POROUS TITANIUM DIOXIDE NANOMATERIALS FOR PHOTOCATALYTIC AND
PHOTOVOLTAIC APPLICATIONS
By
Yan Li

A DISSERTATION
Submitted to
Michigan State University
in partial fulfillment of the requirements
for the degree of
Chemical Engineering – Doctor of Philosophy
2016

ABSTRACT
POROUS TITANIUM DIOXIDE NANOMATERIALS FOR PHOTOCATALYTIC AND
PHOTOVOLTAIC APPLICATIONS
By
Yan Li
The huge demand for fossil fuels and the risks of the environmental crisis have spurred
an interest in renewable energies. Using the clean and abundant solar energy,
semiconductor photocatalysis and photovoltaics have evoked tremendous interest.
Titanium dioxide (titania, TiO2) has been the focus of the research trend because of its
excellent crystallinity, photoreactivity, chemical and thermal stability, and low cost. The
research objective presented in this dissertation is to fabricate titania nanomaterials with
tunable porosities, large surface area, unique morphologies, and enhanced capacities of
adsorption, electron transport, diffusion, and then apply them in photocatalysis and
photovoltaics.
A modified non-hydrolytic sol-gel system with calcium carbonate templating was
developed to create macro/micro/nano porous anatase titanium dioxide. The hydrolysis
rate was lowered by chelating ligands of valeric acid for slow and sufficient precursor
coating. CaCO3 was completely removed via acidification, resulting in titania powders with
a surface area ranging from 197 to 239 m2/g. The templated TiO2 with a surface area of 239
m2/g and pore diameters of 6-109 nm showed a promising 27% photocatalytic
improvement compared to commercial particles, and a 180% increase compared to
template-free TiO2. This increase is attributed to the increased catalyst loading capacity
and active photocatalytic sites.

From hard templates to soft biodegradable natural templates, the rate-controlled solgel method was combined with homogenized micro/nano-fibrillated cellulose (MFCs) with
an average diameter below 50 nm. Cellulose was removed completely by thermal
treatment, and an in-situ coating technique created thin titania films on substrates with a
porous structure. The degradation efficiency of the photocatalytic films was related to film
thickness and to the Ti(IV)-to-cellulose ratio. Photocatalyst on film eliminated the post
separation treatment related to powder catalyst and simplified the purification process.
The sol-gel/MFCs precursor was also coated in-situ as the photoanode for dye-sensitized
solar cells. It was found that the thickness of the anode film was a dominant factor to the
overall performance and efficiency. The 6-layer cell showed a ~40% increase in solar-toelectricity efficiency (1.75%) compared to commercial paste at the same thickness under a
simulated solar light irradiation of 100 mW cm−2 (AM 1.5).
From non-hydrolytic to hydrolytic, a modified liquid phase deposition (LPD) approach
was combined with MFCs. An optimized solvent composition of isopropanol/water ratio of
4 to 1 was found to yield coatings with uniform spherical TiO2 possessing a chain-like
morphology oriented along the axis of the decomposed cellulose fibers. The average rate
constant and degradation percentage were 0.72Âą0.09 min-1, 95% for TiF4-cellulose-4IPA1Water films, which increased by 1.88 times over the film prepared without cellulose
templates due to the beneficial surface area, pore size, and the unique morphology. The
three-dimensional web structure with pseudo one-dimensional sphere-chain could retard
the recombination of photogenerated electron-hole pairs and improve the charge transport.

Copyright by
YAN LI
2016

ACKNOWLEDGEMENTS
I would like to express my heartiest gratitude and sincere appreciation to my advisor,
Prof. Lawrence T. Drzal, for his wisdom and guidance through my whole graduate school
life. Prof. Drzal has been impressing me with his enthusiasm for research, his attitude to life
and work, and his kindness and patience to every student and scholar. I really appreciate
how he cared and understood during my time of injury, surgery, and recovery in 2015. I am
grateful for his help on my career development in 2016. I was lucky to be a part of his 30year devotion to composites research celebration. I would like to thank my entire
committee members: Prof. Richard Lunt, Prof. Thomas Hamman, Prof. Ilsoon Lee, for their
help during my completion of courses, research, and all my questions and ideas. I am very
grateful for your providing of equipment regarding photovoltaic testing and measurements.
It is very kind of Prof. Tim Hogan to join my thesis defense. I would like to thank Professor
McCusker and his student Dr. Lisa Harlow, from Department of Chemistry, for their
generous support as well.
I would like to acknowledge Dr. Per Askland, Ed Downey, Brian Rook, Mike Rich for
teaching me about microscopes, spectroscopies, and all the other equipment in the
Composite Materials and Structures Center. Per deserves special mention for his
knowledge, kindness, patience, humor, and friendship. I would like to say thank you to our
diligent administrators, Karen and Shirley, for everything they accomplished for students,
professors, and the composite center.
Next, I would like to thank all my colleagues, previous or present, Pat, Xian, Jinglei,
Anchita, Deb, Dee, Markus, Keith, Nick, Zeyang, Mariana for your company in the office,

v

your help in labs, and the inspiring conversations we had about research and everything
else. I also appreciate my roommate, and all my friends I got to know at Michigan State
University. Your support always gave me encouragement and motivation to carry on.
In the end, I would like to thank my parents and sister, who have been in China for the
past few years, for their endless love and support. I can never thank them enough for the
powerful words and insights they brought to me. I am lucky to have a family like them. I
appreciate their belief in me, which has and will always be influencing my future life.

vi

TABLE OF CONTENTS

LIST OF TABLES....................................................................................................................................... x
LIST OF FIGURES ....................................................................................................................................xi
KEY TO SYMBOLS............................................................................................................................. xxiii
1 Introduction ......................................................................................................................................1
1.1 Overview ........................................................................................................................................... 1
1.2 Solar Energy, Photocatalysts and Dye-Sensitized Solar Cells ........................................ 2
1.2.1 Solar energy, a clean and safe bet .................................................................................................. 2
1.2.2 Background of solar cells and the dye-sensitized solar cells ............................................. 4
1.2.3 Historical overview of the development of photocatalysis ............................................... 15
1.3 Nano Titania Materials and Synthesis Methods............................................................... 20
1.3.1 Sol-gel method ..................................................................................................................................... 20
1.3.2 Hydrothermal method...................................................................................................................... 23
1.3.3 Liquid phase deposition method ................................................................................................. 25
BIBLIOGRAPHY ....................................................................................................................................... 29

2 Titania from a Hard Template CaCO3 and its Photocatalytic Performance ............. 37
Background ................................................................................................................................... 37
2.1.1 Porous materials and their benefits ........................................................................................... 37
2.1.2 Templating methods for porous TiO2 ........................................................................................ 38
2.1.3 A modified sol-gel method with calcium carbonate template ......................................... 40
Experimental ................................................................................................................................ 45
2.2.1 Materials and equipment ................................................................................................................ 45
2.2.2 Synthesis of porous titania particles with CaCO3 templates............................................. 49
2.2.3 Photocatalytic performance experiment .................................................................................. 50
Results and Discussions ............................................................................................................ 52
2.3.1 Titania particles morphology from different templates .................................................... 52
2.3.2 Photocatalytic performance of porous TiO2 templated with CaCO3.............................. 75
Conclusions ................................................................................................................................... 79
Future Work .................................................................................................................................. 80
2.5.1 Adjusted CaCO3 size and structures to make TiO2 in different shapes ......................... 80
2.5.2 Mixed phase TiO2 ................................................................................................................................ 81
BIBLIOGRAPHY ....................................................................................................................................... 82

3 Templating with Natural Cellulose and Micro-Fibrillated Cellulose (MFC) ............ 88
3.1 Background ................................................................................................................................... 88
3.1.1 Natural templates for inorganic materials............................................................................... 88
3.1.2 Titania materials fabrication through cellulose templating ............................................. 94
3.1.2.1
3.1.2.2

Tubular and fibril titania............................................................................................................................95
Immobilized TiO2 on cellulose.................................................................................................................96

3.2 Experimental ................................................................................................................................ 98
3.2.1 Materials and equipment ................................................................................................................ 98
3.2.1.1
3.2.1.2

Materials: ..........................................................................................................................................................98
Equipment: .......................................................................................................................................................99

vii

3.2.2 Preliminary synthesis of titania hollow ribbons with macro-size cellulose .......... 101
3.2.3 Preparation of homogenized MFCs .......................................................................................... 102
3.3 Results and Discussions ......................................................................................................... 103
3.3.1 Titania hollow ribbons properties ........................................................................................... 103
3.3.2 Homogenized MFCs characterizations and benefits ......................................................... 114
3.4 Conclusions ................................................................................................................................. 123
BIBLIOGRAPHY ..................................................................................................................................... 126

4 Synthesis of Porous Web-structured Titania with Micro/nano-fibrillated
Cellulose as Templates by Sol-Gel Method .............................................................................. 130
4.1 Background ................................................................................................................................. 130
4.1.1 Utilizing nano-scale cellulose for inorganic materials design ...................................... 130
4.1.2 Controlled hydrolysis in a sol-gel process for MFCs templating ................................. 132
4.2 Experimental .............................................................................................................................. 133
4.2.1 Materials and equipment ............................................................................................................. 133
4.2.2 Synthesis of a porous TiO2 network as thin film ................................................................ 138
4.2.3 Measurements of photocatalytic performance of porous titania thin films ........... 140
4.3 Results and Discussions ......................................................................................................... 141
4.3.1 Fabrication of titanium dioxide precursor with MFCs templates ............................... 141
4.3.2 Photocatalysis performance of immobilized titania thin films .................................... 151
4.4 Conclusions and Future Work .............................................................................................. 162
BIBLIOGRAPHY ..................................................................................................................................... 164

5 Synthesis of Porous Titania network with Micro/nano-fibrillated Cellulose as
Templates by Liquid Phase Deposition .................................................................................... 167
5.1 Background ................................................................................................................................. 167
5.1.1 Liquid phase deposition ............................................................................................................... 167
5.2 Experimental .............................................................................................................................. 175
5.2.1 Materials and equipment ............................................................................................................. 175
5.2.2 Synthesis of TiO2 bulk material via TiF4- MFCs LPD method ........................................ 179
5.2.3 Synthesis of porous TiO2 thin films via TiF4- MFCs LPD method ................................ 180
5.2.4 Measurements of photocatalytic performance of porous titania thin films ........... 181
5.2.5 Synthesis of porous titania network via ammonium hexafluorotitanate and boric
acid LPD method ...................................................................................................................................................... 182
5.3 Results and Discussions ......................................................................................................... 183
5.3.1 TiF4- MFCs bulk materials ........................................................................................................... 183
5.3.2 TiF4- MFCs induced TiO2 with tunable sphere chain morphology .............................. 187
5.3.2.1
5.3.2.2

Surface properties of TiO2 films and related materials ............................................................. 188
Surface and elemental properties of TiO2 films ............................................................................ 204

5.3.3 Photocatalysis performance of tunable titania thin films .............................................. 215
5.3.4 Porous titania network via ammonium hexafluorotitanate boric acid method .... 221
5.4 Conclusions and Future Work .............................................................................................. 227
BIBLIOGRAPHY ..................................................................................................................................... 230

6 Nanoporous TiO2 Films Templated with Micro/nano-fibrillated Cellulose and
Applications in Dye-Sensitized Solar Cells .............................................................................. 233
6.1 Background ................................................................................................................................. 233
6.1.1 Nano-structuring of the photoanode of DSSCs .................................................................... 233
6.1.2 Controlled hydrolysis in a sol-gel process for MFCs templating ................................. 234
6.2 Experimental .............................................................................................................................. 235

viii

6.2.1 Materials and equipment ............................................................................................................. 235
6.2.2 Synthesis of porous TiO2 network as thin film .................................................................... 237
6.2.3 DSSC assembly and measurements ......................................................................................... 238
6.3 Results and Discussions ......................................................................................................... 242
6.3.1 Applications of cellulose-templated TiO2 as photoanode in DSSC .............................. 242
6.3.2 Photovoltaic performance test with ttip-vacid films and other films........................ 243
6.4 Conclusions and Future Work .............................................................................................. 248
6.4.1 Conclusions ........................................................................................................................................ 248
6.4.2 Future Work ...................................................................................................................................... 248
BIBLIOGRAPHY ..................................................................................................................................... 250

ix

LIST OF TABLES
Table 2-1 Characteristic peaks of anatase TiO2 obtained with three different CaCO3
templates ......................................................................................................................................................... 67
Table 2-2 Atomic concentrations of C, O, Ca and Ti by spectra of XPS and EDS ....................... 73
Table 2-3 Atomic concentrations of high-resolution fitted peaks by XPS of template-induced
titania ................................................................................................................................................................ 73
Table 2-4 BET surface area and pore size summary.............................................................................. 74
Table 3-1 Summary of dimensions of cellulose structures ................................................................. 93
Table 3-2 EDS atomic percentage of C, O, Ti and other elements in TiO2 materials .............. 113
Table 4-1 Atomic concentration comparison between variable films made from TTIP/Vacid/MFCs precursor ............................................................................................................................... 152
Table 4-2 Degradation and decomposition rate for various films ................................................. 161
Table 5-1 Summary of particle size formed of the I: W= 4:1 sample, with different reaction
time and treatments ................................................................................................................................. 193
Table 5-2 Atomic concentration of related elements in TiF4- MFCs induced films detected
by XPS. ............................................................................................................................................................ 205
Table 5-3 Summary of kinetics of photocatalytic degradation of MO by variable TiF4induced TiO2 films ..................................................................................................................................... 219
Table 6-1 Photocurrent, photovoltage, fill factor and efficiency of TTIP-cellulose cells...... 244
Table 6-2 Photocurrent, photovoltage, fill factor and efficiency of Solaronix TiO2-cellulose
cells .................................................................................................................................................................. 245
Table 6-3 Photocurrent, photovoltage, fill factor and efficiency of DSSCs w/ or w/o compact
layers ............................................................................................................................................................... 246
Table 6-4 DSSCs trials on alternative TiO2 films prepared from aqueous systems ................ 247

x

LIST OF FIGURES
Figure 1-1 Primary energy consumption by fuel (unit of vertical axis: quadrillion Btu)3 For
interpretation of the references to color in this and all other figures, the reader is
referred to the electronic version of this dissertation. .................................................................. 3
Figure 1-2 Electricity generation by fuel (trillion kilowatt hours)3 ................................................... 4
Figure 1-3 Latest solar cell efficiencies research in labs, updated on Mar 7th 2016.7 ................ 5
Figure 1-4 Scheme of a typical dye-sensitized solar cell. ........................................................................ 6
Figure 1-5 Kinetic competitions among multiple reactions in DSSCs under operation.11........ 8
Figure 1-6 Plot of the J-V curve of an operating DSSC. The blue curve is dark current, and
the red curve is light current. ................................................................................................................. 10
Figure 1-7 Structures of polypyridine Ru(II) dyes. (a) N3, (b) Black dye, (c) CYC-B115, 7 .... 11
Figure 1-8 Schematic view of the mechanism of photocatalytic process with metal oxide
semiconductor catalysts ........................................................................................................................... 18
Figure 1-9 Formation mechanism of TiO2 nanotubes using hydrothermal method.63 ........... 24
Figure 1-10 Scheme of the synthesis procedures of silica from H2SiF6.67 (Reproduced with
permission from J. Electrochem. Soc., 150, H205 (2003). Copyright 2003, The
Electrochemical Society) .......................................................................................................................... 25
Figure 2-1 SEM images of (a) a cleaned sea urchin skeletal shell (b) sea urchin shell coated
with gold and annealed.18 ........................................................................................................................ 41
Figure 2-2 SEM images of gold-coated sea urchin skeleton after acid dissolution (a) at lower
magnification, (b) at higher magnification. 1 marks the smooth inner surface, while 2
marks the rough outer surface of the gold structure.18 .............................................................. 41
Figure 2-3 SEM images of titania produced by hydrolysis-templating method on polymer
replica of sea urchin by (a) titanium tetrachloride, (b) titanium ethoxide.14.................... 42
Figure 2-4 Scheme of synthesis procedures of spherical hollow silica .24 .................................... 42
Figure 2-5 TEM images of hollow silica particles using different types of CaCO3: (a) cubic
CaCO3, (b) needle-like CaCO3. 24 ............................................................................................................ 43
Figure 2-6 TEM images of (a) nano-sized calcium carbonate particles, (b) TiO2 with hollow
structure. (c) SEM images of hollow structure of titania.25 ...................................................... 43

xi

Figure 2-7 Scheme of calcium carbonate templating synthesis method for porous titania
particles ........................................................................................................................................................... 45
Figure 2-8 (a): the scheme of the UV lamp system (Ozone was not applied in this work). (b):
the schedule of the pulsed wave. (c): wavelength range of the outputs from the pulsed
xenon UV lamps. (d): UV transmittance of Surasil, quartz, and Pyrex glass
filters[source: Xenon Corp. ...................................................................................................................... 49
Figure 2-9 (a) 1 mL syringe with syringe filter connection. (b) A simple presser for
separation of liquid and solid in MO solution/ TiO2 catalyst mixture. The clear solution
was collected at the bottom of the syringe in a 5 mL vial. ......................................................... 51
Figure 2-10 SEM images of titania products induced with and without CaCO3 W4
templating (5% w.t.). (a) precursor only, no CaCO3 added; (b) with 5% w.t. W4
templated and (c) at higher magnification, (d) sample (b) at cross-sectional view
assisted by FIB. ............................................................................................................................................. 53
Figure 2-11 A closer view of the cross-sectional inside the porous particles, resulted from
5% w.t. W4. ..................................................................................................................................................... 53
Figure 2-12 SEM images of calcium carbonate templates: (a) W4, (b) W3N, (c) OPT. (all
samples were mounted on carbon glue tape on metal stubs) ................................................. 54
Figure 2-13 Comparison between 10% w.t. and 15% w.t. W4 templates in the titania
fabricating process. (a) (c) (e) represents 10% w.t., and (b) (d) (f) are 15% w.t. ......... 55
Figure 2-14 SEM images of films made with a mixture of Ti precursor and W4 template by
dip coating: (a-d) with 5% w.t. W4 CaCO3, (e-f) with 10% w.t. W4 CaCO3. ........................ 57
Figure 2-15 Spin coating working mechanism of a small molecule in solution started with a
static dispense ............................................................................................................................................... 58
Figure 2-16 SEM images of films made with a mixture of Ti precursor and 5% w.t. W4
template at different spin speeds for 45 seconds: (a) 3000 RPM, (b) 1000 RPM, (c) 500
RPM. ................................................................................................................................................................... 59
Figure 2-17 SEM images of (a-b) TiO2 powders and the cross sectional view with 5% w.t.
nano-sized CaCO3 templating; (c-d) TiO2 powders with 10% w.t. nano-sized CaCO3; (ef) films by dip coating with 5% w.t. nano-sized CaCO3. .............................................................. 61
Figure 2-18 SEM images of 15-40 nm CaCO3 particles , (a): as-received, and (b): after
calcination 450 °C for 12 hours. ............................................................................................................ 61
Figure 2-19 Raman spectra of CaCO3 templates of nano-sized type and W4 type. According
to Calcite 282, 713, and 1086 cm-1, and Aragonite 155, 207, 704, 1085 cm-1 ................... 62

xii

Figure 2-20 TGA of the top red curve: W4 CaCO3 particles, middle green curve: nano-sized
CaCO3 after calcination, and bottom blue curve: nano-sized CaCO3 as-received without
any treatment. ............................................................................................................................................... 63
Figure 2-21 TGA of titania particles made with: (top blue)-5% w.t. W4 CaCO3, (middle-red)10% w.t. nano-sized CaCO3, and (bottom green)- 5% w.t. nano-sized CaCO3. (Here the
nano-sized CaCO3 was not pre-treated.)............................................................................................ 63
Figure 2-22 Raman spectra of anatase TiO2 made from the different amount of W4 CaCO3
templates. ........................................................................................................................................................ 65
Figure 2-23 SEM images of titania particles templated from: (a-b) 5% w.t. Opt CaCO3 and
(c-d) 5% w.t. W3N CaCO3 . ....................................................................................................................... 66
Figure 2-24 Raman spectra of anatase TiO2 made with three different CaCO3 templates.
From top to bottom: W4, W3N, Opt, all in 5% w.t. content ....................................................... 67
Figure 2-25 TEM images of 5% w.t. W4 CaCO3 templated TiO2: (a) at low magnification, (b)
at increased magnification, (c) selected-area electron diffraction pattern, (d) individual
particles in an area with crystal size ~ 5 nm, (e) HRTEM image noted with anatase
lattice spacing ~0.35 nm, and 0.36 nm, (f) HRTEM image noted with anatase lattice
spacing ~0.24 nm. ....................................................................................................................................... 69
Figure 2-26 HRTEM images of 5% w.t. W3N CaCO3 templated TiO2: (a) pore formation in a
particle (b) at increased magnification, noted with anatase lattice spacing ~0.35nm,
(c) corresponding FFT pattern of the single crystal area (d) selected-area electron
diffraction pattern ....................................................................................................................................... 70
Figure 2-27 HRTEM images of 5% w.t. Opt CaCO3 templated TiO2: (a) pore formation in a
particle (b) at increased magnification, (c) corresponding FFT pattern of the [101] and
[004] facets, (d) noted with anatase lattice spacing ~0.35 nm ............................................... 71
Figure 2-28 XPS data of TiO2 fabricated with 5% w.t. OPT CaCO3 templates (a) survey
spectra, (b-c) high-resolution XPS spectrum for: Ti 2p and O 1s ........................................... 72
Figure 2-29 Isotherm linear plot, isotherm log plot, isotherm pressure composition and
BET surface area plot via BJH adsorption/desorption of titania made with 5% w.t. Opt
calcium carbonate templates.................................................................................................................. 74
Figure 2-30 Pore size distribution of three titania powders made by calcium carbonate W4,
W3N, and Opt. ............................................................................................................................................... 75
Figure 2-31 Chemical formula of methyl orange ..................................................................................... 76
Figure 2-32 Blank dye without catalysts tested with UV light and the comparison of the
Pyrex lid effects ............................................................................................................................................ 77

xiii

Figure 2-33 Comparison of the photocatalytic capacities through [concentration C/ initial
concentration C0] vs. [time], from top to bottom, regarding to: (green) – no catalyst,
(olive) – pristine TiO2 without templates, (red) – TiO2 from 5% w.t. W4 CaCO3
templating, (black) - TiO2 from 5% w.t. W3N CaCO3 templating, (pink) – commercial
anatase TiO2 nanoparticles, and (blue) - TiO2 from 5% w.t. Opt CaCO3 templating. ...... 78
Figure 2-34 Relationship between lnC and time, based on ln(C) = -kt+ln(C0) (k has unit of
min-1) ............................................................................................................................................................... 79
Figure 3-1 SEM images of nanomaterials made with bio-structure templates: (a) hollow
silica made with bacteria Escherichia coli (KP7600);7 (b) porous titania made with
dandelion pollen grains;8 (c) TiO2 hollow fibers made with cotton wools;1 (d)
hierarchical interwoven titania prepared with eggshell membrane. (reprinted with
permission from reference6) .................................................................................................................. 89
Figure 3-2 Structure of cellulose. (reprinted with permission from reference10) .................... 90
Figure 3-3 Cellulose sources, cell wall, and fibers. The structure level demonstration of
cellulosic fiber. (reprinted with permission from reference 11).............................................. 91
Figure 3-4 TEM images of cellulose nanocrystals derived from (a) tunicate, (b) bacterial, (c)
ramie, and (d) sisal. (reprinted with permission from reference10 ....................................... 92
Figure 3-5 TEM image of MFC gel showing cellulose nanofibrils and nanofibrils bundles
(reprinted with permission from reference10,13............................................................................. 93
Figure 3-6 SEM images of (a) TiO2 fibers after template removing (b) cross section of one
hollow fiber (c) TiO2 nanofibers after milling and ultra-sonication to make a paste (d)
XRD diagram of the fibers.1 ..................................................................................................................... 95
Figure 3-7 (a-d) Low magnification SEM image, photograph, magnified SEM image and
diameter distribution with Energy-dispersive X-ray spectrum of pure bamboo
cellulose fiber. (e-h) Low magnification SEM image, photograph, and magnified SEM
image and diameter distributions of TiO2 product calcined at 500 °C.2 .............................. 96
Figure 3-8 SEM images of (a) pristine cotton fiber (b-d) coated cotton fibers, sol prepared
at 25, 40 and 60°C 3 ..................................................................................................................................... 97
Figure 3-9 Scheme of preliminary TiO2 deposition on cellulose fiber templates...................... 98
Figure 3-10 Scheme of MFCs suspension preparation ........................................................................ 103
Figure 3-11 SEM images of (a) filter paper pieces, and (b) cotton rounds pieces .................. 103
Figure 3-12 SEM images of TiO2 made from deposition of TiF4 on cotton rounds pieces: (a)
lower magnification, and (b) higher magnification of the framed section in (a). .......... 104

xiv

Figure 3-13 SEM images of cotton rounds coated with AHFT and BA without further
calcination: (a) at lower magnification with selected dimensions (b) at higher
magnification with surface features .................................................................................................. 105
Figure 3-14 SEM images of TiO2 made from cotton rounds coated with AHFT and BA after
calcination: (a) hollow ribbon structure of as-prepared TiO2 ; (b) at higher
magnification with the cross section of hollow ribbon wall, and the rough ribbon
surface. ........................................................................................................................................................... 105
Figure 3-15 SEM images of TiO2 made from filter paper coated with TiF4 after calcination:
(a) fractures of titania (b) surface of the as-prepared TiO2 at higher magnification. .. 106
Figure 3-16 SEM images of filter paper coated with AHFT and BA: (a-b), before calcination,
at low and high magnifications highlighting the fiber layout and the nucleation of
amorphous TiO2 on the surface of the preliminary coating. (c-d), after calcination, at
low and high magnification demonstrating long hollow fibers and the cross section of
two TiO2 tubes............................................................................................................................................. 107
Figure 3-17 TGA of (a) cotton rounds, (b) filter paper ....................................................................... 111
Figure 3-18 Thermal analysis through TGA on non-calcined soaked cellulose substrates
including cotton round fibers and filter paper. ............................................................................ 112
Figure 3-19 EDS spectrum of AHFT-BA coated cotton round fibers, after calcination ......... 113
Figure 3-20 Raman spectra of TiO2 templated from AHFT and TiF4 systems on both cotton
round fibers and filter paper ................................................................................................................ 114
Figure 3-21 Cellulose fibers treatment effects: (a) after the kitchen blender and a high
circumferential speed disperser treatments, (b) after 30 passes on the homogenizer, (c)
after 60 passes on the homogenizer.................................................................................................. 116
Figure 3-22 TGA of MFCs made by Celish KY100G (90 wt% in water) purchased from Daicel,
with water evaporated by 12-hour vacuum oven heating. TGA was conducted in air
through a high-resolution ramp program at a heating rate of 25 °C/min. ....................... 117
Figure 3-23 TGA of non-homogenized Celish KY100G (90% w.t. in water) purchased from
Daicel, without further treatment. TGA was conducted in air through a high-resolution
ramp program at a heating rate of 25 °C/min............................................................................... 118
Figure 3-24 TGA of dried MFCs purchased from the University of Maine process
development center with a water content of 97% w.t. To eliminate the water
evaporation step during thermal analysis, sample was dried in vacuum oven overnight
at 80C before the test. TGA was conducted in air through a high-resolution ramp
program at a heating rate of 25 °C/min ........................................................................................... 119

xv

Figure 3-25 TGA of dried CNCs purchased from the University of Maine process
development center with a water content of 93.8% w.t. To eliminate the water
evaporation step during thermal analysis, sample was dried in vacuum oven overnight
at 80C before the test. TGA was conducted in air through a high-resolution ramp
program at a heating rate of 25 °C/min. .......................................................................................... 120
Figure 3-26 FTIR spectra of MFCs................................................................................................................ 121
Figure 3-27 (a) XPS spectrum of MFCs ; (b) and deconvolution of its C1s peak...................... 123
Figure 4-1 Scheme of applying MFCs as a template for porous TiO2 network synthesis .... 133
Figure 4-2 (a): the scheme of the UV lamp system (Ozone was not applied in this work).
(b): the time schedule of the pulsed wave. (c): wavelength range of the outputs from
the pulsed xenon UV lamps. (d): UV transmittance of Surasil, quartz, and Pyrex glass
filters[source: Xenon Corp. .................................................................................................................... 138
Figure 4-3 Scheme of MFCs-templating slow sol-gel method .......................................................... 139
Figure 4-4 Scheme of photocatalytic degradation experiment ....................................................... 141
Figure 4-5 (a) Homogenized MFCs. (b) titanium precursor without cellulose, calcined at
450 °C for 2 h. (c) titanium precursor coated cellulose fibers, before calcination. (d)
titanium precursor with cellulose (ratio is 1: 1), calcined at 450 °C for 2 h. ................... 142
Figure 4-6 (a-c) titanium precursor with cellulose, calcined with a two-stage method at 250
°C for 1h and 450 °C for 1.5 h at increasing magnifications. ................................................... 144
Figure 4-7 TGA results of different Ti: cellulose ratios: 1 (green) Ti: cellulose = 1: 1, reacted
at room temperature. 2 (blue) Ti: cellulose = 1: 1, reacted at 50 °C. 3 (brown) Ti:
cellulose = 0.5: 1, reacted at room temperature. 4 (pink) Ti: cellulose = 2: 1, reacted at
room temperature. All reactions lasted 14 hours with same volume of MFC
suspensions at 0.5mg/mL concentration. ....................................................................................... 145
Figure 4-8 Raman spectra of blue: calcined precursor-cellulose; Red: the blank glass
substrate; Black: uncalcined precursor-cellulose........................................................................ 146
Figure 4-9 X-ray photoelectron spectra of TiO2 thin film after calcination. .............................. 147
Figure 4-10 TGA results of (a) PEG and (b) PVDF through thermal degradation in air ....... 149
Figure 4-11 Calcined Ti-MFCs thin films with different binders. With 10% w.t. PVDF, spin
coated: (a)1 layer, (b)4 layers; doctor blade coated (c) 1 layer. With 10% w.t. PEG:
(c)spin coated, 1 layer (d)doctor bade coated, 1 layer.............................................................. 150
Figure 4-12 (a-b) Calcined Ti-MFCs thin films with ~30% w.t. PVDF, grain-surface riceshaped TiO2 . (c) An anatase tetragonal {101} bi-pyramid. .................................................... 151

xvi

Figure 4-13 XPS spectra of TiO2 thin film after calcination, made from TTIP/V-acid/MFCs,
with Ti: cellulose = 2:1. (a) XPS spectrum with element labeling. (b-d) High-resolution
data fitting. .................................................................................................................................................... 153
Figure 4-14 XPS spectra of TiO2 thin film after calcination, made from TTIP/V-acid/MFCs,
with Ti: cellulose = 4:1. (a) XPS spectrum with element labeling. (b-d) High-resolution
data fitting. .................................................................................................................................................... 154
Figure 4-15 XPS spectra of TiO2 thin film after calcination, made from TTIP/V-acid/MFCs,
with Ti: cellulose = 8:1. (a) XPS spectrum with element labeling. (b-d) High-resolution
data fitting. .................................................................................................................................................... 155
Figure 4-16 TEM images of anatase TiO2 film generated by modified sol-gel method with
TTIP/V-acid/MFCs, with Ti: MFCs = 2:1. (a) and (c) porous structure, at lower
magnifications (b) and (d) HRTEM image noted with anatase lattice spacing ~0.352
nm of the plane (101). ............................................................................................................................. 156
Figure 4-17 TEM images of anatase TiO2 film generated by modified sol-gel method with
TTIP/V-acid/MFCs, with Ti: MFCs = 4:1. (a) showed the porous nanoparticles, (b) and
(c) HRTEM image noted with anatase lattice spacing ~0.352 nm of plane (101) and
~0.19 nm of plane (200). (d) SAED diffraction pattern shows the polycrystalline rings.
........................................................................................................................................................................... 157
Figure 4-18 TEM images of anatase TiO2 film generated by modified sol-gel method with
TTIP/V-acid/MFCs, with Ti: MFCs = 8:1. (a-b) showed the porous nanoparticles, and
(c) HRTEM image noted with anatase lattice spacing ~0.352 nm of plane (101). (d)
SAED diffraction pattern shows the polycrystalline rings....................................................... 158
Figure 4-19 Reflectance change over time of UV-illuminated TiO2 film with 1×10-5 M MO
solution adsorbed. Titania film was made with TTIP/V-acid/MFCs and Ti: cellulose = 2:
1. (a) Reflectance change from clean TiO2, MO solution adsorbed (after dye), and
treated with UV lamp for a total time of 1 min, 3 min, 5 min, and 7 min. (b) Reflectance
difference over time, from dye adsorption, to photocatalytic degradation. .................... 160
Figure 4-20 For the film made with TTIP/V-acid/MFCs, and two layer coating, reflectance
difference change at wavelength 395 nm v.s. time. (a-c): three sets of data, where Ti:
cellulose = 2: 1 (black), Ti: cellulose = 4: 1 (red) and Ti: cellulose = 8: 1 (blue) ............ 160
Figure 4-21 For the film made with TTIP/V-acid/MFCs, and one layer coating, reflectance
difference change at wavelength 395 nm v.s. time. : three sets of data with different
volume of coating precursor. ................................................................................................................ 161
Figure 5-1 (A-C) SEM and AFM images of different LPD bath treatment times. (A) 3 h, (B) 6
h, (C) 24 h. (D) Growth of LPD titania layer thickness. 4........................................................... 169
Figure 5-2 Influence of (a) cycles (b) time (c) molar ratios to the thickness of titania films. 5
........................................................................................................................................................................... 170
xvii

Figure 5-3 SEM images of annealed titania films made by LPD AHFT and BA with various
pH values. 7 ................................................................................................................................................... 170
Figure 5-4 The relationship between Ti concentration, and pH level, with the deposition
scale on the substrate.8 ........................................................................................................................... 171
Figure 5-5 Relationship between thickness of the deposited film and reaction time.8 ........ 172
Figure 5-6 (a): the scheme of the UV lamp system (Ozone was not applied in this work). (b):
the schedule of the pulsed wave. (c): wavelength range of the outputs from the pulsed
xenon UV lamps. (d): UV transmittance of Surasil, quartz, and Pyrex glass
filters[source: Xenon Corp. .................................................................................................................... 179
Figure 5-7 Scheme of TiF4- MFCs precursor preparation .................................................................. 180
Figure 5-8 Scheme of MFCs-templating TiF4 LPD thin films on glass........................................... 181
Figure 5-9 Scheme of photocatalytic degradation experiment ....................................................... 182
Figure 5-10 Scheme of AHFT-BA/MFCs synthesis................................................................................ 183
Figure 5-11 0.04 M TiF4 in 20 mL IPA/water suspension, calcined as bulk material. (a), (b),
(c) are the same area with increasing magnifications. .............................................................. 184
Figure 5-12 0.04 M TiF4 in 20 mL IPA/water suspension, calcined as bulk material. pH
adjusted at 4 during deposition reaction. ....................................................................................... 184
Figure 5-13 0.04 M TiF4 in 20 mL IPA/water suspension, calcined as bulk material. FIB-cut
cross section (a), (b) are taken from the same area with increasing magnifications. . 185
Figure 5-14 0.04 M TiF4 in 20 mL water, calcined as bulk material. (a) and (b) are surface
features at same area with increasing magnifications. (c-d) FIB-cut cross-sectional
view at increasing magnifications. ..................................................................................................... 186
Figure 5-15 XRD patterns of (a) TiF4 and water precursor, calcined bulk material, (b) TiF4
and MFCs/Water, calcined bulk material. ....................................................................................... 186
Figure 5-16 SEM images of 4 samples without and with MFCs templates, at different
isopropanol to water ( I: W) ratios. All reactions last for 6 hours, and films were coated
and calcined. (a) Without MFC templates, 0.04 M TiF4 in water. The inset image scale
bar reads 5 um. (b-d) With MFC templates, at I: W ratios 0:1, 1:1, and 4:1 respectively.
The scale bar of the inset image of (d) reads 1 um. .................................................................... 189
Figure 5-17 0.04 M TiF4 reacted in H2O for (a) 6 hours, (b) 15 hours, and (c) 40 hours. ... 190
Figure 5-18 0.04 M TiF4 reacted in 5 mg/mL MFC water suspension for (a) 6 hours, (b) 15
hours, and (c) 40 hours. .......................................................................................................................... 191

xviii

Figure 5-19 0.04 M TiF4 reacted in 5 mg/mL MFC water and isopropanol suspension
mixture at I:W = 4:1 ratio for (a) 2 hours, (b) 4 hours, (c) 6 hours, (d) 15 hours, and (e)
40 hours. ........................................................................................................................................................ 192
Figure 5-20 The three-dimensional in-depth view of titania film made with the TiF4- with
MFCs in I: W=4:1 suspension. (a) The reverse side of the film, from the inset image of
(a). (b) The cross-sectional view of the film. ................................................................................. 194
Figure 5-21 0.04 M TiF4 reacted in 5 mg/mL MFC isopropanol suspension at I: W = 1:0 ratio
for 6 hours. Calcined films morphologies. ...................................................................................... 194
Figure 5-22 0.04 M TiF4 reacted in 5 mg/mL MFC aqueous suspension at I: W = 0: 1 ratio
for 6 hours, coated on glass substrate, and dried in vacuum oven, without calcination
applied. ........................................................................................................................................................... 195
Figure 5-23 0.04 M TiF4 reacted in 5 mg/mL MFC aqueous suspension at I: W = 1: 1 ratio
for 6 hours, coated on glass substrate, and dried in vacuum oven, without calcination
applied. ........................................................................................................................................................... 195
Figure 5-24 0.04 M TiF4 reacted in 5 mg/mL MFC aqueous suspension at I: W = 4: 1 ratio
for 6 hours, coated on glass substrate, and dried in vacuum oven, without calcination
applied. ........................................................................................................................................................... 196
Figure 5-25 SEM image of TiO2 hollow fibers formed after film calcination from sample
TiF4 with MFCs in aqueous suspension............................................................................................ 196
Figure 5-26 TiF4 reactions in IPA and water mixture at I:W ratio of 4:1 for (a) 2 hours, (b) 4
hours, (c) 6 hours....................................................................................................................................... 198
Figure 5-27 TiF4 reactions in IPA and water mixture at a I: W ratio of 1:1 for (a) 2 hours, (b)
4 hours, (c) 6 hours. .................................................................................................................................. 199
Figure 5-28 TEM images of anatase TiO2 film generated by TiF4 liquid phase deposition,
with a solvent ratio of IPA/H2O =0:1. (a) and (c) porous structure, at low and high
magnification. (b) SAED diffraction pattern and (d) HRTEM image noted with anatase
lattice spacing ~0.352 nm of facet (101) and 0.148 nm of (213)........................................ 201
Figure 5-29 TEM images of anatase TiO2 film generated by TiF4 liquid phase deposition,
with a solvent ratio of IPA/H2O =0:1. (a-b) tangled tubular-shaped TiO2, at low and
high magnification. (b) another TiO2 tube showing the hollow inner and outside wall,
lattice spacing is 0.352 nm in the inset image (d) corresponding HRTEM image of (c)
highlighting the anatase lattice spacing of facet (101) with spacing 0.355 nm and (004)
with spacing 0.24 nm. .............................................................................................................................. 202
Figure 5-30 TEM images of anatase TiO2 film generated by TiF4 liquid phase deposition,
with a solvent ratio of IPA/H2O =4:1. (a-b) spherical TiO2, at low and high
magnification. (c) interface between two spherical particles, and the SAED pattern

xix

(inset) labeled (d) corresponding HRTEM image of (c) highlighting the anatase lattice
spacing of facet (101) and (004), and the lattice angle between is 119.6°....................... 203
Figure 5-31 EDS information of a calcined TiO2 film (with MFCs in suspension of I: W =1:1)
on a glass substrate. The inset tables explained the atomic composition of oxygen and
titanium, and the background elements from the glass substrate. ...................................... 204
Figure 5-32 XPS spectra of thin titania films made with TiF4 precursors with MFCs, at
different solvent compositions: (a) IPA: water= 0:1, (b) IPA: water=1:1, (c) IPA: water=
4:1, (d) IPA: water= 1:0. .......................................................................................................................... 206
Figure 5-33 XPS convolution results of thin titania dried films without calcination, made
with TiF4 precursors with MFCs at IPA: water= 4:1. (a) C 1s, (b) O 1s, (c) Ti 2p, (d) F 1s.
........................................................................................................................................................................... 207
Figure 5-34 XPS convolution results of thin titania films after calcination, made with TiF4
precursors with MFCs at IPA: water= 4:1. (a) C 1s, (b) O 1s, (c) Ti 2p, (d) F 1s. ............ 208
Figure 5-35 Raman spectra of two thin titania films, made from TiF4 -MFCs precursor with
different solvent compositions: IPA: water= 0:1 and IPA: water= 4:1. .............................. 209
Figure 5-36 XRD of titania films made with TiF4/ MFCs templates, with different IPA: water
solvent compositions ............................................................................................................................... 210
Figure 5-37 TGA of dried MFCs. To eliminate the water evaporation step during thermal
analysis, sample was dried in vacuum oven overnight at 80 °C before the test. TGA was
conducted in air through a high resolution ramp program at a heating rate of 25
°C/min. ........................................................................................................................................................... 211
Figure 5-38 TGA of uncalcined TiF4 precursor with template MFCs in a solvent of I: W at
ratio 1: 1. To eliminate the solvent evaporation step during thermal analysis, sample
was dried in vacuum oven overnight at 80C before the test. TGA was conducted in air
through a high-resolution ramp program at a heating rate of 25 °C/min ........................ 212
Figure 5-39 TGA of uncalcined TiF4 precursor with template MFCs in a solvent of I: W at
ratio 0: 1. To eliminate the solvent evaporation step during thermal analysis, the
sample was dried in vacuum oven overnight at 80C before the test. TGA was conducted
in air through a high-resolution ramp program at a heating rate of 25 °C/min. ........... 213
Figure 5-40 TGA of calcined TiF4 precursor with template MFCs in two different solvent
systems: (1) top curve
was prepared in I: W at ratio 0: 1; (2) bottom curve
with ☐ was prepared in I: W at ratio 1: 1. TGA was conducted in air through a highresolution ramp program at a heating rate of 15 °C/min. ....................................................... 214
Figure 5-41 Concentration change of the MO solution on glass substrate measured by UVVis absorbance, over a 20-minute UV lamp irradiance, without any catalyst. ............... 215

xx

Figure 5-42 Reflectance measurement of TiO2 film made from TiF4/ MFCs illuminated by
UV lamp over time. (a) the reflectance of initial TiO2, UV treated by 1 min, 2 min, and 3
min. (b) Subtracted from initial reflectance, v.s wavelength. ................................................. 216
Figure 5-43 Reflectance change over time of UV-illuminated TiO2 film with 1×10-5 M MO
solution adsorbed. Titania film was made with TiF4/ MFCs and IPA: water = 1: 1. (a)
Reflectance change from clean TiO2, MO solution adsorbed (after dye), and treated
with UV lamp for a total time of 1 min, 2 min, 4 min, 6 min, and 9 min. Inset image
showed details of the same curves. (b) Reflectance difference over time, from dye
adsorption, to photocatalytic degradation. .................................................................................... 217
Figure 5-44 For film made with TiF4/ MFCs and IPA: water = 1: 1. (a) Reflectance
difference change at wavelength 395 nm v.s. time. (b) pseudo-first-order kinetic model
of ln(reflectance) v.s. time. .................................................................................................................... 217
Figure 5-45 Reflectance change over time of UV-illuminated TiO2 film with 1×10-5 M MO
solution adsorbed. Titania film was made with TiF4/ MFCs and IPA: water = 4: 1. (a)
Reflectance change from clean TiO2, MO solution adsorbed (after dye), and treated
with UV lamp for a total time of 1 min, 3 min, 5 min, and 7 min. Inset image showed
details of the same curves. (b) Reflectance difference over time, from dye adsorption,
to photocatalytic degradation. ............................................................................................................. 218
Figure 5-46 For film made with TiF4/ MFCs and IPA: water = 4: 1. (a) Reflectance
difference change at wavelength 395 nm vs. time. (b) pseudo-first-order kinetic model
of ln(reflectance) vs. time. ..................................................................................................................... 219
Figure 5-47 Variable films (a) reflectance change (b) kinetics with linear modeling and
equations. ...................................................................................................................................................... 220
Figure 5-48 Comparison between “before calcination” and after calcination of AHFT and BA
with MFCs composite. (a, c, e, g) are before calcination while (b, d, f, h) are calcined.
(a,b) has a molar ratio of BA: AHFT= 0.15:0.05. Similarly, (c,d)’s BA: AHFT= 0.30:0.05;
(e,f)’s BA: AHFT= 0.30:0.30; (g,h)’ BA: AHFT= 0.15:0.30. ........................................................ 222
Figure 5-49 Calcined titania film from AHFT and BA with MFCs precursor, and the molar
ratio of BA: AHFT is 0.15:0.05. ............................................................................................................. 223
Figure 5-50 BA: AHFT= 0.15:0.15 films, calcined. (a-b) low and high magnification SEM
images of 24h-reaction at room temperature; (c-d) low and high magnification SEM
images of 24 h-reaction at 50 °C. ........................................................................................................ 224
Figure 5-51 SEM images of titania films prepared from BA-AHFT/MFCs precursor with BA:
AHFT = 0.15M: 0.15M. From (a) to (c), PEG weight percentages were 1%, 5%, and
10%. ................................................................................................................................................................. 225
Figure 5-52 SEM images of titania films prepared from BA-AHFT/MFCs precursor with low
concentrations. (a-b) BA: AHFT=0.075 M: 0.075 M, (c-d) BA: AHFT=0.025 M: 0.025 M,
xxi

