<%BANNER%>

Improvements in Photocatalytic Oxidation

Permanent Link: http://ufdc.ufl.edu/UFE0022087/00001

Material Information

Title: Improvements in Photocatalytic Oxidation Photocatalyst Composition and Reactor Design
Physical Description: 1 online resource (120 p.)
Language: english
Publisher: University of Florida
Place of Publication: Gainesville, Fla.
Publication Date: 2008

Subjects

Subjects / Keywords: dioxide, graywater, magnetic, oxidation, photocatalyst, photocatalytic, recovery, silica, sio2, space, surface, tio2, titanium, water, zno
Environmental Engineering Sciences -- Dissertations, Academic -- UF
Genre: Environmental Engineering Sciences thesis, Ph.D.
bibliography   ( marcgt )
theses   ( marcgt )
government publication (state, provincial, terriorial, dependent)   ( marcgt )
born-digital   ( sobekcm )
Electronic Thesis or Dissertation

Notes

Abstract: Advanced oxidation processes including UV/H2O2, ozone, and photocatalysis have demonstrated the ability to solve complex environmental problems. These processes derive their power to oxidize recalcitrant organic compounds and microorganisms from powerful hydroxyl radicals. For the case of photocatalysis, hydroxyl radicals are generated when a wide-bandgap semiconductor like titanium dioxide (TiO2) is exposed to ultraviolet light of sufficient energy to generate an electron-hole pair. A favorable reaction occurs when the electron reacts with oxygen to form a superoxide radical and the hole reacts with water or a hydroxide ion to form a hydroxyl radical. Oxidation of organic compounds by hydroxyl radicals can lead to complete mineralization (i.e., conversion to carbon dioxide, water and mineral acids). First photocatalysis, it has the potential for complete conversion of a compound to non-toxic inorganic products. This is in contrast to other technologies that simply transfer the compounds to another phase, like distillation or adsorption. Second, the oxidant is generated in-situ by a catalyst so storage or purchase of a possibly hazardous and expensive oxidant is not necessary. Photocatalytic oxidation of organic contaminants has received much attention in academic literature, but commercial implementation has been slow due to several factors, including the difficulty in scaling up photocatalytic processes for water treatment from a laboratory scale to an industrial scale. One particular difficulty is reactor containment of highly photoactive nanoparticulate titanium dioxide, which is popular due to its low-cost and stability relative to other photocatalysts. Nano TiO2 is putatively nontoxic, however, this property is a current topic of active research. A solution proposed by Purifics, a Canadian engineering company, utilizes a patented separation device utilizing a backpressure cycled membrane and automated system (Photo-Cat?). It was imperative to examine the performance and successful implementation of this technology and compare this to the alternative approaches. Another popular approach has been to immobilize the nano-sized titanium dioxide, which will likely decrease the mass transfer of contaminants to the photocatalyst with respect to slurry, but avoid the separation difficulty. A more effective approach, and that used in this research, was to immobilize the titanium dioxide on magnetic core particles that could be both agitated and contained using a two-component magnetic field. This provided increased mass transfer seen in slurry-based systems and provided a potential option for avoiding photocatalyst separation altogether. A magnetically agitated photocatalytic reactor (MAPR) has been developed and assessed for oxidation of aqueous organic contaminants. While using phenol and organic red dye to simulate waste components is consistent with past literature and important to explore, it ignores factors that may hinder a photocatalytic performance including competitive adsorption of various organic compounds and their oxidation byproducts and the demonstrated detrimental effect of inorganic compounds such as ammonium bicarbonate on photocatalytic oxidation. To assess these effects, this work looked at photocatalytic oxidation of air evaporation subsystem (AES) ersatz water while modifying the photocatalyst mass magnetic field current and frequency to find the optimal conditions. After performing these preliminary experiments, several proposals were formulated to improve the performance of the MAPR involving both the photocatalyst synthesis and the reactor design.
General Note: In the series University of Florida Digital Collections.
General Note: Includes vita.
Bibliography: Includes bibliographical references.
Source of Description: Description based on online resource; title from PDF title page.
Source of Description: This bibliographic record is available under the Creative Commons CC0 public domain dedication. The University of Florida Libraries, as creator of this bibliographic record, has waived all rights to it worldwide under copyright law, including all related and neighboring rights, to the extent allowed by law.
Thesis: Thesis (Ph.D.)--University of Florida, 2008.
Local: Adviser: Mazyck, David W.
Electronic Access: RESTRICTED TO UF STUDENTS, STAFF, FACULTY, AND ON-CAMPUS USE UNTIL 2010-05-31

Record Information

Source Institution: UFRGP
Rights Management: Applicable rights reserved.
Classification: lcc - LD1780 2008
System ID: UFE0022087:00001

Permanent Link: http://ufdc.ufl.edu/UFE0022087/00001

Material Information

Title: Improvements in Photocatalytic Oxidation Photocatalyst Composition and Reactor Design
Physical Description: 1 online resource (120 p.)
Language: english
Publisher: University of Florida
Place of Publication: Gainesville, Fla.
Publication Date: 2008

Subjects

Subjects / Keywords: dioxide, graywater, magnetic, oxidation, photocatalyst, photocatalytic, recovery, silica, sio2, space, surface, tio2, titanium, water, zno
Environmental Engineering Sciences -- Dissertations, Academic -- UF
Genre: Environmental Engineering Sciences thesis, Ph.D.
bibliography   ( marcgt )
theses   ( marcgt )
government publication (state, provincial, terriorial, dependent)   ( marcgt )
born-digital   ( sobekcm )
Electronic Thesis or Dissertation

Notes

Abstract: Advanced oxidation processes including UV/H2O2, ozone, and photocatalysis have demonstrated the ability to solve complex environmental problems. These processes derive their power to oxidize recalcitrant organic compounds and microorganisms from powerful hydroxyl radicals. For the case of photocatalysis, hydroxyl radicals are generated when a wide-bandgap semiconductor like titanium dioxide (TiO2) is exposed to ultraviolet light of sufficient energy to generate an electron-hole pair. A favorable reaction occurs when the electron reacts with oxygen to form a superoxide radical and the hole reacts with water or a hydroxide ion to form a hydroxyl radical. Oxidation of organic compounds by hydroxyl radicals can lead to complete mineralization (i.e., conversion to carbon dioxide, water and mineral acids). First photocatalysis, it has the potential for complete conversion of a compound to non-toxic inorganic products. This is in contrast to other technologies that simply transfer the compounds to another phase, like distillation or adsorption. Second, the oxidant is generated in-situ by a catalyst so storage or purchase of a possibly hazardous and expensive oxidant is not necessary. Photocatalytic oxidation of organic contaminants has received much attention in academic literature, but commercial implementation has been slow due to several factors, including the difficulty in scaling up photocatalytic processes for water treatment from a laboratory scale to an industrial scale. One particular difficulty is reactor containment of highly photoactive nanoparticulate titanium dioxide, which is popular due to its low-cost and stability relative to other photocatalysts. Nano TiO2 is putatively nontoxic, however, this property is a current topic of active research. A solution proposed by Purifics, a Canadian engineering company, utilizes a patented separation device utilizing a backpressure cycled membrane and automated system (Photo-Cat?). It was imperative to examine the performance and successful implementation of this technology and compare this to the alternative approaches. Another popular approach has been to immobilize the nano-sized titanium dioxide, which will likely decrease the mass transfer of contaminants to the photocatalyst with respect to slurry, but avoid the separation difficulty. A more effective approach, and that used in this research, was to immobilize the titanium dioxide on magnetic core particles that could be both agitated and contained using a two-component magnetic field. This provided increased mass transfer seen in slurry-based systems and provided a potential option for avoiding photocatalyst separation altogether. A magnetically agitated photocatalytic reactor (MAPR) has been developed and assessed for oxidation of aqueous organic contaminants. While using phenol and organic red dye to simulate waste components is consistent with past literature and important to explore, it ignores factors that may hinder a photocatalytic performance including competitive adsorption of various organic compounds and their oxidation byproducts and the demonstrated detrimental effect of inorganic compounds such as ammonium bicarbonate on photocatalytic oxidation. To assess these effects, this work looked at photocatalytic oxidation of air evaporation subsystem (AES) ersatz water while modifying the photocatalyst mass magnetic field current and frequency to find the optimal conditions. After performing these preliminary experiments, several proposals were formulated to improve the performance of the MAPR involving both the photocatalyst synthesis and the reactor design.
General Note: In the series University of Florida Digital Collections.
General Note: Includes vita.
Bibliography: Includes bibliographical references.
Source of Description: Description based on online resource; title from PDF title page.
Source of Description: This bibliographic record is available under the Creative Commons CC0 public domain dedication. The University of Florida Libraries, as creator of this bibliographic record, has waived all rights to it worldwide under copyright law, including all related and neighboring rights, to the extent allowed by law.
Thesis: Thesis (Ph.D.)--University of Florida, 2008.
Local: Adviser: Mazyck, David W.
Electronic Access: RESTRICTED TO UF STUDENTS, STAFF, FACULTY, AND ON-CAMPUS USE UNTIL 2010-05-31

Record Information

Source Institution: UFRGP
Rights Management: Applicable rights reserved.
Classification: lcc - LD1780 2008
System ID: UFE0022087:00001


This item has the following downloads:


Full Text

PAGE 1

IMPROVEMENTS IN PHOTOCATALYTIC OXIDATION: PHOTOCATALYST COMPOSITION AND REACTOR DESIGN By WILLIAM L. KOSTEDT, IV A DISSERTATION PRESENTED TO THE GRADUATE SCHOOL OF THE UNIVERSITY OF FLOR IDA IN PARTIAL FULFILLMENT OF THE REQUIREMENTS FOR THE DEGREE OF DOCTOR OF PHILOSOPHY UNIVERSITY OF FLORIDA 2008 1

PAGE 2

2008 William L. Kostedt, IV 2

PAGE 3

To my mother 3

PAGE 4

ACKNOWLEDGMENTS First, I thank my parents and brother for th eir support, guidance, and love. I would not have conquered the life challenges I have without them. I thank my fiance Gretchen Simpkins for making every day more enjoyable. In additi on, I thank my friends all over the country, especially Daniel Brown for allowing me to shar e lifes triumphs and tragedies. I would like to thank my research advisor, Dr. David Mazyck, for providing support and facilitating my personal growth and contributions. I have been blesse d with a wonderful research group who have supported my research efforts tremendously incl uding Dr. Ameena Khan, Heather Byrne, Vivek Shyamasundar, Jennifer Stokke, Dr. Morgana Bach, Alec Gruss, and Matt Joiner in addition to the many undergraduate students who have contri buted to my research. I thank Gill Brubaker, Gary Scheiffele and everyone else at the Particle Engineering Research Center for facilitating a greater understanding of characteri zation techniques and their theo retical basis. Finally, I would like to thank my committee members including Dr. Paul Chadik, Dr. Ben Koopman, Dr. Brij Moudgil, Dr. John Sansalone, and Dr. Chang-Yu Wu for their efforts to help me improve the impact of my work and for f acilitating my personal growth. 4

PAGE 5

TABLE OF CONTENTS page ACKNOWLEDGMENTS...............................................................................................................4 LIST OF TABLES................................................................................................................. ..........8 LIST OF FIGURES.........................................................................................................................9 ABSTRACT...................................................................................................................................12 CHAPTER 1 INTRODUCTION................................................................................................................. .14 2 LITERATURE REVIEW.......................................................................................................22 Magnetism: Source and Properties.........................................................................................22 Photocatalysis and Motion......................................................................................................25 Applications of a Magnetic Field Gradient for Particulate Systems......................................26 Magnetic Photocatalyst Synthesis..........................................................................................29 Dye Removal Using a ZnO-TiO2 Composite.........................................................................35 3 MATERIALS AND METHODS...........................................................................................37 Common Methods..................................................................................................................37 Magnetic Photocatalys t Synthesis Method......................................................................37 Silica coating............................................................................................................37 Titania coating..........................................................................................................37 Incorporation of a Permanent Magnet a nd Influence of Ball Milling on MAPR Performance........................................................................................................................38 Experimental Setup.........................................................................................................38 Photodegradation of Phenol and Procion Red MX-5B Dye............................................39 Enhanced Photocatalyst Synthesis an d Reactor Parameter Exploration................................40 Photocatalytic Test Stand................................................................................................40 Gas Chromatograph/nDOC Analysis of Phenol..............................................................43 Magnetic Photocatalyst Characterization a nd Reactor Performance Using AES Ersatz Water...................................................................................................................................43 Characterization of the Magnetic Photocatalyst..............................................................43 Photocatalytic Test Stand................................................................................................44 Analysis of Photocatalytic Oxidation..............................................................................44 Photocat Experiments Using Graywater Forumula................................................................45 Simulated NASA Graywater Preparation........................................................................45 Biological Water Processor Effluent Preparation............................................................45 Photocatalyst/UV Lamps.................................................................................................47 5

PAGE 6

Supervisory Control and Data Acquis ition Interface, Control Systems, and Programmed Operation Modes....................................................................................48 Effluent Sampling and Testing........................................................................................48 Photocatalyst Comparison Us ing the Photocat Reactor.........................................................48 Biological Water Processor and Air Ev aporation System Effluent Ersatz Preparation...................................................................................................................48 Photocatalyst/UV lamps..................................................................................................49 Scanning Electron Microscopy (SEM)............................................................................49 Effluent Sampling and Testing........................................................................................49 Impact of Heat Treatment and Com position of ZnO-TiO2 Nanoparticles for Photocatalytic Oxidation of an Azo Dye............................................................................50 Materials..........................................................................................................................50 Synthesis..........................................................................................................................50 Characterization...............................................................................................................51 Dye Photocatalytic Oxidation..........................................................................................51 A High Surface Area Magnetic Photocatal yst with Controlled Pore Size.............................52 High Surface Area Magnetic Photocatalyst Synthesis....................................................52 Characterization...............................................................................................................53 Assessment of Photocatalytic Oxidation.........................................................................53 4 INCORPORATION OF A PERMANENT MAGNET AND INFLUENCE OF BALL MILLING ON MAPR PERFORMANCE..............................................................................55 Phenol Photodegradation........................................................................................................55 Ball-Milling Impact on Catalyst Performance........................................................................56 ESM Estimation................................................................................................................. .....58 5 ENHANCED PHOTOCATALYST SYNT HESIS AND REACTOR PARAMETER EXPLORATION....................................................................................................................60 Catalyst Synthesis............................................................................................................. ......60 Phenol Photolysis....................................................................................................................60 Phenol Adsorption.............................................................................................................. ....61 Phenol and nDOC: Effect of Lamp Number and a Permanent Magnet.................................61 Catalyst Mass Optimization....................................................................................................6 3 Solenoid Frequency Optimization..........................................................................................64 Solenoid Current Optimization...............................................................................................65 6 MAGNETIC PHOTOCATALYST CHARACTERIZATION AND REACTOR PERFORMANCE USING AES ERSATZ WATER..............................................................67 Photocatalyst Characterization...............................................................................................67 Magnetic Field and Catalyst Mass Impact on Ex tent of Photocatalytic Oxidation of AES Ersatz......................................................................................................................... ..........71 6

PAGE 7

7 PHOTOCAT EXPERIMENTS US ING GRAYWATER FORUMULA...............................75 Biological Water Processor Effluent......................................................................................75 Graywater Testing.............................................................................................................. ....78 Size Requirement Estimation.................................................................................................79 8 PHOTOCATALYST COMPARISON US ING THE PHOTOCAT REACTOR...................80 Composite Characterization....................................................................................................8 0 Biological Water Processor Effluent......................................................................................84 9 IMPACT OF HEAT TREATMENT AND COMPOSITION OF ZnO-TiO2 NANOPARTICLES FOR PH OTOCATALYTIC OXIDATION OF AN AZO DYE...........90 Impact of Composition.......................................................................................................... .90 Impact of Adsorption..............................................................................................................91 Influence of Zeta Potential......................................................................................................91 Specific Surface Area.............................................................................................................94 X-ray Diffraction....................................................................................................................96 Diffuse Reflectance............................................................................................................ ....98 10 A HIGH SURFACE AREA MAGNETIC PH OTOCATALYST WITH CONTROLLED PORE SIZE...................................................................................................................... .....100 Characterization....................................................................................................................100 Effect of Mass Loading........................................................................................................105 Effect of Preparation Technique on Ph enol and Non-Purgeable Organic Carbon Removal............................................................................................................................106 Effect of Repeated Use.........................................................................................................108 11 CONCLUSIONS................................................................................................................. .110 Contributions to Science.......................................................................................................110 Future Work..........................................................................................................................110 LIST OF REFERENCES.............................................................................................................112 BIOGRAPHICAL SKETCH.......................................................................................................120 7

PAGE 8

LIST OF TABLES Table page 3-1. Simulated graywater formula............................................................................................... ..46 3-2. Biological water processor formula........................................................................................46 4-1. Equivalent systems mass (kg) for internatio nal space station baseli ne and photocatalytic systems........................................................................................................................ .......59 8

PAGE 9

LIST OF FIGURES Figure page 2-1. Hysteresis loop of ferromagnetic materials...........................................................................24 2-2. Comparison of magnetic photocatalyst synthesis procedures...............................................31 3-1. Magnetically agitated photocatalytic reactor.........................................................................39 3-2. Two-dimensional drawing of quartz reactor.........................................................................41 3-3. Three-dimensional dr awing of quartz reactor.......................................................................41 3-4. Alternating magnetic field generator..................................................................................... 42 3-5. Supervisory control and data acquisition schematic of photocat..........................................47 4-1. Phenol photodegradati on in a quartz reactor.........................................................................56 4-2. Phenol concentration decrease vs. frequency for lower catalyst loading..............................56 4-3. Normalized color decrease by catalyst sy nthesized using barium ferrite ball milled...........57 4-4. Scanning electron microscope images of ca talyst prepared using different ball-milling procedures..........................................................................................................................58 4-5. Energy dispersive x-ray spectroscopy spect ra of catalyst prepared using different ballmilling procedures............................................................................................................. 58 5-1. Comparison of number of UV lamps and presence of permanent magnet for degradation of phenol........................................................................................................62 5-2. Non-purgeable dissolved organic carbon concentration as a f unction of number of lamps and presence of permanent magnet in the MAPR...................................................62 5-3. Catalyst mass effect on phenol photocatalysis in the MAPR................................................64 5-4. Solenoid signal frequency effect on phenol photocatalysis in the MAPR............................65 5-5. Solenoid signal current effect on photocatalysis...................................................................66 6-1. Scanning electron microscope image of magnetic photocatalyst..........................................69 6-2. Scanning electron microscope image of magnetic photocatalyst..........................................70 6-3. X-ray diffraction pattern of synthesized titanium dioxide.....................................................70 6-4. Normalized concentration decrease vs. oscillating magnetic field frequency.......................72 9

PAGE 10

6-5. Normalized concentration decrease vs. catalyst mass...........................................................73 6-6. Normalized concentration decrease vs. oscillating magnetic field current...........................74 7-1. Biological water processo r effluent nDOC vs. kWh/m3........................................................76 7-2. Biological water processo r effluent nDOC vs. kWh/m3 (natural log scale)..........................76 7-3. Biological water processo r effluent nDOC vs. kWh/m3 to 3500 min (natural log scale).....77 7-4. Low-flow nDOC data for simulated graywater without bicarbonate....................................77 7-5. Low-flow nDOC data for simulated graywater with bicarbonate.........................................77 7-6. Low-flow nDOC data for biological water processor efflue nt with bicarbonate..................78 8-1. Scanning electron microscope imag e of 45-90 m silica-titania composite photocatalyst surface.......................................................................................................... 81 8-2: Secondary electron image and backsca tter complement of 4590 m silica-titania composite photocatalyst surface........................................................................................83 8-3. Energy dispersive x-ray spectroscopy of 45-90 m silica-titania composite photocatalyst surface.......................................................................................................... 83 8-4. Energy dispersive x-ray spectroscopy of 45-90 m silica-titania composite photocatalyst surface.......................................................................................................... 83 8-5. Biological water processor effluent normalized nDOC vs. time...........................................84 8-6. Biological water processor effluent pseudo-first order rate models......................................85 8-7. Biological water processor effluent UV-254 vs. time...........................................................86 8-8. Biological water processor effluent normalized nDOC vs. time...........................................87 8-9. Biological water processor efflue nt pseudo-first order rate models......................................87 8-10. Biological water proce ssor effluent UV-254 vs. time.........................................................87 8-11. Air evaporation system normalized nDOC vs. time............................................................88 8-12. Air evaporation system ps eudo-first order rate models.......................................................89 9-1. Comparison of photocatalytic oxidation fo r calcined and hydrothermally prepared ZnO-TiO2 photocatalyst of varying composition..............................................................92 9-2. Comparison of color changes for calcine d and hydrothermally prepared ZnO-TiO2 photocatalyst of varying composition................................................................................92 10

