inkjet-printed light-emitting devices - t-space - university of toronto
TRANSCRIPT
Inkjet-Printed Light-Emitting Devices: Applying Inkjet Microfabrication to Multilayer Electronics
by
Peter D. Angelo
A thesis submitted in conformity with the requirements for the degree of Doctor of Philosophy
Department of Chemical Engineering & Applied Chemistry University of Toronto
Copyright by Peter David Angelo 2013
ii
Inkjet-Printed Light-Emitting Devices: Applying Inkjet
Microfabrication to Multilayer Electronics
Peter D. Angelo
Doctor of Philosophy
Department of Chemical Engineering & Applied Chemistry
University of Toronto
2013
Abstract
This work presents a novel means of producing thin-film light-emitting devices, functioning
according to the principle of electroluminescence, using an inkjet printing technique. This study
represents the first report of a light-emitting device deposited completely by inkjet printing. An
electroluminescent species, doped zinc sulfide, was incorporated into a polymeric matrix and
deposited by piezoelectric inkjet printing. The layer was printed over other printed layers
including electrodes composed of the conductive polymer poly(3,4-ethylenedioxythiophene),
doped with poly(styrenesulfonate) (PEDOT:PSS) and single-walled carbon nanotubes, and in
certain device structures, an insulating species, barium titanate, in an insulating polymer binder.
The materials used were all suitable for deposition and curing at low to moderate (<150°C)
temperatures and atmospheric pressure, allowing for the use of polymers or paper as supportive
substrates for the devices, and greatly facilitating the fabrication process.
iii
The deposition of a completely inkjet-printed light-emitting device has hitherto been unreported.
When ZnS has been used as the emitter, solution-processed layers have been prepared by spin-
coating, and never by inkjet printing. Furthermore, the utilization of the low-temperature-
processed PEDOT:PSS/nanotube composite for both electrodes has not yet been reported.
Device performance was compromised compared to conventionally prepared devices. This was
partially due to the relatively high roughness of the printed films. It was also caused by energy
level misalignment due to quantization (bandgap widening) of the small (<10 nm) nanoparticles,
and the use of high work function cathode materials (Al and PEDOT:PSS). Regardless of their
reduced performance, inkjet printing as a deposition technique for these devices presents unique
advantages, the most notable of which are rapidity of fabrication and patterning, substrate
flexibility, avoidance of material wastage by using drop-on-demand technology, and the need for
only one main unit operation to produce an entire device.
iv
Acknowledgements
The Pulp & Paper Centre in the Department of Chemical Engineering & Applied Chemistry at
the University of Toronto has provided me with an environment in which I have been able to
develop and refine both my project and my understanding of the fundamentals of chemical and
materials engineering. The Department, the Estate of W.H Rapson, the School of Graduate
Studies, the University of Toronto, the SENTINEL Bioactive Paper Network, and the
Government of Ontario have all financially supported my continued perseverance in my doctoral
studies, and I am deeply grateful to each. I would also like to acknowledge the generosity and
companionship of all my colleagues within the Pulp & Paper laboratories and other laboratory
groups within the Department. My understanding of several new aspects of chemical
engineering was also deepened by my supervising committee members, Dr. Edgar Acosta and
Dr. Timothy Bender. The laboratory of Dr. Ning Yan in the Faculty of Forestry was particularly
welcoming, and always willing to provide access to both analytical equipment and technical
expertise. The encouragement of my supervisor, Dr. Ramin Farnood, has been instrumental in
keeping me focused on my work. His continued forward drive and enthusiasm, as well as all of
the knowledge he has readily imparted to me, have led me to the successful conclusion of this
long process. I feel very fortunate to have worked with such a helpful and driven person.
My family’s perpetual support has buoyed me from the first day of my degree and throughout. I
cannot express my thanks sufficiently to them for their love and help. My father’s ready
willingness to discuss my work and give advice taken from his long-time familiarity with
electrical engineering has given me insight and inspiration on many occasions. Finally, my wife
Emily has kept me level, confident, and determined throughout my studies, while bringing me
joy each and every day. Her love has made persevering to the end worth every moment spent,
and taught me to look to the future with excitement and hope.
v
Nomenclature
Abbreviations
2D two-dimensional
3D three-dimensional
AC alternating current
AFM atomic force microscopy
ALD atomic layer deposition
ASTM American Society for Testing & Materials
BSE backscattered electron
CRT cathode ray tube
CVD chemical vapour deposition
DC direct current
DLS dynamic light scattering
DOD drop-on-demand
DPI dots per square inch
EL electroluminescence
ELD electroluminescent display
FPD flat-panel display
HOMO highest occupied molecular orbital
IR infrared
LED light-emitting device/diode
LUMO lowest unoccupied molecular orbital
OLED organic light-emitting diode/device
PDP plasma display panel
PEL powder electroluminescent device
PL photoluminescence
PLE photoluminescent excitation
PVD physical vapour deposition
QD quantum dot
R2R roll-to-roll
SEM scanning electron microscope
TAPPI Technical Association of the Pulp & Paper Industry
TEM transmission electron microscope
TFEL thin-film electroluminescent device
TFT thin-film transistor
ToF-SIMS time-of-flight secondary ion mass spectrometry
UV ultraviolet
XRD X-ray diffraction
vi
Terminology
dynamic viscosity (cP)
surface tension/energy (mN/m)
zeta-potential (mV)
dp particle diameter (nm)
dielectric constant (unitless)
resistivity ( m)
R resistance ()
conductivity (S/cm)
wavelength (nm)
c contact angle (°)
Oh Ohnesorge number (unitless)
Z-1
inverse Ohnesorge number (unitless)
En Energy number (unitless)
V voltage (V)
I current (A)
C capacitance (F)
Eg energy band gap (eV)
work function (eV)
L luminance (cd/m2)
luminous efficiency (lm/W)
M molarity (mol/L)
Mw polymer molecular weight (g/mol)
h Planck’s constant (4.13566733 × 10−15 eV·s)
ħ reduced Planck’s constant (6.58211899 × 10−16 eV·s)
0 vacuum permittivity (8.854187817620 × 10−12
F/m)
c speed of light (2.99792458 × 108 m/s)
vii
Materials
3-MPA 3-mercaptopropionic acid
AA acrylic acid
Ac acetate ion (CH3COO-)
AKD alkylketene dimer
ATO antimony tin oxide
C60 fullerene
CdS cadmium sulfide
CdSe cadmium selenide
DMSO dimethyl sulfoxide
HW hardwood
ITO indium tin oxide
MMA methyl methacrylate
MWCNT multi-walled carbon nanotube
PAA polyacrylic acid
PAc polyacetylene
PAni polyaniline
PDADMAC poly(diallyldimethylammonium chloride)
PEDOT poly(3,4-ethylenedioxythiophene)
PEDOT:PSS poly(3,4-ethylenedioxythiophene):poly(styrene sulfonate)
PEG polyethylene glycol
PEI poly(ethyleneimine)
PET polyethylene
PMMA poly(methyl methacrylate)
PPV poly(p-phenylene vinylene)
PPy polypyrrole
PT polythiophene
PTFE polytetrafluoroethylene (Teflon)
PVDF polyvinylidene fluoride
PVK poly(N-vinylcarbazole)
PVP polyvinylpyrrolidone
SHMP sodium hexametaphosphate
SLS sodium lauryl sulfate
SW softwood
SWCNT single-walled carbon nanotube
TGA thioglycolic acid
S-SWCNT ultra-short single-walled carbon nanotube
ZnS:X doped zinc sulfide, X being dopant atom
viii
Units
a.u. arbitrary units
at. % atomic percentage
wt. % weight percentage, weight basis
vol. % volume percentage, volume basis
mol. % molar percentage, molar basis
VAC alternating current voltage
VDC direct current voltage
RPM revolutions per minute
/ ohm/square, sheet resistance for a square sample
ix
Table of Contents
Acknowledgements ........................................................................................................................ iv
Nomenclature .................................................................................................................................. v
Table of Contents ........................................................................................................................... ix
List of Figures .............................................................................................................................. xiv
List of Tables ................................................................................................................................ xx
List of Appendices ....................................................................................................................... xxi
List of Original Papers ................................................................................................................ xxii
Chapter 1 ......................................................................................................................................... 1
1 Introduction ................................................................................................................................ 1
1.1 Motivation ........................................................................................................................... 4
1.2 Related work on printed LEDs ........................................................................................... 6
1.3 Approach ............................................................................................................................. 9
1.4 Thesis structure ................................................................................................................. 11
Chapter 2 ....................................................................................................................................... 14
2 Inkjet printing ........................................................................................................................... 14
2.1 Inkjet printer types ............................................................................................................ 14
2.1.1 Continuous (CIJ) ................................................................................................... 16
2.1.2 Thermal (DOD) ..................................................................................................... 16
2.1.3 Piezoelectric (DOD) .............................................................................................. 17
2.2 Piezoelectric inkjet printing: fluid dynamics .................................................................... 20
2.2.1 Ink ejection ............................................................................................................ 20
x
2.2.2 Droplet formation .................................................................................................. 21
2.2.3 Droplet impact with substrate ............................................................................... 22
2.3 Inks .................................................................................................................................... 24
2.3.1 Typical composition .............................................................................................. 24
2.3.2 Fluid properties ..................................................................................................... 26
2.3.3 Orthogonal solvent systems .................................................................................. 33
2.4 Print quality ....................................................................................................................... 34
2.4.1 Resolution ............................................................................................................. 34
2.4.2 Roughness & topography ...................................................................................... 36
Chapter 3 ....................................................................................................................................... 40
3 Materials for inkjet-printed electronics .................................................................................... 40
3.1 Conductors ........................................................................................................................ 41
3.2 Semiconductors ................................................................................................................. 49
3.2.1 Inorganics .............................................................................................................. 50
3.2.2 Organics ................................................................................................................ 55
3.3 Insulators ........................................................................................................................... 57
3.4 Encapsulants ..................................................................................................................... 60
3.5 Substrates .......................................................................................................................... 60
Chapter 4 ....................................................................................................................................... 64
4 Light-emitting devices ............................................................................................................. 64
4.1 Light-emitting diodes (LEDs) ........................................................................................... 65
4.2 Electroluminescent devices (ELDs) .................................................................................. 69
4.3 Suitability for printing ....................................................................................................... 71
4.4 Characterization ................................................................................................................ 73
xi
Chapter 5 ....................................................................................................................................... 75
5 Approach & method development ........................................................................................... 75
5.1 Materials selection model: semiconductor ........................................................................ 75
5.2 Ink formulation model: conductor .................................................................................... 76
5.3 Film formation model: insulator ....................................................................................... 77
5.4 Multilayer device model: LED ......................................................................................... 78
Chapter 6 ....................................................................................................................................... 80
6 Materials & methods ................................................................................................................ 80
6.1 Materials selection ............................................................................................................ 80
6.1.1 Electrodes .............................................................................................................. 81
6.1.2 Insulators ............................................................................................................... 81
6.1.3 Charge-transporters ............................................................................................... 82
6.1.4 Emitters ................................................................................................................. 82
6.1.5 Substrates .............................................................................................................. 86
6.2 Ink formulation ................................................................................................................. 88
6.2.1 PEDOT:PSS inks .................................................................................................. 90
6.2.2 BaTiO3 ink ............................................................................................................ 92
6.2.3 ZnS inks ................................................................................................................ 92
6.3 Jetting ................................................................................................................................ 94
6.4 Drop spacing and film formation ...................................................................................... 95
6.5 Ink distribution and print quality ...................................................................................... 96
6.6 Functional testing of individual layers .............................................................................. 96
6.6.1 Conductive ink ...................................................................................................... 96
6.6.2 Emissive ink .......................................................................................................... 97
xii
6.6.3 Insulating ink ........................................................................................................ 98
6.7 Device fabrication and testing .......................................................................................... 99
6.7.1 Interlayer dissolution .......................................................................................... 101
6.7.2 Electrical characterization ................................................................................... 101
Chapter 7 ..................................................................................................................................... 102
7 Results & discussion .............................................................................................................. 102
7.1 ZnS synthesis .................................................................................................................. 102
7.1.1 Mn2+
loading ....................................................................................................... 102
7.1.2 Zn2+
to S2-
ratio ................................................................................................... 107
7.1.3 Post-synthesis capping ........................................................................................ 108
7.1.4 Reaction temperature and time ........................................................................... 110
7.1.5 Reduction of particle size & improvement of dispersion ................................... 113
7.1.6 Optimized synthesis procedure ........................................................................... 115
7.1.7 Synthesis of ZnS:Cu nanoparticles ..................................................................... 116
7.1.8 Characterization of ZnS:Mn, ZnS:Cu nanoparticles ........................................... 118
7.2 Other materials ................................................................................................................ 123
7.2.1 BaTiO3 ................................................................................................................ 123
7.2.2 PEDOT:PSS & CNTs ......................................................................................... 124
7.2.3 Substrates ............................................................................................................ 129
7.3 Ink formulation ............................................................................................................... 137
7.3.1 Conductive ink .................................................................................................... 137
7.3.2 ZnS inks .............................................................................................................. 152
7.3.3 BaTiO3 ink .......................................................................................................... 155
7.3.4 Optimized ink formulations ................................................................................ 156
xiii
7.4 Ink film formation ........................................................................................................... 158
7.4.1 Drop spacing ....................................................................................................... 158
7.4.2 Film topography .................................................................................................. 161
7.4.3 Interlayer interactions ......................................................................................... 166
7.5 Functional testing of individual layers ............................................................................ 172
7.5.1 PEDOT:PSS/SWCNTs ....................................................................................... 172
7.5.2 ZnS ...................................................................................................................... 173
7.5.3 BaTiO3 ................................................................................................................ 174
7.6 EL device testing ............................................................................................................. 175
Chapter 8 ..................................................................................................................................... 182
8 Conclusions ............................................................................................................................ 182
8.1 Major findings ................................................................................................................. 183
8.2 Recommendations & future work ................................................................................... 186
References ................................................................................................................................... 190
xiv
List of Figures
Figure 1.1. Example of a subtractive photolithography process ..................................................... 2
Figure 1.2. Example of an additive printing process ...................................................................... 2
Figure 1.3. Flowchart outlining general experimental approach. ................................................ 10
Figure 2.1. Schematic representation of the three main inkjet printer types. .............................. 15
Figure 2.2. Sample model fluid waveform and a typical waveform used in this study. ............... 17
Figure 2.3. Fujifilm-Dimatix DMP2831 Dimatix Materials Printer ............................................ 19
Figure 2.4. Schematic of acoustic wave propagation in a piezoelectric nozzle. ........................... 21
Figure 2.5. Drop formation from a piezoelectric inkjet nozzle. .................................................... 22
Figure 2.6. Jetted drop behaviour on a substrate, showing the “coffee-ring” effect. .................... 23
Figure 2.7. Fluid properties, and effects of deviation on jetting. .................................................. 28
Figure 2.8. Schematic representation of “peak and valley” topography formed during printing. 37
Figure 2.9. Schematic representations of common line morphologies. ........................................ 38
Figure 2.10. Drop spacing of QD/polymer/water ink on slide glass. ........................................... 38
Figure 3.1. Carbon nanostructures. ............................................................................................... 44
Figure 3.2. PAni structure ............................................................................................................. 47
Figure 3.3. PEDOT (left) and PSS- (right) structures. .................................................................. 47
Figure 3.4. Simplified electronic band structures. ........................................................................ 50
Figure 3.5. Film formation in inks containing inorganic nanoparticles. ....................................... 57
xv
Figure 3.6. Microcracking in sol-gel-derived BaTiO3 spun films on glass .................................. 59
Figure 4.1. Schematic of an OLED and its energy level diagram. ............................................... 65
Figure 4.2 Typical planar LED structures. .................................................................................... 67
Figure 4.3. Improvement of carrier mobility by polymer embedding of QDs used in QDLEDs. 67
Figure 4.4. ELD structures. ........................................................................................................... 69
Figure 4.5. PM LED array for testing: schematic (left) and setup (right). ................................... 74
Figure 5.1. Printed-coated ACPEL on paper ............................................................................... 79
Figure 6.1. Solvent selection for different device component inks. ............................................ 89
Figure 6.2. Light-emitting device builds prepared in this study. .................................................. 99
Figure 6.3. Schematic drawing showing a bird's-eye view of device construction. ................... 100
Figure 7.1. PL emission from ZnS:Mn nanoparticles (1.5 at.% Mn). ........................................ 103
Figure 7.2. PL intensity vs. Mn added (at.%) ............................................................................. 104
Figure 7.3. PL of uncapped ZnS:Mn nanoparticles suspended in water. ................................... 105
Figure 7.4. PL intensity at 608 nm vs. actual Mn content. ......................................................... 106
Figure 7.5. PL emission in uncapped ZnS:Mn particles (50% Mn), different Zn:S ratios in
reaction solution. ......................................................................................................................... 107
Figure 7.6. Effect of capping agents added after synthesis on PL of ZnS:Mn. .......................... 109
Figure 7.7. PL spectra of ZnS:Mn nanoparticles capped with SHMP ........................................ 110
Figure 7.8. PL spectra of ZnS:Mn nanoparticles capped with SHMP aged for varying times ... 111
xvi
Figure 7.9. PL spectra of ZnS:Mn nanoparticles capped with varying amounts of SHMP. ....... 112
Figure 7.10. ZnS:Mn nanoparticles in water (0.1 w/w%), different stabilizers .......................... 114
Figure 7.11. Finalized synthesis method of water-dispersible 3-MPA-capped ZnS:Mn ............ 115
Figure 7.12. Comparison of ZnS:Mn dispersion in water (1 w/w%). In all of the images, the vial
on the left contains ZnS:Mn capped with SHMP; the right vial, ZnS:Mn capped with 3-MPA. 116
Figure 7.13. PL and PLE spectra of ZnS nanoparticles capped with 3-MPA. ........................... 117
Figure 7.14. Mechanisms of light emission in ZnS:Mn and ZnS:Cu. ....................................... 119
Figure 7.15. XRD spectra of ZnS nanoparticles and bulk material ............................................ 121
Figure 7.16. TEM micrographs of ZnS:Mn nanoparticles. ......................................................... 122
Figure 7.17. DLS scans of ZnS:Mn and ZnS:Cu nanoparticles in water (ZnS:Mn) and toluene
(ZnS:Cu). .................................................................................................................................... 123
Figure 7.18. DLS-obtained particle size distribution of BaTiO3 (5 w/w%). ............................. 124
Figure 7.19. DLS-obtained particle size distribution of aqueous PEDOT:PSS .......................... 125
Figure 7.20. STEM micrographs of MWCNTs .......................................................................... 126
Figure 7.21. Zeta-potential of CNT/SLS solutions and 50/50 (v/v) PEDOT:PSS/CNT mixtures.
..................................................................................................................................................... 127
Figure 7.22. Printed PEDOT:PSS conductivity differences between different commercial paper
types. ........................................................................................................................................... 129
Figure 7.23. Conductivity of printed PEDOT:PSS (single layer) as a function of added filler. . 130
Figure 7.24. ToF-SIMS maps of relative distribution of PEDOT, PSS, PEI, and TiO2 ............. 132
xvii
Figure 7.25. Estimated conductivity of PEDOT-SWCNT ink on SW fibres. ............................ 133
Figure 7.26. Relative distribution of PEDOT:PSS and PDADMAC in HW sheet ..................... 134
Figure 7.27. Estimated conductivity of PEDOT-SWCNT printed ink on HW fibres ................ 134
Figure 7.28. Cross-sections of printed SW handsheets (30% filler) showing PEDOT:PSS ink
penetration. .................................................................................................................................. 135
Figure 7.29. Greycale ToF-SIMS images of PEDOT distribution on unfilled (0% TiO2) ......... 136
Figure 7.30. Effect of added water on PEDOT:PSS suspension viscosity ................................. 138
Figure 7.31. Effect of added DMSO on PEDOT:PSS/glycerol mixture’s viscosity .................. 139
Figure 7.32. 2-point resistance of cast-coated PEDOT:PSS/glycerol/water films ..................... 139
Figure 7.33. Surface tension and viscosity in PEDOT:PSS inks with different surfactants ....... 141
Figure 7.34. PEDOT:PSS ink droplet formation during ejection from DMP2831 cartridge
nozzles. ........................................................................................................................................ 142
Figure 7.35. Printed patterns of PEDOT:PSS inks on acetate .................................................... 143
Figure 7.36. 2-point estimated conductivity of printed PEDOT:PSS inks on cellulose acetate . 145
Figure 7.37. Raman spectra (excitation wavelength = 785 nm) of PEDOT:PSS inks................ 147
Figure 7.38. Conductivity of printed PEDOT:PSS-SWCNT ink (SLS surfactant) at varying
SWCNT loadings ........................................................................................................................ 149
Figure 7.39. Conductivity of inkjet-printed PEDOT:PSS-carbon composites on acetate .......... 150
Figure 7.40. ToF-SIMS maps of PEDOT and substrate component distribution ....................... 151
Figure 7.41. Printed ZnS:Mn/AA ink on cellulose acetate, 10 printed layers. ........................... 153
xviii
Figure 7.42. Droplet formation of ZnS inks at 10 µs intervals. .................................................. 154
Figure 7.43. Optical microscope imaging of printed ZnS:Mn/PVP ........................................... 155
Figure 7.44. Droplet formation of BaTiO3 ink at 5 µs intervals. ................................................ 156
Figure 7.45. (a) Jetted drops of BaTiO3 ink on ITO PET; (b) edge of BaTiO3/PMMA film on
ITO glass. .................................................................................................................................... 156
Figure 7.46. Drop sizes of BaTiO3/PMMA ink on various substrates. ....................................... 158
Figure 7.47. Printed lines of BaTiO3 ink (single jet, single layer) at different drop spacings. ... 159
Figure 7.48. 3-D profile (left) and 2-D linescan (right) of printed BaTiO3 on slide glass. . ..... 162
Figure 7.49. SEM micrographs of printed BaTiO3 ink ............................................................... 162
Figure 7.50. BaTiO3/PMMA average dried ink film thicknesses on slide glass. ....................... 163
Figure 7.51. Topography of single printed layers of all inks on slide glass. ............................. 165
Figure 7.52. Multiple layers of ZnS:Mn/PVP ink showing edge splattering with excessive
overprinting. ................................................................................................................................ 167
Figure 7.53. ITO dissolution by ZnS:Cu/PVK ink, containing 3-MPA, and ZnS:Mn/PVP ink,
containing TGA .......................................................................................................................... 168
Figure 7.54. Optical profilometry of printed ink layers, on the surfaces they would cover in
printed devices ............................................................................................................................ 170
Figure 7.55. Reduction of “peak-and-valley” topography in ZnS:Cu/PVK films with successive
overprints on slide glass. ............................................................................................................. 171
Figure 7.56. PEDOT:PSS/SWCNT ink conductivity when printed on slide glass. .................... 172
Figure 7.57. Estimated relative dielectric constants of printed BaTiO3 films ............................ 174
xix
Figure 7.58. Current-voltage (I-V) and luminance-voltage (L-V) characteristics of
PEDOT:PSS/SWCNT – ZnS:Cu/PVK – Al LED ...................................................................... 176
Figure 7.59. Current-voltage (I-V) and luminance-voltage (L-V) characteristics of
PEDOT:PSS/SWCNT – ZnS:Cu/PVK – PEDOT:PSS/SWCNT LED ....................................... 176
Figure 7.60. Electronic band structures (top), device architecture (middle) and EL emission
(bottom) of printed ZnS:Cu DC-LEDs ....................................................................................... 177
Figure 7.61. Current-voltage (I-V) and luminance-voltage (L-V) characteristics of
PEDOT:PSS/SWCNT – ZnS:Mn/PVP – Al DCPEL ................................................................. 180
xx
List of Tables
Table 2.1. Typical inkjet ink composition ................................................................................... 25
Table 4.1. ELD/LED materials and layer properties. .................................................................. 72
Table 4.2. Characteristic properties of light-emitting devices. .................................................... 73
Table 6.1. Capping agents for ZnS:Mn nanoparticles, aqueous synthesis. .................................. 83
Table 6.2. Selected properties of commercial substrates. ............................................................ 87
Table 6.3. Selected properties of lab-made handsheets. ............................................................... 88
Table 6.4. Surfactants tested in PEDOT:PSS ink and their CMCs. .............................................. 91
Table 7.3. PEDOT:PSS inks’ fluid properties. .......................................................................... 142
Table 7.5. Finalized ink formulations. ........................................................................................ 157
Table 7.6. Summary of drop sizes and line spacing for all inks. ................................................ 160
Table 7.7. Printed film roughnesses and peak-to-valley differences in ZnS and
PEDOT:PSS/SWCNT inks. ........................................................................................................ 164
Table 7.8. Ink layer thicknesses and layers required for device construction. ........................... 166
xxi
List of Appendices
APPENDIX A. ZnS nanoparticle synthesis & dispersion ............................................. 215
APPENDIX B. Procedure for conductivity estimation in PEDOT/SWCNT films ....... 223
APPENDIX C. Film thickness estimation ..................................................................... 226
APPENDIX D. Drop and line spacing optimization ...................................................... 228
APPENDIX E. Paper substrate preparation ................................................................... 231
APPENDIX F. Impermeable substrate preparation ....................................................... 232
APPENDIX G. Paper substrate characterization ........................................................... 233
APPENDIX H. Detailed ink formulations ..................................................................... 234
APPENDIX I. Ink iterations ......................................................................................... 237
APPENDIX J. Jetting waveforms ................................................................................. 251
APPENDIX K. ToF-SIMS fragments analyzed and construction of molecular maps .. 253
APPENDIX L. Printed PEDOT:PSS/SWCNTs on paper ............................................. 255
xxii
List of Original Papers
1) Angelo, Peter & Farnood, Ramin (2010): Poly(3,4-ethylenedioxythiophene):
poly(styrene sulfonate) inkjet inks doped with carbon nanotubes and a polar solvent: the
effect of formulation and adhesion on conductivity, Journal of Adhesion Science &
Technology. 24(3), 643.
2) Angelo, P.; Farnood, R. (2012): Inkjet-printed PEDOT:PSS-SWCNT films: the effect of
surfactants on jetting and electrical performance, submitted (Journal of Materials
Chemistry).
3) Angelo, P.; Cole, G.; Sodhi, R.; Farnood, R. (2012): Conductivity of inkjet-printed
PEDOT:PSS/SWCNTs on uncoated papers, Nordic Pulp & Paper Research Journal
27(2), 486.
4) Angelo, P.; Farnood, R. (2012): Conductivity of PEDOT-CNT composites, submitted
(Journal of Materials Research).
5) Angelo, P.; Farnood, R. (2012): Inkjet-printed BaTiO3/PMMA dielectric films, Ceramics
International, in press.
6) Angelo, P.; Sweeney, C.; Farnood, R. (2012): Paper-based electroluminescent devices
prepared using conventional printing/coating, submitted (Journal of Display Technology).
7) Angelo, P.; Farnood, R. (2011): Photoluminescent inkjet ink containing ZnS:Mn
nanoparticles as pigment, Journal of Experimental Nanoscience, 6(5), 473.
8) Angelo, P.; Chlebowski, M.; Farnood, R. (2012): Mn2+
incorporation into ZnS
nanoparticles prepared in aqueous solution, submitted (Journal of Nanoparticle
Research).
9) Angelo, P.; Kronfli, R.; Farnood, R. (2012): Synthesis and inkjet printing of aqueous
ZnS:Mn nanoparticles, Journal of Luminescence, in press.
1
Chapter 1
1 Introduction
Electronic devices of all types are ubiquitous in today’s society. The production of electronics
has therefore been established over the past several decades as an integral industry for the
sustenance and development of the modern world in virtually all of its aspects (Sedra & Smith,
1997). Devices such as integrated circuits, sensors, displays, and lighting systems comprise a
tremendous fraction of current technology and are critical to industrial, social, and scientific
endeavours. In particular, display devices of all types are necessary to establish an interface for
technological interaction and for the conveyance of information. Markets have shifted away
from conventional cathode-ray-tube (CRT) technology towards flat-panel displays (FPDs) such
as liquid-crystal displays (LCDs), plasma display panels (PDPs), and electroluminescent displays
(ELDs), pushing peripheral technologies involved in display fabrication forward (DFF 2008,
Mentley 2002). The global market for these displays is expected to reach a value of over US
$100 billion by the year 2015, with the ELD market in particular experiencing the largest
compound annual growth rate out of the FPD classes – 5.3% over a 4-year analysis period, based
on studies of 249 different companies (Electronics.ca Publications 2011). It is self-evident from
these figures that the production and sale of FPDs and their components represents a
tremendously lucrative market. Research into the reduction of materials or manufacturing costs,
as well as the development of novel platforms for display applications, is therefore an area
attracting well-deserved interest. Furthermore, many of the components used to construct
displays, such as the conductive electrodes and the semiconducting phosphors, may also be used
in the construction of other electronic devices (Dimitrijev 2005, Chen et al. 2011), meaning that
their development can contribute to markets beyond that of FPDs.
The gradual development of increasingly complex devices, as well as their miniaturization, has
necessitated constant invention of new fabrication techniques and equipment (Jaeger & Balock
2003). Naturally, these methods have also themselves become complex and involved as
electronic components and their materials of construction shrink to the micro- or nano-scale
(Mahalik 2006). Production has become an intricate process involving multiple deposition steps
2
of a variety of high-quality materials, often at high temperature or vacuum, utilizing aggressive
or toxic chemicals, in a cleanroom environment (Campbell 2001). For the production of FPDs,
which are capable of demonstrating ever-increasing pixel counts, extremely careful patterning to
improve the display resolution necessitates the use of labour-intensive photolithography methods
(Jaeger 2002). In all cases of electronics manufacturing, some degree of patterning is required,
which generally is achieved by either masking areas to prevent material from being deposited on
them, or by selective removal of material after deposition (Figure 1.1). Both of these methods
Figure 1.1. Example of a subtractive photolithography process requiring both masking and etching.
imply material wastage; in the former case, wasted material is deposited on the mask, and in the
latter, a “subtractive” method, material is etched away and removed. They also imply additional
processing stages – namely mask fabrication, swapping out of masks between stages,
maintenance of masks, and etching. Throughput, efficiency, and economic return would all
benefit from the use of an “additive” method, where material is deposited in the correct pattern at
each stage of fabrication, with little to no material wastage (Figure 1.2).
Figure 1.2. Example of an additive printing process using digital patterning rather than masking.
substrate
oxide
photoresist
1) clean substrate
2) coat photoresist
3) expose to UV through photomask
mask/UV
4) rinse solubilised photoresist
5) acid-etch oxide
6) dissolve remaining photoresist
bare substrate
printhead
wet precursor
carrier solvent
1) clean substrate 2) print pattern (computer-
controlled) 3) heat or UV cure
3
The standard substrates of silicon or glass upon which much of the microelectronic industry is
built cannot take advantage of the unique opportunities offered by roll-to-roll (R2R)-type
processing, despite their generally desirable electronic or surface properties. Naturally, their stiff
and brittle nature makes them unsuitable for R2R printing in its conventional sense. However,
even without considering the issue of flexibility, silicon and glass suffer from one major
technological chokepoint – they are limited in terms of size. Silicon wafers generally cannot be
made larger than ~300 mm diameter (O’Mara et al. 1990), and the most advanced new 10th
-
generation display-grade glass cannot be cut in sizes larger than ~2 3 m (American Ceramic
Society 2010). As large as this glass might be, its growth is limited by the rate of advancements
in materials science and practical considerations of handling. Scale-up requires extensive work
on improvement of the physical properties of the substrates, and tremendous investment in
infrastructure for processing, including new dies, chambers, masks, and so forth, and it is
generally more common simply to reduce microelectronic feature size while retaining smaller
substrates (Moore 1965). Scale-up in R2R processing, conversely, is almost entirely digital:
patterns can be remade and scaled with a keystroke, and substrates of practically limitless size
can be readily handled.
With interest in flexible electronics constantly increasing (Whitmarsh 2005), the possibility of
R2R fabrication is a particularly attractive one, due to its high-throughput, continuous process
model. The use of R2R processing implies the use of flexible substrate materials to support
electronic devices, which in turn implies certain restrictions on the process, such as a limited
temperature range, usually < 200°C (to accommodate polymers, textiles, and paper). These
substrates present both advantages and challenges, ranging from mechanical flexibility to
biodegradability in the former, and from mechanical robustness to porosity in the latter.
Innovative materials selection and integration, for both the liquid precursors and the substrates,
therefore, is the dominant factor in the realization of R2R-printed electronics.
For increasingly complex printed devices, more and different materials will be required,
necessitating the establishment of ink “libraries” applicable to different substrates and underlying
layers at a variety of processing conditions (Magdassi 2010). These materials must, by
definition, be suitable for application in the liquid phase, forming features of tightly-controlled
4
size and morphology, while retaining functionality. Also, in most cases, these materials are
nano-sized structures or particles due to the restrictions on particle size imposed by the use of
inkjet nozzles (Tekin 2008). So, it becomes immediately evident that a thorough understanding
of a wide range of subjects, from microelectronic construction to colloid chemistry, is involved
in the realization of printed electronic devices.
1.1 Motivation
Inkjet printing is currently only sporadically used for the production of electronic devices,
generally to provide a single layer, most often an electrode. However, the ever-growing library
of jettable materials presents the possibility of combining inkjetted layers into more complex
devices. Naturally, the specific material requirements determine which materials must be
incorporated into inks. The ink formulation process may be very complex, depending on the
material being used, its compatibility with carrier solvents and other ink components, and its ease
of dissolution or dispersion. Ink formulation therefore becomes an often-overlooked but
absolutely crucial aspect of functional material deposition and device fabrication. This is an
involved process, requiring careful control of fluid properties, wetting behaviour, drying
behaviour, interaction with a given substrate, maintenance of dispersion, and above all, retention
of functionality. A consideration of the level of retained functionality, relative to conventionally-
produced electronics, is a natural component of this process. To summarize, there is a large gap
between a simple suspension or solution and an actual ink suitable for producing electronic
structures (Caglar 2009, Magdassi 2010). An effort to establish and refine the ink formulation
process would prove to be an invaluable tool for the rapid development of the aforementioned
material libraries.
The potential applications of such materials are nearly limitless. Certain applications might
require simple single-layer printed structures, some multilayer stacks. Some concrete examples
of the former type include radio frequency identification (RFID) tags, antennas, or simple printed
wiring using conductive material. They might also include printed patterns of light-emitting or
absorbing species. If biomolecules are printed, they might function as pathogen-detecting or
5
even pathogen-deactivating sensors. Multilayer devices could include any conventionally-
prepared planar electronic device, ranging from thin-film-transistors (TFTs) to photovoltaics
(PVs) to light-emitting devices (LEDs). In this work, the LED served as an ideal model device
for the demonstration of the inkjet deposition process as it pertained to microelectronic
fabrication.
Therefore, there was one overriding objective in this study – the production of a functional LED
using inkjet printing as the sole unit operation, as a demonstration of the possibility of inkjet
deposition of a range of materials. Within this larger objective, several smaller goals were
delineated. Firstly, the development of several customized inkjet inks, each bearing a particular
functional material, was necessary. A methodology for analyzing their performance during
jetting, upon contact with the substrate, and upon curing or drying was integral to this
development. While these analyses of performance occasionally represented relatively simple
characterization (e.g. fluid properties), the interconnectedness of each variable in ink formulation
meant that this first objective was not a trivial one. Other work has reported the use of simple
dispersions or solutions of materials as inks (as is discussed in the following sub-section);
optimal jetting, drying, and electronic behaviour were highly dependent on thorough analysis of
more detailed ink formulations, however.
A second sub-objective was to utilize these optimized inks to produce films and test their
functionality, particularly in comparison to materials not deposited (or able to be deposited) by
inkjet. As will be discussed in more detail in Chapter 3, these structures included conductive,
insulating, and semiconducting layers. So, in a multilayer device, it was a necessity to establish
each layer’s electrical performance as sufficient for device application. Furthermore, in certain
cases, the substrate was expected to have a bearing on layer performance. So, any materials
likely to be in contact with an underlying layer with which they might physicochemically interact
had to be tested for functionality on a variety of different substrates to establish the optimal
surface upon which to deposit them.
The final sub-objective was to stack those sufficiently functional layers into an LED. There
being many different types of LED, multiple structures were expected to be tested. Also, due to
6
the likelihood of interlayer interactions, certain materials were only applicable in combination
with certain other materials. Emission of light was considered proof of concept for a printable
LED – optimization of its electrical properties was beyond the scope of this work.
1.2 Related work & state of the art
The notable contributions of this work were to further develop the science behind inkjet-ready
materials development, and to examine the potential and performance of such materials in fully-
printed devices. The challenges associated with such a novel product as a printed LED
necessitated the development and optimization of specialized processes for materials preparation
and handling.
1.2.1 Inkjet-ready electronic materials
As will be discussed in detail in the body of the thesis, inkjettable materials are specialized in
several ways – the two most notable being sub-micrometre particle size and liquid dispersibility.
Of course, many materials with these properties have already been prepared and studied:
conductive metallic/polymeric colloids, carbon nanostructures, ceramic nanoparticles/sol-gels,
and colloidal semiconductors (quantum dots). Quantum dots, in particular, have been
extensively studied as useful materials for optoelectronic devices. In order to control particle
size and dispersibility, these materials are generally prepared using a “bottom-up” approach –
chemically reacting their component atoms into nanoparticle seeds that are grown in solution,
with growth being arrested by the presence of a compound coating their surface, referred to as a
“cap”. The cap, depending on its chemistry, then determines the stability of a dispersion of the
quantum dots in a given carrier solvent.
Many different types of quantum dots have been prepared for solution-based processing. The
most famous of these are the CdS or CdSe quantum dots, which can be tailored to produce light
emission/absorption at a variety of wavelengths. For displays or light-emitting diodes,
conventional technology has often relied on a different type of semiconductor, however – one
containing a dopant (impurity) atom, from which a single characteristic wavelength is emitted.
Some very common examples of this include yttrium-based phosphors (such as Y2O3) and zinc-
7
based phosphors, including ZnS, which is ubiquitous in electroluminescent display technology,
and was therefore considered a prime candidate for development of a light-emitting device.
The science behind inkjet deposition, however, is more complex than the simple production of
liquid droplets. As will be discussed in the following chapter, inkjet fluids must meet certain
rigorous criteria in order to jet at all, and to form smooth films on a given substrate. Moreover,
the dispersion of quantum dots in a given solvent – which must then be generally mixed with
other components to produce an ink with suitable properties for jetting – is paramount to
maintain printer function. In the literature, the myriad synthesis methods for quantum dots rarely
consider redispersion, let alone redispersion into a multicomponent mixture. There has not been
a work, to date, in which as-synthesized quantum dots are dispersed into a stable ink and jetted
while retaining photoluminescent and electroluminescent brightness. In this work, such
methods as exist for the production of doped nanoparticles are more closely examined. More
importantly, a refined synthesis drawing upon these previous methods is presented, by which
true, monodisperse quantum dots of doped ZnS are produced, with size and surface functionality
suitable for jetting (in aqueous and organic media) and subsequent retention of function. This
has hitherto not been demonstrated for doped ZnS nanoparticles of any type.
1.2.2 Inkjet-printed displays
Inkjet printing of electronics has attracted interest in the past several years. It is a useful means
for the complete fabrication of several film-based electronics, such as transistors (Chung et al.
2010, 2011; Tseng & Subramanian 2011, Liu et al. 2005), sensors (Molina-Lopez et al. 2012),
and capacitors (Lim et al. 2010) which often not only require careful patterning, but also involve
the use of expensive or exotic materials. Some of the advantages afforded by inkjet printing
include single-step processing and low firing and curing temperatures, with relatively high
device resolution (Calvert 2001, Tekin et al. 2004, 2008; Sirringhaus & Shimoda 2003, de Gans
et al. 2004, Yoshioka & Jabbour 2006). Attempts at similarly preparing LEDs have also been
undertaken (Wood 2009, Haverinen 2010); however, fully-printed LEDs have not been realized.
Because inkjet printers require very small particle sizes in the inks (Tekin et al. 2004, Meixner et
al. 2008), printable LEDs require specialized materials as light-emitting species. Often, these
8
take the form of nanoparticulate semiconductors, also often referred to as quantum dots (QDs)
when below a certain size (Sun 2005), or certain organic polymers or molecules. In the existing
research and development conducted on printed displays, the emissive layer alone was printed,
with the remainder of the device fabricated by more conventional means.
The types of materials suitable for inkjet printing of such displays have been solution-processed
by means other than printing in several studies. Well-known cadmium-based nanoparticulate
semiconductors, such as CdS, CdSe, or CdTe, have been used in LEDs studied by Gaponik et al.
(1999), Kumar et al. (1997), Colvin et al. (1994), Yang & Holloway (2003), Mattousi et al.
(1998), and many others. In these cases, the nanoparticles were suspended in a conductive or
semiconducting binder material, or densely packed, by solution processing – usually spin-
coating. Similarly, zinc-based nanoparticles (such as ZnO and ZnS) have been used. ZnS:Mn
has been applied by dispersing it in a cyanoresin paste and screen printing the composite (Adachi
et al. 2007, 2008). Schrage et al. reported a spin-coated single-layer ZnS:Cu/PVK composite
LED (2010). A similar structure using ZnS:Mn, spin-coated with no polymer binder but on top
of a spin-coated PVK layer, was also reported by Yang et al. (2003). A more complex structure,
using PVP-capped ZnS quantum dots with multiple vacuum-deposited charge transport layers
was reported by Manzoor et al. (2003). Another simple single-layer structure using ZnS
quantum dots in polymeric matrices was reported in a patent by Hieronymas (2002). The
motivation for the use of nanoparticulate inorganic semiconductors rather than organic emitters
has generally remained the same: their narrow emission peaks with ~20 nm full-width at half-
maximum (Dabbousi et al. 1997) versus the wider emission spectra of organics, at ~100 nm full-
width at half-maximum (Xing et al. 2005). Besides this, organics may be extremely expensive,
difficult to synthesize, and challenging to process and maintain (Kwong et al. 2005).
With the efficacy of such materials established in these studies, a more elegant and rapid means
of patterning – inkjet printing – appeared to be well-suited to applying them. However, the
realization of printed LEDs was not as simple as expected. Initial attempts simply used the
photoluminescent properties of printed QDs to produce light of various colours, by exciting them
through electroluminescent emission of an adjacent bulk phosphor layer (Wood 2009, Taylor
2007). Although this method did not actually cause light emission from the QDs themselves, and
9
furthermore, required deposition of a bulk phosphor material, it did allow for the use of
alternating-current (AC) drive. Eventually, Haverinen et al. (2009, 2010) applied CdSe QDs by
inkjet as the sole emissive layer in a complex device stack incorporating several charge transport
layers, showing promising device performance with several different colours of QDs. While it is
evident from that study that the inkjet production of an LED from such luminescent materials has
been at least partially realized, this work attempts to further describe the role of the printing
process in such a fabrication. Moreover, in one of the iterations of the devices reported in this
thesis, the entirety of the device is printed, rather than only the single emissive layer. This
represents the first demonstration of a simple light-emitting device prepared entirely by inkjet at
low-moderate temperatures and atmospheric pressure. This clearly represents a major
advancement in the field of printed electronics and opens the door to the realization of solution-
based R2R systems as the sole unit operations in electronics manufacturing.
1.3 Approach
Initially, the focus of the work was to use conventional printing and coating techniques to
produce LEDs (in particular, ELDs) using paper as a substrate. This entailed several limitations
on materials which would be pertinent throughout the remainder of the study. These included,
most notably, the use of liquid precursors and low processing temperatures. Therefore, when the
decision to attempt inkjet printing of different functional materials was made, the constraints on
the experimental approach of outlining a roadmap for ink development and application were
well-understood. When studying printable materials which might be applicable on a variety of
substrates (potentially including paper), these two specifications of liquid-phase precursors and
low-temperature processing guided the research plan.
Because several different materials and variations of those materials were used, as well as their
respective inks, studies were carried out in a “layer-by-layer” fashion. In other words, the
formulation, jetting, and functional testing of each material’s ink was initially carried out one ink
at a time. When layers of inks were stacked together, a similar approach was used to observe
jetting performance, film formation, and so forth, when in contact with the other materials.
Figure 1.3 shows a simplified flowchart of the research method for this study. Because of the
10
large number of interconnected variables associated with ink formulation and deposition, the
process was often an iterative one.
Figure 1.3. Flowchart outlining general experimental approach.
Firstly, suitable materials were chosen to produce the requisite layers. Secondly, an attempt was
made to incorporate those materials into a liquid carrier to serve as the basis of an ink. If the
Materials selection Establishment of functionality
Synthesize materials to jettable specifications
Establishment of dispersion/solution
Ink formulation
Printing studies
Droplet formation
Film optimization
Layer characterization
Interlayer interaction
Printed device Device characterization
Suitable material
available?
Dispersible
/soluble?
Jettable?
Forms desired
structures?
Functional as
desired?
Functional testing
Yes
No
Yes
No
Yes
No
Yes
No Yes
No
11
materials could not be sufficiently well-incorporated by dispersion or dissolution, the first step of
materials selection was repeated. Next, the dispersion was reformulated into an ink with specific
fluid properties for jetting. Chemical incompatibilities became a major concern at this stage, as
certain ink components and additives compromised dispersion or dissolution. The ink was then
introduced to the inkjet printer, where printer and cartridge settings had to be tailored to each
specific ink to produce ideal droplets. If an ink was unable to jet properly, ink formulation was
repeated. The exact print specifications required to produce a printed structure of the desired
conformation (a film) were then established. If the structure could not be successfully formed,
reformulation and possibly materials reselection were warranted. Finally, successfully jetted
materials were studied for functionality, which was affected by the various iterations described
above. Lastly, the interaction of adjacent layers was considered, as deleterious effects in this
regard would have a bearing on LED function.
It would be redundant to outline in this document all of the iterations of all of the inks formulated
(although they are appended to the work in APPENDIX I). Indeed, the individual inks served
unique purposes in outlining the transition from a bulk material to an inkjet-printed layer. Their
exact roles will be elaborated upon within the body of the thesis. In brief summary, a conductive
ink served as a model for the intricacies of precise formulation and its effects on function, a
semiconducting ink as an example of the difficulty of materials selection and incorporation, and
an insulating ink as a model for film formation. Finally, as has been indicated above, an LED
structure was used to study the challenges posed by interlayer interaction, film structure, and so
forth, and as a proof-of-concept of the suitability of this novel development scheme for inkjet-
printed electronic materials.
1.4 Thesis structure
The thesis is divided into eight chapters. The first four of these chapters are intended to review
the fundamental knowledge required to develop inkjet-printed electronic devices. This first
chapter has served to give an overview of the motivation and objectives of the work, as well as a
brief background into the idea of printed electronics. Chapter 2 is concerned with outlining the
12
scientific concepts behind inkjet printing and inks. Chapter 3 provides an overview of some of
the materials which are suitable for use in printed electronic applications, dividing them into the
most common classes, such as electrodes, insulators, and so forth. Substrates, including flexible
ones (polymers, paper) are also discussed here. Chapter 4 discusses the principle of light
emission from LEDs of various types, their typical structures, and means of characterizing them.
The ELD and the actual light-emitting-diode LED are focused on in particular, as they were the
best-suited to inkjet application.
The discussion of the inks prepared and deposited for this study begins in Chapter 5. This
chapter deals with the development of a process to move from a desired structure and its
component materials (covered in Chapters 3 & 4) to actual inks containing those materials, ready
for assembly into an ELD/LED. The experimental methods necessary for this development are
elaborated upon in Chapter 6, taking each material into account in particular for functional
testing of printed films. Each material used had a particular purpose which required a very
different test platform to observe its performance. Certain of the experimental methods requiring
more detailed descriptions are elaborated upon in the Appendices. Chapter 7 describes the
results of these experimental methods, comparing the observed performance of the various inks
both in isolation and in the context of a full ELD/LED. Demonstrations of device functionality
are also presented in this chapter. Finally, Chapter 8 summarizes the major conclusions, the
findings presented in Chapter 7, the limitations of the methods and devices presented, and
recommendations for future research. The Appendices illuminate some of the more detailed
experimental methods, exact ink formulations, novel experimental setups, and so forth.
Finally, the original publications listed in the prefatory material can be considered complements
to the thesis. Certain portions of these publications are included in the body of the thesis; these
sections of the thesis may be considered a distilled version of the work conducted in these
publications.
PAPER 1 describes a first attempt at formulating a conductive inkjet ink, and its application to a
flexible substrate (paper and acetate). The mechanical resilience of the resulting layer is then
studied, and the possibility of using a flexible substrate to support this layer established.
13
PAPER 2 describes the more detailed aspects of refining the formulation of the conductive ink,
comparing the effects of different surfactants in an otherwise constant formulation. Changes in
conductivity attributed to both the print quality and chemical interactions induced by different
surfactants are discussed.
PAPER 3 outlines the introduction of carbon nanotubes (CNTs) into the conductive film and
their role in conductivity enhancement. .
PAPER 4 examines the wide variability in electrical resistance in the conductive printed layer on
different paper substrates, by a comprehensive analysis of several different hand-made paper
sheets’ effect on conductivity. An optimal surface for conductive ink deposition is established.
PAPER 5 describes the formulation and application of an insulating ink, to function as a
dielectric layer. Topography and dielectric behaviour are observed to establish the feasibility of
using such an ink in an ELD.
PAPER 6 summarizes the fabrication and testing of a preliminary ELD on a flexible substrate,
using a printed conductive layer and coated insulating and emissive layers.
PAPER 7 describes an initial attempt at synthesizing and depositing ZnS nanoparticles to
function as a light-emitting layer in a fully-printed ELD. Nanoparticles are characterized for
photoluminescence, crystallinity, and size, and an ink is formulated and printed using an acrylate
polymer as a binder.
PAPER 8 further explores the synthesis of these nanoparticles, with a focus on characterizing
and improving the doping level of the nanoparticles to increase luminescent yield. Several
different pre- and post-synthetic treatments are considered to improve emission from the
nanoparticles, with an optimal synthesis method being described.
PAPER 9 describes a refined version of the synthesis method used in PAPER 8, which is used to
produce truly monodisperse nanoparticles of doped ZnS that are readily dispersible in water and
suitable for inkjet printing. Issues of print quality and substrate interaction are briefly explored,
and the ink is considered a suitable candidate for application to a rudimentary LED.