(e-f) 0.025 M: 0.008 M. (a, c, e) was recorded without addition of binder, while (b, d, f)
was taken with addition of PEG........................................................................................................... 225
Figure 5-53 Raman spectrum of titania films made from AHFT-BA/MFCs precursor: (a)
BA:AHFT = 0.15: 0.05 (b) BA:AHFT = 0.15:0.15 ........................................................................... 226
Figure 5-54 TEM images of the titania films made from AHFT-BA/MFCs precursor, and the
ratio of BA: AHFT = 0.15M : 0.05M. From (a) to (d) magnification increased.
Measurement of lattice (d) shows the lattice space of 0.35 nm and 0.24 nm,
respectively referred to plane (101) and (004). .......................................................................... 227
Figure 6-1 Scheme of applying MFCs as a template for porous TiO2 network synthesis .... 235
Figure 6-2 Scheme of MFCs-templating slow sol-gel method .......................................................... 238
Figure 6-3 Four-layer TTIP-cellulose precursor, calcined, for DSSCs. SEM images: (a) top
view, (b) cross section view. TEM images: (c-d) at higher magnifications and
resolutions, lattice plane measured 0.3 nm (e) selected area diffraction pattern. ....... 243
Figure 6-4 J-V curves of TTIP-cellulose deposited DSSCs in dark and luminance .................. 244
Figure 6-5 J-V curves of commercial Solaronix paste deposited photoanode for DSSCs in
dark and luminance .................................................................................................................................. 244
Figure 6-6 J-V curves of different DSSCs in dark and luminance, with compact block layer
deposited or without. ............................................................................................................................... 246
Figure 6-7 J-V curves of boric-cellu and TiF4-cellu-induced DSSCs under luminance ......... 247
Figure 6-8 Cell performance over the time span. This is a TTIP-cellulose based 6-layer
photoanode cell. ......................................................................................................................................... 248

xxii

KEY TO SYMBOLS

TiO2

Titanium Dioxide

DSSC

Dye-Sensitized Solar Cell

MFC

Micro/nano-fibrillated Cellulose

AM 1.5G

Air Mass 1.5 Global

LPD

Liquid Phase Deposition

TCO

Transparent Conductive Oxide

FTO

Fluorine-doped Tin Oxide

ITO

Indium-doped Tin Oxide

Jsc

Photocurrent density measured at short circuit

Voc

Photovoltage at open circuit

Is

Incident light intensity

FF

Fill factor

Ru(II)

Ruthium

N3

Tris(2,2’-bipyridyl-4,4’-carboxylate) ruthium (II)

IPCE

Incident photo-to-current efficiency

OSE

Organic solvent electrolyte

ILE

Ionic liquid electrolyte

TBP

4-tert-butylpyridine

NMBI

N-methylbenzimidazole

I-/I3-

Iodine/triiodide

CuI

Copper Iodide

CuSCN

Copper Thiocyanate

xxiii

CoS

Cobalt Sulfide

Eg

Band gap energy

Nb2O5

Niobium Pentoxide

SnO2

Tin Oxide

TEOA

Triethanolamine

TTIP

Tetraisopropoxide

ED

Ethylenediamine

AHFT

Ammonium Hexafluorotitanate

BA

Boric Acid

TBOT

Tetrabutoxytitanium

SAED

Selected-area electron diffraction

FFT

Fast Frourier Transfer

VA

Valeric acid

MO

Methyl Orange

EDTA

Ethylenediaminetetraacetic acid

AGU

Anhydroglucose unit

CNC

Cellulose nanocrystal

A

Frequency Factor

Ea

Activation Energy

R

Universal Constant

K(T)

T-dependent reaction rate constant

Α

Degree of conversion

IPA

Isopropanol

xxiv

PEG

Polyethylene glycol

PVDF

Polyvinylidene fluoride

xxv

1 Introduction
1.1 Overview
The research objective presented within this dissertation is to fabricate nano titanium
dioxide with unique morphologies, porosities, and other properties through various
templating methods, and apply them as photocatalysis and photovoltaics.
In this introductory chapter, the motivation of this research will focus on the solar
energy, a renewable source for industry and daily life. Two promising applications as
photocatalysts to remove pollutants and dye-sensitized solar cells (DSSCs) for electricity
generation and the advantages of titanium dioxide nanomaterials for above applications
will be introduced. Then various synthetic routes for synthesizing titanium dioxide will be
presented such as sol-gel chemistry, hydrothermal method, liquid phase deposition,
templating methods and so on.
The second chapter will discuss the controlled hydrolysis of a sol-gel process utilizing
valeric acid and titanium tetraisopropoxide as the base precursor. It will describe the use of
a simple but effective hard-template calcium carbonate as a sacrificial structure to create
macro/micro/nano porous titania particles, and their applications as powder
photocatalysts to degrade the organic dye methyl orange.
The third chapter will focus on the introduction of the cost-effective, biodegradable
cellulose, started with brief background knowledge about this natural-existing polymer,
followed by experiments proving the successful templating effects for titania micro hollow
fiber fabrication. Then micro/nano-fibrillated cellulose (MFCs) fibers will be discussed
regarding the scale effect and procedures to prepare homogenized MFCs in our lab.
1

Chapter 4 will describe a novel combination between MFCs and non-hydrolytic
precursor for a controlled hydrolysis of a sol-gel process. The function of MFCs is evaluated
through various characterizations on unannealed and annealed samples. The addition of
binders such as polyethylene glycol and polyvinylidene fluoride is compared. Films
adjusted with different Ti(IV) and cellulose ratios show different photocatalytic capacities
when applied as photocatalytic films to degrade organic dye methyl orange.
Chapter 5 will provide more trials using MFCs as the templates but in hydrolytic
systems. Liquid phase deposition (LPD) methods will be discussed in detail with two
modified reaction systems: (1) titanium tetrafluoride method, in which the manipulation of
solvent polarity brings out controllable morphology and thus different photocatalytic
capacities; (2) ammonium hexafluorotitanate and boric acid method, where both bulk
materials and thin films will be discussed through characterization approaches.
The sixth chapter will turn to dye-sensitized solar cells, starting with an introduction of
previous work review on cellulose templated TiO2 as photoanodes in DSSCs. The detailed
preparation of electrodes, electrolyte, dye sensitizer and assembly will be described next.
Photovoltaic performances of the cells made with different TiO2 films will be compared.

1.2 Solar Energy, Photocatalysts and Dye-Sensitized Solar Cells
1.2.1 Solar energy, a clean and safe bet
The report of Intergovernmental Panel on Climate Change has pointed out that global
warming brought on by increasing greenhouse gases has had a detrimental impact on
oceans and the atmosphere.1 The huge demand for fossil fuels and the risks of
environmental crisis have spurred an interest in renewable energies. The world energy

2

consumption is estimated to grow by 56% between 2010 and 2040.2 As Figure 1-1 shows,
total primary energy consumption grows by 8.9% from 97.1 quadrillion Btu in 2013 to
105.7 quadrillion Btu in 2040.3 The consumption of marketed renewable energy increases
from 9.0 quadrillion Btu in 2013 to 12.5 quadrillion Btu in 2040, with most of the growth in
electric power sector.3

Figure 1-1 Primary energy consumption by fuel (unit of vertical axis: quadrillion Btu)3 For
interpretation of the references to color in this and all other figures, the reader is referred to the
electronic version of this dissertation.

Figure 1-2 illustrates the electricity generation amount from 2000 to 2040, where
petroleum, coal, and nuclear show a decrease while renewable energies and natural gas
increase. In the long term, natural gas fuels more than 60% of the new generation, and
growth from renewable energy supplies most of the remainder. Among major types of
renewable energy systems, such as solar energy, wind energy, geothermal, biomass,
hydroelectric, nuclear and ocean,4 solar is capable of converting the abundant, free, clean
and cost-effective solar energy directly into electricity - the energy is quantitatively large
enough to meet the growing demand.
3

Figure 1-2 Electricity generation by fuel (trillion kilowatt hours)3

Although PV only accounts for 0.04% of the overall energy production, there has been
a rapid increase of 33%/year since 1997 and an anticipated annual increase of 30% in the
next decades.5 Based on the newest data,3 wind and solar generation account for almost 67%
of the increase in total renewable generation, and solar photovoltaic is the fastest-growing
energy source for renewable generation. Solar stands out with a 6.8%/year average annual
growing rate when compared to other renewable energy resources for electric generation
such as wind (2.4%/year), biomass (3.1%/year), and geothermal (5.5%/year).
1.2.2 Background of solar cells and the dye-sensitized solar cells
Solar cell is an electric device converting phonons directly into electricity through the
photovoltaic effect. The first solar cell was developed by Bell Labs in 1950s6 by Daryl
Chapin et al. with an efficiency of 6%, utilizing silicon as the light absorber. During the next
few decades, solar cells have been tremendously researched and silicon-based solar cells

4

become the dominant products in the market with high efficiency 27.6% according to
National Renewal Energy Laboratory report7 (shown in Figure 1-3). As a large and growing
family, solar cells also have a variety of members, including mono/polycrystalline silicon
solar cells, thin-film solar cells such as amorphous Si cells, polycrystalline CdTe cells,
chalcopyrite cells, III-V semiconductor cells, and highly efficient multijunction
concentrators.8 Similar with silicon purifying requirements, the high processing cost and
raw material scarcity hinder the development of these solar cells.9 This dissertation will
focus on one of the third-generation solar cells, dye-sensitized solar cells (DSSCs) as a
potential competitor to traditional silicon solar cells.

Figure 1-3 Latest solar cell efficiencies research in labs, updated on Mar 7th 2016.7

The first DSSC was created in 1991, with an efficiency of 7% by O’Regan and Grätzel.10
The breakthrough relied on the application of wide band gap nanocrystalline
semiconductor network with large surface area for dye sensitizer adsorption and light
harvest. Unlike p-n junction solar cells, DSSC separates the photon harvesting and charge

5

transport processes.10 DSSCs stand out for the following reasons: (i) no significant
requirement on the grade of the semiconductor as silicon cell does, resulting in lower cost;
(ii) the materials used are environmentally friendly and present in adequate quantities;
(iii) has stable performance under varying thermal conditions.
A dye-sensitized solar cell is composed of four main parts: a nanostructured
semiconductor photoanode, dye sensitizers, electrolytes with redox couples, and a catalystcoated counter electrode, forming a sandwich-like cell assembly. Both of the electrodes use
transparent conductive oxides (TCO) as substrates, usually fluorine-doped tin oxide (FTO),
or indium doped tin oxide (ITO) films. Nanoporous titania nanocrystalline (TiO2),
ruthenium-based metal-ligand compounds (N3 or N791), iodine/triiodide (I-/I3-) in
organic solvents, and platinum-coated FTO represent one typical set of components for
DSSCs.

Figure 1-4 Scheme of a typical dye-sensitized solar cell.

6

The DSSCs operate as follows (Figure 1-4): The dye sensitizer, which is adsorbed as a
molecular monolayer, will be excited from state S0 to S* by illumination, ejecting one
electron very quickly into the conduction band of TiO2. The electron will then diffuse within
the domain of the porous TiO2 layer and reach the FTO film. Simultaneously, the oxidized
dye S+ will be recovered to the original state S0 by the redox couple I-/I3- in the electrolyte.
Once an external load is applied, the electron will travel through the circuit and arrive at
the counter electrode surface, where it will react with the oxidized electrolyte and refresh
the redox couple. Thus, the whole cycle of operation is fulfilled.
From photon to electron, diffusion to redox, the working mechanism of DSSCs is
kinetic-based with competitions between steps. Possible back reactions are: (a) injected
electrons recombine with the oxidized dye, (b) injected electrons react with the oxidized
electrolyte, and (c) relaxation to the ground state.9 The structural optimization of the
photoanode is to offer opportunities for improvement in not only the quantity of dye,
transport of electrons, and diffusion of electrolytes but also the suppression of
recombinations.

7

Figure 1-5 Kinetic competitions among multiple reactions in DSSCs under operation.11

The processes mentioned above are kinetic-based, indicating that the competition
between different steps and reactions will determine the performance of the whole cell. As
shown in Figure 1-5, k1~k6 stand for the rate constants for separate reaction steps in a full
cycle of operation. It is clearly demonstrated that the injection of the photoexcited electron
is ultrafast, proceeding on a femtosecond time scale10 (k2 =1010-1012s-1), while the
relaxation to ground state processes on a nanosecond time scale (k1=108-1010s-1). Three
possible back reactions are undergoing here: injected electron recombines with oxidized
dye (k3), injected electron reacts with oxidized electrolyte (k6), and the relaxation to the
ground state (k1). Focusing on the electron generated by the very first step photoexcitation,
we may notice that either relaxation to the ground state or injection to the semiconductor
will occur. The fact is that k2 is almost 100 times larger than k1, thus typically the electron
prefers to undergo injection. Similar is the situation when the competition between
injection and combination with the electrolyte is considered. In addition, since the diffusion
rate (k5) is much smaller than the back reaction rate (k3 and k6), even the recovery of dye

8

by the mediator in electrolytes (k4) is faster than the recombination between electron and
dye (k3). Although minimal, the back reactions still occur and minimizing them remains an
important problem to be solved.11
The overall efficiency of a solar cell, Ρ, is represented by
𝜂=

𝐽𝑠𝑐 ∙ 𝑉𝑜𝑐 ∙ 𝐹𝐹
𝐼𝑠

𝐽𝑠𝑐 , 𝑉𝑜𝑐 and 𝐼𝑠 are the photocurrent density measured at short circuit current,
photovoltage at open circuit and incident light intensity respectively, while FF is the “fill
factor” which is defined as 𝐹𝐹 =

𝑃𝑚𝑎𝑥
𝐽𝑠𝑐 𝑉𝑜𝑐

. As Figure 1-6 shows, the area underneath the black

dashed line is the product of 𝐽𝑠𝑐 and 𝑉𝑜𝑐 , and the area beneath the red dashed line is the
maximum power 𝑃𝑚𝑎𝑥 . FF ranges from 0 to 1, and the maximum electrical power is 𝑃𝑚𝑎𝑥 =
(𝐽𝑠𝑐 𝑉𝑜𝑐 )𝑚𝑎𝑥 .
The key parameters to enhancing the photo-to-electricity conversion efficiency are Jsc,
Voc, and FF. Theoretically, the photocurrent is controlled by the sensitizer’s light absorption
and electron injection processes, while photovoltage is determined from the difference
between the semiconductor Fermi level and the Nernst potential of the redox couple.
However, recombination reactions could neutralize both Jsc and Voc. As for the FF, its
influence occurs in a more complex way.

9

Figure 1-6 Plot of the J-V curve of an operating DSSC. The blue curve is dark current, and the red
curve is light current.

Along with the discoveries of more powerful dye sensitizers,12

13

redox shuttles,14

semiconductor compositions15, and electrolyte carriers,16 the efficiency of DSSCs has
increased from 7 %10 to 11.5 %17. Recent latest DSSCs have reached an efficiency of 15%
by

using

an

inorganic-organic

composite

structure

of

glass/FTO/TiO2

/CH3NH3PbI3/HTM/Au for improved reproducibility and enhanced photo-to-electricity
conversion.18 Apparently, the optimization of the whole DSSCs could be divided into four
major categories based on the component to be modified. Detailed discussions about the
modifications will be covered in Chapter 6.
(A) Dye sensitizers
The primary considerations for optimum dyes are a wide absorption spectrum range
and the capacity of attaching firmly to the surface of electrode nanomaterials and high
electron injection quantum yields.19 Also, having good thermal and chemical stability is
significant for cell performance and long-term running especially for commercial
applications. Last but not least, a high enough redox potential can ensure that the dye
would be recovered at a fast speed by the electrolyte. 20

10

Figure 1-7 Structures of polypyridine Ru(II) dyes. (a) N3, (b) Black dye, (c) CYC-B115, 7

Metal complexes and ruthenium (Ru(II)) have been intensively researched because of
the high stability, excellent redox properties, proper absorption spectra. Typically, a Rubased dye is composed of a central metal ion with ancillary ligands having certain
anchoring groups,21 such as carboxylate or phosphonate group, enabling the electron
injection. Other parts of the dye such as ancillary ligands could also be tuned to improve
the overall photovoltaic capacity. One of the best-performing ruthenium polypryidyl
complex compounds is tris(2,2’-bipyridyl-4,4’-carboxylate) ruthium (II) (N3 dye). It is
shown that the electron injection occurs in femtosecond time scale, with an incident phototo-current efficiency (IPCE) of 80-85%.20 N3 dye’s tetrabutylammonium salt, tri(cyanato2,2’,2’’-terpyridyl-4,4’,4’’-tricarboxylate) Ru (II) (black dye), was developed to extend 100
nm further into IR region regarding to spectral response.20 Recently, an even more efficient
sensitizer CYC-B11 with a high molar extinction coefficient heteroleptic ruthenium
complex is synthesized by incorporating electron-rich hexylthio-bithiophene antennas.17
The molecular structures of the most successful Ru-based dye sensitizers are shown in
Figure 1-7. There haven been tremendous work focusing on exploitation of panchromatic
sensitizers to cover the whole visible spectrum to enhance the spectral response up to 920
nm,11 searching for economical and diverse alternatives such as organic dyes, and cosensitization instead of one single dye to improve the photo-absorption.22 In addition, no
11

matter which kind of modification is done to the dye, an essential purpose is to promote
hole charge transfer to the redox couple.
(B) Electrolytes
Redox species are dissolved in the electrolyte, serving as the mediator for regeneration
of oxidized dye and reduction at the counter electrode. The electrolyte needs to have longterm chemical and thermal stability. The absorption of sunlight by the dye will reduce if the
electrolyte has significant absorption of visible light.
Organic solvent electrolyte (OSE) and ionic liquid electrolyte (ILE) are two general
types of liquid electrolytes. OSEs are widely used because of their low viscosity, fast ion
diffusion, and high efficiency.23 OSE usually includes organic solvent like acetonitrile, redox
couple I-/I3-, and additives such as 4-tert-butylpyridine (TBP) and N-methylbenzimidazole
(NMBI), which contribute to the suppression of dark current.23 Redox couple
iodine/triiodide (I-/I3-) has been demonstrated to have the best performance for dye
regeneration, even though other kinds of redox couples have also been tried by different
groups.16 However, a major disadvantage of I-/I3- is the corrosion of sealing materials for
cell assembling. ILE has a negligible vapor pressure, non-flammability, and high ionic
conductivity, so it has attracted a lot of attention recently. However, since the pure ionic
liquid has higher viscosity, the transport speed of redox couple is very slow.16 So the ionic
liquid with both low-viscosity and high-conductivity would be a promising alternative to
OSE.
The emerging trend of solid-state electrolytes avoids the leakage and volatilization
problem of liquid ones by using a p-type hole transport material to replace the liquid

12

electrolyte. Besides of the general requirements for electrolytes, qualified hole transport
material should have energy levels aligned with the dye molecules, high hole mobility, and
good penetration into porous photoanode structure.9,21 A variety of materials including
inorganic compounds (CuI, CuSCN), molecular hole conductors, and conducting polymeric
materials have been explored as the solid-state “electrolyte”. Major problems remain in the
inefficient pore filling for dye molecules to be in contact with both anode and the hole
conductor, as well as the losses caused by recombination reactions.21
(C) Catalytic counter electrodes
The TCO-coated glass is used as a substrate of DSSC electrode, but itself alone has too
high charge transfer resistance to be a qualified counter electrode.21 Platinum is thus
widely used as a catalyst for redox reaction on the cathode, prepared simply by deposition
of a thin layer of Pt onto the TCO selected. Depositing nanoclusters of an organo-platinum
compound through thermal treatment has been proved to be stable and effective.24
The high cost of Pt triggered more research on alternatives including carbon family
such as CNTs,9 graphene nanoplatelets25, conducting polymers and cobalt sulfide, etc.21
Carbon materials have been applied to improve the electronic conductivity and catalyze the
reduction of triiodide. Researchers started with just graphite and carbon black, then
reduced the carbon black size to 30 nm with largely reduced charge transfer resistance,
and utilized single wall carbon nanotubes and graphene as well in recent years.21 As for
conducting polymers, poly(3,4-ethylenedioxythiophene) (PEDOT) has been a strong
candidate charge transfer resistance less than 1 Ί cm2 for 1 um thick film on TCO

13

substrates. Electrodeposited CoS on a flexible substrate was proved to catalyze
iodide/triiodide redox couple as well.21
(D) Nanostructured metal oxide photoanodes
Photoanode, composed of semiconductor nanomaterials, acts as the support for dye
sensitizer, the media for electron transport and electrolyte diffusion in the solar cell. The
semiconductor nanomaterial film is required specifically to have a large surface area,
highly porous structure, and fast electron transport capacity.
Among various wide band gap metal oxide semiconductors, TiO2 is the most studied
and effective material because of its surface area, porosity, crystallinity, and electron
transport capacity. Additionally, titania, is available in large quantities, cost-effective,
environmentally friendly, stable chemically and thermally.
Titania exists in three crystal phases, which are rutile, anatase, and brookite. Rutile has
compact aggregates resulting in low surface area, while anatase and brookite phases
contain porous aggregates and more open structure, which are more suitable for
photocatalysis and DSSC applications.26 The commonly investigated phases are anatase
(bandgap 3.2 eV) and rutile (bandgap 3.0 eV), of which the wide bandgaps ensure the
photoelectric excitation of TiO2, followed by scavenging of electrons and holes by surface
adsorbed species.

27

Besides TiO2, other wide band gap semiconductors such as ZnO

(Eg=3.2 eV) 28,29, Nb2O5 30 and SnO231 have also been tried, and a very recent investigation
of ZnO flower-shaped scattering layer improved the efficiency to 3.2%. However, these
efficiencies are still very limited compared to titania-based electrodes.

14

The design of nanocrystalline semiconductor electrode is significantly important not
only because there is relatively a larger range for modifying and optimizing, but also the
semiconductor serves as a central part of the cell and could possibly have the bigger
contribution to enhancing the overall efficiency of the solar cell. From the big picture of
research into the different components of DSSCs provided above, more focus on the
structural optimization of the photoanode will be discussed. Structural considerations offer
many opportunities for improvements in loading of dyes, transport of electrons, and
diffusion of electrolytes and suppression of recombination.
All these years, researchers have made innovations to improve the morphology, shape,
and composition, etc. of electrode materials. From different points of views, the
advancements could be classified in various groups.32 In this dissertation, TiO2
advancements could be separated into categories below: (1) nanoparticle with high surface
area; (2) 1D nanotubes or nanowires; (3) 3D hierarchically ordered structure with a
combination of large pores, large particles and smaller particles; (4) 3D template-induced
macroporous scaffolds; (5) hybrid of TiO2 with highly conductive carbon materials.32 More
details about the state-of-art advancement about titania electrode will be covered in
Chapter 6.
1.2.3 Historical overview of the development of photocatalysis
Efficient approaches to remove contaminants in water have been extensively
researched in the past decades due to the environmental pressure aroused from industrial,
agricultural, and daily life pollutions into the aqueous systems. Organic contaminants,
especially colored waste from textile industry, if not pre-treated, cause tremendous harm
to aquatic life.33 Compared to traditional methods including direct precipitation, or
15

adsorption on activated carbon, heterogeneous photocatalysis requires no further
treatment on the separated or adsorbed secondary pollutes. Additionally, photocatalysis is
superior to conventional biological treatments which usually involve a longer period of
time and more cost. 34
The earliest photocatalysis work started back to 1921 when Renz reported a series of
partial reduction of metal oxides such as TiO2, CeO2, Nb2O5 by illumination with sunlight, in
the presence of organic compounds.35 Noble metal salts reduction was found on the surface
of TiO2 and ZnO, as well as hydrogen peroxide formation on ZnO.34 The first decomposition
of a dye on TiO2 was realized in 1938,36 where initial absorption spectra and quantum
yields were measured. During the several decades afterwards, researchers focused on the
theories and mechanisms. Fujishima group found that photocatalysis capacities varied
among different TiO2 crystal phases, and anatase was more active than rutile.34
The term “photocatalyst” first appeared in work done by Kato and Mashio in 1956.37
The landmark year for photocatalysis research might be 1972, when Fujishima and Honda
discovered water splitting on TiO2.38 The idea of purifying water with illuminated titania
was proposed in 1977 by Frank and Bard,39 where they proved the effectiveness by
decomposing cyanide ions under illumination of a xenon lamp. In the 1990s, more
attention was paid to solid surfaces specialized in self-cleaning, self-sterilizing, and organic
decomposing functions, that relied on less powerful UV photons.40 Since the 21st century,
tremendous research has been done concerning photocatalytic water splitting,41 photoinduced super-hydrophilicity, and artificial photosynthesis.37

16

In short, photocatalysis is a cost-effective, environmentally friendly, versatile solution
to convert solar energy into other forms of chemical energy. It can degrade organic and
biological pollutants, aqueous or gaseous, leading to a broad range of potential applications
in pollutant cleaning, drinking water treatment, clinical usages, and industrial installations.
It is generally known that for a heterogeneous photocatalytic process, five steps could
be summarized as below: (1) diffusion of the reactants to the surface of the catalyst, (2)
adsorption of reactants onto the surface of the catalyst, (3) reaction on the surface of the
catalyst, (4) desorption of products from the surface of the catalyst, (5) diffusion of the
products from the surface of the catalyst.
Heterogeneous photocatalytic reactions begin when a semiconductor catalyst is
irradiated by UV light source (wavelength below 400 nm) that has higher energy than the
semiconductor band gap energy.42 When a photon hits the photocatalyst, electrons are
excited from the valence band to the conduction band, forming electron-hole pairs. On the
surface of a metal oxide (MO) semiconductor, the oxidative reactions initiate, followed by
reaction with moisture in air resulting in hydroxyl radicals; the reductive reactions involve
oxygen in the atmosphere and form hydroxyl radicals as well. Thus generated hydroxyls
are very oxidative and non-selective with a redox potential of E0=3.06 eV.37
Three main steps represent the photocatalytic reactions on titania:
a) photoexcitation of MO generates electrons (e- ) and holes (h+ )
b) electrons (e- ) and holes (h+ ) transfer to the TiO2 surface
c) electrons (e- ) and holes (h+ ) react with adsorbed electron acceptors and donors.43
Equations and a related scheme are shown below:
17

UV + MO → MO (h+ + e−)
h+ + H2O → H+ + •OH
2 h+ + 2 H2O → 2 H+ + H2O2
H2O2→ HO• + •OH
e− + O2 → •O2− superoxide anions
•O2− + HO•2 + h+ → H2O2 + O2
HOOH → HO• + •OH

Figure 1-8 Schematic view of the mechanism of photocatalytic process with metal oxide
semiconductor catalysts

Parameters influencing the effectiveness of the photocatalytic reaction include: catalyst
loading, catalyst surface area, and catalyst charges; solution pH, temperature and target
concentration; light wavelength and intensity; chemical additives and even the reacting
vessel system design. This is why Ryu and Choi44 conducted substrate-specific tests, which
confirmed that one photocatalyst could not meet all photocatalytic requirements and a
standardization of photocatalytic activity is needed but also realistic to be made.
18

A variety of materials such as TiO2, ZnO, V2O5, SnO2, Fe2O3, ZrO2, CdS, WO337 have been
investigated. However, TiO2 remains the most popular material for photocatalysis. For
example, photo-corrosion problems will happen during the excitation of ZnO in solution,
and hematite suffers from rapid charge recombination and short charge carrier diffusion
length.27 The success of TiO2 could be attributed to its excellent UV absorption capacity,
high photoreactivity, chemical and thermal stability, low cost, suitable band gap position to
oxidize organic molecules, and long-term photostability.45 The commonly investigated
phases are anatase (bandgap 3.2 eV) and rutile (bandgap 3.0 eV), of which the wide
bandgaps ensure the photoelectric excitation of TiO2, followed by scavenging of electrons
and holes by surface adsorbed species.27 Titanium dioxide photocatalysts could be used as
powders that suspend in solutions or immobilized on thin films. Large scale reactor design
includes slurry-type or immobilized-type reactors. Drawbacks of titania have been
discussed for years due to the facts that it only absorbs wavelength lower than 387 nm in
the solar light spectrum, and the charge recombination, hole trapping, electron trapping
from defected nanoparticles limit the electron transport properties.
Briefly, current modifying trends on titania photocatalysts include: (1) expanding the
response range of TiO2 to visible light; (2) increasing the adsorption of pollutants for more
active reaction sites, (3) improving the charge separation to retard charge recombination
for a higher overall efficiency; (4) changing the selectivity or yield of a particular reactant.
Surface modifications have been applied to optimize the performance of titania
photocatalysts.

19

1.3 Nano Titania Materials and Synthesis Methods
To improve the solar energy conversion efficiency and meet the current modifying
trends, TiO2 nanomaterials have been researched in various directions, such as spatial
structuring and doping,37 size reduction and shape control,46 graphene supporting47
surface engineering48, etc. To start, general methods to fabricate appropriate titania
nanomaterials for photovoltaic and photocatalytic applications should be discussed. Thus, a
few representative synthesis methods will be introduced as follows.
1.3.1 Sol-gel method
The sol-gel method has been used to prepare a variety of ceramic materials. 49-52 Based
on hydrolysis and condensation reactions, sol-gel chemistry is an attractive wet chemistry
approach to synthesize metal oxides.53 Starting with inorganic metal salts, or metal organic
compounds,45 sol-gel synthesis involves a simple preparation of precursor/solvent system,
lower temperature processing, easy coating of large surfaces with variable thickness and
high optical/electrical qualities.54
Precursors, such as metal alkoxides, metal acetylacetonates, and inorganic metal
compounds, go through a hydrolysis and condensation polymerization to form a sol, while
further cross-linking combined with solvent loss results in a metal oxide gel. Variations in
extent of polymerization and loss of solvent will result in liquid sol, wet gel, or solid gel,
while powders, thin films, aerogel, ceramic fibers can further be obtained depending on
how the gel phase transfers next.45
A typical sol-gel reaction of metal alkoxide initiates with hydrolysis,
𝑀 − 𝑂𝑅 + 𝐻2 𝑂 → 𝑀 − 𝑂𝐻 + 𝑅𝑂𝐻

20

Condensation begins as soon as hydroxyl groups are generated. There might be two
competing mechanisms:53
(i) Olation: removal of a water or alcohol molecule and formation of hydroxyl bridges.
𝑀 − 𝑂𝐻 + 𝑀 − 𝑂𝐻𝑋 → 𝑀 − 𝑂𝐻 − 𝑀 + 𝑋𝑂𝐻
(𝑋 = 𝐻 𝑜𝑟 𝑎𝑙𝑘𝑦𝑙 𝑔𝑟𝑜𝑢𝑝)
(ii) Oxolation: removal of a water or alcohol molecule and formation of oxygen bridges.
𝑀 − 𝑂𝐻 + 𝑀 − 𝑂𝑋 → 𝑀 − 𝑂 − 𝑀 + 𝑋𝑂𝐻
(𝑋 = 𝐻 𝑜𝑟 𝑎𝑙𝑘𝑦𝑙 𝑔𝑟𝑜𝑢𝑝)
The relationship between these three reactions will determine the final structure and
morphology of the resulting network. The rate constants for hydrolysis and condensation
depend on water-to-metal ratio, temperature, solvent and precursor composition, etc.55
Livage defined the hydrolysis ratio as h=H2O/M. As h controls the mechanism of
condensation and affects the size, structure, and physical properties of macromolecular
species, three circumstances were discussed. When h<1, condensation is dominated by
oxolation with alkoxy bridging, leading to molecular oxo-alkoxides. When 1 < h < n (n is the
coordination number of M), a polymeric gel may be formed. When h>n, cross-linked
polymers will be formed, and precipitates can be obtained once excessive water is added.53
Titanium, being a transition metal, differs with the extensively studied silicon
precursors since it has a lower electronegativity and a higher electrophilicity, which makes
titanium alkoxides much more reactive. The molecular complexity of the sol depends on
the alkoxy group in the precursor Ti(OR)4. The bulkier the OR group is the smaller the

21

complexity.53 The hydrolysis rate will increase with temperature and the value of h.
Primary alkoxides hydrolyze easier than more complex alkoxides. With an acetic catalyst,
the hydrolysis rate increases while condensation decreases, this facilitates the separation
between hydrolysis and condensation steps.
Titania nanomaterials have been synthesized from hydrolysis of a Ti precursor with a
sol-gel method in numerous ways due to its versatile synthetic procedures and fine control
over purity, homogeneity, and crystallization.56-59 According to the aforementioned theory,
limiting the amount of water will suppress the hydrolysis rate and develop Ti-O-Ti chains
with the excess titanium alkoxide into a closely packed polymeric skeletons. As h increases,
Ti(OH)4 forms loosely packed primary particles without formation of polymeric skeletons.
While h reaches a large excess, tightly packed primary particles are created through a gel
skeleton and with extended times or higher temperatures, the average radius of
nanoparticles grow linearly with time, and more secondary particles will be formed.45
Water content is not the only parameter that guided the sol-gel chemistry for TiO2
fabrication. Additives and surfactants such as alkylamines, carboxylic acids, and
alkylphosphonic acids are effective as shape controllers and act by changing the surface
energies of the crystallographic planes.58 Amines, such as triethanolamine (TEOA),
ethylenediamine and diethylenetriamine have been applied to prepare TiO2 nanoparticles
with different aspect ratios. Sodium oleate and sodium stearate could alter TiO2 cubes from
round-corner to sharp-edge.60 A standard sol-gel procedure61 to fabricate TiO2 utilized
titanium alkoxide like titanium tetraisopropoxide (TTIP) and tetramethylammonium
hydroxide. TTIP was added to 2 °C alcoholic solvents in a three-neck flask and heated at
around 50 °C for 13 days. Lee et al.62 developed a two-step sol-gel method where they
22

applied triethanolamine (TEOA) with a molar ratio of TTIP: TEOA = 1:2 and
ethylenediamine (ED) further as the shape controllers because amine groups strongly bond
to TiO2 surfaces when pH > 10.5, yielding anisotropic growth with high aspect ratios.
Longer low-temperature heating of the gel could prevent the agglomeration of TiO2
particles during the high-temperature annealing crystallization process,45 and templates
have also been widely applied with sol-gel method to create TiO2 with ordered structures
like pores, or hollow tubes.
1.3.2 Hydrothermal method
Usually prepared with different types of precursors in steel pressure vessels
(autoclaves) under controlled temperature and pressure, the hydrothermal method has
been extensively studied to fabricate nanoscale ceramics.45 The temperature is usually
above the boiling point of water, or as high as the self-generated saturated vapor pressure.
The pressure in the autoclaves is controlled by the reactants and operating temperature.63
The hydrothermal method has been widely utilized to synthesize spherical particles or
1D titania because it requires less cost, mild reaction conditions, and simple equipment
compared to chemical vapor deposition, anodic oxidation, seeded growth and other
approaches.63 The broad manipulation opportunities on each parameter (pH, concentration,
temperature, reaction time) control the morphology and shape.

23

Figure 1-9 Formation mechanism of TiO2 nanotubes using hydrothermal method.63

TiO2 nanoparticles can be synthesized at 240 °C for 2 hours with the peptized
precipitates in water.64 With added acidity, pure anatase spherical particles with adjustable
diameter were made by changing the Ti(IV) concentration and solvent composition.65 1D
titania nanorods, nanowires have also been prepared through acidic or basic aqueous
precursors with hydrothermal methods.45 When the medium is turned to alkaline (usually
concentrated NaOH), titania nanotubes could be easily prepared from TiO2 powders. The
acid washing step introduced by Kasuga et al.66 was to remove the sodium ions, replace
with h+ ions, and form new Ti-OH bonds for further dehydration into Ti-O-Ti. (as Figure
1-9 illustrates)
However, more studies reveal that even without washing step, the hydrothermal step
is enough to create nanotube structure.63 Factors influencing the TiO2 nanotubes formation
were discussed based on the recent literatures in Wong’s review63 in the following
manners: (1) Starting materials could vary among titania pure phase powders, commercial
Degussa P25 TiO2 powders (a mixture of anatase and rutile phases), Ti metal, Ti (IV)

24

alkoxide, layered titanate Na2Ti3O7, and titania sol, leading to various sizes and shapes. (2)
The temperature and duration of the hydrothermal step will also determine the
morphology and dimensions of the nanotubes. (3) Whether or not applying a sonication
pretreatment has influence on the dispersion of the particles, and the crystal growth. (4)
Washing step will determine the elemental composition and surface area of the nanotubes.
1.3.3 Liquid phase deposition method
Liquid phase deposition (LPD) method is a simple route to synthesis nano-sized metal
dioxides via a ligand exchange equilibrium reaction on a substrate. Early research on LPD
dated back to 1980s, when Nagayama innovated a new coating method for silica as shown
in Figure 1-10.67

Figure 1-10 Scheme of the synthesis procedures of silica from H2SiF6.67 (Reproduced with
permission from J. Electrochem. Soc., 150, H205 (2003). Copyright 2003, The Electrochemical
Society)

Hydrofluosilicic acid was diluted with water, followed by the addition of silica gel at
35 °C and kept stirring for 16 hours at the same temperature. The addition of silica led the
equilibrium (a) from right to left. However, boric acid was added as a reactant to fluoric
25

acid, since BF4- is more stable than the noncoordinated F- ions, changing (a) and leaving
the solution supersaturated with silica. Cleaned glass substrates were immersed in the
solution for a varied period of time. The deposition of silica on the surface of glass would
start and proceed steadily with time. Further washing and annealing were conducted to
crystallize the deposited nanoparticles.
𝐻2 𝑆𝑖𝐹6 + 2𝐻2 𝑂 ↔ 6𝐻𝐹 + 𝑆𝑖𝑂 2 (a)
𝐻3 𝐵𝑂3 + 4𝐻𝐹 ↔ 𝐵𝐹4− + 𝐻3 𝑂+ + 2𝐻2 𝑂 (b)
Briefly, this paper pointed out several rules applied in silica LPD process. (1) The
deposition rate is influenced by temperature and boric acid amount, in other words, the
degree of supersaturation. (2) Lower temperature solution coated film is more dense and
stable, and further calcination fortifies this effect.
As for titania synthesis, a very representative system utilizing ammonium
hexafluorotitanate (AHFT) and boric acid (BA) has been applied to prepare titania
depositions by the ligand exchange of metal fluorocomplex ions [TiF6]2- and the
consumption of F- with boric acid as a scavenger. In this process, the hydrolysis of the
titanium fluoro complex yields titanium hydroxide and free F- ions, while the addition of
boric acid drives the reaction towards right. The properties of the titania product and film
thickness could be determined by several factors such as the substrate,68 concentration of
AHFT and BA,

69

reaction temperature, immersion time,

and length. 72,73

26

70,71

and calcination temperature

Another type of precursor came from a simple titanium salt TiF4, which has been
deposited on different substrates such as glass slides, polymers, cotton fibers, filter paper,
porous alumina membranes, by a similar mechanism in acidic aqueous environment. 74,75
𝑇𝑖𝐹 4 + 𝑛𝐻 2 𝑂 ↔ 𝑇𝑖𝐹4−𝑛 (𝑂𝐻)𝑛 + 𝑛𝐻𝐹
𝑇𝑖𝐹4−𝑛 (𝑂𝐻)𝑛 + (4 − 𝑛)𝐻2 𝑂 → 𝑇𝑖(𝑂𝐻)4 + (4 − 𝑛)𝐻𝐹
𝑇𝑖(𝑂𝐻)4 → 𝑇𝑖𝑂2 + 2𝐻2 𝑂 (f)
The hydrolysis and condensation go through stepwise reactions as above equations
illustrate. Different from using a F- scavenger to adjust the supersaturation level, for this
method, pH is usually adjusted by adding acidic or basic reagents. When pH is below 1,
inefficient hydrolysis prevents the deposition. When pH is above 3, precipitation is
preferred so the solution is highly supersaturated with titania rather than forming a
uniform coating on substrates. When pH is between 1 and 3, due to the stability of Ti-F
bonds, rapid hydrolysis and polymerization of the precursor do not occur; instead, the
moderate reaction leads to a stable deposition through heterogeneous nucleation.75
There are various ways to develop a thin TiO2 film, such as vacuum evaporation,
sputtering, chemical vapor deposition and sol-gel based spin coating, doctor blade, or dip
coating. These methods can not win regarding large scale industrial use. Compared to the
above-mentioned methods, LPD is straight-forward, lower-temperature assisted, low-cost,
environmentally friendly and does not require any special equipment, besides, works with
various types of substrates such as glass, quartz,72 mica,76 parylene composites,69 silica
fibers or spheres,77, polyimide surface on spacecraft,78 polycarbonate membranes, and

27

even inorganic salts like MnCO3,79 cotton fibers80 and even pre-made TiO2 nanorods,81
vanadium oxide nanobelts,82 graphene,83 layer-by-layer made PAH/PAA (polyallylamine
hydrochloride/polyacrylic acid) films.84 Most room temperature methods require hightemperature annealing to remove the organic precursor and enhance the crystallinity. Last
but not least, this aqueous one-step reaction yields particles with anatase crystallinity
directly. 74,81
The chemical bonding between the deposited nanomaterials and the substrates
ensured the strong adhesion of the coating, besides, the chemical and physical properties of
the film can be finely tuned by adjusting the parameters (concentration of the reactants,
the duration, and temperature of the reaction, pH, solution composition) of the reacting
system

easily,

which

will

be

discussed

28

in

the

following

chapters.

