the absorption of photons by the catalyst, then the process is photocatalytic. In this case, the reaction is give by the equation: Catalyst B Catalyst hv A where hv is a quantum of energy from the incident photons sufficient to cause the reaction. In regards to this thesis, the photocatalyst is a nano-particulated TiO2 powder, Degussa P-25. For solid material photocatalysts, the pollutant is first adsorbed onto the surface of the catal yst. Here the pollutant undergoes a chemical transformation as it comes into c ontact with the reactive surface sites of TiO2. Finally, the intermediate or final product is desorbed from the catalyst.
7 The adsorption mechanism of the polluta nt onto the surface of the catalyst can be accomplished by either chemisorpti on or physisorption. Chemisorption is adsorption of the pollutant due to a strong chemical bond at the interface of the catalyst and pollutant. Alternatively, physisorption is a non-chemical bond between the pollutant and catalyst that is due to weaker forces such as van der Waals forces. In the event that the adsorption is an unusually strong chemisorption bond, the photocatalyt ic effect can be degraded or even eliminated, and is instead c onsidered stoichiometric ca talysis. At the other extreme, if the bond is too weak, the pollutant is poorly adsorbed, not allowing for bond rupturing or bond making, which also degrades or eliminates the reaction. The bond between the catalyst and pollutant must be strong enough to allow for this bond braking and making within the re sidence time of the intermediates, while allowing for both adsorpt ion and desorption to occur. From the reactions given above, the process of photocatalysis can be seen from two distinct positions. First, progressing from a chemical reaction to catalysis and then to photocatalysis. In th is case, the photocatalysis is caused by catalysis of a thermal reaction. W hen light is absorbed by the catalyst it creates an excited state pr oducing active surface site s. The second possible cycle is from a chemical reaction to a phot ochemical reaction to photocatalysis. After the catalyst absorbs photons, surfac e photochemical processes are said to be catalysis of a photoreaction as light is her e considered as one of the reagents. This photochemical reaction is a func tion of the absorbed photons creating
8 intermediate species, the excited state of the catalyst, which in turn reacts with another reagent forming the final reaction species. As it relates to this paper, photogenerat ed catalysis will be the focus, and specifically the LangmuirHinshelwood process. This process occurs at photochemically active sites at the surf ace of the catalyst when the necessary quantum of energy is absorbed by the catalys t. This in turn causes the creation of electron-hole pairs enabling electrons and holes to migrate to the surface. At the surface electrons and holes are reacti ve species capable of reducing and or oxidizing a pollutant. The mechanism of the Langmuir-Hinshelwood process is given by : 1. adsM S M 2. S M Mads 3. h e hv Catalyst 4. ads adsM h M 5. ads adsM e M 6. products S Mads where M is a reagent and S is a surface si te. This mechanism starts with the adsorption of the reagent ont o the catalyst surface site. The second step shows desorption of the reagent. This adsorp tion-desorption process creates what is
9 referred to as a Langmuir equilibrium, which was a topic of considerable study for this thesis. The third step shows the absorption of some necessary quantum of energy from a photon by the catalyst cr eating an electron-hole pair. Step four shows the creation of a reactive specie s due to hole trapping. Step five shows the decay of this species through the recombination of this species with an electron. Ultimately, step six shows t hat the photocatalytic process concluded with the formation of some products and the surface of the catalyst returning to its original state . In contrast to photogenerated catalysis, if the photocatalytic reaction was categorized as catalytic photolysis it could be considered a catalytic photoreaction if the initial photoexcitation occurred in the adsorbed molecule that interacted with the ground state of the catalys t. Alternatively, it is categorized as a sensitized photoreaction if the photoexcitation occurred initially in the catalyst substrate, which in turn transferred an electron into the ground state of the reagent molecule. Regardless of the proces s in which the initial excitation of the system occurred, the transfer of the elec tron or energy transfer, it is the deexcitation that causes the chemical reaction as shown in steps three and four above . The results of this study sugges t that more than one possible process occurs, and it was possible to speculate t hat sensitized photoreactions took place during this investigation under certain conditions. The transfer of an electron that causes the initial excitation in either the donor or acceptor molecule is a one el ectron reaction where the electron jumps
10 from the donor reactant to t he empty orbital of the accept or reactant. For this to be possible, the orbitals of the accept or and donor must over lap with either an empty or half-filled state availa ble resulting in an ion pair. When the initial excitation is due to an energy transfer, the process is caused by electron exchange or dipole-dipol e resonant coupling. In either case, here the orbitals must again overlap for this excitation. Further, if electron exchange is the mechanism, two independent one-electr on transfer steps, one in each direction, is required. In contrast, the dipole-di pole resonant coupling is an interaction between the Â“oscillating dipole of an excited state molecule coupled with the induced dipole in a ground state quencher moleculeÂ” . 2.4 Electronic Excitation The first step in photocatalysis is t he absorption of a photon with sufficient energy by the reagent or substrate. This photon absorption creates highly reactive species with electronically excited states. The absorpti on rate of energy is on the order of 10-15s, and occurs considerably quick er that the de-excitation event. The rate of de-excitation is det ermined by the pathway that will minimize the lifetime of the reactive species. This pathway can be by emission of radiation, or radiation-less decay, and if it is emitted as radiation, as in fluorescence, the carrier lifetime of the excited st ate is on the order of 10-9 to 10-5s .
11 2.5 Band Gap Excitation Semiconductors are identified as so lids whose electrical conductivity is determined by the amount of energy that is required to move electrons from the valence band to the conduction band, whereas metals have a Â“sea of electronsÂ” available for conduction. In traditional se miconductors, such as silicon, hybrid orbital (sp3 orbital) of silicon can overlap with other atoms to form two different molecular orbitals as seen in Figure 1. Th ese orbitals are said to be in phase if both are negative or positive (also commonl y referred to as up or down), which forms a bonding molecular orbital with energy Eb. The orbital can be out of phase if one is positive and one negative, which forms an anti-bonding orbital with energy Ea. The bonding orbital of the so lid will interact splitting the Eb energy level into discrete energy levels forming a valence band Vb by virtue of the valence electrons that are contained in this band. The energy level Eb is full, and therefore the valance band of the se miconductor is full of electrons. Similarly, the anti-bonding orbital overl ap and also split into discrete energy levels Ea, which forms the energy band that is completely emptied. The energy band formed by the anti-bonding orbital is referred to as the conduction band Cb. This conduction energy band is separated from the valence band by the defined energy gap Eg, which in the ideal sense has no energy states. The stateless energy gap Eg is said to extend from the top of the valence band to the bottom of the conduction band. Again, in the ideal ized sense, electrons cannot be present in the energy gap Eg, and therefore there is no reco mbination of electrons and holes that may be photo-activated in the solid .
12 Figure 1 Formation of Energy B ands Involving Hybridization  For photoexcitation of an electron from the valence band to the conduction band, a photon of sufficient energy must be absorbed by the electron. This excitation causes the creation of the electron-hole pair (e-, h+) and is shown in Figure 2. The lifetime of this electronhole pair is on the order of nanoseconds, which is sufficient time to undergo a charge transfer with the adsorbed species on the surface of the semic onductor catalyst. This can be done in solution or in the gas phase. The term heterogeneous phot ocatalysis is used to describe the charge transfer to the adsorbed species if the semiconductor catalyst remains intact during a continuous exothermic process .
13 Ec+ + + + + + + + + + Ev hv dn Figure 2 Formation of an Electron-Hole Pair Due to the Absorption of a Photon The breakout of Figure 3 shows t he photoexcitation of an electron and the creation of an electr on-hole pair. Photocatalysis of either organic or inorganic compounds begins with the absorption of a quantum of energy greater than, or equal to, the energy gap of the semiconducto r photocatalyst. Once created, this electron-hole pair has several de-excita tion pathways which the electrons and holes can follow . For photocatalysis to occur, thes e photogenerated electrons and holes must migrate to the surface of the catalyst where they can be transferred to the adsorbed organic or inorganic pollutant as shown in Figure 3. However, in competing de-excitation pathways, both surface and volume recombination can occur, as denoted by pathways A and B. T he electron-hole recombination rate, if too high can degrade, or even halt photocat alysis . The concept of charge separation, by any number of means, is an important idea as it relates to doped semiconductor catalysts, and is explored further in this study. Pathway C depicts the transfer of an el ectron that has migrated to the surface of the catalyst to an acceptor, whic h in turn reduces the acceptor. In the aqueous batch reactor used in this study, th is acceptor is oxygen. Pathway D
14 shows the migration of a hole to the catalyst surface that is then used to oxidize a donor species by the transfer of an electr on from the donor. The occurrence and rate of these two mechanisms are functions of probability. This is determined by their band edges relative to the valence and conduction bands in relation to the redox potentials of the adsorbed pollutant . h v + CB VB + h v + + + + + A AD D+ C A B D Figure 3 Photoexcitation and De-Excitation Pathways The efficiency of the photocatalyst in degrading the pollut ant is therefore dependent upon the diffusion of the electr on-hole pair to the surface. Additionally, in a non-idealized semiconduc tor catalyst, surface traps exist and the concentration of holes and electrons at the surface are not equal . Methods are employed for surface tr apping of these charge carriers at the surface, which will be discussed below.
15 2.6 Band Edge Position As noted, the mechanism for a photo-i nduced electron to transfer to an adsorbed pollutant on a semiconductor cata lyst is controlled by the band energy position and the redox potent ial of the adsorbed species. Referring to the Normal Hydrogen Electrode (NHE) scale on Figur e 4, for an oxidation reaction to occur the potential level of the donor species must be above (more negative) then the valence band position of the semiconductor catalyst. Likewise, for a reduction reaction to take place the potentia l level of the acceptor species must be below (more positive) that the conduc tion band position of the semiconductor catalyst. It should be noted that the band edge position of the semiconductors catalysts shown in Figure 4 are given for a pH=1, and that the pH of the solution influences the band edge position of the semiconductor .
16 VacuumNHE 1.0 -4.5 -1.0 0.0 2.0 3.0 -3.5 -5.5 -6.5 -7.5 -4.0 -5.0 -6.0 -7.0 -8.0 -0.5 0.5 1.5 2.5 3.5 WO32.70 eV TiO23.20 eV ZnO 3.20 eV ZnTiO33.06 eV ZnFe2O41.90 eV SrTiO33.40 eV Fe2O32.20 eV CdS 2.40 eV ZnS 3.60 eV -3.0-1.5 H+/H2Reduction Potential O2/H2O Oxidation Potential Figure 4 Band Edge Positions of Common Semiconductor Photocatalyst 2.7 Charge Separation, Trapping and Recombination Given that two of the pathways out lined above lead to surface or volume recombination, charge separation or charge trapping is needed to reduce the probability of recombination and increase the photocatalyt ic effect. Since the crystal structure of the photocatalyst is not pure, but instead has both surface and bulk defects, it is expected that surf ace states (or charges) exist across the surface. These surface states, which di ffer in energy from the bulk, serve as charge carrier traps. The carrier lifet imes of the electrons and holes are therefore increased since these traps stif le the recombination of electrons and holes .
17 The problem of recombination is es pecially acute for doped materials, one of the topics of this stud y. For doped semiconductor catalysts both lattice defects and additional electronic states greatly in crease the probability of recombination. Figure 5 shows a simplified view of electron trapping in the bulk and at the surface of the semiconductor. Both the localized bulk and surface traps in this diagram are located in the band gap. When an electron is trapped at one of these sites it becomes localized at that site on the surface or in the bulk. It is the energy difference between the traps and the bottom of the conduction band and the decrease in entropy due to electron trapping that determines the population of the charge carriers in the traps . Similarly, valence band holes can also be trapped. Oxygen is commonly considered to scavenge electrons and is a process that is investigated and reported on in this study.
18 Energy Surface Traps Bulk Trap + h vVB CB Eg Figure 5 Surface and Bulk Traps  2.8 Quantum Size Effects Quantum size effects are considered for particles that range in size from 1 to 10 nm. The average particle size for pure Degussa P-25 TiO2, which is used in this study, is on the order of 10 to 30 nm, and should not experience these effects. However, this is an import ant aspect as many studies suggest an optimum particle size between 10 and 20 nm, and will therefore be considered here. When the particle size becomes comparable to the de Broglie wavelength of the charge carriers in the photocat alyst, certain phenomena occur. The effective mass of the quantum particles of the photocatalyst is the determining factor on what size of particles experi ence quantum effects. Unlike particles larger than 10 nm, the electrons and holes of quantum sized particles do not experience the electronic delocalizat ion present in a bulk semiconductor possessing a conduction band and valence band. This is due to the fact that the electrons and holes are confined in a pot ential well that has small geometrical dimensions. It is this confinement that creates the quant ization of discrete
19 electronic states. The result is an in crease in the effective band gap of the semiconductor . It has been shown through modeling an d experiment that the effective band gap can be significantly increased through quantization. This increase is most pronounced in semiconductors with small band gaps and can be quite significant, increasing by a fa ctor of six for materials su ch as lead sulfide (PbS). Therefore, considerations of quantum size effects are im portant in photocatalytic materials [6, 8].
20 CHAPTER 3: TITANIUM DIOXID E AS A MODEL PHOTOCATALYST 3.1 The Importance of Titanium Dioxide Titanium dioxide (TiO2) is an indirect gap, n-type semiconductor with electrons as the majori ty carriers . TiO2 has been considered an ideal photocatalyst due to its photocatalytic pr operties, chemical stability, impact on the environment and cost. However, its application has been primarily limited to ultraviolet (UV) environments due to it s high band gap (anatase = 3.2 eV). This high band gap limits the harvest ing of photons to approximately 4% of the solar radiation. The importance of TiO2 as a photocatalyst lies in its ability to oxidize a large number of organic compounds into harmless compounds such as carbon dioxide and water using ultraviolet li ght, and potentially, visible-light. TiO2 is highly photoactive as the H+/H2 reduction potential and O2/H2O oxidation potential lie within its band gap . This is accomplished by the redox energy of electron-hole pairs on the adsorbed pollutan t. The important po int is that the conduction and valence bands of TiO2 lie in energetically favorable positions to both reduce and oxidize the adsorbed s pecies. This means that TiO2 not only has the oxidation potential to degrade pollu tants, but also the reduction potential
21 necessary for splitting water molecules to create hydrogen gas, a topic of great importance today. 3.2 The Lattice Structure of Anatase and Rutile TiO2 Of the numerous forms of TiO2, only rutile, anatase and brookite occur in nature. Of the thr ee, only anatase and rutile can be utilized as photocatalysts. The lattice structure for anatase and rutile is described in terms of distorted TiO6 octahedra . As seen in Figure 6, this configuration consists of Ti4+ ions surrounded by six O2 ions. The crystal structure of rutile has each octahedron in contact with 10 neighboring octahedrons. The rutile structure has two octahedrons sharing the edge oxygen pairs and eight octahedrons sharing the corner oxygen atoms. In contrast, the anatase crystal structure, shown in Figure 7, has each octahedron in contact wit h eight neighboring octahedrons. The anatase structure has four octahedron sharing the e dge oxygen pairs and four octahedron sharing a corner oxygen pair [6 ]. The structural differences of the anatase and rutile structures are small, but electronically significant. The crystal structures differ by the amount of dist ortion of each octahedron, and by the pattern of the octahedra chains . The structure of anatase, as seen in Figure 7, is significantly distorted compared with rutile which results in a gr eater TiÂ–Ti distances than rutile, 3.79 and 3.04 angstrom versus 3.57 and 2.96 angs trom, but a shorter TiÂ–O distance, 1.934 and 1.980 angstrom versus 1.949 and 1.980 angstrom .
22 Figure 6 Stick and Ball Model of Anatase Titanium Dioxide The structure of rutile, depicted in Figure 7, is de scribed as having a slight orthorhombic distortion . It is these small struct ural differences that are responsible for the differences in elec tronic band properties. As mentioned, the band gap of anatase is 3.2 eV whereas the band gap of rutile is 3.0 eV. Despite the lower energy required for electron-hol e pair creation, anatase is reported to be more active than rutile . Titanium Oxygen
23 Figure 7 Stick and Ball Model of Rutile Titanium Dioxide One of the most important features of the crysta l structure is the three types of oxygen vacancy sites shown in Figure 8 [6, 10]. These oxygen vacancies play an important role in phot ocatalysis and their effects are explored in detail below. Titanium Oxygen
24 Figure 8 Oxygen Vacancies in TiO2 Lattice Structure  3.3 Electronic Structure of TiO2 The filled valence band is compos ed of the O2p orbi tal and the empty conduction band is composed of the Ti3d, Ti4s, and Ti4p orbital. The Ti3d orbitals dominate the lower portion of the conduction band . 3.4 Photoadsorption and Photodesorption of Oxygen on Ti02 The photoadsorption a nd photodesorption of O2 on TiO2 is an important mechanism of TiO2 photocatalysis. Adsorbed ox ygen participates in electron charge transfer from the substrat e, and electron scavenging during photooxidation of organics. In separate studies by Bickley et al.  and Nakamura et al.  it was concluded that adsorbed water enhances the photoadsorption of oxygen on rutile TiO2 surfaces. This is believed to be done by trapping the photogenerated holes at OHsites. The formation of surface
25 2HO by electron trapping and 3O by hole trapping reactions is the primary route for oxygen photoadsorption . 3.5 Photooxidation at the Li quid-Solid Interface of TiO2 When a photon with the energy required to excite an electron from the valence to conduction band is absorbed, an electron-hole pair is created and potentially can participate in a redox reaction. The reduction potential and oxidation potential is equiva lent to the band energy, which can be estimated. It is also possible that the elec tron-hole pair may recombine by several mechanisms as described above. For a photocatalytic reaction to take place however, the electrons and holes must migrate to the surface of the catalyst and react with adsorbed species. In a steady state photocat alytic reaction, the rate of oxidation by the holes has to be balanced by the rate of reduction by the electrons . After the initial creatio n of the electron-hole pair the actual process of oxidation taking place at the interface of the catalyst and solution is an area of significant debate. There are two genera lly accepted oxidat ion mechanisms. One is the direct oxidati on of the adsorbed pollutant by the hole. The other is oxidation by the adsorbed hydroxyl r adical attacking the adsorbed organic species. While research has found hydr oxyl radicals and hy droxylated oxidation intermediates in support of that mechanism it is not possible to distinguish this from the direct hole oxidat ion mechanism given that the reaction intermediates are similar .
