<%BANNER%>

Record for a UF thesis. Title & abstract won't display until thesis is accessible after 2012-12-31.

Permanent Link: http://ufdc.ufl.edu/UFE0042370/00001

Material Information

Title: Record for a UF thesis. Title & abstract won't display until thesis is accessible after 2012-12-31.
Physical Description: Book
Language: english
Creator: Lee, Yang-Yao
Publisher: University of Florida
Place of Publication: Gainesville, Fla.
Publication Date: 2010

Subjects

Subjects / Keywords: Materials Science and Engineering -- Dissertations, Academic -- UF
Genre: Materials Science and Engineering thesis, Ph.D.
bibliography   ( marcgt )
theses   ( marcgt )
government publication (state, provincial, terriorial, dependent)   ( marcgt )
born-digital   ( sobekcm )
Electronic Thesis or Dissertation

Notes

Statement of Responsibility: by Yang-Yao Lee.
Thesis: Thesis (Ph.D.)--University of Florida, 2010.
Local: Adviser: Powers, Kevin W.
Local: Co-adviser: El-Shall, Hassan E.
Electronic Access: INACCESSIBLE UNTIL 2012-12-31

Record Information

Source Institution: UFRGP
Rights Management: Applicable rights reserved.
Classification: lcc - LD1780 2010
System ID: UFE0042370:00001

Permanent Link: http://ufdc.ufl.edu/UFE0042370/00001

Material Information

Title: Record for a UF thesis. Title & abstract won't display until thesis is accessible after 2012-12-31.
Physical Description: Book
Language: english
Creator: Lee, Yang-Yao
Publisher: University of Florida
Place of Publication: Gainesville, Fla.
Publication Date: 2010

Subjects

Subjects / Keywords: Materials Science and Engineering -- Dissertations, Academic -- UF
Genre: Materials Science and Engineering thesis, Ph.D.
bibliography   ( marcgt )
theses   ( marcgt )
government publication (state, provincial, terriorial, dependent)   ( marcgt )
born-digital   ( sobekcm )
Electronic Thesis or Dissertation

Notes

Statement of Responsibility: by Yang-Yao Lee.
Thesis: Thesis (Ph.D.)--University of Florida, 2010.
Local: Adviser: Powers, Kevin W.
Local: Co-adviser: El-Shall, Hassan E.
Electronic Access: INACCESSIBLE UNTIL 2012-12-31

Record Information

Source Institution: UFRGP
Rights Management: Applicable rights reserved.
Classification: lcc - LD1780 2010
System ID: UFE0042370:00001


This item has the following downloads:


Full Text

PAGE 1

1 PHOTOCATALYTIC ACTIVITY AND DYE SENSITIZED SOLAR CELL PERFORMANCE OF HIGH ASPECT RATIO TITANIUM DIOXIDE NANOFLAKES By YANGYAO LEE A DISSERTATION PRESENTED TO THE GRADUATE SCHOOL OF THE UNIVERSITY OF FLORIDA IN PARTIAL FU LFILLMENT OF THE REQUIREMENTS FOR THE DEGREE OF DOCTOR OF PHILOSOPHY UNIVERSITY OF FLORIDA 2010

PAGE 2

2 2010 Yang Yao Lee

PAGE 3

3 To my wife, Hung Ju, who companie s me with endless support.

PAGE 4

4 ACKNOWLEDGMENTS I would first thank my advisor, Dr. Kevin Powers, for his patience, enthusiasm, and wisdom. He always guided me to observe and think my work scientifically. I also would like to thank my co chair, Dr. El Shall, for introducing the concept of statistics design of experiment to me. He helped me develo p the engineering aspect of scientific work. I would like to acknowledge my committee members, Dr. Rajiv Singh, Dr. Ronald Baney, and Dr. Myoseon Jang for their invaluable discussions and suggestions. Special thanks to Dr. Arun Ranade at Particle System, L LC. for his compassion and support through the project. I express my gratitude to my parents who endlessly support ed me throughout my Ph.D. and always encouraged me to continue my dreams. I am always grateful for having your support with me. I especially t hank my wife, Hung Ju, for lasting encouragement and support. She helped me in every as pect of life in United States. Thanks for her selfless sacrifice. I gratefully acknowledge my group members including Dr. Nate Stevens, Dr. Steve Tedeschi, Ming Chi Tim Yang, Gary Scheiffele, Gill Brubaker, Kerry Siebein, Jiaqing Zhou, Paul Carpinone, David W i ggins and all the staff of the Particle Engineering Research Center for their help in all aspects of my research. Finally, I acknowledge the Particle Engineering R esearch Center for financial support and friendly environment for carrying out research.

PAGE 5

5 TABLE OF CONTENTS page ACKNOWLEDGMENTS ...............................................................................................................4 LIST OF TABLES ...........................................................................................................................7 LIST OF FIGURES .........................................................................................................................9 ABSTRACT ...................................................................................................................................15 CHAPTER 1 INTRODUCTION ..................................................................................................................17 The Most Essential Component of Life: Water ......................................................................17 Traditional Wastewater Treatments ................................................................................17 Advanced Oxidation Processes (AOPs) ..........................................................................18 The Coming Global Energy Crisis .........................................................................................18 Finite Energy Source: Fossil Fuels ..................................................................................19 Alternative Clean Energy Demand ..................................................................................20 The Remedy for Energy Crisis: Nanotechnology? ..........................................................20 Gap Analysis and Statement of Problem ................................................................................22 2 BACKGROUND ....................................................................................................................24 P hotocatalysis on T itanium D ioxide S urface .........................................................................24 P hotocatalyst L oading .....................................................................................................27 L ight I ntensity .................................................................................................................28 pH ....................................................................................................................................28 T emperature .....................................................................................................................29 O xygen P ressure ..............................................................................................................29 C harge separation .....................................................................................................30 G eneration of active species .....................................................................................30 M aintenance of the stoichiometry of titania .............................................................30 E nhanced of P hotocatalys is ....................................................................................................31 S ize R eduction .................................................................................................................31 S hape C ontrol ..................................................................................................................32 Doping (Narrowing Band Gap) .......................................................................................34 3 HIGH ASPECT RATIO titanium dioxide FLAKES AS PHOTOCATALYSTS ..................41 S ynthesis of H igh A spect R atio of TiO2 F lakes .....................................................................41 Benchmark: Degussa P25 .....................................................................................................49 E xperiment A pparatus ............................................................................................................49 P hotocatalytic D egr adation of D ye M olecules under UVA L ight I llumination .....................50 Statistical Design of Experiments ...........................................................................................52

PAGE 6

6 Synthesized Flakes ..........................................................................................................53 Calcined Flakes ...............................................................................................................55 P25 ...................................................................................................................................56 4 TRANSITION METAL DOPING OF TITANIA PARTICIELS AS VISIBLE PHOTOCATALYSTS ............................................................................................................87 S ynthesis of V anadium D oped T itania F lakes ........................................................................87 C haracterization of Vanadium Doped TiO2 F lakes ................................................................87 XPS Spectra of Series Vanadium Doped Flakes .............................................................92 E xperiment A pparatus ............................................................................................................94 Statistical Design of Experiments ...........................................................................................97 5 TITANIA NANOFLAKES BASED DYE SENSITIZED SOLAR CELLS ........................125 Introduction ...........................................................................................................................125 Basic Principles and Components of DSSCs ................................................................125 The Energetic Aspect of DSSCs ....................................................................................126 Experiment ............................................................................................................................127 Photoelectrodes Characterization ..................................................................................128 Photovoltaic Measurements ...........................................................................................129 Assembling DSSCs using High Aspect Ratio Titania Particles ...........................................129 Characterization of Titania Photoelectrodes .................................................................130 Photovoltaic Performance of Titania Photoelectrodes ..................................................134 6 SUMMARY, CONCLUSIONS, AND FUTURE WORK ...................................................149 Summary ...............................................................................................................................149 Conclusions ...........................................................................................................................150 Future Work ..........................................................................................................................151 APPENDIX A ANALYSIS VARIANCE F OR UV PHOTOCATAL YSIS USING SYNTHESIZED FLAKES ...............................................................................................................................152 B ANALYSIS VARIANCE FOR UV PHOTOCATALYSIS USING CALCINED FLAKES ...............................................................................................................................154 C ANALYSIS VARIANCE FOR UV PHOTOCATALYSIS USING P25 NANOPARTICLES .............................................................................................................155 D ANALYSIS VARIANCE FOR VISIBLE PHOTOCATALYSIS USING VANADIUM DOPED FLAKES .................................................................................................................156 LIST OF REFERENCES .............................................................................................................157 BIOGRAPHICAL SKETCH .......................................................................................................167

PAGE 7

7 LIST OF TABLES Table page 21 Bulk propertie s of three TiO2 polymorphs [ 24] ................................................................37 22 Oxidation power of species [ 26] .......................................................................................37 31 Particle diameter statistics for synthesized and calcined flakes. ........................................57 32 Grain size calculation by the Scherer equation for both nanoflakes .................................57 33 Physisorption measurements of P25, synthesized and calcined titania nanoflakes ..........57 34 Optical and electrostatic properties of titaia samples. .......................................................57 35 Pseudofirst order rate constants of methylene blue under photocatalytic degradation. ...57 36 The 23 factorial design used to investigate the most important factors for UV light photocatalysis. ....................................................................................................................58 42 Grain size calculation by the Scherer equation for series calcined vanadium doped nanoflakes .......................................................................................................................100 43 Physisorption measurements of P25, and series calcined vanadium doped titania nanoflakes .......................................................................................................................100 44 V/Ti molar ratio and band gap energy of series vanadium doped samples. ....................100 45 Pseudofirst order rate constants of methylene blue under visible photocatalytic degradation. ......................................................................................................................101 46 The 23 factorial design used to investigate the most important fa ctors for visible light photocatalysis. ..................................................................................................................101 51 Particle diameter statistics for P25, synthesized and calcined flakes after ultrasonication. .................................................................................................................136 52 Dye loading and physisorption measurements of P25, synthesized and calcined titania nanoflakes ............................................................................................................136 53 Photovoltaic properties of the dye sensitized solar cells assembled by using anodes made from P25 nanoparticles and calcined titania flakes of different thickness. ............136 A 1 Analysis of variance table ................................................................................................152 A 2 Statistical results ..............................................................................................................153 B 1 Analysis of variance table ................................................................................................154

PAGE 8

8 B 2 Statistical results ..............................................................................................................154 C 1 Analysis of variance table ................................................................................................155 D 1 Analysis of variance table ................................................................................................156 D 2 Statistical res ults ..............................................................................................................156

PAGE 9

9 LIST OF FIGURES Figure page 21 Schematic representation of the photoelectrochemical mechanism taking place inside titania: ................................................................................................................................38 22 Schematic photoexcitation in a solid followed by deexcitation events .............................38 23 Second ary reactions with active oxygen species in the photoelectrochemical mechanism .........................................................................................................................39 24 The schematic diagram of conduction and valence bands positions of several semiconductors. The left hand scale is the internal energies to the vacuum level and the right hand scale represented the comparison to normal hydrogen electrode (NHE) ...39 25 Schematic energy level diagram of metal doped TiO2 ......................................................40 31 Photographs of titania slurries. ..........................................................................................59 32 O ptical micrographs of titania samples. .............................................................................59 33 SEM images of titania samples. ........................................................................................60 34 SEM images of edge view of titania samples. ...................................................................60 35 P article size distribution for synthesized and calcined flakes. ...........................................61 36 XRD patterns of titania samples. .......................................................................................61 37 HR TEM images o f titania samples (the SAD pattern as inset) .......................................62 38 XPS analysis o f the titania samples. ..................................................................................63 39 UVVisible spectra of titania samples. A) Diffuse reflectance spectra. B ) the dependence of ( h )2 on the photon energy for synthesized and calcined titania flakes .................................................................................................................................64 311 True absorption spectra of titania samples using the modified integration sphere met hod. (note: reflectivity in an integrating sphere is provided by a coating of pigment grade TiO2 or ZnO. Thus the absorbance above represents only the increased absorbance over the control.) .............................................................................65 312 Zeta potential versus solution pH for three titania samples. ..............................................66 313 Photocatalytic decomposition of 50 M methylene blue with synthesized titania flakes under UVA illumination. .........................................................................................66

PAGE 10

10 314 Photocatalytic decomposition of methylene blue by using synthesized and calcined flakes A ) without bubbling treatment B ) with bubbling treatment .................................67 315 Linear transforms ln(C/C0) vs. time for methylene blue decomposition for different photocatalysts A ) without bubbling treatment B ) with bubbling treatment. Note the difference in scales of the normalized concentrations. ......................................................68 316 Experimental design decomposition% results of synthesized flakes photocatalysis. ........69 317 The 3D response surface plot of dye decomposition% of synthesized flakes as a function of the factor A and B under different light intensity. A) 2. B) 4. C) 6. ...............70 318 Model graphs of synthesized flakes as the function of A and B at constant light intensity (C = 2). A ) Contour plot of decomposition% B) Interaction plot of response data against factor A for both levels of factor B. ...............................................................71 319 Model graphs of synthesized flakes as the function of A and B at constant light intensity (C = 6). A ) Contour plot of decomposition%. B) Interaction plot of response data against factor A for both levels of factor B. ................................................72 320 The 3D response surface plot of dye decomposition% of synthesized flakes as a function of the factor B and C under different aeration. A) 0. B) 5. C) 10 ft3/hr. .............73 321 Model graphs of s ynthesized flakes as the function of B and C at constant flow rate (A = 0). A ) Contour plot of decomposition%. B) Interaction plot of response data against factor B for both levels of factor C. .......................................................................74 322 Model graphs of synthesized flakes as the function of B and C at constant flow rate (A = 10). A ) Contour plot of decomposition% of synthesized flakes B) Interaction plot of response data against factor B for both levels of factor C. .....................................75 323 Experimental design decomposition% results of calcined flakes photocatalysis. .............76 324 The 3D response surface plot of dye decomposit ion% of calcined flakes as a function of the factor A and B under different light intensity. A) 2. B) 4. C) 6. .............................77 325 Model graphs of calcined flakes as the function of A and B at constant light intensity (C = 2). A ) Contour plot of decomposition%. B) Interaction plot of response data against factor A for both levels of factor B. .......................................................................78 326 Model graphs of calcined flakes as the func tion of A and B at constant light intensity (C = 6). A ) Contour plot of decomposition%. B) Interaction plot of response data against factor A for both levels of factor B. .......................................................................79 327 The 3D response surface plot of dye decomposition% of calcined flakes as a function of the factor B and C at different aeration rate. A) 0. B) 5. C) 10 ft3/hr. ...........................80

PAGE 11

11 328 Model graphs of calcined flakes as th e function of B and C at constant flow rate (A = 0). A ) Contour plot of decomposition% of calcined flakes. B) Interaction plot of response data against factor B for both levels of factor C. ................................................81 329 Model graphs of calcined flakes as the function of B and C with constant flow rate (A = 10). A ) Contour plot of decomposition%. B) Interaction plot of response data against factor B for both levels of factor C. .......................................................................82 330 Experimental design decomposition% results of P25 photocatalysis. ...............................83 331 The 3D response surface plot of dye decomposition% of P25 as a function of the cataly st concentration and Flow rate under different light intensity. A) 2. B) 4. C) 6. .....84 332 The 3D response surface plot of dye decomposition% of P25 as a function of the catalyst concentration and light intensity at different aeration rate. A) 0. B) 5. C) 10 ft3/hr. ..................................................................................................................................85 333 Interaction plot of response data against factor B for both levels of factor C at different aeration rate for P25. A) 0. B) 5. C) 10 ft3/hr. .....................................................86 41 Photographs of synthesized vanadium doped titanium dioxide flakes. A) 1. B) 5. C) 10 atomic% vanadium (note the pearlescence of the samples). .......................................102 42 Photographs of calcined vanadium doped titanium dioxide flakes. A) 1. B) 5. C) 10 atomic% vanadium (note the pearlescence of the samples). ............................................102 43 Optical micrographs of vanadium doped titanium dioxide flakes. A) synthesized. B) calcined 5 atomic% vanadium doped flakes. ...................................................................103 44 Optical micrographs of calcined vanadium doped flakes. A) 1. B) 10 atomic% vanadium. .........................................................................................................................103 45 SEM images of synthesized 5 atmoic% vanadium doped titania nanoflakes. A) surface morphology. B) thickness. C) EDX. D) mapping. ..............................................104 46 SEM images of calcined 5 atmoic% vanadium doped titania nanoflakes. A) surface morphology. B) thickness. C) EDX. D) mapping. ...........................................................105 47 SE M images of calcined 1 atomic% vanadium doped titania nanoflakes. A) surface morphology. B) thickness. C) EDX. D) mapping. ...........................................................106 48 SEM images of calcined 10 atomic% vanadium doped titania nanoflakes. A) surface morphology. B) thickness. C) EDX. D) mapping. ...........................................................107 49 Volume based particle size distribution for synthesized and calcined vanadium doped flakes. A) 1. B) 5. C) 10 atomic% vanadium. ..................................................................108

PAGE 12

12 410 XRD patterns of vanadium doped flakes. A) 5 atomic% vanadium. B) Anatase (101) peak for series calcined vanadium doped flakes (at slow scanning mode: 0.01/step). ..109 411 Zeta potential of the vanadium doped flakes as the function of pH. A) 0, 1, 5, 10 atomic% vanadium B) The IEP versus the amount of vanadium in the series vanadium doped flakes. ...................................................................................................110 412 Diffuse reflectance spectra of vanadium doped titania flakes A) 5 atomic% vanadium. B ) the absorbance spectra for synthesized and calcined vanadium doped titania flakes ....................................................................................................................111 413 The true absorption spectra of vanadium doped flakes. A) series synthesized vanadium doped flakes. B) series calcined vanadium doped flakes. C) photoresponse of series vanadium doped flakes in the visible range. .....................................................112 414 XPS spectra of vanadium doped flakes. A) O 1s and V 2p3/2 peaks for series vanadium doped flakes (b) V 2p3/2 peak of 10 atomic% vanadium doped flakes. ..........113 415 XPS spectra of Ti 2p3/2 and Ti 2p1/2 peaks for vanadium doped flakes. A) series vanadium doped flakes. B) Ti 2p3/2 peak of 10 atomic% vanadium doped flakes. .........114 416 XPS spectra of O 1s peak fitting. .....................................................................................115 417 Visible p hotocatalytic d ecomposition of methylene blue by using vanadium doped titania flakes with three vanadium doping levels under bubbling condition. ..................116 418 Linear transforms ln(C/C0) vs. time for methylene blue decomposition for vanadium doped titania flakes with three vanadium doping levels under bubbling. .......116 419 Linear transforms ln(C/C0) vs. time for methylene blue decomposition for vanadium doped titania flakes with three vanadium doping levels under bubbling condition. ( Note the difference in scales of the normalized concentrations ). ..................117 420 Comparison of the ratio of V4+/V5+ and photocatalytic reaction rate for series vanadium doped flakes ...................................................................................................117 421 Experimental design visible light decomposition% results of vanadium doped titania flakes photocatalysis with three different doping levels. .................................................118 422 The 3D response surface plot of dye decomposition% of series calcined vanadium doped flakes as the factor A and B under different aeration rate ....................................119 423 Model graphs of series vanadium doped flakes as the function of B and C at constant aeration rate (A = 0). A ) Contour plot of decomposition% of series vanadium doped flak es. B) Interaction plot of response data against factor B for both levels of factor C. ......................................................................................................................................120

PAGE 13

13 424 Model graphs of series vanadium doped flakes as the function of B and C at constant aeration rat e (A =10). A ) Contour plot of decomposition% of series vanadium doped flakes. B) Interaction plot of response data against factor B for both levels of factor C. ......................................................................................................................................121 425 The 3D response surfac e plot of dye decomposition% of series calcined vanadium doped flakes as the factor B and C ..................................................................................122 426 Model graphs of calcined 1 atomic% vanadium doped flakes as the function of A and B. A ) Contour plot of decomposition%. B) Interaction plot of response data against factor A for both levels of factor B. .................................................................................123 427 Model graphs of calcined 5 atomic% vanadium doped flakes as the function of A and B. A ) Contour plot of decomposition%. B) Interaction plot of response data against factor A for both levels of factor B. .................................................................................124 51 Schematic structure of DSSCs [163]. ..............................................................................137 52 Energy levels in the typical DSSCs [164]. .......................................................................137 53 Typical I V curves of the solar cells. (The fill factor, FF, could be calculated by the ratio of the area A to the area B). .....................................................................................138 54 Schematic cross sectional view of electron transport through titania layer. A) Degussa P25 nanoparticles. B) calcined titania nanoflakes (two dimensional nanostructures). ................................................................................................................139 55 Optical micrographs of sintered titania photoelectordes made from Degussa P25 nanoparticles. A) 5. B) 50 magnifications. ......................................................................140 56 Optical micrographs of sintered titania photoelectordes made from calcined titania flakes. A) 5. B) 50 magnifications. ..................................................................................140 57 SEM micrographs of sintered photoelectrodes made from P25 nanoparticles under different magnifications. ..................................................................................................141 58 Cross section of sintered photoelectrodes made from P25 nanoparticles unde r different magnifications. ..................................................................................................141 59 SEM micrographs of sintered photoelectrodes made from calcined titania flakes under different magnifications. ........................................................................................142 510 Cross section of sintered photoelectrodes made from calcined titania flakes under different magnifications. ..................................................................................................142 511 SEM images of platinum counterelectrode. A) surf ace morphology. B) thickness. C) EDX. D) mapping. ...........................................................................................................143

PAGE 14

14 512 Particle size distribution of titania slurries for P25, synthesized and calcined titania flakes after ultrasonication at 100W for 15 minutes. .......................................................144 513 Absorption spectrum of 0.3 mM Ru complex dye solution (also as known N 719 dye). .................................................................................................................................144 514 Diffuse reflectance spectra of the titania films prepared from Degussa P25 nanoparticles, syntheszied and calcined titania flakes of similar thickness. ....................145 515 I V curves of dye snesitized solar cells with diffe rent film thickenesses. A) calcined titania flakes. B) P25. .......................................................................................................146 517 Comparsion of the Jsc and Voc of P25 and calcined flakes cells as a function of film thickness. ..........................................................................................................................148

PAGE 15

15 Abstract of Dissertation Presented to the Graduate School of the University of Florida in Partial Fulfillment of the Requirements for the Degree of Doctor of Philosophy PHOTOCATALYTIC ACTIVITY AND DYE SENSITIZED SOLAR CELL PERFO RMANCE OF HIGH ASPECT RATIO TITANIUM DIOXIDE NANOFLAKES By Yang Yao Lee December 2010 Chair: Kevin William Powers Cochair: Hassan El Shall Major: Materials Science and Engineering Protection of clean water sources and exploring clean energy resources ar e two of t he most important issue s in the 21st century. Although advanced oxidation processes with titanium dioxide (TiO2) photocatalytic nanoparticles have been a somewhat effective alternative in this regard, the low quantum efficiency of the TiO2 semico nductor and the difficulty of separation from water prevent practical applications. Therefore, there is a need to develop a novel material with higher photocatalytic activity to develop viable commercial products. The first goal of this research project is to develop a cost efficient, and easily separated photocatalyst for water purification. This goal was achieved by synthesizing TiO2 nanoflakes with enhanced quantum yield. The synthesis of nanoflakes has been chosen for increased quantum efficiency and their ease of separation. The results showed that titania flakes have higher photocatalytic activity than the reference commercial product Degussa P25. Ultimately, more than 99% of methylene blue molecules ( 150 ml of 50M) in simulated wastewater w ere decom posed within 2 hours by adding 100 ppm of TiO2 nanoflakes under UVA light illumination F ro m the energy aspect, a more energy efficient way for purify ing the polluted water need s to be developed.

