Synthesis of Nanomaterials For Biomedical Applications

Synthesis of Nanomaterials for Biomedical
Applications
Selina Vi Yu Tang, BSc.
Thesis submitted to the University of Nottingham

for the degree of Doctor of Philosophy
January 2014
Abstract
This Thesis describes an investigation into the use of a continuous

hydrothermal reactor for the synthesis of nanomaterials with potential use in
three different biomedical applications – bone scaffolds, fluorescent biomarkers
and MRI contrast agents.
In Chapter 1, an introduction to nanotechnology is given, followed by a brief

account of the most prominent routes of nanomaterial synthesis. This includes a
literature review of the key publications to date, which use different reactor
designs for the continuous hydrothermal synthesis of nanoparticles. This Chapter
also describes some problems associated with nanomaterial synthesis, along
with the aims and motivations of this PhD.
Chapter 2 details the principles behind the main characterisation techniques

which are used in the work throughout this Thesis. Some advantages and
limitations of each technique are given.
Chapter 3 reports the synthesis of three different morphologies of

hydroxyapatite. Attempts to introduce a drug and cations into the structure of
hydroxyapatite are also detailed. Finally, cell proliferation and bone nodules
assays showed that hydroxyapatite synthesised via this route exhibits no
cytotoxic effect on mouse embryonic stem cells.
The synthesis of ZnS is the focus of Chapter 4, with the production of uncoated
particles along with the implementation of three different stabilisers. The
addition of 3-Mercaptopropionic acid was shown to decrease cluster size, shift
the isoelectric point of particles, and alter their photoluminescent properties.
In Chapter 5, the synthesis of Fe2O3, Fe3O4, Gd(OH)3 and Co3O4 are described.
The structural properties of these materials are analysed, as well as their
capacity as contrast agents for magnetic resonance imaging. While relaxivity
data for Co3O4 proved poor, Fe3O4 based materials showed relaxivity values
comparable to some commercial contrast agents.
To conclude, a summary of the work detailed in this Thesis is delivered in

Chapter 6. The most pertinent findings from this work are put forward, followed
by a discussion of future work which could lead on from this Thesis.
i
Acknowledgments
There have been many people who have helped me through the course of
my PhD research, and I am grateful to every one of them.
First and foremost, I’d like to thank my supervisors Prof. Ed Lester and Dr Andrei
Khlobystov for their support and advice over the last 4 years.
I’d like to thank Dave Clift for his help with SEM; Dr Stephen Briddon for his help
and advice with fluorometry analysis; Dr Michael Fay for obtaining the TEM
images used in this Thesis; Dr Nigel Neate for his help in using the D8 XRD
diffractometer, Vanessa Loczenski Rose for her help with the cell assays
described in Chapter 3; and Don Yee for his help obtaining the SEM data in
Chapter 4. I’d also like to thank Dr Peter Dunne for his help and advice with XRD
analysis.
Huge thanks to my colleagues who have helped or advised me during my PhD

studies: Dr Graham Rance, Dr Sherif Elbasuney, Dr Tom Huddle, Miquel Gimeno-
Fabra, and Haiping Shen.
The funding for this work came from the EPSRC, with additional funding from
Promethean Particles through a CASE award – thanks go to them.
I am grateful to Drs Pete Gooden, Sandy Reid and Helen Hobbs of Promethean
Particles for giving me the opportunity to work on an Industrial Placement with
them during my PhD – the knowledge I gained was invaluable. Additional thanks
to Pete and Sandy, along with Dr Sam Tang and Ste Ambrose for taking the time
to proofread my Chapters.
I’d like to extend an additional thank you to Ste Ambrose for his patience, for
helping me navigate computer software (Figure 3.8 appears courtesy of his
awesome rendering skills), and for his culinary skills which ensured I never ran
on empty while working (MFEO).
Finally, I’d like to thank Sam, Tim, Rotion, Violet-Rosabelle, Xavier and my
parents, Helen and David, for their unconditional love and support.
ii
Affirmation
The work reported in this Thesis has not been published elsewhere, with the
exception of the following publications:
Journal Publications
Q. Wang, S. V. Y. Tang, E. Lester & D. O'Hare, Synthesis of ultrafine layered

double hydroxide (LDHs) nanoplates using a continuous-flow hydrothermal
reactor. Nanoscale, 2013, 5, 114-117, DOI: 10.1039/c2nr32568c
E. Lester, S. V. Y. Tang, A. Khlobystov, V. L. Rose, L. Buttery & C. J. Roberts,

Producing nanotubes of biocompatible hydroxyapatite by continuous
hydrothermal synthesis. CrystEngComm, 2013, 15, 3256-3260, DOI:
10.1039/c3ce26798a
F. Moro, S. V. Y. Tang, F. Tuna & E. Lester, Magnetic properties of cobalt oxide

nanoparticles synthesised by a continuous hydrothermal method. Journal of
Magnetism and Magnetic Materials, 2013, 348, 1-7. DOI:
http://dx.doi.org/10.1016/j.jmmm.2013.07.064
Conference Proceedings
Selina Tang, Edward Lester, Andrei Khlobystov, Producing Multifunctional

Nanocomposites from Different Morphologies of Hydroxyapatite, 2nd International
Solvothermal and Hydrothermal Association (ISHA) Conference, Beijing, China,
July 2010
Selina Tang, Edward Lester, Andrei Khlobystov, Producing Hydroxyapatite

Nanocomposites for Multifunctional Bone Scaffolds, Technology World
NanoForum, London (Excel), UK, December 2010
Selina Tang, Andrei Khlobystov, Stephen Briddon, Edward Lester, Continuous

Hydrothermal Synthesis of Stabilized Zinc Sulphide Nanoparticles with
th
Fluorescent Properties, 10 International Symposium of Supercritical Fluids
(ISSF), San Francisco CA, USA, May 2012
iii
Selina Tang, Alexandra Martín-Cortes, Andrei Khlobystov, David Grant, Edward
Lester, The Impact of Blending Novel Forms of Hydroxyapatite from Continuous
Hydrothermal Synthesis into Bone Scaffolds, 10th International Symposium of
Supercritical Fluids (ISSF), San Francisco CA, USA, May 2012
Selina Tang, Andrei Khlobystov, Stephen Briddon, Edward Lester, Continuous

Hydrothermal Synthesis of Stabilized Zinc Sulphide Nanoparticles with
nd
Fluorescent Properties, 2 International Solvothermal and Hydrothermal
Association (ISHA) Conference, Austin TX, USA, January 2013
Edward Lester, Selina Tang, Andrei Khlobystov, Steve Briddon, Christopher

Starkey, Peter Dunne, The Continuous Production of Nanoparticles for Biomedical
Applications, 7th International Conference on Materials for Advanced
Technologies (ICMAT), Singapore, July 2013
iv
Table of Contents
1 Background ......................................................................................... 1
1.1 Introduction to Nanotechnology ...................................................... 1
1.2 Advantages of Nanoscale ................................................................ 2
1.2.1 Melting Point ........................................................................... 3
1.2.2 Magnetic Properties ................................................................. 4
1.2.3 Catalytic Properties.................................................................. 5
1.2.4 Advantages in Biomedicine ....................................................... 5
1.3 Safety of Nanomaterials ................................................................. 6
1.4 Synthesis of Nanomaterials............................................................. 7
1.4.1 Top-Down Approaches ............................................................. 7
1.4.2 Bottom-Up Approaches ............................................................ 8
1.5 Hydrothermal Synthesis of Nanomaterials .......................................12
1.5.1 Supercritical Water (ScW) and Sub-critical Water .......................13
1.6 Solvothermal Synthesis of Nanomaterials ........................................17
1.7 Batch and Continuous-flow systems ................................................18
1.7.1 Continuous-flow Hydrothermal Reactors ....................................19
1.7.2 The University of Nottingham Counter-Current Reactor ...............41
1.8 Commercialisation of Hydrothermal Synthesis ..................................51
1.8.1 Hanwha Chemical ...................................................................51
1.8.2 Promethean Particles Ltd. ........................................................51
1.9 Challenges of Nanomaterial Synthesis .............................................54
1.9.1 Particle Stability .....................................................................54
1.9.2 Methods of avoiding Particle Coagulation ...................................55
1.9.3 Methods of Determining Particle Stability...................................56
1.10 Aims of this PhD ...........................................................................58
2 Structural Characterisation Techniques ..................................................60
2.1 Scanning Electron Microscopy (SEM) ...............................................60
2.2 Transmission Electron Microscopy (TEM) .........................................63
2.3 Energy Dispersive X-Ray Spectroscopy (EDX/EDS) ...........................65
2.4 X-Ray Diffraction (XRD) ................................................................65
2.5 Time-of-Flight Secondary Ion Mass Spectrometry (ToF-SIMS) ............69
2.6 Dynamic Light Scattering (DLS) .....................................................70
2.7 Zeta (ζ ) Potential .........................................................................73
v
3 Hydroxyapatite ...................................................................................76
3.1 Introduction .................................................................................76
3.1.1 Hydroxyapatite (HA), Ca10(PO4)6(OH)2 ......................................76
3.1.2 Current applications using HA ..................................................77
3.1.3 Potential Applications for HA ....................................................79
3.1.4 Current Methods of Producing HA .............................................80
3.1.5 Chapter Aims .........................................................................84
3.2 Methodology ................................................................................86
3.2.1 Synthesis of HA ......................................................................86
3.2.2 Characterisation Methods ........................................................98
3.2.3 Material Application Methods ....................................................98
3.3 Results and Discussion ................................................................ 100
3.3.1 HA Platelets (HA-1) and Rods (HA-2) ...................................... 100
3.3.2 The Effect of Precursor pH ..................................................... 102
3.3.3 The Effect of Precursor Concentration ..................................... 110
3.3.4 Addition of Simvastatin ......................................................... 113
3.3.5 Addition of Zinc .................................................................... 115
3.3.6 In Vitro Cell Assays............................................................... 117
3.4 Conclusions ............................................................................... 120
4 Fluorescent Nanomaterials ................................................................. 122
4.1 Introduction ............................................................................... 122
4.1.1 Fundamentals of Fluorescence ............................................... 122
4.1.2 Quantum Dots ..................................................................... 123
4.1.3 Applications using Quantum Dots ........................................... 125
4.1.4 Synthesis of Quantum Dots ................................................... 127
4.1.5 Disadvantages of Cadmium-Based Materials ............................ 129
4.1.6 Zinc Sulphide, ZnS ............................................................... 130
4.1.7 Capping Agents and Stabilisers .............................................. 130
4.1.8 Chapter Aims ....................................................................... 131
4.2 Methodology .............................................................................. 132
4.2.1 Synthesis of CdS using the Nozzle Reactor .............................. 132
4.2.2 Initial Synthesis of ZnS using the Nozzle Reactor ..................... 133
4.2.3 Refined Method for ZnS Synthesis .......................................... 134
4.2.4 Synthesis of ZnS using Thiourea ............................................ 135
4.2.5 Synthesis of ZnS with Surfactants/Stabilisers ........................... 135
4.2.6 Structural Characterisation .................................................... 140
4.2.7 Fluorometry Analysis ............................................................ 141
vi
4.3.1 Cadmium Sulphide, CdS ........................................................ 142
4.3.2 Initial synthesis of ZnS.......................................................... 144
4.3.3 Synthesis with Thiourea as the Sulphur Source ........................ 147
4.3.4 Synthesis of ZnS with Surfactants .......................................... 148
4.4 Conclusions ............................................................................... 167
5 Magnetic Resonance Imaging (MRI) Contrast Agents ............................. 171
5.1 Introduction ............................................................................... 171
5.1.1 Principles of MRI .................................................................. 171
5.1.2 MRI Scans ........................................................................... 174
5.1.3 Relaxation Time, Rate and Relaxivity ...................................... 175
5.1.4 Role of Contrast Agents ......................................................... 176
5.1.5 Commercial Contrast Agents .................................................. 180
5.1.6 Current Methods of Synthesising Contrast Agents ..................... 183
5.1.7 Project Aims ........................................................................ 184
5.2 Methodology .............................................................................. 186
5.2.1 Hydrothermal Synthesis of Nanoparticles................................. 186
5.2.2 MRI Characterisation of Samples ............................................ 196
5.3.1 Hematite, Fe2O3 ................................................................... 199
5.3.2 Magnetite, Fe3O4 .................................................................. 207
5.3.3 Gadolinium Hydroxide, Gd(OH)3 ............................................. 214
5.3.4 Cobalt Oxide, Co3O4 .............................................................. 218
5.4 Conclusions ............................................................................... 224
6 Conclusions and Future Work .............................................................. 227
6.1 Conclusions ............................................................................... 227
6.2 Future Work .............................................................................. 231
7 Bibliography ..................................................................................... 233
8 Appendix.......................................................................................... 248
vii
List of Figures
Figure 1.1. A summary of the different fields in which nanoparticles have be
applied (Tsuzuki, 2009). ......................................................................... 3
Figure 1.2. Illustration showing the increase in specific surface area with smaller
particle size (Nano.gov, 2013). ................................................................ 4
Figure 1.3. A graph illustrating changes to ionic product, density and dielectric
constant of water with increasing temperature, at a constant pressure of 24
MPa (Kritzer et al., 1999). .....................................................................14
Figure 1.4. Schematic diagram of the apparatus typically used by Adschiri et al.
for continuous hydrothermal synthesis of nanoparticles (Adschiri et al., 2000).
..........................................................................................................20
Figure 1.5. Image showing the internal structure and water flow in the T-piece
reactor used by Adschiri et al., collected using neutron radiography (Takami
et al., 2012). ........................................................................................21
Figure 1.6. Schematic of the reactor system used by Arai et al. (Kawasaki et al.,
2010). .................................................................................................24
Figure 1.7. Cross sectional photographs and drawings of the T-piece reactors
used by Arai et al. (Kawasaki et al., 2010). ..............................................25
Figure 1.8. (a) Reactor system design used by AIST, incorporating (b) the
Central Collision Mixer (CCM) (Sue et al., 2011). ......................................26
Figure 1.9. (a) Diagram of the equipment used by Darr et al. (b) A diagram of
the counter-current reactor geometry (Tighe et al., 2012). ........................27
Figure 1.10. Diagram showing the geometry of the Confined Jet Reactor (CJR)
used by Darr et al (Gruar et al., 2013). ...................................................30
Figure 1.11. A schematic diagram of the continuous flow reactor system used by
Kim et al. for the synthesis of metal oxide nanoparticles in supercritical water
or methanol. B: balance; C: condenser; F: 0.2 μm metal filter; L: liquid
container; P: pressure gauge; R: supercritical reactor; T: thermocouple; V1:
viii
needle valve; V2: relief valve; V3: three-way valve; V4: safety valve; FT1: DI
water/methanol feed tank; FT2: precursor solution feed tank; HP1: high
pressure pump for DI water/methanol; HP2: high pressure pump for
precursor solution; PH: preheater; PR: back pressure regulator; RH: reactor
heater; WP: water preheater line (Kim et al., 2008). .................................32
Figure 1.12. CFD simulations of the three tee reactors investigated at the Korea
Institute of Science and Technology for the synthesis of LiFePO 4 nanoparticles
(Hong et al., 2013). ..............................................................................34
Figure 1.13. A schematic diagram of continuous hydrothermal reactor system
built by Aymes et al. in 2001 (Aimable et al., 2009b). ...............................37
Figure 1.14. Diagram showing the X configuration of the mixing point, drilled into
a block of Inconel 625, used in the systems by Aymes et al. ......................38
Figure 1.15. Revised system design by Aymes et al., built in 2006. .................39
Figure 1.16. Model depicting the geometry of the reactor patented by Aymes et
al. (Aymes et al., 2011).........................................................................40
Figure 1.17. Schematic diagram of the continuous hydrothermal reactor system
used by Demoisson et al. (Demoisson et al., 2011). ..................................41
Figure 1.18. Diagram showing the configuration of the counter-current Nozzle
reactor, invented at The University of Nottingham (Lester and Azzopardi,
2005, Lester et al., 2006). .....................................................................43
Figure 1.19. Orientation of parts used to construct the counter current Nozzle
Reactor (Lester et al., 2006). .................................................................44
Figure 1.20. Schematic diagram of the counter current reactor system at the
University of Nottingham. ......................................................................45
Figure 1.21. A photo of the counter-current reactor system. ..........................47
Figure 1.22. Photograph of the Pilot Scale reactor system which incorporates the
counter-current Nozzle reactor technology, courtesy of Promethean Particles
Ltd (see Section 1.8.2). .........................................................................50
ix
Figure 1.23. Ratios of surfactant to nanoparticle where (a) shows insufficient
surfactant concentration resulting in partial coverage, (b) displays optimum
surfactant coverage and maximum stability, and (c) shows too much
surfactant leading to layer collapse and unstable particles (Lubrizol, 2013). .56
Figure 2.1. An illustration showing the components in a typical scanning electron
microscope (Flegler et al., 1995). ...........................................................61
Figure 2.2. (a) A simplified depiction of how a typical TEM operates; (b) a
schematic showing a 3-dimensional tube is viewed as a 2-dimensional image
under a TEM.........................................................................................63
Figure 2.3. A diagram showing the basis of XRD analysis. ..............................66
Figure 2.4. A diagram of the hydrodynamic diameter of a particle which is
reported by DLS analysis (Malvern, 2013)................................................71
Figure 2.5. A schematic representation of what is measured by zeta potential
(Malvern, 2013). ..................................................................................74
Figure 3.1 Images illustrating the continuous hydrothermal reactor system. The
reactor section is highlighted in (b), showing the superheated aqueous
(NH4)2HPO4 (red) solution mixing with the ambient Ca(NO3)2.4H2O (yellow) to
form particles at the interface (c). ..........................................................87
Figure 3.2. Chemical structure of (a) Simvastatin in its inactive lactone form, and
(b) Simvastatin in its active hydroxyl acid form. .......................................94
Figure 3.3. (a) SEM and (b) TEM images of particles in sample HA-1. ............ 100
Figure 3.4. (a) SEM image and (b) TEM image of particles in HA-2. ............... 102
Figure 3.5. SEM image of HA-3. ................................................................ 103
Figure 3.6. HRTEM images of (a) HA-4a; (b) HA-4b; (c) HA-4c; (d) HA-4d; (e)
HA-4e; and (f) HA-4f........................................................................... 105
Figure 3.7. (a) SEM image of ‘open-ended rods’, (b) darkfield imaging in TEM
and (c) TEM image show tube structures while (d) is an SEM image showing a
tube in the middle of the rolling process (highlighted with a red arrow). .... 106
x
Figure 3.8. Images illustrating the scrolling of a sheet, so that the corners meet.
........................................................................................................ 107
Figure 3.9. SEM images of (a) HA-6, and (b) HA-7. ..................................... 109
Figure 3.10. XRD patterns for samples HA-1 to HA-7. The expected peak pattern
for hexagonal hydroxyapatite (ICCD PDF 74-566) is shown at the bottom.. 109
Figure 3.11. SEM images of (a) HA-8, (b) HA-9, (c) HA-10, (d) HA-11. ......... 111
Figure 3.12. SEM image of HA-12. ............................................................ 112
Figure 3.13. XRD patterns for HA-8 to HA-Zn. Patterns for HA-1 and HA-2 are
also shown for comparison. The expected peak pattern for hexagonal
hydroxyapatite (ICCD PDF 74-566) is shown at the bottom. For HA-9, peaks
indicated with a black square match monoclinic brushite (ICCD PDF 9-77)
while other peaks match monetite; for HA-Zn, peaks labelled with black
triangles correspond to monoclinic parascholzite (ICCD PDF 35-495)......... 113
Figure 3.14. (a) SEM Image and (b) TEM image of HA-SIM. ......................... 114
Figure 3.15. ToF-SIMS Data for HA-1, HA-2 and HA-SIM. ............................ 115
Figure 3.16. (a) TEM image of sample HA–Zn, and an elemental distribution map
for (b) Ca, (c) O, (d) P, and (e) Zn. ...................................................... 116
Figure 3.17. XRD pattern of HA-Zn; the predominant phases are hexagonal
hydroxyapatite (ICCD PDF 74-566, shown at the bottom) and monoclinic
parascholzite (ICCD PDF 35-495, peaks shown with blue triangles). Additional
smaller peaks match calcium hydrogen phosphate (ICCD PDF 77-128). ..... 117
Figure 3.18. Results from the cell proliferation assay performed using three
different concentrations of nanoparticles from sample HA-1 and sample HA-
Zn, after 1, 3 and 6 days of incubation. ................................................. 118
Figure 3.19. Images from the bone nodule assay - Cells cultured in osteogenic
media with hydroxyapatite from Sample HA-1 at various concentrations, (a)
100 µg/ml; (b) 10 µg/ml; (c) 1 µg/ml; (d) control .................................. 119
xi
Figure 4.1. Illustration of a typical QD structure; a CdSe core is surrounded by a
ZnS shell and functionalised with siloxane and methacrylate (MMA) (Park et
al., 2010). ......................................................................................... 124
Figure 4.2. A schematic diagram showing the reactor setup for the refined
synthesis method of ZnS. .................................................................... 134
Figure 4.3. A schematic diagram displaying the reactor setup for the synthesis
reaction of ZnS with capping agents. ..................................................... 136
Figure 4.4. Photos of (a) CdS 240, (b) CdS 300, and (c) CdS 400. ................ 142
Figure 4.5. (a) and (b) show TEM images of sample CdS 400. ...................... 144
Figure 4.6. XRD pattern for sample ZnS 400, with expected peak positions of
cubic ZnS (ICCD PDF 80-20) shown in black, and hexagonal ZnS (ICCD PDF
80-7) shown in red. ............................................................................ 145
Figure 4.7. a) and b) TEM images and c) HRTEM images of sample ZnS 400. . 146
Figure 4.8. EDX Spectrum collected for sample ZnS 400 .............................. 147
Figure 4.9. XRD pattern for Sample ZnS-TU-400. Peaks match zinc sulphate
hydrate and zinc hydrogen phosphate. .................................................. 148
Figure 4.10. XRD Patterns of the four ZnS samples with the matched peak
positions of different ZnS phases shown below. Numbered labels indicate the
ICCD PDF file number. ......................................................................... 150
Figure 4.11. SEM images of (a) Uncapped ZnS; (b) ZnS:PVP 5/5; (c) ZnS:NaOH
5/5; and (d) ZnS:3-MPA 5/5. ............................................................... 154
Figure 4.12. Zeta potential profiles for the four samples, as a function of pH. . 157
Figure 4.13. A schematic diagram depicting ZnS particles surface modified with
3-MPA. .............................................................................................. 158
Figure 4.14. (a) The excitation spectra (λem = 540 nm) and (b) emission spectra
(λex = 360 nm) for Uncapped ZnS and ZnS produced with three different
surfactants using 5:5 flow ratios. The spectra of the precursors and water
blank were obtained as a control. ......................................................... 161
xii
Figure 4.15. (a) The excitation spectra (λem = 540 nm) and (b) emission spectra
(λex = 360 nm) for ZnS:3-MPA samples produced using different precursor
flow ratios. The spectra of the precursors and water blank were obtained as a
control. ............................................................................................. 164
Figure 5.1. An image of an MRI scanner depicting the different components
(Action, 2013). ................................................................................... 173
Figure 5.2. A diagram showing the z (longitudinal) and xy (transverse) planes of
magnetisation. The relationship between T2 relaxation time and spin recovery
is also summarised (Oldendorf, 1988). .................................................. 175
Figure 5.3. (a) Native MR Image (no contrast agent present) of liver in a patient
with cholangiocellular carcinoma (cancer), and (b) MR image after T 2
weighted contrast agent has been administered. Differentiation of the lesion is
vastly improved after contrast agent is introduced (AG, 2013). ................ 179
Figure 5.4. A schematic diagram showing the reaction setup for synthesis of
Fe2O3. ............................................................................................... 187
Figure 5.5. A schematic diagram showing the reaction setup for Fe 3O4 synthesis.
........................................................................................................ 190
Figure 5.6. A schematic diagram showing the reaction setup for Co 3O4 synthesis.
........................................................................................................ 193
Figure 5.7. A schematic diagram showing the reaction setup for Gd(OH) 3
synthesis. .......................................................................................... 195
Figure 5.8. Photo of the Halbach 0.5 T Table-Top MRI Scanner at Nottingham
Trent University. ................................................................................. 197
Figure 5.9. Photos of the Bruker 2.35 T MRI scanner at Nottingham Trent
University. ......................................................................................... 198
Figure 5.10. Photograph of Fe2O3 Samples in First Temperature Set. ............. 199
Figure 5.11. Photograph of Fe2O3 Samples in Second Temperature Set. ......... 200
Figure 5.12. Photograph of Fe2O3 Samples in Third Temperature Set. ............ 200
xiii
Figure 5.13. A bar chart summarising the r2 relaxivities of the 12 samples of
synthesised Fe2O3, calculated using the T2 times acquired using the 0.5 T
Halbach scanner, and Equation 2. ......................................................... 202
Figure 5.14. Images collected from the 2.35 T MRI scanner showing T 2 relaxation
times (in ms) and R2 relaxation for a) Fe(NO3)3.9H2O, b) HEM A1, c) HEM A2,
d) HEM A3, e) HEM A4. ........................................................................ 203
Figure 5.15. A chart summarising the particle size data obtained from DLS
analysis, for the 12 synthesised Fe2O3 samples. ...................................... 204
Figure 5.16. XRD patterns obtained for samples HEM A1, HEM B1, and HEM C1.
At the bottom are the expected peak positions for Fe2O3 (ICCD PDF 86-550).
........................................................................................................ 206
Figure 5.17. Photo of Fe3O4 Samples in Second Temperature Set. ................. 207
Figure 5.18. Images collected from the 2.35 T MRI scanner showing T 2 relaxation
times (in ms) and R2 relaxation for (a) water blank, (b) 0.025 M Ammonium
Iron Citrate (precursor), (c) MAG G1, (d) MAG G2, (e) MAG G3, and (f) MAG
G4. ................................................................................................... 209
Figure 5.19. XRD patterns obtained for samples MAG G1 and MAG G3. At the
bottom are the expected peak positions for cubic Fe3O4 (ICCD PDF 88-315).
........................................................................................................ 212
Figure 5.20. (a) and (b) are TEM images of MAG G1, uncoated Fe3O4. ........... 213
Figure 5.21. TEM image of MAG G3, Fe3O4 nanoparticles coated in dextran. ... 214
Figure 5.22. Photo of the Second Set of Gd(OH) 3 Samples. .......................... 215
Figure 5.23. XRD Pattern obtained for GAD E1, with the expected peak positions
of hexagonal Gd(OH)3 (ICCD PDF 83-2037) shown below. ....................... 218
Figure 5.24. Darkfield TEM image of sample Co3O4 showing particles with cubic
morphology........................................................................................ 220
Figure 5.25. Graph showing the particle size distribution of Co3O4 particles. ... 220
xiv
Figure 5.26. The XRD pattern obtained for sample Co3O4. Peaks correspond with
the expected pattern of cubic Co3O4 (ICCD PDF 009-0418), which are shown
below. ............................................................................................... 222
Figure 5.27. High magnification HRTEM image of Co3O4 nanoparticles illustrating
the interplanar spacing. ....................................................................... 223
Figure A1. SEM images of (a) ZnS:NaOH 9/1, (b) ZnS:NaOH 8/2, (c) ZnS:NaOH
7/3, (d) ZnS:NaOH 6/4. ...................................................................... 248
Figure A2. SEM images of (a) ZnS:3MPA 9/1, (b) ZnS:3MPA 8/2, (c) ZnS:3MPA
7/3, (d) ZnS:3MPA 6/4. ....................................................................... 249
Figure A3. SEM images of (a) ZnS:PVP 9/1, (b) ZnS:PVP 8/2, (c) ZnS:PVP 7/3,
(d) ZnS:PVP 6/4. ................................................................................ 250
Figure A4. (a) The excitation spectra (λem = 540 nm) and (b) emission spectra
(λex = 360 nm) for ZnS:PVP samples produced using different precursor flow
ratios. The spectra of the precursors and water blank were obtained as a
control. ............................................................................................. 253
Figure A5. (a) The excitation spectra (λem = 540 nm) and (b) emission spectra
(λex = 360 nm) for ZnS:NaOH samples produced using different precursor
flow ratios. The spectra of the precursors and water blank were obtained as a
control. ............................................................................................. 254
xv
List of Tables
Table 1.1. The critical temperatures and pressures of three solvents. (Fang,
2010). .................................................................................................17
Table 3.1. Reactor system temperatures for synthesis of HA-1 .......................87
Table 3.2. System temperatures during synthesis of HA-4 sample ..................90
Table 3.3. A summary of experimental details for synthesis of various HA
samples. ..............................................................................................96
Table 4.1. Reaction temperatures for each ZnS Sample. .............................. 133
Table 4.2. Reactor System temperatures for samples synthesised using thiourea.
........................................................................................................ 135
Table 4.3. System temperatures for synthesis of ZnS with surfactants. ......... 137
Table 4.4. Flow ratios for the different ZnS:3-MPA Samples. ........................ 138
Table 4.5. Flow ratios for the different ZnS:PVP Samples. ............................ 139
Table 4.6. Flow ratios for the different ZnS:NaOH Samples. ......................... 140
Table 4.7. Summary of crystal phase and crystalline size for each sample, as
derived from XRD analysis. .................................................................. 152
Table 4.8. A table showing the average cluster size for each sample, as
measured from SEM images. ................................................................ 155
Table 4.9. A summary of the average crystallite size, average cluster size, and
wavelength of peak emission for each ZnS sample. ................................. 165
Table 5.1. Summary of Temperatures and Flows for first set of Fe2O3 Samples.
........................................................................................................ 188
Table 5.2. Summary of Temperatures and Flows for second set of Fe 2O3
Samples. ........................................................................................... 188
Table 5.3. Summary of Temperatures and Flows for third set of Fe 2O3 Samples.
........................................................................................................ 189
Table 5.4. Summary of Temperatures and Flows for First Set of Fe 3O4 Samples.
........................................................................................................ 191
xvi
Table 5.5. Summary of Temperatures and Flows for Second Set of Fe 3O4
Samples. ........................................................................................... 191
Table 5.6. A Summary of Temperatures for the Synthesis of Co3O4 Sample. ... 193
Table 5.7. A Summary of Temperatures and Flows for the Synthesis of the First
Set of Gd(OH)3 Samples. ..................................................................... 195
Table 5.8 . A Summary of Temperatures and Flows for the Synthesis of the
Second Set of Gd(OH)3 Samples. .......................................................... 196
Table 5.9. Summary of the measured T2 relaxation times for the Fe2O3 samples,
precursor and water blank. The r2 values, calculated using Equation 2, are
also given. ......................................................................................... 201
Table 5.10. Summary of the T2 relaxation times for the Fe3O4 samples, precursor
and water blank, measured using the Halbach 0.5 T MRI Scanner. The r2
values, calculated using Equation 2, are also given. ................................ 208
Table 5.11. Summary of the measured T2 relaxation times for the Gd(OH)3
samples, precursor and water blank. The r2 values, calculated using Equation
2, are also given. ................................................................................ 216
Table 5.12. Summary of the measured T 2 relaxation times for the sample of
Co3O4, its precursor, and water blank. The r2 values, calculated using
Equation 2, are also given. ................................................................... 219
Table 5.13. The d-spacing for each plane in the Co3O4 crystal, and the respective
peak position, as calculated from the XRD pattern. ................................. 222
Table A1. Zeta Potential data for Uncapped ZnS. ........................................ 251
Table A2. Zeta Potential data for ZnS:3MPA 5/5. ........................................ 251
Table A3. Zeta Potential data for ZnS:PVP 5/5. ........................................... 252
Table A4. Zeta Potential data for ZnS:NaOH 5/5. ........................................ 252
xvii
List of Abbreviations
3-MPA 3-mercaptopropionic acid

AIST (National Institute of) Advanced Industrial Science and
Technology
BPR back pressure regulator
BSE backscattered electrons
CaP calcium phosphate
CCM central collision mixer
CFD computational fluid dynamics
CFHS continuous flow hydrothermal synthesis
CJR Confined Jet Reactor
cm centimetre
DI deionised
DLS dynamic light scattering
DNA deoxyribonucleic acid
EDX energy dispersive X-ray spectroscopy
EU European Union
FP7 Seventh Framework Programme
FTIR fourier-transform infra-red
FWHM the full-width of the peak at half maximum
g gram
h hour
HA hydroxyapatite
HitCH High-throughput Continuous Hydrothermal
HPLC high-performance liquid chromatography
HRTEM high resolution transmission electron microscopy
ICCD PDF International Centre for Diffraction Data powder diffraction
file
IEP isoelectric point
kg kilogram
KIST Korea Institute of Science and Technology
kV kilovolt
LAI light adsorption imaging
LCA life cycle analysis
M molar
mESC mouse embryonic stem cells
min minute
xviii
ml millilitre
mm millimetre
mol moles
MPa mega pascal
MRI magnetic resonance imaging
ms millisecond
MS1 metal salt pump 1
MS2 metal salt pump 2
NADP/NADPH nicotinamide adenine dinucleotide phosphate
nm nanometre
NP nanoparticle
PL photoluminescent
PVP polyvinylpyrrolidone
PZC point of zero charge
QD quantum dots
QELS quasi-elastic light scattering
RAMSI Rapid Automated Materials Synthesis Instrument
RF radio frequency
RFU relative fluorescence units
SChEME School of Chemical, Environmental and Mining Engineering
ScW supercritical water
SE secondary electrons
SEM scanning electron microscopy
SHYMAN Sustainable Hydrothermal Manufacture of Nanomaterials
SMEs small and medium enterprises
SNU Seoul National University
SPION superparamagnetic iron oxide nanoparticle
STEM scanning transmission electron microscopy
TCH thiocarbohydrazide
TEM transmission electron microscopy
TGA thermal gravimetric analysis
ToF-SIMS time-of-flight secondary ion mass spectrometry
TTCP tetracalcium phosphate
UCL University College London
USPION ultrasmall superparamagnetic iron oxide nanoparticle
UV ultraviolet
XRD X-ray diffraction
YAG yttrium aluminium garnet
xix
List of Symbols
% percent
˚ degrees
“ inch
< less than
> more than
µL micro litre
µM micro molar
µm micro metre
µmol micro moles
Å angstroms
C centigrade
D crystallite size
Hc magnetic coercivity
Kw dissociation constant of water
mM milli molar
Pc critical pressure
r2 relaxivity
s second
T tesla
T1 relaxation time for the longitudinal component of MRI
T2 relaxation time for the transverse component of MRI
Tc critical temperature
w
/w% mass fraction
β full-width of the peak at half maximum
ζ zeta
θ incident angle
λ wavelength
xx
Chapter 1
1 Background
This Chapter will introduce the concept of nanotechnology and define
some key terms. The relevance of this research field will be discussed, thereby
providing the fundamental motivations for this PhD project. The history of
nanotechnology will only be described briefly; instead this Chapter will focus
more on the current technologies employed in the synthesis of nanomaterials,
and how these can be, or currently are, applied at a commercial level.
1.1 Introduction to Nanotechnology
It can be argued that whilst nanoparticles (NPs) have existed naturally for
millions of years, the ‘notion’ of nanotechnology was first introduced in 1959 by
Richard Feynman in his talk “There’s Plenty of Room at the Bottom”, where he
described the potential to develop a system of manipulating individual atoms and
molecules. Since then, the field of nanotechnology has acquired great interest,
particularly in the last thirty years or so, as techniques to study and analyse
particles have advanced. The field of nanotechnology encompasses the
synthesis, study and manipulation of particles which are between one and one
hundred nanometres in at least one dimension; where a nanometre is defined as
one billionth (or 10-9) of a metre.
1
Chapter 1. Background
1.2 Advantages of Nanoscale
What makes nanomaterials such an interesting area of research is that it
is not simply the miniaturisation of particles. At the nanoscale, some material
properties are affected by the laws of atomic physics, and so behave in a
different way or exhibit properties unlike those of the traditional bulk material.
This is not the case for a macro to micro size transition and is described as the
quantum size effect phenomenon (Bhushan, 2010). Depending on the material
in question, a range of properties can be enhanced at the nanoscale – a few
examples are highlighted in the subsequent Sections. As a result of these
enhanced properties, nanomaterials have been utilised in a plethora of
applications, as summarised in Figure 1.1.
Furthermore, the enhanced properties are often seen for significantly lower
weight loadings of nanoscale material compared to the bulk scale equivalent. In
an applied sense, this translates to less nanomaterial additive required in a
product, compared to the same micron - or larger - scale compound, to give the
same effect. This means, for most applications, lower production and processing
costs for the manufacturers which can be passed on to consumers; for some
niche fields, such as electronics, devices can be made smaller without
compromising on performance.
2
Figure 1.1. A summary of the different fields in which nanoparticles have
be applied (Tsuzuki, 2009).
1.2.1 Melting Point
At bulk scale, gold has a melting point of 1064 ˚C, but for gold
nanoparticles with a diameter of 5 nm, the melting point decreases to about 600
˚C (Buffat and Borel, 1976). This is due to the increase in surface area to
volume ratio from bulk material to nanoscale particles, as depicted in Figure 1.2.
In this instance, the gold nanoparticles are potentially easier to melt and
manipulate as they require a lower energy input than the bulk material, which in
3
turn, means costs are lowered and any detrimental effect on the environment
(through mass extraction and unsustainable consumption) is reduced.
Figure 1.2. Illustration showing the increase in specific surface area
with smaller particle size (Nano.gov, 2013).
1.2.2 Magnetic Properties
It was first speculated in 1930 that ferromagnetic particles (those which
form permanent magnets or are permanently attracted to magnets) could
display super-paramagnetic behaviour (this is where the magnetisation is at an
average of zero, but particles can become magnetised by an external magnetic
field), if the particles were small enough (Frenkel and Dorfman, 1930). Since
then, this theory was proved and ferrite particles of particle size 10 nm were
shown to exhibit super-paramagnetic behaviour with low magnetic coercivity, Hc,
that is the intensity of field required to demagnetise the particles (Sato et al.,
1987). It is believed that such phenomena seen in magnetic nanoparticles is due
to the large proportion of atoms at the particle surface, a magnetic dead layer,
where atoms cannot maintain ferromagnetism over a distance (Willard et al.,
2004).
4
1.2.3 Catalytic Properties
Due to the increased surface area to volume ratio of nanoparticles,
compared to bulk or even micron-scale particles, nano-catalysts have
demonstrated higher efficiencies or rates of conversion with a smaller mass of
material. For example, TiO2 nanoparticles were shown to be more effective than
larger particles in the photocatalytic degradation of pollutants (Loryuenyong et
al., 2013).
1.2.4 Advantages in Biomedicine
Many of the building blocks of biological systems, namely DNA and
proteins, are sized in the nanometre range; for example, haemoglobin has a 5.5
nm diameter while insulin measures 3 nm in width and a DNA double helix has
intra-strand spacing of 2 nm (Alberts et al., 2002). As pharmaceutical research
begins to focus on developing drugs at a nanoscale level, the ability to directly
target erroneous proteins, usually the cause of many biochemical disorders,
becomes feasible (Liu and Webster, 2007). Furthermore, nanoparticles – due to
their small size – have the potential to penetrate cell membranes and bind to
DNA (Geinguenaud et al., 2012) or penetrate the blood brain barrier (Krol et al.,
2013). While much research exploits this possibility as an opportunity to treat
genetic mutations or abnormalities, this also gives rise to the thought that
nanomaterials could pose a massive health and safety risk, an issue which is
discussed in Section 1.3.
5
1.3 Safety of Nanomaterials
While the advantages of nanomaterials in biomedicine have been reported
in the literature, and highlighted in Section 1.2.4, there has been much
speculation in the scientific community, and in the wider public, over the hazards
involved with nanotechnology. As this field of research receives more attention,
particularly in the public eye, more studies are being conducted globally to
ascertain a definitive answer on whether ‘nano’ should be considered
synonymous with ‘toxic’, and what it means to have nanoparticles present in our
everyday products. One such study was initiated by BMBF, the German Federal
Ministry of Education and Research, and led by BASF - one of the biggest
chemicals companies in the world. The three year project concluded that
chemical composition remains the most important factor contributing to toxicity
and health risks, and the particle size is of secondary importance. The study also
found that functionalization of particles could reduce toxicity as the surfaces
became shielded (BASF, 2013). It is generally agreed that a degree of caution
should be exercised when handling nanomaterials, as it is a relatively new area
of research, and handling dry powders should be avoided to minimise the risks of
particle inhalation. Nevertheless, more emerging nanotoxicology studies are
finding that nanoparticles should not be feared in the way that mass media tend
to portray them.
6
1.4 Synthesis of Nanomaterials
The size, morphology and surface area of the nanomaterial are crucial to
its function and application. Generally speaking, these characteristics are
determined by the method employed to prepare the material. As such, the
synthesis method is a highly significant step in determining the characteristics
and properties of the nanomaterial product. Typically, the methods of
nanoparticle synthesis are categorised as one of two approaches: top-down or
bottom-up. These terms will be elaborated upon in Sections 1.4.1 and 1.4.2.
1.4.1 Top-Down Approaches
The most common example of top-down production of nanoparticles is
mechanochemical milling of micron-scale or bulk scale materials, where larger
particles are broken down to smaller ones. There are several reports in the
literature describing the use of different types of milling to produce magnetic
materials (Manh et al., 2009, Sampathkumaran et al., 2011); materials for
electronic applications (Chen et al., 2010, Hallmann et al., 2011, Zaghib et al.,
2011); and for the synthesis of nanoscale drugs in the pharmaceutical industry
(Laaksonen et al., 2011, Basa et al., 2008). In these papers, the length of time
of milling and the milling speed is described as affecting the structure of the final
product. As well as milling, laser ablation has also been used for NP synthesis
(Amendola et al., 2011).
However, compared to bottom-up approaches, the volume of reports using top-
down methods to synthesise nanoparticles is severely lacking; this is likely to be
due to the inherent drawbacks of top-down methods relative to bottom-up
procedures. Examples of such disadvantages include limited control over particle
7
size and morphology, wide particle size distributions and, in some cases, can be
time intensive where processes can take up to 20 hours (Manh et al., 2009).
These shortcomings mean most researchers tend to prefer bottom-up
approaches instead.
1.4.2 Bottom-Up Approaches
Converse to top-down methods, the bottom-up approach generally builds
larger structures from the smaller building blocks, namely atoms (Rotello, 2004).
The main examples of bottom-up synthesis are chemical, and can be further
classified depending on phase; these will be subsequently elaborated upon.
1.4.2.1 Solid State Reactions
In these reactions, solid reactants tend to be mixed then heated at high
temperatures either in air, or another gas flow of choice (Sohrabnezhad and
Valipour, 2013). Reports have described the use of solid-state reactions to
produce CuO (Sohrabnezhad and Valipour, 2013) and Si3N4 (Dasog and Veinot,
2012) amongst other materials. These types of reactions have the disadvantages
of being inherently energy intensive and producing dry powders which can pose
health risks if inhaled.
1.4.2.2 Vapour-Phase Methods
Vapour-phase approaches, such as spray pyrolysis, involve a salt solution
which is mixed with a carrier gas (e.g. nitrogen) and nebulised or atomised to
produce droplets, into a furnace at high temperatures – typically >800 °C. The
high temperature leads to solvent evaporation and chemical reaction. This
method has reportedly been used to synthesise GaN (Ogi et al., 2009) and TiO2
8
(Figgemeier et al., 2007) nanoparticles, amongst other materials. Although this
method is continuous and produces particles with a narrow size distribution, its
energy intensive nature is a drawback. In addition, like solid state reactions,
products are typically nanopowders, which must be handled with care.
1.4.2.3 Liquid-Phase Methods
This classification of synthesis technique for the manufacture of
nanomaterials is perhaps the most relevant and important because the method
used in this PhD (and which will be described in subsequent Sections) falls under
this category. As a result, liquid phase processes will be described in a little more
detail than for previous techniques.
Co-Precipitation
Co-precipitation techniques are widely utilised due to the facile nature of the
process. Typically, aqueous metal salts are mixed, at ambient or raised
temperatures, with a base which acts as a precipitating agent. Co-precipitation is
a widely used technique for the synthesis of ferrite nanomaterials (Rahman et
al., 2013, Amiri and Shokrollahi, 2013) and in some cases, the process is
performed in an inert atmosphere (Lu et al., 2007). While the method benefits
from its simplicity, it has been seen that particle size, morphology and
composition of the product is heavily reliant on the type of precursor used, pH,
and the reaction temperature. Furthermore, particles produced by co-
precipitation have been hindered by a wide particle size distribution. Usually, a
short nucleation phase followed by a slow growth period leads to a narrow size
distribution – which is not the case with this method (Lu et al., 2007).
9
Microemulsion
Microemulsions are systems which contain two immiscible liquids - usually water
and oil - and a surfactant (Fanun, 2010). The water generally contains dissolved
metal salts, and small droplets of the water in oil are stabilised by the surfactant.
Likewise, droplets of oil can be stabilised in the water, depending on the ratio of
the two. The droplet size can be manipulated by altering different parameters,
and can fall within the nanometer range. These tiny droplets then act as
‘nanoreactors’ inside which particle synthesis can occur. The size of the droplet
subsequently dictates, and limits the size of the final particle (López-Quintela et
al., 2004). The microemulsion method has been utilised to synthesise BaTiO3,
CaCO3, CeO2, ZrO2, and SnO2, amongst other materials (Ganguli et al., 2008).
The reaction mechanisms to control particle size and size distribution are still not
fully understood, limiting the wider use of this method.
Sol-Gel
Sol-Gel methods are a commonly reported route for nanoparticle synthesis,
particularly for metal oxides (Bezzi et al., 2003, Senthilkumaar and Selvi, 2008,
Isasi-Marín et al., 2009). Typically, precursors are converted to a colloidal
solution, or ‘sol’, which gradually develops into a ‘gel’ network - a matrix of solid
nanoparticles dispersed in a liquid phase - over time (or gelling agents may be
used to speed the process up). Through centrifugation, filtration or evaporation,
the nanoparticles can then be isolated to yield the product. As a low cost, facile
method which does not require any specialist equipment, sol-gel remains one of
the most popular methods of nanoparticle synthesis.
However there are some disadvantages to these techniques, such as the need
for templates, which can be costly and need to be removed after synthesis (Chon
Chen et al., 2013, Tang et al., 2013); low purity; or require additional steps to
process the material (such as calcination and milling steps) (Dawson, 1988,
10
Hakuta et al., 1998). Moreover, the high cost of some processes (these can
either be high cost of equipment, precursors, or running costs) coupled with low
yield and long reaction times can mean that, even if the method produces
extremely high quality nanoparticles, there is little to no potential of scaling-up
the technology to a commercially-viable level (Tighe et al., 2013).
Conversely, hydrothermal synthesis has been demonstrated to allow greater
control over the composition and morphology of products. Furthermore,
additional processing steps are minimal, and products are dispersed in water
rather than as a nanopowder, making this a preferential method for
nanomaterial synthesis.
11
1.5 Hydrothermal Synthesis of Nanomaterials
Hydrothermal synthesis has been described as the use of aqueous
solvents in conditions of high temperature and pressure to dissolve and
recrystallise materials which, under standard conditions, are relatively insoluble
(Ballman and Laudise, 1963). In addition, ‘high temperature’ has been defined
as a temperature exceeding the boiling point of the solvent being used as the
reaction medium, and as a result, elevated pressure must be involved
(Demazeau, 2010).
Within the field of hydrothermal synthesis, processes can be further classified
depending on the temperature and pressure of the water reaction media. Where
the water has a temperature and pressure above its critical point (Tc=374 ˚C
and Pc=22.1 MPa), it is termed supercritical. The properties of supercritical water
(ScW) are described in Section 1.5.1.
In the last three decades, there has been an increasing interest in using
hydrothermal processes to synthesise nanomaterials (Ioku et al., 1988, Somiya
and Roy, 2000, Adschiri et al., 1992, Adschiri et al., 2000). Because the main
reaction media is water, rather than noxious solvents, the technology attracts a
lot of interest as a sustainable or “green” technology. Furthermore, the diverse
research devoted to hydrothermal synthesis has shown a wide range of
nanomaterials can be produced via this route – including, but not limited to:
metals (Aksomaityte et al., 2013), metal oxides (Adschiri et al., 2001, Cabanas
et al., 2000, Cabañas et al., 2007, Leybros et al., 2012, Lester et al., 2012),
metal organic frameworks (Gimeno-Fabra et al., 2012), hydroxides (Wang et al.,
2013), carbonates (Nassar, 2013, Ni et al., 2011), phosphates (Chaudhry et al.,
2006, Lester et al., 2013), and sulfides (Liu et al., 2001). This demonstrates the
12
versatility of hydrothermal synthesis, and that it is relevant to a wide range of
applications.
1.5.1 Supercritical Water (ScW) and Sub-critical Water
Under ambient conditions, water is polar - making it a good solvent for
dissolving ionic salts, but not for non-polar compounds. Yet, when water is
heated in a pressurised environment towards its critical point (Tc=374 ˚C and
Pc=22.1 MPa), the hydrogen bonds between water molecules break down. This
leads to a decrease in dielectric constant and an increase in dissociation
constant, Kw, meaning the breakdown of water molecules to H+ and OH- ions
(Lester et al., 2006). At and above the critical point, water becomes a
supercritical fluid and exhibits unique behaviour. It has greatly reduced surface
tension and acts as a solvent but diffuses as a gas (Wandeler and Baiker, 2000).
Figure 1.3 summarises the changes in density, dielectric constant and ionic
product of water with changes to temperature, at a constant pressure of 24 MPa.
13
Figure 1.3. A graph illustrating changes to ionic product, density and
dielectric constant of water with increasing temperature, at a constant
pressure of 24 MPa (Kritzer et al., 1999).
Unlike water at ambient conditions, ScW is weakly-polar; it can dissolve non-
polar molecules while having low solubility for inorganic ionic salts.
Consequently, the latter precipitates out of ScW; this principle forms the basis
for the synthesis of fine inorganic particles in ScW. Due to this principle, and the
interesting properties of ScW, it has been increasingly used in recent years as a
reaction medium for the production of nanoscale metals and metal oxides. The
reaction mechanism for particle formation in ScW is given in Section 1.5.1.1.
Still, it is important to note that not all hydrothermal processes require or use
water above its critical point; many described methods use temperatures as low
as 150 ˚C and pressures below 24 MPa (Wang et al., 2013). Thus, pressurised
water at temperatures between 100˚C and 374˚C is termed sub-critical or near-
critical.
14
It has been shown that for some materials, particle size, particle morphology,
and product yield are affected by the reaction temperature. In the case of ZrO2,
particle size and the precursor conversion rate increase with elevated reaction
temperatures (Hobbs et al., 2009). For hydroxyapatite, Ca10(PO4)6(OH)2,
different particle shapes are produced with varying reaction temperature (Lester
et al., 2013). Therefore, depending on the requirements of the particle
properties, or in the interest of balancing yield with energy consumption, lower
reaction temperatures may be utilised.
1.5.1.1 Reaction Mechanisms and Metal Oxide Formation in ScW
As described by Adschiri et al. (Adschiri et al., 1992) and subsequently
Lester et al. (Lester et al., 2006), the proposed formation of metal oxides begins
with a metal salt in ScW readily undergoing hydrolysis without addition of any
base, due to the high Kw, and high OH- concentration in the solvent. The metal
hydroxide consequently experiences a dehydration step which results in the
formation of metal oxide nanoparticles as described by the following, where ML
is a metal salt:
Step 1, Hydrolysis : MLx + xOH− → M(OH)x +xL−
Step 2, Dehydration : M(OH)x → MOx/2 + x/2 H2O
15
1.5.1.2 Other Applications of Supercritical Water
The properties of ScW described in Section 1.4.1 make it a solvent of
interest for applications other than for the synthesis of nanoparticles.
One of the major applications of ScW is the oxidation and breakdown of
hydrocarbons (for example, to treat waste water). Due to the non-polar nature
of ScW, it becomes completely miscible with organic compounds meaning it can
transport molecules more effectively. Compared to other methods of
hydrocarbon breakdown – such as incineration – ScW oxidation is less energy
intensive and is proving to be a sustainable “green” technology (Al-Duri et al.,
2008).
ScW has also reportedly been used to chemically recycle carbon fibre by
dissolving epoxy resins which were bound to the fibres (Piñero-Hernanz et al.,
2008). This method of chemical recycling showed the recovered fibres to have
up to 98% of the tensile strength, compared to virgin fibres. Again this shows
that ScW is fast becoming a media for many sustainable applications.
16
1.6 Solvothermal Synthesis of Nanomaterials
Where organic solvents, such as alcohols, are used in the place of
aqueous solvents, the synthesis is termed solvothermal instead (Xu and Wang,
2012). In the same way as water, the solvent is termed supercritical when it is
at its critical point. This point varies for different solvents; the critical
temperatures and pressures for methanol, ethanol, and acetone or given in Table
1.1.
Table 1.1. The critical temperatures and pressures of three solvents.
(Fang, 2010).
Critical Temperature Critical Pressure