26 The study by Draper et al. concl uded that the primary mechanism of oxidation was by the direct hole oxidat ion route . In the study they hypothesized that trapped electrons reac t with pre-adsorbed molecular oxygen and produce 2O and 2 2O anions. The study itself does not reach a conclusion regarding the actual role these species pl ay. It is possible that these anion directly oxidize the pollutant, create hy droperoxide radicals and hydroxyl radicals, or continue reacting with other trapped electrons whic h may produce water . In addition to scavenging electrons duri ng photooxidation, oxygen may play a larger role in photocatalysis. Other el ectron scavengers achieve only a fraction of the oxidation and photom ineralization effect as oxygen. This has led researchers such as Linsebigler et al. to speculate on a larger role for oxygen in photocatalysis . 3.6 Trapping, Recombination, a nd Interfacial Electron Transfer Carrier lifetime is an impor tant limiting factor in t he photocatalytic rate and illustrates the importance of the electron or hole tr apping species to be preadsorbed on the surface of the catal yst. Conduction band electrons may be trapped within 30 ps . This is compar ed with the trapping of the valence band holes that require an average time of 250 ns is very fast. The recombination rate of trapped electrons with free or trapped holes occurs between 10-11 to 10-6 s . The mean lifetime of a si ngle electron-hole pair is approximately 30 ns for a low charge carrier concentration and hole trapping can compete with the recombination process . Trapped holes are relatively un-reactive with
27 electrons. For high charge carrier c oncentrations, the electron-hole pairs recombine within a fraction of one nanosecond . This illustrates that charge carrier trapping must happen very fast to achieve an effective photochemical conversion . The rate of photooxidation was found by Gerischer et al. to be equal to, and limited by, the reduction rate of dissolved oxygen in the solution . Their study theoretically predicted that if O2 is not reduced at a sufficiently high rate, electrons will accumulate on the surface of the catalyst. This was later confirmed in the study by Wang et al. who dete rmined that the rate of radiation-less recombination of electrons and holes is Â“enhanced until the sum of the electronhole recombination and t he electron transfer to O2 is equal to the rate of the hole photogenerationÂ” . This is the basis fo r using metals such as palladium or platinum for scavenging to elimin ate electron accumulation on the Ti02 . Although not reported in this study, the effect of copper ions from fresh copper surfaces may have played a role in t he photocatalytic rate of the initial experiments. What was found was an in crease in the photocatalytic rate for fresh copper surfaces, which degraded wit h time. These copper ions could potentially have come from the copper c ooling coils originally designed for the photocatalytic reactors. Stainless st eel replaced these copper coils and no variations in the photocatalytic rate were recorded subsequently.
28 3.7 Photomineralization Reactions in Aqueous Photocatalytic Suspensions As possibly the most heavily researched photocatalyst, TiO2 has been tested with virtually every class of or ganic compound . It has been concluded that valence band holes can oxid ize any organic compound to CO2, H2O and some form of mineral acid. The photoo xidation and photomineralization process are non-selective. The Â“ photogenerated holes in the semi conductor particles, the hole-trapping radical species, and the ac tivated oxygen species (by electron trapping) are all strong oxidation agents for organic compoundsÂ” . 3.8 Modifications of TiO2 for Improved Photocatalytic Performance The premise of this study is the use of TiO2 as a visible-light photocatalyst. To achieve this goal it is necessary to modify TiO2 to improve the visible-light absorption, prevent or delay charge ca rrier recombination and improve its surface properties. To achieve these goals, it is necessary to modify TiO2 to achieve each particular goal. 3.8.1 Surface Modification Using Metals Modifying TiO2 using a metal to change the surface properties can improve the photocatalytic reaction ra te, and also change the intermediate products [6, 10]. Figure 9 illustrates a Sc hottky barrier that is created at the interface of the catalyst and metal. Platinum has been studied extensively and found to form particle clusters on the ca talyst surface covering approximately 6%
29 of the surface area allowing for a large TiO2 surface area for adsorption of the pollutant to the catalyst surface. Figure 9 Metal as an Electron Trap The required quantum of energy creat es an electron-hole pair and the metal acts as an electron trap for the elec tron migrating to the surface. This suppresses the recombination of the electron-hole pair. The hole is then able to migrate to the surface of the ca talyst and oxidize the adsorbed organic compound . Further, me tals such as platinum and silver have their own catalytic effect. Metals change the photocatal ytic properties of TiO2 by changing its electrical properties due to the distribut ion of electrons that occurs. At the heterojunction, the Fermi levels of the metal and semiconductor align resulting in a flow of electrons from the catalyst to t he metal. This leads to an increase in hydroxyl groups that play an important role in the p hotocatalytic reaction as described above .
30 3.8.2 Transition Metal Doping Doping TiO2 using transition metals ions has also been widely studied. Similar to metal doping, the use of transition metal ions allows for electron traps that suppress electron-hole recombinati on. Preliminary work using iron (Fe3+) was done during this study. As was noted above, the original photocatalytic reactor used copper cooling coils in the so lution. The effect of copper ions from the experimental setup on the photocatalytic rate was found to be a rate determining factor. Doping TiO2 using Fe3 + results in an increase in Ti3+ intensity by trapping electrons thereby suppressing electr on-hole recombination . It was concluded that a doping threshold exis ts where only small concentrations produce a positive effect on t he photocatalytic rate. It s hould be noted that not all transition metals produce a positive resul t. Some transition metals actually decrease the photocatalytic rate due to an increase in electron-hole recombination by creating re combination centers . 3.8.3 Coupled Photocatalysts Coupling or co-doping was also brie fly studied using tungsten oxide (WO3) during this study. This method is used to exploit the low band gap of one material to produce a photocatalytic effe ct in a wide gap material such as TiO2 by increasing the charge separation and extending the energy range of photoexcitation for the system. Figure 10 shows the valence band and
31 conduction band positions for TiO2 and WO3 prior to contact. If a photon that is not energetic enough to excite TiO2, but is energetic enough to excite WO3 is incident, the hole that is created in the WO3 valence band is excited to the conduction band of TiO2, while the electron is transferred to the conduction band of TiO2. It is this electron transfer that increases the charge separation and increases the efficiency of the photocatalytic process. After separation, the electron is free to reduce the adsorbed pollutant and the hole is available to oxidize as described above . hv TiO2 WO3 + Figure 10 Coupled TiO2-WO3 3.9 Degussa P-25 TiO2 Degussa P-25 powder titanium dioxi de was used for this study, and is considered to be the benchmark phot ocatalytic material for TiO2 powders. An exhaustive study on TiO2 was performed by Diebold et al. , and it is an excellent manuscript for detailed analys is. P-25 is primarily composed of anatase and rutile phases. Published values for the composition of P-25 typically
32 range from 20-25% for rutile and 75-80% for anatase. Ohno et al. also report that an amorphous phase also accounts for approximately 1% . The BET surface area for P-25 is approximately 50 m2/g, which is considerably smaller than some studies that report surface areas as high as 200 m2/g for TiO2 prepared by a titania precursor. Degussa P-25 TiO2 is produced by an aerosol process using titanium tetrachloride as a precursor. The v apor phase hydrolysis method is used in a hydrogen flame to produce the powder . The sizes of the particles vary from 25 nm to 85 nm in diameter, and grow unt il their size, or the agglomerateÂ’s size, causes them to be removed from the reaction due to their weight. The morphology and relationship of the anatase and rutile phases was a topic of considerable discussion during this study. The initial hypothesis was that anatase particles were converted to rut ile particles at elevated temperatures during either calcination or doping proc esses, especially at temperatures in excess of 600C. The review of the literature showed that Bickley et al. performed one of the original studies on the morphology of Degussa P-25 and in fact made a similar conclusion. Bickley et al. concluded that some of the anatase phase particles were covered with a thin layer of rutile phase TiO2. They reported that the increased photocatalytic activity was, in part, due to this layered anatase-rutile particle. They speculated that a potential difference across the space charge region of the
33 two phases, and localized electronic states from the amorphous phase particles were the mechanism for this increase . However, both Datye et al., and later Ohno et al., concluded that anatase phase particles do not have an over-layer of rutile phase on the surface. Instead, anatase phase and rutile phase particles exist completely separate from one another [17, 18]. It has been reported in many places that the combined presence of anatase and rutile phases enhances the photoc atalytic effect of Degussa P-25. As noted, the average surface ar ea of P-25 is approximately 50 m2/g, and anatase particles have a larger surface area than rutile particles. Ohno et al. concluded that the larger surface area of anatase particles Â“improves the efficiency of decomposition of the pollutant in air and waterÂ”. Further, that the Â“presence of both anatase and rutile phases is important for some photocatalytic reactions where oxygen is used as the electron acceptorÂ” . Datye et al. concluded that the Â“well developed crysta llinityÂ” was responsible for the high efficiency. They reasoned that this a llowed for a low density of recombination centers . Figure 11 shows the TEM images of P-25 powder from Ohno et al. They concluded that the images clearly indicate that the two phases exist separately. Ohno et al. also concluded that anatase and rutile particles grow on different nuclei, which also was not the original hypothesis of this study. As stated previously, the original hypothesis was that rutile was purely a thermal phase
34 transformation of anatase . This author concurs with the conclusions of Ohno et al., and has used this model throughout this study. Figure 11 TEM Images of Degussa TiO2 P-25 Powder, (a) Anatase Particles (b) Rutile Particles Experiments were conducted for calc inated P-25, which will be detailed below, that show improved photocatalytic efficiency. The improvement in efficiency can be attributed to reduction of TiO2, which was explained in detail above. This reduction causes a red-sh ift to wavelengths longer than 400 nm, which is attributed to the reduction of Ti4+ to Ti3+. Hence, the rutile particles in the P-25 powder are considered to contain the Ti3+ ions that create electron donors. Ohno et al. conclude that a Â“fairly la rge band bending is generat ed in the rutile particles.Â” They attribute the higher efficiency of P-25 to this .
35 CHAPTER 4: METHYL ORANGE AS A MODEL POLLUTANT 4.1 The Real World Probl em of Methyl Orange In this thesis, methyl orange (MO, C14H14N3SO3Na) was used as the model pollutant. Methyl Orange is a common industrial dye favored for its stability and is categorized as an azo-dy e. Azo-compounds, which are synthetic inorganic chemical compounds, account for up to 70% of the dyes in use today. It is estimated that between 10-15% of the dye used in textile processing is lost and released as effluent . The releas e of this effluent is considered Â“nonaesthetic pollutionÂ” as amounts smaller than 1 ppm is visible in water sources. Although this is the primary motivati on for degrading methyl orange, the dye waste water can also produce dangerous by-products during various chemical reactions such as oxidation and hydrolysis . As stated, these azo-co mpounds are very stable, which is due to the large proportion of aromatics in the dye. Biol ogical treatments may only de-color the dye effluents as opposed to degrading the effluent. A similar de-coloration phenomenon was discovered during this st udy for thermochemically ammonia treated Degussa P-25 TiO2, and is discussed in detail below. Common physicochemical treatments are effective in disco loration, but are also non-destructive.
36 Instead, these treatments transfer the or ganic compounds fr om the water to another phase . Due to the difficulty in degrading pollutants such as MO, a process referred to as an Advanced Oxidation Pr ocess was proposed as an alternative to water purification. This process differs from the traditional oxidation by holes, with oxidation by a very reactive species such as hydroxyl radicals OH These reactive species can non-selectively oxidize a broad range of pollutants. It has been determined that heterogeneous ph otocatalysts, such as TiO2, are the most destructive with regard to azo-co mpounds . This destruction can be promoted by both artificial solar s ources and sources employing solar technologies. The noted advanta ge to this method is that it is destructive, and that it can occur under ambient condition s. Further, it has been shown that it may lead to the complete mineralization of organic carbon into CO2 . 4.2 Chemical Composition of Methyl Orange The chemical formula and molecular composition for methyl orange, depicted in Figures 12 and 13, is C14H14N3SO3Na. It is the presence of the benzene rings which keeps this pollutant from decomposing easily by chemical or biological methods .
37 Na+, -O3S N=N N CH3 CH3 Figure 12 Chemical Composition of Methyl Orange Figure 13 Molecular Structure of Methyl Orange Initial experiments were conducted using de-ionized water as the solvent, to determine the optical absorption spec tra of methyl orange. Figure 14 shows the spectra for a 20 ppm solution of me thyl orange. These spectra show two absorption peak maxima, one at approx imately 272 nm and a second with a higher absorption magnitude at 451 nm. C onsistent with published work in this area, the second maxima peak at 451 nm was used to calculate the concentration changes as a function of ti me for methyl orange [20, 22]. These peaks and corresponding spectra are in line with published works depicting absorption peaks at 270 nm and 458 nm .
38 0 0.2 0.4 0.6 0.8 1 1.2 1.4 240290340390440490540590Wavelength (nm)Absorbance MO Init Figure 14 Spectra of a 20 ppm Methyl Orange Solution 4.3 Chemical Stability of Methyl Orange As noted earlier, methyl orange is considered to be a very stable compound. To test this stability methyl orange was prepared in varying concentrations from 5 ppm to 20 ppm. Samples were syringed into centrifuge tubes and exposed either to ambient light or shielded from a ll light sources. Figure 15 shows the measured spectra for th ose samples measured at time zero, after 24 and after 144 hours. As is rout inely reported elsewhere, there is a negligible change in concentration for all samples.
39 0 0.2 0.4 0.6 0.8 1 1.2 1.4 240340440540640Wavelength (nm)Absorbance 20 ppm Baseline 20 ppm Room 144h 20 ppm Dark 20 ppm Room 24h Figure 15 Stability Spectra of Methyl Orange Without Irradiation To test the stability of MO in the presence of TiO2, but without a light source, an experiment was conducted. A 20 ppm MO solution prepared in a glass beaker was loaded with 1 g/L of TiO2 and stirred in the dark with samples taken after 30 minute intervals. The results are depicted in Figure 16. Consistent with the literatur e, there was no degradation of the MO stirred in the dark, and in the presence of TiO2 [20, 22]. However, there is an adsorptiondesorption process that along with experimental error acco unt for the insignificant changes in concentration. This adsorpti on process, referred to as the Langmuir equilibrium, will be discussed below.
40 0.8 0.85 0.9 0.95 1 1.05 1.1 050100150200Time (m)C/C0 01-27-06 MO 20ppm Figure 16 Concentration as a Function of Time for Methyl Orange in the Presence of Degussa P-25 TiO2 But Without an Irradiation Source Guettai et al. found that the complete disappearance of the dye could only be observed in the simu ltaneous presence of TiO2 and UV-A (320 nm to 400 nm) light. They concluded that the system is therefore Â“working in a pure photocatalytic regimeÂ” . These st ability findings are the baseline of the photocatalytic study included in this thesis. 4.4 Calculation of Rate Constants Consistent with published works r egarding the degradation of methyl orange by TiO2, the peak at 451 nm was used as the evaluation point [20, 22]. Figure 17 shows the spectra for MO as a fu nction of irradiation time for the whole spectrum.
41 0 0.2 0.4 0.6 0.8 1 1.2 1.4 240290340390440490540590Wavelength (nm)Absorbance 0 min 30 min 60 min 90 min 120 min 150 min 180 min 210 min 240 min 300 min 360 min MO Init Figure 17 Absorbance Spectra as a Function of Time for Methyl Orange in the Presence of Calcinated TiO2 and Simulated Solar Spectrum (SSS) Irradiation The value for 451 nm was derived us ing a box car smoothing method for the values between 449-453 nm to account for the fluctuations of data points recorded by the optical spectrometer. The ratio of the concentration versus initial concentration for 451 nm was then plotted as a function of time as depicted in Figure 18. Typically, this curve wa s a good approximation of a first order chemical reaction for simulated solar s pectrum (SSS) irradiation as shown in Figure 18. However, for visible-light irradiation the curve more accurately reflected a zeroth order chemical reaction. These results are the first indication of the difficulty in characterizing the photoc atalytic chemical kinetics as strictly first order reactions, and therefore t he motivation for developing the LangmuirHinshelwood equation.
42 y = 0.972e-0.0083xR2 = 0.9830 0.2 0.4 0.6 0.8 1 1.2 020406080100120140160180200Wavelength (nm)C/C0 08-23-06 TiO2 525C N2 3h Whole MO 20ppm NR NC Expon. (08-23-06 TiO2 525C N2 3h Whole MO 20ppm NR NC) Figure 18 C/C0 Versus Time for Methyl Orange Discoloration in the Presence of Calcinated TiO2 and SSS Irradiation From the plot of concentration versus initial concentration a rate constant for first order kinetics was calculat ed. The method used here mirrors the technique employed by both Guettai et al and Al-Qaradawi et al. for determining rate constants for first or der kinetics. The negativ e log plot of ratio of concentration to initial concentration (s ee Figure 19) is plotted and the tangent provides the apparent rate constant for each material.