PAGE 16

16 Transition metal doping has been shown to extend the a bsorp tion edge of TiO2 from the UV in to visible region (~ 2.8 eV) I n these studies, it is proposed to investigate the factors affecting their photocatalytic activities under visible light. Ultimately it was shown that vanadium doped flakes at 1 atomic % exhib ited the highest visible photocatalytic activity. Due to the depletion of fossil fuel in the future, many researchers focused on finding renewable energy sources. O f the many different types of solar cells, the dyesensitized solar cell (DSSC) is a promis ing device with sufficiently low manufacturing cost to compete in the commercial market place. In this study titania nanoflakes were used to prepare the semiconductor layer of DSSC. The highest efficiency of nanoflake based cell was 7.4% and 5 fold improvement was achieved by replacing P25 nanoparticles with calcined titania flakes in the photoelectrodes on the basis of film thickness.

PAGE 17

17 CHAPTER 1 INTRODUCTION The Most Essential Component of Life: Water Water is one of the most essential substances for lif e. The total amount of water in the world is about 1.4 billion km3 and only 2.5% or 35 million km3 is consider ed freshwater suitable for human usage [ 1] The continued explosive growth in the human population has increased the demand for this limited supply of freshwater. According to the Wo rld Health Organization (WHO) [ 2] more than 40% of the world population suffers from the chronic shortage or lack of w ater due to political, economical and c limatological reasons. H ealth and hygienic problems related to water affects more than 25% of the world s population. Despite the plans carried out by United Nations (UN) in recent years, 1.1 billion people still do not have access to adequate water suppl ies and sanitation. This condition is especially concentrated in the underdeveloped countries of Africa, Asia, and Latin America [ 2] Moreover, the increasing domestic and industrial activities generate high amounts of polluted water which directly flow into natural channels of water bodies. Therefore, protecting the cleanliness of our water sources is one of the most important environmental issues in the 21st century. Traditional Wastewater T reat ments There are many different pollutants involved with the contamination of environmental water such as detergents, pesticides, herbicides, s olvent and chemical wastes produced by industries The conventional wastewater treatments used can be classified as physical, biological, and chemical. Physical water treatment methods include UV light irradiation and filtration. Biological treatment uses bacteria to treat leachates. Chemical treatment uses chlorine to eliminate microorganisms. Although the most popular methods for wastewater treatment are biological methods due to low costs, they have been largely unsuccessful in the removal of many

PAGE 18

18 contam inants such as pesticides since these tend to be toxic to the microorganism s used. Chemical treatments are limited due to high costs and slow reaction times. Advanced Oxidation Processes (AOPs) Advanced oxidation processes (AOPs) present a promising way of decomposing biologically and chemically stable molecules in waste. Reviews on this topic ha ve been published recently [ 24] Although there are many different systems, the common mechanism of AOPs is the oxidation and m ineralizing of organic species using hydroxyl radials ( OH ) and other reactive oxygen species Although the strong ability of AOPs for eliminating hazard compounds in the wastewater is recognized, the operating costs for total oxidation of the organic molecules are still more expensive than the conv entional biological treatments. However, in many instances, they ar e suitable for pre treatment or enhancement of the biodegradation. The use of solar energ y can improve not only the economics but also contribute in process sustainability by saving electr icity. In addition, the AOPs can be used to remediate almost any kind of wastewater containing harmful organic compounds. Among these different systems of AOPs, titanium dioxide (TiO2) is a well studied, chemical stable, environmentally friendly, reusable, and most importantly inexpensive material. Therefore, we propose to use TiO2 as the water treatment reagent in this study. The Coming Global Energy Crisis The booming economic growth in the World over the past few decades was mainly supported by the affordable energy cost s The majori ty of energy sources, about 8085%, come from fossil fuel s which are a product refined from the ancient biomass under the earths surface for more than 200 million years. And the electricity, one of the most vital parts of modern life, is primarily dependent on the combustion of coal, natural gas, and oil. Worldwide population has increased at least four times in the 20 th century and consequent energy consum ption increment is

PAGE 19

19 up by 16 fold. Currently, the total energy demand per year is about 13 terawatts (TW = 1012 watts ) for the world s 6.5 billion people [ 5 ] In another 40 years, it is expected to increase to 23 TW for the increasing population to sustain the current lifestyle [ 5] Except for the tremendous energy demand in the future, the energy production based on fossil fuels also has the significant impact on the environment. When we produce the energy from burning fossil fuels, pollutants and greenhouse gases were inevitably generated. The current global warming phenomenon has been attributed by many to the emis sion of greenhouse gases in the atmosphere [ 6]. The historic high mean surface and ocean temperatures were recorded and stronger and stronger tropical storm s and hurricanes have occurred in recent years. The United Nations Framework Convention on Climate Change calls for stabilization of greenhouse gas concentrations in the atmosphere at a level that would prevent dangerous anthropogenic interference with th e climate system [ 79]. In one possible proposed scenario aggressive policy intervention used to limit greenhouse gas emissions has the potential to ass 10 TW of c arbon emission free power by the year 2050, equivalent to the power provided by all today's energy sources combined [ 6] Finite Energy Source: Fossil Fuels Hubbert proposed a model to predict the production of oil and found that the current exponential growth in the last century is a transient phenomenon [ 10,11]. It will not continuously increase in the future and this predictive model fits the peak fossil fuel production in 1970. The total storage of global oil is 2000 billion barrels and 1800 billion barrels have been found [ 12]. Moreover, we are u sing 4 billion barrels of fossil fuels for discovering every new billion barrel of oil and total of 875 billion barrel have been consumed. Therefore, we will be forced to diversify our energy supply especially alternative clean energy resources in order to keep not only global economic growth but also healthy environment.

PAGE 20

20 Alternative Clean Energy Demand There are two major types of energies proposed to fully or partially replace the fossil fuel consuming in the world: nuclear and renewable energies. If we u se nuclear power as the main source of energy, 1 Gigawatts of nuclear fission plant will be required to build per day for the following 50 years usage of energy [9 ] Furthermore, the undesired high level radioactive waste in the nuclear reactor remains the most significant pollution problem and had to handle and store with extreme care. According to the United States Environment Protection Agency standard, the radioactive decay life of these spent fuels is up to 10,000 years. Renewable energies mainly come from natural resources such as wind, ocean currents, gravity (hydroelectric), geothermal heat, and sun. Therefore, the renewable energies are the most environmental friendly energy resources in the world due to the nature of the production. They are natura lly replenished and estimated to be the amount of electricity as the progressively increasing sequences: hydroelectricity (0.5 TW), ocean currents (2 TW), extractable wind power (24 TW), geothermal over the whole surface of earth (12 TW), and solar energy (120000 TW) [9] Among these energy options, solar energy is the most abundant resource for our energy demand. Total amount of energy consumed annually by the world is equivalent to one hour illumination of sun light irradiation on earth. In spite of the tremendous amount of energy supplied by the sun, only a very small portion of total energy produced (~ 0.01%) comes from the solar energy [ 9] T he major hurdles of solar energy application are the cost issues and not sufficient efficiency. Therefore, we ne ed more advanced technology with substantially reduced manufacturing cost and improved energy conversion efficiency to fully utilize the solar energy. The Remedy for Energy Crisis: Nanotechnology? During the last decade, nanomaterials have been attracted w ide attention owing to unique chemical and physical properties. Therefore, nanotechnology has been involved in a vast range

PAGE 21

21 of applications such as medicine, electronics, biomaterials, and energy production. Novel nanomaterials provide advanced strategies for designing next generation energy conversion devices and serve as the building blocks to harvest solar power. The best known energy device for directly utilizing solar energy is the solar cell which is able to convert sun light directly into electricity by the photovatalic effect. There are three major classifications for solar cell based on the generation: silicon bulk heterojunction cells (first generation), the CuInGaSe2 (CIGS) cells (second generation), and organic and inorganic solar cells (third ge neration). Single crystalline silicon cells currently provide the highest efficiency (up to 15%) but suffer from high manufacturing cost [7] With such efficiency, the sale price of grid generated electricity is about $0.25 to $0.30 per kilowatt hour (kWh) which is much higher than the current price of utility scale electricity production ($0.03 to $0.05 kWh) [7] The second generation devices made with CIGS thin film bring down the cost significantly but the efficiency is not high enough for practical appl ication. Among the third generation of solar cells, dye sensitized solar cells (DSSCs) provide an efficient method to mimic the natural photosynthesis process. Although the energy conversion efficiency of DSSCs is not as high as it of conventional silicon based solar cell, the price/performance ratio should be high enough to compete with fossil based electr i city The estimated manufacturing cost of the DSSCs is 1/3 1/5 of the conventional solar cells. One of the key components in the DSSCs is an electron tr ansport layer consisting of the semiconductor materials such as TiO2, ZnO, and SnO2. Titanium dioxide is a well known low toxicity, chemically stable, and cost effective material used in many industrial applications. Recently researchers have focused on sy nthesis of TiO2 nanostructures with desired shapes such as nanotubes, nanorods, and nanowires. Using these low dimensional nanostructures as the building blocks, new electron transportation pathway could be established and opened a new strategy of

PAGE 22

22 improving energy conversion efficiency of the DSSCs. In this study, we will use our synthesized titania nanoflakes as the starting material and investigate the photon energy conversion efficiency of titania nanoflakebased DSSCs. Gap Analysis and Statement of Prob lem Research focused on the remediation of wastewater has been very limited and most use advanced oxidation processes (AOPs). AOPs are defined as processes that generate hydroxyl radicals with sufficient concentration to oxidize most organic chemicals pres ent in the effluent water. Although AOPs with TiO2 photocatalysts have been shown to be an effective alternative in this regard, the two major obstacles preventing the use of TiO2 semiconductors in practical applications is its low quantum efficiency ( 1% ) and the requirement of UV light irradiation. There is extensive literature consider ing the effect of size on photocatalytic activity especially the use of ultrafine particles. Only a few researchers have studied shape control of titania photocatalysts. Although the photocatalytic efficiency is increased due to the quantization effect of ultrafine particles, absorption in the visible region for nanoparticles tends to decrease. Hence, the first objective of this research is to study the effect of changing shape on photocatalytic behavior. TiO2 nanoflakes with micron size width but nanoscale thickness create a short diffusion path for charge carriers and reduce surface recombination. In addition the effect of transition metal doping level on the visible ligh t absorption characteristics and subsequent influence on the photocatalysis needs to be investigated systematically. The final goal of this research is to eliminate > 95% of 50 ppm of organic pollutant in 150 ml wastewater by adding 0.1 g of transition metal doped titania nanoflakes under visible light illumination To achieve this goal, experimental and fundamental studies are proposed. In these studies, t he effect of changing shape on photocatalytic efficiency under UV and visible irradiation are

PAGE 23

23 compared to conventional photocatalysts. The factors include: light intensity, catalyst loading, aeration rate, and vanadium dopant concentration. The effect of these factors on the final photodegradation behavior has been systematically and experimentally investi gated. To do so, undoped and vanadium doped titanium dioxide flakes with high aspect ratios were synthesized by a modified surface hydrolysis method. Then the flakes are directly injected into the stimulated wastewater under UV and visible light. This will facilitate an understanding of the significance and magnitude of each factor on the UV and visible light photocatalysis behavior. DSSC is a promising solar cell technology with potentially high enough efficiency and cost effective fabrication. However, t he overall ef ficiency has been not maximiz ed and the current technology is not yet cost competitive with the current electrical power generation. The purpose of this study is to investigat e the effect of using low dimen s ional nanomaterials as the alternative materials in the semiconductor layers of the DSSCs. A titania nanoflake based DSSC with higher efficiency is compared to a P25 based cell. To achieve this goal, calcined titania nanoflakes were used to prepare the titania slurry to deposit on the transparent glass substrate and the photon energy conversion efficiency of the nanoflake based DSSC was measured under the simulated sun light illumination (AM 1.5). The effect of thickness of semiconductor layer on the photovoltaic performance was also systemat ically and experimentally investigated.

PAGE 24

24 CHAPTER 2 BACKGROUND P hotocatalysis on T itanium D ioxide S urface In 1972, Fujishima and Honda discovered the electrochemical photolysis of water on TiO2 surface [ 13] This significant event opened the era of heterogeneous photocatalysis on the environmental application. Titanium dioxide has since attract ed much attention because of it s unique properties such as high photocatalytic activity, moderate band gap and suitable band position, nontoxicity, availabi lity, and low cost [ 14,15] Furthermore, TiO2 powders also have versatile optical properties in tinting strength, hiding power, and ultraviolet (UV) light shielding. These properties have many industrial applications such as pigments, cosmetics, and so on [ 14,15] Figure 21 schematically represents the steps of photocatalysis. The main reactions [ 13,1620] and the required time measured by laser flash photolysis [ 21,22] are listed as follows: Charge carrier generation s e h h TiOCB VB 15 210 Charge carrier trapping s OH Ti OH Ti hIV IV VB 910 10 s OH Ti OH Ti eIII IV CB 1210 100 s Ti Ti eIII IV CB 910 10 Charge carrier recombination s OH Ti OH Ti eIV IV CB 910 100 s OH Ti OH Ti hIV III VB 910 10 Oxidation or reduction s d OH Ti d OH TiIV IV 9 010 100 Re Re s O OH Ti O ex IV x TR 310

PAGE 25

25 w here CBe is a conduction band electron, VBh is a valence band hole, TiOH represents the hydrated surface of TiO2, } { OH TiIVis the surfacetrapped hole, OH TiIII is the surfacetrapped electron, Red is an electron donor (i.e., reductant), Ox is an electron acceptor (i.e., oxidant). According to the above equations, the overall photocatalytic efficiency depends on the competition between the interfacial chargetransfer and the electron hole recombination. Therefore, the photocatalytic reaction will be accelerated by increasing either the lifetime of electron hole pairs (retard ing recombination) or the rate of interfacial charge transfer. Fr om the results of laser flash photolysis, t he dominant reaction is the recombination of the e lectron and hole (1 ns) followed by the reduction reaction (10 ns) and oxidation (1 ms) [ 21,22] Figure 2 2 illustrates some of the recombination pathw ays for electrons and holes [ 19] Depending on the location of the sites, there are two types of possible recombinations: surface recombination (path ( A ) ) and volume recombination (path ( B ) ). Both recombinations are detrimental to the efficiency of photocatalys ts as and ener gy is released as heat. Several approaches have been tried to suppress the recombination of electron hole pairs including addition of metals, dopants, or combinati ons with other semiconductors [ 19] More detailed informatio n will be discussed in the following section. The efficiency of photocatalytic process is measured as the quantum yield ( ) which is defined as the rate of reaction induced by photon absorption / flux of absorbed photon. However, it is very difficult to me asure the precise quantity of absorbed light in the heterogeneous system since scattering of light usually occurs at the semiconductor surface. Therefore, most researchers used an assumption that all incident light is absorbed and calculat e a so called ap parent quantum yield. T he apparent quantum yield will be calculated and compared in this study to elucidate the influence of different samples on a photocatalysis

PAGE 26

26 process. There are eight TiO2 polymorphs in principle, which are anatase, rutile, brookite columbite, baddeleyite, cotunnite, py rite, and fluorite structures [ 23] In nature, titania exists three main phases which are rutile, anatase, and brookite, in order of abundance The basic properties of these three polymorphs are listed in Table 2 1 [ 24] Only anatase and rutile showed significant photocatalysis under ultraviol et light irradiation; the former generally has higher photocatalytic activity in general Only a few investigators have investigated the photocatalysis with the brookite phase, which may be attributed to the difficulty of highly pure brookite synthesis [ 25] Since one of the important applications for photocatalysis is water purification, several specific mechanisms fo r interfacial charge transfer in water are illustrated and listed in the following reactions. Specific exampl es of i nterfacial charge transfer OH OH Ti OH OH Ti H OH OH Ti O H OH TiIV IV IV IV 2 OH OH OH Ti O H e O OH Ti O eIV TR IV TR 2 2 2 2 From the above reactions, highly reactive species such as hydroxy l radicals ( OH ) and superoxide ions ( 2O ) could either oxidize or reduce a wide variety of organic compounds in waste water. Elizardo compared many oxidizing agents with chlorine and found that hydroxyl radicals hav e the highes t oxidation power (Table 2 2) [ 26] Many researchers assumed that the primary oxidiz ing species in the photocatalytic reaction are hydroxyl radicals especially in water [ 17,26] One of the reasons is the abundant source of hydroxide ions exists in the environment when photocatalysis proceeding in water. However, there is still an argument that the oxidation process occurs at the titania surface formed directly by direct react ion with holes instead of

PAGE 27

27 hydroxyl radicals [ 2729] The quantum yield c an be directly related to the generation of hydroxyl radicals and can be rewritten as: = generation rate of hydroxyl radicals / flux of absorbed photon. Q uantum yields under typical reactor conditions are frequently below 1% [ 30] Many researchers have attempted to enhance the photocatalytic efficiency of TiO2. However, the previous research has focused on either scavenging the electrons away from the system to prevent re combination, or by retarding the recombination so that a greater proportion of holes will reach the surface and generate OH [ 31 ] Furthermore, there are many parameters which were found to affect the photocatalytic degradation rate Some of them will be discussed in the following sections. P ho tocatalyst L oading The stoichiometry of a photocatalytic reaction in solution is comparable to that of any chemical reaction. It is straightforward that higher dosage concentration leads to more efficient photocatalytic reaction because of more adsorption and decomposition at higher available specific surface area [ 32] In any given experiment, an optimum value of photocatalyst loading should be explored to avoid the waste of materials a nd fully utilize total incident photons. When adding excess catalysts into the system, reduction of quantum yield was observed due to undesired light scattering of titania particles and then diminution of light penetration in the solution. There is also a relationship between photocatalyst concentration and type of reactor. Two main kinds of reactors have been used in the photocatalytic application: immobilized and slurry reactors. In the immobilized system, an optimum thickness of photocatalyst film ins tead of optimum loading is determined according to the geometry of reactor. For the slurry reactors, optimum concentration of catalysts have been reported in a wide range from 0.15 to 8 g/l

PAGE 28

28 depending on the specific set up of photoreactors [ 3337] The effective penetration of light result ing from optimum loading is especially important for a slurry system. The maximum utilization is achieved by adjusting the best combination of illumination intensity and catalyst concentration. L ight I ntensity Light absorption of the photocata lysts strongly affect s the photocatalytic degradation rate since it is the first step of photocatalysis [ 38] Generally, the degradation rate of organic compounds increases with the illumination intens ity during a photocatalysis process. However, the nature of incident light does not alter the react ion pathway of photocatalysis [ 39] In other words, there is no direct relationship between the electron hole excitation mechanism and decomposition of organic species during photocatalysis. Although photon flux is proportional to degradation rate, there are two regimes of decomposition process which could be classified with respect to light intensity. One is a first order reaction at low light intensities, especially laboratory scale (usually up to ~ 25 mW /cm2). The other is a half order reaction for even higher intensity. The reason why these two regimes occur is related to the recombination process in photocatalysis. Recombination is limited by the electrochemical reaction for the former case, but it beco mes dominate by the photocatalytic reaction in the latter case. p H The behavior of metal oxide particles in water i s well known to be amphoteric [ 4042] The surface of TiO2 particles can be protonated or deprotonated under acidic or alkaline condition respectively. The basic mechanisms were described in the following reactions: O H TiO OH TiOH TiOH H TiOH2 2

PAGE 29

29 Therefore, the titania surface can be either positively or negatively charged depending on the pH value of aqueous solution. There are two aspects of the influence of surface charges at titania surface s. One is the degree of dispersion for particles in water. Measuring the magnitude of the zeta potential is a common way to evaluate the st ability of colloid al suspension. Generally, a value of more than 25 mV or less than 25 mV for zeta potential will be taken as a st abilized system [ 4044] Another important parameter with respect to surface charge is the isoelectric point (IEP). IEP is the pH value of slur ry with zero surface charge. For example, the IEP of Degussa P25 was reported to be 6.56.7 [ 17,45] Thus, better dispersibility of a P25 aqueous suspension could be achieved by adjusting the pH value to higher or lower than 6.5. Surface charges of titania surfaces also st rongly affected the adsorption of organic species which is a crucial step for photocatalysis. In short, cationic molecules will be favored for the photocatalysis system with higher pH (> IEP) and anionic molecules will be favored at lower pH (< IEP). T empe rature Temperature rise usually occurred in the system during photocatalytic reaction due to when the high intensity of illumination unless appropriate cooling system is used An increase in temperature tends to intensif y the recombination of electron hole pairs and promotes desorption of the organic species at the titania surface [ 4549] Therefore, the temperature of a given photocatalysis system must be monitored and maintained in order to achieve optimum performance. O xygen P ressure Dissolved oxygen plays an important role in photocatalytic reaction. The molecular oxygen is electrophilic and thus captures the photogenerated electrons to prevent unfavorable recombination [ 5053] Furthermore, it also generated some active species such as superoxygen ions and may involve in the photocatalytic reaction to enhance the mineralization rate of organic

PAGE 30

30 pollutant in water ; it help s to maintain the stoichiometry of titanium dioxide during photocatalysis as well. These effects have been extensively studied, and will be listed in the following sections. C harge separation It was found tha t dissolved oxygen not only trap the photoassisted electrons but al so capture the holes [ 31,54] In other words, the present of molecular oxygen in water tend to suppress the undesired recom bination. Researchers design two ex periments to demonstrate that [ 55] They record the EPR spectra of titania photocatalysis with and without the present of dissolved oxygen. In the latter case, the number of trapped holes was proportional to the oxygen pressure i n water. The results illustrate the rel ationship between photogenerated holes and dissolved oxygen molecules. In these experiments, oxygen is supplied to the system by aeration with ambient air. G eneration of active species Depending on the number of electron reduction by dissolved oxygen, supe rox ygen ions or hydrogen perox ide are the two main products [ 17,56,57] Some researchers suggest that electron trapping by dissolved oxygen may be the rate limiting step in semicond uctor photocatalysis [ 5860] Figure 2 3 summarized the formation of active oxygen speci es in the secondary reactions [ 17] It is worthy to note that both superoxygen ions and hydrogen peroxide help to accelerate the photocatalytic reaction. M aintenance of the stoichiometry of titania Because of oxygen consuming during photocatalysis, the oxygen depleti on leads to reduction of stoichiometry of TiO2 and thus the photocatalytic ability [ 6165 ] Muggli et al. designed an experiment of photocatalysis of formic acid and acetic acid under inert gas atmosphere [ 62,63] The results showed the rapid deactivation of photocatalysis and then fast

PAGE 31

31 recovery of photocatalytic activity after providing the system with oxygen again. Based on this experiment, they suggested that the oxygen in the titania lattice is gradually lost during the photocatalytic reaction and could be compensated by aeration process. E nhanced of P hotocatalysis Titanium dioxide can be only excited by the UV light irradiation since it is a wide band gap material (3.0 or 3.2 eV for rutile or anatase phase respectively) However, the UV range is only about 3% of the solar spectrum. T here are two pathways for improving the photocatalytic efficiency of TiO2: (1) enhancing the photoabsorption in the near UV region [ 66] and (2) shift the absorption threshold towa rds visible light region [ 6669] To enhance photoabsorption, researchers have found that the enhanced photodegradation rate of organic species was achieved by using TiO2 nanoparticles under UV light irradiation. This method is relatively cost and energy intensive since the photochemical reaction has to be generated by a UV lamp. On the other hand, extending the photoresponse to the visible light range will enable the materials to better utilize the abundant and free solar energy. For this purpose, many attempts have been made by incorpora ting a nother semiconductor material or by dye sensitization. Nevertheless, the short term stability, low interparticle electron transfer rate, and low photocorrosion resistance have been difficult to overcome. S ize R eduction During the last decade, the int erest in the semiconductor nanoparticles originated from their unique phy sical and chemical properties [ 70] One of the advantages of using nanoscaled photocatalysts is the extension of the effective band gap due to the quantum confinement effects. When the particles size becomes smaller or comparable to the wavelength of electron (the DeBroglie wavelength), the charge carriers will be constrained within a potential well [ 71,72] And the bands of the semiconductor will spilt into severa l discrete electronic states (quantized