14
Chapter 2
2 Inkjet printing
Printing has been used as a patterning technique since ancient times, and until recently, has been
entirely based on physical contact between a transfer part and a substrate (Eisenstein 1997). The
transfer part, whether it is a patterned roll, stamp, block, or screen, functions on the basis of some
physical phenomenon – usually pressure, surface energy gradients, or masking (Skotheim 2006).
In terms of printing technologies, inkjet printing is unique in that it is a non-contact method,
dispensing material directly from reservoir to substrate in a predefined pattern (Le 1998,
Bucknall 2005). Because this pattern is generally defined using software, inkjet printing is
considered a “digital” printing method, which allows for limitless flexibility in pattern definition
and rapid pattern adjustment. The substrate or the nozzle(s) can be moved in two or three
dimensions as drops are formed to form the pattern. By selecting exactly where to deposit
material, the printer can avoid wastage almost entirely and can deposit very small amounts of ink
in a tightly controlled fashion (Zhouping et al. 2010, Haverinen 2010). The digitally-controlled
inkjet printer can also readily be realigned to a particular location on a substrate, allowing for
over-printing of new materials and patterns with pinpoint accuracy (Curling 2006). The latter
two features, in particular, make inkjet-printing particularly well-suited to the fabrication of
many different types of electronic devices.
2.1 Inkjet printer types
Inkjet printers can be broadly divided into two classes: continuous (CIJ) and drop-on-demand
(DOD). Each of these classes can then be further subdivided into several subclasses, including
Hertz and microdot continuous printers, and thermal and piezoelectric DOD printers (Caglar
2010). The primary difference between the two main classes involves the production of droplets;
as is evident by their name, CIJ printers produce a continuous stream of droplets from their
nozzles, whereas DOD printers eject drops only when impelled to do so. Figure 2.1 shows a
schematic of the three main printer types – CIJ, thermal-DOD, and piezoelectric-DOD.
15
Figure 2.1. Schematic representation of the three main inkjet printer types.
Certain similarities exist between the printer types. Of course, all have some system of nozzles
to allow the flow of ink and its formation into droplets. Generally, all also have some type of
pump, which can provide suction to prevent ink from dripping out of the nozzles when not
printing, and positive pressure to flush the nozzles with ink or another cleaning fluid. More
importantly, though, all require a set of electrical motors and servos to correctly position the
nozzles above the substrate to produce a pattern, according to the digital specifications of that
pattern. These motors can either move the nozzle assembly (printhead) itself, or move the
substrate underneath a fixed nozzle. Commonly, there is motion in both the printhead and the
substrate – a desktop printer being a good example, where the printhead traverses the sheet
laterally, printing a line at a time, and the paper sheet moves forward through the printer.
voltage
source
heating element
vapour bubble
droplet
voltage
source
piezo
element
substrate
nozzle
charging
plates
deflector
plate
ground
gutter
thermal (bubble-jet) piezoelectric continuous
16
2.1.1 Continuous (CIJ)
In CIJ systems, a nozzle connected directly to a reservoir is used to produce a continuous stream
of droplets, either by simple Plateau-Rayleigh instability (Papageorgiou 1995) or by induction of
a high-frequency vibration using a piezoelectric crystal (Croucher & Hair 1989), or both. As the
drops fall, a fraction of them are electrostatically charged by passing between charging plates.
As they approach the substrate, the charged droplets are selectively forced towards a grounded
deflector plate by a large potential applied across a second plate; these deflected droplets are
collected in a gutter and recycled while undeflected droplets impact the substrate (Le 1998).
This is the most mature inkjet technology, likely because of the simplicity of the nozzle assembly
and lack of microfabricated parts. Its relatively complex apparatus, including pumps, reservoirs,
and several charged plates and power sources, however, means that DOD systems are often more
compact and simple, and therefore more commonly used (Caglar 2010).
2.1.2 Thermal (DOD)
Thermal inkjet printers function by rapidly (i.e. in a few µs) heating a small amount of ink inside
a reservoir with an open nozzle using a resistor connected to a power supply. As the heat is
applied, the ink vapourizes, creating a bubble which applies a high pressure (> 1 MPa) wave to
the fluid inside the reservoir, forcing a drop out through the nozzle (Le 1998, Croucher & Hair
1989, Rembe et al. 1999). The heater then cools rapidly, before another drop is produced. An
advantage of this technology is the relatively low cost, as it involves only a resistor and a
reservoir, easily fabricated by existing MEMS techniques with low-cost materials (Molesa 2006).
Therefore, thermal inkjet systems are commonly used as household printers, where printer cost
and regular printhead replacement are relevant. An issue associated with thermal systems is their
limited applicability – usually, inks must be water-based and low-viscosity (~1 cP) to be easily
vapourized and jetted (Zhouping et al. 2010). On this same note, the high thermal stresses to
which inks are subjected may destroy sensitive ink components, such as biomolecules or
polymers, or otherwise alter the functionality of suspended materials (Haverinen 2010). Thermal
inkjet printers are therefore better-suited to the printing of conventional inks for graphics and
text.
17
2.1.3 Piezoelectric (DOD)
Piezoelectric printers, by avoiding thermal cycling issues, are suitable for a broader range of ink
formulations and materials (Clymer & Asaba 2004, Magdassi 2010). A mechanical deformation
of a small ink reservoir induces a pressure wave in the ink, forcing out a droplet (Le 1998, Ridley
et al. 1999). The component of the printhead which causes the deformation is a piezoelectric
crystal, which changes shape upon the application of potential; the use of this specialized
material means that the manufacture of piezoelectric printheads is significantly more difficult
and costly than thermal ones (Molesa 2006). Another difference between this printhead type and
a thermal printhead is the degree to which drop size and shape can be controlled – careful
adjustment of peak voltage, voltage pulse length, and even the shape of the voltage-time
waveform applied to each piezo element can alter drop formation (Fujifilm-Dimatix 2006,
Molesa 2006, Haverinen 2010, Caglar 2010). An example of such a waveform is shown in
Figure 2.2.
Figure 2.2. Sample model fluid waveform (blue) and a typical waveform used in this study (red).
t (µs)
V
rise time echo time
pulse width
slope
(“slew rate”)
step width
typical waveform
model waveform
18
The “rise time” refers to the length of the waveform which increases up to a peak voltage; the
“echo time” or “fall time” refers to the portion in which voltage drops to its minimum value.
These two regions of the waveform affect the fluid mechanics inside the print nozzles, which are
described in Section 2.2.1. If a waveform is segmented into several increases/decreases and
plateaus in voltage, each of these segments has a temporal width referred to as the “step width”.
The slope of the rising or falling segments is referred to as the “slew rate”, generally expressed in
units of V/µs. Finally, the temporal length of the entire waveform is called the “pulse width”.
Piezoelectric nozzles may also include features such as temperature control for more viscous
fluids, high-frequency vibration to prevent liquid “skins” from forming, and purging settings that
flush the nozzles with fluid. They may also include cameras to assist in realignment of the
printer between layers (Curling 2006), and to observe and optimize drop formation by adjusting
the voltage waveform. Because almost any material can be used in the piezoelectric printhead
without concern over its thermal stability or heat of vapourization, fluids with relatively high
viscosity – up to 40 cP, in some cases – can be jetted (Zhouping et al. 2010). Nozzles are usually
arranged in a line in numbers ranging from less than ten up to several hundred (Le 1998, Clymer
& Asaba 2004), and can be individually driven by different waveforms (Fujifilm-Dimatix 2006).
The flexibility and almost universal applicability of the piezoelectric inkjet printer to functional
materials made it ideal for a study of ink development for electronics deposition. As will be
discussed in Section 2.3, however, the ink itself must meet certain criteria before the unique
advantages offered by piezoelectric DOD printing can be realized.
2.1.3.1 Fujifilm-Dimatix DMP2831 Dimatix Materials Printer
An example of a testbed piezoelectric printer, produced by Fujifilm-Dimatix (formerly Dimatix),
is the DMP2831, shown in Figure 2.3 (images courtesy of Fujifilm-Dimatix). This bench-scale
printer offers piezoelectric inkjet printing capability for a limitless number of inks and substrates.
The 1.5 mL cartridges of this printer may be filled and refilled with any type of ink, based on any
solvent system; more rugged cartridges exist for aggressive solvents like toluene and mineral
acids. The printheads’ sixteen 21.5-µm-diameter, 254-µm-spaced nozzles deposit droplets of
either 1 pL or 10 pL in volume, depending on type, allowing for the printing of relatively high-
19
resolution patterns. The droplets’ size, shape, and speed can all be controlled by adjusting the
voltage waveform applied to the nozzles. Spacing of droplets can be controlled by angling the
line of nozzles along the printhead. The printer also provides control over suction (meniscus) to
prevent ink from dripping, high-frequency “tickling” to keep the nozzles wet, temperature of the
nozzle plate (from ambient to 80°C), temperature of the substrate (from ambient to 60°C), and
the type and frequency of cleaning cycles. A mounted fiducial camera allows for observation of
the printed surface, landmarking for alignment, and adjustment of the print origin and angle. A
second camera, the drop watcher, provides real-time imaging of drop ejection from the nozzles;
voltage waveforms and other cartridge settings can be actively adjusted while monitoring drop
formation in order to find optimal values. Finally, the substrate support platen can be angled
and tilted as desired to print more complicated structures. The DMP2831 therefore contained all
of the features required for ink development and application in this study and was an ideal tool.
Indeed, the considerable capability of the DMP2831 has been widely and successfully used in
several other studies on printed electronics. Some of these include the deposition of
PEDOT:PSS by Mire et al. (2011) and Lopez et al. (2008), PEDOT:PSS/SWCNTs by Mustonen
et al. (2007), silver by van Osch et al. (2008), quantum dots by Haverinen et al. (2009) and Small
et al. (2010), and a polymer-fullerene blend by Hoth et al. (2007).
Figure 2.3. Fujifilm-Dimatix DMP2831 Dimatix Materials Printer; magnification shows a disassembled cartridge and
printhead assembly.
20
2.2 Piezoelectric inkjet printing: fluid dynamics
The fluid dynamics occurring during printing involve the formation of pressure (acoustic) waves
due to the movement of the piezo element, which force droplets out of the nozzles, where they
fall to the surface. Each of these steps – drop formation, detachment, and impact – involves
unique fluid mechanics.
2.2.1 Ink ejection
The effect of the movement of fluid inside the ink chamber to produce jetting was described by
Shield et al (1986). A schematic of their model is presented in Figure 2.4. The motion of the ink
begins with the application of potential to the piezo element. Generally, potential is applied such
that the ink chamber enlarges, creating a negative pressure inside it, and generating two negative
pressure waves inside it, moving in opposite directions from the piezo element. These waves are
reflected from the ink reservoir and the nozzle opening, the former changing its sign during
reflection to a positive acoustic wave and the latter reflecting back as a negative acoustic wave.
As the second part of the voltage waveform (the voltage drop) occurs, the piezo element
expands, producing two positive pressure waves in opposite directions. One of these annihilates
the reflected negative pressure wave, and the other doubles the amplitude of the reflected
positive pressure wave. This propagating pressure wave forces out a droplet of ink when it is
sufficiently strong to overcome the viscous drag and surface tension acting on the ink at the
nozzle.
The entire process – and therefore, the length of a typical jetting waveform – is about 20 µs
(Wijshoff 2008). As the length of the channel and the speed of the pressure wave (i.e. the
applied voltage) will have some bearing on how long this process takes, the waveform length
will vary with applied voltage and the exact printer being used. In some cases, an “echo”
waveform is used to dampen the pressure waves in the ink chamber after the drop is ejected,
allowing the chamber to refill for the next drop; an unfilled chamber will fail to jet properly
(Molesa 2010, Wijshoff 2008). An example of an “echo” waveform is shown in Figure 2.2.
21
Figure 2.4. Schematic of acoustic wave propagation in a piezoelectric nozzle.
2.2.2 Droplet formation
As ink is ejected from the nozzle, it forms elongated droplets which fall to the surface as
spherical drops, ideally (Lee 2003, Meixner et al. 2008, Wijshoff 2008), shown in Figure 2.5.
The surface tension of the ink determines how easily it can form a drop: too low, and it will
simply spread onto the nozzle plate and cover the nozzle with a skin of liquid, also known as
overfill (Shin & Smith 2008); too high, and the pressure wave in the nozzle will be insufficient to
overcome the surface tension and eject a droplet. Surface tension will be elaborated upon in
Section 2.3.
Similarly, other fluid properties like viscosity determine how well droplets form. High-viscosity
inks form droplets with long, pronounced tails – tail length scales with viscosity, which is why
inks containing large proportions of polymers tend to display the extreme of this effect – the so-
called “bead on a string” (Mauthner et al. 2008). As long as the tail merges with the body of the
droplet during flight, or at least detaches and descends directly behind the main droplet, the
1)
2)
3) V
t
+ve potential: contraction, -ve pressure wave reflection from reservoir, tip
amplification of +ve pressure wave, jetting
1) First voltage rise, +ve potential 2) Dwell time
3) Voltage fall, -ve potential
piezo element
nozzle reservoir
22
droplet should strike the substrate where intended. However, excessively long tails may break
into several satellite droplets, which then follow different trajectories to the surface, completely
compromising print resolution (Dong et al. 2006, Meixner et al. 2008, Lee 2003). Viscosity
alone is not solely responsible for satellite formation; excessively high driving voltage or
unsuitable voltage waveforms may also cause splattering of satellite droplets by ejecting the
main droplet at too high of a speed (Haverinen 2010).
Figure 2.5. Drop formation from a piezoelectric inkjet nozzle. Top: schematic of drop detachment and formation into
spherical drop. Bottom: BaTiO3/PMMA ink drop forming.
2.2.3 Droplet impact with substrate
When the drops strike the substrate, they will wet it according to Young’s Equation, which
describes the relationship between the surface energy of the substrate and that of the fluid:
c represents the contact angle, shown in Figure 2.6, between the fluid and the substrate, and
the surface energy. As drops impact the surface, they will deform (because of their speed), and
𝛾𝑠𝑜𝑙𝑖𝑑 −𝑙𝑖𝑞𝑢𝑖𝑑 + 𝛾𝑙𝑖𝑞𝑢𝑖𝑑 −𝑣𝑎𝑝𝑜𝑢𝑟 𝜃𝑐 = 𝛾𝑠𝑜𝑙𝑖𝑑 −𝑣𝑎𝑝𝑜𝑢𝑟
t1 t2
t3 t8 t7
t6 t5
t4
0 µs 5 µs
10 µs 35 µs 30 µs
25 µs 20 µs
15 µs
100 µm
23
eventually, begin to evaporate. So, again, the fluid properties and voltage waveform come into
play, with regards to jetting speed, wetting ability, and finally, boiling point and volatility.
Figure 2.6. Jetted drop behaviour on a substrate, showing the “coffee-ring” effect.
Upon impact, the droplet will spread to a diameter several times its original diameter (Tekin et
al. 2008) and deform into a shape not unlike that shown in Image (3), in Figure 2.6, with thick
edges and a thin centre, due to outward deflection of momentum (Soltmann & Subramanian
2008). If the ink dries rapidly after impact, this morphology will remain. So, an ink with high
volatility and low boiling point – such as one containing a volatile organic solvent as its carrier
medium – will likely suffer from this morphological problem. This is, of course, if the ink is
able to jet, as a highly volatile solvent will also evaporate in the nozzles, clogging them with the
solids it leaves behind.
This is one manifestation of the “coffee-ring effect” (Fukai 1993, 1995). Another mechanism by
which it might occur is more rapid solvent evaporation at the edge of the droplet, and capillary
flow of solvent to the edge to replace it. The thinner layer of liquid at the edge of the droplet
compared to the centre of the droplet causes the rate of evaporation there to increase, and solvent
to migrate by capillary flow to the that region (Craster et al. 2009). Subsequent Maragoni flow
of solute to the region results in a ring with a raised edge of concentrated solute, and almost no
solute at the centre of the drop. In both cases, the mixing in of co-solvent(s) with different
boiling points than the primary carrier medium may assist in reducing this effect, by evening
evaporation rates throughout the entire droplet and reducing evaporation rates at the edge of the
droplet.
photoresist
1) Impact with substrate 2) Drying 3) Dried droplet
Capillary flow or
momentum deflection “Coffee-ring” formation
c
Solvent evaporation
24
In any case, the control of ink rheology and surface tension determines the morphology of
resulting printed features. A higher surface tension means less spreading, resulting in thickner
films with higher resolution, but poorer adhesion; a higher viscosity means less-pronounced edge
thicknening, but a generally thicker film; and so forth. Depending on the application, inks must
be carefully formulated to achieve a desired outcome on a particular substrate.
It is also worthwhile to mention that porous substrates, with paper being a prime example, have
entirely different properties when it comes to drop impingement. These structures will absorb
the ink into their bulk almost immediately, as soon as inertial forces in the drop dissipate
(Croucher & Hair 1989). After this point, the ink will sink into the surface, at a rate determined
by the pore size, pore structure, and the surface energy of the pore walls themselves (Gane 2004,
Holman 2002). These substrates present difficulties for printed electronics, as the functional
material is no longer in a cohesive film, but distributed into the substrate itself.
2.3 Inks
As has already been discussed in the two previous sections, inkjet printers require particular fluid
properties to jet reliably and consistently. Although these may vary somewhat between printers,
viscosity, surface tension, specific gravity, and particle size are always controlled (Magdassi
2010). Stable inkjet inks may also require control over dispersion (zeta-potential), foaming,
microbial growth, and pH (Karsa 2003). An ink formulation often must juggle successful jetting
and drop formation, wetting and drying behaviour on the substrate, and functional performance.
However, there are certain guidelines to ink formulation that can prove useful in designing a
suitable inkjet ink for a given application.
2.3.1 Typical composition
Although inks may be formulated in many different ways, the composition of the ink is generally
similar to that given in Table 2.1. The primary components of the ink are the “pigment” – which,
when applied to electronically functional inks, implies the functional material – and the carrier
medium or solvent. It is unlikely that a functional material will be readily available and
dispersible in a solvent with appropriate viscosity and surface tension for printing, so usually, the
25
ink must contain at least one other component to make it jettable. In some cases, this can be as
simple as a second solvent – for example, mixtures of water and common alcohols like EtOH,
MeOH, or IPA can produce solutions with low surface tension, one high-boiling solvent (water)
and one low-boiling solvent (alcohol) (Vazquez et al. 1995). However, this example of a
mixture does not address the issue of viscosity, which would be too low. Therefore, another co-
solvent is often added, with a relatively high viscosity. Such a material may also function as
what is called a humectant, which prevents rapid drying of the ink in the printer nozzles
(Croucher & Hair 1989). A surfactant may also be required to adjust surface tension to a suitable
level. So, as the ink components become more and more sensitive to the presence of additives,
or more specialized materials are used, the ink formulation may become increasingly complex.
In certain cases, such as the printing of biomolecules (Di Risio & Yan 2007), formulation and
ink functionality become very closely entwined, and an iterative approach to achieving the
proper viscosity and surface tension is necessary.
Table 2.1. Typical inkjet ink composition, as described by Tekin et al. (2004, 2008); Le (1998), Croucher & Hair
(1989), Zhouping et al. (2010), Magdassi (2010), and Calvert (2001).
Component Function Loading (w/w%)
Dye, pigment, or functional material
Key component 0.1 – 10
Solvent Dispersion/dissolution medium 50 – 90
Co-solvent(s) Controls drying (“coffee-ring”)
Viscosity modification Surface tension modification
0 – 50
Surfactant Modifies surface tension
Improves wetting 0 – 5
Viscosity modifier (dissolved)
Generally, increases viscosity < 1
Humectant Low volatility, prevents ink drying in nozzles 0 – 20
Other pH buffer; biocide; fungicide; dispersant; defoamer; binder
(polymer) < 1
26
There are some pitfalls that commonly arise during the selection of ink components. Firstly, the
use of dissolved polymers can be problematic, as they tend to modify the rheology of the ink at
high concentrations and molecular weights, causing poor jetting, spreading, and film formation
(Tekin et al. 2004, 2008; de Gans et al. 2004, Mauthner et al. 2008). Dissolved polymers also
tend to increase viscoelasticity, producing long filaments of ink during jetting which
subsequently may break up into small droplets and “splatter” the substrate. Secondly, inks that
require the use of aggressive solvents, including toluene/tetrahydrofuran (THF) mixtures, or
concentrated acid, can damage the print apparatus, and are to be avoided. Finally, and most
importantly, any suspended particles in the ink must be below a critical size to jet uniformly and
avoid clogging of printer nozzles. The packing of particles into a nozzle, even particles that are
smaller than the nozzle diameter by many times, may still result in an irreversible blockage of
that nozzle if they are above a certain size (Valero et al. 2007). As a large amount of particles is
passing through the nozzle at a given time, the maximum allowable particle size is generally
significantly lower than the actual nozzle diameter. Therefore, not only particle size, but
dispersion of the particles becomes an issue in pigmented inks, which are generally
thermodynamically unstable and only held in suspension by chemical or electrostatic treatments
(Wang 2002). Added chemical dispersants keep particles separated by steric or electrostatic
hindrance, overcoming the attractive London or van der Waals forces. Dispersion may therefore
be compromised by chemical or energetic stressors (Spasic & Hsu 2006), so once again, ink
formulation becomes a balance of fluid properties and ink stability. Unstable inks are not
suitable for jetting, especially when using expensive and non-recoverable piezoelectric
printheads as the deposition mechanism.
2.3.2 Fluid properties
As has been discussed in Section 2.2, viscosity and surface tension have a major impact on drop
size, drop velocity, satellite formation, and droplet morphology upon impact. Through repeated
attempts at ink formulation, certain restraints have been determined for fluid properties. Ranges
of values for these key properties are shown in Table 2.2; the effects of deviating from these
values are also listed in this table, Figure 2.7, and are further discussed below.
27
Table 2.2. Restrictions on fluid properties in jettable inks based on the work of Fujifilm-Dimatix 2006, Croucher &
Hair 1989, Magdassi 2010, Calvert 2001, Meixner et al. 2008, Zhouping et al. 2010, Wang 2002.
Property Value Problems Controlled by
Minimum Maximum
Viscosity (µ) 2 cP 20 cP
<2 cP: Droplet breakup Excessive spreading >20 cP: Droplet tails Cannot exit nozzle
Solvent Co-solvent
Viscosity modifier
Surface tension () 30 mN/m 40 mN/m
<30 mN/m: Nozzle plate wetted (cannot jet through film) >40 mN/m: Poor drop formation Unable to wet most surfaces
Solvent Co-solvent Surfactant
Particle size (dp) 0 nm 1% of nozzle
diameter >1% nozzle diameter: Rapid clogging
Dispersant Surfactant
Specific gravity (SG)
~1 1.5
<1: Ink backflows under meniscus pressure >1.5: Ink drips out (insufficient meniscus pressure)
Solvent Co-solvent Humectant
Zeta-potential (), absolute value
40 mV As high as possible
<40 mV: Precipitation of solids
Dispersant Surfactant
Energy number (En) 0 1 >1: Ink splatters on impact with substrate
Drop velocity Other fluid properties
Inverse Ohnesorge number (Z
-1)
1 10
<1: Same issues as µ > 20 cP >10: Satellite formation Poor drop formation
Nozzle size Other fluid properties
Many inkjet inks are water-based (Le 1998), and so achieving the relatively high viscosities ideal
for piezoelectric inkjet printing is problematic. Even organic-solvent-borne inks face the
challenge of increasing viscosity, as many common solvents such as toluene have viscosities of
<1 cP as well. Increasing viscosity can best be accomplished by introducing a co-solvent of
higher viscosity (Calvert 2001) or some sort of polymeric additive, like a cellulose compound
(Di Risio & Yan 2007). However, in electronic devices, the presence of a dried polymer after
ink curing may interfere with or completely compromise device function. Furthermore, high-
28
viscosity co-solvents are often high-boiling materials as well (e.g. glycerol, ethylene glycol),
which may require high-temperature or low-pressure curing to completely remove from the
printed layer. In certain special cases, this may be acceptable, such as with glycerol in printed
PEDOT:PSS films, where it actually serves to improve electrical conductivity (Lia et al. 2003).
Figure 2.7. Fluid properties, and effects of deviation on jetting.
In general, however, foreign compounds are detrimental to electronic performance. This implies
that viscosity modifiers must be cleverly chosen so as to be readily removed, or to remain as
integral parts of the device. For example, a conjugated polymer such as poly(n-vinylcarbazole)
might be dissolved in an organic solvent-based ink as a host material for an emissive molecule or
nanoparticle, while also increasing viscosity. If high-temperature curing is not an issue for the
substrate or the functional material, all of these concerns are nullified – any liquid viscosity
modifier may be chosen that can be burned off by heat treatment after deposition.
Surface tension reduction presents a unique challenge for aqueous inks, as was mentioned in the
previous section. Most other solvents have surface tension values somewhere in the jettable
range specified in Table 2.2, although some have lower surface tension values than desired, such
as acetone and certain alcohols (Colclough 1968). Therefore, water-based inks must contain
dp (nm) (mV)
unstable
drops bead-on-
a-string
flooded
nozzles
no droplets
form
ideal
jetting
immediate
clogging
precipitation
stable
dispersion
µ (cP) (mN/m)
29
either a large proportion of some co-solvent, or a surfactant. The ionic charge present on many
surfactants must be suited to the dispersed species in the ink – if the species is dispersed by
electrostactic repulsion, as is often the case, a surfactant with the opposite charge will interact
strongly with it, sometimes even causing flocculation (Karsa 2003). A surfactant with the same
charge as the dispersant material will compete for adsorption sites on the dispersed particle,
which may be beneficial or detrimental to dispersion. In many cases, the most reliable means of
reducing surface tension without compromising dispersion is the addition of non-ionic
surfactants, which only weakly interact with charged particles or their adsorbed dispersant layers
(Karsa 2003). As the non-ionic surfactants remain dissolved in solution without much
interaction with any dispersed materials, they are generally loaded in very small amounts, well-
below their critical micelle concentrations (Tadros 1987, Capek 2006). Another advantage
offered by non-ionic surfactants is that they cause considerably less foaming of the ink than their
ionic counterparts, especially anionic surfactants (Davison & Lane 2003). This can help to
prevent refill issues in ink nozzles that become flooded with air and cannot jet (Haverinen 2010),
and avoid the addition of generally immiscible silicon-based defoamers (Karsa 2003), although
even non-ionic surfactants can cause some degree of foaming. Foaming can be readily observed
by a simple shake test, followed by measuring the height of foam formed and its lifetime before
settling. If a large volume of foam is formed which does not dissipate rapidly, either a different
surfactant or a defoamer must be added to the ink (Tracton 2005, Karsa 2003).
Dynamic surface tension, the surface tension of a surfactant-laden liquid at different times after
the formation of a new interface, is often characterized in inks as well. Surface tension, as
described in Table 2.2, is equilibrium or static surface tension – i.e., the surface tension value of
an ink as a long-term asymptote of its dynamic surface tension values. However, dynamic
surface tension also plays a role in droplet formation in inks containing surfactants. As the
droplets are forming a new liquid-vapour interface (with the surrounding air) when jetting
occurs, and a certain amount of time is required for surfactant to diffuse to the interface, the
surface tension upon leaving the nozzle and that upon impact with the substrate will differ
(Eastoe & Dalton 2000). The higher value of dynamic surface tension (at time 0, or immediately
upon drop formation) compared to equilibrium surface tension may result in poor drop formation
30
or decreased substrate wetting. Furthermore, as the drops impact, the velocity of the spreading
fluid along the plane of the substrate may be faster than the diffusion velocity of the surfactant
molecules to the new interface formed with the solid (Aytouna et al. 2010). Therefore, to ensure
the best droplet formation and surface wetting in inks containing surfactants, it is ideal to utilize
surfactants which have dynamic surface tension values close to their equilibrium values at a
given concentration (Titcomb 1981). This often entails a careful consideration of the chemistry
of the surfactants in question, as the head group, molecular weight, and charge of a surfactant
may affect its rate of diffusion through the solvent to the interface (Manglik et al. 2001).
Perhaps the most critical fluid property is particle size. For piezoelectric printers – and definitely
for the DMP2831 – a good rule of thumb for maximum particle size is 1% of the nozzle diameter
(Fujifilm-Dimatix 2006a, 2006b). With particles larger than this size, clogging of the nozzles
will inevitably occur, requiring replacement of the printhead and process downtime. It is
theoretically possible to print larger particles with larger-diameter nozzles, but they will produce
larger droplets, which may be undesirable for high-resolution printing (Lee 2003). In general,
the smallest particles possible are ideal, to avoid any issues with clogging of nozzles, which has
spurred the development of inks containing such miniscule particles as quantum dots (QDs) and
metal nanoparticles (Tekin et al. 2004, 2008; de Gans et al. 2004).
Specific gravity (SG) of an ink has less bearing on jetting performance and more on the
maintenance of good printer function. Most printers contain a pump or similar means by which
the meniscus of the ink while in the nozzle can be controlled (Molesa 2006). Meniscus refers to
the curvature of the ink layer in the nozzle before jetting – a high meniscus pressure provides a
convex bulge of ink out of the nozzle, and a low meniscus pressure draws the ink further back
into the reservoir. Most printheads are designed with an SG of ~1 in mind, as most common
solvents – and particularly water – have SG values of approximately 1. So, for the meniscus
pump to function properly on an inkjet printer, preventing both the backflow of ink into the
reservoir and the outflow of ink onto the nozzle plate, SG should be ~1, or slightly greater
(Fujifilm-Dimatix 2006a, 2006b; Molesa 2006).
31
Zeta-potential () is an expression of the magnitude of repulsive or attractive forces acting
between individual particles, and generally increases in absolute value as the thickness of the
charged layer around the particles increases (Davison & Lane 2003). A larger absolute value
(negative values applying to negatively charged systems) correlates with better dispersion
stability. At || of around 12-14 mV, immediate flocculation no longer occurs and a dispersion
begins to become stable. However, a dispersion is not considered indefinitely stable until || >
40 mV (Spasic & Hsu 2006). These values are generally only applied to suspensions of roughly
spherical particles; particles with different shapes and hydrodynamic radii will be stably
dispersed at different values of . A detailed discussion of the theory behind is not necessary
for this study; let it suffice to say that provides a good indication of the stability of an inkjet ink
containing dispersed particles. Flocculation of dispersed particles may lead to irreversible
damage to printheads through clogging. It may also reduce the amount of functional material
being delivered to the substrate, as some of it is settled out of the ink before jetting, and removed
during pre-printing filtration stages. In some cases, it may even be a result of a chemical reaction
in which the functional material agglomerates or grain size grows. An example of this was
observed by Small et al. (2010) when attempting to print ZnS:Cu dispersed in water with an
organic acid containing a sulfide group, resulting in the formation of CuS, a water-insoluble
compound which rapidly precipitated out of solution. In any of these cases, jetting performance
is compromised, and functionality after deposition may also be compromised. Therefore,
measurement of and maintenance of jetting conditions (temperature, pH) at a value that keep
in the stable range is important for replicable, high-quality inkjet deposition.
It is important to note that is generally a term applied to aqueous (or, more generally, polar)
suspensions, which comprise the vast majority of inks. The measurement and definition of in
hydrocarbons is less well-defined. The dielectric constants of hydrocarbons are so low as to
prevent any ionization of solvated species, which in turn prevents the development of
electrostatically charged layers around dispersed particles (McGown & Parfitt 1967). In the
absence of charged layers, adsorbed molecules with hydrocarbon-soluble tails provide dispersion
and sterically prevent agglomeration (van der Waarden 1950). Weak van der Waals/London
forces also play a role in maintaining dispersion stability (Hamaker 1937).
32
There are three more ink properties which are often calculated to determine whether or not an ink
is suitable for jetting. These have been defined as the dimensionless quantities We, En, and Oh,
which refer to the Weber number, Energy number (Meixner et al. 2008) and the Ohnesorge
number (Ohnesorge 1936, Derby & Reis 2003, Jang et al. 2009), respectively. The Weber
number establishes the likelihood of liquid splashing on the substrate; We expresses the ratio of
inertial to surface tension forces (Bergeron et al. 2000). This is very similar to the Energy
number, which has been specifically derived for inkjet inks. En is defined as the ratio of the
kinetic energy of a falling ink droplet to its surface energy, i.e.
where is the ink density, is the ink’s surface tension, and r and v are the droplet’s radius and
downwards velocity, respectively. En also determines the likelihood of splashing of a droplet on
the substrate, which may result in reduced printed resolution. A drop may split into two (or
more) droplets upon impact if the gain in surface energy is lower than the kinetic energy of the
flying drop, so En should be <1 to prevent this from happening – although droplet splitting may
not always occur, if a sufficient amount of energy is dissipated thermally (Cibis & Krueger
2005).
The Ohnesorge number is in turn defined as a relationship between capillary and viscous forces
acting on a droplet (Ohnesorge 1936). It can be expressed as the ratio of the Reynold’s number,
which expresses the ratio of inertial to viscous forces, and the root of the Weber number. Oh –
or, more commonly, the inverse of Oh, often denoted by Z or Z-1
– is a means of expressing
droplet stability for a fluid of given properties being ejected from a nozzle of a certain size by the
following expression:
𝐸𝑛 =
12𝑚𝑣2
4𝜎𝜋𝑟2 =
23𝜌𝜋𝑟3𝑣2
4𝜎𝜋𝑟2 =
𝜌𝑟𝑣2
6𝜎
33
where µ is the ink viscosity, and d is the diameter of the nozzle. This dependence on nozzle size
implies that the same ink may not jet from every different printer or printhead type, and ink
reformulation may be necessary for transferring inks between inkjet platforms. What is meant by
droplet stability has been previously outlined in Section 2.2.2 – the establishment of a spherical
droplet with no pronounced satellite droplets or a very long tail. Meixner et al. (2008), building
on the work of Derby & Reis (2003), suggested that an ink/nozzle combination with an inverse
Ohnesorge number with a value between 1 and 10 would be suitable for the formation of stable
droplets. Jang et al. (2004), as later outlined by Zhouping et al. (2010), more rigorously
specified that values of Z-1
in the range of 2 to 4 resulted in droplets where the tail had
sufficiently high velocity to recombine with the droplet quickly, whereas when 6 < Z-1
< 13, the
tail detached from the droplet, forming a secondary drop. However, in this latter case, the tail
and its resultant droplet still retained high enough velocity to recombine with the larger droplet
before impact. Values of Z-1
> 14 represented printing systems where ink droplets fell at a high
velocity, leaving behind tails which dispersed into several satellite droplets and splattered
irregularly on the substrate. The greater degree of viscous dissipation of energy (and thereby,
velocity) in inks with lower values of Z-1
allows the droplet tails to recombine with droplets,
preserving print quality. Therefore, a lower value of Z-1
– which often implies a higher viscosity
– is ideal for printing.
2.3.3 Orthogonal solvent systems
A final consideration to make during ink formulation is of the ink’s compatibility with adjacent
layers, if a multilayer structure is to be deposited. Each solvent must preferably undergo
minimal interaction with the layer below it during jetting. What this usually implies is that each
layer must alternate its solubility; a water-soluble material, for example, cannot be overprinted
𝑍−1 = 𝑅𝑒
𝑊𝑒 12
= 𝜌𝑣𝑑𝜇
𝜌𝑣2𝑑𝜎
12
= 𝜌𝜎𝑑
𝜇
34
with a water-based ink without some damage to the underlying layer and resulting compromise
of layer integrity. This often requires selection of dispersants for functional materials which can
solubilize the materials in different solvents, and may even preclude the use of certain materials
which cannot be made suitably soluble. Where alternating aqueous/organic layers are not
feasible, the issue of interlayer dissolution may be also be alleviated by selecting materials which
are minimally soluble (e.g. a hexane-borne layer overprinted with an alcohol-borne layer), or by
post-treatment of deposited layers, such as sintering or cross-linking, to reduce solubility.
2.4 Print quality
The concept of print quality encompasses many different characterizations of printed media, such
as sharpness, raggedness, optical density, print mottle, and so forth, as are described by Oitennen
& Saarelma (1998). These characterizations are applied to conventional printed media, typically
on paper surfaces, where they can be summarized into two basic descriptors of a quality print –
resolution and colour intensity/trueness (Oitennen & Hannu 1998). With printed electronics, the
latter descriptor is not particularly relevant – printed layer functionality is a more valuable
measure of print quality. However, the former issue, of resolution, is of major concern when
manufacturing printed electronics, as the feature size will be limited to the maximum resolution
deliverable by the printer. An issue that arises with printed electronics, which is of no concern in
conventional printed matter, is the alignment of individual layers when overprinting a multilayer
device. Finally, although the roughness of printed media can have a bearing on their optical
properties (Daniel & Berg 2006), roughness is a matter of vital concern in printed electronics,
where functional layers are often only a handful of nanometres thick, or have a tendency to
catastrophically fail at localized thin or thick points.
2.4.1 Resolution
Resolution, in its most rigorous definition as applied to printing, refers to the number of distinct
“dots” or drops printed along a fixed length, such as dots per inch (DPI) or pixels per inch (PPI)
(Lee et al. 1981). Resolution may be observed by image analysis of printed films – optical
microscopy usually being sufficient, as drop sizes are generally in the micrometre range. For a
35
printed display, pixel count may be an important measure of quality, but for other printed
electronics, the only meaning of resolution that is significant is that of feature size. Many
conventionally-processed electronic devices have feature sizes of 50 nm or smaller (Mahalik
2006), allowing the packing of many transistors, for example, onto a single silicon wafer.
It is unlikely that a DOD printing process will be able to match this miniscule feature size in the
foreseeable future. The main reason behind this is that the minimum feature size producible by
an inkjet printer is limited to the size of the impacted droplet that it can deposit. Moreover,
unless the desired feature is circular, the minimum printable feature size must be larger than the
droplet size, utilizing multiple droplets to provide lines, corners, and so forth. The impacted
droplet size will be larger than (but related to) the nozzle diameter (Tekin et al. 2008), so
narrower nozzles can deliver smaller droplets and therefore smaller feature sizes. The narrowing
of nozzles causes changes in jetting behaviour, and fluid properties, including particle size, must
be accordingly adjusted to produce stable drops and avoid clogging. Realistically, for printing
suspended nanoparticles such as QDs, which have minimum particle sizes around 1 nm (Haider
et al. 2009, Borovitskaya & Shur 2002), the minimum nozzle diameter would be 100 nm
(Fujifilm-Dimatix 2006a, 2006b). If a droplet formed with diameter 100 nm, it would produce a
circular film of ~200 nm (Tekin et al. 2008), still significantly larger than photolithography-
produced features – not to mention any issues arising with jetting, as indicated by the very small
value of Z-1
associated with such a small orifice diameter. Factors like surface energy of the
substrate would also have an effect on spreading, and the degree of spreading would have a
bearing on resolution in the vertical dimension as well as in the horizontal plane.
Therefore, in its current incarnation, inkjet printing is not particularly well-suited to the
production of extremely high-resolution electronics on the order of those currently produced by
conventional means. However, by delivering droplets that are a handful of micrometres in
diameter, reasonably well-resolved rudimentary electronics may be produced. Furthermore, the
high throughput and limitless substrate size associated with R2R manufacturing somewhat
alleviate the need for extremely small, tightly-packed arrays of pixels, transistors, and so forth.
36
2.4.2 Roughness & topography
Conventional processing methods for electronics can produce extremely smooth films – almost
atomically smooth in some cases, such as with atomic layer deposition, or ALD (Jaeger 2002).
There is good reason for this: electrical properties can be affected by local variations in
thickness, when films are used in electronic devices (Campbell 2001). Key properties such as
resistance, mobility, and dielectric breakdown strength can all be changed by changing the
geometry of a layer. Moreover, localized regions of lower thickness may lead to electrical arcing
or shorting in devices driven by an elevated voltage (Ono 1995). Devices containing films of
rough or irregular topography may therefore either fail immediately, or not function at all. Films
with localized fluctuations in thickness may allow current to channel through the thin regions, or
direct current flow to thicker regions, tunnelling through adjacent films (Wood 2000). As a rule
of thumb, the films should be as smooth as possible to prevent such issues from causing a device
to lose functionality (Haverinen et al. 2009, 2010). As significant a problem as surface
roughness is the presence of pinholes. Any holes in a printed layer present a localized channel
across which the driving current may flow without actually driving the device, or where an arc or
short may cause catastrophic failure (Ono 1995). Therefore, there must be some degree of
overlap between printed drops to prevent the presence of holes. Too much overlap, however,
and the thickness of the overlapping regions will increase dramatically (Haverinen 2010).
Producing smooth, uniform films is relatively simple when using methods like ALD, CVD,
PVD, epitaxy/MBE, and so forth, because the target substrate is exposed to a beam or cloud of
molecules or atoms that deposit all over the surface at the same rate, ideally (Campbell 2001,
Seshan 2002). Film quality and reproducibility are best when using slow (Å/s) deposition rates
with high-purity materials, rather than introducing relatively thick films of materials containing
solvents and surfactants. However, the inkjet printer, by definition, is a solution-processing tool.
With an inkjet printer, the film is formed of individual droplets deposited one at a time from each
nozzle, producing overlapping discs on the substrate (Figure 2.9). In an ideal situation, each disc
merges with its neighbours, forming continuous lines, which in turn merge with adjacent lines.
However, depending on fluid properties, substrate surface properties, drying rates, ink
37
formulation, and so forth, the same issues with capillary/Marangoni flow and “peak and valley”
formation may manifest themselves in the printed films (Figure 2.9).
Different line morphologies, including “stacked coins”, scalloped, and “bulging”, may appear
(Soltmann & Subramanian 2008), as well as the ridged structure caused by coffee-ring
formation. For a given substrate, the different morphologies are dependent primarily upon drop
spacing and drying rate. Some of these morphologies are schematically shown in Figure 2.8.
Figure 2.8. Schematic representation of “peak and valley” topography formed during printing.
When qualifying an ink’s performance, printed film morphology may be the most important
determinant in whether or not it can be used for a functional layer. Therefore, establishing
proper drop spacing and drying conditions is very important in ensuring layer (and device)
function. If smooth layers are not formed under any variations of these conditions, then the ink
may have to be completely reformulated. An example of drop spacing testing is shown in Figure
2.10, where an aqueous ink containing QDs and a polymer was printed onto glass and observed
with an optical microsope. Drops did not begin to coalesce into continuous lines until the drop
spacing was < 75 µm, and the adjacent lines did not coalesce until the drop spacing was < 65 µm.
Below this spacing, the lines overlapped excessively, leading to the “peaks and valleys”
topography shown in Figure 2.8. There is a narrow range of drop spacing values which will
allow smooth film formation, and these will be different for every ink and on every substrate or
underlying layer.
1) Printed lines, birds-eye
Solvent flow
Drop spacing
2) Wet printed ink, x-section 3) Dry printed ink, x-section
Solvent flow during drying
Spreading droplet
Peak and valley formation
Pri
nte
d l
ine
38
Figure 2.9. Schematic representations of common line morphologies (Soltmann & Subramanian 2008). “T” refers to
the substrate temperature, and thus the drying rate of the ink.
Figure 2.10. Drop spacing of QD/polymer/water ink on slide glass.
Thickness is another issue to consider during inkjet deposition. Film thickness is often a
determinant of exactly how a device functions (Dimitrijev 2005, Wood 2000). All of the layers
of a thin-film light-emitting device, for example, are generally of controlled thickness to
maximize performance while minimizing the amount of material deposited and the resistive
losses across the device (Adachi et al. 2007, 2008; Schrage et al. 2010, Manzoor et al. 2003,
Hieronymas 2002, Cho & Cha 2009). Increasing thickness may lead to poorer device
performance and higher power draw (Schrage et al. 2010).
The inkjet printer is capable of controlling thickness via ink formulation (wetting and spreading
of drops) and by depositing multiple layers, or “print passes”. Although the volume of ink
deposited by the inkjet printer can be easily observed from the size of individual droplets, the
actual film thickness after printing and drying is entirely dependent on the spreading of droplets
Normal Stacked coins Bulging Ridged Scalloped
Proper drop
spacing, T
Drops too close, or T
too high
Drops too far apart, or T too
low
Fluid proper-ties/T causing
coffee-ring
Drops too
far apart
100 µm
90 µm drop spacing 80 µm drop spacing 75 µm drop spacing 65 µm drop spacing
39
on the surface, the density of the dried film, the amount of solvent removed, and so on. The
volume of individual drops is usually significant enough that film thickness is on the order of a
few tens of nanometres, at least. Consider, for example, a 10 pL drop from a 25 µm nozzle,
producing a 50 µm circle on the substrate. Assuming that 90% of the drop is solvent, 1 pL of
solids remain after drying. This results in a thickness of ~500 nm. Functional layers in PVs,
LEDs, TFTs, and so forth are often significantly thinner than this (Mahalik 2006, Chen et al.
2011). A droplet that spreads wider to reduce thickness also reduces resolution, and may
encounter issues with coffee-ring formation. Coffee-ring or peak-and-valley formation presents
difficulties not only in increasing the potential for device failure, but also in estimating film
thickness, as it is not uniform. A droplet with reduced solids content will form a thinner film, but
may not contain enough functional material to operate as desired. So, as was the case with
balancing fluid properties, compromises must be made to produce high-quality films by inkjet.
40
Chapter 3
3 Materials for inkjet-printed electronics
Although theoretically any material that can be dispersed or dissolved into an ink can be used as
a functional species in an inkjet ink, the constraints on fluid properties and film-forming
behaviour limit materials selection. In particular, the necessarily small dispersed particle size
restricts available materials to colloids (Singh et al. 2010, Tekin et al. 2008, Zhouping et al.
2010, Caglar 2006), polymers – either dissolved or micellar (Tekin et al. 2004, de Gans et al.
2004, Meixner et al. 2008), and other nanostructures such as nanotubes (Mustonen et al. 2007) or
fullerenes (Hoth et al. 2007). There are, then, several materials which are particularly well-
suited to inkjet printing of thin-film electronics which will be described in this section. These
can be broadly divided into three classes – conductors, semiconductors, and insulators (Ohring
1991). There are other materials as well which serve auxiliary or supplementary purposes, such
as sealing films or barrier layers. Finally, a very important material to be considered as part of a
device is the substrate upon which it rests, which in the case of printed electronics, may be any
conventional rigid substrate, or a R2R-processable flexible substrate.
Dissolution of inorganic dielectrics, semiconductors and metals, is not usually possible at
atmospheric conditions with solvents suitable for printing. Conversion of a precursor to the
desired material during/after jetting can be accomplished, as with AgNO3-based inks producing
Ag films (Liu et al. 2005) or sol-gels for ceramics and oxides (Zhou et al. 2008, Lima et al. 2007,
Sharma et al. 2000, Atkinson et al. 1997, Harizanov et al. 2004, Sharma & Sarma 1998, Sharma
& Mansingh 1998), but the production of sintered or crystalline films requires heat or
hydrothermal treatment (Zeng et al. 1999). Metallic colloids often require high temperatures
after printing, for sintering, to demonstrate any electrical conductivity (Tekin et al. 2008). In the
case of polymers, inkjet deposition of dissolved material is possible, although loading of an ink
with more than a small amount of polymer can cause undesirable effects in drop formation,
especially the bead-on-a-string effect mentioned in the previous chapter (Mauthner 2008,
Magdassi 2010, Fujifilm-Dimatix 2006a). So, dispersed particles of either inorganics or
polymers which require no aggressive post-treatments offer the simplest deposition route.
41
In this work, materials which were functional upon deposition were used. What this means is
that the materials were dispersed in the ink with their desired properties already established – i.e.
no chemical, thermal, hydrothermal, or other post-treatments were to be used to induce
functionality in the printed films. Drying of the inks was of course necessary, but high-
temperature annealing or sintering were not used. The reasoning behind this decision was that
the inks were meant to be usable on a variety of substrates, with minimal unit operations (ideally,
just the inkjet printer) involved in their processing, under atmospheric conditions.
3.1 Conductors
Electronic devices, by definition, utilize the passage of electrons to provide a desired function.
In order to convey electrons to and through a device, an electrode material which conducts them
well, with low losses in voltage or current due to resistance, is required. In virtually all cases, a
set of two electrodes or interconnects is used to provide a complete circuit for electrons flowing
to and from a power source. These electrodes/interconnects may be composed of the same
material, or, in some cases, two different materials, depending on the electrical demands of the
device in question. There are many different electrically conductive materials in existence, but
only some of them are readily available in a form suitable for inkjet printing. The three primary
groups of these materials are metallic colloids, carbon nanostructures, and conjugated polymers.
The former two groups are dispersible species, whereas conjugated polymers may be either
dissolved or dispersed.
Metallic colloids have become increasingly common as inkjet-printable materials, and include
such conventionally-used electrode materials as Ag (Dearden et al. 2005, Wu et al. 2007, van
Osch et al. 2008, Perelaer et al. 2008, Nguyen et al. 2007, Jeong et al. 2010, Kim & Kim 2010,
Jung et al. 2007, Ryu et al. 2005), Au (Jensen et al. 2011 Chow et al. 2009, Ko et al. 2007, Cui et
al. 2010, Lee et al. 2008b), Cu (Li et al. 2009, Hong 2000, Jang et al. 2010, Lee et al. 2008a), Ni
(Li et al. 2009), Mg (Aguey-Zinsou & Ares-Fernandez 2008), and Al (Meziani et al. 2009), as
well as conductive oxides, such as indium tin oxide (ITO) (Yarema et al. 2012, Gan et al. 2006,
Cho et al. 2006, 2009; Al-Dahoudi & Aegerter 2006) and antimony tin oxide (ATO) (Cho et al.
42
2009, Yang et al. 2007). In many of the works listed, conductivity comparable to bulk values
was achieved with proper post-treatment of the colloids. Usually, this post-treatment entails a
high-temperature curing phase to remove solvent and decompose the organic ligands on the
nanoparticles, which were used to stabilize the colloidal suspensions (Caruso 2004). The solvent
and encapsulating ligand or polymer (also referred to as the “cap” or “capping agent”) which
stabilized the particles in solution prevent sintering from occurring until they are removed. The
solvent should ideally evaporate or decompose at a relatively low temperature, as should the cap,
leading to the use of such materials as nitrocellulose (Nguyen et al. 2007), which decomposes
around 135°C (Selwitz 1988). After the removal of the cap, the particles are generally still non-
conductive, being physically separate and therefore preventing effective electron transfer (Tekin
et al. 2008, Mei et al. 2005, Zhouping et al. 2010). They are usually then sintered at high
temperature to induce grain growth and particle merging. The high surface energy to volume
ratio of these extremely small particles can cause a reduction in melting temperature upwards of
500°C (Huang et al. 2003, Hostetler et al. 1998) – however, this still implies relatively high
sintering temperatures for most metals and even higher for oxides, well above 200°C, the usual
cutoff point for flexible substrates. If temperature is not a concern – for example, on a glass
substrate – these colloidal metals and metal oxides make excellent candidates for inkjet-printed
conductive layers. Conductivity in these materials approaches bulk performance, and they are
largely chemically inert, although atmospheric oxidation may be a problem over time for Al and
Cu in particular. There are also several commercial inks available which contain Ag or Cu, and a
few others containing ITO or similar transparent conductive oxides (TCOs); this completely
eliminates the need for ink formulation and voltage waveform tailoring. However, these inks are
expensive – sometimes upwards of $100 US/mL – and even in-house synthesized nanoparticles
for custom-formulated inks require expensive chemical precursors containing ionized metals.