43 y = 0.0074x + 0.1622 R2 = 0.99780 0.2 0.4 0.6 0.8 1 1.2 020406080100120140Time (m)-ln(C/C0) 08-23-06 TiO2 525C N2 3h Whole MO 20ppm NR NC Linear (08-23-06 TiO2 525C N2 3h Whole MO 20ppm NR NC) Figure 19 Calculation of Apparent Rate Cons tant for a First Order Chemical Reaction Yates et al. implored that Â“great ca reÂ” must be used in the design of the experimental setup in order to different iate between the discoloration of the dye due to photocatalytic activity, and the disco loration of the dye due to reduction as a result of electron transfer reactions . This caution again illustrates the difficulty in properly characterizing the reaction kinetics. The spectra in Figure 20 show the pr oblem with using the value at 451 nm as the only point for determining MO degrad ation. As can be seen, the complete degradation of the pollutant has not occu rred, but instead intermediates have been formed that show absorption at shorter wavelengths. It is obvious that such a result cannot be used for evaluating a ra te constant for the photocatalyst.
44 Further, it cannot be concluded that the resulting pollutant species will be fully degraded. The spectra only show an increa se in absorbance at 350 nm due to the creation of intermediates that corres ponds to the decrease at 451 nm of the original pollutant. It is not possible to conclude that the species identified at 350 nm will degrade. Although rare, similar s pectra were found for both visible and whole light irradiation expe riments during this study. Although not concluded, it is hypothesized that the abs orption of ultraviolet li ght has been retarded by the thermochemical ammonia treatment. T he result here and in other experiments shows degradation in species in the visible range, and an increase and no degradation of the intermediat es formed in the ultraviolet wavelengths. This result could be of furt her research interest.
45 0 0.2 0.4 0.6 0.8 1 1.2 1.4 1.6 1.8 240290340390440490540590Wavelength (nm)Absorbance 0 min 30 min 60 min 90 min 120 min 150 min 180 min 210 min 240 min 300 min 360 min MO Init Figure 20 Creation of Intermediate Products Du ring Photocatalytic Reaction of Ammonia Thermochemically Treated TiO2 Under SSS Irradiation 4.5 Catalyst Loading and Discoloration The effect of catalyst loading was tested to determine the optimal loading for both the visible simulated solar spectrum (VSSS) and simulated solar spectrum (SSS). The specifics of the light sources are det ailed below in the section on the characterizati on of the light sources. All of the experiments in this st udy were conducted in a natural pH solution of TiO2 (which is commonly reported as pHNAT = 6.3), with differing catalyst loadings. In line with the findings of Al-Qaradawi et al. and Guettai et al., the de-colorization of MO can be approx imated as pseudo-first order kinetics for simulated solar spectrum irradiati on as shown in Figure 21 [20, 22].
46 y = e-0.0043xR2 = 0.9989 y = e-0.0062xR2 = 0.9965 y = e-0.0069xR2 = 0.9926 y = e-0.0091xR2 = 0.97940 0.2 0.4 0.6 0.8 1 050100150200Time (m)C/C0 0.10 0.25 0.50 1.00 Expon. (0.10) Expon. (0.25) Expon. (0.50) Expon. (1.00) Figure 21 Effects of Catalyst Loading (g/L) on the Rate of Discoloration for Untreated Degussa P-25 TiO2 Under SSS Irradiation By considering the kinetics of the reac tion to be first order, we can plot Â– ln(C/C0) to arrive at the rate constant for the reaction, which is shown in Figure 22. It is this rate constant that is used as the point of comparison to Degussa P25 TiO2 for simulated solar spectrum experiments.
47 y = 0.0043x + 0.0014 R2 = 0.9982 y = 0.0063x 0.0148 R2 = 0.9949 y = 0.0072x 0.038 R2 = 0.9907 y = 0.0101x 0.1274 R2 = 0.98150 0.2 0.4 0.6 0.8 1 1.2 1.4 1.6 1.8 2 050100150200Time (m)-ln(C/C0) 0.1 0.25 0.5 1 Linear (0.1) Linear (0.25) Linear (0.5) Linear (1) Figure 22 Calculation of Rate Constants for Ca talyst Loading (g/L) for Untreated Degussa P-25 TiO2 Under SSS Irradiation Figure 23 is a plot for the apparent ra te constants for differing catalyst loadings for Degussa P-25 TiO2 under SSS irradiation.
48 0.004 0.005 0.006 0.007 0.008 0.009 0.01 0.011 00.20.40.60.811.2Loading (g/L)Apparent Rate Constant (1/min) ErRate Er+ Figure 23 Apparent Rate Constant for Catalyst Loading (g/L) for Untreated Degussa P-25 TiO2 Under SSS Irradiation Similar experiments were conduc ted to find rate constants VSSS irradiation. In Figure 24, the degradat ion rate for visible-light increases significantly as the catalyst loading increas es from 0.5 and1.0 g/L. This is in line with the literature where other s reported similar results. Guettai et al. reported that the degradation rate increased rapi dly as the catalyst concentration increased from 0.2 to 0.8 g/ L. They hypothesized that this was, Â“probably due to the increase of active sites with the suspension of catalyst loadingÂ”. The optimum catalyst loading found in this study was determined to be 1.0 g/L, with no further increase or decrease in the per formance due to increased loading. This is in contrast to the findings of G uettai et al., who found that the 0.8 g/L gave the most effective decomposition rate .
49 0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9 1012345678910Loading (g/L)Decoloration Rate (ppm/min) Rate Figure 24 Apparent Rate Constant for Catalyst Loading (g/L) for Untreated Degussa P-25 TiO2 Under Visible Simulated Solar Spectrum Irradiation More significantly, they report ed a decrease in discoloration with an increase in catalyst loading, which was also found in this study as shown in Figure 24. Guettai et al. speculated that light scattering was the cause of this decrease in efficiency. Figure 25 shows the light intensity fo r a given loading. As can be seen, increasing the catalyst load ing limits the light penetration through the solution. The result is a smaller volume that can be irradiated and theref ore the effects of catalyst loading reach their maximum. T hese results are similar to the results reported in the literature. It is hypothesized that t he higher concentration of suspended catalyst serves as a shield for th e catalyst in bulk of the solution. The
50 result is that part of the catalyst is not able to ads orb photons and participate in the photocatalytic process . Figure 25 shows the light penetration fo r a catalyst loading of 4 g/L. Contrast this with Figure 26 with a catalyst loading of 1 g/L. 0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9 1 00.10.20.30.40.5Distance (cm)Intensity 265-390nm MO Conc (ppm) 20 Loading (g/L) 1 400-490nm MO Conc (ppm) 20 Loading (g/L) 1 Figure 25 SSS and VSSS Light Intensity as a Functi on of Distance from the Source for a 20 ppm Methyl Orange Solution and 1 g/L of Untreated Degussa P-25 TiO2 As can be seen, the fall off in intensit y is dramatic. It should be noted here however that experiments for Degussa P-25 TiO2 in VSSS irradiation showed complete degradation of the po llutant during this study. This too, could be the source of an interesting study.
51 0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9 1 00.10.20.30.40.5Distance (cm)Intensity 265-390nm MO Conc (ppm) 20 Loading (g/L) 4 400-490nm MO Conc (ppm) 20 Loading (g/L) 4 Figure 26 SSS and VSSS Light Intensity as a Functi on of Distance from the Source for a 20 ppm Methyl Orange Solution and 4 g/L of Untreated Degussa P-25 TiO2 It also must be noted that the act ual catalyst loading concentration required is a function of type of reacto r used, its geometry and the incident radiation. As shown in Figures 23 and 24, the optimal loading differed for simulated solar spectrum and visible simula ted solar spectrum light sources. The reported values for optimum loading vary from 0.15 g/L to 8. 0 g/L for suspension photocatalytic processes . 4.6 pH and Discoloration In this study, no attempt was made to control the pH level of the photocatalytic process although pH values were measured for varying loading
52 levels for some experiments. However, the pH level is co nsidered to be the major factor influencing the rate of the photocatalytic process . Guettai et al. conducted experiment s in a pH range from 2 to 10 using NaOH and H2SO4 solutions to control the pH val ues. They reported that the best discoloration results were produced in an acidic solution; 98.58% (pH = 2), 91.85% (pH = 3) and 56.42% (pH = 5) after 5 hours of irradiation. For a neutral pH, they reported a discoloration of 28.58% (pH = 7.04). For a basic solution, they reported 16.67% (pH = 9.01). Their re sults show a significant difference in the discoloration of MO in strong acid ic solutions (pH = 2Â–3) and near neutral (pH 6) solutions. They concluded that the initial pH has a strong influence on the adsorption of the pollutant, and therefore a significant e ffect on the discoloration of the pollutant . Al-Qaradawi et al. reported similar results showing the best degradation occurring at a pH of 3 . However, they also report ed that a pH of 9 produced the next best result, which is in c ontrast to the value reported above. Additionally, it was determined that the pH of the solution changes over time to become more acidic, with the excepti on of a pH of 3, which remained unchanged. One interesting property of TiO2 is that it is amphot eric, meaning that it can behave as an acid or base depending on t he pH of the solution. It is this property that is used to explain the diffe rence in behavior in relation to differing
53 pH values. Below are the reactions that Guettai et al. propose for the reactions with pH values around its point of zero charge (pHPZC) : 2TiOH H TiOH pH < pHPZC O H TiO OH TiOH2 pH > pHPZC These equations show how the pH changes can thus influence the adsorption of the pollutant on the surface of TiO2, which is an important step if photooxidation is to take place. The literature shows the measured pHPZC value for Degussa P-25 TiO2 to be around 5.8 to 6.8. Fo r acidic solutions, pH < 6, MO is strongly adsorbed to the TiO2 surface. This is a result of the electrostatic attraction between the positively charged TiO2 particles and the pollutant. In a basic solution, pH > 7, t he MO molecules ar e negatively charged. In this case the adsorption of the pollutant is also affected due to an increase in the density of TiOon the catalysts surface. The resu lting coulombic repulsion retards the adsorption of the pollutant on the catalyst surface . Further, for basic solutions with high pH values, the scavengi ng of hydroxyl radicals happens so quickly that there is no reaction with the po llutant. The photocatalytic reactivity of sulphonated dyes, to which MO belongs, are maximum in acidic solutions, and decrease as the pH increases . The literature also shows that interpreti ng the effects of pH is very difficult due to the differing reaction mechanisms. There are three reaction mechanisms
54 that can be involved in t he photodegradation process. Fi rst, hydroxyl radicals may participate: OH Dyedegradation product. Second, holes may directly oxidize the pollutant: VBh Dyeoxidation products. And, third, electrons may reduce the pollutant: CBe Dyereduction products. The difficulty in interpretation, and a poi nt of debate, is that at a low pH holes are thought to be the major oxidizi ng species, but at neutral and high pH levels the hydroxyl radicals are believed to be the major oxidizing species . In contrast, Al-Qaradawi et al. conclude d that the degradation rate increases with a high pH value. They st ate that a basic solution wi ll have a higher number of hydroxyl radicals at the catalyst surf ace where they will trap the photo-induced holes . VB CBh e TiO A UV hv TiO2 2 OH H TiO O H h TiOVB 2 2 2 OH TiO OH h TiOVB 2 2
55 The effects of pH on the calculati on of photocatalytic rate are seen in Figure 27. A 20 ppm solution of MO wa s prepared with 1 g/L of Degussa P-25 TiO2. The pH of the solution was adjust ed to produce acidic and basic solutions and the spectra were measur ed. As can be seen in fi gure 27, the peak at 451 nm is shifted due to the change in pH This must be considered when considering the photocatalytic rate. 0 0.5 1 1.5 2 2.5 240290340390440490540590640Wavelength (nm)Absorbance pH = 1.83 pH = 3.63 pH = 5.87 pH = 7.00 pH = 10.08 pH = 11.16 Figure 27 Spectra of 20 ppm Solutions of Methyl Orange and 1 g/L of Degussa P-25 TiO2 for Varying pH Levels 4.7 Initial Concentration and Discoloration Rates The initial concentration of MO is also a factor in the discoloration rate. In this study, it was found that significant adsorption of the pollutant occurs during
56 the first 15 minutes. This was firs t discovered when seemingly similar experiments resulted in di fferent degradation rates. It was discovered that a sonication step meant to de-agglomerate the catalyst produced differing results. Subsequent study of the cause showed that the time used for sonication, not the sonication itself, was responsible for the adsorption of the pollutant on the catalyst surface. This initial adsorpt ion changed the initial concentration, and therefore seemingly the order of the reaction from firs t order to zeroth order. Figure 28 shows the change in MO conc entration due to adsorption of the pollutant on the catalyst over 30 minutes. This adsorption-desorption process is known as the Langmuir equilibrium.
57 19.3 9.7 40.4 15.5 60.4 29.5 80.1 45.60 10 20 30 40 50 60 70 80 90 05101520253035Time (m)Concentration (ppm) 20ppm 40ppm 60ppm 80ppm Figure 28 Change in Concentration as a Function of Time Due only to the Adsorption of Methyl Orange to Untr eated Degussa P-25 TiO2 (3 g/L) The effect of initial MO concentrati on on the photocatalytic effectiveness of the process was studied by Guettai et al. The study was conducted for initial MO concentrations from 5 to 75 mg/L for a pH =3, and a catalyst loading of 0.8 g/L. They made two conclusions from their st udy. First, that the degradation of MO satisfactorily follows the Langmuir-Hinsh elwood model. Sec ond, the percentage of degradation ultimately decreases with an increasing initial concentration of the pollutant. For levels above 75 mg/L, t he degradation rate becomes very slow . Initially, an increase in the concentrati on of MO increases the probability of a reaction between the pollutant and the oxidizin g species. This in turn results in
58 an increase in the discoloration rate. As the concentration of MO increases, the active sites on the catalyst surface are more fully covered reducing the photogeneration of holes or hydroxyl radicals . It has also been speculated in the literature that Â“UV scr eeningÂ” (the adsorption of incident photons by MO molecules) is another possibility regarding the decrease in rate . The general conclusion is that the optim um catalyst loading is a function of active surface area and the pollutant concentra tion, and therefore is not nec essarily static . Similarly, Al-Qaradawi et al. also found that an increase in initial catalyst concentration ultimately decreases the ov erall degradation efficiency overall . 4.8 Photocatalytic Intermediates of MO Methyl orange has an orange color in a basic medium and a red color in an acidic medium. Its color is a resu lt of absorbing cert ain wavelengths of visible-light and either transmitting or re flecting other wavelengths. As an azocompound, MO derives its color from the azo-bonds N N and their associated chromophores and auxochromes. A chromophore is a region in a mo lecule where the energy difference between two different molecular orbitals falls within the range of the visible spectrum. Visible-light t hat hits the chromophore can be absorbed by exciting an electron from its ground stat e into an excited state. An auxochrome is a group of atoms attached to a chromophore which modifies the ability of that chromophore to absorb light. If these groups are in
59 direct conjugation with the pi-system of chromophore, they may increase the wavelength at which the light is absorbed and as a result intensify the absorption. As stated previously, there are tw o peaks with maxima at 272 nm and 451 nm. The peak at 451 nm has the highest absorbance magnitude, and was used as the reference point for calculating the discoloration of the methyl orange. The discoloration of the MO Â– catalyst solu tion indicates that MO has been degraded. Figure 29 shows the change in spectra ov er time, and depicts the degradation of the pollutant. Note that the band at 272 nm is increasing with time, and that a new band appears at 334 nm, which can be attributed to the formation of intermediates. This is similar to th e findings of Al-Qaradawi et al.  and Guettai et al. who state that interm ediates are formed during photocatalytic degradation, and they, too, absorb light. It is also clear that these intermediates also degrade over time  Al-Qaradawi et al. concl uded that the aromatic rings are responsible for this absorption and that this is responsible for the increase in the intensity and new band which forms at 330 nm. Ultimately, the magnitude of these bands decreases, which indica tes the degradation of MO to CO2 and H2O .