PAGE 11

9-3. Comparison of procion red MX-5B adsorption on ZnO-TiO2 photocatalyst prepared by calcination and hydrothermal heat treatment.....................................................................93 9-4. Comparison of zeta potential for a ZnO-TiO2 photocatalyst prepared hydrothermal and calcination technique.........................................................................................................9 4 9-5. Comparison of BET specific surface area for of ZnO-TiO2 photocatalyst prepared by hydrothermal and calcination technique............................................................................95 9-6. Comparison of XRD patterns of ZnO-TiO2 photocatalysts prepared using a calcination technique............................................................................................................................95 9-7. Comparison of XRD patterns of ZnO-TiO2 photocatalysts prepared using a hydrothermal technique.....................................................................................................97 9-8. Modified Kubelka-Munk spectra for calcined ZnO-TiO2 photocatalyst...............................97 9-9. Modified Kubelka-Munk spectra for hydrothermal ZnO-TiO2 photocatalyst.......................98 9-10. Estimated bandgap for calcined and hydrothermally prepared ZnO-TiO2 photocatalyst materials...................................................................................................................... .......99 10-1. Silica coated barium ferrite............................................................................................ ....101 10-2. Estimated pore size distribution estimate for HSAMP1, 4, and 8.....................................103 10-3. Vibrating sample magnetometry hysteresis loop of barium ferrite...................................104 10-4. Vibrating sample magnetometr y hysteresis loop of HSAMP1-8......................................105 10-5. X-ray diffraction patterns for HSAM P, Degussa P25 and barium ferrite.........................105 10-6. X-ray diffraction patterns for HSAM P, Degussa P25 and barium ferrite.........................106 10-7. Normalized phenol concentr ation as a function of time....................................................107 10-8. Normalized nDOC concentr ation as a function of time....................................................107 10-9. Proposed reaction mechanism for photocatalytic oxidation of phenol..............................108 10-10. Effect of repeated use on the extent of photocatalysis....................................................109 11

PAGE 12

Abstract of Dissertation Pres ented to the Graduate School of the University of Florida in Partial Fulfillment of the Requirements for the Degree of Doctor of Philosophy IMPROVEMENTS IN PHOTOCATALYTIC OXIDATION: PHOTOCATALYST COMPOSITION AND REACTOR DESIGN By William L. Kostedt, IV May 2008 Chair: David W. Mazyck Major: Environmental Engineering Sciences Advanced oxidation processes including UV/H2O2, ozone, and photocatalysis have demonstrated the ability to solve complex enviro nmental problems. These processes derive their power to oxidize recalcitrant organic compounds and microorganisms from powerful hydroxyl radicals. For the case of photocatalysis, hydroxyl radicals are generated when a wide-bandgap semiconductor like titanium dioxide (TiO2) is exposed to ultraviolet light of sufficient energy to generate an electron-hole pair. A favorable reaction occurs when the electron reacts with oxygen to form a superoxide radical a nd the hole reacts with water or a hydroxide ion to form a hydroxyl radical. Oxidation of organic compounds by hydroxyl radicals can lead to complete mineralization, i.e., conversion to carbon dioxide water and mineral acids. First photocatalysis, it has the potential for complete conversion of a compound to non-toxic inorganic products. This is in contrast to other technol ogies that simply transfer the compounds to another phase, like distillation or adsorption. Sec ond, the oxidant is generated in-s itu by a catalyst so storage or purchase of a possibly hazardous and e xpensive oxidant is not necessary. Photocatalytic oxidation of orga nic contaminants has received much attention in academic literature, but commercial implementation has been slow due to several factors, including the difficulty in scaling up photocatal ytic processes for water treatment from a laboratory scale to an 12

PAGE 13

13 industrial scale. One particular difficulty is reactor containmen t of highly photoactive nanoparticulate titanium dioxide, which is popular due to its low-cost a nd stability relative to other photocatalysts. Nano TiO2 is putatively nontoxic, however, this property is a current topic of active research. A solution proposed by Purifics a Canadian engineering company, utilizes a patented separation device utilizing a backpres sure cycled membrane and automated system (Photo-Cat). It was imperative to examine th e performance and successful implementation of this technology and compare this to the altern ative approaches. Another popular approach has been to immobilize the nano-sized titanium dioxi de, which will likely decrease the mass transfer of contaminants to the photocatal yst with respect to sl urry, but avoid the separation difficulty. A more effective approach, and that used in this research, was to immobilize the titanium dioxide on magnetic core particles that could be both agitated and contained using a two-component magnetic field. This provided increased mass transf er seen in slurry-based systems and provided a potential option for avoiding photocatalyst sepa ration altogether. A magnetically agitated photocatalytic reactor (MAPR) has been developed and assessed for oxidation of aqueous organic contaminants. While using phenol and or ganic red dye to simulate waste components is consistent with past literature and important to explore, it ignores factors that may hinder a photocatalytic performance including competitive adsorption of various organic compounds and their oxidation byproducts and the de monstrated detrimental effect of inorganic compounds such as ammonium bicarbonate on photocat alytic oxidation. To assess thes e effects, this work looked at photocatalytic oxidation of ai r evaporation subsystem (AES) er satz water while modifying the photocatalyst mass magnetic field current and fr equency to find the optimal conditions. After performing these preliminary experiments, seve ral proposals were formulated to improve the performance of the MAPR involving both the pho tocatalyst synthesis a nd the reactor design.

PAGE 14

CHAPTER 1 INTRODUCTION Drinking water supplies are becoming increasingly polluted with biological and chemical compounds by anthropogenic activitie s including inadequate or improperly designed water and wastewater treatment fac ilities, industrial waste, and the result of trying to mediate other environmental pr oblems. For example, in Milwaukee, a waterborne Cryptosporidium outbreak affecting 403,000 people was implicated in the deaths of many immunocompromised residents [1]. Domestic industry is regulated by the EPA to prevent human and environmental exposure to known hazards when they are discharged into the potential water supplies. This approach has likely prevented a great deal of hazardous discharge as well as reduc ing the negative envir onmental effects when an accidental discharge occurs. Unfortunately, some hazardous waste includes organic compounds that were of unknown hazards during discharge, or that were discharged before regulations were established. Endocri ne disrupting compounds represent a class of chemicals difficult to eliminate using traditio nal biological or thermal oxidation although their health risk has been proven [2]. Anothe r source of water pollution is the result of trying to reduce air pollution by adding methyl tert-butyl ether (MTB E) to gasoline. The addition to gasoline in high popul ation density areas was intend ed to reduce emissions of carbon monoxide and ozone [3], but ultimatel y caused widespread contamination of groundwater supplies due to le aking underground storage tank s [4]. While searching for the most effective technology to overcome these environmental hazards, it is important to consider the robustness, or ability to handl e unforeseen requirements without sacrificing performance. 14

PAGE 15

Several technologies offer effective solutions for complex water treatment problems, but have significant weaknesses. On e of the most widely used solutions is sorption to activated carbon or zeolite surfaces These materials function by transferring the pollutant from solution to relatively high surface area materials that can be tailored to remove the most high priority pollutants [5]. The primary drawback to this method is that the pollutant still needs to be disposed of once the adsorbent is e xhausted. Additionally, the process of manufacturing activated carbon may contribute pollution [6], yet this must be weighed against alternatives that ma y pollute more including titanium dioxide. Another option is thermal oxidation, which is very robust yet energy intensive. Finally, biological treatment provides a low-energy approach to treat a variety of challenging pollutants [7]. The disadvantages are that the microbial community may be difficult to maintain and may not be effective in trea ting a fluid stream with large transient fluctuations in nutrients or pollutants. Furthermore, biological treatment may not be effective for certain compounds [8] and may not be effective for low, yet environmentally significant concentrations. Advanced oxidation allows mineralizati on (conversion to carbon dioxide, water, and mineral acids) of almost all organi c compounds and inactivation of pathogenic microorganisms [9]. The key component to advanced oxidation is the production of highly reactive hydroxyl radicals at ambient temperature using one of several approaches [10]. One approach involves using ultraviolet (UV) light to irradiat e a dissolved oxidant including ozone, hydrogen peroxide, or both. The UV photons have sufficient energy to initiate chemical reactions that lead to the formation of th e hydroxyl radicals [11]. Another approach is to mix hydrogen peroxi de and ozone. Although this chemistry has 15

PAGE 16

been studied for at least 40 years [12], th e commercial technology for application is nascent. Several companies including H ydroxyl Industrial Systems and Calgon Carbon Corporation have developed commercially succ essful technologies to generate and utilize the power of hydroxyl radicals for treatment of water. Hydroxyl Industrial Systems has successfully demonstrated the efficacy of their design for inclusion in municipal wastewater treatment as well as industrial treatment of MTBE and cyanide. Calgon Carbon Corporation has licensed their UV systems to inactivate Cryptosporidium, Giardia, and other pathogenic microorganisms in three municipal water treatment plants. This thriving commercial activity is indicativ e of an advanced oxidation industry that is able to provide unique solutions to challenging problems. During long-term space missions, access to safe water is a key component to success. A unique advanced life support syst em is being developed to close the loop between water and wastewater. Hence, gr aywater is collected from shower, food preparation, humidity condensat e and urine waste sources an d treated to potable water quality standards. The process typically co nsists of several subsystems including a primary treatment technique that will be eith er a physical-chemical process or biological and a post-processor that will likely involve a catalyst or some other physical-chemical treatment. The challenge of a closed-loop water recovery system for long-term space missions has motivated the development of some cutting-edge technology that has applications beyond providing safe drinking wa ter for space inhabitants. The engineering specifications involve an optim ization of power, mass, volume, and crew time that are unlike typical terrestrial requirements. The pr imary mission of these explorations will be to conduct experimentation; hence, anything th at detracts from this mission such as an 16

PAGE 17

unnecessarily power intensive advanced life s upport process should be avoided. Many of the technologies developed have focused on rege nerative systems that use a catalyst, or microbes to treat water. These types of systems are advantag eous because they have the potential to reduce the mass and volume necessa ry with respect to a system that uses consumables. Additionally, systems that use ultraviolet (UV) light to drive chemical reactions in water can utilize solar UV, thus reducing power requirements with respect to a system that uses a powered light source. A water recovery solution that combines the advantages of a catalytic system with the utilization of UV is photocatalysis. Photocatalytic systems for a water post-p rocessor have been developed for space applications [13], but also have been commerc ially successful for terrestrial applications [14]. In photocatalysis a wide bandgap semic onductor such as titanium dioxide (TiO2) is irradiated with UV light. The TiO2 acts as a photocatalyst by generating hydroxyl radicals on its surface when it is exposed to UV light of sufficiently low wavelength (380 nm) [9]. The process occurs when a UV photon excites an electron from the valence band to the conduction band creating an elec tron-hole pair. In order to induce photocatalysis this electron-hole pair must migrate to the surface of the TiO2 crystal. If during this migration the electron and hole recombine, thermal ener gy will be released and the photocatalytic reaction will not occur. The hole is a powerfu l oxidant and can reac t with water or a hydroxyl ion to form a hydroxyl radical. At th e same time, dissolved oxygen can react with the electron to form a superoxide ra dical. This reaction between oxygen and the electron is important because it follows slower kinetics than the re action at the hole and thus can prevent electron-hole recombination. 17

PAGE 18

The titanium dioxide used for photocatalys is is typically in slurry (aqueous suspension of fine particles) or immobilized on a substrate. The most photoactive form of titanium dioxide is nanometer sized primary particles that form slightly larger agglomerates in solution. Degussa P25 is pa rticularly effective commercially available TiO2 that consists of 30 nm prim ary particles [15]. It is advantageous to use slurry because it allows maximal utili zation of the available spec ific surface area, though the small particle size makes separation challengi ng, but not impossible. Purifics ES, Inc. has developed and patented a slurry separation de vice that solves this problem and applied their technology to military and industrial markets [14]. Immobilization of the TiO2 on a substrate prevents the need for a complex separation device, but suffers from a mass transfer disadvantage that nega tively affects the photocatalytic performance. This mass transfer limitation can be mediated, but not eliminated, by immobilizing the TiO2 on a high specific surface area substrate such as amorphous silica prepared using an acid catalyzed sol-gel procedure [16]. To combine the advantages of the slurry-based and immobilized photocatalytic systems a magnetic photocatalyst has been used in the past [17, 18]. The magnetic photocatalyst is synthesized by coati ng small magnetic particles (nm or m) first with silica (SiO2) and then with TiO2. The silica layer acts as an insulator to prevent photodissolution or separation of the TiO2 from the magnetic particle when irradiated by UV light [17]. Additionally, the titanium dioxide has a high affinity for the silica, making the particle more mechanically robust. The magn etic property has been used in the past to provide a simple, effective method of separati on, but this requires the photocatalytic reactor to be operated in batch. 18

PAGE 19

It has been demonstrated that a dynamic magnetic field can be used to agitate and confine the catalyst within a magnetically agitated photocatal ytic reactor (MAPR) such that its exposure to UV light and thus photocatalytic oxid ation is optimized [19]. The general principle is that a magnetic catalyst particle experiences a force due to the magnetic field gradient. The magnetic field consists of two components. First, a permanent magnet is located above the react or to apply a force opposing gravity thus distributing the photocatalyst more evenly in the vertical direction. S econd, an alternating magnetic field is generated by feeding a sinusoi dal signal to a solenoid located below the reactor. This alternating magne tic field generates a magnetic fi eld gradient of alternating direction that induces agitat ion of the catalyst, which ha s been proven to increase the extent of photocatalytic oxidation of phenol [13, 19]. Adsorption and absorption are imperative processes to consider for a magnetically agitated photocatalytic reactor (MAPR). The process of adsorption involves transferring a pollutant (sorbate) material from solution to the catalyst (sorbent). Both the physical and chemical characteristics of th e system, including the chemical nature of the sorbent and sorbate, the particle size, and motion of the sorbent with respect to the bulk, will determine the extent and kinetics of adsorption. The motion of the catalyst is determined not only by the magnetic field, but by the inhe rent magnetic properties of the catalyst. This motion will determine whether there is significant film or bulk diffusion limitations for adsorption. In order to understand these ma gnetic properties, it will be important to examine their origin from an atomic perspective. Several authors have used oscillating magnetic field gradient induced motion of particles for applications other than photocatalysis [20-23], but their knowledge is pertinent to understa nding why the chaotic 19

PAGE 20

rotation and translation of particles might be useful for a photocatalytic reactor particularly for enhancing UV light absorp tion. Similarly, several authors have used magnetic photocatalysts in the past for their simple separation capability using a magnetic field [17, 18, 24-33]. They have synthesized th ese photocatalysts usi ng a wide variety of materials and methods, each of which can influence their appli cability for a MAPR. Additional work combines the advantages of the silica coating with the synergistic impact of a high surface area silica support material with an easily modified pore size distribution. Past research has indicated a synergy between high surface area silica support materials and titanium dioxide for phot ocatalytic oxidation [34]. In addition, the use of magnetic materials for photocatalyst s upports is widespread, but until now the two concepts have not been combined. By first coa ting barium ferrite with silica, it is possible to disperse it in a high surface area silicatitania composite developed previously [34] while avoiding adverse impacts to the ma gnetic properties due to the corrosive environment. When ground, sieved, and rinse d, this material is a highly versatile particulate magnetic photocatalyst. Hypotheses: Substantial improvement in magnetic phot ocatalyst synthesi s can be achieved, resulting in higher rates of photocatalytic oxidation. A photocatalytic reactor with slurry separation device and/or a magnetically agitated photocatalytic reactor can act as a secondary or post-processor for treating graywater during long-term space missions. Improvements to the magnetic field in a magnetically agitated photocatalytic reactor will result in a greater reaction rate of photocatalytic oxidation. An optimal ZnO-TiO2 composite composition and heat treatment will result in an increased reaction rate of photocatalytic oxidation. A high surface area magnetic photocatalyst can be designed with a modifiable pore size distribution. 20

PAGE 21

Objectives: Determine the impact of magnetic core material and repeated coating on photocatalyst performance. Assess photocatalytic reactor performance for photocatalytic oxidation of organic contaminants in simulate d graywater formulations Improve the magnetic field in a way that counteracts gravity and induces confinement while improving photoc atalytic oxidation kinetics. Assess hydrothermal and calcination heat treatment of various compositions of a ZnO-TiO2 composite and understand the impact of primary particle size and bandgap energy. Incorporate a magnetic material and a phot ocatalyst into a hi gh surface area silica support material while maintaining both th e magnetic and photocatalytic properties of the material. 21

PAGE 22

CHAPTER 2 LITERATURE REVIEW Magnetism: Source and Properties Magnetic materials have properties that are both fascinating and useful. As children, many of us are captivated by the apparently magical forces of attraction and repulsion between two materials possessing the proper magnetic characteristics. However, the explanation of the source of thes e magnetic characteristics is better left to scientific experts. Th ese experts have discovered that microscopic current loops, much like the series of current loops in a solenoid, also occur within magne tic materials. This current is due to the electron motion and spin [35]. The macroscopic behavior of materials within a magnetic field can be simply defined by equation 2-1. HMm (2-1) Where M is the magnetization density, m is the magnetic susceptibility and H is the applied magnetic field. This relation allows us to quantify the ability to manifest magnetic changes within a material using an external magnetic field. When m is low ( 10-4 to 10-6), the material is considered paramagnetic or diamagnetic, which means that much more powerful external magnetic fields are necessary to induce changes in the magnetization density [35]. For ferromagnetic ma terials, equation 2-2 is used instead to define the relationship between external magnetic field and magnetization density to make the coefficient a more reasonable value. HH Mr m )1 ( (2-2) Where r is defined to be the relative magnetic permeability of the material. Antiferromagnetism occurs when a couple of atomic moments oppose each other in a 22

PAGE 23

magnetic field such that the net magnetization density is zero. An in teresting variant are ferrimagnets in which a similar behavior occurs with opposing, but unequal magnetic moments leading to a net magnetization that is typically lower than ferromagnetic materials. Magnetite is an example of a ferri magnetic material where the iron is found in the crystalline structure as Fe2+ and Fe3+ depending on the lattice location [36]. This net magnetic moment of magnetite has been used for compass de signs for at least the last thousand years. For ferromagnetic materials, there are several other important quantitative parameters that define the material and thus its applicability for engineering solutions. First, there is a temperature above which the ferromagnetic behavior disappears, which is referred to as the Curie temperature. It is qu ite evident that if the ferromagnetic behavior is important to the function of an engineered device, then the opera ting temperature range should be well below the Curie temperature. The parameters of coercivity, remanent magnetization, and saturation magnetization can be best understood by examining the hysteresis loop plot (Figure 21). By following the series of events on the curve starting with the origin and moving to the right, the magnetic flux density of the magnet (B) increases as the applied magnetic field (H ) increased until it reaches a saturation magnetization value (MS, not shown). Subsequently, the applied magnetic field direction is reversed and the flux density decreases, but not at the same slope as it increased. At the point where the applied magnetic field is zero, the magnetic flux density is still present as a remanent magnetization (BR). As the applied magnetic field is reversed further, the magnetic flux density continues to decrease until it reaches zero at the value referred to as the coercivity (HC). 23

PAGE 24

Figure 2-1. Hysteresis loop of ferromagnetic materials [35]. For the magnetic photocatalyst each of these parameters has a potential impact on performance. First, the coercivity for hard magnetic materials like barium ferrite is has a relatively high magnitude whereas for soft magnetic materials like magnetite it is correspondingly low. This determines both how difficult it is to magnetize the material and likewise how difficult it is to demagneti ze the material. For a magnetically agitated photocatalytic reactor (MAPR) app lication, it is critical that this coercivity be relatively high to assure that the oscillating magne tic field does not demagnetize the magnetic particle. The MAPR uses a combination of os cillating and static magnetic fields to induce agitation and confinement. Additionally, if this value is too high it will not be possible to magnetize the magnetic particles by moving a magnet near them a more powerful electromagnet may be necessary. The re manent magnetization determines what magnitude of magnetic flux density remains af ter the particles have been magnetized. If this value is too low, the magnetic particle wi ll not be agitated within the magnetic field. If the remanent magnetization is too high, th e interparticle attraction might be large 24

PAGE 25

enough that the photocatalyst will form clusters that will not be broken apart in the presence of an oscillating magnetic field gradient. Photocatalysis and Motion For successful photocatalytic oxidation a very distinct series of events must occur involving the almost simultaneou s absorption of light and adso rption of contaminants or oxidant precursors. Perhaps the simplest wa y to understand the phe nomenon is to follow the energy through the process. Light is the s ource of energy necessary to oxidize organic contaminants. This energy is transferred from a photon to a valence band electron in the titanium dioxide crystal. If the photon energy is sufficient, the electron is promoted to the conduction band. The electron and the vacancy it left behind, the electon-hole, can migrate to the surface of the crystal. If, on the way to the surface the electron and electron-hole recombine, the energy is releas ed in the form of heat. If the electron electron-hole pair successfully makes it to th e surface, they can participate in reduction and oxidation reactions respect ively. The electron can reduce adsorbed oxygen to form a superoxide radical and the hole can oxidize ad sorbed organics or oxi dant precursors like water or hydroxyl ions to hydr oxyl radicals [9]. It is possi ble that the hydroxyl radical may migrate into solution, but its high reactiv ity will likely prevent it from traveling far [9]. When applying these principles to a MAPR system, it is clear that the motion of the photocatalyst can play a significan t role in enhancing both the absorption of light and the adsorption of contaminants. First, the light absorption is related to the distribution and motion of the catalyst within the reactor. Since the magnetic core is prone to absorb light (no UV transparent hard ferromagnetic mate rial is known although a composite core could be semi-transparent), light will not likely reach a photocatalyst part icle if its path is 25