So, colloidal metals and TCOs are acceptable for use in small amounts, as in interconnects, but
usually only on glass or comparable substrates. However, recent attempts have been made to
reduce the sintering temperatures of such materials, opening the door for their use on flexible
substrates. Some of the techniques used include the preparation of a metallorganic Ag precursor
rather than nanoparticles with a low reduction temperature of 125°C (Dearden et al. 2005); the
use of more weakly adsorbent ligand species to cap Ag nanoparticles (Perelaer et al. 2008); and
43
avoiding sintering entirely by encouraging close packing of negatively charged PAA-capped Ag
nanoparticles using cationic PDADMAC solution (Magdassi et al. 2010). Also, photonic or
electrical methods such as infrared exposure (Tobjork et al. 2012, Denneulin et al. 2011) or high-
voltage treatment (Allen et al. 2008) can also produce low-temperature-sintered films, suitable
for use on more thermally fragile substrates, at the cost of introducing more complex processing
to the deposition process.
Carbon nanostructures – including nanotubes, nanosheets (graphene) and nanoballs (fullerene) –
are another major category of printable electrode materials. Amorphous carbon is intrinsically
electrically conductive due to the delocalization of its valence electrons within covalent C-C -
bonds (Cutnell & Johnson 2008). Different allotropes of carbon may lead to different values of
conductivity: diamond, for example, having very low electrical conductivity, whereas graphitic
structures have moderate conductivity. Conductivity in carbon is often significantly lower than
in metals. Electrical conduction in carbon relies on electron transfer along the C-C bonds in a
single plane, whereas in many metals electrons are delocalized in every direction/plane within a
metallic crystal lattice, providing many adjacent conduction sites (Collins & Avouris 2000).
Within the aforementioned groups of nanostructures there is also wide variation in conductivity,
attributable to changes in the electronic wavevectors of the structure in question. If a gap exists
between the occupied and unoccupied energy states of the material, which may occur with a
certain size, bonding structures, and chirality, semiconducting behaviour will result; an
infinitesimally small gap will result in metallic behaviour (Dresselhaus et al. 2001). Observation
and control of the metallic-to-semiconducting ratio of carbon nanostructures has been reported
by groups with an interest in producing either highly conductive (Wang et al. 2008b, Blackburn
et al. 2008) or purely semiconducting nanomaterials (Naumov et al. 2009, Kanungo et al. 2009).
Each of these types may be useful in different electronic devices – the former being ideal for
electrode formation. The mechanical strength and flexibility of carbon nanostructures also
makes them well-suited to flexible substrates, increasing their appeal over more brittle
conductive species like ITO (Chen et al. 2002).
The most commonly inkjet-printed carbon species are carbon nanotubes (CNTs), both single-
(SWCNTs) and multi-walled (MWCNTs), fullerene, and graphene. Graphene is a monolayer-
44
thick sheet of carbon atoms bonded hexagonally by hybridized sp2 bonds (Geim & Novoselov
2007). SWCNTs are formed by “rolling up” a graphene sheet into a hollow tube; MWCNTs are
several concentric SWCNTs of increasing diameters (Dresselhaus et al. 2001).
Fullerenes are spherical carbon nanostructures of a “soccer-ball” shape, comprising both
hexagons and pentagons of C atoms – based on the original buckminsterfullerene (C60)
synthesized by Kroto et al. (1985). Schematics of these three structures are shown in Figure 3.1.
Figure 3.1. Carbon nanostructures: (a) SWCNT; (b) MWCNT; (c) graphene; (d) buckminsterfullerene.
Nanotubes and graphene are more widely used as printed electrode materials than fullerenes
because they can layer into stacked sheets or networks, whereas fullerenes must be packed
tightly (like metallic colloids) to allow electrons to hop between them. This behaviour was
(a) (b)
(c) (d)
45
observed in printed layers of MWCNTs (Fan et al. 2005) where overprinting of successive layers
was sufficient to improve conductivity greatly, as more MWCNTs filled in the network of
overlapping conductive paths. Several other groups also successfully printed MWCNTs (Wei et
al. 2007, Sumerel et al. 2006), SWCNTs (Song et al. 2009, Nobusa et al. 2011, Gracia-Espino et
al. 2010), and graphene (Te et al. 2011, Torrisi et al. 2012, Huang et al. 2011) as electrodes or
interconnects in different electronic devices, including thin-film transistors (TFTs) and field-
effect transistors (FETs). In most cases, the sheet resistance of CNT or graphene films was still
several k/square, 2-3 orders of magnitude higher than that of TCOs like ITO and many orders
higher than metals. However, carbon does present some unique advantages over these
conventional materials. Firstly, the extremely strong nanostructures allow for a degree of
flexibility in printed devices. Secondly, the processing of CNT or graphene dispersions does not
require high-temperature sintering like TCO or metal nanoparticle suspensions – only the solvent
has to be removed to allow the materials to conduct electricity, although the dispersant or
surfactant may also need to be decomposed to improve conductivity. Indeed, carbon will not
sinter under atmospheric conditions regardless of temperature (Saavatimskiy 2005). Surface
functionalization of nanotubes or graphene sheets with organic functional groups may achieve
dispersion in a variety of solvents without the need for surfactants, eliminating this problem.
Thirdly, the small, controllable size of carbon nanomaterials means that they can be functional as
conductors or semiconductors while retaining optical transparency, presenting a potential
replacement for ITO or ATO. Finally, while carbon nanostructures are hardly inexpensive due to
their relatively complex purification procedures (Dresselhaus et al. 2001), they are certainly less
expensive than colloidal TCOs, and many times less expensive than colloidal metals. So, CNTs
and graphene are ideal inkjet materials to produce conductive layers, although perhaps best-
suited to the replacement of TCOs.
A point worth addressing about CNTs and graphene sheets is their aspect ratio. These materials
are considered “nanomaterials” due to the fact that at least one of their dimensions is in the nm-
range. However, they may be quite large in the other dimensions – nanotubes are often several
µm long, as are graphene sheets, in both length and width. This does not present a problem for
their functionality, but it is of concern for inkjet printing, where particle sizes are restricted.
46
Also, their stability in solution is often poor (Kim & Ma 2011). Both of these issues should be
taken into consideration when including these materials in an inkjet ink.
The final significant class of jettable conductors is that of conductive polymers. Like metals,
there are many different types of conductive polymers that are suitable for jetting. Unlike metals
(and carbon), these are not always in the dispersed state, but may be dissolved into an ink. The
basis of electrical conduction in these materials is the same as that in carbon nanostructures –
delocalization of overlapping p-electrons along conjugated -C=C-C=C- (or similar) bonds within
the polymer’s “backbone” (Skotheim & Reynolds 2006). The conjugated structure may also
contain other atoms, such as N, O, or S, which also have p-orbitals. When oxidized or reduced
by doping or electrochemical treatment, the mobility of the delocalized orbitals is greatly
increased, resulting in their relocation to other energy states, partially emptying out an electronic
band and allowing electrical conduction along the backbone (Skotheim 1997). The disordered
nature of these polymers, when compared to more highly ordered inorganics, can lead to
structural irregularities within the polymer’s conjugated chain, comparatively reducing mobility
and conductivity (McGinness 1972). Therefore, these polymers are generally not nearly as
conductive as metals, TCOs, or even carbon, as carbon structures are materially homogeneous,
although improving synthesis and purification methods may help to close this gap.
Some of the best-known conductive polymers are polyaniline (PAni), polypyrrole (PPy),
polyacetylene (PAc), poly(p-phenylene vinylene) (PPV), and polythiophene (PT). These
polymers have been extensively modified, substituted, and combined to produce many different
conductive species. The two types which have arisen as the most commonly used, due to their
ease of processing and relatively high conductivities, are PAni and PT – and in particular, a
variant of PT known as poly(3,4-ethylenedioxythiophene), or PEDOT (Molesa 2006). The
structure of these polymers is shown in Figures 3.2 and 3.3. Processing of these types of
polymers is greatly simplified by their ability to be synthesized and dispersed as a micellar
suspension in a solvent by using a charged surfactant molecule, whereas other conductive
polymers must be dissolved (Skotheim 1997). This limits the use of these other polymer types,
since conductive polymers, and conjugated polymers in general, are difficult if not impossible to
dissolve in most solvents (Nguyen et al. 2001). Furthermore, as was mentioned above, suitably
47
high conductivity is only achieved in conductive polymers by doping (usually oxidative, i.e. p-
doping), and these doped polymers are generally nearly insoluble in most solvents (Nalwa 2000).
In the case of doped PAni (shown in Figure 3.2) a variety of surfactants may be used to disperse
nanoparticles prepared by emulsion polymerization methods (Eftekhari 2010), including an acid
species which induces solubility in water (Lee et al. 2005b), although the dispersions often have
short shelf lives (Li & Kaner 2006). PEDOT (shown in Figure 3.3) has been cleverly dispersed
in an aqueous suspension using a material which also functions as a dopant – the anionic
polymer poly(styrene sulfonate), or PSS- – a combination discovered at Bayer (Bayer AG,
Leverkusen, Germany), producing gelled particles of PEDOT/PSS- with an average particle size
ranging from 30 nm (Changneng et al. 2005) to an average of 100 nm (Lee et al. 2005a).
Figure 3.2. PAni structure, where y = 1-x. If y = 1, leucoemeraldine; y = 0.5, emeraldine; y = 0, pernigraniline.
Figure 3.3. PEDOT (left) and PSS- (right) structures.
48
Since the discovery of PEDOT:PSS, as it is now called, it has been extensively studied as an
electronic material due primarily to its ease of processing by wet methods, including inkjet
printing (Tekin et al. 2008, Yoshioka & Jabbour 2006, Ouyang et al. 2004). PEDOT:PSS being
an aqueous dispersion, solvent removal and film curing is a relatively simple matter, and
definitely feasible on flexible substrates. The conductivity of PEDOT:PSS is also relatively high
for a polymer, and has been reported as more than 500 S/cm in a cured PEDOT:PSS film
(Crispin et al. 2003, 2006), with Heraeus claiming conductivity up to 1000 S/cm in films of their
CLEVIOSTM
product, which contains PEDOT:PSS as the conductive species. Some other
advantages offered by PEDOT:PSS are its transparency in the oxidized state, making it an
appropriate material for devices requiring a transparent electrode (Ouyang et al. 2005), the
stability of both the dispersion and films (Crispin et al. 2003), its mechanical flexibility (Polasik
& Schmidt 2005), and its relatively low cost.
A final useful feature of PEDOT:PSS is that its conductivity can be improved by adding certain
polar solvents to the dispersion before printing, or post-treating films with certain oxidizing
species. Some of the solvents that have been found to improve conductivity are dimethyl
sulfoxide (DMSO) (Fehse et al. 2007, Dobbelin et al. 2007, Kim et al. 2002, Xue & Su 2005),
tetrahydrofuran (THF) (Kim et al. 2002), N-methyl pyrrolidone (NMP) (Jonsson et al. 2003), and
glycerol (Lia et al. 2003). Anionic surfactants (Fan et al. 2008) and certain salt solutions (Xia &
Ouyang 2009) have also been shown to improve conductivity. The mechanism of conductivity
enhancement has been described in several ways, including conformational change of the
PEDOT molecule (Ouyang et al. 2004, 2005) and washing away of non-conductive PSS- anions
(Hsiao et al. 2008). Post-treatment of printed films with formic acid was also observed to
improve conductivity by further oxidation of the polymer (Daniel & Fotheringham 2007).
Finally, combining two of the conductive material classes, the addition of a small proportion of
carbon nanotubes (CNTs) to the PEDOT:PSS suspension significantly increases conductivity in
the resulting films (Kymakis et al. 2007, Mustonen et al. 2007, Wang et al. 2008a, 2008b; Ham
et al. 2008, Moon et al. 2005, Bhandari et al. 2009), although the films were only printed in the
work by Mustonen et al. For inkjet printing of electrodes, of all of the conductive polymers
available, PEDOT:PSS suspension is by far the best-suited due to its small dispersed particle size
49
and ready ability to mix with a variety of solvents and surfactants, many of which are even
beneficial to its conductive performance.
3.2 Semiconductors
Semiconductors usually form the key functional layers of an electronic device. Light-emitting
layers, light-collecting layers, diode junctions, and transistor stacks are all formed of different
semiconductors. The property which makes these materials suitable for use in all of these
applications is their small-to-moderate discrete electronic bandgap. The bandgap represents a
region of forbidden energy levels which cannot be occupied by electrons in that particular
material, lying between the unoccupied conduction band and the occupied valence band
(Neamen 2002). Upon some sort of energetic stimulus, electrons can be excited across the band
gap into higher energy states in the conduction band or higher energy bands. These electrons are
free to move in the conduction band, resulting in limited electrical conductivity through the
semiconductor. Electrons that jump to the valence band leave behind an unoccupied space which
is usually referred to as an “electron hole” or simply a “hole” (Neamen 2002). When this
happens, an electron adjacent to a hole in the valence band will fill the hole, leaving a hole,
which will then be filled, and so on – so an apparent flow of holes in the valence band and actual
flow of electrons in the conduction band occurs. The excited electrons can relax to a lower
energy state by jumping back across the bandgap and recombining with holes, losing energy
which is released as either heat or light (Sedra & Smith 1997). As is shown in Figure 3.4,
insulators also have a bandgap, albeit a wide one, meaning that a very large amount of energy is
required to excite an electron into the conduction band – rendering them effectively non-
conductive (Huang 2009). Metals, conversely, have overlapping bands, allowing facile transfer
of electrons and high conductivity. It is also worth noting that the terms conduction and valence
band are generally applied to inorganic semiconductors, whereas organic semiconductors’ bands
are characterized as the lowest unoccupied molecular orbital (LUMO) and highest occupied
molecular orbital (HOMO).
The bandgap (Eg) is an important property in all semiconductor-based microelectronics, where it
determines the amount of current or potential required to emit light (LEDs), the current needed to
50
modulate conductivity (TFTs), or the wavelength and intensity of light required to excite
electrons into the conduction band, where they can flow and be collected (PVs) (O’Mara et al.
1990). Semiconductors are defined as materials with moderate bandgaps (> 0 – 4 eV). Certain
materials are better-suited to each of these applications, with silicon (Eg = 1.11 eV) often being a
“workhorse” material due to its applicability to many types of microelectronics (O’Mara et al.
1990). However, when inkjet printing is the deposition method, rather than vapour deposition
techniques most commonly used for silicon processing (Jaeger 2002), other materials which are
solvent-dispersible or soluble (which silicon is not) must be considered. These materials can be
inorganic or organic semiconductors, and each type displays unique properties and useful
qualities.
Figure 3.4. Simplified electronic band structures.
3.2.1 Inorganics
Inorganic semiconductors are made up of compounds of different inorganic atoms, often
classified by the periodic group in which the element is located – e.g. GaAs would be a III-V
semiconductor. They may also include C, which is not technically an inorganic species, such as
in SiC. They may include any number of inorganic molecules in any proportion, in so-called
“solid solutions” of inorganic atoms (Berger 1996).
When considering inkjet printing of such materials, the likelihood of their synthesis and
dispersion as nanoparticles with size below the cutoff specified by the printer becomes remote as
Work function
Electron affinity
Bandgap
Fermi level
Vacuum level
Conduction band
Valence band
Ele
ctr
on
en
erg
y (
ev
)
Conductors Semiconductors Insulators
51
the complexity of the molecule increases. A controlled synthesis method for producing solution-
dispersible nanoparticles with good semiconductive performance is a difficult prospect even for
simple binary compounds. It is likely partially for this reason that certain materials have
attracted a great deal of attention as solution-processable inorganic semiconductors – most
notably, the chalcogenides of Zn, Cd, Pb, and Y. In particular, CdS and ZnS have been prepared
as nanoparticulate suspensions by controlled precipitation of nanocrystals (Chander 2005). The
crystalline nature of these materials means that they are theoretically functional upon wet
deposition, with no post-treatment, which was specified earlier as a key property of ink
components in this study. However, to allow charge transfer from nanocrystal to nanocrystal,
these materials need to be either sintered at high temperature into monolithic films, incorporated
into a conductive polymeric binder, or self-assembled into a packed film (Bakueva et al. 2003).
CdS and ZnS (as well as their selenides) are generally prepared using a competitive precipitation
process, in which soluble salts of the metal cation and chalcogen anion are mixed in the presence
of a solvent, a ligand, and sometimes, a dopant ion. Under controlled conditions of temperature
and pH, nuclei of the semiconductor material are formed and gradually grown until their growth
is arrested by double-layer repulsion, which prevents free ions still in solution from reacting on
the surface of the colloids (Chander 2005). The positively-charged counterion from the
chalcogen source, which does not participate in the reaction, forms a positively-charged
boundary layer which prevents initial agglomeration of the particles (Warad et al. 2005). The
presence of ligands which attach strongly to the colloids’ surface also assists in keeping them
dispersed and preventing Ostwald ripening or agglomeration (Capek 2006). In some cases, more
conventional colloidal chemistry, such as reverse micelle synthesis, has been used to prepare
these particles (Yang & Bredol 2008). The feature of these synthesis methods that makes them
so well-suited to preparing inkjet-printable materials is control over the ligand species and
solvent system. With the proper ligands, nanoparticles can be made soluble in a variety of
solvents and solvent mixtures, greatly facilitating ink formulation.
Of the species mentioned above, one of the most widely studied is ZnS – and specifically, doped
ZnS. Doped ZnS is a well-known photoluminescent and electroluminescent material (Yen 2004)
which is almost ubiquitous in ELDs and some LEDs. Doped ZnS can also be prepared in a
52
nanocrystalline form with a minimal number of reagents in a variety of solvents, including water.
Doped ZnS nanoparticles have been prepared in aqueous or alcoholic solutions by many groups
(Althues et al. 2006, Igarashi et al. 1997, Adachi et al. 2007, 2008; Konishi et al. 2001, Hwang et
al. 2005, Karar et al. 2004, Takahashi & Isobe 2005, Mu et al. 2005, Vogel et al. 2000, Manzoor
et al. 2003, Warad et al. 2005, Yang et al. 2003, Yu et al. 1996, Jindal & Verma 2008) using
acetate or chloride salts of the metallic precursors. Doping ions in these studies included Cu+,
Mn2+
, Al3+
, and several halogens. The reason ZnS is particularly notable is that its synthesis is
inherently facile – it requires only a Zn/dopant soluble salt, a suitable ligand or cap, and easy-to-
handle solvent system. This stands in contrast to the more complex methods of synthesizing
such colloidal semiconductors as CdSe/CdS/CdTe (Hines & Guyot-Sionnest 1996, Talapin et al.
2001) or PbS (Bakueva et al. 2003). These usually require elevated temperatures (and therefore
non-aqueous solvents), multiple injections of precursor materials, expensive or exotic salts and
ligands, and, of course, the use of biologically and environmentally harmful Cd and Pb. The use
of non-aqueous solvents precludes their inclusion in a water-based ink without some sort of
ligand exchange, as well. However, the aqueous synthesis of doped ZnS has been observed to
yield unpredictable doping levels (Peng et al. 2005), variable ratios of Zn:S (Althues et al. 2006),
and clustering of dopant atoms at the surface of the nanoparticles (Yu et al. 1996, Bulanyi et al.
2002, Bulanyi et al. 1998), if temperature and pH are not controlled similarly to the Cd-based
synthesis. So, preparing inkjet-ready materials is a process which must be carefully adapted to
each ink type and particular application.
Another consideration when preparing or choosing jettable inorganic semiconductors is their
greatly increased surface area to volume ratio, due to small particle sizes. In certain processes,
such as light emission or absorption, the splitting or recombination of charge carrier pairs
determines device function. So-called surface states – different or “bent” electronic band
structures at the semiconductor’s surface, resulting from the transition between the crystal lattice
and the vacuum/atmosphere – tend to trap charge carriers with energy barriers (Shik 1997). The
larger relative surface area in nanomaterials means that surface states are abundant. If charge
carriers are trapped, they may not recombine (emitting light) or move (providing current), or may
recombine inside another material, away from a light emitting centre (Vilms & Spicer 1965).
53
For this reason, another surface layer, called a “shell” or “cap” is often applied to Cd- and Zn-
based nanoparticles, in a process commonly referred to as “passivation”. This secondary surface
layer prevents the electronic band structure at the underlying particle (usually called the “core”)
surface from changing and forming surface trap states (Bhargava et al. 1994). In order to keep
the charge carriers confined to the core material and prevent them from travelling or recombining
in the shell or on the shell’s surface, a wide bandgap material is often chosen for the shell, such
as ZnS (Eg = 3.6) for CdS (Eg = 2.4) (Yang & Holloway 2004). This wide bandgap material
presents an energy barrier which prevents carriers from leaving the core. Shelling can present a
problem in materials with already wide bandgaps, such as ZnS. However, shelling with an
insulator like ZnO or SiO2 has been successfully attempted by several groups, resulting in
improved photoluminescent emission from ZnS:Mn nanoparticles (Karar et al. 2004a, Jiang et al.
2009, Haranath et al. 2005). The insulating nature of many polymers also makes them useful for
capping nanoparticles, such as poly(acrylic acid)-capped ZnS:Mn (Igarashi et al. 1997, Konishi
et al. 2001, Althues et al. 2006).
3.2.1.1 Quantum behaviour
A unique feature of semiconducting materials (and technically, all nanomaterials) is observed
when one or more of their physical dimensions become smaller than some characteristic length –
such as the de Broglie wavelength, the electron mean-free-path, the Bohr exciton radius, and so
on (Shik 1997). Carriers become confined within potential wells with physical sizes smaller than
these characteristic lengths in either one dimension (i.e. a thin film), two dimensions (i.e. a
nanoribbon or nanowire), or all three dimensions (i.e. a nanoparticle, or more commonly,
quantum dot or QD) (Kelly 1995, Harrison 2000). When so confined, the wavefunctions of
carriers in the confined dimensions are no longer represented by continuous probability density
functions, but rather by discrete energy states occurring at integer multiples of the
wavefunctions’ wavelength, much like states within a single atom (Kelly 1995). Therefore, as
the confined dimensions change in size, and the wavefunctions’ wavelengths also change, so do
the discrete energy levels: this alters the electronic band structure of the material (Gaponenko
1998). In other words, by using nanosized semiconductors, bandgap and electron affinity can be
54
altered by adjusting one or more physical dimensions below a critical value. This is referred to
as “energy quantization” or simply “quantization”; materials below whatever characteristic
length is required to achieve quantization are called “quantized”.
In the case of the nanoparticles typically incorporated into inkjet inks, the radius of the
nanoparticles determines whether or not they are quantized. The critical radius below which
strong quantization occurs is referred to as the exciton Bohr radius, or ab
* (Harrison 2000). If the
particle radius, R, is smaller than ab
*, the bandgap of the material will change according to the
equation given below (Sun 2005, Caruso 2004):
where Eg is the electronic bandgap of the material, and h is Planck’s constant. mh* and me
* refer
to the effective masses of electrons and holes in a given material – for example, in ZnS, mh* is
0.61me, and me* is 0.40me (Vogel et al. 2000), where me is the effective mass of a charge carrier
(9.11 10-31
kg) (Berger 1996). As the radius shrinks, the bandgap of the material widens.
Because the emission of light, for example, occurs when an excited electron from the conduction
band recombines with a hole in the valence band, losing energy Eg in the process, the
magnitude of Eg will determine the wavelength of the light emitted () (Neamen 2002):
where c is the velocity of light travelling in a vacuum (3 108 m/s). Therefore, wider-bandgap
(smaller diameter) nanomaterials will emit light upon electrical or photonic excitation at a higher
energy, and a lower wavelength. As has been demonstrated extensively with Cd-based QDs, the
𝐸𝑔𝑛𝑎𝑛𝑜𝑠𝑡𝑟𝑢𝑐𝑡𝑢𝑟𝑒= 𝐸𝑔𝑏𝑢𝑙𝑘
+𝜋2ℎ2
2𝜇𝑅2
1
𝜇=
1
𝑚ℎ∗ +
1
𝑚𝑒∗
𝐸 =ℎ𝑐
55
full visible range of colours, as well as UV and IR, may be emitted by QDs of varying sizes
below ab*. This property may be useful for LEDs and PVs, but the small nanoparticle sizes
required – usually below 10 nm (Shik 1997) – are not necessary for inkjet printing. However,
minimization of particle size is helpful for preventing clogging and maintaining dispersion,
meaning that QDs are a good starting point for ink formulation.
3.2.2 Organics
The electrical properties of organic semiconductors are very similar to those of inorganics, with a
few subtle differences in terminology and carrier transport mechanisms. As was mentioned
previously, the conduction band and valence band do not exist in organics; instead, the
comparable values of LUMO and HOMO, respectively, refer to the edges of organics’ electrical
bandgaps. Also, charge transport within organics is not simply a matter of electrons moving
through a relatively empty conduction band; it is often modeled as being the result of electrons
hopping between different trap states and overcoming energy barriers to do so (Vissenberg &
Matters 1998). This implies behaviour unlike that in inorganics with respect to temperature and
input potential, which both assist electron hopping by providing energy (the opposite occurring
in inorganics).
Organic semiconductors present certain advantages over inorganic semiconductors, not the least
of which is the relative ease of incorporating polymers and small molecules into solution when
compared to inorganic nanoparticles. Of course, issues with solubility already raised in the
section on conductive polymers also apply to semiconductive polymers, which have similar
conjugated chemical structures, often including aromatic rings (Skotheim & Reynolds 2006).
Small molecules may be soluble in particular solvents; however, to allow charge transport
between molecules, a conductive polymer binder (or “host”) is still required, as with inorganic
nanoparticles. Also, the tight molecular packing of solid-phase small molecule semiconductors,
due to interactions between the delocalized charged regions characteristic of conjugated species,
can sometimes mean that they are surprisingly difficult to dissolve (Dimitrakolopous & Mascaro
2001). Functionalization of the molecules or polymers by adding side groups and chains can
assist in improving solubility, but the reduced packing caused by such treatment can then
56
compromise carrier mobility (Chang 2005). A less-packed, more-soluble layer is also more
prone to oxidation and moisture damage. Further functionalization to improve packing and
ordering upon deposition and heat-treatment may improve mobility somewhat, but the more
chemical pretreatment steps that are required, the “dirtier” the final product (Vissenberg &
Matters 1998). Ideally, a molecule or polymer must be built from the ground up, adding
functional groups to the initial structure that allow solubility and encourage packing upon
deposition and drying.
Common semiconductive polymers and molecules are typically built around aromatic rings or
their derivatives (thiophene, pyrrole, etc.) A few have already been listed in the section on
conductive polymers – most semiconductive polymers are similar to these, being substituted
polythiophenes or poly(phenylenevinylene)s, but others include polypyrroles and the commonly-
used poly(n-vinylcarbazole) or PVK (Yang et al. 2003). Some of the most common small
molecules are pentacene, copper phthalocyanine, tris(8-hydroxyquinolinato)aluminum (Alq3),
and fullerene – all of which include numerous aromatic rings, and can be substituted and
functionalized into many different variations (Mullen & Scherf 2006, Cheng et al. 2009).
In many cases, the relatively simple processing and infinite customizability of organic
semiconductors makes them a superior choice for electronics fabrication. However, inorganic
materials do present several advantages, the most important of which are their better
environmental stability (Ono 1995) and, in LEDs, narrower emission spectra (Xing et al. 2005).
In the following chapter, the intricacy of organic LEDs when compared to conventional
inorganic LEDs or ELDs will also be outlined. When constructing an all-printed device, the
number of different inks – all of which must be based in orthogonal solvents – directly correlates
to the difficulty of fabrication and the likelihood of device failure. Therefore, realistically, a 3-
layer device has a better chance of being successfully fabricated by inkjet than an 8-layer one,
which is a significant advantage for inorganic-based inks.
57
3.3 Insulators
Insulators are materials that allow almost no current to flow through them. Similarly to the
previous two classes of materials, there are both inorganic and organic insulators; and again,
inorganic insulators can be made inkjettable by their dispersion or synthesis as colloids, whereas
organics can be dissolved or dispersed. However, inorganic insulators have dielectric constants
that can be orders of magnitude larger than the best-insulating organic materials (Ono 1995,
Matsumodo 1997, Vetanen 1997). In many cases, devices do not require high-dielectric (“high-
k”) layers, and so the moderate dielectric constants of many readily soluble polymers are
sufficient. Most polymers which are not intrinsically conductive (i.e. conjugated) conduct
electricity very poorly, and function well as insulating layers – such materials as
poly(vinylidene) fluoride (PVDF) (Vetanen 1997, Jung et al. 2010), poly(methyl methacrylate)
(PMMA) (Jung et al. 2010), cyanoethylcellulose (Saad 1994), and polyimide (Park et al. 2005).
The dielectric constants for these materials are usually in the range of 5-11 (Reddish 1962).
These and many other polymers can be dissolved in common solvents and deposited using inkjet
printing, although loading/molecular weight may both have a bearing on jetting quality.
Figure 3.5. Film formation in inks containing inorganic nanoparticles.
Improved insulating performance can be achieved by using inorganic materials, with dielectric
constants sometimes in the thousands when crystalline. Crystallinity is a prerequisite for good
insulating behaviour (Jia et al. 2000), meaning that crystalline nanoparticles are needed.
Sintering Binder Self-assembly
Solvent evaporation
Grain merging
Interparticular forces
(>1000°C)
(<200°C)
(<200°C)
58
However, like inorganic semiconductors, inkjet-printed layers must be sintered at high
temperature to form a monolithic film, suspended in a binder, or self-assembled (Figure 3.5)
Inorganic insulators, usually being oxides, have extremely high sintering temperatures
(>1000°C) which are not suitable for R2R processing, particularly not on conventional
substrates. If a binder is used, that binder must also be relatively insulating to prevent localized
tunneling of current. Self-assembly requires monodisperse particles with controlled
intermolecular or interparticular forces on an extremely flat substrate (Aubry et al. 2008).
Another means of producing an inorganic insulating film is to use sol-gel chemistry. Sols are
monodisperse colloidal suspensions, prepared from precursor materials, which may be gelled by
the removal of solvent, and further cured and densified by the elimination of the liquid phase
(Brinker & Scherer 1990). The resulting solids are glassy and amorphous, requiring post-
treatment to become crystalline. Post-treatments can be as simple as high-temperature heating,
or may utilize techniques such as lower-temperature hydrothermal exposure (Xu et al. 2006) or
chemical post-treatment (Zeng et al. 1999, Matsuda et al. 2000). Insulating films of several
species have been prepared and sintered using sol-gel chemistry, particularly high-dielectrics like
barium titanate (BaTiO3 or simply BT), lead titanate (PbTiO3), lead-zirconate titanate (PZT), and
barium strontium titanate (BST) (Yakovlev 2004, Zhou et al. 2008, Lima et al. 2007, Sharma et
al. 2000, Atkinson et al. 1997, Harizanov et al. 2004, Sharma & Sarma 1998, Sharma &
Mansingh 1998). However, sintering temperatures in all of these films were well above 200°C,
making them unsuitable for many flexible substrates. Also, the relatively low solids content of
sols means that they shrink a great deal when dried (Brinker & Scherer 1990), causing cracks
(shown in Figure 3.6). These cracks would only be worsened by the bending of a flexible
substrate. Finally, exposure to a strong base or an alcohol over several hours for hydrothermal
sintering is a technical hassle, requiring a specialized chamber. In some cases, these materials
can potentially solvate or damage other layers and the substrate itself.
59
Figure 3.6. Microcracking in sol-gel-derived BaTiO3 spun films on glass. These films were produced in the laboratory
using sol-gels prepared similarly to those described by Sharma & Mansingh (1998).
Self-assembly is an interesting prospect for forming films of nanoparticles, since very small
nanoparticles can be densely packed into structures resembling monolithic solid films. A
fundamental requirement of self-assembled layers, however, is that they must be made up of
monodisperse particles. Although sol-gel chemistry can produce colloid-sized – or smaller –
particles of insulators, these are not crystalline, and so their dielectric constants are not much
better than polymers. Normally, ceramic particles are ground down from bulk materials, so their
minimum size is limited. Synthesis of sufficiently small/monodisperse insulating nanoparticles
for self-assembly is not yet well-researched, although it has recently been accomplished by Ould-
Ely et al. using a sol-gel/hydrothermal reaction setup (2011). Redispersion of these nanoparticles
with suitable ligands to allow self-assembly has not been reported, however; there is substantial
work to be done before self-assembled insulating films become a reality.
Therefore, inorganic insulating inks generally contain suspended ceramic nanoparticles of
insulators such as titanium dioxide (TiO2), BaTiO3, or barium strontium titanate (BaSrTiO3) with
an accompanying binder polymer (Kim & McKean 1998, Sakai et al. 2006, Ding et al. 2004,
Tseng et al. 2006). This means of producing moderately insulating films is robust, simple, and
applicable to many different substrates – although the usual challenges associated with ink
formulation and film morphology, of course, are present.
200 µm 20 µm
60
3.4 Encapsulants
Electronic devices sometimes include other, inert layers for physical protection. These layers
may be used to prevent moisture from infiltrating the device, oxygen (and other oxidants) from
reacting with device components, and physical/thermal/electrical shock (Ardebili & Pecht 2009).
Devices using organic components, which have many functional groups with which oxygen or
water can react, require particularly effective encapsulation. Conventional encapsulants are most
often two-part epoxies or laminated polymer films, which are rolled on to planar devices (Ono
1995). Jettable encapsulants are generally UV-cured thermoset resins with low permeability and
some degree of mechanical toughness, some of which are commercially available as inks.
Similar materials may also be used in the novel field of paper-based electronics, where the rough
and permeable surface of paper presents a problem for device deposition, function, and
environmental stability. The paper surface may be coated with a material like spin-on-glass
(Kim et al. 2010) or another impermeable resin to improve smoothness and reduce permeability.
3.5 Substrates
Technically, inkjet printers can deposit materials on any substrate, from the usual paper sheet to a
three-dimensional object to a textile. For microelectronic fabrication, the same substrates as are
commonly used in conventional electronics processing – glass and silicon – can be inkjet-
printed. In an R2R process, flexible substrates may also be printed, including paper, fabric, and
polymers. Each substrate presents unique advantages and limitations; as a rule of thumb, the
rigid substrates already used in electronics manufacturing support the best-functioning devices,
and flexible/permeable substrates currently lag behind them in terms of device functionality.
The traditional support for many types of electronics is silicon. Silicon is a useful substrate
because it can also be a functional layer – a TFT, for example, can be prepared by oxidizing the
surface of the Si in a defined area, producing an insulator (SiO2), and source and drain electrodes
applied with evaporated metal (Bucknall 2005). Silicon wafers may also be made almost
atomically smooth by controlled crystal growth and cleaving (O’Mara et al. 1990). Silicon can
also be easily doped with an electron donor (n-doped) or electron acceptor (p-doped) with high
61
spatial resolution (e.g. by ion implantation or epitaxy) to produce diodes and other p-n junction-
based devices (Jaeger 2002). For many types of electronics, and especially for integrated circuits
and TFTs, silicon wafers are the ideal substrate, providing the necessary electronic and physical
properties for maximal functionality. However, many of these advantages of silicon are nullified
when using atmospheric-condition inkjet printing as a fabrication technique. The first and most
important reason why silicon is not particularly useful for printed electronics is that printing is an
additive technique. Functional layers are deposited on an inert substrate, and the substrate itself
is not deliberately altered by the inks resting on it. A hybrid method of fabrication involving
wafer doping or oxidation followed by printing of interconnects, insulators, and so on is of
course feasible; however, the necessity to use both high-vacuum deposition techniques and
printing somewhat defeats the purpose of either deposition method. If a device is going to be
partially constructed at high vacuum by conventional means, a likely better-functioning device
could just be constructed using only conventional means; likewise, if a device is being printed to
increase throughput and facility, having vacuum stages nullifies both of these gains. In other
words, silicon has its place as a substrate in conventionally-processed electronics, but it is not a
sensible candidate for printed electronic substrates.
Glass, on the other hand, still retains its usefulness, even when used in printed electronics. Glass
is widely used in the fabrication of many types of device, especially those requiring light input or
emission (silicon being opaque). The smooth topography and high-temperature processability of
silicon are also present in electronics-grade glass. However, glass offers a further advantage for
printed electronics: it is largely chemically inert, meaning that almost any ink system can be
deposited on its surface. This chemical inertness also means that glass can be effectively
surface-treated to control wetting by any given ink (Menawat et al. 1984). Of course, the major
limitation of glass is that it is brittle and inflexible, like silicon, making it useless for R2R
applications. However, for processing of displays and solar arrays in a batch fashion, inkjet
printing onto glass has been established as a rapid means of production which may displace
conventional fabrication techniques (Caglar 2006, Singh et al. 2010, Kim & Han 2010). ITO-
coated glass is also almost ubiquitous with LEDs and PVs, presenting similar advantages and
limitations as bare glass, but of course, with an added conductive layer. The types of glass
62
available for electronics fabrication are usually borosilicates or soda-limes, depending on the
processing conditions (Chen et al. 2011).
The next logical step after glass in alternate substrates is towards polymers. Polymers offer
many of the advantages of glass, with relatively smooth surfaces, tunable surface chemistry, and
optical transparency. Like glass, ITO-coated versions of certain polymers – particularly
polyethylene (PET) – are available. Unlike glass, they offer mechanical flexibility, light weight,
low cost, and, as a result, R2R processability. All of these advantages come at one major cost:
reduced temperature tolerances. Not many transparent, inexpensive polymers exist with glass
transition temperatures (Tg) much greater than 200°C, and many have decomposition
temperatures even lower than that (Zhouping et al. 2010). Some high temperature-tolerant
polymers do exist, such as Mylar and silicone, but these are generally not transparent or
inexpensive. Furthermore, unlike glass, polymers are soluble in many different solvents and can
be chemically reactive. So, when using polymers, solvent and additive selection during ink
formulation is crucial to ensure that a device will function and not damage the substrate. Several
research groups and corporations have produced different types of flexible electronics,
sometimes using inkjet printing, on many different polymer types (Carter & Gardiner 2009).
Paper is an emerging substrate for electronics, well-regarded due to its renewable and
biodegradeable nature, flexibility, and relatively low cost. Certain film-based devices, such as
transistors, batteries, and ELDs, have been successfully prepared on paper (Tobjork &
Osterbacka 2011). Although offering these advantages, paper is a problematic substrate for
electronics. In terms of processability, it is similar to many polymers: most paper will scorch
below 200°C, and may burn only a few degrees above this temperature (FPL 1964), while being
lightweight, inexpensive, and R2R-processable. However, it is also opaque, rough, irregular, and
porous. The opacity may be overcome in PVs and LEDs by reversing the order of the layer stack
in the device. The surface properties, however, require paper sheets to be treated to minimize the
deleterious effects of their rough, absorbent surfaces. As ink penetrates into a paper sheet, non-
functional paper components become ingrained in the functional layers of the device, and fibres
swell and roughen (Xie et al. 2008) – both of these occurrences might cause a device to fail
completely. Also, physicochemical interactions, like adsorption, may occur between functional
63
materials and paper components (Montibon et al. 2009, 2010; Di Risio & Yan 2009). There are
certainly means of smoothening and sealing paper, usually with mechanical calendering, sizing,
or coating (Smook 1992). Despite the difficulty of using paper to support thin-film electronic
devices, it has presents certain unique features which make it particularly attractive. Paper is a
biodegradeable, disposable, renewable material upon which much of the existing infrastructure
for R2R printing is based. Therefore, using paper as a substrate for simple, rudimentary
electronics is economically and environmentally desirable. However, because the roughness of
most sheets exceeds the typical thickness of LED, PV, or TFT films, its use for producing such
devices is limited. Nevertheless, TFTs (Lim et al. 2009, Kim et al. 2004), primitive capacitors
(Pushparaj et al. 2007), and displays (Kim et al. 2010, Andersson et al. 2002) have been
produced on paper, albeit with few or no inkjet-printed layers.
64
Chapter 4
4 Light-emitting devices
Light-emitting devices come in many different forms and utilize a myriad of materials, many of
which were described in the previous chapter. Some of these devices were listed in Chapter 1.
Several display types – in particular, the LCD, CRT, and PDP – are not usually classified as
light-emitting displays, as they do not emit light solely from electrically exciting the
semiconducting species within the device, but rather by filtration of backlight, excitation of
phosphors with an electron beam, or with plasma, respectively. Although CRTs and PDPs are
technically “emissive”, with the phosphor (luminescent species) being excited by an energy
source and luminescing, the phosphors do not luminesce due to simple applied current. The
architecture of these displays also requires either a vacuum tube or small gas-filled cells,
meaning that thin-film processing – e.g. inkjet printing – cannot be applied to their manufacture
(Chen et al. 2011). The solid-state, planar display types, ELDs and LEDs, can be built from
films of materials, and utilize the phenomenon known as electroluminescence (EL) to produce
light. EL is the emission of light from certain materials when exposed to an electric field or
electrical current. Unlike thermoluminescence, bioluminescence, or chemiluminescence, EL is
entirely a result of electrical excitation of materials, wherein high-energy electrons fall to lower-
energy states, releasing energy as a characteristic wavelength of light (Matsumoto 1984). The
means of exciting the electrons varies depending on the device type and voltage drive type, and
therefore so does the device structure and component materials. It is for this reason that there
exist several different configurations of functional layers for devices which utilize EL.
For inkjet-printed display devices, stacked-film structures are ideal. The relatively simple quasi-
2D structures of ELDs and LEDs allow them to be much more easily fabricated using solution
processing, by overlaying successive layers of functional materials. Structurally, ELDs and
LEDs are similar, but ELDs are excited by accelerated electrons and LEDs are excited by current
flow. This distinction is particularly important depending on the materials used, as certain
excitation mechanisms are only available to some materials. The different structures and typical
materials are outlined in this chapter.
65
4.1 Light-emitting diodes (LEDs)
The LED is a well-established technology based on an extremely simple principle of operation.
A diode is a junction between a p-doped and an n-doped semiconductor (a p-n junction), with an
anode connected to the p-doped material, and a cathode connected to the n-doped material. As
direct current (DC) flows from the cathode into the n-doped side, high-energy electrons in the
conduction band (or LUMO) recombine with holes in the valence band (or HOMO) flowing
from the p-doped side, causing the electron energy to drop across the bandgap and releasing light
of a characteristic wavelength (Sedra & Smith 1997). In the simplest LEDs, a single layer which
transports both electrons and holes and emits light can be used. The use of both a hole-
transporting (p-doped) and electron-transporting (n-doped) layer, one of which is the light-
emitting material, improves device function and efficiency (Crone et al. 1998). However, a
range of additional layers may be included to provide optimal efficiency (see Figure 4.1),
including carrier-blocking layers to keep carriers confined to the emissive region (Adamovich et
al. 2003) and current-limiting layers to prevent device burnout (Ono 1995). To further improve
functionality, electrodes may be chosen which have specific carrier energy levels. The
difference between the energy level and the vacuum level (0 eV) is the electrode’s work
function. The anode, for example, which conveys holes, should have a work function large
enough to contain holes with energy close to the valence band/HOMO of the layer into which the
holes are being injected (He et al. 2004).
Figure 4.1. Schematic of an OLED (right) and its energy level diagram (left), showing movement of charge carriers.
e-
VDC source
cathode
emissive molecule film
transparent anode
electron-blocking layer
substrate
electron-transport layer hole-blocking layer
hole-injection layer
HOMO values
LUMO values
cathode work function
e- e
- e-
h+
h+
h+
e-
h+
blocking
blocking
En
erg
y (
ev
)
work
function
anode
66
The complexity of such devices can be significant. Using any fabrication method, this increases
processing time and cost. In particular, with inkjet printing, each additional layer entails
formulation and functional testing of another inkjet ink. More significantly, the need for an
orthogonal solvent system dictates that each layer must not solvate the layers below it; this can
be problematic for a seven-layer stack in which several conjugated polymers are used, as not
many of these are water-soluble. This problem was addressed by Haverinen (2010), who used
cross-linked layers to minimize interlayer dissolution in a seven-layer stack – however, in that
work, only one of the layers was inkjet-printed and another was spin-coated and cross-linked, the
rest being vacuum-deposited. This problem presents the greatest hurdle to the application of
multilayer LED technology by inkjet. Single-layer devices (which are in fact three layers,
including electrodes) require only one insoluble layer to be placed between the other layers,
greatly simplifying fabrication, at the cost of device functionality. So, depending on the
materials used and the facility of their incorporation into orthogonal inks, a suitable device
structure may be chosen. Some of the common structures are shown in Figure 4.2; note that the
they may not include all of the layers shown, but definitely must include electrodes and an
emissive species.
The classic LED structure is very similar to another structure which is particularly well-suited to
inkjet deposition: the quantum dot LED or QDLED. This structure does not necessarily have to
include QDs as the emissive species when inkjet-printed, as long as the particles are sufficiently
small to be printed. When using inorganic materials, the advantage of this structure over a
standard LED is that it does not require a monolithic semiconductive film, but rather
nanoparticles in a conductive binder. A self-assembled single film of QDs has been deposited in
a QDLED-type structure without a binder by Haverinen et al. (2009, 2010), Yang et al. (2003),
and Hieronymas (2002), but such structures suffer from issues with charge transfer between
QDs. The ligands (and sometimes, shells) of the nanoparticles are usually wide-bandgap
materials – often insulators – preventing charge carriers from moving between nanoparticles
without being provided with significantly more energy (i.e. potential). By incorporating the QDs
into a polymer with a higher LUMO value than the conduction band/LUMO of the ligand and
shell materials, it becomes energetically favourable for electrons to transfer to the QD to
67
recombine with holes (Coe-Sullivan et al. 2003, 2005). A schematic of this improved charge
transfer is shown in Figure 4.3, based on a similar diagram produced by Haverinen (2010).
Figure 4.2 Typical planar LED structures.
Figure 4.3. Improvement of carrier mobility by polymer embedding of QDs used in QDLEDs.
large energy barrier
QD
e-
h+
En
erg
y (
ev
)
sh
ell
lig
an
d
core
En
erg
y (
ev
)
e-
e-
h+
energetically favourable
e-
h+
ch
arg
e t
ran
sp
ort
lay
er
(CT
L)
Packed QD-only layer CTL-embedded QD layer
LUMO
HOMO
VB
CB
+ - + -
VDC source
cathode
emissive polymer film
transparent anode hole-transport layer
substrate
electron-transport layer
VDC source
cathode
emissive molecule/host film
transparent anode
hole-transport layer
substrate
electron-transport layer hole-blocking layer
hole-injection layer
PLED OLED
+ -
VDC source
substrate
cathode
transparent anode
LED
emitter
+
VDC source
substrate
cathode
transparent anode
QDs/conductive binder
QDLED
-
current-limiting layer
68
Layer thicknesses in conventional LEDs are on the order of a few hundred nanometres (Ohring
1991). Film thickness is often a determinant of exactly how a device functions (Dimitrijev 2005,
Wood 2000). All of the layers of any light-emitting device are generally of controlled thickness
to maximize performance while minimizing the amount of material deposited and the resistive
losses across the device (Adachi et al. 2007, 2008; Schrage et al. 2010, Manzoor et al. 2003,
Hieronymas 2002, Cho & Cha 2009). Increasing thickness may lead to poorer device
performance and higher drive voltage and current (Schrage et al. 2010). The topography of LED
films also plays a major role in their performance. Films with localized fluctuations in thickness
may allow current to channel through the thin regions, or direct current flow to thicker regions,
tunnelling through adjacent films (Wood 2000). As a rule of thumb, the films should be as
smooth as possible to prevent such issues from causing a device to lose functionality (Berger
1996, Haverinen et al. 2009, 2010).
LEDs are composed of a wide variety of materials, especially organic LEDs (PLEDs and
OLEDs), which can incorporate many of the conjugated polymers/small molecules discussed in
the previous chapter. Inorganic LEDs can contain any semiconductor with a suitable bandgap
for emitting the desired colour of light. In glass-based LEDs, the light emission occurs through
the “front” of the device, i.e. the glass layer. Therefore, the front electrode or anode must be
made of a transparent or translucent conductive film. The most widely used film for this purpose
is ITO, which has a relatively high conductivity of 2.5 103 – 5 10
3 S/cm (Phillips et al. 1995)
and a high work function of ~ 4.65 eV (Parker 1994, Bakasybramanian & Subrahmanyam 1991).
Other conductive oxides such as zinc-indium-oxide (ZIO) or antimony-tin-oxide (ATO) are also
used for this application, both of which have comparable resitivities. PEDOT:PSS is also
commonly used, although sometimes only as an overlying layer on ITO for hole transport, where
its smooth surface assists in hole injection (Yoshioka & Jabbour 2006). Lithium fluoride is often
employed in a similar role at the cathode, where it assists in electron injection (Nalwa 2003).
The cathode is traditionally composed of evaporated metal with high conductivity and low work
function (Ono 1995). The many other possible layers included in LEDs are composed of
specialized polymers, for the most part; some of the more common materials used for charge
transport and blocking layers are listed in Table 4.1.
69
4.2 Electroluminescent devices (ELDs)
The mechanism of inducing EL is different in what are classically referred to as ELDs than the
mechanism in diodes. Rather than allowing electrons and holes to recombine radiatively inside a
semiconductor by passing a current through a diode, ELDs rely on a strong electric field to
accelerate electrons through the entire device stack. Above a certain voltage, known as the
threshold voltage or Vth, electrons will tunnel through the device layers, gathering enough kinetic
energy to strike and excite luminescent dopant centres in the emissive layer, similarly to a CRT’s
electron gun but in the solid state (Ono 1995). This requires a high voltage; therefore, these
devices contain insulating rather than charge-transporting layers to prevent catastrophic dielectric
breakdown (Kitai 2008, Hart 1999). This also requires specialized materials: light emission
occurs at the dopant sites in the emissive layer, meaning that the emissive materials – or
phosphors – must be doped (Bredol & Dieckhoff 2010, Vlasenko & Popkov 1960). Typical
phosphors are chalcogenides doped with <5 at.% of impurity, often rare-earth metals, transition
metals, or halogens (Yen 2004, Yen 2006). A summary of ELDs is shown in Figure 4.4.