60 0 0.2 0.4 0.6 0.8 1 1.2 1.4 1.6 240290340390440490540590Wavelength (nm)Absorbance 0 min 30 min 60 min 90 min 120 min 150 min 180 min MO Init Figure 29 Absorbance Spectra as a Function of Time for Methyl Orange in the Presence of Untreated Degussa P-25 TiO2 and Simulated Solar Spectrum Irradiation Further, it was concluded that the tw o aromatic rings in MO start to degrade thereby creating mono substituted ar omatic molecules. Similar to the results of this study, the absorption maxima for both MO and the intermediate products disappear with the exce ptions that were previous ly noted. As a result, the photocatalytic degr adation destroys both the conjugate system, which includes N N, and either partially or comp letely destroys the intermediate product(s) converting them to CO2 and H2O . 4.9 Adsorption of Methyl Or ange on Titanium Dioxide It is widely agreed that the photocatalyt ic degradation of a pollutant can be modeled using the Langmuir-Hinshelwood model. The Langmuir-Hinshelwood
61 model describes the mechanism that allows for a reaction to take place at the surface of a catalyst and the adsorbed pollu tant in an aqueous solution. In doing so a number of assumptions must be made: 1. The surface of the catalyst has a limited number of adsorptions sites. 2. The surface of the catalyst is covered only by a monolayer, and the adsorption site can only adsorb one molecule. 3. The molecule can both adsor b and desorb from the surface. 4. The surface of the catalyst is homogeneous. 5. There is no interaction between t he adsorbed molecules on the catalyst surface. The Langmuir adsorption model can be given for compounds in an aqueous suspension according to the following equation : C K C K Q Qads ads ads 1max with, t adsC C m V Q 0, where is the ratio of the TiO2 surface covered by the adsorbed pollutant, Qads is the actual quantity adsorbed on the surface, and Qmax is the maximum quantity that can be adsorbed on the surface. Kads is the Langmuir adsorption constant, and C is the concentration of the pollu tant. The quantity of the pollutant adsorbed on the surface, Qads, is the ratio of the r eaction volume, V, and the
62 mass of the catalyst, m, multiplied by the differ ence of the initial dye concentration, C0, and the dye concentration at time t, Ct. The adsorbed equilibrium quantity (Qeq) can be measured at equilibrium using the follo wing equation: eq ads eq ads eqC K C K Q Q 1max, where Ceq is the adsorption equilibrium concentration. Although adsorption of an organic pollu tant is not required due to the ability of OH radicals diffusing into the bulk of the solution and oxidizing the pollutant, it was concluded in the study by Guettai et al. that no photodegradation took place without the adsorpt ion of MO on the catalyst surface. It is likewise conjectured here to be the same . 4.10 Photocatalytic Degradation of Methyl Orange It has been concluded that the hydroxyl radical is responsible for most heterogeneous photocatalytic oxidations. T he hydroxyl radical is formed by the reduction reactions of holes with eit her water or hydroxide ions . There are many factors that affect the photocatalytic activity of TiO2. These include the lattice structure and phase of the ma terial, specific surface area, adsorption of the pollutant, elec tron-hole generation and recombination, carrier lifetime and trapping, solution pH, method of synthesis, catalyst loading, and initial pollutant concentration, among m any others. The research for this
63 study also showed the profound difference that the design and materials of the reactor itself can have on the photocatalyt ic effect or rate. Although outside of the scope of this work, it was evident by this study that metals ions would be an interesting area of further research as it relates to carrier scavenging . The degradation decay for this stud y agrees with the lit erature and follows first order kinetics for simulated sola r spectrum irradiation. Figure 30 is representative of the general results obtained showing th is first order decay . y = e-0.0043xR2 = 0.9989 y = e-0.0062xR2 = 0.9965 y = e-0.0069xR2 = 0.9926 y = e-0.0091xR2 = 0.97940 0.2 0.4 0.6 0.8 1 050100150200Time (m)C/C0 0.10 0.25 0.50 1.00 Expon. (0.10) Expon. (0.25) Expon. (0.50) Expon. (1.00) Figure 30 Effects of Catalyst Loading (g/L) on the Rate of Discoloration for Untreated Degussa P-25 TiO2 Under SSS Irradiation The first order rate constant is th en calculated from t he slope of the log plot of the ratio of t he measured concentration and or iginal concentration with respect to time. An example of the first order rate const ant is illustrated in Figure 31.
64 y = 0.0043x + 0.0014 R2 = 0.9982 y = 0.0063x 0.0148 R2 = 0.9949 y = 0.0072x 0.038 R2 = 0.9907 y = 0.0101x 0.1274 R2 = 0.98150 0.2 0.4 0.6 0.8 1 1.2 1.4 1.6 1.8 2 050100150200Time (m)-ln(C/C0) 0.1 0.25 0.5 1 Linear (0.1) Linear (0.25) Linear (0.5) Linear (1) Figure 31 Calculation of Rate Constants for Ca talyst Loading (g/L) for Untreated Degussa P-25 TiO2 Under SSS Irradiation As stated previously, the results of the photocatalytic degradation for this study fits the Langmuir-Hinsh elwood model, which is in line with published work in this area. Simply stated, the ox idation rate, r, can be expressed as: dt dC r where C is the concentrati on of the pollutant, and t is time. However, the Langmuir-Hinshelwood model must be modified to account for the fact that the catalyst particles are in solution. In solution, both hydroxyl radicals and water molecules have the ability to be adsorbed on the catalyst surface, which means that the active sites of the catalyst c ould be covered by eit her, which will impact
65 the reaction rate . The result is that the expression mu st account for the adsorption of the pollutant. The expr ession therefore must be re-written as: W S LH LH r rC K C K C K k k dt dC r 1, where kr the reaction rate constant, is the ratio of the TiO2 surface covered by the adsorbed pollutant, KLH is the pollutant adsorption constant, KS is the adsorption constant of t he solvent (water), and Cw is the concentration of the its solvent. Guettai et al. concluded that gi ven that the concentration of the solvent was much greater than the concentration of the pollutant, the co ncentration of the solvent remains constant and the coverage of the catalyst by water molecules also remains constant. Holding all experimental conditions constant, the concentration of the pollutant is the onl y variable and the expression can now be re-written as : C K C K k dt dC rLH LH r 1, Intermediates formed during the photocat alytic process make it particularly difficult to determine the actual photocatal ytic degradation rate. Figure 32 shows another example of this phenomenon. The traditional absorption peak at 451 nm continuously shifts to lower wavelengths indicating that inte rmediate products are forming and degrading. Further, the abs orbance spectra, Figure 32, differ significantly from Figure 29 above for wa velengths below 440 nm. The classical
66 cleavage, described above, into two m ono substituted aromatic molecules, cannot be seen in this spectrum. 0 0.2 0.4 0.6 0.8 1 1.2 1.4 1.6 240290340390440490540590Wavelength (nm)Absorbance 0 min 30 min 60 min 120 min 180 min MO Init Figure 32 Shift in Spectra Due to Intermediates for 20 ppm Methyl Orange Solution in the Presence of 3 g/L of Degussa P-25 TiO2 With No Oxygen Source and Under VSSS Irradiation These intermediates will ultimately compete in the adsorption-desorption process and will complicate the determination of the actual photocatalytic rate. Guettai et al. chose to make calc ulations at the beginning Â“illuminated conversion.Â” The rationale was that any changes could be negated. From this, they could re-write the photocatalytic degr adation rate only as a function of initial concentration and adsorption constant of the pollutant, as shown in the expression below : 0 0 01 C K C K k k rLH LH LH LH,
67 where r0 is the initial rate of phot ocatalytic degradation of MO, is the ratio of the TiO2 surface covered by the adsorbed pollutant, and C0 the initial concentration of the pollutant, which is said to be Â“equal as possible to the concentration at the adsorption equilibrium, CeqÂ” . Assuming a small value for C0, this equation is simplified to the first or der equation shown below: app rk Kt k C C Ln 0 or t k tappe C C0, where app rk K k From this equation for the appar ent first order rate constant, kapp, the initial degradation, r0, rate is given by the equation: 0 0C k rapp Consistent with the literatur e, this study found that t he plot of Ln(Ce/C) versus time produces a straight line. T he literature conclu ded that the kapp is given by the slope of the linear regre ssion of this plot . Guettai et al. concluded that the degradation rate is a function of the initial concentration of the pollutant. The init ial concentration causes the rate to increase until a concentration threshold is reached, and that beyond this point, the degradation rate is independent of the concentration . Using the reciprocal of the initial rate, 0 01 1 1 1 C kK k rr r Guettai et al. also concluded that the plot of the reciprocal confirms the Langmuir-Hinshelwood model. This again
68 points to the photocatalytic process beginning with the ads orption of the pollutant. Further, they concluded that giv en that the reacti on does follow the Langmuir-Hinshelwood model, and KHL is the actual adsorpt ion affinity of the pollutant on the surface of the catalyst, then Kads and KLH should be identical. From this relationship, they concluded t hat the Â“photocatalytic degradation of MO under optimized working conditions fo llows satisfactory the LangmuirHinshelwood modelÂ” .
69 CHAPTER 5: FUNDAMENTALS OF NITROGEN-DOPING 5.1 Decreasing the Band Gap of TiO2 Untreated TiO2 can utilize less than 4% of the available solar energy for photocatalysis. Given the exceptional photocatalytic properties of TiO2, a great deal of study has gone into decreasi ng the band gap to allow visible-light activated photocatalysis. Doping TiO2 by metals or transition metals, and anion doping have dominated this area of res earch. Research has shown that the substitution of oxygen by nitrogen in me tal oxides can cause a red-shift of the absorption edge and facilitate the absorption of visible-light [ 25]. One of the primary goals of this research was to inve stigate a relatively inexpensive way to lower the band gap of Degussa P-25 TiO2 by doping it with nitrogen. Anion doping is a method that has rece ived a great deal of attention and is the focus of this study. Numerous st udies using carbon, sulfur, iodine, and nitrogen have been conducted showin g that the band gap of TiO2 can be reduced [3, 8, 9, 19, 22-49]. Doping TiO2 with anions alters the conductivity and optical properties by creati ng new surface states that are believed to lie near the conduction band or valence band of TiO2 . It was found that these materials had high photocatalytic efficiencies in t he visible region. As a result, the improvement was attributed to a shift in the absorption edge and narrowing of the
70 band gap of the material. Subsequent studi es show that the actual mechanism for the improvement is of great debate. The doping of oxides such as TiO2 creates a set of pr oblems of its own, such as the creation of recombination cent ers in the created electronic states. It has been shown that the electrical properti es are strongly influenced by defects and dopants . Therefore a great deal of research has focused on delaying the recombination of photogenerated electron-hole pairs by adjusting microstructure, processing r outes and composition of TiO2 . Previous work in this area has shown that substitutionally doping TiO2 with nitrogen to produce TiO2-xNy is an effective method of creating a visible-light active photocatalyst [26-30]. By calcinating TiO2 in an ammonia atmosphere, or by a wet chemical method, TiO2 can be easily modified for utilization of the visible spectrum . This study accomplis hes substitutional nitrogen-doping of the catalyst using a gas phase impregnation method using anhydrous ammonia as a nitrogen source. 5.2 Mechanism of Visible-Light Absorption and Photocatalytic Activity There appear to be as many conclusi ons as there are studies regarding the result of nitrogen-doping. In general, it is agreed that doping TiO2 with nitrogen creates new surface states that re sult in new energy levels in the band gap. These new levels provide pathways for electron transfer between the
71 catalyst and the acceptor species in t he solution . However, there is no consensus on where these new states are, or how this is accomplished. In general, the results of nitrogen-dopi ng can be categorized as follows: 1. Nitrogen-doping decreases the band gap by the overlap of nitrogen 2p orbital (N2p) and the oxyg en 2p orbital (O2p) . 2. Nitrogen-doping does not decrease the band gap, but instead creates new electronic state(s) above the valence band . 3. Nitrogen-doping does not decrease the band gap, but instead creates new electronic state(s) below the conduction band . 4. Oxygen vacancies, which are a result of the thermal effe ct of the nitrogendoping process, are responsible for photocatalytic improvement [35, 36]. 5.2.1 Decrease in Band Gap Due to the Overlap of N2p and O2p Orbitals A number of studies regarding the density of states for nitrogen-doped TiO2 concluded that the N2p states and t he O2p states overlap due to the substitutional doping of oxygen with nitrogen. Liu et al., Irie et al., and Miyauchi et al., all performed calculations of the densities of states for anion doping in anatase TiO2. They concluded that nitr ogen-doping was the most effective method of achieving visible-light phot ocatalytic activity [25, 33, 37].
72 Irie et al. concluded, along with a number of authors, that the band gap decreases due to the creat ion of an isolated nitrogen 2p (N2p) band above the oxygen 2p (O2p) valance band by substitu tionally doping nitrogen atoms in place of oxygen atoms in the lattice of TiO2. Further, they concluded that this N2p band is responsible for the visible-light photoc atalytic activity [ 33]. In contrast, Diwald et al. concluded t hat the Â“co-doping effectÂ” of nitrogen and hydrogen is the cause for the visible-light photocatalytic activity . Although Pan et al. agreed that b and gap narrowing is a result of the mixing of N2p and O2p states; they draw a different conclusion regarding visiblelight absorption. They concluded that visible-light absorption is due to the formation of Ti3+ from the reduction by H2 of Ti4+ during the decomposition of ammonia during treatment at 600C. T hey found that the hi gher the temperature of ammonia treatment, the gr eater the red-shift was [39 ]. This study has found similar results. In the study by Madhsudan et al ., they proposed how nitrogen-doping reduces the band gap of TiO2. Figure 33, from their study, depicts the effect of nitrogen-doping on the creation of new surface states. The band bending illustrates the initiation of redox charge transfer reaction s . Diagrams a, b, and c represent the band pinni ng at the semiconductorÂ–el ectrolyte interface. Figure (a) shows the accumulation of char ge at the interface when the pH of the solution is 2. Figure (b) depicts t he flat band where the Fermi level (EF) is in equilibrium. Figure (c) shows the depletion layer for a solu tion of pH of 10. The
73 band diagram (d) illustrates intermediate levels that are produced in TiO2 after nitrogen-doping. Figure (e) shows the band bending at the nitrogen-doped TiO2Â– electrolyte interface. Figure 33 Hypothetical Band Gap of Nitrogen-Doped TiO2 Proposed by Madhsudan et al.  5.2.2 Creation of New Electronic States Above the Valence Band  Asahi et al. concluded that the narrowing of the band gap by nitrogendoping results in visible-light photocatalytic activity . Whereas Irie et al. concluded that it was not the narro wing of the band gap, but instead a narrow band above the valence band that is responsible for the visible-light activity .
74 In the study by Irie et al., TiO2-xNy irradiated with ultrav iolet light produced a higher photocatalyt ic rate than TiO2-xNy irradiated with visibl e-light. They claim that nitrogen was substitutionally doped into the TiO2 oxygen lattice sites, and this resulted in an isolated narrow band above the valence band. Ultraviolet radiation excites the electrons from bot h the valance band and the narrow band. However, if visible-light is used to irr adiate the catalyst, only electrons in the narrowband are excited to higher energy levels. From this difference an important conclusion is dr awn. If the band gap of TiO2 is narrowed by substitutional nitrogen-doping, then the phot ocatalytic rate should be the same for both ultraviolet and visible-light, which of c ourse it is not. Therefore, Irie et al. concluded that the band gap of TiO2 is not narrowed, but instead the isolated narrow band above the valence band which is created by nitrogen-doping is responsible for the visible-light photocatalytic activity . 5.2.3 Creation of New Electroni c States Below the Conduction Band Torres et al. concluded that nitrog en-doping created a new set of electron trap states just below the conduction band positioned energetically in a broad range of 1.3 eV . These states, wh ich they calculated to have a similar density of states as nitrogen atoms also acted as electron-hole recombination centers. The new states responded to ul traviolet, visible and infrared light at wavelengths greater than 700 nm and worked as long-lived electron traps .