PAGE 26

blocked by another particle. Given this inform ation, and that the phot ocatalyst itself or the agglomerates formed in solution are la rge enough to have a diffusion limitation, it is clear that the ideal si tuation for maximum light absorpti on is for the photocatalyst to be moving and well distributed within the reactor. This ideal mo tion and distribution will, of course, depend on the geometry of the reac tor and the light source. Adsorption of contaminants will be affected similarly by both the distribution and motion of the photocatalyst. To be adsorbed, the contam inant must first diffuse through the bulk solution then through any boundary layer before it can reach the surface of a non-porous solid [37]. If the photocatalyst resides in a stationary heap on the bottom of a reactor, diffusion will be impeded by both the diffusi on distance and boundary layer contribution to diffusion resistance. The reactor conditions can be optimized for increased adsorption by distributing the catalyst more evenly throughout the reacto r to reduce diffusion distance and by increasing its motion to reduce the film diffusion resi stance. Ideally, this motion would be chaotic to increase the pr obability of a succe ssful photocatalytic oxidation event. Applications of a Magnetic Field Gradient for Particulate Systems Applied magnetic field gradients have b een used in a diverse group of magnetic particulate system applications. These appl ications typically fall into one of two categories: confinement or agitation. The firs t category includes a magnetic field gradient induced stabilized fluidized beds while the se cond category comprises particle coating, particle deagglomeration for powder flow a nd fluidized beds and finally a magnetically agitated photocatalytic reactor and utilizes an oscillati ng magnetic field gradient. A magnetic field gradient can be utilized to provide an additional force for a fluidized bed that can be useful for both terrestrial and microgravity conditions. 26

PAGE 27

Jovanovic et al. have developed the Gradient Magnetically Assisted Fluidized Bed (GMSFB) to perform this task [38]. Their de sign utilizes a series of Helmholtz rings supplied with direct current that increases from coil to coil as one moves up the column such that the coil at the bottom of the column has the highest current and the coil at the top has the lowest current. Th is variation in coil current gives rise to a magnetic field gradient which acts much like gravity. Th ey found that they could both enhance fluidization under normal gravity conditions and induce fluidization in microgravity conditions (aboard a NASA KC135 aircraft) where it does no t normally occur due to the absence of gravity because the media can not be contained within the reactor [39]. For a MAPR application, this gradient induced force using a direct current could be useful for both improving the distribution and containm ent of the photocatalyst. In fact, the versatility of the electromagnetically generate d field gradient will allow tailoring of the field to a specific gravitationa l environment or flowrate. An oscillating magnetic field gradient can cause small magnetic particles to spin and fluidize [20], which is a useful type of motion for severa l applications. The discharge of powders from hoppers is a relatively simple matter as long as the particles are not small or cohesive. However, cohesion betw een the particles can cause arching or ratholing, which occurs when the powder flows more quickly down the sides or middle respectively. To combat the adhesion, typical means include vibr ating (vibrating hopper wall or vibrating outlet), pneumatic (air blas ts, aeration, etc.), and mechanical methods (rotating screw). Dave et al. proposed the us e of a magnetically assisted powder flow (MAPF) system, which consists of 0.85-3 mm ba rium ferrite particle s that were agitated within the hopper near the ou tlet by an oscillating magnetic field gradient. The particles 27

PAGE 28

were contained by a 635 m wire mesh at the outlet. By using the MAPF they were able to achieve a controlled discharge of a 15 m mean diameter cornstarch powder, which is known to be a very cohesive and hence diffi cult powder to deliver accurately using a hopper. Another application used by the same research group who developed the MAPF was to effect dry particle coating using a ma gnetically assisted impaction coating (MAIC) technique [21]. The general c oncept was to coat larger core (host) particles with fine particles (guest) to prevent cohesiveness and hydrophilicity by causing intraparticle collisions as the magnetic particles were ag itated in an oscillati ng magnetic field. They chose to coat cornstarch and cellulose w ith silica particles an d polymethylmethacrylate (PMMA) with alumina particles. They were successful at reducing hydrophilicity and flowability of cornstarch and cellulose and attributed this success to a thorough coating by silica particles. The SEM evidence suggested that the magnetically agitated particles were not only effective for the impaction coa ting process, but also breaking up larger agglomerates of guest particles to form a more thorough and uniform coverage. They noted that surface coverage was most thor ough at 45 Hz and between 80 and 100 Hz, but there was a decrease in coverage between 60 and 70 Hz. Additionally, they found that as they increased current between 1 and 5 A there was a corresponding increase in surface coverage. A similar magnetically agitated sy stem was used to enhance fluidization of silica nanoparticles was developed by Yu et al. using an oscill ating magnetic field gradient and millimeter sized barium ferrite particles [23]. They found that the chaotic spinning and translation of the particles wa s able to break up the soft nanoparticle agglomerates enough to allow enhanced flui dization even after the magnetic particles 28

PAGE 29

were removed. A follow-up article was also recently published that went into a more detailed explanation of the dry particle coating procedure using the MAIC method and comparing it to other conventional dr y particle coati ng techniques [22]. Both of these oscillating magnetic field gr adient applications demonstrate that a chaotic spinning and translation particle mo tion can be induced. Even though this motion was not developed with the inte nt of enhancing photocatalytic oxidation, it is likely that it will have this effect due to the enhanced ab sorption and adsorption expected as explained in the Photocatalysis an d Motion section and seen in previous work [13]. Magnetic Photocatalyst Synthesis Magnetic titanium dioxide-based photocatal yst synthesis has evol ved since the first conference proceedings by Beydoun et al. [26] in 1998 (Figure 2-2), yet the basic coreshell structure has remained the same. This st ructure consists of a magnetic core material that is coated with titanium dioxide and po ssibly an electrically insulative material between the two. The motivation for the magnetic photocatalyst developed by Beydoun et al. was to provide a simple method of catalyst separation using a magnet. Their photocatalyst design also allowed them to take advantage of the high specific surface area and diminished film diffusion resistance of a particulate catalyst with respect to an immobilized photocatalyst. The magnetic core materials used by past researchers include both hard and soft magnetic materials. The most common core material is magnetite (Fe3O4), a soft ferrimagnetic material that can become oxidized to maghemite ( -Fe2O3), which is also ferrimagnetic if the particles are heated over 150-170C according to Beydoun et al. Fortunately, the saturation ma gnetization of maghemite is decreased by only 12% (from 84 to 74 emu g-1), yet at higher temperatures of about 550C and depending on the 29

PAGE 30

30 particle size and shape the maghemite is c onverted to hematite, which is so weakly ferromagnetic that it is no longer practical to separate using a magnet [17]. Beydoun et al. attempted to reduce the length of the heat treatment and added a silica insulating layer [17] to prevent this core oxidation from adversely affecting the separation capability. Hard magnetic materials are more difficult to magnetize, but once magnetized can maintain their magnetic characteristics as oppos ed to the spontaneously induced magnetic behavior of soft magnetic mate rials. However, they may suff er from some of the same oxidation-related decrease in magnetic propertie s [40]. Materials that are hard magnets can be recovered using a magnet just as soft magnets, but they have the additional capability of being agitated using an osci llating magnetic field gradient. Two hard magnetic materials, barium ferrite and zinc -nickel ferrite have been used by several researchers with the intent of magnetic agitation [19, 41, 42] and magnetic recovery [43, 44].

PAGE 31

31Figure 2-2. Comparison of magnetic phot ocatalyst synthesis procedures.

PAGE 32

The magnetic core sizes found in literature range from 4 to 2000 nm and many were produced in the lab. The smallest measured co re was developed by Beydoun et al. in 1999 using a coprecipitation of iron (II) a nd iron (III) chloride with amm onium hydroxide, which produced a particle size range of 4-10 nm [28]. The same group of researchers also produced 40-70 nm magnetite particles by oxidizing Fe(OH)2 gels with potassium nitrate at 90C [26]. In contrast to the predominantly wet chemistry methods, Chung et al. chose to use an ultrasonic spray pyrolysis method to produce 720 nm nickel ferr ite core particles [45]. Other researchers produced cobalt (26-74 nm) and zinc-nickel (1 200 nm) ferrites using a hydrolysis and precipitation method [43, 44, 46]. The researchers who chose barium ferrite purchased their core materials from Arnold Engineering [41, 42] and Alfa Aesar [19]. Possibly the most elaborate core synthesis method was that developed by Rana et al. utilizing a microemulsion technique [47]. Essentially, this technique involves making a mixture of nanometer sized water droplets in oil using a surfactant that act as individual chemi cal reactors. Two of these microemulsions were prepared: one containing the metal chlorides an d the other containing ammonium hydroxide. The microemulsions were mixed allowing the contents to slowly react with ea ch other as the micellar structures periodically rupture a nd transfer contents to another water droplet thus slowing the rate of particle nucleat ion. In this case, the particle size was approximately 10-20 nm including the titanium dioxide shell. The first use of a silica insulating layer in a magnetic photocatalyst was described by Chen and Zhao in 1999 [48] and later elaborated by Beydoun et al. in 2001 [17]. Chen and Zhao found that they could increase the phot ocatalytic activity of their magnetic photocatalyst by adding a silica membrane deposited using a pH adjusted s odium silicate solution. They proposed that this increase in photocatalytic activity was due to the prevention of electron-hole recombination at 32

PAGE 33

the -Fe2O3-TiO2 interface. Shortly thereafter Beydoun et al. published a series of articles describing the negative impact of the -Fe2O3-TiO2 junction on photocatalytic activity due to the same electron-hole recombination effect and th e occurrence of photodissolution, which causes titanium dioxide particles to be ejected from the surface of the magnetic substrate [28]. Beydoun et al. also proposed to solve th is problem using a silica insulating layer, but instead performed this coating using the Stber method with a tetrae thylorthosilicate (TEOS) silica precursor [17]. Lee et al. [42] and Kostedt et al. [19] used the Stber method as well, but chose to use isopropanol instead of ethanol as a solvent, whic h is likely to produce la rger silica particles. Additionally, Lees method indicated a single coa ting of silica whereas Kostedt coated twice using the same procedure. Along the same lines as their aerosol core synthesis method Chung et al. produced a silica coating using ultr asonic spray pyrolysis of TEOS [45]. Titanium dioxide nanoparticles have been predominantly deposited using a wet chemistry approach with an alkoxide TiO2 precursor. Chen and Zhao [48] were the only authors to use Degussa P25 to coat their magnetic photocatalyst. They did so by first abrading the mixture of magnetic particles and TiO2 then sonicating and annealing the dried particulate mixture. Over half of the authors who synt hesized magnetic photocatalysts used a method involving the hydrolysis and condensation of titanium n-butoxi de (TBOT). This typically involves heating a mixture of water and TBOT in an ethanol so lvent to produce a slightly crystalline TiO2 sol. These titanium dioxide partic les can be deposited on the ma gnetic substrate surface using agitation by a magnetic stir bar or by sonication. Since the titanium dioxide particles deposited using this method are only sligh tly crystalline, and the crystallinity and crystalline phase are critical to high photocatalytic ac tivity, a subsequent heat treatm ent is necessary to generate primarily anatase ph ase nanoparticles. 33

PAGE 34

Most authors have used temperatures between 450-550C for between 30 and 180 minutes to accomplish this increase in crystallinity and ph ase change to anatase. The positive effects of the crystalline phase change ma y be offset by decreases in surface hydroxyl group density or particle size increase if the temperature is too high or the treatmen t duration is too long. Additionally, Gao et al. [49] found that heat treating to between 700 and 900C caused a phase change to rutile and formation of pseudobrookite (Fe2TiO5), which decreased photoactivity as well. It should be noted that Gaos study did not utilize any silica in sulating layer. Another approach was that developed by Watson et al. [ 18] to directly deposit anatase phase using a titanium isopropoxide precursor with a high water to alkoxide ratio and acidified with nitric acid. This direct deposition avoids the necessity of the heat treatment step, which was shown by their previous work to decrease the magnetic propert ies and increase the cost and time required for synthesis. The synthesis procedures can have a dr amatic impact on both the adsorption and absorption properties of the magnetic photocatalyst. The core size and material will determine its adsorption properties. Silica was used by ma ny researchers as a membrane to prevent unfavorable electronic interactions with the co re material, but it also has some promising properties for adsorption due to the surface silanol functional groups [50]. Additionally, the size of the core material as well as the size of the titanium dioxide particles will influence the specific surface area of the photocatalyst. Ge nerally, smaller core material will lead to a greater specific surface area, but it is important to consider the possibility of several core particles being included in a single silica membrane [19, 42]. In light of the concepts presented in this wo rk, it is apparent that the optimization of a MAPR requires knowledge of adsorption and absorption. Specifically, it is important to consider 34

PAGE 35

the properties of the photocatalyst and its behavior in an oscilla ting magnetic field gradient. The impact of photocatalyst dist ribution and motion on oxidation wa s elucidated. It was also demonstrated that the magnetic characteristics of th e material used for the core has an impact on its ability to be fluidized. Dye Removal Using a ZnO-TiO2 Composite Textile dyes represent a signifi cant environmental hazard to aquatic biota and humans due to their toxicity [51] and tendency to cause eutrophication [9]. Improvements in chemical structure have led to increased longevity of dyed fabrics, yet ma ke the textile wastewater more difficult to treat. Traditional biol ogical methods often fail due to th e increased stability of highly aromatic dye molecules. Additionally, methods su ch as adsorption, reverse osmosis, coagulation, etc. simply transfer the dye from one phase to another leaving a se condary waste problem. Several options for advanced oxidation ex ist including wide-bandgap semiconductors. These materials have the ability to generate hydroxyl radicals when exposed to sub-bandgap energy photons (typically in the ultraviole t range). Frequently used wide-bandgap semiconductors include TiO2 [52-54], ZnO [52, 55, 56], CdS [57], and others. However, TiO2 is most popular due to its relatively low toxicity, low cost, and re sistance to photocorrosion [58]. The other materials function as photocatalysts, but exhibit photocorrosion in the presence of UV light that drives the photooxidative reactions. This is particularly troublesome for compounds such as CdS as photocorrosion leads to polluti on with a heavy metal [57]; however several studies have reported superior perf ormance by using ZnO instead of TiO2 [59, 60]. While the most frequently used photocatalyst materials are composed of a single semiconductor, other authors have found benefits of using mixed oxides with TiO2 and one of the following: ZnO, WO3, SiO2, or ZrO2 [61-68]. Some explanations for this phenomenon include increased surface acidity and favorable electronic interactions between the different 35

PAGE 36

metal oxide materials. This study included a co mparison between two heat treatment methods, hydrothermal and calcination afte r co-precipitation of ZnO/TiO2 and explored the role of surface area, heat treatment, and zeta potential, a nd bandgap energy on photocatalytic oxidation and adsorption of Procion Red MX-5B, an azo dye. 36

PAGE 37

CHAPTER 3 MATERIALS AND METHODS Common Methods Magnetic Photocatalyst Synthesis Method Silica coating Several barium ferrite core materials were used. For Incorporation of a Permanent Magnet and Influence of Ball Milling on MAPR Performance the raw mate rial for the magnetic core was ball-milled barium ferrite with an average particle size of 2.2 m. For all other studies, the barium ferrite was a powder obtaine d from Alfa Aesar. Barium fe rrite was coated with silica using a base catalyzed sol-gel procedure [17, 42, 69] to provide insulation between the magnetic core and the photocatalyst layer for the purpose of preventing photodissolu tion and corrosion of the core. The silica coating proce dure is as follows: 300 mg of 2.2 m barium ferrite particles were dispersed in 25 mL of isopropanol (Acr os Organics) and sonicated for 5 min. After sonication, 1.1 mL of tetraethylor thosilicate (TEOS, Aldrich), 4 mL of deionized (DI) water (18.2 M cm, Nanopure), and 0.147 mL of ammonium hydroxide (Aldrich) were added to the sonicated solution. The prepared solution was vigorously stirred fo r 3 hr at room temperature. The silica-coated barium ferrite particles were recovered from the solution by magnetic separation, washed with isopropano l and DI water, and finally dried at room temperature for 2 days. This was a single coating for Incorpora tion of a Permanent Magnet and Influence of Ball Milling on MAPR Performance and a double coating was performed for all other work. Titania coating An amorphous layer of titanium dioxide was deposited using a hydrolysis and condensation technique [ 18, 70]. The prepared silica-coated barium ferrite particles were dispersed in a mixture of 100 mL of ethanol (Aldrich) and 2.7 mL of DI water and sonicated for 37

PAGE 38

10 min. Next, 0.47 mL of titanium n-butoxide (TBO T, Acros Organics) was dissolved in 10 mL of ethanol and this solution was gradually added with a pipette to the mixture of particles, ethanol, and water. The solution wa s vigorously stirred for 10 min at room temperature and then heated to 90 C at a rate of 5 C/min, where it stayed for the ne xt 2 hr while stirring vigorously. The prepared sample was recovered from the solution by magnetic separation and washed with ethanol and DI water. The wet sa mple was then dried in an oven at 60 C for 2 days. The dried sample was heat-treated in a box furnace with a 5 C /min ramp to 500 C, then held for 1 hr in an ambient atmosphere to develop th e anatase crystalline phase of TiO2, which was verified by XRD [42]. The particles were ma gnetized by passing a magnet over th em several times. Presence of the silica layer was verified via energy dispersive x-ray (E DS), scanning electron microscopy (SEM), and transmission elec tron microscopy (TEM) [42]. TiO2 mass loading using the double coating technique was determined to be 28.0% by digesting and comparing dried sols prepared with and without silica-coated barium ferrite pa rticles. The digestion method was adapted from Korn et al. [71] and the samples were analyzed with a Perkin Elmer 3200 inductively coupled plasma optical emission spectrometer (ICP-OE S). A Brunauer, Emmett, and Teller (BET) surface area analysis using a Qu antachrome Nova 2200e revealed a specific surface area of 31.3 m2/g. Incorporation of a Permanent Magnet a nd Influence of Ball Milling on MAPR Performance Experimental Setup The experimental setup consisted of th ree primary components: a magnetic field generation system, a curved quartz reactor (to in crease transmission of UV wavelength light with respect to the previously work with a Pyrex re actor), and three ultravio let lamp fixtures. The magnetic field generation system consists of a f unction generator whose signal is fed to an audio 38

PAGE 39

amplifier and finally to a series of two solenoids (Figure 3-3) In addition, a 2.54 cm3 permanent magnet is suspended on a sheet of plexiglass above th e reactor as can be seen in Figure 3-1. This addition of a static magnetic fiel d counteracts the gravitational fi eld that causes the catalyst to remain clumped near the bottom of the reactor thus increasing UV irradiation of the catalyst. It is important to note that this static field may n eed to be modified depending on the gravitational requirements of the application. For instance, on the moon, where th e gravity field is weaker, a weaker magnet could be used to generate the st atic field. The reactor is constructed of three quartz segments oriented at right angles to each other with the middle segment curved to follow the polar symmetry of the alternating magnetic field. The three UV lamp fixtures are located above and on both sides of the reactor. Figure 3-1. Magnetically agitate d photocatalytic reactor. Photodegradation of Phenol and Procion Red MX-5B Dye Phenol and reactive red were chosen as simula ted pollutants for their ease of testing. Stock solutions were prepared with liquid phenol (Fisher) or Procion Red MX-5B powder (Aldrich) diluted with 18.2 M /cm Barnstead E-Pure water. An 8watt 254 nm UV lamp (Spectronics Corp.) was chosen to test the hypothesis of previo us work [72] that the lack of photocatalytic 39

PAGE 40

activity with the 254 nm lamp was due to sub-visible wavelength light absorption by the Pyrex reactor. To prevent this problem, a quartz r eactor was used instead. All experiments were performed in batch with 10 mL of simulated pollutant, 100 mg of catalyst unless specified otherwise and the reactor was seal ed with rubber caps (fluid was not in contact with the caps and adsorption was found to be negligible). Analysis of phenol concentrations was performed with a GC/MS (Varian 3900/2100T) and non-purgeable dissolved organic carbon (nDOC) with a Tekmar Dohrmann Apollo 9000. Procion Red dye was analyzed using a Hach DR4000U spectrophotometer with a 1 cm quartz cell to measure the maximum wavelength absorbance of 538 nm. Enhanced Photocatalyst Synthesis a nd Reactor Parameter Exploration Photocatalytic Test Stand The experimental setup consisted of a magne tic field generation device, UV lamps and a batch reactor. The 10 mL curved reactor was constructed of 3 fused quartz tubes as can be seen in Figures 3-2 and 3-3. The straight upright segm ents have an inner diameter of 1.27 cm and the curved segment has an inner diameter of 1.05 cm. The 8-watt UV lamps had a 365 nm wavelength peak output and were manufactured by Spectronics Corp. (Westbury, NY). During experimentation, the reactor was sealed with Te flon caps that were not in contact with the solution. 40