Figure 4.4. ELD structures.
+ - + -
cathode
insulator
phosphor/insulating binder
transparent anode
VAC source
substrate
ACPEL
VAC source
substrate
cathode
insulator
phosphor film
transparent anode insulator
ACTFEL
+ - + -
VDC source
substrate
cathode
transparent anode
phosphor/insulating binder
VDC source
substrate
cathode
electron barrier
phosphor film
transparent anode
current-limiting layer
DCPEL DCTFEL
70
Unlike LEDs, ELDs can be both AC- and DC-driven. This presents several logistical
advantages, but also means that the same materials rarely function in both types of device.
However, a unique advantage offered by AC drive is that the work function of the two electrodes
can be the same (Wang et al. 1996), offering the possibility of a completely transparent device.
The two common structures (four total, since each can be both AC- and DC-driven) are
described below.
A common and robust ELD type is the AC-driven powder ELD, or ACPEL. The term “powder”
refers to the inclusion of particulate phosphor material within an insulating binder, rather than a
solid film of phosphor material. The core of this device structure is an insulating polymer resin
incorporating phosphor grains, protected from dielectric breakdown by an insulating layer, all
sandwiched between two electrodes. The other AC-driven device, the AC thin-film ELD or
ACTFEL, contains no binders, but solid films of phosphors and insulators. The core of the
device is a micrometre-thick stack composed of a transparent insulating layer, followed by a
phosphor layer, and finally a second transparent insulating layer. The necessity for solid,
crystalline films of both phosphors and insulators again means that this structure is not well-
suited to solution processing. The DC-driven equivalents of the ACPEL and ACTFEL are
similar in structure; however, current is limited in these devices rather than voltage. In the
DCTFEL, for example, an electron barrier and current limiting layer – of ZnSe and MnO2
respectively – are typically coated between the phosphor layer and the cathode in order to
prevent catastrophic failure (Ono 1995). In the DCPEL, no insulating layer is included, and the
phosphor grains are considerably smaller (0.5 – 1 µm diameter) and more densely packed to
allow electron transfer between the individual particles. These devices generally only work with
Cu-doped phosphors, as the chalcogenides of Cu are still moderately conductive and form
conductive surface layers on the particles when exposed to a strong electric field (Ono 1995).
The magnitude of the applied field can be quite large – on the order of 107 V/m, requiring typical
drive voltages well over 100 V. However, the current required is comparatively low, and
ACPELs, for example, draw only a few mW of power. This high operating voltage has limited
the application of ELDs in many types of electronics, not only due to power supply issues, but
also because any layer faults will lead to rapid dielectric breakdown.
71
Doped semiconductors and high-k dielectrics therefore make up the majority of ELD materials.
For powder EL structures, the phosphor material can be suspended in a dielectric polymer resin
with a high dielectric constant (in the range of 10-15) to form a luminescent layer. Some suitable
resins are cyanoethylcellulose or fluorinated polymers such as PTFE, PVDF, or PVDF:TFE
(Matsumodo 1985). Electrode materials remain the same as those used for LEDs. A summary
of common ELD (and LED) materials is given in Table 4.1.
4.3 Suitability for printing
In general, the DC-driven LEDs are more widely reported when using nanoparticles, which are
required for inkjet printing. There are several reasons for this. The principle of DC drive is
based on the flow of current through the device, and the recombination of electrons and holes
within the emissive material. In the case of AC drive, emission occurs due to the acceleration of
high-energy or “hot” electrons within the electric field and their collision with dopant sites.
Brighter emission is achieved when recombination occurs within the bulk of an emissive
material, rather than at its surface in both cases, due to previously described surface states. To
counter this effect, all ACPEL phosphors are encapulsated with inorganic oxide shells (such as
Al2O3 or SiO2), and TFEL phosphor layers are sandwiched between ceramic layers (Bredol &
Dieckhoff 2010) – this is similar to nanoparticle shelling. In nanoparticles, which have an
extremely high specific surface area, the likelihood of dopant clusters resting on the surface of
the particles is proportionally higher. There are two major problems with encapsulation of
nanoparticles. In bulk ZnS, a charged group on the surface of the phosphor particle is not
necessary to maintain dispersity. This is not the case with ZnS quantum dots, which are almost
always covered with an organic cap or ligand – which is rarely electrically conductive – in order
to maintain dispersion. The electric field is therefore not strong enough around nanoparticles to
induce electroluminescence by injection of hot electrons (Adachi et al. 2007, 2008; Bredol &
Dieckhoff 2010). Even if the particles were dispersed without a ligand, there would still be
issues with encapsulation. Firstly, encapsulation of nanoparticles with ceramics is almost always
achieved using sol-gels (Bredol & Dieckhoff 2010). However, sol-gels require high-temperature
72
Table 4.1. ELD/LED materials and layer properties.
Layer Present in Typical thickness Materials Properties
Cathode All devices > 100 nm Metals (Al, Mg, Ca, Ag)
Carbon (AC-drive) Polymers (AC-drive)
High conductivity, low work function
Insulator AC-ELDs 500 nm (TFEL) > 5 µm (PEL)
Perovskites (BaTiO3, PbZrTiO3 and variants)
High
Phosphor ELDs, LED 1000 nm (TFEL)
grains > 20 µm (PEL) 50-200 nm (LED)
Doped: ZnS, Y2O3, CdS, Gd2O2S, ZnO
Undoped: GaAs, CdS Electroluminescent
Insulating binder
ACPEL > 20 µm Fluorinated polymers Cyanoethylcellulose
Moderate
Conductive binder
QDLED 50-200 nm Conjugated polymers Moderate
conductivity
Current-limiting layer
DCTFEL 500 nm MnO2 Resistive
Electron barrier DCTFEL 500 nm ZnSe Large bandgap
Emissive polymer
PLED OLED 100-200 nm MEH-PPV, P3HT, PVK,
polyfluorene Electroluminescent
Emissive molecule
OLED 50-100 nm Alq3, PVK containing
fluorescent dye Electroluminescent
Electron transport layer
LED PLED OLED
< 100 nm TPBi, Alq3, LiF High electron
mobility (n-doped)
Hole blocking layer
LED PLED OLED
50-100 nm BCP, PBD, Liq Poor hole mobility
Hole transport layer
LED PLED OLED
50-100 nm PVK, TPB, NPB High hole mobility
(p-doped)
Hole injection layer
LED PLED OLED
< 100 nm PEDOT, CuPc Smooth film-forming,
conductive
Anode All devices 15-50 nm ITO
PEDOT CNTs
Moderate conductivity, high
work function
treatment to achieve crystallinity, which leads to irreversible nanoparticle sintering, migration of
atoms like sulphur, and oxidation of metallic ions. Moreover, sol-gel-derived shells suffer from
extensive pore networks and cracking left behind by reactants removed during drying and the
73
escape of gases produced by the reaction (Bredol & Dieckhoff 2010). Closing the pores requires
high temperature, which leads again to the problems described above. DC-LEDs only require
that the nanoparticles be dispersible in the binder polymer matrix, to facilitate energy transfer
from the matrix to the particle, allowing current flow and resulting luminescence. Although AC
EL has been reported from ZnS:Mn nanoparticles in a typical ACPEL structure by Adachi et al.
(2007, 2008), the emission was very weak and required extremely high voltage to induce a
strong enough electric field for charge injection into the nanoparticles. Therefore, for a printed
device, which by definition uses nanoparticles, the LED structure must be used.
4.4 Characterization
Table 4.2. Characteristic properties of light-emitting devices.
Parameter Symbol Unit Significance
Luminance L cd/m2
Intensity of visible light emission from LED at a given voltage
Threshold voltage Vth V Voltage at which 1 cd/m2 is observed
Turn-on voltage V0 V Voltage at which diode behaviour is
observed
Power density Pin W/m2
Power required to light the device at a given voltage over unit area
Luminous efficiency lm/W L ouput per unit power, taking into
account the solid angle
Contrast ratio CR unitless Ratio of the device illuminance to
ambient illuminance
Lifetime t1/2 h Length of time “on” before a device
loses 50% of its initial L value
Light-emitting devices are characterized using several specific parameters. These parameters
define the quality of the visual output, as well as the overall efficiency and durability of the
device. There are generally typical values for these parameters which are highly dependent upon
the devices’ structures and materials. Table 4.2 provides a summary of these parameters and
74
their significance. All of these properties may depend on ambient conditions, such as
temperature and humidity, and device structure, such as layer thickness, so these values are
recorded when reporting on the characteristics of the device.
A primary distinction between devices is the matrix type. Some devices are active-matrix (AM),
some are passive matrix (PM), and some are simply lights or “lamps”, with no pixels assigned.
AM devices have individual circuits behind each luminescent pixel, which can alter colour and
brightness – the ubiquitous TFTs. PM devices simply operate by addressing one row and one
column at a time, lighting up the pixel where the two cross. Lamps do not have any rows or
columns, and the entire luminescent area is one large pixel. The PM architecture is utilized for
testing many light-emitting devices, in order to keep the connections to the power source
removed from the fragile layers of the device itself. An example of such a setup is shown
schematically and photographically in Figure 4.5.
Figure 4.5. PM LED array for testing: schematic (left) and setup (right).
anode
substrate
emissive layer stack
Cath
od
e
+ -
75
Chapter 5
5 Approach & method development
The background research into the technique of inkjet printing, as well as the materials and
electronic structures best suited to it, strongly suggested certain materials and structures as the
best candidates for printing an LED. A detailed study of each ink and device to be created was
considered necessary for the realization of this project. However, each material presented unique
opportunities for study; some of the materials, such as PEDOT:PSS suspension, for example,
were readily available, but had not been comprehensively studied when included in a printable
ink. Others, such as quantum dots, were not readily available, and so had to be studied from the
ground up, including methods of their synthesis. Thus, the concept of using different materials
and electronic devices as “models” for stages of a complete method of developing all-printed
electronics was envisioned. The major stages in this method were briefly outlined in Section 1.3.
In this chapter, a more comprehensive consideration of the first major contribution of the work is
undertaken, discussing the stages of ink development and the “models” used to comprehensively
study each of these. A description of the experimental methods used for each study is provided
in the following chapter. It must be noted that each of these stages was included in the
development of all of the inks; however, the inks which were not the “models” for that stage
followed the groundwork laid by the “model” material rather than beginning from scratch.
5.1 Materials selection model: semiconductor
The prevalence of conductive and insulating materials in dispersion, as was described in Chapter
3, meant that these materials were readily available for incorporation into inks. However,
semiconducting materials were not similarly available. Also, the requirement for materials with a
certain bandgap – either to emit or absorb light at a fixed wavelength, or operate at a certain
voltage to behave as a transistor – meant that in-house tailoring of such materials was preferable
to using commercial substitutes. Perhaps most importantly, direct control over particle size and
dispersibility could be exercised if semiconductors were prepared in the lab. Therefore, the use
76
of the semiconducting species as a model for materials selection/synthesis was a sensible
approach. This model study included the following component parts:
i. Development of a method by which liquid-dispersible semiconductor nanoparticles,
polymers, or molecules with a desired bandgap could be reliably prepared;
ii. Control of the particle size (where applicable) to <200 nm for clog-free jetting;
iii. Characterization of the relevant property or properties of the semiconductor, such as
mobility, PL intensity/spectrum, EL intensity/spectrum, quantum efficiency, etc.
iv. Adjustment of synthesis method or post-treatment to optimize the relevant properties;
v. Dispersion of the material in a liquid carrier and testing of its compatibility with other ink
components, at a desired loading in the ink.
5.2 Ink formulation model: conductor
Because many different dispersions of conductive materials exist, many of which were described
in Chapter 3, the conductive species served as a better model for ink formulation. Indeed, many
conductive species are already included in commercial inks of unknown formulations. So, the
steps required to produce such an ink from a ready-made material (like Ag nanoparticles or
SWCNTs) were studied using the conductive species as the model.
Ink formulation was an involved and iterative process, originally undertaken after the early
stages of the project, when PEDOT:PSS and ATO inks were being used to produce AC-driven
PELs on paper. It was noted that both reformulation of the PEDOT:PSS ink and the use of
different substrates, which PEDOT:PSS wet differently, had a major effect on device
performance. Therefore, the conductive ink became the first to undergo an in-depth study of its
formulation, and the effect of even minor components on functionality. The methods used in this
study would then be carried over to all of the other inks which were prepared. The major steps
involved in the ink formulation study are as follows:
77
i. Selection of a suitable solvent system; measurement and optimization of viscosity within
a target range (2-12 cP for the DMP2831);
ii. Addition of surfactant (or another modifier) to adjust surface tension to an appropriate
value for jetting – this may also change viscosity (Jansen et al. 2001), so reexamination of
viscosity is appropriate here;
iii. Treatment of ink with other necessary components, such as humectants, biocide, pH
buffer, binder polymer, etc. – re-examine viscosity and surface tension and return to Step
(i) if no longer suitable for jetting;
iv. Particle size examination and particle stability study (-potential) – again, if unsuitable,
return to Step (i);
v. Jetting: optimization of voltage waveform and drop formation;
vi. Print quality: examination of printed patterns and distribution of ink components, where
necessary.
5.3 Film formation model: insulator
The sole remaining ink to use as a model was the insulator ink. Therefore, this ink was used to
outline the process for optimizing film structure on the substrate, for functional layer deposition.
This role was particularly appropriate to this material because any flaws in an insulating film
might cause catastrophic failure of a printed device, unlike in the other films where they might
simply reduce functionality. Also, a key aspect of film formation – film thickness – was of
major concern in an insulating layer, which might prevent sufficient charge carrier transfer if too
thick (Ono 1995). The issues with drop spacing, line merging, coffee-ring effect, and so forth
that were described in Chapter 2 were studied using this ink as model, and expanded to be
carried out on the other printed layers, as overall device smoothness and uniformity was expected
to be paramount in determining whether or not a device would function. The steps in the film
formation study are outlined as follows:
78
i. Establishment of droplet volume, drop size, and drop topography upon impact with the
substrate – if unsuitable, reformulation would be needed;
ii. Using the drop size obtained in Step (i) as a starting point, establish ideal drop spacing to
produce uniform films with no holes and the smoothest possible topography;
iii. Establishment of exact film topography using profilometry, microscopy, etc.;
iv. Reformulation of ink to improve topography and film formation if necessary –
improvement of leveling, wetting, etc.;
v. Establishment of film thickness as a single layer, and when overprinted several times, and
optimization of print parameters to yield films of appropriate thickness for device
application
5.4 Multilayer device model: LED
As was mentioned in Section 5.1 above, the initial work conducted in this study was the
preparation of an LED on a paper substrate (as shown in Figure 5.1). This was a proof-of-
concept approach, to demonstrate that conventional electronic materials and technologies could
be adapted to an unusual substrate and processing conditions. However successful this study
was, it was still fraught with difficulties in terms of the deposition process. In that work, an
ACPEL device prepared by inkjet patterning of the electrodes and Meyer-rod coating of the other
two layers was constructed on paper. Both of these deposition techniques are continuous
processes that could be readily applied to roll-to-roll fabrication of paper-based ACPELs. They
also avoided the limitations and waste associated with the batch process of screen-printing,
which is more commonly used (Satoh et al. 2007, Kim et al. 2010). However, they also required
masking of the substrate, as coated layers cannot be patterned. They also required two separate
unit operations. Both of these issues made fabrication very tedious, and the ease of deposition
offered by inkjet printing – which had already been applied to the two electrodes in the ACPELs
– was an attractive prospect. Furthermore, the coatings, although viscous and containing large
particles, were still in the liquid-phase, and could be used as models for inkjet inks.
79
Figure 5.1. Printed-coated ACPEL on paper. Top: schematic of device. Bottom: Device under driving voltage; off,
on, and on (darkened room)
Therefore, a similar, inkjet-printed structure – either an ELD or an LED – was used as the final
“model” in the method described in this chapter. The ELD or LED would serve as a
demonstration of the potential of inkjet printing an entire electronic device. The utility and
visual appeal of such a device also made it ideal as a stepping stone for the development of all-
printed electronics. The approach taken to prepare and characterize this ultimate “model” is
outlined below:
i. Determination of suitable structures that utilize the materials/inks studied in the previous
stages of the project;
ii. Deposition/characterization of variants of these devices, using conventional materials
(ITO, vacuum-deposited metals) to determine the functionality of individual layers before
printing entire ELDs/LEDs.
PEDOT:PSS (printed)
BaTiO3 /PVDF (coated)
ZnS:Cu,Cl/PVDF (coated)
PEDOT:PSS (printed)
paper substrate
VAC source
EL emission
80
Chapter 6
6 Materials & methods
Production of planar electrolumenscence light-emitting devices using inkjet printing techniques
required the formulation, jetting, and functional testing of several materials’ ink on various
substrates. Each series of experiments was focused on a particular ink that served as a model
species for one of the stages in ink development, as was explained in the previous chapter. Each
ink underwent the basic steps of ink formulation, however; the model inks simply expanded upon
each stage. The experimental methods used were also somewhat different for each ink, as each
ink had a different intended function. Each of the experimental techniques will be outlined in
this chapter.
6.1 Materials selection
Materials selection was based upon the typical materials used in ELDs/LEDs (described in
Chapter 4) and refined by choosing materials that were also jettable (as discussed in Chapter 3).
A summary of the characteristics of these materials follows. All reagents were supplied by
Sigma-Aldrich Canada, except where otherwise noted.
Several other materials were considered for use with little success, either due to difficulty of
incorporation into an ink, unsuitable processing conditions, non-functionality upon jetting, or a
variety of other reasons. In the interest of brevity, these materials will not be discussed here in
any detail. Let it suffice to say that materials selection, like ink formulation, is something of an
iterative process, and that theoretically suitable materials may not be practically useful. Some of
these other materials included colloidal Ag (processing temperature too high), aqueous CdS QDs
(unstable in the presence of certain ink additives, especially alcohols), sol-gel-derived emitters
and insulators (non-crystalline), and TiO2 (poor insulating performance). TiO2’s relatively low
dielectric constant of 86 (Sears et al. 1982) meant that it was not sufficiently insulating to
perform well under the conditions of this study.
81
6.1.1 Electrodes
The electrodes were to be inkjet-processable and suitable for curing at a low temperature.
Therefore, the primary conductive species used was aqueous PEDOT:PSS suspension (Mw =
8073, 1.3 w/w% in H2O, PEDOT:PSS ratio = 1:2). This deep blue-black suspension contained
PEDOT particles, stabilized by PSS- sodium salt, swollen with water. The dark colour of the
PEDOT:PSS was also ideal for establishing ink penetration depth on porous substrates.
As was mentioned in Section 3.1, carbon nanostructures have been incorporated into
PEDOT:PSS layers to improve conductivity. In this work, SWCNTs, MWCNTs, and C60 were
all mixed with PEDOT:PSS for this purpose. SWCNTs (1.2-5 nm diameter, 2.5 µm length),
MWCNTs (7-15 nm diameter per bundle, 0.5-10 µm length), C60, and shortened single-walled
nanotubes (S-SWCNTs) were used. The ratio of semiconductive CNTs to metallic CNTs was
2:1, as specified by the supplier. The S-SWCNTs were prepared from the as-supplied SWCNTs
using a chemical cleaving method described in by Chen et al. (2006). All of the carbon species
(SWCNTs, MWCNTs, S-SWCNTs, and C60) were added to water at 0.04 w/w% and stabilized
with 0.5 w/w% sodium lauryl sulfate (SLS) as suggested by Kymakis et al. (2007). Any
undispersed agglomerates were removed by centrifugation. These dilute carbon solutions were
mixed into the electrode inks as needed.
In some cases, ITO was used as the anode. Ready-made ITO glass slides were used, of two
varieties: the first being an unpatterned 1” 1” slide (Rsheet = 70-100 /), and the second being
a patterned 2” 2” slide (Rsheet = 15 /, Kintech Company). The patterned slides had 38 strips
of ITO, in two columns of 19 each, with dimensions of 1 mm 22 mm per strip, and 1.5 mm
strip spacing. ITO PET was also used for some tests (Rsheet = 100 /). Finally, when needed,
evaporated Al was used as a cathode material.
6.1.2 Insulators
The high-k dielectric BaTiO3, being a widely-used insulating layer in ELDs, was utilized. The
existence of a ready-made BaTiO3 nanopowder (≤ 25 nm particle size) made this a relatively
simple material selection to make. Other insulating materials included as binders were methyl
82
methacrylate (MMA), which could be polymerized in-situ into PMMA, and poly(vinyl
pyrrolidone) (PVP), with a molecular weight of 1 300 000.
6.1.3 Charge-transporters
Both PEDOT:PSS and PVK were used as charge-transporting species. PVK was used in two
forms: from Sigma-Aldrich (Mw = 25 000 – 50 000) and synthesized in-house by Brett Kamino,
a fellow student, using a living polymerization method described by Higashimura et al. (1980).
6.1.4 Emitters
The emissive layer was the “model” for materials selection, where a desired functional material
had to be tailored to inkjet application by controlling its particle size and solubility/dispersibility.
This implied the use of dispersible, nanosized, dispersed semiconductor particles, or a dissolved
semiconductive polymer/molecule. Because of its successful application in the printed/coated
paper-based ACPEL described in the previous chapter, doped ZnS was the material primarily
considered for inkjet deposition. As has been mentioned in Section 3.2, doped ZnS has been
widely prepared in a nanoparticulate form, with many of the notable syntheses summarized by
Chander (2005). However, inkjet deposition of these nanoparticles has not been as well-
documented. Cd-based nanoparticles, which have been inkjet printed, were also considered as
possibilities, but their more difficult synthesis processes and aqueous indispersibility were major
impediments to their use. Even a surface-functionalized aqueous CdSe QD solution was
incompatible with viscosity modifiers/surfactants necessary for printing, rapidly flocculating.
The inkjet fabrication of an entire device also precluded the consideration of most organic
materials, as printing of several different orthogonal layers would be an extremely difficult task.
More importantly, the extreme sensitivity of organic materials to contaminants – such as
surfactants and other ink additives – made them problematic for inkjet processing. A single-layer
inorganic or hybrid device structure was thus considered ideal.
Firstly, doped ZnS had to be synthesized. The synthesis method, using hydrated acetate salts of
Zn2+
, Cu2+
, and Mn2+
as the cation/dopant source and Na2S or thiourea as the S2-
source, is
described in detail in APPENDIX A. In order to arrest particle growth and maintain nanoscale
83
sizes for jetting, as well as to passivate surface states to improve PL and EL, different capping
agents were used, to varying degrees of success. Some of these capping agents were polymers:
sodium hexametaphosphate (SHMP), chitosan, poly(acrylic acid) (PAA), and PVP. The
remaining agents were polar molecules: mercaptoethanol, thioglycolic acid (TGA), 3-
mercaptopropionic acid (3-MPA), unpolymerized acrylic acid (AA), and citrate ion (from
sodium tricitrate). Bare particles, with no capping agent, were also prepared, relying on double-
layer repulsion to keep them dispersed (Warad et al. 2005). Capping agents likely to yield a
transparent, nearly monodisperse suspension were considered preferable, as they would be more
thermodynamically stable (Capek 2006); prior work mentioning monodispersity was of
particular interest to this study. Table 6.1 lists the capping agents used in this study and their
relative amounts in the reaction solution.
Table 6.1. Capping agents for ZnS:Mn nanoparticles, aqueous synthesis.
Capping agent Zn2+
:cap in solution Source
SHMP 10:1 (w:w) Warad et al. 2005
PVP Mw = 10 000
10:1 (w:w) Manzoor et al. 2003
Porambo & Marsh 2009
Citrate ion 5:3 (w:w) Peng et al. 2005
Chitosan 20:1 (w:w) Warad et al. 2011
Mercaptoethanol Not specified: 1:3 used (mol:mol) Vogel et al. 2000
TGA 0.5:1 (mol:mol) Zhang & Lee 2010
3-MPA 1:3-4 (mol:mol) Klausch et al. 2010, Schrage et al.
2010, Zhuang et al. 2003
AA 1:36 (mol:mol) Konishi et al. 2001
PAA Mw = 100 000
~4:1 (w:w) Konishi et al. 2001 Hwang et al. 2005
84
It is worthwhile to note that each of these capping agents provided water dispersibility, allowing
particles to be synthesized in an aqueous medium. However, the particles were then only water-
dispersible after synthesis. In order to produce organic-dispersible ZnS, the capping agents had
to be chemically altered. This was not feasible in all cases, but in species with a reactive acid or
alcohol group in solution, phase-transfer was accomplished by reaction of the ligand with
octylamine, producing a long lipophilic tail. This phase transfer has been previously described
by Klausch et al. (2010).
Besides the capping agent, other variables were considered during synthesis. One of the more
important of these was dopant concentration. For example, 1.5 at. % Mn2+
has been shown
repeatedly to be the optimum loading to achieve the maximum emission intensity from the
nanoparticles, whereas for Cu2+
doping, 1 at.% is desirable (Chander 2005). However, dopant
ions primarily populate the surface of the resulting nanoparticles (Igarashi et al. 1997, Adachi et
al. 2007, 2008; Yu et al. 1996, Bulanyi et al. 1998, 2002), where light emission is quenched at
surface states. Light can also be emitted by different mechanisms if excessive dopant is present
on the surface, which emit at different wavelengths and intensities than expected. One such
mechanism is the Mn–Mn exchange interaction which occurs in ZnS:Mn doped too heavily with
Mn2+
(Yang et al. 2003). With aqueous synthesis, the solubility differences between dopant and
Zn2+
ions in water has been reported to lead to unpredictable atomic percentages of dopant (Peng
et al. (2005). The entire amount of dopant dissolved in the precursor solution may therefore not
be effectively incorporated. Similarly, solubility differences between Zn2+
and S2-
in aqueous
solution may lead to non-ideal ratios of Zn2+
to S2-
, where insufficient S2-
is present in the
particles and deep blue emission dominates from S2-
vacancies (Yen 2004). So, different
amounts of dopant were added to the reaction solution to determine an ideal doping level.
Subsequently, the ratio of Zn2+
:S2-
in the reaction solution was varied (keeping dopant
concentration fixed) to achieve (Zn2+
+ dopant):S2-
= 1:1. The concentration of each species in
the final product was determined using inductively coupled plasma atomic emission
spectrometry (ICP-AES). Dried ZnS:dopant nanoparticles were dissolved in 25 v/v%
concentrated (15 M) HNO3 and 75 v/v% concentrated (12 M) HCl – aqua regia – at 80°C. 200
mg of the nanoparticles were added to 4 mL of acid mixture, dissolved, and diluted to 1 L with
85
deionized water. The Zn2+
, S2-
, and dopant percentages of the resulting solution were determined
using an Optima 7300 ICP-AES.
Reaction conditions for preparation of doped ZnS were initially ambient, and the nanoparticles
were removed from the reaction broth for isolation by centrifugation immediately after mixing
the reagents. However, it has been reported that in the synthesis of ZnS nanoparticles, time and
temperature play a role in the development of luminescent intensity (Manzoor et al. 2003,
Zhuang et al. 2003). Therefore, after having established the optimal proportions of reagents,
reaction time and temperature were both varied and optimized.
When the ZnS nanoparticles – using different capping agents, different amounts of dopants,
different reaction conditions, and so forth – were prepared, they all possessed several different
characteristics which were ultimately important to device function. These were: crystallinity and
crystalline structure, PL emission intensity, and perhaps most importantly for printing, particle
size. Dried nanoparticles’ crystalline structures were observed using a Philips PANalytical
PW1830 X-ray diffractometer (XRD) and compared to reference spectra. PL emission was
measured in solution: particles were redispersed in water at 0.1 w/w% using the pigment
dispersing additive ZetaSperse 1200 (tetramethyl-5-decyne-4,7-diol-2,4,7,9-propylene glycol),
provided as a sample by Air Products (0.05 w/w%). The PL of these solutions was observed
qualitatively first using a UVP UVM-57 ultraviolet lamp with a 302 nm emission wavelength.
PL intensities and spectra were then quantified using a Perkin-Elmer LS-55 spectrofluorophoto-
meter and polystyrene cuvettes. Finally, particle size was measured in four separate ways. It
was initially estimated from the XRD spectra using Scherrer’s Equation,
where L is the particle diameter, is the wavelength of the X-rays (1.5406 Å), is the width of
the diffraction peak at the Bragg angle at half-maximum intensity, K is 0.9 and is the Bragg
angle at which the peak is located (Cullity & Stock 2001). Secondly, particles in aqueous
solution were passed through a 0.2 µm syringe filter to determine if they made the cutoff for
printing on the DMP2831. Filtered solutions were observed under UV to determine if any
𝐿 =𝐾
𝛽 cos𝜃
86
particles had passed the filter. Thirdly, particle size distributions were obtained using a dynamic
light scattering (DLS) apparatus, a Malvern Zetasizer Nano ZS. Lastly, particle size was directly
observed using transmission electron microscopy (TEM) with a FEI Tecnai 20, by casting a drop
of dispersant-free nanoparticles onto a holey carbon grid.
Particles were also treated with additional capping agents post-synthesis to improve
dispersibility. These were AA monomer, PAA, ZnO, TGA, and SHMP. To prepare particles
capped with AA polymerized in-situ, 200 mg of dried ZnS nanoparticles were added to 22.5 mL
DI H2O and 5.5 g AA. The resulting solution was heat-treated at 80°C for 24 hours while
stirring vigorously to polymerize AA on the nanoparticles’ surface (Konishi et al. 2001, Althues
et al. 2006, Liu et al. 2008). Particles capped with PAA were prepared similarly, but 0.32 g of
PAA solution (35 w/w% PAA in H2O) diluted in 5 mL of DI H2O was added instead of AA as
described by Konishi et al. (2001), followed by ultrasonication for 2 hours. ZnO-capped ZnS
nanoparticles were prepared by adding 200 mg of dried nanoparticles to 200 mL DI H2O, which
were dispersed using ultrasonication and vigorous stirring. 10 mL of 0.05 M Zn(NO3)2 × 6H2O
solution was then added to the suspension, followed, dropwise, by 10 mL of 0.05 M NaOH
(Karar et al. 2004, Jiang et al. 2009). For TGA treatment, particles were dispersed in 0.023 M
TGA, and the pH was adjusted to ~9 with 0.5 M NaOH to obtain a nearly transparent dispersion
(Yang & Bredol 2008). Cu-based nanoparticles were not dispersible using TGA (it reacted with
the Cu dopant), so 1% SHMP was used instead. A detailed description of these dispersion
methods can be found in APPENDIX A.
6.1.5 Substrates
Several different substrates were used upon which to build the devices. These included slide
glass, ITO glass (patterned and unpatterned, described in Section 6.1.1), cellulose acetate
(Avery), and several paper types. The paper substrates were used after observing that the
functionality of the ACPELs deposited on paper changed depending on paper type, which was
attributed to the interaction of the sheet with the printed electrode layer, as that layer was in
direct contact with the paper sheet. Therefore, PEDOT:PSS-sheet interactions were considered
in more detail. Both commercial paper sheets and lab-prepared handsheets were used. The
87
detailed handsheet preparation procedure is given in APPENDIX E, along with the handling
procedure for commercial sheets. Handling and cleaning procedures for the impermeable (non-
paper) substrates are outlined in APPENDIX F. Sheet thicknesses were measured using a TMI
micrometer at 10 different points on the sheet. Contact angle was estimated using an aqueous
solution of crystal violet dye (test ink) prepared according to TAPPI Standard Method T431 for
measuring ink absorbency into paper. The test ink had a surface tension of 62 mN/m. 30 µL of
this ink was dropped with a calibrated micropipette onto a handsheet, and the resulting drop was
immediately photographed from the side using a Canon Rebel XT-ME DSLR camera with a MP-
E 65 mm macro lens. Finally, absorbency of the sheets was observed by measuring the time for
complete absorption of a 30 µL sample of the same test ink into the surface. During this test, the
samples were placed directly under a 60 W incandescent lamp elevated 30 cm from the test
specimen’s surface. Complete absorbency was defined according to Test Method T431 as the
point at which light reflection from the droplet on the surface was no longer visible. A summary
of the sheets’ properties is given in Table 6.2.
Table 6.2. Selected properties of commercial substrates.
Type Brand Thickness (µm) Contact angle (°) Absorbency
(µL/min)
Cellulose acetate Avery Inkjet
Transparency 159 26 0
Slide glass VWR Plain Microslide
960 51 0
ITO (glass, PET) Sigma-Aldrich Glass: 1090 PET:
140 48 0
Photo-paper Epson Premium
Photo 259 25 0.63
In the handsheets, different components were added to examine their effect on the performance
of printed films on their surface. These included filler (TiO2), fixation agent (PDADMAC),
internal sizing (alkylketene dimer, or AKD), and filler retention aid (poly(ethyleneimine), or
PEI), as well as two fibre types – kraft chemical hardwood (HW) and softwood (SW). A
summary of the handsheets’ properties is given in Table 6.3, and a description of characterization
methods used on these sheets is provided in APPENDIX G.
88
Table 6.3. Selected properties of lab-made handsheets.
Pulp Filler (w/w%) AKD PDADMAC Thickness (µm) Contact angle (°) Absorbency (µL/min)
SW
0
93 81 19
102 100 0.34
110 59 33
102 108 0.33
8.2
124 82 14
119 102 0.46
113 44 16
109 107 0.37
13.6
133 47 18
111 74 4.1
123 30 14
119 94 1.7
HW
0
92 52 51
97 100 0.35
105 31 43
101 107 0.33
10.3
108 50 62
98 114 0.39
90 40 30
85 102 0.36
16.1
108 57 25
116 92 0.46
139 53 28
127 97 0.37
6.2 Ink formulation
Ink formulation required iterative testing of fluid properties while adjusting the specific
components and proportions of those components. Because each ink contained different
functional materials, the ink components also varied in order to achieve and maintain dispersion
while retaining functionality. The choice of the most voluminous ink component, the primary
solvent, had the most bearing on the remainder of the ink components, which had to be chosen to
be compatible with that solvent system. Incompatibility was defined as the situation where the
addition of a particular component caused either an (undesired) chemical reaction, precipitation
of suspended/dissolved material, or liquid-phase separation (i.e. immiscibility). However, the
choice of the solvent, was restricted by the necessity for orthogonal solvents in a layer-by-layer
89
printed LED (Figure 6.1). In the most practicable device structures (summarized in Chapter 4),
namely the DC-LED or PEL, this implied different choices of solvent. In both cases, the
electrode material of choice was PEDOT:PSS, which was water-borne. Therefore, in the single-
layer DC-LED structure, the ZnS layer had to be insoluble in water. In the PEL structure, the
ZnS layer was more problematic, as the front and rear electrodes were both water-borne. Ideally,
the rear electrode would not be printed onto a layer which was water-soluble, as it required many
print passes to build up sufficient thickness (i.e. conductivity), which would prolong the time for
the PEDOT:PSS ink to solubilize the underlying layer. Therefore, an insoluble water-borne ZnS
layer and an organic-borne BaTiO3 layer were necessary for a quasi-orthogonal structure. To
avoid the ZnS layer solvating the water-soluble PEDOT:PSS anode, non-printed ITO was
considered as an alternate anode.
*No BaTiO3 layer necessary in the DCPEL.
Figure 6.1. Solvent selection for different device component inks, to produce an orthogonal structure.
These considerations necessitated the formulation of an aqueous PEDOT:PSS ink, an organic
BaTiO3 ink, and both an organic and an aqueous ZnS ink, containing a conductive and an
insulating binder, respectively.
With the solvent systems so defined, the remaining ink components were added to adjust the
fluid properties of the ink. The necessary ink properties for jetting on the testbed of choice, the
DMP2831, were given in Chapter 2. Many different reagents were used to adjust the relevant
properties of the inks to the acceptable values summarized in Table 2.2, including multiple
solvents, surfactants, and stabilizers. The exact materials used for each ink will be described in
the following sub-sections. As these materials were added, the fluid properties were monitored
using viscometry, tensiometry, zeta-potential analysis, and particle size analysis.
PEDOT:PSS
BaTiO3
ZnS + insulating binder
PEDOT:PSS or ITO substrate substrate
PEDOT:PSS
PEDOT:PSS
ZnS + conductivebinder
aqueous
aqueous
aqueous
aqueous
aqueous organic
organic
DC-LED DC* or ACPEL
90
Prior to testing, inks were ultrasonicated in an ice bath to thoroughly mix their components.
Viscosity was measured using a capillary viscometer (Kimax
, size 100, grade B17) at 25°C.
Equilibrium (not dynamic) surface tension was measured using a Sigma 700 Wilhemy platinum
plate tensiometer (KSV Instruments), at 25°C. Particle size was determined using a DLS
apparatus (Brookhaven Instruments) and a 2 W Ar laser, after 10 dilution to allow the laser to
pass through the darker-coloured inks, such as PEDOT:PSS. DLS was used for particle sizing
because the drying of some ink components – particularly surfactants – on a TEM grid would
obscure the particles of interest. Stability of the similarly-diluted inks was observed using a
Brookhaven ZetaPlus zeta-potential analyzer. Stability was also qualitatively observed by
monitoring the formation of precipitates either over time or immediately upon addition of an
incompatible ink component.
6.2.1 PEDOT:PSS inks
PEDOT:PSS served a similar role to ZnS’s role in materials selection: it was a model species for
ink formulation, upon which the ink formulation method was developed and applied to the other
inks. Therefore, the effect of precise adjustments to a formulation on its functional performance
was studied in greater depth for the PEDOT:PSS inks. Also, the indication of its printed
performance – electrical conductivity – was measured, as described later in this chapter.
PEDOT:PSS inks were formulated using previously described aqueous PEDOT:PSS suspension
as the basis. The viscosity of this suspension was 14.4 cP, and its surface tension was slightly
lower than that of water (61 mN/m). Therefore, it required treatment with both diluent and
surfactant. Inks were diluted with deionized water (the primary solvent) and DMSO to control
viscosity, improve conductivity, and to incorporate a co-solvent to avoid undesirable effects in
film topography, such as the coffee-ring effect. Water and DMSO loadings were systematically
varied to obtain a suitable viscosity for printing and maximum conductivity enhancement from
the DMSO. To prevent premature drying of the ink in the printhead nozzles, glycerol was added
to the mixture as a humectant. Glycerol has also been shown to be a mild conductivity enhancer
for spin-coated films of PEDOT:PSS (Lia et al. 2003). To adjust surface tension for improved
91
jetting and substrate wetting, several surfactants were added to the above ink formulation to
provide several ink types. SLS, Triton X-100 (Union Carbide, laboratory grade), Zonyl FS-300
fluorosurfactant (E.I. DuPont de Nemours et. co.), Igepal CA720 (General Dyestuff Co.), and
ZetaSperse 3700 (Air Products) were added to the ink in amounts ranging from 0.01 w/w% to 2.0
w/w%. An ink containing no surfactant was also studied. Cationic surfactants were not studied;
the addition of benzalkonium chloride caused flocculation through suspected reaction with the
PSS- anion. Each surfactant was loaded into the ink until its critical micelle concentration
(CMC) was reached. The CMC was estimated in the PEDOT:PSS/water/glycerol/DMSO system
as the concentration above which further addition of surfactant yielded no decrease in surface
tension. Surfactant CMCs (in this system) and their respective inks are summarized in Table 6.4.
All inks were treated with Surfynol DF-110D defoamer (Air Products) at 0.5 w/w% after final
formulation to control foaming caused by the surfactants.
Table 6.4. Surfactants tested in PEDOT:PSS ink and their CMCs.
Surfactant Type CMC (w/w%)
None n/a n/a
Triton X-100 Non-ionic 0.1
Zonyl FS-300 Fluoro- 0.02
Igepal CA720 Non-ionic 0.2
ZetaSperse 3700 Anionic 0.2
SLS Anionic 0.5
As was mentioned in Section 6.1.1, carbon species (SWCNTs, S-SWCNTs, MWCNTs, and C60)
were used as conductivity enhancers. The 0.04 w/w% solutions of these materials were also
added to the PEDOT:PSS inks, displacing water (the C solutions were mostly water). Loadings
from 0-10 w/w% C solution were tested in the PEDOT:PSS inks.
The effect of different components on the conductivity was considered in several ways – as a
function of drop formation, print quality, and chemical interaction. These were each examined in
turn using the printer-mounted camera, optical microscopy (Leica DM-LA), and Raman
spectroscopy (Raman Micro 300). Alterations in conductivity were predicted based on changes
in the distribution of double bonds in the PEDOT structure, and consequent shifting of the
92
primary PEDOT peak at Raman shift 1450 cm-1
, (Ouyang et al. 2004, 2005), or changes in
physical conformation of PEDOT chains (Fan et al. 2008).
6.2.2 BaTiO3 ink
The BaTiO3 ink was organic-based in order to be suitable for use in the ACPEL structure. The
BaTiO3 nanopowder (≤ 25 nm particle size) was dispersed in a several solvents, including water,
isopropanol, ethanol, MMA, and ethylene glycol, of which the best-suited organic solvents were
ethanol and MMA. Several dispersants were also tested, including oleic acid, terpineol,
Disperbyk 111 (Byk Chemie), Surfynol CT-324 (Air Products). , and ZetaSperse 1200 (Air
Products). A mixture of Surfynol CT-324 and Disperbyk 111, both dispersants for inorganic
oxides, was used to aid in dispersion. Methyl methacrylate (MMA) was also added to the ink to
be polymerized in situ into an insulating binder after printing. Finally, as a humectant and
viscosity increaser, poly(ethylene glycol) with an average molecular weight of 300 (PEG 300)
was added. The low surface tensions of the organic solvents meant that no surfactant was
necessary in this ink. The use of MMA precluded the use of immiscible water-based materials.
A dissolved binder polymer was not used as it was expected that a relatively large amount (>5
w/w%) of BaTiO3 would be required in the ink to demonstrate acceptable insulating
performance, and adding a dissolved polymer would likely be problematic for jetting.
6.2.3 ZnS inks
Where ZnS nanoparticles have been used in electroluminescent devices, they are generally
loaded into the precursor solvent for spin-coating at a concentration of ~1 w/w% (Yang et al.
2003, Schrage et al. 2010, Hieronymas 2002). Therefore, this concentration was targeted as an
optimal value for the inks. It is worth noting that as the ink is not composed entirely of colloidal
suspension but also of additives to control viscosity, surface tension, and drying rate, the “stock”
ZnS colloidal suspension forming the basis for the ink would necessarily be of a higher
concentration than 1 wt. %. The functional material, ZnS, was dispersed after synthesis either
into toluene, using oleylamine and 3-MPA, or into water, using TGA and NaOH (see
93
APPENDIX A and Section 6.1.4). These dispersions were stable when they contained 2.5 w/w%
ZnS. Therefore, the inks had to contain 40 w/w% ZnS “stock solution” as a starting point.
The organic ZnS ink was loosely based on a model CdS ink described in a patent by Cho & Cha
(2009). The patent suggested the use of a high-boiling organic solvent such as toluene or
chlorobenzene as a primary solvent, a low-boiling co-solvent like tetrahydrofuran (THF) or
cyclohexane, and a viscous surfactant to adjust surface tension and viscosity. So, the organic
ZnS ink used chlorobenzene as a primary solvent, and cyclohexane as a co-solvent. Because a
conductive binder was required, PVK was dissolved into the ink at a low concentration (<0.1
w/w%). The dissolved polymer was also expected to increase viscosity. The choice of this
polymer was important for device function: without dispersibility in the polymer layer, energy
transfer between the binder matrix and the nanoparticles is limited or non-existent (Schrage et al.
2010, Hieronymas 2002). Again, no surfactant was required, due to the already-low surface
tension of the organic solvents.
The aqueous ZnS ink was formulated with a similar template to the organic one. Unlike
PEDOT:PSS, the viscosity of the ZnS stock solution was not already high – it was approximately
that of water. Therefore, co-solvents like glycerol and PEG were considered. Ethylene glycol
butyl ether (or butoxyethanol), with its boiling point of ~170°C, moderate viscosity of 3 cP, and
low volatility, was an ideal candidate. Viscosity was also strongly affected by the presence of
the dissolved polymer, which was the insulator PVP in the case of the aqueous inks. In order to
avoid issues with “bead-on-a-string” jetting, a low concentration (<1%) of PVP (Mw = 1 300
000) was used. To reduce surface tension, a surfactant was needed – however, the tendency of
surfactants to interact with suspended particles and polymers (Jansen et al. 2001) was of concern
in an ink containing both. So, isopropanol was introduced to control surface tension (Vazquez et
al. 1995). This avoided issues with foaming and introduced a third solvent (along with water and
butoxyethanol) to further limit coffee-ring formation.
94
6.3 Jetting
Finalized inks were ultrasonicated in an ice bath for 30 minutes, filtered at 0.45 µm and then at
0.2 µm, and injected with a syringe into 1.5 mL DMP cartridges. Using the Dimatix Model
Fluid Waveform as a starting point, drop formation from the nozzles was observed using the
Drop Watcher software and camera included with the DMP2831 printer. Drop formation was
controlled by adjusting the voltage waveform applied to the piezoelectric nozzles – see
APPENDIX J for the jetting waveforms. In certain cases, an ink formulation failed to jet at all
and the formulation step had to be repeated. As in the previous section, for the sake of brevity,
these attempts are not described here. Ideal drop formation, described in Section 2.2.2, was the
goal. Once a suitable waveform was found, the ink was ready for jetting and observation of
jetted drops. All inks were jetted from a print height of 1 mm onto the substrate, and dried on a
hotplate at 150°C for 30 minutes.
Because particle size was carefully controlled during the formulation stage, theoretically none of
the functional materials were removed during filtration. However, the long CNTs were
hypothesized to have been at least partially removed. Passage of the as-supplied SWCNTs and
MWCNTs through the filters and print nozzles was evaluated using UV-visible spectrometry.
For filtration tests, the aqueous solutions of CNTs/SLS were passed through a 0.2 µm nylon
filter. For print nozzle tests, the previously filtered solutions were jetted using the DMP2831 at a
jetting frequency of 8 kHz into a collection well for 10 minutes. Changes in CNT concentration
across these steps were evaluated by obtaining absorbance spectra of the filtered and jetted CNT
solutions with a UV-visible spectrometer (Perkin-Elmer Lambda 20), utilizing the strong visible-
range absorption of CNTs at an arbitrary wavelength of 500 nm.
Isolated CNTs which passed filtration/jetting were observed using scanning transmission
electron microscopy (Hitachi HD-2000 STEM), using dried droplets of CNT solution on copper
grids. It was also expected that zeta-potential (ζ) of a CNT dispersion would increase in absolute
value as more CNTs were removed and the stability of the dispersion thereby increased, stability
often being closely related to the concentration of the dispersed phase (Tadros 1987). A certain
absolute value of ζ – usually 30 mV (Lin et al. 2010) – can be considered to represent a stable
95
colloidal solution, although a lower value of 25 mV has been reported (Kim & Ma 2011). ζ of
the filtrates and jetted solutions was expected to become increasingly negative in this case as
CNTs were removed, due to the anionic surfactant.
6.4 Drop spacing and film formation
Drop size and spread had a bearing on how films formed (see Section 2.4.2). Drop size testing
was performed by printing a spread of droplets of each ink at a spacing of 254 µm onto each
substrate. Films of the finalized inks were to form functional layers in an eventual working
device. Therefore, their wetting performance was important not only on a model substrate like
glass, but also on the surface of other materials. For example, the emissive layer was to rest on
PEDOT:PSS or ITO. So, glass substrates covered in each of these materials (except ITO) were
prepared by spin coating the inks (2000 RPM, 30 s). The resulting drops were imaged (without
drying) on the substrate using the optical microscope and their diameters measured.
With the drop size established, the drop spacing can be determined. Using a single nozzle,
several lines of each ink were printed onto each substrate at different drop spacings in
decrements of 5 µm, starting at the drop size measured in the previous step. The resulting films
were observed using the microscope. Ideal drop spacing was considered to be when lines were
fully merged (i.e. no holes) but not overlapping, as discussed in Section 2.4.2 and APPENDIX D.
Using the optimum drop spacing, films of 1-10 layers of each ink were printed on each substrate
and dried as described above. The films were then characterized for thickness and topography
using a Veeco WYKO optical profilometer at the edge of each film. The detailed procedure for
measuring thickness from the WYKO profiles is given in APPENDIX C. Given the roughness of
the printed films, thickness of film thickness varied from “valley” to “peak”. For the sake of
consistency, thickness was measured in the valleys of these layers. The optical profilometry
measurement also gave an idea of film morphology and the presence of pinholes, which could
compromise film function.
96
6.5 Ink distribution and print quality
A major factor in the electrical performance of any functional layer is its contiguity. As was
discussed in Chapter 2, inkjet printing produces sequential drops which form lines or blocks,
repeating this mechanism until the print is finished. The contiguity of the printed patterns in this
study was dependent on the properties of the ink, the nature of the substrate, and the successful
establishment of drop spacing. The patterns were optically examined for print quality in terms of
edge resolution, the presence of line overlap or inter-line holes, and minimum feature size. To
observe the print quality of ink layers, images were taken of the layers on their relevant
substrates using both the Canon Rebel XT-ME DSLR (MP-E 65 mm macro lens) and the Leica
DM-LA optical microscope. In the case of the PEDOT inks, which were in direct contact with
the substrates (including paper), a closer examination of the distribution of conductive material
was undertaken, in order to better understand the mechanism of conductivity change between
different substrates. These ink layers were examined using time-of-flight secondary-ion mass
spectrometry, or ToF-SIMS, which provided a molecular map of the substrate surface. An ION-
ToF ToF-SIMS IV apparatus was used to perform the measurements. A Bi3 primary ion gun was
used to induce ion ejection and fragmentation. The detailed ToF-SIMS analysis procedure is
described in APPENDIX K.
