<%BANNER%>

Electrospinning of Nanofibers for Filtration Media

Permanent Link: http://ufdc.ufl.edu/UFE0041620/00001

Material Information

Title: Electrospinning of Nanofibers for Filtration Media
Physical Description: 1 online resource (121 p.)
Language: english
Creator: Park, Hyoung
Publisher: University of Florida
Place of Publication: Gainesville, Fla.
Publication Date: 2010

Subjects

Subjects / Keywords: anatase, aqueous, ceramics, electrospinning, filter, filtration, nanocomposite, nanofiber, nanomaterial, tio2
Materials Science and Engineering -- Dissertations, Academic -- UF
Genre: Materials Science and Engineering thesis, Ph.D.
bibliography   ( marcgt )
theses   ( marcgt )
government publication (state, provincial, terriorial, dependent)   ( marcgt )
born-digital   ( sobekcm )
Electronic Thesis or Dissertation

Notes

Abstract: Since particulate impurity is regarded as the primary cause of lung diseases, purification of air has been a crucial issue. Filtration is the most conventional method to obtain clean air, whereby particulate matter is collected on a fibrous media. The use of fibrous filters is prevalent because of their high filtration efficiency and low pressure drop. Fibrous filters were fabricated via the electrospinning process which can be used to produce continuous submicron-diameter sized fibers. Polyacrylonitrile (PAN) nanofibers with a mean fiber diameter of 224 nm were electrospun to form fibermats. Filtration tests on fibermats of PAN were conducted to confirm that filters of thinner fibers result in higher collection efficiencies and lower pressure drops than that of thicker fibers as predicted by the theoretical filtration mechanism. Results showed that electrospun PAN nanofibermats had a superior quality factor of 0.067plus or minus0 compared to 0.031plus or minus0.001 by the current state-of-the-art microfiber-based high particulate air (HEPA) filtration media. The verified theory implies that nanofibermats of other types of materials could also be considered as promising filtration media since filtration performance is independent of the material used. As materials for advanced next-generation filtration media, ceramics are favored over polymeric materials due to their robustness against environmental factors such as ultraviolet rays, abrasive particles, and high temperature all of which degrade and damage the fibrous structure. Amidst various ceramic materials, the anatase phase of TiO2 was selected due to its mechanical property and versatility as a photocatalyst and microwave-absorbing material. Anatase TiO2 fibers were fabricated by electrospinning followed by heat treatment at 500degreeC for 3 hours. However, early precipitation or gelation of the organic solvent-based TiO2 sol posed a practical challenge in the sample preparation. In order to enhance stability of the precursor sol, a novel aqueous sol with titanium alkoxide was developed. As the result, the time taken for gelation or precipitation was elongated from 4 hours for the organic solvent-based sol to 4 months with the novel aqueous sol. In seeking the proper chemical composition to attain electrospinnability and maximize the period for storage before gelation, the reaction paths of hydrolysis and condensation for one of the components of the aqueous sol were investigated by nuclear magnetic resonance (NMR) spectroscopy. After hydrolysis and condensation reactions, Si-O-Ti bonds were validated to be formed by the reaction mechanism. TiO2-SiO2 composite fibers were successfully electrospun from the aqueous sol system by addition of a spinning agent followed by heat treatment. In contrast to TiO2 fibers in which anatase phase was observed after heat treatment at 500degreeC, anatase phase was formed at 1100degreeC in TiO2-SiO2 composite fibers. The formation of Ti-O-Ti bonds was retarded due to the formation of Si-O-Ti bonds, as evidenced by the NMR results. In regard to the microstructure of TiO2 fibers and TiO2-SiO2 composite fibers with anatase phases, the TiO2-SiO2 composite fibers were observed to have no voids or cleavages on the surface than TiO2 fibers which have coarse structures created upon crystallization at magnification of x330,000 by transmission electron microscopy. The coarse structure of TiO2 fibers characterized as having cleavages at exposed surface grain boundaries is anticipated to adversely affect the mechanical stability by enhancing crack formation and propagation which will lead to failure of the fiber. In contrast, amorphous SiO2 fills in the spaces that have been created by the development of anatase phase for TiO2-SiO2 composite fibers. Smoother surfaces were observed as well in contrast with TiO2 fibers due to the amorphous SiO2 in the continuous phase of the composite material. Based on the observations, TiO2-SiO2 composite fibers are expected to have better mechanical stability by reducing the possibility of crack formation and blockage of crack propagation. The mean fiber diameter of TiO2-SiO2 composite fibers achieved was 243 nm, which is 8% thicker than the PAN fibers achieved and 54% thinner than fibers of the HEPA filter. Hence, the quality factor of the TiO2-SiO2 composite fibers is predicted to be between those of PAN fibermats and the HEPA filter by filtration theory; however it would be closer to that of PAN fibers. Moreover based on transmission electron microscopy (TEM) observation, the mechanical stability was improved as well by achieving denser structures in the fiber than in pure TiO2 fibers.
General Note: In the series University of Florida Digital Collections.
General Note: Includes vita.
Bibliography: Includes bibliographical references.
Source of Description: Description based on online resource; title from PDF title page.
Source of Description: This bibliographic record is available under the Creative Commons CC0 public domain dedication. The University of Florida Libraries, as creator of this bibliographic record, has waived all rights to it worldwide under copyright law, including all related and neighboring rights, to the extent allowed by law.
Statement of Responsibility: by Hyoung Park.
Thesis: Thesis (Ph.D.)--University of Florida, 2010.
Local: Adviser: Sigmund, Wolfgang M.

Record Information

Source Institution: UFRGP
Rights Management: Applicable rights reserved.
Classification: lcc - LD1780 2010
System ID: UFE0041620:00001

Permanent Link: http://ufdc.ufl.edu/UFE0041620/00001

Material Information

Title: Electrospinning of Nanofibers for Filtration Media
Physical Description: 1 online resource (121 p.)
Language: english
Creator: Park, Hyoung
Publisher: University of Florida
Place of Publication: Gainesville, Fla.
Publication Date: 2010

Subjects

Subjects / Keywords: anatase, aqueous, ceramics, electrospinning, filter, filtration, nanocomposite, nanofiber, nanomaterial, tio2
Materials Science and Engineering -- Dissertations, Academic -- UF
Genre: Materials Science and Engineering thesis, Ph.D.
bibliography   ( marcgt )
theses   ( marcgt )
government publication (state, provincial, terriorial, dependent)   ( marcgt )
born-digital   ( sobekcm )
Electronic Thesis or Dissertation

Notes

Abstract: Since particulate impurity is regarded as the primary cause of lung diseases, purification of air has been a crucial issue. Filtration is the most conventional method to obtain clean air, whereby particulate matter is collected on a fibrous media. The use of fibrous filters is prevalent because of their high filtration efficiency and low pressure drop. Fibrous filters were fabricated via the electrospinning process which can be used to produce continuous submicron-diameter sized fibers. Polyacrylonitrile (PAN) nanofibers with a mean fiber diameter of 224 nm were electrospun to form fibermats. Filtration tests on fibermats of PAN were conducted to confirm that filters of thinner fibers result in higher collection efficiencies and lower pressure drops than that of thicker fibers as predicted by the theoretical filtration mechanism. Results showed that electrospun PAN nanofibermats had a superior quality factor of 0.067plus or minus0 compared to 0.031plus or minus0.001 by the current state-of-the-art microfiber-based high particulate air (HEPA) filtration media. The verified theory implies that nanofibermats of other types of materials could also be considered as promising filtration media since filtration performance is independent of the material used. As materials for advanced next-generation filtration media, ceramics are favored over polymeric materials due to their robustness against environmental factors such as ultraviolet rays, abrasive particles, and high temperature all of which degrade and damage the fibrous structure. Amidst various ceramic materials, the anatase phase of TiO2 was selected due to its mechanical property and versatility as a photocatalyst and microwave-absorbing material. Anatase TiO2 fibers were fabricated by electrospinning followed by heat treatment at 500degreeC for 3 hours. However, early precipitation or gelation of the organic solvent-based TiO2 sol posed a practical challenge in the sample preparation. In order to enhance stability of the precursor sol, a novel aqueous sol with titanium alkoxide was developed. As the result, the time taken for gelation or precipitation was elongated from 4 hours for the organic solvent-based sol to 4 months with the novel aqueous sol. In seeking the proper chemical composition to attain electrospinnability and maximize the period for storage before gelation, the reaction paths of hydrolysis and condensation for one of the components of the aqueous sol were investigated by nuclear magnetic resonance (NMR) spectroscopy. After hydrolysis and condensation reactions, Si-O-Ti bonds were validated to be formed by the reaction mechanism. TiO2-SiO2 composite fibers were successfully electrospun from the aqueous sol system by addition of a spinning agent followed by heat treatment. In contrast to TiO2 fibers in which anatase phase was observed after heat treatment at 500degreeC, anatase phase was formed at 1100degreeC in TiO2-SiO2 composite fibers. The formation of Ti-O-Ti bonds was retarded due to the formation of Si-O-Ti bonds, as evidenced by the NMR results. In regard to the microstructure of TiO2 fibers and TiO2-SiO2 composite fibers with anatase phases, the TiO2-SiO2 composite fibers were observed to have no voids or cleavages on the surface than TiO2 fibers which have coarse structures created upon crystallization at magnification of x330,000 by transmission electron microscopy. The coarse structure of TiO2 fibers characterized as having cleavages at exposed surface grain boundaries is anticipated to adversely affect the mechanical stability by enhancing crack formation and propagation which will lead to failure of the fiber. In contrast, amorphous SiO2 fills in the spaces that have been created by the development of anatase phase for TiO2-SiO2 composite fibers. Smoother surfaces were observed as well in contrast with TiO2 fibers due to the amorphous SiO2 in the continuous phase of the composite material. Based on the observations, TiO2-SiO2 composite fibers are expected to have better mechanical stability by reducing the possibility of crack formation and blockage of crack propagation. The mean fiber diameter of TiO2-SiO2 composite fibers achieved was 243 nm, which is 8% thicker than the PAN fibers achieved and 54% thinner than fibers of the HEPA filter. Hence, the quality factor of the TiO2-SiO2 composite fibers is predicted to be between those of PAN fibermats and the HEPA filter by filtration theory; however it would be closer to that of PAN fibers. Moreover based on transmission electron microscopy (TEM) observation, the mechanical stability was improved as well by achieving denser structures in the fiber than in pure TiO2 fibers.
General Note: In the series University of Florida Digital Collections.
General Note: Includes vita.
Bibliography: Includes bibliographical references.
Source of Description: Description based on online resource; title from PDF title page.
Source of Description: This bibliographic record is available under the Creative Commons CC0 public domain dedication. The University of Florida Libraries, as creator of this bibliographic record, has waived all rights to it worldwide under copyright law, including all related and neighboring rights, to the extent allowed by law.
Statement of Responsibility: by Hyoung Park.
Thesis: Thesis (Ph.D.)--University of Florida, 2010.
Local: Adviser: Sigmund, Wolfgang M.

Record Information

Source Institution: UFRGP
Rights Management: Applicable rights reserved.
Classification: lcc - LD1780 2010
System ID: UFE0041620:00001


This item has the following downloads:


Full Text

PAGE 1

1 ELECTROSPINNING OF NANOFIBERS FOR FILTRATION MEDIA By HYOUNGJUN PARK A DISSERTATION PRESENTED TO THE GRADUATE SCHOOL OF THE UNIVERSITY OF FLORIDA IN PARTIAL FULFILLMENT OF THE REQUIREMENTS FOR THE DEGREE OF DOCTOR OF PHILOSOPHY UNIVERSITY OF FLORIDA 2010

PAGE 2

2 2010 Hyoungjun Park

PAGE 3

3 To Jin

PAGE 4

4 ACKNOWLEDGMENTS I thank to Dr. Sigmund, my officemates, friends and family for the support.

PAGE 5

5 TABLE OF CONTENTS page ACKNOWLEDGMENTS .................................................................................................. 4 LIST OF TABLES ............................................................................................................ 8 LIST OF FIGURES .......................................................................................................... 9 LIST OF ABBREVIATIONS ........................................................................................... 13 ABSTRACT ................................................................................................................... 16 CHAPTER 1 INTRODUCTION .................................................................................................... 19 Ceramic Fibermats as Filtration Media ................................................................... 19 Objectives and Hypothesis ..................................................................................... 20 Approach to Hypothesis .......................................................................................... 21 2 B ACKGROU ND ...................................................................................................... 24 Filtration .................................................................................................................. 24 Sol gel Chemistry of Ti alkoxide ............................................................................. 26 Hydrolysis and Condensation ........................................................................... 26 Acid and Base Catalysts ................................................................................... 27 Electrospinning ....................................................................................................... 28 Principle ............................................................................................................ 28 Advantage of Electrospinning over Other Fiber Fabrication Methods for Filters ............................................................................................................ 31 3 F ABRICATION OF FIBERMATS ............................................................................ 37 Electrospinning PAN and Filtration Test ................................................................. 37 Polyacrylonitrile ................................................................................................ 37 Polymer Solution Preparation and Electrospinning .......................................... 37 Set up for Filtration Test ................................................................................... 38 TiO2 ......................................................................................................................... 38 Property ............................................................................................................ 38 Sol Preparation ................................................................................................. 39 Electrospinning and Heat Treatment ................................................................ 39 TiO2SiO2 from Aqueous Sol ................................................................................... 40 Reasoning of Adopting Aqueous Sol ................................................................ 40 Sol Preparation ................................................................................................. 41 Electrospinning and Heat Treatment ................................................................ 41

PAGE 6

6 4 EXPERIMENTAL METHODS ................................................................................. 45 Fiber Diameter ........................................................................................................ 45 Scanning Electron Microscopy ......................................................................... 45 Image Analysis ................................................................................................. 45 Filtration Test .......................................................................................................... 46 Sol gel Chemistry of Aqueous Sol .......................................................................... 47 Nuclear Magnetic Resonance .......................................................................... 47 Chemical Shift .................................................................................................. 49 Sample Preparation .......................................................................................... 50 Reproduction of Spectrum ................................................................................ 51 Microstructure ......................................................................................................... 51 X ray Diffraction ................................................................................................ 51 Principle ..................................................................................................... 51 Sample Preparation and Operation ............................................................ 53 Transmission Electron Microscopy ................................................................... 53 Principle ..................................................................................................... 53 Sample Preparation ................................................................................... 54 5 R ESULT AND DISCUSSION .................................................................................. 58 Fiber Diameter ........................................................................................................ 58 Results and Discussion .................................................................................... 58 Fiber Diam eter and Coefficient of Variation ............................................... 58 Conclusion ........................................................................................................ 63 Filtration Test .......................................................................................................... 64 Results and Discussion .................................................................................... 64 Conclusion ........................................................................................................ 65 Chemistry of Aqueous Sol ...................................................................................... 66 Results and Discussion .................................................................................... 66 Hydrolysis Reaction ................................................................................... 66 Condensation Reaction .............................................................................. 69 Conclusion ........................................................................................................ 71 Microstructure of Electrospun Ceramic Fibers ........................................................ 71 Results and Discussion .................................................................................... 71 TiO2 ............................................................................................................ 71 TiO2SiO2 from Aqueous Sol ...................................................................... 74 Conclusion ........................................................................................................ 77 Discussions ............................................................................................................. 77 6 C ONCLUSIONS AND FUTURE WORK ............................................................... 10 8 Conclusions .......................................................................................................... 108 Future Work .......................................................................................................... 110 LIST OF REFERENCE S ............................................................................................. 114

PAGE 7

7 BIOGRAPHICAL SKETCH ...................................................................................... 12 121

PAGE 8

8 LIST OF TABLES Table page 2 1 Partial charge distribution within the Ti oxopolymer shown in Figure 24 18. .......... 33 4 1 Reference of the chemical shift for 13C a nd 1H NMR. ............................................ 55 5 1 Result of analysis on fiber diameter of PAN and TiO2. ............................................ 81 5 2 Result of analysis on fiber diameter of TiO2SiO2. ................................................... 81 5 3 Results of filtration tests 36, 51. PAN (a) ( b) means (b) layers of PAN fibermat fabricated by (a) minute deposition via electrospinning. If (b) is unspecif ied, its a single layer sample. ................................................................................... 81 5 4 Values of electronegativity of atomic species in GPTMS 53. .................................... 81 5 5 Integration of block A, B and C (Figure 5 21, Figure 522, Figure 523 and Figure 524). ....................................................................................................... 82 5 6 Calculated [MeOH] and [H2O] based on values in Table 55 and experimental errors of [MeOH]+[H2O]. ..................................................................................... 82 5 7 Time required for gelation of sol of [0.005N HNO3]:[GPTMS]:[metal alkoxide]=2:1:x before gelation or precipitation. Precipitated sols are indicated as (p). .................................................................................................. 82 5 8 Grain size of different TiO2 phases calculated by Scherrer's equation on spectra shown in Figure 5 32. N/A means that the peaks were too weak to be analyzed. ............................................................................................................ 82

PAGE 9

9 LIST OF FIGURES Figure page 2 1 Filtration mechanisms of particulates in the aerosol by a fiber, (A) interception, (B) impaction, (C) diffusion 10. Reprinted by permission from Hinds, William C. 1999. Aerosol technology (Page 192, Figure 9. 5., Page 193, Figure 9. 6. Page 194, Figure 9. 7.). John Wiley and Sons, Inc., New York. ......................... 34 2 2 A) Hydrolysis, B) alcoxolation, C) oxolation of Ti alkoxide. R means the alkyl group. ................................................................................................................. 35 2 3 Hydrolysis in (A) acidic condition and (B) basic condition ....................................... 35 2 4 Different Ti sites in Ti oxo polymers. ....................................................................... 36 2 5 (A) Schematic set up of electrospinning, (B) forces applied to the liquidic sol or solution at the syringe tip. ................................................................................... 36 3 1 Chemical structure of polyacrylonitrile. .................................................................... 42 3 2 Electrospun PAN fibermat sandwiched by two ACF mats. ...................................... 42 3 3 Crystal structure of anatase phase of TiO2 43, 46. Gray spheres (or brighter spheres in the black/white print) are Ti atoms while red spheres (or darker spheres in the black/white print) are O atoms. Reprinted by permission from Pyrgiotakis, Georgios. 2006. Titania Carbon Nanotube Composites for Enhanced Photoc atalysis (Page 7, Figure 21). University of Florida, FL. .......... 43 3 4 Preparation procedure of the aqueous sol. ............................................................. 44 4 1 Experimental set up for the filtration test 36. Reprinted by permission from Zhang, Qi et al. 2010. Improvement in Nanofiber Filtration by Multiple Thin Layers of Nanofiber Mats (Page 3, Figure 1). Elsevier. ...................................... 55 4 2 (A) Nuclear spinning moment, (B) Discrete nuclear spin states of nuclei with different spinning momentum. ............................................................................ 55 4 3 Diffraction of X ray with of the incident angle by a crystal with the spacing d between diffraction planes. ................................................................................. 56 4 4 Design of X ray spectrometer, top view. ................................................................. 57 5 1 An SEM image of electrospun PAN fibers from PAN 6% w/v DMF. ........................ 83 5 2 Fiber diameter distribution of electrospun PAN fibers. ............................................ 83 5 3 (A) Thermogravimetric analysis of poly vinyl pyrrolidone degradation, and (B) its first derivative to locate the peak. Dried air was flown to the TGA chamber at