PAGE 41

Figure 3-2. Two-dimensional drawing of quartz reactor. Figure 3-3. Three-dimensional drawing of quartz reactor. 41

PAGE 42

The magnetic field was generated with an altern ating current signal fed to a series of two solenoids (Figure 3-4). Each solenoid consis ted of 152 m of 18 AWG wire wound around an 8.9 cm diameter wire spool. The wire was wound in 20 layers with 23 turns per layer as a result of modeling to maximize the magnetic field gradient [72]. The signal was generated by a BK Precision 4011A function generator. The output from the function generator was delivered to a Guitammer BKA 1000-4 power amplifier with outputs attached to the leads of solenoids in series (Figure 3-4). Solenoid A was used to increase the impedance of the amplifier load to prevent overloading while solenoid B was used for the ma gnetic field generation be neath the reactor as seen in Figure 3-2. Finally, a 2.54 cm3 permanent magnet was suspended on a sheet of plexiglass above the reactor to provide a magnetic field that counteracted gravity. With the permanent magnet in place, the catalyst within the reactor was both horizontally and vertically dispersed. During all experiments the test stand was cove red with a black box to prevent UV light from entering or leaving. A fan assisted in transfer ring heat generated by the solenoid and UV lamps to maintain a steady state sample temperature of 29 5 C. Figure 3-4. Alternating magnetic field generator. 42

PAGE 43

Gas Chromatograph/nDOC Analysis of Phenol Liquid phenol from Fisher Chemicals (90.9%) was used to simulate aqueous organic pollutants. All experiments treated 10 mL of an 11 mg/L phe nol solution prepared for each experiment by diluting the concentrated liquid phenol with Barnstead E-Pure 18.2 M /cm water. The initial dissolved oxygen was found to be 7.5 0.2 mg/L using a YSI 52 dissolved oxygen meter (used according to manufacturer guidelines) and was not augmented with an external oxygen supply during experiments. The reactor was operated in batch such that data for each time represents a discrete experiment. Sample s for GC and nDOC (non-purgeable dissolved organic carbon) analysis were obtained by separating the catalys t from solution with a magnet and storing it in a VOA vial with Teflon septa at 4 C until testing. Non-purgeable dissolved organic carbon was measured on a Tekmar D ohrmann Apollo HS 9000 with autosampler. Purging and mixing was accomplished by sparging the sample, which had been pH adjusted with phosphoric acid, with zero grade air for 3 minut es. Analysis of phenol concentrations was performed with a GC/MS (Varian 3900/2100T) usi ng a solid phase microextraction (SPME) sampling technique with an 85 m polyacrylate co ated fiber assisted by an addition of reagent grade sodium chloride. The column was a PTE-5 and the temperature profile included maintaining 40C for 4 minutes, then elevating to 260C at a rate of 12C/min The relative standard deviation for (RSD) for replicate analys is of samples was less than 5% for both reported phenol concentrations and nDOC values. Magnetic Photocatalyst Characterization and Reactor Performance Using AES Ersatz Water Characterization of the Magnetic Photocatalyst The magnetic photocatalyst was characterized for both crystalline size and to determine what type of assembled structures formed duri ng exposure to the magnetic field. Crystalline size 43

PAGE 44

was determined by performing an x-ray diffrac tion (XRD) analysis using a Phillips APD 3720 then using the Scherrer equation to estimate the size (eq. 3-1). Assemb led structures were examined using a JEOL 6400 scanning electron mi croscope (SEM) with an accelerating voltage of 5 kV before and after exposi ng the sample to the magnetic field at 1 A at 60 Hz with the permanent magnet present. cos B k t (3-1) Photocatalytic Test Stand The photocatalytic test stand c onsisted of components to generate the magnetic field and provide UV light to the curved qua rtz reactor. The oscillating ma gnetic field was generated using a solenoid powered by an amplified sinusoidal wa ve function while the static magnetic field was the result of a permanent magnet located above the reactor. By using a function generator to modify the solenoid frequency, and an audio amp lifier to alter the solenoid current, the magnetic field was altered to determine its effect on phot ocatalytic oxidation of AES (Air evaporation system). All experiments were performed with 100 mg catalyst, 10 mL of AES ersatz and an oscillating magnetic field provided at 60 Hz with 1 A current for 1 hour in batch unless specified otherwise. Also, all experiments us ed three-254 nm peak output lamps. Analysis of Photocat alytic Oxidation Air evaporation system ersatz water was prepared using the procedure developed by Verostko et al. [73] and dilution water wa s provided by a Nanopure Diamond Analytical TOC water purifier. Because the oxidation byproducts are numerous and not easily identifiable, nonpurgeable dissolved organic carbon (nDOC) was used to assess the extent of photocatalysis. The nDOC was measured with a Tekmar Dohrma nn Apollo 9000 HS. The nDOC analyzer first adjusted the pH of the sample with phosphoric acid, then purged with air to remove inorganic 44

PAGE 45

carbon. Subsequently, a sample was injected into a furnace filled with a platinum doped alumina catalyst at 680C to convert orga nic chemicals to carbon dioxide (CO2) and water. The water was removed by cooling to condense some of the wa ter, then by a Nafion membrane to remove remaining water. Finally, a scrubber was used to remove corrosive gas before the stream was directed through a non-dispersive infrared det ector (NDIR) to quantify the amount of CO2 produced. Photocat Experiments Us ing Graywater Forumula Simulated NASA Graywater Preparation Several simulated wastewater types were prepared for testing in the Photo-Cat. The first type, simulated transit mission wastewater was prep ared using the chemicals listed in Table 3-1. The chemicals were quantitatively transferred to the reservoir of the Photo-Cat after it had been filled with 15 L of tap water (containing typical organic and inorganic materials at acceptable levels for potable water) and titanium dioxide Mixing took place as the water was re-circulated through the system. Biological Water Processor Effluent Preparation Biological water processor e ffluent (BWP) was prepared according to the NASA Ersatz [73]. The 1 L concentrated solutions were prepared with the specified chemicals and allowed to remain at room temperature in sealed containe rs. To prepare a working solution, the specified quantity of each concentrate was transferred to th e Photo-Cat reservoir and allowed to mix as the system re-circulated (Table 3-2). 45

PAGE 46

Table 3-1. Simulated graywater formula. Constituent Amount Per Liter Ammonium bicarbonate 2726 mg Sodium chloride 850 mg Potassium bicarbonate 378 mg Hippuric acid 174 mg Potassium dihydrogen phosphate 173 mg Potassium bisulfate 111 mg Citric acid 92 mg Glucuronic acid 60 mg Johnsons Baby Shampoo 1.2 g Ammonium hydroxide (14.8 N) 1 mL Vegetable oil 667 L Water (Tap) 868.4 mL Table 3-2. Biological water processor formula Constituent Amount Per Liter Amonium bicarbonate 724.28 mg Amonium chloride 100.02 mg Sodium chloride 632.28 mg Sodium nitrite 24.60 mg Sodium nitrate 285.42 mg Potassium monobasic phosphate 171.82 mg Potasium bisulfate 165.72 mg Potasium chloride 273.62 mg Hydrochloric acid(37%) 158.00 L Dextran 97.99 mg Glucuronic acid 9.68 mg Creatinine 8.46 mg Urea 9.00 mg 46

PAGE 47

Fructan 4.60 mg Tyrosine 6.09 mg Acetic acid 0.22 L Lactic acid 0.22 L 2-Propanol 0.20 L Acetone 0.19 L Ethanol 0.27 L Photocatalyst/UV Lamps Degussa P-25 (particle size approximately 30 nm) was chosen as the photocatalyst and was used as received. For each experiment, 10 grams of catalyst was added to the reservoir after it was filled with 15 L of tap water providing a TiO2 slurry concentration in the system of 667 mg/L. The UV source was 8-64 75 watt bulbs with a peak output wa velength of 254 nm. Figure 3-5. Supervisory control and data acquisition schematic of photocat. SLP = slurry loop pump, DPCV = discharge pressu re control valve, SLCV = slurry loop control valve, SFCV = slurry feed control valve, BPD = backpulse device, PWSV = purge water solenoid valve, DSV = drain solenoid valve, ISV = influent solenoid valve. 47

PAGE 48

Supervisory Control and Data Acquisition In terface, Control Systems, and Programmed Operation Modes The supervisory control and data acquisiti on (SCADA) interface allowed complete control of system functions including UV lamp operation, pump speed, and solenoid valves (Figure 3-5). A programmed purge mode was used to clean the sy stem between experiments. In this mode, all portions of the Photo-Cat were flushed with tap water at least 3 times. The interface also monitored temperature, pH, and various flowra tes in real-time. Programmed operation modes were used for batch mode, where the slurry separation device was bypassed and default mode where it was not. Finally, a programmed low-flow mode allows a low flowrate of contaminated water to be delivered to the reactor via an externally-controlled pump. Effluent Sampling and Testing Samples were collected in 40 mL acid-washed glass vials with Teflon septa and were promptly refrigerated at 5 C. An initial sample was taken to verify the effectiveness of cleaning between experiments. A portion of each samp le was analyzed with a Hach DR4000U for absorbance at 254 nm using a 1 cm path length quartz cell. Using this va lue as an estimate of non-purgeable dissolved organic car bon (nDOC), the sample was diluted with Barnstead E-Pure water (18.2 M /cm) to maintain a nDOC level below 10 mg/L. The diluted sample was analyzed for nDOC using a Tekmar Dohrmann Apollo 9000 HS TOC analyzer. Photocatalyst Comparison Using the Photocat Reactor Biological Water Processor and Air Evaporation System Effluent Ersatz Preparation Biological water processor efflue nt (BWP) and air evaporation system effluent (AES) were prepared using the NASA ersatz definition [73] developed by Verostko et al. To prepare a working solution, the specified quantity of each concentrate was transferred to the Photo-Cat 48

PAGE 49

reservoir and allowed to mix as the system re-circulated. Dilution water was from a Nanopure Infinity water purifier. Photocatalyst/UV lamps Both Degussa P25 and silica-titania composite photocatalysts were used. Degussa P25 was used as received. Silica-titani a composite photocatalyst was sy nthesized using the procedure described previously for 140 mean pore size and 30% TiO2 by mass [16]. Instead of making pellets in 96 well assay plates, the gel was aged as a monolith in a 3 L beaker. After aging, the xerogel was dried at 103C for 18 hours then ramped to 180C where it dwelled for 6 hours. Once dry, the gel was heat-treated with a 5 C/min ramp to 500C to remove any residual organics were it remained for 1 hour. After the heat treatment, the gel was ground with a mortar and pestle and the 45-90 m fraction was used for experimentation. A Brunauer, Emmett, and Teller (BET) surface area analysis using a Quan tachrome Nova 2200e revealed a specific surface area of 400 m2/g. The UV source was 8-64 75 watt bulbs with a peak output wavelength of 254 nm. Scanning Electron Microscopy (SEM) SEM images and energy dispersive x-ray (E DS) spectra were obtained using a JEOL JSM 6400 with an accelerating voltage of 15 kV. Backscatte r electron (BSE) images were obtained using a solid-state detector located on the polepiece. All specimens were coated with a conductive carbon layer prior to analysis using a vapor deposition technique. Effluent Sampling and Testing Samples were collected in 40 mL amber glass vi als that had been soaked in 10% sulfuric acid for 24 hours and subsequently heated to 550C for at least 1 hour. The samples were promptly refrigerated at 4 C. An initial sample was taken to verify the effectiveness of cleaning between experiments. A portion of each samp le was analyzed with a Hach DR4000U for 49

PAGE 50

absorbance at 254 nm using a 1 cm path length quartz cell. Using this va lue as an estimate of non-purgeable dissolved organi c carbon (nDOC), the sample was diluted with Nanopure DIamond (18.2 M /cm) to maintain a nDOC level below 10 mg/L. The diluted sample was analyzed for nDOC using a Tekmar Dohrmann Apollo 9000 HS TOC analyzer. Impact of Heat Treatment and Composition of ZnO-TiO2 Nanoparticles for Photocatalytic Oxidation of an Azo Dye Materials All chemicals used in this study were analyt ical grade reagents. Zinc nitrate hexahydrate Zn(NO3)26H2O and titanium tetrachloride TiCl4 (Fisher) were chosen as precursors of ZnO and TiO2 respectively. Procion Red MX-5B was used for photocatalytic oxidation and adsorption experiments. Synthesis Zn(NO3)2H2O and TiCl4 solutions were mixed at room temperature using magnetic stirring. The samples were synthesized at different compositions, which will be referred by the percentage of TiO2 with the balance ZnO followed by C for calcined or H for hydrothermal (95H corresponds to 95% TiO2 that was hydrothermally prepared ). The pH was adjusted to 8.5 by slowly adding an ammonium hydroxide solution and continuing to stir for two hours. White Zn(OH)2/ Ti(OH)4 precipitate formed instantaneously. The precipitated samples were divided into two portions to study the effect of direct calcination and hydrothermal techniques. One portion was separated by filtration and washed by water several times until the pH reached 7. The first part was dried overnig ht at 110 C and calcined at 500 C for 5 hours. The second portion was hydrothermally treated (high pressure water vapor environment) in a Teflon-lined autoclave at 150 C for 10 hours. The powder was th en filtered, washed and dried at 110 C for 24 hours. 50

PAGE 51

Characterization Material characterization including specifi c surface area was estimated using nitrogen adsorption isotherms measured on a Quantachrome Nova 2200e and mode led using the Brunauer Emmet and Teller (BET) model, x-ray diffraction (XRD) 2 patterns were derived with a Phillips APD 3720, zeta potential was determin ed using a Brookhaven ZetaPlus, and diffuse reflectance spectra were estimated using a Labs phere PELA-1000 integrating sphere module in a Perkin-Elmer Lambda 800 UV/Vis spectrophotometer. Dye Photocatalytic Oxidation Performance of the photocatalyst materials wa s assessed by measuring both adsorption and photocatalysis of Procion Red MX-5B dye. All experiments were conducted using a Glas-Col Rugged Rotator. Teflon sealed borosilicate vials were prepared with 20 mg of photocatalyst and 40 ml of 15 mg/l Procion Red MX-5B (RR) dye solution (Figure 1) where pH was unadjusted and measured to be 6.8 0.2. Photolytic degrad ation of this dye was found to be negligible in previous work by Khan [74]. Dye adsorption was quantified by preparing vials and rotating for 24 hours in a dark black box before analysis. Photocatalytic oxidati on experiments involved preparing samples in Teflon sealed borosilicate vials in the same manner as in the adsorption experiments except that an array of four-8 wa tt 365 nm peak output UV bulbs were placed 7.6 cm from the rotator for 3 hours. Ferrioxalate actinometry determined the total flux to be einsteins/min. RR concentration was measured by first vacuum filtering the samples using a 0.2 m nitrocellulose membrane filter and then determining the absorbance of the sample using a Hach DR/4000Uspectrophotometer at 538 nm. 51001.1 51

PAGE 52

A High Surface Area Magnetic Photocat alyst with Controlled Pore Size High Surface Area Magnetic Photocatalyst Synthesis Barium ferrite (Alfa Aesar) particles were coat ed with silica using a Stber technique [33]. Specifically, 4000 mg of BaFe12O19 were dispersed in 500 ml of 0.51:0.79:15:0.037 mol ratio of H20:NH4OH:EtOH:TEOS. The particles were stirred for 24 hr with a PTFE overhead mixer at 475 RPM. The resulting particles were dried at 65 C for 24 hr resulti ng in a total mass of approximately 4.40 g. The silica-coated barium ferrite particles and 18.9 g titanium dioxide (Degussa P25) were then dispersed in an acid ca talyzed silica sol-gel [75]. The sol-gel consisted of 113 ml of Nanopure water (18.2 M -cm), 225 ml ethanol, 156 ml TEOS, 18 ml 1M HNO3, and varying volumes of 3% hydroflu oric acid. The volume ratio of HNO3 to HF was used to denote the 3 types of high surface area magnetic photocatalyst (HSAMP) expl ored in this work. For example, HSAMP8 indicates that a volume ratio of 8:1 and a volum e of 2.25 ml HF. The gelation time varied by hydrofluoric acid concentrati on between 7 and 24 hrs. This time was determined as the time necessary to reach a viscosity such that visible particle movement was negligible and the vortex near the overhead mixer paddles was no longer apparent. The gel was then poured into plastic specimen containers, seal ed and aged at 25 C for 48 hrs then 60 C for another 48 hrs. To dry the gel and evaporate resi dual ethanol, the aged gel was dried in a PTFE container with a small hole on the top. The dryi ng followed a 3 C ramp to 103 C, held for 18 hr, then followed another 3 C ramp to 180 C and was held for 6 hr. After drying, the xerogel was heat treated with a 5 C ramp to 450 C, held for 1 hour and cooled naturally. The resulting xerogel was ground with a mortar and pest le and wet sieved to retain the 45-90 m fraction. The powder was dried at 180 C and stored under vacuum until use. 52

PAGE 53

Characterization Characterization techniques included scanning electron microscopy (SEM), x-ray diffraction (XRD), diffuse reflec tance, vibrating sample magnetometry (VSM), and inductively coupled plasma (ICP) analysis. Scanning electron microscope images were obtained with a JEOL JSM 6335F with a field emission element at a 5 kV accelerating voltage. X-ray diffraction patterns were found using a Phillips APD 3720 with Cu K radiation with 2 of 20-80 with a step size of 0.02. Particle size comparisons were made using the Scherrer equation [76]. Vibrating sample magnetometry was performed using a Princeton Measurements Corporation MicroMag 3900. Finally, a Perkin-Elmer Plasma 3200 with optical emission spectrometer (OES) was used for ICP analysis of the HSAMP. Assessment of Photocatalytic Oxidation Phenol was chosen as a surrogate contamin ant due to its extensively studied reaction mechanism and the widespread use for assessme nt of photocatalytic oxidation. A 6.5 mg/l phenol solution was prepared with Nanopure water (18.2 M /cm) and reagent grade phenol (90%, Fisher). The dissolved oxygen was measur ed to be 7.5 0.2 mg/l using a YSI 52 dissolved oxygen meter and was not augmented with an external oxygen supply during experiments. The solution pH was unadjusted and measured to be 6.8 0.2 prior to use during experimentation. Phenol adsorp tion experiments were carried out by preparing Teflon sealed borosilicate vials with specified masses of HS AMP and 40 mL of phenol solution and rotating for 24 hours in a dark black box. Vials were prepared the same manner for the photocatalytic oxidation experiments except that an array of four-8 watt 365 nm peak output UV bulbs were placed 7.6 cm from the rotator for various times. Ferrioxalate actinometry determined the total flux to be einsteins/min. Subsequent to reco vering the HSAMP using a magnetic decantation technique, phenol concentrati on was determined with a GC/MS (Varian 51001.1 53

PAGE 54

3900/2100T) using a solid phase microextraction (SPME) sampling technique with an 85 m polyacrylate coated fiber assist ed by addition of reagent grade sodium chloride. Non-purgeable dissolved organic carbon (nDOC) was measured on a Tekmar Dohrmann Apollo HS 9000 with autosampler. Purging and mixing was accomplished by sparging the sample, which had been pH adjusted with phosphoric acid, with zero grade ai r for 3 minutes. The relative standard deviation for (RSD) for replicate analysis of sample s was less than 5% for both reported phenol concentrations and nDOC values. 54

PAGE 55

CHAPTER 4 INCORPORATION OF A PERMANENT MAGN ET AND INFLUENCE OF BALL MILLING ON MAPR PERFORMANCE The magnetic core of a magnetic photocatal yst forms the basis for the structure and growth of subsequent coatings with silica and titania. Ball milling provided a method for reducing the size of larger partic les. In addition, while the magne tic field initially consisted of only an oscillating component for agitation, experimentation with other external magnetic fields including electromagnets and severa l types of permanent magnets resu lted in the discovery of the enhancement effect possible with a permanent magnet located above the reactor. Phenol Photodegradation In contrast with previous experiments [72], th e current fed to the solenoids in series was held constant and the correspondi ng photocatalytic activity was also found to remain relatively constant (Figure 4-1). The composite catalyst, ir radiated by the 8-watt 254 nm lamp, was able to decrease the 10 mg/L initial phenol concentratio n by an average of 92.2% after one hour. The frequency range was from 0 to 80 Hz in incr ements of 10 Hz. The nDOC decrease was an average of 50.0% by carbon. The lower removal of carbon than phenol is thought to be the result of formation of oxidation byproducts from inco mplete mineralization. Since the treatment objective is defined in terms of a final nDOC goal of 0.5 mg/L, the nDOC value is most important. 0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9 1 020406080100Frequency (Hz)Normalized Concentration Decrease nDOC Phenol 55

