Citation
Development of a Magnetically Agitated Photocatalytic Reactor

Material Information

Title:
Development of a Magnetically Agitated Photocatalytic Reactor
Creator:
DRWIEGA, JACK ( Author, Primary )
Copyright Date:
2008

Subjects

Subjects / Keywords:
Barium ( jstor )
Composite particles ( jstor )
Ferrites ( jstor )
Lamps ( jstor )
Magnetic fields ( jstor )
Magnetism ( jstor )
Magnets ( jstor )
Particle mass ( jstor )
Phenols ( jstor )
Wavelengths ( jstor )

Record Information

Source Institution:
University of Florida
Holding Location:
University of Florida
Rights Management:
Copyright Jack Drwiega. Permission granted to the University of Florida to digitize, archive and distribute this item for non-profit research and educational purposes. Any reuse of this item in excess of fair use or other copyright exemptions requires permission of the copyright holder.
Embargo Date:
4/30/2005
Resource Identifier:
436097569 ( OCLC )

Downloads

This item is only available as the following downloads:


Full Text

PAGE 1

DEVELOPMENT OF A MAGNETICALLY AGITATED PHOTOCATALYTIC REACTOR By JACK DRWIEGA A THESIS PRESENTED TO THE GRADUATE SCHOOL OF THE UNIVERSITY OF FLOR IDA IN PARTIAL FULFILLMENT OF THE REQUIREMENTS FOR THE DEGREE OF MASTER OF ENGINEERING UNIVERSITY OF FLORIDA 2004

PAGE 2

Copyright 2004 by Jack Drwiega

PAGE 3

ACKNOWLEDGMENTS I would like to thank Dr. David Mazyck for giving me an opportunity to work with his group on a project that was both challenging and frustrating. His guidance has helped me overcome many obstacles, and I have learned much from this experience and have come away with one of the great learning experiences of my life. I would like to thank Dr. Mark Meisel of the University of Florida Physics Department for lending his expertise in magnetism to help me solve the magnetic fluidization aspect of the project. This collaboration has greatly helped me in understanding the dynamics of magnetic fields and their effects on particle movement. Dr. Meisel provided a simple model to predict the frequency at which the particles would resonate and also contributed ideas for coil winding geometries, all of which have been invaluable to the success of this project. I would like to thank Dr. Wolfgang Sigmund, Dr. Chang-Yu Wu, Dr. Paul Chadik, and Dr. Kevin Powers for all their suggestions and input throughout the course of my project. I would also like to thank the students in my research group for all their help along the way, particularly Matt Tennant, Ameena Khan, Thomas Chestnutt, Christina Ludwig, Jennifer Hobbs, and Morgana Bach. I thank Seung-woo Lee for conducting SEM, TEM, EDS, XRD, and particle size distribution analyses. I also thank Ju-Hyun Park and Dr. Meisel for performing SQUID analysis on the magnetic particle along with Rick Loftis at Engineering Performance Solutions, Gainesville, for performing GC/MS analysis. iii

PAGE 4

Lastly, I thank my parents for instilling a strong work ethic in me, which is the key to all success. iv

PAGE 5

TABLE OF CONTENTS page ACKNOWLEDGMENTS.................................................................................................iii LIST OF TABLES............................................................................................................vii LIST OF FIGURES.........................................................................................................viii ABSTRACT.........................................................................................................................x CHAPTER 1 INTRODUCTION.........................................................................................................1 2 LITERATURE REVIEW..............................................................................................4 2.1 Photocatalysis.........................................................................................................4 2.2 Photocatalytic Mechanism......................................................................................6 2.3 TiO2 Coating on Magnetic Substrates....................................................................8 2.4 Magnetic Fluidization...........................................................................................10 2.5 Magnetic Agitation Principles and Theory...........................................................11 3 MATERIALS AND METHODS.................................................................................15 3.1 Particle Synthesis..................................................................................................15 3.1.1 Mechanical Coating....................................................................................16 3.1.2 Titanium (IV) Isopropoxide Sol-gel Coating.............................................16 3.1.3 Titanium n-Butoxide Sol-gel Coating........................................................17 3.2 Magnetic Reactor Configuration..........................................................................19 3.3 Particle and Reactor Testing.................................................................................23 3.3.1 Frequency Optimization.............................................................................23 3.3.2 Photomeasurement.....................................................................................24 3.3.3 Dye Removal..............................................................................................25 3.3.4 Digestion.....................................................................................................26 3.3.5 Initial Batch Testing...................................................................................26 3.3.6 Flow Through Studies................................................................................27 4 RESULTS AND DISCUSSION..................................................................................29 4.1 Initial Iterations.....................................................................................................29 v

PAGE 6

4.2 Particle Characterization.......................................................................................32 4.2.1 300 – 600 m Particles...............................................................................32 4.2.2 2.2 m Particles..........................................................................................35 4.3 MAPR Specifications and Characterization.........................................................45 4.4 Magnetic Reactor Frequency Optimization..........................................................48 4.5 Reactor Characterization and Particle Performance.............................................50 4.5.1 Batch Tests.................................................................................................50 4.5.2 Flow Through Tests....................................................................................56 5 CONCLUSION............................................................................................................64 APPENDIX A TRANSMITTANCE CURVES..................................................................................66 B UV INTENSITIES......................................................................................................68 LIST OF REFERENCES...................................................................................................69 BIOGRAPHICAL SKETCH.............................................................................................73 vi

PAGE 7

LIST OF TABLES Table page 2-1. Photocatalytic reaction mechanisms (adapted from Turchi et al., 1990).....................7 3-1. Summary of barium ferrite particle coating methods.................................................15 4-1. Ball-milled barium ferrite attributes...........................................................................36 4-2. UV intensities for MAPR 2 at 3.5 cm........................................................................48 vii

PAGE 8

LIST OF FIGURES Figure page 2-1. Axial transmitted radiation distribution as a function of catalyst concentration for Degussa P25. Concentrations in mg/L. () 0, () 3, () 5,() 10, () 15, (-) 20, (—) 30, () 50, () 80, () 100 (adapted from Salaices et al., 2002)........5 3-1. MAPR 1 configuration...............................................................................................20 3-3. 8 mm Teflon fittings with o-rings...............................................................................21 3-4. MAPR 2 configuration...............................................................................................22 3-5. Cell 3 for Reactor 2....................................................................................................23 3-6. Photomeasurement setup on MAPR 2........................................................................25 4-1. 300 – 600 m particle magnetism..............................................................................33 4-2. Particle 3 before testing..............................................................................................34 4-3. Particle 3 after testing with magnetic agitation..........................................................35 4-4. Ball-milled barium ferrite particle size distribution...................................................36 4-5. 2.2 m particle magnetism.........................................................................................37 4-6. SEM image of uncoated ball-milled barium ferrite....................................................38 4-7. EDS of ball-milled barium ferrite...............................................................................38 4-8. SEM image of silica coated ball-milled barium ferrite..............................................39 4-9. EDS of silica coated ball-milled barium ferrite..........................................................40 4-10. TEM of silica coated barium ferrite.........................................................................40 4-11. SEM image of TiO2-silica coated ball-milled barium ferrite...................................41 4-12. EDS of TiO2-silica coated ball-milled barium ferrite...............................................42 4-13. TEM of TiO2-silica coated barium ferrite................................................................42 viii

PAGE 9

4-14. XRD of uncoated ball-milled barium ferrite...........................................................43 4-15. XRD of TiO2-silica coated barium ferrite...............................................................44 4-16. XRD of TBOT TiO2 precipitate heat treated at 500 C............................................44 4-17. Magnetic field profile and gradient field for MAPR 1............................................46 4-18. Magnetic field profile and gradient field for MAPR 2............................................47 4-19. Agitation measurement as a function of light transmittance....................................49 4-20. Photocatalytic decolorization of 10 mg/L reactice red dye with 200 mg of particle after 30 min..............................................................................................................50 4-21. Four hour photocatalytic decolorization of a 10 mg/L RR solution with Particle 2................................................................................................................................51 4-22. Four hour adsorption of a 10 mg/L RR solution with Particle 3..............................53 4-23. Two hour photocatalytic decolorization of a 5 mg/L RR solution with 100 mg of Particle 4...................................................................................................................54 4-24. Photocatalytic decolorization of a 5 mg/L RR solution over time with 100 mg of Particle 4...................................................................................................................55 4-25. Photocatalytic degradation of a 10 mg/L phenol solution over time with 100 mg of Particle 4 as measured by the change in solution absorption at 270 nm..............56 4-26. Adsorption of 5 mg/L RR solution with 2 g of Particle 3........................................57 4-27. Photocatalytic degradation of 200 mL of a 10 mg/L phenol solution......................58 4-28. One hour removal of a 10 mg/L phenol solution with 300 mg of Particle 4 at different UV wavelengths using a Pyrex glass cell. Pyrex glass is only about 10% transmittant to 254 nm wavelengths explaining the low removal rates for the 254 nm lamps.....................................................................................................59 4-29. Photocatalytic degradation of 200 mL of a 120 g/L phenol solution at a flow rate of 5 mL/min using either three 365 nm lamps or one 312 nm lamp and different particle masses...........................................................................................60 4-30. Photocatalytic degradation of 200 mL of a 120 g/L phenol solution with one 312 nm lamp.............................................................................................................61 4-31. Photocatalytic degradation of 200 mL of a 120 g/L phenol solution with 500 mg of Particle 4 and a 4 mg/L Degussa P-25 TiO2 slurry with 3 lamps........................62 ix

PAGE 10

Abstract of Thesis Presented to the Graduate School of the University of Florida in Partial Fulfillment of the Requirements for the Degree of Master of Engineering DEVELOPMENT OF A MAGNETICALLY AGITATED PHOTOCATALYTIC REACTOR By Jack Drwiega May 2004 Chair: David Mazyck Major Department: Environmental Engineering Sciences Titanium dioxide photocatalysis has been proven effective in mineralizing organic water pollutants. Titanium dioxide is a nontoxic semiconductor making it an ideal medium for water treatment. A photocatalytic system for water treatment incorporating TiO2 coated magnetic particles in a magnetically agitated reactor scheme is desired. The purpose of this research was to develop and characterize magnetic, photocatalytic particles and a magnetic reactor that will utilize these particles. Photocatalytically active TiO2-silica-barium ferrite composite particles with an average size of 2.2 m have been developed and characterized. A magnetically agitated photocatalytic reactor has been developed with the necessary properties to agitate and control the magnetic photocatalysts. The treatment system has been optimized for particle loading (2.5 g/L), agitation frequency (50 Hz), UV wavelength (312 nm), and number of lamps (3). Removal of phenol has been demonstrated in batch and flow through in the system. x