75 5.3 Creation of Oxygen Vacan cies Due to Thermal Effects What also must be considered is the thermal effect of the treatments on oxygen deficiencies in the lattice of TiO2, as oxygen vacancies alone can cause TiO2 to respond to visible-light. This leaves open a debate whether the oxygen vacancies are created by thermal effect s associated with doping methods, or doped nitrogen atoms are responsible for visible-light activity. Thermal treatments alone have been f ound to create sub band gap levels in their electron structures. These elec tron structures are believed to allow low energy excitation pathways. Further, ox ygen deficient sites on the catalyst surface are highly reactive sites that reac t with oxygen molecules. Ihara et al. determined that 400C is the Â“transition temperatureÂ” for the creation of the maximum number of oxygen vacancies [35 ]. While calcination alone increases oxygen vacancies, it also has the negat ive impact of reducing the adsorption affinity of the catalyst as a result of a decrease in specific surface as seen in BET surface area measurements . From two separate studies, Nakamu ra et al. concluded that the photocatalytic activity of nitrogen-doped TiO2 is caused by surface intermediates of oxygen reduction, or by H2O oxidation and not by dire ct reaction with the holes trapped at nitrogen induced midgap levels , and, that the oxygen vacancies that are formed during the ni triding process act as elec tron traps that improve the visible-light reactivity [ 40]. Yates et al. concluded that this occurs Â“when an electron is excited to the oxygen vac ancy state from the valance band (even
76 under visible-light). Â“These electrons (or the holes formed) react with atmospheric oxygen (or an oxygen related species) producing reactive species (O or atomic oxygen), which participate in the oxidation of surface speciesÂ” . Prokes et al. however, believe that th is same phenomenon act ually reduces the photocatalytic activity since trapped photoelectrons cannot easily reach the reactive surface sites of the catalyst . 5.4 Nitrogen Concentration an d Phases Changes Due to Thermal Effects Similar to the results of Pan et al. reported above, Wei et al. also observed an increase in the absorption of visibl e-light with increas ing nitrification temperature up to 550C. However, raisi ng the temperature to 600C resulted in a decrease in the visible-light absorpt ion. They attributed the observed phenomena to a change in the crystal st ructure caused by the high doping concentration of nitrogen in the TiO2. As was found and reported in this study, another explanation may be transition from anatase to rutile phase that takes place at higher temperatures, which is al so the conclusion by Mozia et al. They reasoned that if the crystall ographic structure of TiO2 changes, that this could change the doping mechanism. The result is change in the ultraviolet and visible-light absorption of the catalyst . The conclusion of this study is that there is no significant change in anatase to rutile ratio for nitrogen-doped TiO2 for temperatures below 625C. Silveyra et al. made a similar conclusion for nitrogen-doped TiO2 at temperatures
77 below 600C. Both studies found that signi ficant differences in anatase to rutile ratio between untreated Degussa P-25 and nitrogen-doped Degussa P-25 begin treatment temperatures over 700C . The study by Wang et al. concluded t hat the transition from the anatase to rutile phase during nitrogen-doping is rest rained by the doping of nitrogen into the TiO2 lattice during the preparation of TiO2 xNy. This is the basis of their conclusion that TiO2 xNy has a higher photocatalytic activity than pure TiO2. They reason that since the anatase phase has a higher band gap than the rutile phase, the photocatalytic activity must be attributed to visiblelight absorption of TiO2 xNy, and not to an increase in rutile phase . Similarly, surface area measurement s show the same relationship for nitrogen-doped materials. The results of th is study (see Figure 34) also coincide with the findings of Silveyra et al., wh ich find that the change is surface area is negligible for nitrogen-doped TiO2 treated at 600C and 625C . Both studies found that the surface area remains moderately constant until treatment temperatures greater than 700C where a reduction of surface area increases rapidly with increasing temperature. S ilveyra et al. refer to Degussa P-25 TiO2 as a Â“nonporous material which consists of sp herical particles with a mean diameter of 30 nmÂ”. They concluded t hat sintering is responsibl e for the loss of surface area. This study found a reduction of approximately 40% in surface area for nitrogen-doped Degussa P-25 at 675C, whereas Silv eyra et al. found no
78 significant difference between treat ed Degussa P-25 and nitrogen-doped Degussa P-25 at 600C. . 0 10 20 30 40 50 60 300400500600700800900Temp (C)Surface Area (m2/g) TiO2 Thermally Treated N2 Pure Degussa TiO2 TiO2 Thermochemically Treated NH3 Figure 34 BET Surface Area Measurements for Pure, Calcinated and Thermochemically Ammonia Treated Degussa P-25 TiO2 The accepted view of this study is that substitutionally doped nitrogen is responsible for the visiblelight absorption. A number of studies have used XPS to make conclusions regarding the composition of TiO2 xNy Yuan et al. concluded that molecularly chemis orbed nitrogen is removed during high temperature doping processes such as annealing . Although some chemisorbed nitrogen does remain and does contribute to visible-light absorption, it was determined that the photoc atalytic activity is primarily promoted by substitutionally doped nitrogen . However, it must be again noted that
79 calcinated Degussa P-25 TiO2 also has higher photocatalytic activity which might result from the slight phase transformation of TiO2 from anatase to rutile . 5.5 Effect on Photoc atalytic Activity Similar to the effects of nitrogen-dopi ng on the electronic states of TiO2, there is also debate whether or not nitrogen-doping improves photocatalytic activity. Generally speaking, there are three schools of thought regarding the effect of nitrogen-doping: 1. Positive Effect on Photoc atalytic Activity [32, 45] 2. No Effect on Photocatalyt ic Activity [35, 36, 46, 47] 3. Negative Effect on Photocatal ytic Activity [19, 29, 55] What must be noted is that photocatalytic effect is application specific, and that mistakes may be made in casting ov erreaching conclusions. It is also apparent that the methods used to synt hesize and test the catalyst have a significant effect on the results as simila r studies offer conflicting conclusions as even the authors themselves note. 5.5.1 Positive Effect on P hotocatalytic Activity It was concluded in some studies that the visible-light photocatalytic efficiency of nitrogen-doped TiO2 is a function of doping level. These studies found that the photocatalytic effect increases with increasing nitrogen concentration [22, 58]. This conti nues until a threshold concentration is
80 exceeded at which point the photocatalytic efficiency decreases. Further, the concentration level was determined to be doping method specific . The study by Asahi et al. concluded that visible-light photocatalytic activity is a result of the reduction in the band gap due to nitrogen-doping . 5.5.2 No Effect on Photocatalytic Activity In contrast, Ihara et al. and Marty anov et al. concluded that the oxygen vacancies created by thermal effects ar e responsible for the photocatalytic activity of TiO2 xNy [28, 52]. Ihara et al. conc luded that oxygen-deficient sites formed in the grain-boundaries are import ant to visible-light photocatalytic activity, and that nitrogen-doped in oxygen-deficient sites in TiO2 is important as a blocker for reoxidation . The commonly held assertion that ni trogen-doping causes the overlap of the N2p and O2p states, and t herefore, is responsible for the visible-light photocatalytic activity is challenged by Xu et al. who concluded that impurity states are localized and lie slightly above the top of the O2p valence band. Their conclusion implies that the mixing of N2p with O2p states is small even for high concentrations of nitrogen . Huang et al. concluded that nitrog en-doping only improves the absorption in the visible region, and t hat quantum yield, or photocatalytic rate, is still very low and not necessarily useful for practical applications .
81 5.5.3 Negative Effect on P hotocatalytic Activity Yates et al. concluded that photogener ated holes are trapped at the midgap states that are created by doping and as a result decrease their oxidation power and reduce the photocatalytic activity of the catalyst. It is also speculated that doping may induce structural instabilities in TiO2. This is due to lattice distortions and bond weakening that result from doping . It is also believed that increasi ng the nitrogen concentration actually lowers the photocatalytic activity. Ther e are two conclusions regarding this. First, that doping sites could also serve as recombination sites . Second, that photocatalytic activity is actually reduced since trapped photoelectrons cannot easily reach the reactive surface sites of the catalyst as a result of oxygen vacancies that are created in the doping process . Diwald et al. also concluded that the r eduction in photocatalytic activity of substitutionally doped nitrogen is from Â“induction of ionic charge compensationÂ” . What takes place is that the nitrogen acceptors are compensated by donor levels that are related to additional oxy gen vacancies. This results in an increase in electronically reduced Ti3+ states. Their conclusion is that a blue shift is caused as a result of substitutional ni trogen-doping due to the partial filling of electrons in the conduction band . Yate s et al. alluded to other studies which also draw conclusions based on the reduc tion of photocatalytic activity of nitrogen-doped TiO2 .
82 Regarding the reduction of TiO2, Irie et al. determined that the photocatalytic activity reduces with in creasing nitrogen concentration. They found that TiO2 xNy powders formed using a NH3 atmosphere not only has oxygen replaced with nitrogen, but that TiO2 is simultaneously reduced. The result is an increase in oxygen vacancies and the amount of Ti3+. If correct, this conclusion has a dramatic effect on the phot ocatalytic rate as this suggests that the oxygen vacancies act as a recombi nation center for holes and electrons. They further concluded that an increase in the NH3 annealing temperature resulted in an increase in TiO2 reduction, and a higher number of oxygen vacancies. This is the basis of thei r conclusion regarding the reduction in photocatalytic activity for increas ing nitrogen concentration . 5.6 Reducing the Recombination Rate of Electron-Hole Pairs Lindgren et al. also found a larger density of Ti3+ states in nitrogen-doped TiO2 compared with undoped TiO2. They concluded that those states, together with slow hole transport in the nitr ogen induced band, are the reasons for the high rate of recombination in TiO2 xNy . It is believed that doping creates new band structures that contribute to the recombination of photogenerated electr onÂ–hole pairs. However, it is also believed that by using specific dopants that it is possible to improve the efficiency of the photocatalytic activity by creati ng new band structures or by suppressing the effects of existing band structures to improve quantum efficiency. Coupling two semiconductors is also seen as a technique to enhance charge separation
83 given that the semiconductors have c onduction and valence band potentials that would enhance the separation . Transition metals such as zinc iron, and tungsten have been doped to substitute Ti4+. Transition metals enable the separation of photogenerated electronÂ–hole pairs and increase the light absorption . This study did introductory work regarding the coupling, or co-doping of TiO2 with tungsten but found no evidence of improvement and no mo tivation for continuing that work. 5.7 Current Methods for Pr oducing Nitrogen-Doped TiO2 As an initial remark regarding the selection of anhydrous ammonia as the nitrogen source for this study, urea w ould be more environmentally friendly nitrogen source than anhydrous ammonia. This has been studied by Yuan et al. and should be considered fo r future wo rk . There are numerous other methods t hat are being utilized to synthesize nitrogen-doped TiO2. Those methods include oxidat ion of titanium nitride , calcination of hydrol ysis products of Ti(SO4)2 previously treated with an aqueous NH3 solution , calcination of TiO2 under an ammonia atmosphere , substitutional doping of metal ions , DC magnetron sputtering the TiO2 target in N2(40%)/argon gas atmosphere , i on implantation [ 53], organic dye sensitization , meta l organic chemical vapor deposition (MOCVD) method using titanium-tetra-isopropoxide as a precursor and ammonia as the nitrogen source , hydrogen plasma reduction of TiO2 , pulsed laser deposition
84 method , hydroxide or surface c oordination , and high temperature treatment of single-crystal TiO2 in an argon atmosphere (870K) and nitrogendoped when argon is replaced by NH3 . Yates et al. categorized them into three broad areas : 1. Modification of existing TiO2 in the forms of powder, films and single crystal, or TiN via gas phas e chemical impregnation. 2. Modification of existing TiO2 via ion bombardment. 3. Growth of TiNxOy from liquid or gaseous precursors. 5.8 Formation of Titanium Nitrid e as a Result of Nitrogen-Doping The visible-light absorption affinity of TiO2 xNy is understood to be in direct proportion with the nitrogen concent ration of the catalyst  In their study, Li et al. concluded that nitrogen exists only as an impurity due to its low concentration. This low concentration is too low to form a new phase that would shift the fundamental absorption edge of TiO2 . The black and brow n materials in this study however, clearly suggest movement of the fundamental absorption edge as noted in Figure 70. For colored nitride compounds such as the black and brown materials, this shift was attributed to the substitutional doping of nitrogen atom s . Although studies suggest that increasing concentra tions of nitrogen are proportional to color, Liu et al. found that visible-light activity was not directly proportional to the red-shift of the second absorption edge . Diwald et al. supposed it is due to a
85 form of nitrogen that is likely to be interstitial and chemically bound to hydrogen in the first few hundred angstroms below the TiO2 surface . Miyauchi et al. found that the crea tion of TiN crystal phases by high temperature TiO2 annealing in ammonia had a detrime ntal effect on catalyst activity, which is the conclusion of this study . 5.9 Decreasing the Band Gap The optical absorption data was deduc ed from the KubelkaÂ–Munk function calculated for untreat ed Degussa P-25 TiO2 and the thermochemically ammonia treated Degussa P-25 TiO2 samples. The Kubelka-Munk method is a diffuse reflectance technique that uses a salt, NaCl in this case, mixed with the powder being measured. This technique accounts for the difference in transmissi on and reflectance measurements due to absorption of certain wavelengths by powder s. The material was diluted to about 1% by weight in NaCl and ground usi ng a mortar and pestle. Transmission measurements were made of the powders, which were lightly packed into small sample holders, using an optical spectrom eter. A Kubelka-Munk conversion was applied to a diffuse reflectance spectrum to compensate for differences. The Kubelka-Munk equation is: s k R R R f 2 12,
86 where: R is the absolute diffuse reflectanc e of the sampled layer, k is the molar absorption coefficient, and s is the scattering coefficient. A linear relationship is created between the spectral intensity relative to the sample concentration. This equation assumes that the diluting salt is nonabsorbing, that the scattering coefficient of the salt is constant, and that the sample thickness is infinite. These assumptions can be made for samples greater than1 mm of highly diluted sm all particles. Given that we used nanoparticles and had a sample thickness of 3mm, our packing technique was the only variable that could affect the sca ttering coefficient. This technique was perfected over time by a fellow CERC student, Ms. Paula Algarin. The UV-Vis diffuse reflectance spectra for thermochemically ammonia treated Degussa P-25 TiO2 are shown in Figure 35. These materials were thermochemically treated with anhydrous amm onia at a flow rate of 87.3 mL/min. as a nitrogen source at 675C for six hours. This was followed by annealing in a nitrogen atmosphere at a rate of 41.7 mL/min. for six hour s. Finally, the material was oxidized at 400C for 15 minutes in an oxygen atmosphere.
87 0 0.2 0.4 0.6 0.8 1 1.2 200300400500600700800Wavelength (nm)Diffuse Reflectance 161A TiO2 675C NH3 87ml/min 6h 163A TiO2 675C 87 ml/min 6h 164B TiO2 675C NH3 87ml/min 6h TiO2 Pure Degussa (06-16-06) Figure 35 Diffuse Reflectance of Pure, Thermochemically Ammonia Treated Degussa P25 TiO2 The absorption edge of the samples is determined by the following equation: 8 1239gE, where Eg is the band gap, measured in elec tron volts, of the sample, and is the wavelength of the onset of the spectra measured in nanometers . For TiO2 to produce active photocatalytic species, it must absorb photons with energy that is equal to or greater than the band gap. Fo r visible-light photocatalysis this requires band gaps of less than 3.1 eV. Ho wever, for practical considerations, band gaps around 2.5 eV, which corresponds to wavelengths of 500 nm, have been the goal of this study.
88 Anatase TiO2 shows a single sharp edge, while TiO2-xNy shows a second band edge as can be seen in Figure 36. T he first band edge is attributed to the oxide at 390 nm and the weaker shoulder due to nitrogen-doping is at 450 nm. The absorption is probably due to the O2+ Ti4+ charge transfer. This is a result of the excitation of electrons from the valence band O2p to the conduction band Ti(dxy) [8,15]. Jang et al. concluded that th is second peak is responsible for the visible-light catalytic activity . The presence of the second absorption edge could be due to the substitutional N-doping Asahi et al. concluded this edge is due to the excitation of electrons from the valence band of N2p to the conduction band Ti(dxy) instead of from O2p as in TiO2 .
89 0 1 2 3 4 5 6 200300400500600700800Wavelength (nm)Optical Absorption (KubelkaMunk) 164A TiO2 675C NH3 87ml/min 6h 400C 02 9ml/min 15m Inlet Boat Right Right Section 164B TiO2 675C NH3 87ml/min 6h 400C 02 9ml/min 15m Inlet Boat Middle Section 164C TiO2 675C NH3 87ml/min 6h 400C 02 9ml/min 15m Outlet Boat Figure 36 Optical Absorption (Kubelka-Munk ) of Thermochemically Ammonia Treated Degussa P-25 TiO2 The displacement of the first absor ption edge is probably caused by the transformation of anatase to rutile due to high temperature heating. From the presence of the additional edge in the visi ble-light range it can be supposed that the modified catalysts should be active under visible-light illumination . Li et al. concluded that the second absorption band is due to doped nitrogen atoms. They reasoned that: 1. The band does not exist in untreated TiO2. 2. The yellow color of the catalyst is due to position of the absorption edge in the visible spectrum .
90 The color of untreated Degussa P-25 is brilliant white. The color of the catalysts synthesized in this study varied, and was dependent on temperature and ranged from a very pale yellow to vivid yellow, and to brown and black materials. Materials treat ed at temperatures below 600C remained white, and materials were colored as noted. Acco rding to Mozia et al., one can conclude that the higher treatment temperatur es facilitate nitrogen bonding to the TiO2 matrix . For temperatures below 600 C the increase in performance, if any, is attributable only to thermal effects. The shift in the absorption edge in Figure 36 was attributed to Ti3+ by Irie et al. This reduction of TiO2 is accomplished by H2. During thermal ammonia treatments, NH3 is decomposed into N2 and H2, which is the reducing gas . This second absorption edge indicates the formation of a new N2p band located above the O2p valence band accord ing to Yin et al. .
91 CHAPTER 6: EXPERIMENTAL SYSTEMS 6.1 Thermal and Thermochemical Treatment System Like most of the systems used duri ng this study, the system used for the thermal and thermochemical treatments we re continually upgraded and grew in complexity as greater and greater control was sought. The block diagram in Figure 37 shows that the final thermochem ical treatment system consisted of two subsystems, a gas control subsystem and a furnace reactor subsystem. Figure 37 Block Diagram of Tube Furnace Reactor System The gas control subsystem used ul tra-high purity (UHP) nitrogen (N2), product grade anhydrous ammonia (NH3) and ultra pure (UP) carrier oxygen
92 (O2). Nitrogen was used as a dry inert gas for system purgi ng, calcination, sintering and annealing. Nitrogen-doping was accomplished using a gas phase impregnation process utilizin g anhydrous ammonia (NH3) as a nitrogen source Oxygen or a mixture of nitrogen and oxyg en was used for oxidation of the thermochemically treated catalyst. A system of valves and gas specific flow meters were configured to meter specific rates of gas into a quartz reaction tube (Corning 77481). The exhausted gasses were first bubbled through silicone oil, acting as a check valve, to an exhaust snor kel. Although the positive pressure of the gasses made the silicone oil trap unnecessary it was kept as a visual safety measure due to the use of anhydrous ammonia. The tube furnace reactor system desig ned and implemented by the author consisted of a Barnstead Thermoly ne 21100 (model F21135) Tubular Furnace and a notebook computer running a custom-made Proportional-IntegralDifferential (PID) control program developed in Nati onal Instruments, LabVIEW 8.0. The computer and tube furnace were interfaced using a National Instruments Model USB6210 data acquisition device. The original end-cap insulation of the tube furnace was replaced and the temperature profile of t he furnace was studied to insure a uniform working reaction area to be used during the study Figure 38 is the profile of the temperature gradient along the horizontal axis of the tube furnace. A minimum treatment area of four inches from center was required for this study. This profile was obtained to ensure that a working reaction area could be defined within the
93 quartz tube, and as can be seen, it ext ended more than five inches from the center. Coors glazed porcelain boats (L x W x H: 4 x 3/4 x 1/2 in.; Coors #60036) were used as TiO2 treatment receptacles. 0 100 200 300 400 500 600 700 800 -10-50510Distance from Center of Furnace (inches)Temperature (C) Temperature Figure 38 Temperature Profile for Tube Furnace Reactor 6.2 Photocatalytic Reactor The photocatalytic experiments were performed using a set of three glass suspension batch reactors that was fabr icated by Mr. Chuck Garretson, who is the CERCÂ’s lab technician. Figure 39 shows the general configuration of the system. A one liter Pyrex beak er was used for the housing of the batch reactor and a plastic mounting plate, as seen in Figure 40, was used for attaching the necessary connections. A cooling coil, using tap water for heat transfer, was made of inch stainless steel tubing, which fit loosely around the interior
94 circumference of the beaker. An aerat ing stone was used to diffuse either compressed air or oxygen into the su spension. Experiments were conducted using breathing quality compressed air and ul tra pure carrier oxygen, which were metered through gas specific flow meters. Four Pyrex (9820) culture tubes housed 150 watt halogen light sources that were mounted through holes in the mounting plate (Figure 40). Simulated solar spectrum (SSS) experiments were c onducted using this setup. For visible simulated solar spectrum (VSSS) experim ents, four Pyrex (9820 Â– 18) culture tubes were inserted inside of the larger tubes using stainless steel wire as a spacer for the ultraviolet filter solu tion. A five molar solution of NaNO2 was used as a filter for ultraviolet (UV) light thereby allowing only the transmission of wavelengths in the visible spectrum. A thermocouple, also mounted through the top plate, monitored the temperature of the soluti on. A magnetic stirrer and stirring rod were used to suspend t he particles throughout the solution.