PAGE 56

Figure 4-1. Phenol photodegrad ation in a quartz reactor. Additional experiments were performed to de termine if the consistent photocatalytic activity was due to a high concentration of catalyst, so 50 mg was used for a 30 minute experiment instead. As expected, the decrease in the phenol concentra tion was significantly reduced, but the photocatalytic activity remained relatively constant through the frequency range (Figure 4-2). 0 0.1 0.2 0.3 0.4 0.5 02040608010Frequency (Hz)Normalized Concentration Decrease0 Figure 4-2. Phenol concentra tion decrease vs. frequency fo r lower catalyst loading. Ball-Milling Impact on Catalyst Performance The raw material for catalyst synthesis is ball-milled barium ferrite according to the procedure described in previous work [72]. The pr ocess involves rolling a jar filled with water, large barium ferrite particles and zirconia grinding media. Collisions between the media, walls of the jar, and barium ferrite break the material into progressively smaller pieces. If the barium ferrite loading is too high, the milling process will be incomplete and both the particle size distribution and morphology will be drastically different. In the past, 3 grams of barium ferrite was used for each ball-milling batch. To increase output, 7 grams of barium ferrite was used with the same ball-milling and catalyst manufacturing procedure. The resulting change in 56

PAGE 57

photocatalytic activity was quant ified by decolorizing a 10 mg/L solution of Procion Red dye with catalyst synthesized using both the 3 gram and 7 gram ball-mill batches (Figure 4-3). 0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9 1 7 g 17 g 27 g 33 g 13 g 23 g 3Ball-Mill Type Batch NumberNormalized Color Decrease Figure 4-3. Normalized color decrease (1-C/Co) by catalyst synthesized us ing barium ferrite ball milled in different quantities. There is clearly a dramatic impact of the raw material on the photocatalytic activity of the composite catalyst with an average increase of 70% for the material ball-milled in smaller batches. Scanning electron microsope images of the finished catalyst show a more angular morphology indicating a lack of coating by silica and titania on the 7 g batches (Figure 4-4). In support of the supposition th at the change in morphology i ndicates a lack of coating by silica and titania, EDS spectra indi cated a decrease in silicon and a slight decrease in titanium on the surface of the catalyst using the different raw material (Figure 4-5). Previous work has shown that the silica layer is critical to prevent photodissolution at the titanium dioxide magnetic core junction [17] thus the decrease in silicon indicated by the EDS spectra could mean that there is an increase in photodissolution, which might expl ain the decreased performance shown in Figure 4-3. 57

PAGE 58

Figure 4-4. Scanning electron microscope images of catalyst prepared using different ballmilling procedures 7 g batch, A and 3 g, B NOTE: the magnification is different for each image. Figure 4-5. Energy disper sive x-ray spectroscopy spectra of catalyst prepared using different ball-milling procedures 7 g batch (upper) and 3 g on (lower). ESM Estimation The NASA Specialized Center of Research and Training (NSCORT) at Rutgers University in New Jersey provides an example calculation for a subsystem based on the ISS data from 58

PAGE 59

mission documents. Although the data are presented for a 6-member crew for a 3650-day mission, one can calculate the equivalent system s mass (ESM) for a 4-member crew for 100and 1000-day missions. Similarly, based on reasonable estimations for the MAPR one can calculate the ESM for these units. The ESM of the MAPR was estimated to be 289 and 380 kg for 100and 1000-day missions (Table 4-1). Since the small mass, small volume, and essentially no requirement for crew time or consumables (except potentially UV lamps) significantly reduce the MAPRs ESM, a method to control power consumption could furt her accent the benefits of this technology. It is difficult to compare a single water recove ry system to an entire water treatment train with respect to ESM because the calculation fo r the post processor alone does not account for systems integration. Since this task is not simplified by the data provided on the NJ-NSCORT website, the ESM for an Integrated PhotoCatal ytic Water Recovery System (IPCWRS) based on Honeywells cascade distiller and the MAPR is provided for compar ison purposes. This system consisted of a urine preprocessor, cascade disti ller, and MAPR. These data are presented in Table 1 and clearly show that the IPCWRS ESM is anticipated to be considerably less ESM than the ISS baseline. Although all of these calcula tions were estimates, it was proposed that the major difference between the ISS and IPCWRS was that the IPCWRS significantly reduced logistics, mass, volume, and power consumption. Table 4-1. Equivalent systems mass (kg) fo r international space station baseline and photocatalytic systems. 100 Day mission 1000 day mission ISS Baseline 1365 3898 MAPR 289 380 IPCWRS 886 1207 59

PAGE 60

CHAPTER 5 ENHANCED PHOTOCATALYST SYNT HESIS AND REACTOR PARAMETER EXPLORATION While the previous chapter introduced impr ovements in basic parameters including photocatalyst synthesis and magnetic field configuration, this chapter is focused on a more detailed assessment of reactor parameters and further explores the influence of photocatalyst synthesis including double co atings and titanium di oxide particle size. Catalyst Synthesis The surface morphology of the ba rium ferrite core indicated sharp edges and a dominant flat hexagonal shape. The magnetic property cause d the particles to aggl omerate into stacks. After the catalyst was coated with double layers of silica and titania, the sharp edges are less evident. In addition, the agglomerates are more irregularly shaped possibly due to silica necking between the particles of barium ferrite. A superconducting quantum interference device (SQUID) analysis of the barium ferrite core and the composite photocatalyst demonstrated several important features of the catalyst [72]. First, the catalyst and barium ferrite had a clear hysteresis loop indicated by the lag in magnetism of the sample with changes in applied magnetic field intensity. This hysteresis loop is indicative of a ferromagnetic material possessing a remanent magnetization as was expected for barium ferrite. This remanent magnetization wa s approximately 23 emu G/g for the composite photocatalyst and 24 emu G/g for the barium ferri te core. The SQUID anal ysis confirmed that coating the barium ferrite caused only a slight decrease in remanent magnetization and the ferromagnetic property was preserved. Phenol Photolysis Previous studies have indicat ed that phenol undergoes photolys is when exposed to 185 or 254 nm irradiation [77]. Similarly, 254 nm has been found to be more effective for mineralizing 60

PAGE 61

phenol [13]. In our experiments, phenol showed no photolytic mineralization during seven hours of exposure from three 365 nm lamps using a quar tz reactor. Therefore da ta shown herein are a result of photocatalysis and not photolysis. Indeed, a full-scale MAPR system may take advantage of photolytic degradati on, yet it is important here to distinguish betw een photolytic and photocatalytic effects because the two reaction mechanisms may produce different byproducts. Phenol Adsorption It is important to quantify the extent of adsorption to the catalyst in the absence of UV to isolate what phenol and nDOC concentration changes may be due to oxidation. For this experiment, 100 mg of catalyst wa s agitated with 1 A current at 60 Hz signal frequency. The data indicate that phenol concentration and nDOC did not change over a period of 7 hours. Phenol and nDOC: Effect of Lamp Number and a Permanent Magnet Initial experiments employed 100 mg of the composite photocatalyst, 1 A current, and 60 Hz solenoid frequency. When the permanent magnet was present, the phenol concentration after 5 hours was not detected via GC analysis (Fig ure 5-1), while close to 70% of the nDOC remained (Figure 5-2). Without the permanent magnet (NPM) and with one 365 nm lamp (1L) the nDOC concentration decrease was less than the variation in replicate analysis of samples (Figure 5-2), but the phenol concentration at 7 hours decreased to 29% of its original value (Figure 5-1). Therefore, a larg e fraction of the phenol was simp ly being converted to byproducts. In order to increase the rate and efficiency of photocatalysis, two 365 nm lamps were added (i.e., for a total of 3 lamps or 24 W). Although the de crease in phenol concentr ation after 7 hours was approximately the same with and without the pe rmanent magnet (Figure 5-1), the initial pseudofirst order rate constant for phenol ox idation was 3 times greater (0.80 min-1 vs. 0.27 min-1) with the permanent magnet. Data for nDOC (Figure 52) indicated that after 7 hours, 0.6 mg/L C 61

PAGE 62

remained with a permanent magnet whereas when it was not present during the experiment, 3.1 mg/L was still remaining. The improvement in phot ocatalytic degradation from the fixed magnet is believed to be due to its ab ility to counteract grav ity with a magnetic field gradient. Under the influence of this magnetic field gradient, the cat alyst was more effectively spread throughout the reactor thus increasing its exposure to UV. 0 2 4 6 8 10 12 01234567Time 8 ( hr ) Phenol Concentration (mg/L) WPM, 1L NPM, 1L WPM, 3L NPM, 3L Figure 5-1. Comparison of nu mber of UV lamps and presence of permanent magnet for degradation of phenol in the MAPR (W PM = with permanent magnet, NPM = no permanent magnet, 1L = 1-365 nm lamp, 3L = 3-365 nm lamps). 0 1 2 3 4 5 6 7 8 9 02468 Tim e ( hr ) nDOC (mg/L C) WPM, 1L NPM, 1L WPM, 3L NPM, 3L Figure 5-2. Non-purgeable dissolved organic carbon concentration as a function of number of lamps and presence of permanent magnet in the MAPR (WPM = with permanent magnet, NPM = no permanent magnet, 1L = 1-365 nm lamp, 3L = 3-365 nm lamps). 62

PAGE 63

Catalyst Mass Optimization The quantity of catalyst in the reactor was optimized for phenol and nDOC reduction using 2 hours of exposure with one 365 nm lamp. With a 365 nm lamp, which was shown earlier not to cause photolytic mineralization, the phenol re moval was greater than nDOC removal for the entire range of catalyst mass (F igure 5-3). This was most likely an indication that several oxidation byproducts remain in so lution after 2 hours of irradiation. The trend in phenol removal showed a slight increase as mass increased up to 400 mg and remained constant at higher mass loadings. The nDOC trend demonstrated an increase in removal with mass until 400 mg where there was a peak and then a descent as the mass in creased further. This optima is believed to be the result of two phenomenon. An increase in phenol and nDOC re moval is expected with an increase in TiO2 concentration (as a functi on of catalyst mass increasing) due to an increase in the number of reaction sites and thus the quantity of hydroxyl ra dicals produced. A competing phenomenon is expected as the catalyst concen tration reaches a level where the UV light is blocked, reducing the proportion of catalyst receiving photons of sufficient energy to generate hydroxyl radicals [78]. 0 0.2 0.4 0.6 0.8 1 02004006008001000Catalyst (mg)Normalized Concentration Decrease nDOC Phenol 63

PAGE 64

Figure 5-3. Catalyst mass effect on phenol photocatalysis in th e MAPR (Normalized Concentration Decrease 0 0C CC ). Solenoid Frequency Optimization The sinusoidal signal frequency fed to the soleno id affected the rate at which the magnetic field gradient, and thus force on the catalyst, alternated directi on [72]. This oscillating force caused the catalyst to move within the reactor thus increasing the mass transfer of phenol and photodegradation byproducts to the TiO2 on the surface of the composite catalyst. For this experiment, each data point represents a variatio n in frequency where the current was fixed at 1.25 A, and 100 mg of catalyst was irradiated for 2 hours with three 365 nm UV lamps. The data demonstrate that nDOC is eliminated most eff ectively between 20 and 80 Hz where the variation in removal over the frequency range is less th an the sample RSD from the methods section (Figure 5-4). The elimination of oxidation byproduc ts, which is indicated by nDOC decreases as frequency increases beyond 80 Hz. This effect may be due to a decrease in the ability of the catalyst to effectively respond to changes in the magnetic field at higher frequencies. As the catalyst motion response is diminished, there wi ll be a corresponding decrease in mass transfer and thus photocatalytic degradation. In contrast to the nDOC result, phenol removal showed a less dramatic decrease at higher frequencies. This discrepancy in oxidation byproduct removal and phenol removal may be due to a higher ma ss transfer dependence of byproduct oxidation over phenol oxidation. This would mean that the first few steps of phenol oxidation to hydroxylated compounds [79] occur at the same rate for higher frequencies and lower mass transfer, but subsequent reaction steps such as loss of aromaticity and mineralization occur at a slower rate because they may be more dependent upon mass transfer Since the catalyst is heat treated to 500 C during synthesis, an increa se in hydrophobicity is lik ely to occur due to 64

PAGE 65

dehydroxylation of the silanol groups [80], t hus decreasing adsorption of polar compounds (compounds with more hydroxyl groups such as th e oxidation byproducts). Finally, at 0 Hz (a control with the alternating magnetic field absent and permanent magnet present), the least photocatalysis was observed confirming that the presence of an alternating magnetic field enhanced photocatalysis. 0 0.2 0.4 0.6 0.8 1 0 50 100 150Frequency (Hz)Normalized Concentration Decrease nDOC Phenol Figure 5-4. Solenoid signal fre quency effect on phenol photocatalys is in the MAPR (Normalized Concentration Decrease 0 0C CC ). Solenoid Current Optimization The magnetic flux of the alternating magnetic fi eld generated by the solenoid is directly proportional to current according to the Biot-Sav art law [72]. As such, the increase in magnetic flux is expected to correlate with an increase in catalyst motion and thus mass transfer of phenol to catalyst. The experimental data agreed with th is expected increase from 0 to 1 A; there was a corresponding increase in both reduction of nDOC and phenol concentration (Figure 5-5). As seen in previous experiments, the normalized c oncentration decrease for phenol was higher than nDOC for all conditions except 0 A. This is be lieved to be the result of organic oxidation byproducts. As the current increased above 1 A, the photocatalytic degradation remained 65

PAGE 66

constant. This change in the trend indicates that the system performs optimally with 1 A current as further increases in current increase ener gy requirements and heat generation without increasing the extent of photocatalytic oxidation. 0 0.2 0.4 0.6 0.8 1 00.511.522.5Current (A)Normalized Concentration Decrease nDOC Phenol Figure 5-5. Solenoid signal cu rrent effect on photocatalysis (Normalized Concentration Decrease 0 0C CC ). 66

PAGE 67

CHAPTER 6 MAGNETIC PHOTOCATALYST CHAR ACTERIZATION AND REACTOR PERFORMANCE USING AES ERSATZ WATER For long-term space missions, the MAPR provided a technique for reducing nonpurgeable dissolved organic carbon in the air eva poration system effluent. In addition, magnetic photocatalyst characterization provided informa tion about morphology, primary particle size, and the influence of a magnetic field on the formation of assembled structures. Photocatalyst Characterization It is readily apparent from visual observation of the MAPR when the magnetic field was present that assembled structures continually formed and broke apart. It was not, however, known what these structures looked like on a micr on scale. The assembled structures of the catalyst determined how much of the total available surface area could be exposed to UV, and how much was readily accessible for contaminant ad sorption. The structures formed were due to forces applied by both the static and oscillating components of the magnetic field in addition to the interparticle force due to the remanent magnetization. According to observation, the static field tended to induce very ordere d needle-like structures. Simila rly, observation indicated that the oscillating magnetic field cause d the assembled needle-like structures to break and re-form. The rate of this breaking and re-formation was apparently determ ined by the oscillating frequency, with higher frequencies resulting in higher rates of formati on and subsequent break up. By observing the structures with SEM before and after exposure to th e magnetic field, it was possible to see the development of structures on a micron scale that are not apparent with visual observation. In the SEM images before exposure to the magnetic field (Figure 6-1), the catalyst appeared relatively planar and ra ndomly distributed. In contrast, in the image taken after the magnetic field exposure the catalyst appeared to form a hollow assembled structure about 20 m in diameter that comes out of the plane (Figure 6-2). Since this structure appeared to be only a 67

PAGE 68

few catalyst particles in thickness, it would most likely not interfere with absorption of UV light or adsorption of contaminants with respect to much larger structures. Also, its hollow nature would provide additional available surface area for adsorption, which could subsequently be oxidized as the structur es break and re-form. XRD of the TiO2 powder allowed an assessment of both the crystallinity and particle size. For the highest photoactivity, it is important that the TiO2 be crystalline and primarily anatase phase because it has a higher photoactivity than rutile or other crystalline phases. Previous authors have found that TiO2 coatings using organic precursor s without adding acid to catalyze the particle formation produce primarily amorphous structures [25]. Since the method used for the TiO2 synthesis in this work was the result of an organic precursor (titanium n-butoxide) and did not utilize acid, it was expected that the resul ting coating would be primarily amorphous. From Figure 6-3 A, the XRD show s very broad low peaks that are not much higher than the noise. This type of XRD output is associated with materials that do not have significant crystallinity, providing eviden ce for confirmation of the cont ention that the coating would initially be amorphous. To improve photoactivity, the double coating was heat -treated with a 5C ramp to 500C where it was held for 1 hour. As can be see in Figure 6-3b, the XRD for the heattreated material has much higher and narrower peaks, which indicates an increase in crystallinity. Since the most intense anatase peak was at 25.4 2 it was used to estimate the crystalline size using the Scherrer equation (eq. 6-1) cos B k t (6-1) Using a shape factor (k) of 0.9, of 0.1541 nm, B of 0.0140 radian 2 of 25.4, the particle size was estimated to be 10.1 nm. 68

PAGE 69

Figure 6-1. Scanning electron microscope imag e of magnetic photocatalyst (not exposed to magnetic field). The XRD analysis indicated that the double coating method and variation lowering the water to alkoxide ratio (130:1 to 110:1) from previous work increas ed the particle size from 5.9 nm [42]. Since the nanoparticles from the previo us coating may have act ed as nucleation sites during the double coating procedur e, it was important to assess a ny changes in crystallinity or particle size. This increase in particle size likely decreased the specific surface area of the photocatalyst, yet may have avoided quantum confinement effects observed in smaller TiO2 nanoparticles that can lead to a decrease in photoactivity due to bandgap widening [9]. 69

PAGE 70

Figure 6-2. Scanning electron microscope image of magnetic photocatalyst (exposed to magnetic field). Figure 6-3. X-ray diffraction patte rn of synthesized titanium di oxide before (A) and after (B) heat treatment. 70

PAGE 71

Magnetic Field and Catalyst Mass Impact on Extent of Photocatalytic Oxidation of AES Ersatz The magnetic field was used to control the photocatalyst motion and distribution in the reactor. By modifying the charac teristics of the magnetic field, it was possible to assess their impact on the extent of photocatalysis. The fr equency of the oscillat ing magnetic field was controlled by modifying the fr equency output of the function generator while maintaining a consistent alternating current (1 A) to the so lenoid array (Figure 6-4) Since the oscillating magnetic field is applying a force on the particles, it is hypothesized that the rate of change or frequency of the oscillation would determine both the distance traveled and the velocity of the photocatalyst particles in solution. The data indi cate that there was in in crease in performance as estimated by normalized concentration decrease between 0 and 40 Hz where the highest peak occurred. The decrease above 40 Hz was not obser ved in a previous study using phenol [19] but similar frequency-related performa nce behavior occurred in work focused on dry particle coating by Ramlakhan [21]. This work used oscillating magnetic fields to agitate magnetic particles, which induced dry impaction coating. Above 80 Hz, there was a slow decrease in performance, yet it was not as dramatic as the previous st udy with phenol. There are several possibilities for explaining this change in perfor mance with respect to the phenol work. First, the nDOC in AES ersatz is comprised primarily (74%) of dextra n, a polymer, but also has inorganic compounds that were not present in the previous work. Al so, the light source in this work was 254 nm as opposed to the 365 nm source used previously. The decrease in wavelength corresponds to an increase in the energy per photon, which can i nduce photolysis of certain compounds and also production of ozone that were shown to either not be present or to not have an impact with the 365 nm light sources. 71

PAGE 72

0 0.1 0.2 0.3 0.4 0.5 0 50 100 150Frequency (Hz)Normalized Concentration Decrease Figure 6-4. Normalized concentration decrease vs. oscillating magne tic field frequency. The photocatalyst mass determined the numbe r of potential photo active reaction sites available and the light distributi on within the reactor. Since the core material was barium ferrite, which absorbs UV light, an increase in catalyst mass corresponded to an increase in the light absorption by the magnetic core, which would ha ve a detrimental effect on the extent of photocatalysis. In contrast, the increase in reaction sites would have a potentially beneficial effect on photocatalytic degradation, given that these sites were ac tually irradiated with UV. By incorporating both of these trends, it is clear th at for most photocatalytic reactors there should be an optimal catalyst mass whereby the number of reaction sites is maximized while the light absorbing effect is controlled. In the experime nts performed to assess this effect, evidence to support these hypotheses was found (Figure 6-5). Th e normalized concentration decrease for 50 mg and 100 mg catalyst mass loadings were found to be within the relati ve standard deviation (RSD) of replicate analysis of samples. The reason for this equality may be due to an increase in photolytic and ozone effects at 50 mg because of the lower light absorp tion by the photocatalyst 72

PAGE 73

cores and an increase in photoactiv e sites to compensate for the greater light absorb ing effect at 100 mg. The steady increase in the extent of photocatalys is continued with increase in catalyst mass continued until 400 mg where it reached a p eak. As the mass increased further, the extent of photocatalysis decreased, which was consistent with th e initial hypothesis. 0 0.1 0.2 0.3 0.4 0.5 02004006008001000Catalyst mass (mg)Normalized Concentration Decrease Figure 6-5. Normalized concentrat ion decrease vs. catalyst mass. The current fed to the solenoid provided an oscillating force on the particles and was critical to improving mass transfer of pollutants to the photocatalys t with respect to stationary particles. The Biot-Savart law suggested that the force exerted on the particles is proportional to the current in the solenoid. For long-term space mi ssions, the implication of increasing current to an advanced life support system is disadvantageous because it will likely be at the expense of other support systems or possibly re search activities. It is criti cal that the power used by these systems is adequate to provide satisfactory pe rformance, but not greater than necessary to conserve power. In other work [19], it was f ound that there was an optimal operating current of 1 A and further increases in current did not l ead to corresponding incr eases in photocatalytic 73