PAGE 11

CHAPTER 1 INTRODUCTION Water recovery through heterogeneous photocatalysis has become a viable option for removal of organics in water. The treatment efficiencies of photooxidation have been well documented (Blake, 1999). Titanium dioxide photocatalysis has been proven effective in mineralizing organic water pollutants, and being a nontoxic semiconductor makes TiO2 an ideal medium for water treatment. New technologies exploiting the properties of nanosized photocatalytic particles are of interest to such entities as the U.S. Environmental Protection Agency (EPA) and the National Aeronautics and Space Administration (NASA) for providing novel solutions to unique water related problems. The EPA desires a photocatalytic particle on the nanoscale capable of removing organics for emerging pollutants (e.g., endocrine disrupting chemicals). Nanoscale particles have high surface area exposure to light, thereby allowing for increased production of hydroxyl radicals yielding higher oxidation rates. NASA’s objective is to develop a compact and safe reactor that can be used as a post-processor for organics removal as part of their water treatment system. The goal is to provide product water that meets or exceeds NASA’s potable water quality requirements. This research is also of interest for commercialization potential in terrestrial applications (e.g., disaster response, developing nations, military operations). This magnetically agitated photocatalytic reactor (MAPR) research had many objectives. A multidisciplinary design/build/test strategy was implemented to meet each 1

PAGE 12

2 research objective. The major objectives of the research can be separated into the following three primary areas of development: Particle development, synthesis, and characterization; Magnetic reactor development and characterization; Reactor optimization for pollutant removal. Photocatalysis in water treatment has been most effectively achieved with a titanium dioxide slurry (Chen and Jeng, 1998). Using a slurry takes advantage of the high surface area offered by the highly dispersed TiO2 nanoparticles. A disadvantage of using this method is the difficulty of separating the TiO2 nanoparticles from the treated water. Immobilizing TiO2 on stationary substrates to eliminate the need for separation has been researched (Kobayakawa et al., 1998; Sunada et al., 1998), but this poses other problems including the contact with pollutants and inefficient exposure to UV light. These obstacles were overcome by synthesizing TiO2 coated magnetic microparticles, which consisted of a ferromagnetic core coated with a thin film of TiO2 nanoparticles. The magnetic properties of the core were used to fluidize and recover the particles by the application of an external magnetic field. The size scale of the particles offered a high surface area of TiO2 for efficient production of hydroxyl radicals. Hereby, the goal to develop a particle that had similar or better properties than TiO2 nanoparticles, but could be easily removed after water purification was achieved. Magnetic particles will be used for contaminant removal in a magnetically agitated photocatalytic reactor. Magnetic reactor and particle properties are desirable for magnetic particle control with oscillating magnetic fields, retention, and effective

PAGE 13

3 separation. Variable frequency, alternating current magnetic fields will be provided by a specifically wound solenoid. The oscillating field will be operated at an optimum frequency to agitate the magnetic particles, providing the most mixing for maximum contaminant removal. Reactor characterization will determine the optimum operating parameters for maximum removal. Particle mass, fluid flow rate, magnetic field frequency, and ultraviolet light intensity and wavelength will be varied to demonstrate their effect on contaminant removal.

PAGE 14

CHAPTER 2 LITERATURE REVIEW 2.1 Photocatalysis Heterogeneous photocatalysis has been extensively studied as a treatment option for environmental contaminants in water. Many reviews have been published covering the principles, mechanisms, and applications of photocatalysis (Linsebigler et al., 1995; Hoffmann et al., 1995; Blake, 1999; Serpone, 1995; Herrmann, 1999). Photocatalytic reactions carried out to completion have the capability of mineralizing organics to CO2 and water plus the corresponding mineral acids. The use of TiO2 as a photocatalyst has been studied in pure slurry form as well as attached to a substrate support. TiO2 slurry is the most efficient use of the catalyst because of its small particle size and high surface area allowing for efficient use of available photons and high mass transfer. A TiO2 slurry must be used at an optimum loading because a sufficiently high loading will lead to solution opacity and hence, decreased photon utilization efficiency (Gogate et al., 2004). Salaices et al. (2002) demonstrated the reduction in transmitted radiation at increasing TiO2 powder loadings (Figure 2-1). 4

PAGE 15

5 0100200300400500600700051015202530354045z, cmq (,z) r=7.62, mW cm-2 Figure 2-1. Axial transmitted radiation distribution as a function of catalyst concentration for Degussa P25. Concentrations in mg/L. () 0, () 3, () 5,() 10, () 15, (-) 20, (—) 30, () 50, () 80, () 100 (adapted from Salaices et al., 2002). Despite the high efficiency of TiO2 slurry, real world applications are limited because of the difficulty in separating the slurry from the finished water. TiO2 has been coated on substrates to avoid the separation problems with limited success. Immobilizing TiO2 on a support substantially reduces the surface area available for photocatalytic reactions to occur. TiO2 has been coated on smaller substrates such as glass beads, activated carbon, and magnetic particles to make more surface area available for reactions (Dijksta et al., 2001a; Balasubramanian et al., 2004; Lee et al., 2004; Arana et al., 2003; Gao et al., 2003). Magnetic substrates are ideal for these applications because they offer a simple mechanism for separation, and can improve mass transfer between contaminants and TiO2. Photocatalytic degradation has been shown to decrease with immobilized TiO2 due to mass transfer limitations (Ollis et al., 1991; Rachel et al., 2002;

PAGE 16

6 Dijkstra et al., 2001a). Ollis et al. (1991) reported that immobilization of a photocatalyst limited mass transfer decreasing photocatalysis. A study by Dijkstra et al. (2001a) compared a TiO2 slurry to TiO2 immobilized on a reactor wall and TiO2 coated on glass beads. This study also observed mass transfer limitations with the coated reactor and beads, although the photoactivity of the glass covered beads was comparable to low concentrations of TiO2 slurry. 2.2 Photocatalytic Mechanism Titanium dioxide is a semiconductor material that becomes involved in oxidation-reduction processes when excited by ultraviolet light energy. UV light energy activates the catalyst surface by exciting an electron from the valence band to the conduction band, leaving behind an electron hole. An electron scavenger is needed to prevent recombination of the excited electron back to the valence band in a mechanism called electron-hole recombination. The electron hole reacts with hydroxide ion (OH-) or water molecules to create hydroxyl radicals (OH*). Hydroxyl radical attack is assumed to be the primary mechanism for photooxidation (Turchi et al., 1990). Holes are likely to react with OHbecause it is readily adsorbed to the catalyst surface. The following four cases of hydroxyl radical attack are typically considered (Turchi et al., 1990; Pelizzetti et al., 1993): 1. Reaction occurs while radical and target are adsorbed; 2. Reaction occurs between free radical and adsorbed target; 3. Reaction occurs between adsorbed radical and free target; 4. Reaction occurs with both species in solution.

PAGE 17

7 Turchi et al. (1990) suggest that most of the photooxidation occurs at the catalyst surface, although their study also suggested that OH* can diffuse several hundred angstroms into the bulk solution. This process occurs when holes create free radicals or when holes create radicals that desorb from the TiO2 surface and then diffuse. Table 2-1 shows some proposed photocatalytic reaction mechanisms. Table 2-1. Photocatalytic reaction mechanisms (adapted from Turchi et al., 1990). Excitation TiO2 + h e+ h+ Recombination e+ h+ heat Adsorption O2+ TiIV + H2O OH+ TiIV-OHTiIV + H2O TiIV-H2O site + R1 R1,ads OH+ TiIV TiIV-OHTrapping TiIV-OH+ h+ TiIV-OH* TiIV-H2O + h+ TiIV-OH* + H+ Hydroxyl Attack Case 1 TiIV-OH* + R1,ads TiIV + R2,ads Case 2 OH* + R1,ads R2,ads Case 3 TiIV-OH* + R1 TiIV + R2 Case 4 OH* + R1 R2 The photon energy required for photocatalysis must be equal to or greater than the band gap of TiO2 for electron promotion to occur. The bang gap is the energy between the valence band and the conduction band where no energy levels are available to promote recombination (Linsebgler et al., 1995). UV light provides the energy greater than the band gap to excite the electron. A particular wavelength of UV light will correspond to the band gap of TiO2 and the applied wavelength must be equal to or less

PAGE 18

8 than the corresponding wavelength to promote the electron. Variations in the crystal structure and size of the TiO2 lead to differing band gaps, but wavelengths less than 400 nm in the long and medium ultraviolet wavelength range are typically sufficient for commercially available TiO2 powders. Oxygen has been identified as the primary electron acceptor preventing recombination and therefore has an important role in photocatalytic reactions (Yamazaki et al., 2001). Recombination can occur in picoseconds in the absence of an electron acceptor (Rothenberger et al., 1985). A study by Dijkstra et al. (2001b) has shown that oxygen addition increased the efficiency of an immobilized TiO2 reactor system in the degradation of formic acid. The anatase and rutile crystalline structures of TiO2 have been studied in photocatalysis. Anatase TiO2 has been shown to be more photocatalytically active than rutile and therefore is the most commonly used form (Tanaka et al., 1990; Augustynski, 1993). 2.3 TiO2 Coating on Magnetic Substrates Titanium dioxide has been coated on magnetic substrates for use as photocatalysts in water treatment (Watson et al., 2002; Gao et al., 2003; Chen et al., 2001). Magnetic cores are ideal hosts as TiO2 supports because they allow for easy separation of micrometer and submicrometer sized particles by an applied magnetic field (Beydoun et al., 2000). Particles in this size range also offer high surface area for better mass transfer when compared to other catalyst supports. In one case, Watson et al. (2002) reported synthesis of TiO2 coated magnetic iron oxide particles capable of degrading sucrose with the primary motivation of creating an easily separated photocatalyst. These particles included a silica intermediate layer to

PAGE 19

9 prevent photodissolution and were also heat treated to form the anatase phase of TiO2. Photodissolution is when the excited conduction band electron from the photoreaction interacts with the core resulting in its reduction (Beydoun et al., 2000). In two other noteworthy instances, Gao et al. (2003) and Chen et al. (2001) reported the synthesis of iron oxide supported TiO2 for applications as an easily separated photocatalyst. Chen et al. (2001) included a silica intermediate layer to prevent photodissolution and iron oxide absorption of UV light. Gao et al. (2003) and Chen et al. (2001) both studied the photocatalytic ability of their composites to decolorize dyes in batch systems and compared their results to those of commercially available TiO2 powders. These composite particles showed photocatalytic ability but could not reach the efficiency of the commercially available TiO2 particles. The work in this thesis will also took advantage of the magnetic properties of magnetically supported photocatalysts for separation and also further exploited the particle magnetism to enhance particle dispersion and mixing, hence increasing mass transfer and photocatalytic efficiency. TiO2 and silica layers were coated on micron sized barium ferrite magnetic supports, taking advantage of barium ferrite’s magnetic properties in the fluidization application. Barium ferrite has a sufficiently high remnant magnetization that leads to good fluidization in a magnetic field gradient without being too magnetic and not allowing particle dispersion. These particles have a highly developed anatase crystalline phase and are durable enough to withstand the attrition experienced while mixing. Oscillating magnetic field fluidization and control of magnetic photocatalysts is a novel approach to water treatment and was used to degrade organics of environmental concern in flow through systems.