PAGE 10

10 0.04 LPM in 20 psi. Temperature was increased from the room temperature to 1000C at 10C/min. ....................................................................................... 84 5 4 SEM images of electrospu n TiO2 fibers from (A) PVP 4% and (B) 16% w/v EtOH after heat treatment. Scale bars are (A) 2 m and (B) 5 m. .............................. 84 5 5 Fiber diameter distribution of electrospun TiO2 fibers from PVP 4% w/v EtOH. ...... 85 5 6 Fiber diameter distribution of electrospun TiO2 fibers from PVP 16% w/v EtOH. .... 85 5 7 (A) df and, (B) Cv of as spun PAN fiber, heat treated TiO2 fibers with different polymer concentration, heat treated SiC fibers, fibers that compose Millipore HEPA and LydAir HEPA. .................................................................................... 86 5 8 SEM images of as spun TiO2SiO2 composite fibers from aqueous sol with PVP concentration of (A) 0.25%, (B) 0.5%, (C) 1%, (D) 2% w/v sol. Scale bar s are 20 m. ................................................................................................................ 87 5 9 Dependence of fiber diameter for electrospun fibers of PVP 1% and 2% w/v sol on heat treatment. Connecting lines are for guidance of the eye only. ............... 88 5 10 Electrospun TiO2SiO2 composite fiber from full aqueous sol. df is 243 nm. ......... 89 5 11 Fiber diameter distribution of electrospun TiO2 SiO2 composite fibers from PVP 1% w/v sol, depending on different heat treatment profile. ......................... 90 5 12 Fiber diameter distribution of electrospun TiO2 SiO2 composite fibers from PVP 2% w/v sol, depending on different heat treatment profile. ......................... 91 5 13 SEM images of (A ) Millipore HEPA (CAT. NO.: AP1504700, Millipore, MA, USA), (B) LydAir HEPA (LydAir High Alpha HEPA air filtration media HEP A ....................... 91 5 14 Quality factor (qF) of filters with standard deviation calculated by data from Table 53. ........................................................................................................... 92 5 15 Chemical structure of GPTMS. Chemically different (A) carbon atoms and, (B) hydrogen atom s are noted by different alphabet subscripts. .............................. 92 5 16 13C NMR spectra of (from bottom to top) GPTMS, mixture of 0.005N HNO3 and GPT MS at molar ratio of 1:1, 2:1, 4:1, 8:1, 16:1 and methanol. The wedges show the peaks from transition states. Peaks are assigned to carbons of GPTMS molecule in Figure 515 (A) 69. .............................................................. 93 5 17 Proposed hydrolysis reaction of GPTMS in aqueous HNO3. Chemically different atoms are noted by different alphabetical subscripts. ........................... 93

PAGE 11

11 5 18 1H NMR spectrum with peaks assigned to hydrogen atoms of GPTMS molecule in Figure 515 (B). ............................................................................... 94 5 19 1H NMR spectra of (from bottom to top) GPTMS, 0.005N HNO3 and GPTMS at molar ratio of 1:1, 2:1, 4:1, 8:1, 0.005N HNO3 and methanol. (A) Peaks of hydrogen atoms of hydroxyl group and, (B) overlapped peaks of water and hydrogen atom of hydroxyl group of methanol. .................................................. 95 5 20 1H NMR spectra of (from bottom to top) GPTMS, 0.005N HNO3 and GPTMS at molar ratio of 1:1, 2:1, 4:1, 8:1, 0.005N HNO3 and methanol. (A) Overlapped regime of hydrogen of water and hydroxyl group of methanol ( Figure 5 19 (B)), (B) overlapped regime of hydrogen that is bonded to carbon of methanol and He ( Figure 515 (B)), (C) regime of Hg ( Figure 515 (B)) as the reference for the integration. ............................................................................................... 95 5 21 1H NMR spectra of 0.005N HNO3 and GPTMS at molar ratio of 1:1 with partial integration of block A, B and C from left to right. ................................................ 96 5 22 1H NMR spectra of 0.005N HNO3 and GPTMS at molar ratio of 2:1 with partial integration of block A, B and C from left to right. ................................................ 96 5 23 1H NMR spectra of 0.005N HNO3 and GPTMS at molar ratio of 4:1 with partial integration of block A, B and C from left to right. ................................................ 97 5 24 1H NMR spectra of 0.005N HNO3 and GPTMS at molar ratio of 8:1 with partial integration of block A, B and C from left to right. ................................................ 97 5 25 Enlarged 1H NMR spectra of Figure 519 (A). ....................................................... 98 5 26 Transition state of GPTMS in the proposed hydrolysis reaction (Figure 517). (A) One methoxy gr oup is hydrolyzed, (B) two methoxy groups are hydrolyzed and, (C) fully hydrolyzed GPTHS. .................................................... 98 5 27 Enlarged 1H NMR spectrum of 0.005N HNO3 and GPTMS at molar ratio of 2:1 in Figure 5 19 (A). ............................................................................................... 99 5 28 13C NMR spectra of (from bottom to top) [0.005N HNO3]:[GPTMS]=2:1, TB, [0.005N HNO3]:[GPTMS]:[TB]=2:1:1 and 1 butanol. The wedges indicate the peaks of 1butanol. Notations in peak assignment for TB and [0.005N HNO3]:[GPTMS]=2:1 are based on Figure 529 and Figure 515 (A), respectively. ...................................................................................................... 100 5 29 Chemical structure of TB. Chemically different carbon atoms are noted by different alphabet subscripts. ............................................................................ 100 5 30 Proposed condensation reaction between {GPTMS+HNO3} and TB. Chemically different atoms are noted by different alphabetical subscripts. ...... 101

PAGE 12

12 5 31 13C NMR spectra of (from bottom to top) [0.005N HNO3]:[GPTMS]=2:1 and [0.005N HNO3]:[GPTMS]:[TB]=2:1:1. The wedge indicates the peak of methanol. .......................................................................................................... 102 5 32 X ray diffraction pattern of electrospun TiO2 fibers heat treated at various temperatures for 3 hours. Peaks notated as A and R indicate anatase and rutile phase, respectively. ................................................................................. 103 5 33 TEM images of electrospun TiO2 fibers heat treated at (A) 500C, (B) 600C, (C) 700C, (D) 800C, and (E) 900C for 3 hours. ............................................ 104 5 34 X ray diffraction pattern of electrospun TiO2SiO2 composite fibers heat treated at various temperatures. Peaks marked with wedges indicate anatase phase of TiO2. ............................................................................................................. 105 5 35 TEM images of electrospun TiO2 SiO2 composite fibers from aqueous sol heat treated at (A) 500C, (B) 800C, and (C) 1100C for 3 hours. .................. 106 5 36 SEM images of 1100 C heat treated electrospun TiO2SiO2 fibers from sol A. [TiO2]:[SiO2]=1:1. Images are from same batch of sample that show different morphologies, i.e. (A) rough surfaces of fiber and (B) spikes from the fiber. .... 107 5 37 XRD result of electrospun TiO2SiO2 composite fibers before and after heat treatment at 1100 C for 3 hours. ...................................................................... 107 6 1 (A) Fractured anatase TiO2 fibermats composed of fibers of df=149 nm, (B) A TiO2SiO2 composite fibermat composed of fibers of df=243 nm from full aqueous s ol after heat treatment. ..................................................................... 113

PAGE 13

13 LIST OF ABBREVIATION S ACF activated carbon fiber B magnetic field Bp full width at half maximum of the peak BuOH 1 butanol Cc gas slip correction factor CDCl3 deuterated chloroform CHCl3 chloroform Cv coefficient of variation, (standard deviation)/mean d spacing between diffraction planes df mean fiber diameter dp diameter of particles DI water deionized water DMF N, Ndimethylformamide DMSO dimethyl sulfoxide DMSO d6 deuterated dimethyl sulfoxide E collection efficiency applied electric field ER collection efficiency for interception EtOH ethyl alcohol, ethanol FD drag force GPTHS 3 glycidoxypropyltrihydroxysilane GPTMS 3 glycidoxypropyltrimethoxysilane HEPA high efficiency particulate air h jet diameter ht terminal jet radius in electrospinning

PAGE 14

14 I electric current K electric conductivity Ku Kuwabara hydrodynamic factor LPM liter per minute MeOH methyl alcohol, meth anol Mw weight average molecular weight N neutron number, number of neutrons n number, number of data, number of measurement s NMR nuclear magnetic resonance P particle penetration PAN polyacrylonitrile ppm par st per million PSL polystyrene latex PVP polyvinylpyrrolidone Q flow rate, infuse rate qF quality factor R radius of curvature SEM scanning electron microscopy SMPS scanning mobility particle sizer SSA specific surface area t size of crystals TB titanium (V) nbutoxide TEM transmission electron m icroscopy TiPP titanium isopropoxide TMS tetramethylsilane

PAGE 15

15 U face velocity of air V air velocity w/v weight per volume, g/ml XRD X ray diffraction Z atomic number, number of protons ZP zirconium n propoxide ZB zirconium n butoxide porosity of the filter pressure drop magnetogyric ratio, surface tension, gamma chemical shift, delta partial charge of A strain, dielectric permittivity, epsilon viscosity of air B average of lower and upper limits of the peak in radian mean free path of air wavelength, lambda air viscosity frequency nu p particle density stress, shielding constant sigma dimensionless wavelength of the instability in electrospinning, chi angular frequency, omega [A] number of moles of A chemical {A+B} mixture of A and B

PAGE 16

16 Abstract of Dissertation Presented to the Graduate School of the University of Florida in Partial Fulfillment of the Requirements for the Degree of Doctor of Philosophy ELECTROSPINNING OF NANOFIBERS FOR FILTRATION MEDIA By Hyou ngjun Park May 2010 Chair: Wolfgang M. Sigmund Major: Materials Science and Engineering Since particulate impurity is regarded as the primary cause of lung diseases, purification of air has been a crucial issue. Filtration is the most conventional method to obtain clean air whereby particulate matter is collected on a fibrous media. The use of fibrous filters is prevalent because of their high filtration efficiency and low pressure drop. Fibrous filters were fabricated via the electrospinning proc ess which can be used to produce continuous submicron diameter sized fibers. Polyacrylonitrile (PAN) nanofibers with a mean fiber diameter of 224 nm were electrospun to form fibermats. Filtration tests on fibermats of PAN were conducted to confirm that filters of thinner fibers result in higher collection efficienc ies and lower pressure drops than that of thicker fibers as predicted by the theoretical filtration mechanism Results showed that electrospun PAN nanofibermats had a superior quality factor of 0.067 compared to 0.0310.001 by the current stateof theart microfiber based high particulate air (HEPA) filtration media. The verified theory implies that nanofibermats of other types of materials could also be considered as promising filtration media since filtration performance i s independent of the material used.

PAGE 17

17 As materials for advanced next generation filtration media, ceramics are favored over polymeric materials due to their robustness against environment al factors such as ultraviolet rays, abrasive particles, and high temperature all of which degrade and damage the fibrous structure. Amidst various ceramic materials, the anatase phase of TiO2 was selected due to its mechanical property and versatility as a photocatalyst and microwave absorbing material. Anatase TiO2 fibers were fabricated by electrospinning followed by heat treatment at 500C for 3 hours. However, early precipitation or gelation of the organic solvent based TiO2 sol posed a practical chall enge in the sample preparation. In order to enhance stability of the precursor sol, a novel aqueous sol with titanium alkoxide was developed. As the result, the time taken for gelation or precipitation was elongated from 4 hours for the organic solvent bas ed sol to 4 months with the novel aqueous sol In seeking the proper chemical composition to attain electrospinnability and maximize the period for storage before gelation, the reaction paths of hydrolysis and condensation for one of the components of the aqueous sol were investigated by nuclear magnetic resonance (NMR) spectroscopy. After hydrolysis and condensation reactions, Si O Ti bonds were validated to be formed by the reaction mechanism. TiO2SiO2 composite fiber s were successfully electrospun from the aqueous sol system by addition of a spinning agent followed by heat treatment In contrast to TiO2 fibers in which anatase phase was observed after heat treatment at 500C, anatase phase was formed at 1100C in TiO2SiO2 composite fibers. The for mation of Ti O Ti bonds was retarded due to the formation of Si O Ti bonds, as evidenced by the NMR results.

PAGE 18

18 In regard to the microstructure of TiO2 fibers and TiO2SiO2 composite fibers with anatase phases, the TiO2SiO2 composite fibers were observed to have no voids or cleavages on the surface than TiO2 fibers which have coarse structures created upon crystallization at magnification of x330,000 by transmission electron microscopy The coarse structure of TiO2 fibers characterized as having cleavages at exposed surface grain boundaries is anticipated to adversely affect the mechanical stability by enhancing crack formation and propagation which will lead to failure of the fiber. In contrast, amorphous SiO2 fills in the spaces that have been created by the development of anatase phase for TiO2SiO2 composite fibers. Smoother surfaces were observed as well in contrast with TiO2 fibers due to the amorphous SiO2 in the continuous phase of the composite material. Based on the observations, TiO2SiO2 composite f ibers are expected to have better mechanical stability by reducing the possibility of crack formation and blockage of crack propagation. The mean fiber diameter of TiO2SiO2 composite fibers achieved was 243 nm, which is 8% thicker than the PAN fibers achieved and 54% thinner than fibers of the HEPA filter. Hence, the quality factor of the TiO2SiO2 composite fibers is predicted to be between those of PAN fibermats and the HEPA filter by filtration theory; however it would be closer to that of PAN fibers. Moreover based on transmission electron microscopy (TEM ) observation, the mechanical stability was improved as well by achieving denser structures in the fiber than in pure TiO2 fibers.

PAGE 19

19 CHAPTER 1 INTRODUCTION Ceramic Fibermats as Filtration Media In 2002, over 27 million people were estimated to be suffer ing from respiratory diseases such as asthma, bronchitis, emphysema in the United States 1 and over 10 million people are killed by lung diseases each year globally 2. Particulate matters has been considered as one of the major pathogens of lung diseases 3, which embraces pollens, soils, industrial pollutants, vehicle exhausts Smaller particles with diameter less than 2.5 trigger cardiovascular issues as well Besides the harm to the human health, accumulation of dust may cause malfunction of elec tronic d evices and mechanical equipment Hence, removal of these particles has been of great concern especially for the indoor humane facilities i.e. house, hospital, school vehicle where the particles are inclined to be concentrated in the air. Filtration is one of the most commonly applied methods to remove the exterior contaminant s for the air purification. Among numerous different types of air filtration systems, fibrous filter s that consist of a mat or mats of fine fibers facing perpendicular to the direction of air flow are the most adopted one because of their large surface area to collect the particles, the low pressure drop, and light weight from the high porosity. In this dissertation, fibrous filtration media were fabricated via the electrospinning process because it is the only processing method that could produce continuous fibers with submicron diameter Polyacrylonitrile nanofibers were electrospun to verify the filtration theory that shows filters made of thinner fibers have a lo wer drag force and a higher collection efficienc y than those made of thicker fibers. For the wider spectrum of

PAGE 20

20 applications as filtration media in the future, however, ceramic materials are preferred to polymeric materials because of; Hardness that could prevent damage by abrasive particles D urability to the chemical reaction that could be induced by harsh environment such as high temperature, high energy ultraviolet rays. These advantages are sometimes critical especially to habitats with limited resourc es e.g. the shelters during warfare, inhabited facilities on the lunar surface also known as the l unar outpost. The r egeneration by the microwave irradiation, the photocatalytic behavior could be regarded as auxiliary benefits from certain species of the ceramics. In light of that, TiO2based materials were selected based on their mechanical property and additional functionality in the microwave absorption and the photocatalytic activity Nevertheless, l ike most of other metal alkoxides 4, the titanium alkoxide is h igh ly reactiv e with water molecules in the air to induce the precipitation or the gelation via hydrolysis and condensation reactions. Inhomogeneity and viscosity change caused b y the precipitation or the gelation would result in discontinuity in the process, the mechanical failure of fibers and fibermats. In seeking stability of sol without the precipitation or the gelation, the n ovel aqueous sol towards the TiO2SiO2 composite m aterial was developed to achieve nanof ibers via the electrospinning method. Objectives and Hypothesis E lectrospinning of polymeric materials has been studied while el ectrospun ceramic nanofibers have been rarely reported until recent years. Since early 2000, various ceramic materials, including TiO2 by Li et al. 5, have been successfully electrospun, e.g. CeO2 6, ZrO2 7, Al2O3 8, BaTiO3 9. H owever, reported achievements are from the sol based on organic sol vents that does not hinder the high reactivity of ceramic precursors

PAGE 21

21 towards the precipitation 4. To the best of the authors knowledge, the water based metal alkoxide sol with extended peri od for gelation compared to the organic solvent based sol hasnt been published yet to date. And electrospun TiO2 fibers havent been fabricated in a macroscale for a practical use most probably due the mechanical weakness of nanofibers induced by their coarse microstructure 5. Therefore, t he objectives of t his research are; To fabricate nanofibermat s and measure their quality factors To understand the chemistry to achieve the stable aqueous sol. To fabricate anatase TiO2based filtration media from the novel aqueous sol In order to accomplish the objectives, it was hypothesized that; N anofiber based filtration media have the higher quality factor from the high collection efficienc y and the low pressure drop than microfiber based filtration media. In preparation of the aqueous sol, the hydrolysis of 3 glycidoxypropyltrimethoxysilane ( GPTMS ) followed by the condensation reaction with titanium (IV) n butoxide ( TB ) occurs at the site of meth oxy group s of GPTMS. The s ol is stabilized in the condition that water molecules were consumed in the hydrolysis reaction with GPTMS. A d enser microstructure of fibers is accomplished by i ncorporation of the amorphous SiO2 phase with the anatase TiO2 phase from the aqueous sol than the anatase TiO2 fiber from the organic solvent based sol. Approach to Hypothesis According to Hinds 10, the good filtration media is one with a high qualit y factor (equation 1 1) achieved by a high collection efficienc y and a low pressure drop. He also showed that the filtration media comprised of thinner fibers have a higher collection efficienc y as well as a low er drag force. The d rag force is the resistance against the airflow induced by the fiber which causes the pressure drop. It is theoretically shown that the drag force can be exponentially decreased when the fiber diameter is reduced

PAGE 22

22 1012. On the other hand, s ince collecting particulates on the fiber is a very complex process that involves multiple filtration mechanisms at once, the relationship between the fiber diameter and the collection efficienc y induc ed by all filtration mechanisms has not been established yet. Theoretically, however, the smaller fiber diameter tends to increase the collection efficiency in regard with each filtration mechanism. Therefore, the smaller fiber diameter induces not only a weaker drag force that leads to a lower pressure drop, but also the increased collection efficienc y which result in the higher quality factor. Polyacrylonitrile nanofibermats and commercially available microfiber based high efficiency particulate air (HEPA ) filters will be evaluated and compared i n respect to the quality factor of filtration. For various functionalities on top of the filtration, the TiO2based filtration media are desired in the future. In order to stabilize the sol from which fibermats are composed via the electrospinning, the novel aqueous sol system was developed. The chemistry of the novel aqueous sol will be studied for the first time to seek the proper chemical composition for the fabrication of nanofibers Based on the sol gel chemist ry of inorganic materials, hydrolysis and condensation reactions are predicted to be triggered at the meth oxy group s of GPTMS. The reaction will be validated by analyzing the resultants and the byproducts by the nuclear magnetic resonance (NMR) method. The shelf lives of the sol s with different molar ratios will be timed by observation on the chemical with the mechanical stirring. The m icrostructures of electrospun TiO2based fibers will be analyzed by the X ray diffraction ( XRD ) and the transmission electron microscopy (TEM) study. By comparing results of TiO2 and TiO2SiO2 composite fibers, it will be clarified how insertion of SiO2 affects TiO2 system in regard with the crystallization behavior and the microstructure.