PAGE 74

performance. A similar trend was found with an increase in the extent of photocatalysis between 0 and 1 A, while further increases in current did not have an impact greater the sample analysis RSD (Figure 6-6). 0 0.1 0.2 0.3 0.4 0.5 00.511.52Solenoid current (A)Normalized Concentration Decrease Figure 6-6. Normalized concentration decrea se vs. oscillating magnetic field current. 74

PAGE 75

CHAPTER 7 PHOTOCAT EXPERIMENTS USING GRAYWATER FORUMULA While previous chapters involve d a MAPR, the work in this ch apter explored the use of a Photocat or slurry-based photocatalytic reacto r with slurry separati on device for oxidation of organic contaminants. Instead of using the air evaporation system effluent, the Photocat was assessed for its ability to act as a post-processor after a biological water processor. The influence of bicarbonates on kinetic photocatalytic degradation rates was explored. Biological Water Processor Effluent The biological water processor effluent had an initial nDOC of approximately 51 5.1 mg/L C. The treatment objective is to meet or exceed NASA specification of 500 g/L C nDOC for potable water. Ammonium bicarbonate had a concentration of 724 mg/L in the BWP. The Photo-Cat was operated with and without amm onium bicarbonate to determine the impact of high alkalinity on photocatalysis. The Photo-Cat ran in batch mode with all contaminants and TiO2 circulating through the system at time zero when the UV lamps were turned on. The data are plotted as nDOC vs. kWh/m3 or cumulative energy per unit volume treated. The kWh/m3 value for each data point was calculated by dividing the produ ct of total UV lamp power (t he sum of all 75 watt bulbs used) and cumulative time by the total volume of c ontaminated water in the system. This power is necessary for testing with UV bulbs howev er; the flight hardware will use solar UV significantly reducing the power requirements and ultimately the ESM. The system was tested with and without ammonium bicarbonate to de termine its effect on the rate of nDOC disappearance. With ammonium bicarbonate incl uded in the BWP formula the endpoint nDOC was 0.390 mg/L C. Without ammonium bicarbonate, the endpoint is 0.360 mg/L C (Figure 7-1). 75

PAGE 76

If the same data are plotted with a logari thmic scale on the nDOC axis, the change in pseudo-first order rate constant becomes more clear (Figure 72). The reaction rate for nDOC disappearance in the biological effluent without ammonium bi carbonate is approximately 40% greater than when it is present, or 0.0014 min-1 and 0.0010 min-1 respectively (Figure 7-3). 0 10 20 30 40 50 60 05001000150020002500300035004000 kWh/m3nDOC (mg/L) Without ammonium bicarbonate With ammonium bicarbonate Figure 7-1. Biological water pro cessor effluent nDOC vs. kWh/m3. -2 -1 0 1 2 3 4 5 01000200030004000500060007000 Time (min)ln [nDOC (mg/L)] Without ammonium bicarbonate With ammonium bicarbonate Figure 7-2. Biological water pro cessor effluent nDOC vs. kWh/m3 (natural log scale). 76

PAGE 77

y = -0.0014x + 3.2407 R2 = 0.8848 y = -0.0010x + 3.4053 R2 = 0.8832 -2 -1 0 1 2 3 4 5 0500100015002000250030003500 Time (min)ln[nDOC (mg/L)] Without ammonium bicarbonate With ammonium bicarbonate Figure 7-3. Biological water pro cessor effluent nDOC vs. kWh/m3 to 3500 min (natural log scale). 0 3 6 9 12 15 18 0500100015002000250030003500 Time (min)nDOC (mg/l C ) Figure 7-4. Low-flow nDOC data for simulated gr aywater without bicarbonate. 0 5 10 15 20 25 30 35 40 0500100015002000250030003500 Time (min)nDOC (mg/L C ) Figure 7-5. Low-flow nDOC data for simulated graywater with bicarbonate. 77

PAGE 78

0 0.5 1 1.5 2 2.5 3 050010001500200025003000 Time (min)nDOC (mg/L C ) Figure 7-6. Low-flow n DOC data for biological water processor effluent with bicarbonate. Graywater Testing Although batch testing provides gr eater reaction kineti cs than single pass testing, a singlepass, low flow mode was used for testing graywa ter. The purpose of trea ting the graywater in a low flow mode is to destroy the surfactant prope rty of the shampoo so that it does not build up in the system. This is in contrast to the batch mode testing where the surfactant concentration is at a maximum at time 0. In this type of experiment an equilibrium effluent nDOC is obtained once the system is at steady-state. This value varied by input water type and presence of ammonium bicarbonate. For the simulated graywater without ammonium bicarbona te, at a flow rate of 4.4 mL/min (initial TOC of 447 mg/L) the nDOC from the Photo-Cat discharge increased from 0 to 16 mg/L C in approximately 2500 minutes at which time it was at steady state. (Figure 7-4). In contrast, when ammonium bicarbonate is presen t in the graywater the nDOC increased at a slightly faster rate unt il it reached a final concen tration of 34 mg/L C (Figure 7-5) due to the hydroxyl radical scavenging by th e bicarbonate ion [81]. As a comparison, the BWP with ammonium bicarbonate had a similar linear increase in nDOC, but its final value is only 2.5 mg/L C (Figure 7-6) due to the lowe r feed concentration of 50ppm nDOC. More testing is required to further quantify si ngle-pass destruction of organic contaminants in graywater. Additional treatment is required to reduce the TOC of the graywater below 78

PAGE 79

0.5ppm. It is also recommended that other type s of soaps/shampoo material s be investigated for photocatalytic efficiency including NASA Whole Body Shower Soap (Geropon TC-42). Choosing a soap that mineralizes read ily in the system is desirable. Size Requirement Estimation Given the UV output of the Photocat system us ed for this paper and an estimate of the UV flux on the moon it is possible to estimate the size of a solar collector of similar flux. According to the manufacturer, each lamp pr ovides approximately 25 W UV radiation to the slurry or 200 W total. In addition, if we a ssume that the solar UV radiation on the moon is primarily below 380 nm as is necessary for photo catalytic oxidation using titanium dioxide there would be a flux of 101 W/m2 [82]. Using both of these valu es and assuming the quantum efficiency of the lunar photocatal ytic reactor is equivalent to it s terrestrial counterpart, the lunar solar collector would need to be approximately 1.98 m2. 79

PAGE 80

CHAPTER 8 PHOTOCATALYST COMPARISON USING THE PHOTOCAT REACTOR In the previous chapter Degu ssa P25 was used as a photocatalyst. Degussa P25 is a benchmark material that is frequently used in literature to establish improvements due to variations in photocatalyst composition due to incorporation of a support material or doping with carbon, nitrogen, or metals. This chapter focused on the performance of a silica-titania composite with respect to Degussa P25. Composite Characterization The silica-titania composite was synthesized by impregnating a mesoporous silica gel with Degussa P25 and its morphological properties were elucidated by examining secondary electron SEM images. In Figure 8-1 it is clear that the ph otocatalyst that had been ground and sieved was composed of larger 45-90 m particles whos e surface is clearly covered by much smaller particles. The smaller particles provide an inco mplete and sporadic coverage of the surface, which may be the result of van der Waals forces between the smaller particles and the larger plate-like structures formed duri ng the grinding and sieving process. SEM secondary electron images do not provide any direct information about composition; however, if one examines the b ackscatter electrons it is possible to see average atomic mass contrast. Hence, if a material is composed of two elements of sufficiently different atomic number and sufficiently large e nough domain size, it is possible to distinguish between regions of each element. While secondary electrons ar e liberated from the specimen, backscatter electrons are the original beam electrons that are deflected back to the polepiece where a solidstate detector is located. The amount of energy which the electr ons possess is a function of atomic mass and can be measured by the solid-state detector. 80

PAGE 81

Figure 8-2 is the result of imaging the 45-90 m composite photocatalyst using secondary electrons (top) and backscatter electrons (bo ttom). The secondary electron image shows some charging behavior at both edge sites and in pa rticular sites on the su rface of a few of the particles. This behavior was expected due to the non-conductive nature of the silica in the photocatalyst. The same type of partial coverage of the larger plate-like structures by the smaller particles seen on the higher magnification image in Figure 8-1 is evident. The backscatter image Figure 8-1. Scanning electron microscope im age of 45-90 m silica-titania composite photocatalyst surface. clearly showed regions of varying intensity electron energy indicating regions composed of different average atomic mass elements. Since the higher atomic number elements absorb less of the beam electron energy, the portions of the imag e that are brighter indicate a higher average atomic mass region. Based on the composite photocat alyst synthesis method, it is clear that the elements present should include titanium, silic on, and oxygen with atomic masses of 47.87, 28.09, and 16.00 g/mol respectively. From the atomic mass data, it was clear that if the contrast is sufficient, regions rich in ti tanium dioxide should appear br ightest. An examination of the backscatter image indicates that there are several bright spots on the surfa ce of the particle that 81

PAGE 82

are approximately 1 m in diameter. To further verify whether the spots were indeed an indication of titanium dioxide an EDS spectra was obtained for both one of the bright spots and a darker region nearby. Figure 8-3 shows the spectra of the region indicated by the circle in Figure 8-2, which is a relatively dark spot whereas the spectra in Figure 8-4 are indicated by the square, which is a relatively bright spot. A comparison of the two spectra showed that the peak for the alpha energy level for titanium is less than 500 fo r the dark region; it is greater than 1000 for the bright region. This increase in titanium indicat ed by EDS is further evidence that the regions indicated are titanium dioxide. Silica is likely present in both spectra because the interaction volume includes a portion of the silica matrix for both regions. It is important to understand how the titanium dioxide is distributed within the composite photocatalyst for optimization pu rposes. If the titanium dioxi de forms prohibitively large agglomerates, it is possible that the reactive sites near the center of these agglomerates will be physically inaccessible to either UV light or contam inants. Previous work with titanium dioxide slurries has shown that they also form agglomer ates in water such that the approximately 30 nm primary particles of Degussa P25 will form aggl omerates of about 1.34 m [81]. The backscatter electron images provide evidence supporting the fo rmation of agglomerates, but quantitative data concerning particle size is not easily determined due to the qualitative na ture of the backscatter technique and the inability to determin e precise size measurements using SEM. 82

PAGE 83

Figure 8-2: Secondary electron image and backsc atter complement of 45-90 m silica-titania composite photocatalyst surface. Figure 8-3. Energy dispersive x-ray spectro scopy of 45-90 m silica-titania composite photocatalyst surface (circle region). Figure 8-4. Energy dispersive x-ray spectro scopy of 45-90 m silica-titania composite photocatalyst surface (square region). 83

PAGE 84

Biological Water Processor Effluent The biological water processor effluent was a simulated formula that is intended to replicate the probable contaminants leaving a biological water processor. The formula contains polymers, volatile organic carbon compounds, and inor ganic salts. Of the constituents, dextran, a high molecular weight polymer (15k-20k g/mol) contributes almost 63% of the organic carbon content of the ersatz water. The primary tr eatment objective is to meet or exceed NASA specification of 0.5 mg/L nDOC for potable water. 0 0.2 0.4 0.6 0.8 1 1.2 0100020003000400050006000Time (min)Normalized nDOC Composite Degussa P25 Figure 8-5. Biological water pr ocessor effluent normalized nDOC vs. time (equal total mass basis). As a primary comparison, 10 grams of composite photocatalyst and Degussa P25 respectively were added to the Photocat after preparation of the BWP ersatz. It is readily apparent from Figure 8-5 that the decrease in nDOC happens at a faster rate and reaches a lower endpoint using the Degussa P25 w ith respect to the composite phot ocatalyst. Additionally, the samples from the experiment using the co mposite photocatalyst contained small opaque particles. 84

PAGE 85

y = -0.00049x + 0.05208 R2 = 0.98823 y = -0.00150x + 0.01428 R2 = 0.98880 -3 -2.5 -2 -1.5 -1 -0.5 0 0.5 0 500 1000 1500 2000Time (min)ln(nDOC/nDOCo) Composite Degussa P25 Figure 8-6. Biological water proces sor effluent pseudo-first order rate models. (equal total mass basis). The reaction rate is important because it will a ffect the reactor volume necessary to provide a particular conversion. If th e reactor is designed based on a plug flow model, which is reasonable given the proposed reactor design of a tubular reactor that is exposed to solar UV, then the reactor volume is inversely proportiona l to the reaction rate [83]. To quantify the difference in reaction rates, a pseudo-first orde r initial rate constant was determined for both experiments as seen in Figure 8-6. The slope of the line indicates the reaction rate determined by the model. In this case, the Degussa P25 has a reaction rate that is over three times that of the composite photocatalyst. It is im portant to note that the com posite photocatalyst was composed of approximately 70% silica and 30% titanium dioxide by mass. This means that the Degussa P25 had over three times the mass of TiO2 for the experiment shown in Figure 8-6. The increased mass also corresponded to an incr ease in potential reactive sites fo r oxidation of organic material and may have contributed to the discrepancy in pseudo-first order init ial rate constant. One method for estimating changes in aromatic carbon content is to measure the spectrophotometric absorbance of a solution at 254 nm. For the BWP ersatz, dextran, which is the primary contributor to nDOC is composed of aromatic monomer units that will absorb light at 254 nm. As dextran was removed from the efflue nt by adsorption or oxidation, the absorbance 85

PAGE 86

should also decrease given that no reactions are taking place which might produce UV absorbing species. The UV-254 data for the equal total mass basis comparison shown in Figure 8-7 illustrates a more similar trend for both photocatal yst types than that in Figure 8-5. A plausible explanation is that the dextran is not being oxidized as readily by the composite as Degussa P25, but rather is being adsorbed a nd possibly interfering with the subsequent photocatalytic oxidation of organic material. 0 0.05 0.1 0.15 0.2 0.25 0.3 0100020003000400050006000Time (min)UV-254 nm ABS Composite Degussa P25 Figure 8-7. Biological water processor effluent UV-254 vs. time (equal total mass basis). As a method to determine the efficacy of the hypot hesis that the deficiency in reaction sites was a cause for the diminished reaction rate fo r the composite, an experiment was performed using an equal mass basis of TiO2 (Figure 8-8). In contrast to the earlier experiment with equal total mass basis, the reaction rates were quite similar for the composite and Degussa P25. In fact, the estimated pseudo-first order reaction rate fo r the composite is 24% greater than for the Degussa P25 (Figure 8-9), which provides evidence to support the contenti on that providing an equal number of potential reaction sites would improve the reaction rate. A trend similar to that of the equal total mass comparison for the UV-254 absorbance can be seen in Figure 8-10. Since the UV-254 and the nDOC trends were similar for both photocatalysts, it is likely that the proposed oxidation interference effect observed in the previous experiment is not taking place. It 86

PAGE 87

is possible that the increased ra tio of reaction sites to dextran concentration for the equal mass TiO2 basis is sufficient to prevent an y significant interference effects. 0 0.2 0.4 0.6 0.8 1 1.2 0100020003000400050006000Time (min)Normalized nDOC Composite Degussa P25 Figure 8-8. Biological water pr ocessor effluent normalized nDOC vs. time (equal mass TiO2 basis). y = -0.00104x + 0.01787 R2 = 0.99425 y = -0.00129x + 0.10457 R2 = 0.99425 -3.5 -3 -2.5 -2 -1.5 -1 -0.5 0 0.5 050010001500200025003000Time (min)ln [nDOC]/[nDOCo] Composite Degussa P25 Figure 8-9. Biological water pro cessor effluent pseudo-first orde r rate models (equal mass TiO2 basis). 0 0.05 0.1 0.15 0.2 0.25 0100020003000400050006000Time (min)UV-254 nm ABS Composite Degussa P25 Figure 8-10. Biological wate r processor effluent UV-254 vs. time (equal mass TiO2 basis). 87

PAGE 88

While the BWP ersatz has an nDOC of 51 mg/L, the AES ersatz is only 9.5 mg/L. However, it also contains inorganic contamin ants including ammonium carbonate, which may have a similar detrimental effect on photocatalysis as demonstrated in previous work [14]. The primary contribution to nDOC in AES is dextran (74%) as it was in the BWP, yet the total dextran concentration was 15 mg/L whereas it was 98 mg/L for BWP. This decreased concentration was expected to have an infl uence on the reaction ra te comparison if the interference effect was occu rring. Figure 8-11 shows the result of the equal mass TiO2 basis experiment, which indicates that there is a significant improvement in the extent of photocatalysis over the experimental time peri od shown. Additionally, Figure 8-12 shows the result of pseudo-first order initial rate regr ession modeling. The modeling suggested that the reaction rate is almost twice as high for the com posite as for the Degussa P25. This increase in reaction rate was expected due to the synergis tic effect of the high specific surface area adsorbent combined with TiO2 that may have been obfuscated by the interference effect due to the dextran adsorption. -0.2 0 0.2 0.4 0.6 0.8 1 1.2 0 50010001500200025003000Time (min)Normalized nDOC Composite Degussa P25 Figure 8-11. Air evaporation system nor malized nDOC vs. time (equal mass TiO2 basis). 88

PAGE 89

y = -0.0026x + 0.0622 R2 = 0.993 y = -0.0051x 0.2673 R2 = 0.916 -2.5 -2 -1.5 -1 -0.5 0 0.5 0 100 200 300 400 500 600Time (min)ln(nDOC/nDOCo) Composite Degussa P25 Figure 8-12. Air evaporation system pseudofirst order rate models (equal mass TiO2 basis). 89

PAGE 90

CHAPTER 9 IMPACT OF HEAT TREATMENT AND COMPOSITION OF ZnO-TiO2 NANOPARTICLES FOR PHOTOCATALYTIC OXI DATION OF AN AZO DYE In contrast to previous work, this chapter focused on how the synthesis technique of a coprecipitated composite material impacted performa nce for oxidation of a textile dye. It explored the influence of elemental composition and heat treatment including calcination and hydrothermal techniques. Specifically, the influen ce of particle size, zeta potential and bandgap energy estimates were explored for their influe nce on photocatalytic oxidation and adsorption. Impact of Composition For hydrothermally prepared samples, the tre nd for the extent of photocatalytic oxidation shows that the highest activity occurs in the samples that are near 100% ZnO or TiO2 (Figures 91 and 9-2). Two parameters were chosen to asse ss the progress of photocat alytic oxidation: dye absorbance and non-purgeable dissolved organi c carbon (nDOC). Absorbance was measured at 538 nm, which is an indication of color. Since oxid ative attack of the azo bond is preferential, it is important to also follow the oxidation bypr oducts with non-purgeable dissolved organic carbon as these byproducts may be t oxic even if the solution appear s to be clear. Since it was demonstrated earlier that photolytic oxidation of the dye was neg ligible, it is thought that the primary mechanism for mineralization is photocat alytic oxidation. Bleaching of the dye as measured by absorbance changes was complete in the 5, 95 and 100% TiO2 samples of both heat treatments, and the 90% TiO2 hydrothermal treatment. The highe st extent of photocatalytic oxidation (PCO) was for the calcined sample with 100% TiO2, which still contained 30% of the original organic carbon, presumably as oxi dation byproducts. For these samples with 100% absorbance decrease, kinetic information could not be inferred, yet information about the extent of mineralization as determined by non-purgeab le dissolved organic carbon was used for 90

PAGE 91

comparison. Additionally, the phot ocatalytic activity was neglig ible for samples between 10 and 70% TiO2. Impact of Adsorption The relationship between adsorp tion capacity and photocatalysis is complex and there is not always a clear correlation between the tw o phenomena for material s with photocatalytic activity. For instance, previous work with activated carbon de monstrated an increase in adsorption of methylene blue dye on activated car bon that is thought to be due to a change in electron distribution in the dye molecule, which increases adsorption to the aromatic activated carbon basal planes [74]. The photocatalyst prepared using a ca lcination technique demonstrated a trend of low adsorption and low photocat alytic oxidation between 20 and 90% TiO2 (Figure 93). It was also interesting to note that the adso rption for the calcined sample roughly followed the zeta potential trend (Figure 9-4) and was highest at 95% TiO2 whereas for hydrothermally prepared it was relatively constant between 5 and 95%. These trends demonstrate that the higher adsorption capacity of the hydrothermally prepared materials did not tran slate to higher PCO. Influence of Zeta Potential Since the Procion Red is an anionic dye and its adsorption and photocat alytic activity have been proven to be related to the pH of solution for TiO2, [84, 85] it was expected that a similar trend might be seen for the mixed oxide material s observed in this work Electrostatic repulsion by the negatively charged photocat alyst at a given pH has a tende ncy to prevent close approach of dye molecules to the reaction sites on the particle surface. Sin ce the PCO reactions occur at or near the surface of the photocat alyst because of the high reac tivity of hydroxyl radicals and hence short diffusion distance, it is believed that phenomena which prevent this close approach such as a boundary layer diffusion resi stance[19] or elec trostatic repulsion [86] have a tendency to reduce PCO. Figure 9-4 shows the trend for zeta potential for the various mixed oxide 91