PAGE 32

32 level) in the valence and conduction bands [ 73,74] The other benefit of using nanomater ials is the enhanced photocatalytic activity due to the high surface/volume ratio [ 32] Many researchers investigated the role of particle size in titania photocatalysis [ 71,7577] Some of them cla imed that nanoparticles possess higher photoactivites than bulk materials. The other observed the opposite trend of reducing size on photocatalytic reaction rate [ 78,79] The possible explaining of that could be attributed to the competition between the positive factor (i.e. interfacial charge transfer) and the negative factor (i.e. surface recombination) for photocatalysis. When the size of titania particles falls below a certain level, surface recombination start s to overcome the fast interfacial charge transfer due to the short diffusion path of charge carriers in nanoparticles. S hape C ontrol Recently, nanoscale functional materials, especially low dimensional inorganic semiconductor materials have attracted great interest because of their sizedependent optical and electronic properties and potential applications in nanoscal e electronics and photonics Therefore, there has been great interest in these kinds of materials. Research has been focused on the morphology control, suc h as synthesis of nanotubes [ 52] nanowires [ 80,81] nanoribbons [ 35] diskettes [ 82 ] nanobelts, nanosaws, nanowalls, nanomultipods, nanorings, nanocages, nanohelixes, nanopropellers, and many others [ 8385] While one dimensional nanowires and nanorods have been extensively studied, twodimensional (2D) nanostruct ured materials have attracted much less attention until recently. However, 2D nanomaterials show strong potential as chemical and biological sensors, nanoelectronic devices, and catalysts with high surface areas and large pore volumes. Great interest in na nostructured materials have been focused on controlling the shapes of materials a nd finding novel properties [ 86] The preparation of nanoscale photocatalysts is also of great interest, because nanomaterials offer a high surfaceto volume ratio and short distance from the bulk material to the surface [ 32] Many small band gap

PAGE 33

33 materials such as ZnO, ZnS, and Bi2WO6 with flake like or plate like morphologies have been already synthesized successfully [ 8789] For example, Ye et al. indicated that thinner ZnO nanoplates had hig her photocatalytic activity [ 90] Although the above materials possess visible light absorption, the relatively expensive precursors hinder practical application. Since titanium dioxide is a versatile and low cost material in many industrial a pplications, many scientific works have been focused on particle size control down to the order of tens of nanometers. Quantization effects result in a shift in the absorption edges to longer wavelengths (blue shift) of absorption and therefore the UV light absorption is enhanced [ 42,71,91] Except for size control, shape control of particulates, especially on a nanometer scale, is more difficult and receives less attention. The most common shape of the fine titanium dioxide particles is sp herical in many processes. Thin films or fibers have been fabricated by being supported on a substrate or in the interstices in some three dimension network [ 92] However, the practical application is limited by the complexity and cost of the synthesis process. Sasaki fabricated thin titania flakes through exfolia tion of a layered precursor [ 93] Although the specific surface area of the flakes is about 110 m2/g, the photocatalytic activity is still less than commercial product, Degussa P25 (49 m2/g) Li et al. in 2007 synthesized Brookite phase titania nanoplates by using titanium trichloride (TiCl3) precursor thr ough hydrothermal processes [ 94] Under the same surface area of loaded TiO2, the brookite nanoplates exhibit the highest efficiency in the beaching of methyl orange solution under UV irradiation. Therefore, this research is focused on the effect of shape on photocatalysis with titania. Using TiO2 nanoparticles for water t reatment is limited in practical application since it is very difficult to remove these ultrafine particles due to very small mass. Therefore, the conventional separation methods such as centrifuging, filtration, and sedimentation are difficult

PAGE 34

34 and expensi ve to implement. In addition, the efficiency of photodegradation by using nanoparticles is not very high because of the poor accessibility of the organic pollutants to catalyst surface caused by the agglomeration of particles [ 95] However synthesizing larger flakelike titania with nanosized thickness will alleviate this problem. These titania flakes can be easily separated from the treated water by simply filtration or sedimentation. Because the flakes are nano thin, superior photocataly tic properties are retained due to high surface to volume ratios and short di ffusion paths, which are favor able for the migration of electrons and holes. This reduces the probability of the recombination of photogenerated electrons and holes. At the same t ime, the common agglomeration problem caused by nanoparticles can also be mitigated and therefore maintain the advantages of micro and nanostructure. Doping (Narrowing Band Gap) The most abundant energy source on earth is the solar energy which is mostly in the visible and near infrared ranges with a very small part of UV. T he incoming solar irradiation at the upper earth surface is 174 petawatts and about half of them will be reflected back to the space. Despite of that, one hour sun light irradiation was still more than the energy consuming of entire human being for one year according to 2002 statistics. Unfortunately, titanium dioxide can only absorb the UV light which is the small portion of solar energy due to the large band gap (Figure 24) [ 72]. Therefore, there is a need to develop the smaller band gap of photocatalysts associated with high visible light activities in order to utilize a greater portion of solar spectrum. Nonmetallic do ping in titania is one of the effective method s to narrow the band gap of the photocata lyst. In 2001, Asahi et al. reported that nitrogen doped titanium dioxide showed photoresponse at wa velengths longer than 400nm [ 96] They found that nitrogen atoms substituted the oxygen atoms in the TiO2 lattices and narrowed the band gap by mixing the N 2p and O 2p states. After that, the synthesis of titanium dioxide with high photoactivity under

PAGE 35

35 visible light was attempted by doping different nonmetallic ions (S, N, C, P ions etc.) into TiO2 [ 67,9699] However, the nonmetallic doping methods usually required either complicated equipments or extreme process conditions such as high temperature, high pressure or poisoned dopant sources and therefore were not a cost effective way to fabricate the visible light photocatalyst. Except for nonmetallic elements, doping titania with transition metal ions was found to cause the red shift of band gap and therefore was a promising way to produce visible photocatalysts. There were many different metal elements successfully doped into the lattice of TiO2 such as vanadium, chromium, manganese, iron, copper, zirconium, and tungsten [ 100] Figure 2 5 shows the schematic representation of energy levels of several dopant ions in TiO2 [ 100]. Among these dopants, vanadium is the most frequently investigated candidate elements because of its promising red shift a bility, comparable ion radius of titanium and outstanding catalytic ability of titania vanadia composite Anpo et al. demonstrated the doping effect on red shi ft of photocatalyst absorption edge as the following sequence: V > Cr > Mn > Fe > Ni by bombarding ion beam s into titania lattice [ 101,102 ] Several approaches have been used to fabricate titania vanadia mixed photoc atalyst such as co precipitation [ 103], ionimplantation [ 104,105], liquid phase deposition [ 106 ], wet impregnation [ 107], and sol gel [ 108]. The sol gel pr ocess provides a new approa ch for the synthesis of novel nanomaterials Starting from molecular precursors, a threedimensional oxide network is formed by hydrolysis and polymerized reaction. These reactions occur in solution and are usually termed so l gel processing which is used t o describe the preparation of inorganic oxide by wet chemistry routes. Furthermore, the incorporation with metal ions occurs in the sol gel formation step which provides better homogeneity and feasible control of the final product shape. Therefore, unique catalytic properties could be expected using this technique Wu et al. developed two modified

PAGE 36

36 sol gel methods for synthesis of series vanadium doped titania samples with pertained anatase phase after 40 0 C heat treatment [ 108]. Kiosek et al. investigated the photocatalysi s process of ethanol on titania vanadia mixed catalysts under visible light using 13C NMR method and found that the active species was V4+ instead of V5+ [ 109 ]. Zhao et al. synthesized vanadium doped titania photoelectordes using sol gel process and generated photocurrents under visible light irradiation [ 110]. So far, vanadium doping was proved to be an effective method to not only narrow the band gap of titania but also enhance the photocatalytic efficiency under visible light illumination. In this study, we utilized this method to produce visible light photocatalysts and investigated the doping effect of vanadium on the visible photocatalysis efficiency of vanadium doped titania nanoflakes.

PAGE 37

37 Table 2 1. Bulk properties of three TiO2 polymorphs [ 24] Rutile Anatase Brookite Crystal structure Tetragonal Tetragonal Ort horhomic Space group mnm P Dh/4 2 14 4 amd I Dh/4 1 19 4 Pbca Dh15 2 Lattice constant ( ) a = 4.584; c = 2.953 a = 3.733; c = 9.370 a = 5.436 ; b = 9.166; c = 5.135 Density (g/cm3) 4.24 3.83 4.17 Electron mobility (cm2/Vs) 1 10 Dielectric constant 6.62 6.04 7.89 Refractive index (np) 2.621 2.561 2.583 Band gap (eV) 3.0 (indirect) 3.2 (indirect) Table 2 2. Oxidation power of species [ 26] Species Relative oxidation power Hydroxyl radical 2.06 Singlet oxygen radical 1.7 8 Hydrogen peroxide 1.31 Perhydroxyl radical 1.25 Chlorine dioxide 1.15 Chlorine 1.00

PAGE 38

38 Figure 2 1. Schematic representation of the photoelectrochemical mechanism taking place inside titania: (1) light irradiate the semiconductor. (2) forming of electron hole pair. (3) charge carriers migrate to the surface. (4) initiation of an oxidative pathway by a valence band hole. (5) initiation of a reductive pathway by a conductionband electron. (6) electron and hole recombination to heat or light. Figure 22. Schematic photoexcitation in a solid followed by deexcitation events [1 9].

PAGE 39

39 Figure 2 3. Secondary reactions with active oxygen species in the photoelectrochemical mechanism [ 17]. Figure 2 4. The schematic diagram of conduction and valence bands positions of several semiconductors The left hand scale is the internal energies to the vacuum level and the right hand scale represented the comparison to normal hydrogen electrode (NHE) [ 72].

PAGE 40

40 Figure 2 5. Schematic energy level diagram of metal dope d TiO2 (a) Pure TiO2 (b) Iron doped (c) Copper doped (d) Zirconium doped (e) Cerium doped (f) Vanadium doped (g) Tungsten doped TiO2 [ 100 ].

PAGE 41

41 CHAPTER 3 HIGH ASPECT RATIO TI TANIUM DIOXIDE FLAKE S AS PHOTOCATALYSTS S ynthesis of H igh A spect R atio of TiO2 F lake s Titania nanoflakes were synthesized by a modified sol gel method which was based on surface controlled hydrolysis of titania precursor on water surface. A mixed oil phase contained stearic acid (C18H36O2), hydrocarbon and titanium tetraisopropoxide (Ti(O C3H7)4) w as delivered by spreading it on the surface of an aqueous phase which was consisted of high purity water (Barnstead Nanopure Infinity, 18M /cm1), nitric acid (HNO3) and sodium dodecyl sulfate (C12H25SO4Na). Stearic acid and hydrocarbon were used to decrease the viscosity of titania precursor and enhance the spreading ability of the mixture. After spreading, a titania precursor film spontaneously hydrolyzed at the oil/water interface leading to the formation of thin titanium dioxide flakes. The t hickness of nanoflakes could be easily manipulated by varying the volume ratio of titania precursor and hydrocarbon. Typically, a 1:8 volume ratio of titania tetraisopropoxide to hydrocarbon w as used to synthesize titania nanoflakes with thickness of about 40 nm. The resulting slurry was wash ed with Nanopure water and then centrifuged (Be ck man JA 21 Centrifuge ) at 3000 rpm for 15 minutes to concentrate the slurry. The above cleaning process was repeated five times to get rid of organic residuals. After solv ent exchanging with isopropanol alcohol ((CH3)2CHOH) the nanoflakes were dried by a supercritical fluid drying process [ 111] The following heat treatments were obtained by using a programmable electric furnace with desired heating rate. For convenience, we will call t he nonheat treated and heat treated materials as synthesized and calcined flakes respectively. C haracterization of TiO2 F lakes When an appropriate amount of flakes are disper sed in water, significant visual pearle s cen c e occur s when shaking the solution, a qualitative indication that particles in water are

PAGE 42

42 flaky (Figure 3 1A and B ). It is well known that titania is a high refractive index material within the visible light range, and therefore is commonly used as the basic pigment in paintings. When the lar ger side of flakes faced to the incident light, the strong reflection of light occurr s at the surface of the particles and generates the luster effect. However, the luster is reduced when the incident light diffusely scattered by the edge of the flakes [ 112,113]. Alternative brightness and darkness occurred when flakes are randomly orientated in the solvent. This pearl escent effect does not occur in the P25 slurry owing to the isotropic shape of the particles (Figure 3 1C ). The low magnification optical microscopic (OM) images of both fl akes indicated flat particles with major dimensio n in micrometer range (Figure 3 2 ). The strong interference color observed under dark field microscopy indicates a thin minor dimension, confirmed by SEM to be in the 3040 nanometer range. The flakes were f urther characterized with Scanning Electron Microscopy (SEM), Laser Diffraction for particle size analysis, X Ray Diffraction (XRD), Transmission Electron Microscopy (TEM), Physisorption Techniques, UV Vi sible Spectrometer (UVVis) X ray Photoelectron Spe ctroscopy (XPS), and E lectrophoretic M obility (Zeta Potential) The surface morphology of synthesized nanoflakes was investigated using SEM images, shown in Figure 3 3. The major diameter of the flakes w as found to be of the order of 203 ure 3 3), and the thickness of these flakes was approximately 40 nm (Fig ure 3 4 A ). The aspect ratio of major dimension to thickness ranged from 250:1 to 500:1. The flakes were further treated by calcination at 400 C in air for 2 hours. Aggregation was not apparent when comparing these treated flakes with those obtained directly from synthesis. Moreover, the thickness of calcined flakes did not change by the heat treatment (Figure 3 4 B ) A comparison of particle size distributions measured on the synthesized and c alci ned flakes is shown in Fig ure 35. Particle size distributions of synthesized and calcined flakes were measured using laser diffraction. When

PAGE 43

43 we were using a differential volume distribution for flaky particles that move d in turbulence and pass ed the l aser beam the maximum diameter was measured through the average random orientation of the flakes. The la rger dimension of nanoflakes were estimated by dispersing both flakes in deionized water with liquid modules and were fairly close to the results taken from the SEM images (Figure 3 3) The D10, D50, D90, mean and standard deviation values for the particle size distributions are shown in Table 31. The synthesized material has a broad size distribution spanning from 150 of 39.1 and 23.5 m for the synthesized and calcined flakes respectively (Figure 3 5) Comparing the volume distributions it is evident some larger flakes break or crack during dehydration and crystallization. Crystalline structure changes of the titanium dioxide flakes we re monitored by powder XRD (Figure 3 6). The synthesized flakes show broadening and weak Bragg peaks which indicates the nature of flakes consist of partially amorphous material with a presence of the anatase phase (Figure 3 6 A ). After heat treatment, the expected phase transformation from amorphous to crystalline titanium dioxide was confirmed by seven characteristic diffraction peaks (Figure 3 6 B ). The heat treated flakes were converted to a pure anatase phase [JCPDS: 211272] which is the most photoact iv e phase of titanium dioxide [ 16,17,72] There is no indication of the rutile phase by XRD. The intensity of characteristics peaks is comparable to commercially available pure anatase standard materials (Figure 3 6 C ) which indicate complete conversion of the amorphous titanium dioxide. The average crystal grain size can be calculated by the Scherer equat ion (equation 31): BB k d cos .....(31) where d is the calculated grain size, is the wavelength of X ray (Cu K 1.54), B is the full width at half maximum intensity, and B is the Bragg diffraction angle. The gr ain size is

PAGE 44

44 determined to be 4 and 9 nm for synthesized and calcined flakes respectively (Table 3 2). The photocatalytic activity of nanocrystallite titanium dioxide was shown to be strongly dep endent on the grain size [ 114,115] Generally, the enhancement of interfacial charge transfer rate occurs at smaller grain due to the decrease of volume recombination process and the increase of the surface area. However, t he photocatalytic efficiency is not continuously increasing with decreasing grain size since surface recombination process tends to dominate the photocatalysis. Therefore, a critical size for optimum photoc atalytic efficiency is about 10 15 nm has been rep orted by several researchers [ 114116] The crystalline structure of the flakes was further investigated under HR TEM. The images show the flakes are comprised of circular crystalline platelets of about 5 8 nm in diameter (Figure 3 7). The interference lattice fringes can be seen in the TEM images and has a separation distance of 0.35 nm, corresponding to the interplanar spacing of t he (101) planes for anatase [ 117] Random orientation of individual grains over both flakes is suggested f rom the concentric diffraction rings in the select area diffraction mode and consistent with the anatase (101), (004), (200), (105) for circles 1 to 4 respectively (the insets of Figure 37). On closer inspection, an amorphous layer can be seen surrounding the smaller circular crystallites in s ynthesized sample (Figure 3 7 A ) After calcination, pores developed due to local rearrangement and growth of crystal grains. Consequently, it is thought that the flakes are polycrystalline consisting of fine grai ns of anatase titanium dioxide. Surface area and porosity measurements were perform ed to verify the changes in the structure of the titanium dioxide due to the calcination treatment. This was measured using nitrogen absorption isotherms in conjunction with the Brunauer EmmettTeller (BET) model. It is desirable for photocatalytic materials to have higher specific surface area as the reaction rate is proportional to the number of available surface sites. However, high surface area powders are

PAGE 45

45 usually associated with large amounts of crystal defects favoring fast recombination of electrons and holes, ultimately leading to lower photoactivity Compared to a commercial photocatalyst, Degussa P25, the surface areas of nanoflakes were 3 6 times higher (Table 3 3). Ex cept for specific surface area, the surface chemistry of titanium dioxide is another surface property which significantly affects the photocatalytic reaction. This study is mainly focused on the photocatalysis in water. Therefore, the interaction between t itanium dioxide particles and water is an essential problem. It was well known that the chemistry of metal oxides surface is mainly dominated by the hydroxyl groups arise from the interaction with the environment such as moisture in atmosphere or water in aqueous solution [ 118,119] Surface OH groups were determined by taking XPS spectra of O 1s signal for all of the samples under the same preparation condition. Using the Gaussian mixture peak fitting technique, two components are shown in the typical O 1s XPS spectrum (Figure 38 A ). One rep resentes the oxygen element in titania lattice (529. 9 eV) and the other corresponde s to the surface hydroxyl species (531.9 eV) [ 120] The ratio of surface OH to total oxygen 1s signals for the samples is shown in Figure 38 B It should be noted that both samples have higher surface hydroxyl concentration tha n commercial product, P25, on the same mass basis. Calcined flakes have less OH groups than synthesized samples due to the dehydroxylation proce ss after heat treatment. It ha s been widely reported that the removal of surface absorbed water was carried out by annealing the sample at 450500 C [ 118,121] Higher density of OH groups for synthesized and calcined fla kes were attributed to the higher specific surface area comparing to the Degussa P25. Hydroxyl radical, one of the most reactive species in the photocatalysis process, is strongly related to the hydroxyl group concentration at titania surface. Therefore, w e could expect that better photocatalytic performance could be achieved by the sample showing the larger OH component in the XPS oxygen 1s signal.

PAGE 46

46 The photocatalytic activity of these flakes was investigated by performing dye degradation experiments under ultraviolet activation. Figure 3 9 A shows typical UV visible diffuse reflectance spectra in the wavelength range of 250 500 nm for the synthesized flakes together with the calcined flakes. The sharp decrease in the diffuse reflectance in the UV region res ults from the fundamental light absorption of the titanium dioxides flakes and a blue shift of the onset of reflectance occurred at calcined sample due to heat treatment. In semiconductor physics, the general relation between the absorption coefficient and the band gap energy is given by g mE h h ..(32) where m is an index depending on the nature of the electron transitions, is the absorption coefficient, h is the Planck constant, is the frequency of electromagnetic radiation, and Eg is band gap energy of the semiconductor. To estimate band gap energies of both nanoflakes, data in Figure 3 9 B is plotted as (h)2 versus h and the optical absorption energy is determin ed via extrapolation. An increase in band gap from 3.25 to 3.33 eV, i.e. a blue shift, is mostly likely due to the quantum confinement effect of higher crystallinity after calcination and thin flaky morphology [ 117] Kim et. al suggest ed that the differences in atomic structure between the grain boundary and the amorphous area potentially leads to larger concentrations of electrons and holes as well as the existence of potential barriers at these interfaces [ 122] Hence it is possible for an electric field to be generated, the band gap energy to increase, and the absorption limit shift to a narrower wavelength range. It is important, but also difficult to measure the fraction of light absorbed in the TiO2 suspensions due to the inevitable scattering effect. In this study, a modified method based on the use of the integrating sphere was used to evaluate the true absorb ance in the titani a slurries [ 123] The schematic of assembling integrating sphere in this method is shown in Figure 310. The fundamental principle of this method is described as

PAGE 47

47 the followings: First of all, the re ference solution (for example, water) was placed between the incident beam and reflection plate which consisted of fine barium sulfate particles (Figure 310(a)). The absorbance of the standard could be calculated by equation 33. o a RI I I A 2 log0 .(33) Where AR is the instrumental reading, Ia is the light flux absorbed by the reference solution, and the coefficient comes from the incident beam pass through the cell and back scatter again. When the titania samples wer e put in the same position of reference solution (Figure 310(b)), the instrumental reading could be described as equation 34. 0 02 2 log I I I I Asa a s (34) Where As is the instrumental reading of titania sample, Isa is the photon flux absorbed by titania particles in water. According to these two equation, the true absorbance of titania particles is a function of the ratio of Isa to Ia and could be expressed by equation 35. 2 10 10 1 log 1 logs RA A a saI I A (35) Figure 311 show s the comparison of the true absorbance for three different titania samples using the modified method on the same mass basis. Calcined titania flakes have the highest optical absorption within the wavelength range from 250 to 350 nm. The synthesi zed titania flakes have similar absorbance on the shorter wavelength (250300 nm) regime compared to P25. The higher photon flux absorbed by titania sample implied that more active species generated during the photocatalytic process and thus higher quantum yield could be achieved.

PAGE 48

48 The surface charges of a metal oxide surface within an aqueous solution are highly affected by the dissociation of the surface hydroxyl groups [ 118,121] The equilibrium of an oxide in water could be expressed by equation 36. H O M OH M OH MpH g increa pH g decrea sin sin 2 ...(36) The above equation indicated that the surface of metal oxide may car ry either positive or negative charges according to the pH of the solution. In terms of the chemical structure of the methylene blue, it is a cationic dye and therefore adsorbed preferentially at t he negatively charged surface [ 124] was determined to be about 6.66.7 using the pH meter. In order to evaluate the surface charge of the titania samples, the electrophoretic mobility of the particles were measured to determine the zeta potential of the surface (Brookhaven Zetaplus Zeta Potential Analyzer). Zeta potential of the titania samples as the function of solution pH is plotted in Figure 312 and the isoelectric points (IEP) are listed in Table 3 4. The IEPs were estimated to be 5.8, 5.1, 5.0 for P25, synthesized and calcined titania flakes respectively. Both titania flakes have lower IEPs than P25 nanoparticles owing to the higher concentration of surfac e hydroxyl groups (Figure 38 B ). It was well known that the surface charge of the parti cles is very sensitive to the surface density of OH group especially in the aqueous solution. It is noteworthy that the IEP of the flakes is not severely shifted af ter heat treatment which implie s that there is no difference in the nature of the surface hy droxyl group between both flakes. However, the lower magnitude of the surface charge at the same pH value for calcined flakes was due to the dehyroxylation process during heat treatment. The surface charges of both flakes at pH= 6.6 were higher than that of P25 (Table 33) and indicated the more favorable adsorption of methlyene blue molecules at the surface of the both flakes.