Solvent
(Tc, ˚C) (Pc, MPa)
Methanol 239 8.1
Ethanol 241 6.1
Acetone 235 4.7
As with hydrothermal methods, the solvent can also be used as a reaction media
below its critical point, while increases in reaction temperature beyond T c may
lead to increased yield. Although not as sustainable or cost-effective as using
water, the use of organic solvents the synthesis of materials which are not
favoured in water – in particular, metal nanoparticles without the need for toxic
reducing agents (Choi et al., 2010, Shin et al., 2010), or provides better
compatibility for phase transfer or formulation into other media (Veriansyah et
al., 2009).
17
1.7 Batch and Continuous-flow systems
Chemical processes can generally be categorised as being conducted
using either batch or continuous flow systems – this is not exclusive to
hydrothermal synthesis of nanomaterials, or indeed to the synthesis of
nanomaterials overall. Like many researchers in their respective chemical fields,
the nanotechnology community is divided over which type of system is superior
for the production of nanomaterials.
For hydrothermal synthesis of nanoparticles, batch systems typically employ an
autoclave-type reactor into which the reactants are placed; this is then heated
by an oven, or a sand or oil bath. Conversely, continuous systems use pumps to
drive reactants through a heated, pressurised pipework reactor system. While
batch systems are arguably simpler and easier to design, operate and maintain,
continuous systems are known for high reaction rates within a very short
residence time (< 1 minute) (Fang, 2010). While short reaction times carry
several benefits, there are some reactions which appear to require long reaction
times and, to date, have only been produced by batch methods (Kasuga et al.,
1998, Sekino et al., 2006, Bavykin et al., 2004). In addition, batch reactors
allow the use of reagents which are water or solvent insoluble, but continuous
reactors require salts to be dissolved to prevent blockages at the pumps or other
points of the system. Nevertheless, batch systems are more difficult to
reproduce at larger scale compared to continuous-flow systems. Thus, if batch
synthesis reactions can be tailored for a continuous-flow system, the products
could be manufactured at a volume more meaningful for commercial use.
Continuous-flow hydrothermal systems which are currently used at a commercial
scale are described in Section 1.8.
18
In general, during the synthesis of nanomaterials, particle size, size distribution,
morphology and crystallinity of the product can be controlled by altering
individual process parameters; these being temperature, pressure, concentration
and pH of precursors, and flow rates (Lester et al., 2012, Adschiri et al., 2000,
Sue et al., 2006). Continuous systems allow more control over these
experimental parameters, and carry more potential to be scaled up (Tighe et al.,
2013, Fang, 2010). With these benefits over batch processes, continuous-flow
systems which have been described in literature for the synthesis of
nanomaterials will be discussed in more detail in Section 1.7.1.
1.7.1 Continuous-flow Hydrothermal Reactors
Many researchers see the benefits of continuous flow reactors, which
were discussed in Section 1.7. It is generally understood that the method of
mixing reactant streams in continuous flow systems plays a vital role in
determining the size and size distribution of the synthesised particles. As such,
several groups have built and experimented with different mixer/reactor designs
for the hydrothermal synthesis of nanomaterials. Each research group has
produced different materials, and for varying applications. A brief history of how
research into continuous hydrothermal synthesis began, and a selection of the
most prominent work will be discussed in the subsequent Sections. The reactor
design used for the work described here will then be discussed in Section 1.7.2.
1.7.1.1 Adschiri et al. (Tohoku University)
The process of hydrothermal fine particle synthesis was first transferred
from batch to continuous-flow reactors by Professor Tadafumi Adschiri and his
research group at Tohoku University in Sendai, Japan (Adschiri et al., 1992).
This pioneering work described the use of ten different water-soluble metal salts
19
for the synthesis of seven different metal oxides, with residence times below 2
minutes. Later publications also focussed on the use of their continuous flow
reactor for the production of CeO2 (Hakuta et al., 1998), ZrO2 and TiO2 (Adschiri
et al., 2000), and AlOOH (Adschiri et al., 2001). The equipment set up used for
synthesis is given in Figure 1.4, where the mixing point was a T-piece fitting. As
well as being a readily available, off-the-shelf part, this configuration allowed the
rapid heating of precursor streams, especially compared to batch processed. The
published works typically used metal salts of concentrations up to 0.1 M, total
flow rates up to 10 ml/min and thus produced dry weight equivalents of metal
oxide nanoparticles up to 10 g/h (Adschiri et al., 2000). Since these publications,
other researchers have discussed the drawbacks of using T-piece mixers – these
will be elaborated upon in Section 1.7.2.
Figure 1.4. Schematic diagram of the apparatus typically used by
Adschiri et al. for continuous hydrothermal synthesis of nanoparticles
(Adschiri et al., 2000).
20
Adschiri et al. recently published some very interesting research, where they
used neutron radiography to visualise the flows of water inside their mixing point
(Takami et al., 2012). The mixing patterns inside continuous flow reactors have
long plagued researchers in the field; due to the high temperatures and
pressures required for synthesis, reactors are typically constructed from stainless
steel tubing and fittings thus the internal flows cannot be seen. As such,
researchers have had to rely on computational fluid dynamics (CFD) and other
forms of modelling to visualise flow arrangements. However, the use of a
neutron beam is a breakthrough in the field and Adschiri et al. have shown that
it is possible to use this form of radiography to observe flow arrangements within
the reactor, as seen by the image shown in Figure 1.5.
Figure 1.5. Image showing the internal structure and water flow in the
T-piece reactor used by Adschiri et al., collected using neutron
radiography (Takami et al., 2012).
21
Still, despite the development of this continuous flow reactor, the research group
of Prof. Adschiri has recently published work using batch type reactors to
synthesise nanoparticles of HfO2 (Sahraneshin et al., 2012a), yttrium aluminium
garnet (YAG) (Sahraneshin et al., 2012b), and zirconium dioxide (ZrO2) (Taguchi
et al., 2012). Unlike many reported batch methods of hydrothermal synthesis,
the described protocols required a relatively short reaction time of 10 minutes.
Conversely, the disadvantage of this method is the small reactor volume; with
only 5 ml of product yielded per reaction, there is little scope for scale up. This
small reactor volume appears to be by design but may be a requirement for
successful synthesis. In addition, reverting to a batch process despite the
availability of a hydrothermal flow system suggests these nanomaterials may not
be producible on their flow reactor.
Other work published by Adschiri et al. in recent years, described the use of both
their 5 ml batch reactor and flow reactor to synthesise cobalt aluminate
nanoparticles (Lu et al., 2013). Their results showed that they were able to
synthesise their intended material, CoAl 2O4, using their flow reactor while their
batch system produced Co1-xAlx layered double hydroxide (Co-Al-LDH) as the
principle phase. This was due to the discrepancies between heating in the two
systems; the reactants in the autoclave reach their reaction temperature after 5
minutes while in the flow reactor, heating takes <1 second. The slower heating
in the batch reactor produced the LDH as an intermediate and lead to a growth
phase for the remainder of the reaction. The authors speculate that a much
longer reaction time (>10 hours) is required to dehydrate the LDH to CoAl 2O4.
Meanwhile the rapid heating in the flow reactor lead to total dehydration of the
precursors, and therefore highlights the further benefits of continuous flow
reactors with rapid heating systems.
22
1.7.1.2 Hakuta et al. (AIST)
Yukiya Hakuta, Hiromichi Hayashi and Kunio Arai are researchers in the
field of hydrothermal synthesis of nanoparticles, using both batch and
continuous-flow reactors to produce materials. They are associated with the
National Institute of Advanced Industrial Science and Technology (AIST) in Japan
and Tohoku University, and have often collaborated with Prof. Adschiri. As such,
they conduct much of their research using flow reactors with a similar T-piece
mixer geometry as Prof. Adschiri. A diagram of the reactor system typically used
for their work is shown in Figure 1.6. After the T-piece, the flow passes into a
heated ‘reactor’ which allows particle development. After the reactor, the product
stream is quenched by mixing with a flow of deionised water at ambient
temperature.
Between the three researchers, they have published works using continuous flow
hydrothermal synthesis to generate AlOOH (Hakuta et al., 2005), CuO, Fe2O3,
NiO, ZrO2 (Sue et al., 2006), BaTiO3 (Matsui et al., 2008), ferrites (Sato et al.,
2008), perovskite oxide (Lu et al., 2008) and Al2O3 (Noguchi et al., 2008).
23
Figure 1.6. Schematic of the reactor system used by Arai et al.
(Kawasaki et al., 2010).
Although this work is similar to that previously described by Adschiri, Arai et al.
have also developed microreactor T-pieces for the synthesis of NiO (Kawasaki et
al., 2010). Cross sectional images of the three microreactors used can be seen in
Figure 1.7. This study found that quicker heating rates resulted from smaller
internal volumes of the T-pieces, which in turn, lead to small particles being
produced. The microreactor T-piece has also been employed for the synthesis of
Ni nanoparticles (Sue et al., 2009).
While the production of smaller particles may be advantageous, the
microreactors used have internal diameters as small as 0.3 mm, which may lead
to problems with blockages if the particles begin to aggregate (see Section 1.9
for more details on particle aggregation). With such small reactor volumes being
vital to determine particle size, there is also a limitation on the quantity of
material that can be manufactured.
24
Figure 1.7. Cross sectional photographs and drawings of the T-piece
reactors used by Arai et al. (Kawasaki et al., 2010).
In 2011, a publication from Hakuta and others (including some from AIST)
recognised some problems associated with T-piece mixers – particularly the
creation of nanoparticles with a wide size distribution. As a result, a new mixing
configuration was described, named the Central Collision Mixer (CCM), and is
depicted in Figure 1.8 (b). The metal salt solution is pumped in a downflow and
is met by four streams of heated water flowing in horizontally. This reactor
configuration was used to synthesis nickel ferrite nanoparticles and the authors
found that the CCM yielded smaller particles which had a narrower size
distribution compared to using a conventional T-piece mixer (Sue et al., 2011).
This mixer design is a recent development, with only a few publications
describing its use, and so the details of its reliability and reproducibility are
currently unclear. Still, the CCM geometry appears to show it as a custom made
25
part which could provide problems if reproducing this reactor, and increase
costs.
Figure 1.8. (a) Reactor system design used by AIST, incorporating (b)
the Central Collision Mixer (CCM) (Sue et al., 2011).
In addition to the reactor designs described previously, Hakuta et al. introduced
a swirling micro mixer reactor in 2007, which created a vortex to mix the metal
salt flow with the superheated water flow (Wakashima et al., 2007). However,
future publications by the authors revert to the use of T-piece mixers and
discussions using this swirling mixer are limited.
26
1.7.1.3 Darr et al. (University College London)
Prof. Jawwad Darr and his research group are based at University College
London (UCL) and have been conducting work in the field of continuous
hydrothermal synthesis since 2007. Until recently, Darr et al. have been using
the counter current reactor design (Lester et al., 2006) which was used for the
course of the work described in this Thesis, and which will also be discussed in
Section 1.7.2. Schematic diagrams of the reactor configuration are given in
Figure 1.9. Darr et al. also draw reference to the original research article in
several of their papers (Boldrin et al., 2007, Tighe et al., 2012).
Figure 1.9. (a) Diagram of the equipment used by Darr et al. (b) A
diagram of the counter-current reactor geometry (Tighe et al., 2012).
Darr et al. have published a series of papers describing the use of this counter
current reactor in the continuous flow hydrothermal synthesis (CFHS) of an array
of nanomaterials. These include hydroxyapatite (Chaudhry et al., 2006), spinel
27
structures (Boldrin et al., 2007), titanium dioxide (Zhang et al., 2009b), sodium
titanate (Zhang et al., 2010), indium oxide (Elouali et al., 2012) and
nanophosphors (Gruar et al., 2012).
Additionally, Darr et al. have also developed a high throughput system for
producing Zn-Ce oxides based on the counter current reactor, named the High-
throughput Continuous Hydrothermal (HitCH) synthesis. Using the HitCH system
at lab scale (total flow rates up to 40 ml/min), the research group showed that
they were able to produce a large array of samples in a short time frame, which
they were able to rapidly obtain X-ray diffraction patterns for, using the
beamline at Diamond Light Source (Weng et al., 2009, Parker et al., 2011).
Furthermore, the group have converted the process to be automated; using a
robot, the system produced, cleaned and printed lab scale samples of
nanoparticles. Named RAMSI (Rapid Automated Materials Synthesis Instrument),
the equipment was used to produce europium-doped yttrium oxide, Y2O3:Eu3+
phosphors. In some cases, the authors found that the produced nanoparticles
had to be heat treated in air at up to 1200 °C for them to display fluorescent
properties (Lin et al., 2010). As well as being time consuming, and energy
intensive, heat treatment at such high temperatures is known to cause
densification, grain growth (Ostwald ripening) and in turn, eliminate the
nanoscale aspect to the particles (German, 2010). Currently, only this material
has been reportedly synthesised using RAMSI.
In 2012, Darr et al. published a paper detailing observations (from in situ
temperature measurements) of jetting effects in the counter current reactor; this
is where the superheated flow of water infiltrates the flow of metal salt and may
lead to non-homogenous mixing and in turn, a wide particle size distribution.
This jetting effect was only noted for certain flow ratios and may be affected by
the types of pumps used in the system. More recently Darr et al. developed a
28
new reactor configuration, which they named the Confined Jet Reactor (CJR) –
shown in Figure 1.10 (Tighe et al., 2012).
Contrary to the previous counter current reactor used by the group, the CJR uses
a co-current mixing regime. Here, the heated water stream is pumped through
an inner tube, as an upflow, into the bottom of the reactor. The precursors (at
ambient temperature) are pumped into the sides of the reactor cross piece,
entering below the outlet of the water stream. The formed nanoparticles then
flow upwards to the heat exchanger. Whilst Darr et al. explained that particle
accumulation and blockages in pipes “often occurs due to substantial preheating
of the incoming metal salt precursor stream prior to being brought into contact
with the much hotter water stream” (Gruar et al., 2013) the CJR does exactly
this, i.e. the cool precursor feeds come in to contact with the hot inner pipe
carrying the water stream, before they actually meet and mix with the heated
water. This suggests that precipitation and particle accumulation, to some
degree, could occur on the surface of the heated inner tube.
29
Figure 1.10. Diagram showing the geometry of the Confined Jet Reactor
(CJR) used by Darr et al (Gruar et al., 2013).
The CJR has apparently been scaled up from lab (total flow rates up to 40
ml/min to produce 1-10 g/h) to pilot scale (total flow rates up to 800 ml/min >1
kg/h) by Darr et al. to produce cerium and zinc oxides (Gruar et al., 2013, Tighe
et al., 2013). The researchers showed high reproducibility of particles at the two
different scales.
1.7.1.4 Kim and Lee et al. (KIST and SNU)
Between the Korea Institute of Science and Technology (KIST) and Seoul
National University (SNU), Jaehoon Kim and Youn-Woo Lee have published a
selection of papers describing continuous hydrothermal and solvothermal
methods to produce a wide range of nanomaterials.
30
In the published collaborative works of Kim and Lee, they recognise that
clogging or blocking in reactors is a common problem associated with continuous
hydrothermal synthesis of nanoparticles. However, rather than addressing the
issue by changing their reactor or mixer geometry, they opted for a change in
reaction media from water to methanol instead (Kim et al., 2008, Veriansyah et
al., 2010b). They found that using methanol led to a greater availability of
soluble surface modifiers which could be used to bind to particles, while
methanol itself acted as a surface modifier as well. However, the biggest
drawbacks of using methanol instead of water are the implication of increased
running costs and health and safety risks.
The continuous flow reactor used in these works is shown in Figure 1.11. Like
the reactor setups used by the Japanese research groups described in Sections
1.7.1.1 and 1.7.1.2, their reactor system consists of a T-piece mixing point
where the heated water or methanol stream flows from the top of the T-piece as
a downflow, and meets the cooler metal salt stream which flows in from the side
arm of the T-piece. The mixed feeds then flow down into a heated cylinder with a
volume of 14.58 cm3, which is the reactor. This vessel extends the residence
time of the reaction, which can aid particle growth or increase reaction
conversion.
This reactor has been used by KIST and SNU to synthesise CeO2 (Kim et al.,
2008), ZnO (Veriansyah et al., 2010b) and Fe3O4 (Veriansyah et al., 2010a) in
supercritical methanol. For the study of ZnO synthesis, the products from the
solvothermal process were compared to those from hydrothermal; the
researchers found that using methanol as a solvent affected the surface groups
of particles, as well as size, morphology and dispersability into ethylene glycol.
31
Figure 1.11. A schematic diagram of the continuous flow reactor system
used by Kim et al. for the synthesis of metal oxide nanoparticles in
supercritical water or methanol. B: balance; C: condenser; F: 0.2 μm
metal filter; L: liquid container; P: pressure gauge; R: supercritical
reactor; T: thermocouple; V1: needle valve; V2: relief valve; V3: three-
way valve; V4: safety valve; FT1: DI water/methanol feed tank; FT2:
precursor solution feed tank; HP1: high pressure pump for DI
water/methanol; HP2: high pressure pump for precursor solution; PH:
preheater; PR: back pressure regulator; RH: reactor heater; WP: water
preheater line (Kim et al., 2008).
32
Kim et al. have recently published work which investigated the mixing patterns
inside T-piece reactors with three different geometries, and how these affected
the properties of synthesised LiFePO4 particles for use in battery materials (Hong
et al., 2013); the three different T-piece geometries are shown from the CFD
images in Figure 1.12. While the 90° Tee is an off-the-shelf part from Swagelok,
The 50° and Swirling Tees are both custom made by the research group. The
study also examined the effect of different temperatures, flow rates and
concentrations on the structure and properties of the product.
33
Figure 1.12. CFD simulations of the three tee reactors investigated at
the Korea Institute of Science and Technology for the synthesis of
LiFePO4 nanoparticles (Hong et al., 2013).
Data from X-ray diffraction (XRD) obtained during the study found that the
Swirling Tee mixer produced LiFePO4 particles with a higher degree of
crystallinity than the other two mixers, as well decreasing the ratio of Fe 3O4
34
impurity detected. The authors put this down to a more efficient mixing regime
in the Swirling Tee, which in turn provides more rapid nucleation of LiFePO 4
particles while inhibiting the formation of impurities. This effect was accentuated
with higher flow rates, hence likely to cause more turbulent mixing.
Images collected using Scanning Electron Microscopy (SEM) found that, at
various flow rates tested, the Swirling Tee produced much smaller particles
(100-400 nm diameter spheres) compared to the 90° Tee (platelets with lengths
of 200-600 nm) and the 50° Tee (platelets with lengths between 300-700 nm).
Moreover, using higher flow rates for each mixer type saw a decrease in particle
size – again, likely due to more rapid fluid mixing and faster heat transfer
between fluids.
As a result of the smaller particle size, higher crystallinity, and lower degree of
impurities, the LiFePO4 particles synthesised using the Swirling Tee showed
much more promise as a battery material than those made using the other Tee
geometries. As such, this reactor design could be promising for production of
other nanomaterials. On the contrary, a disadvantage of the Swirling Tee is the
custom-made design, which will inherently be more expensive than off-the-shelf
parts. Furthermore, the products described in this paper exhibit wide particle
size distributions, which may be due to a large range of residence times in the
reactor, rather than rapid heating of the precursor stream followed by quick
cooling. One possible reason for this could be that the product stream is hotter,
and therefore less dense, than the cold precursor stream entering the top of the
Swirling Tee. Therefore, the less dense product stream will naturally rise to the
top of the mixer, rather than being drawn to the reactor outlet at the bottom. In
between pulses of incoming precursor (pulsing caused by the pumps), a degree
of flow (and particle) recirculation will occur - leading to more growth of these
particles compared to those which are quickly pushed out of the reactor due to
35
the positive pulse from the precursor being pumped in. Of course, this theory
depends on the ratio of buoyancy and inertia forces acting on the product
stream, but provides speculative reasoning for particle size distribution.
1.7.1.5 Aymes et al. (University of Burgundy)
Based in Dijon, France, research into continuous hydrothermal synthesis
of nanoparticles by the research group of Daniel Aymes began when a reactor
was built in 2001, as shown in Figure 1.13 (Aimable et al., 2009b). This reactor
setup used an “X geometry” to mix the superheated water flow with two
precursor solutions. To create this geometry, the flows were created by drilling
into a solid block of Inconel 625, shown in Figure 1.14. After the mixing point,
the fluids then flowed into a 2 metre long Inconel reactor tube in a “zigzag”
configuration. The paper states that this geometry maintains a constant
temperature throughout the reactor length, while minimising particle
agglomeration. Using this reactor design, Aymes et al. were able to synthesise
BaZrO3 (Aimable et al., 2008); LiFePO4 (Aimable et al., 2009a); ZrO2; anatase
TiO2 and a range of spinel ferrites (Aimable et al., 2009b).
36
Figure 1.13. A schematic diagram of continuous hydrothermal reactor
system built by Aymes et al. in 2001 (Aimable et al., 2009b).
Inconel 625 is a strong alloy, allowing high temperatures and pressures to be
used and, unlike stainless steel 316, is not corroded through exposure to halides.
This advantage of Inconel also means it is significantly more expensive than
stainless steel, and drives setup costs up; this means the overall cost of
nanoparticle production is greater than using stainless steel equipment. While
using Inconel parts provides more corrosion resistance if using halide precursors
(which are generally cheap, and so the cost of Inconel can be justified), the
published works of Aymes et al. describe the use of non-halide based precursors,
therefore negating the need for this expensive alloy. Furthermore, the Inconel
block used by Aymes et al. is a specialised design, as opposed to using off-the-
shelf parts, which can require additional pressure safety steps (prior to
commissioning) and drive costs up further.
37
Figure 1.14. Diagram showing the X configuration of the mixing point,
drilled into a block of Inconel 625, used in the systems by Aymes et al.
After conducting heat transfer studies, the research group found that they could
reach the necessary temperatures for supercritical water by using shorter
lengths of pre-heater and reactor tubing, thereby reducing residence times as
well. A diagram of their revised reactor system is given in Figure 1.15. However,
the Inconel block used as the mixing device (Figure 1.14) remained the same.
38
Figure 1.15. Revised system design by Aymes et al., built in 2006.
Despite the published papers describing the reactor designs shown in Figure
1.13, Figure 1.14 and Figure 1.15, a paper was later published by Demoisson et
al. (contemporaries of Aymes at the same institution) describing a very different
reactor design, which they patented with Aymes (Aymes et al., 2011, Demoisson
et al., 2011); this suggests that the previous X reactor geometry had operation
issues which were not discussed in the patent. Interestingly, the authors draw
reference to the benefits of the Nottingham counter-current reactor design which
will be discussed in Section 1.7.2.
Their new reactor design, shown in Figure 1.16, was patented in 2011 and like
their earlier designs, uses a block of Inconel 625 into which holes are drilled and
bored. Therefore the reactor has the inherent advantages and drawbacks
described previously for their earlier designs. However, the main difference from
their previous “X” reactor geometry is the adoption of a counter-current system.
39
Figure 1.16. Model depicting the geometry of the reactor patented by
Aymes et al. (Aymes et al., 2011).
A schematic diagram of the whole continuous reactor system used by Aymes et
al. is given in Figure 1.17. The principle differences between this and earlier
designs are the reactor, and the elimination of a separate mixing point and
reactor; instead, these two entities are combined into one chamber.
40
Figure 1.17. Schematic diagram of the continuous hydrothermal reactor
system used by Demoisson et al. (Demoisson et al., 2011).
1.7.2 The University of Nottingham Counter-Current
Reactor
Since the transition from batch to continuous-flow hydrothermal synthesis
by Prof. Adschiri et al. in the 1990s, the process has been hindered by blockages
in the reactor, as highlighted in previous Sections. In the late 1990s and early
2000s, research into continuous hydrothermal synthesis of nanoparticles was
started at The University of Nottingham by the Clean Technology Group. They
employed a T-piece reactor for the synthesis of cerium-doped, zirconium oxides
(Cabanas et al., 2000) and ferrites (Cabañas and Poliakoff, 2001). However, the
use of this reactor type led to problems with inconsistent products, and
blockages which meant experiments had to be aborted before completion. In
41
response to this, work was conducted by SChEME (School of Chemical,
Environmental and Mining Engineering) at The University of Nottingham to
investigate the fluid dynamics and mixing mechanism within the T-piece reactor
(Blood et al., 2004). Where previous studies used CFD to model mixing regimes,
Blood et al developed light adsorption imaging (LAI) - a method which used
fluids of similar density and viscosity to supercritical water and the metal salt
precursor feed, to visualise their interaction when mixed. By adding a dye to the
stream representing the metal salt feed, the researchers were able to use image
analysis techniques to quantitatively conclude the positions of ‘stagnant zones’,
where localised mixing or recirculation occurred, or ‘flow partitioning’, where the
two streams would flow separately with improper mixing. The study concluded
that, regardless of the orientation of the flows in the T-piece, the reactor
geometry would lead to poor mixing conditions and, in turn, particle build-up and
blockages. This was consistent with experimental observations.
Continuing on from this work, the researchers within SChEME used LAI and CFD
analysis, alongside experimental data to model alternative reactor designs
(Lester et al., 2006). This paper presented a novel reactor design which
exploited the natural convection forces in the system to drive turbulent mixing,
and prevented the previously seen problems of flow partitioning, or development
of stagnant zones. The new reactor design, termed the counter-current Nozzle
Reactor, was then patented (Lester and Azzopardi, 2005) and is shown in Figure
1.18. This reactor was used for the work described in this Thesis, the details of
which will be described in Sections 1.7.2.1.
42
Figure 1.18. Diagram showing the configuration of the counter-current
Nozzle reactor, invented at The University of Nottingham (Lester and
Azzopardi, 2005, Lester et al., 2006).
1.7.2.1 Reactor Geometry
During the research conducted for this work, the equipment setup
displayed in Figure 1.20 was used. A photo of the system is also provided in
Figure 1.21. There are some slight differences from the setup previously
described in literature (Lester et al., 2006). Firstly, the in-line filter before the
back pressure regulator (BPR) was removed. A filter was originally incorporated
to remove large aggregates from the sample, and to preserve the BPR. However,
it was found that, compared to previous reactor geometries, fewer large
aggregates were produced. Moreover, without the filter, the collected particles
are more representative of the reaction product. While the filter prevented
blockages at the BPR (which would cause loss of pressure), the filter itself served
as a potential point of blockage if large aggregates were present. As such, it was
better to conduct experiments without it.
43
In addition, two extra pumps have been introduced for a second metal salt feed,
and a “capping point” feed to introduce surfactants (typically organic
compounds) after the reactor, but before the BPR. Depending on the reaction,
these pumps are not necessarily required so are not used, without affecting the
rest of the experimental setup.
In the initial publication describing the counter current Nozzle reactor, some
experiments described the addition of a cone (a Swagelok ferrule) to the end of
the inner pipe, while others omitted it. The function of the cone was to mimic a
spinning disc reactor and dissipate flow of the heated water, therefore
maximising mixing. However, for the reactor used here, the cone was not used
in order to maximise the flow area available.
Figure 1.19. Orientation of parts used to construct the counter current
Nozzle Reactor (Lester et al., 2006).
44
A diagram of the reactor geometry is given in Figure 1.19. It is a pipe-in-pipe
configuration where the heated water feed enters through the thinner inner tube
(1/8” outer diameter), and the cool metal salt feed is flowed up into the outer
pipe (3/8” outer diameter). Mixing and particle formation occurs at the interface,
then the product is carried up and out of the side arm of the cross piece to the
cooling loop.
Figure 1.20. Schematic diagram of the counter current reactor system at
the University of Nottingham.
For nanoparticle synthesis experiments, a vessel of water (or if required, a
thermally-stable precursor) was pumped at a pre-set flow rate up to 25 ml/min,
using a Gilson HPLC pump (Model 305 equipped with a 25 SC pump head). This
then flowed through a check valve and into a coiled tube around an electrical
heater block, where the heater set point could be controlled up to 500 °C. A
thermocouple within the water heater block provided an electronic reading
45
(referred to as the “Water Heater Temperature”) as well as acting as a feedback
control. After the heater, the feed flowed past a thermocouple to give the “After
Heater Temperature”, then into the inner pipe of the reactor.
Meanwhile, the metal salt feed(s) were pumped, each with a maximum flow rate
of 10 ml/min, using Gilson HPLC (high-performance liquid chromatography)
pumps (Model 305, equipped with a 10 SC pump head) through check valves
and into the bottom of the counter current reactor as an upflow. As the product
rose and exited the reactor through one side arm of the cross piece, a
thermocouple attached at the other cross piece side arm monitored the
temperature. This gave the “Post Mixing Point Temperature”.
Upon leaving the reactor, the product stream was partially cooled by one cooling
loop. Here, a flow of mains tap water moved over the pipe carrying the product
stream without directly coming into contact with it. After this first cooling loop,
there was a cross piece at which another feed (e.g. a surfactant or capping
agent), could be pumped in. A check valve on this feed meant the process was
unaffected if this pump was not operating. A thermocouple was present on the
other arm of the cross piece, allowing the “Capping Point Temperature” to be
monitored.
The product stream subsequently flowed through a second cooling loop which
brought the temperature to around ambient. The two cooling loops could be
operated independently, permitting the temperature of the capping point to be
‘tuned’. The product stream passed through the back pressure regulator, BPR,
(manufactured by Tescom, USA) and the product suspension flowed out at
ambient pressure into a collect vessel.
46
The BPR constricted the flow to an adjusted level or set point and maintained it,
hence controlling the system pressure. Being a continuous system, the pressure
throughout the entire equipment was consistent. BPRs can be pneumatic or
manual, where the former is air-actuated and the latter is spring loaded. The air
or the spring pushes a needle down onto a seal causing restriction in the flow.
To measure the system pressure, a piezoelectric pressure transducer and
bourdon gauge were attached in the water feed. The piezoelectric transducer
was attached to a control box to give an electronic reading.
Figure 1.21. A photo of the counter-current reactor system.
1.7.2.2 Advantages of the Counter Current Design
Much of the early work into continuous hydrothermal synthesis used a T-
piece fitting as a mixing device, as described in previous Sections of this
Chapter. In many cases, the T-piece was upright with the heated water feed
47
entering through the side arm and the cold metal salt flowing down; the bottom
arm of the Tee then formed the product outlet. When Blood et al. performed LAI
analysis on their T-piece pseudo reactor, as described in Section 1.7.2, they
found that the water would rise into the top arm of the Tee. This was due to the
lower density and consequent higher buoyancy of the superheated water. By
flowing upwards, the metal salt feed became heated, causing the nucleation of
particulates. Without a strong flow to transport the particles away, this led to
accumulation and blockage at the top arm of the Tee. This observation caused
researchers to see the importance of differences in density between the two
flows.
The counter-current Nozzle reactor has the superheated water feed flowing down
the inner pipe before meeting the cold metal salt feed. The counter current
geometry ensures the metal salt stream is not heated prior to mixing, as this
could cause precipitation of the salt and in turn, blockage. Furthermore, as the
two streams meet, mixing and heating of the metal salt feed is rapid and
immediate - leading to the nucleation of particulates. After particle formation,
the product stream is hotter, less dense and thus more buoyant than the metal
salt upflow; therefore particles inherently flow upwards toward the reactor
outlet. The flow from both the water and metal salt also drives particle
movement, thereby preventing particle build-up, as well as creating a short
residence time which can prevent particle growth. Subsequent to leaving the
reactor, particles are quickly cooled in the cooling loop to inhibit particle growth.
Unlike the reactor designs from Adschiri, Arai and others (Adschiri et al., 2000,
Hakuta et al., 1998, Kawasaki et al., 2010, Aimable et al., 2009b), the Nozzle
reactor combines the mixing device and reactor into one entity which also
promotes instant and rapid heating of the precursor stream, and short residence
times.
48
1.7.2.3 Continuous Hydrothermal Scale Up
While developing the counter current Nozzle reactor, Lester et al. had the
aim of designing a reactor system which was capable of making the process of
synthesising nanoparticles industrially viable. This meant it needed to be capable
of being scaled up without having a detrimental effect on product quality.
Moreover, the scale up process needed to be economically feasible, as any
nanomaterial synthesis (or any emerging technology) that presents high
manufacturing costs will have difficulty in gaining widespread industrial interest.
In 2007, Lester et al. succeeded in building a pilot scale reactor system based on
the technology of the counter current Nozzle reactor, a photo of which is shown
in Figure 1.22, and is owned by Promethean Particles Ltd. (see Section 1.8.2).
The water and precursor flows are driven by two dosing pumps from Milton Roy
(Milroyal B) – each with a maximum flow rate of 400 ml/min. The water stream
is heated using a custom-built electrical heater (Watlow, output of 9 x 4.3 kW).
All pipes and fittings within the system are from Swagelok; where the lab scale
reactor uses a Swagelok cross-piece with an outer diameter of 3/8” (0.9525 cm),
the pilot scale reactor contains a cross-piece with an outer diameter of 11/2”
(3.81 cm) and a wall thickness of 1/4” (0.635 cm). Likewise, at the lab scale, the
inner pipe which carries the heated water stream has an outer diameter of 1/8”
(0.3175 cm). The equivalent pipe on the pilot reactor measures ½” as its outer
diameter with a wall thickness of 0.083” (0.211 cm). The system pressure is
maintained by a pneumatic back pressure regulator (Tescom) and the product
stream is cooled to ambient by a mains water fed heat exchanger built using
Swagelok parts. Typical synthesis experiments on the lab scale equipment
employ a total flow rate ca. 30 ml/min to generate g/h of equivalent dry
material, while the pilot reactor system is capable of flow rates up to 800
ml/min, to produce kg/day.
49
Figure 1.22. Photograph of the Pilot Scale reactor system which
incorporates the counter-current Nozzle reactor technology, courtesy of
Promethean Particles Ltd (see Section 1.8.2).
Currently, The University of Nottingham are involved in an EU-funded, Seventh
Framework Programme (FP7) project entitled SHYMAN (Sustainable
Hydrothermal Manufacture of Nanomaterials). One of the main objectives within
this project is the design and build of a commercial scale plant (production at
1000 tonnes per annum) based on the counter-current Nozzle reactor - making
it the largest scale production plant, using continuous hydrothermal technology,
in the world. The SHYMAN project will be discussed further in Section 1.8.2.1.
50
1.8 Commercialisation of Hydrothermal Synthesis
1.8.1 Hanwha Chemical
Hanwha Chemical is a Korean company established in 1965. Traditionally,
they have produced plastics, such as polyethylene (PE) and polyvinyl chloride
(PVC), and battery materials, as well as having biopharmaceutical departments
dedicated to drug development and production (Hanwha, 2013). In recent years
they have turned their attention to nanotechnology and built a pilot plant for the
hydrothermal synthesis of a range of metal oxide nanoparticles, via a continuous
process. Their pilot scale facility has the capacity to produce the equivalent of 4
kg of dry powder per hour, and was built to gather scale-up data in order to
develop an even larger plant.
At the end of 2010, Hanwha Chemical built a commercial plant for the
manufacture of lithium iron phosphate, LiFePO 4 – a material gaining much global
interest for battery applications. This plant employs the process of continuous
supercritical hydrothermal synthesis, and has a capacity of >200 tons per annum
– the current largest plant of this type in the world (Adschiri et al., 2011).
1.8.2 Promethean Particles Ltd.
A spin-out company was formed in 2008 called Promethean Particles to
take advantage of the patented Nozzle reactor. The company conducts
gram/hour scale feasibility studies to develop bespoke nanoparticle dispersions
for industrial customers in a diverse range of applications. In addition, a Pilot
scale reactor (as described in Section 1.7.2.3) is capable of producing
nanomaterials at a scale of 1 ton per annum, to meet larger demands.
51
As well as commercial projects, Promethean Particles participate in EU-funded
FP7 Projects, providing nanomaterials to Consortium Partners. At the time of
writing, there are 3 FP7 projects running in parallel – POINTS (Printable Organic-
Inorganic Transparent Semiconductor Devices) started in May 2011 and runs
until May 2014; NanoMILE (Engineered nanomaterial mechanisms of interactions
with living systems and the environment: a universal framework for safe
nanotechnology) began in March 2013 and runs until March 2017; and SHYMAN
(Sustainable Hydrothemal Manufacture of Nanomaterials), which started in May
2012 and will conclude at the end of April 2016. The SHYMAN Project revolves
around the ‘Nozzle’ reactor.
1.8.2.1 The SHYMAN Project
The SHYMAN Project is an EU funded Seventh Framework Programme
(FP7) study which started in May 2012 and will run over the course of 4 years
(SHYMAN, 2013). With 16 Consortium partners from academia, industry and
SMEs (small and medium enterprises), the aims of the project fall into three
main categories:
 The design, commission and build of a commercial scale plant, with the
ability to produce nanomaterials at a capacity of up to 1000 tons per
annum. As such, it will become the largest continuous hydrothermal
production plant in the world. Modelling (e.g. computational fluid
dynamics, CFD, and pseudo-fluids modelling) will be employed to
determine the deviations in fluid mixing with increasing scale while
engineers from academia and industry will consider and design solutions
for minimising the energy consumption of the plant, and maximising
product yield.
52
 The development and characterisation of existing and novel
nanomaterials (produced by the University of Nottingham and
Promethean Particles) which can be formulated by industrial partners into
their current products, in order to replace less sustainable additives, or
enhance their existing properties. Industrial partners include multi-
national companies Solvay, Repsol, PPG, and Fiat.
 The analysis of waste products generated during the manufacturing
process, and development of waste water recycling schemes, in order to
reduce waste emissions and other detrimental effects on the
environment. Moreover, Life Cycle Analysis (LCA) will be performed to
assess if nanomaterials produced using this technology are more
sustainable compared to from other existing technologies. As with many
chemical and engineering methods, particularly at pilot or commercial
scale, a lot of focus falls on environmental impact, and whether the
process adheres to the Principles of Green Engineering (Tang et al.,
2008).
By reaching these aims, the SHYMAN Project will develop commercial products
containing nanomaterials, as well as a plant capable of manufacturing these
nanomaterials to meet global demand. A highly interesting and exciting project,
SHYMAN demonstrates the real potential of the technology described in this
Thesis.
53
1.9 Challenges of Nanomaterial Synthesis
In previous Sections of this Chapter, the wide variety of benefits from
nanoparticles has been described. However, the synthesis of nanoparticles is not
without its challenges. This is true for most methods of synthesis, not just
continuous hydrothermal routes. A major challenge facing the synthesis of
nanomaterials, is achieving particles which are stable and do not agglomerate.
This issue will be discussed in Section 1.9.1.
1.9.1 Particle Stability
Due to the small size of nanoparticles, a high proportion of the
component atoms within each particle are present on the surface. Coupled with
the high surface area-to-volume ratio, nanoparticles have a much higher surface
energy compared to bulk material (Nanda et al., 2003). This results in
nanoparticles inherently being drawn to each other to minimise their surface
energy. The net result, if no opposing forces prevent them from doing so, is that
nanoparticles will coagulate into larger clusters. There are two forms of
coagulation: flocculation, which is reversible; and agglomeration, which is
irreversible. While flocculation can usually be solved by agitation of particles and
dispersing them in high viscosity media, agglomeration is not as simple. It poses
a huge disadvantage as, in many cases, the particles no longer move as
individual particles, but as larger micron-scale aggregates. The properties of the
particle can also be affected, no longer exhibiting the advantageous properties
seen at the nanoscale; in these cases, measures should be taken to prevent it
from occurring, otherwise the principle of having nanoparticles becomes void.
Preventing particle aggregation is particularly pertinent in continuous systems,
54
as large clusters are more likely to cause blockages – especially at bench scale
where pipes are narrow.
1.9.2 Methods of avoiding Particle Coagulation
Depending on the degree of coagulation between nanoparticles, a few
different methods can be used to break particles up or keep them apart. These
can be classified into mechanical or chemical methods. For flocculated particles,
mechanical methods can be employed; here, equipment is used to physically
break up flocculates, such as sonicators, homogenisers or dispersers. These use
physical force, such as sound energy or shear forces, to disrupt clusters and
disperse them into a carrier media. If the media has a high viscosity, the
particles disperse more slowly through the fluid, reducing the chance of re-
flocculation.
However, without altering the surface charge or chemistry of the particles, they
are able to re-coagulate. Hence, chemical methods are often more effective at
preventing particle coagulation. Capping agents or surfactants (surface active
agents) can be added to bind to the particle surface which, in turn, provides
either steric stabilisation (generally the case when polymeric surfactants are
used) or charge stabilisation (such as the use of acids or bases). Choosing a
suitable surfactant for each nanoparticle species can be difficult, as it should be
compatible with the solvent media, as well as able to chemically bind to the
particle surface. In addition, the concentration ratio between nanoparticles and
surfactant needs to be balanced; insufficient surfactant means incomplete
coverage and exposed particle surfaces are still able to bind and coagulate, while
too much surfactant can lead to ‘overcrowding’ on the particle surface, a
compression of surfactant layer, and a similarly detrimental effect. This is
depicted in Figure 1.23. Of course, this ratio is also dependent on particle size -
55
smaller particles have a larger surface area and therefore require more
surfactant molecules to bind to it (Lubrizol, 2013).
Figure 1.23. Ratios of surfactant to nanoparticle where (a) shows
insufficient surfactant concentration resulting in partial coverage, (b)
displays optimum surfactant coverage and maximum stability, and (c)
shows too much surfactant leading to layer collapse and unstable
particles (Lubrizol, 2013).
It has also been shown that surface modification can be used to transfer
nanoparticles into different solvent phases (Ohara et al., 2008); improve particle
dispersability into fluids (Kim et al., 2008); limit crystal growth in a particular
direction which then affects particle morphology (Mousavand et al., 2006,
Mousavand et al., 2007); or alters particle size (Taguchi et al., 2012).
1.9.3 Methods of Determining Particle Stability
If particle coagulation occurs, the clusters typically become denser than
the carrier solvent (in most studies, this solvent is water or basic alcohols) and
fall to the bottom of the vessel resulting in phase separation. Therefore, it is
very easy to observe by eye if coagulation is evident. Consequently, many
assessments of particle stability are empirical.
56
Still, the particles in a suspension may have sufficient surface charge to remain
stable in the short term, but will coagulate over time. Therefore the degree of
surface charge (attractive or repulsive forces acting on a particle) should be
measured to ascertain if the particles are truly stable. Zeta potential is a widely
used method to quantify particle charge. Although the technique does not