BIBLIOGRAPHY

29

BIBLIOGRAPHY

(1)

Team, C. W.; P, R. K.; Reisinger, A. Assessment Report (AF4) of the United Nations
Intergovernmental Panel on Climate Change (IPCC); Geneva, Switzerland; Vol.
2007.

(2)

U. S. E. I. A. International Energy Outlook 2013. 2013, 1–312.

(3)

U. S. E. I. A. Annual Energy Outlook 2015 with Projections to 2040. 2015, 1–154.

(4)

Turner, J. A. A Realizable Renewable Energy Future (Vol 285, Pg 687, 1999);
Science, 1999.

(5)

Doney, S. C.; Schimel, D. S. Annual Review of Environment and Resources; 2007;
Vol. 32, pp. 31–66.

(6)

Goetzberger, A.; Hebling, C. Photovoltaic Materials, Past, Present, Future. Solar
Energy Materials and Solar Cells 2000, 62, 1–19.

(7)

National Renewable Energy laboratory. Best Research-Cell Efficiencies; 2016.

(8)

Green, M. A.; Emery, K.; Hishikawa, Y.; Warta, W.; Dunlop, E. D. Solar Cell
Efficiency Tables (Version 43). Progress in Photovoltaics 2014, 22, 1–9.

(9)

Dye-Sensitized Solar Cells: a Safe Bet for the Future. 2008, 1, 655–667.

(10)

Measuring the Thermal Conductivity of Thin Films: 3 Omega and Related
Electrothermal Methods

(11)

Caramori, S.; Cristino, V.; Boaretto, R.; Argazzi, R.; Bignozzi, C. A.; Di Carlo, A.
New Components for Dye-Sensitized Solar Cells. International Journal of
Photoenergy 2010, 2010, 1–16.

(12)

Nazeeruddin, M. K.; Kay, A.; Rodicio, I.; Humphry-Baker, R.; Mueller, E.; Liska, P.;
Vlachopoulos, N.; Graetzel, M. Conversion of Light to Electricity by CisX2bis(2,2“-Bipyridyl-4,4-”Dicarboxylate)Ruthenium(II) Charge-Transfer
Sensitizers (X = Cl-, Br-, I-, CN-, and SCN-) on Nanocrystalline Titanium Dioxide
Electrodes. J. Am. Chem. Soc. 1993, 115, 6382–6390.

(13)

Han, L.; Islam, A.; Chen, H.; Malapaka, C.; Chiranjeevi, B.; Zhang, S.; Yang, X.;
Yanagida, M. High-Efficiency Dye-Sensitized Solar Cell with a Novel CoAdsorbent. Energy Environ. Sci. 2012, 5, 6057–6060.

(14)

Nazeeruddin, M. K.; PĂŠchy, P.; Renouard, T.; Zakeeruddin, S. M.; Humphry-Baker,
R.; Comte, P.; Liska, P.; Cevey, L.; Costa, E.; Shklover, V.; et al. Engineering of
Efficient Panchromatic Sensitizers for Nanocrystalline TiO 2-Based Solar Cells. J.
30

Am. Chem. Soc. 2001, 123, 1613–1624.
(15)

Palomares, E.; Clifford, J. N.; Haque, S. A.; Lutz, T.; Durrant, J. R. Control of Charge
Recombination Dynamics in Dye Sensitized Solar Cells by the Use of Conformally
Deposited Metal Oxide Blocking Layers. J. Am. Chem. Soc. 2003, 125, 475–482.

(16)

Wu, J.; Lan, Z.; Hao, S.; Li, P.; Lin, J.; Huang, M.; Fang, L.; Huang, Y. Progress on the
Electrolytes for Dye-Sensitized Solar Cells. Pure and Applied Chemistry 2008, 80,
2241–2258.

(17)

Chen, C.-Y.; Wang, M.; Li, J.-Y.; Pootrakulchote, N.; Alibabaei, L.; Ngoc-le, C.-H.;
Decoppet, J.-D.; Tsai, J.-H.; Grätzel, C.; Wu, C.-G.; et al. Highly Efficient LightHarvesting Ruthenium Sensitizer for Thin-Film Dye-Sensitized Solar Cells. ACS
Nano 2009, 3, 3103–3109.

(18)

Burschka, J.; Pellet, N.; Moon, S.-J.; Humphry-Baker, R.; Gao, P.; Nazeeruddin, M.
K.; Graetzel, M. Sequential Deposition as a Route to High-Performance
Perovskite-Sensitized Solar Cells. Nature 2013, 499, 316–.

(19)

Morandeira, A.; LĂłpez-Duarte, I.; MartĂ­nez-DĂ­az, M. V.; O'Regan, B.; Shuttle, C.;
Haji-Zainulabidin, N. A.; Torres, T.; Palomares, E.; Durrant, J. R. Slow Electron
Injection on Ru−Phthalocyanine Sensitized TiO 2. J. Am. Chem. Soc. 2007, 129,
9250–9251.

(20)

PĂŠchy, P.; Renouard, T.; Zakeeruddin, S. M.; Humphry-Baker, R.; Comte, P.; Liska,
P.; Cevey, L.; Costa, E.; Shklover, V.; Spiccia, L.; et al. Engineering of Efficient
Panchromatic Sensitizers for Nanocrystalline TiO 2-Based Solar Cells. J. Am.
Chem. Soc. 2001, 123, 1613–1624.

(21)

Hagfeldt, A.; Boschloo, G.; Sun, L.; Kloo, L.; Pettersson, H. Dye-Sensitized Solar
Cells. Chem. Rev. 2010, 110, 6595–6663.

(22)

Clifford, J. N.; Palomares, E.; Nazeeruddin, M. K.; Thampi, R.; Grätzel, M.; Durrant,
J. R. Multistep Electron Transfer Processes on Dye Co-Sensitized Nanocrystalline
TiO 2Films. J. Am. Chem. Soc. 2004, 126, 5670–5671.

(23)

Kong, F.-T.; Dai, S.-Y.; Wang, K.-J. Review of Recent Progress in Dye-Sensitized
Solar Cells. Advances in OptoElectronics 2007, 2007, 1–13.

(24)

Papageorgiou, N. An Iodine/Triiodide Reduction Electrocatalyst for Aqueous
and Organic Media. J. Electrochem. Soc. 1997, 144, 876.

(25)

Kavan, L.; Yum, J. H.; Grätzel, M. Optically Transparent Cathode for DyeSensitized Solar Cells Based on Graphene Nanoplatelets. ACS Nano 2010, 5, 165–
172.

(26)

Phase-Pure TiO2 Nanoparticles: Anatase, Brookite and Rutile. 2008, 19, 145605.

31

(27)

Ibhadon, A.; Fitzpatrick, P. Heterogeneous Photocatalysis: Recent Advances and
Applications. Catalysts 2013, Vol. 3, Pages 455-485 2013, 3, 189–218.

(28)

ZnO Nanotube-Based Dye-Sensitized Solar Cell and Its Application in SelfPowered Devices. 2010, 21.

(29)

Xu, J.; Fan, K.; Shi, W.; Li, K.; Peng, T. Application of ZnO Micro-Flowers as
Scattering Layer for ZnO-Based Dye-Sensitized Solar Cells with Enhanced
Conversion Efficiency. Solar Energy 2014, 101, 150–159.

(30)

Le Viet, A.; Jose, R.; Reddy, M. V.; Chowdari, B. V. R.; Ramakrishna, S. Nb2O5
Photoelectrodes for Dye-Sensitized Solar Cells: Choice of the Polymorph. The
Journal of … 2010, 114, 21795–21800.

(31)

Snaith, H. J.; Ducati, C. SnO2-Based Dye-Sensitized Hybrid Solar Cells Exhibiting
Near Unity Absorbed Photon-to-Electron Conversion Efficiency. Nano Lett.
2010, 10, 1259–1265.

(32)

Maçaira, J.; Andrade, L.; Mendes, A. Review on Nanostructured Photoelectrodes
for Next Generation Dye-Sensitized Solar Cells. Renewable and Sustainable
Energy Reviews 2013, 27, 334–349.

(33)

Bizani, E.; Fytianos, K.; Poulios, I.; Tsiridis, V. Photocatalytic Decolorization and
Degradation of Dye Solutions and Wastewaters in the Presence of Titanium
Dioxide. Journal of Hazardous Materials 2006, 136, 85–94.

(34)

A Fujishima, X. Z. D. A. T. TiO2 Photocatalysis and Related Surface Phenomena.
2008, 63, 515–582.

(35)

Renz, C. Lichtreaktionen Der Oxyde Des Titans, Cers Und Der Erdsäuren.
Helvetica Chimica Acta 1921, 4, 961–968.

(36)

Goodeve, C. F.; Kitchener, J. A. Photosensitisation by Titanium Dioxide. Trans.
Faraday Soc. 1938, 34, 570–579.

(37)

Xu, H.; Ouyang, S.; Liu, L.; Reunchan, P.; Umezawa, N.; Ye, J. Recent Advances in
TiO2-Based Photocatalysis. J. Mater. Chem. A 2014, 2, 12642–12661.

(38)

Fujishima, A.; HONDA, K. Electrochemical Photolysis of Water at a
Semiconductor Electrode. Nature 1972, 238, 37–38.

(39)

Frank, S. N.; Bard, A. J. Heterogeneous Photocatalytic Oxidation of Cyanide Ion in
Aqueous Solutions at Titanium Dioxide Powder. J. Am. Chem. Soc. 1977, 99, 303–
304.

(40)

LINSEBIGLER, A. L.; LU, G. Q.; YATES, J. T. Photocatalysis on Tio2 Surfaces Principles, Mechanisms, and Selected Results. Chem. Rev. 1995, 95, 735–758.

32

(41)

Hartmann, P.; Lee, D.-K.; Smarsly, B. M.; Janek, J. Mesoporous TiO 2: Comparison
of Classical Sol−Gel and Nanoparticle Based Photoelectrodes for the Water
Splitting Reaction. ACS Nano 2010, 4, 3147–3154.

(42)

Rodgher, V.; Moreira, J.; de Lasa, H.; Serrano, B. Photocatalytic Degradation of
Malic Acid Using a Thin Coated TiO 2-Film: Insights on the Mechanism of
Photocatalysis. AIChE J. 2014, 60, 3286–3299.

(43)

Vautier, M.; Guillard, C.; Herrmann, J.-M. Photocatalytic Degradation of Dyes in
Water: Case Study of Indigo and of Indigo Carmine. Journal of Catalysis 2001,
201, 46–59.

(44)

Ryu, J.; Choi, W. Substrate-Specific Photocatalytic Activities of TiO 2and
Multiactivity Test for Water Treatment Application. Environ. Sci. Technol. 2008,
42, 294–300.

(45)

Chen, X.; Mao, S. S. Titanium Dioxide Nanomaterials: Synthesis, Properties,
Modifications, and Applications. Chem. Rev. 2007, 107, 2891–2959.

(46)

Liu, N.; Chen, X.; Zhang, J.; Schwank, J. W. A Review on TiO2-Based Nanotubes
Synthesized via Hydrothermal Method: Formation Mechanism, Structure
Modification, and Photocatalytic Applications. Catalysis Today 2014, 225, 34–51.

(47)

Lui, G.; Liao, J.-Y.; Duan, A.; Zhang, Z.; Fowler, M.; Yu, A. Graphene-Wrapped
Hierarchical TiO2 Nanoflower Composites with Enhanced Photocatalytic
Performance. J. Mater. Chem. A 2013, 1, 12255–12262.

(48)

Chen, F.; Cao, F.; Li, H.; Bian, Z. Exploring the Important Role of Nanocrystals
Orientation in TiO 2Superstructure on Photocatalytic Performances. Langmuir
2015, 31, 3494–3499.

(49)

HU, L. L.; YOKO, T.; KOZUKA, H.; SAKKA, S. Effects of Solvent on Properties of Sol
Gel-Derived Tio2 Coating Films. Thin Solid Films 1992, 219, 18–23.

(50)

Ohya, Y.; Saiki, H.; Tanaka, T.; Takahashi, Y. Microstructure of TiO2 and ZnO
Films Fabricated by the Sol-Gel Method. Journal of the American Ceramic Society
1996, 79, 825–830.

(51)

Schattka, J. H.; Wong, E.; Antonietti, M.; Caruso, R. A. Sol-Gel Templating of
Membranes to Form Thick, Porous Titania, Titania/Zirconia and Titania/Silica
Films. J. Mater. Chem. 2006, 16, 1414–1420.

(52)

Wang, H.; Ding, J.; Xue, Y.; Wang, X.; Lin, T. Superhydrophobic Fabrics From
Hybrid Silica Sol-Gel Coatings: Structural Effect of Precursors on Wettability and
Washing Durability. J. Mater. Res. 2011, 25, 1336–1343.

(53)

sanchez, C.; livage, J. Sol-Gel Chemistry From Metal Alkoxide Precursors. New

33

journal of Chemistry 1990, 513–521.
(54)

Sakka, S.; Yoko, T. Sol-Gel-Derived Coating Films and Applications. Structure and
Bonding 1992, 77, 89–118.

(55)

Askeland, P. A. Preparation and Characterization of Sol-Gel Derived Metal Oxide
Thin Films and Powders for Coatings and Catalysts, 1997.

(56)

Antonelli, D. M.; Ying, J. Y. Synthesis of Hexagonally Packed Mesoporous TiO2 by
a Modified Sol–Gel Method. Angewandte Chemie International Edition in English
1995, 34, 2014–2017.

(57)

BarbĂŠ, C. J.; Arendse, F.; Comte, P.; Jirousek, M.; Lenzmann, F.; Shklover, V.;
Grätzel, M. Nanocrystalline Titanium Oxide Electrodes for Photovoltaic
Applications. Journal of the American Ceramic Society 1997, 80, 3157–3171.

(58)

Kim, J. Y.; Choi, S. B.; Kim, D. W.; Lee, S.; Jung, H. S.; Lee, J.-K.; Hong, K. S.
Surfactant-Assisted Shape Evolution of Thermally Synthesized TiO
2Nanocrystals and Their Applications to Efficient Photoelectrodes. Langmuir
2008, 24, 4316–4319.

(59)

Burunkaya, E.; Akarsu, M.; Çamurlu, H. E.; Kesmez, Ö.; Yeşil, Z.; Asiltürk, M.;
Arpaç, E. Production of Stable Hydrosols of Crystalline TiO2 Nanoparticles
Synthesized at Relatively Low Temperatures in Diverse Media. Applied Surface
Science 2013, 265, 317–323.

(60)

Li, W.; Wu, Z.; Wang, J.; Elzatahry, A. A.; Zhao, D. A Perspective on Mesoporous
TiO 2Materials. 2014, 26, 287–298.

(61)

Chemseddine, A.; Moritz, T. Nanostructuring Titania: Control Over Nanocrystal
Structure, Size, Shape, and Organization. European Journal of Inorganic
Chemistry 1999, 1999, 235–245.

(62)

Lee, S.; Cho, I.-S.; Lee, J. H.; Kim, D. H.; Kim, D. W.; Kim, J. Y.; Shin, H.; Lee, J.-K.;
Jung, H. S.; Park, N.-G.; et al. Two-Step Sol−Gel Method-Based TiO 2Nanoparticles
with Uniform Morphology and Size for Efficient Photo-Energy Conversion
Devices. Chem. Mater. 2010, 22, 1958–1965.

(63)

Wong, C. L.; Tan, Y. N.; Mohamed, A. R. A Review on the Formation of Titania
Nanotube Photocatalysts by Hydrothermal Treatment. J. Environ. Manage. 2011,
92, 1669–1680.

(64)

Yang, J.; Mei, S.; Ferreira, J. M. F. Hydrothermal Synthesis of TiO2 Nanopowders
From Tetraalkylammonium Hydroxide Peptized Sols. Materials Science and
Engineering: C 2001, 15, 183–185.

(65)

Seung Yong Chae; Myun Kyu Park; Sang Kyung Lee; Taek Young Kim; Sang Kyu
Kim, A.; Lee, W. I. Preparation of Size-Controlled TiO2 Nanoparticles and
Derivation of Optically Transparent Photocatalytic Films. Chem. Mater. 2003, 15,
34

3326–3331.
(66)

Kasuga, T.; Hiramatsu, M.; Hoson, A.; Sekino, T.; Niihara, K. Formation of
Titanium Oxide Nanotube. Langmuir 1998, 14, 3160–3163.

(67)

Nagayama, H.; Honda, H.; Kawahara, H. A New Process for Silica Coating. J.
Electrochem. Soc. 1988, 135, 2013–2016.

(68)

Yu, J.; Fan, J.; Cheng, B. Dye-Sensitized Solar Cells Based on Anatase TiO2 Hollow
Spheres/Carbon Nanotube Composite Films. Journal of Power Sources 2011,
196, 7891–7898.

(69)

Malvadkar, N.; Dressick, W. J.; Demirel, M. C. Liquid Phase Deposition of Titania
Onto Nanostructured Poly-P-Xylylene Thin Films. J. Mater. Chem. 2009, 19,
4796–4799.

(70)

Li, S.; Liu, J.; Feng, T. Low Temperature Coating of Anatase Thin Films on Silica
Glass Fibers by Liquid Phase Deposition. J. Wuhan Univ. Technol. 2007, 22, 136–
139.

(71)

Lee, M.-K.; Fan, C.-H.; Wang, H.-C.; Chang, C.-Y. Fluorine Passivation of Titanium
Oxide Films on ITO/Glass Grown by Liquid Phase Deposition for Electrochroism.
J. Electrochem. Soc. 2011, 158, D511.

(72)

Yu, J.-G.; Yu, H.-G.; Cheng, B.; Zhao, X.-J.; Yu, J. C.; Ho, W.-K. The Effect of
Calcination Temperature on the Surface Microstructure and Photocatalytic
Activity of TiO 2Thin Films Prepared by Liquid Phase Deposition. J. Phys. Chem. B
2003, 107, 13871–13879.

(73)

Maeda, M. Preparation of Titania Films with Cohered Nanosized Particles Using
Improved Liquid Phase Deposition Process. international journal of
electrochemical science 2015, 10, 2988–2996.

(74)

Shimizu, K.; Imai, H.; Hirashima, H.; Tsukuma, K. Low-Temperature Synthesis of
Anatase Thin Films on Glass and Organic Substrates by Direct Deposition From
Aqueous Solutions. Thin Solid Films 1999, 351, 220–224.

(75)

Imai, H.; Takei, Y.; Shimizu, K.; Matsuda, M.; Hirashima, H. Direct Preparation of
Anatase TiO2 Nanotubes in Porous Alumina Membranes. J. Mater. Chem. 1999, 9,
2971–2972.

(76)

Pourmand, M.; Taghavinia, N. TiO2 Nanostructured Films on Mica Using Liquid
Phase Deposition. Materials Chemistry and Physics 2008, 107, 449–455.

(77)

Wu, J.-H.; Li, X.-S.; Zhao, Y.; Zhang, W.; Guo, L.; Feng, Y.-Q. Application of Liquid
Phase Deposited Titania Nanoparticles on Silica Spheres to Phosphopeptide
Enrichment and High Performance Liquid Chromatography Packings. Journal of
Chromatography A 2011, 1218, 2944–2953.

35

(78)

Gouzman, I.; Girshevitz, O.; Grossman, E.; Eliaz, N.; Sukenik, C. N. Thin Film Oxide
Barrier Layers: Protection of Kapton From Space Environment by Liquid Phase
Deposition of Titanium Oxide. ACS Appl. Mater. Interfaces 2010, 2, 1835–1843.

(79)

Lee, H.-K.; Sakemi, D.; Selyanchyn, R.; Lee, C.-G.; Lee, S.-W. Titania Nanocoating
on MnCO3 Microspheres via Liquid-Phase Deposition for Fabrication of
Template-Assisted Core-Shell- and Hollow-Structured Composites. ACS Appl.
Mater. Interfaces 2014, 6, 57–64.

(80)

Ghadiri, E.; Taghavinia, N.; Zakeeruddin, S. M.; Grätzel, M.; Moser, J.-E. Enhanced
Electron Collection Efficiency in Dye-Sensitized Solar Cells Based on
Nanostructured TiO 2Hollow Fibers. Nano Lett. 2010, 10, 1632–1638.

(81)

Yu, H.; Yu, J.; Cheng, B.; Lin, J. Synthesis, Characterization and Photocatalytic
Activity of Mesoporous Titania Nanorod/Titanate Nanotube Composites. Journal
of Hazardous Materials 2007, 147, 581–587.

(82)

Yu, H.; Yu, J.; Cheng, B.; Liu, S. Novel Preparation and Photocatalytic Activity of
One-Dimensional TiO 2hollow Structures. Nanotechnology 2007, 18, 065604.

(83)

Zhang, H.; Xu, P.; Du, G.; Chen, Z.; Oh, K.; Pan, D.; Jiao, Z. A Facile One-Step
Synthesis of TiO2/Graphene Composites for Photodegradation of Methyl
Orange. Nano Res. 2010, 4, 274–283.

(84)

Kim, J.-H.; Lee, M. J.; Kim, S.; Hwang, J.; Lim, T.-Y.; Kim, S.-H. Fabrication of
Patterned TiO2 Thin Film by a Wet Process. Met. Mater. Int. 2012, 18, 833–837.

36

2 Titania from a Hard Template CaCO3 and its Photocatalytic
Performance
Background
2.1.1 Porous materials and their benefits
Porous material is a continuous and solid network filled with voids.1 Porous materials
may have high surface area, low density, customizable structures and electronic properties
which are important for the fields of adsorption, separation, catalysis and renewable
energy.2 According to IUPAC definition, nanoporous materials have pores smaller than 100
nm and can be subdivided into three categories as: microporous materials (0.2-2 nm) such
as zeolites and metal organic framework, mesoporous materials (2-50 nm) such as silica
and most metal oxides and macroporous materials (50-1000 nm) such as sponge, cotton
and some metal oxides.1 When considered from the material compositions, they can be
classified as purely inorganic, organic-inorganic hybrid and purely organic porous
materials. Inorganic porous materials are well known as metal oxides, silica, while organic
porous materials are represented by porous carbon and porous polymers. Hybrids are
usually formed by strong bonds between the inorganic parts such as metallic ions, clusters,
layers or three-dimensional scaffolds, and organic parts such as polycarboxylates,
phosphonates, and sulfonates.3
Chapter 1 already introduced the five steps of heterogeneous photocatalytic reactions.
Nanoporous materials will be beneficial for the first two steps - diffusion and adsorption of
reactants on the surface of photocatalysts. Using nanoporous titania rather than its bulk
form provides increased surface-to-volume ratio, higher surface area, and diffusion path
37

for reactants. In addition, the increase of the surface energy for each nanoparticle will
improve the contaminant removal, compared to bulk materials without porous structures.
Last but not least, less amount of catalyst will be needed because of the increase of the
active surface sites. These factors will positively influence the third step – reaction on the
surface of the catalyst- as well.
2.1.2 Templating methods for porous TiO2
In Chapter 1, several synthetic approaches were introduced, such as sol-gel,
hydrothermal, and liquid phase deposition. As for sol-gel method, the liquid sol could be
transformed into thin films, wet gel, dense ceramics, aerogel, fibers, and powders through
various techniques.4 However, it generally produces TiO2 with irregular and disordered
structures when no surfactant additives or templates are used.5 Nanoparticles prepared by
hydrothermal usually require high temperature and pressure. One of the traditional
methods to prepare mesoporous/nanoporous structures of titania is through the
hydrothermal method where TTIP was added into nitric acid and went through
precipitation, hydrolysis, 8-hour peptization, filtration and hydrothermal treatment for 12
hours more in autoclaves. The final product was rotary-evaporated to a TiO2 slurry.6 An
alternative approach7 is to prehydrolyze titanium n-butoxide/ethanol mixture in a dilute
aqueous nitric solution, aging for 14 days, drying, grinding, washing and annealing. An
improved method reported by Yu et al. used high-intensity ultrasonication to control the
condensation and agglomeration of pre-made titania sol particles. Spherical particles with
sizes of 100-200 nm packed with 10-nm small particles and narrow channels were
fabricated, but still lacked the porosity order.8 Liquid phase deposition is a versatile

38

method to prepare titania coatings, but could hardly control the porosity of the particles
grown on the film.
Traditional methods are time-consuming and lacking ordered structures, while
template-directed methods generally pick up structural and morphological properties
duplicated from numerous templates. Templates offer the possibility of creating 1D, 2D,
and 3D9 materials such as wires, tubes, rods, sheets, and interconnected hierarchical
structures. Just for nanotubes alone, it offers controlled scale of diameters via different
templates. There are still some controversial issues such as the extra cost of the template
material, complicated fabrication process, and the possibility of destroyed tube or rod
morphology during the synthesis process. This work will focus on improving templating
methods and overcoming current difficulties for a better solution.
Diblock or triblock copolymers templates are a type of extensively-studied organic
templates which make ordered mesostructures, directing the growth of metal oxides in
non-hydrolytic environment.4 Common amphiphilic poly(alkylene oxide) block polymers
such as Pluronic P-123 and Pluronic F-127 are typically dissolved in ethanol, followed by
the addition of a TiCl4 precursor and gelation in air at low temperature for several days.
Calcination will remove the organic templates, leaving a mesoporous structure. However,
these methods usually require control of the cooperative assembly, longer gelation time,
complicated pH manipulations and the collapse of TiO2 frameworks due to the phase
transition.4
Hard-templating, or nano casting, has been developed to use preformed mesoporous
solids as templates via precursor infiltration, conversion of the precursor to metal oxide,

39

and removal of the hard template. Preformed solids could be silicates,10 aluminas,11 or zinc
oxide,12 with mesoporous structure of replicated nanowires or nanospheres. The precursor
must be able to be incorporated into the channels or pores of the template by adsorption,
ion exchange, or covalent grafting.13 Difficulties are existing in filling the micro channel
vacancies of the templates and finding a suitable precursor with high solubility and low
thermal stability.5 For example, titanium precursors tend to precipitate and crystallize in
aqueous solutions, clogging penetration channels. On the other hand, if the precursor
concentration is too high or too low, either bulk material with a low surface area or
insufficient template coating will be formed. At the right amount of precursor to template
ratio, it’s possible to obtain continuous framework of TiO2 porous structures from a
successful duplicate of the hard template removal by dissolving or etching.
2.1.3 A modified sol-gel method with calcium carbonate template
As mentioned above, various techniques have been developed for the synthesis of
meso/nano porous titania. However, macro-structured solids remain more limited.14 The
goal of the project is to develop a comprehensive titania structure with macro/meso/micro
pores available for photocatalytic applications. In this chapter, a new series of hard
templates were proposed for macro/meso/micro porous titania synthesis. Calcium
carbonate has been prepared through some other templates,15-17 and also worked as
templates for synthesis of gold,18 DNA capsules,19 and protein microspheres,20 now it is
used as an easy-to-remove template itself with acidification dissolving.
The best example of natural calcium carbonate composite is sea urchin skeletal plates,
which have a bicontinuous morphology with 10-15 micrometer pores. A great amount of
work has been done with sea urchin plates by first filling the porous network with
40

materials “precursor” to form a composite structure and then dissolving away the CaCO3,
leaving a duplicate morphology of the original plate. The sea urchin skeletal plate could be
used directly as the template, or used to prepare a duplicate polymer scaffold as a
chemically stable polymer template for future use.14

Figure 2-1 SEM images of (a) a cleaned sea urchin skeletal shell (b) sea urchin shell coated with
gold and annealed.18

Figure 2-2 SEM images of gold-coated sea urchin skeleton after acid dissolution (a) at lower
magnification, (b) at higher magnification. 1 marks the smooth inner surface, while 2 marks the
rough outer surface of the gold structure.18

An early report18 used echinoid (sea urchin) skeletal plates as templates for porous
gold preparation, with pore diameters in the order of 15 um as Figure 2-1 (a) shows. Gold
paint made with organo-stabilized gold particles was coated and annealed with the
skeleton, of which the rough surface and structures were duplications from the original

41

bio-skeleton with similar pore size and the scaffold (Figure 2-1(b)). Removal of CaCO3 was
fulfilled by dissolution of the composite in acid. (Figure 2-2)
To prepare a stable polymer duplicate, methyl methacrylate, and ethyl acrylate at a
ratio of 13: 7 was mixed with 1% w.t. initiator to immerse the sea urchin plates. After
vacuum treatment, curing and template removal, a sponge-like polymer membrane (Figure
2-3) was produced with identical sea urchin structure.14

Figure 2-3 SEM images of titania produced by hydrolysis-templating method on polymer replica of
sea urchin by (a) titanium tetrachloride, (b) titanium ethoxide.14

Through hydrolysis method, silica and titania could also be synthesized from either
direct method with sea urchin plate, or indirectly with polymer duplicates (Figure 2-5). As
for titania, the plate could not be used if the precursor is acidic. In this case, the polymer
scaffold was immersed in titanium ethoxide or titanium tetrachloride, followed by at least 3
days of hydrolysis in a sealed desiccator.14

Figure 2-4 Scheme of synthesis procedures of spherical hollow silica .24

42

Commercial CaCO3 at smaller scales have been applied as templates for porous silica24
and titania synthesis25,26 as well. Calcium carbonates with cubic and needle-like shapes in
nano-size ranges were applied as the sacrificial templates. The hydrolysis of sodium silicate
left a continuous coating of SiO2 on the surface of CaCO3 cores to form a core-shell structure.
With further product calcination and dissolving of CaCO3 in HCl dilution solution,24 hollow
silica structures were obtained.(Figure 2-4)

Figure 2-5 TEM images of hollow silica particles using different types of CaCO3: (a) cubic CaCO3, (b)
needle-like CaCO3. 24

Titania hollow spheres were prepared with the same type of nano-sized calcium
carbonate, through a sol-gel method with tetrabutoxytitanium (TBOT), annealing,
acidification, washing and further high-temperature calcination.25

Figure 2-6 TEM images of (a) nano-sized calcium carbonate particles, (b) TiO2 with hollow
structure. (c) SEM images of hollow structure of titania.25

43

Templating with biomineral sea urchin provides a route to produce various inorganic
solids with unique and complex morphologies. The drawbacks of these patterned
structures are the non-adjustable shape and structure, too large pore size, and the fragility.
Applying industrial-level calcium carbonate with smaller sizes will provide more
opportunities for scale-down titania and by varying the type of CaCO3 used, the surface
area, pore distribution and related applications could be adjusted.
Based on inorganic polymerization reactions, sol-gel chemistry is an attractive wet
chemistry approach to synthesize metal oxides with the utilization of hard templates.

5

Metal alkoxides of transition metals such as Ti, V, Zr are highly reactive and typically tend
to form oxyhydroxide precipitates when exposed to water. Thus, controlled hydrolysis and
condensation rates will be beneficial for a templating method that involves chemical
bonding or physical deposition. To control the hydrolysis rates, utilizing chelating ligands
such as glycols, organic acids, and beta-diketones is a very promising and simple route.
Chelating ligands provide occupations of the coordination sites of the alkoxide, thus
reducing the hydrolysis rates.28 A variety of carboxylic acids were tested to study their
chelating capacities on titanium(IV) isopropoxide to produce sol-gel-derived thin films.28
Valeric acid was proved effective to modify transition metal alkoxides and help produce
uniform thin films from the precursor on glass substrates.29 Thus a titanium valerate
became our promising candidate for titanium resource.
In this chapter, a new method (Figure 2-7) with modified sol-gel approach utilizing
titanium tetraisopropoxide and valeric acid, was combined with three different types of
commercial calcium carbonate as templates, to prepare various titania porous particles
with high surface area and diffusion path for photocatalytic degradation of organic
44

pollutants. Potential practical usage is very promising due to the low cost of calcium
carbonate, and the feasible process for large-scale fabrication and industrialization.

Figure 2-7 Scheme of calcium carbonate templating synthesis method for porous titania particles

Experimental
2.2.1 Materials and equipment
Materials:
All chemicals were used without pre-treatments unless noted. Three types of calcium
carbonates, Hubercarb W4, W3N, and Optfill JSU, were obtained from Huber Engineered
Materials. All three Hubercarb products have very few silicas and silicates (≤0.5%)
compared to other series, median particle sizes were 5.9 um, 2.8 um, 1 um respectively for
W4, W3N, and Optfill JSU (Opt). Titanium(IV) isopropoxide (≥ 97%), valeric acid (≥ 99%),
methyl orange (≥ 85% dye content, certified by Biological Stain Commision), titanium(IV)
oxide nanopowder (anatase, 99.7%, metals basis, <25 nm), ethanol, (absolute 200 proof, ≥
99.5%, A.C.S. reagent), 2-propanol (A.C.S. grade, ≥ 99.5%) were purchased from Sigma
Aldrich. Hydrochloric acid (analytical reagent grade) was from Mallinckrodt. Plain, precleaned micro glass slides were purchased from VWR International, in a size of 1 inch by 3
inch, with a thickness of 1.2 mm. Free sample of nanosized CaCO3 powder, was from
generous Ruicheng Xintai Nanomaterials Technology Co. Ltd, with average size 15-40 nm.

45

Syringes (1mL and 5mL) and PTFE (polytetrafluorethylene) syringe filters (0.2 um filter
size) were purchased from Corning.
Equipment and Characterizations:
(1) The structures of the calcium carbonates templates and titania porous particles
were characterized by scanning electron microscopy (Carl Zeiss Auriga CrossBeam SEMFIB). Samples were prepared by mounting onto a sample holder with carbon paste/tape
before being tungsten-coated. EDAX™ EDS Analysis System for SEM–X-ray Microanalysis
was applied together with SEM characterization for element analysis on TiO2 powders as
prepared. When a cross-sectional view was necessary, a focused ion beam (FIB) was used
to mill surfaces of the materials.
(2) The transmission electron microscopy (TEM) images, high-resolution transmission
electron microscopy (HRTEM) images, and selected-area electron diffraction (SAED)
patterns were taken on a JEOL 2200FS equipment. TiO2 samples were dispersed in ethanol
by bath sonication for 0.5 h. The ethanol/TiO2 suspension was dropped and dried on a
carbon-coated 400 mesh copper grid overnight prior to TEM analysis. Under HRTEM mode,
fast Fourier transfer (FFT) patterns were utilized to optimize image quality with
DigitalMicrograph (DM) software.
(3) Flackteck shear mixing machine was used for rapid mixing and grinding of
materials such as colloids, fluids, powders, resins, inks and paints. A single wafer spin
processor (WS-400A-6NPP/LITE) was purchased from Laurell Technologies Corp and used
for film coating. The doctor blade method was completed the MSK-AFA-III compact tape
casting coater with vacuum chuck. This doctor blade machine has an adjustable

46

micrometer film applicator (thickness 0.01-3.5 mm, or fixed thickness at 5 um, 10 um, 15
um and 20 um) with a 100-mm width, as well as a dryer cover that heats up to 100 °C.
(4) UV-Vis spectrophotometer (Perkin Elmer Lambda 900 UV Spectrometer) was
applied for methyl orange dye characterization to monitor the absorption of the dye
solution, for monitoring the degradation level during photocatalytic experiments.
(5) A LabRAM ARAMIS laser Raman spectrometer (Horiba Scientific) with 532 nm 50
mW DPSS laser was used to inspect sample crystal phases as well. All the samples were
tested as is with a 50X objective lens. The resolution in Raman spectrometer could be
decided by the spectrometer focal length, diffraction grating, laser wavelength, and the
detector. The higher the groove density of the grating is, the higher the spectral resolution.
The gratings used in this work is 1800 gr/mm and results in a spectral resolution of about
1.5 m-1. The laser power at the sample ranged in 1-3 mW, and the 1/e laser spot size was
about 2 Îźm at the sample.
(6) Thermogravimetric analysis (TA Instruments Q500) used a high-resolution ramp of
25 °C/min to ~800 °C in air or nitrogen was conducted on TiO2 samples. Surface area and
pore size data were collected with Brunauer–Emmett–Teller (BET) Micromeretics ASAP
2020.
(7) X-ray photoelectron spectroscopy (XPS) was carried out on Perkin Elmer Phi 5400
using KÎą line of Mg ( 1254 eV) X-ray source with a beam voltage of 15 kV and current of
20 mA for the survey spectra scan (pass energy 187.85 eV, and 1 eV step size). For detailed
Ti 2p or O 1s high-resolution spectra, a pass energy 29.35 eV and step size 0.1 eV were
used. Chamber vacuum was maintained below 1×10-7 Torr (1.33 Pa). The sample to be

47

tested was mounted onto the sample holder with double-sided copper tape, and as for
powder form samples, a thin but dense layer of powder was spread on the tape with excess
removed. After the measurement, data processing was performed by MultiPak V8.0. If not
mentioned, the spectra have been resolved into chemically shifted components by a nonlinear least squares procedure.
(8) Photocatalytic reactions were conducted with an RC-500 Xenon UV lamp (Xenon
Corp., Wobrun, WA) that has 300 W power, 120 Hz frequency, and a 5-inch linear bulb
shape. The lamp was equipped with aluminum reflectors and fused quartz windows to
provide optimum transmission of UV radiation. The lamp housing was made of metal, and a
forced air convection removed the generated heat. The pulsed xenon arc lamp has a
continuous radiation spectrum as shown in 30. The active time span of each pulse is 100 us,
while the interval rest time is 8.3 ms (Figure 2-8 (b)). The range of active UV wavelength
lies between 230 and 360 nm. Applying a Pyrex glass cover as a filter will remove
ultraviolet wavelength region beneath 290 nm.

48

Figure 2-8 (a): the scheme of the UV lamp system (Ozone was not applied in this work). (b): the
schedule of the pulsed wave. (c): wavelength range of the outputs from the pulsed xenon UV lamps.
(d): UV transmittance of Surasil, quartz, and Pyrex glass filters[source: Xenon Corp.

2.2.2 Synthesis of porous titania particles with CaCO3 templates
A simple and versatile method was developed to fabricate size-controllable titania
particles through a sol-gel deposition on calcium carbonate templates. At first, a sol was
prepared with titanium(IV) isopropoxide (TTIP), valeric acid (VA) and water, where valeric
acid was used as a chelating agent to moderate the hydrolysis and precipitation rates
during the sol-gel reactions. Typically, a molar ratio of 1: 9: 2 was used for TTIP: VA: H2O.
As for a 0.004 mol Ti content, while 1.97 mL TTIP was being stirred, slowly add 3.95 mL VA,
followed by 0.144 mL water. The mixture was kept stirring at room temperature for more
than 1 hour. Then 5% w.t. calcium carbonates were added into the mixture and stirring
continued until homogeneity formed. To further improve mixing comprehensively, the
mixture was treated with Flackteck shear mixing machine by consequently 1 minute at
1000 RPM, 1 minute at 2000 RPM, 1 minute at 3000 RPM, and 2000 RPM for another 2

49

minutes. Then the mixture was moved forward to bath sonication for 10 minutes. After a
final 2-minute Flackteck mixing at 2000 RPM, the Ti(IV)-CaCO3 mixture was annealed in
crucibles at 450 °C for 5 hours. The calcined mixture was added into 1 M hydrochloric acid
solution and kept stirring overnight. By filtering the dissolved Ca2+ away with 0.22 um
filtering membrane, and washing, rinsing for several times, clean TiO2 was dried in a
vacuum oven at 80 °C overnight.
Film casting trials were made by both dip coating and spin coating with the precursortemplate mixture. For dip coating, a 1-inch square glass slide was immersed in the mixture,
and withdrawn for layer formation. Spin-coated films were finished on the spin coater with
different time and spin speed settings.
2.2.3 Photocatalytic performance experiment
A calibration curve for the concentration-absorption relationship of methyl orange
(MO) solution was obtained by conducting a series of UV-Vis measurements on diluted MO
solutions. Methyl orange standards were prepared from a 5×10-4 M aqueous solution. The
analysis used water as a reference and monitored the dynamic range between 200 and 800
nm. Absorption data for calibration was taken at peak 465 nm and a calibration curve was
generated
The photocatalytic performance measurements were conducted in a Pyrex glass
crystallizing dish (70 mm in diameter, 50 mm in height, 180 mL in capacity), equipped with
a magnet, and stirring plate. The whole process was conducted in open air at room
temperature in a light blocking box. An Approximate 1×10-5 M MO solution was prepared
and used as the initial solution. For 50 mL as-prepared MO/water solution, 0.2 mg catalyst

50

was added to the solution, followed by a one-hour stirring in darkness to reach the
adsorption-desorption equilibrium before any UV irradiation. The initial starting point was
taken as t_0, at the end of the equilibrium period, and before the UV lamp was turned on.
The change of the concentration of MO/water solution was monitored by taking 3mL
solutions out every 5-minute UV treatment. Solids of catalyst were removed by utilizing a
set of syringe/filter (Figure 2-9) so that clear solution could be ready for UV-Vis spectrum
collecting. Most experiments were run for 3 times to achieve statistically stable data for
error calculations.

Figure 2-9 (a) 1 mL syringe with syringe filter connection. (b) A simple presser for separation of
liquid and solid in MO solution/ TiO2 catalyst mixture. The clear solution was collected at the
bottom of the syringe in a 5 mL vial.

The blank dye test was performed in order to prove that UV light would not damage
methyl orange. As-prepared new MO solution was tested at t_0, and measured again after
5-minute UV treatment with a Pyrex petri dish lid covered (t_1), and then 5-minute UV
treatment without any coverage (t_2), and 5 minutes more as t_3. This experiment was to
prove the low-wavelength UV damage to the dye, and the filtering effect from Pyrex glass
for an proper photocatalytic experiment setup.

51

Results and Discussions
2.3.1 Titania particles morphology from different templates
As Figure 2-7 shows, the Ti (IV) sol was prepared with TTIP and chelating agent VA
with very limited amount of water. During this process, hydrolysis was controlled at a very
slow rate, which facilitated the even coating of calcium carbonate during the mixing steps.
Direct calcination was used to crystallize the Ti sol/calcium carbonate hybrid, and porous
structures were created after the acid dissolution of the calcined product. Without CaCO3
templates, a simple calcination yielded very dense nanoparticle chunks as seen in Figure
2-10 (a), while the addition and removal of template resulted in highly porous structures as
(b-d) indicated. The acidification removed calcium carbonate away, and the titania
displayed as large particles (diameter around 20 um) with smaller pores (< 5 um) on the
surface and in the inner connected structure based on the (d) cross-sectional image. H+ was
able to penetrate into the inner part of the calcined hybrid along previously formed pores
or channels to remove all of the templates, leaving quite pure titania with high crystallinity.
Pores came not only from the original W4 particles but also broken W4 particles due to
mixing and sonication before calcination. The controlled hydrolysis rate was the key to the
successful coating of calcium carbonate templates, which allowed enough time for the
mixing and depositing.
In order to study the porosity distribution within the particles, the focused ion beam
was used to cut into the larger particles. Figure 2-11 further reveals the structure of the
internal pores at the milled cross-sectional surface, pore sizes varied between 1 um to 5 um,
and the inner surface was very rough, offering an even higher surface area for
photocatalytic reactant adsorption.
52

Figure 2-10 SEM images of titania products induced with and without CaCO3 W4 templating (5%
w.t.). (a) precursor only, no CaCO3 added; (b) with 5% w.t. W4 templated and (c) at higher
magnification, (d) sample (b) at cross-sectional view assisted by FIB.