PAGE 20

10 The most common coating techniques are conventional sol-gel techniques using alkoxide hydrolysis processes (Gao et al., 2003; Watson et al., 2002; Beydoun et al., 2000). These processes usually yield precipitated, amorphous TiO2 requiring heat treatment to form anatase crystals. Beydoun and Amal (2002) found a correlation between heat treatment and observed photoactivities of the composites. An increase in the heat or the duration of heat treatment would result in an increase in surface charge and a decrease in surface hydroxyl groups, thereby decreasing photoactivity. Beydoun and Amal (2002) concluded that heat treatment at 450 C for 20 min was sufficient to create anatase crystalline phase. Gao et al. (2003) reported that heat treatment at 500 C yielded the optimum anatase phase. Direct deposition of TiO2 onto a magnetic core is cautioned because of the potential for photodissolution. This problem can be overcome by coating the magnetic core with a passive intermediate layer. Beydoun et al. (2002) coated iron oxide with a silica layer and succeeded in preventing electrical contact and photodissolution. 2.4 Magnetic Fluidization Magnetic mixing has been investigated for many applications in microchannel flow with biofluids where the Reynolds number is low (Re < 1) (Suzuki and Ho, 2002). For highly laminar flow applications, turbulence becomes a nonfactor as a possible mechanism for mixing. Molecular diffusion becomes the dominant mechanism and the degree of mixing becomes highly time dependant. Systems displaying these characteristics tend to have some degree of micro-functionality (Lu et al., 2002), that is, systems dealing with microfluid flow or magnetic microparticles. Mixing with magnetic fields is an attractive method because it requires no mechanical parts within the system other than the magnetic substrate used for mixing. Magnetic particles have been

PAGE 21

11 developed with the specific application of separation after their intended use and are well suited for mixing. Therefore, magnetic particle dispersal is an ideal mechanism for micro and nanoparticle fluidization and control offering many advantages to conventional mixing techniques. 2.5 Magnetic Agitation Principles and Theory Magnetic agitation principles and theory were reviewed to discern how to agitate and control ferromagnetic particles. First, it was desired to use an alternating magnetic field to agitate the particles, which have a magnetic dipole moment. When the magnetic field is applied, the dipoles will align with the magnetic field due to a torque. This alignment occurs on a short time scale. In addition to the dipole alignment, the particles experience a force due to the gradient of the magnetic field. This effect continues as long as the field gradient is present. The oscillating magnetic field is generated by a set of wire windings, called a solenoid, which carry an alternating current. As the current alternates, so does the magnetic field. The continuous oscillation of the magnetic field at some frequency, f, creates a continuous alternating force on the particles causing them to move. A solenoid is used as the magnetic field generator. A solenoid is simply a tightly wound coil of wire. The magnetic field profile of the coil is modeled using the Biot-Savart law that may be expressed in general scalar form as dB = (o/4) (I ds sin/r2), (1) where B is the magnetic field strength, o is the permeability constant, I is the current, ds is the differential length element tangent to the wire in the direction of the current, is the angle between the current-length element I ds and the point of magnetic field

PAGE 22

12 contribution, and r is the distance between I ds and the point of magnetic field contribution For a single circular planar loop of wire, Eq.1 can be simplified to its on-axis form as B(z) = o I R2/2(R2 + z2)3/2 , (2) where R is the radius of the circular wire loop and z is the distance along the z axis of the point in question. Eq. 2 is used to calculate the field produced by each loop of wire at a point along the z axis. This gives the magnetic field strength (B) in gauss divided by the current (I) in amperes. Multiplying points on the curve by the applied current will give the magnetic field strength at each point along the z axis. Taking the derivative of the curve with respect to z yields the magnetic field gradient, (B/I)/z (Gauss/Amp/cm), along the z axis. A simple model can be used to calculate the resonant frequency as a function of particle magnetization and magnetic field gradient. If the magnetic dipoles of the particles are assumed to align instantaneously with the AC magnetic field, their force can be described as F = m B (3) where F is the force, m is the dipole moment, and B is the magnetic field strength.

PAGE 23

13 The AC magnetic field can be written as B = Bs cos(t) (4) where B is the magnetic field strength, Bs is the maximum static value of the magnetic field, is the angular frequency of motion, and t is time. The total forces are combined and result in an equation that also describes damped harmonic motion, and the generic form of this equation is )cos(2222tAzdtdzdtzdo (5) where 22dtzd is the particle acceleration, dtdz is the particle velocity, z is the particle position, is the damping constant, and A is the amplitude of oscillation. Neglecting gravity and assuming negligible damping ( = 0), the damping term, 2 dtdz , goes to zero. In addition, one can assume a linear spatial dependence of the magnetic field gradient over the particle movement distance, meaning one can write Bs ~ dzdBs zz . The solution of the simple harmonic motion form is then z + M dzdBs zz = 0 , (6) where the resonant frequency may be written as

PAGE 24

14 f ~ 1/2 (M ( dzdBs z1 1/z))1/2 (7) and f is the particle resonant frequency, M is the particle magnetization, dzdBs is the magnetic field gradient, and z = particle displacement. In these equations, M is the magnetic moment per unit mass and is the same quantity plotted in Figure 4-1 and 4-5. In this simple model the magnetization of the particle, along with the magnetic field gradient of the applied field, are the parameters governing the resonant frequency needed for optimal particle movement, and in this case, the resonance is expected to be between 10 and 100 Hz.

PAGE 25

CHAPTER 3 MATERIALS AND METHODS Development of the MAPR system required many steps. First, a coating technique had to be developed to coat TiO2 onto the magnetic core, barium ferrite (BaFe12O19). As composite particles were being synthesized, a magnetic fluidization system was being conceptualized and then fabricated. While the magnetic agitators were under iterative development, composite particles were being tested for initial photocatalytic activity. In a continuous iterative cycle, all phases of the project converged to complete a working system ready for characterization and final testing. 3.1 Particle Synthesis Three different methods were used to coat titanium dioxide onto barium ferrite particles. The first process used physical attachment by impaction and the next two methods employed chemical attachment techniques following sol-gel protocols. Table 3-1 summarized the coating techniques used for each particle. Table 3-1. Summary of barium ferrite particle coating methods. Particle Coating Technique Particle Size 1 Mechanical attachment (Theta Composer), 1 % TiO2 layer 300-600 m 2 Mechanical attachment (Theta Composer), 5% TiO2 doped silica-gel layer 300-600 m 3 Chemical attachment (Sol-gel, TTIP) 300-600 m 4 Chemical attachment (Sol-gel, TEOS and TBOT) 2.2 m 15

PAGE 26

16 3.1.1 Mechanical Coating Titanium dioxide coated barium ferrite particles were first synthesized using a Theta Composer (Tokuju Corporation). The Theta Composer mechanically coats particles by impaction. In the Theta Composer, a rotor made of zirconia rotates within a stainless steel chamber, which is rotating in the opposite direction. There is a one millimeter clearance between the rotor and the chamber. Particles are forced through this gap and the high mechanical stress coats the particles. Two types of particles were prepared with the Theta Composer. Particle 1 consisted of 300 – 600 m barium ferrite particles coated with a 1% by weight intermediate silica layer and a 1% by weight Degussa P-25 TiO2 layer. Particle 2 consisted of 300 – 600 m barium ferrite particles coated with a 5% by weight ground silica gel composite containing 12% Degussa P-25 TiO2. These particles will be referred to respectively as Particle 1 and Particle 2. Particle 1 was synthesized by adding 10 g of 300 600 m barium ferrite into the stainless steel chamber along with 100 mg of silica powder and 100 mg Degussa P-25 titanium dioxide powder. The rotor speed was set at 2000 and the Theta Composer was turned on for 30 minutes. Particle 2 was coated with the same method as Particle 1 except 500 mg of the ground silica gel composite (Londeree, 2002) was added instead of the silica and TiO2 powders. 3.1.2 Titanium (IV) Isopropoxide Sol-gel Coating Barium ferrite particles were coated with TiO2 using a TiO2 precursor from a method provided by Seung-woo Lee of the University of Florida Materials Science and Engineering Department, adapted from Hague and Mayo (1994) and Gopal et al. (1997). This coating technique created a thin film of amorphous TiO2 on the barium ferrite surface. Subsequent heat treatment was required to create the desired anatase phase.

PAGE 27

17 Four grams of barium ferrite particles (300 – 600 m) were mixed with approximately 3 mL of polyethyleneimine (PEI) (Alfa Aesar) in 80 mL of nanopure water until all of the PEI was dissolved. The solution was then stirred for 10 min with a magnetic stir bar. Next, the particles and stir bar were removed from the solution and added to 250 mL of nanopure water in a 500 mL 4-neck flask. The refluxing glassware, including condenser, funnel, gas adapter and thermometer, was then assembled onto the 4-neck flask. Nitrogen was then turned on for 10 min to purge the atmosphere. The condenser was activated by flowing water through it. Five milliliters of 2-propanol (Aldrich) were added to the funnel along with 100 L of titanium (IV) isopropoxide (TTIP, Acros). The stir bar was operated at medium speed and the funnel mixture was slowly dropped into the flask and mixed at high speed for 10 min. The stirring was reduced to a medium mixing speed and heated to 95 C at a rate of 5 C/min. The solution refluxed for 20 – 25 hours. When refluxing was completed, the particles were rinsed three times with deionized water. Then the particles were dried for one hour at 55 C. Finally, the particles were heat treated in a muffle furnace at 500 C for one hour in an ambient atmosphere. The particles were magnetized by passing a neodymium iron boron magnet (Bunting Magnetics Co.) over them several times. Particles coated with this method will be referred to as Particle 3. 3.1.3 Titanium n-Butoxide Sol-gel Coating Barium ferrite particles with an average particle size of 2.2 m were also coated with TiO2 using a TiO2 precursor by a procedure adapted from Chen et al. (1996). These particles were ball-milled in water with zirconia balls to achieve the smaller particle size than previously used (300-600 m). The particles were coated with a silica intermediate layer using a silica solgel process and then coated with TiO2. This technique creates

PAGE 28

18 some anatase phase TiO2 but requires further heat treatment for increased anatase crystal formation. Silica was coated onto the barium ferrite to add an insulating layer to hinder photodissolution, conduction band electron interaction, and to provide an attachment surface. The silica coating procedure is as follows: 300 mg of 2.2 m barium ferrite particles were dispersed in 250 mL of isopropanol (Acros Organics) and sonicated for 5 min in an ultrasonic cleaning bath. After sonication, 1 mL of tetraethylorthosilicate (TEOS, Aldrich), 4 mL of deionized (DI) water (18.2 Mcm, Barnstead E-pure), and 0.15 mL of ammonium hydroxide (Aldrich) were added to the sonicated solution. The prepared solution was vigorously stirred for 3 hr at room temperature. The silica-coated barium ferrite particles were recovered from the solution by magnetic separation, washed with isopropanol and DI water, and finally dried at room temperature for 7 days. The prepared silica-coated barium ferrite particles were dispersed in a mixture of 140 mL of ethanol (Aldrich) and 3 mL of DI water and sonicated for 10 min. Next, 0.45 mL of titanium n-butoxide (TBOT, 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 was 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 next 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 sample was then dried in an oven at 60 C for 2 days. The dried sample was taken and heat-treated in a box furnace at 500 C for 1 hr in an ambient atmosphere. The particles were magnetized by passing a