6.6 Functional testing of individual layers
6.6.1 Conductive ink
The main property of concern in the case of the conductive ink was conductivity. This was not
directly measured; resistance was instead measured using a multimeter, and by determining the
physical dimensions of the layer, conductivity could be obtained. Square samples were printed
and bus bars of carbon paste (DuPont MCM) were applied to two of their sides. 2-point
measurement across the sample yielded bulk resistance (R). This value is independent of sample
geometry. Therefore, to obtain a more meaningful figure – conductivity – resistance was used to
calculate resistivity (), the inverse of which is conductivity (). Resistance is related to
conductivity by the following expressions (Heaney 2000):
97
where w, h, and l are the width, thickness, and length of the conductive layer, respectively. With
a square sample, w = l and = (Rh)-1
. Therefore, conductive layer thickness was required to
estimate conductivity. This was obtained by cross-sectioning or profilometry of the printed
samples. A detailed description of these methods is provided in APPENDIX B and APPENDIX
C.
When using paper as a substrate, it is worth noting that “conductivity” in the case of a void-
filled, absorbent substrate, containing macropores, micropores (in the fibres themselves), and
hygroscopic, hollow fibres, was an estimate at best. Conductive material was not present as a
film, per se, but rather as a layer coated onto the fibres, and occasionally filling the voids
between them. Furthermore, the “film” was filled with a large volume fraction of non-
conductive regions – fibres, filler, and empty voids. Surface resistance/sheet resistance might
appear to have been a better metric by which to measure the electrical properties of such sheets,
as a result. However, the penetration of the ink into the bulk of the sheets suggested that not only
the sheet surface was involved in the conduction of electricity. Rather, the conductive volume
was less of a film and more of a composite block. For this reason, the bus bar carbon material
was diluted with DMSO to facilitate deeper penetration into the sheets and permit conduction
over the entire volume.
6.6.2 Emissive ink
The first means of ensuring that an emissive species was present and still functional in an ink
was observing its PL. This was observed using the aforementioned UV lamp (302 nm
excitation) and quantified using the fluorimeter, for inks in solution. Both PL and PLE (PL
excitation) spectra were obtained for the inks, as well as time-resolved PL, to confirm that
emission occurred from the dopant centres (Bube 1953). PL was also observed after printing and
drying of the ink. However, although this confirmed the presence of successfully synthesized
ZnS, it did not establish whether or not it was electroluminescent (the characteristic of interest
𝜌 =𝑅𝑤ℎ
𝑙 𝜎 =
1
𝜌
98
for an LED). In order to establish that the as-synthesized ZnS quantum dots were indeed
electroluminescent, the “stock solution” dispersions, 2.5 w/w% ZnS in water or toluene, were
diluted to 1 w/w%, mixed with 0.1 w/w% of their binder polymers (PVP and PVK, respectively),
and spin-coated onto unpatterned ITO slides using a Laurell WS-400-6NPP 150 mm spin coater.
These were spun at 2500 RPM for 15 seconds, followed by 200 RPM for 45 seconds, and dried
at 150°C for 30 minutes. 100 nm-thick circular Al cathodes were applied on top of the spin-
coated films by vacuum deposition. EL was observed by connecting the ITO and Al electrodes
to a power source (DC in the case of PVK-bound ZnS, and both AC and DC for PVP-bound
ZnS) and increasing voltage. The DC power source was an MPJA 0-50V, 3A benchtop power
supply. The AC source was built in-house; it was capable of delivering 50-200 VAC and 0.1A at
160 Hz frequency, using a near-square waveform with rise and fall times of <100 µs, a period of
16 ms, and a pulse width of 8 ms. Inks in devices displaying EL were considered suitable for
printed device fabrication.
6.6.3 Insulating ink
The primary measure of an insulating film’s function is its dielectric constant. This is closely
related to the film’s geometry and capacitance. BaTiO3 ink was printed onto ITO glass with
different numbers of overprinted layers to observe film structure, thickness, and capacitance.
Samples had to be dried at 250°C for 60 minutes on a hotplate in air to ensure the evaporation of
all of the glycol and to polymerize the MMA. Capacitance of the resulting films was estimated
by first applying an Ag electrode using an Ag pen, covering the surface of the samples. Then,
capacitance was measured between the Ag and the ITO using an Agilent U1701A capacitance
meter, treating the sample as a parallel plate capacitor. Dielectric constant () was estimated
from capacitance according to the following formula:
= 𝐶𝑑
𝜀0𝐴
99
where C is capacitance, d is the distance between plates (film thickness), o is the vacuum
permittivity, and A is the sample area (ASTM International 2004). In this case, the sample area
was the size of the printed samples, which were 2 mm 2 mm squares.
6.7 Device fabrication and testing
Figure 6.2. Light-emitting device builds prepared in this study. Layer thicknesses were based on the following: Al
cathode, Hieronymas (2002); ZnS/PVK, Hieronymas (2002), Schrage et al. (2010), and Yang et al. (2003); BaTiO3
and ZnS/PVP, Adachi et al. (2007, 2008); PEDOT:PSS anode, Yang et al. (2003). ITO thickness was specified by
the supplier. PEDOT:PSS cathode thickness was arbitrary.
The device structures most suitable for the materials and printing method – the DC-LED and the
two PEL structures – have already been discussed in Section 6.2 on ink formulation. The
simplicity of these structures and their incorporation of binder polymers (rather than self-
PEDOT:PSS
BaTiO3/PMMA
ZnS:dopant/PVP
ITO substrate
substrate
Al
ITO
ZnS:dopant/PVK
DC-LED
substrate
Al
PEDOT:PSS
ZnS:dopant/PVK
substrate
PEDOT:PSS
PEDOT:PSS
ZnS:dopant/PVK
Al
BaTiO3/PMMA
ZnS:dopant/PVP
substrate
Al
PEDOT:PSS
BaTiO3/PMMA
ZnS:dopant/PVP
substrate
PEDOT:PSS
substrate
Al
ITO
ZnS:dopant/PVP
substrate
Al
PEDOT:PSS
ZnS:dopant/PVP
substrate
PEDOT:PSS
PEDOT:PSS
ZnS:dopant/PVP
300 nm
200 nm
~100 nm 100 nm
1 µm
1 µm
20 µm
ACPEL
DCPEL
100 nm
~100 nm
20 µm
100 nm
~100 nm
100 nm
100 nm
1 µm
1 µm
100
assembled or monolithic layers) made them ideal for inkjet deposition. In order to minimize the
number of variables which might affect device function after printing, glass was used as the
substrate for all the device builds.
The two electrodes of the printed devices could be deposited conventionally as well – i.e. ITO as
the anode (as supplied), and metal (typically Al) as the cathode, using vapour deposition. So,
DC-LED devices utilizing either one printed layer (the ZnS/binder), two printed layers
(PEDOT:PSS anode and ZnS/binder), or all three layers printed were prepared. ACPEL devices
were prepared similarly, as were DCPELs. The resulting printed structures and the thicknesses
of their layers are shown in Figure 6.2. The layer thicknesses for DC-LEDs were based on those
described by Hieronymas (2002), Schrage et al. (2010), and Yang et al. (2003). The layer
thicknesses for the PELs were based on those described by Adachi et al. (2007, 2008). Layer
thickness was controlled by setting the number of print passes, based on the thickness of
individual films determined by optical profilometry. These previous works also utilized ZnS
nanoparticles as the emissive species. Each of the layers of each device, besides the ITO or
vacuum-deposited Al, was printed using its optimum line spacing, with thickness controlled by
the number of layers deposited (see Section 6.4). Layers were all dried at 150°C for 30 minutes
in air and the full devices were again dried at 200°C for 30 minutes to ensure the removal of any
remaining solvents. Figure 6.3 shows a bird’s-eye view of a device being built.
Figure 6.3. Schematic drawing showing a bird's-eye view of device construction.
substrate
ZnS/binder
BaTiO3/PMMA
PE
DO
T:
PS
S o
r A
l
1) Print PEDOT:PSS anode, dry
(or prepare ITO)
2a) Print ZnS/binder
2b) Overprint BaTiO3/binder
3) Print PEDOT:PSS cathode,
dry (or vacuum-deposit Al)
101
6.7.1 Interlayer dissolution
The simplest method for establishing whether overprinted layers dissolved one another was a
solvent rub test, specifically the test outlined in ASTM Standard D5402-06 for the solvent
resistance of organic coatings (2006). In this case, the method was adapted to use printed films
of each layer, and the solvent rub was replaced with the overprinted ink (e.g. PEDOT:PSS ink
films were rubbed with ZnS ink, ZnS films with PEDOT:PSS and BaTiO3 inks, and so forth).
Single layers of each ink were printed onto slide glass and rubbed once across with a wipe which
had been soaked in the overlying ink (s). Rub resistance was considered acceptable when the ink
film remained with no visible damage after the rub. Any visible damage to the film constituted a
failed rub test. In the case of the aqueous ZnS/PVP ink, overprinting with both BaTiO3 ink and
PEDOT:PSS ink (for the PEL builds) was necessary. PVP is highly soluble in many solvents;
even the orthogonal BaTiO3 solvent system contained ethanol, which solvates PVP. To prevent
this from occurring, the aqueous ZnS films containing PVP were UV-treated to induce cross-
linking to prevent their dissolution by overprinted layers according to the method described by
D’Errico et al. (2008). UV treatment was carried out with a Trojan UV lamp (254 nm) over 4
hours to achieve complete cross-linking without decomposing the samples.
6.7.2 Electrical characterization
For the DC-driven devices, a programmed DC power source (Keithley 647) and its coupled
lightmeter (Minolta LS-110) were used to detect I-V characteristics, luminance, and efficiency
simultaneously on a HP 4140B multimeter and its accompanying software. Voltage was
increased at 0.1 V/s to a maximum of 50 V, with a maximum drawn current of 10 mA. AC-
driven devices were connected to the lab-made AC power source described in Section Emissive
ink6.6.2. Because this voltage source did not have a coupled lightmeter, luminance was
measured with a Sekonic Flashmate L-308S light meter at a distance of 1 cm directly above the
sample in a darkened room.
102
Chapter 7
7 Results & discussion
7.1 ZnS synthesis
ZnS nanoparticles were prepared based on competitive precipitation method with the reaction of
hydrated acetate salt of zinc and Na2S, Na2S2O3, or thiourea, using two different dopants. The
first was Mn2+
, producing ZnS:Mn, which emits in the orange range at 585-595 nm (Brus 1986).
The second was Cu2+
, producing ZnS:Cu, which emits in the green range at 470-480 nm (Bowers
& Melamed 1955). Doping with Ag+ to produce blue-emitting nanoparticles was attempted, but
the reaction mixture inevitably formed insoluble Ag2S, which precipitated out and was
presumably not incorporated into the ZnS crystal lattice. As is described in APPENDIX A, the
synthesis for both Mn- and Cu-doped nanoparticles was similar; for the sake of brevity, a study
of the synthesis of Mn-doped nanoparticles is described in this section in more detail. For
clarification, it should be noted that atomic percent (at.%) refers to the percentage of Zn atoms
substituted with Mn atoms in the ZnS lattice. In a material like ZnS, with one atom of Zn per
atom of S, at. % and mol % are interchangeable terms. The at. % Mn2+
reported here represents
the mol % of Mn2+
in the ZnS:Mn theoretically yielded by the reaction.
7.1.1 Mn2+ loading
Synthesis of ZnS with several different capping species, including citrate, chitosan, and PVP, at
the suggested Mn2+
loading of 1.5 at.% (in the reaction solution), resulted in weak PL. This
weak PL from Mn2+
centres was expected to be in the characteristic orange range, corresponding
to the energetic transition of Mn2+
from the 4T1 state to the
6A1 state (Karlin 2005). In Figure
7.1, weak, red-shifted PL was observed, suggesting both poor Mn2+
incorporation and poor
surface passivation. Red-shifting of PL emission in ZnS:Mn nanoparticles due to lower quantum
efficiency has been observed in the past (Bhargava et al. 1994); moreover, the lack of a
sufficiently passivating layer on the nanoparticles may have led to surface recombination events,
and caused a reduction in PL emission energy (i.e. longer wavelength). The occurrence of a
similar shift even in the particles passivated with polymers suggested that polymeric capping
103
failed to sufficiently cap the particles. Particle size also appeared to be large, as none of the
particles could pass through a 0.2 µm filter, and they presented as a sedimented white precipitate.
Figure 7.1. PL emission from ZnS:Mn nanoparticles (1.5 at.% Mn). (a) no cap; (b) citrate cap; (c) PVP cap; (d)
chitosan cap. Spectra are stacked in order to compare emission at 595 nm.
The PL of the nanoparticles was drastically improved for all the polymeric caps when using
higher concentrations of Mn in the reaction mixture, although it was also red-shifted and
eventually dropped off as more non-luminescent MnS was formed (see Figure 7.2, Figure 7.3).
In general, below 10% Mn in the precursor solution, no notable PL emission was observed, and
PL intensity did not peak until 50% Mn or higher at 608 nm. Stronger luminescence in the red
region at higher loadings was attributed to Mn-Mn interactions, as well as surface recombination
events resulting from an excess of Mn2+
ions and MnS on the surface of the nanoparticles (Yang
et al. 2003). The high surface to volume ratio of the nanoparticles would favour the population
of the surface with a large number of MnS molecules and Mn2+
ions in highly Mn-doped
precursor solutions.
530 540 550 560 570 580 590 600 610 620 630
Ph
oto
lum
ine
sc
en
t in
ten
sit
y (
arb
itra
ry u
nit
s)
Emission wavelength (nm)
PL spectra (stacked plot)
(d)
(a)
(b)
(c)
104
Figure 7.2. PL intensity vs. Mn added (at.%), comparing uncapped and PVP-capped particles. Spectra are
normalized to the highest-intensity PL emission (100 at.%).
520 540 560 580 600 620 640 660 680
Ph
oto
lum
ine
sc
en
ct
inte
ns
ity,
no
rma
lize
d
Emission wavelength (nm)
100%
80%
60%
50%
40%
20%
10%
5%
2%
520 540 560 580 600 620 640 660 680
Ph
oto
lum
ine
sc
en
ct
inte
ns
ity,
no
rma
lize
d
Emission wavelength (nm)
100%
80%
60%
50%
40%
20%
10%
5%
2%
PVP cap
No capping agent
1
0
1
0
105
Figure 7.3. PL of uncapped ZnS:Mn nanoparticles suspended in water under 302 nm UV excitation.
The actual loading of Mn in the particles, measured using ICP-AES, confirmed that the peak PL
at ~608 nm in the orange range (i.e. not red-shifted to 630 nm, as emission was in the case with
loadings > 50%) occurred when Mn loading was ~1.5% (Figure 7.4). As expected from the PL
spectra, the actual Mn2+
content of the nanoparticles was significantly different from that added
during synthesis (Table 7.1). The increased solubility and higher activity of the Zn2+
ion relative
to the Mn2+
ion made Mn2+
incorporation difficult in an aqueous solution at room temperature.
The inclusion of citrate, which has been reported to assist in solubilizing Mn2+
(Peng et al. 2005),
naturally led to higher concentrations of Mn in the citrate-capped particles.
Table 7.1. Actual Mn content of ZnS:Mn nanoparticles determined by ICP-AES.
Mn2+
added (at. %)
Mn2+
retained in particles (at. %)
No capping agent PVP cap Citrate cap
0.5 0.14 0.02 0.10
1 0.17 0.04 0.20
2 0.30 0.30 0.28
5 0.61 0.62 0.65
10 0.74 0.42 1.95
20 1.24 0.79 6.27
30 0.74 0.85 4.77
40 1.24 1.26 3.46
50 1.47 1.85 6.14
60 2.10 1.82 8.09
80 2.15 2.60 8.96
100 5.10 3.80 10.74
UV Visible
0% 1.5% 10% 20% 50% 100%
Mn loading
0% 1.5% 10% 20% 50% 100%
106
It was hypothesized that the inclusion of citrate in the precursor solution may have resulted in
nanoparticles so heavily loaded with Mn that non-emissive MnS would form in large clusters
rather than incorporating as a sporadic dopant in the ZnS lattice. Emission from Mn-Mn
interactions would therefore be very strong, which explained the increasing PL intensity in the
red region shown in Figure 7.4. This hypothesis was supported by the XRD studies, presented
subsequently, which showed small shoulder peaks in the citrate-bearing nanoparticles’ spectra, as
well as by the texture and morphology of the particles. The citrate-bearing ZnS:Mn was not a
dry powder even after vacuum drying, but retained a waxy consistency. It was expected that this
was a result of the presence of citrate complexes of either Mn2+
or Zn2+
, which may have affected
results of ICP-AES and PL measurements.
Figure 7.4. PL intensity at 608 nm vs. actual Mn content. Inset: 50% Mn (added during synthesis) containing ZnS:Mn
nanoparticles in water under 302 nm UV excitation. PL intensity is normalized to the highest value (no capping agent,
~1.5 at.%).
0.01 0.1 1 10 100
Ph
oto
lum
ine
sc
en
ct
inte
ns
ity,
no
rma
lize
d
Actual Mn content of ZnS:Mn nanoparticles (at.%)
No capping agent
PVP cap
Citrate cap
None PVP Cit.
1
0
107
7.1.2 Zn2+ to S2- ratio
Besides the amount of Mn2+
doped into the nanoparticles, the ratio of Zn2+
to S2-
in the precursor
solution has been noted to affect the PL intensity and spectrum. Generally, a ratio of Zn2+
:S2-
of
2:1 has been recommended for optimal particle size and emission intensity (Adachi et al. 2007,
2008; Althues et al. 2006). With the particles produced in this study, the same trend in higher
emission intensity was observed for particles with this precursor ratio, as can be seen in Figure
7.5. Nanoparticles prepared with ratios greater than 2:1 generally showed little to no PL
emission in the orange range, although they did still emit in the blue range (characteristic of
undoped ZnS (Yen 2004, 2006). This was due to the large excess of Zn2+
in solution which
consumed the S2-
before it reacted with Mn2+
; 3:1 retained weak orange PL, but 4:1 and above
had insignificant orange PL. Also, the large amount of Zn2+
resulted in numerous S2-
vacancies
in the particles, which are responsible for blue emission (Yang & Bredol 2008). Nanoparticles
prepared with ratios less than 1:1 emitted very weakly in all ranges, suggesting that this ratio was
the lower limit to achieve PL in ZnS:Mn prepared in aqueous solution, because insufficient host
material (i.e. ZnS) was present to allow PL emission from luminescent centres.
Figure 7.5. PL emission in uncapped ZnS:Mn particles (50% Mn), different Zn:S ratios in reaction solution.
360 410 460 510 560 610 660
Ph
oto
lum
ine
sc
en
t in
ten
sit
y,
no
rma
lize
d
Emission wavelength (nm)
0.25:1
1:1
2:1
3:1
5:1
10:1
1
0
108
The difference in emission intensity between 2:1 and 1:1 Zn2+
to S2-
ratios was small and within
the range of experimental or instrumental uncertainty. However, the 2:1 emission colour was
also slightly blue-shifted relative to the 1:1 spectrum. This could be because of the expected
effect of Zn2+
on particle size, observed in a previous study (Althues et al. 2006). Excess Zn2+
preserved a smaller particle size in that study, leading to a slight blue-shift in the emission
spectrum. Also, the smaller size of the crystallites implied that they were more likely to remain
suspended in water (for PL testing), so it is possible that the slightly brighter PL observed for 2:1
Zn2+
:S2-
nanoparticles was simply because they were better dispersed than 1:1 nanoparticles.
Also, the smaller particle size would lead to less diffuse scattering from large agglomerates and a
higher perceived PL intensity.
7.1.3 Post-synthesis capping
Regardless of the Zn/S/Mn ratio, all of the particles synthesized with the caps initially tested (i.e.
PVP, citrate, and chitosan) displayed red-shifted emission and large particle size (i.e. were not
able to pass through a 0.2 µm nylon filter). The red-shift was due to poor surface passivation,
which also likely resulted in excessive particle growth. So, other caps were applied after
synthesis to see if they could reduce this effect. Capping of the nanoparticles with AA or PAA
did have a beneficial effect on PL intensity (Figure 7.6). The effect may have been the result of
better passivation of surface states, as was noted in our previous work (Angelo & Farnood 2011)
and by other groups using AA (Igarashi et al. 1997, Adachi et al. 2007, Konishi et al. 2001,
Althues et al. 2006). AA capping has been highlighted as ideal for ZnS:Mn due to the
excitation at 350 nm of C=O bonds which are bonded to S2-
in the ZnS structure, resulting in
enhanced PL through additional energy transfer to the Mn2+
luminescent centres (Konishi et al.
2001). PAA capping did not show this effect due to the lack of C=O double bonds. So, in this
case, where PAA capping improved PL the most of any treatment, it is more likely that PAA
simply served as a better dispersant, retaining a larger fraction of ZnS:Mn in solution and
increasing apparent emission intensity, or reducing scattering. PAA-capped nanoparticles also
emitted with a slight blue shift (to 605 nm) compared to bare ZnS:Mn. However, AA-capped
nanoparticles showed a further blue shift (to 595 nm emission) compared to all the other capped
particles as well as bare ZnS:Mn nanoparticles. The ZnS:Mn-AA solution was also significantly
109
more transparent than the others (see inset of Figure 7.6). This effect closely resembled that
observed with the 2:1 Zn2+
:S2-
nanoparticles versus the 1:1 nanoparticles. Therefore, it was
hypothesized that AA served to better water-disperse and isolate the nanoparticles than the other
capping agents.
Figure 7.6. Effect of capping agents added after synthesis on PL of ZnS:Mn nanoparticles (Zn:S:Mn = 2:1:0.5, i.e.
50% Mn in solution). Inset: Capped ZnS:Mn nanoparticles in water under 302 nm UV excitation.
The addition of an inorganic shell, ZnO improved PL likely due to, passivation of surface states,
as was previously reported (Karar et al. 2004, Jiang et al. 2009). However, reduction of surface
recombination events was not sufficient to blue-shift the particles as in the case of PAA or AA.
Also, the further addition of PAA and AA over the ZnO layer did not compound the effects of
the two caps. PL was quenched somewhat; the reason for quenching of PL due to the ZnO shell
was not well-understood, although it was suggested that the presence of ZnO precluded AA or
PAA molecules from chemically bonding with the ZnS:Mn metal or S2-
ions, thereby
diminishing their beneficial effects on PL due to bond excitation.
360 410 460 510 560 610 660
Ph
oto
lum
ine
sc
en
t in
ten
sit
y,
no
rma
lize
d
Emission wavelength (nm)
No capping agent
AA cap
PAA cap
ZnO cap
ZnO AA PAA
1
0
110
7.1.4 Reaction temperature and time
Most syntheses of ZnS reported in the literature were undertaken at room temperature, and when
these procedures were replicated (as described above with PVP, citrate, and so on), they
immediately produced visible precipitates of unsuitably large size. However, increased
temperature during particle nucleation, along with increased pH and higher reagent
concentrations was known to possibly yield smaller nanoparticles (Capek 2006). Therefore,
temperature was maintained at an increased value for the duration of the reaction.
Figure 7.7. PL spectra of ZnS:Mn nanoparticles capped with SHMP at RT and 70°C, reacted 16 h. Inset shows
particles redispersed in water under 302 nm UV excitation. The spectra are stacked to compare emission at 595 nm.
As is shown in Figure 7.7, particles prepared with another capping agent, SHMP, demonstrated
greatly increased PL intensity when held at 70°C throughout their synthesis. The synthesis was
also carried out for 16 h (see the following section on reaction time). This was likely the result
of both better dispersion (smaller particle size) and better Mn2+
integration due to faster diffusion
through the boundary layers. More rapid diffusion of Mn2+
towards the particles’ centres also
likely reduced the number of surface states, resulting in emission which was not as red-shifted as
that observed with room-temperature (RT)-prepared ZnS:Mn. Increases in temperature below
520 540 560 580 600 620 640
Ph
oto
lum
ine
sc
en
t in
ten
sit
y (
arb
itra
ry u
nit
s)
Emission wavelength (nm)
PL spectra (stacked plot) 70°C
RT
111
70°C did not have much notable effect, and rapid evaporation became a problem at more
elevated temperatures. A similar result observed for ZnS:Cu nanoparticles was described by
Klausch et al. (2010).
Figure 7.8. PL spectra of ZnS:Mn nanoparticles capped with SHMP aged for varying times. All were prepared at
70°C and neutral pH in H2O. Inset shows particles in water under 302 nm UV excitation. The spectra are stacked to
compare emission at 595 nm
It had been suggested that increasing the reaction time of doped ZnS nanoparticles leads to major
improvements in luminescent intensity (Klausch et al. 2010, Zhuang et al. 2003). This is likely
due to the longer time available for dopant ions to diffuse into the bulk of the particle. Allowing
longer time and higher temperature to improve diffusion of Mn was necessitated by the solubility
difference between ZnS and MnS (Yang et al. 2005). ZnS, having a solubility constant of 1.1
10-24
, will immediately form as insoluble nanoparticle “seeds”, not allowing MnS (Ksp = 2.6 10-
13) to co-precipitate inside the crystal lattice. Instead, MnS forms as a surface layer, quenching
luminescence. Control of temperature and time may have overcome this effect somewhat.
Furthermore, oxidation of the surface by dissolved oxygen may also have passivated the surface
520 540 560 580 600 620 640
Ph
oto
lum
ine
sc
en
t in
ten
sit
y (
arb
itra
ry u
nit
s)
Emission wavelength (nm)
PL spectra (stacked plot) 0.5 h
1 h
4 h
16 h
112
(Zhuang et al. 2003); passivation (and therefore capping) playing a major role in development of
PL intensity. In this vein, by observing PL emission from ZnS:Mn nanoparticles capped with
SHMP at different times after the mixing of the S2-
and Zn2+
/dopant solutions, the positive effect
of longer reaction time on photoluminescence was easily observed (Figure 7.8). Dopant-related
orange emission at ~595 nm was drastically increased by aging of the reaction broth before the
removal and isolation of the nanoparticles. Also, the emission was initially red-shifted (after 4
hours) due to the presence of a large number of Mn2+
defects on the surface of the nanoparticles
(Yang et al. 2003). After aging for 16 hours, the emission was blue-shifted in the more typical
orange range, resulting from the quenching of these effects by Mn2+
diffusion into the particle
and, possibly, by improved dispersity and reduced particle size. Further aging of the particles led
to excessive particle growth and agglomeration, as is predicted by the La Mer model for particle
nucleation and growth (the Ostwald ripening phase). After >17 h aging, the majority of the
particles had settled to the bottom of the reaction flask, making them unsuitable for inkjetting.
Figure 7.9. PL spectra of ZnS:Mn nanoparticles capped with varying amounts of SHMP. Ratios are in wt:wt. All
were prepared at 70°C for 16 h in water. Inset shows particles in water under 302 nm UV excitation. Spectra are
stacked to compare PL emission at 595 nm.
520 540 560 580 600 620 640
Ph
oto
lum
ine
sc
en
t in
ten
sit
y(a
rbit
rary
un
its
)
Emission wavelength (nm)
PL spectra (stacked plot) 5:1 Zn:SHMP
10:1 Zn:SHMP
100:1 Zn:SHMP
200:1 Zn:SHMP
113
In many of the works describing previous studies of this type, the amount of capping agent used
is often left unmentioned. How much capping agent is used can determine PL intensity, as well
as dispersion and stability, since it is directly responsible for maintaining dispersion and
passivating the particle surfaces (Chander 2005, Caruso 2004). However, it was noticed that
using large amounts of any cap led to early precipitation of the particles, and that the solubility
limit of materials like SHMP or PVP was quickly reached.
In fact, decreasing SHMP concentration was sufficient to greatly increase PL (Figure 7.9). This
came at the cost of dispersion and passivation, as the particles containing little SHMP – i.e. only
1/200 of the Zn2+
loading – began to show red-shifted emission spectra, Emission at 585-590 nm
was strongest in particles containing 100:1 weight ratio of Zn2+
precursor to SHMP. The
presence of additional capping agent was thought to retard Mn2+
diffusion and incorporation into
the nanoparticles, and to prevent S2-
from filling sulphur vacancies on the particle surface. This
lead to a large number of surface states and enhanced undesirable blue emission. Solubility
differences between the component ions, especially when passing through a thick boundary layer
of dissolved SHMP, also may have prevented ZnS from crystallizing properly. A similar effect
was seen when adding excessive amounts of Zn2+
(in Figure 7.5).
7.1.5 Reduction of particle size & improvement of dispersion
After establishing the ideal reaction conditions, molar ratios, and cap amounts to produce the
most highly luminescent nanoparticles, it was found that they were universally still unable to
pass a 0.2 µm filter. Particles capped with PVP, chitosan, citrate, PAA, AA, and ZnO were all
brightly photoluminescent, but not suitable for printing. Furthermore, the maximum loading of
any of the nanoparticles to achieve stability (i.e. no settling or precipitates) was < 1 w/w%. It
was evident even during synthesis that the particles were forming large agglomerates that were
not redispersible – the reaction solution became opaque and white upon the addition of the S2-
solution, instead of transparent and unaffected by light scattering, as would be the case with a
monodisperse nanoparticle suspension (Figure 7.10). It was at this point that the addition of a
more suitable capping agent was considered, with the highly-reactive mercaptans being prime
candidates. As Zhuang et al. (2003) suggested, the use of mercaptans might provide better
114
passivation by actively filling S2-
vacancies on the nanoparticles’ surfaces with the thiolated end
of the molecule, forming a strong bond.
Therefore, an attempt was made to optimize particle size by using the mercaptan stabilizers.
However, the issue of complexes forming between the reactive S-bearing ends of the molecules
and the Zn2+
ion before the S2-
/dopant sources were introduced was still present. The key to
successfully dissolving the mercaptans in the Zn2+
/dopant precursor solution and leaving the
cations free to react with the S2-
source lay in the protonation of the thiol group. According to
Adachi (2008), by adjusting the solution pH to a basic value to induce re-protonation of the thiol
group, the Zn2+
/dopant-mercaptan complex could be dissociated , and as suggested in Klausch et
al. (2010) and Zhuang et al. (2003), the Zn2+
complex became soluble and the precursor solution
clarified.
Figure 7.10. ZnS:Mn nanoparticles in water (0.1 w/w%), different stabilizers: (a) none; (b) chitosan; (c) PVP 10000;
(d) SHMP; (e) 3-MPA.
Using all three mercaptans (mercaptoethanol, TGA, and 3-MPA) at pH values > 8 (adjusted
using NaOH and buffered), reaction mixtures were formed which showed no white colour or
agglomeration upon mixing of the Zn2+
and S2-
precursors but displayed strong
photoluminescence after reaction. 3-MPA was settled upon as the easiest to use of the three
UV (302 nm) unfiltered filtered (0.2 µm)
Visible
(a) (b) (c) (d) (e) (a) (b) (c) (d) (e)
(a) (b) (c) (d) (e) (a) (b) (c) (d) (e)
115
stabilizers simply because of its significantly reduced respiratory toxicity compared to the other
two. Particles were so well-dispersed that they could not be removed from the reaction broth by
simple centrifugation, but rather had to be “crashed” out of solution using acetone, in which the
stabilizing ligands were insoluble. Figure 7.10 shows the comparative dispersity and PL
intensity qualitatively between the 3-MPA-capped particles and those using the other caps.
7.1.6 Optimized synthesis procedure
Figure 7.11. Finalized synthesis method of water-dispersible 3-MPA-capped ZnS:Mn quantum dots.
A finalized synthesis method, to produce these 3-MPA-capped particles, is given in APPENDIX
A. The use of the mercaptan species allowed the S-bearing end of the molecule to react with
Mn(Ac)2
Na2S
H2O
Zn[SCH2CH2COOH]2
H2O S
2-
Na+
NaOH
+ ZnS
ZnS
ZnS
ZnS
MnS
Ac-
S2-
Ac- Ac
-
Ac-
Ac-
S2-
S2-
S2-
S2-
Mn2+
Zn2+
Zn2+
Zn2+
Zn2+
Zn2+
Mn2+
Zn2+
CH3COO-
3-MPA
Mn[SCH2CH2COOH]2
pH 10
buffer
Zn2+
CH3COO-
Mn2+
(H+)(SCH2CH2COOH
-)
pH = 10.3
pH = 10 Δ S
2-
Na+
pH = 10
T = 70°C
(1)
(2)
(1) (2) Δ
S
O
OH
16 h
T = 70°C
ZnS:Mn
O
OH
O
OH
O
HO O
OH
O
OH
Zn(Ac)2
116
Zn2+
during particle growth, thus filling the S2-
vacancies and forming a strong passivating layer,
as is shown in Figure 7.11 (Zhuang et al. 2003, Vogel et al. 2000, Small et al. 2010).
Another advantage offered by using a thiolated acid was that the tail end of the molecule, being a
carboxylic acid group, was water soluble. Furthermore, the electronegativity of the O atoms
present in the group polarized the molecule, providing a highly negatively-charged region around
each nanoparticle to prevent agglomeration. Finally, the acid group could be reacted with an
amine (in this case, octylamine) to produce a long, lipophilic tail, rendering the particles
dispersible in non-polar solvents, such as hexane, chlorobenzene, and toluene (Klausch et al.
2010). A detailed explanation of the phase-transfer process is also provided in APPENDIX A.
Improvement in the dispersion of 3-MPA-capped ZnS:Mn versus SHMP-capped ZnS:Mn can be
easily seen in Figure 7.12 where both are dispersed using 0.023 M TGA at pH 9.
Figure 7.12. Comparison of ZnS:Mn dispersion in water (1 w/w%). In all of the images, the vial on the left contains
ZnS:Mn capped with SHMP; the right vial, ZnS:Mn capped with 3-MPA.
7.1.7 Synthesis of ZnS:Cu nanoparticles
An important advantage of this synthesis method was that it was readily applied to another
dopant, Cu2+
. In experiments conducted with both of Mn2+
and Cu2+
dopants, the solution
became immediately black and unstable upon the addition of the Na2S solution due to
irreversible reaction of these dopants with S2-
before incorporation into the ZnS lattice. The
introduction of a different S2-
source, thiourea was suitable for forming Cu-doped ZnS due to its
slower dissociation into S2-
ions (pKa = 2.03) and lower reactivity (Klausch et al. 2010, Schrage
et al. 2010). The method used was identical to that outlined in Figure 7.11, with a few notable
exceptions.
Visible 302 nm UV
117
Figure 7.13. PL and PLE spectra of ZnS nanoparticles capped with 3-MPA (2.5 w/w% in water). Insets show 3-MPA-
capped particles in water/SHMP or water/TGA/NaOH solutions under 302 nm UV excitation. PLE and PL spectra are
normalized to their maximum values to show emission colour clearly.
Firstly, the Na2S was replaced with excess thiourea, due to its slow release of S2-
anion.
Secondly, the Mn2+
dopant was replaced with Cu2+
, also in the form of a hydrated acetate salt.
Lastly, the pH was adjusted to 8 and buffered at 8. Finally, based on the work of Klausch et al.
(2010), heat treatment was performed at 95°C rather than 70°C to maximize PL intensity for
ZnS:Cu. In that work, 95°C was used as a reaction temperature presumably to maximize the
diffusion rate of dopant into the nanoparticles, while avoiding boiling of the aqueous solvent.
Therefore, this temperature was used for this synthesis as well. ZnS:Cu nanoparticles of similar
dispersity to the ZnS:Mn nanoparticles were thus also prepared. The PL spectra of both of these
materials dispersed in water are shown in Figure 7.13. Both materials were readily dispersed in
200 250 300 350 400 450 500 550 600 650 700
Ph
oto
lum
ine
sc
en
t in
ten
sit
y,
no
rma
lize
d
Emission wavelength (nm)
ZnS:Mn emission ZnS:Cu emission
ZnS:Mn excitation ZnS:Cu excitation
1
0
118
water at high loadings (2.5 w/w%) and passed through a 0.2 µm filter with no reduction in PL
intensity. The use of TGA/NaOH, that proved to be successful for the dispersion of ZnS:Mn
nanoparticles in water, caused rapid precipitation of ZnS:Cu due to the immediate formation of
highly insoluble CuxS (Ksp = 1.27 10-36
). Hieronymas (2002) suggested the use of
polyphosphates to disperse ZnS nanoparticles; based on those proportions and methods, a
transparent suspension of ZnS:Cu was readily formed using a solution of 1% SHMP in water.
7.1.8 Characterization of ZnS:Mn, ZnS:Cu nanoparticles
Luminescent emission from doped nanoparticles is slightly different than that from pure
materials. While “true” quantum dots like pure CdS emit characteristic colours due to exciton
formation across their bandgap, doped materials emit colours associated with certain
characteristic energy levels, rather than just the conduction and valence band levels. In ZnS:Mn,
the emission results from the d-electron states in the dopant centres interacting with the s-p
electron states in the ZnS host, and the resulting transition of an electron in Mn’s 4T1 energetic
state to the 6A1 energetic state (Karlin 2005). The energy difference between these states is
approximately 2.1 eV, resulting in characteristic orange emission at 585 nm. These energetic
states both lie within the bandgap of the host material, ZnS. In ZnS:Cu, emission is attributed to
the transition of an excited electron from the conduction band to the lower energy 2T2 state,
above the valence band (Srivastava et al. 2010). A simplified schematic of these mechanisms is
shown in Figure 7.14. So, although the mechanism is similar to that for excitonic emission
across the bandgap, the energy levels between which carriers move are not the same – and
therefore, widening of the bandgap due to quantization will not have the same bearing on
emission colour as it does in undoped materials.
From Figure 7.13, PL spectra of the ZnS:Mn/water/TGA/NaOH and ZnS:Cu/water/SHMP
suspensions peaked at 591 nm and 495 nm, respectively – having bulk emission wavelengths of
585 and 520 nm, respectively. The large blue shift in emission from the ZnS:Cu nanoparticles is
attributed to their quantization. As the bandgap widens, the conduction band tends to shift while
the valence band remains essentially fixed (Robel et al. 2007). Since the mechanism of emission
119
in ZnS:Cu involves carrier transition from the conduction band to the 2T2 state, movement of the
conduction band widens this energetic gap and causes the observed blue-shift. Using the energy
quantization expression discussed in Section 3.2.11 for ZnS:Cu, the blue shift corresponded to a
bandgap of 4 eV (versus 3.7 for bulk ZnS), or a primary particle size of 2 nm. However, the
emission from ZnS:Mn was actually red-shifted. This has been hypothetically attributed to a
large amount of either surface states or electron-phonon coupling, in previous studies on ZnS:Mn
(Yang et al. 2005). With the polymeric capping agents reported above, where the surface states
were not as efficiently passivated, the red shift was more pronounced, as is discussed in Paper 7.
Figure 7.14. Mechanisms of light emission in ZnS:Mn and ZnS:Cu.
The small differences in peak locations for the ZnS:Mn particles dispersed in water vs.
NaOH/TGA vs. ink were attributed to improved passivation by TGA. No blue shift in the
characteristic blue emission (at ~430 nm) from S2- vacancies on the nanoparticles’ surface was
observed either. Indeed, since the Bohr exciton radius of ZnS:Mn is estimated to be
approximately 2.5 nm (Bhargava et al. 1994), the particle size of the ZnS:Mn in this study may
have been insufficiently small to observe quantum effects. The location of the PLE peak (at 338
Valence band
Conduction band
ZnS:Mn
Shallow trap states (Zn2+ vacancies)
e-
h+
Shallow trap states (S2- vacancies)
4T1 state
6A1 state
hv (2.7 eV)
e-
h+
e-
hv (2.1 eV)
ZnS:Cu
2T2 state
hv (2.5 eV)
(tunable)
e-
e-
h+
120
nm) was similar to that of bulk ZnS:Mn. A brief consideration of the UV-visible absorbance of
the nanoparticles yielded an estimated bandgap of 3.71 eV; again, similar to that of bulk
ZnS:Mn. Regardless, both the ZnS:Mn and ZnS:Cu particles remained suitable for jetting due to
their relatively small size. Although the particles had already passed through a filter and were
therefore suitable for jetting, their small size was ideal for clog-free jetting and the production of
uniform printed films.
As was mentioned above, PLE peaked at 348 nm for the ZnS:Mn nanoparticles and 322 for the
ZnS:Cu nanoparticles, representing a large Stokes shift. For the purpose of obtaining visible
emission using an interrogating UV lamp or laser, this large Stokes shift is ideal, as the emission
colour is very different from the relatively narrow range of UV wavelengths suitable to excite it,
and the excitation spectra are narrow, requiring very specific wavelengths. The blue emission at
~425 nm in the ZnS:Mn nanoparticles, attributed to S2-
vacancies rather than the Mn2+
4T1–
6A1
energy level transition (Yen 2004), was weak compared to the orange emission, suggesting a
well-passivated surface with S2-
vacancies occupied by 3-MPA or TGA molecules (Small et al.
2010). The extremely bright visible PL from the dispersions supported this suggestion, as PL is
quenched by unpassivated surface S2-
vacancies – such as that seen in Figure 7.10, where
emission from ZnS:Mn not capped with 3-MPA is visibly blue or violet.
Aside from the strong PL emission visible in the 3-MPA-capped nanoparticles, the successful
synthesis of the desired cubic ZnS was confirmed by XRD (Figure 7.15). Broadened peaks at 2
~ 28.5°, 47.5°, and 56.6°, characteristic of nanosized ZnS particles (Takahashi & Isobe 2005, Mu
et al. 2005) were clearly visible. According to the Scherrer Equation, the peak widths
corresponded to a mean crystallite size of 2.7 nm for ZnS:Mn and 2.3 nm for ZnS:Cu. This and
the PL spectra implied that the ZnS:Cu dispersion was actually close to a dispersion of individual
primary crystals, while the ZnS:Mn dispersion contained agglomerates of a few crystals.
121
Figure 7.15. XRD spectra of ZnS nanoparticles and bulk material. Inset: cubic crystal structure of luminescent ZnS.
TEM imaging of the 3-MPA-capped ZnS confirmed the primary crystallite size as approximately
2-3 nm (Figure 7.16). The particles were generally clumped into loose nanometer-sized
agglomerates on the TEM grid, where they had likely grouped during the drying process. The
dispersion of the particles in solution was obviously not visible using TEM, so DLS was used to
establish the degree of dispersion in solution – both aqueous and organic (Figure 7.16). DLS
suggested that the particles are present as small agglomerates in solution, although the
hydrodynamic radii of the dispersed particles were likely somewhat larger than the particles
themselves, due to the presence of the 3-MPA or 3-MPA/octylamine ligands. However, the
hydrodynamic particle size was still an order of magnitude smaller than the maximum allowable
size (200 nm) for inkjet filtration and printing. This was a marked contrast to the ZnS:Mn
nanoparticles capped with the other agents and those prepared using in-situ polymerized acrylic
20 30 40 50 60
Co
un
ts (
arb
itra
ry u
nit
s)
Bragg angle, 2 (°)
ZnS:Mn
ZnS:Cu
Bulk ZnS
28.5°
[111] lattice plane
47.5°
(220) lattice plane
56.6°
(311) lattice plane
122
acid (Angelo & Farnood 2011), of which barely, if any, passed through a 200 nm filter and those
that did pass through the filter were generally not emissive at the characteristic 595 nm peak for
ZnS:Mn. The dispersity and small size of the particles, even if somewhat agglomerated in the
ink, allowed for bright PL before and after filtration and jetting.
It may be inferred that the appearance of a white precipitate during synthesis of ZnS precludes
the redispersion and eventual jetting of ZnS nanoparticles, as only the 3-MPA-capped ZnS:Mn’s
reaction broth remained transparent during synthesis. Successful redispersion of ZnS for inkjet
printing where the broth contained precipitate was only reported by Small et al (2010) –
however, a mercaptan-derived acid was also used in that case as a stabilizer, although it was
added after synthesis, citrate being used during synthesis. Also, in that work, inks did not remain
well-dispersed over time, and ZnS:Cu inks succumbed to reaction with the thiol group in the
stabilizer (mercaptosuccinic acid, in that case). In this work, the dispersion remained after many
weeks of shelf life at ambient conditions; in fact, the DLS measurements shown in Figure 7.17
were actually carried out on ZnS dispersions that were several weeks old. So, through careful
study of the reaction conditions and reagents, reliably inkjet-printable ZnS-based luminescent
nanoparticles were produced for the first time.
Figure 7.16. TEM micrographs of ZnS:Mn nanoparticles, dried from dispersion in water. (a) Agglomerate, ~10-20 nm
diameter; (b) primary crystallites, 2-3 nm diameter.
20 nm
10 nm
(b)
(a)
123
Figure 7.17. DLS scans of ZnS:Mn and ZnS:Cu nanoparticles in water (ZnS:Mn) and toluene (ZnS:Cu).
7.2 Other materials
7.2.1 BaTiO3
The crystalline BaTiO3 nanopowder was best dispersed in an ethanol/MMA mixture using a
mixture of commercial dispersants, Disperbyk 111 and Surfynol CT324, yielding the particle
size distribution shown in the DLS results (Figure 7.18). Similar dispersion was achieved in
water with the same dispersants, but the immiscibility of water and MMA precluded its use.
Furthermore, the high surface tension of water would have necessitated the addition of a
surfactant to render the dispersion jettable. Using ethanol, with an already-low surface tension,
avoided the use of a surfactant, which may have adverse effects on dispersion and shelf life of a
given ink. According to the specifications of the nanopowder used, this solvent and dispersant
mixture produced a nearly monodisperse suspension of BaTiO3. Dispersions loaded with more
BaTiO3 nanopowder demonstrated sedimentation of some of the BaTiO3 over time; the 5 w/w%
ink was stable without sedimentation for several weeks. Although some sedimentation in the ink
did eventually occur, ultrasonication of the ink for 30 minutes was sufficient to redisperse it and
re-establish similar shelf life to that observed in the fresh ink.
1 10 100
Inte
ns
ity (
arb
itra
ry u
nit
s)
Hydrodynamic diameter (nm)
ZnS:Mn
ZnS:Cu
(a)
(c)
(b)
124
Figure 7.18. DLS-obtained particle size distribution of BaTiO3 (5 w/w%) dispersed with Disperbyk 111 and CT-324 in
ethanol/MMA and in water.
7.2.2 PEDOT:PSS & CNTs
Although inkjet printing of PEDOT:PSS has been widely reported, the PEDOT:PSS suspension
was still tested for particle size to make sure that it would pass through the printer without
significant loss of conductive material. Figure 7.19 shows the DLS-obtained PSD of the as-
received PEDOT:PSS, confirming that the PEDOT:PSS particles were sufficiently small to meet
the 200 nm cutoff for particle size. However, when compared to the other materials, the PSD
showed quite large particles, and some degree of clogging was expected when printing. Indeed,
it was noted when the inks were finally formulated that PEDOT:PSS did cause more clogging
than the other inks; however, the fact that PEDOT:PSS was a polymer rather than a rigid particle
meant that clogs could be more easily cleaned by re-dissolving the polymer with a suitable
solvent, alleviating this problem somewhat.
1 10 100
Inte
ns
ity (
arb
itra
ry u
nit
s)
Hydrodynamic diameter (nm)
in ethanol/MMA
in water
125
Figure 7.19. DLS-obtained particle size distribution of aqueous PEDOT:PSS suspension, diluted 100x.
A more pressing concern was the size and shape of the CNTs being added to the PEDOT:PSS
ink. DLS was not suitable for establishing their PSDs, because of their irregular hydrodynamic
radii. However, with both SW- and MWCNTs, 0.2 µm filtration removed a large proportion of
the nanotubes, due to the small pore size of the filters relative to the length of the nanotubes
(Table 7.2). Even more of the nanotubes were removed when passing the CNT solution through
the printhead. In the case of SWCNTs, fewer than 10% of the nanotubes passed through the
filter, although a greater number of the nanotubes were able to pass through the print nozzle
(without prior filtration) than were MWCNTs. The increased passage through the printhead
without filtration was likely a result of the relatively large nozzle size of 25 µm as compared to
the CNTs (5 µm length), and the low concentration of the CNTs. Passage of higher
concentrations of particles with sizes greater than 200 nm is compromised by crowding and
packing of the particles in the nozzles (Fujifilm-Dimatix 2006).
Table 7.2. Concentration of CNTs measured using UV-visible spectrometry.
CNT type Concentration (ppm)
Untreated Filtered only Filtered & jetted
Single-walled 32.0 2.8 2.5
Multi-walled 32.0 6.8 6.7
1 10 100 1000
Inte
ns
ity (
a.u
.)
Hydrodynamic diameter (nm)
126
Figure 7.20. STEM micrographs of MWCNTs (grey lines) (a) before filtration/jetting; (b) after filtration; (c) after
filtration/jetting. Images were obtained with the assistance of Ilya Gourevich of the Centre for Nanostructure Imaging,
Department of Chemistry, University of Toronto.
Although more MWCNTs were able to pass through the filter, more were also retained in the
jets. The likely explanation for these phenomena is the difference in shape between the nanotube
types. The MWCNTs are generally more rigid and stiff due to their multilayered structure
(Chang et al. 2005). The SWCNTs, because of their extremely small diameter, may bend or
“crumple” into different shapes, such as balls or bundles. In both cases, passage through the
filter was contingent upon orientation relative to the filter fibres; transverse orientation facilitated
the penetration of the CNTs through the fibre pores, which were significantly larger than the
CNTs’ diameter. The SWCNTs, being potentially bent and crumpled, were less likely to pass
through the fine pores of the filter. Conversely, the rigid, straight MWCNTs passed through the
filter when oriented correctly. However, it appeared from the UV-vis study that a larger
proportion of MWCNTs passed when undergoing both filtration and printing stages. The reason
for this phenomenon was hypothesized to be the presence of differing amounts of soot and metal
catalyst in the two CNT samples. The MWCNT sample used contained only 7.5% CNTs, with
the remainder being miscellaneous amorphous carbon soot or residual catalyst from CNT
production. The SWCNT sample contained 50-70% CNTs. With soot particles sufficiently
small to pass through the filter, the 500 nm absorbance of the carbon soot, being identical to that
of CNTs themselves (Zeng et al. 1999) was stronger in the MWCNT samples. This suggests that
this method of analyzing CNT passage may be somewhat limited by the CNT sample type, and
that purification of the nanotubes may be a necessary precursor for future studies. Regardless of
(a) (b) (c) 1 µm 1 µm 200 nm
127
the composition of the sample, these filtration and jetting steps removed >80% of the CNTs,
greatly limiting the practicality of incorporation of CNTs into a composite with PEDOT.
The apparently improved passage of MWCNTs over SWCNTs was more directly examined
using STEM imaging (Figure 7.20) of dried MWCNT solution. A drop of solution observed
under STEM without any filtration indicated a relatively large amount of MWCNTs distributed
across the sample grid (Figure 7.20a), along with SLS crystals and other particles, possibly of the
aforementioned soot or catalyst. Upon filtration of the solution, only small clumps of MWCNTs
were visible sporadically (Figure 7.20b). Finally, after filtration and passage through the
printhead, almost no MWCNTs were located, and only clumps of surfactant crystals or metal
catalyst/soot particles were visible (Figure 7.20c). X-ray fluorescence linescans of the visible
particles indicated that they were primarily composed of carbon, transition metals (particularly
nickel), and sodium from the SLS surfactant. Qualitatively, it appeared that the MWCNTs did
not pass through the printhead, and the material observed using UV-vis spectrometry was soot.