Figure 2-11 A closer view of the cross-sectional inside the porous particles, resulted from 5% w.t.
W4.

To compare with the resulted titania and the original templates, the Hubercarb W
series calcium carbonates were imaged as well. Figure 2-12 (a) (b) (c) show the size
distribution of three types Hubercarb W series CaCO3 with median sizes 5.9 um for W4, 2.8
um for W3N, and 1 um for Opt. All the calcium carbonates used here share low silica

53

content and are rounded particles with low moisture and chalk-like softness. The material
has a magnesium carbonate content of 0.4%, while the rest composition is made of 0.02%
crystalline silica and 0.05% silicates. Thus, the majority of the pores created by templates
matched the templates’ sizes.

Figure 2-12 SEM images of calcium carbonate templates: (a) W4, (b) W3N, (c) OPT. (all samples
were mounted on carbon glue tape on metal stubs)

Increasing the carbonate loading led to smaller particles and pores at the same time.
Figure 2-13 shows the structures and morphologies of titania induced with 10% w.t. and
15% w.t. W4 template usage. On the left side, with 10% w.t. CaCO3, the particle size was
reduced to 5 um, and the pore size became ~1 um. On the right side, the particle size was 2
um and pore size remained similar to ~1 um. The decrease of particle and pore size were
attributed to the fact that a large amount of templates existing in the acid dissolution
process would be removed and resulted in more fractures than when 5% w.t. of template
was used. Pores smaller than the template sizes were due to the mixing and sonication of
the mixture precursor before calcination. Additionally, it is suspected that increased
amount of templates tended to be removed together as a whole, leaving more fractured
particles with smaller templates which are easier to hold onto the duplicate shapes.

54

Figure 2-13 Comparison between 10% w.t. and 15% w.t. W4 templates in the titania fabricating
process. (a) (c) (e) represents 10% w.t., and (b) (d) (f) are 15% w.t.

As introduced in Chapter 1, immobilized films also act as efficient photocatalysts. To
find out whether this mixture is suitable for film formation, two coating methods were
applied on clean glass substrates. The Ti (IV) sol mixed with 5% w.t. and 10% w.t. W4
CaCO3 were deposited as films, dried in air for 1 hour and then calcined in the same

55

procedure as with the bulk materials. Figure 2-14 (a-d) demonstrate the morphologies of
the dip-coated films at lower magnification (a), higher magnification (b-c), and tilted crosssectional surface (d) with an approximate thickness of 1.5 um. Film formed with many
large cracks. A porous structure was observed on the surface and from the cross-section. It
was hard for the templating precursor mixture to form a uniform homogeneous film due to
the size of the templates, weak adhesion between the carbonates and the substrate, and the
removal of the template during acid digestion. This effect was intensified when the
templates loading was increased to 10% w.t. (Figure 2-14 (e-f)). Films were not able to
form due to the removal of calcium carbonate, though the porous vacancy was duplicated
from the templates.

56

Figure 2-14 SEM images of films made with a mixture of Ti precursor and W4 template by dip
coating: (a-d) with 5% w.t. W4 CaCO3, (e-f) with 10% w.t. W4 CaCO3.

Spin coating was also tried with several different speeds for a 45-second period. The
thickness of the films seemed to follow the spin coating thickness equation ∝

1
√𝑤

, where w

is the angular velocity. Spin coating mechanism could be explained based on Figure 2-15,
where the substrate was coated with drops of the suspension (molecules dissolved in a
solvent). The high speed applied next will drive most of the solvent off the side, and the

57

airflow will make the majority of the solvent evaporate. The centripetal force and the
surface tension of the solution keeps an even coating layer. The spin speed is a critical
parameter to the radial force applied to the liquid, the velocity, and the turbulence of the
air above the film surface. The exact thickness of a film is also determined by the material
and solvent evaporation.

Figure 2-15 Spin coating working mechanism of a small molecule in solution started with a static
dispense

The effect of spin coating speed on film quality is seen in Figure 2-16. At 300 RPM, the
rotation was too slow to form a complete film. While at 500 and 1000 RPM, both films were
coated. When the rate was increased to 3000 rpm, the film became very thin and very few
templates were still in the film. However, after calcination, none of the films show a good
duplication of template morphology. This indicated that the centrifugal force was too
strong for the materials to maintain a good template dispersion.

58

Figure 2-16 SEM images of films made with a mixture of Ti precursor and 5% w.t. W4 template at
different spin speeds for 45 seconds: (a) 3000 RPM, (b) 1000 RPM, (c) 500 RPM.

The elemental composition was also studied for spin-coated film structures and
surprisingly, there is a quite larger amount of Ca (2.1% atomic) residue compared to the
bulk-prepared porous material (0.27% atomic). This might be attributed to the insufficient
contact of acid with film due to the adhesion, or the background glass substrate’s
contribution to Ca element (1.0% atomic from averaged data). The calcium and other

59

contaminants should be maintained at a minimum amount, due to the need for pure
crystalline titania for more active sites on the surface.
Experiments on very fine calcium carbonate were conducted to find out whether nanosize of the template would bring new beneficial features of the porous titania. Thus a new
kind of calcium carbonate with only 15-40 nm size was tried in the same protocol
described to synthesize porous TiO2 particles at a 5% w.t. content. Nano-sized CaCO3
particles were very fine, but not providing a stable reproducibility of the materials. Some
batches (10% w.t. of nano-sized CaCO3 as templates) of materials had a brown color in
some areas with a very high Ca content of 10.0% atomic. Images of the microstructures are
shown in

32,

where (a-b) and (c-d) came from 5% and 10% w.t. Increasing the nano-sized

carbonates loading led to different pore structures and particle distribution. Structures
shown in (c-d) had finer particles and porosities, while (a-b) only existed in large chunks
and pores in random shapes. When films were deposited by dip coating, larger cracks were
observed.
It was found that the highly dispersed nano-sized CaCO3 had been modified with
certain unknown surfactants to be hydrophobic. After being calcined at a higher
temperature (450 °C) for 12 hours, the calcium carbonate particles turned out to be
hydrophilic again, which meant the surfactant was heat-removed. The removal of the
surfactant led to the aggregation of the particles as shown in

32.

After the surfactant was

removed, a new batch of TiO2 was synthesized with aggregated TiO2 and was not ideal for
porous materials due to the much larger chunks of TiO2 and pore limitations.

60

Figure 2-17 SEM images of (a-b) TiO2 powders and the cross sectional view with 5% w.t. nano-sized
CaCO3 templating; (c-d) TiO2 powders with 10% w.t. nano-sized CaCO3; (e-f) films by dip coating
with 5% w.t. nano-sized CaCO3.

Figure 2-18 SEM images of 15-40 nm CaCO3 particles , (a): as-received, and (b): after calcination
450 °C for 12 hours.

The difference between W4 CaCO3 and nano-sized CaCO3 was compared through
Raman and TGA. Figure 2-19 demonstrated the crystallinity of the CaCO3 samples. Nano61

sized calcium carbonate has aragonite phase mostly, while W4 shared a very similar peak
distribution but with an elevated curve trend.34 TGA data from Figure 2-20 showed that at
around 380 °C, around 5% w.t. dropped for nano-sized CaCO3 while W4 sample maintained
the mass until the temperature reached 655 °C. Nano-sized CaCO3 started to degrade at
600 °C, and the residue was almost 53-55% w.t., while W4 CaCO3 was left a little above, at
56% w.t. This decomposition at ~650 °C transformed calcium carbonate into calcium oxide
and carbon dioxide.35

Figure 2-19 Raman spectra of CaCO3 templates of nano-sized type and W4 type. According to
Calcite 282, 713, and 1086 cm-1, and Aragonite 155, 207, 704, 1085 cm-1

Thermal analysis(Figure 2-21) was also performed on the titania particles made with
W4 and nano-sized CaCO3. It turns out there was a weight loss potentially from nano-sized
CaCO3 at around 600 °C. The CaCO3 made with W4 went through dehydration at the
beginning, possible loss of minor contaminations without apparent loss observed at either
380 °C or 600 °C, indicating the absence of suspicious organic surfactant in W4 and the
complete removal of calcium in titania particle products.

62

Figure 2-20 TGA of the top red curve: W4 CaCO3 particles, middle green curve: nano-sized CaCO3
after calcination, and bottom blue curve: nano-sized CaCO3 as-received without any treatment.

Figure 2-21 TGA of titania particles made with: (top blue)-5% w.t. W4 CaCO3, (middle-red)- 10%
w.t. nano-sized CaCO3, and (bottom green)- 5% w.t. nano-sized CaCO3. (Here the nano-sized CaCO3
was not pre-treated.)

63

The crystallinity of titania particles is an important factor regarding the properties of
titania and the activity of photocatalysis. Luttrell et al. found that anatase is a better
photocatalyst than rutile because the charge carriers excited deeper in bulk and
contributed to the surface reactions. Raman spectroscopy was used to characterize titania
particles prepared with W4 CaCO3 at three different content ratios. The inelastic scattering
on titania happens with energy exchange between phonon and the substance, resulting in a
change in the wavelength. The features in Raman spectra were due to the vibrations of
molecules (atoms/ions). For crystals, the crystal lattice vibrates as a whole, and if the phase
is amorphous, the spectra turned to be more blurred rather than sharp and distinct due to
the disordered structure and reduced areas of spatial correlations between the oscillations
of the particles. Broadening of the peaks might be due to the defective crystals. Thus Raman
spectroscopy could be used to identify and quantify the amorphous and crystalline TiO2
phases, also the oxygen defects, stress states, and size effects.
Both anatase and rutile titania has a tetragonal structure. For anatase (space group
D194h-I41/amd), there are six Raman active modes 144 cm-1 (Eg), 197 cm-1 (Eg), 399 cm-1
(B1g), 513 cm-1 (A1g), 519 cm-1 (B1g), and 639 cm-1 (Eg). Rutile (space group D144hP42/mnm) has four active modes 143 cm-1 (B1g), 447 cm-1 (Eg), 612 cm-1 (A1g), and 826
cm-1 (B2g).36 Raman scattering of different morphologies differs as well. The scattering of
1D materials provides information about the subtle changes to the structure and the
symmetry of the primitive unit.37 A new peak at 266 cm-1 for titania nanotubes due to the
Ti-OH bonds within a tubular structure as reported by Qian et al.37 This is explained as the
intermediate formation of hollow tube structures.

64

As shown in Figure 2-22, all three samples have peaks representative of the anatase
phase. Peaks at the wavenumber of 149, 399, 513, 639 cm-1 were detected, while the peaks
at 197 and 519 cm-1 were weaker to be observed. Anatase is tetragonal with two TiO2 units
per primitive cell. The six Raman modes are related to symmetries A1g+2B1g+3Eg.
Sharper peaks were observed for 5% w.t. W4 sample, which could be explained by more
pristine TiO2 corresponding to the EDS result mentioned above.

Figure 2-22 Raman spectra of anatase TiO2 made from the different amount of W4 CaCO3 templates.

65

Figure 2-23 SEM images of titania particles templated from: (a-b) 5% w.t. Opt CaCO3 and (c-d) 5%
w.t. W3N CaCO3 .

After the investigation of the influences induced from W4 and nano-sized CaCO3 on the
morphologies and weight percentage directing porosity, two more templates W3N and Opt
CaCO3 were applied as well. Following same experimental procedures, porous TiO2 was
created after acid dissolution, washing, filtering and drying. Figure 2-23 shows different
morphologies compared to TiO2 made from W4 templates and nano-sized CaCO3. As for (ab), particle sizes ranged from 5 um to 20 um, with pore size around 1-2 um; while for (c-d),
particle sizes ranged from 5 um to 10 um, with pore size around 2-4 um.

66

Figure 2-24 Raman spectra of anatase TiO2 made with three different CaCO3 templates. From top to
bottom: W4, W3N, Opt, all in 5% w.t. content

Raman was collected to confirm the anatase phase of the TiO2 particles from three
templates, and the main peaks were summarized in Table 2-1. The peak at 197 cm-1 was
again too weak to be detected, especially for W4 and W3N templated samples. The peak at
around 149 cm-1 refers to the lowest frequency Eg phonon mode of anatase and it blue
shifted from W3N, to Opt, to W4 templated TiO2. This blue shift might be due to oxygen
deficiency.37
Table 2-1 Characteristic peaks of anatase TiO2 obtained with three different CaCO3 templates

TiO2
templated with

Eg(cm-1)

Eg(cm-1)

B1g (cm-1)

A1g+B2g(cm-1)

Eg(cm-1)

W4

149.3

-

399.1

517.8

639.8

W3N

148.4

-

400

516.9

640.6

Opt

149.1

199

398.5

516

643

67

Based on transmitted electrons rather than scattered electrons, TEM could provide not
only morphology information at much higher resolution but show the crystallization,
diffraction, crystal faces and lattice space. The surface atomic information is relevant to the
photocatalytic performance of titania. As described previously, three main types of TiO2
exist as anatase, rutile, and brookite. Anatase is the most stable phase for nanoparticles
smaller than 11 nm, while rutile is the most stable phase for nanoparticles above 35 nm.38
Anatase in tetragonal bipyramidal shape typically has two main surfaces of ~94% (101)
and ~6% (001). Faces (101) and (001) have theoretical surface energy as 0.44 J m-2 and
0.90 J m-2. Low surface energy of the face could be maintained during the crystal growth to
keep the total surface energy minimum.39 There are only 50% Ti atoms consisting of fivecoordinated Ti5c for face (101), while 100% for (001). Ti5c sites contribute actively to
photocatalytic reactions.40 It was found that the photogenerated h+ preferred to accumulate
on rutile (101) and anatase (001), while the photogenerated e- tended to rest on rutile (110)
and anatase (101)41 In summary, the difference of energy levels in conduction band or
valence band led to various atomic arrangement in facets, and eventually controlled the
electron-hole pair and influenced the charge separation.39
TEM images were obtained with each type of particle dispersed in ethanol. Figure 2-25
(a-b) show the particle and porosity of the TiO2 and (c) is the diffraction pattern at selected
area obtained at lower magnification. The dimensions of the particles shown in TEM
matched the data found in the SEM observations. Clear rings of polycrystallinity were
displayed from inner to outer circle as (101), (004), (200), (105), and (204) planes. In the
image (d), particles were connected in an ordered pattern with a mean crystal size of

68

5.8Âą0.5 nm, while (e) and (f) further reveal the lattice spacing of 0.36 nm, and 0.24 nm,
corresponding to facets (101) and (004).

Figure 2-25 TEM images of 5% w.t. W4 CaCO3 templated TiO2: (a) at low magnification, (b) at
increased magnification, (c) selected-area electron diffraction pattern, (d) individual particles in an
area with crystal size ~ 5 nm, (e) HRTEM image noted with anatase lattice spacing ~0.35 nm, and
0.36 nm, (f) HRTEM image noted with anatase lattice spacing ~0.24 nm.

As shown in Figure 2-26 (a-b), HRTEM images of detailed pores in TiO2 nanoparticles
were shown. The pore size observed in TEM ranged from 30 nm to 100 nm and mean
crystal size is 5.3Âą0.4 nm. Image (b) again matched the 0.352 nm spacing of lattice (101).
Fast Fourier transform (FFT) is used to transform an image from the spatial domain to
reciprocal space.42 For single crystals, the FFT image will show spots, and for
polycrystalline, it will show diffuse spots or rings with lattice periodicity. When defects are
present, FFT will show random spots together along the crystallographic directions, while
when the sample is all amorphous, the FFT will be non-defined diffuse halos. In Figure 2-26

69

(c), the FFT pattern shows the (101) and (004) diffractions of anatase TiO2, and the 101
planes have a ring shape due to random orientation.43

Figure 2-26 HRTEM images of 5% w.t. W3N CaCO3 templated TiO2: (a) pore formation in a particle
(b) at increased magnification, noted with anatase lattice spacing ~0.35nm, (c) corresponding FFT
pattern of the single crystal area (d) selected-area electron diffraction pattern

70

Figure 2-27 HRTEM images of 5% w.t. Opt CaCO3 templated TiO2: (a) pore formation in a particle
(b) at increased magnification, (c) corresponding FFT pattern of the [101] and [004] facets, (d)
noted with anatase lattice spacing ~0.35 nm

Last but not least, TiO2 particles made with Opt CaCO3 was found to have porosities
within particles, plus with smaller sizes. The mean crystal size of TiO2 was 5.0Âą0.5 nm.
Similar with Figure 2-26, the corresponding FFT pattern and the lattice spacing of 0.35 nm
further confirmed the crystal phase of anatase.

71

Figure 2-28 XPS data of TiO2 fabricated with 5% w.t. OPT CaCO3 templates (a) survey spectra, (b-c)
high-resolution XPS spectrum for: Ti 2p and O 1s

The surface chemical states of Ti and O atoms in the titania particles were determined
by XPS. Ti 2p was presented as peak at 457.8 eV and O 1s at 529.4 eV(75.8%), 530.9
eV(16.6%) and 532.0 eV(7.6%). The largest portion of the O 1s at 529.4 eV stands for metal
oxide bond, confirming that the sample is TiO2. The peaks at 530.9 and 532.0 eV might
come from C-O, or C=O bond due to adventitious carbon. Ca 2p3/2 at around 347 eV was
zero from the spectrum obtained, and the further high-resolution analysis shows the
atomic concentrations of C 1s, O 1s, and Ti 2p are 38.2%, 44.9%, and 16.9%. The peak at
around 295 eV was possibly because of the adventitious carbon. XPS has an analysis depth
of 4-5 nm, so the presence of atoms was only detected when they are present at the surface
72

thickness, while the other atoms will be eliminated. EDS result of W3N samples showed
that the amount of calcium left in the sample is 0.2 atomic %. The summary of the XPS and
EDS concentrations was concluded in Table 2-2. Calcium existed as a small amount of
residue and the difference between XPS and EDS data was due to the surface depth of each
equipment.
Table 2-2 Atomic concentrations of C, O, Ca and Ti by spectra of XPS and EDS

C (%)

O (%)

Ca (%)

Ti (%)

TiO2 prepared
with templates

XPS

EDS

XPS

EDS

XPS

EDS

XPS

EDS

5% w.t. W4

47.0

45.8

42.5

36.8

0.5

0.4

10.0

13.6

5% w.t. W3N

36.5

42.4

49.7

20.2

0.5

0.3

13.23

14.0

5% w.t. Opt

38.2

48.6

44.9

37.0

0

0.3

16.9

11.7

Table 2-3 Atomic concentrations of high-resolution fitted peaks by XPS of template-induced titania

TiO2 prepared with
templates
5% w.t. W4

5% w.t. W3N

5% w.t. Opt

O 1s peak and integrated
percentage (eV)
529.46

74 %

531.14

18.8 %

532.62

7.2 %

529.45

77.5 %

530.88

17.1 %

532.05

5.5 %

529.37

75.8 %

530.87

16.6 %

532.00

7.6 %

Ti 2p (eV)

457.85

457.85

457.79

The N2 adsorption–desorption experiment demonstrated that the surface area of the
TiO2 ranges 196-238 m2/g, and the average pore width was 46.5-47.7 Å. TiO2 synthesized
with OPT templating has the highest surface area, finest nanoparticles, small average pore

73

size, and also small most distributed pore size, as summarized in Table 2-4, Figure 2-30
and Figure 2-29.
Table 2-4 BET surface area and pore size summary

BJH adsorption Highest distributed
average pore pore diameter from
diameter (Å)
adsorption(Å)

Average
nanoparticle
size (Å)

Sample

BET surface
area, m²/g

TiO2, from
5% W4

196.7

47.7

55.9

305.1

TiO2, from
5% W3N

224.5

47.5

48.5

267.3

TiO2, from
5% OPT

238.8

46.5

42.6

251.3

Figure 2-29 Isotherm linear plot, isotherm log plot, isotherm pressure composition and BET surface
area plot via BJH adsorption/desorption of titania made with 5% w.t. Opt calcium carbonate
templates.

74

Figure 2-30 Pore size distribution of three titania powders made by calcium carbonate W4, W3N,
and Opt.

2.3.2 Photocatalytic performance of porous TiO2 templated with CaCO3
Methyl Orange (MO) (C14H14N3NaO3S)(327.34g/mol) is a very standard model
compound for accurate evaluation of TiO2 photocatalytic activity. It is also a general pH
indicator because the color will change from red to orange from acidic to basic medium.
The UV-Vis spectrum of aqueous MO solution shows two major peaks at around 270 nm
and 465 nm, and in this research, the 465 nm peak was utilized as the tracker of
concentration change. The theory is based on the Beer-Lambert’s law, and different
standard concentrations of MO aqueous solutions were tested to make a calibration curve
between absorption intensity and MO concentration.

75

Figure 2-31 Chemical formula of methyl orange

During the process of MO decomposition, the peak at 465 nm will decrease. MO could
be adsorbed on the surface of TiO2 through van der Waals force because it’s a polar
molecule. Besides, due to the electronegativity of the N, O, S atoms, hydrogen bonds could
be formed with the hydroxyls on TiO2 surface. The concentration of hydroxyl groups will
depend on the pH of the semiconductor and dye suspension. When it’s acidic, the surface
hydroxyl groups mainly exist as Ti-OH and/or Ti-H2+, providing chances for hydrogen bond
formation. When it’s basic, the surface hydroxyl groups mainly exist as TiO- and TiOH,
preventing the hydrogen bonds forming between MO and TiO2 surface. Thus only when pH
is lower, there are both van der Waals force and hydrogen bonds contributing to the
adsorption of MO, while when pH is higher, there are no hydrogen bonds helping. In this
research, the pH of the MO solution applied was kept as it is without the addition of acid or
base, and measured to be between 6 and 7.
Calibration curve was obtained and fitted into a linear line. The blank dye tests showed
that when no photocatalyst was added, there is very minimal degradation of the dye under
UV light with a Pyrex glass petri dish lid. However, when the lid was removed, UV
irradiation will degrade the dye significantly as shown by the blue curve in Figure 2-32, and
further degradation continued as the green curve showed. This matched Figure 2-8 (c-d),
due to the fact that wavelength < 275 nm was filtered by Pyrex glass lid. Previous work
mentioned that it was noted that when no photocatalyst was added, only 4% of the

76

degradation of methyl orange solution with UV irradiation could be observed,44 which
could be considered as a small enough amount to be neglected compared to the effects of
the photocatalysts.

Figure 2-32 Blank dye without catalysts tested with UV light and the comparison of the Pyrex lid
effects

By repeating photocatalytic experiments for each sample for at least 3 times, the
comprehensive comparison was concluded in Figure 2-33 and Figure 2-34. As shown in 45,
almost no concentration change was observed for the blank dye sample, during the 20minute treatment, and the only minimal change was probably due to the evaporation of
some solvent. The degradation levels were compared between the same weight of different
titania samples, such as pristine TiO2 prepared with only TTIP and VA sol, the commercial
anatase titania nanoparticles, and the TiO2 particles made with three types of CaCO3
templates.

77

Figure 2-33 Comparison of the photocatalytic capacities through [concentration C/ initial
concentration C0] vs. [time], from top to bottom, regarding to: (green) – no catalyst,
(olive) – pristine TiO2 without templates, (red) – TiO2 from 5% w.t. W4 CaCO3
templating, (black) - TiO2 from 5% w.t. W3N CaCO3 templating, (pink) – commercial
anatase TiO2 nanoparticles, and (blue) - TiO2 from 5% w.t. Opt CaCO3 templating.

By utilizing a pseudo-first-order kinetic model, the log[C] vs. time was further plotted
in Figure 2-34. Linear fitting results gave the values of the reaction rate constant k as 0.07
min-1 for TiO2 from W4, 0.11 min-1 for TiO2 from W3N, 0.14 min-1 for TiO2 from Opt, 0.05
min-1 for TiO2 made with sol only with templates, and 0.11 min-1 for commercial anatase
TiO2.

78

Figure 2-34 Relationship between lnC and time, based on ln(C) = -kt+ln(C0) (k has unit of min-1)

The rate constant of the best-performed TiO2 templated by 5% w.t. Opt was 27%
higher than the commercial anatase TiO2 particles. The W3N templated product also
reached the same level as the commercial one. The highest surface area and largest pore
volume explain the reason for the best performance of 5% w.t. Opt templated TiO2. The
macro/micro/meso porosities of the particles were different from traditional particles, and
probably provided more reflection, scattering of light harvesting during the photocatalytic
process, while at the same time, the improved surface area offered larger dye adsorption
and active spots for electron transport.

Conclusions
A simple and cost-effective modified sol-gel CaCO3 templating method was developed
to create macro/micro/nano porous anatase titanium dioxide particles in different sizes,

79

morphologies, and porosities. A complete removal of CaCO3 removal was confirmed
through various of characterizing methods such as SEM, TEM, Raman, XPS, and EDS.
Applications in photocatalysts were investigated with UV irradiation and compared among
TiO2 made without templates, with three different templates, and the commercial TiO2
particles. The results showed a promising 27% improvement compared to commercial
particles, and 180% increase compared to no-template induced TiO2.

Future Work
2.5.1 Adjusted CaCO3 size and structures to make TiO2 in different shapes
According to the previous report, CaCO3 could also be lab-made according to a
procedure developed for cross-linked DNA capsules.19 0.33M aqueous sodium carbonate
(Na2CO3) solution was rapidly poured into 0.33 M calcium chloride dehydrate (CaCl22H2O
) solution at same volume under vigorous stirring and mixed for 30 seconds. The product
CaCO3 was washed twice by centrifugation, filtration, and drying process. Further removal
of CaCO3 was obtained in a mild condition with ethylenediaminetetraacetic acid (EDTA).
The resulting porous CaCO3 microparticles have a diameter around 4 um, and worked as an
adsorbent for diverse macromolecules such as polyelectrolytes, proteins, and
polysaccharides. FE-SEM images confirmed that after the acid etching, a hollow crosslinked DNA capsules were made without collapsing. One thing to notice is that homemade
CaCO3 size and morphology may vary due to the solution concentration, reactant addition
speed, and the mixing method.

80

2.5.2 Mixed phase TiO2
Degussa P25 is a type of titania powder composed of 70% anatase and 30% rutile TiO2.
The surface area of P25 is around 50 m2g-1. The photocatalytic efficiency of P25 is among
the top among commercial products, and this has been attributed to the different charge
separation of mixed phase was enhanced compared to anatase only.46 The energy of
anatase trapping site is lower than both anatase and rutile conduction band, so the
photogenerated electrons were observed to migrate from rutile to anatase phase. It is the
interfacial sites between two phases that lead to a more efficient charge separation and
photo-efficiency.
The current results were compared with anatase pure-phase commercial product due
to the consistency of the material crystallinity. Further work should be focusing on
adjusting the ratio of anatase and rutile phase of the products made from CaCO3 templating
and found the optimum parameters such as calcination temperature, time, through the
analysis of Raman, XRD, and photocatalytic degradations.

81

BIBLIOGRAPHY

82

BIBLIOGRAPHY

(1)

Pal, N.; Bhaumik, A. Soft Templating Strategies for the Synthesis of
Mesoporous Materials: Inorganic, Organic–Inorganic Hybrid and Purely
Organic Solids. Advances in Colloid and Interface Science 2013, 189-190,
21–41.

(2)

Shin, Y.; Exarhos, G. J. Template Synthesis of Porous Titania Using
Cellulose Nanocrystals. Materials Letters 2007, 61, 2594–2597.

(3)

FĂŠrey, G.; Haouas, M.; Loiseau, T.; Taulelle, F. Nanoporous Solids: How Do
They Form? an in Situ Approach. Chem. Mater. 2014, 26, 299–309.

(4)

Chen, X.; Mao, S. S. Titanium Dioxide Nanomaterials: Synthesis,
Properties, Modifications, and Applications. Chem. Rev. 2007, 107, 2891–
2959.

(5)

Li, W.; Wu, Z.; Wang, J.; Elzatahry, A. A.; Zhao, D. A Perspective on
Mesoporous TiO 2Materials. 2014, 26, 287–298.

(6)

BarbĂŠ, C. J.; Arendse, F.; Comte, P.; Jirousek, M.; Lenzmann, F.; Shklover,
V.; Grätzel, M. Nanocrystalline Titanium Oxide Electrodes for
Photovoltaic Applications. Journal of the American Ceramic Society 1997,
80, 3157–3171.

(7)

Liu, C.; Fu, L.; Economy, J. A Simple, Template-Free Route for the
Synthesis of Mesoporous Titanium Dioxide Materials. J. Mater. Chem.
2004, 14, 1187–1189.

(8)

Yu, J. C.; Zhang, L.; Yu, J. Rapid Synthesis of Mesoporous TiO 2 with High
Photocatalytic Activity by Ultrasound-Induced Agglomeration. New
journal of Chemistry 2002, 26, 416–420.

(9)

Kim, W.; Choi, S. Y.; Jeon, Y. M.; Lee, S.-K.; Kim, S. H. Highly Ordered,
Hierarchically Porous TiO 2Films via Combination of Two SelfAssembling Templates. ACS Appl. Mater. Interfaces 2014, 6, 11484–
11492.

(10)

Fuertes, AB. Low-Cost Synthetic Route to Mesoporous Carbons with
Narrow Pore Size Distributions and Tunable Porosity Through Silica
Xerogel Templates. Chem. Mater. 2004, 16, 449–455.

(11)

Lakshmi, B. B.; Dorhout, P. K.; Martin, C. R. Sol-Gel Template Synthesis of

83

Semiconductor Nanostructures. Chem. Mater. 1997, 9, 857–862.
(12)

Zhuge, F.; Qiu, J.; Li, X.; Gao, X.; Gan, X.; Yu, W. Toward Hierarchical TiO2
Nanotube Arrays for Efficient Dye-Sensitized Solar Cells. Adv. Mater.
2011, 23, 1330–1334.

(13)

Bagheri, S.; Mohd Hir, Z. A.; Yousefi, A. T.; Abdul Hamid, S. B. Progress on
Mesoporous Titanium Dioxide: Synthesis, Modification and Applications.
Microporous and Mesoporous Materials 2015, 218, 206–222.

(14)

Yue, W.; Park, R. J.; Kulak, A. N.; Meldrum, F. C. Macroporous Inorganic
Solids From a Biomineral Template. Journal of Crystal Growth 2006, 294,
69–77.

(15)

Guo, F.; Li, Y.; Xu, H.-X.; Zhao, G.-Q.; He, X.-J. Size-Controllable Synthesis of
Calcium Carbonate Nanoparticles Using Aqueous Foam Films as
Templates. Materials Letters 2007, 61, 4937–4939.

(16)

Chen, L.; Hu, F.; Zhou, Q.; Sun, J.; Zhang, H. Growth of Hierarchical Calcium
Carbonate Crystals Templated by Biomolecules of Lotus Root. Crystal
Research and Technology 2009, 44, 1289–1292.

(17)

Jiang, J.; Zhang, Y.; Yang, X.; He, X.; Tang, X.; Liu, J. Assemblage of NanoCalcium Carbonate Particles on Palmitic Acid Template. Advanced Powder
Technology 2014, 25, 615–620.

(18)

Meldrum, F. C.; Seshadri, R. Porous Gold Structures Through Templating
by Echinoid Skeletal Plates. Chem. Commun. 2000, 29–30.

(19)

Fujii, A.; Maruyama, T.; Ohmukai, Y.; Kamio, E.; Sotani, T.; Matsuyama, H.
Cross-Linked DNA Capsules Templated on Porous Calcium Carbonate
Microparticles. Colloids and Surfaces A: Physicochemical and Engineering
Aspects 2010, 356, 126–133.

(20)

Volodkin, D. V.; Klitzing, von, R.; Moehwald, H. Pure Protein Microspheres
by Calcium Carbonate Templating. Angew. Chem. Int. Ed. 2010, 49, 9258–
9261.

(21)

Yu, H.; Yu, J.; Cheng, B.; Lin, J. Synthesis, Characterization and
Photocatalytic Activity of Mesoporous Titania Nanorod/Titanate
Nanotube Composites. Journal of Hazardous Materials 2007, 147, 581–
587.

(22)

Yu, H.; Yu, J.; Cheng, B.; Liu, S. Novel Preparation and Photocatalytic

84

Activity of One-Dimensional TiO 2hollow Structures. Nanotechnology
2007, 18, 065604.
(23)

Pourmand, M.; Taghavinia, N. TiO2 Nanostructured Films on Mica Using
Liquid Phase Deposition. Materials Chemistry and Physics 2008, 107,
449–455.

(24)

Chen, J.-F.; Wang, J.-X.; Liu, R.-J.; Shao, L.; Wen, L.-X. Synthesis of Porous
Silica Structures with Hollow Interiors by Templating Nanosized Calcium
Carbonate. Inorganic Chemistry Communications 2004, 7, 447–449.

(25)

Shao, L.; Chen, J.-F.; Luo, L.-L.; Gao, L.-D. Synthesis of Hollow Titania Using
Nanosized Calcium Carbonate as a Template. Chemistry Letters 2005, 34,
138–139.

(26)

Hou, Q.; Tao, X.; Yang, Y.-J.; Ma, Y. Optimal Synthesis of Mesostructured
Hollow Titania Nanotubes Templated on CaCO3 Nanoparticles. Powder
Technology 2010, 198, 429–434.

(27)

Kumar, P. M.; Badrinarayanan, S.; Sastry, M. Nanocrystalline TiO2 Studied
by Optical, FTIR and X-Ray Photoelectron Spectroscopy: Correlation to
Presence of Surface States. Thin Solid Films 358, 122–130.

(28)

DUNUWILA, D. D.; GAGLIARDI, C. D.; BERGLUND, K. A. Application of
Controlled Hydrolysis of Titanium(Iv) Isopropoxide to Produce Sol GelDerived Thin-Films. Chem. Mater. 1994, 6, 1556–1562.

(29)

Askeland, P. A. Preparation and Characterization of Sol-Gel Derived Metal
Oxide Thin Films and Powders for Coatings and Catalysts, 1997.

(30)

Dozzi, M.; Selli, E. Specific Facets-Dominated Anatase TiO2: FluorineMediated Synthesis and Photoactivity. Catalysts 2013, Vol. 3, Pages 455485 2013, 3, 455–485.

(31)

Imai, H.; Takei, Y.; Shimizu, K.; Matsuda, M.; Hirashima, H. Direct
Preparation of Anatase TiO2 Nanotubes in Porous Alumina Membranes. J.
Mater. Chem. 1999, 9, 2971–2972.

(32)

Al-Qaradawi, S.; Salman, S. R. Photocatalytic Degradation of Methyl
Orange as a Model Compound. “Journal of Photochemistry & Photobiology,
A: Chemistry” 2002, 148, 161–168.

(33)

Ulery, A. L.; Drees, R. Methods of Soil Analysis: Mineralogical Methods.

(34)

Parker, J. E.; Thompson, S. P.; Lennie, A. R.; Potter, J.; Tang, C. C. A Study of
85

the Aragonite - Calcite Transformation Using Raman Spectroscopy ,
Synchrotron Powder Diffraction and Scanning Electron Microscopy.
CrystEngComm 2010, 12, 1590–1599.
(35)

L’vov, B. V. Mechanism and Kinetics of Thermal Decomposition of
Carbonates. Thermochimica Acta 2002, 386, 1–16.

(36)

Frank, O.; Zukalova, M.; Laskova, B.; KĂźrti, J.; Koltai, J.; Kavan, L. Raman
Spectra of Titanium Dioxide (Anatase, Rutile ) with Identified Oxygen
Isotopes (16, 17, 18). Physical Chemistry Chemical Physics 2012, 14,
14567–14572.

(37)

Qian, L.; Du, Z.-L.; Yang, S.-Y.; Jin, Z.-S. Raman Study of Titania Nanotube
by Soft Chemical Process. Journal of Molecular Structure 2005, 749, 103–
107.

(38)

A Fujishima, X. Z. D. A. T. TiO2 Photocatalysis and Related Surface
Phenomena. 2008, 63, 515–582.

(39)

Xu, H.; Ouyang, S.; Liu, L.; Reunchan, P.; Umezawa, N.; Ye, J. Recent
Advances in TiO2-Based Photocatalysis. J. Mater. Chem. A 2014, 2,
12642–12661.

(40)

Xu, H.; Reunchan, P.; Ouyang, S.; Tong, H.; Umezawa, N.; Kako, T.; Ye, J.
Anatase TiO2 Single Crystals Exposed with High-Reactive {111} Facets
Toward Efficient H2 Evolution. Chem. Mater. 2013, 25, 405–411.

(41)

Ohno, T.; Sarukawa, K.; Matsumura, M. Crystal Faces of Rutile and
Anatase TiO2 Particles and Their Roles in Photocatalytic Reactions. New
journal of Chemistry 2002, 26, 1167–1170.

(42)

Mathew, S. S.; Ma, S.; Kretzschmar, I. Three-Dimensionally Ordered
Macroporous TiO2 Electrodes: Fabrication of Inverse TiO2 Opals for
Pore-Size-Dependent Characterization. J. Mater. Res. 2012, 28, 369–377.

(43)

Masuda, Y.; Kato, K. Anatase TiO2 Films Crystallized on SnO2:F Substrates
in an Aqueous Solution. Thin Solid Films 2008, 516, 2547–2552.

(44)

Li, Y.; Li, X.; Li, J.; Yin, J. Photocatalytic Degradation of Methyl Orange by
TiO2-Coated Activated Carbon and Kinetic Study. Water Research 2006,
40, 1119–1126.

(45)

Phase-Pure TiO2 Nanoparticles: Anatase, Brookite and Rutile. 2008, 19,
145605.

86

(46)

Li, G.; Chen, L.; Graham, M. E.; Gray, K. A. A Comparison of Mixed Phase
Titania Photocatalysts Prepared by Physical and Chemical Methods: the
Importance of the Solid–Solid Interface. Journal of Molecular Catalysis A:
Chemical 2007, 275, 30–35.

87

3 Templating with Natural Cellulose and Micro-Fibrillated
Cellulose (MFC)
3.1 Background
3.1.1 Natural templates for inorganic materials
As introduced in Chapter 2, diblock or triblock co-polymers have their limitations, and
nano-scale hard templates are facing some crucial challenges as well. Applying calcium
carbonate as the hard template yields macro/meso/micro pores within TiO2 particles and
turned out to be excellent photocatalysts. For better defined nanostructures and easier
template removal, new interest was transferred to natural biodegradable templates at
nano scale.
Using nature as guide to solve scientific challenges in material design is not a new
thing. The complex and elaborate structures of biological materials provide tremendous
opportunities in the preparation of functional inorganic materials with special
morphologies and shapes for novel applications. Utilizing natural templates to build up
specific 2D or 3D structures has become an emerging attempt in “bottom-up” synthesis of
inorganic micro/nano-materials.
Nanocrystals, nanowires and other hierarchical architectures with sizes ranging from
1-25 nm, match perfectly with the scale of certain types of biological molecules. Natural
templates such as cotton fibers1-3, viruses4, butterfly wings,5 eggshell membranes,6
bacteria,7 pollen grains,8 etc. served as templates, and are easily removed by elevated
thermal degradation. Hollow inorganic materials made using these templates have low

88

density, high surface area, and controllable optical and electric properties. Figure 3-1
demonstrates several examples of silica and titania nanomaterials derived from various
natural templates.

Figure 3-1 SEM images of nanomaterials made with bio-structure templates: (a) hollow silica made
with bacteria Escherichia coli (KP7600);7 (b) porous titania made with dandelion pollen grains;8 (c)
TiO2 hollow fibers made with cotton wools;1 (d) hierarchical interwoven titania prepared with
eggshell membrane. (reprinted with permission from reference6)

Among different kinds of natural templates, cellulose fibers have many advantages.
They are the most abundant renewable and biodegradable polymer derived from various
sources with an annual production of 1011-1012 tons.9 Since its first discovery and isolation
in 1838, cellulose have been extensively researched and applied in paper, composite,
pharmaceutical, and food industries.10 As a potential template, it has multiple conveniences
including: an abrupt thermal decomposition, being environmentally friendly, a large
surface area for material deposition, abundant surface hydroxyl groups for chemical

89

bonding and very low cost. Being the most abundant organic compound derived from
biomass on earth,11 cellulose is of interest to scientists due to its availability,
biocompatibility, renewability, biological degradability, and sustainability.
Cellulose (C6H10O5)n is a long-chain polysaccharide consisting of several hundred to
over ten thousand β-1,4-linked D-glucose rings. Each D-glucose ring is corkscrewed 180°
with respect to its neighbors, and cellubiose is the repeating segment with non-reducing
end and reducing end on each side as Figure 3-2 shows.10 The monomer anhydroglucose
unit (AGU), carries three free hydroxyl groups: one primary hydroxyl on C6, and two
secondary hydroxyls on C2 and C3. All these hydroxyl groups along the chain help form the
intra- and intermolecular hydrogen bonds,11additionally, provide numerous sites for
further chemical modification. The polymer chains of cellulose are made of fibrous
structures called microfibrils or nanofibrils, which are associated by hydrogen bonds and
van der Waals forces.12 The hydrogen bond network gives cellulose structures rigidity and
is used as a structural building block for many plants. Figure 3-3 is a brief synopsis of how
cellulose fibers are harvested from plants.

Figure 3-2 Structure of cellulose. (reprinted with permission from reference10)

Nano-scale cellulose has gained tremendous attention because it retains the original
bulk material properties as mentioned above, and also presents high strength, high
stiffness comparable to those of inorganic fibers, low weight,13 and nano-size effects.