PAGE 49

49 Therefore, higher photocatalytic performance of the flakes could be expected due to the stronger surface adsorption of the methylen e blue molecules. Benchmark: Degussa P25 Degussa P25 is a commercial titanium dioxide nanoparticles synthesized by using a high temperature flame pyrolysis process with titanium chloride precursor. The primary particle size is 30 nm under TEM investigati on [ 125]. It is a mixed phase material which contains anatase and rutile phases in a ratio of 3:1. However, rutile particles exist separately from larger anantase particles instead of covering the surface of P25 nanoparticle as surface layer. The specific surface area is relatively high, ~ 50 m2/g (Table 3 3). Because of its high activity for many kinds of different photocatalytic reactions, it was used as the benchmark for activity comparison. Researchers summarized the reasons why P25 showed high photocatalyt ic efficiency for two points: (1) slow recombination due to stable charge separation by electron transfer between the mixed two phases (2) the smaller band gap of rutile phase (3.0 eV) extend the photorespons e to the visible light region [ 125,126] E xperiment A pparatus Dye decomposition experiments were performed inside a cylinder UV reactor equipped with 6 UVA lamps (Southern New England Ultra V iolet Company, Branfield, CT). A 150ml Erlenmeyer flask contained dye solution with TiO2 particles were placed inside the UV reactor and continuously agitated by a magnetic stir. House air was delivered by a glass tip connected to the plastics hoods from a ir outlet for investigating the aeration process. The amount/volume of input air flow was controlled by a flow meter. A cold air flow generated by a cooling fan circulated through the cylinder UV reactor was to prevent the heating effect during the photoca talysis process. 100 ml of dye solution with concentration of 50 M were prepared. Three photocatalysts were added to the dye solution respectively and then stirred for 10 minutes

PAGE 50

50 in the dark before UV light illuminating. Prior to turn on the UV light, abo ut 2 ml of dye solution was pipetted out and stored in a plastic vial. After starting the photocatalytic process, the same amounts of samples were taken by every 15 minutes for decoloration study. For avoiding the artificiality caused by reflection of tita nia dioxide particles during the UV Visible spectrometer study, each collected sample were centrifuged at 10000 g three times to settle down the particles. Then the centrifuged sample was carefully transferred into a double sided cuvette without breaking t he bottom titania cakes. P hotocatalytic D egradation of D ye M olecules under UVA L ight I llumination The photocatalytic activity of titanium dioxide nanoflakes were probed by decomposition of methylene blue solutions. Degussa P25, one of the most efficient photocatalysts was used as benchmark for comparison. The UV Visible spectra were obtained to evaluate the photocatalytic degradation of dye solution. Figure 313 is a typical methylene blue dye absorbance spectra resulting from 2 hours photocatalysis with sy nthesized titania flakes. The maximum absorption wavelength ( m) of methylene blue, 662 nm, was chosen for calculation the decomposition rate of dye solution under photocatalytic process. Figure 3 14 compares the photocatalytic activity of P25, synthesize d and calcined titanium dioxide nanoflakes by removing the methylene blue from water under UV light irradiation. Degussa P25 was used as a reference material. It has a mean diameter of 30 ~ 40 nm which is comparable to the minor dimension of nanoflakes (40 nm). It was observed that calcined flakes exhibited the highest photocatalytic efficiency on the degradation of methylene blue among all tested samples. A first order rate re action showing in Figure 314 A indicates that dye concentration is the limiting factor. In contrast, significant enhancement is observed when introducing air bubbles into the system cont inuously (Figure 314 B ). It was well known that surface adsorption of organic species is very crucial to the titania photocatalysis. The whole

PAGE 51

51 photoc atalysis process could be separated by several parts. First, organic molecules need to be a dsorbed at the available sites of titania surface. Then the UV light strikes the titania particles which generate an electron hole pair. The charge carriers diffuse to the surface and form either hydroxyl radicals or s uper oxide ions. Finally, the a dsorbed molecules contact the above strong reductants or oxidants and then photocatalysis occurs It has been demonstrated experimentally by several researchers that dye de gradation rate of UV/TiO2 photocatalysis follows a pseudo first order expressed by the Langmuir Hinshelwood (L H) mechanism [ 16,20,127131] .Therefore, this model will be used to analyze and calculate the kinetics of dye degradation with photocatalysis in this study. In the L H model, the reaction rate is proportional to the surface cover age which becomes proportional to the concentration C at low concentrations range and could be written as equation 36. dt dC C k kKC KC KC k k rpseudo 1 ..(36) Where r is the reaction rate, k is the kinetics constant, K is the absorption constant, C is the concentration of dye solution and kpseudo is the pseudo first order constant. The linear transform of degradation curve is plotted as ln( C/C0) versus irradiation time and the rate constants are calculated from the slopes of Figure 3 15. The rate constants are shown in Table 35. The reaction with added turbulence present a pseudo first order reaction and much higher efficiency especially for the flake systems. One possible explanation of these differences may be correlated to oxygen depletion dur ing the photocatalysis process. From the results of XRD and TEM, both flakes consisted of very small nanocrystallites which imply a large number of defects (grain boundaries) within both materials. Consequently, fast recombination of photoassisted charge c arriers preferentially occurs at these local defect sites and dominates the reaction. The photocatalytic performance is not proportional to surface area without supplying oxygen to the

PAGE 52

52 system. By introducing air into the system, dissolved oxygen will prima rily become an electron accepter and may form superoxide radicals [ 132] More importantly, fast recombination, the rate limiting step, could be depressed by eliminating excited conduction band electron for higher efficiency. Statistic al Design of Experiments There are several possible parameters that affect the heterogeneous photocatalysis of this system such as light wavelength, light intensity, irradiation time, catalyst loading, target pollution concentration, pH reactor geometry, and dissolve d oxygen concentration [ 45] It is impossible to systematically study all of the parameters in limited time. However, from the preliminary experiments (Figure 3 14), light intensity, catalyst loading and aeration rate were determined to significantly affect the photocatalytic efficiency. Therefore, a statistic al design of experiments was performed add ressing these three parameters by using the Design Expert software. There are at least three advantages for doing so. First, the interaction between multiple parameters could be easily investigated by limited experiments. It saves time, cost, and mater ials. Second, the experiment error could be separated form results leading to more accurate evaluation of these parameters. Third, fewer e x periments are needed to obtain statistic ally relevant results More importantly, this method allows the researcher to generate mathematical models and map the experiment al space by identifying which variables are significant and fitting the data to the developed model. The optimiz ed response is identified by using the developed model leading to the conditions necessary t o achieve the highest performance The parameters investigated in the design of experiments were A: bubbling flow rate (cubic f eet per hour), B: mass concentration of photocatalysts (ppm), and C: Ultraviolet light intensity (number of lamps). A full 23 factorial design with three center points was used to obtain the developed model and estimate the experiment error. The detail of design of e xperiments is

PAGE 53

53 shown in Table 36. This statistic al design was performed for three different materials as listed in the preliminary results for further comparison. The results for three photocatalysts are outlined in the following sections. Synthesized Flakes The results for the synthesized flakes are shown as a 3D representation of design space (Figure 3 16) Three experi ment variables are presented on each side of the cube respectively and each corner indicated one point in the design matrix. The dye degradation% obtained by synthesized flakes photocatalysis is pointed out as the number located at each corner inside the c ube. The best results for synthesized flakes could be seen using high level of light intensity, catalyst loading, and flow rate. However, the complete degradation of dye molecules within 2 hours is not found in the whole range of design points (Figure 316 ). A developed model expressed by the equation 3 7 was developed by the software. Dye degradation% = 42.45 +5.63A (Flow rate) +17.48B (Catalyst loading) +10.32C (Light intensity) +5.30AB +8.25BC..(37) The significant factors for d ye degradation with synthesized flakes are shown in equation 37 with the magnitude of their effect. The mathematical equation indicate s that light intensity and mas s concentration of catalysts have the largest effect on photocatalytic efficiency. There were some small interactions between these factors and the model did not include the AC (Flow rate Light intensity) interaction because it was insignificant compared to the decomposition%. It is reasonable that these two factors are controlled by individua l devices (i.e. flow meter and UV lamps) and there are not any intrinsic effects between these two factors on the photocatalytic efficiency. The surface response plots, the contour lines of the constant response and interaction curves are constructed to gi ve the graphical representations of the decomposition% as a function

PAGE 54

54 of two factors when the third factor is fixed. The surface response of factor A and B under different intensity of light are shown in Figure 3 17. These cur ves show that there i s only lim ited benefit on degradation rate at low concentration of catalyst no m atter how high the flow rate is Howeve r, the dramatic increase occur s when increasing the amount of photocatalysts. The results indicate that there were no enough active species (or photocatalyst) to oxidize the organic pollutants and therefore increasing bubbling rate did not speed up the photocatalytic reaction. Moreover, the aeration process further improved the overall efficiency only with moderate to high catalyst concentration (Fig ure 3 17 B and C ). The 2D contour lines of response data and interaction plots are shown in Figure 3 18 and 319 for two level of light intensity (C = 2 and 6 lamps ) respectively. Th e curvature of these contour lines suggests interactions between factor s A and B (Figure 318 A ). Both the interaction plot s show similar tend s : degradation efficiency increases by increasing the amount of catalyst and bubbling rate further increase s the efficiency especially for higher loading s of titania (Figure 3 18 B and 319 B ). It is worthy to note that increasing mass of catalyst did not improve the efficiency at lo w light intensity and flow rate, presumably due to insufficient photons activating the reaction at the excess available surface area (Fig ure 3 18 B ). Figure 3 20 show s the 3D surface response of decomposition% as the function of factor B and C. It is apparent that the interactions between these two parameters are significantly stronger than the AB interaction (Figure 3 17) especially for high level of both factor s. T his result is also confirmed by the magnitude of BC term in equation 37. There is only minor enhancement on reaction rate when increasing the light intensity at low catalyst loading. In contrast, the increase the amount of photocatalyst largely improv ed the efficiency without changing the intensity of light. This is consistent with both the experiment observation (equation 37) and fundamental theory: the photocatalyst is the most significant factor in the phot ocatalysis reaction. Figure 3 21 and 3 22 show both the contour lines and interaction plots as the function

PAGE 55

55 of B and C for two levels of flow rate (A = 0 and 10) respectively. When fixing the flow rate, the BC (catalyst loading* light intensity) interaction shows the same trend but stronger intera ction. The 3D surface response of factor A and B shows an incremental trend for both factors and similar trend also presents in response of factor B and C (Figure 319). The maximum degradation%, 86.7%, could be predicted using the developed model and furt her proved by experiment. The full ANOVA analysis is listed in Appendix A. Calcined Flakes The cube plot of calcined flakes photocatalysis is show n in Figure 3 23. The complete degradation of dye indicates that calcined flakes have higher photocatalytic a ctivity than synthesized flakes and P25. The software generated model for calcined flakes is described by equation 38. Dye degradation% = 62.10 +2.18A (Flow rate) +14.55B (Catalyst loading) +16.85C (Light intensity) +1.07AB +3.30BC.(38) The most significant factors are the mass concentration of catalysts and light intensity for calcined flakes. Again there is no AC interaction owing to the nature of these two parameters. Figure 3 24 show s the surface response as the function of fa ctor A and B. There is a dramatic increase at constant light intensity when adding more catalysts into the system. Aeration process did not improve the efficiency at low catalyst concentration (50ppm), but helped to enhance the reaction rate when higher am ount of photocatalyst presented in the system (Figure 3 25 and 326 ). It is worthy to note that there is an enhancement of catalyst loading on overall efficiency for calcined flakes under weak illumination compared to no apparent effect for synthesized sam ple under the same condition (comparing Figure 3 18 B a nd 325 B ). When the 3D response plot is considered as the function of factor B and C, factor C (light intensity) is the most significant factor and the strong interaction is shown in the Figure 3 2 7. It is more obvious

PAGE 56

56 in the contour and interaction pl ot (Figure 328 and 3 29). These results suggested that current light source was not enough for the excess surface area of calcined flakes. Therefore, the number or intensity of photon flux is the limitin g factor for photocatalysis with calcined flakes in this case. Back to the characterization part, we know that calcined flakes have good crystallinity, high specific surface area, appropriate grain size and higher light absorption ability which implied hig h photocatalytic activity. F rom both characteristics and experiment design aspects, we could conclude that the calcined flakes are the most efficient materials for photocatalysis in this study. P25 The maximum decomposition% of P25 was slightly higher than synthesized flakes under the same condition. Complete degradation of dye molecules was not achieved using Degussa P25 as photomediated agent in the current d esign of experiment (Figure 3 30). Factor C is again the most significant factor and weaker inter actions also present in the computer developed model equation 39. Dye degradation% = 50.10 +2.08A (Flow rate) +14.15B (Catalyst loading) +19.03C (Light intensity) +2.02AB +0.77BC.(39) The similar trends respect to calcined sample are shown when the surface responses are plotted as the function of AB and BC for P25 (Figure 331 and 3 32). It is worthy to note that there is a very weak interaction between factor B and C compared to both flakes (equation 39). It is more apparent whe n the parallel lines presented in the interaction plots under the constant flow rate (Figure 3 33). The results suggested that the amount of catalyst is the limiting factor for photocatalysis. In order to achieve higher efficiency, we need to increase the amount of P25 instead of adding more lamps.

PAGE 57

57 Table 3 1. Particle diameter statistics for synthesized and calcined flakes Sample D10 ( m) D50 ( m) D90 ( m) Me di an ( m) Standard Deviation ( m) Synthesized flakes 5.2 39.1 81.6 20.8 58.9 Calcined flakes 5.1 23.5 50.1 18.4 18.8 Table 3 2. Grain size calculation by the Scherer equation for both nanoflakes Sample B (degree) B (degree) d (nm) Synthesized flakes 2.72 25.91 4.1 Calcined flakes 0.93 25.35 8.7 Table 3 3. Physisorption measurements of P25, synthesized and calcined titania nanoflakes Sample Specific surface area (m 2 /g) Specific pore volume (cm 3 /g) Average pore diameter (nm) Synthesized flakes 323 Calcined flakes 151 0.511 7.2 P25 49 Table 3 4. Optical and electrostatic properties of titaia samples Sample Band gap energy (eV) IEP Surface charge at pH= 6.7 (mV) Synthesized flakes 3.25 5.1 17.9 Calcined flakes 3.33 5.0 12.7 P25 5.8 7.5 Table 3 5. Pseudofirst order rate constants of methylene blue under photocatalytic d egradation Without bubbling treatment With bubbling treatment Sample k pseudo (10 3 min 1 ) Sample k pseudo (10 3 min 1 ) Synthesized flakes 1.0 Synthesized flakes 10.0 Calcined flakes 3.8 Calcined flakes 32.5 P25 2.0 P25 11.9

PAGE 58

58 Table 3 6. The 23 fact orial design used to investigate the most important factors for UV light photocatalysis Factor 1 Factor 2 Factor 3 A: Light intensity (number of lamps) B: Catalyst loading (ppm) C: Bubbling flow rate (ft 3 /hour) (2) (50) (0) + (6) (50) (0) (2) + ( 100) (0) + (6) + ( 100) (0) (2) (50) + ( 10) + (6) (50) + ( 10) (2) + ( 100) + ( 10) + (6) + ( 100) + ( 10) *Center point: A (flow rate): 5; B (catalyst loading): 75; C (light intensity): 4

PAGE 59

59 Figure 31. Photographs of titania sl urries A) synthesized flakes. B) calcined flakes. C ) P25 on the same mass basis (note the pearlescence of the samples) Figure 3 2. O ptical micrographs of titania samples A) synthesized flakes. B) calcined flakes A B C A B

PAGE 60

60 Figure 3 3. SEM images of titania samples A ) synthesized flakes. B ) calcined flakes Figure 3 4. SEM images of edge view of titania samples A ) synthesized flakes. B ) calcined flakes A B A B

PAGE 61

61 0.1 1 10 100 1000 0 1 2 3 4 5 Calcined flakesDifferential volume %Particle diameter m ) Synthesized flakes Figure 3 5. P artic le size distribution for synthesized and calcined flakes 30 40 50 60 70 JCPDS 21-1276 (Rutile) JCPDS 21-1272 (Anatase) C Anatase standard nanoparticles B calcined titania flakesIntensity (Arbitrary unit)2 Theta A synthesized titania flakes Figure 3 6. XRD patterns of titania samples A ) synthesized flakes B ) calcined flakes C ) Anatase nanoparticles (Titanium(IV) oxide nanopowder, 99.9% anatase, Fisher Chemical).

PAGE 62

62 Figure 3 7. HR TEM images of titania samples (the SAD pattern as inset) A ) synthesized flakes B ) calcined flakes The diffraction rings are indexed as (1) 101 (2) 004 (3) 200 (4) 105 for anatase 0.35 nm ~5 nm 0.35 nm ~8nm A B

PAGE 63

63 528 530 532 534 536 Intensity (a.u.)Binding energy (eV) O 1s Surface OH group 0.00 0.05 0.10 0.15 0.20 0.25 Calcined flakes Synthesized flakes OH/OtotP25 Figure 3 8. XPS analysis of the titania s amples. A ) O 1s peak fitting of synthesized flakes B ) OH/Otot XPS surface ratio of the titania samples. A B

PAGE 64

64 250 300 350 400 450 500 0 10 20 30 40 50 60 70 80 90 100 % ReflectanceWavelength (nm) synthesized flakes calcined flakes 2.5 2.6 2.7 2.8 2.9 3.0 3.1 3.2 3.3 3.4 3.5 3.6 3.7 3.8 0 20 40 60 80 100 calcined flakeshv (eV) synthesized flakes Figure 3 9. UVVisible spectra of titania samples. A) Di ffuse reflectance spectra. B ) the dependence of ( h )2 on the photon energy for synthesized and calcined titania flakes A B

PAGE 65

65 Figure 3 10. Schematic diagram of integrating sphere set up in the modified method for measuring the true absorption of titania suspension. A) reference standard B ) sample [123]. 300 325 350 375 400 0 2 4 6 8 10 12 Synthesized flakes P25 Calcined flakesAbsorbance ( *102)Wavelength (nm) Figure 3 11. True absorption spectra of titania samples using the modified integration sphere method. (note: reflectivity in an integrating sphere is provided by a coating of pigment grade TiO2 or ZnO. Thus the absor bance above represents only the increased a bsorbance over the control .) A B

PAGE 66

66 -40 -30 -20 -10 0 10 20 30 40 2 3 4 5 6 7 8 9 10 11 12 pH Zeta potential (mV) Synthesized titania flakes Calcined titania flakes P25 Figure 3 12. Zeta potential versus solution pH for three titania samples. 0 0.2 0.4 0.6 0.8 1 250 350 450 550 650 750 850 Absorbance Wavelength (nm) 0 minute 30 minutes 1 hour 1.5 hours 2 hours Figure 3 13. Photocatalytic decomposition of 50 M methylene blue with synthesized titania flakes unde r UVA illumination.

PAGE 67

67 0 0.2 0.4 0.6 0.8 1 0 20 40 60 80 100 120 Irradiation time (minutes) Normalized concentration, C/C0 Pure MB synthesized flakes P25 calcined flakes 0.0 0.2 0.4 0.6 0.8 1.0 0 20 40 60 80 100 120 Irradiation Time (minutes) Normalized concentration, C/C0 Pure MB synthesized flakes P25 calcined flakes Figure 3 14. Photocatalytic decomposition of methylene blue by using synthesized and calcined flakes A ) without bubbling treatment B ) with bubbling treatment A B

PAGE 68

68 0 0.05 0.1 0.15 0.2 0.25 0.3 0.35 0 10 20 30 40 50 60 70 80 90 100 Irradiation time (minutes) -Ln(C/C0) synthesized flakes calcined flakes P25 0 0.5 1 1.5 2 2.5 3 3.5 0 10 20 30 40 50 60 70 80 90 100 Irradiation time (minutes) -Ln (C/C0) synthesized flakes calcined flakes P25 Figure 3 15. Linear transforms ln(C/C0) vs. time for methylene blue decomposition for different photocatalysts A ) without bubbling treatment B ) with bubbling treatment. Note the difference in scales of the normalized concentrations. A B

PAGE 69

69 Figure 3 16. Experimental desig n decomposition% results of synthesized flakes photocatalysis.

PAGE 70

70 Figure 3 17. The 3D response surface plot of dye decomposition% of synthesized flakes as a function of the factor A and B under different light intensity A) 2. B ) 4 C ) 6. A B C

PAGE 71

71 Figure 3 18. Model graphs of synthesized flakes as the function of A and B at constant light intensity (C = 2) A ) Contour plot of decomposition% B ) Interaction plot of response data against factor A for both levels of factor B A B 100 ppm 50 ppm

PAGE 72

72 Figure 3 19. Model graphs of synthesized flakes as the function of A and B at constant light intensity (C = 6). A ) Contour plot of decomposition% B ) Interaction plot of response data against factor A for both levels of fac tor B. A B 50 ppm 100 ppm

PAGE 73

73 Figure 3 20. The 3D response surface plot of dye decomposition% of synthesized flakes as a function of the factor B and C under different aeration. A ) 0 B ) 5 C ) 10 ft3/hr. A C B

PAGE 74

74 Figure 3 21. Model graphs of synthesized flakes as the function of B and C at constant flow rate (A = 0). A ) Contour plot of decomposition% B ) Interaction plot of response data against factor B for both levels of factor C. A B 6 lamps 2 lamps

PAGE 75

75 Figure 3 22. Model graphs of synthesized flakes as the function of B and C at constant flow rate (A = 10) A ) Contour plot of decomposition% of synthesized flakes B ) Interaction plot of response data against factor B for both levels of factor C. A B 2 lamps 6 lamps

PAGE 76

76 Figure 3 23. Experimental design decomposition% results of calcined flakes photocatalysis.

PAGE 77

77 Figure 3 24. The 3D response surface plot of dye decomposition% of calcined flakes as a function of the factor A and B under different light intensity A ) 2 B ) 4 C ) 6. A B C

PAGE 78

78 Figure 3 25. Model graphs of calcined flakes as the function of A and B at constant light intensity (C = 2). A ) Contour plot of decomposition% B ) Interaction plot of response data against factor A for both levels of factor B. A B 100 ppm 50 ppm

PAGE 79

79 Figure 3 26. Model graphs of calcined flakes as the function of A and B at constant light intensity (C = 6). A ) Contour plot of decomposition% B ) Interaction plot of response data against factor A for both levels of factor B. A B 50 ppm 100 ppm

PAGE 80

80 Figure 3 27. The 3D response surface plot of dye decomposition% of calcined flakes as a function of the factor B and C at different aeration rate. A ) 0 B ) 5 C ) 10 ft3/hr A C B

PAGE 81

81 Figure 3 28. Model graphs of calcined f lakes as the function of B and C at constant flow rate (A = 0) A ) Contour plot of decomposition% of calcined flakes B ) Interaction plot of response data against factor B for both levels of factor C. A B 2 lamps 6 lamps

PAGE 82

82 Figure 3 29. Model gra phs of calcined flakes as the function of B and C with constant flow rate (A = 10). A ) Contour plot of decomposition%. B ) Interaction plot of response data against factor B for both levels of factor C. A B 6 lamps 2 lamps

PAGE 83

83 Figure 3 30. Experimental design decomposition% res ults of P25 photocatalysis.

PAGE 84

84 Figure 3 31. The 3D response surface plot of dye decomposition% of P25 as a function of the catalyst concentration and Flow rate under different light intensity. A ) 2 B ) 4 C ) 6. A C B