Figure 7.21. Zeta-potential of CNT/SLS solutions and 50/50 (v/v) PEDOT:PSS/CNT mixtures. (a) Unfiltered; (b)
filtered at 0.2 µm; (c) filtered at 0.2 µm and jetted through DMP2831 printhead. Error bars represent standard
deviation.
-70
-60
-50
-40
-30
-20
-10
0
Ze
ta-p
ote
nti
al (m
V)
bare CNTs (no SLS)
CNT threshold of stability
PEDOT:PSS stock
(a)
(b)
(c)
128
The SWCNTs were very small in diameter and therefore difficult to image using STEM; instead,
their passage was correlated to of the solutions, according to the theory that increasingly
negative values of corresponded to lower concentrations of the generally unstable (Kim & Ma
2011) aqueous CNT dispersions. Therefore, solutions containing larger numbers of CNTs were
hypothesized to have lower absolute values of ζ. This behaviour was observed in both the neat
CNT solutions and 50/50 mixtures of CNT solution and stock PEDOT:PSS suspension (Figure
7.21).
After filtration and jetting, | ζ | of the resulting solutions increased, indicating improved
dispersion stability. In the case of MWCNTs jetted after filtration, ζ decreased to almost 0,
suggesting that almost all of the MWCNTs had been removed, leaving behind only distilled
water. This result corresponded to that observed using STEM. This approach still did not
address the actual amount of SWCNTs removed by the filtration and jetting steps; in order to
quantitatively establish the passage of SWCNTs through the printer, a purified SWCNT sample
would be required.
Another related observation was the improved stability of PEDOT/CNT mixtures, likely due to
the presence of PSS- anion as a secondary dispersant for the CNTs. The stability did not
markedly differ for SWCNTs and MWCNTs in the PEDOT:PSS mixture; however, | ζ | was
higher for pure PEDOT:PSS stock than for that mixed with SWCNTs, suggesting that dispersion
was somewhat compromised by the presence of CNTs, as expected. Flocculation of the PEDOT
particles in similar PEDOT:PSS/CNT inks over time was observed in our previous work (Angelo
& Farnood, 2010a). Again, after filtration and jetting, the MWCNT-containing mixture
approached the | ζ | value measured for pure PEDOT:PSS suspension, indicating the removal of
nearly all of the MWCNTs during these processes. Considering these results, only SWCNTs
were retained as potential additives for conductivity improvement in the inks. As discussed in
the following sections, however, SWCNTs were similarly removed, although not to the same
extent, and so the use of smaller carbon species was considered as a possible means of delivering
more conductive material to the substrate during jetting.
129
7.2.3 Substrates
The use of non-porous, non-rough substrates for preparing electronic layers has been widely
studied; their impermeability, smoothness, and chemical inertness makes them ideal for the
deposition of thin layers of functional materials. Therefore, a study on the effects of a substrate
like glass, ITO, PET, acetate, or so forth on the functionality of the layers it supported was
considered redundant to this study. However, as was discussed in Section 3.5, porous substrates
– specifically paper – have attracted some interest for certain electronics; and, of course, paper
was used in the first stages of this work to support an ACPEL (Section 5.4). In this study, where
PEDOT:PSS/SWCNTs form the electrodes, these layers would be the only ones in direct contact
with the paper surface. When printing a PEDOT:PSS ink (whose exact formulation will be
discussed in Section 7.3) onto paper, the effect of ink absorption into the sheets was immediately
evident. On all of the sheets examined, 3 layers of PEDOT:PSS ink were printed at a drop
spacing of 25 µm. With a drop volume of 12 pL (shown in Figure 7.34) at S.G. ~ 1, the
grammage of wet ink deposited was 58 g/m2 on each substrate. The large volume of ink did
result in dimensional instability due to fibre swelling on unsized sheets. The grammage of dried
ink (containing 0.45% PEDOT:PSS) was therefore 0.25 g/m2.
Figure 7.22. Printed PEDOT:PSS conductivity differences between different commercial paper types.
High-yield paper
Inkjet paper
Cardstock Glossy paper
Photo- paper
Cellulose acetate
Co
nd
uc
tivit
y (
S/c
m)
10-3
10-2
10-1
100
130
Figure 7.23. Conductivity of printed PEDOT:PSS (single layer) as a function of added filler. Dashed lines (a) and (b)
represent the conductivity of handsheets containing retention aid but no TiO2: the amount of retention aid is equal to
that of 15% TiO2 handsheets for (a) and 30% TiO2 handsheets for (b). Error bars represent standard deviation.
Conductivity was affected greatly by both chemical and physical means, varying widely between
sheets of commercial paper (Figure 7.22). Indeed, the thickness of printed ink layers and their
bulk resistances were widely varied across all of the sheets, both commercially prepared and lab-
made (see APPENDIX L). Printing of the liquid PEDOT:PSS-SWCNT ink onto the porous,
absorbent sheets of paper naturally resulted in the penetration of the conductive species into the
paper. Certain sheets absorbed the majority of the ink, drying quickly; others retained an ink
film at the surface and dried slowly; and all manner of print effects (feathering, mottle) were
visible in the prints. Also, the penetration of ink into the paper may have induced a
“chromatographic” effect, where certain ink components were adsorbed or absorbed at different
spatial locations than others, resulting in a concentration gradient of conductive material. The
distribution of ink components in this fashion has been previous observed by Filenkova et al.
(2010). Depending on the location of the PEDOT and any conductivity enhancers, apparent
conductivity could have been improved or compromised by this distribution effect. According to
the cross-sectional imaging of printed PEDOT:PSS/SWCNT layers discussed previously (Angelo
0 2 4 6 8 10 12 14 16
Co
nd
uc
tivit
y (
S/c
m)
TiO2 filler content (w/w%)
Hardwood pulp
Softwood pulp
10-3
10-4
10-5
(a)
(b)
10-2
131
& Farnood, 2010c), PEDOT:PSS tended to be more concentrated near the surface of the sheets,
potentially improving conductivity versus a completely uniform “film”.
Conductivity of these samples was four (4) orders of magnitude smaller than the bulk
conductivity of PEDOT (550 S/cm as reported by Crispin et al. 2003), and an order of magnitude
smaller than that on cellulose acetate. Despite slight differences in ink absorption rate and
surface properties (Table 6.3), there was little if any difference between SW and HW handsheets
in terms of conductivity. However, it was observed that the physical differences between
unfurnished sheets (i.e. sheets containing no sizing or fixation agent) and furnished sheets (which
contained either or both of these materials) affected conductivity. Figure 7.23 shows that with
the addition of TiO2 and its accompanying filler retention aid (Polymin SK), printed conductivity
on handsheets decreased by nearly two orders of magnitude. However, this decline was
primarily due to the presence of retention aid, as the control sheets – containing appropriate
amounts of Polymin SK but no TiO2 (dashed lines) – performed similarly to those with both
filler and retention aid. It is also worth noting that conductivity was slightly higher in sheets
containing filler over those with only retention aid because of the adsorption of Polymin SK to
TiO2, which reduced the amount of Polymin SK reacting or interacting with the PSS– stabilizer,
as discussed below.
The above observations suggest that an interaction between the Polymin SK and the
PEDOT:PSS-SWCNT ink took place resulting in compromised electrical performance. The
most likely interaction is the neutralization of the excess ionic charge provided by the PSS–
counterion in the ink, a function for which polyethyleneimide (PEI), an active component of this
retention aid, is known (Neimo 1999). Furthermore, due to this strong interaction with the
PEDOT:PSS complex, non-conductive PEI was likely incorporated within the conductive layer,
dramatically reducing conductivity of the layer as a whole. This type of interaction has been
previously exploited by Lin et al. (2007), modulating PEDOT:PSS conductivity by five orders of
magnitude at a ratio of PEDOT:PSS to PEI of 1:1. However, in the case of printed handsheets,
PEI was dispersed throughout the sheet in a relatively low concentration, and hence had a less
drastic effect on conductivity. These results suggest that the effect of the filler itself on
conductivity was evidently negligible compared to PEDOT interaction with the PEI retention aid.
10-2
132
ToF-SIMS mapping of PEDOT, TiO2, and PEI (Figure 7.24) confirms that the presence of
isolated clusters of TiO2 did not disturb the spatial distribution of PEDOT in the handsheets. This
observation implies that these clusters of TiO2 were also saturated with the conductive ink, and
the filler species did not interrupt the conductive path. Hence, in the absence of retention aid,
finely divided filler particles are expected to have little or no effect on the conductivity of paper-
based printed PEDOT-SWCNT films. However, with the addition of a strongly positively-
charged retention aid, filler can increase conductivity by binding to the retention aid, thereby
diminishing its interaction with the PSS- counterion. ToF-SIMS images of printed handsheets
also revealed that PEDOT and PSS were located in exactly the same regions but were less
concentrated where PEI was localized (Figure 7.24). The resulting non-uniformity of the
conductive layer would have an adverse effect on the conductivity of filled handsheets. It
appears that a very small amount of PEI is actually interacting with the TiO2, and the majority
resides in or on the fibres, where it can readily interact with the PSS–. In fact, ToF-SIMS peaks
for Ti4+
-PEI (m/z = 91) and TiO2-PEI (m/z = 123) were very weak, suggesting that little PEI was
adsorbed or bonded to the surface of the filler, leaving the rest to freely interact with PSS– in or
on the fibres. These observations are consistent with the hypothesis that PEI-PSS– interaction
was primarily responsible for increased electrical resistance.
Figure 7.24. ToF-SIMS maps of relative distribution of PEDOT, PSS, PEI, and TiO2 on SW/30% TiO2 sheet.
PEDOT PSS PEI TiO2
20 µm
133
Figure 7.25. Estimated conductivity of PEDOT-SWCNT ink on SW fibres: (a) unfurnished (b) PDADMAC fixation
agent. Error bars represent standard deviation.
The addition of a cationic ink fixation agent, PDADMAC, decreased conductivity of the printed
handsheets, and this effect was more pronounced at higher filler addition levels (Figure 7.25).
With a similar effect to the retention aid, this was likely caused by the interaction of the cationic
PDADMAC with the PSS– counterion and the introduction of non-conductive PDADMAC into
the PEDOT film. However, the lower concentration of PDADMAC versus PEI resulted in a
smaller decrease in estimated conductivity. In addition, the highly-charged, cationic nature of
PDADMAC may also have caused agglomeration of the PEDOT or SWCNTs after printing due
to destabilization of these suspensions through interaction with the PSS- and lauryl sulfate
dispersants. Such an interaction would create a non-uniformly distributed conductive ink and
hence a lower sheet conductivity. The non-uniform distribution and the apparent agglomeration
of both PEDOT and PSS in the ToF-SIMS maps of handsheets containing PDADMAC
supported this theory (Figure 7.26). With increasing filler levels, the detrimental effect of
PDADMAC also became more pronounced as the contact angle decreased. A lower contact
angle implied better penetration of PEDOT:PSS into the sheet, and hence reduced conductivity.
0 15 30
Co
nd
uc
tivit
y (
S/c
m)
TiO2 filler added during sheet forming (w/w%)
(a) (b)
10-3
10-4
10-5
10-6
10-2
134
Figure 7.26. Relative distribution of PEDOT:PSS and PDADMAC in HW sheet (30% TiO2, no sizing).
Figure 7.27. Estimated conductivity of PEDOT-SWCNT printed ink on HW fibres: (a) unfurnished; (b) with internal
AKD sizing; (c) with internal AKD sizing and fixation agent. Error bars represent standard deviation.
The most notable effect on conductivity aside from that of PEI was that of internal sizing (Figure
7.27). In every case, the addition of sizing agent resulted in an increase in conductivity. As
shown in Table 6.3, internal sizing decreased ink spreading and absorption by increasing the
contact angle of the ink. Therefore, a more uniform ink layer (containing fewer non-conductive
fibres/filler particles) with higher connectivity and conductivity was obtained. Moreover, the ink
absorption rate was reduced by several orders of magnitude in the sized sheets, allowing a longer
time for the ink to rest on the surface during drying of the PEDOT-SWCNT ink. Cross-sectional
0 15 30
Co
nd
uc
tivit
y (
S/c
m)
TiO2 filler added during sheet forming (w/w%)
(a) (b)
(c)
10-2
10-3
10-4
10-5
PEDOT PSS PDADMAC Total ion
20 µm
135
images confirmed the presence of a relatively thin PEDOT-SWCNT layer on the sized sheets.
Figure 7.28 shows the cross-sections of printed SW handsheets (30% filler) for an internally
AKD sized handsheet and an unsized handsheet. In the case of the sized sheet, the PEDOT-
SWCNT ink (blue-coloured) is concentrated near the surface of the sample while in the latter
case ink is distributed throughout the sheet thickness.
Figure 7.28. Cross-sections of printed SW handsheets (30% filler) showing PEDOT:PSS ink penetration.
The positive effect of AKD sizing appeared to be more pronounced for handsheets containing
retention aid and/or fixation agent. In other words, AKD sizing appeared to largely eliminate the
adverse effects of PDADMAC and PEI that are distributed throughout the sheet, by reducing the
amount of contact of the conductive ink with these molecules. Because of this correlation
between ink “holdout” in the printed sheets and the performance of the conductive layers, paper
with minimal porosity and high hydrophobicity (for aqueous inks, at least) should be considered
ideal for electrically conductive paper production. The performance of such materials might be
improved by the deposition of larger amounts of ink, as the initial print passes would fill the
pores and become absorbed into the fibres and filler clusters, providing a less absorbent surface
for subsequent print passes. However, this approach would of course consume more ink and
require more processing time to achieve similar performance to appropriately sized sheets. The
significant improvement in conductivity across all the sheets resulting from internal sizing is
likely more easily and economically achieved by sheet pretreatment.
50 µm
inked region
inked region
AKD sized unsized
136
Figure 7.29. Greycale ToF-SIMS images of PEDOT distribution on unfilled (0% TiO2) sheets with no PDADMAC.
% coverage by PEDOT was estimated using grey-level thresholding.
ToF-SIMS mapping of the ink components on the sized handsheets further confirmed the
improved ink retention on the sheet surface. It is evident in Figure 7.29 that internal sizing
allowed the PEDOT ink to coat the surface of the sheets to a greater degree, rather than being
absorbed into them, resulting in a larger proportion of interconnected PEDOT-SWCNT regions.
However, there was still not a completely contiguous PEDOT/PSS/SWCNT layer on the surface
of the sized sheets, and fragments of cellulosic materials and filler were clearly visible in the
spectral maps. As is also shown in Figure 7.28, even sized sheets absorbed the ink to a certain
degree. However, it is worth noting that the intensity of the PEDOT signal in the ToF-SIMS
maps of sized sheets was significantly higher indicating a higher concentration of conductive ink
on the surface of the sized samples.
Another variable that might assist in improving conductivity, in the same vein, is sheet
smoothness. In fact, a brief examination of printed PEDOT-SWCNT layers on rough,
uncalendered handsheets revealed that their conductivity was close to those of the base sheets
themselves. The universal calendering of the handsheets used in these experiments provided a
high degree of smoothness, improving the ink holdout and conductivity of printed samples.
So, if paper were to be considered as a substrate for electronics deposition, the type of paper
itself would play a major role in the functionality of the device. Because many planar electronic
devices involve the deposition of interconnects and electrodes as a base layer, the conductive
ink’s interaction with the surface would determine its conductivity and the device’s performance.
% covered by PEDOT (a) sized HW: 45.9% (b) unsized HW: 25.6% (c) sized SW: 40.2% (d) unsized SW: 25.1%
(a) (b) (c) (d)
20 µm
137
The use of paper with good ink holdout (i.e. low absorption), chemically compatible additives,
and similarly-charged species – avoiding cationic materials – is critical to maximizing
conductivity in an electrode ink.
7.3 Ink formulation
7.3.1 Conductive ink
The initial ink formulation was based on a PEDOT:PSS solution described by Bronzyck (2003)
used for spin-coating onto paper, containing a mixture of glycerol (20%), isopropanol (12%), and
PEDOT:PSS suspension (68%). Because of DMSO’s beneficial effect on conductivity in
PEDOT:PSS films, this formulation was refined to contain DMSO, yielding a similar ink to that
presented by Garnett & Ginley (2005). Given that the viscosity of the DMSO-containing ink
was >12 cP at room temperature, water was used to dilute the PEDOT:PSS to a suitable viscosity
before treating it with other additives (IPA, DMSO, etc). To clarify the following discussion,
“water content” in the inks refers to the amount of water added to the PEDOT:PSS suspension,
not the total amount of water in the ink (which included the water contained in the original
PEDOT:PSS suspension).
It was observed during preliminary testing that moderate-viscosity (i.e. below 10 cP)
PEDOT:PSS inks could be jetted for extended periods of time without clogging, whereas jetting
of PEDOT:PSS inks with viscosity greater than 10 cP, while stable for a short period, resulted in
rapid clogging of the printer nozzles. Also, the higher viscosity inks displayed an increased
amount of splattering onto the substrate due to the formation of long “tails” on the droplets that
formed satellite droplets, due to increased viscoelastic forces. This issue could not be resolved
by increasing the drive voltage to its maximum (40 V) nor by increasing the printhead
temperature (which aggravated the issue by boiling off the solvent prematurely and causing
curing of the ink in the nozzle and on the nozzle plate). Therefore, considering these factors, a
more moderate viscosity in the range of 5-6 cP was targeted. This range also corresponded to the
suggested viscosity (5 cP) of piezoelectric inkjet inks found in literature (Di Risio & Yan 2007,
Dong et al. 2006, Calvert 2001).
138
Figure 7.30. Effect of added water on PEDOT:PSS suspension viscosity. Error bars represent standard deviation.
Figure 7.30 shows the effect of added water on viscosity, indicating that a water content of
approximately 35 w/w% provided a suitable viscosity (~4.75 cP). The choice of a slightly
smaller value of viscosity was prompted by the necessity for the subsequent addition of the more
highly viscous DMSO and glycerol co-solvents. Adding glycerol as a co-solvent/humectant,
while expected to change viscosity, did not cause any significant variability, even at the highest
typical humectant loading of 20 w/w% (Magdassi 2010). The viscosity of the undiluted ink
containing 20 w/w% glycerol was 12.1 cP, versus 11.4 cP for 0% glycerol. At 17 w/w%
glycerol, the viscosity was still 11.9 cP (i.e. in the jettable range). In order to avoid clogging of
the piezoelectric nozzles with dried and crusted ink, 17 w/w% glycerol was added to the ink.
The ink containing both 35 w/w% water and 17 w/w% glycerol (balance PEDOT:PSS
suspension) had a viscosity of 4.9 cP.
Using the glycerol-treated ink, DMSO addition was studied. According to Figure 7.31,
increasing the DMSO content did not demonstrate any trend in terms of viscosity variation. The
effect of DMSO addition on the electrical resistance was also examined by cast coating 0.25 mL
of the ink on cellulose acetate with a calibrated pipette. It was noticed that adding DMSO up to
0
2
4
6
8
10
12
14
16
18
0 5 10 15 20 25 30 35 40 45 50
Vis
co
sit
y (
cP
)
Added water content (w/w%)
139
10 w/w% reduced electrical resistance in cast-coated conductive films PEDOT:PSS while
beyond 10 w/w% DMSO, resistance remained nearly unchanged (Figure 7.32). Therefore, a
DMSO content of 10 w/w% was chosen, resulting in a mixture viscosity of 4.8 cP.
Figure 7.31. Effect of added DMSO on PEDOT:PSS/glycerol mixture’s viscosity, at different water loadings
Figure 7.32. 2-point resistance of cast-coated PEDOT:PSS/glycerol/water (48%/17%/35%) films with different DMSO
contents (displacing water in the solution).
0
2
4
6
8
10
12
14
0 10 20 30 40 50
Vis
co
sit
y (
cP
)
Added water (w/w%)
0 w/w% DMSO
1 w/w% DMSO
5 w/w% DMSO
10 w/w% DMSO
15 w/w% DMSO
20 w/w% DMSO
0 2 4 6 8 10 12 14 16 18 20
Res
ista
nc
e (
Ω)
DMSO content (w/w%)
140
To optimize the surface tension, the surfactants discussed in Table 6.4 were then added to the ink
(displacing water) to reduce its surface tension to a jettable level. Because of the addition of
DMSO/glycerol, even the ink containing no surfactant had acceptable surface tension for jetting
(33.1 mN/m). Nevertheless, as the PEDOT:PSS ink was needed to wet the surface of the
hydrophobic polymer PVK in some of the device structures, further reduction of surface tension
to 30 mN/m was desirable. Surface tension generally decreased up to a certain critical loading of
surfactant, and then either plateaued (Zonyl FS-300, ZetaSperse 3700) or began to increase again
(Triton X-100, Igepal CA-720, SLS) as the surfactant likely detached from the particles’ surface
and formed double-layer micelles of pure surfactant in solution (Spasic & Hsu 2006). Each
surfactant except ZetaSperse 3700 was capable of reducing the surface tension to 30 mN/m (or
significantly lower), as is shown in Figure 7.33. The addition of surfactant also affected
viscosity. Certain surfactants were observed to decrease viscosity markedly, SLS being the best
example. The effect on viscosity was expected, as each surfactant was likely responsible for a
different thinning effect (Jansen et al. 2001). In no case did the viscosity fall so low that the ink
was no longer jettable, although the surface tension dropped below the jettable range, particularly
for Zonyl FS-300 (expected for a fluorosurfactant).
The measured maximum particle sizes of each ink varied with the surfactant type as well, but
were generally close to the cutoff of 200 nm. This was simply by virtue of selecting
PEDOT:PSS as the starting material. Also, the addition of surfactants served to improve
dispersion stability (| | > 40 mV) in most cases, or at least maintain it above the threshold of
stability. Igepal CA-720 appeared to reduce | |, by a mechanism that was not well-understood,
as it is a non-ionic molecule with no positively-charged regions. It is possible that the PSS- was
displaced by the OH-bearing end of the molecule to some degree, and the Igepal served as a
poorer stabilizer for the PEDOT micelles. A summary of the dispersion and particle size
properties, as well as the rheological properties of the inks (with surfactants loaded to their
CMCs), is given in Table 7.3. Each ink, again, contained 17% glycerol, 35% water (minus the
amount replaced by surfactant), 10% DMSO, and the balance as PEDOT:PSS suspension.
141
Figure 7.33. Surface tension and viscosity in PEDOT:PSS inks with different surfactant types. Error bars represent
standard deviation.
20
25
30
35
40
0 0.25 0.5 0.75 1 1.25 1.5 1.75 2
Su
rfa
ce
te
ns
ion
(m
N/m
)
Surfactant content (w/w%)
Triton X-100 Zonyl FS-300 Igepal CA-720 Zetasperse 3700 SLS
0
1
2
3
4
5
6
7
8
0 0.25 0.5 0.75 1 1.25 1.5 1.75 2
Vis
co
sit
y (
cP
)
Surfactant content (w/w%)
142
Table 7.3. PEDOT:PSS inks’ fluid properties.
Ink (cP) (mN/m) dp, max (nm) Z-1
En ζ (mV)
No surfactant 4.1 33.1 212 7.0 ~0 -68.8
Igepal CA-720 (0.2%) 4.9 26.3 235 5.2 ~0 -55.3
SLS (0.5%) 2.2 30.2 200 12.5 ~0 -67.5
Triton X-100 (0.1%) 5.7 30.3 236 4.8 ~0 -67.9
Zonyl FS-300 (0.02%) 4.3 22.5 82 5.5 ~0 -63.8
ZetaSperse 3700 (0.2%) 3.7 31.4 178 7.6 ~0 -65.8
Because of the wide variation in surface tension between inks, inks with identical formulations
except for the surfactant type exhibited markedly different jettability and film formation. Images
of the droplet and film formation can be seen in Figure 7.34 and Figure 7.35, respectively. For
reference, an example of good drop formation is that of the SLS-bearing ink; poor drop
formation, with nozzle plate splattering, is visible with Zonyl-bearing ink. Good film formation
is visible in the Triton and Igepal-bearing inks, whereas the ink containing no surfactant formed
a poor film.
Figure 7.34. PEDOT:PSS ink droplet formation during ejection from DMP2831 cartridge nozzles.
Figure 7.34 shows that the drop size was similar for various inks, with an average diameter of
about 28 µm (or droplet volume of 12 pL), close to the expected 10 pL delivered by the
(a) (b) (b)
(a)
(b)
(d)
(e)
200 µm
No surfactant
Igepal CA-720
SLS
Triton X-100
Zonyl FS-300
ZetaSperse 3700
143
DMP2831 cartridge. Within the range of this study, viscosity had less of a bearing on drop
formation and jettability, although it did affect Z-1
, by definition (which still remained in the
jettable range). Surface tension had much greater bearing on jetting performance. The inks with
the lowest surface tension, containing Zonyl FS-300 and Igepal CA-720, demonstrated poorer
drop formation than the other inks. With these two inks, the formation of small droplets which
detached from the nozzle rapidly and dropped satellites onto the nozzle plate is clearly visible.
Figure 7.35. Printed patterns of PEDOT:PSS inks on acetate. Films consist of a single printed layer of PEDOT:PSS
ink (25 µm drop spacing). Images were adjusted for colour and contrast.
Each of these inks was jetted onto cellulose acetate. The quality of the printed films varied
widely among samples. Triton- and SLS-bearing inks, both having a surface tension around 30
mN/m, formed contiguous films while Igepal CA-720, with a slightly lower surface tension,
created a more uniform but “striated” film, with visible printed lines and unprinted or only
lightly inked areas. These results suggest that surface tension plays a major role in printed film
quality with a higher surface tension better tolerated than a lower one. Poor quality films in
No surfactant Igepal CA-720 SLS
Triton X-100 Zonyl FS-300 ZetaSperse 3700
2 mm
2 mm
144
general exhibited either such striations or “island” formation. Striations in the film,
characteristic of the failure to deliver ink to the entire printed area either due to clogging or poor
drop formation, were visible in the Zonyl and Igepal-containing ink films. The lower surface
tension may have produced smaller droplets which were not spaced closely enough. In the case
of Zonyl FS-300, the entire printed pattern failed to be delivered to the substrate, which is typical
when nozzles clog during printing, or, more likely, when the nozzle plate becomes wet with fluid
through which drops cannot penetrate. These results also suggest that the use of potent
surfactants such as fluorosurfactants is unsuited to this application. A similar effect (poor drop
formation and print uniformity) was observed with a ZnS:Mn ink containing Zonyl FS-300
(Angelo & Farnood 2011). However, a higher surface tension produced isolated “islands” and
unwetted regions on the substrate. Without the capability to wet the substrate, drops tended to
merge into thick puddles, resulting in their characteristic “islands” upon drying.
The print quality issues resulting from the use of different surfactants had a major bearing on
conductivity of printed film. Regardless of surfactant type, the conductivity of spin-coated – not
printed – films on glass were fairly similar at just over 1000 S/cm, with all values within the
same statistical envelope, indicating that the surfactants themselves did not have any significant
effect on the conductivity. Compared to spin- coated samples, conductivity of inkjet-printed
films was 3-5 orders of magnitude lower and varied widely from 0.016 S/cm for Zonyl-bearing
ink to 1.03 S/cm for SLS-bearing ink, when resistance was measured parallel to the print
direction (Figure 7.36). As would be expected, printed films that exhibited striated structures
had lower conductivity than those with more uniform print quality. The inks containing no
surfactant and ZetaSperse, which both were non-uniform in terms of print quality as well (the
“islands” previously mentioned due to poor wetting), still had higher conductivity than the
striated films, where definite gaps between printed lines were visible. The presence of large
patches or blobs of PEDOT:PSS in these inks likely served as bridges between adjacent printed
lines. Measuring conductivity in the direction normal to print (i.e. across rather than along the
print lines) confirmed this hypothesis, as the striated inks (Igepal and Zonyl) effectively were not
conductive in this direction. In every case, conductivity was reduced by measuring across the
samples, suggesting that the printed lines merged poorly regardless of the surfactant. It was also
145
thought that viscosity may play an important role in determining the conductivity of inkjet-
printed films, and in the case of conductive films, a lower viscosity might actually be beneficial.
The lowest-viscosity ink containing SLS, where scalloping and spreading across the printed lines
occurred (Figure 7.36), achieved the highest conductivity. The spread of the ink more rapidly
due to less viscoelastic resistance to flow was expected to help fill in some of the inter-line gaps
and produce better conductivity. The use of a strong anionic surfactant has also been linked to
improved conductivity in PEDOT:PSS films due to “unzipping” of the PEDOT-PSS chain
through surfactant exchange with PSS- and lengthening of the PEDOT chain as a result (Fan et
al. 2008). This might explain the improved conductivity seen in both SLS and ZetaSperse-
bearing inks, when printed, but there was no similar improvement observed in spin-coated films,
casting doubt on the likelihood of this mechanism. Furthermore, the large improvement in
conductivity provided by DMSO likely overshadowed any benefit provided by SLS or
ZetaSperse.
Figure 7.36. 2-point estimated conductivity of printed PEDOT:PSS inks on cellulose acetate (single printed layer, 25
µm drop spacing).
No surfactant Igepal CA720 SLS Triton X-100 Zonyl FS-300 ZetaSperse 3700
Es
tim
ate
d c
on
du
cti
vit
y (
S/c
m)
Printed (parallel direction) Printed (normal direction) Spin-coated
10-3
10-2
103
104
102
101
100
10-1
10-4
146
Nevertheless, the unexpectedly high conductivity of the anionic surfactant-bearing films was
examined more closely, to determine if a chemical or conformational change was occurring in
the films. Raman spectrometry indicated that the chemical structure remained the same in the
presence of most of the surfactants except ZetaSperse (Figure 7.37). All spectra were red-shifted
from that of stock PEDOT:PSS, due to the presence of DMSO as a conductivity enhancer,
representing the benzoid-to-quinoid transition of the PEDOT structure (Figure 7.37 inset). The
quinoid structure has higher electron mobility (Ouyang et al. 2004, 2005), due to the loss of
aromaticity (and stabilization energy) in the ring, which in turn reduces the energetic bandgap
(Cheng et al. 2009) and forms a conjugated structure in the polymer chain. A smaller bandgap is
synonymous with improved carrier mobility. A peak at ~1460 cm-1
of Raman shift is
representative of the benzoid thiophene ring, whereas the red-shifted peak at 1425 cm-1
represents the conjugated quinoid structure. The distinct shoulder peak at 1460 cm-1
(most easily
visible in the PI spectrum), as well as the broadness of the other peaks (besides PZS, which will
be discussed further below), suggests that the addition of 10% DMSO resulted in the transition of
a certain fraction of the ethylenedioxythiophene repeating units. The shoulder represents
remaining benzoid units. As the intensity of the red-shifted peak at 1425 cm-1
increased, this
shoulder became less pronounced.
It was notable that the slope of the ZetaSperse 3700-treated film was the greatest, with a very
minor shoulder peak present. The shape of this peak, being very similar to that of untreated
PEDOT:PSS, suggested that all of the benzoid structures had been converted to quinoid
structures in this ink. Either the functional molecule in ZetaSperse 3700 itself (which was
proprietary) or the solvent system bearing it may have furthered the complete transition to the
quinoid structure. Since ZetaSperse is water/propylene glycol-borne, it may have simply been
the glycol which caused the transition, as has been reported previously (Kim et al. 2003).
Another notable point was the further red shift in the Raman spectrum for this ink compared to
the other inks. This result suggested that the addition of ZetaSperse 3700 may have caused a
further conformational change in the PEDOT molecule. Examination of this surfactant by FTIR
indicated the presence of primary and secondary amine groups in the ZetaSperse 3700
formulation. Therefore, it is possible that the peak in the PZS spectrum at 1415 cm-1
could
147
represent the deformational vibtrations attributed to primary methylated amine groups, masking
the PEDOT:PSS peak.
Figure 7.37. Raman spectra (excitation wavelength = 785 nm) of PEDOT:PSS inks. The inset shows the benzoid-to-
quinoid structural transition.
Based on the above consideration, the addition of either 0.5% SLS or 0.2% ZetaSperse 3700 was
found to result in optimum conductivity of the pinted PEDOT film. The much superior
conductivity and print quality offered by the SLS-bearing ink versus the ZetaSperse-bearing ink
made it the more attractive option. Foaming was an issue with the use of SLS, which produced a
large volume of foam upon agitation – a problematic situation for piezoelectric jets. However, a
small amount (0.05%) of defoamer (Surfynol DF-110D) was sufficient to reduce foam in the
PEDOT:PSS ink.
The consideration of CNT addition also was made to further improve printed conductivity to a
value useful for electronics fabrication. Because the CNTs were suspended in water using SLS,
1275 1325 1375 1425 1475 1525 Raman shift (cm-1)
No surfactant Igepal CA-720 SLS Triton X-100 Zonyl FS-300 ZetaSperse 3700 PEDOT:PSS
148
the use of the same surfactant in the ink was ideal. At this point, treatment of the ink with CNT
solution (replacing water) was considered. Because MWCNTs appeared not to pass through the
filter/printhead in any significant amount, SWCNTs were used. Figure 7.38 shows that there was
a gradual increase in conductivity with SWCNT inclusion in printed PEDOT:PSS ink, although
this increase was not consistent at higher loadings. Conductivity increased significantly with the
addition of a smaller fraction of SWCNTs (i.e. from 0-2 w/w%) and then varied irregularly,
although it remained at ~1 S/cm at 2% CNTs. It was evident from the CNT passage studies
discussed in the previous section that only a very small fraction of CNTs passed through the
printer to the substrate. However, as more nanotubes were added, the nanotubes were delivered
to the substrate in larger amounts, indicated by increasing conductivity. Even in pure SWCNT
stock, however, less than 8% of the CNTs (i.e. 3 × 10-2
w/w%) remained after filtration, which
likely included some soot and catalyst materials. This implied that the highly CNT-loaded
PEDOT-SWCNT inks were limited to approximately this number of nanotubes passing, so even
if all of the SWCNTs in the 8% CNT ink passed through the printer, the 9 w/w% and 10 w/w%
inks would be limited to 3 × 10-2
w/w% CNTs passing to the substrate. This maximum number
of CNTs passing through the printer provided a limit above which the addition of CNTs served
no further purpose, hence the plateau effect observed at 5-6% SWCNT solution. Also, at higher
CNT concentrations, the likelihood of interaction and agglomeration of the CNTs is increased,
reducing the passage of CNTs through the nozzles and causing nozzle clogging. This may offer
another explanation for the increased variability of the conductivity data at higher CNT loadings.
The shape of the curve describing conductivity dependence on SWCNT content on acetate also
suggested a logarithmic or power-law relationship between CNT concentration and probability of
passing through the printer before the plateau conductivity value.
The mechanism of conductivity enhancement might be attributed to the presence of more-highly
conductive SWCNTs within the PEDOT:PSS matrix . As was described by Kerner (1956), the
incorporation of a highly conductive species in a composite matrix may enhance the conductivity
of the composite. However, the extremely small proportion of CNTs relative to PEDOT could
not reasonably account for the measured increase in the overall conductivity of the composite,
where conductive “filler” materials generally have to be in contact with one another to allow
149
charge transport (Kerner 1956, Kim & Ma 2011). As an alternative hypothesis, the physical
conformation of the CNTs may assist in “bridging” non-conductive regions – i.e. that their length
would connect nearby or adjacent printed regions of PEDOT and thereby decrease resistive
losses. However, because the probability of any SWCNTs passing through the printer was
extremely low after filtration, the “bridging” hypothesis was unlikely. A more realistic
mechanism was increased doping of the PEDOT backbone by anionic species present in the CNT
soot/catalyst. The presence of additional dopant (supplementing the effect of PSS-) would
improve carrier mobility in the LUMO of the conjugated PEDOT, enhancing conductivity.
Figure 7.38. Conductivity of printed PEDOT:PSS-SWCNT ink (SLS surfactant) at varying SWCNT loadings (single
printed layer, 25 µm drop spacing).
Firstly, C60 and S-SWCNTs, which did not have a high enough aspect ratio to physically bridge
gaps between conductive regions, were jetted in the inks. Both C60, with a diameter of 0.7 nm
(Yeo et al. 2009) and chemically-shortened SWCNTs, with an average length <60 nm (Chen et
al. 2006), were sufficiently small to pass through the filters and printheads. Therefore, it was
expected that if the higher conductivity of carbon relative to PEDOT was improving the
composite’s conductivity, not only would inks containing these materials demonstrate similar
trends in conductivity to those containing SWCNT solution, but that these trends would be more
0 1 2 3 4 5 6 7 8 9 10
Co
nd
uc
tivit
y (
S/c
m)
SWCNT solution content (w/w%)
Acetate
Photo paper
10-1
100
101
10-2
150
pronounced due to their wholesale passage through the printer. However, this was not the case,
and neither C60 nor S-SWCNTs had any significant effect on conductivity when printed at 25 µm
drop spacing for a single print pass (Figure 7.39). The x-axis of this figure shows the amount of
C-species solution that was added to the ink, where the solutions contained 0.04 w/w% of their
respective C-species. S-SWCNTs, in particular, which were chemically shortened from the same
batch of SWCNTs used in the ink, did not improve conductivity at all. The S-SWCNTs had
identical properties to the SWCNTs, with the only difference being length; this directly
supported the idea that SWCNTs were “bridging” materials. There was a small improvement in
conductivity due to C60 addition, and a similar trend of irregular increase above 5 w/w% C60
solution. This suggested that perhaps a component of the phenomenon of increased conductivity
in SWCNT-containing PEDOT layers was simply the presence of conductive carbon. The lack
of a similar effect in S-SWCNTs was attributed to the chemical process by which they were
shortened, which involved the use of strong mineral acids; dopant species may have been
consumed during this reaction or removed by rinsing.
Figure 7.39. Conductivity of inkjet-printed PEDOT:PSS-carbon composites on acetate (single printed layer, 25 µm
drop spacing).
0
0.2
0.4
0.6
0.8
1
1.2
1.4
1.6
0 1 2 3 4 5 6 7 8 9 10
Co
nd
uc
tivit
y (
S/c
m)
C species sol'n (w/w%)
SWCNTs
US-SWCNTs
C60
SWCNTs
S-SWCNTs
C60
151
Spatial location of the SWCNTs was not feasible using SEM, partially due to their small size, but
more importantly because of their incorporation into an organic matrix and resulting lack of
contrast. Again, ToF-SIMS mapping confirmed the physical behaviour of the conductive ink on
the various substrates in terms of ink penetration and retention (Figure 7.40). The shallow
penetration depth of the ToF-SIMS primary ions indicated significant retention on the acetate
and photo paper substrates even at the very surface of their respective coating layers.
Figure 7.40. ToF-SIMS maps of PEDOT and substrate component distribution, single printed layer. Al3+
comprises
the majority of the coating layer of the photo-paper (as alumina).
ToF-SIMS images show that on both acetate and photo-paper the printed PEDOT-SWCNT films
were not uniform. In the case of the photo-paper, the PEDOT-SWCNTs and surface coatings
formed composite layers with micro- or nano-scale non-uniformities in the form of finely divided
coating pigments and binder. In the cases where non-uniformity was most pronounced on a
spatial level, with permeable substrates, conductivity was the lowest and the least improved by
the addition of SWCNTs. As PEDOT-laden regions moved closer together, in the coated sheets,
conductivity increased, as did the effect of CNT addition. However, there was no visible means
by which the “bridging” of conductive regions might have improved conductivity, and the
practical invisibility of CNTs using ToF-SIMS examination meant that this mechanism could not
PEDOT Acetate
Al3+
PEDOT PSS
PSS Total ion
Total ion
Photo-paper
Cellulose acetate
40 µm
152
be reliably confirmed. Therefore, increased PEDOT doping by impurities in the CNTs remained
the most sensible mechanism of conductivity enhancement. Further examination of the means of
CNT enhancement of conductivity was discussed in Paper 4; in short, however, the introduction
of CNTs into inkjet inks is of limited utility, due to their nearly wholesale removal from the ink
during the printing process.
In any case, the slight beneficial effect of SWCNT/soot addition on conductivity was evident, to
a point. Referring again to Figure 7.38, as well as work described in Paper 1, the addition of 3-4
w/w% SWCNT solution increased conductivity of the printed PEDOT:PSS films, and further
addition did not yield any noticable improvement, as conductivity plateaued at about 1 S/cm at
this point (for a single layer printed at 25 µm drop spacing) on cellulose acetate.
The detailed and rigorous formulation procedure outlined above for the PEDOT:PSS/SWCNT
ink encompassed a large part of the work. This procedure was adapted to each respective ink
depending on the mechanism of that ink’s function. In this case, physical connectedness was
vital to proper ink performance. However, other key variables might affect the performance of
various inks, such as topography, thickness, roughness, and so forth. The basic tenets of the
formulation procedure remained universally the same, however; modification of the fluid
properties of a dispersion in a linear fashion, establishing one and moving on to the next, while
always making sure that each new additive did not have any adverse effects on the previously
optimized properties. Again, because PEDOT:PSS/SWCNTs served as a “model” ink for the
formulation procedure, all of the details of the formulations of the other inks will not be included
here, for the sake of brevity – but unique challenges encountered with the formulation of each
will be addressed. In each case, however, at least ten different ink formulations were tested
before a suitable mixture was found (all of which are listed in APPENDIX I).
7.3.2 ZnS inks
Each of the aqueous and organic ZnS inks had one primary component – the nanoparticle
suspension. A solids concentration of 1 wt.% ZnS was targeted. An initial attempt (described in
Paper 7) at making an ink using AA-capped ZnS:Mn in an ink composed primarily of AA and
other aprotic solvents yielded a very dilute suspension (< 0.8% in an ink almost completely
153
composed of AA) of weakly photoluminescent nanoparticles. The almost pure acid comprising
the ink meant that it damaged the printhead, cartridge, substrate, and underlying layers very
rapidly, and the slow heat-polymerization of the AA monomer allowed the solids to move to the
edges of the printed pattern (Figure 7.41). Also, the capping agent (AA) did not quench surface
states as well as 3-MPA, resulting in red-shifted emission and blue emission from S2-
vacancies.
The reformulation of ZnS:Mn and ZnS:Cu inks capped with 3-MPA are described below.
Figure 7.41. Printed ZnS:Mn/AA ink on cellulose acetate, 10 printed layers.
Table 7.4. Ink components incompatible with ZnS nanoparticle suspensions (aqueous and organic).
Component Purpose Issue Ink
Mercaptosuccinic acid (Small et al. 2010)
Dispersant, viscosity modifier Immediate precipitation Aqueous ZnS:Mn, Cu
PVP (>1wt%) Binder Gradual precipitation Aqueous ZnS:Mn, Cu
Air Products DF-110D defoamer
Defoamer (req’d in Triton-containing inks)
Gradual precipitation All
Carboxymethylcellulose Viscosity modifier Gradual precipitation, insoluble in organic
solvents All
Isopropanol Surface tension modifier Immediate precipitation
(SHMP insoluble) Aqueous ZnS:Cu
Butoxyethanol Viscosity modifier Immediate precipitation
(SHMP insoluble) Aqueous ZnS:Cu
3-amino-1-propanol Viscosity modifier
PL enhancer (Wang et al. 2009) Reaction with TGA Aqueous ZnS:Mn
Dimethyl sulfoxide Co-solvent Gradual precipitation All
As was mentioned in Section 6.2.3, the a mixture of isopropanol, butoxyethanol, water, and PVP
1 300 000 were suitable to prepare an aqueous ink containing ZnS:Mn. The high viscosity and
Visible 302 nm UV
154
low volatility but moderate boiling point of butoxyethanol made it a good co-solvent and
humectant, and isopropanol served to reduce surface tension without having to introduce
surfactants – the potential issues which have been elaborated upon in Section 7.3.1. However,
the addition of certain materials during formulation caused precipitation of the nanoparticles for
a variety of reasons and imposed limitations on ink formulation (Table 7.4). The inability to add
any solvents other than water to the ZnS:Cu aqueous suspension due to the insolubility of
SHMP, for example, meant that it was not suitable to use in an aqueous inkjet ink.
While ZnS:Cu was precluded from use in an aqueous inkjet ink, ZnS:Mn encountered problems
with application in an organic ink. The first issue that arose was a practical one: the phase
transfer process failed to produce as stable a suspension of ZnS:Mn in toluene, whereas in
ZnS:Cu, no such problem was encountered. More importantly, however, the organic-dispersible
ZnS nanoparticles were to be bound with PVK in a DC-LED structure, where charge transfer
was vital between the matrix and the nanoparticles. The semi-conductive CuS which forms in
DC-driven devices is integral to this charge transfer (Ono 1995); MnS does not similarly
conduct. Therefore, ZnS:Cu was the only choice for the emitter in the PVK layer. So, one ink
was made containing ZnS:Mn as an aqueous suspension, and one ink was made containing
ZnS:Cu as an organic suspension, meaning that the DC-LEDs were to be built using ZnS:Cu and
the PELs to be built using ZnS:Mn.
Figure 7.42. Droplet formation of ZnS inks at 10 µs intervals.
Jetting waveforms producing stable, spherical droplets were created for both inks (Figure 7.42);
see APPENDIX J for the waveforms. No significant bead-on-a-string effects were observed
even with the loading of polymers into the inks. Even with the low viscosity observed in the
200 µm
(a)
(b) (c)
ZnS:Mn
(aqueous)
ZnS:Cu
(organic)
155
toluene-based ZnS:Cu ink, drop formation was uniform and stable, although with some
splattering.
There was no means by which to increase viscosity in the ZnS:Cu layer except by adding more
PVK (typical viscosity modifiers were insoluble). However, at higher PVK loadings, film
structure was observed to be poor; however, inks with a lower PVK loading actually formed
superior films due to their lower viscosity. Although other viscosity modifiers were considered,
the desire for a semiconducting film free of high-boiling solvents in DC-LED structure precluded
their use. This was the only ink which did not meet the jetting criteria specified for the printer,
with µ = 0.97 cP; however, inks containing toluene at low µ values have occasionally been
printed on the DMP2831 (Sumerel et al. 2007). Regardless, film formation on the surface was
successful with both inks, producing gap-free layers (Figure 7.43). The optimization of the
layers’ structure will be discussed in Section 7.4. PL emission was immediately visible after
deposition, indicating that not only had the ink jetted properly, but the formulation allowed for
the retention of functionality and had no chemical compatibility issues.
Figure 7.43. Left: optical microscope imaging of a printed ZnS:Mn/PVP (aqueous) film on glass (single layer); centre:
PL on glass, 1-5 printed layers; right: 10 printed layers on photo-paper. PL was induced by excitation with a 302 nm UV source
As described in Section 6.2.2, the BaTiO3 ink utilized an ethanol/MMA solvent mixture and PEG
300 to reduce volatility and increase viscosity. 5% BaTiO3 was added to the ink, as this was the
limit of stability (see Section 7.2.1). No surfactant was required due to the low surface tension
values of the ethanol and MMA. Drop formation was ideal with this ink (Figure 7.44).
500 µm
100 µm
302 nm UV excitation
156
Figure 7.44. Droplet formation of BaTiO3 ink at 5 µs intervals.
The relatively high viscosity of the ink contributed to stable, uniform drop ejection from the
piezoelectric nozzles. There was no evidence of the formation of satellite droplets, wetting of the
nozzle plate, nozzle clogging, or other undesirable print effects. The drops were approximately
10 pL in volume, estimating from the drop formation images, indicating ideal operation of the
printhead and successful voltage waveform design. The uniformity of the jetted drops allowed
moderately high resolution lines of the ink to be printed, with drop sizes as small as 25 µm
(Figure 7.45).
Figure 7.45. (a) Jetted drops of BaTiO3 ink on ITO PET; (b) edge of BaTiO3/PMMA film on ITO glass.
7.3.3 Optimized ink formulations
A summary of all the finalized ink formulations is given in Table 7.5. Summaries of these
formulations and the iterations attempted to reach them are given in APPENDIX H and
APPENDIX I, respectively.
200 µm
100 µm 200 µm (a) (b) (b) (a)
157
Table 7.5. Finalized ink formulations.
Ink Solvent system
Composition (w/w)
Fluid properties
µ (cP)
(mN/m)
dp, max (nm)
Z-1
En
PEDOT:PSS/SWCNTs
Aqueous
34% PEDOT:PSS susp’n (1.3% in water)
17% glycerol 10% DMSO 10% SWCNT sol’n
(0.4 % in water/SLS)
0.5% SLS 0.5% DF-110D 28% H2O
2.2 30.2
~180 (no CNTs)
~400 (due to CNTs)
12.5 ~0
BaTiO3 PMMA
Organic
5% BaTiO3 33% ethanol 28% MMA 0.5% Surfynol CT-
324 0.5% Disperbyk
111 33% PEG 300
9.6 31.7 ~30 2.9 ~0
ZnS:Mn PVP
Aqueous
40% ZnS:Mn (2.5% in water/TGA/NaOH)
15% butoxyethanol
10% isopropanol 0.1% PVP
1,300,00 34.9% H2O
12.1 32.1 ~16 2.3 ~0
ZnS:Cu PVK
Organic
40% ZnS:Cu (2.5% in toluene/olelyamine/3-MPA)
10% cyclohexane 0.075% PVK 49.925%
chlorobenzene
0.97 28.4 ~33 27.5 ~0
158
7.4 Ink film formation
Following ink formulation, the procedure for determining ideal drop spacing, film topography,
and film thickness for inkjet printing of the inks is presented in this section. In this case,
BaTiO3/PMMA ink was used as the model ink. Again, it must be noted that each ink had several
substrates upon which it was deposited. These included an impermeable substrate (glass, in this
case) for topography, functionality, and thickness observation, and any layers lying below it in
the actual device structure. The surfaces upon which each ink was deposited are summarized in
Table 7.6.
7.4.1 Drop spacing
BaTiO3 was to be deposited on top of the ZnS:Mn/PVP layer in the ACPEL stack, as well as on
ITO for capacitance testing and on glass for thickness testing. The drop sizes of BaTiO3 on each
of these surfaces varied widely, as can be seen in Figure 7.46. On the surface of the
ZnS:Mn/PVP layer, some interlayer mixing appeared to occur, and the droplet naturally widened
as it dissolved into this layer. This was problematic for the construction of the ACPEL devices
(see Section 6.7.1). Also, the different surface energies of each substrate meant that drops dried
different ways; the ITO-based drop, for example, produced a pronounced coffee-ring effect,
while the glass-based drop spread evenly. The different surface energies and roughnesses of
these underlying layers were the determinants of drop spreading.