conclusively give information about whether particles have a positive or negative
charge, it is a reliable method for measuring long term particle stability. Zeta
Potential, and other structural characterisation techniques, will be discussed in
more detail in Chapter 2.
57
1.10 Aims of this PhD
Through the course of this Chapter, the benefits of using nanoparticles
have been described for a vast range of applications. While nanotechnology as a
field of research has existed for decades, its emergence into industrial
application remains in its infancy. With increasing numbers of studies
highlighting nanoparticles as safe to use, coupled with the exceptional data
obtained from research, more and more industries want to test and apply
nanoparticles into their products. This developing demand calls for synthesis
procedures which stretch far beyond the lab-scale, while keeping costs and
negative environmental impacts down.
The merits of continuous hydrothermal synthesis have been described in this
Chapter, with the importance of optimised reactor design being highlighted. In
response to this need, the counter-current Nozzle reactor was designed and
constructed at the University of Nottingham at both lab and pilot scale. While
studies have been conducted using the Nozzle reactor for the synthesis of a few
different nanomaterials (Hobbs et al., 2009, Lester et al., 2012, Aksomaityte et
al., 2013), the immense range of nanomaterials which could potentially be
produced by this method, largely remains unchartered territory. Because reactor
geometry plays such a vital role in the structural properties (size, shape,
crystallinity) of the synthesised particles, the products from the Nozzle reactor
may be vastly different to those made using a different continuous hydrothermal
reactor, even though synthesis is based on the same principle method. Since
particle structure ultimately dictates functional performance, products from two
different reactors could exhibit very different properties.
58
Therefore the invention of the Nozzle reactor opens up a vast field of research -
comprising the synthesis of a material, optimisation of the process, structural
characterisation of the products, and application testing. The latter provides
assessment of how materials can be applied to give real-world meaning. All the
while, this technology has been proven to be scalable, with high product
reproducibility. This principle forms the motivation of this PhD work.
The aims of this work are to use a novel continuous hydrothermal reactor (the
counter current Nozzle reactor) in the synthesis of a range of nanomaterials for
three different biomedical applications, to characterise the structure of those
nanomaterials, and finally to test the particles for properties which are specific to
that application. Testing performance properties and characteristics is key to
assessing the viability of applying the nanoparticles in a specific end application.
59
Chapter 2
2 Structural Characterisation
Techniques
This chapter will discuss the analysis techniques used in Chapters 3, 4
and 5 to characterise the structural properties of nanomaterials. For each
procedure, the principles of each technique will be discussed, along with the data
which can be obtained from it. The methods of sample characterisation which
are more specific to each individual Chapter, such as testing of application
properties, will be discussed in the relevant Chapter.
2.1 Scanning Electron Microscopy (SEM)
SEM is a surface imaging technique which allows the user to study the
particle size, topography, and morphology of samples using magnifications
between 100 and 100,000 times. With standard optical microscopes, light is
used to form an image. With SEM, images are formed with an electron beam
generated from an electron gun. Electrons are a type of ionizing radiation; using
them not only means that higher magnifications are possible due to their shorter
wavelength compared to visible light, but there is also a larger field of depth
compared to light microscopy. This means a large area of the sample is in focus
at one time, giving an image which is a good 3-dimensional representation of the
sample.
60
Chapter 2. Structural Characterisation Techniques
Figure 2.1. An illustration showing the components in a typical scanning
electron microscope (Flegler et al., 1995).
Samples of powders or particles are held on an aluminium stub using adhesive
carbon tape. If required, the surface of the sample and stub can be coated with a
very thin layer of metal such as gold, to enhance conductivity and in turn,
improve the flow of electrons across the sample surface. The stub is then loaded
into the sample chamber. The electron beam travels down the microscope under
vacuum from the gun, through a series of apertures and electromagnetic lenses
to control and modify the beam, as well as prevent electron spray, before
reaching the sample (the beam can be controlled to scan across a specific area
of the sample). The orientation of these components is shown in Figure 2.1.
When the beam of electrons hits the sample, they interact and different signals
are generated, including secondary electrons (SE), backscattered electrons
61
(BSE), and X-rays. Detectors are in place to pick up each of these signals. The
detected electron intensity is converted to light intensity which gives a visual
image that can be digitally viewed. SE are electrons expelled from the sample
atoms as a result of interactions with the (primary) electrons of the beam. SE
generally have low energy and are used to give the best imaging resolution,
where contrast in the image is principally due to sample topography. On the
other hand, BSE are electrons from the beam that have been scattered back
after collisions with the nuclei of sample atoms, where nuclei with a higher
atomic number cause more electrons to be scattered back. As a result, these
atoms appear brighter in BSE images and in turn, contrast in these images can
give information about sample composition (Griffith and Danilatos, 1993).
Simple sample preparation, rapid data acquisition, high magnification imaging,
and being a non-destructive technique (meaning samples can be analysed
repeatedly) make SEM a very useful technique. However, disadvantages are that
only solid material can be analysed, which has to fit in the sample chamber
(usually a maximum of 10 cm). Samples usually have to be stable under high
vacuum, although microscopes with low vacuum or environmental modes are
available. In addition, some samples (particularly organic samples) are beam
sensitive and can become damaged (Flegler et al., 1995).
The SEM kit used in this study is a FEI Quanta 600 with a lanthanum hexaboride
(LaB6) filament to generate an electron beam. In most cases, an accelerating
voltage of 25 or 30 kV is used.
62
2.2 Transmission Electron Microscopy (TEM)
TEM operates using a similar principle to SEM, where electrons are used
as a “light” source. However, unlike SEM, the electron beam travels through the
sample (Figure 2.2 (a)). Some electrons may become scattered depending on
the density of the material, but those that are not, hit a screen at the bottom of
the microscope to form a “projection” of the sample where parts may appear
darker than others depending on density. Thus, TEM can be useful for visualising
hollow structures (Figure 2.2 (b)) (Williams and Carter, 2009).
(a)
(b)
Figure 2.2. (a) A simplified depiction of how a typical TEM operates; (b)
a schematic showing a 3-dimensional tube is viewed as a 2-dimensional
image under a TEM.
63
There are different types of TEM, for example Scanning TEM (STEM), high
resolution TEM (HRTEM) and analytical electron microscopy (AEM), providing a
choice over image analysis. TEM also allows higher magnifications than SEM and
can determine, to an extent, the internal structure of particles. In addition, the
crystal lattice of particles can even be observed with HRTEM. Nevertheless, the
trade-off of higher magnifications is the decreased sample area, which questions
if an image obtained using TEM is truly representative of the full sample.
Furthermore, there can be confusion when interpreting images; after all, these
are 2D images of 3D samples (Williams and Carter, 2009). Like SEM, some
samples can be sensitive to the ionizing radiation of the electron beam; as a
result, some materials (particularly organic materials) may suffer heat or
chemical etching damage.
In this work, samples were prepared for TEM analysis using one of two methods;
nanoparticles produced in water from the reactor were washed and diluted with
deionised water before being loaded onto the copper-grid mounted ‘holey’
carbon films and allowed to air dry on filter paper. Dry powder samples were
suspended in alcohol, typically ethanol or methanol, and sonicated. A few drops
of the suspension were loaded onto the grids and air dried. Alcohol is used
because it tends not to react with the samples, but has relatively low surface
tension (compared to water, for example) thereby allowing the particles to be
well distributed over the sample grid.
Images were obtained using a JEOL 2100F HRTEM equipped with a field emission
electron gun (FEG) and operating at 100 kV accelerating voltage.
64
2.3 Energy Dispersive X-Ray Spectroscopy
(EDX/EDS)
EDX analysis is an additional feature of most SEM and TEM instruments.
As previously mentioned in Section 2.1, electron interaction with a sample can
also produce X-rays. This signal can be specifically detected to give information
on the elements present within a sample. In the case of the FEI Quanta 600 SEM
equipment mentioned in Section 2.1, collected EDX data was interpreted using
Genesis software by EDAX.
EDX analysis can be a very helpful confirmatory tool, but very light elements
such as H, He and Li cannot be detected.
2.4 X-Ray Diffraction (XRD)
This analysis method is used to ascertain information about the crystal
structure (or arrangement of atoms) in a sample. Figure 2.3 illustrates how data
can be obtained; X-rays of known wavelength (λ) are fired at the sample, which
then interact with the electrons of atoms on each crystal plane. As a result, the
X-rays are scattered to give a signal (elastic scattering) and are identified by the
detector component of the equipment.
During analysis, X-rays are directed at the sample at a range of angles (incident
angle θ seen in Figure 2.3). As the rays hit the sample, they are diffracted by the
lattice structure of the crystal. A detector moves around the sample and
measures the intensity of these peaks at each angle. Thus, signal intensity at
each angle depends on the orientation and spacing between crystal planes.
65
Figure 2.3. A diagram showing the basis of XRD analysis.
The relationship between the X-ray wavelength, the d-spacing, and the angle θ
can be summarised by Bragg’s Law:
where: is an integer,
is the wavelength of the beam,
is the spacing between diffracting planes, and
is the incident angle
Each crystal structure has a unique diffraction pattern, giving signal peaks of
different intensity at differing angles of 2θ; once the diffraction pattern (or
corresponding d-spacings if calculated using Bragg’s Law)) of a sample has been
measured and the constituent elements are known, it can be matched to an
existing library of known structures in order to identify the crystal phase and the
component compound(s) in the sample.
66
With specific analysis, the diffraction pattern of a sample can be used to infer
much more information than just compound identification. The shape of peaks
in a diffraction pattern can be an indication of the degree of crystallinity in a
sample. For example, a highly crystalline sample will have consistent d-spacings
and in turn, high signal intensity for a narrow range of angles. Therefore, the
resulting diffraction pattern will have sharp narrow peaks. Conversely, samples
with a lower degree of crystallinity will have more irregular d-spacings and the
subsequent diffraction pattern will show broader, less intense peaks. Amorphous
samples (those which contain atoms arranged in a random order with no distinct
crystal planes) will not create a detectable signal and so, do not give any peaks
or diffraction patterns.
It is important to note, however, that peak shape and size can also be ascribed
to factors other than the degree of crystallinity in the sample. The instrument
used can play a role in the peak profile obtained, and for small particles
(principally <200 nm) - crystallite size. The Scherrer equation establishes the
relationship between peak broadening and crystallite size:
Where: = crystallite size
= the Scherrer constant, which takes into account the shape of
the crystal
= X-ray wavelength
= the full-width of the peak at half maximum (FWHM), in
radians
= the Bragg angle
67
The Scherrer equation demonstrates that crystallite size is inversely proportional
to peak broadening, i.e. a wider peak indicates a smaller crystallite. Because
this effect is mainly seen for small particles, size analysis by other techniques
such as TEM should be used concurrently with XRD analysis, in order to infer the
cause of peak broadening. It should also be noted that crystallite size is not the
same as particle size, as a particle may be made up of different crystallites.
In this work, dry powder samples were analysed using a Bruker D8 Advance
diffractometer, using Cu Kα1 radiation (λ = 0.15406 nm) with a step size of 0.04
and step time of 5 seconds. Typically, scans were taken for a 2θ range between
15 and 65˚. Diffraction patterns were analysed using EVA processing software
for phase identification, while Xfit (Cheary and Coelho, 1996) and Celref
(Laugier, 1999) software were used to calculate crystallite size and cell
parameters, respectively.
68
2.5 Time-of-Flight Secondary Ion Mass
Spectrometry (ToF-SIMS)
ToF-SIMS is a method for analysing the chemical composition of the
surface of a solid sample. It uses a pulsed ion beam (Bi 3+) to remove molecules
from the very outermost surface of the sample in a chosen square area. Particles
are removed from atomic monolayers on the surface (secondary ions). These
particles, which could be full or fragmented molecules, are then accelerated into
a "flight tube" and their mass is determined by measuring the time taken to
reach the detector (i.e. time-of-flight). Three operational modes are available
using ToF-SIMS: surface spectroscopy, surface imaging and depth profiling
(Vickerman and Briggs, 2001).
This method has the advantages of being non-destructive, easy to use and with
little sample preparation, and high sensitivity for trace elements or compounds.
Nevertheless, the equipment collects a spectrum for each pixel of the chosen
square area of analysis. This can provide too much data which can be very
time-consuming to interpret.
Data obtained from ToF-SIMS analysis will be described in Chapter 3.
69
2.6 Dynamic Light Scattering (DLS)
DLS is a particle sizing technique, also referred to as photon correlation
spectroscopy (PCS) or quasi-elastic light scattering (QELS), which measures the
rate of particle diffusion in a suspension by Brownian motion and correlates this
to particle size. In a sample of a liquid dispersion, solvent molecules move
around in solution due to their thermal energy. In turn, these bombard or “kick”
particles that are also present in the suspension and this is known as Brownian
motion. As such, particles must be suspended in a solvent, to be analysed by
DLS.
For DLS analysis, a monochromatic laser illuminates the particles in a liquid
dispersion and the light is scattered. The intensity of this scattered light varies
depending on the size of the particles because smaller particles are pushed
further by solvent molecules and move faster. By measuring the changes in
intensity of scattered light, the velocity of the Brownian motion can be found;
the particle size is then calculated using the Stokes-Einstein equation:
( )
where: ( ) = hydrodynamic diameter
= Boltzmann’s Constant
= absolute temperature
= viscosity
= translational diffusion coefficient
The particle size calculated by DLS relates to how the particle diffuses in the
surrounding dispersant and is called the ‘hydrodynamic diameter’ (Figure 2.4).
70
This value is dependent on the particle core, as well as any surface modifications
to the particle, and the surrounding ions present in the solvent. The particle
shape affects the reported size as well, and the measured hydrodynamic
diameter for non-spherical particles is given as “the diameter of a sphere with
the same translational diffusion speed as that particle” (Malvern, 2013).
Consequently, the measured particle size can be larger than that seen using
electron microscopy (where a particle may be out of its natural environment).
Figure 2.4. A diagram of the hydrodynamic diameter of a particle which
is reported by DLS analysis (Malvern, 2013).
The application of DLS analysis will be principally discussed in Chapters 4 and 5;
DLS data was obtained using a Zetasizer Nano-ZS by Malvern Instruments,
which can measure particles in the range of 0.3 nm to 10 μm. This instrument
uses a He-Ne laser with a wavelength of 633 nm. Where required, the Refractive
Indices for sample materials were taken from the “Dispersion and Refractive
Index Guide” from Malvern Instruments. Data analysis was performed using
Zetasizer software Version 7.20 from Malvern Instruments. The obtained particle
size data is initially given as an intensity distribution for the scattered light. In
this instance, where I is the signal intensity and d is the particle diameter;
71
larger particles provide a much stronger signal than smaller particles meaning
data can be skewed if larger particles, dust or agglomerates are present in the
sample.
For samples which contain different sized particles, i.e. where there is more than
one peak in the intensity distribution graph, it may be necessary to convert the
data to a volume based distribution. To do this, the refractive index and
adsorption value of the particles needs to be added to the software. Using Mie
theory, the software then converts the data to a volume-basis so that the
aforementioned data-skewing due to larger particles is not seen.
This technique benefits from being a non-invasive method that requires
practically no sample preparation. Moreover, data acquisition takes only a few
minutes – significantly shorter than most other particle sizing techniques, and a
fraction of the running costs compared to electron microscopy. On the other
hand, particles must be stable and well dispersed in the solvent in order to
obtain a viable measurement; if particles are coagulated, the size of the clusters
is measured and no information on the primary particle size can be gathered.
Thus, samples with particles which have coagulated and phase separated due to
gravity can be visually identified, and analysing these with DLS is redundant.
72
2.7 Zeta (ζ ) Potential
The Zetasizer Nano ZS from Malvern Instruments can also measure the
zeta potential of colloidal systems, which gives an indication of the particle
stability i.e. if the particles are likely to coagulate over time.
In aqueous media, particles in dispersion carry a surface charge. This is
dependent on the particle’s surface groups. For example, dissociation of acidic
surface groups will lead to a negative change, while basic groups dissociate and
give rise to positive charge. In addition, the presence of anionic or cationic
surfactants around particles affects the surface charge. Figure 2.5 shows an
example particle with a negative surface charge. Depending on the pH and ionic
strength of the solution, ions of opposite charge to the surface are strongly
bound to the particle, called the Stern layer. Around this layer are more loosely
bound ions – part of the slipping plane. As a result, each particle has an
electrical double layer which is considered a stable entity and moves in solution
with the particle. Outside of this electrical double layer, the ions are considered
as part of the bulk dispersant and do not move with the particle. Zeta potential
is a measure of the net electrical charge for a particle and its electrical double
layer.
73
Figure 2.5. A schematic representation of what is measured by zeta
potential (Malvern, 2013).
To measure zeta potential, an electric field is applied across a sample. Charged
particles will move towards the electrode of opposite charge at a speed that is
proportional to the size of its charge. The Zetasizer measures this velocity and,
considering the viscosity of the dispersant, can calculate the zeta potential value.
This data can give information on the attractive or repulsive forces present in a
dispersion and in turn, how likely particles are to coagulate. A high zeta potential
is considered as <-30 mV or >+30 mV (Malvern, 2013), and the particles in that
suspension can be considered to be stable.
The zeta potential of particles varies with the pH and conductivity of the
surrounding media, so measurements can be taken over a pH range in order to
ascertain the conditions where particles begin to aggregate. This is known as the
point of zero charge (PZC) or iso-electric point (IEP).
74
Zeta potential is crucial analysis for understanding colloidal systems, and in the
development of stable, non-aggregating dispersions. This technique holds the
advantage of rapid data acquisition as measurements generally only take a few
minutes, but drawbacks include the need to prepare samples with accuracy as
pH, conductivity and presence of surfactants can significantly affect zeta
potential.
The application of zeta potential measurements will be discussed in Chapter 4.
Measurements were taken for dilute suspensions of each sample (0.1 mg/ml)
over a pH range, while the background electrolyte (sodium chloride, NaCl) was
maintained at 0.01 M. Hydrochloric acid, HCl, or sodium hydroxide, NaOH, were
used to alter the pH. Samples were loaded into disposable zeta cells and for
each sample, three measurements were obtained using a Malvern Zetasizer
Nano ZS, and standard deviations calculated. Data were analysed using Malvern
Zetasizer software.
75
Chapter 3
3 Hydroxyapatite
3.1 Introduction
Prior to this project, hydroxyapatite had been established as a material of
interest in the Nottingham group and some preliminary experiments have been
described previously (Li, 2008). As a result, one of the objectives of this work
has been to continue the synthesis of hydroxyapatite using the Nozzle reactor, to
explore the experimental parameters governing formation, and to test material
properties for biomedical application.
3.1.1 Hydroxyapatite (HA), Ca10(PO4)6(OH)2
Apatites are a group of phosphate minerals which are widely found in
nature. These are isomorphic compounds which exhibit the same hexagonal
dipyramidal crystal structure. HA has unit cell dimensions of a = 9.432 Å and c =
6.881 Å (Posner et al., 1958) (See Chapter 2 for general details on crystal
structure and how it is analysed). The OH- ion of HA can undergo substitution
with carbonate (CO32-), fluoride (F-) or chloride (Cl-) ions to give carbonate-
apatite, fluorapatite and chloroapatite respectively. In addition, the molar ratio
of calcium to phosphorus can vary from the stoichiometric ratio of Ca/P = 1.67;
as such, calcium-deficient HA exists. These isomorphs display different
properties to HA. For example, fluorapatite is chemically more stable than HA
(Montazeri et al., 2011). Such property changes can be exploited to tailor apatite
76
Chapter 3. Hydroxyapatite
materials suitable for specific applications. Hydroxyapatite, and chemically
similar forms of HA, make up the inorganic component of teeth and bones in
mammals. For this reason, HA has been widely utilised in the biomedical industry
for many decades. This is described in more detail in Section 3.1.2.
3.1.2 Current applications using HA
The biocompatible and bioactive nature of HA, namely that it can interact
with living tissue without being rejected or causing infection, makes it a very
desirable material for biomedical applications. It is already widely used in
dentistry (Whitters et al., 1999), plastic surgery (Gladstone et al., 1995) and
orthopaedics (Block and Thorn, 2000). In this Section, applications of HA in
tissue engineering will be discussed further.
3.1.2.1 Bone Scaffolds
Ideally, bone scaffolds should be temporary structures which are
implanted onto or into damaged bone, gaps caused by trauma or infection, or
where fusion of joints is needed. Once implanted, the scaffold should facilitate
the formation of new bone by being either osteoinductive (recruiting and
stimulating immature cells to become young bone cells, or preosteoblasts) or
osteoconductive (allowing the formation of new osteoblasts - cells that regulate
bone tissue synthesis on the surface of or within the structure of a bone scaffold
or graft). The scaffold will eventually decompose once the bone has healed.
Both osteoinduction and osteoconduction are vital for the bone healing process;
however, subsequent to many bone fracture incidents, much of the bone healing
procedure is reliant on osteoinduction (Albrektsson et al., 2001). In craniofacial
implantology, high rates of osseointegration have also been reported (Gladstone
77
et al., 1995), where direct bone-to-implant contact is achieved to provide the
implant with a firm basis, and have led to efficient bone healing.
Through introducing or increasing porosity of HA scaffolds, an improvement in its
osteoconductive properties has been demonstrated (Mastrogiacomo et al.,
2006). Upon implantation, osteoblasts are able to weave into and through the
HA scaffold, promoting new cell, blood vessel and bone matrix development.
Nevertheless, the disadvantage of these scaffolds is the brittle nature of HA. The
porosity has a further detrimental effect on the mechanical strength of the
scaffold. As a result, it is not possible to use these bone grafts alone in a
position where load bearing is required. In these cases, HA is applied as a
coating instead.
3.1.2.2 Coating on Prosthetics
Many materials used for prosthetic implants such as titanium, titanium
alloys, copper, and bone cements are exploited for their capability of load
bearing duties. They are however, limited by their low biocompatibility. In turn,
they demonstrate little or no osteoconduction. To solve this problem, implants
are coated with HA to increase biocompatibility and in some cases,
osseointegration. Implants are principally coated with HA by the plasma spray
process. Here, HA powder is heated in a plasma jet until molten and then
propelled towards the substrate. As it solidifies, the HA hardens onto the implant
material (Zhou and Lee, 2011).
3.1.2.3 Other Applications
Hierarchical microspheres of HA have been used as drug delivery
vehicles, namely in the slow release of ibuprofen (Yang et al., 2012b), while HA
78
has also been used for protein purification (Orlovskii et al., 2002), separation of
biomolecules (Bio-Rad, 2013) and as a component of dietary supplements (Shen
et al., 2007).
3.1.3 Potential Applications for HA
Researchers have recently been focussing their studies on HA with other
materials incorporated into the structure, such as cations (Kalita and Bhatt,
2007, Oh et al., 2004, Webster et al., 2004), or drugs (Leprêtre et al., 2009) –
these studies report on the synthesis and characterisation of such materials.
The doping of different cations into the HA structure has been found to improve
upon the existing biocompatible and osteoconductive nature of HA, afterall
natural apatite tends to contain a range of cations other than calcium (Webster
et al., 2004). For example using silver as a dopant can induce antimicrobial
activity (Oh et al., 2004, Stanic et al., 2011), while magnesium-doped HA has
been shown to display improved cell adhesion (Mroz et al., 2010), and HA doped
with zinc was shown to be biocompatible after in vivo tests (Saraswathi Amma et
al., 2008), as well as possessing antimicrobial activity (Stanić et al., 2010) and
can be used for the controlled release of bone growth proteins (Azizian-
Kalandaragh and Khodayari, 2010). In addition, drugs such as statins have been
used to produce composite materials with HA in order to enhance osteogenic
effects (Griffiths and Cartmell, 2007, Jadhav and Jain, 2006).
This research shows that although HA is a widely used material, and has been
for several decades, it remains the focus for current research as there are still
ways to manipulate the structure and further enhance its beneficial properties.
79
3.1.4 Current Methods of Producing HA
Due to its excellent properties and wide range of applications, several
methods of generating HA have been reported, where products are of the nano-
or micron- scale. A few of these techniques will be detailed in this Section,
including some advantages and disadvantages of each.
3.1.4.1 Naturally Occurring
Many HA based ceramics and cements that are already commercially
available for bone graft implants are derived from natural sources. For example,
Endobon® and Pro-Osteon® are produced from bovine and coral HA respectively,
and have a structure like that of human bone (Habraken et al., 2007). However,
there are limitations in altering their mechanical strength and porosity, which in
turn limits their applicability (e.g. as bone scaffolds). Moreover, naturally
occurring HA is a limited resource.
3.1.4.2 Precipitation Method
Synthetic HA powders which are analogous to naturally occurring HA can
be produced from a range of different precursors and using several different
synthetic routes. A commonly used technique for manufacturing HA is the wet
chemical method of precipitation. As described by Osaka’s research group
(Osaka et al., 1991), orthophosphoric acid solution was added to calcium
hydroxide in a controlled, dropwise manner with stirring. This was conducted
with various reaction temperatures between 0 and 100 ˚C with the pH monitored
throughout. The study showed that higher reaction temperatures led to the
formation of a more crystalline product, and processing at a pH lower than 9 led
to the formation of calcium deficient HA.
80
Another study (Zhang and Lu, 2008) described the use of calcium nitrate
tetrahydrate and diammonium hydrogen phosphate, Ca(NO 3)2.4H2O and
(NH4)2HPO4 respectively, which were mixed and kept at 85 ˚C for 48 hours in
the presence of urea, which underwent hydrolysis and then yielded calcium
phosphate (CaP) crystals. The group found that the urea concentration dictated
the type of CaP produced, either dicalcium phosphate or octacalcium phosphate.
They went on to immerse these two CaPs in ammonium hydroxide solution (pH
12) at 60 ˚C for 48 hours, and HA particles were obtained with nanobelt
morphology. With this method, the reactants were continually stirred after the
precursors were mixed, and left to age. This allowed the calcium to be slowly
integrated into the apatite structure, thus aiding the material to reach the
stoichiometric Ca/P ratio. This method yielded particles with a narrow size
distribution; however, the lengthy time required for synthesis forms the major
disadvantage of the precipitation method as both the described procedures took
several days before HA powder could be obtained. In turn, this creates difficulty
in producing material at commercial scale, problems with batch to batch
reproducibility, and the implication of high cost.
3.1.4.3 Sol-Gel Method
This procedure essentially uses a solution (“sol”), which is left to
gradually form a gel-like substance, containing a liquid and solid phase. The
solvent is then removed by drying or centrifuging and decanting, to yield a
powder.
Vijayalakshmi and Rajeswari (Vijayalakshmi and Rajeswari, 2006) reported the
use of sol-gel method to produce microcrystalline HA. Here, they dissolved
calcium acetate, Ca(C2H3O2)2, in water as the calcium source. To this they added
81
a triethyl phosphate solution, (C2H5)3PO4, dropwise until a calculated
stoichiometric Ca/P ratio of 1.67 was achieved. After stirring at room
temperature, the solution was left to age for 24 hours. This allowed gel
formation which was then dried at 120 ˚C for 16 hours to give a powder product,
which was then washed and dried again. It was found that the samples needed
to be calcined at temperatures up to 700 ˚C before highly crystalline products
could be obtained. The methodology is relatively simple but, again, this
procedure is time-consuming, with limitations to scalability. Moreover, the
calcination step can be energy intensive.
3.1.4.4 Microwave Synthesis Method
In recent years, various groups have used microwave radiation to
produce different sizes and morphologies of hydroxyapatite powders, where the
reported microwave power has ranged between 175 W and 900 W. For example,
Kalita and Verma (Kalita and Verma, 2010) stirred calcium nitrate tetrahydrate
with EDTA (ethylenediaminetetraacetic acid, which acts as a chelating reagent)
and sodium phosphate dibasic, Na2HPO4, for 240 minutes at room temperature.
After the pH was adjusted to 9 using ammonium hydroxide, NH4OH, the
suspension was loaded into a 600 W domestic microwave oven with a
customised reflux system, and heated on a cycle for 19 minutes. The product
was reported as being highly crystalline HA with particles having a size
distribution between 5-30 nm and a mixture of elliptical and rod morphologies.
Unlike precipitation and sol-gel, microwave synthesis is less time-consuming, but
the main disadvantages are the high equipment costs and problems associated
with scaling up the technology.
82
3.1.4.5 Hydrothermal Synthesis of HA
Hydrothermal synthesis is a commonly used method for producing HA. A
process using a Teflon-lined batch hydrothermal reactor has been described
previously (Earl et al., 2006). In this study, calcium nitrate tetrahydrate and
ammonium phosphate dibasic of stoichiometric ratios were dissolved into
distilled water. The solution was placed in the reactor and oven heated to 200 °C
for a range of time between 24 and 72 hours, then subsequently washed with
distilled water by centrifuging and decanting several times. Where reaction time
exceeded 24 hours, monetite (CaHPO4), was found to be produced as well as HA.
Most reported hydrothermal methods of HA production, like the aforementioned,
are batch processes, where pressure is a difficult parameter to monitor and alter.
This is because most off-the-shelf batch vessels do not incorporate pressure
transducers. Instead, these are typically custom-designed and manufactured;
even then, the pressure inside the autoclave is autogeneous and dependent on
the reaction temperature.
Conversely, the reaction temperature and pressure are arguably easier to control
and monitor in continuous systems. The full benefits of continuous hydrothermal
processing is outlined in Chapter 1; in brief, the ability to use this method for
reliable synthesis has been dependent on a reactor geometry which allows
homogeneous mixing without causing blockages in the system – the counter-
current Nozzle reactor has been shown to demonstrate these qualities (Lester et
al., 2006, Lester and Azzopardi, 2005)
There have been publications from a different research group using a continuous
hydrothermal system to produce HA (Chaudhry et al., 2006). The reactor used in
this study is described in Chapter 1, Section 1.7.1.3. In their report, the
83
researchers used water in one flow which was pressurised and heated to
supercritical conditions. In the reactor, it was met by a cold aqueous flow of
NH4HPO4 and Ca(NO3)2, leading to the production of HA nanorods with diameters
measuring 40 nm and lengths up to 140 nm.
Following this publication by Chaudhry et al., our research group at Nottingham
chose to experiment with the synthesis methodology, altering various process
parameters. It was found that, rather than having water in the downflow and
two metal salts in the upflow, the aqueous NH4HPO4 could be pumped into the
heater and be used as the superheated downflow. This precursor solution
remains stable at high temperatures (i.e. the precursor does not precipitate out
of solution), and does not cause blockages in the heater. Moreover, HA synthesis
was at successful at sub-critical conditions, with the heater temperature at 200
˚C, instead of supercritical conditions of 400 ˚C (Li, 2008). These optimisation
measures meant the need for an extra pump was eliminated, and a more
energy-efficient synthesis temperature could be adopted. These preliminary
experiments also found that the synthesis parameters could be tuned to produce
three different morphologies of HA – sheets, rods and tubes (Li, 2008).
3.1.5 Chapter Aims
Subsequent to the results obtained by other researchers in the group (Li,
2008), this current work set out to explore the effect of each experimental
parameter on the morphology and crystallinity of the resulting product, in order
to shed light on the formation mechanism of the different morphologies. As
mentioned previously in Section 3.1.3, the incorporation of cations and drugs
into HA is desirable. Because the continuous hydrothermal method employed in
this project is a “bottom-up” approach, the work reported in this Chapter aimed
to demonstrate the synthesis of HA composites (containing drugs or dopants),
84
and to assess the effect these nanomaterials have on mouse embryonic stem
cells (Lester et al., 2013).
85
3.2 Methodology
3.2.1 Synthesis of HA
The fundamental experimental conditions for the synthesis of
hydroxyapatite were taken from Section 4.7.2 of Li’s Thesis (Li, 2008). The
production of HA sheets or platelets was described, and so the protocol was
replicated and is reported in Section 3.2.1.1.
The precursors used were ammonium phosphate dibasic, (NH4)2HPO4 (≥98%
purity, Sigma Aldrich, UK); calcium nitrate tetrahydrate, Ca(NO3)2.4H2O (≥99%
purity, Sigma Aldrich, UK); ammonium hydroxide, NH 4OH (35%, Fisher
Scientific, UK); sodium hydroxide, NaOH (≥97%, Sigma Aldrich, UK); zinc nitrate
hexahydrate, Zn(NO3)2 · 6H2O (98%, Sigma Aldrich, UK); simvastatin, C25H38O5
(≥97%, Sigma Aldrich, UK) and deionised (DI) water. All reagents were used
without further purification.
3.2.1.1 Synthesis of HA Platelets
An aqueous solution containing 0.015 M (NH4)2HPO4 was pumped at a
rate of 20 ml/min through a preheater, and into the counter-current Nozzle
reactor as a downflow. Simultaneously, aqueous Ca(NO3)2.4H2O at a
concentration of 0.05 M was used as the upflow and pumped at a rate of 10
ml/min to meet the downflow. The temperatures at different points in the
system are outlined in Table 3.1, where the thermocouple positions were stated
in Chapter 1, Section 1.7.2.1. The system pressure was maintained at 240 bar
by the BPR. Images demonstrating the experimental setup are shown in Figure
86
3.1 while Table 3.1 shows the measured temperatures at different points within
the system during HA synthesis.
The product appeared as a white aqueous mixture, where the white particles
settled over time. This sample will be referred to as HA-1.
Figure 3.1 Images illustrating the continuous hydrothermal reactor
system. The reactor section is highlighted in (b), showing the
superheated aqueous (NH4)2HPO4 (red) solution mixing with the
ambient Ca(NO3)2.4H2O (yellow) to form particles at the interface (c).
Table 3.1. Reactor system temperatures for synthesis of HA-1
Water Heater After Heater Post Mixing Capping Point Post Cooler
Set Point (°C) (°C) Point (°C) (°C) (°C)
220 215 116 52 23
87
Other experiments conducted for the synthesis of HA were based on the above
protocol, with slight deviations. With this reactor system and the experimental
setup, several different parameters can be changed – including precursor type,
precursor concentration and ratios, precursor pH, flow rates and ratios,
temperature, pressure, reaction solvent, and the addition of dopants or capping
agents. To investigate the effect each of these parameters has on the product
(be it particle size, morphology, crystallinity, stability and/or functional
properties) would lead to a massive matrix of experiments and would be too
time consuming for the scope of this work. Thus, a limited set of parameters
have been selected for experimental investigation, based on previous work in the
group. These being precursor concentration and ratios, precursor pH,
temperature, and the addition of dopants and are described in Sections 3.2.1.2
to Section 3.2.1.5.
3.2.1.2 Investigating the Effect of Precursor pH
Synthesis of HA with rod morphology was described previously by Dr Li.
This involved the same reaction conditions as mentioned in Section 3.1.6.1., but
with one difference – the pH of the downflow (NH4)2HPO4 solution was increased
to 10 (compared to an as-prepared pH of 8). This was achieved by the dropwise
addition of NH4OH, whilst being monitored by a pH meter (Hanna Instruments).
All other reaction parameters remained unchanged. By eye, the product in this
scenario looked the same as sample HA-1 – with the white particles settling over
time. This sample shall be referred to as HA-2.
To determine if morphology changes are solely caused by pH alterations to the
downflow solution, or if the same product is yielded from upflow precursor
changes, a sample was produced using as-prepared (NH4)2HPO4 in the downflow
88
(at pH 8), and Ca(NO3)2.4H2O with pH increased to 10 (compared to its
unaltered pH of 5), using NH4OH. The product was labelled as HA-3.
To establish whether rod formation is base-specific (i.e. if rods are only formed
when NH4OH is used as a pH buffer), or more generally pH-specific (i.e. if any
base can be used to alter the pH of the downflow and still result in this
morphology being produced), experiments were conducted using NaOH as a
buffer to increase downflow solution pH. To do this, a 1 M solution of NaOH was
prepared using solid pellets (4g NaOH in 100 ml deionised water) and added
dropwise to the (NH4)2HPO4 solution whilst the pH was constantly being
monitored.
The precursors were pumped into the system at the flow rates stated in Section
3.1.6.1, with the preheater set point ranging between 200 and 450 ˚C, at 50 ˚C
increments, giving rise to samples HA-4a, 4b, 4c, 4d, 4e and 4f. Table 3.2
presents the thermocouple measured temperatures at the points in the system.
A summary of experimental parameters for each sample is given in Table 3.3.
89
Table 3.2. System temperatures during synthesis of HA-4 sample
Water
After Post Post
Heater Capping
Sample Heater Mixing Cooler
Set Point Point (°C)
(°C) Point (°C) (°C)
(°C)
HA-4a 215 210 115 51 21
HA-4b 265 261 179 106 28
HA-4c 315 314 218 124 30
HA-4d 365 365 266 144 31
HA-4e 410 405 362 178 35
HA-4f 455 453 378 202 35
Following on from the work previously conducted in the group, experiments were
performed using the precursors and flow rates as described for synthesis of HA-2
(i.e. downflow solution had been adjusted to pH 10 with NH 4OH), but using a
higher preheater temperature of 350 ˚C, to give sample HA-5. This temperature
was chosen because it was established that using reaction temperatures near the
supercritical point of water yielded novel and interesting morphologies –
specifically nanotubes (Li, 2008). Nevertheless, little work had been done to
explore the factors governing the formation of this morphology, and so this work
aims to test the reproducibility of those findings, and to elucidate the formation
mechanism.
To investigate if these tube structures are reliant on pH alterations in a specific
precursor feed, the upflow (Ca(NO3)2.4H2O) had NH4OH added to it until a pH of
10 was reached. The experimental temperature was held at 350 °C, consistent
90
with experiment HA-5 where tube structures were produced. This yielded sample
HA-6.
With the goal of identifying how morphology is affected when the pH of both
precursors are as-prepared (as for HA-1) but reaction temperature is increased
to 350 °C (as for tube synthesis), the sample HA-7 was produced. Table 3.3
gives a summary of experimental parameters used for the synthesis of samples.
3.2.1.3 Investigating the Effect of Precursor Concentration
There are several questions surrounding the effect of precursor
concentration on the manufactured product:
1. What are the limits of reaction conversion?
2. Will HA remain the principle species produced when concentrations are
increased?
3. How will particle size be affected, as the dynamic of particle nucleation
versus particle growth changes?
4. In terms of reactor design, what are the limitations of the reactor
system?
The latter point exists because the reactor system is composed of several
constricted pipes and a BPR, all of which can become blocked if there are simply
too many particles which are coagulating. To address these questions, a range of
experiments were conducted where precursor concentrations were changed from
those specified in Section 3.2.1.1.
In the first experiment, the stoichiometric ratio of Ca:P was retained at 1.67, but
reactant concentrations were doubled from the protocol described for HA-1;
91
namely 0.03 M (NH4)2HPO4 and 0.1 M Ca(NO3)2.4H2O solutions were used. Here,
all other parameters were unchanged – giving rise to sample HA-8.
Similarly for the second experiment, reactant concentrations were increased
five-fold from the protocol for HA-1, using 0.075 M (NH4)2HPO4 and 0.25 M
Ca(NO3)2.4H2O solutions. This resulted in the synthesis of sample HA-9.
Subsequent to this, a question exists: what morphological effect is observed if
the pH of the downflow was increased to 10 at these increased concentrations?
To solve this, the precursor concentrations described for HA-9 were used, while
the downflow pH was adjusted to 10 using NH4OH - yielding sample HA-10. This
leads to another question – what happens to product morphology if, at these
increased concentrations, the downflow pH is adjust using NH 4OH and the
system temperature is increased to 350 °C? These parameters were hence
employed in the synthesis of sample HA-11.
Finally, to address a change in ratios of precursor concentrations, the Ca:P ratio
was changed from the stoichiometric to 2. This ratio was chosen, not only
because it is a simple ratio, but with a secondary motive of producing
tetracalcium phosphate (TTCP). With a formula of Ca4(PO4)2O, TTCP has been
shown to improve resorption of bone scaffolds and cements (the process
whereby a scaffold is broken down by osteoclasts to release minerals for the
formation of new bone), making it more beneficial to the patient in the long term
(Moseke and Gbureck, 2010).
For this experiment, reactant solutions were prepared as 0.0125 M (NH4)2HPO4
in the downflow, and 0.05 M Ca(NO3)2.4H2O in the upflow. Precursor pH, flow
rates, system temperatures and pressure were maintained as for HA-1. The
yielded sample was called HA-12.
92
The HA samples described in this Section, and their corresponding reaction
conditions, are summarised in Table 3.3. All samples were white in colour and
the particles settled over time. SEM and XRD analysis was conducted for each
sample.
3.2.1.4 Incorporation of Simvastatin into HA
As described in Section 3.1.4.5, an aim of this sub-project is to produce
HA nanocomposites containing drugs or cations which will enhance its properties.
Also stated previously in Section 3.1.4.5, statins are a group of drugs which,
although primarily used to tackle high blood pressure, have been shown to
exhibit secondary osteogenic effects. In the literature, simvastatin is described
as an example of such a statin (Griffiths and Cartmell, 2007). For this reason,
and because a supply of this drug can be readily obtained commercially,
simvastatin was chosen as a drug candidate for incorporation into HA. It is
important to note that the chemical structure of simvastatin (depicted in Figure
3.2.) can degrade under high temperature or high pH (Park, 2009) – therefore,
it will be used in the manufacture of HA platelets (HA-1), rather than other
morphologies.
93
Figure 3.2. Chemical structure of (a) Simvastatin in its inactive lactone
form, and (b) Simvastatin in its active hydroxyl acid form.
The simvastatin was activated using the procedure and concentrations described
by Griffiths and Cartmell (2007), transforming it from its inactive lactone form
(Figure 3.2. (a)) to its active hydroxyl acid form (Figure 3.2. (b)). Following this,
a 5 µM solution of the statin was added to a prepared solution of Ca(NO3)2.4H2O
and the experiment was conducted as described for sample HA-1 in Section
3.2.1.1. The sample was called HA-SIM.
3.2.1.5 Incorporation of Zinc into HA
Of the potential cations to be used as a dopant, zinc was chosen due to
its availability, its low cost and because it has previously been reported as a
dopant into HA by other research groups (Webster et al., 2004, De Lima et al.,
2010, Stanić et al., 2010). For this synthesis, a 0.05 M Ca(NO3)2.4H2O solution
containing 0.0046 mol Zn(NO3)2. 6H2O (equivalent to 15 w/w% Zn) was prepared
and the experiment carried out as described for HA-1 in Section 3.2.1.1. The
sample will be referred to as HA-Zn.
94
All the aforementioned samples were white products, where particles flocculated
and settled to the bottom of the vessel. By eye, all the products looked the
same. A summary of the reaction conditions for each sample is shown in Table
3.3.
95
Table 3.3. A summary of experimental details for synthesis of various HA samples.
(NH4)2HPO4 (NH4)2HPO4 Ca(NO3)2.4H2O Ca(NO3)2.4H2O Reagent for Reaction Particle