90

Figure 3-3 Cellulose sources, cell wall, and fibers. The structure level demonstration of
cellulosic fiber. (reprinted with permission from reference 11)

Typically, nanofibrils have a size of 2-10 nm in diameter (varies depending on the
origin) and several tens of micrometers in length.12 From various cost-effective sources
(wood, cotton, potato, wheat, soy, coconut husk), they are obtained by methods including
physical treatments such as cryocrushing, grinding, high-pressure homogenizing, highspeed blundering, ultra-sonicating etc., and chemical treatments such as concentrated acid
hydrolysis, biological treatments and oxidations.12 A combination of the physical and
chemical treatments removes wax, extractives, hemicelluloses and lignin chemically prior
to grinding or homogenizing afterwards. 14
Although there are numerous ways to describe nano-scale cellulose, e.g., cellulose
nanofibers, microfibers, nanocrystals, nanowhiskers, etc., a recent review of cellulose11

91

described the classification in regards to the processing steps to prepare them. Similarly, a
standard terminology will be used for the fibers used in this work.
(1) cellulose nanocrystals, termed as “CNC” in this thesis. These CNCs have a diameter
at 2-20 nm, and length ranging from 100 nm to 2 Îźm, which are prepared from
wood, sea plants, or sea animals. From the original cellulose fiber suspensions, it
goes through chemical treatment and dialysis, yielding different sizes of
nanocrystals. (Figure 3-4)

Figure 3-4 TEM images of cellulose nanocrystals derived from (a) tunicate, (b) bacterial, (c) ramie,
and (d) sisal. (reprinted with permission from reference10

(2) microfibrillated cellulose, termed as “MFC”, stands for cellulose fibers with a
diameter of 20-60 nm due to a higher extent of aggregation of those 2-10 nm
microfibrils. For MFCs that are only processed mechanically, the length is beyond 10
Îźm, while for MFCs with pre-treatment and mechanical treatment, the length is ~2
Îźm.(Figure 3-5)11

92

Figure 3-5 TEM image of MFC gel showing cellulose nanofibrils and nanofibrils bundles (reprinted
with permission from reference10,13

In the coming chapters, terms CNC and MFC are distinguished with “microfibrils” or
“nanofibrils”, the non-aggregated version.13 A brief summary of the diameter, length and
aspect ratio data for above cellulose structures is seen below.
Table 3-1 Summary of dimensions of cellulose structures



Cellulose structure

Diameter (nm)

Length (nm)

Aspect ratio (L/d)

Microfibril

2-10

>10,000

>1000

Microfibrillated cellulose
(MFC) *

20-60

>10,000

>100

Cellulose nanocrystal (CNC)

2-20

100-20,000

10-100

The microfibrillated cellulose discussed in this table is processed by mechanical treatments
only. partially adapted from.13

Thus, MFC materials may be composed of (1) nanofibrils, (2) fibrillar fines, (3) fiber
fragments and (4) fibers. This implies that MFC is not necessarily synonymous with
nanofibrils, microfibrils or any other cellulose nano-structure. However, properly produced
MFC materials contain nano-structures as a main component, i.e. nanofibrils.15

93

3.1.2 Titania materials fabrication through cellulose templating
Cellulose contains 31.48% w.t. hydroxyl groups, with one primary and two secondary
in each AGU. The hydroxyls on the surface of cellulose are the starting point of any chemical
functionalization of cellulose. When a cellulose molecule is fully extended, it forms a flat
ribbon with –OH groups protruding laterally. The hydrogen bonds are formed inter- and
intramolecularly, where the intramolecular bonding between adjacent anhydroglucose
rings could improve the linear integrity of the polymer chain and affects the reactivity of
the hydroxyl groups especially of the C-3.9 Due to the existence of hydrogen bonding,
cellulose fibrils tend to become aggregated. As Figure 3-3 shows, the amorphous regions
are more reactive than the crystalline regions. The degree of crystallinity (ratio between
amorphous and crystalline) of cotton cellulose, for instance, is about 40-45%. In addition,
the reactivity of the hydroxyls will be influenced by the reaction medium as well. 9
Cellulose is usually suspended in water or organic solvents such as ethanol or
isopropanol for further interaction with the inorganic precursor for titania deposition. Also,
it could start as a dry substrate for immersion or soaking in the precursor systems. For presuspended cellulose, an ordered layer of water molecules cover the cellulose fibers via
hydrogen bonding, and the outside layer was composed of disordered water molecules. If
the cellulose was treated with solvent exchange, with up to 99% of the original water
removed, the outside layer would be replaced by alcohol, and the inside first layer will
retain an ordered network of hydrogen-bonded water molecules.16 For the second situation,
the dry cellulose surface hydroxyls especially the three active ones will interact with the
hydroxyls in the precursor and form bonding between the Ti(IV) precursor and the
template.

94

Inorganic material templated with MFCs or CNCs obtain desired porosities via
precursor coating and calcination owing to the rigidity and the thermal stability of the
cellulose. Different types of titania or titania/cellulose composites have been reported as
follows.
3.1.2.1 Tubular and fibril titania
Titania hollow microfibers were prepared by coating TiO2 onto hydrophilic cellulose
fibers through heterogeneous nucleation1 with ammonium hexafluorotitanate and boric
acid through an acid-assisted supersaturated aqueous solution. Pure cotton wool was
immersed into this solution, washed, dried, and calcined into hollow titania fibers (Figure
3-6).

Figure 3-6 SEM images of (a) TiO2 fibers after template removing (b) cross section of one hollow
fiber (c) TiO2 nanofibers after milling and ultra-sonication to make a paste (d) XRD diagram of the
fibers.1

An extra step of mixing the above-prepared TiO2 into an applicable doctor blade paste
was required for dye-sensitized solar cell assembly. An efficiency of 7.2% was reached
under solar illumination and electron transport properties were enhanced compared to
traditional spherical nanoparticles. Another work discussed how titania nanotubes were
derived from laboratory filter paper by surface sol-gel coating process. A further
95

hydrothermal step created a needle-like deposition of rutile TiO2 along the surface of
anatase nanotubes.17
Bamboo cellulose fibers were used as a template to make a millimeter-long TiO2 fibers
composed of 30 nm spherical nanoparticles2 through titanium tetrachloride in an aqueous
system. With a molar ratio of TiCl4: (NH4)2SO4: HCl at 1: 2: 10, the additives here served as
promoter for the anatase phase. A finishing annealing at different temperatures (500, 600,
700, 800, 900 °C) in air removed the cellulose fibers and produced TiO2 microfibers (Figure
3-7). Superior photocatalytic activity was demonstrated in degradation tests of a phenol
solution. Materials that were calcined at 500, 600, and 700 °C achieved higher
photodegradation efficiency compared to standard TiO2 material Degussa, P25 at the same
conditions.

Figure 3-7 (a-d) Low magnification SEM image, photograph, magnified SEM image and diameter
distribution with Energy-dispersive X-ray spectrum of pure bamboo cellulose fiber. (e-h) Low
magnification SEM image, photograph, and magnified SEM image and diameter distributions of TiO2
product calcined at 500 °C.2

3.1.2.2 Immobilized TiO2 on cellulose
Coatings of self-cleaning titania can be used as anti-pollutants and anti-bacterial
materials especially with its versatile applications on substrates like metal, glass, or even
plastic, papers, and cotton fabrics which have very low thermal resistance. Qi and
96

colleagues developed a simple dip-pad-dry-cure process to obtain anatase titania coated
cotton fabric at room temperature in a titanium tetraisopropoxide/acid based sol-gel
system.3 In a typical procedure, cleaned cotton fibers were dipped into a TiO2 sol (TTIP
added drop wise into 0.7% nitric acid containing ethanol and acetic acid, and heated to
different temperatures), neutralized in ammonia gas and dried at 80 °C for 5 min and
finally cured at 120 °C for 3 min. A uniform layer of coating was formed (Figure 3-8) and
XRD results revealed pure anatase phase. The bactericidal activities, colorant
decomposition, degradation of red wine and coffee stains under UV irradiation
demonstrated the excellent self-cleaning capacities of this organic/inorganic hybrid
nanomaterials.

Figure 3-8 SEM images of (a) pristine cotton fiber (b-d) coated cotton fibers, sol prepared at 25, 40
and 60°C 3

Chauhan et al. successfully immobilized 2.5 to 21.0% w.t. titania (diameter 10-20 nm)
on cellulose fibers through single-step hydrothermal method instead of multiple steps of
synthesis of TiO2 and then the incorporation onto the cellulose paper. Briefly, the cellulose
fibers were dispersed in water/ethanol, followed by a dropwise addition of titanium
isopropoxide, stirring, and autoclave treatment. Paper matrices were further prepared with

97

TiO2 -coated cellulose and degraded methyl blue and formaldehyde for photocatalytic
performance evaluation and antibacterial capacity through E. coli inhibition.
In next section, a simple and brief experimental trial would be conducted on easily
accessible lab cellulose materials with the scheme shown in Figure 3-9 to find the working
mechanism of the TiO2 deposition on cellulose surfaces and possible factors that govern the
templating and growth. In detail, cellulose (green cylinder) would be coated with a layer of
Ti (IV) precursor, and then went through high-temperature cellulose removal, ended up
with TiO2 hollow tubes/fibers. The Ti(IV) precursors used were based on sol gel, and liquid
phase deposition discussions from Chapter 1, and will be investigated in further details in
future chapters.

Figure 3-9 Scheme of preliminary TiO2 deposition on cellulose fiber templates

3.2 Experimental
3.2.1 Materials and equipment
3.2.1.1 Materials:
All chemicals were used without pre-treatments unless noted. Cotton rounds were
purchased from Delon Laboratories Inc., with 100% cotton content, bleached, chlorine-free,
pure without brighteners. Filter paper (medium porosity with 60 mL/min flow rate and 510 Îźm particle retention) was from Fisher with a diameter of 11 cm. Titanium(IV) fluoride,

98

titanium(IV) isopropoxide (≥ 97%), valeric acid (≥99%), boric acid (≥99.5%), ammonium
hexafluorotitanate (99.99%) were purchased from Sigma Aldrich. Hydrochloric acid
(analytical reagent grade) was from Mallinckrodt. anhydrous 2-propanol were obtained
from Sigma Aldrich. Ethanol (200 proof, 99.5% ACS grade) and isopropanol (99.5%) were
purchased from Michigan State university stores. Cellulose Celish KY100G (90% w.t. in
water) was purchased from Daicel, and has been refined only by physical processing
without chemical treatment, from highly refined, pure fiber raw materials. Another source
of cellulose fibers was purchased from the University of Maine process development center
with a water content of 97% w.t. Mineral oil (laboratory grade, heavy) was from Columbus
Chemical Industries. Plain, pre-cleaned micro glass slides were purchased from VWR
international, in a size of 1 inch by 3 inch, with a thickness of 1.2 mm.
3.2.1.2 Equipment:
(1)Homogenization of cellulose fibers was performed by two sets of instruments. IKA
T25 digital ULTRA-TURRAXÂŽ package with an input energy of 50W at 50/60Hz. A Mini
DeBEE ultrahigh-pressure homogenizer has a single-jet parallel flow pattern (flow rate up
to 250 mL/min) with an adjustable operating pressure of 2000 to 45000 psi, and a back
pressure of 0 to 5000 psi, which helps process fluids producing uniform particles size
reduction to nanoparticles, and high yield cell disruption. Each pass is cooled with a heat
exchanger.
(2) The structures of the macro-scale cellulose templates, titania hollow ribbons, MFCs
and CNCs were characterized by scanning electron microscopy (Carl Zeiss Auriga
CrossBeam SEM-FIB). Samples were prepared by mounting onto a sample holder with

99

carbon paste/tape prior to being tungsten coated. EDAX™ EDS Analysis System for SEM –
X-ray Microanalysis was applied together with SEM characterization for element analysis
on TiO2 powders as prepared.
(3) Thermogravimetric analysis (TA instruments Q500) used a high-resolution ramp of
25 °C/min to ~800 °C in air or nitrogen was conducted on macro-scale cellulose, MFCs,
CNCs, and cellulose-TiO2 sol samples. For instance, MFC samples were dried at 80-100 °C in
a vacuum oven overnight, and 0.2 mg sample was tested.
(4) A LabRAM ARAMIS laser Raman spectrometer (Horiba Scientific) with 532 nm 50
mW DPSS laser was used to inspect sample crystal phases as well. All the samples were
tested as is with a 50X objective lens.
(5) Fourier transform infrared (FTIR) absorption spectra were collected with Perkin
Elmer 2000 Spectrum One FT-IR spectrometer at room temperature. KBr Pellets were
prepared by Carver Laboratory press (model 2518) with grounded and pre-dried KBr
powder. For MFCs suspension, a drop was cast between two pellets forming a sandwich
structure with the excess liquid squeezed out and wiped off. The spectra of the composites
were scanned from 4000 to 400 cm-1.
(6) ) X-ray photoelectron spectroscopy (XPS) was using a Perkin Elmer Phi 5400. The
KÎą line of Mg ( 1254 eV ) X-ray source with a beam voltage of 15 kV and current of 20 mA
was the excitation source. The survey spectra scan were acquired with a pass energy
187.85 eV, and a 1 eV step size. For detailed C 1s or O 1s high-resolution spectra, a pass
energy of 29.35 eV and step size 0.1 eV were used. Chamber vacuum was maintained below
1×10-7 Torr (1.33×10-5 Pa). The cellulose suspension sample was prepared by drop

100

casting on a piece of clean glass substrate and then mounted onto the sample holder with
double-sided copper tape. After the measurement, data processing was performed by
MultiPak V8.0.
3.2.2 Preliminary synthesis of titania hollow ribbons with macro-size cellulose
A 0.04 M TiF4 aqueous solution was prepared by stirring for one hour at room
temperature with the pH adjusted to ~2.5 with diluted hydrochloric acid or ammonium
hydroxide aqueous solution. Around 0.07 g cotton rounds that were cut into 1 cm by 2 cm
small pieces were added into 50 mL TiF4 solution. In a separate vial, filter paper was cut
into 1 cm by 1 cm sheets and soaked in the solution as well. Then the vial was capped and
transferred to a pre-heated mineral oil bath at 65 °C on a stirring hot plate. The reaction
lasted 6 hours and the mixture was cooled to room temperature before further use.
Aqueous ammonium hexafluorotitanate (AHFT) (0.30 M) and aqueous boric acid (BA)
(0.10 M) were prepared into 20 mL solution each. The pH of the boric acid solution was
adjusted to ~2 by adding appropriate amount of HCl solution. Then the AHFT solution was
added to BA solution, with vigorous stirring for 1 h at room temperature, resulting in final
concentrations of 0.15 M for AHFT and 0.05 M for BA. The prepared stock solution (pH
~2.5) was added to the cut cotton wool and filter paper in different capped vials, and then
moved to a 50 °C oil bath to react for 6 hours.
A gentle rinsing was applied to remove excess Ti(IV) left on the surface of the cotton
wool or filter paper. Typically, three 100 mL beakers were filled with 80 mL RO water, and
the coated filter paper or cotton wool pieces were held by plastic tweezers and rinsed for 5
seconds in each beaker. After the rinsing process, all the coated substrates were dried in a

101

vacuum oven at 100 °C for 1 hour. Finally, the high-temperature annealing was performed
with step-wise temperature program at 200 °C, 300 °C, 350 °C and 500 °C for 0.5 hour each.
The multi-step temperature raising was based on the thermal properties of cellulose.
3.2.3 Preparation of homogenized MFCs
To obtain homogenously-dispersed MFCs, as-received cellulose fibers was treated in a
high-pressure homogenizer and made into isopropanol and aqueous suspensions (Figure
3-10). The process begins with creating a 5 mg/mL aqueous cellulose suspension by
diluting 167 g of cellulose with approximately 833 g water. The mixture was stirred
overnight. To improve dispersion, the mixture was pretreated with a kitchen blender and a
high circumferential speed disperser (IKA T25 digital ULTRA-TURRAXÂŽ ) at an input
energy of 500 W for 10 min. Then MFCs were obtained using a Mini DeBEE homogenizer,
processing the as-prepared suspension with a 0.008-inch zirconium nozzle at 16000 psi
(110 MPa) for 30 passes and subsequently a 0.005-inch diamond nozzle at 32000 psi (220
MPa) for 30 passes. The temperature of the process could reach very high due to the energy
gaining from the shear forces and it was maintained at around 50-70 °C by the heat
exchanger. In order to increase the degree of fibrillation, it might be necessary to repeat the
procedure of homogenization several times based on needs.
The non-aqueous MFC suspension was also prepared with an extra solvent exchange
step after stirring overnight, by filtering water out of the cellulose/water suspension
followed by triple washing/filtering with isopropanol. The subsequent steps remained
unchanged.

102

Figure 3-10 Scheme of MFCs suspension preparation

3.3 Results and Discussions
3.3.1 Titania hollow ribbons properties
Before the deposition experiment of Ti(IV) sol, both types of cellulose were observed
with SEM.

Figure 3-11 SEM images of (a) filter paper pieces, and (b) cotton rounds pieces

In Figure 3-11, the top view (a) of the filter paper was composed of fibers with various
diameter, including large aggregated fiber as thick as ~70 um, and smaller fibers with a
width of ~12 um. As for cotton rounds (b), the more fluffy structure made focusing more
difficult than the 2D filter paper, but it generally showed the diameter of the cotton fiber is
between 12 to 20 um. This is the starting point of the TiO2 deposition and crystal formation.

103

The first set of experiments were conducted on cotton rounds, including aqueous
liquid phase deposition methods TiF4 deposition and AHFT-BA deposition described in
Chapter 1. Results of the titania made with cotton rounds fibers and TiF4 are shown in
Figure 3-12, of which the low-magnification image presented fractured structure while a
high-magnification image (b) further exposed the possibly cracked hollow ribbons (width
~ 11.8 um) derived from the deposition of TiF4. The reason for the cracking of large
amount of coated hollow fibers are attributed to insufficient coating time, removal of
materials through washing, or unstable scaffold built up with this type of cotton fibers. The
starting cotton fibers before immersion was 0.12 g, and 0.1270 g after immersion reaction,
rinsing, and vacuum oven drying. After calcination, 92.7% w.t. of the weight was removed
being the degradable cellulose portion.

Figure 3-12 SEM images of TiO2 made from deposition of TiF4 on cotton rounds pieces: (a) lower
magnification, and (b) higher magnification of the framed section in (a).

The AHFT-BA system deposition on cotton rounds fibers was greatly improved. As
Figure 3-13 (a) shows, separate long fibers of cotton with width around 18 um were evenly
coated with Ti(IV) sol. At higher magnification (b) shows the rough surface of amorphous
nanoparticles of titania before high-temperature annealing.

104

Figure 3-13 SEM images of cotton rounds coated with AHFT and BA without further calcination: (a)
at lower magnification with selected dimensions (b) at higher magnification with surface features

Figure 3-14 SEM images of TiO2 made from cotton rounds coated with AHFT and BA after
calcination: (a) hollow ribbon structure of as-prepared TiO2 ; (b) at higher magnification with the
cross section of hollow ribbon wall, and the rough ribbon surface.

After thermal treatment, the inner cotton fibers were removed as Figure 3-14 shows.
Based on the comparison between Figure 3-13 and Figure 3-14, the width of the calcined
titania hollow ribbons were similar, further confirming the duplicating effect from the
cellulose. The ribbon thickness was estimated to be less than 0.5 um and mostly dependent
on the reaction time. Extra particles were formed on the surface of the ribbon due to the
excess Ti(IV) sol nucleation.
Similar experiments were also conducted on filter paper pieces for both TiF4 and
AHFT-BA systems. As Figure 3-15 showed below, the TiF4 coating on filter paper resulted
105

in fractured TiO2 again and very few long fiber-like TiO2 was observed. The deposition of
TiF4 solution needs to be modified for a better TiO2 formation and this will be discussed in
future chapters. A 67.2% w.t. loss was recorded from the immersed, rinsed and dried
product, compared to the original clean cotton fibers. After calcination, 91.6% w.t. mass
was eliminated.

Figure 3-15 SEM images of TiO2 made from filter paper coated with TiF4 after calcination: (a)
fractures of titania (b) surface of the as-prepared TiO2 at higher magnification.

A more successful coating was realized on filter paper with AHFT-BA solution, as
Figure 3-16 (a-d) shows. Figure 3-16 (a) and (b) show the deposition before calcination,
with excess formation of amorphous particles (diameter = 1-3 um) on the fiber surface. The
particles were formed randomly, due to the nucleation site activity of the fiber surface and
the coating reactivity. After calcination (c), there are bunches of hollow TiO2 tubes, and two
well-formed hollow titania ribbons with diameters of around 5 um and 15 um. The surface
was composed of ~200 nm nanoparticles. The original filter paper gained 83.1% w.t. after
the immersion reaction (rinsed and dried in vacuum oven). After calcination, 90.1% w.t.
was removed as C, O and H.

106

Figure 3-16 SEM images of filter paper coated with AHFT and BA: (a-b), before
calcination, at low and high magnifications highlighting the fiber layout and the
nucleation of amorphous TiO2 on the surface of the preliminary coating. (c-d), after
calcination, at low and high magnification demonstrating long hollow fibers and the
cross section of two TiO2 tubes.

Thermogravimetric analysis (TGA) measures the decrease in sample mass caused by
the release of volatiles or devolatilization during heating, and becomes the most widely
used thermoanalytical technique in solid-phase thermal decomposition filed. Different
types of signals are presented from the obtained raw data:
(a) mass vs. temperature, sample mass change over the applied thermal conditions,
named as TGA.

107

(b) –dm/dt vs. temperature, the first derivative of thermogravimetric curves over time,
known as derivative thermogravimetry (DTG), which provides the maximum
reaction rate.
In the 1970s, thermal decomposition of cellulose was extensively studied with DSC and
the mainstream mechanism consisting of 3 steps: (1) Dehydration of the cellulose which
produces anhydrocellulose, in the temperature range of 200-280 °C. The anhydrocellulose
is very active, and will result in CO2, CO, H2O and most of the residual char. (2) Non-reacted
cellulose goes through depolymerization during which levoglucosan appears as an
intermediate at 280 °C. (3) The levoglucosan might volatilize into gaseous products and
residual char in an endothermic way, or be trapped by the overlay of cellulose and
decompose exothermally, or a combination. 18
Others have discussed studies based on TGA analysis19 revealing that three stages of
mass loss were involved in the thermal decomposition. (1) Water evaporation: from room
temperature to around 120 °C, the loss of physically adsorbed water. (2) Oxidation of
cellulose: in the temperature range of 250-360 °C, the main rapid loss is due to dehydration
and decarboxylation. This will yield combustible gases like aldehydes, ketones, ethers and
so on. The maximum decomposition rate of cellulose shows up in this region. (3) Oxidation
of carbonaceous residues: From 360 to 450 °C, the loss of carbonaceous residue.
Lots of kinetic models have been studied for the cellulose pyrolysis since the early 20th
century, by performing a series of experiments under isothermal conditions at different
temperatures. But no consensus has been reached in terms of issues such as the actual
reaction temperature, the physical and chemical nature of the primary products, the
considerations of the intermediate products, and the order of the reactions etc.20-22
108

In theory, very kinetic model’s rate law obeys the classic fundamental Arrhenius rate
equation:
−𝐸𝑎
)
𝑅𝑇

𝑘(𝑇) = 𝐴 exp(

where T is the absolute temperature, A is the frequency factor, Ea is the activation
energy of the reaction, R is the universal gas constant, and k(T) is the T-dependent reaction
rate constant. When the kinetics of cellulose decomposition is routinely predicated as a
single reaction, the rate of the isothermal process is expressed as,
𝑑𝛼
−𝐸𝑎
= 𝑘(𝑇)𝑓(𝛼) = 𝐴 exp(
)𝑓(𝛼)
𝑑𝑡
𝑅𝑇
where 𝛼 is the degree of conversion;
since,
𝑑𝛼 𝑑𝛼 𝑑𝑡 𝑘(𝑇)
𝐴
−𝐸𝑎
=
=
𝑓(𝛼) = exp(
)𝑓(𝛼)
𝑑𝑇 𝑑𝑡 𝑑𝑇
𝛽
𝛽
𝑅𝑇
where

𝑑𝑡
𝑑𝑇

1

= , describes the inverse of the heating rate;
𝛽

Thus, for a reaction with (1 − 𝛼) remaining fraction of volatile material in the sample
and reaction order n,
𝑑𝛼
= 𝑘(𝑇)(1 − 𝛼)𝑛
𝑑𝑇
The cellulose pyrolysis process is still under debate with controversies till now.
However, three categories could be concluded out of the various pyrolysis kinetic models,
which are, single-step global reaction models, multiple-step models, and semi-global
models.21 Experimental conditions might influence the biomass reaction kinetics, such as
109

heating rate, particle size effects, surrounding atmosphere.21 A higher heating rate results
in lower char yields compared with lower heating rate probably because the increased
heating rate allows enough time for the completion of the decomposition. As for the
ambient gas atmosphere during the reaction, an oxidative environment brings out a more
active degradation reactions than in the inert atmosphere, lowers the active pyrolysis
temperature and boosts the combustion of chars at higher temperatures.21 It is also found
that the addition of inorganic salts has an influence on the decomposition of cellulose.19 It
will lower the temperature range for the second stage and the total heat of combustion for
phosphoramidate treated cellulose 23.
For this research, although not devoted to analyzing the pyrolysis kinetics, it is
important to have an adequate understanding of the basic decomposition theories so that
the templating role of the cellulose will be legible and definite in further discussions.
The thermal properties of cotton round fibers and filter paper were analyzed by TGA
(Figure 3-17). After losing the absorbed water content, cotton rounds began to degrade
rapidly at ~350 °C, and at ~375 °C 80% w.t. of the original weight was removed. A second
decomposition peak appeared at 500 °C and the material was completely removed at 525
°C. The filter paper started to degrade at ~325 °C and at ~375 °C, there was still 30% w.t.
left. The second decomposition peak appeared at 475 °C, then a total degradation was
reached at 500 °C.

110

Figure 3-17 TGA of (a) cotton rounds, (b) filter paper

From the TGA of the TiO2 coated fibers (Figure 3-18), the AHFT-BA coated cotton fibers
had a rapid weight drop at 295 °C from 91.0% w.t., reached 35.3% w.t. at 500 °C, and
remained 33.4% w.t. after temperature reached 600 °C. AHFT-BA coated filter paper began
losing weight at 89% w.t., reached 37.6% w.t. at 500 °C, and resulted in 34.8% w.t. in the
end. In addition, the TiF4 coated filter paper had a much lower percentage of mass as 10.5%
w.t. at 500 °C and 9.8% w.t. in the end. This explains the net TiO2 amount in the dried
coated cellulose, and further confirms a better deposition method as AHFT-BA compared to
TiF4.

111

Figure 3-18 Thermal analysis through TGA on non-calcined soaked cellulose substrates including
cotton round fibers and filter paper.

To understand the elemental composition of the samples before and after calcination,
EDS was conducted on sets of samples. Table 3-2 listed the summarized data for cotton
fibers, cotton fibers after deposition without calcination & with calcination, filter paper
after deposition without calcination, and with calcination. By measuring the X-ray
spectrum emitted by the solid sample resulted from the bombardment with a focused beam
of electrons, EDS could obtain qualitative and quantitative analysis of the element present
in the sample. Depending on the sample density and accelerating voltage of the incident
beam, the analysis depth is usually 0.5-3 um, where the secondary electrons are generally
formed. Cotton rounds were composed carbon and oxygen, and coated ones with AHFT-BA
were still detected with carbon, possibly due to uneven thickness (< 0.5um) of the coating
spot. After calcination, cellulose was removed, resulting in Ti and O, as Figure 3-19 shows.

112

Sharp and intense peak at 4.508 keV referred to Ti KÎą, smaller peak at 4.930 keV
corresponded to Ti Kβ, and peak at 0.452 keV referred to Ti Lι. The energy for O was
located at 0.522 keV. Similarly, when filter paper was used as the substrate for TiO2
deposition, C, O, Ti and N were observed, and this N KÎą (0.395 keV) was likely to come from
AHFT that was not rinsed off. Once the cellulose template was removed by annealing, the
composition became O and Ti as well.
Table 3-2 EDS atomic percentage of C, O, Ti and other elements in TiO2 materials

EDS data

C(%
atomic)

O(%
atomic)

Cotton rounds

73.6

26.4

Cotton rounds coated with
AHFT & BA, uncalcined

17.6

46.7

35.7

58.6

41.4

57.5

20.1

60.8

39.2

Cotton rounds coated with
AHFT & BA, calcined
Filter paper, coated with BA &
AHFT, uncalcined

16.2

Filter paper, coated with BA &
AHFT, calcined

Ti(%
atomic)

N(%
atomic)

6.2

Figure 3-19 EDS spectrum of AHFT-BA coated cotton round fibers, after calcination

The Raman spectra for the calcined titania samples (Figure 3-20) reveal the major
characteristic peaks of anatase TiO2 for all samples. The 145 cm-1, 199 cm-1, 395 cm-1, 515

113

cm-1 and 636 cm-1 corresponded to Eg, Eg, B1g, A1g or B1g, and Eg respectively.24 This
further confirmed the existence of TiO2 and the successful crystallization of anatase phase
with the removal of cellulose since no cellulose peaks were shown in the Raman spectra.
Anatase is the most effective crystal phase for potential applications in photocatalysis and
photovoltaics.

Figure 3-20 Raman spectra of TiO2 templated from AHFT and TiF4 systems on both cotton round
fibers and filter paper

3.3.2 Homogenized MFCs characterizations and benefits
Compared to non-processed cellulose fibers, microfibrillated cellulose (MFCs) has a
much higher surface area and aspect ratio, which make it a good candidate to form rigid
networks.11 A powerful physical method was developed to disintegrate cellulose materials
into a viscous and shear thinning gel-like suspension.
Since the first cellulose homogenizing process in 1985,25 the homogenizer has become
a well-developed instrument applying high shear to obtain viscous MFC suspensions at low
concentrations.26 In one cycle, a cellulose slurry is pumped at high pressure and goes
through a spring-loaded valve assembly, where the high pressure drop facilitates the
114

microfibrillation. By repeating the procedures, the resulting MFC consists of microfibrils
with different levels of entanglement and disorder, increased surface area, and extensive
hydrogen-bonding ability.
As mentioned above, the degree of crystallinity is an important factor of reactivity. It
was shown that ratio of crystalline region to amorphous region decreased as the number of
grinding passes increased from 1 to 30,11 which could be due to the homogenization of the
cellulose nanofibers under a high shear rate.
After homogenization, the original aggregated cellulose fibers were transformed into
20-60 nm wide, several micrometers long, but web-like network. This splitting process
greatly increased the surface area of the fibers. SEM observation of a drop of 5 mg/mL
cellulose water suspension coated and dried on a metal stub shows the web-like structure
in Figure 3-21. The high-pressure homogenizer treatment of the as-received cellulose
fibers help homogenize the large and aggregated fibers into reduced sizes and suspend
extensively in the solvent (water or IPA). Size reduction is observed in Figure 3-21, where
(a) to (c) demonstrate the effects of the homogenization process. Figure 3-21 (a) is at the
stage after the kitchen blender and a high circumferential speed disperser treatments,
where large aggregated bundles were visible. After 30 passes on the homogenizer, bundles
of fibers were gradually removed and growing amounts of fibers were distributed evenly.
At the end of homogenization (c), the cellulose fiber size was reduced to approximately 30
nm, and presented a well-dispersed MFC suspension. Homogenized MFCs are generally 3050 nm in diameter, with length in micrometers or longer. The degree of fibrillation was
increased simply by repeating the procedure of homogenization for more times. With more

115

energy input, the disintegration of cellulose was improved by increasing the number of the
passes from 2, 6, 14, 22 to 30 times.27
It is important to note that the homogenized product, mainly composed of microfibrils
bundles, cannot be redispersed simply by dissolving if it is dried to form microfibril
agglomerates.

Figure 3-21 Cellulose fibers treatment effects: (a) after the kitchen blender and a high
circumferential speed disperser treatments, (b) after 30 passes on the homogenizer, (c) after 60
passes on the homogenizer.

116

Figure 3-22 TGA of MFCs made by Celish KY100G (90 wt% in water) purchased from Daicel, with
water evaporated by 12-hour vacuum oven heating. TGA was conducted in air through a highresolution ramp program at a heating rate of 25 °C/min.

Figure 3-22 shows several stages of the heating of MFCs derived from Celish cellulose,
with mass vs. temperature in green color, and DTG, in blue color. There were rapid weight
losses at 200 °C, 300 °C, and 425 °C. A slight weight loss before 100 °C was due to the
removal of absorbed water molecules, and then between 200 and 300 °C, the dehydration
of the cellulose produced anhydrocellulose, rapidly converted to CO2, CO, and H2O. The
remaining carbonaceous residues went through further oxidation, which matched with the
last peak at 425 °C. At 430 °C, the remaining cellulose mass reached as little as 0.3% w.t.

117

Figure 3-23 TGA of non-homogenized Celish KY100G (90% w.t. in water) purchased from Daicel,
without further treatment. TGA was conducted in air through a high-resolution ramp program at a
heating rate of 25 °C/min.

The first DTG peak referred to the loss of the approximately 90% w.t. water when
temperature reached 62 °C as shown in Figure 3-23. Then no rapid loss was observed until
it was 325 °C. In comparison with homogenized and dried MFCs in Figure 3-22, DTG peaks
varied a lot. Since most of the cellulose researched with were in original forms or
decreased into micro-to-millimeter dimensions, it’s interesting to learn how the MFCs or
CNCs react in pyrolysis process. As a matter of fact, the thermal stability of the dried MFCs
decreased after homogenization and subsequent drying, besides, the char level at the end
of the pyrolysis was higher than for cellulose fibers. 26 It’s concluded that the formation of a
higher number of contacts between the homogenized microfibrils, microfibrils bundles,
and agglomerates, as a network, induces an earlier degradation.

118

Figure 3-24 TGA of dried MFCs purchased from the University of Maine process development
center with a water content of 97% w.t. To eliminate the water evaporation step during thermal
analysis, sample was dried in vacuum oven overnight at 80C before the test. TGA was conducted in
air through a high-resolution ramp program at a heating rate of 25 °C/min

Different types of homogenized cellulose carry different thermal stabilities. Figure 3-24
shows several stages of the calcination of MFCs, with mass vs. temperature in green color,
and DTG, in blue color. Between room temperature and 100 °C, the residue free water
content not removed from drying process was evaporated, giving a major peak in the
derivative weight change curve. The following weight loss before the temperature reached
around 200 °C was attributed to the detachment of the absorbed water on the surface of
porous cellulose structure. At 300 °C, a large derivative weight change peak appears,
resulting from the rapid loss of weight due to oxidation of cellulose. During this stage,
dehydration and decarboxylation will produce anhydrocellulose and combustible gases like

119

aldehydes, ketones, ethers, and so on. From 360 °C to 450 °C, the carbonaceous residues
were oxidized and resulted in CO2, CO and H2O. Around 0.2% w.t. char was left after heating
process till 500 °C.

Figure 3-25 TGA of dried CNCs purchased from the University of Maine process development center
with a water content of 93.8% w.t. To eliminate the water evaporation step during thermal
analysis, sample was dried in vacuum oven overnight at 80C before the test. TGA was conducted in
air through a high-resolution ramp program at a heating rate of 25 °C/min.

The thermal stability of CNCs is quite different compared to MFCs, the major peak of
DTG appeared at ~210 °C, and then at 600 °C there was still 4% w.t. left as the residue
material. Nearly 40% w.t. mass was lost in the 200-210 °C region, followed by another 55%
w.t. mass loss between 210 and 600 °C. This suggest a different mechanism compared to
MFCs. Due to the smaller size of CNCs compared to MFCs and original cellulose materials,
the larger surface area of CNCs opens up more exposure to heat. It has also been reported
120

that the existence of sulfur in most CNCs due to sulfuric acid hydrolysis might influence the
thermo-stability of CNCs.28
The real mass of cellulose in the homogenized suspension was measured by drying
20mL as-prepared suspension in vacuum oven at 100 °C for 24 hours and weighing the
resulting product. The real concentration of the final cellulose suspension product was 4%
w.t.
The surface area was not tested for MFCs for the following reasons: (a) A dried sample
is required for BET testing conditions (b) By drying the MFCs either by vacuum oven
heating, or freeze drying, aggregation would happen (c) The data collected by BET
measurement couldn’t represent the real surface properties of MFCs.

Figure 3-26 FTIR spectra of MFCs

Infra-red spectroscopy was applied to detect the vibrations of typical chemical bonds
in the cellulose materials based on the absorption of infrared radiation. The spectra were

121

collected with 4 acquisitions in the range of 4000-400 cm-1 as shown in Figure 3-26. The
broad peak in the 3000-3650 cm-1 region corresponded to O-H stretching vibrations, and
smaller peak at 2910 cm-1 was due to C-H stretching vibrations. Peak at 1643 cm-1 could be
explained as the adsorbed water in the cellulose. According to Lu et al28, the peak at around
1436 cm-1 was because of the bending of the CH2-, and all the features between 600 and
1600 cm-1 were from the deformation and twisting modes of the anhydroglucopyranose
unit. Peaks representing C-OH stretching (1060 cm-1), C-O-C bending vibrations (1163 cm1),

and C-O or C-C stretching or C-OH bending (1030 cm-1) were also featured in the spectra.
XPS was also applied for the surface composition study on MFCs in a 0.5% w.t. water

suspension (Figure 3-27). The O 1s and C 1s ratio in atomic concentration was 40.6 %: 59.4%
= 0.68, which is lower than theoretical value of 0.83 based on the formula of cellulose. This
further confirmed that the homogenization didn’t induce any new chemical composition
onto the surface of the cellulose. The deconvolution of the C peak showed the presence of
three types of carbon bonds: C-C (C1 284.6 eV), C-O (C2, 286.2 eV), and O-C-O (C3, 287.8 eV).

122

Figure 3-27 (a) XPS spectrum of MFCs ; (b) and deconvolution of its C1s peak.

3.4 Conclusions
In this chapter, a brief literature review on the benefits of using natural bio-template
materials to fabricate inorganic metal oxide nanomaterials especially various titania

123

materials was discussed. To confirm the feasible functions of cellulose as a template for
TiO2 fabrication, cotton rounds and filter paper were utilized, in combination with two
aqueous precursor systems. Tubular or hollow ribbon shaped titania in anatase phase were
produced from above systems. The shape of the cellulose was successfully duplicated in the
ammonium hexafluorotitanate system, while the titanium tetrafluoride system resulted in
more fractured hollow structures. The hydroxyl groups of the cellulose fibers interacted
with the Ti-O through H-bonding, which led to the nuclei growth on the surface and finally
generated TiO2 nanoparticles.
The high surface area, high aspect ratio, crystallinity, flexibility and a large amount of
hydroxyl groups on the surface made micro/nano-fibrillated cellulose (MFCs) an even
better candidate. Homogenization of cellulose was conducted and suspension of cellulose
in both water and isopropanol were prepared for future precursor templating. Thermo
stability analysis indicated the effects MFCs will make in the crystallization of titania. FTIR
and XPS confirmed the chemical structure and surface composition of the homogenized
MFCs as well.
Current tubular titania materials have been made with micro/macro size cellulose
templates, which usually result in fracture of the product, further slurry making for certain
applications such as dye-sensitized solar cell photoanode. The other popular
titania/cellulose composite has been developed as titania supported by paper or sheets.
This extra addition of cellulose might retard the pristine properties of titania in terms of
electron transport, phonon harvest, and so on. In a word, a facile simple with least step
deposition of titania film will be promising no matter for photoanode preparation, or the

124

immobilized –type photocatalytic films. In the coming chapters, the focus will be put on
developing Ti (IV) precursor with MFCs templates for multiple applications.

125

BIBLIOGRAPHY

126

BIBLIOGRAPHY

(1)

Ghadiri, E.; Taghavinia, N.; Zakeeruddin, S. M.; Grätzel, M.; Moser, J.-E. Enhanced
Electron Collection Efficiency in Dye-Sensitized Solar Cells Based on
Nanostructured TiO2 Hollow Fibers. Nano Lett. 2010, 10, 1632–1638.

(2)

Lu, Y.; Sun, Q.; Liu, T.; Yang, D.; Liu, Y.; Li, J. Fabrication, Characterization and
Photocatalytic Properties of Millimeter-Long TiO2 Fiber with Nanostructures
Using Cellulose Fiber as a Template. Journal of Alloys and Compounds 2013, 577,
569–574.

(3)

Qi, K.; Daoud, W. A.; Xin, J. H.; Mak, C. L.; Tang, W.; Cheung, W. P. Self-Cleaning
Cotton. J. Mater. Chem. 2006, 16, 4567.

(4)

Lee, Y. M.; Kim, Y. H.; Lee, J. H.; Park, J. H.; Park, N.-G.; Choe, W.-S.; Ko, M. J.; Yoo,
P. J. Highly Interconnected Porous Electrodes for Dye-Sensitized Solar Cells
Using Viruses as a Sacrificial Template. Adv. Funct. Mater. 2011, 21, 1160–1167.

(5)

Lou, S.; Guo, X.; Fan, T.; Zhang, D. Butterflies: Inspiration for Solar Cells and
Sunlight Water-Splitting Catalysts. Energy Environ. Sci. 2012, 5, 9195.

(6)

Dong, Q.; Su, H.; Cao, W.; Zhang, D.; Guo, Q.; Lai, Y. Synthesis and
Characterizations of Hierarchical Biomorphic Titania Oxide by a Bio-Inspired
Bottom-Up Assembly Solution Technique. Journal of Solid State Chemistry 2007,
180, 949–955.

(7)

Nomura, T.; Morimoto, Y.; Ishikawa, M.; Tokumoto, H.; Konishi, Y. Synthesis of
Hollow Silica Microparticles From Bacterial Templates. Advanced Powder
Technology 2010, 21, 218–222.

(8)

Hall, S. R.; Swinerd, V. M.; Newby, F. N.; Collins, A. M.; Mann, S. Fabrication of
Porous Titania (Brookite) Microparticles with Complex Morphology by Sol-Gel
Replication of Pollen Grains. Chem. Mater. 2006, 18, 598–600.

(9)

Varshney, V. K.; Naithani, S. Chemical Functionalization of Cellulose Derived
From Nonconventional Sources. In Cellulose Fibers: Bio- and Nano-Polymer
Composites; Springer Berlin Heidelberg: Berlin, Heidelberg, 2011; pp. 43–60.

(10)

Habibi, Y.; Lucia, L. A.; Rojas, O. J. Cellulose Nanocrystals: Chemistry, SelfAssembly, and Applications. Chem. Rev. 2010, 110, 3479–3500.

(11)

Lavoine, N.; Desloges, I.; Dufresne, A.; Bras, J. Microfibrillated Cellulose - Its
Barrier Properties and Applications in Cellulosic Materials: a Review.
Carbohydrate Polymers 2012, {90}, {735–764}.

(12)

Chen, W.; Yu, H.; Liu, Y. Preparation of Millimeter-Long Cellulose I Nanofibers
127

with Diameters of 30–80&#Xa0;Nm From Bamboo Fibers. Carbohydrate
Polymers 2011, 86, 453–461.
(13)

SirĂł, I.; Plackett, D. Microfibrillated Cellulose and New Nanocomposite
Materials: a Review. Cellulose 2010, 17, 459–494.

(14)

Xu, C.; Zhu, S.; Xing, C.; Li, D.; Zhu, N.; Zhou, H. Isolation and Properties of
Cellulose Nanofibrils From Coconut Palm Petioles by Different Mechanical
Process. PLOS ONE 2015, {10}.

(15)

Chinga-Carrasco, G. Cellulose Fibres, Nanofibrils and Microfibrils: the
Morphological Sequence of MFC Components From a Plant Physiology and
Fibre Technology Point of View. Nanoscale Res Lett 2011, 6, 417.

(16)

Sun, D.; Yang, J.; Wang, X. Bacterial Cellulose/TiO2 Hybrid Nanofibers Prepared
by the Surface Hydrolysis Method with Molecular Precision. Nanoscale 2010, 2,
287–292.

(17)

Heterogeneous Nanotubular Anatase/Rutile Titania Composite Derived From
Natural Cellulose Substance and Its Photocatalytic Property. 2013, 15, 5586–
5590.

(18)

Arseneau, D. F. Competitive Reactions in the Thermal Decomposition of
Cellulose. Canadian Journal of Chemistry 2011, 49, 632–638.

(19)

Rantuch, P.; Chrebet, T. Thermal Decomposition of Cellulose Insulation.
Cellulose Chemistry and Technology 2014, 48, 461–467.

(20)

Burnham, A. K.; Zhou, X.; Broadbelt, L. J. Critical Review of the Global Chemical
Kinetics of Cellulose Thermal Decomposition. Energy Fuels 2015, 29, 2906–
2918.