PAGE 29

19 magnet over them several times. Particles coated with this method will be referred to as Particle 4. 3.2 Magnetic Reactor Configuration The magnetic reactor consisted of a magnetic field generator, ultraviolet (UV) light tubes, and a glass reactor tube. Two generations of reactor configurations were built, and each one tested two different magnetic solenoids. In each instance, two different glass reactor tube geometries were used for each coil configuration. The first generation coil consisted of 1000 ft of 18 AWG wire wound around a 2.5 in inside diameter polycarbonate tube. The wire was wound in 16.5 layers with 60 turns per layer. Three UV light tubes were installed on the inside of the tube. The coil/UV assembly was mounted on an adjustable pivot for pitch adjustment. The pitch adjustment allowed for gravity to be used to balance out excessive forces from flow and magnetic field gradients. See Figure 3-1 for reactor setup. This reactor configuration will be referred to as MAPR 1. Straight tube glass reactor cells were used in the first generation coil. Two cell sizes were tested, each made of Pyrex glass (see Figure 3-2). Cell 1 was 20 cm long with a 2.54 cm outside diameter. This cell contained a porous glass frit midway through the cell. The cell was capped with one inch Teflon and stainless steel Swagelock compression fittings with 6 mm stainless steel straight hose barbs. Cell 2 was 13.5 cm long with a 10 mm inside diameter. The cell was capped with two Teflon fittings with o-rings containing 8 mm in diameter, 8 m porous glass frits (see Figure 3-3). These cells will be referred to as Cell 1 and Cell 2, respectively.

PAGE 30

20 Figure 3-1. MAPR 1 configuration. Side View Pivot Pitch Adjustment Rail Pin FrontView UV Lamp 20cm 2.54 cm OD 6 mm OD Glassfrit Swagelock compression fitting Cell 1 13.5 c m o-ring 10 mm ID Teflon fittin g Cell 2 Figure 3-2. Cells 1 and 2 for MAPR 1.

PAGE 31

21 10 mm Front View o-ring Bottom View 8 mm D frit Figure 3-3. 8 mm Teflon fittings with o-rings. The second generation coil consisted of 500 ft of 18 AWG wire wound around a 3.5 in diameter wire spool. The wire was wound in 20 layers with 23 turns per layer. The coil fit under a 3 lamp UV assembly. See Figure 3-4 for the reactor setup. This reactor will be referred to as MAPR 2. The second generation coil required a glass cell to fit the spool diameter. A 10 cm long 8 mm inside diameter Pyrex glass tube was bent to match the spool curvature. This sat flat on the horizontally positioned coil. Two 4 cm long, 10 mm inside diameter Pyrex glass tubes were joined to the curved tube in the vertical orientation (see Figure 3-5). The cell was capped with two Teflon fittings with o-rings containing 8 mm in diameter, 8 m porous glass frits (see Figure 3-3). This cell will be referred to as Cell 3.

PAGE 32

22 Top View Coil UV Lamps Front View Side View Figure 3-4. MAPR 2 configuration.

PAGE 33

23 Front View 3 . 5i nD 10 mm ID 8m m I D Figure 3-5. Cell 3 for Reactor 2. Each coil was operated with a Kenwood Sovereign MX-5000 power amplifier and a Hewlett Packard 3314A function generator set at an output signal of 1.2 volts. The UV lamps were provided by 12-inch long 8-watt tubes of 365, 312 and 254 nm nominal wavelengths. UV intensity measurements were made with a UVP UVX radiometer. 3.3 Particle and Reactor Testing 3.3.1 Frequency Optimization The simple model used to calculate the frequency for agitation only provided an approximate value used to determine the range in which the particles will resonate. The optimum frequency was determined through experimentation. Light transmittance through the particle matrix and contaminant removal were used to determine the optimal agitation frequency.

PAGE 34

24 3.3.2 Photomeasurement The photomeasurement for light transmittance was set up on MAPR 2. The materials included 200 mg of the 2.2 m particles and a 1 cm quartz cuvet. The quartz cuvet was filled with 200 mg of the particle and positioned in the area of maximum magnetic field gradient on the coil. The cuvet was clamped to prevent any movement or vibration. A silicon photodetector (UDT Sensors, Inc.) was positioned and clamped on one side of the cuvet. A blue LED was positioned on the opposite side of the detector and immobilized. See Figure 3-6 for an assembly diagram. An Agilent E3614A DC power supply powered the LED with 3.8 V. The silicon detector was connected to a Hewlett Packard 3478A multimeter with BCN style connectors. An Omega Precision Fine Wire Thermocouple was installed next to the cuvet assembly and the temperature was monitored by a Barnant Thermocouple Thermometer. The multimeter was interfaced with National Instruments LabView 5.1 for data acquisition and the program monitored the silicon detector, logging a data point every 0.5 second for each run. All electrical equipment was turned on and allowed to warm up for 30 minutes prior to testing. The entire photomeasurement setup was covered with a black box and then a black cloth to eliminate incidental light. First, an initial baseline reading with no agitation was taken to measure voltage drift. The magnet was then turned on and data were collected at different frequencies for 30 seconds at each frequency. The frequency range tested was 0 to 70 Hz at 5 Hz intervals. Data was collected while ramping the frequency up and then down for each trial. Temperature data were also collected over the entire trial run.

PAGE 35

25 Black Box Cuvet LabView 5.1 Multimeter DC Powe r Blue LED Thermocouple Thermocouple Thermomete r Silicon Detecto r Coil Figure 3-6. Photomeasurement setup on MAPR 2. 3.3.3 Dye Removal The optimum agitation frequency by contaminant removal was determined by experimentation in a batch system. MAPR 2 was used along with 100 mg of the 2.2 m coated particles. A 10 mg/L reactive red dye (RR) (Aldrich) solution was used as the contaminant to be removed. Reactive red dye was used in this experiment because of its ease of measurement. RR can be used as a surrogate for many environmental contaminants and results are assumed to apply to other contaminants. Ten milliliters of the RR solution was added to Cell 3 along with the particles. The coil was then energized at a set frequency along with three 365 nm UV lamps. The system was operated for 30 minutes and then absorbance was measured at 538 nm on a Hach DR/4000U spectrophotometer. The results were shown as decolorization and not removal because the analysis showed only the removal of substances that absorb light at 538 nm. Spectrophotometric analysis does not show complete mineralization of RR, therefore,

PAGE 36

26 intermediate substances may still be present. This process was repeated for each frequency over a frequency range from 0 to 75 Hz at 5 Hz intervals. 3.3.4 Digestion The 2.2 m particles were digested for emission spectroscopy using inductively coupled plasma (ICP) analysis. First, 5 mg of coated particles were added to 50 mL of 36 N sulfuric acid (Fisher) in a 250 mL round bottom, single neck flask with a magnetic stir bar. Then the flask was connected to a condenser and put on a hot plate stirrer. The condenser was activated with water and the hot plate was turned on to heat the solution at approximately 100 C. Samples were digested for 1 week and then diluted to 10 %. 3.3.5 Initial Batch Testing Photocatalytic degradation abilities of all synthesized particle batches were first tested in batch studies. Batch studies were conducted in a rotating mixer containing one 365 nm UV light. The rotating mixer can hold up to ten 40 mL vials and operates by inverting them at a rate of 24 rpm. Batch studies were also conducted in the MAPR 1 and 2 configurations. The following procedure applied to particles tested in the rotating mixer. The 300 – 600 m particles coated by the Theta Composer and the TTIP precursor were tested in this manner. First, a 10 mg/L RR solution was prepared. Particles were then added to 40 mL vials in 10, 100, and 1000 mg increments. The vials were filled with 40 mLs of the 10 mg/L RR solution and then rotated with the UV light on for four hours. The samples were then centrifuged for 15 min to settle out particulates. Each sample was then analyzed for absorbance with a Hach DR/4000U spectrophotometer at 538 nm. The 2.2 m particles coated with the titanium n-butoxide were tested in batch using MAPR 1 and 2. Tests in MAPR 1 used Cell 2 and was capped with rubber stoppers.

PAGE 37

27 Tests in MAPR 2 were carried out in Cell 3. MAPR 1 used a 5 mg/L RR solution. MAPR 2 batch tests used a 5 mg/L RR solution and a 10 mg/L phenol solution. Tests in each reactor configuration were carried out by adding 100 mg of particle to the reactor cell. Next, 10 mL of stock solution (either RR or phenol) was added to the test cell and the magnetic coil was energized at 45 or 50 Hz with three 365 nm UV lamps. MAPR 1 used a frequency of 45 Hz and MAPR 2 used a frequency of 50 Hz. Samples for the first generation reactor were collected after two hours with no UV and magnetic agitation, UV and no magnetic agitation, and UV with magnetic agitation. Samples were also taken with UV and magnetic agitation after 4 hours. Tests for MAPR 2 used RR and phenol. Samples were taken after 30 min. with UV and no magnetic agitation and with UV and magnetic agitation for the RR trials. For the phenol trials, samples were taken at 1, 2, 3, and 4 hours. The RR samples were analyzed for 538 nm absorbance and the phenol samples were analyzed for 270 nm absorbance on the Hach DR/4000U. 3.3.6 Flow Through Studies Flow through degradation studies were conducted for the first generation reactor and the second generation reactor. MAPR 1 was tested only with a 10 mg/L RR solution. Reactor 2 was only tested with phenol with solution concentrations of 10 mg/L and 120 g/L. Flow was provided by a Masterflex L/S standard digital drive pump with a PTFE pump head using 6 mm PTFE tubing. MAPR 1 was tested with Particle 3. Particle 3 was first tested for adsorption. The reactor was operated at 45 Hz with no UV for 3.5 hours. One liter of RR solution was recirculated at a rate of 5 mL/min with 2 g of Particle 3. Samples were taken over time and analyzed with the spectrophotometer. Four grams of Particle 2 were then added to

PAGE 38

28 the reactor and a one liter solution of RR was recirculated with the three 365 nm UV lamps on. Samples were taken over time and analyzed. MAPR 2 was tested with Particle 4. This reactor was tested by varying particle mass, UV wavelength, number of lamps, and changing flow rates. Particle masses of 100, 200, 300, 500, and 1000 mg were tested. Two hundred milliliters of phenol solution was recirculated in each experiment. The reactor was operated at 50 Hz. First, 100 mg of particles were tested with three 365 nm UV lamps at a 10 mg/L phenol concentration and flow rate of 5 mL/min. Next, 200 mg of particles were tested with three 365 nm lamps at 10 mg/L using flow rates of 5 and 10 mL/min. Samples were taken over time and analyzed with the UV spectrophotometer at 270 nm. Samples from the 200 mg, 5 mL/min tests were also analyzed with a GC/MS. Next, 200 and 300 mg of particles were tested on a 120 g/L phenol solution with three UV lamps of 365 nm wavelength. It was then discovered that 312 nm lamps provided more phenol removal then the 365 nm lamps so subsequent tests used this wavelength. The 500 and 1000 mg of particles were tested at a 120 g/L phenol solution using a flow rate of 5 mL/min with one 312 nm UV light. Finally, 500 mg of particle were then tested with one and three 312 nm UV lamps to find the best combination. Samples were taken over time and analyzed with the GC/MS.