PAGE 85

85 Figure 3 32. The 3 D response surface plot of dye decomposition% of P25 as a function of the catalyst concentration and light intensity at different aeration rat e. A ) 0 B ) 5 C ) 10 ft3/hr B C A

PAGE 86

86 Figure 3 33. Interaction plot of response data against fac tor B for both levels of factor C at different aeration rate for P25. A ) 0 B ) 5 C ) 10 ft3/hr (a) B 6 lamps 6 lamps 2 lamps C 2 lamps 6 lamps A 6 lam ps 2 lamps

PAGE 87

87 CHAPTER 4 TRANSITION METAL DOPING OF TITANIA PARTICIELS AS VISIBLE PHOTOCATALYSTS S ynthesis of V anadium D oped T itania F lakes Vanadium doped titania flakes wer e prepared by adding vanadium alkoxide (vanadium isopropoxi de) to the precursor used to produce titania nanoflakes. According to the desired molar ratio of Vanadium/Titanium, the appropriate amount of vanadia precursor was mixed in the oil phase and vigorously stirred by a magnetic stirrer. The oil phase mixture for making va nadium doped samples looks more yellow than the precursor for making titania flakes an intrinsic indicator that its optical properties are different. After two hours mixing, there were no individual particles or separate phase s exist ing in the final mixture which indicates vanadium isopropoxide homogeneously dissolve s in titanium isopropoxide and the hydrocarbon solvent The same synthesis procedure for making titania flakes was used for making vanadium doped flakes. After cleaning and colleting the products, synthesized vanadium doped flakes in water appeared yellowish except for the 1 at omic % doped sample (Figure 4 1) which impl ied these samples may absorb visible light and therefore have distinct color change compared to pure titania flakes. After solvent exchanging with isopropanol alcohol, the nanoflakes were dried by a superc ritical fluid drying process [ 111] The following heat treatments were performed as the same manner in the programmable electric furnace. A fter calcinations, the yellow color became more pronounced for the calcined samples compared to the non annealed samples (Figure 4 2). C haracterization of Vanadium Doped TiO2 F lakes Vanadium doping of titania has been reported as one of the most effective ways to extend the absorption edge of photocatalyst into the visible light range. Many researchers have investigated the optimum doping level of vanadium in TiO2 with diverse results [ 137141, 149 ]

PAGE 88

88 The optimum doping concentration of vanadium has generally been reported as between < 1 atomic% to 5 atomic% depending to the synthesis method and the types of vanadium precursors In order to study the effect of vanadium concentration on visible photocatalysis, vanadium doped titania flakes with three doping levels (1, 5, and 10 atomic%) were synthesized using the modified surface hydrolysis method. Like pure titania nanoflakes the significant p earle s cen c e occurred indicated a large percentage of high aspect ratio particles (Figure 4 1 and 42). Therefore, we could expect the majority of the vanadium doped samples were flakes. The low magnification optical microscopic (OM) images of series vanadium doped samples indicated flat particles with major dimension in micrometer range (Figure 43 and 44). The strong interference color could be correlated to very thin minor dimension. The flakes were further treated by calcination at 400 C for 2 hours. The surface morphology of series synthesized and calcined vanadium doped titania flakes was investigate d using SEM images, shown from Figure 45 to 48. The major diameter of the series vanadium doped flakes was found to be of the order of 110 ure 45 A to 48 A ) ), and the thickness of these flakes was approximately 40 nm (Fig ure 45 B to 48 B ). Aggregation was not apparent when comparing these treated flakes with those obtained directly from synthesis. Moreover, the thickness of the ca lcined samples did not change with heat treatment. Both the synthesized and calcined vanadium doped flakes were transparent under the SEM. Vanadium was detected by Energy Dispersive X ray Spectroscopy (EDX) coupled with the SEM images within the v anadium doped flakes (Figure 45 C and 48 C ). The results showed that vanadium was successfully doped into titanium dioxide lattice using the current synthesis method up to 10 atomic% doping level. The silicon signal in the EDX spectra was form the flat silicon wa fer substrate for depositing the flakes. Further chemical element mapping scanning of titanium, oxygen and vanadium signals are shown in Figures 45 D to 4 8 D Vanadium exist ed mostly along the fringes of the flakes instead of evenly dispersing within

PAGE 89

89 the whole particle. This is reasonable when considering the relatively low doping level of vanadium in samples Particle size distributions of the synthesized and calcined vanadium doped flakes were measured using laser d iffraction (Figure 4 9 ). The measured major dimension of the flakes was similar to the dimensions obtained from the SEM images. The particle diameter statistics for series vanadium doped flakes are shown in Table 41. Unlike undoped titania samples, vanadium doped flakes have smaller diameters after synthesis. This may be attributed to the higher hydroly sis rate of vanadium precursor compared to that of titanium alkoxide. After heat treatment, the diameter does not decrease as much as the annealed titania samples. It has been reported that ther mal stability of vanadium doped titania particles decreases as vanadium content increases [ 133,134] Crystal structure changes of the vanadium doped flakes were identified by powder XRD (Figure 4 10 A ). The XRD pattern s of the synthesized and calcined vanadium doped flakes are found to c orrespond to the anatase phase There is no indication of the presence of other titania phases such as rutile or brookite a nd there are no vanadium oxides or vanadium pentaoxides detected These results indicate that vanadium is evenly dispersed throughout the flakes most likely as a lattice substitution. Therefore, a vanadium doped flakes with single pha se of anatase structure were successfully synthesized by this modified surface hydrolysis method. A slight shift of the (101) peak position form 25.67 of both undoped and 1 at omic % vanadium doped sample to 25.35 of 10 at moic % doped sample was observed usi ng the slow scanning technique at the scannin g rate of 0.01/step (Figure 410 B ). This phenomenon suggested the larger lattice distortion caused by substitution of more vanadium ions into titanium position in titania also obs erved by other researchers [ 107,135,136] The average grain size of vanadium doped flakes was calculated by the Scherer equation (equation 31). The grain size was determined from 8.7 to 10.2 nm for calcined 1 and 10 atomic% doped flakes respectively (Table

PAGE 90

90 42). The crystallite size of vanadium doped flakes increases wit h increasing the doped vanadium concentration. The possible reason of this result is that vanadium is w ell known as a catalyst to fa cilitate the an atase phase transformation [ 106,108,133,134,137] But the grain size of vanadium doped flakes is still within the range of optimum size for photocataly sis which is about 1015 nm t herefore, excellent performance of photocatalysis might be expected using the vanadium doped flakes [ 114116] Apart from the crystallinity and grain size, specific surface area is one of the most important parameters which strongly influences the photocatalysis. Since the photocatalytic reaction mainly occurs at the surface of catalysts, materials possess ing higher surface area are usually expected to be more photocatalytic ally active. According to t he resu lts of nitrogen sorption measurement s shown in Table 4 3, there is a decrease in the specific surface area of catalysts with increasing the amount of vanadium doping. The diminution of specific surface area could be related to the raise of grain size when adding more vana dium in the synthesis, and corresponded well to the results of grain size calculation from the XR D measurements (Table 4 2). S imilar results have been demonstrated for Manganese, Iron, and Vanadium doping by other investigators [ 100,138] Degussa P25 was also used as our benchmark for visible photocatalysis comparison since some researcher s claimed that it could decompose organic species under visible light irradiation [ 67,126 ] All calcine d vanadium doped flakes have ~ 3 times higher surface area than P25 (Table 4 3). The surface potential of the series calcined vanadium doped flakes were measured using the electrophoretic technology (Brookhaven Zetaplus), as shown in t he Figure 4 1 1 A The IEPs of the series calcined vanadium doped flakes s hifted toward lower pH value s with increasing van adium concentration (Figure 4 11 B ). Di Paola et al. observed the same behavior and found the surface Bronsted acid sites formed at th e surface of the titania sample when adding appreciable amount of Mo, V, and W using the Fourier Transform Infrared Spectroscopy (FTIR) [ 139] A similar trend was previously

PAGE 91

91 observed for doped titania samples consist ing of acidic oxides (Mo and V) and suggest s that those oxides significantly increase the surface acidity by developing surface Bronsted acid sites [ 140] F r om the dye adsorbing aspect, series vanadium doped samples had more negatively charges than P25 which could favor the dye loading at surface and hence the overall degradation efficiency (Figure 4 11 B ). Therefore, both the surface properties of photocatalyst and photocatalytic efficiency could be manipulated by altering the amount of vanadium doping. The optical properties of the vanadium doped flakes were determined by taking diffuse reflectance spectrum from the UV Visible spectrometer coupled with an integrati on sphere (Figure 4 12). There is a clear red shift for calcined vanadium doped flakes when comparing to undoped titania flakes. The photoresponse extended into the visible region (400700nm) with vanadium doping. And the extension of absorption edges is attributed to the charge transfer transition from the d orbital of vanadium atom to the conduction band of titania The band gap of vanadium doped flakes could be calculated from the diffuse reflectance spectrum which is listed in Table 4 4. Comparing to previous studies of vanadium doped titania [ 94,106,141143] the same trend is shown in the shift of absorption ba nd edge with increasing amount of vanadium dopant. Furthermore, a modified method was used to accurately measure the true absorption of photocatalyst particles in water in order to mimic the realistic of photoca talytic process (Figure 413) [ 123] The range of band gap of the vanadium doped flakes is f r om 3.33 to 2.61 eV according to the amount of vanadium in samples (Table 44). Higher vanadium doping level of vanadium doped flakes tends to increase the optical absorption in the visible range. However, the absorption behavior seems to be changed when the doping level reached 10 atomic% (Figure 4 13 C ). This phenomenon may be attributed to the chemical state of vanadium cations in the titania lattice. The absorption band in the visible light regime may be correlated to the generation of impurity levels created by both V5+ and V4+ states, which will be shown in the following X -

PAGE 92

92 ray Photoelectron Spectrom eter (XPS) spectrum (Figure 4 14 ). It has been repo rted that the absorption band of V5+ and V4+ cations were shown above and below 550nm respectively in the substitutional sites in anatase phasic titania [ 108,110,144] Hence, the absorption band shifted the range of shorter wavelength when V5+ exists predominantly in the lattice site of titania ( Figure 4 14 B ). Except for the influence of optical absorption, the chemical state of vanadium cations also play s an important role in the visible light photocatalysis and will be discussed in the following section. In order to correlate the amount of dopa nt to the change of optical properties of vanadium doped flakes, the inductively coupled plasma mass spectrometry (ICP MS) was used to analyze the concentration of vanadium in the catalyst. The results of absolute molar ratio of V/Ti were slightly lower th an the nominal molar ratio which was directly calculated from the initial concentration of precursor (Table 4 4). These differences may be attributed to the error of volume measurement for initial precursor or diffe rence in hydrolysis and condensation rate between titania and vanadia precursor and loss of vanadium during washing T he chemical state of vanadium species may have a stronger influence in visible light photocatalysis [ 106,141,145,146] Therefore, detailed XPS spectra has been taken and discussed in the follow ing section. XPS Spectra of Series Vanadium Doped Flakes The core level titanium 2p spectra for series vanadium doped samples (form 0 to 10 atomic% doping level) were shown in figure 4 15 A Two distinct peaks at 458.5 and 464.2 eV were observed and recognized as Ti 2p3/2 and Ti 2p1/2 core electrons respectivel y for the pure titania sample [ 147149] When taking the spectr a for the vanadium doped samples under the same conditions (the samples were prepared on the same mass basis), these two peaks shifted slightly with increasing amount of vanadium dopant. The binding energy of the titanium core electrons moved tow ard to hig her energy about 0.40.5 eV for 110 atomic% vanadium doped

PAGE 93

93 samples. An assumption of decreased electron charge density around Ti4+ due to vanadium doping could be drawn form these positive shift in binding energy of titanium signal [ 141] Beside, an asymmetric part appearing near the Ti 2p3/2 peak especially for higher vanadium doped sample (Figure 4 15 B ), suggested a low energy shoulder corresponded to Ti3+ ion. The lower chemical state of Ti3+ ions occurs in vanadium doped samples were also reported by Robba et al. [ 150] The interaction between dopant (vanadium) and matrix (titanium dioxide) explained why lower state titanium species were forming. In order to get more understanding of this result, a series of vanadium spectra were also taken fro m different samp les (Figure 4 14). A wide range of V 2p2/3 peak from 517 to 522 eV for all samples were identified as V4+ and V5+ states respectively [ 106,146,150157] A increment of vanadium chemical state from +4 to +5 when adding more va nadium was observed (Figure 414 A ) and has also been repor ted by many other researchers [ 106,141,145,146] Therefore, the positive shifting of titanium 2p peak is reasonable because of changing of chemical state of vanadium. When the oxidation state of vanadium increased, the local electron density around the Ti4+ ion have to decrease at the same time to balance the total cha rge of material. Above is also evidence of vanadium ions incorporating into titania matrix instead of forming separate vanadia species. It should be noted that the intensity of V 2p2/3 signal did not change proportionally to the content of vanadium, which implied the discrepancy of vanadium concentration at surface and in bulk (Table 44). There are two important issues that should be addressed here: one is the formation of V4+ respected to the vanadium doping level and the other is the simultaneously posit ive shifting of titanium and vanadium signal for all vanadium doped samples. In the study of V2O5/TiO2 catalysts, Trifiro et al. observed the same trend of change of chemical state of vanadium and proposed a scenario based on reduction of V5+ during anneal ing pro cess [ 158] Hermann et al. also reported the generation of tetravalent vanadium cations after calcined at 450 C in the titania

PAGE 94

94 matrix [ 159] Both studies suggested the formation and incorporation of V4+ ions in the TiO2 lattice were related to the dehydroxylation process. The possible mechanism could be demonstrated by the following reactions. 4 4 5 3 3 ) ( 4 ) ( ) ( ) (Ti V V Ti Ti e Ti e OH OHs s s s A h igh concentration of hydroxyl group was usually generated after the water based sol gel method and therefore facilitated the above reduction reaction of pentavalent vanadium cations. However, the surface hydr oxyl group detected by the XPS spectrometer for all vanadium samples (Figure 4 16) was roughly the same because of the similar synthesis process and specific surface area (Table 4 3). When the amount of vanadium dopant increased, more and more pentavalent ions were detected by the XPS spectrometer as a result of the depletion of surface hydroxyl group. Accordingly, the V5+ signal become s stronger and stronger and predominate in the titania lattice when vanadium content is beyond 5 atomic%. As a result, lowe r chemical state of titanium species, Ti3+, should appear owing to the coexist ance of vanadium pentavalent cations and a weak shoulder that did present in all of the samples except for 1 atomic% va nadium doped flakes (Figure 415 B ). E xperiment A pparatus D ye decomposition experiments with vanadium doped samples were performed in a visible light Pyrex reactor. The light source was a 300 W Xe arc lamp (Varian Eimac Division, Light R 300) equipped with a power supply (Varian Eimac Division, model PS 300) For differentiating photocatalytic efficiency of visible and UV, a UV cutoff filter (< 400 nm) was placed between the light source and the dye solutions. 150 ml dye solutions were made with the same concentration used in the UV degradation experiments (50 and illuminated under

PAGE 95

95 visible light without adding photocatalysts as the background. A n Erlenmeyer flask contained dye solution with TiO2 particles w as placed inside the visible light reactor and continuously agitated by a magnetic stir rer House air was d elivered through a glass t ube for investigating the aeration process. The amount/volume of input air flow was controlled by a rotary flow meter. Photocatalytic Degradation of Dye Molecules under Visible Light Il lumination The photocatalytic activity of van adium doped titania flakes unde r visible light irradiation w as measured by the destruction of the dye molecules (methylene blue) in water with the above set up. The following calculations of degradation rate were obtained by the experiment results after background subtraction following the same procedure of UV degradation test s in the previous chapter, vanadium doped titania flakes with three different vanadium doping level s were added to the reactor for studying the photocatalytic efficiency under only vis ible light irradiatio n. Figure 417 compare s the photocatalytic activity of three vanadium doped samples by decomposition of the methylene blue dye solution under pure visible light illumination. Degussa P25 was also used as a reference material since some researcher claimed that it possesses some v isible photocatalytic ability [ 67,126] It was observed that calcined 1 at omic % vanadium doped tit ania flakes exhibited the highe st photocatalytic efficiency on the degradation of methylen e blue among three samples. Degussa P25 did not show any visible photocatalysis capability in this resear ch. The Langmuir Hinshelwood (L H) model was again used to analyze the kinetics of dye degradation with visible light photocatalysis [ 16,20,127131] The linear transform of degradation curve was plotted as ln( C/C0) versus irradiation time and the rate constants were calculated from the slopes of Figure 419. The rate constants are show n in Table 45. The reaction with added turbulence presented a pseudo first order reaction and much higher efficiency especially fo r the flake systems (Figure 4 18).

PAGE 96

96 From the preliminary results, the amount of vanadium in the doped samples seemed to be th e most important parameter for the visible light photocatalysis. Moreover, the vanadium concentration was also directly related to the chemical state of vanadium atom i n the doped samples (Figure 4 14). Therefore, we expected that there could be a direct r elationship between the chemical state of vanadium atom and visible photocatalytic processes. Bhattacharyya et al. proposed a possible mechanism to explain the effect of chemical state of vanadium on photocatalysis [ 141] First of all, the common mechanism of titania photocatalysis were investigated by the laser flash photolysis met hod and illustrated as the following reactions: III Ti e IV Ti O O e OH OH h e h h TiO 2 2 2 And the main electron hole pairs recombination processes were described as the followings: IV Ti III Ti h OHOH e At this moment, the nature of chemical state of vanadium ion played a role in the efficiency of the visible photocatalysis. The V4+ ions could be oxidized by the photogenerated holes and thus not only affect the charge transfer processes but also inhibit the undesired recombination reactions (equation 41). OH V OH V V h V4 5 5 4 .(41) Furthermore, the combination of above reactions leads to the formation of hydroxyl radical and then the following oxidation reaction in the common photocatalytic reaction. Accordingly, we could expect that the presen ce of V4+ ions could benefit the visible photocatalysis process. When

PAGE 97

97 the vanadium species presented predominantly as the pentavalent ions in the doped sample, some of the tetravalent titanium ions must be reduced to trivalent titanium ions to maintain the charge balance of the system. This was also consistent with the results o f XPS spectra in the Figure 4 15 B a small shoulder represented Ti3+ ions for the 10 atomic% vanadium doped samples which had the highest concentration of V5+ ions (Figure 4 14 A ). And the presence of trivalent titanium ions could promote the detrimental recombination process. Consequently, the higher the pentavalent vanadium ions presented in the sample, then lower the overall photocatalytic efficiency would be (Figure 420). Statis tic al Design of Experiments From the results of preliminary experiments (Figure 417 and 4 18), catalyst loading, and aeration process were determined to significantly affect the visible photocatalytic efficiency. Furthermore, the concentration of doped tr ansition metal seemed to be one of the most important parameters. Therefore, the statistics design of experiments was again performed by using the Design Expert software. The investigated parameters in the design of experiments were A: bubbling flow ra te ( ft3/hr ), B: mass concentration of photocatalysts (ppm), and C: atomic concentration of vanadium (atomic %). A full 23 factorial design with three center points was used to obtain the developed model and estimate d the experiment error. The detail of des ign of experiments is shown in Table 46. The results of visible photocatalysis with series vanadium doped flakes are shown as a 3D representation of design space (Figure 421) The best results for vanadium doped titania flakes could be seen using high le vel of catalyst loading, and flow rate but low level of vanadium doping The complete degradation of dye could be found in the space which indicates that 1 atomic% vanadium doped flakes have the highest photocatalytic activity among the range of

PAGE 98

98 concentrat ion investigated in this research. The software generated model for vanadium doped titania fla kes is described by equation 4 2. Dye degradation% = 50.05 +3.84A (Flow rate) +9.01B (Catalyst loading) 32.81C (Vanadium concentration) 1.44AB 6.61BC..(42 ) No doubt there is no obvious interaction between flow rate and vanadium concentration of sample and therefore no AC interaction term in this m odel. When the surface response of data is considered as the function of factor B and C; the surface was relatively flat, which suggested that there is only a small interaction between these two parameters (Figure 4 22 ). And the prompt increase occurred wh en factor C decreased, which indicated that the vanadium concentration is the most important parameter in the visible light degradation of dye solution. The amount of catalyst only contributed to limited enhancement when the vanadium doping was at the lowe r level regardless of the flow rate (Figure 4 23 and 4 24). The detailed argument of vanadium concentration was given in the previous section (Please see the preliminary part). Figure 425 shows the 3D surface response plot of series vanadium doped samples as the function of factor A and B. It is apparent that the amount of vanadium in the doped sample is the most significant parameter and inversely proportional to the dye decomposition percentage. Figure 426 A represents the contour plot with respect to t he factor A and B, and Figure 4 26 B shows the response of 1 atomic% vanadium doped sample against factor A for different level of factor B. Increasing flow rate gives moderate improvement of overall efficiency for both different levels of catalyst loading A similar trend is shown when the same plots were presented for 5 atomic% vanadium doped sample (Figure 4 27). However, the high vanadium doped sample does not follow the same track owing to the severe charge carrier recombination effect (Figure 4 28). I n short, calcined 1 atomic% vanadium doped sample always had the highest performance within the whole investigation region. From the results of characterization, vanadium doped flakes with

PAGE 99

99 three doping levels possessed very similar physical properties not only with each other but also with calcined undoped sample. Therefore, we could conclude that vanadium concentration of sample is the most important parameter for visible light photocatalysis fr om statistical perspective. T he optimum doping level of vanadi um doped flakes is identified as 1 atomic% within the investigated range of this study for vanadium doped flakes.