Figure 7.46. Drop sizes of BaTiO3/PMMA ink on various substrates.
50 µm
Slide glass ITO glass ZnS:Mn/PVP
55 µm 40 µm 65 µm
159
The size of the isolated droplets determined the line spacing in a printed layer, where line overlap
contributed to film roughness, and too-wide spacing produced holes in the conductive layer.
However, spacing the drops at the diameter of the individual drops did not necessarily result in
the formation of a smooth film (Figure 7.47). The BaTiO3 ink on ITO, for example, did not form
a uniform film at 40 µm drop spacing, and began to overlap excessively at 30 µm drop spacing –
suggesting an ideal drop spacing of 35 µm. However, it formed a smooth film on slide glass at a
drop spacing which was the same as its individual drop size (55 µm). The way the solids
distributed in the drying droplet, as shown in the previous figure, determined the degree of
overlap required to produce a film containing no holes.
Figure 7.47. Printed lines of BaTiO3 ink (single jet, single layer) at different drop spacings.
As the lines are brought closer together, more and more material is deposited in a ridge. This
exact problem was encountered by Haverinen et al. (2010), where it caused localized dimming of
the devices. In some cases, the wetting of the surface was poor (as with ITO), so that close
spacing of the lines was necessary to prevent pinholes. However, excessively closely-spaced
lines were also expected to cause such an increase in film thickness as well as roughness. Ideal
(c) (d)
50 µm
Drop spacing: 40 µm 35 µm 30 µm 25 µm
Drop spacing: 60 µm 55 µm 50 µm
On ITO glass
On slide glass
160
drop spacing was established by jetting several lines of ink at different drop spacings in
increments of 5 µm, ranging from the measured single drop diameter down to 50% of the single
drop diameter (i.e. for a drop size of 60 µm, drop spacing ranging from 30-60 µm was tested.
The resulting films were characterized using optical microscopy for uniformity (see APPENDIX
D). Ideal drop spacing was considered to be when lines were fully merged (no holes) but not
overlapping. Because the printer was limited to 5 µm increments in drop spacing, a compromise
between overlap and merging of lines was often necessary. The drop sizes and spacings for
various inks are given in Table 7.6.
Table 7.6. Summary of drop sizes and line spacing for all inks.
Ink Underlying layer Drop size (µm) Correct line spacing (µm)
PEDOT:PSS SWCNTs (aqueous)
Glass BaTiO3/PMMA ZnS:Mn/PVP ZnS:Cu/PVK
45 30 45 50
25 25 25 45
BaTiO3 PMMA
(organic)
Glass ITO
ZnS:Mn/PVP
55 40 65
55 35 65*
ZnS:Mn PVP
(aqueous)
Glass ITO
PEDOT/SWCNTs
65 50 60
55 45 50
ZnS:Cu PVK
(organic)
Glass ITO
PEDOT/SWCNTs
65 45 50
30 30 45
*the BaTiO3 ink damaged the ZnS:Mn/PVP layer extensively.
There was no consistent trend for any substrate in terms of drop size versus drop spacing, other
than that drop spacing was generally smaller than the average drop size. Perhaps the most
notable observation was that PEDOT:PSS wet most surfaces worse than the other inks (smaller
drop sizes), and was not able to form smooth films except at narrower drop spacings. This may
have been a result of the use of a surfactant rather than a fluid with low surface tension – wetting
became a dynamic process rather than an instantaneous one as the surfactant molecules adhered
to the substrate. As the droplets rapidly dried, they would not have yet had a chance to spread to
their fullest diameter. Also, the (relatively) low solids content likely meant that capillary flow of
solvent/particles to the droplet edges was not a dominant process, and that drops dried more in a
(b) (c) (d) (e)
161
“dome” shape than a coffee-ring or flat morphology. Indeed, in inks with higher solids contents,
the Marangoni effect was much more pronounced, as will be seen in the following section,
whereas PEDOT:PSS/SWCNT ink formed smoother films.
7.4.2 Film topography
When a 3-D examination of film topography was made using optical profilometry, the problems
with solute movement during drying became self-evident. These were observed on both a micro-
and macro-scale. In the latter case, the BaTiO3 ink’s chemistry was problematic upon film
curing. The component that contributed primarily to viscosity and assisted with jetting, PEG
300, also made the jetted films difficult to rapidly cure and therefore affected their topography;
indeed, the images of the BaTiO3 films shown above represent films which were not fully cured.
PEG 300 has a very high boiling point compared to the volatile ethanol solvent, and took much
longer to evaporate; also, the polymerization of MMA took some time. It is possible, in fact, that
it did not evaporate entirely, and remained entrapped within the printed films even after they
appeared to be dry. In any case, the slow curing caused a visible shift in film topography,
inducing a macroscale version of the troublesome coffee-ring effect (Figure 7.48). A portion of
the ink migrated to the edges of the printed pattern, forming a ridge of substantial height,
compared to the average film thickness of the samples. Because of the long drying time of the
PEG 300, the large surface tension gradient between PEG 300 and the other components (43.5
mN/m versus 28 mN/m and 23 mN/m for MMA and ethanol, respectively) caused these ridges to
form (Tracton 2005). The non-ridged regions of the film had very high smoothness and
uniformity; reduction of the effects of surface tension gradients by reformulation or immediate
high-temperature curing would likely alleviate this ridging effect. . The ridges observed
between printed lines in the drop spacing tests were not visible in the fully cured films.
Therefore, the slow drying time and viscous co-solvent did assist in film leveling to a degree, but
produced the ridging problem at the edge of the film, regardless, due to surface tension gradients.
Film thickness averaged at 150 nm (not including the ridged edge of the sample).
1.5 µm
162
Figure 7.48. 3-D profile (left) and 2-D linescan (right) of a single printed BaTiO3 on slide glass. The arrow on the
scale bar on the 3-D plot refers to the level of the glass surface.
Figure 7.49. SEM micrographs of printed BaTiO3 ink (one layer). (a) Surface at 70° tilt; (b) 0° tilt. Yellow circles on
the high-magnification micrographs highlight some of the localized clusters of BaTiO3 where solvent concentrated.
The surface of the BaTiO3 film’s non-ridged regions was very smooth, and apparently pinhole-
free. SEM imaging (Figure 7.49) shows the relatively smooth surface of the BaTiO3 films. More
importantly, deep pinholes were not observed. If the films were to be used as gate dielectrics or
insulators in a display, pinholes could cause catastrophic breakdown of the entire device.
Localized thinness of the insulator might also be problematic, and some shallow pits were
observed in the SEM images. These were likely a result of solvent vapour bubbles escaping
from the film as it cured. The use of lower temperature vacuum-drying might alleviate this
problem to a degree; however, it is unavoidable that bubbling of solvent vapours will occur in
0
0.1
0.2
0.3
0.4
0 200 400 600
Fil
m h
eig
ht
(µm
)
x-dimension (µm)
10 µm
700 (b)
6.3 µm
600 µm 450 µm
5 µm
500
0
300
100
nm
(a) (b)
BaTiO3 cluster
163
any wet-processed film. The relatively shallow depth of the pits in the surface suggested that the
presence of high-boiling PEG 300 assisted in the leveling of such voids, as it flowed in to fill
them after the volatile solvents were removed. In regions where volatile solvents were more
concentrated during curing, clusters of BaTiO3 were left behind – resulting in a non-uniform
distribution of insulating material. Some of these clusters can be clearly seen in Figure 7.49b.
The insulator was generally well dispersed and assembled, but small holes and some non-
uniformity had occurred as a result of rapid ethanol evaporation.
Figure 7.50. BaTiO3/PMMA average dried ink film thicknesses on slide glass.
The thickness of the BaTiO3 layer was examined during over-printing of multiple layers, as well.
Because a thicker layer (~1 µm) was needed for the deposition of an ACPEL, several layers of
BaTiO3 would be required. The thickness was estimated at the centre of the films (not at the
raised edges) using the procedure outlined in APPENDIX C. As the number of layers increased,
the thickness increased in a linear fashion; at 8 printed layers, the film was thick enough at ~1
µm (Figure 7.50) to apply in the ACPEL structure, according to the desired layer thicknesses
specified earlier in Section 6.7. It was expected at this thickness that the film would be
sufficienctly insulating to function in an ACPEL device. The other inks, which did not contain
the high-boiling PEG 300, formed films with varying topographies. In each case, some degree of
“coffee-ring” formation was observed, with the edges of printed lines forming raised edges,
0
200
400
600
800
1000
1200
1400
1600
0 1 2 3 4 5 6 7 8
Fil
m t
hic
kn
es
s (
nm
)
# of printed layers
(a) (b)
164
resulting in a “peak-and-valley” structure. The peak-and valley structure was particularly rough
and uneven, with large variations between the highest and lowest points, listed in Table 7.7.
Table 7.7. Printed film roughnesses and peak-to-valley differences in ZnS and PEDOT:PSS/SWCNT inks.
Ink Underlying layer Valley thickness (nm) Peak thickness (nm) RMS roughness (nm)
PEDOT:PSS/ SWCNTs
Glass BaTiO3/PMMA ZnS:Mn/PVP ZnS:Cu/PVK
130 145 25
110
400 500 300 620
40 30 70 80
ZnS:Mn/ PVP
ITO PEDOT/SWCNTs
55 30
680 300
40 10
ZnS:Cu/ PVK
ITO PEDOT/SWCNTs
65 65
1000 1100
300 230
This problem was the most pronounced in the ZnS:Mn/PVP ink, which contained a higher
loading of large-Mw polymer than either of the other inks. It was noticed in the ZnS:Cu ink and
another trial ink that was prepared using only PVK that these had the worst topography of any –
suggesting that the presence of dissolved polymers caused problems with coffee-ring formation.
PEDOT:PSS/SWCNT ink, which contained no dissolved polymer, had the smoothest topography
of any of the materials used, and no peak-and-valley formation was visible in the BaTiO3 inks
either, supporting the hypothesis that dissolved polymers were responsible for this rough
topography. Again, Haverinen (2010) discussed this precise issue during the printing of QDs;
the approach in that work was to remove a fraction of the solids in the ink, resulting in smoother
film topography. The hypothesis presented was that the larger amount of solvent present in the
ink as a result of reducing solids content would prolong drying time and thereby aid in surface
wetting. This is possible, considering that the worst-offending ink was ZnS:Mn/PVP, which had
the lowest boiling-point solvents of any of the inks. However, by Haverinen’s explanation, only
a very small amount of solvent was added (reducing solids concentration from 0.7 wt.% to 0.5
wt.% and replacing it with an additional 0.2 wt.% of solvent). This trivial amount of added
solvent did not seem to support the claim of enough of an increased drying time to affect film
topography. It was more likely that better leveling occurred in inks with the proper fluid
characteristics, where a lower viscosity allowed more rapid flow of solvent and better film
formation – as evidenced by the large drop sizes observed for ZnS:Cu/PVK (µ = 0.97 cP) (Table
7.6) and the superior film formation of PEDOT:PSS/SWCNTs (µ = 2.2 cP). The high-viscosity
165
inks, containing BaTiO3 and ZnS:Mn/PVP, experienced the most drastic problems with
comparatively rougher surfaces. Examples of ink topography on glass slides are shown in Figure
7.51.
Figure 7.51. Topography of single printed layers of all inks on slide glass.
Besides PEDOT:PSS/SWCNTs, most of the other inks demonstrated a relatively high degree of
roughness, and in some cases, isolated ridges or features of several hundred nanometres in
height. The edges of printed lines observed during optical microscopy to determine drop spacing
were clearly visible as repeated ridges, the extreme example of which is seen in the ZnS:Mn/PVP
ink. In this case, the finished film was to be relatively thick (~20 µm), meaning that topography
was not as much of a concern – so the ink was not reformulated to improve its leveling ability.
The relatively high smoothness of the ZnS:Cu/PVK and PEDOT:PSS/SWCNT films, which
were to be used in the thin (sub-µm) film DC-LEDs was suitable.
PEDOT:PSS/SWCNTs ZnS:Cu/PVK
ZnS:Mn/PVP BaTiO3/PMMA
600 µm 450 µm
166
The film thickness was found to vary almost linearly with the number of printed layers on every
substrate, meaning that films of controlled thickness were readily deposited. Where a single ink
film was thicker than the layer’s targeted thickness, only a single layer was used. Thus, the
parameters for inkjet-printing of all of the device structures outlined in Chapter 6 were finalized
and summarized in Table 7.8. It was noted that the ink layer thicknesses were best-suited to thin-
film devices, as many print passes were required to provide the thick films characteristic of
PELs.
Table 7.8. Ink layer thicknesses and layers required for device construction.
Ink Underlying layer Device type Single layer
thickness (nm) Layer & desired thickness (nm)
Number of printed layers required
PEDOT:PSS SWCNTs (aqueous)
Glass BaTiO3/PMMA ZnS:Mn/PVP ZnS:Cu/PVK
All ACPEL DCPEL DC-LED
130 145 25
110
Anode (100) Cathode (1000) Cathode (1000) Cathode (1000)
1 7 40 9
BaTiO3 PMMA
(organic) ZnS:Mn/PVP ACPEL Dissolved (n/a) Insulator (1000) n/a
ZnS:Mn PVP
(aqueous)
ITO PEDOT/SWCNTs
PEL PEL
55 30
Emitter (20,000) Emitter (20,000)
363 667
ZnS:Cu PVK
(organic)
ITO PEDOT/SWCNTs
DC-LED DC-LED
65 65
Emitter (200) Emitter (200)
3 3
7.4.3 Interlayer interactions
When printing ACPEL structures several issues became apparent. Firstly, the BaTiO3 ink
completely dissolved through the ZnS:Mn ink below it, regardless of the degree of cross-linking.
Rub testing further confirmed that heat- or UV-treated films of ZnS:Mn/PVP were completely
pervious to the BaTiO3 ink’s solvents. A second major issue was with the number of printed
layers necessary to produce the thick film necessary for PEL function, which was exacerbated by
the small average thickness of the ZnS:Mn films (due to the peak-and-valley formation). Time
consumption aside, overprinting with a large number of layers also presented difficulties in
maintaining print resolution. When a large number of ZnS:Mn/PVP layers were jetted onto
glass, the edges of the pattern began to lose sharpness. As droplets impacted the wet ink on the
167
surface during subsequent print passes, ink was ejected from the previous films and landed
randomly on the substrate surface (Figure 7.52).
Figure 7.52. Multiple layers of ZnS:Mn/PVP ink printed on (a) aluminum foil (1-50 layers) and (b) ITO PET (50
layers), showing edge splattering with excessive overprinting.
Because of these issues associated with the production of ACPELs, this structure was dropped.
The BaTiO3 ink developed in this study may find future use as a dielectric in an AC-driven
device using non-printed micron-sized phosphors, or as a gate dielectric in a transistor structure.
The rub-test was carried out on the ZnS:Mn/PVP layer as well with PEDOT:PSS/SWCNT ink –
resulting in its complete dissolution, even after heat- and UV-crosslinking. Overprinting of the
ZnS:Mn/PVP layer with even a single layer of PEDOT:PSS/SWCNT ink caused it to dissolve.
So, with the inks developed in this work, a fully inkjet-printed DCPEL seemed unlikely to be
realistically constructed. However, this issue could be addressed by vacuum-deposition of Al as
top conductive layer.
200 µm Visible UV
1
# layers
2
2
3
4
5
10
20
50
(a) (b)
Visible UV
168
Figure 7.53. ITO dissolution by ZnS:Cu/PVK ink, containing 3-MPA, and ZnS:Mn/PVP ink, containing TGA. Top: 3-D
profiles of the surfaces; bottom: x-direction linescans across the droplets.
Another issue that arose with device deposition was the sensitivity of ITO to the acid (3-MPA or
TGA)-bearing inks – i.e. the emitter species. As is shown in Figure 7.53, the 3-MPA in
particular etched into the ITO’s surface and penetrated through its entire thickness, the ITO only
being ~100 nm thick, as specified by the supplier. The etching effect of the TGA was not as
severe, but it was still present. It is possible that the bases (olelylamine and NaOH) present in
the inks were responsible for the etching effect as well. There was no way to further treat the
ITO to prevent this from occurring; however, the dried PEDOT:PSS/SWCNT layers survived the
rub tests from both of the ZnS inks with no apparent damage. Therefore, PEDOT:PSS was used
as the anode in the place of ITO, further narrowing down the number of device structures that
-0.15
-0.1
-0.05
0
0.05
0.1
0.15
0 200 400 600
To
po
gra
ph
ica
l h
eig
ht
(µm
)
-0.06
-0.04
-0.02
0
0.02
0.04
0.06
0 200 400 600
ZnS:Cu/PVK (3-MPA)
600 µm 450 µm
ZnS:Mn/PVP (TGA)
Ink droplet
ITO
x-dimension (µm)
169
were actually built and tested. While this was initially considered a problematic issue, the
usefulness of PEDOT:PSS as a readily-patterned flexible material (as described in Paper 1)
versus the brittle, photolithography-patterned ITO made it a better-suited candidate for all-
printed devices. Also, the HOMO value of PEDOT:PSS, reported as 5.2 eV (Mihailetchi et al.
2003), is higher than that of plasma-treated ITO, which is typically around 4.7 eV (Shlaf et al.
2001). For the purpose of charge injection, this is energetically favourable, allowing holes to
flow easily into the emissive material.
Based on the above considerations among the device structures described in Table 6.2, only three
structures remained to be constructed and examined for their functionality. To facilitate charge
injection into the anode and to provide a connection point for the power source, patterned ITO
glass slides were used as the substrate to demonstrate the feasibility of the above fully inkjet
printed EL devices. The device structures were:
DC-LED: PEDOT:PSS/SWCNTs – ZnS:Cu/PVK – Al
DC-LED: PEDOT:PSS/SWCNTs – ZnS:Cu/PVK – PEDOT:PSS/SWCNTs
DCPEL: PEDOT:PSS/SWCNTs – ZnS:Mn/PVP – Al
170
Figure 7.54. Optical profilometry of printed ink layers, on the surfaces they would cover in printed devices. The
underlying surfaces were prepared by spin-coating of the respective inks.
The film structures of the individual layers did not change much from those deposited on glass,
when deposited on their respective underlying layers in the three device structures described
above. The topography of the films remained smooth for PEDOT:PSS/SWCNTs (being
(a) (b)
(c) (d)
Al cathode (PEDOT:PSS/SWCNT ink dissolved
ZnS:Mn/PVP layer)
PEDOT:PSS/SWCNTs
on spin-coated ZnS:Cu/PVK ink
ZnS:Cu/PVK
on spin-coated PEDOT:PSS/SWCNT ink
PEDOT:PSS/SWCNTs
on glass
ZnS:Mn/PVP
on spin-coated PEDOT:PSS/SWCNT ink
600 µm 450 µm
171
deposited on glass); while the topography remained as the “peak-and-valley” type for ZnS (
Figure 7.54) on PEDOT:PSS/SWCNTs. The high variability in the thickness of ZnS film
presented the greatest concern for device functionality. Any regions where the film was thicker
(peaks) would present localized points through which current could tunnel and short the device,
or alternatively, would be too resistive to allow the passage of current and prevent device
function. Haverinen (2010) observed that the raised regions in a single printed layer of CdS QDs
would not light in a finished device. In the DCPEL, the thin regions would provide pathways
through which hot electrons could arc between the electrodes, causing device burnout in those
regions (Ono 1995). The pronounced peak-and-valley topography in the ZnS:Mn/PVP films
made this latter concern a major issue. The necessity for overprinting each of the layers to reach
a desired thickness alleviated this concern somewhat, as continued overprinting appeared to
reduce the peak-and-valley formation (Figure 7.55). The usefulness of overprinting was limited
by issues of edge resolution, as described above, and by the desired thickness of the printed film.
However, in driving a printed device, it would most likely be more practical to have a thicker
layer that required more voltage to achieve EL than to have a rough layer which could fail
catastrophically. The eventual application of such a device might be difficult with a high driving
voltage, but as a model for an all-printed device, some of the disadvantages of the inkjet printing
method could be overcome by overprinting. In the following section, the relationship of
overprinting to functionality is illustrated for PEDOT:PSS/SWCNTs and BaTiO3.
Figure 7.55. Reduction of “peak-and-valley” topography in ZnS:Cu/PVK films with successive overprints on slide
glass. The topography of a single film of ZnS:Cu/PVK was presented in Figure 7.51.
5 printed layers
600 µm 450 µm
10 printed layers
172
7.5 Functional testing of individual layers
7.5.1 PEDOT:PSS/SWCNTs
From the examination of the PEDOT:PSS/SWCNT ink during formulation and substrate
interaction studies, functionality for this layer has been already established. Because it was to be
applied to glass, printed layers of PEDOT:PSS/SWCNTs on glass were tested for conductivity,
demonstrating that an acceptable conductivity could be achieved even with a single printed layer
(Figure 7.56). Successive overprinting yielded improvement in conductivity, as any gaps in the
printed pattern or localized topography were smoothed out – this smoothening behaviour was
universally seen for the printed films, as discussed in the previous section. Also, the introduction
of successively more conductive “pigment” (i.e. SWCNT) into the films and the overlaying of a
network of PEDOT:PSS chains and SWCNTs improved the connectivity of the conductive
regions. The single print-pass conductivity was 4.6 S/cm, which is significantly lower than ITO
(averaging at about 100 S/cm – a conductivity not achieved until 7 print passes with
PEDOT:PSS/SWCNTs had been made). However, this was more than sufficient for application
as an anode in an electronic device (Hsiao et al. 2008).
Figure 7.56. PEDOT:PSS/SWCNT ink conductivity when printed on slide glass.
1
10
100
1000
0 1 2 3 4 5 6 7 8 9 10
Co
nd
uc
tivit
y (
S/c
m)
# of printed layers
173
PEDOT:PSS/SWCNTs were also comparable with ITO in terms of their optical transmittance.
ITO’s optical transmittance is generally 80% or lower, depending on the film thickness (Wei et
al. 2001) as is the case with PEDOT:PSS, where more layers of PEDOT:PSS/SWCNTs produce
a darker blue colour. A single printed film of PEDOT:PSS/SWCNT ink on cellulose acetate had
a transmittance (measured by UV-visible spectrometry) in the visible range of ~83%.
Of course, the PEDOT:PSS/SWCNT ink could not match the conductivity of a metallic layer.
When used as a cathode (10 printed layers), the conductivity across the plane of the film was 200
S/cm, several orders of magnitude lower than most metals. Furthermore, the direction of current
flow in the functioning device was both across the plane of the film as well as across the film
thickness. Depending on the alignment of the CNTs and PEDOT chains, conductivity might be
anisotropic. It was expected as a result that a higher drive voltage would be required to observe
uniform EL from the devices, overcoming resistive losses across the PEDOT:PSS/SWCNT
electrode that would be negligible with a metal electrode.
7.5.2 ZnS
The spin-coated devices were prepared to qualify the phosphor materials and to ensure that they
were actually electroluminescent. In these devices, printed-line topography and other pitfalls
associated with inkjet printing were not expected to be relevant. However, because there is no
way to pattern to spin-coated layers, the entire ITO slide was coated with the device, and the
cathode (Al) was deposited directly onto the top of the device, with no point away from the
emissive area to which to connect an electrode. Therefore, when electrodes were contacted to
the device, they tended to penetrate the Al (and, indeed, the rest of the device) very rapidly,
causing shorting. However, short-lived blue-green electroluminescence was observed from films
of ZnS:Cu/PVK deposited on PEDOT:PSS-coated ITO. Momentary orange electroluminescence
was also observed from ZnS:Mn/PVP films. Drive voltage in all cases was >30 VDC. Because
of the poor device architecture, shorting occurred within a few seconds of successful
illumination. However, the experiment did successfully demonstrate that the ZnS quantum dots
were capable of electroluminescence under DC drive. For the sake of completeness, AC-driven
devices were also tested (without the BaTiO3 layer, as it would have dissolved the underlying
174
ZnS:Mn/PVP film. No EL was observed from these devices. Unfortunately, the nature of the
EL material, being nanoparticulate and unencapsulated, as well as the absence of an insulating
layer, prevented these devices from functioning (Bredol & Dieckhoff 2010). Since the
electroluminescent properties of DC-driven ZnS nanoparticles had been demonstrated, testing of
the fully printed devices was undertaken.
7.5.3 BaTiO3
Figure 7.57. Estimated relative dielectric constants of printed BaTiO3 films. Film thickness at the centre of the
sample (in the non-ridged region) is also indicated.
Even though the BaTiO3 film was not used in the LEDs, its functionality was established in order
to demonstrate its usefulness as an insulating layer in other electronic devices. In general, the
BaTiO3/PMMA films were sufficiently thick, uniform, and pinhole-free to demonstrate moderate
dielectric constants (Figure 7.57). Dielectric constant () scaled with film thickness, which
naturally scaled with the number of deposited layers, as per the definition of . In the case of a
single printed film, which had some small voids and surface pinholes (Figure 6), current was able
to tunnel through the film and no capacitance was measurable. However, as the pinholes were
filled by subsequent printed layers and tighter packing of the larger amounts of BaTiO3 on the
surface, the dielectric constant also increased. The occasional irregularities seen in the optical
0
200
400
600
800
1000
1200
1400
1600
0
20
40
60
80
100
120
140
160
180
0 1 2 3 4 5 6 7 8
Fil
m t
hic
kn
es
s (
nm
)
Die
lec
tric
co
ns
tan
t (u
nit
les
s)
# of printed layers
175
profiles of the single BaTiO3/PMMA film were not present in the films with more layers, likely
resulting in not only the increase in at higher film thicknesses, but also the nearly linear
relationship between and film thickness above 5 printed layers. Above a certain number of
layers, pinholes were completely covered by repeated print passes and breakdown became less
and less likely. Therefore, deposition became a compromise between the pronounced ridging
effect at the edge of the films and increased dielectric performance. The estimated value of
and the film smoothness (in the non-ridged region) would be sufficient for application of the
printed BaTiO3 films in a number of electronic devices, including emissive displays (Ono 1995).
7.6 EL device testing
Upon printing of the full LEDs, EL emission was achieved in the DC-LED device builds, with
current-voltage (I-V) characteristics of these devices showing typical diode behaviour (Figure
7.58, Figure 7.59). The I-V curve for the fully-printed (PEDOT:PSS/SWCNT cathode) device
was more linear than that for the Al-cathode device, suggesting that the relatively high resistance
across the cathode was causing resistor-like behaviour. This deviation from diode behaviour
might explain the weaker EL from this device. Weak blue-green emission characteristic of
ZnS:Cu was visible over the 2 mm2 emissive area of the devices (Figure 7.60). In both cases,
sudden shorting ended the functioning, which was visible during testing by their “sparkle”,
representing multiple shorts during testing. Unfortunately, this shorting prevented further testing
of the devices, and the weak emission strength made it difficult to obtain a meaningful EL
spectrum. It was thought that the roughness of the ink layers was sufficient to cause localized
channeling of current and prevent any further radiative recombination at the Cu2+
luminescent
centres during the shorting events. However, the brief functioning of these devices did establish
the feasibility of an all-printed LED using simple processing conditions, even if the EL was
weak.
176
Figure 7.58. Current-voltage (I-V) and luminance-voltage (L-V) characteristics of PEDOT:PSS/SWCNT –
ZnS:Cu/PVK – Al LED. Blue diamonds: I-V curve; red circles: L-V curve.
Figure 7.59. Current-voltage (I-V) and luminance-voltage (L-V) characteristics of PEDOT:PSS/SWCNT –
ZnS:Cu/PVK – PEDOT:PSS/SWCNT LED. Blue diamonds: I-V curve; red circles: L-V curve.
0
0.05
0.1
0.15
0.2
0.25
0.3
0
0.5
1
1.5
2
2.5
3
3.5
4
0 5 10 15 20 25 30 35
Lu
min
an
ce
(c
d/m
2)
Cu
rre
nt
(mA
)
Voltage (VDC)
0
0.02
0.04
0.06
0.08
0.1
0.12
0.14
0.16
0.0
2.0
4.0
6.0
8.0
10.0
12.0
0 5 10 15 20
Lu
min
an
ce
(c
d/m
2)
Cu
rre
nt
(mA
)
Voltage (VDC)
substrate
Al
PEDOT:PSS
ZnS:Cu/PVK
PEDOT:PSS
substrate PEDOT:PSS
ZnS:Cu/PVK
177
Figure 7.60. Electronic band structures (top), device architecture (middle) and EL emission (bottom) of printed
ZnS:Cu DC-LEDs. Left : with Al electrode, Right: with PEDOT electrode.
Another reason for the weak EL, even at relatively high drive voltages (> 30 V), was charge
transfer into the emissive matrix. The bandgaps of the PVK and ZnS were similar (both emitting
light in the blue region, when undoped), so charge injection from the semiconducting matrix into
the nanoparticles themselves was not expected to be a problem. The HOMO value of
PEDOT:PSS (5.1 eV) was also close to that of the PVK (5.3 eV), which in turn was close to that
of the valence band energy of ZnS (5.5 eV), meaning that hole transfer was expected to be
energetically favourable after charge injection into the PEDOT:PSS/SWCNT layer. However,
electron injection through the cathode, be it Al or PEDOT:PSS/SWCNTs, had to overcome a
quantized
Ec = 1.5 quantization
0 eV
8 eV
Al = 4.1
Al PE
DO
T:P
SS
PV
K
ZnS:Cu
20
0
nm
10
0
nm
30
0
nm
Ec = 1.8
Ev = 5.5
HOMO = 5.3
LUMO = 1.9
20
0
nm
10
0
nm
1 µ
m
HO
MO
= 5
.1
LUMO = 2.1
substrate
Al
PEDOT:PSS
ZnS:Cu/PVK
PEDOT:PSS
substrate PEDOT:PSS
ZnS:Cu/PVK
178
large energy barrier due to the wide bandgaps of both PVK and ZnS:Cu. Furthermore, the
quantization of the ZnS:Cu nanoparticles significantly widened their bandgap, increasing the
energetic difference between the Al work function or PEDOT:PSS LUMO and the ZnS:Cu
conduction band. The quantization effect may also have made charge transfer from the PVK to
the ZnS:Cu more difficult, as the quantized ZnS:Cu was a very wide-gap (4.0 eV) semiconductor
– almost an insulator. The energetically unfavourable situation meant that higher voltages were
required to drive the devices, and with their topographical irregularities, high voltage drive
inevitably lead to device breakdown and shorting. Compared to typical LEDs, the drive voltages
of >30 VDC were notably quite high.
With these problems in mind, there exist several possibilities for device improvement. The first
of these concerns the cathode. Even when using Al, the work function is relatively high – and
therefore provides a large barrier to charge injection. Other metallic cathodes with lower work
functions, such as calcium (2.87 eV) or magnesium (3.66 eV) might alleviate this problem.
However, the production of well-dispersed colloids of either of these materials has not been
widely reported; colloidal Mg in THF was prepared by Kalidindi & Jagirdar (2009), and
polydisperse colloidal Ca in THF was similarly prepared by Sanyal et al. (2012). The high
reactivity of both of these metals might preclude them from incorporation into an ink; however,
it is possible that they might be suitable candidates for a cathode material. Once again, the likely
high sintering temperature of these metals would be an issue for any polymeric
components/substrates. A more likely method to improve charge injection from the cathode
would be to include an electron-transport layer of a polymer with a narrower bandgap than PVK.
Of course, this approach necessitates the formulation of another ink, with orthogonal solvents
once again being an added challenge. The presence of wide-bandgap quantized ZnS was the
primary cause for energy band misalignment. The use of CdS, almost ubiquitous with solution-
processed LEDs, avoids this issue (with certain colours, in any case) by controlling bandgap via
particle size. However, for reasons discussed in Chapter 3, CdS was not considered ideal for this
application; moreover, the stability of CdS or CdSe colloids was observed to be compromised
when incorporating them into a multicomponent ink. A sample of orange-emitting (585 nm)
CdSe colloids in water was mixed with the ink components containing in the aqueous ZnS:Mn
179
ink; the result was immediate flocculation of the colloids upon the addition of either isopropanol
or butoxyethanol. PEG and glycerol had the same effect. Addition of the polymer binders (PVP
and PVK) to both water- and toluene-soluble CdSe colloids resulted in immediate flocculation as
well. So, Cd-based nanoparticles, likely due to their surface/cap chemistry, were not suitable for
jetting, unlike the extremely stable ZnS:Mn and ZnS:Cu inks. Because energy level
misalignment was the primary reason for poorer device performance, and the misalignment was
a function of the materials being used, alteration of the cathode material or inclusion of new
charge transport layers was the only practical way to consider improving functionality.
It is also possible that testing apparatus for the devices limited their luminescent intensity
somewhat. The apparatus was capable of delivering 10 mA of current maximum, and the
resistive losses across the devices because of roughness and the use of PEDOT:PSS rather than
ITO or Al meant that only a fraction of that current was able to flow through the layers and
induce radiative recombination. In previous studies with DC-LEDs using ZnS:Cu nanoparticles
– most notably, that of Schrage et al. (2010) – significantly more current was drawn across the
devices. Looking at Figure 7.59, the full 10 mA were drawn at maximum luminance; if more
current had been available at lower voltages, stronger EL may have been produced at a lower
potential. However, the use of self-limiting testing apparati caused immediate shorting, as the
sources delivered as much current as needed to drive the devices, and any topographical flaws
caused a large amount of current to be drawn, breaking down the layers.
The use of PVP as the binder polymer in the DCPEL structure exacerbated the need for high-
voltage drive and the resulting failure of the devices. No significant EL emission was visible
from these devices (Figure 7.61). Again, the I-V curve appeared more linear, suggesting resistor
rather than diode behaviour, due to the thick layer of non-conductive PVP. Although non-
conductive binders such as PMMA have been previously used (Schrage et al. 2010), the amount
of current required to drive such devices far exceeds that output by the testing apparatus (>40
mA), and the devices also required high potential to drive. A different power source may be
sufficient to drive these devices successfully. It would be ideal to be able to use the water-
dispersible ZnS:Mn quantum dots as an emissive layer for ease of processing, and because of the
difficulties mentioned above with phase transferring the ZnS:Mn quantum dots to the organic
180
phase. However, the interlayer dissolution issue which arose with overprinting of PEDOT:PSS
onto the PVP-bound layer remains to be addressed.
Figure 7.61. Current-voltage (I-V) and luminance-voltage (L-V) characteristics of PEDOT:PSS/SWCNT –
ZnS:Mn/PVP – Al DCPEL. Blue diamonds: I-V curve; red circles: L-V curve.
The use of inkjet printing as the preparation method for these devices limited their functionality
in several ways. The first, and most obvious, is the topography of the resulting films, which in
no way approached the smoothness of typical LED layers. The thickness of the films was well-
controlled, but the presence of ridges between printed lines was not. When stacking layers, any
topographical effects between them are amplified, particularly if the droplet spacing is the same
or close to the same for those layers. A slight offset in overprinted layers (i.e. by half of the drop
spacing) might partially alleviate this issue. However, judging by the work presented by
Haverinen et al. (2010) and Wood et al. (2009), some degree of ridging in printed nanoparticle
layers is almost unavoidable, so a better approach to improving functionality is to consider
0
0.02
0.04
0.06
0.08
0.1
0.12
0.14
0.16
0
0.002
0.004
0.006
0.008
0.01
0.012
0.014
0.016
0.018
0.02
0 5 10 15 20 25
Lu
min
an
ce
(c
d/m
2)
Cu
rre
nt
(µA
)
Voltage (VDC)
substrate
Al
PEDOT:PSS
ZnS:dopant/PVP
181
energetic efficiency. Overprinting to improve surface properties resulted in devices through
which no current was able to flow; the I-V plots are not presented here because they simply show
a minimal current flow, no diode behaviour and no luminance. The added thickness of these
layers further reduced the energetic favourability of carrier recombination due to increased
resistance (Schrage et al. 2010).
Energetic efficiency can be improved primarily by clever materials selection. Materials
selection, however, is limited by what is jettable (and orthogonally jettable), and it is here that
the major stumbling block in the way of printed LEDs and PVs is encountered. In this study, the
use of ZnS meant that a material was suitable for jetting which was readily prepared by wet
methods, highly luminescent, and nearly monodisperse. However, it also meant that a wide
bandgap material was the luminescent centre, requiring a large amount of energy to overcome
barriers to charge transfer. Furthermore, the binder matrix was also a wide-bandgap material.
The preparation of a practically applicable fully printed device presents a challenge for these
reasons. The inkjet printing process may be limited or even completely crippled in its capacity to
produce these devices because of these definitive problems. It is likely that single-layer devices
with improved layer smoothness are the most realistic approach to optimizing these devices for
future application.
182
Chapter 8
8 Conclusions
A model electronic device, a DC-driven LED, was prepared using inkjet printing to produce each
of its layers. The device established a proof-of-concept; i.e. that an entire electronic device could
be built using a single unit operation at ambient conditions with solution-based materials. The
device also served to illustrate the many challenges involved with transferring conventionally
vacuum-processed materials to the liquid phase, and further, to jettable inks.
The fundamental contribution of the work can be stated as the establishment of a detailed sample
procedure for the development of a desired functional, printed material from chemical
precursors. Of course, the demonstration of a fully-printed LED was also a major contribution,
but this device requires further optimization before it becomes practically applicable. The many
considerations of materials selection, dispersion, ink formulation, deposition, and layer formation
comprise the greater part of the difficulty of inkjet manufacturing. The functionality of the
resulting devices is entirely dependent on each of these prior steps, and so, this work has sought
to comprehensively outline the means by which each step may be successfully undertaken.
The advantages of having such an outline to inkjet processing are manifold. Firstly, this work
attempted to illustrate the difficulty of moving a material from the bulk phase to the dispersed
liquid phase – particularly with traditionally-used inorganic semiconductors and insulators. This
stage is likely the limiting factor in the success or failure of ink development. Secondly, the
complexity and time-consuming nature of ink formulation is often overlooked in previous reports
on inkjet deposition. Particularly in the case of commercial application and production of inks,
repeatability and reliability of print is absolutely vital; therefore, the stability of ink dispersions
and the maintenance of their fluid properties must be established. As is often the case with
research endeavours, the practical application and scale-up of the technologies being researched
may be overlooked, but in this case, the establishment of industrially viable ink formulations was
considered an important aspect of the work. Also, the use of multiple ink layers pressed home the
necessity for early consideration of device structures; the many different layer stacks that might
183
use the same materials in different orders required the production of inks with different solvent
resistances. Thirdly, having an established routine for determining ideal drop spacing to achieve
a certain topography meant that inks could be rapidly tested and used or discarded depending on
their jetted performance. Lastly, several rapid but effective means for determining layer
functionality before device construction were outlined, including capacitance testing and
conductivity testing, even on porous substrates. It is hoped that by following this procedural
example, electronics printing will become a better-streamlined process and that opportunities for
utilizing all of the advantages of the inkjet printer will be realized.
8.1 Major findings
The use of paper and conventional paper printing and coating techniques was suitable for the
production of paper-based AC-driven powder ELDs, as is described in detail in Paper 6. By
avoiding the use of ITO and Al, the devices were readily produced at ambient conditions.
However, they still presented difficulties in fabrication, due to the need for masking and the
wastefulness of most coating methods. This served as the motivation for the examination of the
exact procedure needed to move from these bulk materials to inkjet-printable materials. There
were numerous findings at each stage of the ink development. To summarize the absolute
milestones of the work: the successful preparation and deposition of a functional conductor,
semiconductor, and insulator, and their integration into a printed LED. However, the study of
each of these inks as models for the stages of the LED’s development yielded several other
findings.
In the case of the PEDOT:PSS/SWCNT ink, conductivity of inkjet-printed films was shown to
not only be a function of the conductivity of its constituents but also depended on the jetting
characteristics and ink-substrate interactions. In particular, ideal droplet formation and ink
spreading were found to play a dominant role in determining the film resistance, with less
pronounced effects resulting from chemical/conformational changes in the PEDOT:PSS
molecule. Optimizing viscosity and surface tension are needed both for acceptable jetting and for
maximizing the connectivity and hence conductivity of such films. However, it was first
184
observed with the PEDOT:PSS ink that the reduction of viscosity to a value significantly lower
than that recommended for inkjet printing resulting in better spreading of the ink – improving
both film topography and conductivity. The addition of SWCNTs assisted in improving
conductivity, as well, by providing long conductive pathways between isolated conductive
regions of PEDOT:PSS. However, the use of high-aspect-ratio materials such as CNTs
necessitates closer examination of their successful delivery to the substrate, due to the filtration
and jetting stages of the process, which remove most large particles.
If paper substrates are to be considered as realistic supports for electronic devices, the lessons
learned by observing their effect on PEDOT:PSS/SWCNT film conductivity must be considered.
Paper’s absorbent and non-conductive nature made it a major hindrance to the transport of
charge through the PEDOT:PSS/SWCNT layers; similar behaviour could be expected with any
conductive material. Internal sizing or surface coating improved the ink holdout and therefore
resulted in a higher conductivity by forming a thin layer of PEDOT:PSS/SWCNTs on the surface
of the paper. Care should be taken in the treatment of paper with charged additives which may
chemically interact with the conductive species, such as PDADMAC and PEI, in this case. The
usually negatively-charged colloids in many inks would encounter similar problems.
However, on paper, values of conductivity were estimated from the dimensions of the ink layers.
It should be emphasized that the key obstacle to objectively quantify the conductivity of printed
PEDOT:PSS on paper arises from the irregularity and unpredictability of the ink distribution.
Therefore, ideally, to optimize the performance of conductive papers, the uniformity of the paper
itself needs to be controlled as tightly as possible. It is likely that the best result that can be
offered for a given paper type is a range of conductivity values, which cannot be compared in an
absolute sense to the conductivity of typical PEDOT:PSS/SWCNT films on impermeable
surfaces. The smooth films formed by PEDOT:PSS/SWCNT ink on such impermeable surfaces,
as well as its good electrical conductivity – suitable for use as an anodic material – suggested that
the details of formulation that were refined had a major bearing on ink function. The resulting
PEDOT:PSS/SWCNT ink emphasized the need to make a careful study of ink properties in order
to maximize its functionality, even in adverse conditions, such as on the surface of paper. It also
emphasized that even minor components – in this case, surfactants – have a major bearing on ink
185
function, once again reiterating the necessity for careful formulation. The ink in its current
incarnation could be widely applied as an anode, charge transport layer, or paper-based electrode
in a wide variety of electronic devices.
The insulating ink, although not included in the final LEDs, performed its intended role as a
standalone material. BaTiO3/PMMA composite films with high dielectric constants were
successfully printed. The smoothness, uniformity, and resolution of the printed layers was
sufficient to consider this method of dielectric film fabrication a realistic alternative to vacuum-
deposition conventionally used. This ink avoided the pitfalls of sol-gel processing and illustrated
a single-step means of depositing a crystalline insulating layer. It also served as an example of
the difficulty of forming a smooth ink film while maintaining jettability. The detrimental effect
of surface tension gradients, even on the macro-scale, was shown. These issues with film
formation served as examples for refinements to ink formulation in the ZnS-bearing inks. Also,
the positive effect of layer overprinting on functionality was shown here, and replicated in the
PEDOT:PSS/SWCNT ink.
Aqueous synthesis has been successfully employed to produce brightly luminescent ZnS:Mn and
ZnS:Cu nanoparticles. However, the use of the aqueous method requires careful tailoring of the
synthesis procedures to ensure good performance – sufficient dopant and small particle size
being the primary concerns. A simple and environmentally friendly synthesis method is
practicable for this optoelectronic material, and can produce nanoparticles not only in an aqueous
system, but also readily provide a means for their redispersion in a variety of solvents. It was
found that the use of 3-mercaptopropionic acid (3-MPA) as a capping agent, at controlled pH,
temperature, and reaction time, was suitable for forming ZnS nanoparticles of sufficiently small
size to be redispersible in an aqueous inkjet ink. The particles produced using this method were
so small and well-passivated that their energy states were quantized due to the confinement of
charge carriers, resulting in bright PL emission that was greatly blue-shifted for ZnS:Cu (510 to
470 nm). Effective passivation was achieved by bonding of the sulphur atom in 3-MPA to the
S2-
vacancy sites, resulting in primarily characteristic dopant emission with minimal emission in
the blue range. Their small size and dispersity were crucial in formulation of an inkjet ink
containing a relatively high loading of solids. The small size of the particles in the ink would
186
allow for use of nozzles of significantly smaller diameter and resulting higher resolution, if
needed. The particles remained so well-passivated that even loading the dispersions with
polymer could not cause them to become unstable, making them ideal for incorporating into a
binder polymer. Printing of these nanoparticles produced brightly photoluminescent films of
controlled thickness, albeit with poor topography due to solids migration during drying. Again,
the method of overprinting assisted in reducing this effect – however, this remained as a problem
for device function. Also, the wide bandgap of the quantized particles made it difficult to
transfer charge from the electrodes and binder matrix into the luminescent centres. If used
outside of LEDs, the ZnS layers have many possible applications – a particular example of this
that was considered was the idea of anti-counterfeiting. The strong PL of the ZnS particles, at a
wavelength determined by the degree of quantization and a peak intensity determined by the
wavelength of the interrogator, could provide a means of identity verification for any printed
surface.
8.2 Recommendations & future work
In order to improve the poor functionality of the LEDs, which was the major limitation of results
produced by the work, there are several approaches which might be taken. Firstly, the
improvement of ink leveling would not only probably improve the electronic characteristics of
the devices, but it would also normalize the layer thicknesses. Having tightly-controlled and
consistent layer thicknesses would assist greatly in applying these films to different electronic
devices, such as thick-film devices like ELDs. This might involve surface treatments of
underlying layers, or minor adjustments to the ink formulations.
Leveling was particularly poor in the aqueous ZnS:Mn/PVP layers. Ideally, aqueous inks would
be used to build the entire device, due to their ease of processing and handling;
PEDOT:PSS/SWNCT layers demonstrated that they were not soluble when overprinted with
aqueous inks, as PEDOT is not water soluble. However, this implies that the emitter layer must
somehow be rendered insoluble. The use of in-place polymerization with a printed initiator and
cross-linker of a suitable binder polymer might serve to alleviate this problem.
187
Regardless of the film topography, however, charge transfer still remains the major hurdle to
better LED function. The use of another narrow-bandgap emitter material such as CdS is a
possibility for encouraging better charge transfer into the luminescent centres. If ZnS is to be
used, however, due to its avoidance of heavy metals and ease of synthesis, a power source
capable of delivering higher current is likely the best way to improve emission intensity. Also,
an examination of the problems with phase-transfer of ZnS:Mn would assist in getting this
material into the printed LED structure as well. To counter the effects of resistive losses, a
printed metal cathode might be used. On glass, a sintered metallic ink would be suitable; on
polymers, a novel means of low-temperature sintering would have to be applied, with the
assumption that it would yield films of greater conductivity than the currently-used
PEDOT:PSS/SWCNTs. Also, careful selection of a material with a lower work function to assist
in electron injection might also bring down the operating voltage. It was concluded that the
physical structure of the device was not so much of an issue to observing function as was the
absence of a well-suited testing platform, and possibly a metallic cathode.
A structure that has been considered recently as an alternative – which should be tested to
confirm the theories presented in this thesis about printed LED function – involves the use of a
silver ink which “self-sinters” in the presence of a cationic polyelectrolyte. The formulation of a
silver ink was undertaken but not included in this thesis, as no low-temperature-curing ink was
successfully produced. However, a stable ink which was jettable was prepared. Using this ink,
and another ink containing a cationic species (such as PDADMAC), a sinter-free Ag layer might
be deposited as a cathode (Magdassi et al. 2010). The remaining layers would remain the same –
ZnS:Cu/PVK and PEDOT:PSS. This would entail the formulation of a PDADMAC ink. The
interesting aspect of this study lies in the use of PDADMAC as an underlying layer for the
printed Ag. If a smooth surface – such as photo-paper – was used as the support for a thick
PDADMAC film, which itself was smooth, the possibility of realizing a paper-based printed
LED would be conceivable.
With an optimized printed LED, other inks could be considered to better its function, to construct
other layers mentioned in this document. These might include charge transport layers, blocking
layers, and an encapsulant. Because the focus of this work was simply to prove that an LED
188
could be printed by following a rigorous ink-formulation scheme, optimization was not
considered. Not only might this involve the production of new inks, it might also involve
changing deposition conditions, to avoid dust, humidity, and excessive handling. At this point,
the fabrication process was very rough and rudimentary, but collaborative efforts with expertise
in physics, materials science, and electrical engineering might assist in elaborating upon the
opportunities and limitations associated with this extremely young technology.
Regardless of the printed LEDs’ mediocre performance, this technology still presents many
unique possibilities and advancements in the fields of ink development and printed electronics.
The limitations on resolution caused by drop size mean that high-resolution electronics
manufacturing, on a scale similar to that of photolithography/shadowmask techniques, is not
practical. Also, the ever-present issue of orthogonality means that overlaying of many layers –
such as, in a display, the emissive material, transistor layer, colour filter, and so forth – becomes
progressively less feasible. Therefore, a good possibility for application of the LEDs prepared in
this work is for low-information-content displays, or for lighting. The flexibility of the printed
materials, and the infinite patterning possibilities, mean that printed passive-matrix, single-pixel,
or lamp-type LEDs are realistic applications. As demands for low-energy lighting and lighted
signage increase, simple, rapid fabrication and especially patterning techniques such as those
outlined here may become dominant. Even in the context of more complex non-emissive LCDs,
low-energy, large-area, durable backlighting is an area of intense focus; most LCDs now contain
LED backlights. With LCDs currently ubiquitous with consumer displays, the rapid fabrication
of simple backplanes represents a unique opportunity to utilize printed LED technology. In
many consumer electronics, also, backlit keyboards, accent lighting, and indicators currently use
filtered fluorescent, incandescent, or LED lighting, behind a patterned stencil. Direct printing of
such accents avoids material and energy wastage. The use of environmentally inert and non-
cytotoxic ZnS as the emissive material also presents a unique advantage over Cd and Pb-based
quantum dots. With a heavy focus on recycling and reuse of electronic components, LEDs based
solely on polymers and inert inorganics are of particular commercial interest. Finally, with the
groundwork laid for novel substrates incorporating these materials – paper being the most
notable – the idea of biodegradeable and disposable electronics may be realized.