PAGE 23

23 = ln ( 1 ) ( 1 1 ) 10

PAGE 24

24 CHAPTER 2 BACKGROUND Filtration Filtration is the common method to capture the aerosol particles The two most impor tant parameters in filtration are the collection efficiency, E, and the pressure drop, The c ollection efficiency is defined as the portion of particles that were caught by the media while the pressure drop is the pressure difference between before and after the filtration media. By def inition, E is equal to 1P where P is the particle penetration. In regard with the quality of filters with different pressure drops and collection efficiencies it is often inconvenient to evaluate filters with them because some filters have high E and hig h while others have low E and low Hence, the parameter called as the quality factor, qF, equation 1 1 10. Better filter s n the greater qF value. Collection of particulates is the purpose of using filtration media. There are largely three mechanisms by which the fibers in the filtration media could collect particulates in the aerosol the interception, the inertial impaction a nd the diffusion. In each mechanism, it was assumed that the collection efficiency is for a single fiber and the particles are captured by van der Waals force between the particle and the fiber. The i nterception occurs when the particle follows the streaml ine of the airflow that approaches the fiber within a distance of the radius of the particle ( Figure 21 ( A) ) The collection efficiency of a single fiber by the interception is expressed as equation 2 1 where ER dp is the dia meter of particles, df is the fiber diameter, Ku is the Kuwabara hydrodynamic factor that can be expressed as equation 22 ER is roughly proportional to an inverse of

PAGE 25

25 square of the mean fiber diameter, i.e. df 2 13. W hen the path of the particle is diverted from the airflow because of its inertia, the particle moves across the streamline and hits the fiber as depicted in Figure 21 ( B). Its called as the i nertial impaction and the collection efficiency of a single fiber by the inertial impaction (Et) is expressed in equation 23 p is the particle density, Cc is the gas slip correction factor as described in equation 24 the viscosity of air Et is roughly proportional to an inverse of the mean fiber diameter to the three, i.e. df 3 14. Equation 2 5 is for the condition of dp/df<0.4 while theres no simple equation set up for dp/df>0.4. And the particles that were on the streamline that goes around the fiber can be captured by particles because particles can diffuse out of the streamline in the irregular paths, i.e. the Brownian motion ( Figure 21 ( C ) ) The c ollection efficiency of a single fiber by the diffusion (ED) is expressed in equation 25 where D is the particle diffusion coefficient. ED i s proportional to the inverse two thirds of the mean fiber diameter, df 2/3 15. By all three major fi ltration mechanism s f or aerosol particulates, thinner fibers are favored for the filtration media towards the higher collection efficiency. The pressure drop is caused by r esistance of airflo w, or the drag force, by the filtration media. The dr ag force can be decreased by using thinner fibers for filtration based on the Stokes law as given in equation 26 Here FD is the drag force, is the air viscosity, V is the air velocity and Cc is the gas slip correction factor 10. The gas slip correction factor is a function of the fiber diameter and the which is 66 nm at the standard temperature and pressure, as expressed in equation 24 1112.

PAGE 26

26 In the experimental set a f iltration system can be designed in different ways i.e. number and kinds of fibermat, fibermat support, filter holder, etc. E of each element of the filtration system can be calculated from the penetration, P, by the equation 27 16. The s ubscript s stands for the filter holder and other structures, e.g. tubing, while a, b means different kinds of fibermat s as many as m, n of each element of the filtration system can be calculated by the Darcys law as described in equation 28 17. Subscripts and symbols mean same to those in equation 27 For further reading, Hinds 10 discussed the filtration theory in detail. =( 1 ) 2 2+ ( 2 1 ) 13 = ln 2 3 4 + 24 ( 2 2 ) 10 = 36 2 ( 29. 6 28 0 62) dpdf 2 27 5 dpdf 2 8 ( 2 3 ) 14 = 1 + 2 34 + 1 05 0 39 ( 2 4 ) 1112 = 2 2 3 ( 2 5 ) 15 =3 ( 2 6 ) 10 = ( 2 7 ) 16 = + + + ( 2 8 ) 17 Sol gel Chemistry of Ti A lkoxide Hydrolysis and Condensation TiO2 can be formulated via the hydrolysis and the condensation reactions that involve water molecules and Ti alkoxides as the precursor In the absence of the

PAGE 27

27 catalyst, both of them occur by the nucleophilic substitution, or the SN2 mechanism, where S N, 2 mean the substitution, nucleophilic, and the bimolecular reaction, respectively, as depicted in Figure 22 Because of the strong electronegativity of the alkyl groups ( OR) of the Ti alkoxide, the Ti atom becomes a favorite target to be attack ed by nucleophiles In the hydrolysis reaction ( Figure 2 2 (A)), the nucleophilei.e. the oxygen atom in the water molecule attacks the Ti atom in the alkoxide from the back. As the OR group leaves, it takes one of hydrogen atoms in the water molecule to complete the reaction to form Ti with a hydroxyl group (HO Ti) and an alcohol (R OH). There are two different paths to the condensation, which are the alkoxolation and the oxolation as shown in Figure 2 2 (B) and (C), respectively. The a lkoxolation is also a kind of the oxolation in that it forms an oxo bridge ( O ) but differentiated from the usual oxolation by the byproduct, an alcohol. In condensation reactions, the nucleophile is the oxygen atom in the hydroxyl group ( OH) of the hydrolyzed Ti Upon the attack of the nucleophile on the other Ti atom from the back the alkoxy group in the alkoxol ation and the hydroxyl group in the oxolation leave and take the hydrogen atom that was bonded to the nucleophile. As the resultant, a Ti O Ti bond, and an alcohol molecule in the alkoxolation or a water molecule in the oxolation are formed. Acid and Base Catalysts The m olecular structure of TiO2 can be controlled by the addi tion of the acid or the base catalysts. In the acidic condition, the leaving group of the Ti alkoxide readily leaves upon the attack of the nucleophile on the Ti atom as shown in Figure 23 (A). The hydrolysis is accelerated because the OR group doesnt need the additional proton transferred from the nucleophile to leave. In the bas ic condition, the Ti alkoxide gets

PAGE 28

28 deprotonated by the base catalyst ( Figure 23 (B)). Due to the protonated alkyl group, the electrophilicity of the Ti atom gets weakened, and being less attacked by the nucleophile. Thus, the condensation becomes relatively fast compared to the retarded hydrolysis reaction. T he structure of the resultant TiO2 depends on the favorite sites for the reaction in the oxo polymer in different conditions. T here a re three kinds of sites in the Ti oxo polymer for the hydrolysis and the condensation reactions as shown in Figure 2 4 and partial charges of each case are listed in Table 21 18. In the acidic condition, the leaving group can be easily protonated ( Figure 2 3 (A)) in the order of C, B, A from the most negative to the most positive partial charge on the OR group. Because the hydrolysis and the condensation reactions occur more often at the site C which is the end of the oxo polymer than the site A and B, the result ant would be the linear oligomeric TiO2. The sol containing the linear TiO2 is observed as clear homogeneous liqu id. In the basic condition, the site A with the most positive partial charge on the OR group is the preferred site to attract the OHfor the deprotonation as described in Figure 2 3 (B). The order of preference of the reaction in the basic condition is A, B, C. Because there are more than 2 sites available for reaction in the site A and B, the Ti oxo polymer expands via formation of the three dimensional network to result in precipitations and the inhomogeneous sol. Electrospinning Principle Electrospinning i s an electrohydrodynamic phenomenon of liquids observed in the high electric field to form the fibrous morphology. In contrast to the conventional spinning methods that push out the fiber s from the material source by the mechanical

PAGE 29

29 force, in electrospinning, difference of electric potential across the space between the source and the collector drives the material to carry electric charges from the source to the collector. In the process, the morphology of the material varies from beads to fibers depending on the concentration of the spinning aid which is commonly polymer s dissolved in the spinning solutio n or sol The c omponents of the experimental set up are high voltage supplies, a syringe pump, a syringe with a tip that has electrically conductive surface, the sol or the solution, and a conductive collector as shown in Figure 25 (A). The collector may be connected to the voltage supply of opposite charge or grounded depending on the experimental condition. The number of the syringe pump and the syringe may vary as well depending on the experimental set up. As the syringe pump infuses the sol and the electric field is applied in the system forces that pull out the sol and forces that resist against them compete as depicted in Figure 25 (B). The former group is the normal electric stress, the tangential electric stress, the electric polarization stress, and the gravity while the latter group is the surface tension and the viscosity. When the applied electric field reaches certain strength, the forces that pull the sol out of the syringe tip overcome the forces that resis t against ejection. At the moment, the droplet at the tip forms a conical shape. It is named as Taylor co ne after the researcher who reported it for the first time 19. At the apex of the Taylor cone, a jet is spouted and breaks down to charged droplets that are sprayed out when the sol doesnt contain any spinning aid. It named the phenomenon as electrospraying. The breakdown of the jet into small droplets is due to th e effect of surface charges that tend to create larger surfaces to decrease the charge density over a larger area while the surface tension tends to keep the droplet large to reduce the specific surface area. At the higher

PAGE 30

30 electric field, the effect of the surface charge is more favored over the effect of the surface tension, which leads to smaller droplets with the larger specific surface area. By adding the polymeric spinning, the viscosity of the sol is increased to retard the formation of droplets by the chain entanglement of the polymer. When the effect of the surface tension and the viscosity at the higher concentration overcomes the breakdown by the surface charge, the jet maintained its fibrous morphology. As the electric field is increased, the ef fect of the surface charge gets stronger to result in thinner fibers with the larger specific surface area. After a short flight from the ejection, the behavior of ejected fibers gets unstable which can be described by three different instability modes i.e the Rayleigh instability, the axisymmetric instability and the nonaxisymmetric instability In an occasion when the system is highly electrically charged as the electrospinning condition, behavior of the jet is governed by the nonaxisymmetric instability in which t he jet fl ies in a random whipping motion at a high frequency. The w hipping instability is of crucial importance in the electrospinning process because the fiber gets elongated during the whipping motion that consequently results in the fiber t hinning as well Rutledge and coworkers estimated the fiber diameter from equation 2 6 that was formulated by calculation for the solution with the conductivity of lower than 1 S/cm 2022, where ht is the terminal jet radius, the the dielectric permittivity of the medium which is air, Q is the flow rate, I is the the dimensionless wavelength of the instability that is an approximation of R/h. R is the radius of curvature and h is the jet diameter. Equation 26, however, couldnt be applied to ceramic sols with the higher conductivity Therefore, Sigmund et al. 9 developed equation 27 based on equation 26

PAGE 31

31 for sols with the higher conductivity than S/cm, where E is the applied electric field, K is the electric conductivity, c is a n empirical constant. The electrospinning phenomenon is so dynamic in the nano scale with many experimental parameters that its difficult to observe and study the behavior insitu Hence the relationship between the fiber diameter and experimental parameters has not been clearly elucidated yet For example, the viscosity was not mentioned in the equations as a parameter affecting the fiber diameter while it was reported as one of the key parameters on the fiber diameter control by other researchers 5, 2327. Thus t he equations cannot be applied to all solution systems but gives us a guideline to see the impact on the fiber diameter by th e parameters. The e xperimental parameters that are commonly used to control the fiber diameter are as following; High viscosity of the sol decelerates the elongation of fibers in the whipping instability, which tends to result in thick fibers. Strong elect ric field requires a larger surface area to decrease the charge density, which tends to result in thinner fibers with larger specific surface area. Slow infusion means less amount of material, which tends t o result in thinner fibers. = 22 2 ( 2 ln 3 ) 1 3 ( 2 6 ) 20 = 2 23 22 2 2 ln 3 2 3 ( 2 7 ) 9 Advantage of Electrospinning over Other Fiber Fabrication Methods for Filters Ele ctrospinning has advantages over the traditional fabrication method of glass fibers, such as the melt blow spinning, in the controllability of the fiber diameter, the large variety of applicable materials from polymers to ceramics 9, 28 as well as the theoretically anticipated nanosized fiber diameter by equation 2 6 20 and equation 27 9.

PAGE 32

32 It has also been reported that electrospun fibers of the different polymeric materials had Cvs of lower than 0.30 2930. Tsai et al. 31 reported that, in a strictly controlled experimental condition, melt blown fibers achieved Cv of as low as 0.02. However, df of the corresponding fibers was 13.3 m Melt blown fibers with the submicron range of df along with Cv data havent been published at the best of authors knowledge at the moment

PAGE 33

33 Table 21 Part ial charge distribution within the Ti oxo polymer shown in Figure 24 18. Site (OR) (Ti) A +0.22 +0.76 B +0.04 +0.71 C 0.08 +0.68

PAGE 34

34 Figure 21 Filtration mechanism s of particulates in the aerosol by a fiber, ( A ) interception, ( B ) impact ion, ( C ) diffusion 10. R eprinted by permission from Hinds, William C. 1999. Aerosol technology (Page 1 92, Figure 9. 5., Page 1 93 Figure 9. 6. Page 1 94 Figure 9. 7.) John Wiley and Sons, Inc. New York.

PAGE 35

35 Figure 22 A) Hydrolysis, B) alcoxolation, C) oxolation of Ti alkoxide. R means the alkyl group. Figure 23 Hydrolysis in (A) acidic condition and (B) basic condition

PAGE 36

36 Figure 24 Different Ti sites in Ti oxo polymers. Figure 25 (A) Schematic set up of electrospinning, (B) forces applied to the liquidic sol or solution at the syringe tip.

PAGE 37

37 CHAPTER 3 FABRICATION OF FIBER MATS Electrospinning PAN and Filtration Test Polyacrylonitrile Polyacrylonitrile, or PAN, is a polymer that has acrylonitrile as the repeating unit ( Figure 31 ) As carbon based materials e.g. carbon fiber, graphite, etc get more and more attention due to their exceptional mechanical and electrical properties, so does PAN as a precursor towards carbon materials 3235. Electrospinning of PAN has been reported 34 as well as the convertibility to carbon material Polymer Solution Preparation and Electrospinning PAN (Mw 150, 000, Pfaltz & Bauer, CT) 6% weight per volume (w/v, g/ml) in N,N dimethylformamide (DMF, 99.8%, SigmaAldrich, MO) solution was prepared after the mechanical stirring at 80C for an hour The solution was loaded in a plastic syringe (5 ml, Luer Lok, Becton, Dickinson and Company NJ ) with a gauge15 needle tip (inner diameter 1.499 mm, outer diameter 1.829 mm, stock No. JG15 1.5X, Jensen Global Inc. CA) attached. The syringe was fixed on the syringe pump (PHD2000, Harvard Apparatus, MA) with the needl e tip connected to the high negative voltage power supply (ES30N 5W, Gamma High Voltage, FL). The circular collector of 5 cm in diameter was covered by a sheet of aluminum foil (Fisher Scientific, NH) and connected to the high positive voltage power supply (ES30P5W, Gamma High Voltage, FL). Nanofibers were electrospun when the needle tip and the collector were charged by 0.75 kV/cm of the electric field over 20 cm of the tip to collector distance at 1 m l /hr of the infusion rate. The electrospinning process was continued to form a fibermat for the intended time. After the electrospinning was completed, the fiber mat s were heated at 160C in the box

PAGE 38

38 furnace ( controlled by model No. F4792580, Barnstead International, IA ) for an hour to evaporate the solvent, D MF 36. Set up for Filtration Test Activated carbon fiber mats (ACF, article No. ACC 507 15, American Kynol, Inc.) were used to sandwich each electrospun PAN fibermat because electrospun fibermats were not mechanically strong enough to be practically handled due to its thickness of less than 100 ACF mats were adopted because they were known to have a low collection efficiency, a low pressure drop and have enough mechanical strength to support electrospun fibers for filtration tests. When the multi layered PAN fibermats were tested, ACF mats were placed between each PAN fibermat e.g. ACF/PAN/ACF /PAN/ACF/ as shown in Figure 32 In regard to the aerosol, the HEPA standard measures the collection efficiency against particles of 300 nm in diameter. Therefore, 300 nm mono dispersed polystyrene latex (PSL Cat. No. 5030, geometric standard deviation~1.05, Thermal Scientific, CA) particles were dispersed in the deionized (DI) water at the concentration of 2 g/l followed by addition of the same volume of ethanol (EtOH) in order to ensure the dispersion of PSL particles and reduce the surface tension for better nebulization 36. TiO2 Property TiO2 has 8 different crystalline phases in principle 37 among which the anatase, the rutile and the brookite are the stable phases while other 5 phases are formed only by high pressure processing 3840. Among them, the a natase phase is the most versatile. Anatase TiO2 can be readily heated up by electromagnetic wave in the range of microwave due to its high dielectric permittivity 41 while it has been the most promising

PAGE 39

39 photocatalytic material 4243 because of the large bandgap of 3.2 eV and slow er recombination rate of the hole and the electron compared to the rutile phase. Moreover, the anatase phase boasts the superb mechanical property e.g. 88 GPa of the elastic modulus 44 and 4 GPa of the hardness 45 by the nano indentation method. The elastic modulus was measured on the 80 nm thick anatase film while the hardness was measured on the 600 nm thick anatase film. The c rystal structure of the anatase phase is tetragonal with a=b=3.784 and c=9.515 46 as depicted in Figure 33 Sol P reparation 1 mole of titanium ( IV) n butoxide (TB, 99%, Acros Organics) was added to the mixture of 4 moles of acetic acid (AA, glacial, Fisher Chemical) and 2 moles of ethanol (EtOH, 99.5%, anhydrous, 200 proof Acros Organics) followed by the mechanical stirring for 10 minutes in a capped bottle. This sol was added to 18 moles of EtOH containing polyvinylpyrrolidone (PVP, Mw 1,300,000, Acros Organics) followed by the vigorous stir for an hour in a capped bottle. The ov erall molar ratio of chemicals in the sol was [ TB ] : [ AA ] : [ EtOH ] =1:4: 20 while PVP concentrations were 4% and 16% w/v EtOH each for different sols. Electrospinning and Heat Treatment After the stirring for an hour the sol was fed to the plastic syringe with a gauge15 syringe tip attached immediately to avoid the precipitation due to the high reactivity of TB to humidity in the air The syringe tip was connected to the negative voltage supply while the collector was connected to the positively voltage su pply The overall electrospinning set up was same to that of PAN fibers. The electric field applied was 1 kV/cm over 20cm tipto collector distance at 0.5ml/hr of infuse rate by the syringe pump. Collected fibers were heat treated at various temperatures fr om 500C to 900C