PAGE 39

CHAPTER 4 RESULTS AND DISCUSSION 4.1 Initial Iterations It must be noted that several iterations of particle and reactor development were required before reasonable progress was made in this research. The materials, methods and results discussed in this thesis include only trials that produced encouraging results that ultimately led to the final iteration of this project. The following is a brief overview of the trials and processes that led to the work represented in this thesis. There were two major points of concern in developing the objectives of this research project: particle development and magnetic agitation. In pursuing this project it was immediately recognized that TiO2 coating durability would largely impact the success of this research. Magnetically agitating particles will subject them to high attrition conditions and heat stress. Concerns of deleterious attrition conditions bring up the question of how to actually magnetically agitate the particles. In contemplating this challenge, particle size, AC or DC magnetic fields, magnet geometry, and configuration were considered. Sol-gel methods were attempted to coat titania onto the barium ferrite surface. The coating procedure coated a thin film of TiO2 onto the surface. The TiO2 was amorphous phase and required heat treatment to produce the anatase phase required for photocatalytic activity. The particles were heat treated at 500, 550, and 600 C for one hour in air and under nitrogen. The particles were tested in the MAPR. Photocatalytic activity was observed as the RR solution was getting clearer. This was evident upon 29

PAGE 40

30 visual inspection since the spectrophotometer gave incorrect readings because of TiO2 detachment. Many factors have been the cause of this outcome. The coating may have come off due to attrition, or due to the universal solvent nature of water, or because most of the TiO2 was not bonded to the surface of the barium ferrite but simply agglomerated. The coatings were studied before and after use with SEM images. The TiO2 was clearly visible coating the entire surface before use. After use, the amount of TiO2 present on the surface was negligible. Another drawback of heat treating the particles is that they become more brittle after heat treatment. At times, some of the particles in the test reactor would disintegrate into fine particles and turn the entire solution in the reactor black, blocking any UV light penetration. To avoid problems arising from the need to heat treat the particles, anatase phase TiO2 nanoparticles (Degussa p-25) were used instead of a precursor in a modified method of the TTIP sol-gel method. This method took advantage of the positive charge imparted on the barium ferrite surface from the PEI. The anatase TiO2 powder was added instead of the precursor with the solution kept above the point of zero charge of TiO2. This insured that the TiO2 had a negative charge and would be electrostatically attracted to the positively charged barium ferrite surface. The solution was refluxed for 20 – 24 hours and dried, revealing a TiO2 thin-film coating. With no need to heat treat the particles, the particle structural integrity would not be compromised. The particles were tested and the results were similar in that the coating was detaching. Still, other coating methods were investigated. Silica gel technology was used as a medium in which to suspend 12% TiO2 and different masses of barium ferrite particles.

PAGE 41

31 Silica gels were thought to be a novel idea because they provided the advantage of adsorption while providing an ideal medium for supporting the TiO2 catalyst and magnets. Initial results showed moderate magnetic agitation. Pellets were made with different magnet ratios in order to find an optimum dosage. However, magnetic agitation proved to be deleterious to the silica structural integrity. The silica pellets disintegrated into a powder after a few hours of magnetic agitation and, therefore, further development was ceased. Several concepts were visualized for use as magnetic particle agitators. First, a General Electric, 1 hp, AC motor armature powered with 12 volts AC at 60 Hz was used to try and fluidize 300 – 600 m barium ferrite particles. Fluidization was very poor and this concept was abandoned in search of more viable concepts. Other concepts were considered in search of a usable alternative ranging from pulsing DC magnets and Helmholtz coils, to solenoids. To test the solenoid idea, a 40 mL vial containing 300 – 600 m barium ferrite particles was directly wound with approximately 20 ft of 18 AWG wire and connected to a variable AC power supply. A small amount of voltage was run through the coil and the particles began to agitate. This simple test provided proof of concept. The driving force behind propelling particles is to maximize the magnetic field gradient, which could be accomplished by winding the solenoid in a gradient. Coil geometries were wound with varying core diameters and wire lengths and gages. Experimentation with these coils led to the conclusion that very large coils requiring high currents would be needed to provide the necessary gradients for optimum fluidization. Due to the constraints of the system, compact size and low power required by NASA, this

PAGE 42

32 was not a feasible option. To stay within the constraints, standard solenoid windings were used as described in Chapter 3. Although these windings provided a more uniform magnetic field, maximizing the strength of the magnetic field, there was sufficient magnetic field gradient along with magnetic field strength at the ends of the solenoid to provide adequate force for agitation. 4.2 Particle Characterization The magnetic particles were characterized using a variety of techniques. Images were taken by a scanning electron microscope (SEM, JEOL 6335F) and a transmission electron microscope (TEM, JEOL 200CX). Energy-dispersive X-ray spectroscopy (EDS, Oxford) and X-ray diffraction (XRD, Philips APD 3720 diffractometer) analyses were conducted for elemental analysis. Particle coating composition was analyzed with inductively coupled plasma spectrometry (ICP, Perkin-Elmer Plasma 3200). Particle magnetization analysis was conducted with a superconducting quantum interference device (SQUID magnetometer, Quantum Design MPMS XL 7T). The particle size distribution was measured with an Aerosizer (Ahmherst Process Instrument Co. Inc.). The specific surface area was measured using a BET analyzer (Quantachrome NOVA 1200). 4.2.1 300 – 600 m Particles Figure 4-1 shows the hysterisis loop from the magnetometer results for the uncoated and the TTIP coated particles. The uncoated particle remnant magnetization was 28 emu G/g, and the coated particle remnant magnetization was 26 emu G/g. From these results it is evident that the TiO2 coating does not significantly impact the magnetization of the particles.

PAGE 43

33 0123452025303540 BaFe TiBaFeT = 300 KBaFe and TiBaFe powder M (emu G / g)B (Tesla) -6-4-20246-40-2002040 Figure 4-1. 300 – 600 m particle magnetism. The smooth, coated surface of the barium ferrite particles is shown in Figure 4-2, which shows an SEM image of the TTIP coated Particle 3 before contaminant removal testing. The agglomerated smooth topography of the TiO2 coating is clearly seen covering the barium ferrite particles. Experiments using these particles with magnetic agitation resulted in TiO2 coating detachment from the barium ferrite surface as shown in Figure 4-3. Attrition is the most likely cause of the TiO2 detachment and this coating

PAGE 44

34 instability led to the development of smaller particles synthesized with different techniques. Coated, agglomerated particles Figure 4-2. Particle 3 before testing.

PAGE 45

35 Uncoated particles after use Figure 4-3. Particle 3 after testing with magnetic agitation. 4.2.2 2.2 m Particles Particle 4 was characterized using all of the aforementioned techniques. The particle size distribution is shown in Figure 4-4, and the average particle size is 2.2 m. Decreasing the particle size leads to a more appropriate particle as the increased surface area will enhance photocatalytic efficiency and improve mass transfer. A TiO2 loading of 11 % was determined from ICP analysis results of the digested particles. Specific surface area analysis revealed a BET surface area of 27 m2/g for the coated particles as compared to 4 m2/g for the uncoated particles. The increase in surface area is provided by the amorphous silica layer and the TiO2 nanoparticle layers. Amorphous silica is used as an adsorbent in some applications and can have a surface area of up to 200 m2/g. TiO2

PAGE 46

36 particles also have an available surface area. Degussa P-25 TiO2 has a reported surface area of 50 m2/g. Particle attributes are summarized in Table 4-1. Table 4-1. Ball-milled barium ferrite attributes. Average particle size (m) 2.2 TiO2 loading (wt %) 11 BET Surface Area (m2/g) 27 01234567890.1110Particle size (microns)Differential volume (%) Figure 4-4. Ball-milled barium ferrite particle size distribution. The magnetization data possesses a hysteresis loop as shown in Figure 4-5. Magnetic fields used in the MAPR are very small (~ 0.004 – 0.008 tesla); therefore, the particle magnetization of interest is close to the zero tesla point in the figure. At zero tesla, the particle magnetization is called the remnant magnetization and this value is used in the particle resonant frequency calculations. The coated particle remnant magnetization is 24 emu G/g, a value that is similar to that obtained for the 300 – 600 m particles.

PAGE 47

37 -0.4-0.20.00.20.4-1.0-0.50.00.51.0 Msat = 42.65 emu G / gramMrem = 23.7 emu G / gram-0.5 T to 0.5 TT = 300 K M / MsatB [T]Oigin DemoOrigin DemoOrigin DemoOmoOin DemoOrigin DemoOrigin DemoOrigin DemoOigin DemoOrigin Demorigin DemoOrigin DemoOrigin DemoOrigin DemoOrigin DemoOrigin DemoOrigin Demo O r rig r Hc = ~ 52 Grigin DemoOrigin De Msat = 42.65 emu G/gram Mrem = 23.7emu G/gram Figure 4-5. 2.2 m particle magnetism. The SEM image of the uncoated ball-milled barium ferrite particles (Figure 4-6) revealed the smooth angular surface of the barium ferrite. Figure 4-7 shows the EDS spectrum of the uncoated barium ferrite, and the elemental composition consists primarily of barium, iron, and oxygen.

PAGE 48

38 Figure 4-6. SEM image of uncoated ball-milled barium ferrite. Figure 4-7. EDS of ball-milled barium ferrite. Figures 4-8 and 4-9 show the SEM image and EDS spectrum of the silica coated 2.2 m barium ferrite particles. The silica coating has a smooth topography covering

PAGE 49

39 agglomerated barium ferrite particles. The particles become agglomerated because the barium ferrite is slightly magnetized when stirred with a magnetic stir bar during the coating procedure. Agglomerates can also form during the silica deposition. The elemental composition shown by the EDS spectrum confirms the presence of silica on the particles (Figure 4-9). A TEM image of these particles is presented in Figure 4-10. The silica silhouette is clearly seen surrounding the barium ferrite particle and is smooth representing an amorphous coating. The SEM and TEM images show complete coverage of the barium ferrite surface by silica. Figure 4-8. SEM image of silica coated ball-milled barium ferrite.