Sample Approximate Particle Size
Concentration pH Concentration pH pH alteration Temp. (˚C) Morphology
HA-1 0.015 M 8 0.05 M 5 - 200 Sheet <2 µm width <80 nm thick
HA-2 0.015 M 10 0.05 M 5 NH4OH 200 Rod Diameter: 30–40 nm
HA-3 0.015 M 8 0.05 M 10 NH4OH 200 Sheet/rod Diameter: ~80 nm
HA-4a 0.015 M 10 0.05 M 5 NaOH 200 Rod Diameter: 20 nm
HA-4b 0.015 M 10 0.05 M 5 NaOH 250 Rod Diameter: 20-70 nm
HA-4c 0.015 M 10 0.05 M 5 NaOH 300 Rod Diameter: 25 - 140 nm
HA-4d 0.015 M 10 0.05 M 5 NaOH 350 Rod Diameter: 30 - 120 nm
Low aspect Diameter: <25 nm

HA-4e 0.015 M 10 0.05 M 5 NaOH 400
Ratio Rods Length: < 110 nm
Low aspect Diameter: <20 nm

HA-4f 0.015 M 10 0.05 M 5 NaOH 450
Ratio Rods Length: <100 nm
Outer Diameter: <0.5 µm

HA-5 0.015 M 10 0.05 M 5 NH4OH 350 Tubes
Inner Diameter: 30-70 nm
Outer diameter: 100-300

HA-6 0.015 M 8 0.05 M 10 NH4OH 350 Rod/tube
nm
96
HA-7 0.015 M 8 0.05 M 5 - 350 Sheet/Rod Diameter: 80–300 nm
<2 µm width <100 nm

HA-8 0.03 M 8 0.1 M 5 - 200 Sheet
thick
HA-9 0.075 M 8 0.25 M 5 - 200 Sheet/rod Diameter: <100 nm

HA-10 0.075 M 10 0.25 M 5 NH4OH 200 Sheet
thick
HA-11 0.075 M 10 0.25 M 5 NH4OH 350 Sheet/rod Diameter: 30-40 nm

HA-12 0.0125 M 8 0.05 M 5 - 200 Sheet
thick

HA-SIMa 0.015 M 8 0.05 M 5 - 200 Sheet
thick

HA-Znb 0.015 M 8 0.05 M 5 - 200 Sheet
thick
a
5 µmol of activated Simvastatin was added to the Ca(NO3)2.4H2O solution prior to synthesis.
b
15 w/w% Zn(NO3)2.6H2O was added to the Ca(NO3)2.4H2O solution prior to synthesis.
97
3.2.2 Characterisation Methods
All samples were washed by centrifuging, decanting the supernatant and
re-suspending in deionised water. They were subsequently freeze dried for
further characterisation, namely SEM and XRD. Where required, HRTEM was also
utilised. Particle size analysis was performed using Digital Micrograph ™ software
(from Gatan Inc., USA). To assess the presence of simvastatin in HA-SIM, ToF-
SIMS was used.
3.2.3 Material Application Methods
An important part of the validation work for this new approach to nano-
HA manufacture has been cell proliferation assays. This is to demonstrate that
HA produced in this reactor system exhibits no cytotoxic effects on mouse
embryonic stem cells (mESC). This cell type is regularly used for such
cytotoxicity studies. Additionally, bone nodule assays were conducted to
highlight that HA synthesised in this work can promote stem cell differentiation
into an osteoblast lineage.
3.2.3.1 Cell Proliferation Assay
These assays were used to ascertain if HA produced in the continuous
hydrothermal reactor system exhibit a cytotoxic effect, rather than determine
cell response to every sample. Therefore, only nanoparticles from sample HA-1,
HA-SIM and HA-Zn were used for this study. The samples were firstly sterilised
by exposure to UV light. The mESC were maintained in either standard culture
media or osteogenic culture media at 37˚C, 5% CO2 at 85% relative humidity.
Standard culture media contained DMEM growth media, supplemented with 10%
Fetal Bovine Serum (FBS), 1% Penicillin/Streptomycin 100x solution, 1% L-
98
Glutamine and 100 µM 2-Mercaptoethanol. Osteogenic media contained standard
culture media and additional supplements (0.284 mM L-ascorbic acid, 50 mM β-
glycerophosphate) which guide stem cells towards differentiation into an
osteoblast lineage.
A 96 well plate was seeded with mESCs (7000 cells/well) and exposed to
nanoparticles at one of three different concentrations (1, 10 or 100 µg/ml). A
control sample was kept with no nanoparticles added. Cell proliferation was
monitored after 1, 3 and 6 days of culture using a 96® AQueous One Solution
Cell Proliferation Assay (Promega, USA). The assay principle is based on the
MTS tetrazolium compound being bioreduced into a coloured formazan product
that is soluble in culture media. This conversion is accomplished by NADPH or
NADH produced by dehydrogenase enzymes in metabolic active cells.
3.2.3.2 Bone Nodule Assay
To study osteogenic differentiation and bone mineralisation, cells were
cultured in 12 well plates (100,000 cells/well) for 21 days in the presence of
nanoparticles at one of three concentrations (1, 10 and 100 µg/ml), as well as a
control. After this time, cells were fixed in 10% formal saline solution for 20
minutes followed by a histological staining with 1% Alizarin red solution for 5
minutes. The cells were washed three times with deionised water to reduce the
background intensity. Clusters which appeared stained with a red-brown colour
were indicative of mineralised nodules.
99
3.3 Results and Discussion
3.3.1 HA Platelets (HA-1) and Rods (HA-2)
The synthesis of HA-1 and HA-2 were primarily to act as confirmatory
work on the previously described results within the group. Indeed platelet
morphology was observed in HA-1, as shown from the SEM and TEM images in
Figure 3.3. The particles have widths and lengths measuring up to 2 µm while
the thickness of the platelets is <80 nm.
Figure 3.3. (a) SEM and (b) TEM images of particles in sample HA-1.
XRD analysis performed on this sample showed that the only crystalline phase
present was hexagonal hydroxyapatite, matching to ICCD PDF 74-566 (Figure
3.10). While the peaks match to the expected pattern, they are not sharp; this
may indicate that the material has low crystallinity, assuming the particle size is
large (See Chapter 2, Section 2.4 for reasoning).
100
The SEM and TEM images in Figure 3.4 show the nanorod morphology identified
in HA-2, again confirming the results previously seen in the group. These
particles have diameters between 30 and 40 nm and lengths measuring up to 1
µm, and demonstrate that by merely altering the pH of one of the precursors, a
different product morphology is observed. The XRD pattern for HA-2 is displayed
in Figure 3.10 and also matches to the expected pattern for hexagonal HA. The
peaks are sharper that those seen for HA-1, suggesting that the particles in HA-2
are more crystalline. These data show that as well as changing particle
morphology, the pH (or indeed the addition of NH4OH itself) is leading to the
formation of a more crystalline product – a process which has been described
before for HA synthesis. Several methods of HA synthesis use NH 4OH to increase
the reaction pH above 9 in order to drive HA formation (Zhang et al., 2009a, Li
et al., 2007). This is because acidic by-products are usually produced during HA
synthesis; the acid can degrade the HA, and compromise the crystallinity. In this
case, a by-product of the synthesis is nitric acid, HNO3, because of the use of
Ca(NO3)2 as a precursor. Without any post-synthesis washing of HA-1, the HNO3
is likely to decompose the HA. For HA-2 however, the addition of NH4OH
neutralises the acid formed, which may lead to higher crystallinity. It is
important to note that it is apparent NH4OH is also playing a role in the
mechanism of synthesis, leading to this change in morphology.
101
Figure 3.4. (a) SEM image and (b) TEM image of particles in HA-2.
3.3.2 The Effect of Precursor pH
It was found that the pH of either precursor solution can be increased to
yield nanorods, not just the downflow (NH4)2HPO4. Figure 3.5 shows that some
rod shaped particles resulted when the Ca(NO3)2 precursor solution was adjusted
to pH 10. However, there was still evidence of some platelet-type particles being
produced, and the rods appear less well defined compared to those in sample
HA-2. The rods in HA-3 were apparently larger than in HA-2 as well, with
diameters in the region of 80 nm.
102
Figure 3.5. SEM image of HA-3.
Samples HA-4a to 4f were produced using NaOH as a pH buffer, as opposed to
NH4OH, over a range of reaction temperatures. This was to investigate if
morphology changes were dependent on pH alone, or were base-specific. The
TEM image in Figure 3.6 (a) shows the presence of rod shaped particles in HA-4a
(reaction temperature ~200 °C); from this it can be inferred that rod
morphology is a result of pH changes, but is not reliant on NH 4OH being used.
Nevertheless, the size of the particles is not identical to those in HA-2. The rods
in HA-4a have diameters measuring ~20 nm – smaller than those seen in HA-2.
The XRD data for HA-4a (shown in Figure 3.10) confirms that HA is the only
crystalline phase present, but the peaks are not as sharp compared to HA-2.
Samples HA-4b to 4f use the same precursors as HA-4a, but reflect increases in
reaction temperature. The TEM images in Figure 3.6 indicate that reaction
temperature does affect particle size and morphology. It can be observed that
with increases in temperature up to 350 °C (HA-4b to 4d), the particle diameter
103
also increases; in HA-4b (reaction temperature ~250 °C) rod diameters measure
between 20 and 80 nm, in HA-4c (reaction temperature ~300 °C) these are 25-
140 nm, and in HA-4d (reaction temperature ~350 °C), these are 30-140 nm. In
the latter two samples, some particles may even be classed as having platelet
morphology rather than rod. It is possible that this trend is due to the increased
thermal energy causing growth of the existing particles, rather than nucleation of
new ones.
Interestingly at even higher reaction temperatures, as is the case with HA-4e
(reaction temperature ~400 °C) and 4f (reaction temperature ~450 °C), the
particles appear to be smaller than in the aforementioned samples, and the
morphology can be described as low-aspect ratio rods. In HA-4e, the rods have
diameters up to 25 nm and lengths up to 110 nm – aspect ratios between 3.69
and 6.22. In HA-4f, the diameters typically observed are no more than 20 nm,
and lengths measure up to 100 nm – giving aspect ratios ca. 5.3. This is
compared to the much higher aspect ratio rods seen in HA-4a (aspect ratio ca.
16.9), HA-4b (ca. 10.9), HA-4c (ca. 10.1), and HA-4d (ca. 9.1). This decrease in
aspect ratio with increasing reaction temperature relates to the increase in
particle diameters mentioned previously. XRD patterns were only obtained for
samples HA-4a and HA-4d; these should be representative of the sample series,
confirming that HA is the sole crystalline phase, rather than looking at specific
changes to crystal structure between the samples. Moreover, HA-4d was chosen
for XRD analysis as a comparison to sample HA-5 because a reaction
temperature of 350 °C is common to both samples.
104
Figure 3.6. HRTEM images of (a) HA-4a; (b) HA-4b; (c) HA-4c; (d) HA-
4d; (e) HA-4e; and (f) HA-4f.
HA-5 is a very interesting sample; collected SEM images (shown in Figure 3.7
(a)) firstly showed ‘open-ended’ rods. Using HRTEM, and darkfield imaging,
these particles were shown to have a tube morphology (Figure 3.7 (b) and (c)).
In these micrographs, the thicker walls are shown darker than the hollow centre
105
(or brighter in the case of darkfield imaging). This morphology has, to date, not
been previously reported – making it a significant finding.
Figure 3.7. (a) SEM image of ‘open-ended rods’, (b) darkfield imaging in
TEM and (c) TEM image show tube structures while (d) is an SEM image
showing a tube in the middle of the rolling process (highlighted with a
red arrow).
There are several potential applications for nanotubes of HA, which could result
from the ability to fill the tubes with drugs or other nanomaterials. For example,
these tubes could act as slow-release drug carriers in bone scaffolds or other
drug delivery systems (Yang et al., 2012b, Netz et al., 2001). To begin this sort
of research, it is important to understand the formation mechanism of these
106
particles. Examining the morphology; in Figure 3.7 (c), the tube looks darker in
the centre, suggesting that it is not completely hollow. By considering that TEM
essentially shows a projection of an image, it is entirely feasible that the tube
has formed from the scrolling of a sheet (Figure 3.8), to yield the image in
Figure 3.7 (c). To reinforce this notion, Figure 3.7 (d) is an SEM image
apparently showing a particle in the middle of the scrolling process. The TEM
images obtained for HA-5 showed these tubes to have an inner diameter or pore
size between 30 and 70 nm, while the outer diameter ranged up to 0.5 µm.
Therefore, by crude estimation, a sheet or platelet particle with a width of 600
nm would roll up with 3-4 layers to give a 50 nm pore. Some initial N2
adsorption data was collected for HA-5, but proved inconclusive. This may be
due to the scrolls opening up during the measurement as a result of changes to
the surface charge (Lester et al., 2013).
As discussed previously, only rod structures, not tubes, are seen in HA-4d, when
NaOH is used as a pH buffer, demonstrating that unlike rod formation, the
synthesis of HA tubes is base-dependent.
Figure 3.8. Images illustrating the scrolling of a sheet, so that the
corners meet.
107
It was found that by adjusting the pH of the upflow precursor rather than the
downflow, while keeping the reaction temperature at 350 °C (as is the case for
HA-6), tube structures are again synthesised. The SEM image in Figure 3.9 (a)
illustrates this. The inner and outer diameters of the particles are not
significantly different from those seen in HA-5; however, the particles appear to
be less well-formed with jagged edges apparent. As a result, it can be said that
addition of NH4OH to the downflow is more effective for the synthesis of tubes,
than in the upflow.
Sample HA-7 was shown to contain a mixture of rods and high aspect ratio
platelets, as displayed in Figure 3.9 (b). In this case, the precursors had not
been treated with any pH alterations, and underwent a reaction temperature of
350 °C. The particles appear to have diameters between 80 and 300 nm, and
variable lengths up to a few hundred nanometers. One possible reason for seeing
these particles with higher aspect ratios is that the higher reaction temperature
causes the fragmentation of sheet particles (evident from sample HA-1) into rods
and smaller platelets. However, the XRD data in Figure 3.10 show that HA-7
produced more intense peaks than HA-1, indicating a higher level of crystallinity
with increased reaction temperature. As such, it is more likely that the difference
lies in the formation of particles, i.e. these rods and platelets are formed via a
different mechanism, in which growth to larger sheets is inhibited, rather than
fragmentation.
108
Figure 3.9. SEM images of (a) HA-6, and (b) HA-7.
Figure 3.10. XRD patterns for samples HA-1 to HA-7. The expected peak
pattern for hexagonal hydroxyapatite (ICCD PDF 74-566) is shown at
the bottom.
109
3.3.3 The Effect of Precursor Concentration
Now focussing on the effects of modifying precursor concentration;
sample HA-8 was the result of doubling precursor concentrations from those
used for HA-1 (i.e. using 0.03 M (NH4)2HPO4 and 0.1 M Ca(NO3)2). The principle
test was to determine if HA remains the sole crystalline phase obtained. The XRD
pattern shown in Figure 3.13 indicates that it is. Nevertheless, the pattern does
not show sharp peaks, so it can be inferred that, like HA-1, the material is not
very crystalline (if particle size is large). This corresponds with the reasons
highlighted earlier in the discussion for HA-1. The obtained SEM images (seen in
Figure 3.11 (a)) present particles with a very similar morphology and size as
those in HA-1. Again, these sheets have lengths and widths up to 2 µm, and
thickness of less than 100 nm.
However, it was found that when the precursor concentration is increased
fivefold that of HA-1, HA is no longer the sole crystalline phase produced. The
XRD pattern contains sharp peaks which correspond to monoclinic brushite,
CaHPO4.2H2O (ICCD PDF 9-77) and some peaks which can be assigned to
monetite, CaHPO4. These minerals are sometimes used as precursors for apatite
formation (Monma and Kamiya, 1987) suggesting that these may be
“intermediates” in the reaction pathway to HA formation during hydrothermal
synthesis. The particles in this sample are a mixture of sheets and rods, the
latter with diameters less than 100 nm. The morphology differences in this
sample cannot be fairly compared to HA-1 and HA-2 because of the different
phases that are present in the sample and it is difficult to distinguish between
the two phases using SEM and TEM analysis.
Interestingly, it was found that by using five times the “standard” precursor
concentration, increasing the downflow pH to 10 using NH4OH, and a reaction
110
temperature of 200 °C (as is the case for HA-10), the sole crystalline phase is
once again HA. As Figure 3.13 shows, peaks for brushite or monetite are no
longer evident. Furthermore, the SEM image displayed in Figure 3.11 (c) shows
the sheet like particles present in HA-10 which look very similar to the particles
in HA-1. In addition, pumping these same precursors at a higher temperature of
350 °C (sample HA-11) produces rods and high aspect ratio platelets (Figure
3.11 (d)). It can be inferred from these results that NH 4OH is playing a wider
role in the formation of these particles. This evidence indicates NH 4OH directs
the reaction towards HA formation over other calcium phosphates; it increases
the crystallinity of the product; and the particle morphology changes.
Figure 3.11. SEM images of (a) HA-8, (b) HA-9, (c) HA-10, (d) HA-11.
111
With the aim of synthesising TTCP, HA-12 was produced in which the precursor
concentrations were altered from the stoichiometric ratio; instead, a Ca:P ratio
of 2 was used. However, the XRD pattern collected for this sample (see Figure
3.13) showed that the only crystalline phase present was HA. Additionally, it was
found that the particles were similar to those seen in HA-1, both in morphology
and size (as seen from the SEM image in Figure 3.12). This data suggests that
HA is preferentially produced in the Nozzle reactor, over other calcium
phosphates, when precursor concentrations do not significantly deviate from the
“standard”. Furthermore, it shows that this method of HA synthesis is not
sensitive to changes in the ratio of precursor concentrations, and the precise
stoichiometric ratio does not need to be maintained to produce HA.
Figure 3.12. SEM image of HA-12.
112
Figure 3.13. XRD patterns for HA-8 to HA-Zn. Patterns for HA-1 and HA-
2 are also shown for comparison. The expected peak pattern for
hexagonal hydroxyapatite (ICCD PDF 74-566) is shown at the bottom.
For HA-9, peaks indicated with a black square match monoclinic brushite
(ICCD PDF 9-77) while other peaks match monetite; for HA-Zn, peaks
labelled with black triangles correspond to monoclinic parascholzite
(ICCD PDF 35-495).
3.3.4 Addition of Simvastatin
The SEM and TEM images obtained for HA-SIM (Figure 3.14) showed
particle size and morphology similar to HA-1. Given the very low concentration of
simvastatin added to the sample, XRD analysis was not performed in HA-SIM to
confirm the crystalline phase of the material. Instead, the aim here was to
assess the presence of simvastatin in the sample. Initially, Fourier-Transform
Infra-Red (FTIR) spectroscopy was performed on HA-SIM and HA-1 and the two
113
spectra were compared. However, there was no observed difference between the
two, so it could not be concluded that the drug was present in HA-SIM. It is
possible the concentration of the statin was too low in the sample, and below the
sensitivity threshold of the FTIR spectrometer, meaning that if present, the drug
would not be detected. As a result, the sample was also analysed using ToF-
SIMS. The obtained data (Figure 3.15) showed that a fragment with a
mass/charge value of 115.08 was present in HA-SIM, but not in HA-1 or HA-2.
This fragment can be assigned to the molecule C 6H12O2, which forms part of
simvastatin. However, this was the only difference observed in HA-SIM – which
again, may be due to the low concentration of simvastatin used in the precursor.
It should be noted that ToF-SIMS operates by bombarding the sample to create
fragments which can then be detected; hence, it is difficult to ascertain from this
data whether the simvastatin has been preserved during the synthesis reaction
and is present as a whole molecule in the sample, or if the heat and pressure of
the reaction has caused the breakdown of the drug and what is seen is actually a
degraded portion of the statin. Ultimately, it is inconclusive whether simvastatin
is wholly present in HA-SIM. An in vitro cell proliferation was also carried out
using HA-SIM. The results will be discussed in Section 3.3.6.
Figure 3.14. (a) SEM Image and (b) TEM image of HA-SIM.
114
Figure 3.15. ToF-SIMS Data for HA-1, HA-2 and HA-SIM.
3.3.5 Addition of Zinc
HA-Zn was created with the aim of introducing zinc as a dopant into HA.
The obtained TEM images indicated the particles had a sheet morphology, which
was expected for the experimental parameters used. An elemental distribution
map was also collected for the sample, during TEM analysis. The data (Figure
3.16) confirmed that zinc was indeed present in the particles, as well as the
expected elemental components of HA. Nevertheless, the concentration of zinc
present appeared to be higher than that of calcium - seen from the larger
115
number of white spots in Figure 3.16 (e) than (b). This is despite only 15 w/w %
of Zn was used compared to Ca; this indicates that Zn may be preferentially
taken up during particle formation. Moreover, the XRD pattern of HA-Zn (shown
in Figure 3.13 and more clearly in Figure 3.17) showed that, while peaks
corresponding to HA were present, there were also quite a few other peaks
detected. The majority of these were ascribed to monoclinic paracholzite,
CaZn2(PO4)2.2H2O, a mineral first described by (Sturman et al., 1981). This
confirms that the calcium to zinc ratio is consistent with the elemental map data.
Since it was first reported, parascholzite has rarely been discussed in literature,
and synthesis details of this material are hard to come by. Still, this material
may have applications in the incorporation of zinc into bone-based materials
(Sepulveda et al., 2000, Dasgupta et al., 2010) and the biocompatibility of HA-
Zn is an important factor. Consequently, the in vitro cell proliferation assay
using HA-Zn is reported in Section 3.3.6.
Figure 3.16. (a) TEM image of sample HA–Zn, and an elemental
distribution map for (b) Ca, (c) O, (d) P, and (e) Zn.
116
Figure 3.17. XRD pattern of HA-Zn; the predominant phases are
hexagonal hydroxyapatite (ICCD PDF 74-566, shown at the bottom) and
monoclinic parascholzite (ICCD PDF 35-495, peaks shown with blue
triangles). Additional smaller peaks match calcium hydrogen phosphate
(ICCD PDF 77-128).
3.3.6 In Vitro Cell Assays
The results from the cell proliferation assay (Figure 3.18) show that for
the duration of the experiment, cell activity (absorbance correlates to cell
activity) continued to increase for all three HA samples, irrespective of
nanoparticle concentration. After 6 days, the cells cultured in control media are
seen to have lower cellular metabolic activity when exposed to higher
concentrations of nanoparticles, relative to the control. It is important to note
that these effects are likely to be due to HA exhibiting an osteogenic effect on
the cells, directing them towards an osteogenic phenotype - typically associated
117
with a decrease in proliferation. This is because once mESCs have differentiated
into a cell line, they stop multiplying. In addition, these effects may be
influenced by confluency initiated cell death (i.e. too many cells competing for
resources) rather than the cytotoxic properties of the nanoparticles, the latter of
which would be evident much earlier in the experiment.
Figure 3.18. Results from the cell proliferation assay performed using
three different concentrations of nanoparticles from sample HA-1 and
sample HA-Zn, after 1, 3 and 6 days of incubation.
118
The bone nodule assay gives an indication of bone mineralisation - a distinctive
marker of osteogenesis. The data (Figure 3.19) shows that the concentration of
nanoparticles added to the media plays an important role in bone nodule
formation. At higher nanoparticle concentrations, more intensive staining results
and bone nodules can be seen. The frequency of intensely stained clusters also
increases with higher concentrations of nanoparticles, indicating the formation of
the extracellular matrix as cells undergo differentiation into the osteoblast
lineage. The risk of calcium from HA being stained was minimised by excessive
washing steps to remove residual nanoparticles prior to staining.
Figure 3.19. Images from the bone nodule assay - Cells cultured in
osteogenic media with hydroxyapatite from Sample HA-1 at various
concentrations, (a) 100 µg/ml; (b) 10 µg/ml; (c) 1 µg/ml; (d) control
119
3.4 Conclusions
Firstly, the work in this Chapter has shown that the counter-current
Nozzle reactor can be used to reproducibly synthesise HA with three different
morphologies: platelets, rods and tubes. The synthesis method is also
significantly faster than other techniques previously reported in literature.
Results of this study suggest that in this hydrothermal system, NH 4OH plays
several roles in HA synthesis – it directs precursors towards the formation of HA
over other CaP minerals, promotes the formation of rods instead of sheets at
200˚C, and tubes instead of rods at 350˚C. This process is more effective when
NH4OH is added to the heated downflow, compared to the cooler upflow.
In a continuous system, the process variables are linked thus making it very
difficult to alter one parameter without affecting the others (Lester et al., 2012).
This creates difficulty in isolating a single cause and effect, thereby making it a
challenge to hypothesise a mechanism of formation for the different HA
morphologies. Still, it is likely that HA sheets are the fundamental morphology of
choice under ‘standard’ conditions. By increasing either the reaction temperature
or pH, the synthesised particles are not only more crystalline, but given the
observed dimensions of the particles, the sheets become fragmented into
nanorods. However, the combination of a higher reaction temperature and the
presence of NH4OH specifically, appear to prevent the fragmentation of sheets,
but cause the sheets to roll up - giving nanotubes. Further work is undoubtedly
required to understand the formation mechanism of the different morphologies
but it is apparent that the mechanisms are linked, rather than three independent
schemes existing.
120
Experiments were conducted to introduce zinc as a dopant and simvastatin as a
drug, into the structure of HA. The morphology of particles synthesised with
simvastatin did not significantly differ from normal HA particles – again
producing sheets. ToF-SIMS data for this sample showed the presence of a
fragment which could be assigned to simvastatin, but it is inconclusive. This is
likely to be due to the low concentration of statin used during synthesis. For
future work, the concentration should be increased, although this does implicate
higher costs. In addition, thermal analysis such as TGA (thermal gravimetric
analysis) could provide additional data to ascertain the presence of simvastatin
in the HA sample.
The addition of a zinc salt into the synthesis protocol yielded parascholzite,
CaZn2(PO4)2.2H2O, a material rarely described in literature but may have
applications in bone based materials. In the scope of this work, experiments with
only one concentration of Zn were attempted. For future work, the ratio of Ca:Zn
can be investigated further, to determine if other calcium-zinc phosphates can
be produced using the Nozzle reactor.
Finally, cell proliferation and bone nodule assays were conducted using HA
produced using the Nozzle reactor. Results indicate that the HA particles have no
cytotoxic effect on mESCs and promote bone nodule formation.
121
Chapter 4
4 Fluorescent Nanomaterials
4.1 Introduction
Following a previously reported hydrothermal method for the synthesis of
cadmium sulphide, CdS, (Liu et al., 2001) the Nottingham research group
established a synthesis procedure for the manufacture of CdS particles using the
hydrothermal Nozzle reactor. Other publications using the Nozzle reactor
included a method for synthesising nanoparticles of zirconium dioxide, ZrO2,
doped with europium, Eu (Hobbs et al., 2009). These particles displayed
fluorescent properties while exhibiting no cytotoxic effects on Chinese hamster
ovary cells; thus showing potential for use in biological cell imaging.
Consequently, the motivation in this Chapter is to explore what other fluorescent
materials are possible.
4.1.1 Fundamentals of Fluorescence
Fluorescence describes the emission of light by particles, upon stimulation
from a form of electromagnetic radiation. There are different categories of
fluorescence depending on the source of radiation.
When a molecule absorbs light photons from the UV light range (10 to 400 nm),
it reaches an excited state with a higher energy level. Hence, this is known as
excitation. Still, the molecule is unstable at this higher energy level and swiftly
122
Chapter 4. Fluorescent Nanomaterials
emits light photons in order to return to its previous, more stable ground state.
The emitted energy tends to be less than is absorbed by the molecule; as this
energy is released in the form of light, the emitted light has a longer wavelength
than the light that is absorbed (Geszke-Moritz and Moritz, 2013). Molecules
which exhibit this behaviour are defined as photoluminescent (PL). Typically, PL
particles emit light within the visible light range (400 to 700 nm); these are
useful as they can be seen by the human eye without the need for additional
detection apparatus. Nevertheless, there are particles which have been tuned to
emit light in the infra-red (>700 nm) range (Hu et al., 2011, Xue et al., 2012,
Poole et al., 2009).
Rather than reaching an excited state through light energy, particles stimulated
by an electrical current are electroluminescent (EL), by a chemical reaction are
chemiluminescent (CL), by sound are sonoluminescent and by a mechanical
action are mechanoluminescent (ML) (Ronda, 2008). While electroluminescence
will be briefly mentioned in subsequent Sections of this Chapter,
photoluminescent properties of particles will remain the primary focus.
4.1.2 Quantum Dots
Quantum dots (QD) are a subset of fluorescent materials which are
typically 1 to 20 nm in size (Geszke-Moritz and Moritz, 2013). They can be
defined as crystalline semiconductor nanoparticles made up of elements from
transition metals and group VI of the periodic table, such as CdSe or ZnS, or
groups III and V, such as InP (Byers and Hitchman, 2011). QDs have gained
much interest because of the potential to tune the PL properties through size
manipulation. It has been reported that the relative brightness of QDs is
dependent on the diameter of the particles and, in turn, the emission wavelength
(Dickerson, 2005). For example, the intensity or brightness of QDs emitting at
123
525 nm (green) was 17 times lower than QDs emitting at 655 nm (red) (Xing et
al., 2007).
Commonly reported QDs are comprised of a heavy metal core, usually cadmium
sulphide (CdS) or cadmium selenide (CdSe), with a zinc sulphide (ZnS) shell to
increase quantum yield (brightness) and avoid oxidation, as well as prevent Cd
from leaching out (Byers and Hitchman, 2011). Polymer coatings and
biomolecules can be attached to the surface to improve particle stability or
solubility, increase biocompatibility, or target specific sites in a biological system
(Jamieson et al., 2007). Section 4.1.3.4 will discuss biological applications of
QDs in further detail. Figure 4.1 illustrates the structure of one example of a QD.
Figure 4.1. Illustration of a typical QD structure; a CdSe core is
surrounded by a ZnS shell and functionalised with siloxane and
methylmethacrylate (MMA) (Park et al., 2010).
124
4.1.3 Applications using Quantum Dots
Nanoparticles with fluorescent properties, particularly QDs, are of great
interest because they can be applied in a wide range of fields - from electronics
to medicine. Some of these applications will be briefly explained in this Section.
4.1.3.1 Electronic Devices
As QDs are semi-conductors, they are often reported for use as
components in electronic devices. For example, SiO2 QDs have been applied for
memory storage devices (Corso et al., 2003), while PbS QDs have been used in
fibre-optics amplifiers (Heo and Liu, 2007). Meanwhile, particles with EL
properties have also found uses in this field. For example, ZnS–based QDs have
been described in the field of printed electronics (Wood et al., 2009); in this
study, an inkjet printer was used to deposit nanoparticles in a structured pattern
onto substrates for use in display panels. Furthermore, InP/ZnS composite QDs
have reportedly been used in light emitting diodes (LEDs) (Kim et al., 2012),
while Poole et al. described the use of InAs/InP QDs in lasers (Poole et al.,
2009).
4.1.3.2 Solar Panels
Increasing population and technological advances have resulted in a rise
in the global demand of energy. Conversely, a limited supply of fossil fuels (coal,
oil, natural gas) has meant sharp increases in energy prices. Twinned with the
emission of greenhouse gases from burning fossil fuels, there has been a
growing need to focus efforts in developing other more sustainable and
renewable energy sources - including wind, bio-fuels, geothermal and solar cells.
125
The latter has generated a lot of interest as the Earth receives over 120 PW
(petawatts) of energy from the Sun each day (Tsakalakos, 2008). As such, many
researchers have developed various materials, including QDs, to capture solar
energy. Much of the initial work in this area was conducted on silicon, but the
need for more efficient materials lead to advances with other elements, namely
Gallium based QDs (Tsakalakos, 2008) and CdS QDs (Santra and Kamat, 2012)
being used.
4.1.3.3 Anti-Counterfeiting
QDs and other fluorescent materials have been considered for anti-
counterfeiting purposes (Zhang and Ge, 2011), where genuine articles are
marked with the fluorescent particles and are invisible to the naked eye. Only
upon excitation with light of a particular wavelength can the distinctive mark be
seen, thereby distinguishing fraudulent copies from the real article.
4.1.3.4 Biological Applications
In 1998, researchers first began to explore using QDs in the field of bio-
imaging (Bruchez Jr et al., 1998, Chan and Nie, 1998). In this ground-breaking
study, proteins in mice were attached to, or labelled with, water-dispersible
fluorescent QDs which enabled them to be detected for the purpose of medical
diagnostics. The semiconductor QDs proved to be significantly more effective
than traditional organic dyes. Since then QDs have been widely used for a whole
host of biological applications. In cell biology, QDs have been used for labelling
and tracking cells (Jamieson et al., 2007, Byers and Hitchman, 2011). The
nanoparticles can be conjugated to antibodies which target specific proteins in a
cell. Once attached to the protein, its path can be tracked and imaged. This
126
process is particularly pertinent in the labelling of proteins specific to cancerous
cells, potentially allowing the prediction of cancer onset.
Furthermore, ZnO QDs have been described as antimicrobial agents (Jin et al.,
2009), while CdTe QDs were functionalised to bind to specific strands of DNA for
the detection of leukaemia (Sharma et al., 2012), and QDs have been
conjugated to antibodies for the detection of microorganisms in water sources
(Zhu et al., 2004). Other studies using functionalised QDs in the detection of
pathogens and toxins have also been conducted (Jamieson et al., 2007). In
terms of live in vivo imaging, studied subjects tend to be mice where cells have
been labelled to study embryogenesis, stem cell therapy and cancer metastasis
(Byers and Hitchman, 2011). Each of these applications depends on the right
biomarker or probe being chosen and attached to the QD, which can then bind to
the relevant site. As such, the QD itself is very versatile.
4.1.4 Synthesis of Quantum Dots
As described in Section 4.1.2, the fluorescent properties of QDs are
affected by their particle size. Furthermore, fluorescent properties can be
optimised if samples contain monodisperse particles with little or no
agglomeration. Hence, the route of synthesis can play a crucial role in
determining the structure and function of QDs.
Given that the umbrella term ‘quantum dot’ covers a plethora of materials,
various synthesis methods have been reported which are dependent on the
chemistry of reactions. Commonly reported methods of QD preparation include
aerosol-gel (Firmansyah et al., 2012), precipitation (Lang et al., 2011),
photochemical (Liu et al., 2012), electrochemical (Li et al., 2013), ultrasound-
127
assisted (Ren and Yan, 2012), microwave (Yang et al., 2012a), solvothermal
(Yang et al., 2011) and batch hydrothermal (Yu et al., 2012).
While these and many other described methods of QD synthesis in the literature
yield high-quality nanoparticles, they are only capable of producing milligrams of
material; consequently, there is little potential for scalability. Nevertheless, there
are processes capable of producing larger quantities of QDs and companies have
been founded on these technologies. One such example will be described in
more detail in Section 4.1.4.1.
4.1.4.1 Commercial Synthesis of Quantum Dots
Nanoco Group Plc is a UK based company established in 2001, which
produces QDs on a commercial scale (approx. 25 kg/annum). They produce both
QDs which contain heavy metals, and those which are free of heavy metals.
Their technology is based on the use of a “molecular seed” which acts as a
nucleation point on which nanoparticle growth can occur. This patented
technology produces batches of QDs, up to 1 kg each batch (Nanoco, 2013).
Because QDs can be so effective, only very small volumes of material are
required for application – as such, 1 kg of material is considered as commercial-
scale.
Nanoco produce QDs with an array of different compounds, and for a range of
applications. Their range of Cadmium Free Quantum Dots (CFQD™) have been
advertised for use as electroluminescent LEDs in lighting, in electronic displays,
and for biological applications such as drug-delivery, cell tracking and clinical
whole-body imaging. Their more recently developed range of CIS, CGS and CIGS
128
QDs are composed of varying elemental ratios of copper, indium, gallium and
selenium. These particles have been developed in a range of solvents which can
then be printed onto substrates then cured as components for solar cells to
capture light energy.
The successes of commercial manufacturers of QDs such as Nanoco highlight the
existing high demand for nanomaterials with PL and EL properties.
4.1.5 Disadvantages of Cadmium-Based Materials
The majority of QDs which are described in the literature contain Cd-
based particles, or other heavy metals such as lead, in the core. However, these
can have highly toxic effects; in humans, Cd can accumulate in the lungs, liver
and kidneys. As a result, essential proteins are excreted from the body while the
metal itself is not so easily removed. The consequences are very serious, and
even fatal.
Meanwhile, there has been much debate over the safety of using QDs containing
heavy metals in biological applications, as the toxic cores tend to be coated with
a several layers which prevent the heavy metal leaching out. Although much
research has gone into the use of QDs for in vitro and in vivo cell imaging,
toxicity studies of these materials are still lacking. Nevertheless, a publication
has shown Cd-based QDs to cause apoptosis (programmed cell death) (Chen et
al., 2012). Consequently, a Cd-free route would have more potential, especially
if the PL properties are equalled. Typically, CdS particles produced with a ZnS
shell exhibit peak PL emission at 438 nm, when excited at 370 nm (Saraswathi
Amma et al., 2008).
129
In addition, the European Commission has been striving to reduce and prohibit
the use of heavy metals (these being cadmium, mercury and lead) in consumer
products in order to reduce human exposure to these toxic elements. Hence,
legislations have been imposed to restrict the use of heavy metals in electrical
and electronic items (RoHS 2002/95/EC) and in batteries (Directive
2006/66/EC). Manufacturers of household appliances (namely electronics) must
limit the concentration of heavy metals to 0.1% (1000 ppm) (Nanoco, 2013).
Furthermore, official analysis techniques have been imposed to test levels of
heavy metals in foodstuffs (Commission Regulation (EC) No 333/2007) to ensure
they do not enter the food chain. These controls indicate the severity of heavy
metal poisoning and if their use can be avoided, these risks will be bypassed.
4.1.6 Zinc Sulphide, ZnS
Several literature reports state the use of ZnS as a shell around CdS or
CdSe for the reasons outlined in Section 4.1.2. Unlike the latter two materials,
ZnS does not carry implications of toxicity. Furthermore, due to the availability
of low-cost zinc precursors, and an established method of CdS synthesis in place,
ZnS was opted for as a material of interest in this PhD. The materials properties
will be primarily assessed with the aim of application in biological fields.
4.1.7 Capping Agents and Stabilisers
Due to the very small size of QDs (<20 nm), the attractive forces on the
surface of the particles are significant and cause the particles to coagulate and
cluster, if there is nothing to prevent them from doing so (Geszke-Moritz and
Moritz, 2013). This can lead to problems if the application calls for non-
agglomerated particles, as most do. Particularly for biomedical applications,
130
particle agglomeration can mean clusters of particles no longer exhibit the
properties seen in the nano-form, and can lead to undesired cell uptake or
excretion (Jamieson et al., 2007). Moreover, it has already been described in
Section 4.1.2 that the fluorescent properties of QDs depend on the size of the
particles, and so the aggregation of these small particles results in the exhibition
of bulk-like behaviour and the loss (or reduction) of fluorescent properties
(Dickerson, 2005).
In this Chapter, three different capping agents have been chosen to stabilise ZnS
particles. They have all been reported in the literature in stabilising ZnS;
additional reasons are also described below.
Polyvinylpyrrolidone (PVP) is a low-cost, non-toxic polymer widely used for
biological applications due to its high biocompatibility (Sharma et al., 2010); 3-
Mercaptopropionic acid (3-MPA) has been described several times as a capping
agent for QDs, especially ZnS (Luan et al., 2012, Rosli et al., 2012, Azizian-
Kalandaragh and Khodayari, 2010, Geszke-Moritz and Moritz, 2013); and sodium
hydroxide (NaOH) has been reported in the synthesis of ZnS (Shen et al., 2007)
to improve the photo luminescent properties of particles. Moreover, all the above
reagents are water-soluble, meaning experiments can be conducted in a
completely hydrothermal environment without the need for organic solvents,
which are potentially detrimental to biological systems.
4.1.8 Chapter Aims
The aim of this Chapter is to use the counter-current Nozzle reactor for
the continuous synthesis of ZnS nanoparticles, and to assess their properties in
terms of crystallinity, particle size and photoluminescence. This will be achieved
by using XRD, SEM and TEM techniques, and fluorometry analysis.
131
4.2 Methodology
The reagents used for synthesis were cadmium nitrate tetrahydrate,
Cd(NO3)2.4H2O (98% purity); thiocarbohydrazide, (NH2NH)2CS (98% purity);
zinc nitrate tetrahydrate, Zn(NO3)2.6H2O (98% purity); thiourea, (NH2)2CS
(≥99% purity); 3-Mercaptopropionic acid (3-MPA), HSCH2CH2CO2H (≥99%
purity); sodium hydroxide, NaOH (≥98% purity); polyvinylpyrrolidone,
(C6H9NO)n (average molecular weight of 10,000). All reagents were purchased
from Sigma Aldrich, UK and used without further purification. DI grade water
water used throughout this work.
4.2.1 Synthesis of CdS using the Nozzle Reactor
As mentioned in Section 4.1, a method was previously established for the
continuous hydrothermal synthesis of CdS. This protocol will be elaborated on in
this Section.
The method described a 0.01 M Cd(NO3)2.4H2O solution being pumped in one
metal salt pump at 5 ml/min. Simultaneously, a 0.01 M (NH 2NH)2CS solution was
pumped at 5 ml/min through a second metal salt pump. These met at a T-piece
before the reactor and subsequently flowed into the reactor section of the
system together. Here, the feed met a stream of water pumped at 10 ml/min.
The method described the water downflow being pre-heated to 400 ˚C, and the
system pressure maintained at 240 bar.
In the work described in this Chapter, a few samples of CdS were generated
following this protocol, but employing a range of temperatures – either 240, 300
or 400 ˚C and the samples were labelled CdS 240, CdS 300 and CdS 400,
132
respectively. Sample CdS 400 was analysed by DLS and TEM. Fluorometry
analysis was also performed.
4.2.2 Initial Synthesis of ZnS using the Nozzle Reactor
For the preliminary synthesis of ZnS, the methodology for CdS synthesis
(Section 4.2.1) was adapted: quite simply Zn(NO3)2.6H2O was chosen as the
metal salt precursor, instead of Cd(NO3)2.4H2O. The precursor concentrations,
flow orientation and rates were maintained, and the system pressure was again
kept at 240 bar. A range of samples were produced using five different reaction
temperatures; these are detailed in Table 4.1. The positions in the system from
which these temperatures were taken, are outlined in Chapter 1.
Table 4.1. Reaction temperatures for each ZnS Sample.
Post
Water After Post
Mixing Capping
Sample Heater Heater Cooler
Point Point
Name Temp Temp Temp
Temp Temp (˚C)
(˚C) (˚C) (˚C)
(˚C)
ZnS 200 215 213 102 54 19
ZnS 250 265 260 120 60 19
ZnS 300 315 313 156 70 19
ZnS 350 365 363 262 80 19
ZnS 400 420 416 304 101 20
Sample ZnS 400 was analysed using TEM, EDX and XRD techniques.
133
4.2.3 Refined Method for ZnS Synthesis
The synthesis method was later simplified to combine both metal salt
precursors in one flask, thereby eliminating the need for a third pump. The
aforementioned temperatures for sample ZnS 400 in Table 4.1 were used with
the experimental setup shown in Figure 4.2 to produce the sample “Uncapped
ZnS” – this is the “basic” sample of ZnS without any surfactants or capping
agents introduced.
SEM, XRD, ζ-potential and Fluorometry analysis was conducted on the sample
Uncapped ZnS.
Figure 4.2. A schematic diagram showing the reactor setup for the
refined synthesis method of ZnS.
134
4.2.4 Synthesis of ZnS using Thiourea
Due to the toxicity of (NH2NH)2CS, it is desirable to use a safer sulphur
source in the synthesis of ZnS. Thiourea, (NH2)2CS, was chosen as a potential
substitute due to its lower toxicity, because literature describes it as working
well (Wageh et al., 2003, Apte et al., 2011), it is water soluble, and was readily
available. The same reaction conditions as described in Section 4.2.3 were used
but using the range of reaction temperatures shown in Table 4.2. The system
positions from which the thermocouple readings were taken, were outlined in
Chapter 1.
Table 4.2. Reactor System temperatures for samples synthesised using
thiourea.
Water After Post Capping Post