PAGE 40

40 for 3 hours in air by the same box furnace in the fume hood used for PAN fibers The furnace was heated from the room temperature to the intended temperature at 10 C/min and left to be cooled down with samples inside after the 3 hour trea tment. In consequence, the specimen that was heat treated at higher temperature was exposed to heat for a longer period. TiO2SiO2 from Aqueous Sol Reasoning of Adopting Aqueous Sol It has been observed that preparation of ethanol based TiO2 sol was failed by the precipitation of TiO2 due to high reactivity of titanium alkoxides to humidity in air. The failure ratio reaches up to 50% that wastes time and resources. Even after the successful sol preparation, the precipitation has been observed in the s yringe during electrospinning to fabricate fibermats. Hence, novel aqueous sol towards TiO2SiO2 composite material with outstretched the time taken for gelation of 4 months was developed as a consequence of the pursuit of stability in the sol without prec ipitation or gelation. The time for completion of gelation is determined when the sol does not flow in the glass vial located upside down. Moreover, the electrospun TiO2 fibermat was reported to have a coarse microstructure inside the fiber with cleavages on the rough surface 5 by crystallization from the amorphous to the anatase phase The microstructure is speculated to deteriorate the mechanical stability of the fiber because cracks would be generated at the tip of the cleavage by focused stress and propagate through the fiber to result in failure of the fiber. On the other hand, TiO2SiO2 composite material is expected to develop a dense microstructure of fiber s due to the amorphous SiO2 phase that would constitute the continuous phase in the composite. When the composite material is heat treated at 500 C for crystallization of the anatase of TiO2 5, 47 -

PAGE 41

41 49, SiO2 that has the crystallization temperature of 1500 C 50 to the quartz phase keeps its amorphous phase. At the elevated temperature, SiO2 is expected to be transported to fill in the spaces created by polymorphic TiO2 phases which brings the dense structure. Sol Preparation 1 mole of 0.005N HNO3 in the deionized water and 0.5 mole of 3glycidoxypropyltrimethoxysilane (GPTMS, 97%, Acros Organics) were mixed followed by addition of 0.5 mole of TB in 10 minutes and mechanically stirred for 1 hour to form the sol A. 22 moles of 0.005N HNO3 in the deionized water with PVP and 0.5 mole of GPTMS were mixed followed by addition of 0.2 mole of tetraethyl orthosilicate (TEOS, 98%, Acros Organics) in 2 hours and mechanically stirred for 2 hours to form the sol B. The s ol A with PVP 12% w/v sol was prepared for electrospinnin g along with the full aqueous sol. The c oncentration of PVP which was adopted as the spinning aid, was varied as 0.25%, 0.5%, 1% and 2% w/v full aqueous sol to control the electrospinnability of the sol. The s ol A and B were mixed followed by the vigorous stirring for 1 hour to achieve the homogeneous aqueous sol. The overall molar ratio of chemicals in the complete aqueous sol is [ 0.005N HNO3]:[GPTMS]:[TB]:[TEOS]= 23:1:0.5:0.2. The p rocedure of sol preparation is depicted in Figure 34 as well. Electros p inning and Heat Treatment The electrospinning procedure and the heat treatment process for TiO2SiO2 composite fiber is same to that for TiO2 fibers aforementioned. The electric field applied was 1 kV/cm over 20 cm distance betw een the tip and the circular collector covered by the aluminum foil. The infuse rate was 0.6 ml/hr. Electrospun samples were heat treated

PAGE 42

42 at 500, 800 and 1100 C After the heat treatment, fiber s from sol A and the full aqueous sol have the molar composition of TiO2SiO2 at 50% 50% and 29% 71% respectively. Figure 31 Chemical structure of polyacrylonitrile. Figure 32 Electrospun PAN fibermat sandwiched by two ACF mats.

PAGE 43

43 Figure 33 Crystal structure of anatase phase of TiO2 43, 46. Gray spheres (or brighter spheres in the black/white print) are Ti atoms while red spheres (or darker spheres in the black/white print) are O atoms R eprinted by permission from Pyrgiotakis, Georgios 2006 Titania Carbon Nanotube Composites for Enhanced Photocatalysis (Page 7 Figure 2 1). University of Florida, FL.

PAGE 44

44 Figure 34 Preparation procedure of the aqueous sol.

PAGE 45

45 CHAPTER 4 EXPERIMENTAL METHODS Fiber Diameter S canning Electron Microscopy Scanning electron microscopy (SEM, JSM 6400, JEOL) is a characterization method to observe the surface morphology of objects. Magnification ranges over 6 orders of the magnitude from 10 to 500,000 depending on the equipment Magnification of x50,000 was the highest for the current SEM that was used to take discernible images of nanosized fiber samples. When the beam of electrons from the electron gun hits the sample, among other signals, secondary electrons are generated as a result of the interaction between the electrons and the atoms at or near the surface of the sample. Based on the detected signal of the secondary electrons in a raster scan pattern, the image of the sample is constructed. T he surface of the s ample must be electrically conductive and grounded in order to avoid the charge accumulation on the surface. When the electronic charge is accumulated at the surface of the sample, electrons are deviated from the path and the interaction between the electrons and the atom s at the surface gets distorted that result s in a damaged image e .g. blur Since electrospun PAN fibers or TiO2based fibers were not conductive enough to avoid the issue, all the samples observed in this dissertation by SEM were coated with carbon by the coater (Ion Equipment Corporation) at 5 105 mmHg for 1 minute. Image Analysis Fiber diameters from the SEM images were measured by ImageJ, the image analysis software ( http://rsb.info.nih.gov/ij National Institutes of Health, USA). Before

PAGE 46

46 any measurement on each image, the num ber of pixels over the scale bar must be calibrated by the given length of the scale bar in the corresponding image. Then the length across the fiber in perpendicular to the fiber axis was measured manually for each fiber. Fiber diameters of all the fibers that were shown in the image were measured. Each single fiber was measured once, and t he spot in a fiber for measurement was randomly selected. Because the areal density of fiber deposition varies by samples the n umber of images t aken by SEM per each sample varied from 5 to 20 in order to get images of about 100 fibers at least. Total number of measurement s (n ) on one sample ranged from 100 to 250 and measured data were saved and analyzed to calculate df and Cv. Filtration Test The experiments on filtr ation were carried out by colleagues at the Aerosol and Particulate Research Laboratory by Dr. Chang Yu Wu in Department of Environmental Engineering Sciences University of Florida. The c ollection efficiency, E, and the pressure drop, p were measured and the q uality factor qF, for filters were calculated based on measured data of E and p. The experimental set up is shown in Figure 41 The c lean and dry air was fl owed out from the gas cylinder and split into two ways one way to the six jet collison nebulizer (Model CN25, BGI Inc., MA) at the flow rate of 5.5 liter per minute (LPM) to generate the aerosol and the other way at 11 LPM to completely dry the aerosol in the dilution dryer Afterwards the aerosol was lead to pass through filtration media with the Magnehelic differential pressure gauge (Model 2010, Dwyer Instrument s, IN) before and after filtration to measure the pressure drop. The face velocity of the air was controlled at 5.3 cm/s for circular filters with the diameter of 47 mm 5.3 cm/s is the minimum face velocity required to test HEPA filters according to

PAGE 47

47 the m ilitary HEPA standard, MIL F 51079D 51. The effective diameter of the filter was 40.5 mm due to the filter holder structure. The c ollection efficiency was measured by analyzing number and size distribution of particles from the aerosol before and after filtration using scanning mobility particle sizer (SMPS, Model 3936, TSI Inc., MN). Theres a split air path to control the flow distribution and pressure over the whole system by a valve. Sol gel Chemistry of Aqueous Sol In order to understand the reason of the extended time taken for gelation in the aqueous sol system, t he chemistry of the aqueous sol gel system was investigated using the nuclear magnetic resonance ( NMR Mercury 300 BB ) spectroscopy. The goal is to get a detailed picture of the reaction path for the aqueous sol gel system. Nuclear Magnetic Resonance An atom consists of nucleus surround by negatively charged electrons. The atomic nucleus contains neutrons and positively charge d protons. The number s of protons and neutrons in the atomic nucleus are called as the proton number (Z) and the neutron number (N), respectively. Atoms that has even Z and odd N, odd Z and even N, or odd Z and odd N has spinning angular momentum that gives the magnetic moment with the magnetic dipole in the magnetic field, B0 ( Figure 42 (A)). The angular frequency of the precession 0, induced by B0, is linearly proportional to B0 the magnetogyric ratio which is the characteristic constant for different nuclei. The rel ationship between these parameters is described in equation 41 the the precession frequency. Nuclei like 1H or 13C have two discrete nuclear spin states developed by B0. Upon absorption of the electromagnetic wave that has with the frequency 0, spins in the lower energy state flip into the upper energy

PAGE 48

48 state. Those excited spins return to the lower energy state when the electromagnetic wave stops being applied. This phenomenon is called as t he nuclear magnetic resonance. For the proton at a field strength of 1.4 T and room temperature, spins in the lower energy state is more than spins in the upper energy state as many as 0.001% 52 when calculated by equation 42 nupper is the number of spins in the upper state, nlower is the number is the energy difference between the lower and the upper state, k is the Boltzmann constant, T is temperature, h is the Planck constant. Due to the small amount of spins that generate signal s, induced NMR signal s have been t oo weak to be detected or analyzed for a long time. Another obstacle that NMR spectroscopy had was scarcity of 13C in the nature. Carbon is very important in the field of organic or organometallic chemistry. However, 12C cannot create the NMR signal becaus e of its even Z and A, hence, no magnetic moment. Alternative target was 13C ; however, only 1.1% 52 of carbon atoms in the nature exist as 13C isotope. That also induces the problem of the weak signal intensity Recent technological advancement like enhanced sensitivity and the Fourier transformation method makes the NMR spectroscopy a viable and rather preferred option for study on the structure and the reaction in the organic and the organometallic chemistry. Adoption of the Fourier transformation dramatically increased the nu mber of measurements in a given time to amplify the signal. 0= 2 0= 0 ( 4 1 ) = exp E kT = exp h 0kT = exp h B02 kT ( 4 2 ) 52

PAGE 49

49 Chemical Shift When the external magnetic field, B0, is applied, the atom with the electron cloud rotates according to the left hand rule. At the same time, the rotation of the electron cloud that surrounds the atom in the same direction induces magnetic field B in the opposite direction of B0 by the right hand rule Then net applied field strength, B, is weaker than B0 because it is shielded by B as expressed in equation 43 where is the shielding constant. The l ower electron density around a nucleus means less shielding which results in higher B and the precession f requency, i. Then, this nucleus interacts with the electromagnetic wave with the higher frequency. However, if the location of the signal is presented in the precession frequency in the spectrum, it would cause confus ion because the frequency is dependent on B, and B is dependent on B0 (equation 4 3 ) which is not a constant value but an arbitrary experimental parameter. Therefore, the peak location needs to be calibrated by a reference material, e.g. tetramethylsilane (TMS, (CH3)4Si) for 1H and 13C NMR. TM S is regarded as one of the most shielding material for its hydrogen and carbon atoms because of the low electronegativity of Si, 1.7 53. The calibrated unit for peak location is called as the 4 The c hemical shift is dimensionless but often expressed as ppm as consequence of the factor of 106. In the spectrum, the chemical shift on x axis decreases from the left to the right. The l ower chemical shift area is called as the upfield or the high field whil e the higher chemical shift area is called as the downfield or the low field. For example, when a proton is close to an atom with a high electronegativity in 1H NMR, the electron density around the proton would be low because electrons are attracted by the adjacent atom with the high electronegativity. Then this proton does not shield B0 much which leads to large

PAGE 50

50 and consequently positioning of the corresponding peak in the downfield. On the contrary, when the proton is far from an atom with the high electronegativity, its electron density is high so that the induced magnetic field by the electron cloud shields B0 much. Then B becomes small with small small and consequently positioning of the corresponding peak in the upfield. depends on the environment of the corresponding nucleus, nuclei in the same environment here is the chemical equivalence which is satisfied when the nuclei have either of the symmetry operation, i.e. an n fold axis or plane symmetry, or the conformation, i.e. the free rotation. For example, protons of ethylene (C2H4) have both of the rotation and the plane symmetry. Protons of the methyl group in toluene do not have any symmetry due to the phenyl group but they are chemically equivalent because of free rotation by the bond between the carbon in the methyl group and the carbon in the phenyl group. Further discussion of the theoretical aspects of NMR may be obtained from th e books by Gunther 52, Breitmaier et al. 54 and Freeman 55. = 0 = 0( 1 ) = 2 ( 4 3 ) = ( ) ( ) = ( ) ( ) 106 ( 4 4 ) Sample Preparation 725 l of the desired sample was loaded to the disposable grade NMR sample tubes (Cat. No. 8971930000, Kimble Chase Kontes, NJ). For a nalysis on the hydrolysis reaction, mixtures with the molar chemical ratio of [0.005N HNO3]:[GPTMS]=0:1, 1:1, 2:1, 4:1, 8:1, 16:1 were prepared. For analysis on the condensation reaction, a mixture

PAGE 51

51 with the molar chemical ratio of [0.005N HNO3]:[GPTMS]:[TB]=2:1:1 was prepared. Different solvents were selected and added to the tube according to the sample to avoid any chemical reaction or immiscibility with the sample and mixed by vortex mixer. For 13C NMR, deuterated dimethyl sulfoxide (DMSO d6) was used as solvent for most samples except for TB with which was immi scible. For 1H NMR, mixture of DMSO d6 and dimethyl sulfoxide (DMSO) was used as the solvent. For TB, deuterated chloroform (CDCl3) and the mixture of CDCl3 and chloroform were used as solvents for 13C NMR and 1H NMR, respectively. DMSO and CHCl3 were added as the reference for the calibration of NMR peaks. Reproduction of Spectrum Raw data were converted to the spectrum and analyzed by KnowItAll Informatics System (ver. 8.0, BioRad Laboratories, Inc. NJ ). All spectra were calibrated by the location of t he carbon peak in deuterized solvents and the hydrogen peak in s ame chemicals to deuterized solvents for 13C NMR and 1H NMR, respectively, by the reference 56 as shown in Table 41 Microstructure X ray Diffraction Principle X ray diffraction (XRD, Philips APD 3720) is a useful tool to characterize the crystalline phases and the grain size of the cor responding phase of any material that has longrange order in atomic level When the incident beam 1 and 2 hit atoms the incident angle, beams interact with atoms and are scattered. Because theyre scattered in all directions, the scattered beams annul one another i.e. the destructive interferencein most cases.

PAGE 52

52 However scattered beams would be completely in phase and reinforce the amplitude each other i.e. the constructive interference when certain conditions are satisfied ( Figure 43 ) ; Atoms are arranged with the longrange order. The path difference, ABC, is equal to a whole number n of wavelength as expressed in equation 45 which called as the Braggs law, where is the wavelength of the beam and d is the spacing between diffraction planes. d The number of scattered beams in phase that strengthen the amplitude each other depends on the degree of the longrange order which is also called as the crystallinity. In a well developed crystal, the scattering is strong and is called as the diffraction. Depending on the value of 23, that satisfies Braggs law with n=2, 3, By analyzing the locations of peaks at 123,, the crystalline phase can be determined because every crystall ine phase has different crystal structures and spacing s d, hence, intrinsic peak locations. Usually incident X ray beam in usual XRD instruments as shown in Figure 44 The grain size was calculated from XRD result by the Scherrers formula ( equation 4 6 ), where t is the size of crystals, Bp is the full width at half maxi B is the average of lower and upper limits of the peak in radian. Details of the theoretical and experimental aspects of XRD may be obtained from the literatures 5758. = 2 sin ( 4 5 ) = 0 9 cos ( 4 6 )

PAGE 53

53 Sample Preparation and Operation The fibermat sample was ground to powder by a mortar and a pestle. Sample powder w as attached to a glass slide (No. 13 0800, petrographic slides, Wards Natural Science Establishment, Inc., NY) by the double sided tape. It was scanned continuously in the desired range of 2 the scan speed of 0.02 2 The w avelength of X ray was 0.15404 nm from 1 of the copper source. Transmission Electron Microscopy The t ransmission electron microscopy (TEM TEM 200CX JEOL ) was used to observe the structure and the size of the grains in TiO2 f ibers By observing the microstructure, macroscopic properties of material e.g. mechanical property can be predicted. The gr ain size s calculated from XRD results would be compared with grains in the image as well. Principle TEM is a microscopic technique of observing samples from the interaction between electrons and the specimen. After the electrons are accelerated from the gun, theyre focused by the condenser lens on the specimen. The electrons interact with, pass through the sample and are detected by the fluorescent screen or the CCD camera. The magnification of TEM images reaches up to several million times which is higher than SEM images because of the quality of the detected signal. In SEM, the detected signals to buil d images are weak er because they are randomly scattered secondary electrons from the sample surface by the incident electron beam Since detected signals are original incident beams after interaction with the sample in TEM, the signals has higher intensity with less noise compared to SEM, which leads to the higher resolution and magnification. Hence, t he thickness of specimen against the

PAGE 54

54 incident electron beam has to be less than a micron in order to let the electron beams transmit the sample and reach the detector. Sample Preparation The sample were immersed in EtOH ( 99.5%, anhydrous, 200 proof Acros Organics) and sonicated to be shattered into pieces by ultrasonicator (S3000, Misonix Inc., NY) for 10 minutes. Pieces of the sample were collected on the TEM grid (Cat. No. 01814F Ted Pella, Inc., CA) by scooping the sample in EtOH with the grid. The specimen was properly deposited on the grid by evaporation of EtOH for an hour in ambient air.

PAGE 55

55 Table 41 Reference of the chemical shift for 13C and 1H NMR Solvent Atom 13 C NMR DMSO d 6 C of DMSO d 6 39.51 CDCl 3 C of CDCl 3 77.16 1 H NMR DMSO d 6 H of DMSO 2.54 CDCl 3 H of CHCl 3 7.26 Figure 4 1 Experimental set up for the filtration test 36. Reprinted by permission from Zhang Qi et al. 20 10 Improvement in Nanofiber Filtration by Multiple Thin Layers of Nanofiber Mats (Page 3, Figure 1 ). Elsevier Figure 42 (A) Nuclear spinning moment, (B) Discrete nuclear spin states of nuclei with different spinning momentum.