PAGE 50

40 Figure 4-9. EDS of silica coated ball-milled barium ferrite. Silica Layer Figure 4-10. TEM of silica coated barium ferrite.

PAGE 51

41 Figures 4-11 and 4-12 show the SEM image and EDS spectrum of the TiO2-silica coated 2.2 m ferrite particles. The TiO2 coating has a smooth topography covering agglomerated silica-barium ferrite particles. The elemental composition shown by the EDS spectrum confirmed the presence of titanium on the particles. A TEM image of these particles is presented in Figure 4-13, and the TiO2-silica silhouette is clearly seen surrounding the barium ferrite particle. The TiO2 layer is granular, representing a crystalline phase coating made up of TiO2 nanoparticles. The SEM and TEM images show complete coverage of the barium ferrite surface by TiO2. Figure 4-11. SEM image of TiO2-silica coated ball-milled barium ferrite.

PAGE 52

42 Figure 4-12. EDS of TiO2-silica coated ball-milled barium ferrite Figure 4-13. TEM of TiO2-silica coated barium ferrite.

PAGE 53

43 XRD analysis showed the presence of anatase phase TiO2 on the coating of Particle 4. Figure 4-14 and Figure 4-15 present the XRD patterns for uncoated barium ferrite TiO2-silica coated particles, respectively. The triangles show the fingerprint of anatase peaks and the strongest peak is at 25.6. The peak is very short and broad because of the low TiO2 loading and crystal size as compared to the barium ferrite. The strong, sharp peaks are from the large crystal size and amount of barium ferrite present as compared to TiO2. To verify the presence of anatase TiO2, extra precipitate of TiO2 particles from the coating process was heat treated and analyzed with XRD. Figure 4-16 presents the XRD pattern. The peaks for anatase phase are clearly stronger and sharper confirming the presence of anatase phase TiO2. Figure 4-14. XRD of uncoated ball-milled barium ferrite.

PAGE 54

44 Figure 4-15. XRD of TiO2-silica coated barium ferrite. Figure 4-16. XRD of TBOT TiO2 precipitate heat treated at 500 C.

PAGE 55

45 4.3 MAPR Specifications and Characterization MAPR 1 was configured so that the UV lamps and the glass cell would pass through the inner diameter of the magnetic coil. Figure 4-17 shows the calculated magnetic field profile and gradient along the centerline of the coil where the magnetic field is the strongest. The distance (z) of zero on the x-axis represents the position in the center of the coil windings and the strength of the magnetic field is maximized at this point. Although the field strength is greatest at this point, it is also relatively uniform with only a weak magnetic field gradient present. Therefore, the reactor cell was positioned toward the end of the coil where the field gradient is larger. This position provided good agitation for Particle 3 as determined by visual inspection. When this position was tested for Particle 4, varying degrees of contaminant removal was observed at different frequencies. Due to the reactor geometry and cell position visual inspection of particle movement was difficult. Particle agitation was assumed similar to that of Particle 3. Upon further investigation, visual inspection of particle movement with a mirror, which provided a means to better view the inside of the coil, showed very low amounts of fluidization. The smaller particles agitated along the bottom of the cell at varying speeds corresponding to the operating frequency. There was almost no fluidization in the vertical direction. The difference in vertical fluidization between Particle 3 and Particle 4, neglecting drag forces, is explained by the difference in the size of the particles. Although both particle sets have comparable magnetizations, it is important to notice that this value is per unit mass of particle. The larger particles (Particle 3) have more mass and therefore have more magnetization as a whole. The force provided by the magnetic field strength and gradient was sufficient in this

PAGE 56

46 arrangement to provide acceptable agitation for the large particles but not for the smaller ones. 0102030405060708090-8-6-4-202468Z (cm)B/I (Gauss/Amp)-15-10-5051015B/Z (Gauss/cm) Figure 4-17. Magnetic field profile and gradient field for MAPR 1. MAPR 2 was designed to provide the necessary fluidization for Particle 4. This could be achieved by maximizing the magnetic field gradient. Because of the power and size constraints of the system, a coil winding similar to that one used in MAPR 1 was considered. It was found that by changing the entire configuration of the reactor, the area of maximum field gradient could be used more effectively (see Figure 3-3). The calculated magnetic field profile and gradient field along the centerline of MAPR 2 is shown in Figure 4-18. The area of maximum field gradient is outside of the inner coil diameter. Cell 3 was designed to fit the curvature of the coil in order to fully utilize the maximum gradient position. This reactor configuration provided vertical fluidization of Particle 4 as determined by visual inspection.

PAGE 57

47 05101520253035404550-8-6-4-202468Z (cm)B/I (Gauss/Amp ) -8-6-4-202468B/Z (Gauss/cm) Figure 4-18. Magnetic field profile and gradient field for MAPR 2. Due to the low flow rates and smooth glass tubing, the flow in the glass cells is assumed to be laminar. The calculated Reynolds number, using equation 8, for Cell 3 is 13. Generally, laminar flow is described by Reynolds numbers less than 2100. Mixing in this system is dominated by turbulence caused by the oscillating particles and molecular diffusion. The Reynolds number is defined as Re = VD (8) where is the fluid density, V is the mean fluid velocity, D is the pipe diameter, and is the fluid viscosity. UV intensities for MAPR 2 are shown in Table 4-2 for the 365 and 312 nm lamps at a distance of 3.5 cm.

PAGE 58

48 Table 4-2. UV intensities for MAPR 2 at 3.5 cm. Intensity (mW/cm2) Wavelength (nm) 1 lamp 2 lamps 3 lamps 365 1.8 3.7 5.0 312 2.5 5.4 7.3 4.4 Magnetic Reactor Frequency Optimization A photomeasurement for light transmittance was conducted to quantify the optimum frequency for maximum agitation for the MAPR 2 configuration using Particle 4. The results of this test are presented in Figure 4-19. Light transmittance measurements were taken by increasing the frequency from zero to 70 Hz in 5 Hz increments and then decreasing the frequency in the same manner. Two data points were logged every second for 30 seconds at each frequency interval. Temperature data was logged over the entire experimental run. Increasing agitation will scatter more light, decreasing the light that excites the photodetector. The optimum agitation frequency will have the largest change in voltage. Figure 4-19 shows that the optimum agitation frequency is approximately 50 Hz. Both increasing frequency and decreasing frequency measurements showed the same trend. Each curve includes one standard deviation error bars and the increasing temperature had no effect on the measurements.

PAGE 59

49 00.020.040.060.080.10.120.14020406080Frequency (Hz) V (volts)0510152025303540Temperature (C) Increasing f Decreasing f Temp. at increasing f Temp. at decreasing f Figure 4-19. Agitation measurement as a function of light transmittance. The maximum amount of light transmittance at 50 Hz is hypothesized to correlate with the maximum potential for contaminant removal. Agitation at this frequency will expose the particles to maximum amount of UV light and also enhance mass transfer by the increased mixing. RR decolorization was measured after 30 min at each frequency as described above. Results of this test are presented in Figure 4-20. RR decolorization was observed at 0 Hz because particle surface area is still being exposed to UV light when the particles are not fluidized. Maximum RR decolorization is observed at 50 Hz as expected.

PAGE 60

50 01020304050607080901000510152025303540455055606570Frequency (Hz)% Decolorization Figure 4-20. Photocatalytic decolorization of 10 mg/L reactice red dye with 200 mg of particle after 30 min. 4.5 Reactor Characterization and Particle Performance Reactive red dye was used as a surrogate pollutant because of its ease in measurement for assessment of initial particle performance. The stable, most functional particle was tested to degrade phenol, a ubiquitous compound occurring naturally and also anthropogenically produced. Phenol is primarily used in the production of phenolic resins for use in plywood, adhesive, automotive, and appliance industries. All particles were first tested in batch with a RR solution. Particles 1, 2, and 3 were tested with the rotating mixer because MAPR 1 was still in development. Particle 4 was tested in batch with MAPR 1 and 2. 4.5.1 Batch Tests Particle 1 was mechanically coated with the Theta Composer with a silica intermediate layer and Degussa P-25 TiO2. After testing these particles for 4 hr, the

PAGE 61

51 solution was cloudy with detached TiO2. The RR solution color was much lighter by visual inspection, but the amount of suspended TiO2 eliminated the need for further analysis, and this particle series was abandoned. The observed removal was due, in part, to slurry photocatalysis and not a result of effective composite particle catalysis. Particle 2 was also coated with the Theta Composer with a silica-TiO2 composite powder. There was slight coating detachment observed in the solution after the 4 hr test with 1000 mg of particle. On visual inspection, the RR solution color was less intense than the 10 mg/L stock but not as intense as observed for Particle 1 as a result of the decreased TiO2 loading. The 10 mg and 100 mg tests showed no color change by visual inspection. The samples were centrifuged for 15 min and analyzed with a spectrophotometer. The results are shown in Figure 4-21. Coating detachment and low photocatalytic efficiencies eliminated the need for further testing with these composite particles. 051015202530354045101001000Particle mass (mg)% Decolorization Figure 4-21. Four hour photocatalytic decolorization of a 10 mg/L RR solution with Particle 2.

PAGE 62

52 Mechanical coating techniques did not create the necessary permanence of bond to the barium ferrite surface. The physical bond characteristics were not strong enough to retain the coatings in the low attrition environment of the rotating mixer. Magnetic agitation would provide substantially more agitation, exposing the particles to more vigorous attrition conditions. For these reasons, chemical coating techniques, specifically sol-gel techniques, were pursued to provide more stable and durable TiO2 thin films. Particle 3 was coated with a sol-gel technique using TTIP as a precursor. Tests were conducted as with Particles 1 and 2. The results are shown in Figure 4-22. Upon visual inspection, no detached TiO2 was observed. The 1000 mg test yielded a 95% RR decolorization, followed by 18% for 100 mg, and 5% for 10 mg. The order in magnitude increase from 100 to 1000 mg of particles substantially increased photocatalysis by providing more catalyst surface for photon absorbance. These results were proved contradictory by later experimentation in a flow through system. As explained later, the removal observed in Figure 4-22 can be attributed to adsorption because the TiO2 coating was amorphous, not crystalline anatase phase, hence negating the possibility of photocatalytic action.