Sample Heater Heater Mixing Point Cooler
Name Temp Temp Point Temp Temp
(˚C) (˚C) Temp (˚C) (˚C) (˚C)
ZnS- TU-150 190 162 79 49 19
ZnS- TU-250 300 260 130 66 20
ZnS- TU-350 390 352 180 83 22
ZnS- TU-400 450 409 279 105 23
4.2.5 Synthesis of ZnS with Surfactants/Stabilisers
An issue affecting the synthesis of ZnS nanoparticles is agglomeration, as
is the case for most QDs and indeed most nanoparticles. The problems arising
from particle aggregation were described in Section 4.1.7.
135
Consequently, three different stabilising agents were chosen as candidates for
capping ZnS in the Nozzle reactor during synthesis: 3-MPA, NaOH and PVP. The
reasoning behind these choices was described in Section 4.1.7. The methodology
described in Section 4.2.3 was used as a basic protocol for ZnS synthesis in this
experiment set. The sample “Uncapped ZnS” described in Section 4.2.3 was
used as a control for comparison.
For the experiments outlined in Sections 4.2.5.1, 4.2.5.2, and 4.2.5.3, Figure
4.3 depicts the schematic setup of the reactor section of the system and the feed
orientations can be seen. The system pressure was maintained at 240 bar by the
BPR, and the recorded temperatures are shown in Table 4.3.
Figure 4.3. A schematic diagram displaying the reactor setup for the
synthesis reaction of ZnS with capping agents.
136
Table 4.3. System temperatures for synthesis of ZnS with surfactants.
Water Post Mixing Capping

After Heater Post Cooler
Heater Temp Point Temp Point Temp
Temp (˚C) Temp (˚C)
(˚C) (˚C) (˚C)
415 405 265 102 23
4.2.5.1 ZnS and 3-Mercaptopropionic Acid (3-MPA)
In this experiment set, an aqueous solution of 0.25 M 3-MPA was
introduced at the second metal salt pump (MS2) while the precursors for ZnS
were pumped through the first metal salt pump (MS1). The flow rate of the H 2O
downflow was maintained, but the flow rates of MS1 and MS2 were altered for
each sample. A summary of flow rates for each of the 5 samples produced is
displayed in Table 4.4.
137
Table 4.4. Flow ratios for the different ZnS:3-MPA Samples.
Sample H2O H2O Flow MS 1 MS 1 MS 2 MS 2 Flow

Name Feed Rate Feed Flow Feed Rate
(ml/min) Rate (ml/min)

(ml/min)
ZnS:3-
9 1
MPA 9/1
ZnS:3-
8 2
MPA 8/2
0.02 M
ZnS:3- Zn(NO3)2 0.25 M
H2O 10 7 3
MPA 7/3 + 0.02 M 3-MPA
TCH
ZnS:3-
6 4
MPA 6/4
ZnS:3-
5 5
MPA 5/5
MS – Metal Salt; TCH – Thiocarbohydrazide; 3-MPA – 3-Mercaptopropionic Acid
4.2.5.2 ZnS and Polyvinylpyrrolidone (PVP)
For this set of experiments, an aqueous solution of 0.005 M PVP was
introduced at MS2. As with the samples produced with 3-MPA described in
Section 4.2.5.1, the flow rates of MS1 and MS2 were altered for each sample. A
summary of flow rates for each of the 5 samples produced is displayed in Table
4.5.
138
Table 4.5. Flow ratios for the different ZnS:PVP Samples.
MS 2
H2O Flow MS 1
Sample H2O MS 1 MS 2 Flow
Rate Flow Rate
Name Feed Feed Feed Rate
(ml/min) (ml/min)
(ml/min)
ZnS:PVP
9 1
9/1
ZnS:PVP
8 2
8/2
0.02 M
ZnS:PVP Zn(NO3)2 0.005 M
H2O 10 7 3
7/3 + 0.02 M PVP
TCH
ZnS:PVP
6 4
6/4
ZnS:PVP
5 5
5/5
MS – Metal Salt; TCH – Thiocarbohydrazide; PVP – Polyvinylpyrrilidone
4.2.5.3 ZnS and Sodium Hydroxide (NaOH)
NaOH has been described in literature as a surfactant for ZnS (Shen et
al., 2007). It was reported that the alkaline environment results in the formation
of core/shell particles of ZnS/Zn(OH)2 where the surface hydroxyl groups provide
charge stabilisation as well as improving PL properties. In these reactions, an
aqueous solution of 0.25 M NaOH was pumped through MS2, with the flow rates
of MS1 and MS2 altered for each sample, as summarised in Table 4.6.
139
Table 4.6. Flow ratios for the different ZnS:NaOH Samples.
H2O Flow MS 1 MS 2 Flow

Sample H2O MS 1 MS 2
Rate Flow Rate Rate
Name Feed Feed Feed
(ml/min) (ml/min) (ml/min)
ZnS:NaOH
9 1
9/1
ZnS:NaOH
8 2
8/2
0.02 M
ZnS:NaOH Zn(NO3)2 0.25 M
H2O 10 7 3
7/3 + 0.02 M NaOH
TCH
ZnS:NaOH
6 4
6/4
ZnS:NaOH
5 5
5/5
MS – Metal Salt; TCH – Thiocarbohydrazide
4.2.6 Structural Characterisation
For each of the ZnS samples described in Section 4.2.5 and produced
with different ratios of precursors and surfactants, a small volume was collected
for SEM analysis and fluorometry measurements (the latter is outlined in Section
4.2.6.1). In addition, XRD and ζ-potential analysis were conducted on each of
the three 5/5 ratio samples. These four analysis techniques were also performed
on the Uncapped ZnS sample, for comparison. The principles of these structural
characterisation techniques were discussed in Chapter 2.
Samples were washed by vacuum filtration using deionised water, and then
either re-suspended in deionised water for SEM and fluorescence analysis, or
oven dried at 70˚C for 24 hours to yield a powder for XRD analysis.
140
4.2.7 Fluorometry Analysis
Excitation and emission spectra were collected for each of the aqueous
CdS and ZnS samples using a Flexstation II fluorometer from Molecular Devices.
For each sample, 200 µL of material was loaded into a UV-transparent 96-well
plate, which was then scanned by the fluorometer.
Broad emission sweeps were firstly collected by scanning between 450 and 650
nm using a step size of 10 nm, with excitation kept at 350 nm. Likewise, broad
excitation sweeps were conducted by measuring emission at 540 nm while
excitation was varied between 260 and 450 nm using 10 nm increments. For
both excitation and emission spectra, each sample was read 6 times.
Subsequently, more specific emission sweeps were performed by measuring
emission between 430 and 650 nm with 2 nm steps. Excitation was held at 360
nm. To confirm the emission data, a detailed excitation sweep was conducted by
varying excitation between 290 and 410 nm with 2 nm increments; the emission
was measured at a constant 530 nm. For both excitation and emission spectra,
each sample well was read 12 times.
141
4.3.1 Cadmium Sulphide, CdS
This Section describes the samples of CdS produced from the synthesis
experiments detailed in Section 4.2.1. At the lower two temperatures, 240 and
300 °C, the collected samples were transparent and colourless – photos are
provided in Figure 4.4 (a) and (b). On testing with a laser pointer, these samples
did not show a beam through the dispersion. If particles were present and
therefore scattering light, a beam path would be visible. Consequently, it was
determined that no particles were apparent in these samples and no further
tests were conducted on them. Sample CdS 400 produced a bright yellow colour
(see Figure 4.4), and the particles remained in suspension for a few days.
Figure 4.4. Photos of (a) CdS 240, (b) CdS 300, and (c) CdS 400.
DLS measurements performed on CdS 400 reported a particle size of 67 nm.
Meanwhile, TEM analysis of sample CdS 400 revealed it to be composed of
particles with diameters between 20 and 50 nm, as seen in Figure 4.5 (a). It is
142
likely that DLS was measuring the hydrodynamic radii of these particles, thus
gave a bigger particle size than observed under TEM. Images obtained with TEM
at a higher magnification (Figure 4.5 (b)) showed that these particles may in fact
be clusters of individual smaller particles because crystal planes can be seen in
different alignments. Nevertheless, the images do not distinctly show particle
boundaries, and another possibility is that the particles are composed of different
crystallites, which explains the presence of differently aligned planes.
Fluorometry analysis was performed on sample CdS 400, but the data was
inconclusive as there was no discernible difference between the excitation and
emission spectra for the sample, compared to that of the control water blank. In
the latter, no excitation or emission activity is expected; instead, the sample
peaks were seen for both the CdS 400 sample and the water blank. It is most
likely that the signal for the water blank is caused by the fluorescence of the
CdS, as the sample well was adjacent to the well holding the water blank during
analysis – a process known as “bleedthrough”. Because CdS was not the main
material of interest in this study, a full data set (XRD analysis etc.) was not
acquired. Instead, most effort was directed towards ZnS analysis.
143
Figure 4.5. (a) and (b) show TEM images of sample CdS 400.
4.3.2 Initial synthesis of ZnS
The samples produced under the methodology described in Section 4.2.2
will be discussed in this Section. With the exception of sample ZnS 400, all
samples were transparent and colourless when collected from the BPR outlet.
When tested with a laser, there was no indication of the presence of particles –
144
therefore these samples were not analysed further. Still, sample ZnS 400 was a
white/grey suspension when collected from the BPR outlet, with the coloured
particles sedimenting at the bottom of the vessel within a few hours. The XRD
pattern collected for this sample is shown in Figure 4.6; the peaks can be
ascribed to ICCD PDF 80-20, cubic phase ZnS, and 80-7, hexagonal phase ZnS.
Figure 4.6. XRD pattern for sample ZnS 400, with expected peak
positions of cubic ZnS (ICCD PDF 80-20) shown in black, and hexagonal
ZnS (ICCD PDF 80-7) shown in red.
The TEM images which were captured of sample ZnS 400 showed clustered
particles where each cluster measured between 50 and 100 nm (see Figure 4.7
(a) and (b)). At higher magnifications, the clusters appeared to be composed of
primary particles with diameters <15 nm (Figure 4.7 (c)). Furthermore, at high
resolution, the crystal lattice of the particles can be observed (Figure 4.7 (d)).
145
Figure 4.7. a) and b) TEM images and c) HRTEM images of sample ZnS
400.
The elements expected from sample ZnS 400 are indeed present, seen from the
EDX spectrum in Figure 4.8. The Cu and C peaks can be assigned to the copper
and carbon of the TEM sample grid. A small peak for oxygen is present which is
likely to have been introduced during sample preparation as the XRD pattern
(Figure 4.6) shows a match to the expected peaks for cubic ZnS, rather than a
crystal structure containing an oxide group.
146
Figure 4.8. EDX Spectrum collected for sample ZnS 400
4.3.3 Synthesis with Thiourea as the Sulphur Source
The experiments described in Section 4.2.4 (using thiourea as a sulphur
source) resulted in clear, colourless products at reaction temperatures up to and
including 350 °C. The samples were also shown to be negative when tested with
a laser. From this, it can be inferred that a minimum reaction temperature of
400 °C is required for particle formation as a cloudy white suspension was
obtained for sample ZnS- TU-400. This sample was dried and the powder
analysed with XRD. The resulting pattern, shown in Figure 4.9, displayed no
obvious peaks matching ZnS but the peaks were matched to zinc sulphate
hydrate and zinc hydrogen sulphate (ICCD PDF 74-1331 and 78-2208
respectively).
This result shows that thiourea cannot be used interchangeably with
thiocarbohydrazide to produce ZnS. To maintain the focus of this project, all
further reactions were conducted using thiocarbohydrazide as the sulphur
147
source. Nevertheless, the method of using thiourea as a sulphur source for ZnS
synthesis could be optimised in future work.
Figure 4.9. XRD pattern for Sample ZnS-TU-400. Peaks labelled with red
triangles match zinc hydrogen phosphate and peaks labelled with blue
crosses match zinc sulphate hydrate.
4.3.4 Synthesis of ZnS with Surfactants
In this Section the results from the experiments described previously in
Section 4.2.5 will be presented and discussed. For each type of analysis
performed, the ZnS products synthesised with the three different surfactants will
be discussed together and compared to each other, and to ZnS particles
produced in the absence of any surfactants – Uncapped ZnS.
For all the samples produced, at each flow ratio of the 3 different surfactants,
samples were collected as cloudy suspensions - white/grey in colour - where the
particles settled to the bottom of the vessel within a few hours.
148
4.3.4.1 XRD Data
XRD analysis performed on ZnS:NaOH 5/5, ZnS:PVP 5/5, ZnS:3-MPA 5/5
and Uncapped ZnS produced only peaks which could be assigned to ZnS.
Therefore, it can be concluded that ZnS is the only crystalline phase present in
the samples. However, each sample is composed of different crystal phases of
ZnS – as such, there are discrepancies between the XRD patterns of the four
samples. Of the four samples, the patterns for Uncapped ZnS and ZnS:PVP 5/5
are the most similar to each other. Both have peak positions matching ICCD
PDFs 80-20 and 80-7 - corresponding to cubic and hexagonal ZnS respectively.
Nevertheless, the peaks consistent with hexagonal ZnS are more distinct in
ZnS:PVP, compared to Uncapped ZnS, which suggests either that the hexagonal
phase is more abundant, or more crystalline. Moreover, the pattern for ZnS:PVP
contains peaks matching ICCD PDF 12-688, the pattern for another phase of
hexagonal ZnS. From this, it can be inferred that the presence of PVP in the
reaction promotes the formation of hexagonal crystal phase over cubic phase.
The XRD pattern for ZnS:NaOH contains peaks which can also be ascribed to
cubic and hexagonal phases of ZnS (ICCD PDFs 80-20, 5-566 and 12-688), as
does the pattern for ZnS:3-MPA. The peaks in the latter sample match the
database pattern file 75-1534. It is difficult to conclude from this data if the
pattern also matches PDF 12-688, which is hexagonal phase, or PDF 5-566,
which is cubic phase, as the peaks lie in similar positions. However, as the
pattern for ZnS:3-MPA is very similar to that of the other ZnS samples, it is
likely to also be a mixture of cubic and hexagonal phases.
149
Figure 4.10. XRD Patterns of the four ZnS samples with the matched
peak positions of different ZnS phases shown below. Numbered labels
indicate the ICCD PDF file number and the corresponding crystal phase
is given.
150
4.3.4.2 Crystallite Size
In terms of peak shape, the XRD pattern obtained for ZnS:NaOH contains
peaks which are slightly narrower and sharper than those seen for the other
three samples. As discussed in Chapter 2, sharper, narrower peaks can either
mean a higher degree of crystallinity in the sample, or, in the case of
nanoparticles, a larger average crystallite size compared to samples which
exhibit broader XRD peaks.
To determine if the crystallite size phenomena is valid in the case of these
samples, particle size analysis was required. As such, SEM images were collected
for the four samples. These are presented in Figure 4.11. The micrographs
indicate the presence of clustered particles where the level of clustering, and the
size of the clusters varies between samples. In all four samples, the clusters are
<300 nm and the compositional particles much smaller than this (<50 nm as
seen under SEM at higher magnification, and based on TEM images collected for
sample ZnS 400). Hence, the XRD peak broadening can be ascribed to the
presence of small crystallites. The crystallite size calculations will be discussed in
this Section, while more in depth analysis of the SEM data will be detailed in
Section 4.3.4.3.
To calculate the average crystallite size for each sample, Xfit software was used.
The results, along with the crystal phase data inferred from XRD patterns, are
presented in Table 4.7. For Uncapped ZnS, the calculated crystallite size is 4.1
nm while for ZnS:PVP it is slightly bigger at 4.6 nm. For ZnS:NaOH, it appears
that addition of NaOH to the ZnS synthesis reaction leads to a significant
increase in the crystallite size, as the calculated value here is 5.8 nm (this
corresponds to the earlier point of sharper narrower peaks seen in the XRD
pattern of this sample). This significant increase in crystallite size may be due to
151
NaOH having a neutralising effect on the synthesis reaction; the nitrate
precursor creates acidic conditions through the generation of HNO 3, which leads
to surface oxidation of the particles and termination of growth. However the
presence of NaOH counteracts this, the surfaces do not become oxidised, and
crystal growth prevails.
Finally, the average crystallite size for sample ZnS:3-MPA was calculated as 3.1
nm – significantly smaller than the other three samples. A possible explanation
for this is the efficient binding of 3-MPA to the surface of the ZnS nanoparticles,
which prevents further crystallite or particle growth.
Table 4.7. Summary of crystal phase and crystalline size for each
sample, as derived from XRD analysis.
ICCD PDF
Sample Crystal Phase^ Crystallite Size*
Match
80-20
Uncapped ZnS Cubic/Hexagonal 4.1 nm
80-7
80-20
ZnS:PVP 80-7 Cubic/Hexagonal 4.6 nm
12-688
80-20
ZnS:NaOH 5-566 Cubic/Hexagonal 5.8 nm
12-688
5-566
ZnS:3-MPA Cubic/Hexagonal 3.1 nm
75-1534
^
Crystal phase identified using EVA software
*Crystallite size calculated using Xfit software
152
4.3.4.3 SEM Analysis
The SEM images shown in Figure 4.11 will be discussed in this Section. As
mentioned previously in Section 4.3.4.2, the images show clusters of particles.
To some extent, the clustering effect is due to the sample preparation step prior
to SEM analysis – where liquid samples are dried on the sample stub, surface
tension may cause particles to stick together. Nevertheless, the effect that each
surfactant has on particle aggregation can be assessed from this data.
SEM images were also collected for the samples synthesised using different flow
ratios of each of the surfactants; these are shown in the Appendix but in the
interest of comparison, only SEM images of the samples produced at 5/5 flow
ratios are presented and discussed here.
153
Figure 4.11. SEM images of (a) Uncapped ZnS; (b) ZnS:PVP 5/5; (c)
ZnS:NaOH 5/5; and (d) ZnS:3-MPA 5/5.
The sizes of 20 clusters in each sample were measured from the SEM images,
and an average calculated along with the standard deviation from the mean. This
data is shown in Table 4.8. The 20 clusters were chosen at random for analysis
from the top of the sample as seen in the micrograph. This was to ensure any
differences in scale, due to the sample depth of view, were minimised.
154
Table 4.8. A table showing the average cluster size for each sample, as
measured from SEM images.
Sample Average Cluster Size (nm)
Uncapped ZnS 201.7 ± 46.5
ZnS:PVP 5/5 207.4 ± 39.2
ZnS:NaOH 5/5 165.3 ± 25.7
ZnS:3-MPA 5/5 157.2 ± 65.7
The clusters seen in ZnS:PVP are slightly bigger than those present in Uncapped
ZnS, while in ZnS:NaOH and ZnS:3-MPA, they are smaller. This difference may
be due to the effectiveness of each of the surfactants in preventing the
aggregation of primary particles. It can be deduced, following the data presented
in Section 4.3.2 for sample ZnS 400, that without the addition of any capping
agents or surfactants, the nanoparticles in sample Uncapped ZnS aggregate and
form the clusters seen in Figure 4.11 (a). However, if the presence of NaOH and
3-MPA are indeed preventing particle aggregation (even if only to some extent),
then the clusters would be smaller than in the control sample, which is what can
be observed in Figure 4.11 (c) and Figure 4.11 (d). Therefore, the SEM data
suggests NaOH and 3-MPA are lessening the degree of ZnS particle aggregation.
On the contrary, the larger average cluster size seen for ZnS:PVP (in Figure 4.11
(b)) may be a result of the PVP polymer coating the clustered particles, as
opposed to the desired effect of coating the individual primary particles.
Unlike the images for the other three samples, the clustered particles present for
ZnS:3-MPA (in Figure 4.11 (d)) are apparently more “textured” and can be
clearly seen, even at lower magnifications, to be formed of smaller particles. This
may be further indication of 3-MPA preventing the formation of tightly
aggregated particles; instead the particles are only loosely coagulated. It is likely
155
that the observed coagulation of particles is due to an insufficient concentration
of 3-MPA present during synthesis, therefore the nanoparticles have not been
completely coated with the surfactant.
To further investigate the effect of each surfactant on the stability of ZnS
particles, zeta potential measurements were taken over a pH range. These data
are presented and discussed in Section 4.3.4.4.
4.3.4.4 Zeta Potential Measurements
Zeta potential measurements were taken over a pH range for the four
samples, to ascertain the point of zero charge (PZC) or iso-electric point (IEP).
These data are shown in Figure 4.12 and with the individual data points for each
sample given in the Appendix. The zeta potential profile for Uncapped ZnS
confirms that in acidic conditions, the particles have a positive zeta potential and
under basic conditions, the particles have a negative zeta potential. From this
profile, it can be deduced that the IEP of Uncapped ZnS is approximately pH 7.8.
This means that at this pH, there are little repulsive forces keeping the particles
apart, and they are likely to coagulate rapidly. The IEP of ZnS particles has
previously been reported at pH 5.5 (Durán et al., 1995); the IEP measured for
sample Uncapped ZnS may be higher due to the surface oxidation reactions
which occur as a result of the acidic conditions of synthesis (due to the nitrate
precursor used).
Only in basic conditions (pH 10 or higher) is the zeta potential of Uncapped ZnS
< -30 mV and the dispersion can be considered stable. These data corresponds
with the problems associated with particle aggregation seen for unmodified QDs,
which were outlined in Section 4.1 and reinforces the need for viable surfactants.
156
Figure 4.12. Zeta potential profiles for the four samples, as a function of
pH.
It can be seen that the zeta potential profiles for the four samples are indeed
different, indicating that the different capping agents are affecting the surface
chemistry of the ZnS particle. In the case of ZnS:3-MPA the IEP is reached ca.
pH 4.1 – a shift compared to Uncapped ZnS – and at neutral pH, the zeta
potential is <-30 mV suggesting the particles are stable. In this sample, the
particles are capped by an acidic species; Figure 4.13 shows the predicted
surface modification of ZnS particles with 3-MPA. In water, the acid group
dissociates to leave COO- at the surface. As a result, the electrical double layer
of the particle (and in turn the zeta potential) is negative at neutral pH. In
addition, more H+ ions must be added (i.e. the pH reduced further) in order to
give a positive zeta potential, compared to the Uncapped ZnS sample.
157
Figure 4.13. A schematic diagram depicting ZnS particles surface
modified with 3-MPA.
On the contrary, sample ZnS:NaOH - which has been capped with an alkaline
species – has a slightly positive zeta potential value at pH 7 and an IEP ca. pH
8.5, marginally more basic compared to Uncapped ZnS. Like the sample
Uncapped ZnS, particles in ZnS:NaOH exhibited a zeta potential <-30 mV, and
are deemed stable, at pH 10 or more.
Unlike NaOH and 3-MPA, PVP is a non-ionic surfactant – where it provides
stabilisation through steric hindrance as opposed to surface charge
modifications. Therefore, the zeta potential profile for ZnS:PVP is largely
unaffected by changes to pH, and over the pH range tested, the sample did not
have a zeta potential which could be concluded as stable. Although the resulting
profile yields an IEP ca. pH 3, it was not possible to measure the zeta potential
at a lower pH.
At pH values less than 2-3 or more than 11-12, the conductivity of the samples
were too high to be measured by the same means and equipment; thus accurate
data could not be obtained at these pH environments.
158
This zeta potential data demonstrates that PVP is not a suitable surfactant for
the charge stabilisation of ZnS particles. The addition of NaOH to the reaction
scheme does appear to affect the surface chemistry of the ZnS particles, but the
zeta potential profile is still similar to that of the Uncapped ZnS particles. The
most effective surfactant of the three tested is 3-MPA, as the IEP is significantly
shifted away from neutral pH. Instead, the particles have a zeta potential of <-
30 mV at pH 7. This means that the ZnS particles are less likely to aggregate at
neutral pH when surface modified with 3-MPA. This is particularly pertinent for
biological applications, where cellular environments tend to be ca. pH 7.
4.3.4.5 Fluorometry Analysis
The specific excitation and emission sweeps for the four samples,
Uncapped ZnS and ZnS synthesised with each of the three surfactants at a 5/5
flow ratio, are shown in Figure 4.14. The intensity of fluorescence is measured in
relative fluorescence units, RFU. It is important to note that the measured RFU is
dependent on the concentration of ZnS present in the sample; however, the ZnS
concentration varies between the samples so the fluorometry data obtained can
only be assessed qualitatively, and not quantitatively, i.e. the peak positions for
each sample are of importance, but the signal intensity cannot be compared in a
meaningful way. For all samples and blanks, the excitation spectra were
collected with the emission held at 540 nm. The emission spectra were collected
with excitation maintained at 360 nm.
The excitation spectrum for Uncapped ZnS shows that the unmodified particles
fluoresce with the highest intensity at 540 nm, when excited with light at 360
nm. Parallel to this data, the emission spectrum shows that when excited with
light at 360 nm, the emission by particles actually covers a broad wavelength
159
range, with the peak intensity at 558 nm – a wavelength corresponding to green
colour in the visible light range.
Samples ZnS:PVP and ZnS:NaOH display very comparable excitation and
emission spectra to Uncapped ZnS, with peaks in similar positions. Nevertheless,
there are subtle differences. For both these modified samples of ZnS, the
excitation peak at 360 nm is less intense than at lower wavelengths. Instead,
the peak falls at 282 nm. However, there is still a fluorescence signal when
excited at 360 nm; as such, emission data was taken at this excitation
wavelength for ease of analysis. Figure 4.14 (b) shows that the emission peak
for ZnS:NaOH is broad, but the most intense emission is at a wavelength of 556
nm – very similar to Uncapped ZnS and again, corresponds to green light. This is
somewhat different to PL data previously reported, where ZnS particles
synthesised with NaOH showed emission at 440 nm when excited at 323 nm
(Shen et al., 2007). This is likely to be because the particles prepared in the
literature by ultrasonication resulted in the formation of ZnS/ZnS(OH) 2 core/shell
nanoparticles where Zn(OH)2 alters the PL properties.
The emission peak for ZnS:PVP also covers a broad range of wavelengths, but
the highest RFU was detected at 572 nm – slightly red shifted compared to
Uncapped ZnS, and correlates to yellow light. Many existing reports using PVP as
a capping agent discuss QDs with a cadmium-based core and a ZnS shell; as
such, the PL data cannot be directly compared (Saraswathi Amma et al., 2008).
160
Figure 4.14. (a) The excitation spectra (λem = 540 nm) and (b) emission
spectra (λex = 360 nm) for Uncapped ZnS and ZnS produced with three
different surfactants using 5:5 flow ratios. The spectra of the precursors
and water blank were obtained as a control.
The excitation and emission spectra for ZnS:3-MPA is unlike that of the other
samples. The excitation spectrum exhibits only one peak, as opposed to the two
161
seen in the other three samples. While the excitation peak does reach a
maximum at 366 nm, similar to that of Uncapped ZnS, the emission spectrum
was significantly different to that of the other three samples. Although the
emission signal was detected over a wide range of wavelengths, the most
intense signal was detected at 450 nm – violet or blue colour in the visible light
range. This emission data is comparable to previous reports of ZnS capped with
MPA; particles synthesised by a chemical precipitation method yielded a broad
emission spectrum with a peak at 425 nm (blue light) when excited at 310 nm
(Rosli et al., 2012). Conversely, introducing a dopant appears to have a vast
effect on PL; ZnS particles doped with Mn and capped with MPA gave broad PL
emission peaks around 598 nm – orange light – when excited at 320 nm (Kong
et al., 2012).
As a control, the excitation and emission spectra were also collected for a water
blank, and for each of the precursors. With the exception of PVP, no fluorescence
signal was detected in these controls. This shows that the fluorescence signal
detected in the samples can indeed be attributed to the ZnS, rather than their
starting materials. In the case of PVP, the excitation and emission spectra are
remarkably similar to that of ZnS:3-MPA, with the emission peak ca. 450 nm. As
PVP is not expected to fluoresce, the signal is most likely due to bleedthrough
fluorescence from the ZnS:3-MPA sample which was in the adjacent well during
analysis. Moreover, the emission peak at 450 nm was not detected in the sample
ZnS:PVP, reinforcing that bleedthrough is the most likely cause.
To further investigate the effect of 3-MPA on the PL properties of ZnS, the
excitation and emission spectra for the ZnS:3-MPA samples produced using
different precursor flow ratios, are displayed in Figure 4.15. The excitation
spectra, seen in Figure 4.15 (a), shows that ZnS:3-MPA 9/1 has a very similar
spectra to that of Uncapped ZnS, with peaks at the same wavelength. Increasing
162
the concentration of 3-MPA only slightly, as is the case for the 8/2 sample, saw a
discernible change in the excitation spectrum from Uncapped ZnS; the secondary
peak seen ca. 280 nm has flattened out in ZnS:3-MPA 8/2, while the most
intense fluorescence signal is still produced when excited at 360 nm. When the
ratio of 3-MPA to ZnS is increased beyond this, the excitation spectra collected
for the samples are significantly different and show only one peak, around 365
nm.
Correspondingly, the emission spectra in Figure 4.15 (b) shows that ZnS:3-MPA
9/1 yields a similar profile to Uncapped ZnS, and there is only a slight red-shift
in the peak to 548 nm. Sample 8/2, however, shows a more obvious shift in the
peak, to 526 nm – although still in the range of green light. With a higher ratio
of 3-MPA in the samples (ratios 7/3, 6/4 and 5/5), the peak shifts to the blue
light range below 495 nm.
163
Figure 4.15. (a) The excitation spectra (λem = 540 nm) and (b) emission
spectra (λex = 360 nm) for ZnS:3-MPA samples produced using different
precursor flow ratios. The spectra of the precursors and water blank
were obtained as a control.
This PL data suggests that there is a critical concentration ratio of 3-MPA to ZnS
which changes fluorescent properties. This work can be further investigated in
future work, to optimise surfactant concentration. The PL data for ZnS samples
produced with different ratios of PVP and NaOH did not show significant
164
differences from the PL data of Uncapped ZnS, so are presented in the Appendix,
but will not be discussed in detail in this Chapter.
The difference in wavelengths at which the particles in each sample fluoresce is
of interest, because there is a correlation to the average cluster size which was
discussed in Section 4.3.4.3. The crystallite size, cluster size and emission data
for the four samples is reviewed in Table 4.9. It can be seen that there is a
trend: the larger the average cluster size, the longer the wavelength at which
particles peak in emission.
Table 4.9. A summary of the average crystallite size, average cluster
size, and wavelength of peak emission for each ZnS sample.
Average Average Wavelength of

Sample Crystallite Size Cluster Size emission peak
(nm) (nm) (nm)
Uncapped ZnS 4.1 201.7 558
ZnS:PVP 4.6 207.4 572
ZnS:NaOH 5.8 165.3 556
ZnS:3-MPA 3.1 157.2 450
The differences in wavelength of peak emission between the different ZnS
samples indicates that the band gap energies of the particles are varied; another
factor which is affected by the choice of surfactant.
This correlation also suggests that the particles are behaving (namely
fluorescing) as clusters, rather than as primary particles or crystallites.
Consequently, this means there is still some degree of particle aggregation in all
the samples. While the addition of the surfactants has been shown to lessen the
aggregation, the process has not been optimised to completely eradicate it. The
165
broad emission peaks seen for all four samples in Figure 4.14 (b) also suggests
that there is a wide particle or cluster size distribution (Bruchez Jr et al., 1998),
as QDs have size-dependent optical properties as explained by quantum
confinement theory (Dickerson, 2005).
166
4.4 Conclusions
In this Chapter, the protocol for CdS synthesis was replicated and bright
yellow particles of diameters 20 to 50 nm were produced at 240 bar and 400 °C
- consistent with what was seen previously. In addition, it was found that at
temperatures below 400 °C, no particles were formed. The TEM images captured
for the sample were not clear enough to distinguish if the particles were
composed of smaller, aggregated particles, or were just formed of different
crystallites. The fluorometry data collected for the CdS sample were inconclusive
within the scope of this Thesis; instead, the aim of the synthesis experiment was
to demonstrate the reproducibility of the method established by previous
researchers.
The methodology for CdS synthesis was adapted to produce ZnS nanoparticles.
Using Zn(NO3)2.6H2O instead of the Cd(NO3)2.4H2O precursor, white/grey
particles were synthesised at 240 bar and 400 °C. As with CdS, no particles were
produced at temperatures below this. The XRD pattern collected for the sample
proved it to contain a mixture of hexagonal and cubic phases of ZnS, but no
other crystalline phases were detected. The sample was also analysed under TEM
which showed primary particle sizes <15nm but had aggregated into clusters –
50 to 100 nm in diameter. Furthermore, EDX data confirmed the presence of
only the expected elements.
Attempts were made to produce ZnS nanoparticles using thiourea as a sulphur
source, rather than the more toxic reagent thiocarbohydrazide. Consistent with
previous experiments, no particles were generated at temperatures up to and
including 350 °C (with the system pressure at 240 bar). However, at 400 °C,
white/grey particles were synthesised. XRD analysis on these particles revealed
167
them to be a mixture of zinc sulphate hydrate and zinc hydrogen sulphate rather
than ZnS. This demonstrates that, in the established protocol for ZnS synthesis,
thiocarbohydrazide and thiourea cannot be used interchangeably to produce
ZnS.
The final set of experiments described in this Chapter yielded unmodified ZnS
particles which were compared to ZnS produced with three different capping
agents – an acid, a base and a polymer. The samples were analysed using SEM,
XRD, zeta potential and fluorometry techniques. The results showed that without
any capping agents, and consistent with previous data, a mixture of hexagonal
and cubic phases of ZnS was produced. The average crystallite size is 4.1 nm,
but the particles aggregate into clusters with an average diameter of 202 nm.
The particles were shown to reach their IEP at a pH ~7.8 and upon excitation
with UV light at 360 nm, they fluoresced in the visible light range; while the
emission range was broad, the strongest signal was detected at 558 nm – green
light.
Introducing NaOH into the reactor during synthesis of ZnS does not appear to
affect the crystal phase of the product, but the average crystallite size was
increased to 5.8 nm. This is likely to be due to neutralisation of the reaction
environment, meaning crystal surfaces are not oxidised and growth continues to
a degree. The base did have an effect on aggregation as the average cluster size
was decreased to 165 nm, and the zeta potential data showed that the pH at
which the IEP of particle was reached had shifted from 7.8 to 8.5. However, the
obtained fluorescence data was very similar to that of Uncapped ZnS, with a
broad emission peak and the optimum fluorescence detected at 556 nm – again,
green light.
168
The addition of PVP to the synthesis method did not appear to prevent primary
particles from aggregating; instead the polymer seemed to coat the clustered
particles. PVP also did not affect the crystal phase of ZnS produced, but the
calculated average crystallite size was marginally bigger at 4.6 nm. As PVP is a
non-ionic surfactant, the zeta potential data collected could not conclusively
show the IEP of ZnS:PVP. The fluorometry analysis did show this sample had a
red shifted emission peak compared to Uncapped ZnS, with the highest emission
at 572 nm – yellow light.
The work described in this Chapter shows that, of the three surfactants tested,
3-MPA has the most potential in the stabilisation of ZnS nanoparticles. The XRD
data indicated that hexagonal and cubic phase crystals were produced with a
smaller calculated crystallite size than Uncapped ZnS – 3.1 nm - which may be
due to the prevention of further crystallite growth after 3-MPA had successfully
bound to the particle surface. While some degree of particle aggregation was still
evident from the SEM data, the average cluster size had decreased to 157 nm.
Moreover, the zeta potential data showed a shift in the profile and the IEP was
reached at pH 4.1. The excitation and emission data was radically different to
the other samples, with emission blue shifted and reaching a peak at 450 nm –
corresponding to violet or blue light. The difference in emission peak position
between the four samples is likely due to differences in particle band gap
energies. In addition, the PL emission data from ZnS:3-MPA is in a comparable
wavelength range to that previously reported for cadmium-based QDs
(Saraswathi Amma et al., 2008), thus there is the potential to develop ZnS as a
heavy-metal free QD alternative.
Prior to this work, continuous hydrothermal synthesis had not been used to
produce fluorescent nanoparticles of ZnS. Although the samples described in this
work have not been fully stabilised to isolate the primary particles and prevent
169
them from aggregating, their fluorescent potential has been demonstrated, and
3-MPA was shown to be the most effective at stabilising these ZnS particles. The
ratio between surfactant and particle concentration must be carefully balanced in
order to achieve stabilisation, and is dependent on several factors, including
particle size. It is therefore likely that the surfactant concentrations used in this
work were insufficient to effectively cap the particles completely. Consequently,
there is scope for additional studies which investigate the optimum surfactant
concentration to coat and stabilise ZnS nanoparticles, while supplementary
analysis could also quantify the quantum yield of these particles. While
fluorometry analysis has been conducted in this Thesis to assess the PL
properties of the ZnS particles, other tests can also be carried out to measure
their electroluminescent properties. This would allow the material to be applied
in several different potential applications, such as in electronic devices.
170
Chapter 5
5 Magnetic Resonance Imaging