PAGE 56

56 Figure 43 Diffraction of X ray with of the incident angle by a crystal with the spacing d between diffraction planes.

PAGE 57

57 Figure 44 Design of X ray spectrometer, top view.

PAGE 58

58 CHAPTER 5 RESULTS AND DISCUSSI ON Fiber Diameter Results and Discussion Electrospun samples of PAN, TiO2, and TiO2SiO2 composite fibers were observed under SEM and analyzed in regard with the fiber diameter. Images taken by SEM were analyzed by the image analysis software. The number of measurement s (n) for each sa mple was 100 at least. Fiber Diameter and Coefficient of Variation Electrospun PAN fibers for the filtration test have the mean fiber diameter (df) of 224 nm and the coefficient of variation (Cv) of 0.25. They have smooth continuous morphology and stacked on each other as shown in Figure 51 Result of analysis on fiber diameter is shown in Figure 52 T he n umber of measurement (n) was 189. The f iber diameter has the uni modal distribution with the center at df that implies a stable electrospinning of fiber s throughout the process. The measured f iber diameter matches the reported result on the same condition of the solution and electrospinning 59. Electrospun TiO2 fibers from sols with different concentrations of the polymer spinning aid, i.e. PVP were analyzed and compared after the heat treatment at 500 C for the degradation of PVP and the phase transformation of TiO2 to the anatase phase. PVP is reported to be thermally degraded at around 420C 6062. Temperature for the thermal degradation of PVP was experimentally shown in result of thermogravimetric analysis (TGA, Mettler Toledo TGA/SDTA 851e, Figure 53 ( A) ). The dried air was flown to the TGA chamber at 0.04 LPM in 20 psi. The temperature was increased from the room temperature to 1000 C at the ramp rate of 10C /min to mimic the experimental

PAGE 59

59 condition of the heat treatment for TiO2based fiber fabrication. Result of 1st derivative of the TG ( Figure 53 ( B)) locates the thermal degradation temperature of PVP at 433 C which is 13 C higher than the literatures However, since it is still 67 C lower than the lowest holding temperature for the h eat treatment for TiO2based fibers, the error of 13 C is not expected to affect the crystallization behavior of TiO2. The phase transformation behavior upon the heat treatment is discussed in the XRD study in one of the following chapters TiO2 from PVP 4 % w/v EtOH has df of 149 nm and Cv of 0.24 (n=166) while TiO2 from PVP 16% w/v EtOH sol has df of 1110 nm and Cv of 0.20 (n=127) as compared in Figure 5 4 The analysis result on the fiber diameter is presented in Figure 55 and Figure 56 df of electrospun TiO2 fiber is increased remarkably mainly due to the increased viscosity by the higher polymer concentration from 4% to 16% w/v EtOH and the infuse rate from 0.5 ml/hr to 2 ml/hr as predicted by equation 26 and the reported result by other researchers 5, 23, 63. df of the electrospun TiO2 fiber from 4% w/v EtOH is 149 nm which is about 50% larger than that of about 100 nm from same sol reported by Li et al 5. It could be attributed by the higher infuse rate of 0.5 ml/hr compared to 0.2 m l/hr in the literature in order to avoid the blockage at the tip. The effect of infuse rate on fiber diameter is shown in equation 26 Due to the high reactivity of TB to water molecules i.e. humidity in air there grows the white cluster which is presume d to be the composite of the amorphous TiO2 and PVP, from the syringe tip during electrospinning of the TiO2 precursor sol in a minute from the beginning of the processing. When the infuse rate is not high enough to overcome the rates of the hydrolysis and condensation reactions of TB, the tip is completely blocked and the sol is irregularly spit out from the tip because the syringe pump still keeps

PAGE 60

60 push ing the sol out. F or Li et al.5 who needed about 100 measurements for the research on df, one minuteelectrospinning would produce 100 nm diameter fibers as long as 3 km which would provide enough number of measurements in SEM images. Length of the fiber was calculated based on the assumption that the fibers had the monodisperse fiber diameter at 100 nm with the cross section of perfect circle, the density of the anatase phase is 3.89 g/cm3 64 and the conversion ratio of TB to TiO2 is 100% For the fabrication of fiberm at s the sol need ed to be electrospun rather continuously because the deposition time towards filtration media ranged from 5 minutes to several hours. 0.5 ml/hr was the lowest infuse rate that was found in the current experimental set up to achieve continuous electrospinning without blockage. The r esult of measurement s on the fiber diameters for both of PVP 4% and 16% w/v EtOH samples show the b i modal distribution ( Figure 5 5 Figure 5 6 ) which purports instability in electrospinning. However, it didnt critically damage the consistency in the fiber diameter because Cvs of 0.24 and 0.20, respectively, are smaller than that of electrospun PAN fibers with the uni modal fiber diameter distribution as shown in Figure 57 and Figure 5 2 Figure 58 shows as spun fibers for TiO2SiO2 composite fibers from the aqueous sol imaged with SEM. The sols conta in PVP at the concentrations of 1% and 2% w/v sol The sols were electrospinnable and formed fibers with submicron diameters stacked on top of each other in a random mesh. Lowering PVP concentration to 0.25% w/v sol caused th e formation of droplets and beads. This may be interpreted as lacking chain entanglement of the polymer. For the higher concentration of PVP 0.5% w/v sol, fibers appeared that evidenced PVP chains began to be entangled to induce formation

PAGE 61

61 of continuous fibers while the beads were still dominant. By increasing the PVP concentration to 1% and 2% w/v sol, beads bec a me rare and smooth fibers were prevalently observed. The here reported observations in morphologic changes of electrospun material based on polymer concentrations was also reported by other researchers with various polymers 5, 28, 63, 65. With existence of none of or small amount of polymer, electrospun material readily forms spherical droplets to reduce the specific surface area. On the contrary, charges on the surface of droplets tend to be distributed over larger area of the surface and drive the electrospun material towards small droplets or thin fibers. By adding more polymers, higher viscosity and more entangled chains retard the breakdown of electrospun jet to droplets. When the effect of the chain entanglement and the viscosity ove rcomes that of the surface tension, the electrospun jet retains its fibrous morphology and deposited on the collector 66. Upon the heat treatment, the fiber diameter from PVP 2% and 1% w/v sol was decrease d by 62% and 48%, respectively, as shown in Figure 5 9 The shrinkage in fiber diameter is acc ounted for by degradation of PVP and mass transport of TiO2 to fill in pores which were created by degradation of PVP 5, 49, 63, 67. Mass transport of the material to fill in pores created by removal of binder is commonly observed in the conventional metal oxide sintering. With pores inside, the total surface energy of the material would be high because of the large area of the interface between the material and pores. In order to reduce the total surface energy, the surface area should be decreased and it is accomplished by filling in pores by mass transport 6768. When the onedimensional fiber diameter shrinkage ratios are powered by three to be converted

PAGE 62

62 to threedimensional PVP volume, the volume shrinkage ratio of fibers from PVP 2% to PVP 1% w/v sol is 2.08 that matc h es well to the PVP concentration ratio. The fiber diameters were increased by 10% and 11% for PVP 1% and 2% w/v sol samples, respectively, after the heat treatment at 1100C compared to 800C heat treated samples. There is not a significant difference in percentage of fiber diameter thickening between the two because they are composed of chemically same material after PVP is degraded at 4 3 3 C. This is attributed by the phase separation of TiO2 from TiO2SiO2 network The lowest mean fiber diameter of composite fibers achieved with anatase TiO2 is 243 nm ( Figure 510 ) from PVP 1% w/v sol a fter heat treatment of 1100C. The crystallization and the phase development of the composite fiber is discussed in the XRD study in one of the following chapters Elect rospun fibers from PVP 1% w/v sol keep the uni modal distribution in the fiber diameter through the heat treatment while as spun fibers from PVP 2 % w/v sol have the bi or multi modal distribution in the fiber diameter ( Figure 5 11 Figure 512 ) Through the heat treatment process, however, the fiber diameter distribution from PVP 2% w/v sol shaped to the uni modal distribution probably because the 50 nm bin is too large to distinguish the drastically reduced fiber diameter in detail for the original bi modal distribution to be observed. It would be difficult to reduce the fiber diameter further down from 243 nm only by varying the concentration of PVP with dominantly smooth fiber morphology. The difference between the lowest PVP concentration for smooth fibers and the highest concentration that has beads as majority in morphology was 0.5%. Varying the concentration of PVP in this small range hardly affects the property of the whole sol. Other options that would help attain thinner fibers are addition of surface

PAGE 63

63 agents for the lower surface tension, addition of electrolytes f or the higher electric conductivity of the sol or modification of the composition of chemicals i.e. [005N HNO3]:[GPTMS]:[TB]:[TEOS]to reset all parameters to completely different range which have much more subjects to investigate. Summarized data of df an d Cv of electrospun fibers and commercial HEPA filters are presented in Figure 5 7 Electrospun fibers of PAN, TiO2, and TiO2SiO2 composite material were compared with two conventional HEPA filters Millipore HEPA ( CAT. NO.: A P1504700, Millipore, MA, USA) and LydAir HEPA ( LydAir High Alpha HEPA air filtration media HEPA Lydair grade 4450H, product that is equipped by U.S. Army Lyd all Filtration, CT ). Figure 5 13 showed the SEM images of the HEPA filters The analysis results show the e lectrospinning can fabricate more cons istent fibers over from hundred nanometer s to a few micrometer in fiber diameter with different materials than HEPA filters given that the fiber fabrication process for HEPA filters was optimize d to produce thinnest and the most uniform fibers in fiber diameter as possible. The enhanced uniformity in fiber diameter leads better quality control towards the intended fiber diameter processing. Conclusion Fiber diameter s of electrospun fibers from the various sols were analyzed and the summarized result is p resented in Table 51 and Table 5 2 df of PAN fibers for the filtration test was 224 nm The f iber diameter of TiO2 fibers from the organic sol was significantly increased from 149 nm to 1110 nm after the heat treatment by increasing the concentration of PVP from 4% to 16% w/v EtOH TiO2SiO2 composite fibers from the aqueous sol were successfully electrospun and the morphology was changed from sprayed beads to smoothly spun fibers by increasing PVP concentration. The lowest df

PAGE 64

64 achieved for TiO2SiO2 fiber with the anatase phase is 2 43 nm Electrospun fibers demonstrate d excelling consistency in the fiber diameter with Cvs less than 0.3 regardless the material and the fiber diameter, compared to microfibers from the commercial HEPA filters with Cvs over 0.7. Filtration Test Results and Discussion The penetration, P and the pressure drop, p of different layered PAN fibermats with ACF mats were measured and substituted into equation 26 and equation 28 respectively, to calculate P and p of PAN fibermats only When Darcy s law was applied, PAN fibermats and ACF mats we re assumed to act independently as filtration media. The e xponential regression for P and the linear regression for p are shown below; Pt otal=1.02 0.931n, or lnPt otal= 0.0719n+0.0195, R2=0.935 pt otal=10.6n+4.17 (Pa), R2=0.999 P and p of the single ACF mat are found to be 0.931 and 10.6Pa, respectively. Based on this result, net effect of different number of ACF mats on P and p could be calculated. For example, when two ACF mats sandwiched a PAN fibermat, effect of P and p are 0.8 71 and 21.2Pa, respectively. The result of the filtration test were recalculated for calibration accordingly and shown in Table 5 3 and Figure 514. qFs of all the PAN fibermats tested are higher than the military standard while qFs of two samples composed of PAN 5 (PAN fibermats that were formed by 5 minute deposition of electrospun PAN fibers) fibermats exceeds that of LydAir HEPA filter that is currently used U.S. Army. PAN 5 3 (3 layered PAN fibermats that were formed by 5minute deposition of electrospun PAN fibers) sample shows superior qF of 0.067 to qF of

PAGE 65

65 0.0230.002 of PAN 15 i n spite of the same amount of PAN fiber deposition. It is due to the inconsistency of deposition over the sample. In the electrospinning process, the fibers are supposed to be randomly deposited which practically does not happen. Therefore, in a single fibermat, there are regimes with lower areal fiber densit ies and regime with higher areal fiber densit ies When the single fibermat is used as the filtration media, the particulates in the air pass through the low fiber density regime to result in low E. At the same time, p is not decreased accordingly because the air stream still hits the high fiber density regime and caused turbulence before it is finally directed to the low fiber density regime. By forming the multi layered fibermat system, this issue could be resolved by overlapping multiple layers of fibermats for the compensation in the fiber density or thickness of the fibermat The effect of multi layred fibermats is clearly shown when filtration data of PAN 5x3 and PAN 15 in Table 53 are compared. PAN 5x3 has low er P, i.e. high er E, and low er p which leads to higher qF than PAN 15. qF of PAN 15 is lower than not only PAN 5 that has same df, but also LydAir HEPA of which df value is higher than twice. It could be accounted for by the further inconsistency as the deposition time gets longer. Conclusion It was hypothesized that the nanofiber based filtration media has higher qF than the microfiber based filtration media as predicted by equation 2 6 equation 24 and Hinds 10. The hypothesis was confirmed by comparing qF values of the uniformly coated PAN nanofibermat PAN 5x3, with HEPA filters. In addition to that, it was found that the multi layered fibermats achieved higher qF than a single layer fibermat with the same total amount of fiber deposition by enhance the uniformity of the fiber deposition over the media. Since the fiber diameter is the factor that affects filtration parameters

PAGE 66

66 according to equation 26 and Hinds 10 but type of material is not the drag force that induces the pressure drop was increased by 13% when 243 nm diameter TiO2SiO2 composite fibers instead of 224 nm diameter PAN fibers based on equation 26 and equation 24 The increment by TiO2SiO2 composite fibers could be regarded to be slight compared to that by LydAir HEPA. FD of LydAir HEPA with df of 527 nm is calculated to be 221% higher than the PAN fibermat. Chemistry of Aqueous Sol Results and Discussion Hydrolysis Reaction The m ixture of GPTMS and HNO3 has been investigated by 13C NMR. Peaks of GPTMS were assigned based on the literature 69. When HNO3 was added, new peaks were appeared near the peaks of carbons a b c, and d ( Figure 515 ). These carbons were affected much compared to carbon e f and g by the hydrolysis because they are either the carbon of the methoxy group that was hydrolyzed or the carbons that are adjacent to the methoxy group. Therefore, those new peaks could be assigned to the carbons a b c, d of partially hydrolyzed GPTMS in the transition state and fully hydrolyzed GPTMS which is GPTHS. The wedges in Figu re 516 show the carbons in the transition state. As more HNO3 was added from the bottom to the top, GPTMS became more hydrolyzed. At the ratio of [0.005N HNO3]:[GPTMS ] =4:1 or higher, peaks from the transition states were not clearly observed which implied full hydrolysis. Shifted peaks of carbon b c and d could be regarded as corresponding carbons of GPTHS. The peak of carbon a in Figure 5 15 (A) shifted to the downfield to form a peak for the methanol which was assumed to be the byproduct of the reaction in Figure 5 17.

PAGE 67

67 For the quantitative analysis on chemicals and, especially, water molecules, 1H NMR was used on the same mixture in order to confirm the hypothesized reaction in Figure 5 17. 1H NMR peak assignment of GPTMS was preceded further experiments because it hasnt been reported yet for the moment. Peaks were assigned to each chemically different hydrogen atom by comparing its relative integration with the number of chemically equivalent hydrogen atoms and the chemical shift of each peak based on the electronegativity of adjacent atoms to hydrogen atoms ( Table 54 Figure 515 ( B), Figure 518 ). The p eak of hydrogen a was clearly recognized by the relative integration of 9. There are three peaks with integration of around 2 at 0.6 ppm 1.6 ppm and 3.4 ppm These were assigned to hydrogen b c and d respectively, in the order of the distance from oxygen that has the relatively high electronegativity in the molecule. Hydrogen d which shields the applied magnetic field the least and has lowest among them ha d its corresponding peak located in the most downfield while the peak of hydrogen b was located in the upfield due to the opposite reason to hydrogen d The inte grities of the rest of the peaks near which new peaks were not developed didnt change either. A ccordingly, they were not involved in any of the hydrolysis and condensation reaction s. The peaks at 3.65 ppm 3.25 ppm were assigned to two hydrogen e atoms and peaks at 3.1 ppm were assigned to hydrogen f based on the same rational process applied to the peak assignment of hydrogen b c, d The peak at 2.54 ppm is for the hydrogen in DMSO as the reference but it is also overlapped with hydrogen either of g or h while one of them was on the peak at 2.7 ppm Although hydrogen g and h are bonded to the same carbon atom carbon g in Figure 5 15 (A), they are not chemically equivalent because there is no plane symmetry between them.