PAGE 63

53 0102030405060708090100101001000Particle mass (mg)% Decolorization Figure 4-22. Four hour adsorption of a 10 mg/L RR solution with Particle 3. Initial tests of Particle 4 were conducted using MAPR 1 and Cell 2. Two hour adsorption tests for 100 mg of the particle resulted in 10% removal. The same set of particle was tested for 2 hrs repeatedly. Figure 4-23 shows the results of the 2 hr tests for this particle set exposed up to 31 hrs of agitation. The average removal is 81%. The particle performance over time suggests that the particle coating is durable although a finite lifetime of the particles is expected due to the attrition conditions in the reactor. The deterioration of the particle is not shown in Figure 4-23 and the absolute lifetime is not yet known. Figure 4-24 shows the capacity of the particle to decolorize the RR solution over time. The particle decolorization rate levels off at about 80 % after one hour. The leveling off of the curve was not expected and this anomaly may be explained as an analytical error at the one hour or two hour points. This experiment was not duplicated but Figure 4-23 shows that there is some variability in the amount of decolorization after two hours. The last point may actually have a greater decolorization or the one hour point may have a lower decolorization. Another explanation that the decolorization did

PAGE 64

54 not follow the linear removal rate of the first hour may be because of low concentration of RR left to be removed after one hour. It will take longer to degrade a low concentration of RR because there are less RR particles to come in contact with the available hydroxyl radicals. 0102030405060708090100Adsorption4 hr10 hr17 hr25 hr31 hrParticle Life% Decolorization Figure 4-23. Two hour photocatalytic decolorization of a 5 mg/L RR solution with 100 mg of Particle 4.

PAGE 65

55 0102030405060708090100020406080100120140Time (min)% Decolorization Figure 4-24. Photocatalytic decolorization of a 5 mg/L RR solution over time with 100 mg of Particle 4. Particle 4 was also tested in batch with a 10 mg/L phenol solution. Figure 4-25 shows the removal over time as measured by the UV-spectrophotometer at 270 nm. Only 38% of the phenol was removed. These results demonstrated that photocatalytic degradation of phenol was possible in this system. Phenol adsorption on the particle surface was negligible as determined by agitating Particle 4 in the phenol solution with no UV exposure and analyzed with the UV-spectrophotometer.

PAGE 66

56 0102030405060708090100060120180240300Time (min)% Removal Figure 4-25. Photocatalytic degradation of a 10 mg/L phenol solution over time with 100 mg of Particle 4 as measured by the change in solution absorption at 270 nm. 4.5.2 Flow Through Tests Flow through tests were conducted in MAPR1 with Particles 3 and 4. The optimum agitation frequency for the 300 – 600 m size particle series was determined by visual inspection to be 45 Hz. Since Particle 3 and Particle 4 have about the same magnetizations, it was assumed that Particle 4 would also agitate well at 45 Hz. As described previously, observation of particle agitation was difficult due to the reactor geometry. Therefore, tests were performed under the assumption of good mixing until, on further inspection with a mirror, poor agitation was observed. Testing ended with this reactor scheme and MAPR 2 was designed to provide better fluidization of particle 4. Particle 3 was tested in Cell 1. The particles were first tested for adsorption. Results are shown in Figure 4-26. These particles showed high levels of adsorption. This is explained because the initial synthesized particles were coated with amorphous phase TiO2. Amorphous TiO2 has no photocatalytic ability so all the removal is

PAGE 67

57 explained by adsorption. The decrease in adsorption after the peak was unexpected and may be due to the desorption of RR. Although the temperature of the system with only the coil on was not measured, the coil did generate heat. When the system warmed up the adsorption characteristics may have changed causing desorption of RR. Some detached TiO2 particles were also observed in Cell 1 after use. 020406080100060120180240Time (min)% Decolorization ` Figure 4-26. Adsorption of 5 mg/L RR solution with 2 g of Particle 3. Particles were then heat treated at 500 C for 1 hr prior to use to develop the anatase phase. When magnetic agitation was applied, TiO2 detachment was observed immediately as a white plume flowing off of the particle surface. Testing continued and after one hour the particles were observed to have disintegrated into a powder. The barium ferrite structure was weakened by sintering and some photodissolution may have occurred. These particles were abandoned in favor of a 2.2 m TiO2-silica-barium ferrite

PAGE 68

58 composite. A silica intermediate layer was incorporated in these particles to prevent barium and iron leaching and also to prevent photodissolution. Flow through tests were conducted in MAPR 2 with Particle 4 and Cell 3. The first series of flow through tests degraded 200 mL of a 10 mg/L solution of phenol with 100 mg and 200 mg of particle 4 at a flow rate of 5 mL/min. The 200 mg batch of particles was also tested at a flow rate of 10 mL/min. The results are shown in Figure 4-27. Photocatalytic performance with 200 mg was better than the performance with 100 mg. This was expected because the larger amount of particles increased the TiO2 loading in the system. An optimum mass loading is expected in the system for the best performance. An increase in flow rate decreased the photocatalytic performance because of the reduced retention time. Subsequent tests used a flow rate of 5 mL/min. The overall removal rate of the solution was low because of the high concentration of phenol used as compared to the amount of titanium dioxide in the system. Subsequent experiments used a more realistic phenol concentration of 120 g/L to show system performance. 012345678050100150200250Time (min)% Removal 200 mg @ 5 mL/min 200 mg @10 mL/min 100 mg @ 5 mL/min Figure 4-27. Photocatalytic degradation of 200 mL of a 10 mg/L phenol solution.

PAGE 69

59 Before further flow through testing continued, a short batch experiment was conducted to see if changing the UV wavelength would increase photocatalytic efficiency. Degradation of a 10 mg/L phenol solution for one hour with 300 mg of Particle 4 was tested with 365, 312, and 254 nm UV lamps. Results are shown in Figure 4-28, and it is clear that the 312 nm lamp had the greatest efficiency. These results are not entirely representative of the 254 nm lamp performance. This wavelength showed a low photocatalytic efficiency because the glass cell was made of Pyrex glass. Pyrex glass is only about 10% transmittant to 254 nm wavelengths (see Appendix A). Testing with fused silica glass could increase the photocatalytic efficiency with this wavelength but those tests were not conducted. 05101520253035365 nm312 nm254 nm% Removal Figure 4-28. One hour removal of a 10 mg/L phenol solution with 300 mg of Particle 4 at different UV wavelengths using a Pyrex glass cell. Pyrex glass is only about 10% transmittant to 254 nm wavelengths explaining the low removal rates for the 254 nm lamps.

PAGE 70

60 Next, 200 mg and 300 mg of particles were tested with three 365 nm lamps and 300 mg of particle was also tested with one 312 nm lamp. This experiment degraded 200 mL of a 120 g/L phenol solution. The results are shown in Figure 4-29. The higher TiO2 loading of the 300 mg particle mass performed better than the 200 mg of particles. One 312 nm lamp also performed as well as three 365 nm lamps with 300 mg of particles even though the intensity of one 312 UV lamp had half the intensity of three 365 nm lamps. Lower wavelengths have higher energy values compared to larger wavelengths and may increase the electron-hole production. Subsequent tests all used 312 nm lamps. 00.10.20.30.40.50.60.70.80.910246Time (hr)C/Co 200 mg, 365 nm 300 mg, 365 nm 300 mg, 312 nm Figure 4-29. Photocatalytic degradation of 200 mL of a 120 g/L phenol solution at a flow rate of 5 mL/min using either three 365 nm lamps or one 312 nm lamp and different particle masses. To find the optimum mass loading of particles in the system, 500 mg and 1000 mg of particles were used to degrade the 120 g/L phenol solution. Each test used one 312 nm lamp. The results are shown in Figure 4-30. The 500 mg particle mass outperformed both the 300 mg and 1000 mg of particle. These mass loadings can be converted to TiO2

PAGE 71

61 loadings per unit volume treated to make the results more adaptable to other systems. The loadings for the 300 mg, 500 mg, and 1000 mg particle masses are 165 mg TiO2/L, 275 mg TiO2/L, and 550 mg TiO2/L, respectively. The 500 mg and 1000 mg particle mass both had the same end result after 5 hr but the 500 mg of particles degraded the phenol at a faster rate. The 1000 mg particle does not degrade the phenol as fast as the 500 mg despite more TiO2 loading because of UV light blocking and scattering. The system opacity reaches a point at which UV light does not reach the maximum amount of particle surface, decreasing photocatalytic efficiency based on TiO2 loading. 00.10.20.30.40.50.60.70.80.910123456Time (hr)C/Co 300 mg 500 mg 1000 mg Figure 4-30. Photocatalytic degradation of 200 mL of a 120 g/L phenol solution with one 312 nm lamp. Using the optimal particle mass loading of 500 mg of Particle 4 for this system, the number of UV lamps used was optimized. Three 312 nm lamps were tested and compared to the removal efficiency of one lamp and the results are shown in Figure 4-31. From Table 4-2, three lamps have a greater intensity than one lamp, so it is expected to

PAGE 72

62 increase photocatalytic efficiency. Using three lamps did increase the rate of removal compared to one lamp after one hour while approaching complete removal after two hours as also observed with one lamp. Table 4-2 also compares the removal of Particle 4 compared to a 4 mg/L TiO2 slurry. From this test 500 mg of Particle 4 with three 312 nm UV lamps out performs the slurry. 00.10.20.30.40.50.60.70.80.910123456Time (hr)C/Co 1 312 nm lamp 3 312 nm lamps 4 mg/L P-25 slurry Figure 4-31. Photocatalytic degradation of 200 mL of a 120 g/L phenol solution with 500 mg of Particle 4 and a 4 mg/L Degussa P-25 TiO2 slurry with 3 lamps. A TiO2 slurry offers more surface area for mass transfer than the 2.2 m average particle size of Particle 4, so the comparative performance of both must be noted. In comparing the two experiments, tests with Particle 4 used magnetic agitation to increase mixing and mass transfer. The slurry was, essentially, circulated in plug flow based on the previous Reynolds number calculation indicating laminar flow and limited mixing decreasing the potential mass transfer. The TiO2 loading was greater in the tests involving Particle 4 at 275 mg/L as compared to the 4 mg/L loading of the slurry system. Although the TiO2 loading was greater for the composite particle tests, all of the TiO2 in

PAGE 73

63 the system was not available for mass transfer due to layering and the two orders of magnitude increase in particle size as compared to the slurry. Using the surface area of both the composite particles (27 m2/g) and the Degussa P-25 (50 m2/g), the total surface area available was calculated for the TiO2 loading in each system. The composite particle system offered 0.256 m2 of total surface and the slurry system offered 0.0726 m2 of total surface. The increased mass transfer due to magnetic agitation and the increased available surface of the composite particle can account for the improved photocatalytic performance as compared to the slurry system.