PAGE 100

100 Table 4 1. Particle diameter statistics for series synthesized and calcined vanadium doped flakes Sample D10 ( m) D50 ( m) D90 ( m) Median ( m) Standard Deviation ( m) 1 at omic % vanadium Synthesized flakes 2.3 12.2 24.7 10.0 10.2 Calcined flakes 1.3 7.8 16.3 6.3 6.3 5 at omic % vanadium Synthesized flakes 2.2 10.1 24.3 7.5 10.7 Calcined flakes 1.8 8.8 19.0 6.8 7.5 10 at omic % vanadium Synthesized flakes 0.3 7.5 17.3 5.7 6.6 Calcined flakes 0.2 4.2 9.7 3.3 3.6 Table 42. Grain size calculation by the Scherer equation for series calcined vanadium doped nanoflakes Sample B (degree) B (degree) d (nm) 1 at omic % vanadium doped flake s 1.04 25.63 8.7 5 at omic % vanadium doped flakes 0.93 25.51 9.2 10 at omic % vanadium doped flakes 0.80 25.35 10.2 Table 43. Physisorption measurements of P25, and series calcined vanadium doped titania nanoflakes Sample Specific surface area (m2/g) S pecific pore volume (cm3/g) Calcined 1 at omic % vanadium doped flakes 156 0.492 Calcined 5 at omic % vanadium doped flakes 152 0.487 Calcined 10 at omic % vanadium doped flakes 132 0.467 P25 49 Table 44. V/Ti molar ratio and band gap energy of series vanadium doped samples Expected V/Ti ( atomic %) V/Ti in bulk (atomic %) Band gap (eV) Expected 1 Measred 2 0 0 0 3.33 1 1 0.92 2.91 5 5 4.82 2.73 10 10 9.79 2.61

PAGE 101

101 Table 45. Pseudofirst order rate constants of methylene blue under visible photocatal ytic degradation. With bubbling treatment Sample kpseudo (104 min1) P25 3 Calcined 1 atomic % vanadium doped flakes 115 Calcined 5 atomic % vanadium doped flakes 70 Calcined 10 atomic % vanadium doped flakes 27 Table 46. The 23 factorial design use d to investigate the most important factors for visible light photocatalysis Factor 1 Factor 2 Factor 3 A: Bubbling flow rate (ft 3 /hour) B: Catalyst loading (ppm) C: Vanadium concentration (atomic%) (0) (50) (1) (0) (50) (1) (0) + ( 100 ) (1) (0) + ( 100) (1) + ( 10) (50) + ( 10) + ( 10) (50) + ( 10) + ( 10) + ( 100) + ( 10) + ( 10) + ( 100) + ( 10) *Center point: A (flow rate): 5; B (catalyst loading): 75; C (vanadium concentration): 5

PAGE 102

102 Figure 41. Photographs of synthesized va nadium doped titanium dioxide flakes A ) 1 B ) 5 C ) 10 atomic % vanadium (note the pearlescence of the samples) Figure 42. Photographs of calcined vanadium doped titanium dioxide flakes A ) 1 B ) 5 C ) 10 at omic % vanadium (note the pearlescence of the samples) A B C A B C

PAGE 103

103 Figure 43. O ptical micrographs of vanadium doped titanium dioxide flakes A ) synthesized B ) calcined 5 atomic % vanadium doped flakes Figure 44. O ptical micrographs of calcined van adium doped flakes. A ) 1. B ) 10 atomic% vanadium. A B A B

PAGE 104

104 Figure 45. SEM images of synthesized 5 at moic % vanadium doped titania nanoflakes A ) surface morphology B ) thickness C ) EDX D ) mapping A B C D

PAGE 105

105 Figure 46. SEM images of calcined 5 at moic % vanadium doped titania nanoflakes A) surface morphology B) thickness C) EDX D) mapping C D A B

PAGE 106

106 Figur e 47. SEM images of calcined 1 at omic % vanadium doped titania nanoflakes A ) surface morphology B ) thickness C ) EDX D ) mapping B A C D

PAGE 107

107 Figure 48. SEM images of calcined 10 at omic % vanadium doped tit ania nanoflakes A ) surface morphology B ) thickness C ) EDX D ) mapping B A C D

PAGE 108

108 0.01 0.1 1 10 100 0 1 2 3 4 5 Differential%Particle diameter (um) Calcined 1at% vanadium doped flakes Synthesized 1at% vanadium doped flakes 0.01 0.1 1 10 100 0 1 2 3 4 5 Synthesized 5at% vanadium doped flakes Calcined 5at% vanadium doped flakesDifferential%Particle diameter (um) 0.01 0.1 1 10 100 0 1 2 3 4 5 Synthesized 10at% vanadium doped flakes Calcined 10at% vanadium doped flakesDifferential%Particle diameter (um) Figure 49. Volume based particle size distribution for synthesized and calcined vanadium doped flakes A ) 1 B ) 5 C ) 10 atomic% vanadium B A C

PAGE 109

109 30 40 50 60 70 JCPDS 21-1272 Anatase JCPDS 44-1426 V2O5JCPDS 44-0251 VO2calcined 5mol% vanadium doped flakes synthesized 5mol% vanadium doped flakesIntensity (a.u.)2 Theta 22 24 26 28 10 at% Vanadium 5 at% Vanadium 1 at% VanadiumIntensity (aArbitrary unit)2 Theta 0 at% Vanadium Figure 410. XR D patterns of vanadium doped flakes A) 5 atomic% vanadium. B ) Anatase (101) peak for series calcined vanadium doped flakes (at slow scanning mode: 0.01/step). A B

PAGE 110

110 -40 -30 -20 -10 0 10 20 30 40 2 3 4 5 6 7 8 9 10 11 12 pH Zeta potential (mV) Calcined titania flakes (0 at%) Calcined 1 at% vanadium doped flakes Calcined 5 at% vanadium doped flakes Calcined 10 at% vanadium doped 3.0 3.5 4.0 4.5 5.0 5.5 6.0 0 1 2 3 4 5 6 7 8 9 10 Vanadium concentration (at%) IEP Figure 411. Zeta potential of the vanadium doped flakes as the function of pH A) 0, 1, 5, 10 at omic % vanadium B ) The IEP versus the amount of vanadium in the series vanadium doped flakes. A B

PAGE 111

111 250 300 350 400 450 500 550 600 650 700 0 10 20 30 40 50 60 70 80 90 100 5mol% calcined vanadium doped flakes% ReflectanceWavelength (nm) synthesized flakes calcined flakes 300 400 500 600 700 0 2 4 6 8 5mol% calcined vanadium doped flakes calcined flakesAbsorbanceWavelength (nm) synthesized flakes Figure 4 12. Diffuse reflectance spectra of vanadium doped titania flakes A) 5 atomic% vanadi um. B ) the absorbance spectra for synthesized and calcined vanadium doped titania flakes A B

PAGE 112

112 250 300 350 400 450 500 550 600 650 700 750 800 0.00 0.05 0.10 0.15 0.20 0.25 0.30 0.35 0.40 0.45 Synthesized titania flakes Synthesized 10 at% vanadium doped flakes Synthesized 5 at% vanadium doped flakes Synthesized 1 at% vanadium doped flakesAbsorbanceWavelength (nm) 250 300 350 400 450 500 550 600 650 700 750 800 0.00 0.05 0.10 0.15 0.20 0.25 0.30 0.35 0.40 0.45 AbsorbanceWavelength (nm) Calcined titania flakes Calcined 10 at% vanadium doped flakes Calcined 5 at% vanadium doped flakes Calcined 1 at% vanadium doped flakes 400 450 500 550 600 650 700 0.00 0.05 0.10 0.15 0.20 0.25 0.30 AbsorbanceWavelength (nm) Calcined titania flakes Calcined 10 at% vanadium doped flakes Calcined 5 at% vanadium doped flakes Calcined 1 at% vanadium doped flakes Figure 413. The true absorption spectra of vanadium doped flakes. A ) series synthesized vanadium doped flakes B ) series calcined vanadium doped flakes C ) photoresponse of series vanadium doped flakes in the visible range. B A C

PAGE 113

113 510 515 520 525 530 535 540 1 at% vanadium doped flakes 5 at% vanadium doped flakesV4+O 1S satelliteIntensity (a.u.)Binding energy (eV) O 1S V5+10 at% vanadium doped flakes 512 514 516 518 520 522 V4+Intensity (a.u.)Binding energy (eV)V5+V 2p2/3 Figure 4 14. XPS spectra of vanadium doped flakes. A) O 1s and V 2p3/2 peaks for series vanadium doped flakes (b) V 2p3/2 peak of 10 at omic % vanadium doped flakes A B

PAGE 114

114 454 455 456 457 458 459 460 461 462 463 464 465 466 Binding energy (eV) Intensity (a.u.) Calcined titania flakes 1 atomic% Vanadium doped flakes 5 atomic% Vanadium doped flakes 10 atomic% Vanadium doped flakes 458.5 eV 458.9 eV 458.9 eV 459.0 eV 464.2 eV 464.6 eV 464.6 eV 464.5 eV 455 455.5 456 456.5 457 457.5 458 458.5 459 459.5 460 460.5 461 Binding energy (eV) Intensity (a.u.) Ti3+ Ti4+ Figu re 4 15. XPS spectra of Ti 2p3/2 and Ti 2p1/2 peaks for vanadium doped flakes A) series vanadium doped flakes. B ) Ti 2p3/2 peak of 10 at omic % vanadium doped flakes A B

PAGE 115

115 528 529 530 531 532 533 Intensity (a.u.)Binding energy (eV) O 1s Surface OH group 0.00 0.04 0.08 0.12 0.16 Otot/OH 1 at% vanadium doped flakes 5 at% vanadium doped flakes 10 at% vanadium doped flakes Figure 4 16. XPS s pectra of O 1s peak fitting A) 1 atomic% vanadium doped flakes B ) Surface ratio of OH/Otot XPS signal for series vanadium doped flakes A B

PAGE 116

116 0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9 1 0 20 40 60 80 100 120 Visible light irradiation time (minutes) Normalized concentration, C/C0 Pure MB P25 Calcined 1 at% vanadium doped flakes Calcined 5 at% vanadium doped flakes Calcined 10 at% vanadium doped flakes Figure 417. Visible p hotocatalytic decomposition of methylene blue by using vanadium doped titania flakes with three vanadium doping levels under bubbling condition. 0 0.2 0.4 0.6 0.8 1 0 30 60 90 120 Visible light irradiation time (minutes) Normalized concentration (C/C0) Pure MB Calcined 5mol% vanadium doped flakes P25 Calcined 5mol% vnadium doped flakes (bubbling) Figure 418. Linear transforms ln(C/C0) vs. time for methylene blue decomposition for vanadium doped titania flakes with three vanadium doping levels under bubbling

PAGE 117

117 0 0.2 0.4 0.6 0.8 1 1.2 0 10 20 30 40 50 60 70 80 90 100 Visible light irradiation time (minutes) -Ln(C/C0) P25 Calcined 1at% vanadium doped flakes Calcined 5 at% vanadium doped flakes Calcined 10 at% vanadium doped flakes Figure 419. Linear transforms ln(C/C0) vs. time for methylene blue decomposition for vanadi um doped titania flakes with three vanadium doping levels under bubbling condition. ( Note the difference in scales of the normalized concentrations ). 0 2 4 6 8 10 12 0 2 4 6 8 10 12 14 0 2 4 6 8 10 12 14 k ( 10-3 min-1) V4+/V5+Vanadium dopant concentration (atomic%) Figure 420. Comparison of the ratio of V4+/V5+ and photocatalytic reaction rate for series vanadium doped flakes

PAGE 118

118 Figure 421. Experimental design visible light decomposition% results of vanadium doped titania flakes photocatalysis with three different doping levels.

PAGE 119

119 Figure 4 22. The 3D response surface plot of dye decomposition% of series calcined vanadium doped flakes as the factor A and B under different aeration rate. A ) 0. B ) 5 C ) 10 ft3/hr A B C

PAGE 120

120 Figure 423. Model graphs of series vanadium doped flakes as the function of B and C at constant aeratio n rate (A = 0) A ) Contour plot of decomposition% of series vanadium doped flakes B ) Interaction plot of response data against factor B for both levels of factor C. A B 1 atomic% V 10 atomic% V

PAGE 121

121 Figure 424. Model graphs of series vanadium doped flakes as the function of B and C at constant aeration rate (A =10) A ) Contour plot of decomposition% of series vanadium doped flakes B ) Interaction plot of response data against factor B for both levels of factor C. A B 10 atomic% V 1 atomic% V

PAGE 122

122 Figure 4 25. The 3D response surface plot of dye decomposition% of series calcined vanadium doped flakes as the factor B and C A ) 1 B ) 5 C ) 10 atomic% vanadium. A C B

PAGE 123

123 Figure 426. Model graphs of calcined 1 atomic% vanadium doped flakes as the function of A and B A ) C ontour plot of decomposition% B ) Interaction plot of response data against factor A for both levels of factor B. A B 50 ppm 100 ppm

PAGE 124

124 Figure 427. Model graphs of calcined 5 atomic% vanadium doped flakes as the function of A and B A ) Contour plot of decomposition% B ) Interaction plot of response data against factor A for both levels of factor B. A B 100 p pm 50 ppm

PAGE 125

125 CHAPTER 5 TITANIA NANOFLAKES BASED DYE SENSITIZED SOLAR CELLS Introduction Dye sensitized solar cells (DSSCs) are the member of th e group of thin film solar cells. The basic components of the DSSCs are a wide bang gap semiconductor film on the conducting glass substrate surrounding between the sensitizer and the electrolyte. In 1991 ORegan and Gratzel invented a nanoporous electrode and made the remarkable breakthrough of energy conversion efficiency up to ~ 7% [ 160] The current ly highest efficiency of this type of cell is ~ 11% [ 161] Th e advantages of DSSCs are as follows : (1) C ost effective manufacture due to the usage of low cost material compared to conventional solar cell (2) F lexible or lightweight products without extra protection are possible (3) Relatively insensitive to impurities (4) Wide range for operation (5) Nonvacuum and low temperature manufacture via continuous processes (for instance, doctor blade, spraying coating, or screen printing) [ 162] Conventional solar cells employ a p n junction to absorb light and generate electronhole pairs. At the same time, the photoexcited electrons have to separate from positive charged holes before they transfer into the outer circuit t o generate electricity. Therefore, the quality of single crystal silicon film that commonly used in the conventional product has to be extremely pure in order to overcome the undesired recombination. The high cost of making defect free silicon is the main obstacle for large scale solar energy application. DSSCs have different operational features compared to the conventional solar cell and will be discussed in the following section. Basic Principles and Components o f DSSC s Schematic diagram of the DSSC s i s shown in Figure 51 [ 163] The detail photoconversion process could be described as the following steps (Figure 5 2) [ 164] :

PAGE 126

126 1. When incident photon with suita ble energy absorbed by the dye molecules (S), an electron were excited from the highest occupied molecular orbital (HOMO) to lowest unoccupied molecular orbital (LUMO). 2. T he electronically excited dye (S*) injected an electron into the conduction band of the wide band gap semiconductor film (the most common material is TiO2) within a very short time period ( typically occurs in several picoseconds ) [ 162] 3. The electrons diffused through the semiconductor film and transported to the external circuit. 4. The liquid electrolyte (iodide complex) reduced the dye cation (S+) to the neutral state for future light absorption. 5. The oxidiz ed electrolyte (I3 -) reached to the c atalytic platinum layer at the counter electrode and then reduced back to I-. It is worthy to note that there is no permanent chemical transformation in the cell when generating electricity. Unlike the conventional solar cell, DSSCs have more common featur es with the phot osynthesis process in the natur al world. For example, the chlorophyll is a common green pigment found in many plants. It serves the same function as the sensitizer in DSSCs which absorb light and transfer energy to the positive and negative charge carriers. Therefore, the main difference between DSSCs and conventional solar cell is light absorption that occurs away from the place where charge carriers separat ed It obviously benefits the initial charge separation. More detail discussion of operation principles c an be found in other review papers [ 165167] T he E nergetic A spect of DSSC s The energetics of the DSSCs are usually described as the relationships between the HOMO/LUMO levels of the dye molecules, the conduction band of the semiconductor and redox potential of the electrolyte (Figure 5 2). In order to function the DSSCs properly, there are several requirements needed to be confirmed. First of all, the LUMO level of the dye must be sufficiently higher than the conduction band of the semiconductor for efficient electron injection. Then the redox potential of the electrolyte should be lower than the HOMO level of the dye for

PAGE 127

127 successful reduction of excited dye. Finally, the charge carrier injection process must be sufficiently faster than the de excitation step of the photon activated dye molecules in order to collect electrons. The typical current voltage (I V) cu rve of a s olar cell is shown in Figure 5 3. In order to introduce the photoenergy conversion efficiency, several important parameters will be caref ully described in the following: 1. Short circuit current ( Isc): T he current flows through in a solar cell of no resistance (or voltage across the cell is zero). It is the maximum current that could be generated and drawn from a solar cell at certain intensity of light. 2. Open circuit voltage ( Voc): In contrast, the voltage developed by a solar cell of very large external resistance (or current is zero). It is also the maximum voltage available from a cell. 3. Fill factor ( FF): T he ratio of the maximum power available along the whole I V curve divided by the product of Isc and Voc (equation 51). Higher the fill f actor, better the cell perform. oc sc M MV I V I FF .(5 1) where IM is the maximum current density; VM is the maximum voltage. 4. S olar energy conversion efficiency ( ): Any energy conversion efficiency could be defined as the ratio of the energy output to the input energy. Therefore, it could be described as the ratio of output electricity generated by the solar cell to the available solar power (equation 52). This efficiency is the most commonly used property when comparing to different s olar cells. % 100 % 100 % 100 %2 2 cm mW P FF V V cm mA I P P P Pin oc sc in M in out .(5 2) where Pout is the output energy; Pin is the input energy which is the available solar power in this case; PM is the maximum power point. E xperiment The titania pastes were prepared by adding 5 volume% of acetic acid solution as stabilizer into 15 wt% of water based titania slurry with three different samples (Degussa P25, synthesized and calcined titania flakes). In order to reduce aggregation, the above slurries were dispersed

PAGE 128

128 using the programmable ultrasonicat or (MISONIX Sonicator 3000 cup horns type) at 100 W for 15 minutes. The photoelectrodes for DSSCs were made by depositing the titania pastes on the transparent indium tin oxide (ITO) coated conducting glass (sheet using the doctor blade method. Before pasting the titania film, there were two additional cleaning steps for the ITO glass. The glass sheet was degreased in a sonication assisted acetone bath and then washed thoroughly with high purity water (nanopure) and isoproponal respectively. After drying the film in air at ambient temperature, the photoelectrode was annealed at 450 C for 1 hour and then cooled. The sintered photoanode was immersed in the ruthenium dye solution (0.3 mM) for 24 hours to completely load the sensitizer. The stained photoelectrode was then rinsed with acetonitrile to remov e the excess dye molecule, and assembled with the platinized counter elect rode made by platinum coating f r om the ion beam coater (Gatan model 681 ion beam coater). The morphology of sintered titania photoanodes and platinum coated counter electrodes were investigated by SEM. The liquid electrolyte consisted of 0.05 M iodine, 0.1 M lithium iodide, and 0.5 M tert butyl pyridine in acetonitrile wa s added to the interspacing between working and counter electrodes via capillary action. The photovoltaic performance of the DSSCs was determined by the Keithley 4200 source meter with a solar simulator ( Oriel sol 1A TH ) which provided the standard AM 1.5 illumination. P hotoelectrodes C haracterization The basic physical properties (such as particle size distribution, thickness, specific surface area etc.) of synthesized and calcined titania flakes were shown in the Chapter 3. The microstructure and thicknes s of electrodes were measured by the SEM technique. The diffuse reflectance spectra of the titania films were investigated by the UV Visible spectrometer equipped with an integrating spher e within the wavelength of 400700 nm. The dye absorption

PAGE 129

129 test was s tudied by adding 50 mg of titania samples in 20 ml 0.3 mM N719 solution at room temperature for 24 hours to mimic the real dye absorption behavior at the titania surface. After separating the solid and dye molecules by centrifugation, the concentration of residual dye solution was determined by the UV Visible spectrometer. P hotovoltaic M easurements The photocurrent voltage (I V) characteristics of the cells were measured under illumination of simulated AM 1.5 irradiation (100 mW/cm2) provide by a solar simulator (Oriel sol 1A TH) calibrated with a standard crystalline silicon solar cell. The photovoltaic performance such as short circuit current ( Isc), opencircuit voltage ( Voc), of the DSSCs were recorded with the Keithley model 4200 digital source meter ; t hen fill factor ( FF ), and energy conversion efficiency ( ) were calculated based on the above results with the corresponding area of the titania film. A ssembling DSSCs using H igh A spect R atio T itania P articles Typical photoelectrodes in the DSSCs or Gratzel cells were made by pasting a layer of titania nanopart icles on an indium tin oxide (ITO) glass or a fluorine doped ti n oxide (FTO) glass substrate [ 164167] Those nanoparticles were synthesized by a hydrothermal process in a batch manner using titanium alkoxide as titania precursor. The shape of nanoparticles is nearly spherical with the primary size of 15 nm and moderate specific surface area of about 100 m2/g. We developed a novel method to fabricate high aspect ratio titania flak es with hig her specific surface area (150 330 m2/g) and similar size in one dimension (40 nm thickness) in a continuous process. Therefore, we expected several advantages of using titania nanoflakes as an alternative materials in DSSCs. First of all, highe r specific surface area tends to absorb more dye molecules and is expected to generate more electricity under the same irradiation condition. In the mean time, the electron diffusion path does not increase too much because one dimension of flakes

PAGE 130

130 (thicknes s) still remain s in the nanometer range. Furthermore, the morphology of film has been proved to be significantly affected the electron transport in the DSSCs [ 168 ]. The most important parameter is the number of interconnect between neighboring particles in the electron diffusion path. It was found that the higher the area of overlapping, the faster the electron transportation. For a nanoparticles film, the average number of neighboring particles (or being called coordinated particles) was calcu lated to be more than two. However, there is significantly larger number of coordinated particles when using calcined titania flakes as the starting materials. Due to the smaller grain size (~ 10 nm) and higher surface area, each flakes have much higher co ntact points with the adjacent flakes compared to the P25 film (Figure 54). Therefore, the photoexcited electron has higher probability to diffuse through the calcined flakes film and then into the outer circuit owing to the much higher contact area betwe en each flake. Finally, the continuous synthesis allows us to scale up the titania flakes production to larger quantity comparing to the low productive batch process. Characterization of Titania Photoelectrodes After doctor blading and sintering, the photoelectrodes made from Degussa P25 nanoparticles usually showed some cracks on the ITO glass substrate (Figure 5 5). It was difficult to maintain the integrity of the semiconductor film when using P25 nanoparticles as titania precursor. The possible expla nat ion is related to the difficulty of nanoparticles dispersion especially in the high solid loading condition (15 wt%). If the start ing slurry for making titania film could not reach the appropriate dispersion in the main solvent (for example, water), the fo llowing agglomeration or coagulation must occur during drying and heat treatment steps and therefore film cracking could be easily observed. However, there were no cracks on the photoanodes made from cal cined titania flakes (Figure 5 6). Because of the nat ure of these high aspect ratio flakes, the better dispersion could be easily achieved under the same process

PAGE 131

131 condition. Two dimensions of these high aspect ratio particles are still in micron range. Therefore, titania flakes could be easily dispersed by th e conventional dispersion techniques such as sonication, surfactant or electrostatic stabilization. And the integrity of semiconductor film could be easily preserved even after drying and the following sintering processes. The surface morphology of sinter ed photoelectrodes made from calcined titania flakes were observed by SEM. Figure 5 7 showed the top views of the P25 based photoelectrodes. A c ontinuously cracking film with many titania chunks could be observed from the low magnification SEM images (Figu re 5 7 A and B ). The result indicated that appropriate dispersion did not achieve in the current process condition and severe shrinkage occurred during the drying and sintering steps. Under higher magnification, the P25 based photoanode consisted of 3040 nm interconnected na noparticles layer (Figure 5 7 C ). The film thickness is about 28 calculated from the cross section of the photoelectrode. In contrast, a crack free surface film with high roughness and high porosity could be observed from the top vie ws of SEM images for calcined flakes based photoelectrodes (Figure 5 9 ). Th e thickness of titania film is measured by taking the side view of titania film and determined to be about 2025 10). It is obvious that titania nanoflakes stack above each ot her and assemble to an integr ated layer (Figure 5 10 B ). The results indicated that calcined flakes films ha d much better adhesion than those made from Degussa P25. The relationship between film thickness and the photovoltaic performance of DSSCs wi ll be discussed in the later section. Smooth and continuous platinized cathode deposited by the ion beam coater was shown in Figure 511. The thickness of platinum layer was about 40 nm and the quality of the platinum film was determined by the EDX spectru m and mapping technique (Figure 511B D ). The result suggested high quality platinum