PAGE 68

68 The epoxy ring of GPTMS restrains the free rotation of bonds of carbon g to carbon f and oxygen atom. In the viewpoint of the plane of epoxy ring, one of hydrogen g or h is at the same side to hydrogen f while the other one is at the same side with the rest of the GPTMS molecule. When HNO3 was added to GPTMS, new peaks appeared at around 6 ppm and 4.3 ppm ( Figure 519 ) and they were analyzed assuming that those chemicals react as hypothesized in Figure 5 17 In regard to peaks at around 4.3 ppm the regime of new peaks was close to hydrogens of water and, especially, splitting of peaks from the sols of [0.005N HNO3]:[GPTMS]=1:1, 2:1 look same to that from the hydrogen of the hydroxyl group in methanol. Also, the shape and the integration of the peak at 3.15 ppm that represents hydrogen bonded to carbon atom in methanol was changed as more 0.005N HNO3 is added. For t hose reasons, it is assumed that; 4 4.5 ppm : the overlapped regime of hydrogen atoms of water and the hydroxyl group in methanol. Figure 5 20 (A). Integration of A= 2x[ H2O]+1x[MeOH] 3.15 ppm : the overlapped regime of one of He in GPTMS and hydrogen bonded to carbon atom in methanol. Figure 5 20 (B). Integration of B =3x[MeOH]+1x[GPTMS] 2.65 ppm : the reference peak of Hg in GPTMS for the integration analysis of overlapped peaks listed above. Figure 5 20 (C). Integration of C =1x [GPTMS] The i ntegrations of regime A, B and C in Figure 5 20 from each sol was measured as shown in Figure 5 21 Figure 5 22 Figu re 5 23 and Figure 5 24 followed by the calculation to get sums of [MeOH] and [H2O] for each sol wh ich is supposed to be same to added [0.005N HNO3]. Based on the integration of each regime presented in Table 55 sums of [MeOH] and [H2O] were calculated for each sol and compared with theoretical values ( Table 5 6 ) The error varies from 1% to 10% which are in the

PAGE 69

69 acceptable range for these overlapped peaks 7071. Therefore, the assumption of peak overlap indicated in Figure 5 20 was confirmed. In regar d to the peaks at around 6 ppm the integrations of peaks from sol of [0.005N HNO3]:[H2O]=1:1 and 2:1 were same in the acceptable error range 7071 to the integrations of the hydrogen peak of the hydroxyl group in methanol ( Figure 525 ) Based on results of Table 5 6 it could be assumed that the peaks at 4.1 ppm are not related to water, but only to methanol. When GPTMS and water molecules react as hypothesized in Figure 5 17 same moles of methanol and the hydroxyl group bonded to the Si atom in GPTMS are form ed Thus, these peaks could be regarded as hydrogen peaks from the silanol (SiOH) group in different env ironments. Three dominant peaks from left to right were assigned to GPTHS and two transition states from GPTMS to GPTHS Chemical structures of each species are depicted as C, B and A in Figure 526 respectively. The peaks with low intensities between these dominant peaks could be assigned to hydrogen atoms of the silanol groups in the different condensed environment based on the electronegativity of adjacent atoms ( Figure 527 ). Comparison of the integration was not useful for these peaks because n either amount of partially and fully hydrolyzed GPTMS nor amount of each hydrolyzed GPTMS involved in the condensation reaction was not discernable by characterization methods used. Although these peaks existed, they were hardly distinguished at the higher molar ratio of [0.005N HNO3]:[GPTMS] due to their relatively low intensity compared to the dominant peak at 3.2 ppm in Figure 5 20 (B). Condensation Reaction Condensation reaction with the addition of TB to sol of [ 0.005N HNO3] : [ GPTMS ] =2:1 were studied by 13C NMR. When TB was added to mixture of

PAGE 70

70 [ 0.005N HNO3] : [ GPTMS ]=4 :1 TiO2 particles were precipitated immediately by the hydrolysis and condensation reactions between TB and water molecules that w ere left over after hydrolysis reaction. When sol of [ 0.005N HN O3] : [ GPTMS ]=1 :1 was used to achieve sol A, the attempt to electrospin fibers from sol A was not successful because of its high viscosity from lack of water. TB was added at the same molar ratio to form sol A with m olar ratio of [0.005N HNO3] : [ GPTMS ] : [ TB ] =2 :1:1 According to the observation on the time taken for gelation of TB in varying molar ratio ( Table 57 ), sol A attains the longest time for gelation when [TB] is twice of [GPTMS]. However, probably due to the excessive amount of TB in sol A, full aqueous sol system became unstable and TiO2 par ticles precipitated when sol B was added. Stable full aqueous sol was success accomplished when sol B was added to sol A with m olar ratio of [0.005N HNO3] : [ GPTMS ] : [ TB ] =2:1:1, which has the second longest time for gelation of around 4 weeks before gelation at which the sol didnt flow down when the capped glass vial was put upside down. In Figure 528 sol of the third spectrum from the bottom is from th e mixture of two below it. Chemical structure of TB along with subscripts for chemically different carbon atoms is presented in Figure 529. Wedges in Figure 528 indicate the formation of 1bu tanol which is the byproduct of hypothesized reac tion in Figure 530 The enlarged spectrum of [ 0.005N HNO3] : [ GPTMS ] : [ TB ] =2:1:1 sol in Figure 5 -31 confirmed the existence of carbons from GPTMS and methanol indicated by a wedge as well Integration of peak does not necessarily match the number of moles of corresponding atom in 13C NMR 52. 1H NMR was not used for this mixture of sol because there were too many peaks in t he spectrum that overlapped each other to identify

PAGE 71

71 Shift of peaks to low field was observed when HNO3 was added to the sol for both of 13C and 1H NMR. It is accounted for by the formation of hydrogen bonding between the corresponding atom and the water m olecule because hydrogen bond is regarded as being electron attractive 52, 5455. This phenomenon is clearly found for the peaks of carbon e and g in Figure 5 16 and regime A in Figure 5 19 Conclusion The hypothesized chemical reactions described in Figure 5 17 and Figure 530 were confirmed by analysis on NMR results. TB is not readily hydrolyzed even with the significant amount of water because water was consumed to hydrolyze GPTMS before TB was added. At a longer time scale, TB reacts with hydrolyzed GPTMS to form the Si O Ti bond via the condensation reaction, which explains the higher temperature for format ion of the anatase phase in TiO2SiO2 composite material 72 than that of pure TiO2 in the result of following XRD study and formation of 1butanol as a byproduct. The sol gels at a much slower rate than the gelation of the organic solvent based sols towards TiO2 5 probably due to the s teric hindrance of the 3 glycidoxypropyl group of hydrolyzed GPTMS 73. Microstructure of Electrospun Ceramic Fibers Results and Discussion TiO2 The behavior of the phase transformation of electrospun TiO2 fibers was studied by XRD and TEM. The results are presented in Figure 5 32 and Figure 533 In the XRD spectra, p eaks noted as A and R are assigned to the anatase and the rutil e phase, respectively. TiO2 samples that were heat treated at 500C and 600 C were found out to be mainly the anatase phase without any rutile peaks observed. The

PAGE 72

72 r utile phase appeared in TiO2 heat treated at 700C along with the anatase and both phase coexisted in TiO2 heat treated at 800 C After heat treatment at 900 C the rutile became the dominant phase while peaks of the anatase were barely observed. The p hase transformation of TiO2 by the heat treatment temperature is as reported by researchers 5, 4849, 67, 74. The grain size for both of the anatase and the rutile phase was calculated by equation 46 based on peaks in Figure 5 32 and the result is shown in Table 58 Two peaks with the highest intensity were selected to be analyzed for both phases, which are peaks at 25 and 48 of 2 for the anatase and peaks at 27 and 54 of 2 the rutile phase. 2 5 and 48 of 2 are (101) and (200) planes of the anatase while 27 and 54 of 2 Grains kept growing as the heat treatment temperature increases as long as peaks of the phase could be observed as r eported by Avci et al 47. It is thermodynamically preferred because, with the grain size increased, the total surface energy decreases due to the smaller specific area of the grain boundary. At the higher temperature, it could be achieved easier with the higher mobility of atoms that results in the grain growth. TEM images in Figure 533 show the morphology of grains of electrospun TiO2 fibers after the heat treatment for 3 hours at 500, 600, 700, 800, 900 C from (A) to (E). W hen compared to XRD results, grains of the 500, 600 C heat treated samples shown in Figure 5 33 (A) and (B), respectively, are the anatase while the rutile phase was developed in the 7 00C heat treated sample. The grain size of the rutile phase increases with the higher heat treatment temperature and the morphology changes as well from the roughly spherical shape to the columnar shape For the rutile fibers in Figure 533 (E), the fibers are formed by the link of single grains. Thickness fringes are

PAGE 73

73 also observed in some crystallites due to the gr adual thinning of the grain in the direction of the incidental beam towards the grain boundary 75. The grain size of the anatase observed as around 20 to 30 nm is comparable to data presented in Table 5 8 while the grain sizes of the rutile phase dont match well. The calculated grain size for the rutil e phase by the Scherrers equation (equation 46 ) is 43 nm for the 900 C heat treated sample instead of diameters around 100 nm and lengths of up to around 250 nm that t he columnar shaped grains have in the Figure 533 (E). In spite that the number of observation is small to be statistically meaningful, the gap between the calculation and observation is 133% at least which could be regarded as a large error The error could be ascribed to the limitation of the Scherrers equation that is dependable in calculation of smaller spherical crystallites than 100 nm in diameter 5758, 76. In regard with the fiber morphology, cleavages between the grains of the anatase phase and coarse structure formed by grains was observed in the 500, 600 C heat treated samples. When fibers with these morphologies are loaded with external stress, it is focused on the apex of the cleavage to initiate cracks while the coarse structure would ease the propagation of the crack. The fiber structure became significantly denser with the development of the large r grains of the rutile phases which allowed less space between grains. The f iber diameter should be reduced as the heat treatment temperature increases because of formation of the denser structure in the fiber as well as the higher density of the rutile th an that of the anatase phase, 4.13 and 3.79, respectively 77. Nonetheless, fiber diameters were not observed to be decreased drastically probably because the number of observation was not many enough to be statistically meaningful to exhibit the differences.

PAGE 74

74 TiO2SiO2 from Aqueous Sol In the XRD diagram in Figure 534, 500C heat treated sample showed a broad peak over early 20 of 2 while the 800C heat treated sample and the 1100C heat treated sample showed weak peaks and well defined peaks, respectively, with the slight increase in the intensity of the broad peak compared to the baseline. If TiO2 was being separated as the amorphous phase in SiO2 matrix at the 500C heat treated sample, the intensity of the broad peak over 25 of 2 should be decreased significantly as TiO2 got crystallized because of ; O verlapped peak position of amorphous TiO2 broadly at around 25 and 30 of 2 78 with amorphous SiO2 7981. T he indicative amount of TiO2 to SiO2 in molar ratio of 5:12 according to the recipe of the sol described in sol preparation of TiO2SiO2 from aqueous sol The precursor of Ti was TB while those of Si were GPTMS and TEOS. Therefore, based on the characteristics of XRD peaks with varying temperatures of heat treatment s, it is hypothesized that Ti atoms were dispersed in the SiO2dominant matrix to form Si O Ti bonds for the 500C heat treated sample and separated to be crystallized to the anatase phase at the higher temperature, 1100C as indicated by wedges. The i ntensity increment of the amorphous SiO2 peak from 800C to 1100 C heat treated samples is accredited to the newly formed SiO2 phase by Si atoms that were incorporated in Si O T i structure in the 500C heat treated sample. It is also found that the crystallization temperature for the anatase TiO2 is between 800 and 1100C which is higher than 500C, usual temperature for the formation of the anatase phase from TiO2 precursor sol without SiO2 precursor 5, 47 49. The i ncrease of the crystallization temperature of TiO2 has been observed in TiO2SiO2 system before 79, 82 and is reported that the Si O Ti bond hinders the formation of the TiO2 phase while the Ti O Ti bonds

PAGE 75

75 and the Si O Si bonds become more stable at the higher temperature that leads the phase separation of TiO2 and SiO2 72, 83, which explains the XRD result as well. The f ormation of the Si O Ti bond was as hypothesized in Figure 530 and evaluated in the NMR study in the previous chapter The grain size of the anatase TiO2 phase in the 1100C heat treated sample was calculated to be 10 nm by equation 46 Two peaks with strongest intensity at 25 and 48 of the grain size by the Scherrers equation. Peaks at 25 and 48 represent (101) an d (200) planes of the anatase phase, respectively. TEM images of electrospun TiO2SiO2 compos ite fibers after the heat treatment at 500, 800, 1100 C for 3 hours are presented in Figure 5 35 (A, B, C), respectively It was observed that the phase development of TiO2 in the fiber through the different heat treatment temperature was in an agreement with the result of XRD an alysis. When the fiber was heat treated at 500 C, it was homogeneously amorphous with smooth surface, no grains or crystalline plane observed ( Figure 535 (A)). After the heat treatment at 800 C, dots estimated to be about 2~3 nm in diameter were appeared in Figure 535 (B). These are regarded as grains of TiO2 by XRD result showing small hump ( Figure 534 ). When the fractured end of the fiber was observed, it was found to be the cross section of the fiber because of the contrast gradation. The grains of TiO2 were observed over the whole regime of graded contrast which means anatase crystallites were developed not only on the surface but well dispersed through the fiber. And TiO2 phases observed as dark spheres which are the anatase phase according to the XRD result, are clearly developed after the heat treatment at 1100 C. In comparison with the TEM image of the anatase fibers ( Figure 5-

PAGE 76

76 33 (A, B)), TiO2SiO2 composite fibers with the anatase TiO2 have a de nse structure without any pores regardless of the heat treatment temperature, as reported 83, probably because the amorphous SiO2 with the high mobility at the elevated temperature was transported to fill up the space created by the grain growth. The grain size is estimated to be around 10 nm It also corresponds well to 10 nm which is calculated by the Scherrers equation from the XRD result. Electrospun TiO2S iO2 fibers from the half sol A have been characterized by SEM and XRD. The half sol was attempted to electrospin the composite fiber with higher TiO2 loading to SiO2 at 50% 50% compared to 29% 71% of the full aqueous sol. Fibers made of the composite mater ial with the anatase TiO2 were successfully electrospun as observed in Figure 536 and Figure 537. Nonetheless the high viscosity due to the higher TB ratio retards the thinning of fiber s in the electrospinning process that leads to the formation of microfibers. However, when c ompared to TiO2 29% SiO2 71% fibers ( Figure 510 Figure 535 (C)), the crystallization growth of the anatase on the surface of TiO2 50% SiO2 50% fibers along with the unique needlelike crystal growth were clearly distinguishable at lower magnification of SEM ( Figure 536 ). It is attributed to the relative abundance of TiO2 that contributes the grain growth. The n eedlelike crystal growth of the anatase TiO2 from the fiber was observed in about 5% of images taken on randomly selected area over the sample by SEM. The grain size cal culated by the Scherrers equation on peaks at 25 and 48 in Figure 537 was 18 nm. Although the measured crystallite size was not confirmed by highresolution TEM images, bright grains observed on the surface of the fiber in Figure 536 (A) have diameter of roughly 25 nm. Provided that it was estimated based on a low resolution SEM image, the 7 nm -

PAGE 77

77 difference is ignorable. Based on the observation, the reason that needlelike large crystallites grown from f ibers didnt affect the XRD result is attributed to their limited amount with that morphology Conclusion Behavior of phase transformation and crystallization was investigated on TiO2 and TiO2SiO2 composite fibers. In regard with TiO2 from the EtOH based sol, the p ure anatase phase was developed via the heat treatment at 500 C for 3 hours. As the heat treatment temperature increases, the rutile phase was developed and prevailed via the heat treatment at 900 C for 3 hours. In regard with TiO2SiO2 co mposite fibers, the anatase was formed after the heat treatment for 3 hours at the temperature higher than 8 00 C which is at least 300 C higher than crystallization temperature for the anatase from TiO2 of the EtOH based sol When Ti atoms were incorporate d in the SiO2 matrix, the Si O Ti bond is preferably formed so that the formation of the Ti O Ti bond gets hindered at 500 C. At higher temperaturei.e. 1100 C TiO2 phase is separated to form the anatase phase from the SiO2 phase because the Ti O Ti bonds and the Si O Si bonds become more stable. TiO2SiO2 fibers established the denser microstructure than pure TiO2 fibers because of the amorphous SiO2 phase to form the continuous phase in the composite material. Discussions In the electrospinning process, the fiber diameter could be controlled by varying concentration of polymers that promotes spinning of continuous fiber s by the chain entanglement. More importantly, Cvs of all electrospun fibers of PAN, TiO2 and TiO2SiO2 composite material were suppressed below 0.3 which is less than half of fibers that compose the commercial HEPA filters ( Figure 5 7 (B)). Cv of fibers is a substantial

PAGE 78

78 factor in the quality control on the fabrication because fibers with the larger fiber distribution, i.e. larger Cv, have the smaller specific surface area than those with smaller Cv. Material system with the smaller specific surface area discourages all the surface driven functionality including the filtration, the photocatalytic activity, the microwave absorption, etc. Furthermore when the fibers have larg er Cv, the pressure drop will increase because the drag force is exponentially proportional to the fiber diameter as shown in equation 26 equation 24 Therefore, in order to fabricate filtration media with a high quality factor accomplishing low Cv in the fiber diameter of filtration media is of paramount importance. E lectrospun PAN nanofibermats with df of 224 nm showed promising results in the filtration test s as enhanced filtration media. PAN fibermats achieved the quality factor over twice higher th an LydAir HEPA filter s that are currently used by U S Army, and over three times higher than the criteria by the Department of Defense 51. If the electrospun TiO2SiO2 composite fiber with the anatase phase was used instead of the PAN fiber, 13% increase in the drag force is anticipated based on the calculation with equation 26 and equation 24 Compared to 221% increase in the drag force that is expected with LydAir HEPA with df of 527 nm it could be considered as a slight increase. Since the relationship between the pressure drop and the drag force heavily depends on empirical conditions, it ca nnot be concluded that the pressure drop will increase 13% and 221% for TiO2SiO2 composite fibers and HEPA filter fibers. However, because it is obvious that a weak drag force causes a low pressure drop, TiO2SiO2 composite fibers will have a lower pressu re drop than HEPA filter fibers.

PAGE 79

79 Also for the additional functionality as a microwave absorbing material or a photocatalytic material, electrospun TiO2, especially anatasebased nanofibers have advantages. As the fiber diameter decreases, D istance of diffusion for excited electrons by photons becomes shorter to reduce the probability of recombination with holes. T he specific surface area increases to induce more contact with reactant chemical s or electromagnetic wave. Towards the TiO2based materials v ia the sol gel method, d ue to the high reactivity to water molecules it has been presumed to be very difficult to handle transition metal alkoxides without strict control of moisture in order to achieve homogeneous sols or gels 4. As the result, enormous amount of organic chemicals has been disposed after processing. For the first time at the best knowledge, control on the reaction rate of metal alkoxides and water molecules was demonstrated by adopting GPTMS as the control agent. With dominant amount of water, i.e. 93% in molar ratio, TB and TEOS were stabilized to keep the transparent homogeneous sol. Adopting the aqueous sol to produce metal oxide materials is expected to decr ease the harm to environment as well as the cost in disposal of toxic organic chemicals. In order to acquire the homogeneous aqueous sol with other transition metal alkoxides, the mechanism between chemicals should be discerned. In this dissertation, the chemical reaction mechanism of half sol A was explained by analyzing results from 1H and 13C NMR. Although there are more to explore to fully understand the aqueous sol system, the result presented here would be the least clue to navigate the direction in the design of experiments. In terms of the microstructure of anatase TiO2SiO2 composite fiber, smooth fibers with the dense structure were achieved in TiO2SiO2 fiber compared to anatase TiO2

PAGE 80

80 fiber that has cleavages at exposed grain boundaries on the surface. These cleavages are where the cracks ar e formed by concentrated stress to result in failure of the fiber. When the external stress is applied on the smoother surface of the fiber, the stress is more evenly distributed and the threshold val ue of the stress to initiate a crack becomes higher which result s in mechanically stable fiber s. The c ontinuous phase of amorphous SiO2 could also block the propagation of the crack from inside of the anatase grains. Furthermore, in the SEM observation, Ti O2SiO2 composite fibers were bonded to each other to form the network due to the low vapor pressure of the solvent, i.e. water in the aqueous sol. The fused network would enhance the mechanical stability of fibers as well because the external force would be distributed over the network while the force is concentrated onto the limited number of fibers to result in failure of those fibers. Consequently, better mechanical stability is anticipated for TiO2SiO2 composite fibers than TiO2 fibers. Among TiO2SiO2 composite fibers, the 1100 C heat treated fiber ha d the roughest surface due to the anatase crystallites formed on the surfac e compared to other fibers heat treated at lower temperature which kept the amorphous phase or smaller grains. In spite of the rough surface that would induce the inferior mechanical stability, however, fibers of anatase TiO2SiO2 phase are still preferred because of versatility of the anatase, e.g. the microwave absorption, the photocatalytic effect. These functionalities could b e of utmost importance in future applications, e.g. the regeneration of filtration media by the microwave absorption, chemically active filtration media with the photocatalytic effect.