(MRI) Contrast Agents
5.1 Introduction
5.1.1 Principles of MRI
While X-rays have been used since their discovery in 1895 to gather
information about the internal structure of human and animal bodies - from
diagnostic to research purposes – they are not without their drawbacks. X-rays
can only image hard tissue, such as bone, clearly; this means soft tissues such
as tendons, ligaments, organs and even blood cannot be detected or imaged
well. Furthermore, X-rays are a form of ionising radiation which can, above
certain thresholds, cause defects in the DNA of individuals exposed to it, and
lead to knock-on effects - including cancer formation (Chang and Hou, 2011). As
such, X-ray exposure in patients must be carefully monitored and only used if
absolutely necessary. From this, there was no doubt that a more effective
alternative was required. In the 1970’s, a technique called Magnetic Resonance
Imaging (MRI) was developed and applied to the field of medicine, which could
image the soft tissues and blood in the body, without the use of any form of
ionising radiation (Mansfield and Maudsley, 1977, Mansfield and Pykett, 1978).
171
Chapter 5. MRI Contrast Agents
An MRI scanner is composed of several parts. The predominant part is a large,
powerful magnet (a clinical scanner usually has a 1.5 T magnet) which forms a
cylinder around the area holding the sample or the patient (see Figure 5.1). MRI
operates by manipulating and measuring the hydrogen atoms present in a
sample or a patient. Because humans are predominantly composed of water,
there are plenty of hydrogen atoms to measure.
The positive charge of the hydrogen nuclei (a single proton), coupled with the
spin it possesses, gives rise to a magnetic field or magnetic moment. Under the
influence of a large external magnetic field (such as a body or sample being
placed into an MRI scanner), the magnetic moments of the hydrogen nuclei
become aligned. Within an MRI scanner, there are Radio Frequency (RF) coils to
produce extra gradients of magnetic field; these can be rapidly switched on and
off to create a non-uniform magnetic field in the body or sample being scanned.
The RF coils emit an RF pulse which is specific to H atoms only, and the energy is
absorbed by the atoms. This causes them to ‘flip’, so that the magnetic dipole
now faces 180˚ from its previous direction. When the RF pulse is switched off,
the H atoms flip back to their previous alignment in the magnetic field and
release a signal or “spin echo”, as well as the energy that it had absorbed. The
spin echo is made up of several frequencies, which correspond to the different
positions along the magnetic field gradient; it is detected by the RF coil and is
converted to its component frequencies. At each frequency, the signal magnitude
is related to the density of hydrogen nuclei at that respective position. As a
result, the spin echo signal also contains spatial information and can be used to
generate a greyscale image (Conn, 2009).
172
Figure 5.1. An image of an MRI scanner depicting the different
components (Action, 2013).
Where there is a high concentration of hydrogen atoms present (due to the large
quantity of water), there will be a large signal, and will show up as bright white
patches in the image. Conversely, where there is a low concentration of H atoms
(such as lipid or fatty tissues), the signal will be small, and therefore manifest as
dark patches in the image (Oldendorf, 1988).
Also present in MRI scanners are gradient coils. These can fine tune the
magnetic field so that only a specific region of the patient or sample is analysed.
For example, if the liver is of interest then only a cross-section, down to a few
millimetres, of the liver can be imaged. This cross-section is known as a ‘slice’.
The presence of these gradient coils also means that different regions can be
173
imaged and analysed without having to move the patient or sample (Conn,
2009).
5.1.2 MRI Scans
As the RF pulse is switched off, the H nuclei recover to their natural
alignment; the magnetisation of the nuclei exists along the direction of the
applied field (termed the longitudinal component or z plane) and perpendicular
to the direction (termed the transverse component or xy plane). These planes
are shown in Figure 5.2. Relaxation times T1 and T2, which are typically
measured during MRI scans, indicate the time required for the recovery of the H
nuclei to equilibrium, in the longitudinal and transverse components,
respectively. This forms the basis of the two main types of MRI scan – T1-
weighted and T2-weighted.
T1 relaxation is also called spin-lattice relaxation and denotes the rate at which
the longitudinal component of the magnetization vector recovers to equilibrium
maximum from zero. It measures the degree to which the spins are being
disrupted by the surrounding tissue, hence it is known as spin-lattice relaxation.
The image produced distinguishes between water, which appears darker, and
fat, which is shown as lighter areas (Oldendorf, 1988).
Conversely, T2-weighted scans measure the time (usually in milliseconds, ms)
taken for the transverse component of the magnetisation vector to recover and
decay the signal. In the transverse direction, magnetisation recovery is
dependent on how spins interact with each other and put each other out of
phase, so it is also called spin-spin relaxation. T2-weighted images are almost an
inverse of T1-weighted images – water and fluids appears lighter while fat,
darker. The T2 relaxation time is actually the time taken for the signal to decay
174
to 37% of its original value (Brown and Semelka, 2010). Figure 5.2 summarises
the relationship between T2 relaxation time and spin recovery.
Figure 5.2. A diagram showing the z (longitudinal) and xy (transverse)
planes of magnetisation. The relationship between T 2 relaxation time
and spin recovery is also summarised (Oldendorf, 1988).
5.1.3 Relaxation Time, Rate and Relaxivity
It is important to note that the relaxation time of a tissue or a contrast
agent is not to be confused with its relaxation rate. The latter, represented as
R2, is an inverse of the T2 relaxation time, or 1/T2. Moreover, contrast agents can
be compared by their relaxivity, or their ability to increase the relaxation rates of
the surrounding water proton spins. The relationship between relaxivity, r2, and
R2 is shown by the following equation (Brown and Semelka, 2010):
175
Equation 1 R2 = + r2 [Fe]
Where: R2 is 1/T2 in s-1;
T2, H2O is the relaxation time of protons in water (s);
r2 is the relaxivity of the contrast agent (mM-1 s-1), and
[Fe] is the concentration of iron (mM)
Rearranging the equation allows for the relaxivity (r2) to be calculated:
Equation 2 r2 = ( - )/ (Fe)
As T2 relaxation times are concentration-dependent, comparing contrast agents
by relaxivity rather than T2 times gives a better indication of how effective each
material is as a contrast agent.
5.1.4 Role of Contrast Agents
In most cases of clinical MRI scans, the region of interest contains tissues
with sufficient contrast to allow the operator or consultant to distinguish between
two different tissues (e.g. normal healthy tissue and a tumour). However, this is
not always true and the different tissues can manifest as two very similar shades
of grey in the image. To overcome this problem, contrast agents can be used.
These are materials which can be introduced to a subject or patient, and improve
visibility of the organ by reducing the T1 or T2 relaxation time of water protons in
the tissue of interest, thereby enhancing the contrast. Organs which are
commonly imaged by MRI include the gastrointentinal tract, the brain, liver, and
spleen (Conn, 2009).
176
Contrast agents are usually inorganic compounds which can be conjugated to a
chosen biomarker - which will target and bind to a specific tissue, such as a
tumour, to ensure the contrast agent highlights only a designated region of
interest (McCann et al., 2011). Contrast agents which are currently used are
administered either orally or intravenously, depending on the target organ.
Reports indicate particles to be used as contrast agents for MRI must be smaller
than 5 µm in order to avoid being trapped in the lungs (Bammer et al., 2005).
Furthermore, particles with superparamagnetic properties (this is where the
magnetisation is at an average of zero, but particles can become magnetised by
an external magnetic field) give the best effects for T 2-weighted imaging; thus,
nanoparticles of iron oxide are typically used for transverse scans. These fall into
two main categories: superparamagnetic iron oxide nanoparticles (SPIONs) and
ultrasmall superparamagnetic iron oxide nanoparticles (USPIONs); ultrasmall
nanoparticles are classed as having a diameter 30 nm or smaller (Thomsen and
Webb, 2009). A few SPION-based contrast agents are commercially available,
including Feridex IV®, and Combidex® (Stephen et al., 2011). These and a few
other commercially available contrast agents will be described in Section 5.1.5.
Meanwhile, T1-weighted scans typically use paramagnetic gadolinium (Gd) or
manganese (Mn)-based agents (Conn, 2009); these are materials which become
slightly magnetised when placed in a magnetic field, but lose this magnetism
when the field is removed.
Coatings may be applied to the surfaces of contrast agents to improve stability
and dispersion of the particles, thereby preventing coagulation (Tsai et al., 2010)
(See Chapter 1, Section 1.9.2); to lengthen the shelf life of the suspension; and
in some cases, to improve the biocompatibility or to provide an anchor on which
other materials can be attached (Chekina et al., 2011). Other researchers have
177
reported the use of silica (Casula et al., 2011), oleic acid (Zhang et al., 2006),
and sodium bis(2-ethylhexyl)sulfosuccinate (AOT) (Zheng et al., 2006) as
coating materials.
Figure 5.3 (a) shows an MR image of the liver of a patient with cancerous
lesions, without any contrast agents being administered. By contrast, Figure 5.3
(b) is an MR image of the same patient’s liver 10 minutes after a T 2 weighted
contrast agent had been injected. The targeted cancerous region can be visibly
seen as the darker area, making it easier for radiologists locate lesions for
treatment.
178
Figure 5.3. (a) Native MR Image (no contrast agent present) of liver in a
patient with cholangiocellular carcinoma (cancer), and (b) MR image
after T2 weighted contrast agent has been administered. Differentiation
of the lesion is vastly improved after contrast agent is introduced (AG,
2013).
After a SPION contrast agent is administered (typically by intravenous injection),
it enters the bloodstream of an animal model or patient and is subsequently
taken up into the liver, spleen, lymph nodes and bone marrow, through
phagocytosis (internalisation and ingestion of foreign materials by phagocyte
cells as a means of cell defence). Once in these organs, the contrast agents can
begin localised decay of MR signal to create contrasted images (Wang, 2011).
USPIONs, on the other hand, are small enough to remain in the blood pool,
179
rather than being taken up into organs (Bammer et al., 2005). Once the contrast
agent takes effect, the window for optimum imaging lasts only a few hours (this
varies depending on the contrast agent used) as the particles begin to be
metabolised by lysosomes. These are cell components which secrete enzymes to
‘digest’ waste materials and remove them; the contrast agent is consequently
broken down and excreted. In the case of SPIONs and USPIONs, these are
metabolised to a non-superparamagnetic form of iron which can then be
‘recycled’ in cells, such as in haemoglobin.
In the current market, the most commonly used contrast agents are gadolinium
(Gd) based molecules. However, as a heavy metal, Gd is extremely toxic and
accumulation in the body can be fatal. Consequently, the contrast agents are
synthesised so that the Gd3+ ion is trapped in a molecular cage or strongly
bound to ligands, so that it does not bind to cell components (Carbonaro et al.,
2011). Other species, including iron oxide-based agents, are arguably better for
use in MRI as the particles can be digested by natural cell processes without
posing toxicity risks. In conclusion, good contrast agents feasible for commercial
application should strike a balance between having high relaxivity and low
toxicity, whilst being economically feasible.
5.1.5 Commercial Contrast Agents
A variety of nanoscale contrast agents are currently used, or have been used in
the past, for clinical MRI scans; the majority of these are based on iron oxide
nanoparticles. A few will be briefly described in this Section.
180
5.1.5.1 Feridex IV®
Developed by AMAG Pharmaceuticals, Feridex IV ® are SPIONs coated with
low molecular weight Dextran. Particle size ranges between 120 and 180 nm,
and its r2 relaxivity is 98.3 mM-1 s-1 (Wang, 2011). Feridex I.V® is used
specifically for MR imaging in the liver to detect lesions and is administered as a
drip infusion which takes roughly 30 minutes. Around 6-8% of the SPIONs are
taken up by the liver starting within 8 minutes of injection, though optimum
imaging (maximum signal decay) occurs after one hour. The length of time for
contrast agent activity varies between patients but is typically between 30
minutes and 6 hours from administration (Wang, 2011). Although Feridex IV®
was, and possibly still is, used in a clinical background; its manufacture has
since been discontinued by AMAG Pharmaceuticals (AMAG Pharmaceuticals,
2013).
5.1.5.2 Resovist®
Produced by Schering AG, the active ingredient in Resovist ® are
carboxydextrane-coated SPIO particles. These have a reported hydrodynamic
diameter of 45-60 nm, where larger particles are taken up faster, and a r2
relaxivity of 151 mM-1 s-1 (Wang, 2011). Resovist® is also used for T1-weighted
MRI, but with mixed results – therefore it is principally used for T2-weighted
imaging in the liver. Quick uptake by the reticuloendothelial system (part of the
body’s defence mechanism) means Resovist® is effective in as little as 10
minutes after injection, with approximately 85% of the administered dose taken
up by cells in the liver. Furthermore, Resovist® has shown a very good safety
profile and is widely used by radiologists (AG, 2013).
181
5.1.5.3 Ferumoxtran-10
Known as Combidex in the USA and Sinerem in Europe, this contrast
agent was developed by AMAG Pharmaceuticals and is based on USPIONs (35
nm in diameter) coated in dextran (Hudgins et al., 2002). Sinerem was used in
Europe on a clinical basis, but Combidex did not advance past trial stages in the
USA. Both products have since been discontinued.
The r2 relaxivity of Sinerem was 60 mM-1 s-1 and due to the small size of
particles, they tended to build up in lymph nodes. However nodes containing
cancerous tissue, and lacking certain cell components, could not uptake Sinerem
and hence did not show a decayed MR signal and could be easily spotted
(Harisinghani et al., 2003). Nevertheless, it took over 24 hours before particles
were properly uptaken and MR images could be obtained.
In 2009, a study on 296 patients with prostate cancer assessed the use of
ferumoxtran-10 and MRI to detect cancerous lymph nodes (Heesakkers et al.,
2009). Findings of a 24.1% false positive rate lead to the clinical development of
ferumoxtran-10 being halted, and AMAG Pharmaceuticals subsequently
discontinued production. While Combidex/Sinerem had some flaws, the
underlying research approach into its development is still considered to be
promising (Wang, 2011).
5.1.5.4 Clariscan™
Composed of SPIONs with a carbohydrate polyethylene glycol (PEG)
coating, Clariscan™ was developed by GE Healthcare (formerly Nycomed
Imaging). Total particle diameters lie in the region of 20 nm, with r2 relaxivity
measuring 35.3 mM-1 s-1 (Bjørnerud et al., 2001). Clariscan™ is principally
182
uptaken to the liver and spleen, and has been used to characterise tumour
development by providing contrast in the MRI of blood vessels (Wang, 2011).
However, production of Clariscan™ was discontinued amid safety concerns.
5.1.5.5 Omniscan®
The active ingredient in Omniscan® is the Gd(DTPA-
bismethylamide)(H2O) complex, which yields an r2 relaxivity of 3.4 mM -1 s-1.
However, it is principally used as a T1-weighted contrast agent due to its good r1
relaxivity of 4.9 mM-1 s-1 (Dutta et al., 2008). Manufactured by GE Healthcare,
Omniscan® has been used in clinical applications, mainly in the USA, for over 20
years. It is administered intravenously, and can be used to image the brain,
spine, kidney, intrathoracic, intra-abdominal and pelvic regions (Healthcare,
2013).
5.1.6 Current Methods of Synthesising Contrast Agents
While there is little information publically available detailing the synthesis
methods for commercial contrast agents, there is a vast array of journal
publications describing the synthesis and testing of nanomaterials for MRI use.
As is the case with many nanomaterials, co-precipitation is one of the most
widely used synthesis methods. One example described this synthetic route for
the production of Fe3O4, NiFe2O4 and CuFe2O4 nanoparticles (Covaliu et al.,
2011). In this study, chitosan and sodium alginate were used to coat the
particles but while magnetic measurements confirmed the superparamagnetic
nature of the coated particles, neither their relaxation times nor relaxivities were
analysed.
183
Other commonly described methods to manufacture magnetic nanoparticles
include thermal decomposition (Bateer et al., 2013, Khodadadi-Moghaddam et
al., 2013), microemulsion (Darbandi et al., 2012, Bao et al., 2006),
hydrothermal synthesis (Haw et al., 2010, Bae et al., 2012) and sonochemical
routes (Zhang et al., 2012).
Regardless of the synthesis route chosen, a common observation which
researchers have identified is the need for particles to be coated in order to
improve biocompatibility and dispersibility in water (Jadhav and Bongiovanni,
2012, Karimi et al., 2013).
5.1.7 Chapter Aims
To avoid these issues of toxicity, this Chapter focuses on non-Gd based
contrast agents, namely iron oxide nanoparticles hematite (Fe2O3) and magnetite
(Fe3O4). Hence the aim of this Chapter is to use the counter-current Nozzle
reactor to synthesise materials of interest as MRI contrast agents, and to
characterise their respective particle sizes, crystallinity, and applicability as a
contrast agent, where the latter will be measured as the T 2 time. These values
will then be used to calculate the material’s relaxivity.
Although the principle focus will be on iron oxides synthesised by the continuous
hydrothermal reactor, and their suitability as T2-weighted contrast agents, cobalt
oxide (Co3O4) and gadolinium hydroxide (Gd(OH)3) will also be synthesised and
analysed. These materials were chosen because they have previously been
manufactured successfully on the counter-current Nozzle reactor (Lester et al.,
2012). Furthermore, cobalt-based nanoparticles have reportedly exhibited
positive results for MRI (Parkes et al., 2008). Although Gd-based contrast agents
after typically used for T1-weighted MR imaging, it is possible these particles also
184
exhibit positive results for T2-weighted scans (Dutta et al., 2008); hence,
Gd(OH)3 will also be tested along with the iron oxide samples.
Products will be produced pure (no coatings or surfactants added), or with one of
three different polymers added as a coating; these being polyvinylpyrrolidone
(PVP), dextran or sodium alginate. PVP was chosen due to its known
biocompatibility, low cost and its widespread availability (Sharma et al., 2010);
Dextran has been widely reported as a surfactant for Fe3O4 particles in MRI
applications (Ding et al., 2010, Mornet et al., 2004); and sodium alginate has
been reported in coating magnetic nanoparticles for biomedical applications (Mu
et al., 2012, Covaliu et al., 2011).
185
5.2 Methodology
5.2.1 Hydrothermal Synthesis of Nanoparticles
The reagents used for synthesis were hydrogen peroxide, H2O2; iron (III)
nitrate nonahydrate, Fe(NO3)3.9H2O (≥98% purity); ammonium iron (III) citrate,
NH3 Fe(C6H5O7) (reagent grade); cobalt acetate tetrahydrate, Co(CH 3COO)2·4H2O
(reagent grade); gadolinium (III) acetate hydrate, Gd(CH 3CO2)3·xH2O (99.9%
purity); polyvinylpyrrolidone (PVP), (C 6H9NO)n (average molecular weight of
10,000); dextran low fraction, (C6H10O5)n, from Acros Organics; and sodium
alginate, NaC6H7O6.
Unless otherwise stated, all chemical reagents were purchased from Sigma
Aldrich (UK) and used without further purification. DI grade water was used for
the work in this Chapter.
5.2.1.1 Hematite, Fe2O3
To produce Fe2O3, a 0.05 M Fe(NO3)3.9H2O aqueous solution was used as
a precursor. A brief reaction schematic with flows and flow rates can be seen in
Figure 5.4. Three different reaction temperatures were employed to produce
samples, ~150, ~200 and ~300 °C, as this has previously been shown to affect
particle size (Li, 2008). For each of these temperatures, a control sample was
produced without any surfactants, and then each of the three capping agents,
PVP 10, dextran or sodium alginate, was introduced. Therefore, twelve samples
of Fe2O3 were synthesised. The temperatures at different points in the reactor
system are summarised in Table 5.1, Table 5.2 and Table 5.3. The positions of
the thermocouples where these temperatures were logged can be found in
186
Chapter 1, Section 1.7.2.1. For all samples, the system pressure was maintained
at 240 bar by the BPR. Previous unpublished work by other researchers using
the Nozzle reactor found that adding surfactants or capping agents at the second
metal salt pump was more effective at coating particles than introducing them at
the capping point. It is for this reason that the capping agents were introduced
in this manner.
Figure 5.4. A schematic diagram showing the reaction setup for
synthesis of Fe2O3.
187
Table 5.1. Summary of Temperatures and Flows for first set of Fe2O3
Samples.
Temperatures (°C)
Sample Capping Water Post

After Capping Post
Name Agent Flow Heater Set Mixing
Heater Point Cooler
Point Point
HEM A1 Water
1 wt% PVP
HEM A2
10
1 wt%
HEM A3 185 154 107 71 29
Dextran
0.25 wt%
HEM A4 Sodium
Alginate
Table 5.2. Summary of Temperatures and Flows for second set of Fe2O3
Samples.
Temperatures (°C)

After Capping Post
Heater Point Cooler
Point Point
HEM B1 Water
1 wt% PVP
HEM B2
10
1 wt%
HEM B3 240 203 135 86 31
Dextran
0.25 wt%
HEM B4 Sodium
Alginate
188
Table 5.3. Summary of Temperatures and Flows for third set of Fe2O3
Samples.
Temperatures (°C)

After Capping Post
Heater Point Cooler
Point Point
HEM C1 Water
1 wt% PVP
HEM C2
10
1 wt%
HEM C3 345 307 206 117 38
Dextran
0.25 wt%
HEM C4 Sodium
Alginate
The twelve synthesised samples were characterised using a 0.5 T MRI scanner to
measure their T2 times (see Section 5.2.2), and using DLS to determine the
particle size of each sample. Samples HEM A1, B1 and C1 were oven dried in air
at 70 °C overnight to obtain powders; these were then analysed using XRD.
Details of the DLS and XRD equipment used were given in Chapter 2.
5.2.1.2 Magnetite, Fe3O4
For the synthesis of Fe3O4, a 0.05 M NH3 Fe(C6H5O7) aqueous solution was
used as a precursor. Figure 5.5 shows a simple reaction schematic of the flow
orientations and flow rates. Samples were produced with the water heater set at
two different temperatures: ~350 and ~400 °C. For each of these temperatures,
a control sample was produced without any surfactants, and then each of the
three capping agents, PVP 10, dextran or sodium alginate, were introduced. In
total, eight samples of Fe3O4 were obtained. The temperatures at different points
189
in the reactor system are summarised in Table 5.4 and Table 5.5. The positions
of the thermocouples where these temperatures were logged can be found in
Chapter 1, Section 1.7.2.1. For all samples, the system pressure was maintained
at 240 bar by the BPR.
Figure 5.5. A schematic diagram showing the reaction setup for Fe 3O4
synthesis.
190
Table 5.4. Summary of Temperatures and Flows for First Set of Fe 3O4
Samples.
Temperatures (°C)

After Capping Post
Name Agent Flow Heater Mixing
Heater Point Cooler
Set Point Point
MAG F1 Water
MAG F2 1 wt% PVP 10
1 wt%
MAG F3
Dextran 385 353 239 127 40
0.25 wt%
MAG F4 Sodium
Alginate
Table 5.5. Summary of Temperatures and Flows for Second Set of Fe3O4
Samples.
Temperatures (°C)

After Capping Post
Heater Point Cooler
Set Point Point
MAG G1 Water
MAG G2 1 wt% PVP 10
1 wt%
MAG G3
Dextran 450 402 340 211 50
0.25 wt%
MAG G4 Sodium
Alginate
191
Samples MAG G1-G4 were characterised using a 0.5 T MRI scanner and a 2.35 T
MRI scanner to measure their T2 times (see Section 5.2.2). Samples MAG G1 and
G3 were oven dried in air at 70 °C overnight to obtain powders, and analysed
using XRD. These two samples were also analysed using TEM, for particle size
and morphology data. Details of the XRD and TEM equipment used were given in
Chapter 2.
5.2.1.3 Cobalt Oxide, Co3O4
The synthesis of Co3O4 was performed for a separate study (Moro et al.,
2013), however a sample was also taken for MRI analysis. Here, a 0.05 M
Co(CH3COO)3 aqueous solution was used as a precursor. Figure 5.6 shows a
simple reaction schematic of the flow orientations and flow rates. Hydrogen
peroxide, H2O2, was added to the water feed during synthesis, to provide a more
oxidising environment for the breakdown of the CH 3COO- ion from the precursor,
and to drive oxidation of the Co3+ cation. Samples were produced with the water
heater set at a temperature of ~400 °C. Unlike the samples of Fe 2O3 and Fe3O4,
no surfactants were added to the synthesis of Co3O4 as it was not required for
the separate study, and so one sample was obtained. The temperatures at
different points in the system during synthesis are summarised in Table 5.6. The
positions of the thermocouples where these temperatures were logged can be
found in Chapter 1, Section 1.7.2.1. For all samples, the system pressure was
maintained at 240 bar by the BPR.
192
Figure 5.6. A schematic diagram showing the reaction setup for Co3O4
synthesis.
Table 5.6. A Summary of Temperatures for the Synthesis of Co3O4
Sample.
Temperatures (°C)
Sample Water Post

After Capping Post
Name Heater Set Mixing
Heater Point Cooler
Point Point
Co3O4 430 427 376 173 32
193
The sample of Co3O4 was analysed using a 0.5 T MRI scanner (see Section
5.2.2). Because this material was produced as part of a separate study, XRD and
TEM was also conducted on the sample and the acquired data will be presented
in this Chapter.
5.2.1.4 Gadolinium Hydroxide, Gd(OH)3
To synthesise Gd(OH)3, a 0.05 M Gd(CH3COO)3 aqueous solution was
used as a precursor. Figure 5.7 shows a simple reaction schematic of the flow
orientations and flow rates. Hydrogen peroxide, H2O2, was added to the water
feed during synthesis to provide a more oxidising environment for the
breakdown of the CH3COO- ion from the precursor, and to drive oxidation of the
Gd3+ cation. Samples were produced with the water heater set at one of two
temperatures. At each temperature, a control sample was produced without any
surfactants, and then each of the three capping agents, PVP 10, dextran or
sodium alginate, was introduced. In total, eight samples were obtained. The
temperatures at different points in the reactor system are summarised in Table
5.7 and Table 5.8. The positions of the thermocouples where these temperatures
were logged can be found in Chapter 1, Section 1.7.2.1. For all samples, the
system pressure was maintained at 240 bar by the BPR.
194
Figure 5.7. A schematic diagram showing the reaction setup for Gd(OH) 3
synthesis.
Table 5.7. A Summary of Temperatures and Flows for the Synthesis of
the First Set of Gd(OH)3 Samples.
Temperatures (°C)

After Capping Post
Heater Point Cooler
Set Point Point
GAD D1 Water
GAD D2 1 wt% PVP 10
1 wt%
GAD D3
Dextran 385 352 191 124 46
0.25 wt%
GAD D4 Sodium
Alginate
195
Table 5.8 . A Summary of Temperatures and Flows for the Synthesis of
the Second Set of Gd(OH)3 Samples.
Temperatures (°C)

After Capping Post
Heater Point Cooler
Set Point Point
GAD E1 Water
GAD E2 1 wt% PVP 10
1 wt%
GAD E3
Dextran 450 406 303 185 60
0.25 wt%
GAD E4 Sodium
Alginate
The four Gd(OH)3 samples produced in the second experimental set were
characterised using a 0.5T MRI scanner (see Section 5.2.2). Sample GAD E1 was
also oven dried at 70 °C overnight to yield a white powder, which was then
analysed using XRD to confirm the crystal phase produced.
5.2.2 MRI Characterisation of Samples
5.2.2.1 Halbach 0.5T Table-Top Scanner
Due to the high sample turnover from the continuous hydrothermal
reactor, a large number of samples were synthesised. A preliminary MRI
screening process was employed to ascertain an average T 2 relaxation time for
each sample. This was conducted using a 0.5 T Halbach MRI system, based at
Nottingham Trent University, using a repetition time of 2000 ms and 4 averages;
thus each scan took 8 seconds. For the majority of samples, 128 echoes were
196
used; however, with the sample of Co3O4 and its precursor, 512 echoes were
chosen to provide a more reliable result. Data analysis was performed with
Prospa v2.1 software. This scanner, a photo of which is given in Figure 5.8,
provided a rapid method for measuring T2 values for each sample. Those which
were most promising were then characterised further using DLS or TEM, and
where appropriate, XRD. Details of these techniques were given in Chapter 2.
Samples with low T2 relaxation times from the 0.5 T Halbach system were also
tested on a more powerful MRI scanner – a 2.35 T Bruker kit (described in
Section 5.2.2.2.).
Figure 5.8. Photo of the Halbach 0.5 T Table-Top MRI Scanner at
Nottingham Trent University.
5.2.2.2 Bruker 2.35 T Scanner
A select number of samples were taken forward to be scanned on the
Bruker Biospin 2.3 T MRI system, based at Nottingham Trent University, and T 2
197
images were acquired. A photo of the equipment is shown in Figure 5.9. For each
scan, a multi slice multi echo (MSME) sequence and the following parameters
were used:
Field of View = 7 cm x 7 cm
Repetition Time = 2000 ms
Slice Thickness = 3 mm
Number of Slices = 5
Echo Time = 10.25 ms
Number of echoes = 24
It should be noted that 1 pixel in the obtained images corresponds to
approximately 0.3 mm of real space.
Figure 5.9. Photos of the Bruker 2.35 T MRI scanner at Nottingham Trent
University.
198
5.3.1 Hematite, Fe2O3
The four samples of Fe2O3 synthesised at ~150 °C are shown in the
photograph in Figure 5.10. Likewise, the four samples produced at ~250 °C are
displayed in Figure 5.11, and the four samples obtained at ~350 °C are shown in
Figure 5.12. When a laser light was passed through each sample vial, a beam
path could be seen – indicative of particles being present in all samples. As the
synthesis temperature increased, the colour of the samples became darker – this
could be an indication of a change to particle size, or a difference in particle
concentration.
Figure 5.10. Photograph of Fe2O3 Samples in First Temperature Set.
199
Figure 5.11. Photograph of Fe2O3 Samples in Second Temperature Set.
Figure 5.12. Photograph of Fe2O3 Samples in Third Temperature Set.
200
The T2 relaxation times of the 12 HEM samples are given in Table 5.9; these
were measured using the Halbach 0.5 T table-top scanner. Also shown in Table
5.9 are the r2 relaxivities of each of the samples, calculated using Equation 2
which was stated in Section 5.1.3. For a more straightforward comparison of the
different T2 times, the data is summarised as a bar chart in Figure 5.13.
Table 5.9. Summary of the measured T2 relaxation times for the Fe2O3
samples, precursor and water blank. The r2 values, calculated using
Equation 2, are also given.
T2 Time
Sample Code Error (±) r2 (mM-1 s-1)
(ms)
HEM A1 7.44 0.01 8.03
HEM A2 6.89 0.01 8.68
HEM A3 6.66 0.01 8.98
HEM A4 8.56 0.01 6.98
HEM B1 8.13 0.01 7.35
HEM B2 7.43 0.01 8.04
HEM B3 7.25 0.009 8.24
HEM B4 8.31 0.01 7.19
HEM C1 8.08 0.008 7.40
HEM C2 9.18 0.009 6.51
HEM C3 8.99 0.01 6.64
HEM C4 9.25 0.01 6.46
0.05 M Iron Nitrate

15.6 0.002 1.27
Nonahydrate
Water 2122 0.01 -
201
Given the T2 and r2 data presented in Table 5.9, the set of samples with the
highest and most promising relaxivities are HEM A – samples synthesised at the
lowest reaction temperature. These samples were therefore taken for addition
analysis using the Bruker 2.35 T scanner. The images obtained from these scans
are presented in Figure 5.14.
Figure 5.13. A bar chart summarising the r2 relaxivities of the 12
samples of synthesised Fe2O3, calculated using the T 2 times acquired
using the 0.5 T Halbach scanner, and Equation 2.
202
Figure 5.14. Images collected from the 2.35 T MRI scanner showing T 2
relaxation times (in ms) and R2 relaxation for a) Fe(NO3)3.9H2O, b) HEM
A1, c) HEM A2, d) HEM A3, e) HEM A4.
The data acquired from the T2 scans indicate that samples HEM A1-4 have the
highest relaxivity, compared to the samples synthesised at different
temperatures. The particle size data obtained using DLS is given in Figure 5.15
and shows that, with the exception of sample HEM A2 (PVP 10 coating), samples
produced at this temperature contain the smallest particles. From this, it can be
gathered that smaller particles result in increased proton relaxation and MR
signal decay, a trend consistent with what has previously been described in
literature. It can be seen from Figure 5.13 that for each of the capping agents
used, r2 decreases as synthesis temperature increases.
203
Figure 5.15. A chart summarising the particle size data obtained from
DLS analysis, for the 12 synthesised Fe2O3 samples.
The DLS data presented in Figure 5.15 shows that, in general, particle size
increases with an increase in temperature. This is a pattern seen previously for
Fe2O3 synthesised in the counter-current Nozzle reactor (Li, 2008). Nevertheless,
sample HEM A2 serves as an exception. One possible explanation for the large
particles measured in HEM A2 is that these large particles are globules of PVP
10, as opposed to Fe2O3. At a low reaction temperature of 150 °C, it is possible
(or altogether probable) that conversion rate of precursor to particles is low. As
such, the concentration ratio of PVP 10 to Fe 2O3 in the product is skewed in
favour of the former. This excess of polymer could result in it binding to itself
rather than to the nanoparticles.
The DLS data for samples containing sodium alginate as a capping agent (HEM
A4, HEM B4 and HEM C4) suggest the particles have a significantly larger
diameter than the uncoated particles. Parallel to this data, these samples have
the r2 relaxivities across their sample sets. Moreover, the MR image obtained for
HEM A4 using the Bruker 2.35 T scanner and displayed in Figure 5.14 e), shows
204
a colour gradient across the circle. This is because the particles, which are
causing signal decay, are sedimenting at the bottom of the sample cell.
Conversely, the particles are more dispersed in the uncoated sample HEM A1.
This data suggests that the addition of sodium alginate creates a “matrix” or
network around the particles, forming a large aggregate rather than coating
individual particles. This amorphous matrix surrounding the particles hinders the
superparamagnetic effect of the Fe2O3 nanoparticles, as well as enveloping
particles into a larger entity – which leads to sedimentation. Moreover, the
photos of samples HEM A4, HEM B4 and HEM C4 in Figure 5.10, Figure 5.11 and
Figure 5.12 respectively, show that the dispersions are more opaque than their
uncoated counterparts; this suggests particle flocculation is occurring which
could be due to a sodium alginate matrix.
In sample sets HEM A and HEM B, the samples containing dextran as a capping
agent (HEM A3 and HEM B3) yield the highest r2 relaxivities – this is
unsurprising as dextran is commonly used to coat superparamagnetic particles
for MRI. Interestingly, this is not the case for sample HEM C3, which has a lower
r2 relaxivity than the equivalent uncoated particles in HEM C1. This could be
attributed to thermal decomposition of dextran in the reactor during synthesis,
which has been reported to occur at temperatures ca. 300 °C (Carp et al., 2010)
– the temperature used to synthesise samples in the HEM C set. The DLS data
for sample HEM C3 also shows smaller particles than in HEM B3, and does not
follow the trend of increasing particle size with increasing temperature. This
result reinforces the notion of thermal breakdown; the resulting fragments of
dextran would not scatter light in the same way as if there were coated particles,
and the latter would have a larger hydrodynamic radius.
The XRD data for Samples HEM A1, HEM B1 and HEM C1 are given in Figure
5.16, along with the expected peak positions for Fe2O3. The three patterns do
205
not exhibit strong peaks, suggesting that the samples are quite amorphous. In
addition, the XRD data may be influenced by small crystallite sizes in the
samples, which would give rise to peaks with a broad width. Still, HEM A1 and
HEM C1 do contain peaks ca. 2θ = 33.1 and 35.7 °, which correspond to Fe2O3.
For sample HEM B1, these peaks are not obvious but signal intensity does
increase a little around these 2θ regions, suggesting that this phase may also be
present in this sample. In HEM C1, the extra peak at 2θ = 20.4° corresponds to
a peak for iron nitrate hydroxide hydrate, Fe 4NO3(OH)11.2H2O (ICCD PDF 44-
519). This composition is similar to the starting precursor, which suggests that
there may be incomplete conversion during the reaction.
Figure 5.16. XRD patterns obtained for samples HEM A1, HEM B1, and
HEM C1. At the bottom are the expected peak positions for Fe 2O3 (ICCD
PDF 86-550).
206
5.3.2 Magnetite, Fe3O4
The samples produced at a temperature ~350 °C (MAG F1-4) were a
similar colour to the precursor. When a laser light was passed through each
sample vial, no beam path could be seen, suggesting that no particles were
produced at this temperature. It is likely that this temperature is insufficient to
drive synthesis of Fe3O4. Therefore, samples MAG F1-4 were not characterised or
analysed further.
The samples synthesised at a temperature of ~400 °C (MAG G1-4) are shown in
Figure 5.17. All four samples were dark brown in colour and the particles
sedimented to the bottom of the vial over time. Due to this sedimentation, the
samples could not be analysed using DLS (the reasons for this were discussed in
Chapter 2 Section 2.6).
Figure 5.17. Photo of Fe3O4 Samples in Second Temperature Set.
207
The four samples, along with the precursor, were tested on the 0.5 T Halbach
MRI Scanner and their T2 times obtained. These values, as well as the T2 time of
water for comparison, are presented in Table 5.10. It can be seen that the T2
times for all four samples are significantly shorter than that of water, and much
shorter than that of the precursor – confirming that the synthesised particles are
having a positive effect on proton relaxation. The data shows that, while the
addition of PVP has little effect on the particle’s T 2 time (MAG G1 compared to
MAG G2), the incorporation of dextran shortens the T 2 time significantly (as seen
from MAG G3). This, in turn, increases the r2 value from 22.68 to 33.47 mm-1 s-
1
. On the contrary, the addition of sodium alginate appears to have a detrimental
effect on the relaxation effect, as the T 2 time for MAG G4 is much higher than
that of MAG G1.
Table 5.10. Summary of the T2 relaxation times for the Fe3O4 samples,
precursor and water blank, measured using the Halbach 0.5 T MRI
Scanner. The r2 values, calculated using Equation 2, are also given.
T2 Time
(ms)
MAG G1 2.642 0.01 22.68
MAG G2 2.634 0.009 22.75
MAG G3 1.791 0.007 33.47
MAG G4 4.84 0.02 12.37
0.05 M Ammonium Iron

44.28 0.02 0.44
Citrate
Water 2122 0.01 -
The Fe3O4 samples were tested using a 2.35 T Bruker MRI scanner; the obtained
images depicting proton T2 relaxation time (ms) and the corresponding R2
208
relaxation are given in Figure 5.18. In Figure 5.18 (a), the T2 relaxation time for
water is so high, that the image appears as if there is no sample present.
Likewise the R2 relaxation value is so low that there is no discernible difference
between the water sample and background.
Figure 5.18. Images collected from the 2.35 T MRI scanner showing T 2
relaxation times (in ms) and R2 relaxation for (a) water blank, (b) 0.025
M Ammonium Iron Citrate (precursor), (c) MAG G1, (d) MAG G2, (e)
MAG G3, and (f) MAG G4.
On the contrary, The T2 relaxation times for samples MAG G1-4 are much
shorter, and correspond to the values shown in Table 5.10. The non-circular
image seen in Figure 5.18 (f) generated by sample MAG G4 is cause by an air
bubble present in the sample vial during analysis, rather than the MR properties
of the sample itself. The colour of the obtained images for MAG G4 are
significantly different to those of MAG G1-3 - ascribed to its longer T2 relaxation
time, and smaller R2, compared to the other three samples. This is consistent
with the data obtained using the Halbach 0.5 T system.
The r2 relaxivity values can be calculated using the T2 times and the theoretical
concentration of Fe in the samples, and are summarised in Table 5.10. The
209
relaxivity of all the synthesised samples of Fe3O4 are lower than the commercial
contrast agents Feridex IV®, which has a relaxivity of 98.3 mM-1 s-1 (Wang,
2011), and Resovist®, which has an r2 value of 151 mM-1 s-1 (Wang, 2011). This
is most likely to be due to effects of particle coagulation, thereby eliminating the
benefits seen at the nanoscale. This is a common problem when synthesising
nanoparticles, especially Fe3O4, and so surfactants or capping agents need to be
introduced.
The addition of PVP to the synthesis reaction has little effect on relaxivity, and a
similar value to the uncapped sample is derived, 22.75 mM-1 s-1. This may be
because PVP is not having a very effective stabilising influence on the Fe 3O4
particles; or that the PVP is indeed preventing particle aggregation to some
degree, but also reducing the surface interactions of the particles.
Sample MAG G4, with sodium alginate as a capping agent, has a much longer T 2
time than the other samples, and consequently a lower relaxivity of 12.37 mM -1
s-1. This could be a result of the capping agent interacting with the surfaces of
the Fe3O4 particles and hindering the superparamagnetic properties.
Lastly, sample MAG G3 shows the most promise as a contrast agent, with a T 2
time of 1.79 ms and corresponding r2 of 33.47 mM -1 s-1 – a relaxivity
comparable to some of the commercial contrast agents highlighted in Section
5.1.5. In sample HEM C3, which was discussed in Section 5.3.1, a detrimental
effect on r2 relaxivity was observed with the incorporation of dextran as a
capping agent – this was attributed to possible thermal decomposition of the
surfactant. Interestingly, this effect is not seen in the case of MAG G3 in
comparison to MAG G1, the uncoated sample, despite reaction temperatures ca.
400 °C. Instead, dextran improves the r2. This difference in results could be due
to the phase change of water into the supercritical region during magnetite
210
synthesis compared to water being subcritical during synthesis of HEM C3. At the
supercritical state, water is less dense and in turn, the residence time of
particles/dextran in the reactor is shorter. This means dextran is exposed to the
high temperatures in the reactor for less time during synthesis of MAG G3 and
may not be enough time for thermal decomposition to occur.
While the r2 value for MAG G3 is promising, but lower than some commercial
contrast agents, it is derived from a theoretical concentration of Fe which
assumes 100% conversion during the synthesis reaction. Therefore, it is possible
that the actual concentration of Fe in the sample is lower, thus yielding a higher
relaxivity still. If the ratio of Fe3O4 and dextran were optimised to ensure there is
no particle aggregation at all, it is likely that the relaxivity could be enhanced
further.
As MAG G3 gave this promising T2 result, it was characterised further. The
sample was dried to a powder and analysed using XRD to ascertain its crystal
structure. XRD was also performed on sample MAG G1, for comparison. The XRD
patterns of both samples are given in Figure 5.19, with the reference peak
positions of cubic phase Fe3O4 also shown. All the peaks corresponding to the
reference phase are present in both samples, confirming that cubic phase Fe 3O4
is the predominant crystal structure for both MAG G1 and G3. Nevertheless,
there are differences between the two samples; the pattern for MAG G3 contains
a lot more noise than that of MAG G1. This may be due to the amorphous
dextran content in MAG G3, which is not present in MAG G1.
211
Figure 5.19. XRD patterns obtained for samples MAG G1 and MAG G3. At
the bottom are the expected peak positions for cubic Fe 3O4 (ICCD PDF
88-315).
The images of MAG G1 collected using TEM analysis are shown in Figure 5.20. It
can be seen that the particles are spherical in morphology with their diameters
measuring between 9 and 30 nm. TEM images collected at a higher
magnification (Figure 5.20 (b)) highlight the crystalline nature of the particles.
212
Figure 5.20. (a) and (b) are TEM images of MAG G1, uncoated Fe 3O4.
A TEM image collected for sample MAG G3 is given in Figure 5.21. It was seen
that particle morphology and size did not deviate significantly from those seen in
MAG G1. Yet the principle difference observed between the two samples was the
apparent presence of an amorphous coating around the Fe 3O4 particles in MAG
G3 – ascribed to dextran. It can be seen that while the dextran sufficiently
covers the entirety of the nanoparticles, it does appear that there is still some
degree of particle aggregation which the dextran has not completely prevented.
Here, the capping agent is enveloping the small clusters of particles, rather than
coating the individual particles as desired. This is an issue which can be
addressed by either changing (increasing) the concentration ratio of dextran to
iron precursor used, or altering the method of how the dextran is introduced to
coat the particles. This provides scope for future work.
213
Figure 5.21. TEM image of MAG G3, Fe3O4 nanoparticles coated in
dextran.
5.3.3 Gadolinium Hydroxide, Gd(OH)3
The first set of Gd(OH)3 samples synthesised with the water heater set
point at 385 °C all gave rise to transparent, colourless samples. When tested
with a laser pointer, no beam path could be seen – indicating no particles were
present. As such, it was concluded that the reaction temperature was not high
enough to produce particles and the samples were not characterised further.
The second set of samples manufactured with the water heater set point at
450 °C yielded the samples shown in Figure 5.22. Sample GAD E1 was
translucent with a white haze, and when tested with a laser pointer, showed a
positive beam path. GAD E3 and E4 were a more opaque white, while GAD E2
appeared to contain a slight yellow tinge. The particles in all four samples
214
sedimented over time, meaning particle sizing by DLS technique was not
possible. A day or so after synthesis, the particles in GAD E1 dissolved; this may
be due to the HNO3 which is produced as a by-product. Still, before this
occurred, a sample of the particles were dried and analysed by XRD.
Figure 5.22. Photo of the Second Set of Gd(OH)3 Samples.
The four samples of Gd(OH)3 produced in the second experimental set were
analysed using the Halbach 0.5 T scanner, along with the precursor, and the
obtained T2 times are shown in Table 5.11. Their r2 relaxivities were calculated
using the measured T2 times and Equation 2 which was listed in Section 5.1.3.
These are also displayed in Table 5.11. It can be seen the T2 time for the
precursor, gadolinium acetate hydrate, is the lowest value obtained for this
sample set, at 1.69 ms. Hence the calculated r2 relaxivity of 11.82 mM-1 s-1 is the
highest for all in this sample set. This is unsurprising, as the Gd 3+ ion has been
reported as having the highest degree of paramagnetism of all metal ions (Dutta
et al., 2008). Nevertheless, the cytotoxicity issues associated with Gd 3+ ions
215
which were previously described in Section 5.1.4 mean it cannot be used as a
free ion for clinical MRI applications.
Table 5.11 also shows that a relatively low T2 time and high r2 value is produced
from GAD E1, uncapped Gd(OH) 3 – 7.32 ms and 8.17 mM-1 s-1 respectively.
However, it was observed that the white particles which were produced in this
sample began to ‘disappear’ after a day of so. This is likely to be due to chemical
decomposition caused by HNO3 which is produced as a by-product.
Consequently, it is probable that particle breakdown occurred before and during
analysis. This means the Gd(OH) 3 particles were being dissolved to yield Gd3+
ions and give rise to the efficient signal decay in MRI scans. While a good T2 and
r2 result, this sample presents the same problem of cytotoxicity as the precursor
and cannot be considered as a potentially viable contrast agent.
Table 5.11. Summary of the measured T2 relaxation times for the
Gd(OH)3 samples, precursor and water blank. The r2 values, calculated
using Equation 2, are also given.
T2 Time
(ms)
GAD E1 7.32 0.01 8.17
GAD E2 17.2 0.02 3.46
GAD E3 9.952 0.01 6.00
GAD E4 16.26 0.02 3.66
0.025 M Gadolinium
1.69 0.001 11.82
Acetate Hydrate
Water 2122 0.01 -
216
Unlike sample GAD E1, samples GAD E2-4 did not exhibit visible signs of particle
decomposition after synthesis. The T2 and r2 values for GAD E2 and GAD E4 are
very similar, with GAD E2 possessing a relaxivity of 3.46 mM-1 s-1 and GAD E4 of
3.66 mM-1 s-1. These values, a fair degree lower than GAD E1 and the precursor,
suggests that PVP and sodium alginate are having an effect in coating particles
and preventing particle decomposition. Still, the extent of this positive effect
cannot be concluded from this data, as the T2 and r2 data may also be affected
by the capping agents themselves either enhancing or pacifying the
paramagnetic behaviour of the Gd(OH)3 particles.
Sample GAD E3, with dextran as the capping agent, shows a promising T 2 time
of 9.952 ms, which translates to an r2 of 6.00 mM-1 s-1. Like the other GAD
samples though, it cannot be determined from this data if the signal decay can
be entirely attributed to the synthesised particles, or if it due (even in part) to
the presence of Gd3+ ions present in the samples. To eliminate this possibility
during analysis in future work, samples should be washed prior to analysis.
The relaxivity of GAD E3, and those of GAD E2 and GAD E4, are quite
comparable to that of the commercial contrast agent Omniscan ®, which has an
r2 of 3.4 mM-1 s-1. As well as the possible presence of Gd 3+ ions in the samples
which has been discussed, it should also be noted that the Gd-based complexes
in Omniscan® contain tightly bound ligands to prevent leaching of toxic ions.
These could be partially diminishing the paramagnetic properties of Gd 3+ ions
and decreasing the r2 value. Nevertheless, the r2 values of the Gd(OH) 3 samples
described in the Chapter show promising r2 data which provides scope for
further future investigation.
To confirm the nanomaterial species being produced in this set of experiments,
XRD analysis was performed on sample GAD E1. The acquired pattern is shown
217
in Figure 5.23, with the peak positions for hexagonal Gd(OH) 3 also given. It can
be seen that all the expected peaks are indeed exhibited by the samples,
denoting that this is principle and sole crystalline phase in GAD E1. In addition,
the peaks are sharp and narrow, which suggests a high level of crystallinity and
probably large crystallites (See Chapter 2 for reasoning). With the exception of
different organic capping agents being introduced, the experimental parameters
for synthesis of GAD E2-4 did not change, so it is gathered that Gd(OH) 3 is
synthesised in the other three samples.
Figure 5.23. XRD Pattern obtained for GAD E1, with the expected peak
positions of hexagonal Gd(OH)3 (ICCD PDF 83-2037) shown below.
5.3.4 Cobalt Oxide, Co3O4
The synthesised sample of Co3O4 was opaque and black in colour, where
the particles remained in dispersion over time, i.e. the particles did not
sediment. The MRI properties of the sample, and its precursor, were tested using
218
the Halbach 0.5 T kit. The T2 times and respective r2 values are given in Table
5.12. The measured T2 time of Co3O4 was 578.1 ms and significantly longer than
those of Fe2O3 and Fe3O4 described in Sections 5.3.1 and 5.3.2 respectively. This
gave rise to a low r2 value of 0.30 mM-1 s-1; consequently, the sample of Co3O4
was not tested on the 2.35 T MRI scanner.
Table 5.12. Summary of the measured T2 relaxation times for the sample
of Co3O4, its precursor, and water blank. The r2 values, calculated using
Equation 2, are also given.
T2 Time
(ms)
Co3O4 578.1 0.2 0.30
0.05 M Cobalt Acetate