95 Figure 39 Batch Reactor for Photocatalytic Experiments
96 Figure 40 One Liter Photocatalytic Batch Re actor with Stainless Steel Cooling Coils
97 6.3 Characterization of Light Source The type of light source used by other researchers in this area varies considerably. Typically, the use of xenon lamps, LEDÂ’s, halogen lamps and fluorescent lamps appear in most studies Many of the studies focus on one particular application, and select their light source in accordance. This study considered both the effects of ultr aviolet and visible-light radiation. The experiments conducted were cat egorized as either simulated solar spectrum (SSS) or visible simulated so lar spectrum (VSSS), and informally as whole light or visible-light. The li ght sources were four 150 watt, 120 volt, halogen lamps (Bright Effects Model #LBPQ 150T4/JCD). The lampsÂ’ irradiance profile was characterized using an Ocean Optics Spectrometer fitted with an Ocean Optics UV-CC3 filter to attenuate the signalsÂ’ intensity and Spectra Suite software. This process wa s then repeated for the total so lar irradiance incident on the filter measured here in Tampa, Florida on Ap ril 26, 2007 at 1:30PM. Figure 41 shows the spectra for wavelengths in the ultraviolet range are similar in intensity for both the lamps and solar radiation. Although the lamps create spectra similar to a black body radiati on of approximately 3000 K, the wavelength intensity in the visible spectrum is cons iderably greater than the solar irradiation measured.
98 0 0.2 0.4 0.6 0.8 1 1.2 1.4 1.6 1.8 2 310330350370390410430Wavelength (nm)uW/cm^2/nm Solar Irradiance Halogen Lamp Figure 41 Spectra for Solar and Simulated Solar Spectrum Irradiation To limit the halogen lampsÂ’ radiation to the visible spectrum, a five mole solution of sodium nitrate was pr epared by heating the solution at low temperature while magnetical ly stirring until a clear solution was formed. The solution was syringed in between the culture tubes as described above. Figure 42 shows the spectra for the lamps with and without the UV filt er. This figure shows that the filter solution begins a ttenuating the radiati on at approximately 410 nm and completely absorbs the radiation at wavelengths below 390 nm. It is on this basis that experiments that are considered VSSS (visible) are said to have no contribution from the ultraviolet radi ation. At the same time, it can also be seen that the filtering solution does not attenuate the wavelengths in the
99 visible spectrum therefor e allowing for direct comparison between SSS (whole) and VSSS (visible) experiments. 0 500 1000 1500 2000 2500 3000 300400500600700800Wavelength (nm)Counts (a.u.) With NaNO2 Filter Without NANO2 Filter Figure 42 Spectra for Halogen Lamps With and Without Ultraviolet NaNO2 Solution Filtering
100 CHAPTER 7: EXPERIMENTAL PROCEDURES 7.1 Preparation of Thermally Tr eated Photocatalytic Materials Two Coors glazed porcelain boats were weighed, then filled with Degussa P-25 TiO2, then weighed again to re cord the initial catal yst weight. Figure 43 shows the configuration of the calcination system A Barnstead Thermolyne 21100 Tube Furnace was fitted with a 36 inch long, 1 inch in diameter Corning quartz tube that served as a treatment chamber. The boats containing the catalyst were centered inside of the reactor.
101 Figure 43 Original Calcination System Nitrogen was used as a dry inert gas for the calcination environment. The temperature was raised 100C per five mi nutes until the calcin ation temperature was reached. Treatment temperatur es ranged from 275C to 825C for three hours. At the conclusion of the treat ment the material cooled to ambient temperature under nitrogen flow. 7.2 Preparation of Thermochemi cally Treated Photocatalytic Materials Two Coors glazed porcelain boats we re weighed then filled with Degussa P-25 TiO2, then weighed again to re cord the initial catal yst weight. Figure 44 shows the configuration of the thermo chemical treatment system. A Barnstead
102 Thermolyne 21100 Tube Furnace was fitt ed with a 36 inch long, 1 inch in diameter Corning quartz tube that served as a reaction chamber. The tube was connected by compression fittings to t he gas delivery and outlet systems. The boats containing the catalyst were c entered inside of the reactor tube. Figure 44 Tube Furnace Reactor System Nitrogen-doping was done by a gas phase impregnation process using anhydrous ammonia (NH3) as the nitrogen source. To begin this process, the atmosphere in the reactor was replaced by ultra-high purity (UHP) nitrogen (N2) to suppress the affects of the ambi ent atmosphere on the thermochemical process. The reaction tube had a volume of approximately 485 cm3, and the concentration of nitrogen was calculated by the following equation: 01V VTe C where C is the concentration of nitrogen, VT is the volume of nitrogen at time T, and V0 is the volume of the reaction tube. The time required to replace the atmosphere at a particular flow rate was ca lculated from this equation. By this
103 calculation, 46 minutes is required to co mpletely replace the ambient atmosphere with nitrogen at a flow rate of 155 mL/min. However, a time of 15 minutes was used, which resulted in replacement of 99.168% of the ambient atmosphere. After 15 minutes, the nitrogen flow was stopped and anhydrous ammonia, at a flow rate between 6.6 mL/min. and 87.3 mL/min., was begun. The ammonia flowed into the reactor for six minutes prio r to the start of t he thermal sequence. The temperature was raised at a ra te of 100C per five minutes with NH3 at the required flow rate until the treatm ent temperature was reached. The 33 minutes required to reach the treatment temperature, combined with the six minutes prior to the warm-up proc ess, provided 39 minutes for NH3 to flow into the reactor. This allowed for 99.91% of the nitrogen atmosphere to be replaced by NH3 at the start of the doping process. Once the treatment temperature was reached the thermochemical tr eatment process continued for three hours. At the completion of the process, nitrogen was us ed to cool the ma terial to ambient temperature. 7.3 Experimental Procedures fo r Photocatalytic Experiments Photocatalytic experiments were per formed using a glass batch reactor system, which was described in detail abov e. A 20 ppm methyl orange solution using de-ionized water was prepared us ing (A.C.S. Reagent) MO from SigmaAldrich. Methyl orange was dissolved into solution using a magnetic stirrer for 15 minutes. Samples were drawn and the initial concentration was measured and
104 calculated in accordance with Be ers Law using an Ocean Optics USB2000 optical spectrometer. Once the 20 ppm solution was prepared, 150 mL of the MO solution was reserved in a beaker. The prepared catalyst was ground in a crucible with 2 mL of MO solution to de-agglomerate the mate rial. Portions of the reserved MO solutions were added to dilute the catal yst paste that was then poured back into the beaker. This process was repeated wit h the remaining MO solution until the maximum amount of the catalyst wa s recovered from the crucible. To further de-agglomerate the particles and achieve a Langmuir equilibrium, the MO-catalyst solution was t hen sonicated using a Fisher-Scientific Sonic Dismembrator Model 100. Sonica tion was performed at 5 watts (RMS) for 15 minutes while being magnetica lly stirred. It is impor tant to note that during this time, adsorption of the pollutant onto the surface of the catalyst also took place altering the initial concentration of the MO solution. This process and its importance were detailed above. The solution was then moved to the experimental area where it again was pl aced on a magnetic stirrer. Figure 45 shows the photocatalytic experimental set up.
105 Figure 45 Photocatalytic Batch Reacto rs for SSS and VSSS Irradiation Experiments The cooling assembly, which housed the light sources, was then placed into the beakers and connected to tubing for cooling and air flow. Cooling was done by heat transfer using tap water ci rculating through the stainless steel tubing. A thermocouple was placed in each reactor and used to monitor the solution temperature. As noted in the literature, ther e is negligible Arrhenius effect, but the temperatur e was controlled by stirring speed and water flow to create consistency of environment for s eparate experiments. The photocatalytic experiments were conducted at approximat ely 40C. Breathing Quality Air was used as an oxygen source. The flow rate was controlled by flow meters and bubbled into the solution by aerating stones at a rate of 0.5 L/min.
106 2.0 mL samples, representing zero time, were drawn using Micromate 5cc glass syringes and placed into micro-cent rifuge tubes. The solution was then irradiated using the light s ources detailed above. Both simulated solar spectrum experiments and visible solar spectr um experiments were conducted for durations of three to six hours. Sample s were drawn at 30 minute intervals for the first three hours, followed by one hour intervals if the experiments duration was six hours. At the completion of the experiment, the samples that were collected were centrifuged using an Eppendorf 5414C centri fuge at 8,000 rpm for 15 minutes. The samples were then syringed to new tubes then centrifuged again at 8000 rpm for 15 minutes. The concentrations of the samples were then calculated by measuring the absorbance of the samples using a spectr ometer. The results were compiled and the rate constant for t he material was calculated. 7.4 Control Experiments Using Degussa P-25 TiO2 Control experiments using Degussa P-25 TiO2 for both simulated solar spectrum and visible solar spectrum we re conducted in the exact manner as the photocatalytic procedures det ailed above. The experimental conditions that affect the photocatalytic process have already been addressed. These included the initial concentration of the pollutant, catalyst loading, catalyst particle size, intensity of the light source, initial pH of the solution and the concentration of
107 oxygen in the solution . The design of these experimental procedures allow for the control or optimization of these parameters. Every effort was made to perform consistent repeatable experiments allowing for the comparison of the results.
108 CHAPTER 8: EXPERIMENTAL RESULTS 8.1 Untreated Degussa P-25 TiO2 Baseline experiments using untreated Degussa P-25 TiO2 were performed under no irradiation, simulated solar spectr um (SSS or whole) for varying times, and visible simulated solar spectrum ( VSSS or visible) irradiation for both sonicated and non-sonicated suspensi ons. Figure 46 shows the normalized change in concentration for each of these conditions. Consistent with t he literature, TiO2 in the presence of a pollutant, but without an irradiation source (labeled dark below) produ ced no de-coloration of methyl orange . According to Guettai et al., this shows that the de-coloration of methyl orange is a purely photocatalytic effect. In Figure 46, the experimental points for VSSS MO degradation can be approximated by a linear dependence from time whereas, for whole spectrum, the experimental points for the initial stage of MO degr adation are fitted well with the exponential function from time. Such a big difference in the degradation kinetics can be explained by di fferent limiting stages and types of photocatalytic reactions. Particularly, the linear dependence can be well described by a zero ordered photocatalytic reaction. For a zeroth order reaction, the rate remains constant throughout the reaction and is i ndependent of the concentration of the
109 reactant(s) . Zeroth order reactions are typically found when the surface of the catalyst is saturated by the reactant s. In photocatalytic reactions, it may be the case when the rate of contaminant degradation on the catalytic surface is limited by the intensity of electron-hole excitation. The plots for simulated solar spec trum (SSS) light can be considered a good first approximation for a first order chemical reacti on. A first order reaction is a reaction whose rate depends on the concentration of only one reactant raised to the first power . Other reactants can be present, but each will be zeroth order. As was shown earlier, the kinetics ar e more complicated and it is difficult to classify the reaction as purely zeroth and/or first order, although this can be a good approximation. This further explains the need for more sophisticated models such as the Langm uir-Hinshelwood model. The plots for both sonicated and non-sonicated TiO2 in visible-light also show a difference in rate. The material used in this study came from large bulk quantities of Degussa P-25 TiO2. The particle size is commonly reported near 30 nm, with a surface area of 50 m2/g. However, due to simple settling, the material compacts over time. This explains why the catalyst was ground prior to each experiment and sonicated. The differenc e in the rates is attributed to the deagglomeration of the particles, thereby exposing a greater surface area to the pollutant to react at.
110 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9 1 050100150200250300350400Time (m)C/C0 Dark VSSS VSSS Sonicated SSS SSS Stir 23h SSS Stir 6h Figure 46 Change in Concentration as a Function of Time for Methyl Orange in the Presence of 1 g/L of Untreated Degussa P-25 TiO2 Under No, SSS and VSSS Irradiation By using the integrated rate law, we are able to convert the first order reaction rate using calculus. As s hown in Figure 47, the negative natural logarithm of the ratio of t he concentration at time (t) to the initial concentration shows what fraction of the original pollutant remains at time (t) .
111 y = 0.0027x + 0.2462 R2 = 0.99060 0.2 0.4 0.6 0.8 1 1.2 1.4 050100150200250300350400Time (m)-ln(C/C0) SSS Stir 6h Linear (SSS Stir 6h) Figure 47 Integrated Rate Law Plot for Untreated Degussa P-25 TiO2 Under SSS Irradiation The slope of the integrated rate la w plot (Figure 47 above) was then used as the apparent rate constant for the reacti on. This rate is given in units of 1/time, which is shown in Figure 48 [20, 22, 60]. It is again im portant to note that a straight line will only occur for truly firs t order chemical reac tions. Therefore the R-squared value is an important indicato r of the validity of categorizing the reaction purely as first order. Again, it is apparent that a good first approximation can be made for the reaction in SSS irradiation to be considered first order. It is also obvious that the complexity of t he chemical kinetics cannot be considered purely first order. This further reinforces the need for models such as the Langmuir-Hinshelwood model. The methods used here are identical to the
112 methods used by Guettai et al. and Al -Qaradawi et al. as both studied the degradation of methyl orange in the presence of TiO2 [17, 18]. The calculated rate constant for Degussa P-25 TiO2 under SSS irradiation is given in Figure 48. -0.001 -0.0005 0 0.0005 0.001 0.0015 0.002 0.0025 0.0030123Light SourceApparent Rate Co nstant (1/min) ErRate Er+SSS Dark Figure 48 Apparent Rate Constants for Untreated Degussa P-25 TiO2 Under No and SSS Irradiation 8.2 Effects of Oxygen Concentr ation on Photocatalytic Rate One of the established parameters a ffecting the photocatalytic rate is oxygen concentration in the suspension. Figure 49 shows the affects of using one and two aerating stones. The air flow rate for each reactor was the same (0.5L/min) along with all other conditions for the experiment. Consistent with the
113 literature, by diffusing a higher concentration of oxygen, more oxygen is available for electron scavenging and therefore a higher photocatalytic effect . 0 0.005 0.01 0.015 0.02 0.025 NRSSRNumber of Aerating StonesApparent Rate Constant (1/min) RateOne Aerating Stone Two Aerating Stones Figure 49 Apparent Rate Constant for Untreated Degussa P-25 TiO2 Under SSS Irradiating Using One and Two Aerating Stones In a similar fashion oxygen and air were again bubbled into the suspension at the same rate (0.5 L/min.) and under t he same conditions. As seen in Figure 50 the use of pure oxy gen increased the photocatalytic activity.
114 0 0.002 0.004 0.006 0.008 0.01 0.012 0.014 0.01600.20.40.60.811.2Loading (g/L)Apparent Rate Constant (1/min) ErRate Er+ ErRate Er+ A ir Oxygen Figure 50 Comparison of Air and Oxygen on the Apparent Rate Constant Throughout this study we were able to control the parameters that have the greatest influence of over photocatalytic efficiency. While doing so we were also able to show the effects of the rmal and thermochemical treatments on the performance of Degussa P-25 TiO2. Figure 51 below is a HRTEM image of pure Degussa P-25 TiO2. This image clearly shows the atomic lines and gr ain boundaries that are understood to be the location of the oxygen deficiencies.
115 Figure 51 HRTEM Showing the Grain Boundaries of Pure Degussa P-25 TiO2 8.3 Effect of Thermal Treatment on th e Photocatalytic Activity of TiO2 The established theory is that the rmal treatments cause the removal of oxygen ions. This in turn creates ox ygen vacancies, which accelerates the anatase to rutile transformation. These oxygen vacancies cause the lattice to contract and the volume to shrink w hen two of the six Ti Â–O bonds are broken . Oxygen vacancies are believed to be formed in the grain boundaries . It is these very oxygen vacancies that are believed to increase the photocatalytic effects of TiO2.
116 The improvement in efficiency can be attributed to the reduction of TiO2, which was explained in detail earlier . Asahi et al. further concluded that these oxygen deficient sites ar e important for the visible response to take place. This reduction causes a red-shift to wa velengths longer than 400 nm, which is attributed to the reduction of Ti4+ to Ti3+. Hence, the rutile particles in the P-25 powder are considered to contain the Ti3+ ions that create electron donors. Ohno et al. concluded that a Â“fairly large band bending is generated in the rutile particles.Â” They attribute the higher efficiency of P-25 to this . The curves in Figure 52 show the concentration as a function of time for calcinated TiO2. As described earlier, the cata lyst was calcinated at varying temperatures from 275C to 625C for three hours under a dry inert nitrogen atmosphere. The curves below show a good first approximation of first order kinetics.