PAGE 132

132 film, which is essential to the reduction of liquid electrolyte during the photovoltaic process, deposited on the substrate. The high loading titania slurries were vigorously sonicated in the ultrasonicator for better dispersion. The subsequent change of particle size especially for flakes is important and was determined by the laser diffraction technique. It should be noted again that larger dimension of flakes after soni cating could be measured under volume distribution basis. The results of three slurries are shown in Figure 512 and the statistics data are listed in Table 5 1. According to the BET measurement s calcined flakes ha ve much higher surface area and porosity than P25 nanopart icles which have relatively low surface area (50 m2/g) and no porosity (Table 3 3). In general, more visible sensitized dye molecules could be adsorbed by the materials with higher specific surface area and porosity and therefore generated more electricity under the same illumination area. The amount of surface adsorbed dye were determined by immersing titania samples in the dye solution used in the photovoltaic measurement at room temperature for 24 hours. After solidliquid separation, the residual dye concentration was monitored using the UV Vis spectrometer and the results are shown in Table 52. It is worthy to note that calcined titania flakes have the highest coated dye molecules per unit surface area among these samples It could be attributed to the preferential adsorption of dye in the small pores at titania surface. The absorption spectrum of the N719 dye was determined by the UV Vis spectrometer (Figure 5 13). The maximum absorption wavelength of the N719 dye is shown at 538 nm and the absorption range covers the whole visible light spectrum. Typical photoanodes made from titania nanoparticles in DSSCs were optically transparent with 10 implied the loss of a portion of the visible light or sunlight due to transmittance [ 169171] Optical properties, especially light scattering effect, are essential for the

PAGE 133

133 DSSCs to enhance the light absorption capability [ 169171]. In order to i mprove the light harvesting efficiency and photon to current conversion efficiency, surface modification of titania films were attempted by several methods such as introducing of scattering centers inside or adding smaller particles on the top. The basic idea of these methods is to increase the abi lity of light scattering of the photoelectrodes. For quantifying the amount of diffusely scattered light f r om an incident beam on the titania film, the d iffuse reflectance spectroscopy were used to measure and co mpare the photoelectrodes made from P25 nanocrystal s and titania flakes (Figure 514). P25 films showed high ability of diffuse scattering especially in the range between 400 and 450 nm. However, a prompt decay in the diffuse reflection capabilities was observed while the wavelength increased from 450 to 800 nm. The weaker light scattering for P25 nanoparticles within the higher wavelength regime could be attributed to the smalle r particle size (nominally 3040 nm) compared to the wavelength of visible light. According to the classic Mie theory, the optimum diameter of light scattering center should be about half of the wavelength of incident light [ 170,171] For the range of wavelength of interest, the optimum particle size for scattering is around 200400 nm whic h is much larger than the nominal diameter of P25 nanoparticles. In contrast, gradually decrease of diffuse reflectance went toward the higher wavelength range for the films composed of synthesized flakes. Moreover, the photoelectrodes made from calcined f lakes possessed significantly higher scattering ability within the whole wavelength range under the same process condition (with similar thickness). Both flakes are micrometer sized wide and this dimension is perpendicular to the path of incident light. Therefore, stronger light scattering could be expected due to the comparable particle size of the wavelength of visible light. Furthermore, the scattering of calcined flakes exceeded that of synthesized flakes could be attributed to phase transformation from low refractive index phase (a morphous, R.I = 1.8) to

PAGE 134

134 anatase which is a high refrac tive index phase (R.I. = 2.5) [ 172,173] In summary, light absorpt ion and scattering ability of photoanodes could be improved by using titania nanoflakes as the starting materials due to the higher specific surface area and comparable size of visible light wavelength. Photovoltaic Performance of Titania Photoelectrodes T o stud y the influence of titania film thickness on the photovoltaic behavior of DSSCs, P25 and calc ined flake based photoelectrod were prepared and compared by their current densi ty voltage properties (Table 53 thick photoanodes made from calcined titania flakes showed a short circuit current density of 15.5 mA/cm2 and an power conversion efficiency of 7.4%, whereas the P25 electrodes with similar thickness reached a current density of 2.8 mA/cm2 and efficiency of 1. 3%. The typical I V curves of P25 and calcined flakes electrodes with different thickness are demonstrated in Figure 5 15. At the comparable film thickness, the photocurrent densities and voltage s generated from calcined flake based cells are higher than those of P25based cells (Figure 5 16) For illustrating the effect of thickness on the photovoltaic proper ties of DSSCs, the dependence of Jsc and Voc on the thickness of photoanodes for P25 and calcined flakes are shown respectively in Figure 5 17. P25 anodes show no significant change on Jsc when increasing the thickness of film compar ing to higher increment for calcined flakes (Figure 5 16 A ). It is reasonable that efficiency and short circuit current of cell increase continuousl y with titania film thickness [ 174] Because the higher surface area associated with thicker layer leads to more absorbed dyes at titania surface and thus improve the probability o f photoexcited electrons that inject into the conduction band of TiO2. Therefore, photogen erated current density increas es correspondingly. In contrast, Voc for both materials show the same trend (decrease) respect to th e film thickness (Figure 5 16 B ). Th e opencircuit voltage usually decreased with increasing thickness of film

PAGE 135

135 because of higher recombination occurring at more available surface sites [ 174] Furthermore, the energy conversion efficiency of DSSCs increased with film thickness could be explained by the competition between short circuit curr ent, opencircuit voltag e and fill factor The raise of short circuit current not only compensate s but also overcome s the loss of opencircuit which results in improved efficiency of DSSCs. In short, the photon energy conversion efficiency of the DSSCs was improved from 1.2% to 7.4% (about 5 times improvement) when replacing P25 nanoparticles with calcined titania flakes in the photoelectrodes on the same basis of film thickness.

PAGE 136

136 Table 51. Particle diameter statistics for P25, synthesized and calcined fl akes after ultrasonication. Sample D10 ( m) D50 ( m) D90 ( m) Me di an ( m) Standard Deviation ( m) P25 3.9 13.3 19.8 9.8 20.9 Calcined flakes 0.2 3.2 7.5 2.2 6.5 Synthesized flakes 0.5 5.6 14.2 3.8 8.4 Table 52. Dye loading and physisorption measur ements of P25, synthesized and calcined titania nanoflakes Sample Dye loading Specific surface area (m2/g) Specific pore volume (cm3/g) Average pore diameter (nm) Synthesized flakes 19.3 323 Calcined flakes 15.3 151 0.511 7.2 P25 3.1 49 Table 53. Photovoltaic properties of the dye sensitized solar cells assembled by using anodes made from P25 nanoparticles and calcined titania flakes of different thickness Sample P25 Calcined titania flakes 10 15 25 10 15 25 Voltage (V oc mV) 776 747 737 810 793 785 Current density (I sc mA/cm 2 ) 2 .6 2.7 2.8 11.9 13.5 15.5 Fill factor (FF) 0.59 0.58 0.59 0.61 0.62 0.61 Efficiency (%) 1.1 1.2 1. 3 5.9 6.6 7.4

PAGE 137

137 Figure 51. Schematic structure of DSSCs [163] Figure 52. E ne rgy levels in the typical DSSCs [164]

PAGE 138

138 Figure 53. Typical I V cur ves of the solar cells (The fill factor, FF, could be calculated by the ratio of the area A to the area B)

PAGE 139

139 Figure 54. Schematic cross sectional view of electron transport through titania layer. A ) Degussa P25 nanoparticles B ) calcined titania na noflakes (two dimensional nanostructures). A B

PAGE 140

140 Figure 55. O ptical micrographs of sintered titania photoelectordes made from De gussa P25 nanoparticles A ) 5 B ) 50 magnifications Figure 56. O pt ical micrographs of sintered titania photoelectordes made from calcined titania flakes A ) 5 B ) 50 magnifications A B (a) ( b )

PAGE 141

141 Figure 57. SEM micrographs of sintered photoelectrodes made from P25 nanoparticles under different magni fication s. Figure 58. Cross section of sintered photoelectrodes made from P25 nanoparticles under different magnification s. A B C A B

PAGE 142

142 Figure 59. SEM micrographs of sintered photoelectrodes made from c alcined titania flakes under different magnification s. Figure 510. Cross section of sintered photoelectrodes made from calcined titania flakes under different magnification s. A B B A

PAGE 143

143 Figure 5 11. SEM images of platinum counterelectrode A ) surface morphology B ) thickness C ) EDX. D ) mapping. A B C D

PAGE 144

144 0.1 1 10 100 1000 0 1 2 3 4 5 6 7 Calcined flakes Synthesized flakesDifferential volume%Particle diameter (um) P25 Figure 512. Particle size distribution of titania slurries for P25, synthesized and calcined titania flakes after ultrasonica tion at 100W for 15 minutes 400 500 600 700 800 0.0 0.5 1.0 1.5 2.0 AbsorbanceWavelength (nm) Figure 513. Absorption spectrum of 0.3 mM Ru complex dye solution (also as known N 719 dye).

PAGE 145

145 400 500 600 700 800 0 10 20 30 40 50 Calcined titania flakes Synthesized titania flakes P25 ITO glassDiffuse Reflection (%)Wavelength (nm) Figure 514. Diffuse reflectance spectra of the titania films prepared from Degussa P25 nanoparticles, syntheszied and calcine d titania flakes of similar thickness

PAGE 146

146 0 100 200 300 400 500 600 700 800 900 0 2 4 6 8 10 12 14 16 10 um 15 umCurrent density ( mA/cm2)Voltage (mV) 25 um 0 100 200 300 400 500 600 700 800 900 0.0 0.5 1.0 1.5 2.0 2.5 3.0 15 um 10 umCurrent density ( mA/cm2)Voltage (mV) 25 um Figure 515. I V curve s of dye snesitized solar cells with different film thickenesses A ) calcined titania flakes B ) P25 B A

PAGE 147

147 0 100 200 300 400 500 600 700 800 900 0 2 4 6 8 10 12 14 16 P25 25 umCurrent density ( mA/cm2)Voltage (mV) Calcined titania flakes 25 um Figure 516. I V characteristics of DSSCs made from P25 and calcined flakes of similar thickness under AM 1.5 simulated sunlight irradiation.

PAGE 148

148 10 15 20 25 0 5 10 15 20 P25Jsc ( mA/cm2)Anode thickness ( um) Calcined titania flakes 10 15 20 25 725 750 775 800 825 P25 Calcined titania flakesVoc (mV)Anode thickness (um) Figure 517. Comparsion of the Jsc and Voc of P25 and calcined flakes cells as a function of film thickness.

PAGE 149

149 CHAPTER 6 SUMMARY, CONCLUSIONS AND FUTURE WORK Summary The first goal of th is study was to develop a new type of photocatalyst able to decompose more than 99% organic molecules in waste water within 2 hours. To accomplish this goal, high aspect ratio titania nanoflakes were successfully synthesized using the surface control hydro lysis method. Many characterization techniques were used to investigate both physical and chemical properties of titania flakes. When using calcined titania flakes as photocataly sts completely photodegradation of the stimulated waste water (50 M methylene blue solution) was achieved under UVA irradiation within 2 hours. At the same time, one of the most successful commercial products, Degussa P25, only showed 87% degradation under the same operational condition. In order to optimize the photocatal y tic per formance of titania flakes, statistics design of experiment were used to develop a predictive model within the investigated range. The results showed that the optimum condition of photocatalysis for calcined flakes was: median catalyst loading (75 ppm), hi gher aeration rate (10 ft3/hr), and higher light intensity (mW/cm2). Although titania flakes showed higher photocatalytic efficiency, higher energy required to activate this new materials due to larger band gap resulted from the quantum confinement effect. The second goal of this study was to develop an energy efficient photocatalyst without scarifying the high photocatalytic activity. In order to take the advantage of the most abundant energy in the world, solar power, narrowing the band gap of titania nanoflakes was the important approach In this study, vanadium doped titania flakes were successfully fabricated using the same synthesis method. The decrease of band gap energy for vanadium doped samples was confirmed by the UV Visible spectrometer. Accordin gly, vanadium concentration, aeration rate, and catalyst loading were used as three main parameters in the statistics design analysis. It was

PAGE 150

150 found that 1 atomic % vanadium doped flakes have the highest photocatalytic activity under visible light illumination. Completely decomposition of the organic species in the same simulated polluted water was achieved using 1 atomic % vanadium doped flakes under pure visible light illumination within 2 hours; on the other hand, P25 nanoparticles did not show significan t visible photocatalytic activity. Dye sensitized solar cell is a promising technique to meet the clean energy demand in the future. However, the further improvement on the efficiency of this device is crucial for competing with the current electricity gen eration technology. Modified surface hydrolysis synthesized titania nanoflakes possessed many unique properties such as high aspect ratio, high specific surface area, slower recombination rate, mass production possibility etc. Therefore, we expected enhan ced photovoltaic performance for DSSCs when replacing the typical semiconductor layers, titania nanoparticles, with calcined titania nanoflakes. The results showed that 5 fold improvements compared to P25 nanoparticles based cell was achieved when assembli ng DSSCs with calcined flakes. Although the efficiency of titania flakes based cell has not been optimized, the preliminary results showed encouraging efficiency (7.4%). Conclusions The main conclusions in this study include: O ver 99% of methylene blue solution (50 M) was degraded by the high aspect ratio calcined titania nanoflakes u nder UVA irradiation within 2 hours whereas not completely decomposition of dye solutions was achieved using P25 nanoparticles as photocatalysts under the same process condition. The s ame concentration of dye solutions were de composed more than 99% under visible light illumination within 2 hours when using 1 atomic % vanadium doped flakes as remediation agent while almost no visible photocatalytic ability was shown for P25 nanoparticles 7.4% of photon energy conversion efficiency of calcined flakes based DSSC which was 5 times improvement compared to P25 based cell was accomplished.

PAGE 151

151 Future Work The following were several recommendations for future work: Form the optical measurements o f titania samples, titania nanoflakes have highly light scattering ability compared to commercial products. Therefore, they have highly potential to be pigment materials. Further experiments such as hiding power, viscosity, and rheology need to be carried out in order to investigate the optimum formula for making paint using titania flakes. Thickness of titania flakes could be manipulated by adjusting the ratio of titania precursor to organic solvent. The desired thickness of titania flakes could be made ac cordingly to specific applications. Further DOE could be very useful to investigate the influence of thickness on certain applications. For example, one of the main parameters in this study, photocatalytic activity, could depend on the thickness of flakes The real properties of visible photocatalysis using vanadium doped flakes could be investigated by carrying out the experiment under sun light illumination. And the optimum concentration of vanadium could be measured using the DOE technique. The optimiz ation of the formula for DSSCs using flakes based materials need to be conducted by using the DOE. And larger area of DSSCs should be made and tested in order to extend the application to commercial scale.

PAGE 152

152 APPENDIX A ANALYSIS VARIANCE F OR UV PHOTOCATALYSIS USING SYNTHESIZE D FLAKES A: Flow rate B: Catalyst loading C: Light intensity Response: Decomposition% Table A 1. Analysis of variance table Source Sum of Squares df Mean Square F Value p value Prob > F Model 4318.20 5 863.64 18.94 0.0069 significan t A 253.13 1 253.13 5.55 0.0780 B 2443.01 1 2443.01 53.58 0.0019 C 852.84 1 852.84 18.70 0.0124 AB 224.72 1 224.72 4.93 0.0906 BC 544.50 1 544.50 11.94 0.0259 Curvature 1.46 1 1.46 0.032 0.8669 not significant Residual 182.39 4 45.60 Lack o f Fit 143.70 2 71.85 3.71 0.2121 not significant Pure Error 38.69 2 19.34 Cor Total 4502.04 10 The Model F value of 18.94 implies the model is significant. There is only a 0.69% chance that a "Model F Value" this la rge could occur due to noise. Values of "Prob > F" less than 0.0500 indicat e model terms are significant. In this case B, C, BC are significant model terms. Values greater than 0.1000 indicate the mode l terms are not significant. If there are many insignificant model terms (not counti ng those required to support hierarchy), model reduction may improve your model. The "Curvature F value" of 0.03 implies the curvature (as measured by difference between the average of the center points and the average of the factorial points) in the desi gn space is not sign ificant relative to the noise. There is a 86.69% chance that a "Curvature F value" this la rge could occur due to noise. The "Lack of Fit F value" of 3.71 implies the Lack of Fit is not si gnificant relative to the pure error. There is a 21.21% chance that a "Lack of Fit F val ue" this large could occur due to noise. Non significant lack of fit is go od -we want the model to fit.

PAGE 153

153 Table A 2. Statistical results The "Pred R Squared" of 0.4699 is not as close to the "Adj R Squared" of 0.9088 as one might normally expect. This may indicate a large block effect or a possible problem with your model a nd/or data. Things to consider are model reduction, respons e tranformation, outliers, etc "Adeq Precision" measu res the signal to noise ratio. A rati o greater than 4 is desirable. Your ratio of 12.410 indicates an adequate signal. This model can be used to navigate the design space. Std. Dev. 6.75 R Squared 0.9595 Mean 42.67 Adj R Squared 0.9088 C.V. % 15 .82 Pred R Squared 0.4699 PRESS 2386.32 Adeq Precision 12.410

PAGE 154

154 APPENDIX B ANALYSIS VARIANCE FOR UV PHOTOCATALYSIS USING CALCINED FLAKE S A: Flow rate B: Catalyst loading C: Light intensity Response: Decomposition% Table B 1. Analysis of variance table Source Sum of Squares df Mean Squ are F Value p value Prob > F Model 4099.21 5 819.84 454.42 < 0.00 01 significant A 37.84 1 37.84 20.98 0.0 102 B 1693.62 1 1693.62 938.73 < 0.00 01 C 2271.38 1 2271.38 1258.96 < 0.00 01 AB 9.24 1 9.24 5.12 0.0 863 BC 87.12 1 87.12 48.29 0.0 023 Curv ature 10.88 1 10.88 6.03 0. 0700 not significant Residual 7.22 4 1.80 Lack of Fit 5.69 2 2.85 3.73 0.2115 not significant Pure Error 1.53 2 0.76 Cor Total 4117.31 10 The Model F value of 454.42 implies the model is significant. There is only a 0.01% chance that a "Model F Value" this la rge could occur due to noise. In this case A, B, C, BC are significant model terms. The "Curvature F value" of 6.03 implies there is curvature in the design space. There is only a 7.00% chance that a "Curvatur e F value" this large could occur due to noise. The "Lack of Fit F value" of 3.73 implies the Lack of Fit is not si gnificant relative to the pure error. There is a 21.15% chance that a "Lack of Fit F va lue" this large could occur due to noise. Table B 2. Statistical results The "Pred R Squared" of 0.9771 is in reasonable agreement with t he "Adj R Squared" of 0.9960. "Adeq Precision" measu res the signal to noise ratio. Your ratio of 62.669 indicates an adequate signal. This model can be used to navigate the design space. Std. Dev. 1.34 R Squared 0.9982 Mean 61.49 Adj R Squared 0.9 960 C.V. % 2.18 Pred R Squared 0. 9771 PRESS 94.47 Adeq Precision 62.669

PAGE 155

155 APPENDIX C ANALYSIS VARIANCE FOR UV PHOTOCATALYSIS USING P25 NANOPARTICLES A: Flow rate B: Cataly st loading C: Light intensity Response: Decomposition% Table C 1. Analysis of variance table Source Sum of Squares df Mean Square F Value p value Prob > F Model 4569.44 5 913.89 1370.83 < 0.00 01 significant A 34.44 1 34.44 5 1.67 0.0 020 B 1601.78 1 1601.78 2402.67 < 0.00 01 C 2895.61 1 2895.61 4343.41 < 0.00 01 AB 32.80 1 32.80 49.21 0.0 022 BC 4.80 1 4.80 7.21 0.0 550 Curvature 62.84 1 62.84 94.26 0. 0006 significant Residual 2.67 4 0.67 Lack of Fit 1.94 2 0.97 2.67 0.2 725 not significant P ure Error 0.73 2 0.36 Cor Total 4 634.95 10 The Model F value of 1370.83 implies the model is significant. There is only a 0.01% chance that a "Model F Value" this la rge could occur due to noise. Values of "Prob > F" less than 0.0500 indicate model terms are significant. In this case A, B, C, AB are significant model terms. Values greater than 0.1000 indicate the mode l terms are not significant. The "Curvature F value" of 94.26 implies there is signifi cant curvature in the design space. There i s only a 0.06% chance that a "Curva ture F value" this large could occur due to noise. The "Lack of Fit F value" of 2.67 implies the Lack of Fit is not si gnificant relative to the pure error. There is a 27.25% chance that a "Lack of Fit F val ue" this large could occur due to noise. Table C 2. Statistical results The "Pred R Squared" of 0.9930 is in reasonable agreement with the "Adj R Squared" of 0.9987. "Adeq Precision" measures the signal to noise ratio. Your ratio of 108.239 indicates an adequate signal. This model can be used to navigate the design space. Std. Dev. 0.82 R Squared 0.9994 Mean 51.56 Adj R Squared 0.9 987 C.V. % 1.58 Pred R Squared 0. 9930 PRESS 32.67 Adeq Precision 108.239

PAGE 156

156 APPENDIX D ANALYSIS VARIANCE FOR VISIBLE PHOTOCATALYSIS USING VANADIUM DOPED FLAKES A: Flow rate B: Catalyst loading C: Vanadium concentration Response: Decomposition% Table D 1. Analysis of variance table Source Sum of Squares df Mean Square F Value p value Prob > F Model 9972.65 5 1954.53 96.26 < 0.00 01 significant A 117.81 1 117.81 5 .80 0.0 610 B 649.80 1 649.80 32.00 0.0024 C 8638.70 1 8638.70 4 25.47 < 0.00 01 AB 16.53 1 16.53 0.81 0. 4082 BC 349.80 1 349.80 17.23 0.0 089 Residual 101.52 5 20.30 Lack of Fit 1 00.51 3 33.50 66.56 0. 0148 not sig nificant Pure Error 1.01 2 0. 50 Cor Total 9874.17 10 The Model F value of 96.26 implies the model is significant. There is only a 0.01% chance that a "Model F Value" this la rge could occur due to noise. In this case B, C, BC are significant mode l terms. Table D 2. Statistical results The "Pred R Squared" of 0.9012 is in reasonable agreement with t he "Adj R Squared" of 0.9794. "Adeq Precision" measu res the signal to noise ratio. A ratio great er than 4 is desirable. Your ratio of 27.438 indicates an adequate signal. This model can be used to navigate the design space. Std. Dev. 4.51 R Squared 0.9897 Mean 51.05 Adj R Squared 0.9 794 C.V. % 8.83 Pred R Squared 0. 9012 PRESS 975.37 Adeq Precision 27.438

PAGE 157

157 LIST OF REFERENCES [ 1] Gleick, P. H. The Worlds Water 2000 2001; Island Press, 2000. [ 2] WHO Global Water Supply and Sanitation Assessment 2000 Report WHO/UNICEF Joint Monitoting Programme for Water Supply and Sanitation, World Health Organization and United Nations Children Fund, 2000. [ 3] Gogate, P. R.; Pandit, A. B. Advances in Environmental Research 2004, 8, 501. [ 4] PeraTitus, M.; Garca Molina, V.; Baos, M. A.; Gimnez, J.; Esplugas, S. Applied Catalysis B: Environmental 2004, 47, 219. [ 5] Bartlett, A. A. Am. J. Phys. 1986, 54, 398. [ 6] Mann, M. E.; Bradley, R. S.; Hughes, M. K. Nature 1998, 392, 779. [ 7] Hoffert, M. I.; Caldeira, K.; Benford, G.; Criswell, D. R.; Green, C.; Herzog, H.; Jain, A. K.; Kheshgi, H. S.; Lackner, K. S.; Lewis, J. S.; Lightfoot H. D.; Manheimer, W.; Mankins, J. C.; Mauel, M. E.; Perkins, L. J.; Schlesinger, M. E.; Volk, T.; Wigley, T. M. L. Science 2002, 298, 981. [ 8] Hoffert, M. I.; Caldeira, K.; Jain, A. K.; Haites, E. F.; Harvey, L. D. D.; Potter, S. D.; Schlesinger, M. E.; Schneider, S. H.; Watts, R. G.; Wigley, T. M. L.; Wuebbles, D. J. Nature 1998, 395, 881. [ 9] Kamat, P. V. The Journal of Physical Chemistry C 2007, 111, 2834. [ 10] Hubbert, M. K. Science 1949, 109, 103. [ 11] Hubbert, M. K. Am. J. Phys. 1981, 49, 1007. [ 12] Deffeyes, K. S. Hubbert's Peak: The Impending World Oil Shortage ; Princeton University Press, 2001. [ 13] Fujishima, A.; Honda, K. Nature 1972, 238, 37. [ 14] Clark, R. J. H. Comprehensive Inorganic Chemistry 1973; Vol. 3. [ 15] Kirk, R. E.; Othmer, D. F. Encyclopedia of Chemical Technology ; John Wiley and Sons, 1963. [ 16] Fox, M. A.; Dulay, M. T. Chemical Reviews 1993, 93, 341. [ 17] Hoffmann, M. R.; Martin, S. T.; Choi, W.; Bahnemann, D. W. Chemical Reviews 1995, 95, 69. [ 18] Kamat, P. V. Chemical Reviews 1993, 93, 267.