PAGE 92

photocalysts. Both photocatalysts demonstrate high negative zeta potentials for compositions between 20 and 70% TiO2. 0 0.2 0.4 0.6 0.8 1 0 20 40 60 80 100TiO2 (%)Normalized nDOC Decrease calcined hydrothermal Figure 9-1. Comparison of photocatalytic oxidation fo r calcined and hydrothermally prepared ZnO-TiO2 photocatalyst of varying composition. 0 0.2 0.4 0.6 0.8 1 0 20 40 60 80 100TiO2 (%)Normalized Absorbance Decrease calcined hydrothermal Figure 9-2. Comparison of color changes for calcined and hydrothermally prepared ZnO-TiO2 photocatalyst of varying composition. 92

PAGE 93

This region also corresponded to relatively lo w adsorption and PCO for calcined photocatalysts as was expected due to electrostatic re pulsion. Additionally, the higher performing photocatalysts had near neutral zeta potentia l, which corresponds to approximately equal quantities of positively and negatively charged sites on the surface. For calcined samples, the zeta potential was near neutral for the 0% TiO2 samples, but was slightly negative for the hydrothermal sample, which may explain the high er PCO for the calcined sample. For the 95 and 100% TiO2 samples, zeta potential was similar for both preparation methods and approximately neutral, which provided a more ideal environment for PCO. 0 0.05 0.1 0.15 0.2 02 04 06 08 0TiO2 (%)Normalized Concentration Decrease1 0 0 hydrothermal calcined Figure 9-3. Comparison of procion red MX -5B adsorption on ZnO-TiO2 photocatalyst prepared by calcination and hydrothermal heat treatme nt (normalized concentration decrease 0 0C CC. 93

PAGE 94

-100 -80 -60 -40 -20 0 20 02 04 06 08 0TiO2 (%)Zeta Potential (mV)1 0 0 Hydrothermal Calcined Figure 9-4. Comparison of zeta potential for a ZnO-TiO2 photocatalyst prepared hydrothermal and calcination technique with varying composition. Specific Surface Area Specific surface area is often a useful tool fo r assessing the photocatalytic activity of a material since the reactions likel y take place at or near the su rface. Higher specific surface areas correspond to more reaction sites available give n a constant surface density and quality of reaction sites. This trend is important, but since other variables such as zeta potential, quantum effects, and surface defects can al so be a function of particle si ze for nanometer sized particles, surface area is not often an ultima te determinant of PCO. From th e data that are shown in Figure 9-5, it is apparent that the highest specific surf ace area materials are the hydrothermally prepared photocatalyst with 70% TiO2. Interestingly, this does not corr espond to the highest photocatalytic activities or adsorption, which ma y be due in part to the highly negative zeta potential. Also, those photocatalysts with the hi ghest photocatalytic activity for hydrothermal and calcined have 94

PAGE 95

specific surface areas of 30.2 and 37.6 m2/g respectively, which was in the middle range of all specific surface areas. 0 50 100 150 200 250 0 25 50 75 100TiO2 (%)BET Surface Area (m2/g) Hydrothermal Calcined Figure 9-5. Comparison of BET specific surface area for of ZnO-TiO2 photocatalyst prepared by hydrothermal and calcination technique with varying composition. 0 1500 3000 4500 6000 7500 9000 203040506070802 ()Intensity (Counts/s) 5C 70C 95C 100C Figure 9-6. Comparison of XRD patterns of ZnO-TiO2 photocatalysts prepared using a calcination technique with varying composition. 95

PAGE 96

X-ray Diffraction X-ray diffraction can provide information about the crystalline structure, composition, and particle size of materials. For photocatalysis, this is important because the crystalline structure and particle size have been proven to have a strong influence on photo catalytic activity [9]. Researchers have demonstrated that calcination of TiO2 produced using alkoxide precursors can generate anatase or at higher temperatures rutile crystalline phase [9]. The anatase phase has a higher photocatalytic activ ity and sintering of TiO2 produces larger particle size rutile, which has a lower photocatalytic activity than anatase. This is in contrast to Degussa P25 where there is apparently a synergy between the anatase and sm aller rutile phase crystals produced by flame synthesis [87]. Figures 9-6 and 9-7 show the XRD patterns for several ZnO-TiO2 samples. The Figures show a similar pattern for both 5% TiO2 samples, with no strong indication of typical anatase or rutile TiO2 peaks. The pure TiO2 samples both have anatase 101 plane peaks at about 25.4 2 yet no rutile 100 plane peak s indicating that only the an atase crystalline phase is present. Using the Scherrer equation (eq. 9-1), particle size (t) can be estimated based on peak width (B). Given a shape factor (k) of 0.9, (wavelength of the CuK1 x-ray source) of 0.1541 nm, B (full peak width at half maximum correct ed for instrumental broadening) of 0.0065 and 0.0090 radian for pure TiO2 calcined and hydrothermally prepar ed samples respectively, and 2 (anatase 101 peak) of 25.4, cos B k t (9-1) The Scherrer equation indicates that particle sizes for the calcined and hydrothermally treated TiO2 were 21.1 and 16.0 nm respectively. The larg er particle size indi cates that sintering has taken place for the calcined TiO2, which may explain its more intense 101 peak. Either this increase in particle size or an increase in crys tallinity is believed to provide a reason for the 96

PAGE 97

higher PCO for the calcined photocatalyst. The 5H and 5C samples show characteristic peaks of ZnO at 31.5, 34.3, 36.1, 47.3, 56.0, 62.7, and 67.6 corresponding to (100), (002), (101), (102), (110), (103), and (112) planes respectivel y. Even though these samples had 5% TiO2, they showed no evidence of the characteristic peaks of either anatase or rutile crystalline phases. 0 1500 3000 4500 6000 7500 9000 203040506070802 ()Intensity (Counts/s) 5H 90H 100H Figure 9-7. Comparison of XRD patterns of ZnO-TiO2 photocatalysts prepared using a hydrothermal technique with varying composition. 0 2 4 6 8 10 2.52.72.93.13.33.53.7hv (eV)[F(R)h ]1/2 0C 5C 10C 20C 50C 70C 90C 95C 100C Figure 9-8. Modified Kubelka-Munk spectra for calcined ZnO-TiO2 photocatalyst. 97

PAGE 98

0 2 4 6 8 10 12 2.52.72.93.13.33.53.7 hv (eV)[F(R)h ]1/2 0H 5H 10H 20H 50H 70H 90H 95H 100H Figure 9-9. Modified Kubelka-Munk spectra for hydrothermal ZnO-TiO2 photocatalyst. Diffuse Reflectance Optical properties of photo catalyst materials can have a st rong influence on performance. For TiO2, the relationship between particle size, bandgap energy and performance is well documented [88]. Particle size incr eases lead to decreases in speci fic surface area and decrease in bandgap energy. For the ZnO-TiO2 composites explored in this work, bandgap energy was estimated with the Kubelka-Munk method using diffuse reflectance spectra. The modified Kubelka-Munk function was determ ined using the equation 9-2 2 1 22 )1( h R R (9-2) where R is the proportion reflected h is Plancks constant, and is the frequency of light. Plots of this parameter were used to determine bandgap energy by determining a linear model for the linear portion of the absorption tr ansition using least s quares regression and ex trapolating to zero at the corresponding photon energy [32]. The modi fied Kubelka-Munk spectra can be seen in Figures 9-8 and 9-9 and the corr esponding bandgap energies can be seen in Figure 9-10. The 98

PAGE 99

increase in bandgap energy for the hydrothermal ly prepared samples between 10 and 70% may be due to an increase in atom ic dispersion and decrease in prim ary particle size as indicated by specific surface area. Additionally, the bandgap d ecrease for the calcined photocatalyst between 20 and 70% TiO2 is thought to be the result of sinter ing-related particle size increases and increased interaction between ZnO and TiO2 induced by calcination. 2.75 2.85 2.95 3.05 3.15 3.25 3.35 3.45 3.55 0 20 40 60 80 100TiO2 (%)Bandgap (eV) Calcined Hydrothermal Figure 9-10. Estimated bandgap for calcined and hydrothermally prepared ZnO-TiO2 photocatalyst materials. 99

PAGE 100

CHAPTER 10 A HIGH SURFACE AREA MAGNETIC PHOTOC ATALYST WITH CONTROLLED PORE SIZE Prior chapters exploring magnetic photocatalys ts used a modified conventional procedure to produce a magnetic photocatalyst with a specific surface ar ea of nearly 30 m2/g. The work reported in this chapter was unique by incorpor ating a high surface area silica support for silica coated barium ferrite and titanium dioxide powde r providing specific surface areas greater than 500 m2/g. Several parameters including synthesis te chnique and catalyst loading were shown to have an impact on photocatalytic oxidation. Characterization SEM images provide information about particle shape and distribution of material. For the HSAMP, it was critical to understand how cha nges in the synthesis method affected these textural properties. The silica coated barium fe rrite demonstrates an irregular platelet like structure very similar to previous work [19] (Figure 10-1). Alternatively, the HSAMP morphology resembles broken pieces of glass that are approximately 45-90 m in diameter that is likely a function of the materi al preparation including grinding a nd sieving. On a smaller scale, differences between the HSAMPs can be distingu ished. HSAMP1, which was synthesized with a greater concentration of hydrofluoric acid, has a greater degree of porosity than the other HSAMP materials. For the purposes of this work, hydrofluoric acid was used to modify the pore structure. Hydrofluoric acid increased the rate of gelation and increased the solubility of silica. It is believed that this increased porosity is due to the combination of these two effects in presence of higher concentrations of HF. However, the resolution of the SEM images is not adequate to see small mesoporous structures. 10

PAGE 101

Figure 10-1. Silica coated barium ferrite (top left ), particle shape HSAMP 8 (top right), SEM of HSAMP 1 (bottom left), HSAMP 8 (bottom right). To assess the distribu tion of smaller mesopores, which ar e thought to have an impact on the transport of contaminants for photocatalytic oxidation, nitrogen ad sorption isotherms were used. First, BET specific surface area was estimat ed using 12 points for each material in the linear range between relative pre ssures of 0.05 and 0.30. The resu lting specific surface areas were 374, 543, and 550 m2/g for HSAMP1, 4, and 8 respectively. The higher surface area for the low hydrofluoric acid concentration was apparen tly due to small mesopores not evident in SEM imaging. To further identify this difference, pore size distributions were estimated using the BJH (Barret-Joyner-Halenda) methodology. The resulting pore size distri bution can be seen in Figure 10-2. The median pore diameter appears to decr ease with a decrease in hydrofluoric acid concentration. In addition, the total pore volume, which is indicated by the area under the curve, 10

PAGE 102

decreases with a decrease in HF. These changes will have the functional impact of changing the adsorption and transport characteri stics of the support material. The magnetic properties of HSAMP are a critical and distingu ishing characteristic of this work with respect to past work on supported pho tocatalysts. The raw magnetic material chosen for this work was barium ferrite due to its strong remanent magnetization and ferromagnetic properties. In the future, it is believed that thes e properties will allow the HSAMP to be used for an agitation/confinement reacto r much like the MAPR. However, for this work, the magnetic properties were necessary for separation using a ma gnetic field. As explor ed in the introduction, magnetic materials are subject to oxidation during processing a nd use, which can diminish or eliminate their beneficial properties. Vibrat ing sample magnetometry (VSM) was used to compare the different HSAMP compositions. The process involves magnetizing and demagnetizing the material by vary ing the externally applied magnetic field and determining the impact of these fluctuations on the magnetic mome nt of the material. The barium ferrites high remanent magnetization can be seen in Figure 10-3 when its moment is close to 0.6 mAm2 in the absence of an externally applied magnetic field. After processing, the rema nent magnetization is diminished, which is primarily due to the de creased mass loading of barium ferrite in the composite. However, the wide hysteresis loop of the HSAMPs in Figure 10-4 indicates that the material maintained its ferromagnetic properties despite being exposed to the highly corrosive environment during the acid catalyzed sol-gel process. 10

PAGE 103

Figure 10-2. Estimated pore size distri bution estimate for HSAMP1, 4, and 8. XRD provides evidence of changes in crystalline structure during synthesis, which is important for both the magnetic and photocatalyt ic properties of HSAM P (Figure 10-5). The particle size of titanium dioxide has a well esta blished link to photocatalytic activity. Sintering during heat treatment and dissolu tion in the corrosive synthesis environment could potentially increase or decrease the particle size. Using the Scherrer equation (eq. 10-1), particle size (t) was estimated based on peak width (B). Given a sh ape factor (k) of 0.9, wavelength of the CuK 1 xray source ( ) of 0.1541 nm, B (full peak width at half maximum corrected for instrumental broadening) of 0.004790, 0.004888, 0.005298, and 0.005444 radian obtained from data for Degussa P25, HSAMP8, HSAMP4, a nd HSAMP1 in Figure 10-4, and 2 (anatase 101 peak) of 25.4, particle sizes for each of the three samples were calculated. cos B k t (10-1) 10

PAGE 104

Using the analysis described, the Scherrer equation indicates that particle sizes for the Degussa P25 anatase crystalline phase decreases slightly from 30.0 for Degussa P25 to 29.4, 27.1, and 26.4 nm in the HSAMP8, 4, and 1 re spectively. The smaller particle size in the composite material may be due to an increase in solubility of Ti during the synthesis process due to the presence of hydrofluoric and nitric acid. Since the decrease in particle size estimate is small this is not believed to have a significan t adverse impact on photocat alytic oxidation, yet it could pose a problem if the acid concentration is increased beyond that us ed by these synthesis methods. Figure 10-3. Vibrating sample magnetometry hysteresis loop of barium ferrite. 10

PAGE 105

Figure 10-4. Vibrating samp le magnetometry hystere sis loop of HSAMP1-8. Figure 10-5. X-ray diffraction patterns for HSAMP, Degussa P25 and barium ferrite. Effect of Mass Loading The catalyst mass loading is an important ope rational parameter for reactor design. The sensitivity to changes in photocatalyst concentr ation was previously explored for the MAPR [19]. In that work, the increase in photocatalyst mass corresponded to an increase in the extent of photocatalysis until an optima was reached and fu rther increases resulted in declining reactor performance. A similar trend can be seen in Figure 10-6. As the photocatalyst mass was 10

PAGE 106

increased from 30 to 120 mg, there was a dramatic increase in both the change in phenol and nDOC concentrations. However, above 240 mg, ther e was a steady decrease in both performance metrics. This decease is believed to be associated with an unfavorable light distribution within the reactor. Understanding this phenomena is cri tical for effective scale-up and a fundamental understanding of photocatalytic reactor performance. Figure 10-6. X-ray diffraction patterns for HSAMP, Degussa P25 and barium ferrite. Effect of Preparation Technique on Phenol and Non-Purgeable Organic Carbon Removal HSAMP synthesis allowed for a customizable pore size distribution by changing the concentration of HF. It was important to understand what impact these changes have on elimination of phenol and its oxidation byproducts measured as nDOC. For these experiments, the reaction was followed over time to determine th e impact of material type on kinetics and the extent of photocatalysis. Complete oxidation, as evidenced by elimination of the nDOC, is most desirable since the oxidation byproducts are known to also have a ssociated toxicity. As can be seen in Figure 10-7, the concen tration of phenol decreases most rapidly for the Degussa P25 and was less rapid for HSAMP4 and HSAMP8 compositions. HSAMP1, with the lowest specific surface area, had the slowest removal of phenol a nd demonstrated almost linear or zero-order 10

PAGE 107

kinetics. When looking at the change in normalized concentr ation of nDOC, the trend was somewhat similar initially to phenol removal (Figure 10-8). However, at longer times the extent of photocatalysis over 7 hours was more complete for HSAMPs than Degussa P25 alone. This results in fewer potentially toxic oxidation byproducts given the proposed photocatalytic oxidation mechanism (Figure 10-9). Figure 10-7. Normalized phenol concentration as a function of time for HSAMP1,4, and 8. Figure 10-8. Normalized nDOC concentration as a function of time for HSAMP1,4, and 8. 10

PAGE 108

Effect of Repeated Use In addition to the recoverability of the magnetic photocatalyst, it is also desirable for the material to be reused with its high photoactivity maintained. For the series of experiments seen in Figure 10-10, the magnetic photocatalyst was recovered using a permanent magnet, and immediately re-used after experimentation wit hout rinsing, drying or cleaning the surface. Separate experiments determined the mass recove ry is nearly 99% using the permanent magnet recovery technique. For the highe r specific surface area materials, HSAMP4 and HSAMP8, there was nearly complete removal of nDOC during all 6 repeated UV exposure periods (Figure 1010). However, for the lower surface area material there was an increase in performance with repeated use. This may be related to adsorp tion of organic fragments during photocatalytic oxidation. The consistent performance shown for the HSAMPs would be advantageous for longterm use in a continuous flow reactor using an agitation confinement magn etic field in addition to recovery using a static magne tic field in a batch reactor. OH OH OH OH OH OH OH OH Aliphatic Compounds Figure 10-9. Proposed reaction mechanism for photocatalytic oxidation of phenol. 10

PAGE 109

Figure 10-10. Effect of repeated use on the ex tent of photocatalysis for HSAMP1,4 and 8. 10

PAGE 110

CHAPTER 11 CONCLUSIONS Contributions to Science Demonstrated the feasibility of reducing to tal organic carbon below the NASA specified limit of 0.5 ppm with a commercial photocatalyt ic reactor for water recovery during longterm space missions using ersatz water using both Degussa P25 and a silica-titania composite material Increased understanding of th e influence of bicarbonates on reducing reaction rates during photocatalytic oxidation using a simulated graywater Developed a higher performance magnetic photocatalyst by modi fying the coating technique Improved the magnetic field of the magnetically agitated photocatalytic reactor such that the reaction rate for disappearance of non-purg eable dissolved organic carbon was at least 2.5 times greater. Elucidated the influence of operational parameters on magnetically agitated photocatalytic reactor performance Increased understanding of ZnO-TiO2 composite material composition and heat treatment influence on dye photocatalytic oxidation performance Developed a new high surface area magnetic photocatalyst that outperforms Degussa P25 for extent of reduction in non-pur geable dissolved organic carbon Future Work Future work should include Development of a modular MAPR reactor that can be easily scaled to a particular application Exploration of the distinc tion between magnetic and mech anical agitation and their influence on photocat alytic oxidation Further develop understanding of th e superior performance of HSAMP Specifically exploring the influence of lamp conditions on MAPR and photocatalytic reactors in general including lamp temperatur e, age, and thermal history using spectral distribution Increasing understanding of competitive oxidat ion in photocatalytic reactors and whether specific conditions including pore size distribution of support or agitation conditions can 11

PAGE 111

11 positively influence total mineralization or specific mineralization of highly toxic compounds Develop a MAPR for air-phase photocatalytic oxidation

PAGE 112

LIST OF REFERENCES 1. Vakil, N. B.; Schwartz, S. M.; Buggy, B. P. ; Brummitt, C. F.; Kherellah, M.; Letzer, D. M.; Gilson, I. H.; Jones, P. G., Biliary cryptosporidiosis in HIV-infected people after the waterborne outbreak of cryptosporidiosis in Milwaukee. New England Journal of Medicine 1996, 334, (1), 19-23. 2. Horikoshi, S.; Hidaka, H., Non-degradable tr iazine substrates of atrazine and cyanuric acid hydrothermally and in supercritical wate r under the UV-illuminated photocatalytic cooperation. Chemosphere 2003, 51, (2), 139-142. 3. Brocard, J. C.; Baronnet, F.; Oneal, H. E., Chemical-Kinetics of the Oxidation of Methyl Tert-Butyl Ether (Mtbe). Combustion and Flame 1983, 52, (1), 25-35. 4. Williams, P. R. D.; Scott, P. K.; Sheehan, P. J.; Paustenbach, D. J., A probabilistic assessment of household exposures to MTBE from drinking water. Human and Ecological Risk Assessment 2000, 6 (5), 827-849. 5. Nowack, K. O.; Cannon, F. S.; Mazyck, D. W ., Enhancing activated carbon adsorption of 2-methylisoborneol: Methane and steam treatments. Environmental Science & Technology 2004, 38, (1), 276-284. 6. Bayer, P.; Heuer, E.; Karl, U.; Finkel, M., Economical and ecological comparison of granular activated carbon (GAC) adsorber refill strategies. Water Research 2005, 39, (9), 17191728. 7. Ang, E. L.; Zhao, H. M.; Obbard, J. P ., Recent advances in the bioremediation of persistent organic pollutants via biomolecular engineering. Enzyme and Microbial Technology 2005, 37, (5), 487-496. 8. Parra, S.; Sarria, V.; Malato, S.; Peringer, P.; Pulgarin, C., Photochemical versus coupled photochemical-biological flow system for the tr eatment of two biorec alcitrant herbicides: metobromuron and isoproturon. Applied Catalysis B-Environmental 2000, 27, (3), 153-168. 9. Hoffmann, M. R.; Martin, S. T.; Choi W. Y.; Bahnemann, D. W., Environmental Applications of Semiconductor Photocatalysis. Chemical Reviews 1995, 95, (1), 69-96. 10. Rosenfeldt, E. J.; Linden, K. G., Degr adation of endocrine disrupting chemicals bisphenol A, ethinyl estradio l, and estradiol dur ing UV photolysis and advanced oxidation processes. Environmental Scie nce & Technology 2004, 38, (20), 5476-5483. 11. Neamtu, M.; Siminiceanu, H.; Yediler, A.; Kettrup, A., Kinetics of decolorization and mineralization of reactive azo dyes in a queous solution by the UV/H2O2 oxidation. Dyes and Pigments 2002, 53, (2), 93-99. 12. Doerffler, W.; Hauffe, K., Heterogeneous Photocatalysis .1. The Influence of Oxidizing and Reducing Gases on the Electrical Conductivity of Dark and Illuminated Zinc Oxide Surfaces. Journal of Catalysis 1964, 3 (2), 156-170. 11