(21)

White, J. E.; Catallo, W. J.; Legendre, B. L. Biomass Pyrolysis Kinetics: a
Comparative Critical Review with Relevant Agricultural Residue Case Studies.
Journal of Analytical and Applied Pyrolysis 2011, 91, 1–33.

(22)

LĂŠdĂŠ, J. Cellulose Pyrolysis Kinetics: an Historical Review on the Existence and
Role of Intermediate Active Cellulose. Journal of Analytical and Applied Pyrolysis
2012, 94, 17–32.

(23)

Gaan, S.; Rupper, P.; Salimova, V.; Heuberger, M.; Rabe, S.; Vogel, F. Thermal
Decomposition and Burning Behavior of Cellulose Treated with Ethyl Ester
Phosphoramidates: Effect of Alkyl Substituent on Nitrogen Atom. Polymer
Degradation and Stability 2009, 94, 1125–1134.

(24)

Frank, O.; Zukalova, M.; Laskova, B.; KĂźrti, J.; Koltai, J.; Kavan, L. Raman Spectra
of Titanium Dioxide (Anatase, Rutile ) with Identified Oxygen Isotopes (16, 17,

128

18). Physical Chemistry Chemical Physics 2012, 14, 14567–14572.
(25)

Turbak, A. F.; Snyder, F. W.; Sandberg, K. R. Micro-Fibrillated Cellulose and
Process for Producing It.

(26)

QuiEvy, N.; Jacquet, N.; Sclavons, M.; Deroanne, C.; Paquot, M.; Devaux, J.
Influence of Homogenization and Drying on the Thermal Stability of
Microfibrillated Cellulose. Polymer Degradation and Stability 2010, 95, 306–
314.

(27)

Siqueira, G.; Bras, J.; Dufresne, A. Cellulosic Bionanocomposites: a Review of
Preparation, Properties and Applications. POLYMERS 2010, {2}, {728–765}.

(28)

Lu, P.; Hsieh, Y.-L. Preparation and Properties of Cellulose Nanocrystals: Rods,
Spheres, and Network. Carbohydrate Polymers 2010, 82, 329–336.

129

4 Synthesis of Porous Web-structured Titania with
Micro/nano-fibrillated Cellulose as Templates by Sol-Gel
Method
4.1 Background
4.1.1 Utilizing nano-scale cellulose for inorganic materials design
As introduced in Chapter 3, micro and nano-scale cellulose are superior among the
large cellulose family, due to their much higher surface area and aspect ratio, cost-effective
price and environmentally friendly properties, adequate surface hydroxyl groups for
chemical bonding and abrupt decomposition behavior with heat treatment. Traditionally,
cellulose fibers in the range of 2-10 um in diameter are usually used as a template at the
reaction stage, then either through thermal removal and separation, or pre-separation and
annealing, the inorganic products are left as random distributed structures, which requires
further paste forming for thin film coating. Emerging research has focused on utilizing
micro/nano-scale cellulose for nano-sized metal particles,1 porous metal oxide materials,2
nanocubic TiO2,3 and porous titania scaffolds.4
Previously, different nano/micro-cellulose types and Ti precursors were combined,
and the resulting morphologies and porosity were different as well. For instance, Shin et al
2

used lab-made cellulose crystals with width of 5-10 nm and length of 150-200 nm to

form a dried free-standing film, coated with Tyzoer-LA 50% solution, followed by 500 °C
calcination. The TiO2 materials went through severe shrinkage of the cellulose film, but still,
with adequate surface area (170-200 m2) promising in catalysis applications. Zhou et al

3

took advantage of the morphology-inducing effect of nano-scale cellulose, to create new
130

nanocube structures. It was indicated that the cellulose nanocrystals influenced the
nucleation and growth of TiO2, as well as served as a hydrophilic polymer chain coordinate
agent. They might promoted favoring of some certain facets of TiO2 particles and induced
the nanocube shape, besides, stabilizing the whole system and preventing aggregation.
Ivanova et al.,4 chose a complicated procedure to assemble pre-synthesized crystal anatase
particles into the nanocrystal cellulose templated titania scaffold. By varying the weight
percentage of the preformed TiO2 particles, the crystallinity of the films could be tuned
without changing annealing temperature or losing surface area and porosity.
The porous photocatalysts discussed in Chapter 2 are in powder form templated from
calcium carbonate. Although efficient, in practice, the reclamation of used powder-form
photocatalysts is difficult due to their small sizes. A further separation by centrifugation or
micropore-sized filtering will be energy-consuming and disobey the initial idea of
economical usage of phonon energy. Thus photocatalysts on film is an optimized concept
since there is no need for separation. Eliminating the separation steps of the previous
methods will simplify processing. A thin film catalyst eliminates the separation procedure
in pollutant treatment when using powder catalysts. When well-dispersed nano cellulose
fiber suspension is applied as templates in a controlled sol-gel approach, the precursor will
become thick, homogeneous, and beneficial to thin film coating.
Thus we propose a novel method to make immobilized titania thin films with
transparency, porosity, and photocatalytic activity by templating with nano-sized cellulose
fibers.

131

4.1.2 Controlled hydrolysis in a sol-gel process for MFCs templating
Chapter 1 has discussed the theory of sol-gel method to synthesize metal oxide
materials, and the specialty of Titanium precursors as transition metal source. Sol-gel
derived materials can be prepared by tailoring parameters with high surface area,
controlled porosity, and desired electric and optical properties. A controlled hydrolysis is
crucial for manipulation of materials size, morphology, and chemical properties. Chapter 2
described a successful method to prepare porous TiO2 with CaCO3 templating in a titanium
tetraisopropoxide-valeric acid based sol gel environment, which utilized strong chelating
ligands of valeric acid for occupation of the coordination sites of the alkoxide to lower
hydrolysis rates. Chapter 3 illustrated how cellulose fibers worked as soft templates for
inorganic materials fabrication in hydrolytic media. To further the exploration of cellulose
templating, the combination of much smaller cellulose MFCs and a hydrolysis-controlled
sol-gel process will be discussed in this chapter.
Micro-fibrillated Cellulose (MFC) is made by a homogenization process, at high shear,
yielding microfibrils with a diameter range of 10-40 nm, a length of several tens of
micrometers, and a web-like structure.5 Different from the other cellulose fibers discussed
earlier, MFCs applied in this investigation have smaller diameters and higher aspect ratios.
The work here will not only give insights on how the type of cellulose influences the metal
oxide deposition, but offers wide applications in photocatalysis, photoelectric, and
inorganic/organic hybrid from in-situ film deposition.
MFCs combined with the titanium precursor in modified sol-gel systems will result in
encapsulation of the MFCs in the growing direction of the TiO2 crystalline nanoparticles.

132

After calcination, the MFC will cleanly decompose resulting in the formation of micro/nano
channels in the TiO2 nanoporous film as Figure 4-1 shows. The benefits are expected to be a
more ordered interconnection and less aggregation with nano/micro channels.

Figure 4-1 Scheme of applying MFCs as a template for porous TiO2 network synthesis

4.2 Experimental
4.2.1 Materials and equipment
Materials:
All chemicals were used without pre-treatments unless noted. MFCs in isopropanol
(IPA) were processed by Mini DeBEE ultrahigh-pressure homogenizer as Chapter 3
described. Titanium(IV) isopropoxide (≥ 97%), valeric acid (≥ 99%), polyethylene glycol
(PEG, Mw=20000), polyvinylidene fluoride (PVDF, Mw=534,000), methyl orange (≥ 85%
dye content, certified by Biological Stain Commision), ethanol (≥ 99.5%, A.C.S. reagent), 2propanol (A.C.S. grade, ≥ 99.5%) were purchased from Sigma Aldrich. Plain, pre-cleaned
micro glass slides were purchased from VWR International, in a size of 1 inch by 3 inches,
with a thickness of 1.2 mm.
Equipment and Characterizations:
(1) The structures of the cellulose (MFCs) templates and titania porous thin films were
characterized by scanning electron microscopy (Carl Zeiss Auriga CrossBeam SEM-FIB).

133

Samples were prepared by drying and gold or tungsten coating before being loaded into
the SEM chamber. EDAX™ EDS Analysis System for SEM – X-ray Microanalysis was applied
together with SEM characterization for element analysis on films before and after
calcination. When cross-sectional views were necessary, a focused ion beam (FIB) was used
to mill surfaces of the materials.
(2) The transmission electron microscopy (TEM) images, high-resolution transmission
electron microscopy (HRTEM) images, and selected-area electron diffraction (SAED)
patterns were taken on a JEOL 2200FS equipment. TiO2 films were scratched off from glass
substrates and then dispersed in ethanol by bath sonication for 0.5 h. The ethanol/TiO2
suspension was dropped and dried on a carbon-coated 400 mesh copper grid overnight
prior to TEM analysis. Under HRTEM mode, fast Fourier transfer (FFT) patterns were utilized
to optimize image quality with DigitalMicrograph (DM) software.
(3) A single wafer spin processor (WS-400A-6NPP/LITE) was purchased from Laurell
Technologies Corp and used for film coating. Part of the doctor blade method was
completed by a manual Scotch tape method with a clean sharp razor blade, while part of
the doctor blade used the MSK-AFA-III compact tape casting coater with vacuum chuck.
This doctor blade machine has an adjustable micrometer film applicator (thickness 0.013.5 mm, or fixed thickness at 5 um, 10 um, 15 um and 20 um) with an 100-mm width, as
well as a dryer cover that heats up to 100 °C.
(4) A UV-Vis spectrophotometer (Perkin Elmer Lambda 900 UV Spectrometer) was
applied for TiO2 films absorption and transmission characterizations. The UV-Vis
absorption spectra of the TiO2 films were recorded in the range of wavelength 300-800 nm.

134

A diffuse reflectance spectroscopy (DRS) was utilized to analyze the degradation of methyl
orange on the surface of the titania films.
Reflectance is the base quantity characterizing the process of reflection, and it’s
defined as the ratio of the reflected radiant flux to the incident radiant flux. One of the
important factors that affects the diffuse reflectance spectra of a sample is the particle size
for powered samples and surface properties.6 When the diameter of the particles is
between the size of the beam cross-section and the light wavelength, diffraction will occur.
Part of the incident light goes at all angles back, so all the reflection, refraction, diffraction,
and absorption by particles are called diffuse reflection. Theoretically, an ideal diffuse
reflection follows Lambert Cosine Law, which states that the remitted radiation per unit
surface and unit solid angle is proportional to the cosine of the angle I of the incident light
and the cosine of the angle of observation e. If the particle size is smaller than the
wavelength, the contributions of reflection, refraction, and diffraction of the remitted
radiation flux are comparable and impossible to separate. Thus this is called scattering.
There are a few theories to interpret single particle scattering, however, when the distance
between particles is too small, multiple scattering becomes dominant. This is primarily a
qualitative technique with limited quantitative applications.
An integrating sphere attachment was used for the diffuse reflectance measurements
to establish the optical band gap and the reflection/transition before and after
photocatalytic dye degradation. The TiO2 films were placed at the entrance port of the
integrating sphere and measured in the 250-800 nm wavelength range.

135

(5) Fourier transform infrared (FTIR) absorption spectra were collected with Perkin
Elmer 2000 Spectrum One FT-IR spectrometer at room temperature. KBr Pellets were
prepared by Carver Laboratory press (model 2518) with ground and pre-dried KBr
powder. For liquid samples, a drop of solution was cast between two pellets forming a
sandwich structure with the excess liquid squeezed and wiped off. For solid samples, KBr
powder was mixed with samples with a sample weight ratio of ~ 0.2% w.t. The spectra of
the composites were scanned from 4000 to 400 cm-1.
(6) A LabRAM ARAMIS laser Raman spectrometer (Horiba Scientific) with 532 nm 50
mW DPSS laser was used to inspect sample crystal phases as well. All the samples were
tested as is with a 50X objective lens. The gratings used in this work is 1800 gr/mm and
results in a spectral resolution of about 1.5 m-1. The laser power at the sample ranged in 13 mW, and the 1/e laser spot size was about 2 Îźm at the sample. The crystallographic
structures of the solid samples were also determined with XRD, Bruker Davinci
Diffractometer. Dry powders could be deposited into the hollow space of a flat plate sample
holder. Thin films were tested with minimal preparations.
(7) Thermogravimetric analysis (TA instruments Q500) used a high-resolution ramp of
25 °C/min to ~800 °C in air or nitrogen was conducted on TiO2 samples. Surface area and
pore size data was collected with Brunauer–Emmett–Teller (BET) Micromeretics ASAP
2020. As for thin film samples, films were scratched off and collected for analysis. To
further remove moisture and impurities, samples were degassed at 150 °C for > 4 hours
under vacuum.

136

(8) X-ray photoelectron spectroscopy (XPS) was using a Perkin Elmer Phi 5400. The KÎą
line of Mg ( 1254 eV ) X-ray source with a beam voltage of 15 kV and current of 20 mA was
the excitation source. The survey spectra scan were acquired with a pass energy 187.85 eV,
and a 1 eV step size. For detailed C 1s or O 1s high resolution spectra, a pass energy of
29.35 eV and step size 0.1 eV were used. Chamber vacuum was maintained below 1×10-7
Torr (1.33×10-5 Pa). The sample to be tested was mounted onto the sample holder with
double-sided copper tape, and as for powder form samples, a thin but dense layer of
powder was spread on the tape with excess removed. After the measurement, data
processing was performed by MultiPak V8.0. If not mentioned, the spectra have been
resolved into chemically shifted components by a non-linear least squares procedure.
(9) Photocatalytic reactions were conducted with a RC-500 Xenon UV lamp (Xenon
Corp., Wobrun, WA) that has 300W power, 120Hz frequency and a 5-inch linear bulb shape.
The lamp was equipped with aluminum reflectors and fused quartz windows to provide
optimum transmission of UV radiation. The lamp housing was made of metal, and a forced
air convection removed the generated heat. The pulsed xenon arc lamp has a continuous
radiation spectrum as shown in Figure 2-8. Top right image also illustrates that the active
time span of each pulse is 100 us, while the interval rest time is 8.3 ms. The range of active
UV wavelength lies between 230 and 360 nm. Applying a Pyrex glass cover as a filter will
remove ultraviolet wavelength region beneath 290 nm.

137

Figure 4-2 (a): the scheme of the UV lamp system (Ozone was not applied in this work). (b): the
time schedule of the pulsed wave. (c): wavelength range of the outputs from the pulsed xenon UV
lamps. (d): UV transmittance of Surasil, quartz, and Pyrex glass filters[source: Xenon Corp.

(10) A HIMONT Plasma Science PS 0500 Plasma Surface Treatment System was used to
clean the glass substrates for ~0.5 h. The oxygen pressure was 0.270 Torr and the power
was 300 W.
4.2.2 Synthesis of a porous TiO2 network as thin film
In order to moderate the hydrolysis and precipitation rates during the sol-gel
reactions, valeric acid was used as a chelating agent for the non-aqueous environment.
Valeric acid (v-acid) was added to titanium isopropoxide (TTIP) followed by water as
shown in Figure 4-3, The atomic ratio of v-acid: water: TTIP is 9: 1.5: 1 respectively.7 The
mixture was vigorously stirred in capped vials at room temperature for 1 h. Then the asprepared titanium valerate was combined with MFC/isopropanol suspension at a ratio of 1:
1 and kept in a 50 °C oil bath for 14 h. The mixture turned from a colorless, transparent
solution to cloudy.
138

Figure 4-3 Scheme of MFCs-templating slow sol-gel method

Titania films were fabricated by drop-casting onto a glass substrate. Spin-coating was
attempted with different amounts of precursor solution, rotating rates and times, however,
the existence of cellulose fibers made it difficult for a uniform film formation, even with
multi-layer spinning. Binders, PEG (10% w.t. ) or PVDF (10% w.t.), were added into the
Ti/cellulose precursor to reach a smooth slurry texture for spin or doctor blade coating as
well.
For photocatalytic films, the precursors were made with a modified ratio between
Ti(IV) and cellulose, to improve the TiO2 content and activity of the films. Three different
ratios were tested as titanium valerate: MFCs (in isopropanol)= 2: 1, 4: 1, and 8: 1, then the
mixtures were kept in a 50 °C oil bath for more than 12 h as well. Specifically, a 67 mL
MFC/IPA suspension was used as a base volume, and 0.605, 1.21, 2.42 mL TTIP were used
with 1.98, 3.95, 7.90 mL valeric acid respectively for the increasing ratios.
Films were made with 0.13 mL precursor drop cast for same thickness, dried in dry
room hood for 1 hour to avoid moisture, followed by step-wise annealing at 200 °C, 300 °C,
350 °C, 450 °C.

139

4.2.3 Measurements of photocatalytic performance of porous titania thin films
The photocatalytic performance was evaluated by monitoring the degradation of
methyl orange (MO) aqueous dye solution with the illumination of Xenon RC500 lamp. MO
solution with a concentration of 5×10-4 M was prepared and kept in a capped glass
container. Diffuse Reflectance Spectroscopy (DRS) was chosen to keep track of the
reflectance change during the whole process. Typically, for a series of tests, the first data
point was taken on a pristine as-prepared TiO2 film (treated with plasma). The second was
taken on the same film with a coating of 10 uL 5×10-4 M solution, left in darkness to reach
equilibrium and solvent evaporation for 0.5 hour. Then UV treatment was started in a
manner of 1 min, 2 min, 2 min, and 2 min, where films were located 2.5 inches underneath
the UV lamp, with a Pyrex petri dish lid to remove the extremely strong short waves below
290 nm (Figure 4-4). After each treatment, a data point was recorded as well.
Plasma treatment of TiO2 films provide opportunities for improving the photochemical
capacities as for solar cells or photocatalysts. The glow-discharge plasma contains reactive
ions, radicals, and electrons to break down the covalent chemical bonds and remove
organic contaminants on the surface.8 Thus a higher contact surface area is created as
active spots for electrode/dye/electrolyte interfacial and also electron transport path with
fewer obstacles.

140

Figure 4-4 Scheme of photocatalytic degradation experiment

4.3 Results and Discussions
4.3.1 Fabrication of titanium dioxide precursor with MFCs templates
Following homogenization, the cellulose was reduced to microfibers with typical fiber
diameters around 30 nm and lengths extending from several to tens of micrometers
uniformly suspended in isopropanol. Figure 4-5 (a) shows the morphology and dimensions
of the homogenized fibers. There are occasional thick fibers with diameters exceeding 100
nm, however, most of the fibers have uniform dimensions. Figure 4-5 (b) and (d) clearly
illustrate the influence of cellulose on the titania morphology after calcination. The image
in Figure 4-5 (b) shows films that were condensed with low porosity, while (d) shows that
the titania has adopted the morphology of tangled cellulose and retained its porous
structure. For comparison, coated MFCs without calcination are shown in Figure 4-5 (c).
The removal of cellulose fibers and shrinkage of the precursor left large holes on the
surface of films.

141

Figure 4-5 (a) Homogenized MFCs. (b) titanium precursor without cellulose, calcined at 450 °C for 2
h. (c) titanium precursor coated cellulose fibers, before calcination. (d) titanium precursor with
cellulose (ratio is 1: 1), calcined at 450 °C for 2 h.

To prevent large amounts of cracks, a two-stage calcination was utilized for the
crystallization of as-prepared titania precursor. Starting from room temperature, the film
was heated to 250°C for 1h, followed by an increase to 450°C for another 1h. A new
morphology was formed as shown in Figure 4-6 (a-c). Through the modification of
precursor, hydrolysis, gelation, precipitation, drying, and heat treatment, the nanoporous
TiO2 network thin film was formed. The fine titania particles were grown along the fiber’s
direction and channel-like pores were produced when the cellulose was decomposed with
heat treatment. The inorganic network was determined by the competition between
hydrolysis and precipitation steps, and the removal of cellulose fibers. When hydrolysis is
fast and condensation is slow, small pores are formed in gels; otherwise, a gel consisting of
142

an interconnected network is formed, leaving enough time and space for template
interactions.
Thermogravimetric analysis (TGA) was performed on the cellulose fibers and titanium
precursor-cellulose composite. As a reference, 30mL of TTIP, valeric acid, and the cellulose
mixture were taken and evaporated in an oil bath at 80°C, and then analyzed with TGA.
This indicates that cellulose fibers alone started to decompose from around 250°C and
were eliminated at ~430°C (Figure 4-7). In the composite sample, at 250°C cellulose
started decomposing and there was approximately 20% w.t. titania left in the tested
specimen. The two-stage calcination was then utilized to balance the temperature
difference between cellulose degradation and crystal transformation.

143

Figure 4-6 (a-c) titanium precursor with cellulose, calcined with a two-stage method at 250 °C for
1h and 450 °C for 1.5 h at increasing magnifications.

144

Figure 4-7 TGA results of different Ti: cellulose ratios: 1 (green) Ti: cellulose = 1: 1, reacted at room
temperature. 2 (blue) Ti: cellulose = 1: 1, reacted at 50 °C. 3 (brown) Ti: cellulose = 0.5: 1, reacted at
room temperature. 4 (pink) Ti: cellulose = 2: 1, reacted at room temperature. All reactions lasted 14
hours with same volume of MFC suspensions at 0.5mg/mL concentration.

Figure 4-7 also indicated that the different ratios of Ti(IV) to MFCs yielded different
amounts of titania. Curve 3, 1, 4 had an increased amount of titania residue since their Ti:
cellulose ratios grew from 0.5: 1, 1: 1, and 2: 1. Curve 2 had higher TiO2 products than
curve 1 because an elevated reaction temperature was used, which promoted the
hydrolysis and condensation. Other chemical molar ratios were tried such as Ti: cellulose =
10:1 or 20:1, however, films after calcination showed a much denser morphology.
Thick films were prepared via repeating the coating-drying process for stronger
Raman spectrum signals. All films were confirmed anatase by Raman. Figure 4-8 shows the
comparison between uncalcined precursor-cellulose, calcined precursor-cellulose, and the

145

blank glass substrate. The Raman spectrum of the calcined film clearly shows that titania is
anatase as evidenced by the characteristic peaks at 144, 197, 399, 513, 519, 639 cm-1.9

25000

Intensity

20000

TiO2/Cellulose after calcination

15000

10000

TiO2/Cellulose no calcination
5000

Blank glass
0
0

500

1000

wavenumber (cm-1)

1500

2000

Figure 4-8 Raman spectra of blue: calcined precursor-cellulose; Red: the blank glass substrate;
Black: uncalcined precursor-cellulose.

146

Figure 4-9 X-ray photoelectron spectra of TiO2 thin film after calcination.

Quantitative XPS analysis was conducted on the annealed thin films. Figure 4-9(a)
shows signals of all peaks, and (b-d) gave the high-resolution evolution of elements C, O, Ti.
The atomic concentration of Ti 2p was 10.5% based on calculations from detailed
evolution. Figure 4-9 (b) illustrates C 1s peaks at 284.7 eV (C-C, or C-H), 286.2 eV (C-O-C, or
C-OH), 287.7 eV (C=O), and 289.2 eV (O-C=O). Some of these peaks exist even after high-

147

temperature annealing, which has been reported by others for TiO2 films and powders
after sintering.10 It is adventitious carbon that possibly came from ex-situ preparation
process. Most samples that have been exposed to atmosphere will get a layer of 1-2 nm
carbon contamination. Also, carbon could be due to small amounts of organic residuals
from titanium isopropoxide and cellulose. From the O 1s peak spectra in Figure 4-9 (c),
binding energy about 529.4 eV presented O2- for TiO2, and 531.7 eV was from hydroxyl
species, considered as defect states. Higher ratios of [OH]/[O2-] would indicate more
trapping states, while in this case, the pristine [O2-] was almost three times of the [OH]
species based on the integrated area. In addition, another source of hydroxyl species might
be from adsorbed alcohols on titania both molecularly and dissociatively at room
temperature.11 Figure 4-9 (d) explained the Ti 2p core level spectrum, at binding energies
457.8 eV referred to TiO2 .
Coating methods matter regarding the results of the adhesion and film morphology.
For the templated precursors, spin coating was not ideal due to the existence of fibrous
cellulose agglomeration. The binders (PEG and PVDF) were incorporated into the samples
to improve the continuity of the films and make thicker films through doctor blade method
or spin coating. TGA analysis of the binders found that PEG began to degrade at ~235 °C,
while PVDF degradation wouldn’t begin until the temperature reached 420 °C.

148

Figure 4-10 TGA results of (a) PEG and (b) PVDF through thermal degradation in air

When using PEG as a binder, larger aggregates and wider cracks tended to appear, due
to the fact that PEG melted, decomposed and interfered with the distribution of the metal
oxide. However, with a much higher decomposition temperature, PVDF helped stabilize the
film structure. Figure 4-11 (a-e) shows the calcined films with different binders and coating
methods. Figure 4-11 (a-b) shows that with increased coating thickness by spin-coating,
dense structures started to form, while (c) shows more porosity on the surface when
doctor blade was applied. Generally, spin coated PEG-added films still presented minimum
cracks, while with PVDF, films were filled with large cracks, and left with large pores as
shown in Figure 4-11(d-e). The residual of PEG binder was around 1.2% w.t. after 450 °C,
while PVDF had a much higher residual content at around 30% w.t. Although the coating
and morphology could be improved to some extent, no organic residue was expected in the
network of titania films for future applications.

149

Figure 4-11 Calcined Ti-MFCs thin films with different binders. With 10% w.t. PVDF, spin coated:
(a)1 layer, (b)4 layers; doctor blade coated (c) 1 layer. With 10% w.t. PEG: (c)spin coated, 1 layer
(d)doctor bade coated, 1 layer.

Preliminary dye adsorptions were conducted with an organic dye (Rose Bengal) on
binder-free calcined TTIP/cellulose films, PEG-assisted, and PVDF-assisted TTIP-cellulose

150

films. UV-vis spectroscopy results demonstrated that the binder does not facilitate more
dye loading.
An interesting fact of F on the growth of TiO2 crystal was revealed while experimenting
with the parameters of binder addition. When the amount of PVDF was increased to ~30%
w.t. to change the low viscosity of the original precursor solution, a new morphology was
created as Figure 4-12 shows. It was reported that when {101} planes of TiO2 nanocrystals
were stabilized by chemical bonds, more selective growth of {001} planes will occur,
resulting in rod-like TiO2 .12 However, fluorine-mediated TiO2 nanocrystals usually have
more stabilized {001} facets in various aqueous TiO2 synthesis routes13, indicating that the
grains on the surface came from the spatial steric hindrance and less available interfacial
contacts as nucleation sites, while the rice-like shape couldn’t be explained completely yet
in this work.

Figure 4-12 (a-b) Calcined Ti-MFCs thin films with ~30% w.t. PVDF, grain-surface rice-shaped
TiO2 . (c) An anatase tetragonal {101} bi-pyramid.

4.3.2 Photocatalysis performance of immobilized titania thin films
The modified sol-gel method with cellulose template created nano/meso porous
titania films as discussed before. It was applied as photocatalytic film in our research, to

151

study the influence of morphology, concentration, and thickness to the performance of
immobilized film photocatalysts.
Films made with three ratios between Ti (IV) and cellulose 2: 1, 4: 1, and 8: 1 were
tested, in the degradation process of methyl orange (MO) aqueous solutions.
Table 4-1 Atomic concentration comparison between variable films made from TTIP/V-acid/MFCs
precursor
Calcined films: TTIP V-acid+MFCs, with
varied Ti: cellulose ratios, and layers

C1s

O1s

Na1s

Si2p

Ti2P

1

1 layer, 2:1 Ti:cellulose film

24.3

46.5

13.7

0

15.5

2

1 layer, 4:1 Ti:cellulose film

24.0

46.4

12.1

2.3

15.2

3

1 layer, 8:1 Ti:cellulose film

22.5

50.3

6.8

8.6

11.8

4

2 layer, 2:1 Ti:cellulose film

23.3

47.1

11.8

1.8

16.0

5

2 layer, 4:1 Ti:cellulose film

25.0

44.9

13.2

1.6

15.3

6

2 layer, 8:1 Ti:cellulose film

26.5

47.9

5.3

6.7

13.6

As shown in Table 4-1, the atomic concentrations of C 1s, O 1s, Ti 2p, Na 1s, and Si 2p
were obtained through XPS. The concentration of Ti(IV) increased from one-layer films to
two-layer films for the same Ti: cellulose ratios. When Ti: cellulose was 2: 1, it gave the
highest Ti concentration, and the 8: 1 ratio films were the lowest among the three types.

152

Figure 4-13 XPS spectra of TiO2 thin film after calcination, made from TTIP/V-acid/MFCs, with Ti:
cellulose = 2:1. (a) XPS spectrum with element labeling. (b-d) High-resolution data fitting.

The C 1s, O 1s, and Ti 2p were fitted and analyzed as shown in Figure 4-13 for Ti:
cellulose = 2: 1 thin films after calcination. As mentioned before, C 1s peaks at 284.8 eV (CC, or C-H), 286.4 eV (C-O-C, or C-OH), 288.6 eV (C=O), and 290.1 eV (O-C=O) were from
adventitious carbon. From the O 1s peak spectra in Figure 4-13 (c), binding energy about
529.4 eV presented O2- for TiO2, and 531.2 eV was from hydroxyl species, the pristine [O2-]
was twice of the [OH] species based on the integrated area. Figure 4-13 (d) explained the Ti
2p core level spectrum, at binding energies 458.0 eV referred to TiO2 .

153

Figure 4-14 XPS spectra of TiO2 thin film after calcination, made from TTIP/V-acid/MFCs, with Ti:
cellulose = 4:1. (a) XPS spectrum with element labeling. (b-d) High-resolution data fitting.

154

Figure 4-15 XPS spectra of TiO2 thin film after calcination, made from TTIP/V-acid/MFCs, with Ti:
cellulose = 8:1. (a) XPS spectrum with element labeling. (b-d) High-resolution data fitting.

Figure 4-14 and Figure 4-15, similarly, illustrated the XPS spectra of films made with
Ti: cellulose = 4:1 and 8:1. After summarizing the ratio of the C 1s, O 1s, and Ti 2p
evolutions, some general ideas were concluded as follows:
(1)

Adventitious carbon was recorded in four peaks representing different C
bonds, and difficult to be distinguished from the film properties. (whether the
film contains carbon). The distribution of 4 fitted peaks were very consistent,
with 56% - 14% - 22% - 8%.

(2)

O 1s fitted peaks have lied in the same binding energy region as 529.4 eV,
531.3 eV, and 532.52 eV. The integrated area of the 529.4 eV peak was 51%,

155

80%, and 78% for sample 2:1, 4:1, 8:1, which corresponds to the Ti(IV)
amount trend summarized in Table 4-1.
(3)

Ti 2p was fitted into 457.9 eV representative of titania and was the only Ti
species observed for all samples.

Figure 4-16 TEM images of anatase TiO2 film generated by modified sol-gel method with TTIP/Vacid/MFCs, with Ti: MFCs = 2:1. (a) and (c) porous structure, at lower magnifications (b) and (d)
HRTEM image noted with anatase lattice spacing ~0.352 nm of the plane (101).

TEM images of each type of films were obtained at high resolution and anatase phase
was further confirmed with selected area diffraction pattern and the lattice spacing
measured at 0.35 nm corresponding to dominate plane (101) (JCPDS No. 21-1272).

156

Figure 4-17 TEM images of anatase TiO2 film generated by modified sol-gel method with TTIP/Vacid/MFCs, with Ti: MFCs = 4:1. (a) showed the porous nanoparticles, (b) and (c) HRTEM image
noted with anatase lattice spacing ~0.352 nm of plane (101) and ~0.19 nm of plane (200). (d) SAED
diffraction pattern shows the polycrystalline rings.

Figure 4-16, Figure 4-17, and Figure 4-18 showed the crystal sizes of TTIP/Vacid/MFCs induced films with Ti: cellulose (2:1, 4:1 and 8:1) ratios were, respectively: 15
nm, 12 nm and 21 nm. The diffraction pattern indicated that TiO2 films were composed of
polycrystalline anatase TiO2 with the Debye-Scherrer concentric rings (101), (004), (200),
(105) and (204).

157

Figure 4-18 TEM images of anatase TiO2 film generated by modified sol-gel method with TTIP/Vacid/MFCs, with Ti: MFCs = 8:1. (a-b) showed the porous nanoparticles, and (c) HRTEM image
noted with anatase lattice spacing ~0.352 nm of plane (101). (d) SAED diffraction pattern shows
the polycrystalline rings.

The surface area of these films was by scraping off the materials from the glass
substrate and collected for Brunauer-Emmett-Teller (BET) test. The surface area of
TTIP/V-acid/MFCs films were 40.9 m2/g (Ti: MFCs = 2:1), and 41.5 m2/g (Ti: MFCs = 1:1).
This procedure, to some extent, damaged the original shape of the film and the adhesion
between the film and substrate. Also, the limited amount of thin film materials that could

158

be obtained hindered the test. Thus, this data could be seen as a reference data with
accuracy worth doubting.
In chapter 2, MO has been briefly introduced as a powerful photocatalytic performance
indicator. During the process of MO decomposition, the peak at 465 nm will decrease.
Previous work14 showed that a new peak t 270 nm will show up in the preliminary stage,
and its intensity started to increase at first. Gradually, both peaks at 465 nm and 270 nm
decreased and disappeared eventually, leaving the only peak at 254 nm increasing with
time. This was explained to be related to the chemistry bonding change during the
degradation of MO. At first, the polyaromatic ring in MO was degraded, but created mono
substituted aromatics, so the 465 nm peak decreased, while the 270 nm peak increased.
Then further degradation converted all the MO and byproduct into CO2, H2O, and mineral
ions, shown as the fading of both peaks.
There is an optimum concentration where the best photodegradation rate constant
was reached.14 Generally, decreasing the concentration ensures an increasing constant k
due to the saturation level.
The method applied here was based on the observation of reflectance spectra change
from pristine TiO2 films, to dye-coated TiO2 films, and the evolution of the photocatalyzed
dye-coated TiO2 films. A simple quantitative method is used on the wavelength of 395 nm
to calculate the reflectance difference between the real-time data and the initial data, thus
by comparing the trend of the difference, the dye amount that was eliminated could be
estimated.

159

Figure 4-19 Reflectance change over time of UV-illuminated TiO2 film with 1×10-5 M MO
solution adsorbed. Titania film was made with TTIP/V-acid/MFCs and Ti: cellulose = 2: 1. (a)
Reflectance change from clean TiO2, MO solution adsorbed (after dye), and treated with UV lamp for
a total time of 1 min, 3 min, 5 min, and 7 min. (b) Reflectance difference over time, from dye
adsorption, to photocatalytic degradation.

Before MO solution was adsorbed onto titania films, the reflectance was measured as
black curve in Figure 4-19 (a), while after dye equilibrium, the red curve was moved away
from the initial stage. Once photocatalytic degradation started, the curve moved back to the
original stage gradually. The difference between real-time reflectance and the original state
was plotted vs. wavelength, and at 395 nm, the characterizing peak was used to represent
the dye decomposition amount.

Figure 4-20 For the film made with TTIP/V-acid/MFCs, and two layer coating, reflectance
difference change at wavelength 395 nm v.s. time. (a-c): three sets of data, where Ti:
cellulose = 2: 1 (black), Ti: cellulose = 4: 1 (red) and Ti: cellulose = 8: 1 (blue)

Three ratios of Ti to cellulose (2:1, 4:1 and 8:1) precursors were prepared into twolayer films, by depositing one layer, drying, depositing again, drying and annealing. Three

160

sets of films are shown in Figure 4-20. Ratio 4: 1 and 2: 1 had won over 8: 1 films in terms
of the degradation percentage. Single-layer films were studied with different volume of
precursor for coating, and the photocatalytic performance is shown in Figure 4-21.

Figure 4-21 For the film made with TTIP/V-acid/MFCs, and one layer coating, reflectance difference
change at wavelength 395 nm v.s. time. : three sets of data with different volume of
coating precursor.

The kinetics model was not fitted into an established model yet, thus a simple (R-R0)/t
calculation was used for parallel comparison between single-layer films, double-layer films,
and different Ti over cellulose ratios. The best candidate for TiO2 photocatalytic film is the
2: 1 and 4: 1 ratio. When the films have composition ratio of 8: 1, although TiO2 amount
was greatly increased, the porosity and film surface area would be lowered down. Table
4-2 summarizes the efficiency and rate scales for reference.
Table 4-2 Degradation and decomposition rate for various films

2Ti-1MFC
two-layer

degradatio
n efficiency

(RR0)%/t(min-1)

0.87

1.64

2Ti-1MFC 0.10
mL one layer

161

degradatio
n efficiency

(RR0)%/t(min-1)

0.83

0.58

Table 4-2 (cont’d)
4Ti-1MFC
two-layer

0.88

1.38

2Ti-1MFC 0.13
mL one layer

0.78

0.72

8Ti-1MFC
two-layer

0.92

1.48

2Ti-1MFC 0.15
mL one layer

0.89

0.40

degradatio
n efficiency

(RR0)%/t(min-1)

degradatio
n efficiency

(RR0)%/t(min-1)

4Ti-1MFC
0.10 mL one
layer

0.86

3.17

8Ti-1MFC 0.10
mL one layer

0.90

0.90

4Ti-1MFC
0.13 mL one
layer

0.85

2.25

8Ti-1MFC 0.13
mL one layer

0.88

0.89

4Ti-1MFC
0.15 mL one
layer

0.86

2.16

8Ti-1MFC 0.15
mL one layer

0.59

0.43

4.4 Conclusions and Future Work
A rate-controlled sol-gel method has been developed with chelating agent valeric acid,
and template micro/nano-fibrillated cellulose. Cellulose was removed cleanly by thermal
treatment and in-situ coating techniques created thin titania films on substrates with
porous structures, created by the vacancy left from cellulose. The theory was further
explored by SEM, TEM, TGA, XPS, EDS and BET. Two-stage calcination was utilized to
improve the interface between organic and inorganic composition, and binders such as PEG
and PVDF were studied as well.
Different ratios of Ti to cellulose were examined as photocatalytic films to degrade
methyl orange adsorbed on film. Degradation efficiencies of the films were related to

162

thickness and Ti(IV) concentration and proved to be a promising immobilized
photocatalytic tool for pollution removal.
Future work could focus on more versatile binder to improve the film quality, without
leaving residual within the film. Surfactants are also affordable and exciting directions to
further control crystal growth and size.

163

BIBLIOGRAPHY

164

BIBLIOGRAPHY
(1)

Shin, Y.; Bae, I.-T.; Arey, B. W.; Exarhos, G. J. Simple Preparation and
Stabilization of Nickel Nanocrystals on Cellulose Nanocrystal. Materials Letters
2007, 61, 3215–3217.

(2)

Shin, Y.; Exarhos, G. J. Template Synthesis of Porous Titania Using Cellulose
Nanocrystals. Materials Letters 2007, 61, 2594–2597.

(3)

Zhou, Y.; Ding, E.-Y.; Li, W.-D. Synthesis of TiO2 Nanocubes Induced by Cellulose
Nanocrystal (CNC) at Low Temperature. Materials Letters 2007, 61, 5050–
5052.

(4)

Ivanova, A.; Fravventura, M. C.; Fattakhova-Rohlfing, D.; RathouskĂ˝, J.;
Movsesyan, L.; Ganter, P.; Savenije, T. J.; Bein, T. Nanocellulose-Templated
Porous Titania Scaffolds Incorporating Presynthesized Titania Nanocrystals.
Chem. Mater. 2015, 150903090311005–150903090311008.

(5)

SirĂł, I.; Plackett, D. Microfibrillated Cellulose and New Nanocomposite
Materials: a Review. Cellulose 2010, 17, 459–494.

(6)

Ulery, A. L.; Drees, R. Methods of Soil Analysis: Mineralogical Methods.

(7)

Askeland, P. A. Preparation and Characterization of Sol-Gel Derived Metal Oxide
Thin Films and Powders for Coatings and Catalysts, 1997.

(8)

biris, A. S. The Structure of Titanium Oxidie Titania (Tio2) Photoactive WaterSplitting Catalysts by Raman Spectroscopy. 2009, 1–7.

(9)

Phase-Pure TiO2 Nanoparticles: Anatase, Brookite and Rutile. 2008, 19,
145605.

(10)

Oekermann, T.; Zhang, D.; Yoshida, T.; Minoura, H. Electron Transport and Back
Reaction in Nanocrystalline TiO 2Films Prepared by Hydrothermal
Crystallization. J. Phys. Chem. B 2004, 108, 2227–2235.

(11)

Kumar, P. M.; Badrinarayanan, S.; Sastry, M. Nanocrystalline TiO2 Studied by
Optical, FTIR and X-Ray Photoelectron Spectroscopy: Correlation to Presence of
Surface States. Thin Solid Films 358, 122–130.

(12)

Kim, J. Y.; Choi, S. B.; Kim, D. W.; Lee, S.; Jung, H. S.; Lee, J.-K.; Hong, K. S.
Surfactant-Assisted Shape Evolution of Thermally Synthesized TiO
2Nanocrystals and Their Applications to Efficient Photoelectrodes. Langmuir
2008, 24, 4316–4319.

(13)

Dozzi, M.; Selli, E. Specific Facets-Dominated Anatase TiO2: Fluorine-Mediated

165

Synthesis and Photoactivity. Catalysts 2013, Vol. 3, Pages 455-485 2013, 3, 455–
485.
(14)

Al-Qaradawi, S.; Salman, S. R. Photocatalytic Degradation of Methyl Orange as a
Model Compound. “Journal of Photochemistry & Photobiology, A: Chemistry”
2002, 148, 161–168.

166

5 Synthesis of Porous Titania network with Micro/nanofibrillated

Cellulose

as

Templates

by

Liquid

Phase

Deposition
5.1 Background
5.1.1 Liquid phase deposition
When controlled hydrolysis of metal ions happens in an aqueous solution, and
generates metal oxides and related precursor species, this deposition process on a
substrate is called liquid phase deposition (LPD). As mentioned in Chapter 1, LPD is a
versatile and cost-effective method to prepare titania films that do not require additional
equipment and works well with various types of substrates. Also, it supports lowtemperature reactions, provides strong adhesion between the film and the substrate and
excellent crystallinity. The reaction mechanism, adjustable parameters and relevant
applications of two advanced approaches - ammonium hexafluorotitanate (AHFT)/boric
acid (BA) LPD, and titanium tetrafluoride (TiF4) acidic LPD - will be discussed
AHFT and BA method:
Emerging research has been conducted on titanium dioxide synthesis through LPD,
typically, by the ligand exchange of metal fluorocomplex ions [TiF6]2- and the consumption
of F- with boric acid as a scavenger.
𝑇𝑖𝐹6 2− + 𝑛𝐻2 𝑂 ↔ 𝑇𝑖𝐹6−𝑛 (𝑂𝐻)𝑛

2−

+ 𝑛𝐻𝐹

𝐻3 𝐵𝑂3 + 4𝐻𝐹 ↔ 𝐵𝐹4− + 𝐻3 𝑂+ + 2𝐻2 𝑂

167

(5-1)
(5-2)

Further hydrolysis of TiF6-n(OH)n2- can be represented as follows
𝑇𝑖𝐹6−𝑛 (𝑂𝐻)𝑛
𝑇𝑖(𝑂𝐻)6

2−

2− 𝐻2 𝑂

→

𝑇𝑖𝐹5−𝑛 (𝑂𝐻)𝑛+1

2−

(5−𝑛)𝐻2 𝑂

+ 𝐻+ + 𝐹− →

+ (6 − 𝑛)𝐻 + + (6 − 𝑛)𝐹 −

𝑇𝑖(𝑂𝐻)6

2−

(5-3)

+ 2𝐻 + → 𝑇𝑖𝑂2 + 4𝐻2 𝑂

(5-4)

Equation (5-1) demonstrates the hydrolysis of the titanium fluoro complex, which
yields titanium hydroxide and free F- ions. It will shift towards right with the addition of Fscavenger boric acid (equation (5-2)). TiO2 formation is achieved as shown equations (5-3
&4).
A type of representative work is to immerse cleaned glass/quartz slides into the
prepared solution for extended hours.1 Concentrations and ratios of AHFT and BA typically
result in different degrees of supersaturation, thus influencing the formation and growth
rate of the TiO2 films. The calcination of LPD-induced films was also comprehensively
studied2 by varying calcination temperatures from 100-900 °C in air for 1 h respectively.
The thickness of the film deposited on a flat surface is dependent on the reactant
concentration, reaction temperature and deposition time. For a 0.1 M AHFT and 0.3 M BA, a
soaking at room temperature for 48 hours at pH 3.9 gave films with average thickness of
300 nm after calcination at 500 °C.2 Another case with a similar concentration ratio
reacted at 60 °C and pH~2.5 for 5 h followed by 500 °C calcination post-treatment resulted
in 800-1000 nm.3 This confirms that the growth and deposition of TiO2 are endothermic, so
that increasing the system temperature will speed the deposition.