PAGE 74

CHAPTER 5 CONCLUSIONS Photocatalytically active magnetic particles were synthesized using a variety of techniques. Coating durability and stability were the primary obstacles to overcome. After many attempts, a working, durable, photocatalytically active composite particle was synthesized using a sol-gel technique. These TiO2-silica-baruim ferrite composites were composed of a hydrolyzed alkoxide (TBOT) to create the TiO2 layer, a silica intermediate layer, and ball-milled barium ferrite as the core. The particles were heat treated to form the anatase crystalline phase and had a mean particle size of 2.2 m. A simple model was used to estimate the resonant frequency as a function of particle magnetization and magnetic field gradient. This model accurately predicted the range of the resonant frequency. A frequency range within the predicted bounds was tested to determine the optimum agitation frequency using light scattering and dye degradation experiments. Both experiments correlated in returning an optimum agitation frequency of 50 Hz. Two iterations of MAPR development were required to produce the best system. These systems were developed simultaneously with particle development. The final particle size (2.2 m) required a MAPR redesign for the required magnetic fluidization (MAPR 2). MAPR 2 is a tightly wound coil used to produce the magnetic field gradient. A glass reactor cell is placed on the coil in the area of maximum magnetic field gradient and the reactor system is completed with three ballasts capable of operating up to three 8-watt 365 or 312 nm UV lamps. 64

PAGE 75

65 The 2.2 m TiO2-silica-barium ferrite composite particles were tested in batch and flow through systems. The batch experiments were conducted with reactive red dye and a 10 mg/L phenol solution to show initial photocatalytic oxidation capacity. The composite particles showed an 81% decolorization of RR after two hours and 35% removal of phenol after four hours with a particle mass of 100 mg. The composite particles were then tested in a flow through system with varying parameters. Particle mass, flow rate, UV wavelength, and the number of UV lights were varied to determine the optimum system operating parameters. The system is optima for phenol degradation are a 2.5 g/L composite particle loading with three UV lamps at a wavelength of 312 nm. The magnetic particles were also easily removed from the solution by applying an external magnetic field.

PAGE 76

APPENDIX A TRANSMITTANCE CURVES Transmittance curve for Pyrex glass (adapted from Corning, 2003). 020406080100230250270290310330350370390Wavelength (nm)% Transmittance 66

PAGE 77

67 Transmittance curve for fused silica glass (Del Mar Ventures, 2003).

PAGE 78

APPENDIX B UV INTENSITIES UV intensities of 8 watt, 12 in UV light tubes. 012345678910111213024681012Distance from light (cm)Intensity (mW/cm2) 365 nm 312 nm 68

PAGE 79

LIST OF REFERENCES Arana, J., Dona-Rodriguez, J. M., Tello Rendon, E., Garriga i Cabo, C., Gonzalez-Diaz, O., Herrera-Melian, J. A., Perez-Pena, J., Colon, G. and Navio, J. A., “TiO2 activation by using activated carbon as a support: Part I. surface characterization and decantability study,” Applied Catalysis B: Environmental, Vol. 44, pp. 161-172, 2003. Augustynski, J., “The role of the surface intermediates in the photoelectrochemical behavior of anatase and rutile TiO2,” Journal Electrochimica Actica, Vol. 38, pp. 43-46, 1993. Balasubramanian, G., Dionysiou, D. D., Suidan, M. T., Baudin, I. and Laine, J., “Evaluating the activities of immobilized TiO2 powder films for the photocatalytic degradation of organic contaminants in water,” Applied Catalysis B: Environmental, Vol. 47, pp. 73-84, 2004. Beydoun, D. and Amal, R., “Implications of heat treatment on the properties of a magnetic iron oxide-titanium dioxide photocatalyst,” Materials Science and Engineering, Vol. B94, pp. 71-81, 2002. Beydoun, D., Amal, R., Low, G. and McEvoy, S., “Novel photocatalyst: titania-coated magnetite. Activity and photodissolution,” Journal of Physical Chemistry B, Vol. 104, pp. 4387-4396, 2000. Beydoun, D., Amal, R., Low, G. and McEvoy, S., “Occurrence and prevention of photodissolution at the phase junction of magnetite and titanium dioxide,” Journal of Molecular Catalysis A, Vol. 180, pp. 193-200, 2002. Blake, D., “Bibliography of work on photocatalytic removal of hazardous compounds from water and air,” Update Number 3, NREL/TP-570-26797, Golden, CO: National Renewable Energy Laboratory, 1999. Chen, P. H. and Jeng, C. H., “Kinetics of photocatalytic oxidation of trace organic compounds over titanium dioxide,” Environmental International, Vol. 24, pp. 871-879, 1998. Chen, F., Xie, Y., Zhao, J. and Lu, G., “Photocatalytic degradation of dyes on a magnetically separated photocatalyst under visible and UV irradiation,” Chemosphere, Vol. 44, pp. 1159-1168, 2001. 69

PAGE 80

70 Chen, J., Gao, L., Huang, J. and Yan, D., “Preparation of nanosized titania powder via the controlled hydrolysis of titanium alkoxide,” Journal of Materials Science, Vol. 31, pp. 3497-3500, 1996. Corning: Life Sciences, http://www.corning.com/lifesciences/technical_information/techdocs/descglasslabware.asp#7740LowExpansion , September 2, 2003. Del Mar Ventures, http://www.sciner.com/Opticsland/FS.htm , September 2, 2003. Dijkstra, M. F. J., de Jong, A. W. F., Michorius, A., Winkelman, J. G. M. and Beenackers, A. A. C. M., “Experimental comparison of three reactor designs for photocatalytic water purification,” Chemical Engineering Science, Vol. 56, pp. 547-555, 2001a. Dijkstra, M. F. J., Michorius, A., Buwalda, H. J., Winkelman, J. G. M. and Beenackers, A. A. C. M., “Comparison of the efficiency of immobilized and suspended systems in photocatalytic degradation,” Catalysis Today, Vol. 66, pp. 487-494, 2001b. Gao, Y., Chen, B., Li, H. and Ma, Y., “Preparation and characterization of a magnetically separated photocatalyst and its catalytic properties,” Materials Chemistry and Physics, Vol. 80, pp. 348-355, 2003. Gogate, P. R. and Pandit, A. B., “A review of imperative technologies for wastewater treatment I: oxidation technologies at ambient conditions,” Advances in Environmental Research, Vol. 8, pp. 501-551, 2004. Gopal, M., Moberly Chan, W. J. and De Jonghe, L. C., “Room temperature synthesis of crystalline metal oxides,” Journal of Materials Science, Vol. 32, pp. 6001-6008, 1997. Hague, D. C. and Mayo, M. J., “Controlling crystallinity during processing of nanocrystalline titania,” Journal of the American Ceramic Society, Vol. 77, pp. 1957-1960, 1994. Herrmann, J., “Heterogeneous photocatalysis: fundamentals and applications to the removal of various types of aqueous pollutants,” Catalysis Today, Vol. 53, pp. 115-129, 1999. Hoffmann, M., Martin, S., Choi, W. and Bahnemann, D., “Environmental application of semiconductor photocatalysis,” Chemical Reviews, Vol. 95, pp. 69-96, 1995. Kobayakawa, K., Sato, C., Sato, Y. and Fujishima A., “Continuous-flow photoreactor packed with titanium dioxide immobilized on large silica gel beads to decompose oxalic acid in excess water,” Journal of Photochemistry and Photobiology A: Chemistry, Vol. 118, pp. 65-69, 1998.

PAGE 81

71 Lee, D., Kim, S. C., Cho, I., Kim, S. J. and Kim, S. W., “Photocatalytic oxidation of microcystin-LR in a fluidized bed reactor having TiO2-coated activated carbon,” Separation and Purification Technology, Vol. 34, pp. 59-66, 2004. Linsebigler, A. L., Guangquan L. and Yates J. T., “Photocatalysis on TiO2 surfaces: Principles, mechanisms, and selected results,” Chemical Reviews, Vol. 95, pp. 735-758, 1995. Londeree, D. J., Silica-titania Composites for Water Treatment. Master’s thesis, University of Florida, Gainesville, 2002. Lu, L., Ryu, K. S. and Liu, C., “A magnetic microstirrer and array for microfluidic mixing,” Journal of Microelectrochemical Systems, Vol. 11, pp. 464-469, 2002. Ollis, D. F., Pelizzetti, E. and Serpone, N., “Photocatalyzed destruction of water contaminants,” Environmental Science and Technology, Vol. 25 (9), pp. 1522-1529, 1991. Pelizzetti, E. and Minero C., “Mechanism of the photo-oxidative degradation of organic pollutants over TiO2 particles,” Journal Electrochimica Actica, Vol. 38, pp. 47-55, 1993. Rachel, A., Subrahmanyam, M. and Boule, P., “Comparison of photocatalytic efficiencies of TiO2 in suspended and immobilized form for the photocatalytic degradation of nitrobenzenesulfonic acids,” Applied Catalysis B: Environmental, Vol. 37, pp. 301-308, 2002. Rothenberger, G., Moser, J., Gratzel, M., Serpone, N. and Sharma, D. K., “Charge carrier trapping and recombination dynamics in small semiconductor particles,” Journal of the American Chemical Society, Vol. 107, pp. 8054-8059, 1985. Salaices, M., Serrano, B. and de Lasa, H. I., “Experimental evaluation of photon absorption in an aqueousTiO2 slurry reactor,” Chemical Engineering Journal, Vol. 90, pp.219-229, 2002. Serpone, N., “Brief introductory remarks on heterogeneous photocatalysis,” Solar Energy Materials and Solar Cells, Vol. 38, pp. 369-379, 1995. Sunada, K., Kikuchi, Y., Hashimoto, K. and Fujishima, A., “Bactericidal and detoxification effects of TiO2 thin film photocatalysts,” Environmental Science and Technology, Vol. 32, pp. 726-728, 1998. Suzuki, H. and Ho, C., “A magnetic force driven chaotic micro-mixer,” http://ho.seas.ucla.edu/Ho_Publications/200-Present/mems2002_magnetic_mixer.pdf , April 1, 2002.

PAGE 82

72 Tanaka, K., Capule, M. and Hisanaga, T., “Effect of crystallinity of TiO2 on its photocatalytic action,” Chemical Physics Letters, Vol. 187, pp. 73-76, 1991. Turchi, C. and Ollis, D., “Photocatalytic degradation of organic water contaminant: mechanisms involving hydroxyl radical attack,” Journal of Catalysis, Vol. 122, pp. 178-192, 1990. Watson, S., Beydoun, D. and Amal, R., “Synthesis of a novel magnetic photocatalyst by direct deposition of nanosized TiO2 crystals onto a magnetic core,” Journal of Photochemistry and Photobiology A: Chemistry, Vol. 148, pp. 303-313, 2002. Yamazaki, S., Matsunaga, S. and Hori, K., “Photocatalytic degradation of trichloroethylene in water using TiO2 pellets,” Water Resources, Vol. 35, pp. 1022-1028, 2001.

PAGE 83

BIOGRAPHICAL SKETCH Jack Drwiega was born on January 3, 1979, in New Britain, Connecticut. He attended middle and high school in Brooksville, Florida. Jack then started off his college career in the Fall of 1997 at the University of North Florida in Jacksonville and later transferred to the University of Florida to pursue an education in environmental engineering. After obtaining his bachelor’s degree in December, 2002, Jack remained at UF to pursue a master’s degree focusing on potable water treatment. His research on magnetically agitated photocatalysis has provided him a test in problem solving and the importance of perseverance in solving the tough problems that will benefit him throughout his career. 73