PAGE 113

13. Kostedt, W. L., IV, Optimization of a magne tically agitated photocatalytic reactor for water recovery. In Proceedings of the 35th International Conference on Environmental Systems, Society of Automotive Engineers (SAE) Rome, Italy, 2005; Vol. 2005-01-2995. 14. Kostedt, W. L., IV; Witwer, M. A.; Mazyck D. W.; Powell, T.; Butters, B. A., A slurrybased photocatalytic reactor with slu rry separation for wa ter recovery. In Proceedings of the 35th International Conference on En vironmental Systems, Society of Automotive Engineers (SAE) Rome, Italy, 2005; Vol. 2005-01-2994. 15. Lee, W.; Gao, Y. M.; Dwight, K.; Wold, A., Preparation and Characterization of Titanium(Iv) Oxide Photocatalysts. Materials Research Bulletin 1992, 27 (6), 685-692. 16. Holmes, F. R.; Chadik, P. A.; Mazyck, D. W.; Wu, C.-Y.; Garton, M. J.; Powers, K. W.; Londeree, D. J., Photocatalytic oxidation of selected organic c ontaminants in a continuous flow reactor packed with titania-doped silica. In Proceedings of the 34th In ternational Conference on Environmental Systems, Society of Automotive Engineers (SAE) Colorado Springs, CO, 2004; Vol. 2004-01-2403. 17. Beydoun, D.; Amal, R.; Low, G.; McE voy, S., Occurrence and prevention of photodissolution at the phase junction of magnetite and titanium dioxide. Journal of Molecular Catalysis a-Chemical 2002, 180, (1-2), 193-200. 18. Watson, S.; Beydoun, D.; Amal, R., Synthesi s of a novel magnetic photocatalyst by direct deposition of nanosized TiO2 crystals onto a magnetic core. Journal of Photochemistry and Photobiology a-Chemistry 2002, 148, (1-3), 303-313. 19. Kostedt, W. L., IV; Drwiega, J.; Mazyck, D. W.; Lee, S. W.; Sigmund, W.; Wu, C. Y.; Chadik, P., Magnetically agitated photocatalytic reactor for photocatalytic oxidation of aqueous phase organic pollutants. Environmental Science & Technology 2005, 39, (20), 8052-8056. 20. Dave, R. N.; Wu, C. Y.; Chaudhuri, B. ; Watano, S., Magnetically mediated flow enhancement for controlled powder discharge of cohesive powders. Powder Technology 2000, 112, (1-2), 111-125. 21. Ramlakhan, M.; Wu, C. Y.; Watano, S.; Dave R. N.; Pfeffer, R., Dry particle coating using magnetically assisted impaction coa ting: modification of surface properties and optimization of system and operating parameters. Powder Technology 2000, 112, (1-2), 137-148. 22. Yang, J.; Sliva, A.; Banerjee, A.; Dave, R. N.; Pfeffer, R., Dry particle coating for improving the flowability of cohesive powders. Powder Technology 2005, 158, (1-3), 21-33. 23. Yu, Q.; Dave, R. N.; Zhu, C.; Quevedo, J. A.; Pfeffer, R., Enhanced fluidization of nanoparticles in an oscillating magnetic field. Aiche Journal 2005, 51, (7), 1971-1979. 24. Amal, R.; Beydoun, D.; Low, G.; McEvoy, S. Photocatalyst. 6,558,553, 2003. 11

PAGE 114

25. Beydoun, D.; Amal, R., Implications of heat treatment on the properties of a magnetic iron oxide-titanium di oxide photocatalyst. Materials Science and Engi neering B-Solid State Materials for Advanced Technology 2002, 94, (1), 71-81. 26. Beydoun, D.; Amal, R.; Low, G.; McEvoy, S., A preliminary inve stigation into the synthesis of titania-coated magne tite as a novel photocatalyst. Proceedings of the Third World Congress on Particle Technology 1998, Paper No. 385 27. Beydoun, D.; Amal, R.; Low, G.; McEvoy, S ., Role of nanoparticle s in photocatalysis. Journal of Nanoparticle Research 1999, 1 (4), 439-458. 28. Beydoun, D.; Amal, R.; Low, G. K. C.; McE voy, S., Novel photocatal yst: Titania-coated magnetite. Activity and photodissolution. Journal of Physical Chemistry B 2000, 104, (18), 4387-4396. 29. Beydoun, D.; Amal, R.; Scott, J.; Low, G.; McEvoy, S., Studies on the mineralization and separation efficiencies of a magnetic photocatalyst. Chemical Engineering & Technology 2001, 24, (7), 745-748. 30. Cohen, F.; Fiaud, C.; Chemla, M., Electro chemical study of the corrosion of dental amalgams. Comptes rendus des sances de l'Academie des sciences. Srie II, Mcanique, physique, chimie, sciences de l'univers, sciences de la terre 2000, 89, 170-171. 31. Coleman, H. M.; Marquis, C. P.; Scott, J. A.; Chin, S. S.; Amal, R., Bactericidal effects of titanium dioxide-based photocatalysts. Chemical Engineering Journal 2005, 113, (1), 55-63. 32. Teoh, W. Y.; Amal, R.; Madler, L.; Pratsinis, S. E., Flame sprayed visible light-active FeTiO2 for photomineralisation of oxalic acid. Catalysis Today 2007, 120, (2), 203-213. 33. Watson, S.; Scott, J.; Beydoun, D.; Amal, R., Studies on the prep aration of magnetic photocatalysts. Journal of Nanoparticle Research 2005, 7 (6), 691-705. 34. Stokke, J. M.; Mazyck, D. W.; Wu, C. Y. ; Sheahan, R., Photocatalytic oxidation of methanol using silica-titania composites in a packed-bed reactor. Environmental Progress 2006, 25, (4), 312-318. 35. O'Handley, R. C., Modern Magnetic Materials: Principles and Applications John Wiley & Sons, Inc.: New York, NY, 1999. 36. Jakubovics, J. P., Magnetism and Magnetic Materials The Institute of Metals: Brookfield, VT, 1987. 37. Weber, W. J.; DiGiano, F. A., Process Dynamics in Environmental Systems Environmental Science and Technology: A Wiley Interscience Se ries of Texts and Monographs: New York, NY, 1996. 11

PAGE 115

38. Jovanovic, G. N.; Somchamni, T.; Atwate r, J. E.; Akse, J. R.; Wheeler, R. R., Magnetically assisted liquidsolid fluidization in normal and microgravity conditions: experiment and theory. Powder Technology 2004, 148, (2-3), 80-91. 39. Sornchamni, T. Magnetically assisted liqui d-solid fluidization in a gradient magnetic field: theory and applicat ion. Ph.D. Dissertation, Or egon State University, 2004. 40. Qiao, S.; Sun, D. D.; Tay, J. H.; East on, C., Photocatalytic oxidation technology for humic acid removal using a nano-structured TiO2/Fe2O3 catalyst. Water Science and Technology 2003, 47, (1), 211-217. 41. Lee, S. W.; Drwiega, J.; Wu, C. Y.; Mazyck, D.; Sigmund, W. M., Anatase TiO2 nanoparticle coating on barium ferrite using titanium bis-ammonium lact ato dihydroxide and its use as a magnetic photocatalys. Chemistry of Materials 2004, 16, (6), 1160-1164. 42. Lee, S. W.; Drwiega, J.; Mazyck, D.; Wu, C. Y.; Sigmund, W. M., Synthesis and characterization of hard magnetic composite pho tocatalyst Barium ferrite/silica/titania. Materials Chemistry and Physics 2006, 96, (2-3), 483-488. 43. Shchukin, D. G.; Kulak, A. I.; Sviridov, D. V., Magnetic photocatalysts of the core-shell type. Photochemical & Photobi ological Sciences 2002, 1 (10), 742-744. 44. Shchukin, D. G.; Ustinovich, E. A.; Sviridov, D. V.; Kulak, A. I., Titanium and iron oxide-based magnetic photocatalysts for oxidatio n of organic compounds and sulfur dioxide. High Energy Chemistry 2004, 38, (3), 167-173. 45. Chung, Y. S.; Park, S. B.; Kang, D. W., Magnetically separable titania-coated nickel ferrite photocatalyst. Materials Chemistry and Physics 2004, 86, (2-3), 375-381. 46. Fu, W. Y.; Yang, H. B.; Li, M. H.; Li, M. H.; Yang, N.; Zou, G. T., Anatase TiO2 nanolayer coating on cobalt ferrite na noparticles for magnetic photocatalyst. Materials Letters 2005, 59, (27), 3530-3534. 47. Rana, S.; Srivastava, R. S.; Sorensson, M. M.; Misra, R. D. K., Synthesis and characterization of nanoparticles with magnetic core and photocatalytic shell: Anatase TiO2NiFe2O4 system. Materials Science and Engineering B-Sol id State Material s for Advanced Technology 2005, 119, (2), 144-151. 48. Chen, F.; Zhao, J. C., Preparation and photocatalytic properties of a novel kind of loaded photocatalyst of TiO2/SiO2/gamma-Fe2O3. Catalysis Letters 1999, 58, (4), 245-247. 49. Gao, Y.; Chen, B. H.; Li, H. L.; Ma, Y. X., Preparation and characterization of a magnetically separated photocatalys t and its catalytic properties. Materials Chemistry and Physics 2003, 80, (1), 348-355. 50. Voumard, P.; Zhan, Q.; Zenobi, R., Ad sorption of Monosubstituted Benzenes on Low Surface-Area Silica Studied by TemperatureProgrammed Desorption and Laser-Induced Thermal-Desorption Methods. Langmuir 1995, 11, (3), 842-848. 11

PAGE 116

51. Boschke, E.; Bohmer, U.; Lange, J.; Constape l, M.; Schellentrager, M.; Bley, T., The use of respirometric measurements to determine the toxicity of textile dyes in aqueous solution and after oxidative decolourisation processes. Chemosphere 2007, 67, (11), 2163-2168. 52. Evgenidou, E.; Fytianos, K.; Poulios, I., Semiconductor-sensitized photodegradation of dichlorvos in water using TiO2 and ZnO as catalysts. Applied Catalysis B-Environmental 2005, 59, (1-2), 81-89. 53. Goncalves, M. S. T.; Oliveira-Campos, A. M. F.; Pinto, E. M. M. S.; Plasencia, P. M. S.; Queiroz, M. J. R. P., Photochemical treatme nt of solutions of azo dyes containing TiO2. Chemosphere 1999, 39, (5), 781-786. 54. Sauer, T.; Neto, G. C.; Jose, H. J.; Moreir a, R. F. P. M., Kinetics of photocatalytic degradation of reactive dyes in a TiO2 slurry reactor. Journal of Photochemistry and Photobiology a-Chemistry 2002, 149, (1-3), 147-154. 55. Curri, M. L.; Comparelli, R.; Cozzoli, P. D.; Mascolo, G.; Agostiano, A., Colloidal oxide nanoparticles for the photocatalytic degradation of organic dye. Materials Science & Engineering C-Biomimetic and Supramolecular Systems 2003, 23, (1-2), 285-289. 56. Daneshvar, N.; Salari, D.; Khataee, A. R., Photocatalytic degradat ion of azo dye acid red 14 in water on ZnO as an a lternative catalyst to TiO2. Journal of Photochemistry and Photobiology a-Chemistry 2004, 162, (2-3), 317-322. 57. Neppolian, B.; Choi, H. C.; Sakthivel, S.; Arabindoo, B.; Murugesan, V., Solar/UVinduced photocatalytic degradation of three commercial textile dyes. Journal of Hazardous Materials 2002, 89, (2-3), 303-317. 58. Kostedt, W. L., IV; Mazyck, D. W., Eval uation of a Photocatalytic Water Treatment Process. Florida Water Resources Journal 2006, 58, (11), 44-48. 59. Peralta-Zamora, P.; Kunz, A.; de Moraes, S. G.; Pelegrini, R.; Moleiro, P. D.; Reyes, J.; Duran, N., Degradation of reactive dyes I. A comparative study of ozonation, enzymatic and photochemical processes. Chemosphere 1999, 38 (4), 835-852. 60. Sakthivel, S.; Neppolian, B.; Shankar, M. V.; Arabindoo, B.; Palanichamy, M.; Murugesan, V., Solar photocatalytic degradati on of azo dye: comparison of photocatalytic efficiency of ZnO and TiO2. Solar Energy Materials and Solar Cells 2003, 77, (1), 65-82. 61. Liao, S. J.; Huang, D. G.; Yu, D. H.; Su, Y. L.; Yuan, G. Q., Preparation and characterization of ZnO/TiO2 SO42-/ZnO/TiO2 photocatalys t and their photocatalysis. Journal of Photochemistry and P hotobiology a-Chemistry 2004, 168, (1-2), 7-13. 62. Wu, C. H., Comparison of azo dye degradation efficiency using UV/single semiconductor and UV/coupled semiconductor systems. Chemosphere 2004, 57, (7), 601-608. 63. Marci, G.; Augugliaro, V.; Lopez-Munoz, M. J.; Martin, C.; Palmisano, L.; Rives, V.; Schiavello, M.; Tilley, R. J. D.; Venezia, A. M., Preparation characteri zation and photocatalytic 11

PAGE 117

activity of polycrystalline ZnO/TiO2 syst ems. 1. Surface and bulk characterization. Journal of Physical Chemistry B 2001, 105, (5), 1026-1032. 64. Marci, G.; Augugliaro, V.; Lopez-Munoz, M. J.; Martin, C.; Palmisano, L.; Rives, V.; Schiavello, M.; Tilley, R. J. D.; Venezia, A. M., Preparation characteri zation and photocatalytic activity of polycrystalline ZnO/TiO2 system s. 2. Surface, bulk ch aracterization, and 4nitrophenol photodegradation in liquid-solid regime. Journal of Physical Chemistry B 2001, 105, (5), 1033-1040. 65. Evgenidou, E.; Konstantinou, I.; Fytianos, K. ; Poulios, I.; Albanis, T., Photocatalytic oxidation of methyl parathion over TiO2 and ZnO suspensions. Catalysis Today 2007, 124, (34), 156-162. 66. Zhang, Z. H.; Yuan, Y.; Fang, Y. J.; Liang, L. H.; Ding, H. C.; Jin, L. T., Preparation of photocatalytic nano-ZnO/TiO2 film and applic ation for determination of chemical oxygen demand. Talanta 2007, 73, (3), 523-528. 67. Peternel, I. T.; Koprivanac, N.; Bozic, A. M. L.; Kugic, H. M., Comparative study of UV/TiO2, UV/ZnO and photo-Fenton processes for the organic reactive dye degradation in aqueous solution. Journal of Hazardous Materials 2007, 148 (1-2), 477-484. 68. Wang, C.; Xu, B. Q.; Wang, X. M.; Zhao, J. C., Preparation and phot ocatalytic activity of ZnO/TiO2/SnO2 mixture. Journal of Solid State Chemistry 2005, 178, (11), 3500-3506. 69. Philipse, A. P.; Vanbruggen, M. P. B.; Pathmamanoharan, C., Magnetic Silica Dispersions Preparation and Stability of Su rface-Modified Silica Part icles with a Magnetic Core. Langmuir 1994, 10, (1), 92-99. 70. Chen, J. Y.; Gao, L.; Huang, J. H.; Yan, D. S., Preparation of nanosized titania powder via the controlled hydroly sis of titanium alkoxide. Journal of Materials Science 1996, 31, (13), 3497-3500. 71. Korn, M. D. A.; Ferreira, A. C.; Costa, A. C. S.; Nobrega, J. A.; Silva, C. R., Comparison of decomposition procedures for an alysis of titanium dioxide us ing inductively coupled plasma optical emission spectrometry. Microchemical Journal 2002, 71, (1), 41-48. 72. Mazyck, D.; Drwiega, J.; Wu, C.-Y.; Chad ik, P.; Sigmund, W.; Lee, S. W.; Meisel, M. W.; Park, J.-H., Development and Characterizatio n of a Magnetically Agitated Photocatalytic Reactor for Water Recovery. International Conference on Envir onmental Systems, Society of Automotive Engineers (SAE) 2004, 2004-01-2404. 73. Verostko, C. E.; Carrier, C.; Finger, B. W ., Ersatz wastewater formulations for testing water recovery systems. In 34th International Conference on En vironmental Systems, Society of Automotive Engineers (SAE) Colorado Springs, CO, 2004; Vol. 2004-01-2448. 74. Khan, A. Y.; Mazyck, D. W., The effect of UV irradiation on adsorption by activated carbon/TiO2 composites. Carbon 2006, 44, (1), 182-184. 11

PAGE 118

75. Byrne, H. E.; Kostedt, W. L., IV; Stokke, J. M.; Mazyck, D. W., Characterization of HFcatalyzed silica gels doped with Degussa P25 titanium dioxide. Submitted, 2008. 76. Serwicka, E. M., Surface area and porosity X-ray diffraction and chemical analyses. Catalysis Today 2000, 56, (4), 335-346. 77. Hu, C.; Wang, Y. Z.; Tang, H. X., Destruction of phenol aqueous solution by photocatalysis or direct photolysis. Chemosphere 2000, 41, (8), 1205-1209. 78. Dijkstra, M. F. J.; Buwalda, H.; de Jong, A. W. F.; Michorius, A.; Winkelman, J. G. M.; Beenackers, A. A. C. M., Experimental compar ison of three reactor designs for photocatalytic water purification. Chemical Engineering Science 2001, 56, (2), 547-555. 79. Sivalingam, G.; Priya, M. H.; Madras, G ., Kinetics of the photodegradation of substituted phenols by solution combustion synthesized TiO2. Applied Catalysis B-Environmental 2004, 51, (1), 67-76. 80. Muster, T. H.; Prestidge, C. A.; Hayes, R. A., Water adsorption kinetics and contact angles of silica particles. Colloids and Surfaces a-Physicoc hemical and Engineering Aspects 2001, 176, (2-3), 253-266. 81. Bhatkhande, D. S.; Kamble, S. P.; Sawant, S. B.; Pangarkar, V. G ., Photocatalytic and photochemical degradation of nitrobenzene using artificial ultraviolet light. Chemical Engineering Journal 2004, 102, (3), 283-290. 82. Cockell, C. S.; Andrady, A. L., The Martian and extraterre strial UV radiation environment 1. Biological and cl osed-loop ecosystem considerations. Acta Astronautica 1999, 44, (1), 53-62. 83. Fogler, H. S., Elements of Chemical Reaction Engineering Third ed.; Prentice Hall: New York, 1999. 84. So, C. M.; Cheng, M. Y.; Yu, J. C.; Wong, P. K., Degradation of azo dye Procion Red MX-5B by photocatalytic oxidation. Chemosphere 2002, 46, (6), 905-912. 85. Hu, C.; Tang, Y. H.; Jiang, Z.; Hao, Z. P.; Tang, H. X.; Wong, P. K., Characterization and photocatalytic activity of nobl e-metal-supported surface TiO2/SiO2. Applied Catalysis aGeneral 2003, 253, (2), 389-396. 86. Hasnat, M. A.; Siddiquey, I. A.; Nuruddin, A., Comparative photocatalytic studies of degradation of a cationic and an anionic dye. Dyes and Pigments 2005, 66 (3), 185-188. 87. Hurum, D. C.; Agrios, A. G.; Gray, K. A. ; Rajh, T.; Thurnauer, M. C., Explaining the enhanced photocatalytic ac tivity of Degussa P25 mixed-phase TiO2 using EPR. Journal of Physical Chemistry B 2003, 107, (19), 4545-4549. 11

PAGE 119

11 88. Lin, H.; Huang, C. P.; Li, W.; Ni, C.; Sh ah, S. I.; Tseng, Y. H., Size dependency of nanocrystalline TiO2 on its optical property a nd photocatalytic reac tivity exemplified by 2chlorophenol. Applied Catalysis B-Environmental 2006, 68, (1-2), 1-11.

PAGE 120

BIOGRAPHICAL SKETCH William L. Beau Kostedt, IV was born in 1980 in St. Louis, Missouri. He lived in the central west end and attended New City School until 1990 when he moved to Clayton. He attended Clayton schools and gra duated high school in 1998. He then began attending Trinity University in San Antonio wher e he studied chemical engineer ing and earned his bachelors degree in 2002. After graduation, he taught at Affton High School in St. Louis for two years before beginning graduate school at the University of Florida. At the University of Florida, he was advised by Dr. David Mazyck in the Depart ment of Environmental Engineering Sciences where he performed research on photocat alysts and photocatalytic reactors.