PAGE 81

81 Table 51 Result of analysis on fiber diameter of PAN and TiO2. Material PAN TiO 2 Polymer concentration w/v DMF/EtOH 6% PVP 4% PVP 16% Heat treatment N/A 500 C 500 C d f 224 149 1110 C v 0.25 0.24 0.20 Table 52 Result of analysis on fiber diameter of TiO2SiO2. Material TiO 2 SiO 2 Polymer concentration w/v sol PVP 1% PVP 2% Heat treatment as spun 500 C 800 C 1100 C as spun 500 C 800 C 1100 C d f 444 228 221 243 953 361 336 375 C v 0.29 0.27 0.25 0.29 0.22 0.18 0.20 0.23 Table 53 Results of filtration tests 36, 51. PAN (a) ( b) means (b) layers of PAN fibermat fabricated by (a) minute deposition via electrospinning. If (b) is unspecified, its a single layer sample. Filter P ACF2 8.66 0.37 10 1 21 Millipore HEPA 7 94 3.71 10 4 1171 LydAir HEPA 1. 35 4.35 10 4 284 PAN 5 3 67 0.07 10 1 18 PAN 15 1.0 3 0.1910 1 98 PAN 53 6 54 0.16 10 2 41 PAN 152 2. 2 5 .7 8 10 2 152 Military standard 310 4 400 Table 54 Values of electronegativity of atomic species in GPTMS 53. At om Electronegativity H 2.2 C 2.5 O 3.5 Si 1.7

PAGE 82

82 Table 55 Integration of block A, B and C (Figure 521, Figure 522, Figure 523 and Figure 524). [0.005N HNO 3 ]:[GPTMS] A B C 1:1 0.97 4.00 1.00 2:1 2.03 7.16 1.00 4:1 5.68 10.37 1.00 8:1 12.88 10.85 1.00 Table 5 6 Calculated [MeOH] and [H2O] based on values in Table 55 and experimental errors of [MeOH]+[H2O] [0.005N HNO 3 ]:[GPTMS] [MeOH] [H 2 O] [MeOH]+[H 2 O] Error Experiment Theory 1:1 1.00 0.02 0.98 1 2% 2:1 2.05 0.01 2.04 2 2% 4:1 3.13 1.27 4.40 4 10% 8:1 3.28 4.80 8.08 8 1% Table 57 Time required for gelation of sol of [0.005N HNO3]:[GPTMS]:[metal alkoxide ]=2:1:x before gelation or precipitation. Precipitated sols are indicated as (p). [Metal alkoxide] X Time taken before Gelation/Precipitation TB TiPP ZP ZB 0.125 6 24 hours 6 24 hours 24 hours 2 18 hours 0.25 <4 hours 6 24 hours 2 18 hours 2 18 hours 0.5 16 20 hours <3 hours 2 18 hours <1 hour 1 4 weeks 4 days <1 hour 2 18 hours 2 4 months 5 months 3 days 3 days 4 3.5 months 3 days (p) 10 days 20 hours (p) Table 58 Grain size of different TiO2 phases calculated by Scherrer's equation on spectra shown in Figure 532 N/A means that the peaks were too weak to be analyzed. Heat treatment temperature (C) for 3 hours 500 600 700 800 900 anatase (nm) 16 22 28 34 N/A rutile (nm) N/A N/A 35 40 43

PAGE 83

83 Figure 51 An SEM image of electrospun PAN fibers from PAN 6% w/v DMF Figure 52 Fiber diameter distribution of electrospun PAN fibers. 0 5 10 15 20 25 30 35 40 100 120 140 160 180 200 220 240 260 280 300 320 340 MoreFrequencyfiber diameter (nm)PAN Fiber Diameter Distribution

PAGE 84

84 Figure 53 (A) Thermogravimet ric analysis of pol y vinyl pyrrolidone degradation, and (B) its first derivative to locate the peak. Dried air was flown to the TGA chamber at 0.0 4 LPM in 20 psi. Temperature was increased from the room temperature to 1000 C at 10 C /min. Figure 54 SEM images of electrospun TiO2 fibers from (A ) PVP 4% and (B ) 16% w/v EtOH after heat treatment. Scale bars are (A ) 2 m and (B ) 5 m.

PAGE 85

85 Figure 55 Fiber diameter distribution of electrospun TiO2 fibers from PVP 4% w/v EtOH Figure 56 Fiber diameter distribution of electrospun TiO2 fibers from PVP 16% w/v EtOH 0 5 10 15 20 2570 80 90 100 110 120 130 140 150 160 170 180 190 200 210 220 230 240 More FrequencyFiber Diameter (nm)dfDistribution of TiO2Fibers from PVP 4% w/v sol 0 2 4 6 8 10 12 14 16 18 20750 800 850 900 950 1000 1050 1100 1150 1200 1250 1300 1350 1400 1450 1500 1550 1600 1650 1700 1750 More FrequencyFiber Diameter (nm)dfDistribution of TiO2Fibers from PVP 16% w/v sol

PAGE 86

86 Figure 57 (A) df and, (B) Cv of as spun PAN fiber, heat treated TiO2 fibers with different polymer concentration, heat treated SiC fibers, fibers that compose Millipore HEPA and LydAir HEPA.

PAGE 87

87 Figure 5 8 SEM images of as spun TiO2SiO2 composite fibers from aqueous sol with PVP concentration of ( A) 0.25%, ( B) 0.5%, ( C) 1%, ( D) 2% w/v sol. Scale bars are 20 m

PAGE 88

88 Figure 59 Dependence of fiber diameter for electrospun fibers of PVP 1% and 2% w/v sol on heat treatment. Connecting lines are for guidance of the eye only. 0 200 400 600 800 1000 1200 as spun 500 800 1100Mean fiber diameter (nm)Temperature( C) of heat treatment for 3 hoursFiber Diameter 2% 1%

PAGE 89

89 Figure 510. Electrospun TiO2SiO2 composite fiber from full aqueous sol. df is 243 nm.

PAGE 90

90 Figure 511. Fiber diameter distribution of electrospun TiO2 SiO2 composite fibers from PVP 1% w/v sol, depending on different heat treatment profile.

PAGE 91

91 Figure 512. Fiber diameter distribution of electrospun TiO2 SiO2 composite fibers from PVP 2% w/v sol, depending on different heat treatment profile. Figur e 513. SEM images of (A ) Millipore HEPA ( CAT. NO.: A P1504700, Mill ipore, MA, USA), (B) LydAir HEPA ( LydAir High Alpha HEPA air filtration media HEPA Lydair grade 4450H Lydall Filtration, CT )

PAGE 92

92 Figure 514. Quality factor (qF) of filters with standard deviation calculated by data from Table 53 Figure 515. Chemical structure of GPTMS. Chemically different (A) carbon atoms and, (B) hydrogen atoms are noted by different alphabet subscript s. 0.007 0.020 0.008 0.031 0.056 0.023 0.067 0 0.01 0.02 0.03 0.04 0.05 0.06 0.07 0.08 ACF 2 Military standard Millipore HEPA LydAir HEPA PAN 5 PAN 15 PAN 5 3 qF(Pa1)

PAGE 93

93 Figure 516. 13C NMR spectra of (from bottom to top) GPTMS, mixture of 0.005N HNO3 and GPTMS at molar ratio of 1:1, 2:1, 4:1, 8:1, 16:1 and methanol. The wedges show the peaks from transition states. Peaks are assigned to carbons of GPTMS molecule in Figure 515 (A) 69. Figure 5 17. Proposed hydrolysis reaction of GPTMS in aqueous HNO3. Chemically different atoms are noted by different alphabetical subscripts.

PAGE 94

94 Figure 518. 1H NMR spectrum with peaks assigned to hydrogen atoms of GPTMS molecule in Figure 515 (B).

PAGE 95

95 Figure 519. 1H NMR spectra of (from bottom to top) GPTMS, 0.005N HNO3 and GPTMS at molar ratio of 1:1, 2:1, 4:1, 8:1, 0.005N HNO3 and methanol. (A) Peaks of hydrogen atoms of hydroxyl group and, (B) overlapped peaks of water and hydrogen atom of hydroxyl group of methanol. Figure 5 20. 1H NMR spectra of (from bottom to top) GPTMS, 0.005N HNO3 and GPTMS at molar ratio of 1:1, 2:1, 4:1, 8:1, 0.005N HNO3 and methanol. ( A ) O verlapped regime of hydrogen of water and hydroxyl group of methanol

PAGE 96

96 ( Figure 519 ( B )), ( B ) o verlapped regime of hy drogen that is bonded to carbon of methanol and He ( Figure 515 (B) ) ( C ) regime of Hg ( Figure 515 (B) ) as the reference for the integration. Figure 5 21. 1H NMR spectra of 0.005N HNO3 and GPTMS at molar ratio of 1:1 with partial integration of block A, B and C from left to right. Figure 5 22. 1H NMR spectra of 0.005N HNO3 and GPTMS at molar ratio of 2 :1 with partial integration of block A, B and C from left to right.

PAGE 97

97 Figure 5 23. 1H NMR spectra of 0.005N HNO3 and GPTMS at molar ratio of 4 :1 with partial integration of block A, B and C from left to right. Figure 5 24. 1H NMR spectra of 0.005N HNO3 and GPTMS at molar ratio of 8 :1 with partial integration of block A, B and C from left to right.

PAGE 98

98 Figure 525. Enlarged 1H NMR spectra of Figure 519 (A). Figure 526. Transition state of GPTMS in the proposed hydrolysis reaction ( Figure 5 17 ). (A) One methoxy group is hydrolyzed, (B) two methoxy groups are hydrolyzed and, (C) fully hydrolyzed GPTHS.

PAGE 99

99 Figure 527. Enlarged 1H NMR spectrum of 0.005N HNO3 and GPTMS at molar ratio of 2:1 in Figure 5 19 (A).

PAGE 100

100 Figure 528. 13C NMR spectra of (from bottom to top) [ 0.005N HNO3] : [ GPTMS ] =2:1, TB, [ 0.005N HNO3] : [ GPTMS ] : [ TB ] =2:1:1 and 1butanol. The wedges indicate the peaks of 1butanol. Notations in peak assignment for TB and [ 0.005N HNO3] : [ GPTMS ] =2:1 are based on Figure 529 and Figure 515 (A), respectively. Figure 529. Chemical structur e of TB. Chemically different carbon atoms are noted by different alphabet subscripts.

PAGE 101

101 Figure 5 30. Proposed condensation reaction between {GPTMS+HNO3} and TB. Chemically different atoms are noted by different alphabetical subscripts.

PAGE 102

102 Figure 5-31. 13C NMR spectra of (from bottom to top) [ 0.005N HNO3] : [ GPTMS ] =2:1 and [ 0.005N HNO3] : [ GPTMS ] : [ TB ] =2:1:1. The wedge indicates the peak of methanol.

PAGE 103

103 20 30 40 50 60 70 80 intensity (a. u.)2theta 900 degree C 800 degree C 700 degree C 600 degree C 500 degree C A-anatase R-rutile R R R R R R R R A A A A A A Figure 5 32. X ray diffraction pattern of electrospun TiO2 fibers heat treated at various temperatures for 3 hours Peaks notated as A and R indicate anatase and rutile phase, respectively.

PAGE 104

104 Figure 533. TEM images of electrospun TiO2 fibers heat treated at (A) 500 C (B) 600 C (C) 700 C (D) 800 C and (E) 900 C for 3 hours

PAGE 105

105 Figure 534. X ray diffraction pattern of electrospun TiO2SiO2 composite fibers heat treated at various temperatures. Peaks marked with wedges indicate anatase phase of TiO2.

PAGE 106

106 Figure 5 35. TEM images of electrospun TiO2 SiO2 composite fibers from aqueous sol heat treated at (A) 500 C (B) 800 C and (C) 1100 C for 3 hours

PAGE 107

107 Figure 5 36. SEM images of 1100 C heat treated electrospun TiO2SiO2 fibers from sol A. [TiO2]:[SiO2] =1:1. Images are from same batch of sample that show different morphologies, i.e. (A) rough surfaces of fiber and (B) spikes from the fiber. 10 20 30 40 50 60 70 Intensity (a.u.)2 theta 1100 degree C-3 hours As-spun Figure 537. XRD result of electrospun TiO2SiO2 composite fibers before and after heat treatment at 1100 C for 3 hours

PAGE 108

108 CHAPTER 6 CONCLUSIONS AND FUTURE WORK Conclusions In the filtration test, it was presented that filtration media compris i ng of thinner fibers have a higher quality factor (qF), expressed as equation 11, than the microfiber based high efficiency particulate air ( HEPA) filters. The mean fiber diameter (df) of electrospun polyacrylonitrile ( PAN ) fibers was 224 nm. The qF of PAN 3x5 was 0.067 while the HEPA filter that is adopted and currently used by the U.S. Army has a qF of 0.0310.001, the military standard being established at 0.020. When 3 layers of 5minutedepositioned PAN fibermats (PAN 5) were used to improve t he uniformity of fiber deposition, the qF was amplified by 69% from that of a single layer of PAN 5. In the electrospinning process, the mean fiber diameter ( df) of anatase TiO2 was increased from 149 nm to 1110 nm by increasing the polyvinylpyrrolidone ( P VP) concentration from 4% to 16% weight per volume ( w/v g/ml) in ethanol after heat treatment to remove of PVP and for crystallization. TiO2SiO2 composite fibers were successfully electrospun from novel aqueous sol with the extended time taken before gelation or precipitation from 4 hours of ethanol based TiO2 sol to 4 months Precipitation was determined when the sol got opaque based on the visual observation by every 10 minutes while gelation was defined when the sol didnt flow down in the glass vial located upside down. The NMR study on the reaction mechanisms in the sol provided a basis upon which to optimize the composition for a stable aqueous sol. Morphologic change by the electrospun TiO2SiO2 composite material was observed going from beads to smooth fibers with increasing PVP concentration. During heat treatment the electrospun fiber shrank in diameter by 48% for the sol with PVP 1% w/v

PAGE 109

109 sol due to thermal degradation of PVP and mass transport by amorphous SiO2 to fill in the space left from the polymer burnout. The df that was achieved for TiO2SiO2 composite fibers with anatase phase of TiO2 was 243 nm. Since the mean fiber diameter of TiO2SiO2 composite fibers is thicker than that of PAN fibers used in the filtration tests by 8% the Ti O2SiO2 composite fibermat would have a lower filtration efficiency and higher pressure drop than that of PAN fibermats. Based on the geometry, the drag force of the TiO2SiO2 composite fiber was calculated to be 13% higher than for the PAN fiber tested. H owever, the composite fiber would have a higher quality factor than the LydAir HEPA filter which has a 221% higher drag force due to having fibers with a mean fiber diameter of 527 nm. Since filtration is a complicated phenomen on, filtration theory involves many complex empirical factors and simplifying assumptions. As a consequence, parameters like collection efficiency and pressure drop are not calculated for the whole filtration system. Nonetheless, the tendency of their changes by fiber diameter is pred ictable. From the NMR study, the hydrolysis and condensation reactions between 0.005N HNO3, 3 glycidoxypropyltrimethoxylsilane ( GPTMS ) and titanium (IV) n butoxide ( TB ) for a component of the aqueous sol system were hypothesized. Methoxy groups of GPTMS are hydrolyzed by water molecules and hydroxyl group formed by hydrolysis reaction react with TB to form Si O Ti bond through a condensation reaction. Reaction paths are described in Figure 5 17 and Figure 5 30 Through the 3 hours of heat treatment process on the TiO2 fibers, the anatase phase was formed at 500C while the pure rutile phase was developed at 700C and became predominant at 900C. Grains of both phases grew as the temperature of heat

PAGE 110

110 treatment was increased. This is attributed to the higher mobility of atoms towards the formation of larger grains wit h smaller specific surface areai.e. grain boundary. For TiO2SiO2 composite fibers, the anatase phase was clearly observed after heat treatment at 1100C while formation of TiO2 grains dispersed in the fiber was observed at 800C Whether the anatase grai ns after heat treatment at 1100C are well distributed in the fiber or aggregated on the surface of the fiber hasnt been clarified by TEM observation. The higher temperature required for the formation of the anatase phase is explained by the formation of Si O Ti bonds during the condensation reaction which then retards the formation of Ti O Ti bonds at temperatures lower than 1100C. Unlike the coarse structure of anatase TiO2 fibers, the TiO2SiO2 composite fibers exhibited dense structures of anatase TiO2 crystallites surrounded by an amorphous SiO2 matrix Because the crystallization temperature of SiO2 to quartz is at 1500 C, the SiO2 maintains its noncrystallinity at 1100 C while the increased molecular mobi lity allows for it to diffuse and fill the space formed during the crystallization of TiO2. More active grain growth, i.e. larger grain size, needle like crystallites rooted on the fiber, was observed in electrospun TiO2SiO2 fibers from the use of aqueous sol A with an abundance of TiO2. The composite fibers from the aqueous sol A have TiO2 molar concentrations of 50% while those fibers from full aqueous sol has a TiO2 concentration of 29% by mole. Future Work With regard to the mechanical stability issue, the TiO2SiO2composite nanofiber mats were tested by handling the samples in the lab similarly to the mechanical conditions that a filter would be expected to sustain. Superior strength by the composite fibers compared to pure anatase TiO2 nanofiber mats was observed. The fibermats

PAGE 111

111 could be handled with tweezers or hands by holding on to a small edge. They did not fracture when doing so. This is in stark contrast to the pure polycrystalline TiO2 nanofiber mats. Those only sustained handling, without fracturing, if they were supported by other materials at all times. Further tests on the mechanical propert ies e.g. elastic modulus, toughness of either single fiber or fibermat s are desired to be done in the future ( Figure 61 ). Fibermat compris ing of TiO2SiO2 composite nanofibers is anticipated to fulfill the performance that the PAN fibermat with 10% thinner df showed as filtration media. However, in contrast to the polymeric fibers, TiO2based fibers dont have smooth surfaces upon crystallization. They have cleavages, spikes, and pores depending on the composition and the heat treatment. The drag force and filtration efficiency by the filtration media with thes e rough microstructures cannot be the same as those by filtration media with smooth surfaces. It is a good material to conduct research on to determine how different microstructures affect filtration behavior. With respect to the fibers, increasing the electrical conductivity and decreasing the surface tension of the electrospinning solution have not been tried yet, but these changes are expected to further reduce the fiber diameter of the TiO2SiO2 composite fibers either from full aqueous sol or sol A Fo r sol A, sols with lower and higher molar ratio of TB or lower PVP concentration were electrospun while fibers with mean fiber diameter less than 1 m were not obtained from any of them. By adding additives like electrolytes and surfactants, a further decr ease in fiber diameters is theoretically possible.

PAGE 112

112 In regards to the chemistry in the aqueous sol, the reaction paths in sol B and the full sol, which is the mixture of sol A and B, were not discussed here due to the limitations of resources. This topic is of great interest for further research because of the unique stabilization of the metal alkoxide in conditions with an abundance of water. Aging of aqueous sol is another subject to explore as well due to its surprisingly elongated shelf life when in contact with an abundance of water molecules. Results from observation on aqueous sols with varying metal alkoxides and their molar ratios are l isted in Table 57 Since the samples were not watched through for 24 hours, the shelf life in Table 5 7 has an error of 1 or 2 hours. Moreover, the point of gelation was ambiguous because it is practically impossible to run the rheometer to measure the change of viscosity of the sample insitu for, sometimes, over a few months. H owever, despite the result being from a limited number of metal alkoxides and molar ratios, clear tendencies for longer shelf life was observed with optimized molar ratio for each metal alkoxide. Discerning the aging mechanism would be of great interest an d would provide clues to improve the stability of sols containing metal alkoxide in other organic system. The chemical mechanisms of full aqueous sol could be understood in parallel as well.