101.95 0.02 0.19
Tetrahydrate
Water 2122 0.01 -
This sample of Co3O4 was found to possess antiferromagnetic properties (Moro et
al., 2013); though T2-weighted contrast agents are usually superparamagnetic,
and T1-weighted agents have paramagnetic properties (Conn, 2009). This
explains the low r2 values. This sample of Co3O4 was produced for analysis of its
other properties which are described elsewhere (Moro et al., 2013) and so,
despite the poor MR properties, TEM and XRD analysis was conducted on the
sample. A darkfield TEM image of the particles is given in Figure 5.24, which
shows the cubic morphology of the particles. Using Digital Micrograph software
from Gatan, the edge lengths of 300 particles were measured and a mean length
of 7.03 nm was calculated. Using this data, the particle size distribution graph
displayed in Figure 5.25 was generated, showing the majority of nanoparticles
have an edge length between 2 and 10 nm.
219
Figure 5.24. Darkfield TEM image of sample Co3O4 showing particles with
cubic morphology.
Figure 5.25. Graph showing the particle size distribution of Co3O4
particles.
220
Figure 5.26 shows the XRD pattern of the sample, which matches the expected
pattern for cubic phase spinel cobalt oxide (ICCD PDF 9-418). No other peaks
were detected, indicating the purity of the sample. Xfit software was used to
calculate an average crystallite size of 12.3 nm, a value slightly larger than that
calculated from TEM analysis; this may be due to the amorphous nature of
smaller particles which are observed by TEM but not detected by XRD. Bragg′s
equation was used to calculate the d-spacing of each plane then checked and
verified with Celref software. Table 5.13 shows the calculated d-spacing and the
hkl plane for each peak. Considering the highest intensity peak at 2θ = 36.8031,
corresponding to the plane (311), the calculated d-spacing is 2.44Å (0.244 nm).
This value is confirmed by the high magnification TEM image shown in Figure
5.27. Furthermore, Celref software was used to refine the unit cell parameters of
the sample and yielded a lattice constant a = 8.0937 ± 0.004Å. This figure is
close to, but slightly larger than the theoretical cell unit size of 8.084 Å. The
expansion of lattice parameters in nanoparticles has been previously reported
and ascribed to the high surface to volume ratio which leads to a “stretched” unit
cell because of surface tension attraction between nanoparticles (Wei et al.,
2007).
221
Figure 5.26. The XRD pattern obtained for sample Co3O4. Peaks
correspond with the expected pattern of cubic Co 3O4 (ICCD PDF 009-
0418), which are shown below.
Table 5.13. The d-spacing for each plane in the Co3O4 crystal, and the
respective peak position, as calculated from the XRD pattern.
Peak Position, 2θ (°) (hkl) Plane d-spacing, dhkl (Å)
18.904 (111) 4.69
31.223 (220) 2.87
36.803 (311) 2.44
38.535 (222) 2.34
44.785 (400) 2.02
55.698 (422) 1.65
59.343 (511) 1.56
222
0.24nm
1 0 n m
Figure 5.27. High magnification HRTEM image of Co3O4 nanoparticles
5 n illustrating
m the interplanar spacing.
223
5.4 Conclusions
In this Chapter, the synthesis of Fe2O3, Fe3O4, Gd(OH)3 and Co3O4 were
described. With the exception of Co 3O4, synthesis was attempted using different
reactor temperatures. For Fe3O4 and Gd(OH)3, it was found that particles were
not produced at 350 °C but were at 400 °C; this indicates supercritical
conditions are required for particle formation. In addition, three different capping
agents (PVP 10, dextran and sodium alginate) were added to coat particles
during synthesis, as well as uncoated particles being made for comparison.
In the case of Fe2O3, the samples were manufactured at three different reaction
temperatures; DLS data showed a general trend of increasing particle size with
increasing synthesis temperature. Another general trend showed that smaller
particles exhibited higher r2 relaxivities. At the lower two reaction temperatures,
the highest relaxivities were produced by particles coated with dextran (HEM A3
had an r2 of 8.98 mM-1s-1, which HEM B3 was calculated as 8.24 mM-1s-1) –
consistent with previous reports in the literature that dextran improves
relaxivity. At the highest reaction temperature, the largest r2 value was
attributed to the sample of uncoated particles (HEM C1) rather than those coated
with dextran (HEM C3); this may be a result of thermal decomposition of dextran
caused by the reactor temperature. These r2 values for Fe2O3 particles are
higher than some commercial contrast agents, such as Omniscan at 3.4 mM-1s-1,
but lower than most.
The synthesised samples of Fe3O4 which contained particles were analysed using
a 0.5 T MRI scanner to acquire T2 relaxation times. A 2.35 T scanner was then
used to acquire MR images showing relaxation. The obtained data showed that
MAG G3, Fe3O4 with dextran as a capping agent, had the lowest T 2 time and
224
consequently, the highest r2 relaxivity of 33.47 mM -1 s-1. This complies with
results from the literature. The r2 value is also comparable to some commercial
contrast agents, such as Clariscan which has a relaxivity of 35.3 mM-1s-1. Unlike
the sample of hematite, use of dextran at a high reaction temperature was not
detrimental to the r2 value; it is plausible that thermal decomposition is not a
problem in this case due to the supercritical nature of water meaning lower
solvent density and lower residence times. Obtained TEM images and XRD
patterns showed that dextran did not affect crystal phase or primary particle
size, but did form a partial coating around particles. The ratio between particles
and capping agent can be optimised in future work to ensure ideal particle
coating, which will then prevent particle coagulation and may enhance relaxivity.
PVP 10 did not appear to affect relaxivity, as the r2 value was similar to that of
uncoated particles, while sodium alginate had a detrimental effect on particle
relaxivity. This may be due to sodium alginate enveloping particles and
diminshing their superparamagnetic properties.
The relaxivity values calculated in this Chapter were based on theoretical Fe
concentrations and assumed 100% reaction conversion; the actual
concentrations of Fe could be experimentally measured to give more accurate r2
values. These may actually be higher than those presented here.
Of the four samples of Gd(OH)3 synthesised, the uncoated particles were not
chemically stable and became digested by the acidic by-product after a few days.
However, the obtained XRD data of the isolated particles confirmed the principle
crystal phase to be hexagonal Gd(OH)3. Gd3+ ions are known to be highly
paramagnetic and so the precursor and uncoated particles displayed high r2
relaxivities. However, the cytotoxic nature of the cation means it cannot be used
as a contrast agent. Sample GAD E3, Gd(OH)3 coated with dextran exhibited a
promising T2 relaxation time; further studies should focus on analysis of washed
225
samples to eliminate the possibility of Gd3+ ions (which could be present in the
unwashed sample) contributing to relaxation. Meanwhile, samples containing
PVP 10 and sodium alginate gave the lowest r2, but these were still values
comparable to those previously reported for Gd-based materials. So, these
maintain potential as contrast agent materials for further investigation. These
materials could be characterised using T1-weighted scans as well, in order to
gather information about their effects on spin-lattice relaxation.
Lastly, the Nozzle reactor was used to synthesise Co 3O4 for a separate study, but
was also analysed for possible application in MRI. From the acquired TEM and
XRD data, particles are crystalline with cubic morphology. TEM analysis yielded a
mean edge length of 7.03 nm while XRD analysis presented an average
crystallite size of 12.3 nm; this discrepancy could be due to smaller amorphous
particles which are not detected by XRD. Analysis of Co 3O4 on the 0.5 T MRI
scanner showed poor T2 relaxation times; the particles were then found to be
antiferromagnetic which explained this data.
To summarise, dextran was generally shown to be the most effective capping
agent in terms of improving r2 relaxivity of samples. Conversely, sodium
alginate had a detrimental effect on relaxivity. Hematite, Fe2O3; magnetite,
Fe3O4; and gadolinium hydroxide, Gd(OH)3 produced using the Nozzle reactor
have been shown to exhibit properties for MRI which are comparable to
commercial contrast agents. As such, these materials have the potential for
further research and development which can serve to enhance their r2
relaxivities further.
226
Chapter 6
6 Conclusions and Future Work
6.1 Conclusions
The work described in this Thesis follows on from that of previous
researchers using the Nozzle reactor, including Li (Li, 2008), Hobbs (Hobbs et
al., 2009) and Aksomaityte (Aksomaityte et al., 2013, Lester et al., 2012). Since
the reactor geometry was invented in 2005 (Lester and Azzopardi, 2005), the
wide range of materials which could potentially be synthesised using it, has only
been touched upon. Furthermore, the industries and applications which could (or
already do) benefit from the implementation of nanomaterials is vast. In this
Thesis, the objective was to investigate the synthesis of, and subsequently
characterise, nanomaterials which could be applied into biological or biomedical
fields.
Firstly, Chapter 1 showed that the Nozzle reactor could be used to reproducibly
synthesise HA with either platelet, rod or tube morphology. Data obtained from
this investigation suggested that NH4OH plays several roles during HA synthesis,
from directing precursors towards the formation of HA over other CaP minerals,
to promoting the formation particular morphologies. In particular, it was found
that in the presence of NH4OH, rods were produced instead of sheets at 200˚C,
and tubes instead of rods at 350˚C. While more data needs to be gathered
before a formation mechanism can be concluded, it appears that the
mechanisms for the three morphologies are linked. Cell proliferation and bone
227
Chapter 6. Conclusions and Further Work
nodule assays performed using HA from the Nozzle reactor indicated that the
nanoparticles did not have cytotoxic effects and appeared to promote bone
nodule formation.
Chapter 1 also detailed experiments to introduce zinc and simvastatin into the
structure of HA. It was found that the morphology of synthesised particles did
not change by adding other compounds into the reaction. However, it was
unclear from the obtained data whether simvastatin was successfully
incorporated into the structure. By adding a zinc salt into the synthesis method,
parascholzite (CaZn2(PO4)2.2H2O) was formed. This could have potential
applications in bone based materials where zinc may provide antimicrobial
properties (Stanić et al., 2010).
In Chapter 4, synthesis of CdS yielded bright yellow particles of diameters 20 to
50 nm - consistent with what was others had observed previously. The
fluorometry data collected for the CdS sample were inconclusive within the scope
of this Thesis. The methodology for CdS synthesis was adapted to produce ZnS
nanoparticles – a material which had not been previously produced by a
continuous hydrothermal route. For both CdS and ZnS, a minimum reaction
temperature of 400 °C was required to yield particles. The ZnS sample was
shown to contain a mixture of hexagonal and cubic phases of ZnS, with primary
particle sizes <15nm; these had aggregated into clusters with diameters of 50 to
100 nm. Under the same experimental conditions, but using thiourea as a
sulphur source, ZnS was not produced – showing that thiocarbohydrazide and
thiourea cannot be used interchangeably.
Also in Chapter 4, the synthesis of ZnS with 3 different capping agents was
reported. Structural and photoluminescent properties of these products were
compared to unmodified ZnS. It was found that the addition of 3-MPA to the
228
synthesis method yielded particles with an average crystallite size of 3.1 nm,
with a PL emission peak in the violet or blue light range (450 nm). This was
compared to a peak at 558 nm, in the green light range, for Uncapped ZnS
nanoparticles. This emission data for ZnS:3-MPA is in a wavelength range
comparable to previous reports for cadmium-based QDs (Saraswathi Amma et
al., 2008), thus there is the potential to develop ZnS as a heavy-metal free QD
alternative. Furthermore, the zeta potential data highlighted a change in the IEP
of particles in response to capping with 3-MPA; uncapped particles had an IEP at
pH 7.8 while ZnS:3MPA 5/5 had an IEP at pH 4.1 and a stable zeta potential at
pH 7.
Coating particles with PVP had little observed effect on structural properties of
particles. In addition, the zeta potential data was inconclusive. However, the PL
data showed a shift in the optimum emission wavelength compared to uncoated
particles, with the peak reached at 572 nm, corresponding to yellow light.
Meanwhile, the obtained data showed that the addition of NaOH to synthesis did
not have any discernible effect on the structural or PL characteristics. It is
inferred that differences in emission peak wavelength between the four samples
are linked to differences in particle band gap energies.
In Chapter5, the synthesis of Fe2O3, Fe3O4, Gd(OH)3 and Co3O4 were described.
With the exception of Co3O4, synthesis was attempted using different reactor
temperatures. For Fe3O4 and Gd(OH)3, it was found that particles were not
produced at temperatures below 400 °C, indicating supercritical conditions are
required for particle formation. Three different capping agents (PVP 10, dextran
and sodium alginate) were added to coat Fe 2O3, Fe3O4 and Gd(OH)3 particles
during synthesis, as well as uncoated particles being made for comparison.
229
Samples of Fe2O3 were synthesised at three different reaction temperatures; a
general trend of increasing particle size with increasing temperature was
observed. Moreover, smaller particles typically exhibited higher r2 relaxivities.
The highest relaxivity seen for the Fe2O3 sample set was ascribed to particles
coated in dextran and synthesised at the lowest reaction temperature of
~150 °C. Still, all of the uncoated particles produced at the three different
temperatures appeared to be quite amorphous when analysed using XRD.
Likewise for Fe3O4, particles coated with dextran displayed the highest relaxivity.
It was shown that the dextran coating did not affect primary particle size or
crystallinity, but did partially coat individual particles.
Without the use of capping agents, Gd(OH) 3 produced in the scope of this Thesis
was chemically unstable and dissolved. As Gd3+ ions are known to be highly
paramagnetic, the uncapped sample exhibited high relaxivity. Particles coated
with dextran also showed high relaxivity. However, samples were not washed
prior to analysis, so the effect of Gd3+ ions - which may be present in the sample
- on relaxivity is unknown.
For all samples, sodium alginate had a detrimental effect on r2 relaxivity. It is
thought that the alginate may have a dampening effect on the
superparamagnetic properties of the iron oxide particles. Hematite, Fe 2O3;
magnetite, Fe3O4; and gadolinium hydroxide, Gd(OH)3 produced using the Nozzle
reactor have been shown to exhibit properties for MRI which are comparable to
commercial contrast agents. As such, these materials have the potential for
further research and development which can serve to enhance their r2
relaxivities even more.
230
Synthesis of Co3O4 was also reported in Chapter 5, but the particles were found
to be antiferromagnetic so did not display good relaxivity data. Still, this material
may find use in other applications (Moro et al., 2013).
6.2 Future Work
Following on from the work on HA reported in Chapter 1, further attempts
to incorporate drugs into the structure of HA can be made. Here, it is important
to use compounds with high thermal stability. The work in this Thesis developed
issues due to the low concentration of statin used, so this should be increased in
future work. As well as FTIR or ToF-SIMS, thermal analysis such as TGA could
provide additional data to ascertain the presence of organic compounds in
samples.
Subsequent to the addition of Zn(NO3)2.6H2O to HA synthesis which resulted in
the novel synthesis of parascholzite, the ratio of Ca:Zn can be investigated
further, to determine if other calcium-zinc phosphates can be produced using the
Nozzle reactor.
The work described in Chapters 4 and 5 began to look at the use of stabilisers or
capping agents to coat nanoparticles and prevent their aggregation. Meanwhile,
the difficulty in choosing suitable surfactants for each particle, which are also
compatible with the synthesis reaction system, was highlighted in Chapter 1.
Nevertheless, the work in this Thesis described the positive effects of using 3-
MPA as a surfactant for ZnS, and dextran for iron oxide nanoparticles (Fe 2O3 and
Fe3O4). There is scope for determining the optimum concentration ratios between
these particles and surfactants. By improving the coating of these nanoparticles,
their properties could be enhanced even further.
231
Also in Chapter 5, the relaxivities of samples was calculated based upon
theoretical concentrations of Fe or Gd. Elemental analysis, such as atomic
adsorption (AA) or inductively coupled plasma (ICP) could provide experimental
concentrations for more accurate relaxivity values.
The functional properties of HA, ZnS, Fe2O3, Fe3O4 and Gd(OH)3 were tested in
this Thesis. However, other such tests could be performed to acquire more
information on how these nanomaterials would perform if applied to commercial
products. For example, more in depth cell assays or using other cell lines would
yield additional cytotoxicity data for HA; tests could be carried out to determine
the electroluminescent properties of ZnS; and T1-weighted scans could be used
on Fe2O3, Fe3O4 and Gd(OH)3, in order to gather information about their effects
on spin-lattice relaxation.
232
7 Bibliography
ACTION, P. C. 2013. Pancreatic Cancer Action - MRI Scans [Online]. Available:

www.pancreaticcanceraction.org.
ADSCHIRI, T., HAKUTA, Y. & ARAI, K. 2000. Hydrothermal synthesis of metal
oxide fine particles at supercritical conditions. Industrial & Engineering
Chemistry Research, 39, 4901-4907.
ADSCHIRI, T., HAKUTA, Y., SUE, K. & ARAI, K. 2001. Hydrothermal synthesis of
metal oxide nanoparticles at supercritical conditions. Journal of
Nanoparticle Research, 3, 227-235.
ADSCHIRI, T., KANAZAWA, K. & ARAI, K. 1992. Rapid and continuous
hydrothermal crystallization of metal-oxide particles in supercritical
water. Journal of the American Ceramic Society, 75, 1019-1022.
ADSCHIRI, T., LEE, Y. W., GOTO, M. & TAKAMI, S. 2011. Green materials
synthesis with supercritical water. Green Chemistry, 13, 1380-1390.
AG, S. 2013. Resovist: Information for Radiologists. Available: http://radiologie-
uni-frankfurt.de/sites/radiologie-uni-
frankfurt.de/content/e43/e2321/e2331/resofinal_eng.pdf.
AIMABLE, A., AYMES, D., BERNARD, F. & LE CRAS, F. 2009a. Characteristics of
LiFePO4 obtained through a one step continuous hydrothermal synthesis
process working in supercritical water. Solid State Ionics, 180, 861-866.
AIMABLE, A., MUHR, H., GENTRIC, C., BERNARD, F., LE CRAS, F. & AYMES, D.
2009b. Continuous hydrothermal synthesis of inorganic nanopowders in
supercritical water: Towards a better control of the process. Powder
Technology, 190, 99-106.
AIMABLE, A., XIN, B., MILLOT, N. & AYMES, D. 2008. Continuous hydrothermal
synthesis of nanometric BaZrO3 in supercritical water. Journal of Solid
State Chemistry, 181, 183-189.
AKSOMAITYTE, G., POLIAKOFF, M. & LESTER, E. 2013. The production and
formulation of silver nanoparticles using continuous hydrothermal
synthesis. Chemical Engineering Science, 85, 2-10.
AL-DURI, B., PINTO, L., ASHRAF-BALL, N. H. & SANTOS, R. C. D. 2008. Thermal
abatement of nitrogen-containing hydrocarbons by non-catalytic
supercritical water oxidation (SCWO). Journal of Materials Science, 43,
1421-1428.
ALBERTS, B., JOHNSON, A., LEWIS, J., RAFF, M., ROBERTS, K. & WALTER, P.
2002. The Shape and Structure of Proteins. Molecular Biology of the Cell.
4th ed. New York: Garland Science.
ALBREKTSSON, ALBREKTSSON, T., JOHANSSON & JOHANSSON, C. 2001.
Osteoinduction, osteoconduction and osseointegration. European Spine
Journal, 10, S96-S101.
AMAG PHARMACEUTICALS, I. 2013. AMAG Pharmacuticals [Online]. Available:
http://www.amagpharma.com/about/partners.php 2013].
AMENDOLA, V., MENEGHETTI, M., GRANOZZI, G., AGNOLI, S., POLIZZI, S.,
RIELLO, P., BOSCAINI, A., ANSELMI, C., FRACASSO, G., COLOMBATTI,
M., INNOCENTI, C., GATTESCHI, D. & SANGREGORIO, C. 2011. Top-down
synthesis of multifunctional iron oxide nanoparticles for macrophage
labelling and manipulation. Journal of Materials Chemistry, 21, 3803-
3813.
AMIRI, S. & SHOKROLLAHI, H. 2013. The role of cobalt ferrite magnetic
nanoparticles in medical science. Materials Science and Engineering C,
33, 1-8.
APTE, S. K., GARAJE, S. N., ARBUJ, S. S., KALE, B. B., BAEG, J. O., MULIK, U.
P., NAIK, S. D., AMALNERKAR, D. P. & GOSAVI, S. W. 2011. A novel
233
Bibliography
template free, one pot large scale synthesis of cubic zinc sulfide
nanotriangles and its functionality as an efficient photocatalyst for
hydrogen production and dye degradation. Journal of Materials
Chemistry, 21, 19241-19248.
AYMES, D., ARIANE, M., BERNARD, F., MUHR, H. & DEMOISSON, F. 2011.
Particle synthesis by means of the thermal hydrolysis of mineral
precursors. PCT/FR2010/051520.
AZIZIAN-KALANDARAGH, Y. & KHODAYARI, A. 2010. Aqueous synthesis and
characterization of nearly monodispersed ZnS nanocrystals. Physica
Status Solidi (A) Applications and Materials Science, 207, 2144-2148.
BAE, H., AHMAD, T., RHEE, I., CHANG, Y., JIN, S.-U. & HONG, S. 2012. Carbon-
coated iron oxide nanoparticles as contrast agents in magnetic resonance
imaging. Nanoscale Research Letters, 7, 44.
BALLMAN, A. A. & LAUDISE, R. A. 1963. Hydrothermal growth. The Art and
Science of Growing Crystals, 231-251.
BAMMER, R., SKARE, S., NEWBOULD, R., LIU, C., THIJS, V., ROPELE, S.,
CLAYTON, D. B., KRUEGER, G., MOSELEY, M. E. & GLOVER, G. H. 2005.
Foundations of advanced magnetic resonance imaging. NeuroRx, 2, 167-
196.
BAO, X., LIN, M., KOH, H. Z. & ZHANG, Q. 2006. Preparation of Iron Oxide and
Iron Oxide/Silicon Oxide Nanoparticles Via Water-in-Oil Microemulsion.
Ceramic Nanomaterials and Nanotechnology II. John Wiley & Sons, Inc.
BASA, S., MUNIYAPPAN, T., KARATGI, P., PRABHU, R. & PILLAI, R. 2008.
Production and in vitro characterization of solid dosage form incorporating
drug nanoparticles. Drug Development and Industrial Pharmacy, 34,
1209-1218.
BASF. 2013. Further research on effects of nanomaterials [Online].
http://www.basf.com/group/pressrelease/P-13-323. Available:
http://www.basf.com/group/pressrelease/P-13-323.
BATEER, B., TIAN, C., QU, Y., DU, S., TAN, T., WANG, R., TIAN, G. & FU, H.
2013. Facile synthesis and shape control of Fe3O4 nanocrystals with good
dispersion and stabilization. CrystEngComm, 15, 3366-3371.
BAVYKIN, D. V., PARMON, V. N., LAPKIN, A. A. & WALSH, F. C. 2004. The effect
of hydrothermal conditions on the mesoporous structure of TiO 2
nanotubes. Journal of Materials Chemistry, 14, 3370-3377.
BEZZI, G., CELOTTI, G., LANDI, E., LA TORRETTA, T. M. G., SOPYAN, I. &
TAMPIERI, A. 2003. A novel sol-gel technique for hydroxyapatite
preparation. Materials Chemistry and Physics, 78, 816-824.
BHUSHAN, B. 2010. Springer Handbook of Nanotechnology, Springer.
BIO-RAD 2013. CHTTM Ceramic Hydroxyapatite. In: BIO-RAD (ed.).
http://www.bio-
rad.com/webroot/web/pdf/psd/literature/Bulletin_5667.pdf.
BJØRNERUD, A., JOHANSSON, L. O. & AHLSTRÖM, H. K. 2001. Pre-clinical
results with clariscan™ (NC100150 Injection); experience from different
disease models. Magnetic Resonance Materials in Physics, Biology and
Medicine, 12, 99-103.
BLOCK, J. E. & THORN, M. R. 2000. Clinical indications of calcium-phosphate
biomaterials and related composites for orthopedic procedures. Calcified
Tissue International, 66, 234-238.
BLOOD, P. J., DENYER, J. P., AZZOPARDI, B. J., POLIAKOFF, M. & LESTER, E.
2004. A versatile flow visualisation technique for quantifying mixing in a
binary system: Application to continuous supercritical water hydrothermal
synthesis (SWHS). Chemical Engineering Science, 59, 2853-2861.
BOLDRIN, P., HEBB, A. K., CHAUDHRY, A. A., OTLEY, L., THIEBAUT, B., BISHOP,
P. & DARR, J. A. 2007. Direct synthesis of nanosized NiCo2O4 spinel and
related compounds via continuous hydrothermal synthesis methods.
Industrial and Engineering Chemistry Research, 46, 4830-4838.
234
Bibliography
BROWN, M. A. & SEMELKA, R. C. 2010. MRI: Basic Principles and Applications,

Wiley-Blackwell.
BRUCHEZ JR, M., MORONNE, M., GIN, P., WEISS, S. & ALIVISATOS, A. P. 1998.
Semiconductor nanocrystals as fluorescent biological labels. Science, 281,
2013-2016.
BUFFAT, P. & BOREL, J.-P. 1976. Size effect on the melting temperature of gold
particles. Physical Review A, 13, 2287-2298.
BYERS, R. J. & HITCHMAN, E. R. 2011. Quantum Dots Brighten Biological
Imaging. Progress in Histochemistry and Cytochemistry, 45, 201-237.
CABANAS, A., DARR, J. A., LESTER, E. & POLIAKOFF, M. 2000. A continuous and
clean one-step synthesis of nano-particulate Ce(1- x)Zr(x)O2 solid
solutions in near-critical water. Chemical Communications, 901-902.
CABAÑAS, A., LI, J., BLOOD, P., CHUDOBA, T., LOJKOWSKI, W., POLIAKOFF, M.
& LESTER, E. 2007. Synthesis of nanoparticulate yttrium aluminum
garnet in supercritical water-ethanol mixtures. Journal of Supercritical
Fluids, 40, 284-292.
CABAÑAS, A. & POLIAKOFF, M. 2001. The continuous hydrothermal synthesis of
nano-particulate ferrites in near critical and supercritical water. Journal of
Materials Chemistry, 11, 1408-1416.
CARBONARO, L. A., PEDICONI, F., VERARDI, N., TRIMBOLI, R. M., CALABRESE,
M. & SARDANELLI, F. 2011. Breast MRI using a high-relaxivity contrast
agent: An overview. American Journal of Roentgenology, 196, 942-955.
CARP, O., PATRON, L., CULITA, D. C., BUDRUGEAC, P., FEDER, M. &
DIAMANDESCU, L. 2010. Thermal analysis of two types of dextran-coated
magnetite. Journal of Thermal Analysis and Calorimetry, 101, 181-187.
CASULA, M. F., CORRIAS, A., AROSIO, P., LASCIALFARI, A., SEN, T., FLORIS, P.
& BRUCE, I. J. 2011. Design of water-based ferrofluids as contrast agents
for magnetic resonance imaging. Journal of Colloid and Interface Science,
357, 50-55.
CHAN, W. C. W. & NIE, S. 1998. Quantum dot bioconjugates for ultrasensitive
nonisotopic detection. Science, 281, 2016-2018.
CHANG, M. L. & HOU, J. K. 2011. Cancer risk related to gastrointestinal
diagnostic radiation exposure. Current Gastroenterology Reports, 13,
449-457.
CHAUDHRY, A. A., HAQUE, S., KELLICI, S., BOLDRIN, P., REHMAN, I., KHALID,
F. A. & DARR, J. A. 2006. Instant nano-hydroxyapatite: A continuous and
rapid hydrothermal synthesis. Chemical Communications, 2286-2288.
CHEARY, R. W. & COELHO, A. A. 1996. XFIT. CCP14 Powder Diffraction Library:
Engineering and Physical Sciences Research Council, Daresbury
Laboratory, Warrington, England.
CHEKINA, N., HORAK, D., JENDELOVA, P., TRCHOVA, M., BENES, M. J., HRUBY,
M., HERYNEK, V., TURNOVCOVA, K. & SYKOVA, E. 2011. Fluorescent
magnetic nanoparticles for biomedical applications. Journal of Materials
Chemistry, 21, 7630-7639.
CHEN, N., HE, Y., SU, Y., LI, X., HUANG, Q., WANG, H., ZHANG, X., TAI, R. &
FAN, C. 2012. The cytotoxicity of cadmium-based quantum dots.
Biomaterials, 33, 1238-1244.
CHEN, Z., CAO, Y., QIAN, J., AI, X. & YANG, H. 2010. Facile synthesis and stable
lithium storage performances of Sn-sandwiched nanoparticles as a high
capacity anode material for rechargeable Li batteries. Journal of Materials
Chemistry, 20, 7266-7271.
CHOI, H., VERIANSYAH, B., KIM, J., KIM, J. D. & KANG, J. W. 2010. Continuous
synthesis of metal nanoparticles in supercritical methanol. Journal of
Supercritical Fluids, 52, 285-291.
CHON CHEN, C., CHENG, C. H. & LIN, C. K. 2013. Template assisted fabrication
of TiO2 and WO3 nanotubes. Ceramics International, 39, 6631-6636.
CONN, P. M. 2009. Essential Bioimaging Methods, Elsevier Science.
235
Bibliography
CORSO, D., CRUPI, I., AMMENDOLA, G., LOMBARDO, S. & GERARDI, C. 2003.
Programming options for nanocrystal MOS memories. Materials Science
and Engineering C, 23, 687-689.
COVALIU, C. I., BERGER, D., MATEI, C., DIAMANDESCU, L., VASILE, E.,
CRISTEA, C., IONITA, V. & IOVU, H. 2011. Magnetic nanoparticles coated
with polysaccharide polymers for potential biomedical applications.
Journal of Nanoparticle Research, 13, 6169-6180.
DARBANDI, M., STROMBERG, F., LANDERS, J., RECKERS, N., SANYAL, B.,
KEUNE, W. & WENDE, H. 2012. Nanoscale size effect on surface spin
canting in iron oxide nanoparticles synthesized by the microemulsion
method. Journal of Physics D: Applied Physics, 45.
DASGUPTA, S., BANERJEE, S. S., BANDYOPADHYAY, A. & BOSE, S. 2010. Zn-
and Mg-doped hydroxyapatite nanoparticles for controlled release of
protein. Langmuir, 26, 4958-4964.
DASOG, M. & VEINOT, J. G. C. 2012. Solid-state synthesis of luminescent silicon
nitride nanocrystals. Chemical Communications, 48, 3760-3762.
DAWSON, W. J. 1988. Hydrothermal synthesis of advanced ceramic powders.
American Ceramic Society Bulletin, 67, 1673-1678.
DE LIMA, I. R., ALVES, G. G., DE OLIVEIRA FERNANDES, G. V., DIAS, E. P., DE
ALMEIDA SOARES, G. & GRANJEIRO, J. M. 2010. Evaluation of the in vivo
biocompatibility of hydroxyapatite granules incorporated with zinc ions.
Materials Research, 13, 563-568.
DEMAZEAU, G. Solvothermal and Hydrothermal Processes: Main Physico-
Chemical Factors Involved and New Trends. International Solvothermal
and Hydrothermal Association Conference, 2010 Beijing, China.
DEMOISSON, F., ARIANE, M., LEYBROS, A., MUHR, H. & BERNARD, F. 2011.
Design of a reactor operating in supercritical water conditions using CFD
simulations. Examples of synthesized nanomaterials. Journal of
DICKERSON, B. D. 2005. Organometallic synthesis kinetics of CdSe quantum
dots. Virginia Polytechnic Institute and State University.
DING, J., TAO, K., LI, J., SONG, S. & SUN, K. 2010. Cell-specific cytotoxicity of
dextran-stabilized magnetite nanoparticles. Colloids and Surfaces B:
Biointerfaces, 79, 184-190.
DURÁN, J. D. G., GUINDO, M. C., DELGADO, A. V. & GONZÁLEZ-CABALLERO, F.
1995. Stability of monodisperse zinc sulfide colloidal dispersions.
Langmuir, 11, 3648-3655.
DUTTA, S., PARK, J. A., JUNG, J. C., CHANG, Y. & KIM, T. J. 2008. Gd-complexes
of DTPA-bis(amide) conjugates of tranexamic acid and its esters with high
relaxivity and stability for magnetic resonance imaging. Dalton
Transactions, 2199-2206.
EARL, J. S., WOOD, D. J. & MILNE, S. J. 2006. Hydrothermal synthesis of
hydroxyapatite. Journal of Physics: Conference Series, 26, 268-271.
ELOUALI, S., BLOOR, L. G., BINIONS, R., PARKIN, I. P., CARMALT, C. J. & DARR,
J. A. 2012. Gas sensing with nano-indium oxides (In 2O 3) prepared via
continuous hydrothermal flow synthesis. Langmuir, 28, 1879-1885.
FANG, Z. 2010. Rapid Production of Micro- and Nano-particles Using Supercritical
Water, Springer Berlin Heidelberg.
FANUN, M. 2010. Microemulsions: Properties and Applications, Taylor & Francis.
FIGGEMEIER, E., KYLBERG, W., CONSTABLE, E., SCARISOREANU, M.,
ALEXANDRESCU, R., MORJAN, I., SOARE, I., BIRJEGA, R., POPOVICI, E.,
FLEACA, C., GAVRILA-FLORESCU, L. & PRODAN, G. 2007. Titanium
dioxide nanoparticles prepared by laser pyrolysis: Synthesis and
photocatalytic properties. Applied Surface Science, 254, 1037-1041.
FIRMANSYAH, D. A., KIM, S. G., LEE, K. S., ZAHAF, R., KIM, Y. H. & LEE, D.
2012. Microstructure-controlled aerosol-gel synthesis of ZnO quantum
dots dispersed in SiO 2 nanospheres. Langmuir, 28, 2890-2896.
236
Bibliography
FLEGLER, S., HECKMAN JR, J. & KLOMPARENS, K. 1995. Scanning and

transmission electron microscopy: an introduction, Oxford University
Press.
FRENKEL, J. & DORFMAN, J. 1930. Spontaneous and induced magnetisation in
ferromagnetic bodies [1]. Nature, 126, 274-275.
GANGULI, A. K., AHMAD, T., VAIDYA, S. & AHMED, J. 2008. Microemulsion route
to the synthesis of nanoparticles. Pure and Applied Chemistry, 80, 2451-
2477.
GEINGUENAUD, F., SOUISSI, I., FAGARD, R., MOTTE, L. & LALATONNE, Y. 2012.
Electrostatic assembly of a DNA superparamagnetic nano-tool for
simultaneous intracellular delivery and in situ monitoring. Nanomedicine:
Nanotechnology, Biology, and Medicine, 8, 1106-1115.
GERMAN, R. M. 2010. Coarsening in sintering: Grain shape distribution, grain
size distribution, and grain growth kinetics in solid-pore systems. Critical
Reviews in Solid State and Materials Sciences, 35, 263-305.
GESZKE-MORITZ, M. & MORITZ, M. 2013. Quantum dots as versatile probes in
medical sciences: Synthesis, modification and properties. Materials
Science and Engineering: C, 33, 1008-1021.
GIMENO-FABRA, M., MUNN, A. S., STEVENS, L. A., DRAGE, T. C., GRANT, D. M.,
KASHTIBAN, R. J., SLOAN, J., LESTER, E. & WALTON, R. I. 2012. Instant
MOFs: Continuous synthesis of metal-organic frameworks by rapid
solvent mixing. Chemical Communications, 48, 10642-10644.
GLADSTONE, H. B., MCDERMOTT, M. W. & COOKE, D. D. 1995. Implants for
cranioplasty. Otolaryngologic Clinics of North America, 28, 381-400.
GRIFFITH, E. M. & DANILATOS, G. D. 1993. Environmental scanning electron
microscopy, Wiley-Liss.
GRIFFITHS, S. L. & CARTMELL, S. H. 2007. Use of statins for enhancing bone-
tissue-engineered grafts. European Journal of Plastic Surgery, 1-7.
GRUAR, R. I., TIGHE, C. J. & DARR, J. A. 2013. Scaling-up a confined jet reactor
for the continuous hydrothermal manufacture of nanomaterials. Industrial
and Engineering Chemistry Research, 52, 5270-5281.
GRUAR, R. I., TIGHE, C. J., MUIR, J., KITTLER, J. T., WODJAK, M., KENYON, A. J.
& DARR, J. A. 2012. Continuous hydrothermal synthesis of surface-
functionalised nanophosphors for biological imaging. RSC Advances, 2,
10037-10047.
HABRAKEN, W. J. E. M., WOLKE, J. G. C. & JANSEN, J. A. 2007. Ceramic
composites as matrices and scaffolds for drug delivery in tissue
engineering. Advanced Drug Delivery Reviews, 59, 234-248.
HAKUTA, Y., ONAI, S., TERAYAMA, H., ADSCHIRI, T. & ARAI, K. 1998.
Production of ultra-fine ceria particles by hydrothermal synthesis under
supercritical conditions. Journal of Materials Science Letters, 17, 1211-
1213.
HAKUTA, Y., URA, H., HAYASHI, H. & ARAI, K. 2005. Effects of hydrothermal
synthetic conditions on the particle size of γ-AlO(OH) in sub and
supercritical water using a flow reaction system. Materials Chemistry and
Physics, 93, 466-472.
HALLMANN, S., FINK, M. J. & MITCHELL, B. S. 2011. The mechanochemical
formation of functionalized semiconductor nanoparticles for biological,
electronic and superhydrophobic surface applications. Houston, TX.
HANWHA. 2013. Hanwha Chemical Co. Ltd [Online]. Available:
http://hcc.hanwha.co.kr/eng/index_eng.jsp 2013].
HARISINGHANI, M. G., BARENTSZ, J., HAHN, P. F., DESERNO, W. M.,
TABATABAEI, S., VAN DE KAA, C. H., DE LA ROSETTE, J. & WEISSLEDER,
R. 2003. Noninvasive Detection of Clinically Occult Lymph-Node
Metastases in Prostate Cancer. New England Journal of Medicine, 348,
2491-2499.
237
Bibliography
HAW, C. Y., MOHAMED, F., CHIA, C. H., RADIMAN, S., ZAKARIA, S., HUANG, N.
M. & LIM, H. N. 2010. Hydrothermal synthesis of magnetite nanoparticles
as MRI contrast agents. Ceramics International, 36, 1417-1422.
HEALTHCARE, G. 2013. Omniscan [Online]. Available:
http://www3.gehealthcare.com/en/Products/Categories/Contrast_Media/
Omniscan#tabs/tab488621DABE7848C7837C18A88CFE9FFF 2013].
HEESAKKERS, R. A. M., JAGER, G. J., HÖVELS, A. M., DE HOOP, B., VAN DEN
BOSCH, H. C. M., RAAT, F., WITJES, J. A., MULDERS, P. F. A., VAN DER
KAA, C. H. & BARENTSZ, J. O. 2009. Prostate cancer: Detection of lymph
node metastases outside the routine surgical area with ferumoxtran-10-
enhanced MR imaging. Radiology, 251, 408-414.
HEO, J. & LIU, C. 2007. Pbs quantum-dots in glass matrix for universal fiber-
optic amplifier. Journal of Materials Science: Materials in Electronics, 18,
135-139.
HOBBS, H., BRIDDON, S. & LESTER, E. 2009. The synthesis and fluorescent
properties of nanoparticulate ZrO(2) doped with Eu using continuous
hydrothermal synthesis. Green Chemistry, 11, 484-491.
HONG, S.-A., KIM, S. J., CHUNG, K. Y., CHUN, M.-S., LEE, B. G. & KIM, J. 2013.
Continuous synthesis of lithium iron phosphate (LiFePO4) nanoparticles in
supercritical water: Effect of mixing tee. The Journal of Supercritical
Fluids, 73, 70-79.
HU, D., ZHANG, P., GONG, P., LIAN, S., LU, Y., GAO, D. & CAI, L. 2011. A fast
synthesis of near-infrared emitting CdTe/CdSe quantum dots with small
hydrodynamic diameter for in vivo imaging probes. Nanoscale, 3, 4724-
4732.
HUDGINS, P. A., ANZAI, Y., MORRIS, M. R. & LUCAS, M. A. 2002. Ferumoxtran-
10, A Superparamagnetic Iron Oxide as a Magnetic Resonance
Enhancement Agent for Imaging Lymph Nodes: A Phase 2 Dose Study.
American Journal of Neuroradiology, 23, 649-656.
IOKU, K., YOSHIMURA, M. & SOMIYA, S. 1988. Post-sintering of apatite ceramics
from fine powders synthesized under hydrothermal conditions. Journal of
the Ceramic Society of Japan. International ed., 96, 111-112.
ISASI-MARÍN, J., PÉREZ-ESTÉBANEZ, M., DÍAZ-GUERRA, C., CASTILLO, J. F.,
CORRECHER, V. & CUERVO-RODRÍGUEZ, M. R. 2009. Structural,
magnetic and luminescent characteristics of Pr 3+-doped ZrO2 powders
synthesized by a sol-gel method. Journal of Physics D: Applied Physics,
42.
JADHAV, S. A. & BONGIOVANNI, R. 2012. Synthesis and organic
functionalization approaches for magnetite (Fe3O4) nanoparticles.
Advanced Materials Letters, 3, 356-361.
JADHAV, S. B. & JAIN, G. K. 2006. Statins and osteoporosis: New role for old
drugs. Journal of Pharmacy and Pharmacology, 58, 3-18.
JAMIESON, T., BAKHSHI, R., PETROVA, D., POCOCK, R., IMANI, M. & SEIFALIAN,
A. M. 2007. Biological applications of quantum dots. Biomaterials, 28,
4717-4732.
JIN, T., SUN, D., SU, J. Y., ZHANG, H. & SUE, H. J. 2009. Antimicrobial efficacy
of zinc oxide quantum dots against Listeria monocytogenes, Salmonella
Enteritidis, and Escherichia coli O157:H7. Journal of Food Science, 74,
M46-M52.
KALITA, S. J. & BHATT, H. A. 2007. Nanocrystalline hydroxyapatite doped with
magnesium and zinc: Synthesis and characterization. Materials Science
and Engineering: C, 27, 837-848.
KALITA, S. J. & VERMA, S. 2010. Nanocrystalline hydroxyapatite bioceramic
using microwave radiation: Synthesis and characterization. Materials
Science and Engineering C, 30, 295-303.
238
Bibliography
KARIMI, Z., KARIMI, L. & SHOKROLLAHI, H. 2013. Nano-magnetic particles used