117 0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9 1 050100150200250300350400Time (m)C/C0 275 325 375 425 500 525 575 Figure 52 Change in Concentration as a Function of Time for Methyl Orange in the Presence of 1 g/L of Calcinated Degussa P-25 TiO2 Under SSS Irradiation As a result of this first order approx imation the curves were plotted using the integrated rate law as shown in Figur e 53. As can be seen, the slopes of these curves have high R-squared values and give a high level of confidence in their use. What should be further cautioned is that first order reactions must not have a build up of intermediates during t he reaction . As illustrated above, there were occasions that intermediat es that could not be degraded, and others that could, were produc ed during some of the phot ocatalytic experiments. Therefore, these results can be used only as an approximation.
118 y = 0.0038x + 0.2266 R2 = 0.9951 y = 0.0084x + 0.183 R2 = 0.9998 y = 0.0075x + 0.2482 R2 = 0.9916 y = 0.0072x + 0.1464 R2 = 0.999 y = 0.0051x + 0.1697 R2 = 0.9981 y = 0.0026x + 0.1184 R2 = 0.99590 0.2 0.4 0.6 0.8 1 1.2 1.4 1.6 1.8 050100150200Time (m)-ln(C/C0) 325 375 425 525 575 625 Linear (325) Linear (375) Linear (425) Linear (525) Linear (575) Linear (625) Figure 53 Integrated Rate Law Plot for Calcinated Degussa P-25 TiO2 Under SSS Irradiation Plotted in Figure 54 are the apparent rate constants for thermally treated (calcinated) Degussa P-25 TiO2. This plot shows a broad range of temperatures that produce a positive effect on the photoc atalytic rate. This study found that temperatures from 375C to 575C had a pos itive effect on the photocatalytic rate, and that temperatures above 600C pr oduced a declining effect. The study by Ihara et al. conclud ed that 400C is the transit ion temperature at which oxygen deficiencies are created. They furt her concluded that this is also the temperature at which the ma ximum number of deficiencie s are created . This decline over 600C in photocatalytic effect is attributed to the increase in particle size and a shift from anatase to rutile fractions in the catalyst. This was supported by Kosowska et al. who f ound that materials prepared below 600C
119 were mostly anatase, whereas temperatur es above 650C produced a mixture of anatase and rutile phases . 0 0.001 0.002 0.003 0.004 0.005 0.006 0.007 0.008 0.009 0.01 250300350400450500550600650Temperature (C)Apparent Rate Constant (1/min) ErRate Er+ ErRate Er+Pure TiO2 SSS Calcinated TiO2 Figure 54 Apparent Rate Constants for Calcinated Degussa P-25 TiO2 Under SSS Irradiation 8.4 Characterization of Therma lly Treated Degussa P-25 TiO2 Figure 55 shows values for anatase and rutile particle size, number of anatase and rutile particles per volu me, photocatalytic efficiency and mass fraction of rutile particles for temperatures in the range of 275C to 625C for calcinated Degussa P-25 TiO2. Rutile particles were found to have an average size of 30 nm and anatase an average size of 10 nm. It can be seen that the size of these particles stays relatively consistent for temperature below 525C. However, at temperatures above 575C t he average particle size increases rapidly.
120 The relative photonic efficiency is a pl ot of the apparent rate constants as shown in Figure 54 above. What can be gle aned from this plot is that the decline in the photonic efficiency (apparent rate co nstant) coincides with an increase in particle size for anatase and rutile particles. It is the rutile particles however that begin to increase much more rapidly than the anatase particles. This is further indicated by the increase in the mass frac tion of rutile particl es increasing over 575C. 0 10 20 30 40 50 60 200300400500600700Temperature (C)Parameters Anatase nm Rutile nm (Relative Photonic Efficiency) x 10 Num An Part*E5 Num Rut Part*E6 [Mass fract rutile]*100 Figure 55 Characterization of Thermally Treated (Calcinated) Degussa P-25 TiO2 Under SSS Irradiation 8.5 Effect of Thermochemical Ammonia Treatment on the Photocatalytic Activi ty of Degussa P-25 TiO2 One of the major objectives of this st udy was the pursuit of a visible-light activated photocatalyst using Degussa P-25 TiO2 as the starting material. As
121 was described in detail earlier, anhydr ous ammonia was used as a nitrogen source for the thermochemically treated photocatalyst consistent with the work done by Kosowska et al. . Doping TiO2 with nitrogen was concluded to improve the visible-light abs orption, which increased t he number of photons that could take place in the r eaction . The overall performance is improved due to this mechanism. The often cited study by Asahi et al. regarding nitrogen-doping, concluded that the visible-light response is due to the formation of an isolated band, which consisted of N2p orbitals above the O2p orbitals in the valence band . Further, they concluded that the doping could not exce ed 1% of the nitrogen. The results below, regarding gas flow ra te point to a similar conclusion. As discussed previously, visible-light experiments produced reactions that could not be considered first order, but could on a first approximation be considered zeroth ordered. Figure 56 shows that photocatalytic effect of thermochemically ammonia treated Degussa P-25 TiO2 on methyl orange is more accurately modeled as zeroth order reaction.
122 0.7 0.75 0.8 0.85 0.9 0.95 1 050100150200250300350400Time (m)C/C0 6.6 12.7 24.8 49.1 87.3 Degussa Figure 56 Change in Concentration as a Function of Time for Methyl Orange in the Presence of 1 g/L of Thermochemica lly Ammonia Treated Degussa P-25 TiO2 Under VSSS Irradiation The calculation of the rate is a si mple average of the slope over time multiplied times the initial concentration. The result is presented as the decoloration rate given in units of pollut ant ppm decay per minute. Figure 57 shows the De-coloration Decay Rate (which can be thought of as the rate constant) for thermochemically treated (nit rogen-doped) Degussa P-25 TiO2.
123 0 0.005 0.01 0.015 0.02 0.025 0.03 0.035 0.04 0.045 400450500550600650700750Temperature (C)Decoloration Decay Rate (ppm/min) Rate Degussa VSSS Figure 57 De-coloration Decay Rate as a Function of Treatment Temperature for Thermochemically Ammonia Treated Degussa P-25 TiO2 Under VSSS Irradiation Figure 57 above shows t hat the highest rate of de-coloration occurred at 675C. Using a similar method but r eactive red and phenol as a pollutant, Kosowska et al. found simila r results for nitrogen-doped TiO2 as shown in Figure 58. By comparing Figures 57 and 58 one can conclude that the optimized temperature of the TiO2 thermochemical treatment is different for different pollutants to be degraded and may al so depend of the precursor.
124 Figure 58 Dye Concentration as a Function of Treatment Temperature for NitrogenDoped TiO2 for Reactive Red and Phenol by Kosowska et al.  8.6 Optimization of Thermochemical Ammonia Treatment Flow Rate The initial study was expanded into an investigation of ammonia flow rate and its effect on the photocatalytic rate. Having determined an optimum temperature of 675C for visible-light activated photocatalyst, experiments with flow rates from 6.6 mL/min. to 87.3 mL/min. were conducted. Figure 59 shows that the highest MO de-coloration rate was obtained for a catalyst thermochemically ammonia tr eated at a flow rate of 12.7 mL/min.
125 0 0.002 0.004 0.006 0.008 0.01 0.012 0.014 0.016 0.018 0102030405060708090100Flow Rate (mL/min)Decoloration Decay (ppm/min) Rate Degussa VSSS Figure 59 De-coloration Decay Rate as a Function of Ammonia Flow Rate at 675C for Thermochemically Ammonia Treated Degussa P-25 TiO2 Under VSSS Irradiation The color of the catalyst ranged from a very pale yellow at 6.6 mL/min. to a more vivid yellow at 12.7 mL/min., show n in Figure 60. A yellow material is commonly reported in published wo rks for doped materials. Figure 60 Thermochemically Amm onia Treated Degussa P-25 TiO2 at 675C for 3 Hours at 12.7 mL/min.
126 For flow rates above 12.7 mL/min. t he catalyst developed a gradient of color across the boats from the gas inlet side being a green, black or brownish color to a pale yellow on the outlet boat si de as shown in Figure 61. Ihara et al. concluded that the reduction reaction t hat takes place during thermal ammonia treatments are responsible for this change to yellow material . Although not used here in this study, further stud y found that a small zone existed for maximum deposition of nitrogen that resulted in heavily nitrided black materials. Figure 61 Thermochemically Amm onia Treated Degussa P-25 TiO2 at 675C for 3 Hours at 24.8 mL/min. Figure 62 depicts the area of the color gradient that is typically experienced for flow rates of 24.8 mL/min. The darken ed material started from the top far right hand side of the boat on the gas inlet side of the tube furnace reactor and penetrated to the bot tom of the boat. From th is point, the materialsÂ’ color gradient followed an angle upward toward s the surface of t he catalyst. At higher rates, this gradient would extend th rough the right side boat into the left side boat.
127 Figure 62 Depiction of Color Gradient After Thermochemically Ammonia Treated At or Above 24.8 mL/min. Figure 63 shows the color gradient of the material when thermochemically ammonia treated at 675C at 87.3 mL/min. fo r 3 hours. Notice the color gradient has now extended across the gas inlet boat and has reached the left hand boat. Figure 63 Thermochemically Amm onia Treated Degussa P-25 TiO2 at 675C at 87.3 mL/min. for 3 Hours 8.7 Effects of Nitrific ation of Thermochemically Ammonia Treated TiO2 A short study was done using the same thermochemical treatment procedures for flow rates of 24.8mL/min. This was done to investigate the photocatalytic effect that titanium nitride (found by X RD analysis, see below) had on the rate of MO discoloration. Figure 64 shows the change in concentra tion for treatment temperatures from 325C to 825C. It can be seen t hat temperatures below 550C are
128 insufficient to dope TiO2 using anhydrous ammonia as a nitrogen source, which is consistent with the findings of Kosowska et al. . 0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9 1 050100150200Time (m)C/C0 325 425 525 625 725 825 Figure 64 Effects of Nitride on the Photocatalytic Effects of Thermochemically Ammonia Treated Degussa P-25 TiO2 Figure 65 further illustrates the effect that the thermoch emical treatment by ammonia has on the photocatalytic effect of TiO2. Temperatures below 625C produced white materials that have essentia lly been calcinated. The increase in effect can only be attributed to the reduction of TiO2 similar to the effects seen in thermal treatments using nitrogen.
129 0 0.005 0.01 0.015 0.02 0.025 0.03 300400500600700800Temperature (C)Apparent Rate Constant (1/ppm) ErRate Er+ ErDegussa SSS Er+ Figure 65 Apparent Rate Constant for T hermochemically Ammonia Treated (24.8 mL/min.) Degussa P-25 TiO2 Under SSS Irradiation The phase structure of the untreated Degussa P-25 and the thermochemically ammonia treated Degussa P-25 TiO2 samples was characterized by x-ray diffraction (XRD). For the quantitative characterization for the phase identification and av erage grain size has been carried out using Philips XÂ’pert pro PreFix powder x -ray diffractometer with CuK radiation ( =1.54060 ). The incident and diffraction slit width used for all the experiments are 1 and 2 respectively and the incident beam mask used corresponds to 10 mm. The sample preparations for the XRD measur ement are strictly followed to obtain maximum signal to noise raise.
130 The anatase TiO2 content (CA) was estimated according to XRD patters based on the following equation: % 100 *R A A AA A A C where AA and AR are areas for anatase (101) peak and the rutile (110) peak (110) respectively. The grain size (D) was estimat ed from the Scherrer equation : ) cos( D, where is the shape constant which was taken as 1.54060 , is the wavelength of the radiation, is the diffraction angle, and is the half-value width of the anatase diffraction peak. The XRD plot shown in Figure 66 s hows the formation of titanium nitride during the thermochemical process at a temperature of 825C. Additionally, it also shows the transformation from anatase to rutile phases during this process. These results support the findings of Kobay akawa et al. who found that TiN was formed when high doping ratios of nitrogen were formed in TiO2 . The samples for heavily treated materials for th is study were brownish to black in color, which is similar to the findings of Kobayakawa et al. Further, TiN was not observed for materials that were white or yellow in color.
131 0 20 40 60 80 100 120 140 160 180 200 20304050602 Theta (degree)Counts NH3 825C 3h TiO2 Deg Rutile Tab Anatase Tab TiN Tab Figure 66 Formation of Titanium Nitride Du ring Thermochemical Ammonia Treatments at 825C of Degussa P-25 TiO2 Figure 67 shows the XRD analysis fo r thermochemically treated Degussa P-25 TiO2 as it pertains to anatase and rutile phase composition. This plot further identifies 625C as a critical tem perature for the transformation of anatase to rutile phase.
132 0 500 1000 1500 2000 2500 3000 3500 4000 242526272829Position (2Th)Counts NH3 3h 325C NH3 3h 425C NH3 3h 525C NH3 3h 625C NH3 3h 725C NH3 3h 825C A natase Peak Rutile Peak Figure 67 XRD Comparison for Thermochemically Ammonia Treated Degussa P-25 TiO2 8.8 Characterization of Thermoch emically Ammonia Treated TiO2 Figure 68 shows measured and calculat ed values as they pertain to thermochemically treated Degussa P-25 TiO2. Rutile particles were found to have an average size of 30 nm and anatase an average size of 10 nm. It can be seen that the size of these particles st ays relatively constant for temperature below 625C. However, at temperatur es above 625C the average particle size increases rapidly, especially for rutile. It is interesting to note the contrast in temperatures between the calcinated treatment and thermochemical ammonia treatments with regar d to their particle size. It can be see that the thermochemical treatment had some influe nce on the temperature point at which the transition towards a more rutile mass fraction occurred.
133 The relative photonic efficiency is a pl ot of the apparent rate constants. What can be gleaned from this plot is t hat the decline in the photonic efficiency (apparent rate constant) coincides with an increase in particle size for anatase and rutile particles. It is the rutile par ticles however that begin to increase much more rapidly than the anatase pa rticles. This is further indicated by the increase in the mass fraction of rutile particles increasing over 625C. 0 50 100 150 200 250 300400500600700800Temperature (C)Parameters Anatase nm Rutile nm (Relative Photonic Efficiency) x 100 Num An Part*E5 Num Rut Part*E6 [ Mass fract rutile ] *100 Figure 68 Characterization of Thermochemic ally Ammonia Treated Degussa P-25 TiO2 Under SSS Irradiation Figure 69 is the HRTEM image of t hermochemically ammonia treated TiO2. The results of HRTEM analysis perfo rmed by Dr. Yusuf Emirov, indicates that the lattice of ammonia treated TiO2 is oxygen deficient and can be described as Ti9O17. No nitrogen compounds were f ound since their content was probably too low to be detected.
134 Figure 69 HRTEM of Thermochemically Ammonia Treated Degussa P-25 TiO2 The optical absorbance was calc ulated using the Kubelka-Munk technique. As described earlier, an absorption edge is formed during the thermochemical ammonia treatments that allows for absorption of visible length radiation as shown in figure 70.
135 0 1 2 3 4 5 6 250300350400450500550600650700Wavelength (nm)Optical Absorption (Kubelka-Munk a.u.) Pure Degussa P-25 TiO2 164B TiO2 NH3 6h Figure 70 Optical Absorbance (Kubelka-Munk ) of Thermochemically Ammonia Treated Degussa P-25 TiO2 In Figure 71, Jang et al. plo tted the absorption for anatase TiO2 and TiO2XNX. TiO2-XNX shows two absorption edges. The first is the edge at 390 nm that is attributed to absorption by the ox ide, and a second weaker edge at 451 nm that is attributed to nitrogen-doping. They concluded that this second shoulder peak is responsible for visible-li ght photocatalytic activity .
136 Figure 71 (A) TiO2-XNX and (B) Pure TiO2 Calcinated at 400C  What needs to be stressed here is that while Figure 70 is a good example of the ability to control the absorption edge, the material itself performed very poorly photocatalytically. Th is is an area of considerab le research today, and a topic that was briefly investigated during this study. Further study of this critical component is necessary to draw any further conclusions.
137 CHAPTER 9: CONCLUSIONS AND RECOMMENDATIONS The results of this study are as follow: 1. Thermal treatments of Degussa P-25 TiO2 improve the photocatalytic effect over untreated Degussa P-25 TiO2 for SSS. 2. The most photoactive catalyst prepared by thermal treatment was at 375C for SSS. Improvement in photocatalytic efficiency was nearly two times that of untreat ed Degussa P-25 TiO2. 3. A simple and inexpensive method of creating VIS-active photocatalyst by modification of Degussa P-25 TiO2 with nitrogen was accomplished. 4. A thermochemical ammonia tr eatment process was optimized for temperature, duration and gas flow rate for nitrogen-doping Degussa P-25 TiO2. 5. The most photoactive catalyst pr epared by thermochemical treatment of Degussa P-25 TiO2 was at 675C at 12.7 mL/min. for 3 hours under VSSS. Improvement in photocatalytic effect is 1.5 times higher than untreated Degussa P-25 TiO2.
138 6. An additional absorption edge in the visible range was observed for thermochemically ammonia treated Degussa P-25 TiO2 for treatment temperatures above 625C. 7. Based on XRD results, the temper ature of the phase transition from anatase to rutile is 625C for thermally treated Degussa P-25 TiO2, and 725C for thermochemically treated Degussa P-25 TiO2. This indicates that nitrogen-doping inhibits the phase tr ansition from anatase to rutile. 8. The results of HRTEM analysis indicate that the lattice of thermochemically ammonia treated Degussa P-25 TiO2 is oxygen deficient and can be described as Ti9O17. No nitrogen compounds were found and it is believed that t heir content was probably too low to be detected. 9. The XRD patterns of thermochemically treated Degussa P-25 TiO2 exhibit diffraction peaks of TiN compound at 825C. 10. The color of the thermochem ically treated Degussa P-25 TiO2 was dependent upon treatment te mperature, duration and gas flow rate. 11. The color ranged from vivid white to gr ey to pale yellow to vivid yellow to green to brown to black, which may be attributed to TiN content on the surface. 12. The phase composition of the most Vis-active photoactive catalyst was 88% anatase and 12% rutile for thermo chemically treated Degussa P-25 TiO2 at 675C.