168

Figure 5-1 (A-C) SEM and AFM images of different LPD bath treatment times. (A) 3 h, (B) 6 h, (C) 24
h. (D) Growth of LPD titania layer thickness. 4

As shown in Figure 5-1, at conditions such as 0.05 M AHFT/0.15 M BA, pH 2.88,
temperature 50 °C with deposition time ranging between 3 h to 24 h, the thickness grew
with prolonged reaction time. Similar observations are shown in Figure 5-2, where
reactions with 0.05 M AHFT/0.2 M BA at 40 °C for up to 8 h were conducted. Film thickness
increased with longer reaction time, while the depositing speed decreased. Once Ti4+ was
consumed, the reaction would stop. Figure 5-2 (c) shows the ratio influence on the film
thickness. A molar ratio of H3BO3/[TiF6]2- =2~4:1 is suitable for the deposition. If the BA
concentration is too low, scavenging of non-coordinated F- will not be sufficient, making the
hydrolysis of [TiF6]2- slow. Thus, an excess amount of BA will lead to more amorphous TiO2
because of the lack of F- ions. 5 The growth rate of films was 43 nm/h at 40 °C while only 13
nm/h at 30 °C,6 indicating that temperature is also an influencing factor.

169

Figure 5-2 Influence of (a) cycles (b) time (c) molar ratios to the thickness of
titania films. 5

At pH lower than 2.5, the deposition amount was very limited because of the
suppression of nucleation and crystal growth. However, when pH increases, the degree of
supersaturation increases since the hydrolysis equilibrium shifts right. The difference is
presented in Figure 5-3. By defining the theoretical and real solubility curves of Ti(OH)4, an
optimum range of pH was found for more efficient nuclei formation, and highly dense and
transparent films.

Figure 5-3 SEM images of annealed titania films made by LPD AHFT and BA with various pH values.
7

TiF4 in acidic media method:
Another type of precursor comes from a single titanium salt TiF4, which has been
deposited on different substrates such as glass slides, polymers, cotton fibers, filter paper,
porous alumina membranes, by a similar mechanism in acidic aqueous environment. 8,9
170

Typically, a 0.005-0.1 M TiF4 aqueous solution is prepared with adjusted pH by
hydrochloric acid (HCl) or ammonia (NH4OH) solutions. Substrates, e.g., clean glass slides,
filter paper, or polypropylene are immersed into the reactants, at temperature ranging
between 25-70 °C for various lengths of time. The following wash in water and alcohol
solvents cleans the excess precursor on the substrates and a gentle drying procedure will
get the coated substrates ready for different further treatments.

Figure 5-4 The relationship between Ti concentration, and pH level, with the deposition scale on
the substrate.8

In a glass substrate experiment,8 the concentration of TiF4 was maintained at 0.03-0.1
M, pH at 1-3.1 and temperature at 40-70 °C. When pH was below 1.0, or the TiF4
concentration was below 0.03 M, there was no precipitation nor film formation. When pH
was larger than 3.1, the solutions were highly supersaturated, so many precipitates were
observed while no film deposition formed on glass. Ideal deposition regions for
heterogeneous nucleation on the substrates was between the highly supersaturated and
the metastable. Different from using the F- scavenger to adjust the supersaturation level,

171

pH is the key parameter for controlled deposition. The F- ions bonded to Ti were assumed
to suppress the ultra-fast hydrolysis conditions.9
The type of substrates for deposition would influence the film adhesion. The more
hydrophobic the substrates were, the more easily the films peeled off. Firmly adherent
titania films are usually formed on hydrophilic surfaces, which provides possibilities for
site-selective growth and formation of patterned films.10 The addition of surfactant nhexadecyltri- methylammonium bromide (C16H33N(CH3)3Br, CTAB) into the solution help
create smaller anatase particles, fewer cracks, and higher transparency, since CTAB
retarded the crystalline growth on the particle surface.

Figure 5-5 Relationship between thickness of the deposited film and reaction time.8

The diameter of the deposited particles was proportional to the deposition time, and
the transparency of the film decreased at longer reaction time. When the reacting
temperature was low (~40 °C), not much deposition was formed during the first 8 h,
confirming that higher temperature did promote the film formation from Figure 5-5. Even
without calcination, the titania showed anatase crystallinity.8

172

Shapes can be well controlled by choice of substrates, resulting in nanorods,9 microhollow fibers,11 mesoporous titania nanorod/titanate nanotube composites,12 and other
hollow structures.13
Some researchers also applied boric acid into this titanium tetrafluoride method.12 The
reaction was performed at 60 °C for 12 hours. With a comparatively lower TiF4
concentration at 0.025 M, the addition of 0.075 M boric acid help shift the hydrolysis
equilibrium to the right. Hydrothermal-made amorphous titanate nanotubes ( d=7-12 nm,
l= several hundred nm) acted as the substrates for liquid phase deposition, and ultrafine
nanorods (d=5-10 nm, l=15-40 nm) were uniformly coated on the surface of the titanate
nanotubes. The resultant composites obtained higher surface area and larger pore volume
compared to titanate nanotubes or P25 particles, leading to an enhancement of the
photocatalytic performance for the oxidation of acetone.
Some other work, interested in removing templating substrates in situ, developed a
vanadium oxide nanobelts directed deposition with one-step HF dissolution of the
template.13 At a TiF4 concentration between 0.005 M and 0.1 M, vanadium oxide nanobelts
were added into the 60 mL TiF4 solution at 60 °C for 12 h. The concentration of the
byproduct HF increased during the immersion time, and gradually dissolved the vanadium
oxide until completely consumed. The concentration of TiF4 is the key factor to determine
the thickness of the hollow wall, in other words, the amount of the deposition.
More and more work reveal that solution concentration and temperature are tightly
related to the crystal growth and deposition rates. Comparatively higher concentration and
temperature resulted in more supersaturated system, leading to faster and complete

173

deposition.10 The pH of the whole system dropped at the beginning due to the byproduct
HF, then changes only very slowly for an extended time. In this same paper, a detailed
three-step deposition procedure was brought up as follows10:
(1) Initial delay: at this stage, the thickness had a very limited increase, but only for
lower temperature and lower concentration conditions. For highly saturated systems, the
initial delay could be hardly observed. (2) Fast growth: during this stage, nuclei on the
surface are ready to grow into crystals, governed by a first order kinetics (3) Growth
saturation: after a deposition thickness of ~500 nm at favorable temperature and
concentration; for low-temperature solutions, this stage could not be reached even after 70
hours.
However, films made in this way tend to be dense, especially denser as the
temperature of annealing goes up. An alternative approach is desired to improve the
surface area and the porosity of the titania nanoparticles on the film. Recently, researchers
have accomplished the immobilization of titania nanoparticles via a TiF4 LPD method on a
silicone nanofilaments-coated glass substrate.14 Silicone nanofilaments were prepared
from trichloromethylsilane (TCMS) in a custom-built chamber with precisely controlled
water ratio. Plasma treatment was followed to induce hydrophilicity on the surface of the
silicone as a prerequisite for further LPD functionalization. TiF4 was the titanium source in
a series of ethanol/water solvents (with adjusted polarity) at different temperatures.
Titania nanoparticles with sizes 50-90 nm were synthesized after calcination and the
photocatalytic capacities were investigated by methyl blue degradation.

174

In Chapter 3, preliminary experiments were conducted on large-scale cellulose fibers
like cotton wool fibers and filter paper with AHFT/BA, and TiF4 acidic conditions. We
discussed the hollow tubular titania prepared and the unique properties of MFCs. In
Chapter 4, sol gel synthesis of porous titania via MFCs proves the benefits of MFCs as
templates. In this chapter, MFCs were studied for bulk material fabrication, and thin film
coating via LPD. The photocatalytic capacities were tested through methyl orange
degradation.

5.2 Experimental
5.2.1 Materials and equipment
Materials:
All chemicals were used without pre-treatments unless noted. MFCs in isopropanol
(IPA) were processed by Mini DeBEE ultrahigh-pressure homogenizer as Chapter 3
described. Titanium(IV) fluoride, boric acid (≥99.5%), ammonium hexafluorotitanate
(99.99%), methyl orange (≥ 85% dye content, certified by Biological Stain Commision),
ethanol (A.C.S. reagent), 2-propanol (A.C.S. grade, ≥ 99.5%) were purchased from Sigma
Aldrich. Hydrochloric acid (analytical reagent grade) was from Mallinckrodt. Ethanol (200
proof, 99.5% ACS grade) and isopropanol (99.5%) were purchased from Michigan State
university stores. Plain, pre-cleaned micro glass slides were purchased from VWR
international, in a size of 1 inch by 3 inches, with a thickness of 1.2 mm. Mineral oil
(laboratory grade, heavy) was from Columbus Chemical Industries.
Equipment and Characterizations:

175

(1) The structures of the organic/inorganic composite films, colloidal crystals and
mesoporous annealed films were characterized by scanning electron microscopy (Carl
Zeiss Auriga CrossBeam SEM-FIB). Film samples were prepared by necessary drying and
gold or tungsten coating before loading into SEM chamber. Free-standing powder samples
were prepared by mounting onto a holder with carbon paste/tape after. EDAX™ EDS
Analysis System for SEM–X-ray Microanalysis was applied together with SEM
characterization for element analysis on films before and after calcination. When a crosssectional view was necessary, a focused ion beam (FIB) was used to mill surfaces of the
materials.
(2) The transmission electron microscopy (TEM) images, high-resolution transmission
electron microscopy (HRTEM) images, and selected-area electron diffraction (SAED)
patterns were taken on a JEOL 2200FS equipment. TiO2 films were scraped off from glass
substrates and then dispersed in ethanol by bath sonication for 0.5 h. The ethanol/TiO2
suspension was dropped and dried on a carbon-coated 400 mesh copper grid overnight
before TEM analysis. Under HRTEM mode, fast Fourier transfer (FFT) patterns were utilized to
optimize image quality with DigitalMicrograph (DM) software.
(3) A single wafer spin processor (WS-400A-6NPP/LITE) was purchased from Laurell
Technologies Corp and used for film coating.

Part of the doctor blade method was

completed by a manual Scotch tape method with a clean, sharp razor blade, while part of
the doctor blade used the MSK-AFA-III compact tape casting coater with a vacuum chuck.
This doctor blade machine has an adjustable micrometer film applicator (thickness 0.013.5 mm, or fixed thickness at 5 um, 10 um, 15 um and 20 um) with a 100-mm width, as well
as a dryer cover that heats up to 100 °C.
176

(4) A UV-Vis spectrophotometer (Perkin Elmer Lambda 900 UV Spectrometer) was
applied for TiO2 films absorption and transmission characterizations. The UV-Vis
absorption spectra of the TiO2 films were recorded in the range of wavelength 300-800 nm.
An integrating sphere attachment was used for diffuse reflectance measurements to
establish the optical band gap and the transition before and after photocatalytic dye
degradation. The TiO2 films were placed at the entrance port of the integrating sphere and
measured in the 250-800 nm wavelength range. Data was arbitrarily normalized before
comparison.
(5) Fourier transform infrared (FTIR) absorption spectra were collected with Perkin
Elmer 2000 Spectrum One FT-IR spectrometer at room temperature. KBr Pellets were
prepared by Carver Laboratory press (model 2518) with ground and pre-dried KBr
powder. For liquid samples, a drop of solution was cast between two pellets forming a
sandwich structure with the excess liquid squeezed out and wiped off. For solid samples,
KBr powder was mixed with samples with a sample weight ratio of ~0.2% w.t. The spectra
of the composites were scanned from 4000 to 400 cm-1.
(6) XRD: The crystallographic structures of the solid samples were determined with
XRD, Bruker Davinci Diffractometer. A LabRAM ARAMIS laser Raman spectrometer (Horiba
Scientific) with 532 nm 50 mW DPSS laser was used to inspect sample crystal phases as
well. All the samples were tested as is with a 50X objective lens. The resolution in Raman
spectrometer could be decided by the spectrometer focal length, diffraction grating, laser
wavelength, and the detector. The higher the groove density of the grating is, the higher the
spectral resolution. The gratings used in this work is 1800 gr/mm and results in a spectral

177

resolution of about 1.5 m-1. The laser power at the sample ranged in 1-3 mW, and the 1/e
laser spot size was about 2 Îźm at the sample.
(7) Thermogravimetric analysis (TA instruments Q500) used a high-resolution ramp of
25 °C/min to ~800 °C in air or nitrogen on TiO2 samples. Surface area and pore size data
were collected with Brunauer–Emmett–Teller (BET) Micromeretics ASAP 2020. As for thin
film samples, films were scraped off and collected for analysis. To further remove moisture
and impurities, samples were degassed at 150 °C for over 4 hours under vacuum.
(8) X-ray photoelectron spectroscopy (XPS) was using a Perkin Elmer Phi 5400. The KÎą
line of Mg ( 1254 eV ) X-ray source with a beam voltage of 15 kV and current of 20 mA was
the excitation source. The survey spectra scan were acquired with a pass energy 187.85 eV,
and a 1 eV step size. For detailed C 1s or O 1s high resolution spectra, a pass energy of
29.35 eV and step size 0.1 eV were used. Chamber vacuum was maintained below 1×10-7
Torr (1.33×10-5 Pa). The sample to be tested was mounted onto the sample holder with
double-sided copper tape, and as for powder form samples, a thin but dense layer of
powder was spread on the tape with excess removed. After the measurement, data
processing was performed by MultiPak V8.0. If not mentioned, the spectra have been
resolved into chemically shifted components by a non-linear least squares procedure.
(9) Photocatalytic reactions were conducted with an RC-500 Xenon UV lamp (Xenon
Corp., Wobrun, WA) that has 300 W power, 120 Hz frequency, and a 5-inch linear bulb
shape. The lamp was equipped with aluminum reflectors and fused quartz windows to
provide optimum transmission of UV radiation. The lamp housing was made of metal, and a
forced air convection removed the generated heat. The pulsed xenon arc lamp has a

178

continuous radiation spectrum as shown in Figure 2-8. The top right image also illustrates
that the active time span of each pulse is 100 us, while the interval rest time is 8.3 ms. The
range of operating UV wavelength lies between 230 and 360 nm. Applying a Pyrex glass
cover as a filter will remove ultraviolet wavelength region beneath 290 nm.

Figure 5-6 (a): the scheme of the UV lamp system (Ozone was not applied in this work). (b): the
schedule of the pulsed wave. (c): wavelength range of the outputs from the pulsed xenon UV lamps.
(d): UV transmittance of Surasil, quartz, and Pyrex glass filters[source: Xenon Corp.

(10) A HIMONT Plasma Science PS 0500 Plasma Surface Treatment System was used to
clean the glass substrates for ~0.5 h. The oxygen pressure was 0.27 Torr, and the power
was 300W.
5.2.2 Synthesis of TiO2 bulk material via TiF4- MFCs LPD method
At the beginning of research, two methods to combine TiF4 and cellulose were studied.
One method was to make TiF4 dissolved in water at a concentration of 0.08 M with a pH
value around 2.5, and then add the prepared solution into the same volume of MFCs/water,

179

capped and reacted in an oil bath at 65 °C for 8 h. A turbid and white sol was generated and
the sol was dried and calcined at 250 °C and 450 °C for 1 hour each. Another approach was
to add the corresponding amount of TiF4 directly into the MFCs/water to achieve a 0.04 M
TiF4 concentration. After heating at 65°C for 8h, the precursor was also annealed at twostage temperature for 2 h in a crucible. Water only without template was tried for
comparison, with all other conditions kept at the same(the process is summarized in Figure
5-7 below).

Figure 5-7 Scheme of TiF4- MFCs precursor preparation

5.2.3 Synthesis of porous TiO2 thin films via TiF4- MFCs LPD method
A 0.04 M TiF4 solution was prepared by adding 0.0991 g TiF4 into a vial with 20 mL 5
mg/mL MFC-water suspension while stirring. After one-hour mixing at room temperature,
the pH was adjusted to ~2.5 by the addition of diluted hydrochloric acid or ammonium
hydroxide aqueous solutions. Then the vial was capped and transferred to a pre-heated
mineral oil bath at 65 °C on a stirring hot plate. The reaction went for 6 hours and the
mixture was cooled down to room temperature before further use.

180

Solvent polarity was manipulated by changing the ratio of the MFC/isopropanol and
MFC/water. A series of ratios of isopropanol to water suspensions (I/W) = 0:1, 1:1, 4:1, and
1:0 was investigated with same procedures as above. A drop cast film was made with 0.1
mL stock solution and dried in vacuum oven for 1 h at room temperature, followed by
increasing temperature to 100 °C for 1 hour and a step-wise annealing at 200 °C, 250 °C,
300 °C, 350 °C for 15 minutes each, with a finishing 450 °C session for 1 hour. The whole
process could be demonstrated in Figure 4-3. To compare between MFC suspension and
solvent only reagents, similar I/W ratios were utilized for a set of TiF4 deposition reactions
as well.

Figure 5-8 Scheme of MFCs-templating TiF4 LPD thin films on glass.

5.2.4 Measurements of photocatalytic performance of porous titania thin films
The photocatalytic performance was evaluated by monitoring the degradation of
methyl orange (MO) aqueous dye solution with the illumination of Xenon RC500 lamp. MO
solution with a concentration of 5×10-4 M was prepared and kept in a capped glass
container. Diffuse Reflectance Spectroscopy (DRS) was chosen to keep track of the
reflectance change during the whole process. Typically, for a series of tests, the first data
point was taken on a pristine as-prepared TiO2 film (treated with plasma). The second was
taken on the same film with a coating of 10 μL 5×10-4 M solution, left in darkness to reach

181

equilibrium and solvent evaporation for 0.5 hour. Then UV treatment was started in a
manner of 1 min, 2 min, 2 min, and 2 min, where films were located 2.5 inches underneath
the UV lamp, with a Pyrex petri dish lid to remove the extremely strong short waves below
290 nm (Figure 4-4). After each treatment, a data point was recorded as well.

Figure 5-9 Scheme of photocatalytic degradation experiment

5.2.5 Synthesis of porous titania network via ammonium hexafluorotitanate and
boric acid LPD method
Additional titania thin films were prepared via liquid phase deposition with MFCs and
drop cast coating. A stock solution of Ti4+ (0.05 M) was prepared by mixing a homogenized
MFCs/water suspension with ammonium hexafluorotitanate (AHFT) (0.05 M) and boric
acid (BA) (0.15 M), followed by vigorous stirring for 1h at room temperature. The prepared
stock solution (pH adjusted to ~2.5) was then moved to a 50°C oil bath to react for 20h.
After the reaction, the solution turned from translucent to white. For comparison, different
ratios were prepared between BA and AHFT as 0.15:0.05, 0.15:0.15, 0.15:0.30, and
0.30:0.15 M. Films were prepared using doctor blade on glass substrate, dried in air at
room temperature, then a vacuum oven at 80°C, and finally annealed at 250 °C and 450 °C
for 1 h each.
182

PEG as a binder could improve the viscosity and continuity of the precursor. After 6hour LPD reaction, various amounts of PEG were added and kept stirring for two more
hours. Films were prepared via doctor blade then.

Figure 5-10 Scheme of AHFT-BA/MFCs synthesis

5.3 Results and Discussions
5.3.1 TiF4- MFCs bulk materials
When the precursor was calcined as bulk materials in crucibles, uniform cubes of
titania were formed. Figure 5-11 shows the bulk calcined materials at different
magnifications. Cellulose provided a substrate for nucleation and growth, and while the
deposition continued, surface crystals kept growing. This is suspected to result from the
fluoride-stabilized {001} facets. 15 Nanoparticle grains were characterized in Figure 5-11 (c)
where the cubes were connected by particle edges as well. The cube length is around 200
to 300 nm. The mechanism of TiO2 formation is explained by equations below:
𝑇𝑖𝐹 4 + 𝑛𝐻 2 𝑂 ↔ 𝑇𝑖𝐹4−𝑛 (𝑂𝐻)𝑛 + 𝑛𝐻𝐹
𝑇𝑖𝐹4−𝑛 (𝑂𝐻)𝑛 + (4 − 𝑛)𝐻2 𝑂 → 𝑇𝑖(𝑂𝐻)4 + (4 − 𝑛)𝐻𝐹
𝑇𝑖(𝑂𝐻)4 → 𝑇𝑖𝑂2 + 2𝐻2 𝑂 (f)

183

Figure 5-11 0.04 M TiF4 in 20 mL IPA/water suspension, calcined as bulk material. (a), (b), (c) are
the same area with increasing magnifications.

Figure 5-12 0.04 M TiF4 in 20 mL IPA/water suspension, calcined as bulk material. pH adjusted at 4
during deposition reaction.

TiF4 went through step-wise hydrolysis and condensation to become TiO2. It does not
use F- as a scavenger, but by controlling pH level, the supersaturation could be manipulated
as well. When pH is below 1, inefficient hydrolysis step prevents the deposition. When pH
is above 3, precipitation is preferred so the solution is highly supersaturated with titania
rather than forming a uniform coating on substrates. When pH is between 1 and 3, due to
the stability of Ti-F bonds, rapid hydrolysis and polymerization of the precursor do not
occur here. Instead, a moderate reaction leads to a stable deposition through
heterogeneous nucleation.9 A pH level of 2.5 was selected through experiments for
cellulose templating to make bulk materials and thin films. The morphology of TiO2

184

induced by precursor with higher pH at 4 is shown in Figure 5-13, where the
supersaturation results in dense films with fewer particles and secondary particle grains.

Figure 5-13 0.04 M TiF4 in 20 mL IPA/water suspension, calcined as bulk material. FIB-cut cross
section (a), (b) are taken from the same area with increasing magnifications.

Figure 5-13 reveals a highly porous surface at the ion beam cut cross section. The inner
part of the bulk materials had cracks and variable pore sizes. It indicated the removal of
cellulose fibers, and the difficulty of preparing a more ordered nanoporous network.
As comparison, samples made with TiF4 only without cellulose templates were tested
as shown in Figure 5-14. When no cellulose was applied, the base precursor was calcined
into randomly distributed particles. When the focused-ion beam was used to cut through
the particles, large dense aggregates without patterned nanopores or mesopores were
revealed.

185

Figure 5-14 0.04 M TiF4 in 20 mL water, calcined as bulk material. (a) and (b) are surface
features at same area with increasing magnifications. (c-d) FIB-cut cross-sectional view at
increasing magnifications.

Figure 5-15 XRD patterns of (a) TiF4 and water precursor, calcined bulk material, (b) TiF4 and
MFCs/Water, calcined bulk material.

The products were in anatase phase from the characteristic peaks labeled in Figure
4-5. Both samples were detected with peaks at around 25.3, 38.2, 48.2, 54.0, 55.2 and 63.0
186

°, which represented plane (101), (004), (200), (105), (211) and (204) [JCPDS 78-2486].
Crystal sizes were calculated using the Scherer Equation:
𝐷=

𝑘∙𝜆
𝛽 ∙ 𝑐𝑜𝑠𝜃

D: crystal size
k: constant=0.89
λ: X-ray wavelength, 1.5406 Å
β: full width at half maximum of the peak
θ: half the Bragg’s diffraction angle of the peak
The results from the XRD spectrum were, 27.9 nm for pure TiF4 induced TiO2, and
18.5 nm for TiF4 with cellulose templating TiO2. Besides, surface area tested by BET was
112 m2/g.
5.3.2 TiF4- MFCs induced TiO2 with tunable sphere chain morphology
To improve the photocatalytic capacity of titania, we consider several aspects such as:
the adsorption of pollutants, the absorption of photons, charge separation, the reactivity of
the surface sites, recombination rate of photogenerated electron-hole pairs. When this
consideration is applied to materials, it brings up higher surface area, better crystallinity,
beneficial morphology and nanostructures.

187

5.3.2.1 Surface properties of TiO2 films and related materials
Scanning electronic microscope (SEM):
As-received cellulose was treated with high-pressure homogenization which
transferred the large and aggregated fibers into reduced sizes and suspended extensively
in the solvent (water and IPA). Homogenized MFCs are generally 30 nm in diameter, with
length in micrometers or longer as shown in Chapter 3.
Figure 5-16 illustrates the influence of cellulose templates and the effect of the
percentage of isopropanol in the solvent for calcined thin films on glass substrates. Figure
5-16 (a) indicates that without the addition of cellulose, a film with huge cracks and large
aggregates of TiO2 particles larger than 10 um was formed. From Figure 5-16 (b) to (d),
TiF4 solutions were prepared with MFC suspensions, respectively, at isopropanol to water
( I: W) ratios 0:1, 1:1, and 4:1. As the isopropanol ratio increased to 4:1, a uniform of
spherical nanoparticles were formed along the chain of MFCs with a diameter around 200
nm.

188

Figure 5-16 SEM images of 4 samples without and with MFCs templates, at different isopropanol to
water ( I: W) ratios. All reactions last for 6 hours, and films were coated and calcined. (a) Without
MFC templates, 0.04 M TiF4 in water. The inset image scale bar reads 5 um. (b-d) With MFC
templates, at I: W ratios 0:1, 1:1, and 4:1 respectively. The scale bar of the inset image of (d) reads 1
um.

Figure 5-17 and Figure 5-18 compare the influence of time and TiF4 suspension on the
film morphologies. When water was the reaction environment, large cracks formed on the
surface of the film, and as reaction time accumulated, the film was denser and thicker,
while cracks larger. When MFCs/water was the reaction environment, the film was covered
with a few aggregations of titania particles, and some porosity was visible. Also, as time
went by, more and more aggregations formed on the surface of the film, even the removal
of cellulose templates could not contribute much to the porosity formation.

189

Figure 5-17 0.04 M TiF4 reacted in H2O for (a) 6 hours, (b) 15 hours, and (c) 40 hours.

190

Figure 5-18 0.04 M TiF4 reacted in 5 mg/mL MFC water suspension for (a) 6 hours, (b) 15 hours,
and (c) 40 hours.

191

Figure 5-19 0.04 M TiF4 reacted in 5 mg/mL MFC water and isopropanol suspension mixture at I:W
= 4:1 ratio for (a) 2 hours, (b) 4 hours, (c) 6 hours, (d) 15 hours, and (e) 40 hours.

Size variations based on different reaction time lengths are summarized in Figure 5-19
and Table 5-1. Sizes were measured based on around 75-150 data points, collected and
calculated by Image-pro Plus 5.0 for each sample. The size of the spherical particles grew
larger as the reaction time increased, and reached the top at around 14 hours, while started
to decrease at prolonged reaction times. As for samples reacted for the same 6-hour period,

192

the uncalcined amorphous particle size was 14% larger compared to the annealed samples.
TiO2 formed using bulk calcination (precursor stock solution calcined directly in a crucible
with the same manner of temperature control) had smaller spheres as well.
Table 5-1 Summary of particle size formed of the I: W= 4:1 sample, with different reaction time and
treatments

TiF4-Cellulose-4IPA-1H2O , with

Spherical particle

Standard

various treatment methods

mean size (nm)

deviation (nm)

2-hour reaction, calcined

101.8

8.4

4-hour reaction, calcined

111.0

8.3

6-hour reaction, calcined

125.4

10.3

14-hour reaction, calcined

187.3

11.8

40-hour reaction, calcined

139.1

13.3

6-hour reaction, dried in vacuum,
uncalcined

142.4

14.4

6-hour reaction, bulk calcination

114.2

10.8

The chain-like uniform sphere structures were present not only on the surface of the
film, but on the back of the film, and the cross-sectional area. Figure 5-20 (a) shows the
back of the film, where the inset is a piece of partially-lifted film. Figure 5-20 (b) shows the
cross-sectional linking between the spherical particles which followed the tangled pattern
of MFCs. When no water was applied in the reaction media, a similar structure with fewer
nanoparticles and porous fibril TiO2 formed on the surface of the calcined films as shown in
Figure 5-21.

193

Figure 5-20 The three-dimensional in-depth view of titania film made with the TiF4- with MFCs in I:
W=4:1 suspension. (a) The reverse side of the film, from the inset image of (a). (b) The crosssectional view of the film.

Figure 5-21 0.04 M TiF4 reacted in 5 mg/mL MFC isopropanol suspension at I: W = 1:0 ratio for 6
hours. Calcined films morphologies.

To understand better how cellulose was guiding the crystal growth of titania, more
SEM images were obtained for various sample films before calcination. Figure 5-22, Figure
5-23 and Figure 5-24 provide the information at different magnifications separately. The
amount of the patterned spherical particles was increased as the I: W ratio changed from
0:1, to 1:1 and finally to 4:1. Spherical particles before calcination had similar sizes at
around 140 nm.

194

Figure 5-22 0.04 M TiF4 reacted in 5 mg/mL MFC aqueous suspension at I: W = 0: 1 ratio for 6 hours,
coated on glass substrate, and dried in vacuum oven, without calcination applied.

Figure 5-23 0.04 M TiF4 reacted in 5 mg/mL MFC aqueous suspension at I: W = 1: 1 ratio for 6 hours,
coated on glass substrate, and dried in vacuum oven, without calcination applied.

195

Figure 5-24 0.04 M TiF4 reacted in 5 mg/mL MFC aqueous suspension at I: W = 4: 1 ratio for 6 hours,
coated on glass substrate, and dried in vacuum oven, without calcination applied.

Figure 5-25 SEM image of TiO2 hollow fibers formed after film calcination from sample TiF4 with
MFCs in aqueous suspension.

196

In addition to the formation of TiO2 spheres, the cellulose also acted as a sacrificial
template to produce titania hollow fibers or tubes after the heat removal of MFCs. Figure
5-25 represents the fiber-like TiO2 material from one of the samples at the I: W ratio of 0:1.
Other series samples such as I: W= 4:1 had intense spheres coverage along the “chain”,
resulting in the difficulty observing the TiO2 hollow fibers. However, after dispersion and a
further preparation of TEM samples of I: W = 4:1 samples, the details could be well
characterized as shown in Figure 5-29.
To further prove the influence of templating MFCs in the system, control was
conducted on TiF4 dissolved in isopropanol and water mixture at two ratios, 4:1 and 1: 1,
without any cellulose. Figure 5-26 shows the morphologies of coatings made on the surface
of glass substrates after reaction time 2 hours, 4 hours, and 6 hours. Only TiO2 clusters
were formed randomly and thick films with cracks were observed. Figure 5-27 shows a
more aggregated surface with TiO2 clusters, but still lacked a controlled morphology and
porosity.

197

Figure 5-26 TiF4 reactions in IPA and water mixture at I:W ratio of 4:1 for (a) 2 hours, (b) 4 hours,
(c) 6 hours.

198

Figure 5-27 TiF4 reactions in IPA and water mixture at a I: W ratio of 1:1 for (a) 2 hours, (b) 4 hours,
(c) 6 hours.

199

Transmission electronic microscopy (TEM):
As Figure 5-28 (a) and (c) show, porous titania film were composed of particles with
size ranging 10-20 nm and pore size ranging in 10-40 nm, attributed to the removal of
cellulose fibers. Particles were observed to be agglomerated due to the thickness of the film
and incompetence of sonication and dispersion. Figure 5-28

(b) SAED ring patterns

confirm that the TiO2 was made of polycrystalline anatase TiO2 with the Debye-Scherrer
concentric rings (101), (004) and (200). The 0.35 nm lattice fringe spacing (Figure 5-28
(d)) further implied the existence of anatase (101) facet (JCPDS No. 21-1272). Titania film
has (101) the most exposed face in the nanocrystal forms.

200

Figure 5-28 TEM images of anatase TiO2 film generated by TiF4 liquid phase deposition, with a
solvent ratio of IPA/H2O =0:1. (a) and (c) porous structure, at low and high magnification. (b) SAED
diffraction pattern and (d) HRTEM image noted with anatase lattice spacing ~0.352 nm of facet
(101) and 0.148 nm of (213)

As Figure 5-29 shows, nanotubular titania with diameter ranging 10-35 nm were found
and it was due to the removal of nano-size MFCs. As discussed before, even after
homogenization, different level of aggregation still existed in the cellulose suspension.
Some of them remained at the finest size, becoming ideal template for nanotubular titania
structures. Figure 5-29 (a-c). The 0.355 and 0.24 nm lattice fringe spacing (Figure 5-29
(d)) implied the existence of anatase (101) and (004) facets (JCPDS No. 21-1272).

201

Figure 5-29 TEM images of anatase TiO2 film generated by TiF4 liquid phase deposition, with a
solvent ratio of IPA/H2O =0:1. (a-b) tangled tubular-shaped TiO2, at low and high magnification. (b)
another TiO2 tube showing the hollow inner and outside wall, lattice spacing is 0.352 nm in the
inset image (d) corresponding HRTEM image of (c) highlighting the anatase lattice spacing of facet
(101) with spacing 0.355 nm and (004) with spacing 0.24 nm.

202

Figure 5-30 TEM images of anatase TiO2 film generated by TiF4 liquid phase deposition, with a
solvent ratio of IPA/H2O =4:1. (a-b) spherical TiO2, at low and high magnification. (c) interface
between two spherical particles, and the SAED pattern (inset) labeled (d) corresponding HRTEM
image of (c) highlighting the anatase lattice spacing of facet (101) and (004), and the lattice angle
between is 119.6°

When the IPA to water ratio was increased to 4:1, the morphology of the film changed
dramatically. As shown in SEM images previously, uniform sphere grew along the chain
and crossed over, resulting in a porous interconnected network. TEM images in Figure 5-30
(a) illustrate the neighboring particles with diameter around 150 nm. The inset image in
(b) revealed the anatase phase, corresponding to the SAED rings (105) (200) (004) (101)
(215) (220) (204) of inset image in Figure 5-30 (c). Similarly, the lattice space shown in (d)
203

confirmed the anatase planes of (101) and (004) again, with the lattice angel measured as
around 120°.
5.3.2.2 Surface and elemental properties of TiO2 films
Elemental analysis:
Energy-dispersive x-ray spectroscopy (EDS) was utilized with SEM for elemental
analysis. It is elucidated in Figure 5-31 that after calcination, no carbon was left in the film.
And the sodium, silicon, and calcium were from glass substrate background based on the
comparison of two inset quantitative results.

Figure 5-31 EDS information of a calcined TiO2 film (with MFCs in suspension of I: W =1:1) on a
glass substrate. The inset tables explained the atomic composition of oxygen and titanium, and the
background elements from the glass substrate.

Table 5-2 is a summary of the element percentages, corresponding to the full scan in
Figure 5-32. Before calcination, C 1s has a much higher atomic ratio in the film, indicating
organic composition in the film before calcination counted around 30%, and C 1s in all
other calcined films came from carbon contamination from atmosphere. The highest Ti 2p

204

concentration came from precursor without any IPA, which possibly was due to the highest
level of precipitation of titania. When IPA was added to adjust the polarity of the precursor,
hydrolysis and precipitation were slowed down, resulting in less TiO2, higher porosity on
the top surface of the film. Na and Si had been detected with repeating experiments, and
they were suspected to be from the chemical reactants contaminations, water used, or glass
substrates.
Table 5-2 Atomic concentration of related elements in TiF4- MFCs induced films detected by XPS.

Sample: TiF4 with or without
cellulose MFCs

C1s

O1s

F1s

Na1s

Si2p

Ti2P

1

0.5% w.t., 1 water + 4 IPA

25.3

50.5

0.5

6.3

7.8

9.5

2

0.5% w.t., 1 water +1 IPA

23.9

50.8

0.5

7.3

10.4

7.2

3

0.5% w.t., 0 water +1 IPA

25.7

49.2

0.4

8.5

2.9

13.3

4

sample (1), non calcined

54.1

39.9

3.8

0.4

-

1.8

5

0.5% w.t., 1water+0 IPA

23.0

50.4

0.5

7.7

4.1

14.3

6

water 10 mL, w/o cellulose

26.4

49.5

0.5

7.1

8.4

8.1

205

Figure 5-32 XPS spectra of thin titania films made with TiF4 precursors with MFCs, at different
solvent compositions: (a) IPA: water= 0:1, (b) IPA: water=1:1, (c) IPA: water= 4:1, (d) IPA: water=
1:0.

Table 5-2 is a summary of the element percentages, corresponding to the full scan in
Figure 5-32. Before calcination, C 1s has a much higher atomic ratio in the film, indicating
organic composition in the film before calcination counted around 30%, and C 1s in all
other calcined films came from carbon contamination from atmosphere. The highest Ti 2p
concentration came from precursor without any IPA, which possibly was due to the highest
level of precipitation of titania. When IPA was added to adjust the polarity of the precursor,
hydrolysis and precipitation were slowed down, resulting in less TiO2, higher porosity on
the top surface of the film. Na and Si had been detected with repeating experiments, and
they were suspected to be from the chemical reactants contaminations, water used, or glass
substrates.

206

Figure 5-33 XPS convolution results of thin titania dried films without calcination, made with TiF4
precursors with MFCs at IPA: water= 4:1. (a) C 1s, (b) O 1s, (c) Ti 2p, (d) F 1s.
Further surface chemical states of each element within the TiO2 films were obtained via XPS peak
fitting as shown in Figure 5-33 and Figure 5-34. C 1s distributed its chemical states bonding energy
as 284.8 eV (25%), 286.3 eV (57%), 287.7 eV (15.8%), and 288.6 eV (2.2%) in Figure 5-33 (a).
While for calcined titania films (Figure 5-34 (a)), the composition is 284.8 eV (71.2%), 286.4 eV
(12.7%), 288.4 eV (10.3%), and 289.9 eV (5.8%). The change from uncalcined to calcined,
corresponding to

Table 5-2 is a summary of the element percentages, corresponding to the full scan in
Figure 5-32. Before calcination, C 1s has a much higher atomic ratio in the film, indicating
organic composition in the film before calcination counted around 30%, and C 1s in all
other calcined films came from carbon contamination from atmosphere. The highest Ti 2p
concentration came from precursor without any IPA, which possibly was due to the highest
level of precipitation of titania. When IPA was added to adjust the polarity of the precursor,
207

hydrolysis and precipitation were slowed down, resulting in less TiO2, higher porosity on
the top surface of the film. Na and Si had been detected with repeating experiments, and
they were suspected to be from the chemical reactants contaminations, water used, or glass
substrates.
Table 5-2, is also similar when compared to other calcined titania films synthesized at
other IPA: water ratios. When the precursor was not calcined, the contribution of C-C
(~284.8 eV), C-O (~286.0 eV), O-C=O (~288.5 eV), C=O (~289 eV) was characteristic of
cellulose.

Figure 5-34 XPS convolution results of thin titania films after calcination, made with TiF4
precursors with MFCs at IPA: water= 4:1. (a) C 1s, (b) O 1s, (c) Ti 2p, (d) F 1s.

After calcination, the ratio of each bond was changed, because it came from
adventitious carbon on the surface. O 1s had dramatic change as demonstrated in Figure
208

5-33 (b) and Figure 5-34 (b). The peak at ~529.9 eV increased from 9 % to 46.5 % after
calcination, implying that metal oxide bond was formed. And the peak at 531-532 eV
dropped from 90.1 % to 30.0 % due to the removal of C-O bonds from the degradation of
cellulose at elevated temperatures. Ti 2p bonding energy transformed from 457.8 eV to
458.0 eV and maintained a similar Ti(IV) state. As for F 1s, after calcination, the chemical
state turned from organic fluorine (~686-688 eV) to metal fluoride (~684 eV), as shown in
in Figure 5-33 (d) and Figure 5-34 (d).
Crystallinity analysis:
Raman spectroscopy, as an easy quick way to detect the crystallinity of TiO2, was
utilized to compare two parallel samples, thin films with IPA: water = 0: 1 and 4: 1. Both
films reveal anatase crystallinity. The Raman spectra of the calcined film clearly show that
titania is anatase by the characteristic peaks at 144, 197, 399, 513, 519, 639 cm-1.16

Figure 5-35 Raman spectra of two thin titania films, made from TiF4 -MFCs precursor with different
solvent compositions: IPA: water= 0:1 and IPA: water= 4:1.

209

All samples were detected with peaks at around 25.3, 38.3, 48.3, 54.0, 55.2 and 63.0 °,
which represented plane (101), (004), (200), (105), (211) and (204) [JCPDS 78-2486].
Anatase phase could be obtained by appropriate annealing at 450-500 °C.

Figure 5-36 XRD of titania films made with TiF4/ MFCs templates, with different IPA: water solvent
compositions

Thermal analysis:
The thermal stability of both MFCs, uncalcined TiF4 with MFCs were studied to explore
the degradation of MFCs at elevated temperatures. Figure 5-37 shows several stages of the
calcination of cellulose. Between room temperature and 100 °C, the residue water content
not removed from drying process was evaporated, giving a major peak in the derivative
weight change curve. The following weight loss before the temperature reached around
200 °C was attributed to the detachment of the absorbed water on the surface of the porous
cellulose structure. At 300 °C, a large derivative weight change peak showed up, resulting
from the rapid loss of weight due to oxidation of cellulose. During this stage, dehydration

210

and decarboxylation will produce anhydrocellulose and combustible gases like aldehydes,
ketones, ethers, and so on. From 360 °C to 450 °C, the carbonaceous residues were
oxidized and resulted in CO2, CO and H2O . Around 0.2% w.t. char was left after heating
process till 500 °C.

Figure 5-37 TGA of dried MFCs. To eliminate the water evaporation step during thermal analysis,
sample was dried in vacuum oven overnight at 80 °C before the test. TGA was conducted in air
through a high resolution ramp program at a heating rate of 25 °C/min.

The thermal stability of MFCs combined with Ti precursor showed the degradation
during the heating process very well. In Figure 5-38,

the weight loss from room

temperature to ~100 °C was from the major evaporation of solvents, in which a small peak
at around 70 °C indicated the existence of organic solvent. Then the weight loss continued
till around 250 °C, during which the absorbed water was removed, and the shrinkage of the

211

TiO2 cross-linked structure started, eventually, the oxidization of cellulose began. At
around 274 °C, dehydration and decarboxylation of cellulose reached the highest rate and
produced anhydrocellulose and combustible gases like aldehydes, ketones, ethers, and so
on. From 360 °C to 450 °C, the carbonaceous residues were oxidized and resulted in CO2,
CO and H2O , while at the same time, titania was further crystallized. After 500 °C, around
46% w.t. residue was left and that was the weight of TiO2 if the 0.2% w.t. cellulose was
neglected.