in biomedicine: Core and coating materials. Materials Science and
Engineering C, 33, 2465-2475.
KASUGA, T., HIRAMATSU, K., HOSON, A., SEKINO, T. & NIIHARA, K. 1998.
Formation of Titanium Oxide Nanotube. Langmuir, 1998, 3160-3163.
KAWASAKI, S. I., SUE, K., OOKAWARA, R., WAKASHIMA, Y., SUZUKI, A.,
HAKUTA, Y. & ARAI, K. 2010. Engineering study of continuous
supercritical hydrothermal method using a T-shaped mixer: Experimental
synthesis of NiO nanoparticles and CFD simulation. Journal of
KHODADADI-MOGHADDAM, M., SALIMI, F., SAHEBALZAMANI, H. & JAFARI, N.
2013. Synthesis of Fe2O3 nanoparticles via various methods and coating
with silica for drug immobilization. Synthesis and Reactivity in Inorganic,
Metal-Organic and Nano-Metal Chemistry, 43, 1224-1227.
KIM, J., PARK, Y. S., VERIANSYAH, B., KIM, J. D. & LEE, Y. W. 2008. Continuous
synthesis of surface-modified metal oxide nanoparticles using
supercritical methanol for highly stabilized nanofluids. Chemistry of
Materials, 20, 6301-6303.
KIM, K., JEONG, S., WOO, J. Y. & HAN, C. S. 2012. Successive and large-scale
synthesis of InP/ZnS quantum dots in a hybrid reactor and their
application to white LEDs. Nanotechnology, 23.
KONG, H. Y., HWANG, C. S. & BYUN, J. 2012. Biological toxicity changes of
mercaptoacetic acid and mercaptopropionic acid upon coordination onto
ZnS:Mn nanocrystal. Bulletin of the Korean Chemical Society, 33, 657-
662.
KRITZER, P., BOUKIS, N. & DINJUS, E. 1999. Factors controlling corrosion in
high-temperature aqueous solutions: A contribution to the dissociation
and solubility data influencing corrosion processes. Journal of
KROL, S., MACREZ, R., DOCAGNE, F., DEFER, G., LAURENT, S., RAHMAN, M.,
HAJIPOUR, M. J., KEHOE, P. G. & MAHMOUDI, M. 2013. Therapeutic
benefits from nanoparticles: The potential significance of nanoscience in
diseases with compromise to the blood brain barrier. Chemical Reviews,
113, 1877-1903.
LAAKSONEN, T., LIU, P., RAHIKKALA, A., PELTONEN, L., KAUPPINEN, E. I.,
HIRVONEN, J., JÄRVINEN, K. & RAULA, J. 2011. Intact nanoparticulate
indomethacin in fast-dissolving carrier particles by combined wet milling
and aerosol flow reactor methods. Pharmaceutical Research, 28, 2403-
2411.
LANG, J., LI, X., YANG, J., YANG, L., ZHANG, Y., YAN, Y., HAN, Q., WEI, M.,
GAO, M., LIU, X. & WANG, R. 2011. Rapid synthesis and luminescence of
the Eu3+, Er3+ codoped ZnO quantum-dot chain via chemical
precipitation method. Applied Surface Science, 257, 9574-9577.
LAUGIER, J. 1999. Celref v3. Laboratoire des Materiaux et du G nie Physique de
l′Ecole Superieure de Physique de Grenoble, France.
LEPRÊTRE, S., CHAI, F., HORNEZ, J. C., VERMET, G., NEUT, C., DESCAMPS, M.,
HILDEBRAND, H. F. & MARTEL, B. 2009. Prolonged local antibiotics
delivery from hydroxyapatite functionalised with cyclodextrin polymers.
Biomaterials, 30, 6086-6093.
LESTER, E., AKSOMAITYTE, G., LI, J., GOMEZ, S., GONZALEZ-GONZALEZ, J. &
POLIAKOFF, M. 2012. Controlled continuous hydrothermal synthesis of
cobalt oxide (Co 3O 4) nanoparticles. Progress in Crystal Growth and
Characterization of Materials, 58, 3-13.
LESTER, E., BLOOD, P., DENYER, J., GIDDINGS, D., AZZOPARDI, B. &
POLIAKOFF, M. 2006. Reaction engineering: The supercritical water
hydrothermal synthesis of nano-particles. Journal of Supercritical Fluids,
37, 209-214.
239
Bibliography
LESTER, E., TANG, S. V. Y., KHLOBYSTOV, A., ROSE, V. L., BUTTERY, L. &
ROBERTS, C. J. 2013. Producing nanotubes of biocompatible
hydroxyapatite by continuous hydrothermal synthesis. CrystEngComm,
15, 3256-3260.
LESTER, E. H. & AZZOPARDI, B. J. 2005. Counter Current Mixing Reactor.
WO2005/077505.
LEYBROS, A., PIOLET, R., ARIANE, M., MUHR, H., BERNARD, F. & DEMOISSON,
F. 2012. CFD simulation of ZnO nanoparticle precipitation in a
supercritical water synthesis reactor. Journal of Supercritical Fluids, 70,
17-26.
LI, B., WANG, X. L., GUO, B., XIAO, Y. M., FAN, H. S. & ZHANG, X. D. 2007.
Preparation and characterization of nano hydroxyapatite. Key Engineering
Materials, 330-332, 235-238.
LI, J. 2008. Engineering Nanoparticles in Near-critical and Supercritical Water.
Ph.D Chemical Engineering The University of Nottingham.
LI, S., ZHAO, H. & TIAN, D. 2013. Aqueous synthesis of highly monodispersed
thiol-capped CdSe quantum dots based on the electrochemical method.
Materials Science in Semiconductor Processing, 16, 149-153.
LIN, T., KELLICI, S., GONG, K., THOMPSON, K., EVANS, J. R. G., WANG, X. &
DARR, J. A. 2010. Rapid automated materials synthesis instrument:
Exploring the composition and heat-treatment of nanoprecursors toward
low temperature red phosphors. Journal of Combinatorial Chemistry, 12,
383-392.
LIU, H. & WEBSTER, T. J. 2007. Nanomedicine for implants: A review of studies
and necessary experimental tools. Biomaterials, 28, 354-369.
LIU, M., ZHAO, H., CHEN, S., WANG, H. & QUAN, X. 2012. Photochemical
synthesis of highly fluorescent CdTe quantum dots for "on-off-on"
detection of Cu(II) ions. Inorganica Chimica Acta, 392, 236-240.
LIU, Y., ZHAN, J., REN, M., TANG, K., YU, W. & QIAN, Y. 2001. Hydrothermal
synthesis of square thin flake CdS by using surfactants and
thiocarbohydrate. Materials Research Bulletin, 36, 1231-1236.
LÓPEZ-QUINTELA, M. A., RIVAS, J., BLANCO, M. C. & TOJO, C. 2004. Synthesis
of Nanoparticles in Microemulsions. In: LIZ-MARZÁN, L. & KAMAT, P.
(eds.) Nanoscale Materials. Springer US.
LORYUENYONG, V., BUASRI, A., POCHANA, J., HOSAWANGWONG, S.,
THAISAUNG, S. & SOOKSAEN, P. 2013. Synthesis of anatase TiO2
nanoparticles by template sol-gel method and its application in
photocatalytic degradation of organic pollutants. Advanced Science
Letters, 19, 2919-2922.
LU, A. H., SALABAS, E. L. & SCHÜTH, F. 2007. Magnetic nanoparticles:
Synthesis, protection, functionalization, and application. Angewandte
Chemie - International Edition, 46, 1222-1244.
LU, J., HAKUTA, Y., HAYASHI, H., OHASHI, T., NAGASE, T., HOSHI, Y., SATO, K.,
NISHIOKA, M., INOUE, T. & HAMAKAWA, S. 2008. Preparation of
Ca0.8Sr0.2Ti1-xFexO3-δ (x = 0.1-0.3) nanoparticles using a flow
supercritical reaction system. Journal of Supercritical Fluids, 46, 77-82.
LU, J., MINAMI, K., TAKAMI, S. & ADSCHIRI, T. 2013. Rapid and continuous
synthesis of cobalt aluminate nanoparticles under subcritical
hydrothermal conditions with in-situ surface modification. Chemical
Engineering Science, 85, 50-54.
LUAN, W., YANG, H., WAN, Z., YUAN, B., YU, X. & TU, S. T. 2012.
Mercaptopropionic acid capped CdSe/ZnS quantum dots as fluorescence
probe for lead(II). Journal of Nanoparticle Research, 14, 1-8.
LUBRIZOL. 2013. Hyperdispersants - Technology and Benfits.
MALVERN. 2013. Dynamic Light Scattering [Online]. Available:
http://www.malvern.com.
240
Bibliography
MANH, D. H., THUAN, N. C., PHONG, P. T., HONG, L. V. & PHUC, N. X. 2009.
Magnetic properties of La0.7Ca0.3MnO3 nanoparticles prepared by
reactive milling. Journal of Alloys and Compounds, 479, 828-831.
MANSFIELD, P. & MAUDSLEY, A. A. 1977. Medical imaging by NMR. British
Journal of Radiology, 50, 188-194.
MANSFIELD, P. & PYKETT, I. L. 1978. Biological and medical imaging by NMR.
Journal of Magnetic Resonance (1969), 29, 355-373.
MASTROGIACOMO, M., SCAGLIONE, S., MARTINETTI, R., DOLCINI, L.,
BELTRAME, F., CANCEDDA, R. & QUARTO, R. 2006. Role of scaffold
internal structure on in vivo bone formation in macroporous calcium
phosphate bioceramics. Biomaterials, 27, 3230-3237.
MATSUI, K., NOGUCHI, T., ISLAM, N. M., HAKUTA, Y. & HAYASHI, H. 2008.
Rapid synthesis of BaTiO3 nanoparticles in supercritical water by
continuous hydrothermal flow reaction system. Journal of Crystal Growth,
310, 2584-2589.
MCCANN, T. E., KOSAKA, N., TURKBEY, B., MITSUNAGA, M., CHOYKE, P. L. &
KOBAYASHI, H. 2011. Molecular imaging of tumor invasion and
metastases: The role of MRI. NMR in Biomedicine, 24, 561-568.
MONMA, H. & KAMIYA, T. 1987. Preparation of hydroxyapatite by the hydrolysis
of brushite. Journal of Materials Science, 22, 4247-4250.
MONTAZERI, N., JAHANDIDEH, R. & BIAZAR, E. 2011. Synthesis of fluorapatite-
hydroxyapatite nanoparticles and toxicity investigations. International
journal of nanomedicine, 6, 197-201.
MORNET, S., VASSEUR, S., GRASSET, F. & DUGUET, E. 2004. Magnetic
nanoparticle design for medical diagnosis and therapy. Journal of
Materials Chemistry, 14, 2161-2175.
MORO, F., TANG, S. V. Y., TUNA, F. & LESTER, E. 2013. Magnetic properties of
cobalt oxide nanoparticles synthesised by a continuous hydrothermal
method. Journal of Magnetism and Magnetic Materials, 348, 1-7.
MOSEKE, C. & GBURECK, U. 2010. Tetracalcium phosphate: Synthesis,
properties and biomedical applications. Acta Biomaterialia, 6, 3815-3823.
MOUSAVAND, T., OHARA, S., UMETSU, M., ZHANG, J., TAKAMI, S., NAKA, T. &
ADSCHIRI, T. 2007. Hydrothermal synthesis and in situ surface
modification of boehmite nanoparticles in supercritical water. Journal of
MOUSAVAND, T., TAKAMI, S., UMETSU, M., OHARA, S. & ADSCHIRI, T. 2006.
Supercritical hydrothermal synthesis of organic-inorganic hybrid
nanoparticles. Journal of Materials Science, 41, 1445-1448.
MROZ, W., BOMBALSKA, A., BURDYNSKA, S., JEDYNSKI, M., PROKOPIUK, A.,
BUDNER, B., SLOSARCZYK, A., ZIMA, A., MENASZEK, E., SCISLOWSKA-
CZARNECKA, A. & NIEDZIELSKI, K. 2010. Structural studies of
magnesium doped hydroxyapatite coatings after osteoblast culture.
Journal of Molecular Structure, 977, 145-152.
MU, B., ZHONG, W., DONG, Y., DU, P. & LIU, P. 2012. Encapsulation of drug
microparticles with self-assembled Fe 3O 4/alginate hybrid multilayers for
targeted controlled release. Journal of Biomedical Materials Research -
Part B Applied Biomaterials, 100 B, 825-831.
NANDA, K. K., MAISELS, A., KRUIS, F. E., FISSAN, H. & STAPPERT, S. 2003.
Higher Surface Energy of Free Nanoparticles. Physical Review Letters, 91,
1061021-1061024.
NANO.GOV. 2013. What's So Special about the Nanoscale? [Online]. Available:
http://www.nano.gov/nanotech-101/special.
NANOCO. 2013. Nanoco Group Plc [Online]. Available:
http://www.nanocotechnologies.com/].
NASSAR, M. Y. 2013. Size-controlled synthesis of CoCO3 and Co3O 4
nanoparticles by free-surfactant hydrothermal method. Materials Letters,
94, 112-115.
241
Bibliography
NETZ, D. J. A., SEPULVEDA, P., PANDOLFELLI, V. C., SPADARO, A. C. C.,

ALENCASTRE, J. B., BENTLEY, M. & MARCHETTI, J. M. 2001. Potential use
of gelcasting hydroxyapatite porous ceramic as an implantable drug
delivery system. International Journal of Pharmaceutics, 213, 117-125.
NI, S., YANG, X. & LI, T. 2011. Hydrothermal synthesis and photoluminescence
properties of SrCO3. Materials Letters, 65, 766-768.
NOGUCHI, T., MATSUI, K., ISLAM, N. M., HAKUTA, Y. & HAYASHI, H. 2008.
Rapid synthesis of γ-Al2O3 nanoparticles in supercritical water by
continuous hydrothermal flow reaction system. Journal of Supercritical
Fluids, 46, 129-136.
OGI, T., KAIHATSU, Y., ISKANDAR, F., TANABE, E. & OKUYAMA, K. 2009.
Synthesis of nanocrystalline GaN from Ga2O3 nanoparticles derived from
salt-assisted spray pyrolysis. Advanced Powder Technology, 20, 29-34.
OH, K. S., KIM, K. J., JEONG, Y. K., PARK, E. K., KIM, S. Y., KWON, J. H., RYOO,
H. M. & SHIN, H. I. 2004. Cytotoxicity and antimicrobial effect of Ag
doped hydroxyapatite. In: MANDAL, H. & OVECOGLU, L. (eds.) Key
Engineering Materials. Istanbul.
OHARA, S., MOUSAVAND, T., SASAKI, T., UMETSU, M., NAKA, T. & ADSCHIRI, T.
2008. Continuous production of fine zinc oxide nanorods by hydrothermal
synthesis in supercritical water. Journal of Materials Science, 43, 2393-
2396.
OLDENDORF, W. 1988. Basics of Magnetic Resonance Imaging, Springer.
ORLOVSKII, V. P., KOMLEV, V. S. & BARINOV, S. M. 2002. Hydroxyapatite and
hydroxyapatite-based ceramics. Inorganic Materials, 38, 973-984.
OSAKA, A., MIURA, Y., TAKEUCHI, K., ASADA, M. & TAKAHASHI, K. 1991.
Calcium apatite prepared from calcium hydroxide and orthophosphoric
acid. Journal of Materials Science: Materials in Medicine, 2, 51-55.
PARK, J. B. 2009. The use of simvastatin in bone regeneration. Medicina Oral,
Patologia Oral y Cirugia Bucal, 14, e485-e488.
PARK, J. J., PRABHAKARAN, P., JANG, K. K., LEE, Y., LEE, J., LEE, K., HUR, J.,
KIM, J. M., CHO, N., SON, Y., YANG, D. Y. & LEE, K. S. 2010.
Photopatternable quantum dots forming quasi-ordered arrays. Nano
Letters, 10, 2310-2317.
PARKER, J. E., THOMPSON, S. P., COBB, T. M., YUAN, F., POTTER, J., LENNIE, A.
R., ALEXANDER, S., TIGHE, C. J., DARR, J. A., COCKCROFT, J. C. &
TANG, C. C. 2011. High-throughput powder diffraction on beamline I11 at
Diamond. Journal of Applied Crystallography, 44, 102-110.
PARKES, L. M., HODGSON, R., LU, L. T., TUNG, L. D., ROBINSON, I., FERNIG, D.
G. & THANH, N. T. K. 2008. Cobalt nanoparticles as a novel magnetic
resonance contrast agent - Relaxivities at 1.5 and 3 Tesla. Contrast Media
and Molecular Imaging, 3, 150-156.
PIÑERO-HERNANZ, R., DODDS, C., HYDE, J., GARCÍA-SERNA, J., POLIAKOFF,
M., LESTER, E., COCERO, M. J., KINGMAN, S., PICKERING, S. & WONG,
K. H. 2008. Chemical recycling of carbon fibre reinforced composites in
nearcritical and supercritical water. Composites Part A: Applied Science
and Manufacturing, 39, 454-461.
POOLE, P. J., KAMINSKA, K., BARRIOS, P., LU, Z. & LIU, J. 2009. Growth of
InAs/InP-based quantum dots for 1.55 μm laser applications. Journal of
Crystal Growth, 311, 1482-1486.
POSNER, A. S., PERLOFF, A. & DIORIO, A. F. 1958. Refinement of the
hydroxyapatite structure. Acta. Cryst., 11, 308.
RAHMAN, S., NADEEM, K., ANIS-UR-REHMAN, M., MUMTAZ, M., NAEEM, S. &
LETOFSKY-PAPST, I. 2013. Structural and magnetic properties of ZnMg-
ferrite nanoparticles prepared using the co-precipitation method.
Ceramics International, 39, 5235-5239.
REN, H. B. & YAN, X. P. 2012. Ultrasonic assisted synthesis of adenosine
triphosphate capped manganese-doped ZnS quantum dots for selective
242
Bibliography
room temperature phosphorescence detection of arginine and methylated

arginine in urine based on supramolecular Mg2-adenosine triphosphate-
arginine ternary system. Talanta, 97, 16-22.
RONDA, C. R. 2008. Luminescence, Wiley.
ROSLI, S. A., ZUBIR, Z. A. & AZIZ, N. M. A. 2012. Characterization of zns
nanoparticles using Mpa as capping agents. Selangor.
ROTELLO, V. M. 2004. Nanoparticles: Building Blocks for Nanotechnology, Kluwer
Academic/Plenum Publishers.
SAHRANESHIN, A., ASAHINA, S., TOGASHI, T., SINGH, V., TAKAMI, S., HOJO,
D., ARITA, T., MINAMI, K. & ADSCHIRI, T. 2012a. Surfactant-assisted
hydrothermal synthesis of water-dispersible hafnium oxide nanoparticles
in highly alkaline media. Crystal Growth and Design, 12, 5219-5226.
SAHRANESHIN, A., TAKAMI, S., HOJO, D., ARITA, T., MINAMI, K. & ADSCHIRI,
T. 2012b. Mechanistic study on the synthesis of one-dimensional yttrium
aluminum garnet nanostructures under supercritical hydrothermal
conditions in the presence of organic amines. CrystEngComm, 14, 6085-
6092.
SAMPATHKUMARAN, E. V., MUKHERJEE, K., IYER, K. K., MOHAPATRA, N. & DAS,
S. D. 2011. Magnetism of fine particles of kondo lattices, obtained by
high-energy ball-milling. Journal of Physics Condensed Matter, 23.
SANTRA, P. K. & KAMAT, P. V. 2012. Mn-doped quantum dot sensitized solar
cells: A strategy to boost efficiency over 5%. Journal of the American
Chemical Society, 134, 2508-2511.
SARASWATHI AMMA, B., MANZOOR, K., RAMAKRISHNA, K. & PATTABI, M. 2008.
Synthesis and optical properties of CdS/ZnS coreshell nanoparticles.
Materials Chemistry and Physics, 112, 789-792.
SATO, T., IIJIMA, T., SEKI, M. & INAGAKI, N. 1987. Magnetic properties of
ultrafine ferrite particles. Journal of Magnetism and Magnetic Materials,
65, 252-256.
SATO, T., SUE, K., SUZUKI, W., SUZUKI, M., MATSUI, K., HAKUTA, Y., HAYASHI,
H., ARAI, K., KAWASAKI, S. I., KAWAI-NAKAMURA, A. & HIAKI, T. 2008.
Rapid and continuous production of ferrite nanoparticles by hydrothermal
synthesis at 673 K and 30 MPa. Industrial and Engineering Chemistry
Research, 47, 1855-1860.
SEKINO, T., OKAMOTO, T., KASUGA, T., KUSUNOSE, T., NAKAYAMA, T. &
NIIHARA, K. 2006. Synthesis and properties of titania nanotube doped
with small amount of cations. Key Engineering Materials, 317-318, 251-
254.
SENTHILKUMAAR, S. & SELVI, R. T. 2008. Synthesis and Characterization of One
Dimensional ZnS Nanorods. Synthesis and Reactivity in Inorganic Metal-
Organic and Nano-Metal Chemistry, 38, 710-715.
SEPULVEDA, P., BINNER, J. G. P., ROGERO, S. O., HIGA, O. Z. & BRESSIANI, J.
C. 2000. Production of porous hydroxyapatite by the gel-casting of foams
and cytotoxic evaluation. Journal of Biomedical Materials Research, 50,
27-34.
SHARMA, A., PANDEY, C. M., SUMANA, G., SONI, U., SAPRA, S., SRIVASTAVA,
A. K., CHATTERJEE, T. & MALHOTRA, B. D. 2012. Chitosan encapsulated
quantum dots platform for leukemia detection. Biosensors and
Bioelectronics, 38, 107-113.
SHARMA, M., KUMAR, S. & PANDEY, O. P. 2010. Study of energy transfer from
capping agents to intrinsic vacancies/defects in passivated ZnS
nanoparticles. Journal of Nanoparticle Research, 12, 2655-2666.
SHEN, L., CUI, X., QI, H. & ZHANG, C. 2007. Electrogenerated
chemiluminescence of ZnS nanoparticles in alkaline aqueous solution.
Journal of Physical Chemistry C, 111, 8172-8175.
243
Bibliography
SHIN, N. C., LEE, Y. H., SHIN, Y. H., KIM, J. & LEE, Y. W. 2010. Synthesis of
cobalt nanoparticles in supercritical methanol. Materials Chemistry and
Physics, 124, 140-144.
SHYMAN. 2013. Sustainable Hydrothermal Manufacturing of Nanomaterials
[Online]. http://www.shyman.eu/. Available: http://www.shyman.eu/
2013].
SOHRABNEZHAD, S. & VALIPOUR, A. 2013. Synthesis of Cu/CuO nanoparticles in
mesoporous material by solid state reaction. Spectrochimica Acta - Part
A: Molecular and Biomolecular Spectroscopy, 114, 298-302.
SOMIYA, S. & ROY, R. 2000. Hydrothermal synthesis of fine oxide powders.
Bulletin of Materials Science, 23, 453-460.
STANIĆ, V., DIMITRIJEVIĆ, S., ANTIĆ-STANKOVIĆ, J., MITRIĆ, M., JOKIĆ, B.,
PLEĆAŠ, I. B. & RAIČEVIĆ, S. 2010. Synthesis, characterization and
antimicrobial activity of copper and zinc-doped hydroxyapatite
nanopowders. Applied Surface Science, 256, 6083-6089.
STANIC, V., JANACKOVIC, D., DIMITRIJEVIC, S., TANASKOVIC, S. B., MITRIC,
M., PAVLOVIC, M. S., KRSTIC, A., JOVANOVIC, D. & RAICEVIC, S. 2011.
Synthesis of antimicrobial monophase silver-doped hydroxyapatite
nanopowders for bone tissue engineering. Applied Surface Science, 257,
4510-4518.
STEPHEN, Z. R., KIEVIT, F. M. & ZHANG, M. 2011. Magnetite nanoparticles for
medical MR imaging. Materials Today, 14, 330-338.
STURMAN, B. D., ROUSE, R. C. & DUNN, P. J. 1981. Parascholzite, a new mineral
from Hagendorf, Bavaria, and its relationship to scholzite. American
Mineralogist, 66, 843-851.
SUE, K., AOKI, M., SATO, T., NISHIO-HAMANE, D., KAWASAKI, S. I., HAKUTA,
Y., TAKEBAYASHI, Y., YODA, S., FURUYA, T. & HIAKI, T. 2011.
Continuous hydrothermal synthesis of nickel ferrite nanoparticles using a
central collision-type micromixer: Effects of temperature, residence time,
metal salt molality, and naoh addition on conversion, particle size, and
crystal phase. Industrial and Engineering Chemistry Research, 50, 9625-
9631.
SUE, K., SUZUKI, A., HAKUTA, Y., HAYASHI, H., ARAI, K., TAKEBAYASHI, Y.,
YODA, S. & FURUYA, T. 2009. Hydrothermal-reduction synthesis of Ni
nanoparticles by superrapid heating using a micromixer. Chemistry
Letters, 38, 1018-1019.
SUE, K., SUZUKI, M., ARAI, K., OHASHI, T., URA, H., MATSUI, K., HAKUTA, Y.,
HAYASHI, H., WATANABE, M. & HIAKI, T. 2006. Size-controlled synthesis
of metal oxide nanoparticles with a flow-through supercritical water
method. Green Chemistry, 8, 634-638.
TAGUCHI, M., TAKAMI, S., ADSCHIRI, T., NAKANE, T., SATO, K. & NAKA, T.
2012. Synthesis of surface-modified monoclinic ZrO 2 nanoparticles using
supercritical water. CrystEngComm, 14, 2132-2138.
TAKAMI, S., SUGIOKA, K. I., TSUKADA, T., ADSCHIRI, T., SUGIMOTO, K.,
TAKENAKA, N. & SAITO, Y. 2012. Neutron radiography on tubular flow
reactor for hydrothermal synthesis: In situ monitoring of mixing behavior
of supercritical water and room-temperature water. Journal of
TANG, G., LIU, S., TANG, H., ZHANG, D., LI, C. & YANG, X. 2013. Template-
assisted hydrothermal synthesis and photocatalytic activity of novel TiO2
hollow nanostructures. Ceramics International, 39, 4969-4974.
TANG, S. Y., BOURNE, R. A., SMITH, R. L. & POLIAKOFF, M. 2008. The 24
Principles of Green Engineering and Green Chemistry: "IMPROVEMENTS
PRODUCTIVELY". Green Chemistry, 10, 268-269.
THOMSEN, H. S. & WEBB, J. A. W. 2009. MR Contrast Media: Organ specific. In:
BAERT, A. L., BRADY, L. W., HEILMANN, H. P., KNAUTH, M., MOLLS, M. &
244
Bibliography
NEIDER, C. (eds.) Contrast Media: Safety Issues and ESUR Guidelines. 2

ed.: Springer.
TIGHE, C. J., CABRERA, R. Q., GRUAR, R. I. & DARR, J. A. 2013. Scale up
production of nanoparticles: Continuous supercritical water synthesis of
Ce-Zn oxides. Industrial and Engineering Chemistry Research, 52, 5522-
5528.
TIGHE, C. J., GRUAR, R. I., MA, C. Y., MAHMUD, T., WANG, X. Z. & DARR, J. A.
2012. Investigation of counter-current mixing in a continuous
hydrothermal flow reactor. Journal of Supercritical Fluids, 62, 165-172.
TSAI, Z.-T., WANG, J.-F., KUO, H.-Y., SHEN, C.-R., WANG, J.-J. & YEN, T.-C.
2010. In situ preparation of high relaxivity iron oxide nanoparticles by
coating with chitosan: A potential MRI contrast agent useful for cell
tracking. Journal of Magnetism and Magnetic Materials, 322, 208-213.
TSAKALAKOS, L. 2008. Nanostructures for photovoltaics. Materials Science and
Engineering R: Reports, 62, 175-189.
TSUZUKI, T. 2009. Commercial scale production of inorganic nanoparticles.
International Journal of Nanotechnology, 6, 567-578.
VERIANSYAH, B., KIM, J. D., MIN, B. K. & KIM, J. 2010a. Continuous synthesis of
magnetite nanoparticles in supercritical methanol. Materials Letters, 64,
2197-2200.
VERIANSYAH, B., KIM, J. D., MIN, B. K., SHIN, Y. H., LEE, Y. W. & KIM, J.
2010b. Continuous synthesis of surface-modified zinc oxide nanoparticles
in supercritical methanol. Journal of Supercritical Fluids, 52, 76-83.
VERIANSYAH, B., PARK, H., KIM, J. D., MIN, B. K., SHIN, Y. H., LEE, Y. W. &
KIM, J. 2009. Characterization of surface-modified ceria oxide
nanoparticles synthesized continuously in supercritical methanol. Journal
of Supercritical Fluids, 50, 283-291.
VICKERMAN, J. C. & BRIGGS, D. 2001. ToF-SIMS: Surface Analysis by Mass
Spectrometry, IM.
VIJAYALAKSHMI, U. & RAJESWARI, S. 2006. Preparation and characterization of
microcrystalline hydroxyapatite using sol gel method. Trends in
Biomaterials and Artificial Organs, 19, 57-62.
WAGEH, S., SHU-MAN, L., YOU, F. T. & XU-RONG, X. 2003. Optical properties of
strongly luminescing mercaptoacetic-acid-capped ZnS nanoparticles.
Journal of Luminescence, 102-103, 768-773.
WAKASHIMA, Y., SUZUKI, A., KAWASAKI, S. I., MATSUI, K. & HAKUTA, Y. 2007.
Development of a new swirling micro mixer for continuous hydrothermal
synthesis of nano-size particles. Journal of Chemical Engineering of
Japan, 40, 622-629.
WANDELER, R. & BAIKER, A. 2000. Supercritical fluids: Opportunities in
heterogeneous catalysis. CATTECH, 4, 34-50.
WANG, Q., TANG, S. V. Y., LESTER, E. & O'HARE, D. 2013. Synthesis of ultrafine
layered double hydroxide (LDHs) nanoplates using a continuous-flow
hydrothermal reactor. Nanoscale, 5, 114-117.
WANG, Y.-X. J. 2011. Superparamagnetic iron oxide based MRI contrast agents:
Current status of clinical application. Quantitative Imaging in Medicine
and Surgery, 1, 35-40.
WEBSTER, T. J., MASSA-SCHLUETER, E. A., SMITH, J. L. & SLAMOVICH, E. B.
2004. Osteoblast response to hydroxyapatite doped with divalent and
trivalent cations. Biomaterials, 25, 2111-2121.
WEI, Z., XIA, T., MA, J., FENG, W., DAI, J., WANG, Q. & YAN, P. 2007.
Investigation of the lattice expansion for Ni nanoparticles. Materials
Characterization, 58, 1019-1024.
WENG, X., COCKCROFT, J. K., HYETT, G., VICKERS, M., BOLDRIN, P., TANG, C.
C., THOMPSON, S. P., PARKER, J. E., KNOWLES, J. C., REHMAN, I.,
PARKIN, I., EVANS, J. R. G. & DARR, J. A. 2009. High-throughput
245
Bibliography
continuous hydrothermal synthesis of an entire nanoceramic phase

diagram. Journal of Combinatorial Chemistry, 11, 829-834.
WHITTERS, C. J., STRANG, R., BROWN, D., CLARKE, R. L., CURTIS, R. V.,
HATTON, P. V., IRELAND, A. J., LLOYD, C. H., MCCABE, J. F.,
NICHOLSON, J. W., SCRIMGEOUR, S. N., SETCOS, J. C., SHERRIFF, M.,
VAN NOORT, R., WATTS, B. C. & WOOD, D. 1999. Dental materials: 1997
literature review. Journal of Dentistry, 27, 401-435.
WILLARD, M. A., KURIHARA, L. K., CARPENTER, E. E., CALVIN, S. & HARRIS, V.
G. 2004. Chemically prepared magnetic nanoparticles. International
Materials Reviews, 49, 125-170.
WILLIAMS, D. B. & CARTER, C. B. 2009. Transmission electron microscopy: a
textbook for materials science., Springer.
WOOD, V., PANZER, M. J., CHEN, J., BRADLEY, M. S., HALPERT, J. E., BAWENDI,
M. C. & BULOVIĆ, V. 2009. Inkjet-printed quantum dot-polymer
composites for full-color AC-driven displays. Advanced Materials, 21,
2151-2155.
XING, Y., CHAUDRY, Q., SHEN, C., KONG, K. Y., ZHAU, H. E., CHUNG, L. W.,
PETROS, J. A., O'REGAN, R. M., YEZHELYEV, M. V., SIMONS, J. W.,
WANG, M. D. & NIE, S. 2007. Bioconjugated quantum dots for
multiplexed and quantitative immunohistochemistry. Nature Protocols, 2,
1152-1165.
XU, B. & WANG, X. 2012. Solvothermal synthesis of monodisperse nanocrystals.
Dalton Transactions, 41, 4719-4725.
XUE, B., DENG, D. W., CAO, J., LIU, F., LI, X., AKERS, W., ACHILEFU, S. & GU,
Y. Q. 2012. Synthesis of NAC capped near infrared-emitting CdTeS
alloyed quantum dots and application for in vivo early tumor imaging.
Dalton Transactions, 41, 4935-4947.
YANG, F., XU, Z., WANG, J., ZAN, F., DONG, C. & REN, J. 2012a. Microwave-
assisted aqueous synthesis of new quaternary-alloyed CdSeTeS quantum
dots; and their bioapplications in targeted imaging of cancer cells.
Luminescence.
YANG, H., LUAN, W., CHENG, R., CHU, H. & TU, S. T. 2011. Synthesis of
quantum dots via microreaction: Structure optimization for microreactor
system. Journal of Nanoparticle Research, 13, 3335-3344.
YANG, L.-X., YIN, J.-J., WANG, L.-L., XING, G.-X., YIN, P. & LIU, Q.-W. 2012b.
Hydrothermal synthesis of hierarchical hydroxyapatite: Preparation,
growth mechanism and drug release property. Ceramics International,
38, 495-502.
YU, Y., XU, L., CHEN, J., GAO, H., WANG, S., FANG, J. & XU, S. 2012.
Hydrothermal synthesis of GSH-TGA co-capped CdTe quantum dots and
their application in labeling colorectal cancer cells. Colloids and Surfaces
B: Biointerfaces, 95, 247-253.
ZAGHIB, K., CHAREST, P., DONTIGNY, M., LABRECQUE, J. F., MAUGER, A. &
JULIEN, C. A new synthetic route of LiFePO 4 nanoparticles from molten
ingot. 2011 Las Vegas, NV. 23-31.
ZHANG, C., YANG, J., QUAN, Z., YANG, P., LI, C., HOU, Z. & LIN, J. 2009a.
Hydroxyapatite nano- and microcrystals with multiform morphologies:
Controllable synthesis and luminescence properties. Crystal Growth and
Design, 9, 2725-2733.
ZHANG, J. & GE, M. 2011. Effecting factors of the emission spectral
characteristics of rare-earth strontium aluminate for anti-counterfeiting
application. Journal of Luminescence, 131, 1765-1769.
ZHANG, L., HE, R. & GU, H.-C. 2006. Oleic acid coating on the monodisperse
magnetite nanoparticles. Applied Surface Science, 253, 2611-2617.
ZHANG, S., ZHANG, Y., WANG, Y., LIU, S. & DENG, Y. 2012. Sonochemical
formation of iron oxide nanoparticles in ionic liquids for magnetic liquid
marble. Physical Chemistry Chemical Physics, 14, 5132-5138.
246
Bibliography
ZHANG, Y. & LU, J. 2008. The transformation of single-crystal calcium phosphate

ribbon-like fibres to hydroxyapatite spheres assembled from nanorods.
Nanotechnology, 19.
ZHANG, Z., BROWN, S., GOODALL, J. B. M., WENG, X., THOMPSON, K., GONG,
K., KELLICI, S., CLARK, R. J. H., EVANS, J. R. G. & DARR, J. A. 2009b.
Direct continuous hydrothermal synthesis of high surface area nanosized
titania. Journal of Alloys and Compounds, 476, 451-456.
ZHANG, Z., GOODALL, J. B. M., BROWN, S., KARLSSON, L., CLARK, R. J. H.,
HUTCHISON, J. L., REHMAN, I. U. & DARR, J. A. 2010. Continuous
hydrothermal synthesis of extensive 2D sodium titanate (Na 2Ti3O7)
nano-sheets. Dalton Transactions, 39, 711-714.
ZHENG, Y.-H., CHENG, Y., BAO, F. & WANG, Y.-S. 2006. Synthesis and magnetic
properties of Fe3O4 nanoparticles. Materials Research Bulletin, 41, 525-
529.
ZHOU, H. & LEE, J. 2011. Nanoscale hydroxyapatite particles for bone tissue
engineering. Acta Biomaterialia, 7, 2769-2781.
ZHU, L., ANG, S. & LIU, W. T. 2004. Quantum Dots as a Novel
Immunofluorescent Detection System for Cryptosporidium parvum and
Giardia lamblia. Applied and Environmental Microbiology, 70, 597-598.
247
8 Appendix
Figure A1. SEM images of (a) ZnS:NaOH 9/1, (b) ZnS:NaOH 8/2, (c)
ZnS:NaOH 7/3, (d) ZnS:NaOH 6/4.
248
Appendix
Figure A2. SEM images of (a) ZnS:3MPA 9/1, (b) ZnS:3MPA 8/2, (c)
ZnS:3MPA 7/3, (d) ZnS:3MPA 6/4.
249
Appendix
Figure A3. SEM images of (a) ZnS:PVP 9/1, (b) ZnS:PVP 8/2, (c)
ZnS:PVP 7/3, (d) ZnS:PVP 6/4.
250
Appendix
Table A1. Zeta Potential data for Uncapped ZnS.
Average ζ potential Standard

pH
(mV) Deviation
2.59 20.1 1.362
4.48 21.12 1.126
6.72 3.668 0.991
8.12 -0.843 0.274
9.52 -1.407 0.46
10.38 -34.6 1.589
11.69 -55.1 5.704
Table A2. Zeta Potential data for ZnS:3MPA 5/5.

pH
(mV) Deviation
2.09 19.92 1.184
3.62 13.5 0.731
4.31 -3.992 0.767
5.92 -23.86 1.036
6.44 -26.74 2.348
9.33 -22.34 1.074
10.97 -52.28 1.875
251
Appendix
Table A3. Zeta Potential data for ZnS:PVP 5/5.

pH
(mV) Deviation
3.05 0.443 0.423
3.53 -5.926 1.318
4.98 -6.492 1.37
7 -11.34 0.811
8.24 -18.34 0.879
10.14 -22.1 0.675
11.03 -18.06 3.469
Table A4. Zeta Potential data for ZnS:NaOH 5/5.
Average ζ Standard
pH
potential (mV) Deviation
3.16 9.622 0.518
5.04 10.598 0.635
7.1 8.52 0.835
8.21 8.244 0.859
9.35 -21.34 1.378
10.96 -38.48 1.813
252
Appendix
Figure A4. (a) The excitation spectra (λem = 540 nm) and (b) emission
spectra (λex = 360 nm) for ZnS:PVP samples produced using different
253
Appendix
Figure A5. (a) The excitation spectra (λem = 540 nm) and (b) emission
spectra (λex = 360 nm) for ZnS:NaOH samples produced using different
254

Synthesis of Nanomaterials For Biomedical Applications

Uploaded by

Copyright:

Available Formats

Synthesis of Nanomaterials For Biomedical Applications

Uploaded by

Document Information

Original Title

Copyright

Available Formats

Share this document

Share or Embed Document

Sharing Options

Did you find this document useful?

Is this content inappropriate?

Copyright:

Available Formats

Synthesis of Nanomaterials For Biomedical Applications

Uploaded by

Copyright:

Available Formats

Synthesis of Nanomaterials for Biomedical

Selina Vi Yu Tang, BSc.

Thesis submitted to the University of Nottingham

This Thesis describes an investigation into the use of a continuous

In Chapter 1, an introduction to nanotechnology is given, followed by a brief

Chapter 2 details the principles behind the main characterisation techniques

Chapter 3 reports the synthesis of three different morphologies of

To conclude, a summary of the work detailed in this Thesis is delivered in

Huge thanks to my colleagues who have helped or advised me during my PhD

Q. Wang, S. V. Y. Tang, E. Lester & D. O'Hare, Synthesis of ultrafine layered

E. Lester, S. V. Y. Tang, A. Khlobystov, V. L. Rose, L. Buttery & C. J. Roberts,

F. Moro, S. V. Y. Tang, F. Tuna & E. Lester, Magnetic properties of cobalt oxide

Selina Tang, Edward Lester, Andrei Khlobystov, Producing Multifunctional

Selina Tang, Edward Lester, Andrei Khlobystov, Producing Hydroxyapatite

Selina Tang, Andrei Khlobystov, Stephen Briddon, Edward Lester, Continuous

Selina Tang, Andrei Khlobystov, Stephen Briddon, Edward Lester, Continuous

Edward Lester, Selina Tang, Andrei Khlobystov, Steve Briddon, Christopher

Figure 1.1. A summary of the different fields in which nanoparticles have be

applied (Tsuzuki, 2009). ......................................................................... 3

particle size (Nano.gov, 2013). ................................................................ 4

constant of water with increasing temperature, at a constant pressure of 24

MPa (Kritzer et al., 1999). .....................................................................14

for continuous hydrothermal synthesis of nanoparticles (Adschiri et al., 2000).

reactor used by Adschiri et al., collected using neutron radiography (Takami

et al., 2012). ........................................................................................21

used by Arai et al. (Kawasaki et al., 2010). ..............................................25

Central Collision Mixer (CCM) (Sue et al., 2011). ......................................26

the counter-current reactor geometry (Tighe et al., 2012). ........................27

used by Darr et al (Gruar et al., 2013). ...................................................30

or methanol. B: balance; C: condenser; F: 0.2 μm metal filter; L: liquid

container; P: pressure gauge; R: supercritical reactor; T: thermocouple; V1:

pressure pump for DI water/methanol; HP2: high pressure pump for

heater; WP: water preheater line (Kim et al., 2008). .................................32

Institute of Science and Technology for the synthesis of LiFePO 4 nanoparticles

(Hong et al., 2013). ..............................................................................34

Figure 1.13. A schematic diagram of continuous hydrothermal reactor system

built by Aymes et al. in 2001 (Aimable et al., 2009b). ...............................37

a block of Inconel 625, used in the systems by Aymes et al. ......................38

al. (Aymes et al., 2011).........................................................................40

Figure 1.17. Schematic diagram of the continuous hydrothermal reactor system

used by Demoisson et al. (Demoisson et al., 2011). ..................................41

Figure 1.18. Diagram showing the configuration of the counter-current Nozzle

reactor, invented at The University of Nottingham (Lester and Azzopardi,

2005, Lester et al., 2006). .....................................................................43

Reactor (Lester et al., 2006). .................................................................44

University of Nottingham. ......................................................................45

Figure 1.21. A photo of the counter-current reactor system. ..........................47

counter-current Nozzle reactor technology, courtesy of Promethean Particles

Ltd (see Section 1.8.2). .........................................................................50

surfactant concentration resulting in partial coverage, (b) displays optimum

Figure 2.1. An illustration showing the components in a typical scanning electron

microscope (Flegler et al., 1995). ...........................................................61

schematic showing a 3-dimensional tube is viewed as a 2-dimensional image

Figure 2.3. A diagram showing the basis of XRD analysis. ..............................66

Figure 2.4. A diagram of the hydrodynamic diameter of a particle which is

reported by DLS analysis (Malvern, 2013)................................................71

Figure 2.5. A schematic representation of what is measured by zeta potential

(Malvern, 2013). ..................................................................................74

reactor section is highlighted in (b), showing the superheated aqueous