139 13. A three step thermochemical method was developed for heavy nitrogendoping of TiO2 using anhydrous ammonia as a nitrogen source. 14. Thermochemical ammonia treatment of Degussa P-25 TiO2 is a very promising and inexpensive method for Vis-active photocatalyst preparation on a commercial scale compared with sputtering and ion implantation techniques. Anhydrous ammonia as a nitrogen sour ce for the doping of Degussa P-25 TiO2 has the potential to create an effi cient and inexpensive visible-light photocatalyst. It has been shown here and in other studies how to create a visible-light photocatalytic effect. As noted, many authors have attributed this to the creation of states c apable of utilizing energy fr om the visible spectrum. An important question is Â“how do we get this material to absorb more and more of the available visible-light spec trum without degrading t he efficiency of the catalyst in both the visible and ultraviolet spectra?Â” Initial work in this area was done during this study. While it was shown it is possible to significantly improve the absorption of visible wa velengths, the pr eliminary results have also shown that this can be done at the expense of photocatalytic efficiency. The improvement in visible-light absorption is not directly related to an improvement in photocatalytic efficiency. Advanc ed semiconductor theory suggests that materials can become electrically inacti ve if over doped due to a saturation effect. Could this possibly be the case here?
140 Future research should include a more rigorous treatment of the photocatalytic rate. The Langmuir-Hin shelwood equation should be used to better model the chemical kinetics that oc curs during the photocatalytic process. A focused study should be done on the optim ization of treatment temperature, flow rate and duration. Additionally, an in depth characterization of the doped catalyst would be beneficial in identifyi ng optimum doping levels, the role of interstitial and substitutional dopants as well as the effects of T iN. Finally, for practical commercial concerns, further research and development of an improved gas phase impregnation system s hould be undertaken. All in all, this is an area with a great deal of work to be done. As a final note, the reader should note that unless otherwise expressly cited, all activities of this study we re done under the direction of Dr. Nikolai Kislov. Dr. Kislov should be primarily credited with the design of experiments, experimental systems, procedures and analytical methods.
141 LIST OF REFERENCES  R. G. Salomon, Tetrahedron, vol. 39, p. 485, 1983.  UNICEF, "Water, environment and sanitation," 2007.  J. T. Chang, Y. F. Lai, and J. L. He, "Photocatalytic performance of chromium or nitrogen doped arc ion plated-TiO2 films," ICMCTF 2005, vol. 200, pp. 1640-1644, 2005.  N. A. Serpone and A. V. Emeline, "Suggested terms and definitions in photocatalysis and radiocatalysis," International Journal of Photoenergy, vol. 4, pp. 91-131, 2002.  O. Carp, C. L. Huisman, and A. Relle r, "Photoinduced reactivity of titanium dioxide," Progress in Solid State Chemistry, vol. 32, pp. 33-177, 2004.  A. L. Linsebigler, G. Lu, and J. T. Yates, "Photocatalysis on TiO2 Surfaces: Principles, Mechanisms, and Selected Results," Chemical Reviews, vol. 95, pp. 735-758, 1995.  S. O. Kasap, Principles of eletroni c materials and devices 2nd ed. New York, NY: McGraw-Hill, 2000.  D. Beydoun, R. Amal, G. Low, and S. McEvoy, "Role of Nanoparticles in Photocatalysis," Journal of Nanoparticle Research, 1999.  K. Madhusudan Reddy, B. Baruwati, M. Jayalakshmi, M. Mohan Rao, and S. V. Manorama, "S-, Nand Cdoped titanium dioxide nanoparticles: Synthesis, characterization and r edox charge transfer study," Journal of Solid State Chemistry, vol. 178, pp. 3352-3358, 2005.
142  U. Diebold, "The surface science of titanium dioxide," Surface Science Reports, vol. 48, pp. 53-229, 2003.  Y. Q. Wang, X. J. Yu, and D. Z. Sun, "Synthesis, characterization, and photocatalytic activity of TiO2-xNx nanocatalyst," Journal of Hazardous Materials, 2006.  R. I. Bickley, T. Gonz alez-Carreno, J. S. Lees, L. Palmisano, and R. J. D. Tilley, "A structural investigation of titanium dioxide photocatalysts," Journal of Solid State Chemistry, vol. 92, pp. 178-190, 1991.  I. Nakamura, N. Negishi, S. Kuts una, T. Ihara, S. Sugihara, and K. Takeuchi, "Role of oxygen vacancy in the plasma-treated TiO2 photocatalyst with visible light activity for NO removal," Journal of Molecular Catalysis A: Chemical, vol. 161, pp. 205-212, 2000.  R. B. Draper and M. A. Fox, "Titanium dioxide photosensitized reactions studied by diffuse reflectance flash photolysis in aqueous suspensions of TiO2 powder," Langmuir, vol. 6, pp. 1396-1402, 1990.  H. Gerischer, "Photoelectrochemical catalysis of the oxidation of organic molecules by oxygen on small semiconductor particles with TiO2 as an example," Electrochimica Acta, vol. 38, pp. 3-9, 1993.  J. Wang, B. Guo, X. Zhang, Z. Z hang, J. Han, and J. Wu, "Sonocatalytic degradation of methyl orange in the presence of TiO2 catalysts and catalytic activity comparison of rutile and anatase," Ultrasonics Sonochemistry, vol. 12, pp. 331-337, 2005.  T. Ohno, K. Sarukawa, K. Toki eda, and M. Matsumura, "Morphology of a TiO2 Photocatalyst (Degussa, P-25) C onsisting of Anatase and Rutile Crystalline Phases," Journal of Catalysis, vol. 203, pp. 82-86, 2001.  A. K. Datye, G. Riegel, J. R. Bolton, M. Huang, and M. R. Prairie, "Microstructural Characterizati on of a Fumed Titanium Dioxide Photocatalyst," Journal of Solid State Chemistry, vol. 115, pp. 236-239, 1995.
143  E. J. Weber and V. C. Stickney, "Hydrolysis Ki netics of Reactive Blue 19Vinyl Sulfone," Water Research, vol. 27, pp. 63-67, 1993.  N. Guettai and H. Ait Amar, "Photocatalytic oxidat ion of methyl orange in presence of titanium dioxide in aqueous suspension. Part I: Parametric study," Desalination, vol. 185, pp. 427-437, 2005.  J. M. Herrmann, "Heterogeneou s photocatalysis: fundamentals and applications to the removal of va rious types of aqueous pollutants," Catalysis Today, vol. 53, pp. 115-129, 1999.  S. Al-Qaradawi and S. R. Salman, "Photocatalyt ic degradation of methyl orange as a model compound," Journal of Photochemistry and Photobiology A: Chemistry, vol. 148, pp. 161-168, 2002.  H. M. Yates, M. G. Nolan, D. W. Sheel, and M. E. Pemble, "The role of nitrogen doping on the development of vi sible light-induced photocatalytic activity in thin TiO2 films grown on glass by chemical vapour deposition," Journal of Photochemistry and Photobiology A: Chemistry, vol. 179, pp. 213-223, 2006.  N. Guettai and H. Ait Amar, "Photocatalytic oxidat ion of methyl orange in presence of titanium dioxide in aq ueous suspension. Part II: Kinetics Study," Desalination, vol. 185, pp. 439-448, 2005.  S. Liu, X. Chen, and X. Chen, "Preparation of N-Doped Visible-Light Response Nanosize TiO2 Photocatalyst Using the Acid-Catalyzed Hydrolysis Method," Chinese Journal of Catalysis, vol. 27, pp. 697-702, 2006.  J. S. Jang, H. G. Kim, S. M. Ji, S. W. Bae, J. H. Jung, B. H. Shon, and J. S. Lee, "Formation of crystalline TiO2-xNx and its photocatalytic activity," Journal of Solid State Chemistry, vol. 179, pp. 1067-1075, 2006.  V. Pore, M. Heikkila, M. Ritala, M. Leskela, and S. Areva, "Atomic layer deposition of TiO2-xNx thin films for photocatalytic applications," Journal of Photochemistry and Photobiology A: Chemistry, vol. 177, pp. 68-75, 2006.
144  S. K. Pradhan and P. J. Reucroft, "A study of growth and morphological features of TiOxNy thin films prep ared by MOCVD," Journal of Crystal Growth, vol. 250, pp. 588-594, 2003.  S. Yin, K. Ihara, M. Komatsu, Q. Zhang, F. Saito, T. Kyotani, and T. Sato, "Low temperature synthesis of TiO2-xNy powders and films wit h visible light responsive photocatalytic activity," Solid State Communications, vol. 137, pp. 132-137, 2006.  S. Yin, H. Yamaki, M. Komatsu, Q. Zhang, J. Wang, Q. Tang, F. Saito, and T. Sato, "Synthesis of visible-light reactive TiO2-xNy photocatalyst by mechanochemical doping," Special issue in honour of C.N.R. Rao, vol. 7, pp. 1479-1485, 2005.  D. Li, H. Haneda, S. Hishita, and N. Ohashi, "Visible-light-driven nitrogendoped TiO2 photocatalysts: effect of nitrogen precursors on their photocatalysis for decomposition of gas-phase organic pollutants," Materials Science and Engineering B, vol. 117, pp. 67-75, 2005.  R. Asahi, T. Morikawa, T. Ohwaki K. Aoki, and Y. T aga, "Visible-Light Photocatalysis in Nitrogen-Doped Titanium Oxides," Science Magazine, vol. 293, pp. 269-271, 2001.  H. Irie, Y. Watanabe, and K. Ha shimoto, "Nitrogen-Concentration Dependence on Photocatalytic Activity of TiO2-xNx Powders," Journal of Physical Chemistry B, vol. 107, pp. 5483-5486, 2003.  G. R. Torres, T. Lindgren, J. Lu, C. G. Granqvist, and S. E. Lindquist, "Photoelectrochemical Study of Nit rogen-Doped Titanium Dioxide for Water Oxidation," Journal of Physical Chemistry B, vol. 108, pp. 59956003, 2004.  T. Ihara, M. Miyoshi, Y. Iriyama, O. Matsumoto, and S. Sugihara, "Visiblelight-active titanium oxide photocatalyst realized by an oxygen-deficient structure and by nitrogen doping," Applied Catalysis B: Environmental, vol. 42, pp. 403-409, 2003.  I. N. Martyanov, S. Uma, S. Rodr igues, and K. J. Klabunde, "Structural Defects Cause TiO2-based Photocatalysts to be Active in Visible Light," ChemComm, vol. 21, pp. 2476-2477, 2004.
145  M. Miyauchi, A. Ikezawa, H. Tobi matsu, H. Irie, and H. Hashimoto, "Zeta potential and photocatalytic activity of nitrogen doped TiO2 thin films," Physical Chemistry Chemical Physics, vol. 6, pp. 865-870, 2004.  O. Diwald, T. L. Thompson, T. Zubkov E. G. Goralski, S. D. Walck, and J. T. Yates, "Photochemical Activi ty of Nitrogen-Doped Rutile TiO2 (110) in Visible Light," Journal of Physical Chemistry B, vol. 108, pp. 6004-6008, 2004.  C. C. Pan and J. C. S. Wu, "Visible-light response Cr-doped TiO2-XNX photocatalysts," Materials Chemistry and Physics, vol. 100, pp. 102-107, 2006.  R. Nakamura, T. Tanaka, and Y. Nakato, "Mechanism for Visible Light Responses in Anodic Photocurrents at N-Doped TiO2 Film Electrodes," Journal of Physical Chemistry B, vol. 108, pp. 10617-10620, 2004.  S. M. Prokes, J. L. Goel, X. Chen, B. Clemmons, and W. E. Carlos, "Defect-Related Optical Behavior in Surface Modified TiO2 Nanostructures," Advanced Functional Materials, vol. 15, pp. 161-167, 2005.  S. Mozia, M. Tomaszewska, B. Koso wska, B. Grzmil, A. W. Morawski, and K. Kalucki, "Decomposition of nonion ic surfactant on a nitrogen-doped photocatalyst under visibl e-light irradiation," Applied Catalysis B: Environmental, vol. 55, pp. 195-200, 2005.  R. Silveyra, L. De La Torre Saenz, W. A. Flores, V. C. Martinez, and A. A. Elguezabal, "Doping of TiO2 with nitrogen to modify the interval of photocatalytic activation towards visible radiation," Selected Contributions of the XIX Ibero American Catalysis Symposium Selected Contributions of the XIX Ibero Americ an Catalysis Symposium, vol. 107-108, pp. 602605, 2005.  J. Yuan, M. Chen, J. Shi, and W. Shangguan, "Preparations and photocatalytic hydrogen ev olution of N-doped TiO2 from urea and titanium tetrachloride," International Journal of Hydrogen Energy, vol. 31, pp. 13261331, 2006.
146  A. Orlov, M. S. Tikhov, and R. M. Lambert, "Application of surface science techniques in the study of environm ental photocatalysis: nitrogen-doped TiO2," Conversion photochimique et st ockage de l'energie solaire 2e partie, vol. 9, pp. 794-799, 2006.  P. Xu, L. Mi, and P. N. Wang, "Improved optical response for N-doped anatase TiO2 films prepared by pulsed laser deposition in N2/NH3/O2 mixture," Journal of Crystal Growth, vol. 289, pp. 433-439, 2006.  D. G. Huang, S. J. Liao, J. M. Li u, Z. Dang, and L. Petrik, "Preparation of visible-light responsive N-F-codoped TiO2 photocatalyst by a sol-gelsolvothermal method," Journal of Photochemistry and Photobiology A: Chemistry, vol. 184, pp. 282-288, 2006.  O. Diwald, T. L. Thompson, E. G. Gora lski, S. D. Walck, and J. T. Yates, "The Effect of Nitrogen Ion Implanta tion on the Photoactivity of TiO2 Rutile Single Crystals," Journal of Physical Chemistry B, vol. 108, pp. 52-57, 2004.  T. Lindgren, J. Lu, A. Hoel, C. G. Granqvist, G. R. Torres, and S. E. Lindquist, "Photoelectrochemical study of sputtered nitrogen-doped titanium dioxide thin films in aqueous electrolyte," International Solar Energy Society World Congress 2003, vol. 84, pp. 145-157, 2004.  A. Glaser, S. Surnev, F. P. Netz er, N. Fateh, G. A. Fontalvo, and C. Mitterer, "Oxidation of vanadium nitr ide and titanium nitride coatings," Surface Science, vol. 601, pp. 1153-1159, 2007.  B. Kosowska, S. Mozi a, A. W. Morawski, B. Grzmil, M. Janus, and K. Kalucki, "The preparation of TiO2-nitrogen doped by calcination of TiO2*H2O under ammonia atmosphere for vi sible light photocatalysis," Solar Energy Materials and Solar Cells, vol. 88, pp. 269-280, 2005.  Y. Sakata, T. Yamamoto, T. Ok azaki, H. Imamura, and S. Tsuchiya, "Generation of Visible Light Respons e on the Photocatalyst of a Copper Ion Containing TiO2," Chemistry Letters, vol. 27, pp. 1253-54, 1998.  M. Iwasaki, M. Hara, H. Kawada, H. Tada, and S. Ito, "Cobalt Ion-Doped TiO2 Photocatalyst Response to Visible Light," Journal of Colloid and Interface Science, vol. 224, pp. 202-204, 2000.
147  B. O'Regan and M. Grat zel, "A low-cost, high-efficiency solar cell based on dye-sensitized colloidal TiO2 films," Nature, vol. 353, pp. 737-740, 1991.  Y. Suda, H. Kawasaki, T. Ueda, and T. Ohshima, "Preparation of high quality nitrogen doped TiO2 thin film as a photocatalyst using a pulsed laser deposition method," Thin Solid Films, vol. 453-454, pp. 162-166, 2004.  M. Anpo and M. Takeuchi, "The desi gn and development of highly reactive titanium oxide photocatalysts operati ng under visible light irradiation," Journal of Catalysis, vol. 216, pp. 505-516, 2003.  R. Marchand, F. Tessier, A. Le S auze, and N. Diot, "Typical features of nitrogen in nitride-type compounds," International Journal of Inorganic Materials, vol. 3, pp. 1143-1146, 2001.  G. Liu, F. Li, Z. C hen, G. Q. Lu, and H. M. Cheng, "The role of NH3 atmosphere in prepar ing nitrogen-doped TiO2 by mechanochemical reaction," Journal of Solid State Chemistry, vol. 179, pp. 331-335, 2006.  S. Z. Chen, P. Y. Zhang, D. M. Zhuang, and W. P. Zhu, "Investigation of nitrogen doped TiO2 photocatalytic films prep ared by reactive magnetron sputtering," Catalysis Communications, vol. 5, pp. 677-680, 2004.  J. McMurray and R. Fay, Chemistry, Fourth Edition Fourth ed. Upper Saddle River, NJ: Prentice Hall, 2004.  K. Kobayakawa, Y. Murakami, and Y. Sato, "Visible-light active N-doped TiO2 prepared by heating of titani um hydroxide and urea," Journal of Photochemistry and Photobiology A: Chemistry, vol. 170, pp. 177-179, 2005.