Figure 5-38 TGA of uncalcined TiF4 precursor with template MFCs in a solvent of I: W at ratio 1: 1.
To eliminate the solvent evaporation step during thermal analysis, sample was dried in vacuum
oven overnight at 80C before the test. TGA was conducted in air through a high-resolution ramp
program at a heating rate of 25 °C/min

When the solvent was simplified to water only for TiF4 precursor, the preliminary
stage slightly differed from Figure 5-37. As Figure 5-38 shows, at 163 °C, around 8% w.t.
weight loss was observed. Although the weight loss percentage of both I: W = 1: 1 and 0 : 1

212

samples were similar, a steep drop of latter might be due to the difference in solvent
composition, or inadequate drying process of the tested sample. Two more peaks in
derivative weight change curve occurred at similar temperatures (310 °C and 400 °C) when
compared with Figure 5-37 due to cellulose degradation and TiO2 crystallization. The
weight maintained constant after 500 °C and resulted in 45% w.t. TiO2 crystals.

Figure 5-39 TGA of uncalcined TiF4 precursor with template MFCs in a solvent of I: W at ratio 0: 1.
To eliminate the solvent evaporation step during thermal analysis, the sample was dried in vacuum
oven overnight at 80C before the test. TGA was conducted in air through a high-resolution ramp
program at a heating rate of 25 °C/min.

The thermal stability of calcined TiF4- induced films was further measured with TGA
and Figure 5-40 shows the weight loss curves of two samples in different ratios of organic
and water solvent. The first drop before 200 °C was due to the removal of free-standing
water and adsorbed water molecules. There was an average of 1.2% w.t. weight loss

213

between 500 °C and 600 °C which should be from either the further degradation of
cellulose char or the transition of titania particles at much higher temperatures.

Figure 5-40 TGA of calcined TiF4 precursor with template MFCs in two different solvent systems: (1)
top curve with  was prepared in I: W at ratio 0: 1; (2) bottom curve with ☐ was prepared in I: W at
ratio 1: 1. TGA was conducted in air through a high-resolution ramp program at a heating rate of 15
°C/min.

XRD, Raman, UV-Vis, AT-IR, BET and TGA measurements on titania-coated silicone
nanofilaments did not provide additional insight beyond the results reported.14 We
attribute this to the thin film nature and to the low absolute amount of material which was
deposited on the glass slides. Similar situations applied for this work as well, the materials
were collected not in the original form of the film, instead, it was scraped from the
substrates, and the amount of materials had limitations for ideal results as we predicted.

214

As for thin films, especially nanoscale thin films, Raman signals could be covered by the
substrate information. To get improved signals, one should consider accumulating the
spectra for more time until getting an appropriate intensity. A general method is to
increase the integration time and number of running cycles. Changing the configuration of
the measured film might also help by turning the film surface 90 degrees, from vertical to
the laser beam, to parallel to the laser beam. In addition, if possible, subtracting the signals
of background substrate from the whole thin film could yield a relatively accurate spectrum
of the film itself as well.
5.3.3 Photocatalysis performance of tunable titania thin films

Figure 5-41 Concentration change of the MO solution on glass substrate measured by UVVis absorbance, over a 20-minute UV lamp irradiance, without any catalyst.

Before the photocatalytic degradation test, 1×10-5 M MO solution was tested with no
catalyst applied at the beginning. Absorbance data from UV-Vis were taken every five
minutes and converted to concentration change. Figure 5-41 clearly shows that over UV
irradiation, the concentration of MO solution has minimum change of ~2%, which also

215

corresponds to the literature.17 Besides of the UV degradation to MO, one of the possible
reasons for the change might be the trace evaporation of the solution volume.
The ideal concentration of the MO solution should no be too low or too high. When 1×
10-5 M concentration was used, the extent of change was too small to be well observed and
compared. When the concentration was too high, the coverage of MO on TiO2 surface was
too saturated to ensure adequate conversion of photons. Similarly, the volume of solution
cast on the TiO2 film matters too. An optimum concentration and volume combination was
set to be 5×10-4 M and 10 uL for an 1 inch by 1 inch titania film.

Figure 5-42 Reflectance measurement of TiO2 film made from TiF4/ MFCs illuminated by UV
lamp over time. (a) the reflectance of initial TiO2, UV treated by 1 min, 2 min, and
3 min. (b) Subtracted from initial reflectance, v.s wavelength.

One set of titania blank films were tested with UV lamp illumination for 1-minute
interval each. From Figure 5-42 (a), the reflectance of the TiO2 film was scanned after each
UV treatment, and the reflectance change based on initial pristine TiO2 was shown in (b),
where a 5% drop was observed from time 0 to 3 minutes at wavelength around 350 nm.
The change is likely caused by the band gap difference brought in from UV radiation.

216

Figure 5-43 Reflectance change over time of UV-illuminated TiO2 film with 1×10-5 M MO
solution adsorbed. Titania film was made with TiF4/ MFCs and IPA: water = 1: 1. (a) Reflectance
change from clean TiO2, MO solution adsorbed (after dye), and treated with UV lamp for a total
time of 1 min, 2 min, 4 min, 6 min, and 9 min. Inset image showed details of the same curves. (b)
Reflectance difference over time, from dye adsorption, to photocatalytic degradation.

Figure 5-44 For film made with TiF4/ MFCs and IPA: water = 1: 1. (a) Reflectance difference change
at wavelength 395 nm v.s. time. (b) pseudo-first-order kinetic model of ln(reflectance)
v.s. time.

Before MO solution was adsorbed onto titania films, the reflectance data is shown by
the black curve in Figure 5-43 (a), while after dye equilibrium, the red curve represented
the change. Once photocatalytic degradation started, the curve moved back to the original
pure TiO2 stage gradually over time, as shown in the inset image in Figure 5-43 (a). Another

217

way to illustrate the degradation change is shown in Figure 5-43 (b) where the reflectance
difference for dye elimination at 395 nm was calculated and plotted. Further calculations in
Figure 5-44 explain the relationship between dye decomposition amount (R0-R) and time,
then generated the pseudo-first-order kinetic model in Figure 5-44 (b). The rate constant
was calculated as 0.30 min-1, and the degradation was completed to 77.7% of the original
dye concentration, for the films made with TiF4/ MFCs and IPA: water= 1:1 solvents.

Figure 5-45 Reflectance change over time of UV-illuminated TiO2 film with 1×10-5 M MO
solution adsorbed. Titania film was made with TiF4/ MFCs and IPA: water = 4: 1. (a) Reflectance
change from clean TiO2, MO solution adsorbed (after dye), and treated with UV lamp for a total
time of 1 min, 3 min, 5 min, and 7 min. Inset image showed details of the same curves. (b)
Reflectance difference over time, from dye adsorption, to photocatalytic degradation.

Similarly, Figure 5-45 and Figure 5-46 show the degradation of MO by the films made
with TiF4/ MFCs and IPA: water = 4:1 solvents. The rate constant was calculated to be 0.52
min-1, and the degradation was improved to 97%.

218

Figure 5-46 For film made with TiF4/ MFCs and IPA: water = 4: 1. (a) Reflectance difference change
at wavelength 395 nm vs. time. (b) pseudo-first-order kinetic model of ln(reflectance)
vs. time.
Table 5-3 Summary of kinetics of photocatalytic degradation of MO by variable TiF4- induced TiO2
films

TiF4-

TiF4-

TiF4-

TiF4-

TiF4-

TiF4-

TiF4-

water-1

water-2

cellulose1IPA1Water-1

cellulose1IPA1Water-2

cellulose4IPA1Water-1

cellulose4IPA1Water-2

cellulose4IPA1Water-3

Rate constant
(k, min-1)

0.25

0.25

0.32

0.34

0.61

0.78

0.76

Intercept (b)

2.33

2.25

2.84

2.66

3.08

3.13

3.28

R2

0.9417

0.9374

0.956

0.9402

0.9887

0.9877

0.9786

Degradation
percentage(%)

56

63

72

76

92

96

97

y=-k+b

Photocatalytic performance of TiO2 films made with TiF4 and water only without
cellulose templates, and with cellulose but variable IPA: water solvent composition are
shown in Figure 5-47. Table 5-3 summarized the rate constants and linear modeling Rsquared values, as well as the degradation percentage. The average rate constant and
degradation percentage was, 0.25 min-1 , 60% for TiF4- no cellulose films, 0.33Âą0.01 min-1 ,

219

74% for TiF4- cellulose-1IPA-1Water films, and 0.72Âą0.09 min-1, 95% for TiF4- cellulose4IPA-1Water films.

Figure 5-47 Variable films (a) reflectance change (b) kinetics with linear modeling
and equations.

220

5.3.4 Porous titania network via ammonium hexafluorotitanate and boric acid
method
A liquid phase deposition method was used to apply the aqueous titanium precursor to
the cellulose/water suspension. In the acidic environment, the solution supersaturation is
low. The stepwise hydrolysis and dehydration of (NH4)2TiF6 produce TiO2 nanoparticles
coated along the hydrophilic cellulose fiber surface through heterogeneous nucleation.
Figure 5-48 presents the transformation of titanium precursor with cellulose before
and after calcination. Ratios of BA to AHFT varied and directly had an influence on the
product. A growing BA ratio resulted in more cracks within the film surface, while more
AFHT led to dense and non-porous anatase titania films. Specific shaped amorphous TiO2
were formed before calcination. From (a) to (b), primary nucleation was annealed into a
fine web-like film; Particle size shrunk from (c) to (d), and (e) to (f) had a shape change
from cubic to spherical (g) and (h) showed a denser layer of titania probably due to
increased Ti concentration. When boric acid is fixed at 0.15M, by increasing the Ti
concentration from 0.05M- 0.15M-0.30M, different morphologies formed. Higher resolution
SEM images in Figure 5-49 reveal fiber-like titania materials, duplicating the size of the
cellulose templates at ~30 nm. Flower-like TiO2 also surrounded these TiO2 fibers.

221

Figure 5-48 Comparison between “before calcination” and after calcination of AHFT and BA with
MFCs composite. (a, c, e, g) are before calcination while (b, d, f, h) are calcined. (a,b) has a molar
ratio of BA: AHFT= 0.15:0.05. Similarly, (c,d)’s BA: AHFT= 0.30:0.05; (e,f)’s BA: AHFT= 0.30:0.30;
(g,h)’ BA: AHFT= 0.15:0.30.

222

Figure 5-49 Calcined titania film from AHFT and BA with MFCs precursor, and the molar ratio of
BA: AHFT is 0.15:0.05.

To study the parameters influencing the morphology, a series of experiments were
conducted at different reaction temperature, time, concentration and binder content. For
example, the results from set BA: AHFT = 0.15 M: 0.15 M were compared between
temperatures kept at 50 °C or just room temperature. White precipitates appeared after 30
min in 50 °C reaction, while suspension looked the same after 10 h for room temperature.
(Figure 5-50)

223

Figure 5-50 BA: AHFT= 0.15:0.15 films, calcined. (a-b) low and high magnification SEM images of
24h-reaction at room temperature; (c-d) low and high magnification SEM images of 24 h-reaction at
50 °C.

Increasing the mass percentage of binder PEG caused the change of morphologies for
the film deposited from 1: 1 precursor with MFCs, as shown in Figure 5-51. The ratios of BA:
AHFT were lowered to 0.075 M: 0.025 M, or 0.025 M: 0.008M. For the 0.075: 0.025 ratio, no
obvious difference was observed for 6 hours and 15 hours reaction times. However, in
general lowering the concentrations adopted an evolution of morphology from SEM images
(Figure 5-52). The addition of binder PEG created more porous structures when BA:
AHFT = 3: 1.

224

Figure 5-51 SEM images of titania films prepared from BA-AHFT/MFCs precursor with BA: AHFT =
0.15M: 0.15M. From (a) to (c), PEG weight percentages were 1%, 5%, and 10%.

Figure 5-52 SEM images of titania films prepared from BA-AHFT/MFCs precursor with low
concentrations. (a-b) BA: AHFT=0.075 M: 0.075 M, (c-d) BA: AHFT=0.025 M: 0.025 M, (e-f) 0.025 M:
0.008 M. (a, c, e) was recorded without addition of binder, while (b, d, f) was taken with addition of
PEG.

225

Figure 5-53 Raman spectrum of titania films made from AHFT-BA/MFCs precursor: (a) BA:AHFT =
0.15: 0.05 (b) BA:AHFT = 0.15:0.15

Figure 5-53 shows the characteristic peaks for anatase phase at wavenumber around
of 149, 199, 399, 519, 639 cm-1, while the peaks at 513 cm-1 were weaker to be observed.
Anatase is tetragonal with two TiO2 units per primitive cell. The six Raman modes are
related to symmetries A1g+2B1g+3Eg. Both 1:1 and 3:1 ratios of 0.15 M base Ti(IV) were
tested and confirmed anatase.

226

Figure 5-54 TEM images of the titania films made from AHFT-BA/MFCs precursor, and the ratio of
BA: AHFT = 0.15M : 0.05M. From (a) to (d) magnification increased. Measurement of lattice (d)
shows the lattice space of 0.35 nm and 0.24 nm, respectively referred to plane (101) and (004).

5.4 Conclusions and Future Work
Liquid phase deposition was utilized as a simple and effective method to prepare TiO2
nanomaterials with tunable morphologies and porosities. Preliminary experiments
successfully prepared TiO2 coatings on cotton rounds fibers and filter paper with titanium
tetrafluoride precursor and ammonium hexafluorotitanate/boric acid, creating hollow
ribbon materials after the removal of cellulose templates. However, the dimensions of
these titania materials were much larger than nanoscale. To further improve the surface
area, porosity, and potential applications, home-made micro/nano-fibrillated cellulose
fibers were combined with liquid phase deposition method.

227

TiF4 system has generated bulk materials and thin films. For bulk materials, reaction
time, pH level, surface properties and the effects of MFCs templates were studied. For thin
film materials, an optimized solvent composition of isopropanol/water ratio of 4 to 1 was
found to yield coatings with uniform sphere-shaped TiO2 possessing a chain-like
morphology oriented along the axis of the heat-removed cellulose fiber templates. Effects
of cellulose and reaction time were studied as well.
Photocatalytic activity of the nano-chained titanium dioxide was investigated by
monitoring the degradation of methyl orange (MO) with the use of quantitative diffuse
reflectance UV-Vis spectroscopy. The enhancement in photocatalytic performance was
explained regarding the surface area, pore size, and the unique morphology. The threedimensional web structure with pseudo one-dimensional sphere-chain could retard the
recombination of photogenerated electron-hole pairs and improve the charge transport. An
increase of decomposition rate was found when the sample was compared with the
traditional water-based precursor. The average rate constant and degradation percentage
were 0.72Âą0.09 min-1, 95% for TiF4- cellulose-4IPA-1Water films, increased by 1.88 times of
the film prepared without cellulose templates.
Ammonium hexafluorotitanate and boric acid system was prepared. Reaction time,
temperature, reactant concentration and ratio, and binder addition were investigated for
the titania formation. The morphology evolved from aggregated particles to spheres, and
flower-shaped, with increased surface area, compared to non-hydrolytic methods.
To remove cellulose fibers, heating was required in our work. The crystallinity from
liquid phase deposition alone has been proved to be excellent, for photocatalytic activities

228

from others’ work.8 Future directions could be focused on the photocatalytic woven
structures with MFCs as the substrates, thus no additional high-temperature annealing will
be required.
Based on the completed results, it was concluded that the surface chemistry of nanosized MFCs is different from the traditional large-scale (Îźm to cm range) cellulose or other
natural fibers reported on in the literature. It is important to understand what differences
exist in the surface chemistry of the cellulose, either in water or isopropanol, or even dried
paper form.

229

BIBLIOGRAPHY

230

BIBLIOGRAPHY
(1)

Yu, J.; Fan, J.; Cheng, B. Dye-Sensitized Solar Cells Based on Anatase TiO2 Hollow
Spheres/Carbon Nanotube Composite Films. Journal of Power Sources 2011, 196,
7891–7898.

(2)

Yu, J.-G.; Yu, H.-G.; Cheng, B.; Zhao, X.-J.; Yu, J. C.; Ho, W.-K. The Effect of Calcination
Temperature on the Surface Microstructure and Photocatalytic Activity of TiO
2Thin Films Prepared by Liquid Phase Deposition. J. Phys. Chem. B 2003, 107,
13871–13879.

(3)

Maeda, M. Preparation of Titania Films with Cohered Nanosized Particles Using
Improved Liquid Phase Deposition Process. international journal of electrochemical
science 2015, 10, 2988–2996.

(4)

Malvadkar, N.; Dressick, W. J.; Demirel, M. C. Liquid Phase Deposition of Titania
Onto Nanostructured Poly-P-Xylylene Thin Films. J. Mater. Chem. 2009, 19, 4796–
4799.

(5)

Li, S.; Liu, J.; Feng, T. Low Temperature Coating of Anatase Thin Films on Silica Glass
Fibers by Liquid Phase Deposition. J. Wuhan Univ. Technol. 2007, 22, 136–139.

(6)

Lee, M.-K.; Fan, C.-H.; Wang, H.-C.; Chang, C.-Y. Fluorine Passivation of Titanium
Oxide Films on ITO/Glass Grown by Liquid Phase Deposition for Electrochroism. J.
Electrochem. Soc. 2011, 158, D511.

(7)

Maki, H.; Okumura, Y.; Ikuta, H.; Mizuhata, M. Ionic Equilibria for Synthesis of TiO2
Thin Films by the Liquid-Phase Deposition. J. Phys. Chem. C 2014, 118, 11964–
11974.

(8)

Shimizu, K.; Imai, H.; Hirashima, H.; Tsukuma, K. Low-Temperature Synthesis of
Anatase Thin Films on Glass and Organic Substrates by Direct Deposition From
Aqueous Solutions. Thin Solid Films 1999, 351, 220–224.

(9)

Imai, H.; Takei, Y.; Shimizu, K.; Matsuda, M.; Hirashima, H. Direct Preparation of
Anatase TiO2 Nanotubes in Porous Alumina Membranes. J. Mater. Chem. 1999, 9,
2971–2972.

(10)

Pourmand, M.; Taghavinia, N. TiO2 Nanostructured Films on Mica Using Liquid
Phase Deposition. Materials Chemistry and Physics 2008, 107, 449–455.

(11)

Imai, H.; Matsuta, M.; Shimizu, K.; Hirashima, H.; Negishi, N. Preparation of TiO2
Fibers with Well-Organized Structures. J. Mater. Chem. 2000, 10, 2005–2006.

(12)

Yu, H.; Yu, J.; Cheng, B.; Lin, J. Synthesis, Characterization and Photocatalytic Activity
of Mesoporous Titania Nanorod/Titanate Nanotube Composites. Journal of

231

Hazardous Materials 2007, 147, 581–587.
(13)

Yu, H.; Yu, J.; Cheng, B.; Liu, S. Novel Preparation and Photocatalytic Activity of OneDimensional TiO 2hollow Structures. Nanotechnology 2007, 18, 065604.

(14)

Meseck, G. R.; Kontic, R.; Patzke, G. R.; Seeger, S. Photocatalytic Composites of
Silicone Nanofilaments and TiO2 Nanoparticles. Adv. Funct. Mater. 2012, 22, 4433–
4438.

(15)

Dozzi, M.; Selli, E. Specific Facets-Dominated Anatase TiO2: Fluorine-Mediated
Synthesis and Photoactivity. Catalysts 2013, Vol. 3, Pages 455-485 2013, 3, 455–485.

(16)

Phase-Pure TiO2 Nanoparticles: Anatase, Brookite and Rutile. 2008, 19, 145605.

(17)

Li, Y.; Li, X.; Li, J.; Yin, J. Photocatalytic Degradation of Methyl Orange by TiO2Coated Activated Carbon and Kinetic Study. Water Research 2006, 40, 1119–1126.

232

6 Nanoporous TiO2 Films Templated with Micro/nanofibrillated Cellulose and Applications in Dye-Sensitized Solar
Cells
6.1 Background
6.1.1 Nano-structuring of the photoanode of DSSCs
As introduced in Chapter 1, The design of nanocrystalline semiconductor electrode is
significantly important because of its responsibility in the loading of dyes, transport of
electrons, and diffusion of electrolytes but also the suppression of recombination reactions.
Natural templates have been researched as promising candidates to create tunable
titania structures for applications in DSSCs. Nanostructured hollow TiO2 fibers were
prepared via fibers from cotton rounds, and used as porous photoanode for DSSCs,1
resulting in enhanced electron transport properties, compared to randomly packed
spherical nanoparticle photoanode. The transport of photoinjected electrons in the oxide
network got 3-4 times of lifetime, due to the faster travel within fiber morphology. The
efficiency of the cell reached 7.2% under simulated AM 1.5 solar illumination. However, the
fiber size was large still with ~2 um diameter and ~100 um length. The preparation
process was complicated due to the fact that materials had to be prepared, thermally
treated, washed, and then mixed with other chemicals to form a paste.
Another interesting approach was through a virus, in a similar mechanism.2 Highly
interconnected and nano-channeled titania films were fabricated with the M13 virus as a
template. After calcination, viruses were decomposed and left trace channel structure

233

inside TiO2 particles. This not only provided access to electrolyte but also increased surface
area and porosity. The efficiency of the DSSC generated was 6.32%. Commercial TiO2
particles were used in the film with binder PEG for better slurry thickness.
A modified sol-gel based template-controlling approach is proposed to assist the
formation of micro/nano channels in the original random nanoparticle network, to
optimize the dye loading, interfacial contact with the electrolyte, and the electrical contact
efficiency.
6.1.2 Controlled hydrolysis in a sol-gel process for MFCs templating
Chapter 1 has discussed sol-gel to synthesize metal oxide materials and the specialty of
titanium precursors as transition metal source. Sol-gel derived materials can be prepared
by tailoring parameters for high surface area, controlled porosity, and desired electric and
optical properties. A controlled hydrolysis is crucial for manipulation of materials size,
morphology, and chemical properties. Chapter 3, 4, and 5 comprehensively demonstrated
how cellulose fibers worked as soft templates for inorganic materials fabrication in nonhydrolytic and hydrolytic media and the enhancement in photocatalysis. To further the
exploration of cellulose templating and titania film applications, the combination of MFCstemplated titania and the DSSCs anode will be discussed in this chapter.
MFCs will be combined with the titanium precursor in modified sol-gel systems,
resulting in the encapsulation of the MFCs in the growing direction of the TiO2 crystalline
nanoparticles. After calcination, the MFCs will cleanly decompose leading to the formation
of micro/nano channels in the TiO2 nanoporous film. Other systems such as TiF4 and
ammonium hexafluorotitanate and boric acid will be illustrated by similar means as Figure

234

4-1 shows. The benefits are expected to be a more ordered interconnection and less
aggregation with nano/micro channels.

Figure 6-1 Scheme of applying MFCs as a template for porous TiO2 network synthesis

6.2 Experimental
6.2.1 Materials and equipment
Materials:
All chemicals were used without pre-treatments unless noted. MFCs in isopropanol
(IPA) were processed by Mini DeBEE ultrahigh-pressure homogenizer as Chapter 3
described. Titanium(IV) isopropoxide (≥ 97%), valeric acid (≥ 99%), ethanol (≥ 99.5%,
A.C.S. reagent), 2-propanol (A.C.S. grade, ≥ 99.5%), Lithium iodide (99.9% trace metal
basis), iodine (≥ 99.99%), acetonitrile (≥ 99.99%), and anhydrous 2-propanol were
purchased from Sigma Aldrich. Hydrochloric acid (analytical reagent grade) was from
Mallinckrodt. Commercial titania paste Ti-Nanoxide T (transparent, 11% w.t.)，25μm-thick
Surlyn and N,N’-bis(4,4’-dicarboxylate-2,2’-bipyridine)ruthenium(II) (N3 dye) were
235

purchased from Solaronix. Ethanol (200 proof, 99.5% ACS grade) and isopropanol (99.5 %)
were purchased from MSU university stores. Hydrogen hexachloroplatinate (IV) hydrate
(40% Pt) was purchased from Acros Organics. 4-tert-butylpyridine was purchased from
TCI America Inc. The silver conductive epoxy (8331G) was purchased from Fulton Radio
Supply Co. Tapes for the mask frame were 3M Scotch 810 & 811, and the microscope cover
glasses at a thickness of 0.14 mm were from Corning. Plain, pre-cleaned micro glass slides
were purchased from VWR International, in a size of 1 inch by 3 inches, with a thickness of
1.2 mm. TEC-15 (15 Ί/mm2) fluorine-doped tin oxide (FTO) glass was purchased from
Hartford Glass Co.
Equipment and Characterizations:
Details of sample preparation have been stated in previous chapters, and most of the
material characterization methods used in this chapter have been described. Thus, only a
brief description is provided here.
The structure of the cellulose (MFCs) templates and titania porous thin films were
characterized by scanning electron microscopy (Carl Zeiss Auriga CrossBeam SEM-FIB).
Thickness of films were characterized after milling into the cross section of the film by
focused ion beam (FIB). The transmission electron microscopy (TEM) images, highresolution transmission electron microscopy (HRTEM) images, and selected-area electron
diffraction (SAED) patterns were taken on a JEOL 2200FS equipment. UV-Vis
spectrophotometer (Perkin Elmer Lambda 900 UV Spectrometer) was applied for TiO2
films absorption and transmission characterizations. UV-vis absorption spectra of the TiO2
films were recorded in the range of wavelength 300-800 nm. Fourier transform infrared

236

(FTIR) absorption spectra were collected with Perkin Elmer 2000 Spectrum One FT-IR
spectrometer at room temperature. The spectra of the composites were scanned from 4000
to 400 cm-1. Thermogravimetric analysis (TA instruments Q500) used a high-resolution
ramp of 25 °C/min to ~800 °C in air was conducted on TiO2 samples. X-ray photoelectron
spectroscopy (XPS) was carried out on Perkin Elmer Phi 5400 using KÎą line of Mg ( 1254
eV) X-ray source with a beam voltage of 15 kV and current of 20 mA for the survey spectra
scan (pass energy 187.85 eV, and 1 eV step size). A HIMONT Plasma Science PS 0500
Plasma Surface Treatment System was used to clean the glass substrates for ~0.5h. The
oxygen pressure was 0.270 Torr and the power was 300 W. Photoelectrochemical
measurements were performed with a JV-EQE solar cell testing station with an AM 1.5
solar filter used to simulate sunlight at 100 mW cm-2. Atomic layer deposition was finished
with Savannah 100 instrument (Cambridge Nanotech, Inc.).
6.2.2 Synthesis of porous TiO2 network as thin film
To moderate the hydrolysis and precipitation rates during the sol-gel reactions, valeric
acid was used as a chelating agent for the non-aqueous environment. Valeric acid (v-acid)
was added to titanium isopropoxide (TTIP) followed by water as shown in Figure 4-3, The
atomic ratio of v-acid: water: TTIP is 9: 1.5: 1 respectively.3 The mixture was vigorously
stirred in capped vials at room temperature for 1 h. Then the as-prepared titanium valerate
was combined with MFC/isopropanol suspension at a ratio of 2: 1 and kept in a 50 °C oil
bath for 14 h. The mixture turned from a colorless, transparent solution to cloudy.

237

Figure 6-2 Scheme of MFCs-templating slow sol-gel method

Films were then dried in vacuum oven for 0.5 h without heating to avoid moisture,
followed by a step-wise annealing at 200 °C, 300 °C, 350 °C, 450 °C for 0.5 h, 0.5 h, 0.5 h,
and 1 h.
6.2.3 DSSC assembly and measurements
To investigate the potential of MFC-templated TiO2 as a DSSC photoanode material, a
series of solar cells were assembled and current density vs. voltage (J-V) measurements
were conducted. Using a commercial titania paste and lab-made Ti(IV) precursors as
examples, the procedure of assembling a test-ready DSSC is described as below:
Cleaning:
Fluorine-doped tin oxide (FTO) glasses were cut into 0.75 inches by 1 inch (or 0.5
inches), and 1 inch by 1 inch slides for anode and cathode use. A cleaning procedure is
essential to remove dirt and oil from the original packaging. First FTO was cleaned with
soap water (Alconox Powder Detergent) and then rinsed in acetone, isopropanol, and
ethanol in the sonication bath for 20 minutes each. After dried in air, FTO glass slides were
located in a glass or metal petri dish ready for a 30-minute oxygen plasma treatment where
the oxygen pressure was 0.270 torr and the power was 50% of 600 W. Thus, most of the
contaminants were removed and hydroxyl groups were exposed as well. At the same time,

238

tools for the coating step such as glass pipets, razor blades etc. should be cleaned with
acetone and let dry in the hood. Based on previous experience, the furnace used for titania
annealing should be purified at 600 °C for 1 hour to ensure no cross contamination from
materials of other users.
Coating:
A coating frame was made simply by weighing paper and Scotch 810 tape with the
dimensions of 1 cm by 1 cm (or 0.6 cm by 0.6 cm), which would serve as the mold for the
tape mask every time. To make a tape mask, overlay the removable Scotch 811 tape onto
the coating frame, apply pressure, and then cut precisely along the 1 cm by 1 cm edges with
a sharp razor blade. The tape mask was attached onto the conductive surface of the FTO
glass and no ridges or bumps should be created along the tape.
A small amount of commercial titania paste was applied to the bottom edge of the tape
mask, and then doctor-bladed across the 1 cm by 1 cm area using a cleaned glass pipet. For
lab-synthesized titania, the precursor was either doctor-bladed in the same way, or by drop
casting and tilting for accumulating layers. The TiO2 photoanode film was made by drop
casting 20 uL Ti/cellulose (2:1) precursor and spread evenly within the tape mask. After
the film was dried for 10 min, accumulating layers could be applied and dried again when
necessary. The coated slides went through a two-stage calcination. Note that different
procedures were followed when the film was lab-synthesized with or without templates, or
by different systems. After titania coating was dried in air for more than 15 minutes, the
tape mask was carefully peeled off.

239

The coated slides were annealed in air in the furnace at 450°C for 30 minutes for
commercial paste. Different procedures were followed when the film was lab-synthesized
with or without templates, or by different systems. Details about these can be reached in
previous chapters.
Dye loading:
0.3 mM N3 dye in ethanol was prepared and bath-sonicated for 20 minutes. The
annealed TiO2-FTO glass was treated in oxygen plasma chamber again to remove possible
contaminants and quickly heated up to 90 °C on a hot plate. Slides were immersed in a 10
mL beaker with as-prepared dye solution covering at least the top of the titania area.
Beakers were also covered with paraffin films to prevent evaporation. All the slides were
immersed in dye solution for 20-24 h in the dark.
Cathode fabrication:
Pre-cut FTO slides were drilled with two 1 mm-in-diameter holes. The holes should be
aligned in the diagonal direction located at the corner of the active TiO2 area to ensure a
smooth percolation of the electrolyte. To save time, cleaning of the cathode substrates
should be followed after drilling for cathode. Deposition of a thin layer of Pt catalyst was
performed in the following steps: 32.3 uL of 5 mM H2PtCl6 in IPA solution was applied to
the conductive side of FTO and spread on the surface by tilting. It is important to avoid
leaking of solution into drilled holes. After dried for 5-10 min on a flat surface in
the hood, cathode slides were calcined at 400 °C for 15 min (for commercial paste) in a
furnace and then cooled down.

240

Wire electrodes can be applied using 2’’ copper wires and silver epoxy. The polymer
sheathing of copper wires was removed and the wires were cleaned with acetone and
ethanol. Silver epoxy was prepared by mixing double tubes quickly and let set till it was
sticky. Then a string of copper wire was attached to the farther edge away from
holes on FTO slide with a dollop of silver epoxy. The epoxy can be cured by
heating at 65 °C for 15 min or at room temperature for five hours.
Cell assembly and electrolyte injection:
When the dye loading process was finished, the assembly could be started. The dyed
photoanode was washed with ethanol to remove the excess dye, and dried with nitrogen.
Pieces of surlyn were cut 0.5’’ by 0.5’’ with the center cut out leaving a 2-3 mm thick frame.
The surlyn frame was placed around the dyed area with no ridges and then covered with
the platinized counter electrodes with the conductive surface facing down. The assembly
was pressed sealed for 20-30 seconds with another piece of thick glass on a 120 °C hot
plate. A well-sealed surlyn between both electrodes should become translucent with no
bubbles or open channels seen by naked eyes.
Before the electrolyte percolation, another piece of copper wire was attached to the
conductive side of the photoanode, and cured for 15 minutes at 65 °C. The standard
electrolyte consists of 0.3 M lithium iodide, 0.03 M iodine, 0.2 M 4-tert-butylpyridine in
acetonitrile. A 1-mL volumetric flask was used to prepare the fresh solution for each single
time use. The solution was dispersed in sonication bath minimally (both iodine and
acetonitrile easily vaporize) and transferred into a tightly capped vial.

241

The as-prepared cell was placed photoanode-side down, and a 0.1 mL droplet of the
electrolyte was placed close to one hole using a syringe or micropipette. To prevent bubble
formation, the cell could also be vertically positioned so that the electrolyte was drawn in
with the help of gravity. After the excess electrolyte solution was wiped off, immediately a
soldering iron should be used to melt the surlyn beneath the cover glass to seal the holes.
Sometimes, the electrolyte emerges under the surlyn and will cause it not to melt properly.
Re-wipe the glass and dry the contact surface sufficiently if necessary.
Characterization of solar cells was conducted with a JV-EQE solar cell testing station
with an AM 1.5 solar filter used to simulate sunlight at 100 mW cm-2. Voltage was applied in
a two-electrode arrangement, where the working electrode lead was the photoanode, and
the reference and counter electrode leads were the cathode. The solar cell under evaluation
was secured onto an irradiation port, and the solar-simulated light beam could be adjusted
regarding intensity, angle, and position. The same position has been located for the beam
source to ensure the irradiation created was maintained at the same level. A mask with an
open area in the center was also made to apply to the solar cell before testing so that only
the aimed TiO2 area was interacting with the beam source.

6.3 Results and Discussions
6.3.1 Applications of cellulose-templated TiO2 as photoanode in DSSC
SEM images in Figure 6-3 showed the morphology and cross-sectional view of the
calcined 4-layer film. TEM images further demonstrated the particle size of ~20 nm at
higher magnification and resolution, and the diffraction pattern shows an anatase phase.

242

Solar-to-electricity performance was tested on a series of cells made with different
numbers of layers of TiO2 (Figure 6-4).

Figure 6-3 Four-layer TTIP-cellulose precursor, calcined, for DSSCs. SEM images: (a) top view, (b)
cross section view. TEM images: (c-d) at higher magnifications and resolutions, lattice plane
measured 0.3 nm (e) selected area diffraction pattern.

6.3.2 Photovoltaic performance test with ttip-vacid films and other films
Based on Figure 6-4 and Table 6-1, the efficiency of the solar-to-electricity
transformation increased with the increasing number of layers at first and then decreased
for thicker films. Both photocurrent and photovoltage increased first and then decreased
when the thickness went up.

243

Figure 6-4 J-V curves of TTIP-cellulose deposited DSSCs in dark and luminance
Table 6-1 Photocurrent, photovoltage, fill factor and efficiency of TTIP-cellulose cells
Device
1
2
3
4
5
6

Photoanode
material
1 layer
2 layers
4 layers
5 layers
6 layers
10 layers

Photoanode
thickness (nm)
644
727
1435

Jsc (mA/cm2)

Voc (V)

FF

Efficiency Ρ

-0.50
-1.90
-2.77
-2.60
-4.09
-1.76

0.72
0.63
0.82
0.78
0.78
0.75

0.51
0.40
0.52
0.48
0.55
0.60

0.17 %
0.49 %
1.20 %
0.98 %
1.75 %
0.80 %

Figure 6-5 J-V curves of commercial Solaronix paste deposited photoanode for DSSCs in dark and
luminance

Solaronix 3% w.t. TiO2 as a photoanode material:
A commercial paste Solaronix Ti-Nanoxide T (11% w.t.) was applied by the doctor
blade with microscope cover glass and calcined, forming a 4.1 Îźm thick film. To normalize

244

the thickness influence, Ti-Nanoxide T was diluted to 3% w.t. and sonicated thoroughly
with ethanol in a capped vial. Solaronix 3% T 1 layer and 2 layers were made by doctor
blade as well. Solaronix 3% T with cellulose was prepared by adding cellulose/isopropanol
suspension into the original paste and sonicating. Figure 6-5 indicates that at the
approximately same thickness, homemade 6-layer TTIP-cellulose TiO2 photoanode had
higher Jsc increased by ~59% and efficiency improved by ~44% (Table 6-2). While when
the thickness exceeded the TTIP-cellulose films, the performance of the whole cell was
further behind commercial anodes assembly. The addition of cellulose/isopropanol
suspension did not improve the overall efficiency.
Table 6-2 Photocurrent, photovoltage, fill factor and efficiency of Solaronix TiO2-cellulose cells
Device
1
2
3
4

Photoanode
material
Solaronix 3% T_1
layer
Solaronix 3% T_2
layers
Solaronix 3%
T_with cellulose
Solaronix 11% T

Photoanode
thickness (nm)
717

Jsc (mA/cm2)

Voc (V)

FF

Efficiency Ρ

-2.57

0.78

0.60

1.21 %

1350

-5.59

0.80

0.61

2.73 %

-

-2.14

0.8

0.68

1.17 %

4100

-7.90

0.73

0.59

3.26 %

Compact blocking layer (CBL) influence on different cells:
Before any photoanode material deposition, a thin CBL was applied on FTO with 500
atomic layer deposition (ALD) cycles using TTIP and water as precursors. TiO2 was grown
at 225 °C using reactant exposure times of 0.3 s and 0.015 s, respectively, and nitrogen
purge times of 5 s between exposures. Then 4 layers, 6 layers of TTIP-cellulose precursor
and Solaronix 3% TiO2 were used as photoanodes and tested, as Figure 6-6 shows below.

245

Figure 6-6 J-V curves of different DSSCs in dark and luminance, with compact block layer deposited
or without.

As stated above, back reactions suppressed the solar-to-electricity efficiency, especially
the one where photogenerated electrons from anode reacted with the liquid electrolyte.
When illumination is not sufficient, the FTO/electrolyte back reaction is more dominant. An
increase of Voc , Jsc and efficiencies was attributed to the compact pinhole-free thin barrier
layer between FTO glass and the nanoporous titania layer application, which blocked the
loss pathway and retarded the charge recombination.
Table 6-3 Photocurrent, photovoltage, fill factor and efficiency of DSSCs w/ or w/o compact layers
Device

Photoanode material

Jsc (mA/cm2)

Voc (V)

FF

Efficiency Ρ

1

TTIP-cellulose 4 layers,
without CBL
TTIP-cellulose 4 layers, with
500 cycles of ALD CBL
TTIP-cellulose 6 layers,
without CBL
TTIP-cellulose 6 layers, with
500 cycles of ALD CBL
Solaronix 3% T_2 layers,
without CBL
Solaronix 3% T_2 layers,
with 500 cycles of ALD CBL

-2.06

0.77

0.54

0.87 %

-2.83

0.8

0.49

1.12 %

-2.60

0.78

0.48

0.98 %

-2.63

0.8

0.47

1.01 %

-5.41

0.77

0.63

2.69 %

-6.89

0.8

0.52

2.91 %

2
3
4
5
6

Photoanodes made with hydrolytic-induced TiO2 films: Two types of cells made
with anodes deposited by TiO2 from the aqueous system were briefly investigated as well,
respectively, the (BA: AHFT = 0.15M: 0.05M) with binder 2-hydroxyl ethyl cellulose (HEC)
246

precursor, and the (TiF4-cellulose) precursor. The performance of these two cells couldn’t
exceed the TTIP-cellulose system, probably due to the larger particle and pore sizes
limiting the effective dye adsorption and photon-to-electron conversion.

Figure 6-7 J-V curves of boric-cellu and TiF4-cellu-induced DSSCs under luminance
Table 6-4 DSSCs trials on alternative TiO2 films prepared from aqueous systems
Device

Photoanode material

Jsc (mA/cm2)

Voc (V)

FF

Efficiency Ρ

1

(BA: AHFT = 0.15M: 0.05M) with
binder 2-hydroxyl ethyl cellulose (HEC)
TiF4-cellulose

-0.41

0.74

0.66

0.20 %

-0.82

0.67

0.60

0.33 %

2

Cell lifetime: Another study on how storage time influenced the cell performance was
conducted on a series of cells. Storage conditions are light-free, water-free desiccator, and
tests were made at 1 h, 22 h, 52 h and 166 h after electrolyte injection. Figure 6-8 indicates
that as time went by, the performance was significantly influenced, since an 80% decease
was observed for efficiency after 166 h. The degradation is possibly caused by the
electrolyte evaporation during the storage time and sealing defects.

247

Figure 6-8 Cell performance over the time span. This is a TTIP-cellulose based 6-layer photoanode
cell.

6.4 Conclusions and Future Work
6.4.1 Conclusions
To summarize, a facile approach to synthesize nanoporous titania films via microfibrillated cellulose templating and an in-situ coating has been developed to be applied in
dye-sensitized solar cells. As for TTIP-cellulose precursor, the thickness of the anode film is
a dominant factor for the overall performance and efficiency. The 6-layer cell showed a
~40% increase in solar-to-electricity efficiency (1.75%) compared to commercial paste at
the same thickness level under a simulated solar light irradiation of 100 mW cm−2 (AM1.5).
6.4.2 Future Work
Work that has been completed for DSSCs is still at preliminary stage, and the solar cells
could be enhanced in many more aspects, such as adding barrier layer on FTO by titanium
tetrachloride (TiCl4, Aldrich) coating at 70 °C, and increasing the thickness of the titania
anode coating.

248

In order to improve the capacity of cellulose-templating photovoltaic films, one could
manipulate the template in various ways. For example, 3D free-standing cellulose template
will work as a dry scaffold for further coating, especially with methods that have been
developed in previous chapters. Dehydration methods such as freeze drying, supercritical
drying could be combined and new design on the template structures will be worth
researching too.
In terms of flexible and transparent electrode substrates, liquid phase deposition
method will offer low-temperature deposition technique without destroying the flexible
electrode.

249

BIBLIOGRAPHY

250

BIBLIOGRAPHY
(1)

Ghadiri, E.; Taghavinia, N.; Zakeeruddin, S. M.; Grätzel, M.; Moser, J.-E. Enhanced
Electron Collection Efficiency in Dye-Sensitized Solar Cells Based on Nanostructured
TiO 2Hollow Fibers. Nano Lett. 2010, 10, 1632–1638.

(2)

Lee, Y. M.; Kim, Y. H.; Lee, J. H.; Park, J. H.; Park, N.-G.; Choe, W.-S.; Ko, M. J.; Yoo, P. J.
Highly Interconnected Porous Electrodes for Dye-Sensitized Solar Cells Using Viruses
as a Sacrificial Template. Adv. Funct. Mater. 2011, 21, 1160–1167.

(3)

Askeland, P. A. Preparation and Characterization of Sol-Gel Derived Metal Oxide Thin
Films and Powders for Coatings and Catalysts, 1997.

251