PAGE 113

113 Figure 61 (A) Fractured anatase TiO2 fibermats composed of fibers of df=149 nm, (B) A TiO2SiO2 composite fibermat composed of fibers of df=243 nm from full aqueous sol after heat treatment.

PAGE 114

114 LIST OF REFERENCE S 1. Care, A. A. f. R. Lung Health Statistics by State. http://www.aarc.org/ (March, 15, 2010) 2. Foundation, W. L. Annual Report ; World Lung Foundation: New York, 2008. 3. United States. Environmental Protection Agency. Office of, R.; Development; National Center for E nvironmental, A. Integrated science assessment for particulate matter. http://cfpub.epa.gov/ncea/cfm/recordisplay.cfm?deid=201805 4. Brinker, C. J.; Scherer, G. W., Sol gel science : the physics and chemistry of sol gel processing. Academic Press: Boston, 1990; p xiv, 908. 5. Li, D.; Xia, Y. N., Fabrication of titania nanofibers by electrospinning. Nano Letters 2003, 3, (4), 555560. 6. Yang, X. G.; Shao, C. L.; Liu, Y. C.; Mu, R. X.; Guan, H. Y., Nanofibers of CeO2 via an electrospinning technique. Thin Solid Films 2005, 478, (12), 228 231. 7. Shao, C. L.; Guan, H. Y.; Liu, Y. C.; Gong, J.; Yu, N.; Yang, X. H., A novel method for making ZrO2 nanofibres via an electrospinning technique. Journal of Crystal Growth 2004, 267, (12), 380 384. 8. Larsen, G.; Velarde Ortiz, R.; Minchow, K.; Barrero, A.; Loscertales, I. G., A method for making inorganic and hybrid (organic/inorganic) fibers an d vesicles with diameters in the submicrometer and micrometer range via sol gel chemistry and electrically forced liquid jets. Journal of the American Chemical Society 2003, 125, (5), 1154 1155. 9. Sigmund, W.; Yuh, J.; Park, H.; Maneeratana, V.; Pyrgiotakis, G.; Daga, A.; Taylor, J.; Nino, J. C., Processing and structure relationships in electrospinning of ceramic fiber systems. Journal of the American Ceramic Society 2006, 89, (2), 395407. 10. Hinds, W. C., Aerosol technology : properties, behavior, an d measurement of airborne particles 2nd ed.; Wiley: New York, 1999; p xx, 483. 11. Allen, M. D.; Raabe, O. G., Slip Correction Measurements of Spherical Solid Aerosol Particles in an Improved Millikan Apparatus. Aerosol Science and Technology 1985, 4, (3), 269286. 12. Allen, M. D.; Raabe, O. G., Re evaluation of Millikan Oil Drop Data for the Motion of Small Particls in Air. Journal of Aerosol Science 1982, 13, (6), 537 547. 13. Lee, K. W.; Ramamurthi, M., Filter Collection. In Aerosol Measurement: Principles, Techniques, and Applications Willeke, K.; Baron, P. A., Eds. Van Nostrand Reinhold: New York, 1993; p 179.

PAGE 115

115 14. Yeh, H. C.; Liu, B. Y. H., Aerosol filtration by fibrous filters --I. theoretical. Journal of Aerosol Science 197 4, 5, (2), 191204. 15. Kirsch, A. A.; Fuchs, N. A., Studies on fibrous aerosol filters. 3. Diffusional deposition of aerosols in fibrous filters. Ann Occup Hyg 1968, 11, (4), 299 304. 16. Wang, J.; Kim, S. C.; Pui, D. Y., Investigation of the figure of merit for filters with a single nanofiber layer on a substrate. Journal of Aerosol Science 2008, 39, (4), 323334. 17. Darcy, H., Les Fontaines Publiques de la Ville de Dijon. Dalmont: Paris, 1856. 18. Livage, J.; Henry, M.; Sanchez, C., Sol gel Chemistr y of Transitionmetal Oxides. Progress in Solid State Chemistry 1988, 18, (4), 259 341. 19. Taylor, G., Disintegration of Water Drops in Electric Field. Proceedings of the Royal Society of London Series aMathematical and Physical Sciences 1964, 280, (1380), 383 +. 20. Fridrikh, S. V.; Yu, J. H.; Brenner, M. P.; Rutledge, G. C., Controlling the Fiber Diameter during Electrospinning. Physical Review Letters 2003, 90, (14), 144502. 21. Hohman, M. M.; Shin, M.; Rutledge, G.; Brenner, M. P., Electr ospinning and electrically forced jets. I. Stability theory. Physics of Fluids 2001, 13, (8), 2201 2220. 22. Hohman, M. M.; Shin, M.; Rutledge, G.; Brenner, M. P., Electrospinning and electrically forced jets. II. Applications. Physics of Fluids 2001, 13, (8), 2221 2236. 23. He, J. H.; Wan, Y. Q.; Yu, J. Y., Effect of concentration on electrospun polyacrylonitrile (PAN) nanofibers. Fiber. Polym. 2008, 9, (2), 140142. 24. Jarusuwannapoom, T.; Hongroijanawiwat, W.; Jitjaicham, S.; Wannatong, L.; Nithitana kul, M.; Pattamaprom, C.; Koombhongse, P.; Rangkupan, R.; Supaphol, P., Effect of solvents on electro spinnability of polystyrene solutions and morphological appearance of resulting electrospun polystyrene fibers. European Polymer Journal 2005, 41, (3), 40 9 421. 25. Skotak, M.; Larsen, G., Solution chemistry control to make well defined submicron continuous fibres by electrospinning: the (CH3CH2CH2O)(4)Ti/AcOH/poly(N vinylpyrrolidone) system. Journal of Materials Chemistry 2006, 16, (29), 3031 3039. 26. S on, W. K.; Cho, D.; Park, W. H., Direct electrospinning of ultrafine titania fibres in the absence of polymer additives and formation of pure anatase titania fibres at low temperature. Nanotechnology 2006, 17, (2), 439443.

PAGE 116

116 27. Thompson, C. J.; Chase, G. G.; Yarin, A. L.; Reneker, D. H., Effects of parameters on nanofiber diameter determined from electrospinning model. Polymer 2007, 48, (23), 69136922. 28. Burger, C.; Hsiao, B. S.; Chu, B., Nanofibrous materials and their appli cations. Annual Review of Materials Research 2006, 36, 333368. 29. Chen, H. C.; Jao, W.C.; Yang, M.C., Characterization of gelatin nanofibers electrospun using ethanol/formic acid/water as a solvent. Polymers for Advanced Technologies 2009, 20, (2), 98 103. 30. Cengiz, F.; Krucinska, I.; Gliscinska, E.; Chrzanowski, M.; Goktepe, F., Comparative Analysis of Various Electrospinning Methods of Nanofibre Formation. Fibres & Textiles in Eastern Europe 2009, 17, (1), 1319. 31. Tsai, P. P.; Chen, W.; Roth, J. R., Investigation of the Fiber, Bulk, and Surface Properties of Meltblown and Electrospun Polymeric Fabrics. International Nonwovens Journal 2004, 13, (3), 17 23. 32. David, L. I. B.; Ismail, A. F., Influence of the thermastabilization process and soak time during pyrolysis process on the polyacrylonitrile carbon membranes for O 2/N 2 separation. Journal of Membrane Science 2003, 213, (1 2), 285291. 33. Fitzer, E.; Frohs, W.; Heine, M., Optimization of Stabilization and Carbonization Treatment of Pan Fibers and Structural Characterization of the Resulting CarbonFibers. Carbon 1986, 24, (4), 387395. 34. Hou, H. Q.; Ge, J. J.; Zeng, J.; Li, Q.; Reneker, D. H.; Greiner, A.; Cheng, S. Z. D., Electrospun polyacrylonitrile nanofibers containing a high concentration of well aligned multiwall carbon nanotubes. Chemistry of Materials 2005, 17, (5), 967973. 35. Zhang, W. X.; Liu, J.; Wu, G., Evolution of struct ure and properties of PAN precursors during their conversion to carbon fibers. Carbon 2003, 41, (14), 2805 2812. 36. Zhang, Q.; Welch, J.; Park, H.; Wu, C.Y.; Sigmund, W.; Marijnissen, J. C. M., Improvement in nanofiber filtration by multiple thin layers of nanofiber mats. Journal of Aerosol Science In Press, Corrected Proof. 37. Muscat, J.; Swamy, V.; Harrison, N. M., Firstprinciples calculations of the phase stability of TiO2. Physical Review B 2002, 65, (22), 38. Arashi, H., Raman Spectroscopic S tudy of the PressureInduced PhaseTransition in Tio2. J Phys Chem Solids 1992, 53, (3), 355 359. 39. Mammone, J. F.; Nicol, M.; Sharma, S. K., Raman Spectra of Tio2Ii, Tio2 Iii, Sno2, and Geo2 at HighPressure. J Phys Chem Solids 1981, 42, (5), 379384.

PAGE 117

117 40. Mammone, J. F.; Sharma, S. K.; Nicol, M., Raman Study of Rutile (Tio2) at HighPressures. Solid State Communications 1980, 34, (10), 799802. 41. Dervos, C. T.; Thirios; Novacovich, J.; Vassiliou, P.; Skafidas, P., Permittivity properties of thermall y treated TiO2. Materials Letters 2004, 58, (9), 1502 1507. 42. Woan, K.; Pyrgiotakis, G.; Sigmund, W., Photocatalytic CarbonNanotube TiO2 Composites. Advanced Materials 2009, 21, (21), 22332239. 43. Pyrgiotakis, G. Titania carbon nanotube composites for enhanced photocatalysis. http://purl.fcla.edu/fcla/etd/UFE0013482 44. Olofinjana, A. O.; Bell, J. M.; Jmting, A. K., Evaluation of the mechanic al properties of sol gel deposited titania films using ultramicro indentation method. Wear 2000, 241, (2), 174179. 45. Soares, P.; Mikowski, A.; Lepienski, C. M.; Santos, E.; Soares, G. A.; Swinka, V.; Kuromoto, N. K., Hardness and elastic modulus of TiO2 anodic films measured by instrumented indentation. Journal of Biomedical Materials Research Part B Applied Biomaterials 2008, 84B, (2), 524530. 46. Cromer, D. T.; Herrington, K., The Structures of Anatase and Rutile. Journal of the American Chemical Society 1955, 77, (18), 4708 4709. 47. Avci, N.; Smet, P. F.; Poelman, H.; Van de Velde, N.; De Buysser, K.; Van Driessche I.; Poelman, D., Characterization of TiO2 powders and thin films prepared by nonaqueous sol gel techniques. Journal of Sol Gel Science and Technology 2009, 52, (3), 424 431. 48. Madhugiri, S.; Sun, B.; Smirniotis, P. G.; Ferraris, J. P.; Balkus, K. J., Electrospun mesoporous titanium dioxide fibers. Microporous and Mesoporous Materials 2004, 69, (1 2), 7783. 49. Nuansing, W.; Ninmuang, S.; Jarernboon, W.; Maensiri, S.; Seraphin, S., Structural characterization and morphology of electrospun TiO2 nanofi bers. Materials Science and Engineering B Solid State Materials for Advanced Technology 2006, 131, (1 3), 147 155. 50. Kanno, Y., Discussion from the Nucleation for the Crystallization Temperature of Vitreous Silica. J. Mater. Sci. Lett. 1990, 9, (4), 451 453. 51. Defense, D. o., Filter Medium, Fire Resistant, High Efficiency. In MIL F 51079D U. S. Government Printing Office: 1998. 52. Gnther, H., NMR spectroscopy: basic principles, concepts, and applications in chemistry 2 ed.; John Wiley & Sons Ltd. : New York, 1995; p 581.

PAGE 118

118 53. Allen, S. M.; Thomas, E. L., The Structure of Materials 1st edition ed.; John Wiley & Sons, Inc.: New York, 1999; p 472. 54. Breitmaier, E.; Voelter, W., Carbon13 NMR spectroscopy: highresolution methods and applications in organic chemistry and biochemistry 3 ed.; VCH Publishers: New York, 1987; p 515. 55. Freeman, R., A Handbook of Nuclear Magnetic Resonance. 2 ed.; Addison Wesley Longman Ltd.: Harlow, UK, 1997; p 344. 56. Gottlieb, H. E.; Kotlyar, V.; Nudelman, A., NMR chemical shifts of common laboratory solvents as trace impurities. Journal of Organic Chemistry 1997, 62, (21), 7512 7515. 57. Cullity, B. D.; Stock, S. R., Elements of X Ray Diffraction 3 ed.; Prentice Hall, Inc.: Upper Saddle River, NJ, 2001. 58. Klug, H. P.; Alexander, L. E., X ray Diffraction Procedures John Wiley & Sons: New York, 1974; p 966. 59. Yoon, K.; Kim, K.; Wang, X.; Fang, D.; Hsiao, B. S.; Chu, B., High flux ultrafiltration membranes based on electrospun nanofibrous PAN scaffolds and chitosan coating. Polymer 2006, 47, (7), 24342441. 60. Chrissafis, K.; Paraskevopoulos, K. M.; Papageorgiou, G. Z.; Bikiaris, D. N., Thermal and dynamic mechanical behavior of bionanocomposites: Fumed silica nanoparticles dispersed in poly(vinyl pyrrolidone), chitosan, and poly(vinyl alcohol). Journal of Appli ed Polymer Science 2008, 110, (3), 1739 1749. 61. Shibaev, L. A.; Melenevskaya, E. Y.; Ginzburg, B. M.; Yakimanskii, A. V.; Ratnikova, O. V.; Gribanov, A. V., Thermal and tribological properties of fullerene containing composite systems. Part 3. Features of the mechanism of thermal degradation of poly (N vinyl pyrrolidone) and its compositions with fullerene C 60. J. Macromol. Sci. Part B Phys. 2008, 47, (2), 276287. 62. Lau, C.; Mi, Y., A study of blending and complexation of poly(acrylic acid)/poly(vin yl pyrrolidone). Polymer 2002, 43, (3), 823829. 63. Lee, D. Y., Titania nanofibers prepared by electrospinning. Journal of the Korean Physical Society 2006, 48, (6), 16861690. 64. Nakamura, T.; Ichitsubo, T.; Matsubara, E.; Muramatsu, A.; Sato, N.; Tak ahashi, H., On the preferential formation of anatase in amorphous titanium oxide film. Scripta Materialia 2005, 53, (9), 1019 1023.

PAGE 119

119 65. Fong, H.; Chun, I.; Reneker, D. H., Beaded nanofibers formed during electrospinning. Polymer 1999, 40, (16), 4585 4592. 66. Li, D.; Xia, Y. N., Electrospinning of nanofibers: Reinventing the wheel? Advanced Materials 2004, 16, (14), 11511170. 67. Kumar, A.; Jose, R.; Fujihara, K.; Wang, J.; Ramakrishna, S., Structural and optical properties of electrospun TiO2 nanofiber s. Chemistry of Materials 2007, 19, (26), 6536 6542. 68. Reed, J. S., Principles of Ceramics Processing. 2nd edition ed.; John Wiley & Sons, Inc.: New York, 1995. 69. Umeda, J.; Moriya, M.; Sakamoto, W.; Yogo, T., Synthesis of proton conductive inorganic organic hybrid membranes from organoalkoxysilane and hydroxyalkylphosphonic acid. Journal of Membrane Science 2009, 326, (2), 701 707. 70. Yoder, C. H.; Schaeffer, J. C. D., Introduction to Multinuclear NMR The Benjamin/Cummings Publushing Company, Inc. : Menlo Park, CA, 1987; p 335. 71. Keeler, J., Understanding NMR Spectroscopy John Wiley & Sons Ltd.: Chichester, West Sussex, UK, 2005; p 459. 72. Samantaray, S. K.; Parida, K. M., SO42 /TiO2 SiO2 mixed oxide catalyst: 2. Effect of the fluoride ion and calcination temperature on esterification of acetic acid. Applied Catalysis A: General 2001, 211, (2), 175187. 73. Schmidt, H. K., Organically modif ied silicates and ceramics as two phasic systems: Synthesis and processing. Journal of Sol Gel Science and Technology 1997, 8, (13), 557565. 74. Falaras, P.; Xagas, A. P., Roughness and fractality of nanostructured TiO2 films prepared via sol gel technique. Journal of Materials Science 2002, 37, (18), 3855 3860. 75. Williams, D. B.; Carter, C. B., Transmission electron microscopy 3. Imaging. Plenum Pr.: New York [u.a., 1996. 76. Jiang, H. G.; Ruhle, M.; Lavernia, E. J., On the applicability of the x ray diffraction line profile analysis in extracting grain size and microstrain in nanocrystalline materials. Journal of Materials Research 1999, 14, ( 2), 549 559. 77. Mo, S. D.; Ching, W. Y., Electronic and optical properties of three phases of titanium dioxide: Rutile, anatase, and brookite. Physical Review B 1995, 51, (19), 13023.

PAGE 120

120 78. Wu, Q.; Zhao, B. Y.; Fang, C.; Hu, K. A., An enhanced polarization mechanism for the metal cations modified amorphous TiO2 based electrorheological materials. European Physical Journal E 2005, 17, (1), 63 67. 79. Izutsu, H.; Nair, P. K.; Maeda, K.; Kiyozumi, Y.; Mizukami, F., Structure and properties of TiO2---SiO2 prep ared by sol gel method in the presence of tartaric acid. Materials Research Bulletin 1997, 32, (9), 1303 1311. 80. Choi, S. S.; Lee, S. G.; Im, S. S.; Kim, S. H.; Joo, Y. L., Silica nanofibers from electrospinning/sol gel process. Journal of Materials Sci ence Letters 2003, 22, (12), 891893. 81. Lee, S. W.; Kim, Y. U.; Choi, S. S.; Park, T. Y.; Joo, Y. L.; Lee, S. G., Preparation of SiO2/TiO2 composite fibers by sol gel reaction and electrospinning. Materials Letters 2007, 61, (3), 889893. 82. Ren, J.; Li, Z.; Liu, S. S.; Xing, Y. L.; Xie, K. C., Silica titania mixed oxides: Si O Ti connectivity, coordination of titanium, and surface acidic properties. Catal. Lett. 2008, 124, (3 4), 185 194. 83. Viswanath, R. N.; Ramasamy, S., Study of TiO2 nanocrystall ites in TiO2---SiO2 composites. Colloids and Surfaces A: Physicochemical and Engineering Aspects 1998, 133, (1 2), 49 56.

PAGE 121

121 BIOGRAPHICAL SKETCH Hyoungjun Park has worked with Dr. Sigmund in the field of sol gel chemistry, electrospinning and other ceramics processing methods since 2006 at University of Florida. He earned his Bachelor of Science degree in m aterials science and e ngineering at Seoul National University, South Korea.