PAGE 158

158 [ 19] Linsebigler, A. L.; Lu, G.; Yates, J. T. Chemical Reviews 1995, 95, 735. [ 20] Zhu, C.; Wang, L.; Kong, L.; Yang, X.; Wang, L.; Zheng, S.; Chen, F.; MaiZhi, F.; Zong, H. Chemosphere 2000, 41 303. [ 21] Martin, S. T.; Herrmann, H.; Choi, W.; Hoffmann, M. R. Journal of the Chemical Society, Faraday Transactions 1994, 90, 3315. [ 22] Martin, S. T.; Herrmann, H.; Hoffmann, M. R. Journal of the Chemical Society, Faraday Transactions 1994, 90, 3323. [ 23] Muscat, J.; Swamy, V.; Harr ison, N. M. Physical Review B 2002, 65, 224112. [ 24] Mo, S.D.; Ching, W. Y. Physical Review B 1995, 51, 13023. [ 25] Li, J. G.; Ishigaki, T.; Sun, X. The Journal of Physical Chemistry C 2007, 111, 4969. [ 26] Elizardo, K. In Pollution Engineering 1991, p 106. [ 27] Carraway, E. R.; Hoffman, A. J.; Hoffmann, M. R. Environmental Science & Technology 1994, 28, 786. [ 28] Ishibashi, K. i.; Fujishima, A.; Watanabe, T.; Hashimoto, K. Journal of Photochemistry and Photobiology A: Chemistry 2000, 134, 139. [ 29] Kormann, C.; Bahnemann, D. W.; Hoffmann, M. R. Environmental Science & Technology 1991, 25, 494. [ 30] Stafford, U.; Gray, K. A.; Kamat, P. V. Heterogeneous Chem. Rev. 1996, 3, 77. [ 31] Fox, M. A.; Dulay, M. T. Journal of Photochemistry and Photobiology A : Chemistry 1996, 98, 91. [ 32] Rothenberger, G.; Moser, J.; Graetzel, M.; Serpone, N.; Sharma, D. K. Journal of the American Chemical Society 1985, 107, 8054. [ 33] Chen, D.; Ray, A. K. Applied Catalysis B: Environmental 1999, 23, 143. [ 34] Crittenden, J C.; Liu, J.; Hand, D. W.; Perran, D. L. Water Res. 1997, 31, 429. [ 35] Kudo, A.; Hijii, S. Chem. Lett. 1999, 28, 1103. [ 36] Okamoto, K.; Yamamoto, Y.; Tanaka, H.; Itaya, A. Bull. Chem. Soc. Jpn. 1985, 58, 2023. [ 37] Yu, Y.; Leu, R. M.; Lee, K. C. Wate r Res. 1996, 30 1169.

PAGE 159

159 [ 38] Curc, D.; Gimnez, J.; Addardak, A.; Cervera March, S.; Esplugas, S. Catalysis Today 2002, 76, 177. [ 39] Stylidi, M.; Kondarides, D. I.; Verykios, X. E. Applied Catalysis B: Environmental 2004, 47, 189. [ 40] Bahemann, D. W.; Kormann, C.; Hoffmann, M. R. J. Phys. Chem. 1987, 91 3789. [ 41] Faust, B. C.; Hoffmann, M. R.; Bahnemann, D. W. J. Phys. Chem. 1989, 93, 6371. [ 42] Kormann, C.; Bahnemann, D. W.; Hoffmann, M. R. The Journal of Physical Chemistry 1988, 92, 5196. [ 43] R iddick, T. M. Zeta Meter Manual.; Zeta Meter Inc.: New York., 1968. [ 44] Zeta Potential of Colloids in Water and Waste Water, A. S. D.. Materials, A. S. f. T. a., Ed. 1985. [ 45] Gaya, U. I.; Abdullah, A. H. Journal of Photochemistry and Photobiology C: Photochemistry Reviews 2008, 9, 1. [ 46] Evgenidou, E.; Fytianos, K.; Poulios, I. Applied Catalysis B: Environmental 2005, 59, 81. [ 47] Fu, X.; Clark, L. A.; Zeltner, W. A.; Anderson, M. A. Journal of Photochemistry and Photobiology A: Chemistry 1996, 97, 181. [ 48] Muradov, N. Z.; T Raissi, A.; Muzzey, D.; Painter, C. R.; Kemme, M. R. Solar Energy 1996, 56, 445. [ 49] Tunesi, S.; Anderson, M. A. Chemosphere 1987, 16 1447. [ 50] Fujishima, A.; Rao, T. N.; Tryk, D. A. Journal of Photochemistry and Photobiology C: Photochemistry Reviews 2000, 1, 1. [ 51] Heller, A. Accounts of Chemical Research 1995, 28, 503. [ 52] Ljjna, S. Nature 1991, 354, 56. [ 53] Ollis, D. F.; Pelizzetti, E.; Serpone, N. Environ. Sci. Technol. 1991, 25, 1522. [ 54] Mills, A.; Le Hunte, S. Journal of Photochemistry and Photobiology A: Chemistry 1997, 108, 1. [ 55] Berger, T.; Sterrer, M.; Diwald, O.; Knzinger, E. ChemPhysChem 2005, 6, 2104.

PAGE 160

160 [ 56] Berger, T.; Sterrer, M.; Diwald, O.; Knzinger, E.; Panayotov, D.; Thompson, T. L.; Yates, J. T. The Journal of Physical Chemistry B 2005, 109, 6061. [ 57] Nakaoka, Y.; Nosaka, Y. Journal of Photochemistry and Photobiology A: Chemistry 1997, 110, 299. [ 58] Gerischer, H.; Heller, A. J. Elecrtochem. Soc. 1992, 139, 113. [ 59] Gerischer, H.; Helle r, H. J.Phys. Chem. 1991, 95, 5261. [ 60] Wang, C. M.; Heller, A.; Gerischer, H. J. Am. Chem. Soc. 1992, 114, 5230. [ 61] Jun, Y.; et al. Journal of Applied Physics 2006, 99 084908. [ 62] Muggli, D. S.; Falconer, J. L. Journal of Catalysis 1999, 187, 230. [ 63] Muggli, D. S.; Falconer, J. L. Journal of Catalysis 2000, 191, 318. [ 64] Yoshida, K.; Nanbara, T.; Yamasaki, J.; Tanaka, N. J. Appl. Phys. 2006, 99 084908. [ 65] Yoshida, K.; Yamasaki, J.; Tanaka, N. Appl. Phys. Lett. 2004, 84, 2542. [ 66] Carp, O .; Huisman, C. L.; Reller, A. Progress in Solid State Chemistry 2004, 32, 33. [ 67] Ohno, T.; Akiyoshi, M.; Umebayashi, T.; Asai, K.; Mitsui, T.; Matsumura, M. Applied Catalysis A: General 2004, 265, 115. [ 68] Ohno, T.; Tanigawa, F.; Fujihara, K.; Izumi, S.; Matsumura, M. Journal of Photochemistry and Photobiology A: Chemistry 1999, 127, 107. [ 69] Patrick, B.; Kamat, P. V. The Journal of Physical Chemistry 1992, 96, 1423. [ 70] Bahnemann, D. W. Israel Journal of Chemistry 1993, 33, 115. [ 71] Anpo, M.; Shima, T.; Kodama, S.; Kubokawa, Y. The Journal of Physical Chemistry 1987, 91, 4305. [ 72] Hagfeldt, A.; Graetzel, M. Chemical Reviews 1995, 95, 49. [ 73] Henglein, A. Berichte der Bunsen Gesellschaft Physical Chemistry 1997, 101, 1562. [ 74] Weller, H.; Eychmuller, A. Advances in Photochemistry ; John Wiley and Sons. Inc., 1995; Vol. 20. [ 75] Hoffman, A. J.; Carraway, E. R.; Hoffmann, M. R. Environmental Science & Technology 1994, 28, 776.

PAGE 161

161 [ 76] Nedeljkovic, J. M.; Nenadovic, M. T.; Micic, O. I.; Nozik, A. J. The Journal of Physical Chemistry 1986, 90, 12. [ 77] Nosaka, Y.; Ohta, N.; Miyama, H. The Journal of Physical Chemistry 1990, 94, 3752. [ 78] Hoffman, A. J.; Mills, G.; Yee, H.; Hoffmann, M. R. The Journal of Physical Chemistry 1992, 96, 5546. [ 79] Lepore, G. P.; Langford, C. H.; Vchov, J.; Vlcek, A. Journal of Photochemistry and Photobiology A: Chemistry 1993, 75, 67. [ 80] Morales, A. M.; Lieber, C. M. Science 1998, 279, 208. [ 81] Pan, Z. W.; Dai, Z. R.; Wang, Z. L. Science 2001, 291, 1947. [ 82] Dai, Z. R.; Pan, Z. W.; Wang, Z. L. Journal of the American Chemical Society 2002, 124, 8673. (83) Shen, G.; Bando, Y.; Liu, B.; Golberg, D.; Lee, C. J. Advanced Functional Materials 2006, 16, 410. [ 84] Wang, Z. L. J. Phys. Condens. Matter. 2004, 16, R829. [ 85] Ye, C.; Fang, X.; Hao, Y.; Teng, X.; Zhang, L. The Journal of Physical Chemistry B 2005, 109, 19758. [ 86] Burda, C.; Chen, X.; Narayanan, R.; El Sayed, M. A. Chem. Rev., 2005, 105, 1025. [ 87] Amano, F.; Nogami, K.; Abe, R.; Ohtani, B. The Journal of Physical Chemistry C 2008, 112, 9320. [ 88] Jang, E.; Won, J. H.; Hwang, S. J.; Choy, J. H. Advanced Materials 2006, 18, 3309. [ 89] Jang, J. S.; Yu, C.J.; Choi, S. H.; Ji, S. M.; Kim, E. S.; Lee, J. S. Journal of Catalysis 2008, 254, 144. [ 90] Ye, C.; Bando, Y.; Shen, G.; Golberg, D. The Journal of Physical Chemistry B 2006, 110, 15146. [ 91] Serpone, N.; Lawless, D.; Khairutdinov, R. The Journal of Physical Chemistry 1995, 99, 16646. [ 92] Anpo, M.; Aikawa, N.; Kubokawa, Y.; Che, M.; Louis, C.; Giamello, E. The Journal of Physical Chemistry 1985, 89, 5017. [ 93] Sasaki, T.; Nakano, S.; Yamauchi, S.; Watanabe, M. Chemistry of Materials 1997, 9, 602.

PAGE 162

162 [ 94] Li, L.; Liu, C.y.; Liu, Y. Materials Chemistry and Physics 2009, 113, 551. [ 95] T akahashi, R.; Sato, S.; Sodesawa, T.; Arai, K.; Yabuki, M. J. Catal. 2005, 229, 24. [ 96] Asahi, R.; Morikawa, T.; Ohwaki, T.; Aoki, K.; Taga, Y. Science 2001, 293, 269. [ 97] Irie, H.; Watanabe, Y.; Hashimoto, K. Chem. Lett. 2003, 32, 772. [ 98] Nakahira, A.; Konishi, K.; Yokota, K.; Honma, T.; Aritani, H.; Tanaka, K. J. Ceram. Soc. Jpn. 2006, 114, 46. [ 99] Umebayashi, T.; Yamaki, T.; Itoh, H.; Asai, K. Journal of Physics and Chemistry of Solids 2002, 63, 1909. [ 100] Nagaveni, K.; Hegde, M. S.; Madras, G The Journal of Physical Chemistry B 2004, 108, 20204. [ 101] Anpo, M.; Takeuchi, M. Journal of Catalysis 216 505. [ 102] Anpo, M.; Takeuchi, M.; Ikeue, K.; Dohshi, S. Current Opinion in Solid State and Materials Science 2002, 6, 381. [ 103] Briand, L. E.; Bonetto, R. D.; Sanchez, M. A.; Thomas, H. J. Catalysis Today 1996, 32, 205. [ 104] Yamashita, H.; Anpo, M. Catalysis Surveys from Asia 2004, 8, 35. [ 105] Yamashita, H.; Harada, M.; Misaka, J.; Takeuchi, M.; Ikeue, K.; Anpo, M. Journal of Photochemist ry and Photobiology A: Chemistry 2002, 148, 257. [ 106] Gu, D. E.; Yang, B.C.; Hu, Y.D. Catalysis Letters 2007, 118, 254. [ 107] Martin, S. T.; Morrison, C. L.; Hoffmann, M. R. The Journal of Physical Chemistry 1994, 98, 13695. [ 108] Wu, J. C. S.; Chen, C. H. Journal of Photochemistry and Photobiology A: Chemistry 2004, 163, 509. [ 109] Kiosek, S.; Raftery, D. J. Phys. Chem 2001, 105, 2815. [ 110] Zhao, Z.; Yamada, Y.; Teng, Y.; Ueda, A.; Nakagawa, K.; Kobayashi, T. Journal of Catalysis 2000, 190, 215. [ 111] Tedeschi, S.; Stevens, N.; Cepeda, D.; Lee, Y.Y.; Powers, K.; Ranade, M.; El Shall, H. Powder Technology 2009, 191, 188.

PAGE 163

163 [ 112] Maile, F. J.; Pfaff, G.; Reynders, P. Progress in Organic Coatings 2005, 54, 150. [ 113] Pfaff, G.; Reynders, P. Chemical Reviews 1999, 99, 1963. [ 114] Inagaki, M.; Imai, T.; Yoshikawa, T.; Tryba, B. Applied Catalysis B: Environmental 2004, 51, 247. [ 115] Zhang, Z.; Wang, C.C.; Zakaria, R.; Ying, J. Y. The Journal of Physical Chemistry B 1998, 102, 10871. [ 116] Baiju, K. V.; Shukla, S.; Sandhya, K. S.; James, J.; Warrier, K. G. K. The Journal of Physical Chemistry C 2007, 111, 7612. [ 117] Sasaki, T.; Watanabe, M. The Journal of Physical Chemistry B 1997, 101, 10159. [ 118] Kanta, A.; Sedev, R.; Ralston, J. Langmuir 2005, 21, 2400. [ 119] Parfitt, G. D. Pure Appl. Chem. 1976, 48, 415. [ 120] Crist, B. C. Handbook of Monochromatic XPS Spectra: The Elements and Native Oxides ; Wiley, 2000. [ 121] Parfitt, G. D. Progress in Surface and Membrane Science ; Academic Press, 1976. [ 122] Kim, G. M.; Lee, S. M.; Michler, G. H.; Roggendorf, H.; Gosele, U.; Knez, M. Chemistry of Materials 2008, 20, 3085. [ 123] Sun, L.; Bolton, J. R. The Journal of Physical Chemistry 1996, 100, 4127. [ 124] Lachheb, H.; Puzenat, E.; Houas, A.; Ksibi, M.; Elaloui, E.; Guillard, C.; Herrmann, J.M. Applied Catalysis B: Environmental 2002, 39, 75. [ 125] Ohno, T.; Sarukawa, K.; Tokieda, K.; Matsumura, M. J. of Catal. 2001, 203 82. [ 126] Hurum, D. C.; Agrios, A. G.; Gray, K. A.; Rajh, T.; Thurnauer, M. C The Journal of Physical Chemistry B 2003, 107, 4545. [ 127] Aguedach, A.; Brosillon, S.; Morvan, J.; Lhadi, E. K. Applied Catalysis B: Environmental 2005, 57, 55. [ 128] De Heredia, J. B.; Torregrosa, J.; Dominguez, J. R.; Peres, J. A. Journal of Hazardous Materials 2001, 83, 255. [ 129] Kusvuran, E.; Gulnaz, O.; Irmak, S.; Atanur, O. M.; Ibrahim Yavuz, H.; Erbatur, O. Journal of Hazardous Materials 2004, 109, 85.

PAGE 164

164 [ 130] Tanaka, K.; Padermpole, K.; Hisanaga, T. Water Research 2000, 34, 327. [ 131] Tang, W Z.; Huren, A. Chemosphere 1995, 31, 4157. [ 132] Houas, A.; Lachheb, H.; Ksibi, M.; Elaloui, E.; Guillard, C.; Herrmann, J.M. Applied Catalysis B: Environmental 2001, 31, 145. [ 133] Madia, G.; Elsener, M.; Koebel, M.; Raimondi, F.; Wokaun, A. Applied C atalysis B Environmental 2002, 39, 181. [ 134] Reddy, B. M.; Ganesh, I.; Reddy, E. P. The Journal of Physical Chemistry B 1997, 101, 1769. [ 135] Fern'ndez Garc'a, M.; Mart'nez Arias, A.; Hanson, J. C.; Rodriguez, J. A. Chemical Reviews 2004, 104, 4063. [ 136] Kemp, T. J.; McIntyre, R. A. Polymer Degradation and Stability 2006, 91, 165. [ 137] Sene, J. J.; Zeltner, W. A.; Anderson, M. A. The Journal of Physical Chemistry B 2003, 107, 1597. [ 138] Karvinen, S. M. Industrial & Engineering Chemistry Research 2003, 42, 1035. [139] Di Paola, A.; Marc, G.; Palmisano, L.; Schiavello, M.; Uosaki, K.; Ikeda, S.; Ohtani, B. The Journal of Physical Chemistry B 2001, 106, 637. [ 140] Marci, G.; Palmisano, L.; Sclafani, A.; Venezia, A. M.; Campostrini, R.; Carturan, G.; Martin, C.; Rives, V.; Solana, G. Journal of the Chemical Society, Faraday Transactions 1996, 92, 819. [ 141] Bhattacharyya, K.; Varma, S.; Tripathi, A. K.; Bharadwaj, S. R.; Tyagi, A. K. The Journal of Physical Chemistry C 2008, 112, 19102. [ 142] Forzat ti, P.; Tronconi, E.; Busca, G.; Tittarelli, P. Catalysis Today 1987, 1, 209. [ 143] Ivanova, T.; Harizanova, A.; Surtchev, M.; Nenova, Z. Solar Energy Materials and Solar Cells 2003, 76, 591. [ 144] Luca, V.; Thomson, S.; F. Howe, R. Journal of the Chemic al Society, Faraday Transactions 1997, 93, 2195. [ 145] Depero, L. E.; Bonzi, P.; Musci, M.; Casale, C. Journal of Solid State Chemistry 1994, 111, 247. [ 146] Iketani, K.; Sun, R.D.; Toki, M.; Hirota, K.; Yamaguchi, O. Materials Science and Engineering B 2004, 108, 187.

PAGE 165

165 [ 147] Kumar, P. M.; Badrinarayanan, S.; Sastry, M. Thin Solid Films 2000, 358, 122. [ 148] Sleigh, C.; Pijpers, A. P.; Jaspers, A.; Coussens, B.; Meier, R. J. Journal of Electron Spectroscopy and Related Phenomena 1996, 77, 41. [ 149] Stak heev, A. Y.; Shpiro, E. S.; Apijok, J. The Journal of Physical Chemistry 1993, 97, 5668. [ 150] Robba, D.; Ori, D. M.; Sangalli, P.; Chiarello, G.; Depero, L. E.; Parmigiani, F. Surface Science 1997, 380, 311. [ 151] Casaletto, M. P.; Kaciulis, S.; Lisi, L .; Mattogno, G.; Mezzi, A.; Patrono, P.; Ruoppolo, G. Applied Catalysis A: General 2001, 218, 129. [ 152] Kobayashi, M.; Miyaji, K.; Saitoh, M.; Hiramoto, T. In Semiconductor Device Research Symposium, 2005 International 2005, p 173. [ 153] Lars, S.; Ander sson, T. Catalysis Letters 1990, 7, 351. [ 154] Mizushima, K.; Tanaka, M.; Asai, A.; Iida, S.; Goodenough, J. B. Journal of Physics and Chemistry of Solids 1979 40, 1129. [ 155] Moshfegh, A. Z.; Ignatiev, A. Catalysis Letters 1990, 4, 113. [ 156] Sawatzky G. A.; Post, D. Physical Review B 1979 20, 1546. [ 157] Wong, G. S.; Concepcion, M. R.; Vohs, J. M. Surface Science 2003, 526, 211. [ 158] Trifir, F. Catalysis Today 1998, 41, 21. [ 159] Herrmann, J.M.; Disdier, J.; Deo, G.; E. Wachs, I. Journal of the Chemical Society, Faraday Transactions 1997, 93, 1655. [ 160] O'Regan, B.; Gratzel, M. Nature 1991, 353, 737. [ 161] Chiba, Y.; Islam, A.; Watanabe, Y.; Komyia, R.; Koide, N.; Han, L. Japanese Journal of Applied Physics 2006, 45 L638. [ 162] Jose, R.; Thavasi, V.; Ramakrishna, S. J. Am. Ceram. Soc. 2009, 92, 289. [ 163] Snaith, H.; Schmidt Mende, L. Advanced Materials 2007, 19, 3187. [ 164] Lee, K. E.; Charbonneau, C.; Shan, G.; Demopoulos, G. P.; Gauvin, R. JOM 2009, 61, 52. [ 165] Grtzel, M. Progress i n Photovoltaics: Research and Applications 2000, 8, 171.

PAGE 166

166 [ 166] Grtzel, M. Progress in Photovoltaics: Research and Applications 2006, 14, 429. [167] Hagfeldt, A.; Gratzel, M. Accounts of Chemical Res earch 2000, 33, 269. [ 168] Frank, J.; Nikos, K.; Jao van de, L. Coordination Chemistry Reviews 2004, 248, 1165. [ 169] Chen, D.; Huang, F.; Cheng, Y. B.; Caruso, R. A. Advanced Materials 2009, 21, 2206. [ 170] Vargas, W. E.; Niklasson, G. A. Solar Energy Materials and Solar Cells 2001, 69, 147. [ 171] Yang, L .; Lin, Y.; Jia, J.; Xiao, X.; Li, X.; Zhou, X. Journal of Power Sources 2008, 182, 370. [ 172] Jiang, X.; Herricks, T.; Xia, Y. Advanced Materials 2003, 15, 1205. [ 173] Zhang, M.; Lin, G.; Dong, C.; Wen, L. Surface and Coatings Technology 2007, 201, 7252. [ 174] Ito, S.; Zakeeruddin, S.; Humphry Baker, R.; Liska, P.; Charvet, R.; Comte, P.; Nazeeruddin, M.; Pchy, P.; Takata, M.; Miura, H.; Uchida, S.; Grtzel, M. Advanced Materials 2006, 18, 1202.

PAGE 167

167 BIOGRAPHICAL SKETCH Yang Yao Lee was born in Taipei Ci ty, Taiwan in 1977. He graduated from the Tatung in May 2000 and enrolled in the National Taiwan University for the master degree in m aterials s cience and e ngineering. He served the military service in 2002 and was discharged in 2004. He began the graduate studies at University of Florida in 2006 and joined Dr. Kevin Power's group in 2007. He received his Ph.D. from the Department of Materials Science and Engineering at University of Florida in the fall of 2010.