189
This work has ideally served as a stepping-stone for the development of fully-printed electronics
in outlining the challenges and opportunities associated with their production. While not
idealizing either materials or structure, the LEDs produced demonstrated the process of materials
and ink development, and the bearing these stages have on the eventual production of a printed
device. It is anticipated that such incremental advancements in the field, from the perspective of
printing science and technology, will contribute to the refinement and optimization of future
work on such devices, and that the detailed and rigorous map from raw materials to printed
electronic layers will assist others in their respective studies in this fascinating field.
190
References
Adachi, D.; Hama, T.; Morimoto, T.; Toyama, T; Okamoto, H. (2008): Orange
electroluminescence from chemically synthesized zinc sulfide nanocrystals doped with
manganese, J. Non-Crys. Sol. 354, 2740.
Adachi, D.; Hama, T.; Toyama, T; Okamoto, H. (2007): Electroluminescent properties of
chemically synthesized zinc sulfide nanocrystals doped with manganese, J. Mater. Sci.:
Mater. Elec. 10, 1.
Adamovich, V.; Cordero, S.; Djurovich, P.; Tamayo, A.; Thompson, M.; d’Andrade, B.; Forrest,
S. (2003): New charge-carrier blocking materials for high-efficiency OLEDs, Org.
Electron. 4(2-3), 77.
Agarwal, M.; Lvov, Y.; Varahramyan, K. (2006): Conductive wood microfibers for smart paper
through layer-by-layer nanocoating, Nanotech.17, 5319.
Aguey-Zinsou, K.; Ares-Fernandez, J. (2008): Synthesis of colloidal magnesium: a near room
temperature store for hydrogen, Chem. Mater. 20(2), 376.
Al-Dahoudi, N.; Aegerter, M. (2006): Comparative study of transparent conductive In2O3:Sn
(ITO) coatings made using a sol and a nanoparticle suspension, Thin Solid Films 502,
193.
Allen, M.; Aronniemi, M.; Mattila, T.; Alastalo, A.; Ojanpera, K.; Suhonen, M.; Seppa, H.
(2008): Electrical sintering of nanoparticle structures, Nanotechnology 19(17), 1.
Althues, H.; Palkovits, R.; Rumplecker, A.; Simon, P.; Sigle, W.; Bredol, M.; Kynast, U.;
Kaskel, S. (2006): Synthesis and characterization of transparent luminescent
ZnS:Mn/PMMA nanocomposites, Chem. Mater. 18(4), 1068.
American Ceramic Society (2010): A day made of glass – made possible by Corning.
<http://ceramics.org/ceramictechtoday/tag/gen-10-glass/>
Andersson, P.; Nilsson, D.; Svensson, P.; Chen, M.; Malmstrom, A.; Remonen, T.; Kugler, T.;
Berggren, M. (2002): Active matrix displays based on all-organic electrochemical smart
pixels printed on paper, Adv. Mater. 14(20), 1460.
Angelo, P.; Cole, G.; and Farnood, R (2009): The effect of paper components on the conductivity
of paper-based PEDOT:PSS films, Proc. 7th
Int. Paper & Coating Chem. Symp., 283.
Angelo, P.; Farnood, R. (2010a): Poly(3,4-ethylenedioxythiophene):poly(styrene sulfonate)
inkjet inks doped with carbon nanotubes and a polar solvent: the effect of formulation
and adhesion on conductivity, J. Adhes. Sci. Tech, 24(3), 643.
Angelo, P; Farnood, R. (2010b): Deposition of PEDOT/SWCNT composites by inkjet printing,
Proc. Mater. Res. Soc. Symp. T.
Angelo, P; Farnood, R. (2010c): Inkjet-printed PEDOT:PSS/SWCNTs on paper: substrate
effects on conductivity, Proc. Mater. Res. Soc. Symp. G.
191
Angelo, P.; Farnood, R. (2011): Photoluminescent inkjet ink containing ZnS:Mn nanoparticles as
pigment, J. Exper. Nanosci. 6(5), 473.
Angelo, P.; Cole, G.; Sodhi, R.; Farnood, R. (2012): Conductivity of inkjet-printed PEDOT:PSS-
SWCNTs on uncoated papers, Nordic Pulp & Paper Res. J. (in press).
Angelo, P.; Farnood, R.; Chlebowski, M (2012): Mn incorporation into ZnS nanoparticles
prepared in aqueous solution, J. Nanoparticle Res. (submitted).
Ardebili, H.; Pecht, M. (2009): Encapsulation technologies for electronic applications, William
Andrew Publishing, Norwich, New York.
Asberg, P.; Inganas, O. (2003): Hydrogels of a conducting conjugated polymer as 3-D enzyme
electrode, Biosensors & Bioelectronics 19, 199.
ASTM Standard D150 – 98 (2004): Standard Test Methods for AC Loss Characteristics and
Permittivity (Dielectric Constant) of Solid Electrical Insulation, ASTM International,
Subcommittee D09.12, Volume 10.01, West Conshohocken, Pennsylvania.
ASTM Standard D5402-06 (2006): Standard Practice for Assessing the Solvent Resistance of
Organic Coatings Using Solvent Rubs, ASTM International, Subcommittee D01.46,
Volume 06.02, West Conshohocken, Pennsylvania.
Atkinson, A.; Doorbar, J.; Hudd, A.; Segal, D.; White, P. (1997): Continuous inkjet printing
using sol-gel ceramic inks, J. Sol-Gel Sci. Tech. 8, 1093.
Aubry, N.; Singh, P.; Janjua, M.; Nudurpati, S. (2008): Micro- and nanoparticles self-assembly
for virtually defect-free, adjustable monolayers, Proc. Nat. Acad. Sci. 105(10), 3711.
Aytouna, M.; Bartolo, D.; Wegdam, G.; Bonn, D.; Rafai, S. (2010): Impact dynamics of
surfactant laden drops: dynamic surface tension effects, Exp. Fluids 48, 49.
Bakasybramanian, N.; Subrahmanyam, A. (1991): Studies on evaporated indium tin oxide
(ITO)/silicon junctions and an estimation of ITO work function, J. Electrochem. Soc.
138, 322.
Bakueva, L.; Musikhin, S.; Hines, M.; Chang, T.; Tzolov, M.; Scholes, G.; Sargent, E. (2003):
Size-tunable infrared electroluminescence from PbS quantum dot nanocrystals in
semiconductive polymer, App. Phys. Lett. 82(17), 2895.
Berger, L. (1996): Semiconductor materials, CRC Press, Boca Raton.
Bergeron, V.; Bonn, D.; Martin, J. (2000): Controlling droplet deposition with polymer additives,
Nature 405, 772.
Bhandari, S.; Deepa, M.; Srivasta, A.; Joshi, A.; Kant, R (2009) : Electrochemical and optical
properties of the poly(3,4-ethylenedioxythiophene) film electropolymerized in an
aqueous sodium dodecyl sulfate and lithium tetrafluoroborate medium, J. Phys. Chem. B
113, 9416.
Bhargava, R.; Gallagher, D.; Hong, X.; Nurmikko, A. (1994): Optical properties of Mn-doped
nanocrystals of ZnS, Phys. Rev. Lett. 72(3), 416.
192
Biswas, P.; De, A.; Dua, L.; Chkoda, L. (2006): Work function of sol–gel indium tin oxide (ITO)
films on glass, Appl. Surf. Sci. 253, 1953.
Blackburn, J.; Barnes, T.; Beard, M.; Kim, Y.; Tenent, R.; McDonald, T.; To, B.; Coutts, T.;
Heben, M. (2008): Transparent conductive single-walled carbon nanotube networks with
precisely tunable ratios of semiconducting and metallic nanotubes, ACS Nano 2(6), 1266.
Blazdell, P.; Evans, J.; Edirishinghe, M.; Shaw, P.; Binstead, M. (1995): The computer aided
manufacture of ceramics using inkjet printing, J. Mater. Sci. Lett. 14(22), 1562.
Borovitskaya, E.; Shur, M. (eds.) (2002): Quantum dots, World Scientific, Hackensack, New
Jersey.
Bowers, R.; Melamed, N. (1955): Luminescent centres in ZnS:Cu, Cl phosphors, Phys. Rev.
Lett. 99(6), 1781.
Brandrup, J.; Immergut, E. (eds.) (1989): Polymer handbook, 3rd ed., Wiley-Interscience, New
York, New York.
Bredol, M.; Dieckoff, H (2010): Materials for powder-based AC electroluminescence, Materials
3, 1353.
Brinker, C.; Scherer, G. (1990): Sol-gel science: the physics and chemistry of sol-gel processing,
Academic Press, New York, New York.
Bronzcyk, A. (2003): Fabrication of an organic light-emitting diode on a paper substrate,
NSF/SRC Engineering Research Center for Environmentally Benign Semiconductor
Manufacturing, Thrust C.
Brus, L. (1986): Zero-dimensional excitons in semiconductor clusters, J. Quantum Electron. 22,
1909.
Bube, R. (1953): Electronic transitions in the luminescence of zinc sulfide phosphors, Phys. Rev.
90(1), 70.
Bucknall, D. (ed.) (2005): Nanolithography and patterning techniques in microelectronics,
Woodhead Publishing, Cambridge.
Bulanyi, M.; Polezhaev, B.; Prokofev, T. (1998): The nature of Mn luminescence centers in ZnS
single crystals, Semiconductors 32(6), 603.
Bulanyi, M.; Kovalenko, A.; Polezhaev, B. (2002): Spectra of manganese emission centers in
zinc sulphide, J. Appl. Spectroscopy 69(5), 747.
Caglar, U. (2009): Studies of inkjet printing technology with focus on electronic materials,
doctoral thesis, Tampere University of Technology, Finland.
Calvert, P. (2001): Inkjet printing for materials and devices, Chem. Mater. 13, 3299.
Campbell, S. (2001): The science and engineering of microelectronic fabrication, Oxford
University Press, Oxford.
Capek, I. Nanocomposite structures and dispersions. Elsevier (2006): Amsterdam
193
Carter, E.; Gardiner, F. (eds.) (2009): Polymer electronics: a flexible technology, Smithers,
London, England.
Caruso, F. (ed.) (2004): Colloids & colloid assemblies, Wiley, Weinheim, Germany.
Chander, H. (2005): Development of nanophosphors – a review, Mater. Sci. Eng. R49, 113.
Chang, T.; Guo, W.; Guo, X. (2005): Buckling of multiwalled carbon nanotubes under axial
compression and bending via a molecular mechanics model, Phys. Rev. B 72(6), 64101.
Chang, P. (2005): Improving performance in oligothiophene based semiconductor devices via
functionality, PhD thesis, University of California, Berkeley.
Changneng, Z.; Mingtai, W.; Fang, L.; Mingguang, K.; Li, G.; Weiwei, X.; Xiaoguang, Z.;
Kongjia, W. (2005): A polymeric/inorganic nanocomposite for solid-state dye-sensitized
solar cells, Plasma Sci. Tech. 7(4), 2692.
Chen, Z.; Cotterell, C.; Wang, W. (2002): The fracture of brittle thin films on compliant
substrates in flexible displays, Eng. Fracture Mech. 69(5), 597.
Chen, Z.; Kobashi, K.; Rauwald, U.; Booker, R.; Fan, H.; Hwang, W.; Tour, J. (2006): Soluble
ultra-short single-walled carbon nanotubes, J. Amer. Chem. Soc. 128, 10568.
Chen, J.; Cranton, W.; and Fihn, M. (2011): Handbook of visual display technology, Springer,
New York.
Cheng, K.; Yang, M.; Chiu, W.; Huang, C.; Chang, J.; Ying, T.; Yang,Y. (2005): Ink-jet
printing, self-assembled polyelectrolytes, and electroless plating: low cost fabrication of
circuits on a flexible substrate at room temperature, Macromol. Rapid Comm. 26(4), 247.
Cheng, Y.; Yang, S.; Hsu, C. (2009): Synthesis of conjugated polymers for organic solar cell
applications, Chem. Rev. 109(11), 5868.
Chien, A.; Speck, J.; Lange, F.; Daykin, A.; Levi, C. (1995): Low temperature/low pressure
hydrothermal synthesis of barium titanate: powder and heteroepitaxial thin films, Journal
of Materials Research 10(7), 1784.
Cho, K.; Cha, T. (2009): Quantum dot ink composition for inkjet printing and electronic device
using the same, US Patent #20090314991.
Cho, Y.; Kim, H.; Yi, G.; Hong, J.; Jang, S.; Yang, S. (2009): Dispersion stabilization of
conductive transparent oxide nanoparticles, Coll. & Surf. A: Physicochem. Eng. Aspects
336, 88.
Cho, Y.; Yi, G.; Hong, J.; Jang, S.; Yang, S. (2006): Colloidal indium tin oxide nanoparticles for
transparent and conductive films, Thin Sol. Films 515, 1864.
Chow, E.; Herrmann, J.; Barton, C.; Raguse, B.; Wieczorek, L. (2009): Inkjet-printed gold
nanoparticle chemiresistors: Influence of film morphology and ionic strength on the
detection of organics dissolved in aqueous solution, Anal. Chem. Acta. 632(1), 135.
Chung, S.; Jang, J.; Cho, J. (2010): All-inkjet-printed organic thin-film transistors with silver
gate, source/drain electrodes, Jap. J. Appl. Phys. 50(3), S1.
194
Chung, S.; Kim, S.; Kwon, S. (2011): All-inkjet-printed organic thin-film transistor inverter on
flexible plastic substrate, IEEE Electron Dev. Lett. 32(8), 1134.
Cibis, D.; Krueger, K. (2005): DOD printing of conductive silver tracks, 1st Int. Conf. Ceram.
Interconnect Ceram. Microsys. Techn., Baltimore, Maryland.
Clymer, N.; Asaba, S. (2004): A new approach for understanding dominant design: the case of
the ink-jet printer, J. Eng. Tech. Management 25, 137.
Coe-Sullivan, S. (2005): Hybrid organic-quantum dot thin film structures and devices, PhD
thesis, Department of Electrical Engineering and Computer Science, Massachusetts
Institute of Technology, Boston, Massachusetts.
Coe-Sullivan, S.; Woo, W.; Steckel, J.; Bawendi, M.; Bulović, V. (2003): Tuning the
performance of hybrid organic/inorganic quantum dot light-emitting devices, Org.
Electron. 4, 123.
Colclough, W. (1968): The chemistry of solvents used for ultrasonic cleaning, Ultrasonics 6(1),
21.
Collins, P.; Avouris, P. (2000): Nanotubes for electronics, Scientific American 283(6), 62.
Colvin, V.; Schlamp, M.; Alivisatos, A. (1994): Light-emitting diodes made from cadmium
selenide nanocrystals and a semiconducting polymer, Nature 370(4), 354.
Cranton, W.; Stevens, R. Thomas, C. (1998): AC thin film electroluminescent device, US Patent
5742322.
Craster, R.; Matar, O.; Sefiane, K. (2009): Pinning, retraction and terracing of evaporating
droplets containing nanoparticles, Langmuir 25(6), 3601.
Crispin, X.; Jakobsson, F.; Crispin, A.; Grim, P.; Andersson, P.; Volodin, A.; van Haesendonck,
C.; Van der Auweraer, M; Salaneck, W.; and Berggren, M. (2006): The origin of the high
conductivity of poly(3,4-ethylene dioxythiophene)-poly(styrene sulfonate) (PEDOT-PSS)
plastic electrodes, Chem. Mater. 18, 4354.
Crispin, X.; Marciniak, S.; Osikowicz, W.; Zotti, G.; Denier van der Gon, A.; Louwet, F.;
Fahlman, M.; Groenedaal, L.; De Schryver, F.; and Salaneck, W. (2003): Conductivity,
morphology, interfacial chemistry, and stability of poly(3,4-ethylene dioxythiophene)-
poly(styrene sulfonate): a photoelectron spectroscopy study, J. Poly. Sci. B: Poly. Phys.
41, 2561.
Crone, B.; Campbell, I.; Davids, P.; Smith, D. (1998): Charge injection and transport in single-
layer organic light-emitting diodes, Appl. Phys. Lett. 73, 3162.
Croucher M.; Hair, M. (1989): Design criteria and future directions in inkjet ink technology,
Indus. Eng. Chem. Res., 28(11), 1712.
Cui, W.; Lu, W.; Zhang, Y.; Lin, G.; Wei, T.; Jiang, L. (2010): Gold nanoparticle ink suitable for
electric-conductive pattern fabrication using in ink-jet printing technology, Coll. & Surf.
A: Physicochem. Eng. Aspects 358(1-3), 35.
195
Cullity, B.; and Stock, S. (2001): Elements of X-ray diffraction, 3rd edition, Prentice-Hall Inc.,
New Jersey.
Curling, C. (2006): Prototype line for manufacturing flexible active matrix, USDC 5th Flex.
Disp. Microelec. Conf.
Cutnell, J.; Johnson, K. (2008): Physics, 8th
edition, John Wiley & Sons, Hoboken, New Jersey.
D’Errico, G.; Lellis, M.; Mangiapia, G.; Tedeschi, A.; Ortona, O.; Fusco, S.; Borzacchiello, A.;
Ambrosio, L. (2008): Structural and mechanical properties of UV-photo-cross-linked
poly(N-vinyl-2-pyrrolidone) hydrogels, Biomacromolecules 9, 231.
Dabbousi, B.; Rodrigues-Viejo, J.; Mikulec, F.; Heine, J.; Mattoussi, H.; Ober, R.; Jensen, K.;
Bawendi, M. (1997): CdSe-ZnS core-shell quantum dots: synthesis and characterization
of a size series of highly luminescent nanocrystallites, J.Phys. Chem. B 101, 9463.
Daniel, A.; and Fotheringham, A. (2007): Microstructured conducting polymer coatings on
plastic strip for contact electrodes: Formulation and tuning of sheet resistance, J. Appl.
Polym. Sci. 104, 234.
Daniel, R.; Berg, J. (2006): Spreading on and penetration into thin, permeable print media:
application to ink-jet printing, Adv. Coll. Interface Sci. 123-126, 439.
Davison, G.; Lane, B. (eds.) (2003): Additives in water-borne coatings, Royal Society of
Chemistry, Cambridge.
de Gans, B.; Duineveld, P.; and Schubert, U. (2004): Inkjet printing of polymers: state of the art
and future developments, Adv. Mater. 16(3), 203.
Dearden, A.; Smith, P.; Shin, D.; Reis, N.; Derby, B.; O’Brien, P. (2005): A low curing
temperature silver ink for use in ink-jet printing and subsequent production of conductive
tracks, Macromol. Rapid Comm. 26, 315.
Delgado, A.; Gonzalez-Caballero, F.; Hunter, R.; Koopal, L.; Lyklema, J. (2005): Measurement
and interpretation of electrokinetic phenomena, Pure Appl. Chem. 77(10), 1753.
Denneulin, A.; Blayo, A.; Bras, J.; and Neuman, C. (2008): PEDOT:PSS coating on specialty
papers: process optimization and effects of surface properties on electrical performances,
Prog. Org. Coatings 63, 87.
Denneulin, A.; Blayo, A.; Bras, J.; and Neuman, C. (2011): Infra-red assisted sintering of inkjet
printed silver tracks on paper substrates, J. Nanoparticle Res. 13(9), 3815.
Derby, B.; Reis, N. (2003): Inkjet printing of highly loaded particulate suspensions, MRS Bull.
28, 816.
Deslandes, Y.; Pleizier, G.; Poire, E.; Sapieha, S.; Wertheimer, M.; and Sacher, E. (1998) : The
surface modification of pure cellulose paper induced by low-pressure nitrogen plasma
treatment, Plasmas & Polym. 3(2), 61.
Destriau, G. (1936): Research on the scintillations of zinc sulphide to rays, J. Phys. Chem. 33,
587.
196
Di Risio, S.; Yan, N. (2007): Piezoelectric ink-jet printing of horseradish peroxidase: effect of
ink viscosity modifiers on activity, Macromol. Rapid Comm. 28, 1934.
Di Risio, S.; Yan, N. (2009): Adsorption and inactivation behaviour of horseradish peroxidise on
cellulosic fiber surfaces, J. Colloid & Interface Sci. 338, 410.
Dimitrakopoulos, C.;Mascaro, M. (2001): Organic thin-film transistors: a review of recent
advances, IBM J. Res. Dev., 45,11.
Dimitrijev, S. (2005): Principles of semiconductor devices, Oxford University Press, Oxford.
Ding, X.; Li, Y.;Wang, D.;Yin, Q. (2004): Fabrication of BaTiO3 dielectric films by direct
inkjet printing, Ceram. Inter. 30, 1885.
Dobbelin, M.; Marcilla, R.; Salsamendi, M.; Pozo-Gonzalo, C.; Carrasco, P.; Pomposo, J.; and
Mecerreyes, D. (2007): Influence of ionic liquids on the electrical conductivity and
morphology of PEDOT:PSS films, Chem. Mater. 19, 2147.
Dong, H.; Carr, W.; and Morris, J. (2006): An experimental study of drop-on-demand drop
formation, Phys. Fluids 18, 1.
Dresselhaus, M.; Dresselhaus, G.; Avouris, P. (eds.) (2001): Carbon nanotubes, Springer, New
York, New York.
Eastoe, J.; Dalton, J. (2000): Dynamic surface tension and adsorption mechanisms of surfactants
at the air-water interface, Adv. Coll. Interf. Sci. 85, 103.
Eftekhari, A. (2010): Nanostructured conductive polymers, John Wiley & Sons, Ltd, Hoboken,
New Jersey.
Eisenstein, E. (1997): The printing press as an agent of change, Cambridge University Press,
United Kingdom.
Electronics.ca Publications (2011): Flat panel displays: global strategic business report.
Fakhri, N.; Tsyboulski, D.; Cognet, L.; Weisman, R.; Pasquali, M. (2009): Diameter-dependent
bending dynamics of single-walled carbon nanotubes in liquids, Proc. National Academy
of Sci., 106(34), 14219.
Fan, B.; Mei, X.; and Ouyang, J. (2008) : Significant conductivity enhancement of poly(3,4-
ethylenedioxythiophene):poly(styrenesulfonate) films by adding anionic surfactants into
polymer solution, Macromolecules 41, 5971.
Fan, Z.; Wei, T.; Luo, G.; and Wei, F. (2005): Fabrication and characterization of multi-walled
carbon nanotubes-based ink, J. Mater. Sci. 40, 5075.
Fardim, P.; Holmbom, B. (2005): ToF-SIMS imaging: a valuable chemical microscopy technique
for paper and paper coatings, Appl. Surf. Sci. 249, 393.
Fehse, K.; Walzer, K.; Leo, K.; Lovenvich, W. And Elschner, A. (2007): Highly conductive
polymer anodes as replacements for inorganic materials in high-efficiency organic light-
emitting diodes, Adv. Mater. 19, 441.
197
Filenkova, A.; Acosta, E.; Brodersen, P.; Sodhi, R.; Farnood, R. (2010): Distribution of inkjet
ink componenets via ToF-SIMS imaging, Surface & Interface Anal. 43(1-2), 576.
Forest Products Laboratory (FPL) (1964): Ignition and charring temperatures of wood, Forest
Service, U. S. Department of Agriculture. Washington D.C.
Fujifilm-Dimatix (2006a): Dimatix Materials Printer DMP-2800 Series User Manual.
Fujifilm-Dimatix (2006b): Fujifilm-Dimatix ink tutorial.
Fukai, J.; Shiiba, Y.; Yamamoto, T.; Miyatake, O.; Poulikakos, D.; Megaridis, C.; Zhao, Z.
(1995): Wetting effects on the spreading of a liquid droplet colliding with a flat surface:
experiment and modeling, Phys. Fluids 7, 236.
Fukai, J.; Zhao, Z.; Poulikakos, D.; Megaridis, C.; Miyatake, O. (1993): Modeling of the
deformation of a liquid droplet impinging upon a flat surface, Phys. Fluids A, 5, 2588.
Gan, Y.; Liu, J.; Zeng, S. (2006): Transparent conductive indium tin oxide film fabricated by
dip-coating technique from colloid precursor, Surf. Coat. Tech. 201(1-2), 25.
Gane, P. (2004): Absorption properties of coatings: a selected overview of absorption criteria
derived from recent pore network modeling, J. Disper. Sci. Tech., 25, 389.
Gaponenko, S. (1998): Optical properties of semiconductor nanocrystals, Cambridge University
Press, Cambridge, England.
Gaponik, N.; Talapin, D.; Rogach, A. (1999): A light-emitting device based on a CdTe
nanocrystal/polyaniline composite, Phys. Chem. Chem. Phys. 1, 1787.
Garnett, E.; Ginley, D. (2005): Electrical and morphological properties of inkjet-printed
PEDOT:PSS films, J. Undergraduate Res. 15, 24.
Geim, A.; Novoselov, K. (2007): The rise of graphene, Nature Mater. 6(3), 183.
German Flat Panel Display Forum (DFF) (2008): European technology: flat panel displays, 6th
ed.
Gracia-Espino, E.; Sala, G.; Pino, F.; Halonen, N.; Luomahaara, J.; Maklin, J.; Toth, G.; Kordas,
K.; Jantunen, H.; Terrones, M.; Helisto, P.; Seppa, H.; Ajayan, P.; Vajtai, R. (2010):
Electrical transport and field-effect transistors using inkjet-printed SWCNT films having
different functional side groups, ACS Nano 4(6), 3318.
Gullichsen, J. and Paulapuro, H (1999): Papermaking science and technology 6A: chemical
pulping, TAPPI & Fapet Oy, Helsinki.
Hadley, A. (n.d.) CombineZ software homepage.
http://www.hadleyweb.pwp.blueyonder.co.uk/
Haider, M.; Pitters, J.; DiLabio, G.; Livadaru, L.; Mutus, J.; Wolkow, R. (2009): Controlled
coupling and occupation of silicon atomic quantum dots at room temperature, Phys. Rev.
Lett. 102, 046805.
Ham, H.; Choi, Y.; Chee, M.; Cha, M.; and Chung. I. (2008): PEDOT-PSS/singlewall carbon
nanotubes composites, Poly. Eng. Sci. 48, 1.
198
Hamaker, H. (1937): The London-van der Waals attraction between spherical particles, Physica
A 10, 1058.
Hamberg, I.; Granqvist, C. J. Appl. Phys. 60 (1986): R123-R159.
Haranath, D.; Chander, H.; Bhalla, N.; Sharma, P.; Sood, K. (2005): Surface distribution studies
and improved photoluminescence characteristics of silica coated ZnS:Mn nanophosphor
layers, Appl. Phys. Lett., 86(20), 201904.
Harizanov, O.; Harizanova, A.; and Ivanona, T. (2004): Formation and characterization of sol–
gel barium titanate, Mater. Sci. Eng. B106, 191.
Harrison, P. (2000): Quantum wells, wires, and dots: theoretical and computational physics,
J.Wiley & Sons, Hoboken, New Jersey.
Haverinen, H.; Myllyla, R.; Jabbour, G. (2009): Inkjet printing of light-emitting quantum dots,
App. Phys. Lett. 94, 073108.
Haverinen, H.; Myllyla, R.; Jabbour, G. (2010): Inkjet printed RGB quantum dot-hybrid LED, J.
Display Tech. 6(3), 87.
Haverinen, H. (2010): Inkjet-printed quantum dot light-emitting devices – towards display
applications, PhD thesis, Faculty of Technology, Department of Electrical and
Information Engineering, University of Oulu.
He, G.; Pfeiffer, M.; Leo, K.; Hofmann, M.; Birnstock, J.; Pudzich, R.; Salbeck, J. (2004): High-
efficiency and low-voltage p-i-n electrophosphorescent organic light-emitting diodes with
double-emission layers, Appl. Phys. Lett. 85(17), 3911.
Heaney, M. (2000): Electrical conductivity and resistivity, In: The measurement,
instrumentation, and sensors handbook, CRC Press, Oxford, 43-1.
Hebner, T.; Wu, C.; Marcy, D.; Lu, M.; Sturm, J. (1998): Ink-jet printing of doped polymers for
organic light emitting devices, Appl. Phys. Lett. 72, 519.
Hensema, E. (1997): Sizing and ink jet printing of European multi-purpose office paper, TAPPI
Proc.: Eng. Papermakers Conf. 851.
Hieronymas A. (2002): Thin film inorganic light-emitting diode, US Patent 20020153830.
Higashimura, T.; Teranishi, H.; and Sawamoto, M. (1980): Living cationic polymerization of n-
vinylcarbazole with iodine, Poly. J. 12(6), 393.
Hines, M.; Guyot-Sionnest, P. (1996): Synthesis and characterization of strongly luminescing
ZnS-capped CdSe nanocrystals, J. Phys. Chem. 100, 468.
Holman, R.; Cima, M.; Uhland, S.; Sachs, E. (2002): Spreading and infiltration of inkjet-printed
polymer solution droplets on a porous substrate, J. Coll. Interface Sci. 249, 432.
Hong, C. (2000): Inkjet printed copper source/drain metallization for amorphous silicon thin-film
transistors, IEEE Elec. Dev. Lett.21(8), 384.
Hostetler, M.; Wingate, J.; Zhong, C.; Harris, J.; Vachet, R.; Clark, M.; Londono, J.; Green, S.;
Stokes, J.; Wignall, G.; Glish, G.; Porter, M.; Evans, N.; Murray, R. (1998):
199
Alkanethiolate gold cluster molecules with core diameters from 1.5 to 5.2 nm: core and
monolayerproperties as a function of core size, Langmuir 14, 17.
Hoth, C.; Choulis, S.; Schilinksky, P.; Brabec, C. (2007): High photovoltaic performance of
inkjet-printed polymer:fullerene blends, Adv. Mater. 19, 3973.
Hsiao, Y.; Whang, W.; Chen, C.; and Chen, Y. (2008): High-conductivity poly(3,4-
ethylenedioxythiophene): poly(styrenesulfonate) film for use in ITO-free polymer solar
cells, J. Mater. Chem. 18, 5948.
Huang, A. (2009): Recent advances in dielectric materials, Nova Science Publishers, Commack,
New York.
Huang, D.; Liao, F.; Molesa, S.; Redinger, D.; Subramanian, V. (2003): Plastic-compatible low-
resistance printable gold nanoparticle conductors for flexible electronics, J. Electrochem.
Soc.150, 412.
Huang, L.; Huang, Y.; Liang, J.; Wan, X.; Chen, Y. (2011): Graphene-based conducting inks for
direct inkjet printing of flexible conductive patterns and their applications in electric
circuits and chemical sensors, Nano Res. 4(7), 675.
Hutchison, D.; Morril, N.; Aten, Q.; Turner, B.; Jensen, B.; Howell, L.; Vanfleet, R.; Davis, R.
(2010): Carbon nanotubes as a framework for high aspect ratio MEMS fabrication, J.
MEMS, 19(1), 75.
Hwang, J.; Oh, M.; Kim, I.; Lee, J.; and Ha, C. (2005): Preparation and characterization of ZnS
based nanocrystalline particles for polymer light emitting diodes,
Hwang, Y.; Son, Y.; Lee, Y. (2007): Printed circuit patterns of conductive polymer, Mol. Cryst.
Liq. Cryst. 472, 113.
Igarashi, T.; Isobe, T.; Senna, M. (1997): EPR study of Mn2+
electronic states for the nanosized
ZnS:Mn powder modified by acrylic acid, Phys. Rev. B 56(11), 6444.
Jaeger, R. (2002): Introduction to microelectronic fabrication, Prentice-Hall, Upper Saddle River.
Jaeger, R.; Balock, T. (2003): Microelectronic circuit design, McGraw-Hill, New York.
Jang, D.; Kim, D.; Moon, J. (2009): Influence of fluid physical properties on ink-jet printability,
Langmuir 25, 2629.
Jang, S.; Seo, Y.; Choi, J. Kim, T.; Cho, J.; Kim, S.; Kim, D. (2010): Sintering of inkjet printed
copper nanoparticles for flexible electronics, Scripta Materialia 62(5), 258.
Jansen, K.; Agterof, W.; and Mellema, J. (2001): Viscosity of surfactant-stabilized emulsions, J.
Rheol. 45(6), 1359.
Jensen, G.; Krause, C.; Sotzing, G.; Rusling, J. (2011): Inkjet-printed gold nanoparticle
electrochemical arrays on plastic: application to immunodetection of a cancer biomarker
protein, Phys. Chem. Chem. Phys. 13(11), 4888.
200
Jeong, M.; Byeon, C.; Cha, O.; Jeong, H.; Han, J.; Choi, Y.; An, K.; Oh, K.; Kim, K.; Lee, Y.
(2008): Purity measurement of single-walled carbon nanotubes by UV-vis-nIR absorption
spectroscopy and thermogravimetric analysis, Nano Brief Rep. Rev. 3(2), 101.
Jeong, S.; Song, H.; Lee, W.; Choi, Y.; Ryu, B. (2010): Preparation of aqueous Ag Ink with
long-term dispersion stability and its inkjet printing for fabricating conductive tracks on a
polyimide film, J. App. Phys. 108, 1.
Jia, Q.; Miranda, F.; Oates, D.; Xi, X. (2000): Materials issues for tunable RF and microwave
devices, Proc. Mater. Res. Soc. 603, 181.
Jiang, D.; Cao, L.; Liu, W.; Su, G.; Qu, H.; Sun, Y.; and Dong, B. (2009): Synthesis and
luminescence properties of core/shell ZnS:Mn/ZnO nanoparticles, Nanoscale Res. Lett. 4,
78.
Jindal, Z.; Verma, N. (2008): Photoluminescent properties of ZnS:Mn nanoparticles with in-built
surfactant, J. of Mater. Sci. 43, 6539.
Jonsson, S.; Birgerson, J.; Crispin, X.; Greczynski, G.; Osikowicz, W.; van der Gon, A.;
Salaneck, W.; Fahlman, M. (2003): The effects of solvents on the morphology and sheet
resistance in poly(3,4-ethylenedioxythiophene)-polystyrenesulfonic acid (PEDOT-PSS)
films, Synth. Metals 139(1), 1.
Jung, H.; Cho, S.; Joung, J.; Oh, Y. (2007): Studies on inkjet-printed conducting lines for
electronic devices, J. Electron. Mater. 36(9), 1211.
Jung, S.; Baeg, K.; Yoon, S.; You, I.; Lee, J.; Kim, Y.; Noh, Y. (2010): Low-voltage-operated
top-gate polymer thin-film transistors with high capacitance poly(vinylidene fluoride-
trifluoroethylene)/poly(methyl methacrylate) dielectrics, J. Appl. Phys. 108, 102810.
Kalidindi, S.; Jagirdar, B. (2009): Highly monodisperse colloidal magnesium nanoparticles by
room temperature digestive ripening, Inorg. Chem. 48(10), 4524.
Kanungo, M.; Lu, H.; Malliaras, G.; Blanchet, G. (2009): Suppression of metallic conductivity of
single-walled carbon nanotubes by cycloaddition reactions, Science 323, 234.
Karar, N.; Chander, H.; Shivaprasad, S. (2004a): Enhancement of luminescent properties of
ZnS:Mn nanophosphors by controlled ZnO capping, Appl. Phys. Lett. 85(21), 5058.
Karar, N.; Raj, S.; Singh, F. (2004b): Properties of nanocrystalline ZnS:Mn, J. Crys. Growth 268,
585.
Karlin, K. (2005): Progress in inorganic chemistry v. 54, Wiley-Interscience, Hoboken, New
Jersey.
Karsa, D. (ed.) (2003): Surfactants in polymers, coatings, inks and adhesives, Blackwell, Oxford.
Kavoosi, B. (2001): Active imaging and visual information on paper and paper-based products,
Printed Microsystems on Paper, 1-106.
Kelly, M. (1995): Low-dimensional semiconductors, Clarendon Press, Gloustershire, England.
Kerner, E. (1956): The electrical conductivity of composite media, Proc. Phys. Soc. B 69, 802.
201
Kim, I.; Kim, J. (2010): The effect of reduction atmospheres on the sintering behaviors of inkjet
printed Cu interconnectors, J. App. Phys. 108, 102807.
Kim, J.; Cho, J.; Chung, S.; Kwak, J.; Lee, C.; Hong, Y.; Kim, J. (2009): Inkjet-printed silver
gate electrode and organic dielectric materials for bottom-gate pentacene thin-film
transistors, J. Korean Phys. Soc. 54(1), 518.
Kim, J.; Ma, P. (2011): Carbon nanotubes for polymer reinforcement, CRC Press, Boca Raton,
Florida.
Kim, J.; Park, S.; Jeong, T.; Bae, M.; Song, S.; Lee, J.; Han, I.; Jung, D.; Yu, S. (2010): Paper as
a substrate for inorganic powder electroluminescent devices, IEEE Trans. Electron. Dev.
57(6), 1470.
Kim, J.; Jung, J.; Lee, D.; and Joo, J. (2002): Enhancement of electrical conductivity of poly(3,4-
ethylenedioxythiophene)/poly(styrenesulfonate) by a change of solvents, Synth. Metals
126, 311.
Kim, N.; Han, K. (2010): Future direction of direct writing, J. Appl. Phys. 108, 102801.
Kim, S.; McKean, D. (1998): Aqueous TiO2 suspension preparation and novel application of
ink-jet printing technique for ceramics patterning, J. Mater. Sci. Lett. 17, 141.
Kim, Y.; Moon, D.; Han, J. (2004): Organic TFT array on a paper substrate, IEEE Electron
Device Lett. 25(10), 702.
King, C. (1998): Electroluminescent displays, Mater. Res. Soc. Symp. Proc. 508.
King, C. (2000): Electroluminescent displays, Planar Systems Inc.
King, C.; Coovert, R. (1990): Thin film electroluminescent display with improved contrast, US
Patent 4963788.
Kitai, A. (ed). (2008): Luminescent materials and applications, Wiley, Hoboken, New Jersey.
Klausch, A.; Althues, H.; Schrage, C.; Simon, P.; Szatkowski, A.; Bredol, M.; Adam, D.; Kaskel,
S. (2010): Preparation of luminescent ZnS:Cu nanoparticles for the functionalization of
transparent acrylate polymers, J. Luminescence 130, 692.
Ko, S.; Pan, H.; Grigoropolous, C.; Luscombe, C.; Frechet, J.; Poulikakos, D. (2007): All-inkjet-
printed flexible electronics fabrication on a polymer substrate by low-temperature high-
resolution selective laser sintering of metal nanoparticles, Nanotech. 18, 1.
Konishi, M.; Isobe, T.; and Senna, M. (2001): Enhancement of photoluminescence of ZnS:Mn
nanocrystals by hybridizing with polymerized acrylic acid, J. Lumin. 93, 1.
Kroto, H.; Heath, J.; O'Brien, S.; Curl, R.; Smalley, R. (1985): C60:
buckminsterfullerene, Nature 318,162.
Kumar, N.; Joshi, M.; Friend, C.; Prasad, P., Burzynski, R. (1997): Organic-inorganic
heterojunction light-emitting diodes based on poly(p-phenylene vinylene)/cadmium
sulfide thin films, App. Phys. Lett. 71(10), 1388.
202
Kwong, C.; Djurisic, A.; Choy, W.; Li, D.; Xie, M.; Chan, W.; Cheah, K.; Lai, P.; Chui, P.
(2005): Efficiency and stability of different tris(8-hydroxyquinoline) aluminum (Alq3)
derivatives in OLED applications, Mater. Sci. Eng. B 116, 75.
Kymakis, E.; Klapsis, G.; Koudoumas, E.; Stratakis, E.; Kornilios, N.; Vidakis, N.; and
Franghiadakis, Y. (2007): Carbon nanotube/PEDOT:PSS electrodes for organic
photovoltaics, Eur. Phys. J. Appl. Phys. 36, 257.
Le, H. (1998): Progress and trends in ink-jet printing technology, J. Imaging Sci. Tech., 42(1),
49.
Lee, E. (2003): Microdrop generation, CRC Press, Boca Raton.
Lee, I.; Pennathur, S.; Goodnick, S.; Wager, J. (1997): band structure and high-field electron
transport of a ZnS phosphor in AC thin-film electroluminescent devices, J. Korean Phys.
Soc., 31(3), 517.
Lee, C.; Joo, J.; Han, S.; and Koh, S. (2005a): An approach to durable PVDF cantilevers with
highly conducting PEDOT/PSS (DMSO) electrodes, Sens. & Actuators A: Phys., 121,
373.
Lee, K.; Blanchet, G.; Gao, F.; Loo, Y. (2005b): Direct patterning of conductive water-soluble
polyaniline for thin-film organic electronics, Appl. Phys. Lett., 86, 0741021.
Lee, F.; Mills, R.; Talke, F. (1981): Drop-on-demand inkjet printing at high print rates and high
resolution, IBM Research Report 3154, IBM Corporation, New York.
Lee, Y.; Choi, J.; Lee, K.; Stott, N.; Kim, D. (2008a): Large-scale synthesis of copper
nanoparticles by chemically controlled reduction for applications of inkjet printing
electronics, Nanotech. 19(41), 415604.
Lee, S.; Yoon, S.; Shin, H.; Joo, W.; Yi, D.; Choi, J.; Amarnath, C.; Paik, U. (2008b):
Hierarchical organization of Au nanoparticles in a poly(vinyl carbazole) matrix for hybrid
electronic devices, Nanotech. 19, 1.
Leising, G.; Stefan, H.; Meghdadi, F.; Tasch, S. (1999): Electroluminescent display device with
continuous base electrode, US Patent 5949188.
Li, D.; Kaner, R. (2006): Shape and aggregation control of nanoparticles: not shaken, not stirred,
J. Am. Chem. Soc. 128(3), 968.
Li, D.; Sutton, D.; Burgess, A.; Graham, D.; Calvert, P. (2009): Conductive copper and nickel
lines via reactive inkjet printing, J. Mater. Chem. 19, 3719.
Lia, S. M. Chan, M. Fung, C. Lee, and S. Lee. (2003): Concentration effect of glycerol on the
conductivity of PEDOT film and the device performance, Mater. Sci. Eng. B104, 26.
Lim, W.; Douglas, E.; Kim, S.; Norton, D.; Pearton, S.; Ren, F.; Shen, H.; Chang, W. (2009):
High mobility InGaZnO4 thin-film transistors on paper, Appl. Phys. Lett. 94, 072103.
Lim, J.; Kim, J.; Yoon, Y. (2010): All-inkjet-printed metal-insulator-metal (MIM) capacitor,
Curr. Appl. Phys. 12(1), E14.
203
Lima, S.; Cremona, M.; Davolos, M.; Legnani, C.; and Quirino, W. (2007): Electroluminescence
of zinc oxide thin-films prepared via polymeric precursor and via sol–gel methods, Thin
Solid Films 516, 165.
Lin, D.; Liu, N.; Yang, K.; Xing, B.; Wu, F. (2010): Different stabilities of multiwalled carbon
nanotubes in fresh surface water samples, Enviro. Pollution 158, 1270.
Lin, Y.; Li, Y.; Yeh, H.; Wang, Y.; Wen, T.; Huang, L.; and Chen, Y. (2007): Blending of
poly(3,4-ethylenedioxythiophene):poly(styrenesulfonate) with poly(ethyleneimine) as an
active layer in depletion-mode organic thin film transistors, Appl. Phys. Lett. 91, 253501.
Liu, X.; Ni, X.; Wang, J.; Yu, X. (2008): A novel route to photoluminescent water-soluble Mn-
doped ZnS quantum dots via photopolymerization initiated by the quantum dots,
Nanotech. 19, 1.
Liu, Z.; Su, Y.; Varahramyan, K. (2005): Inkjet-printed silver conductors using silver nitrate ink
and their electrical contacts with conducting polymers, Thin Solid Films, 478, 275.
Lopez, M.; Sanchez, J.; and Estrada, M. (2008) : Characterization of PEDOT :PSS dilutions for
inkjet printing applied to OLED fabrication, Proc. 7th
Int. Carribean Conf. Dev. Circ. Sys.
Lua, Y.; Pew, C.; Zhang, F.; Fillmore, W.; Bronson, R.; Sathyapalan, A.; Savage, P.; Whittaker,
J.; Davis, R.; and Linford, M. (2005): Polyelectrolytes as new matrices for secondary ion
mass spectrometry, J. Am. Soc. Mass Spectrom. 16(10), 1575.
Magdassi, S. (ed.) (2010): The chemistry of inkjet inks, World Scientific Publishing, Singapore.
Magdassi, S.; Grouchko, M.; Berezin, O.; Kamyshny, A. (2010): Triggering the sintering of
silver nanoparticles at room temperature, ACS Nano 4(4), 1943.
Mahalik, N. (2006): Micromanufacturing and nanotechnology, Springer, Berlin.
Manglik, R.; Wasekar, V.; Zhang, J. (2001): Dynamic and equilibrium surface tension of
aqueous surfactant and polymeric solutions, Exper. Thermal & Fluid Sci. 25, 55.
Manzoor, K.; Vadera, S.; Kumar, N.; Kutty, T. (2003): Synthesis and photoluminescent
properties of ZnS nanocrystals doped with copper and halogen, Mater. Chem. Phys. 82,
718.
Matsuda, H.; Kobayashi, N.; Kobayashi, T.; Miyazawa, K.; Kuwabara, M. (2000): Room-
temperature synthesis of crystalline barium titanate thin films by high-concentration sol-
gel method, J. Non-Crystalline Solids 271, 162.
Matsumoto, S. (1984): Electronic display devices, Ohmsha Ltd., Tokyo.
Mattoussi, H.; Radzilowski, L.; Dabbousi, B.; Thomas, E.; Bawendi, M.; Rubner, M. (1998):
Electroluminescence from heterostructures of poly(phenylene vinylene) and inorganic
CdSe nanocrystals, J. App. Phys. 83(12), 7965.
204
Mauthner, G.; Landfester, K.; Kock, A.; Bruckl, H.; Kast, M.; Stepper, C.; List, E. (2008): Inkjet
printed surface cell light-emitting devices from a water-based polymer dispersion, Org.
Electron. 9, 164.
McGinness, John (1972): Mobility gaps: a mechanism for band gaps in melanins, Science 177,
896.
McGown, D.; Parfitt, G. (1967): Stability of non-aqueous dispersions, part 5: theoretical
predictions for dispersions in hydrocarbon media, Colloid & Poly. Sci. 219(1), 48.
Mei, J.; Lovell, M.; Mickle, M. (2005): Formulation and processing of novel conductive solution
inks in continuous inkjet printing of 3-d electric circuits, IEEE Trans. Electron.
Packaging Manuf. 28(3), 265.
Meixner, R.; Cibis, D.; Krueger, K.; and Goebel, H. (2008): Characterization of polymer inks for
drop-on-demand printing systems, Microsys. Techn. 14, 1137.
Menawat, A.; Joseph, H., Siriwardane, R. (1984): Control of surface energy of glass by surface
reactions: contact angle and stability, J. Colloid & Interface Sci., 101(1), 110.
Mentley, D. (2002): State of flat panel display technology and future trends, Proc. IEEE 90(4),
453.
Meziani, M.; Bunker, C.; Lu, F.; Li, H.; Wang, W.; Guliants, E.; Quinn, R.; Sun, Y. (2009):
Formation and properties of stabilized aluminum nanoparticles, ACS Appl. Mater.
Interfaces 1(3), 703.
Mihailetchi, V.; Blom, P.; Hummelen, J.; Rispens, M. (2003): Cathode dependence of the open-
circuit voltage of polymer:fullerene bulk heterojunction solar cells, J. App. Phys. 94(10),
6489.
Miki, T.; Ogawa, T.; Isobe, T.; and Sfihi, H. (2004): Low-temperature synthesis of nanosize
Zn2SiO4:Mn2+
phosphor from the precursor sol in supercritical ethanol and its
photoluminescent properties, J. Sol-Gel Sci. & Tech. 31, 73.
Mire, C.; Agrawal, A.; Wallace, G.; Calvert, P.; Panhuis, M. (2011): Inkjet and extrusion
printing of conduction poly(3,4-ethylenedioxythiophene) tracks on and embedded in
biopolymer materials, J. Mater. Chem. 21, 2671.
Miura, N. (2008): Proc. 14th
Int. Workshop Inorg. Org. Electroluminescence, 427.
Molesa, S; Redinger, D.; Huang, D.; Subramanian, V. (2003): High-quality inkjet-printed
multilevel interconnects and inductive components on plastic for ultra-low-cost RFID
applications, Proc. Mater. Res. Soc. 769, Symp. H8.3.
Molesa, S. (2006): Ultra-low-cost printed electronics, PhD thesis, Department of Electrical
Engineering, University of California, Berkeley.
Molina-Lopez, F.; Briand, D.; De Rooij, N. (2012): All additive inkjet printed humidity sensors
on plastic substrate, Sensors & Actuators B 166, 212.
205
Montibon, E.; Jarnstrom, L.; and Lestelius, M. (2009): Characterization of poly(3,4-
ethylenedioxythiophene): poly(styrenesulfonate) (PEDOT:PSS) adsorption on cellulosic
materials, Cellulose 16, 807.
Montibon, E.; Jarnstrom, L.; Lestelius, M. (2010): Electroconductive paper prepared by coating
with blends of poly(3,4-ethylenedioxythiophene)/poly(4-styrenesulfonate) and organic
solvents, J. Appl. Poly. Sci. 117(6), 3524.
Moon, J.; Park, J.; Lee, T.; Kim, Y.; Yoo, J.; Park, C.; Kim, J.; and Jin, K. (2005): Transparent
conductive film based on carbon nanotubes and PEDOT composites, Diamond &
Related Mater. 14, 1882.
Moore, G. (1965): Cramming more components onto integrated circuits, Electronics 38.
Mu, J.; Gu, D.; and Xu, Z. (2005): Effect of annealing on the structural and optical properties of
nano-crystalline particles for polymer light-emitting diodes, Mater. Sci. Forum 475, 1795.
Mullen, K.; Scherf, U. (eds.) (2006): Organic light emitting devices: synthesis, properties, and
applications, Wiley-VCH, Berlin, Germany.
Mustonen, T.; Kordas, K.; Saukko, S.; Toth, G.; Penttila, J.; Heliston, P.; Seppa, H.; and
Jantunen, H. (2007): Inkjet printing of transparent and conductive patterns of single-
walled carbon nanotubes and PEDOT-PSS composites, Physica Status Solidi B 244(11),
4336.
Nalwa, H. (2000): Handbook of nanostructured materials and nanotechnology, Academic Press,
New York, New York.
Nalwa, H. (ed.) (2003): Handbook of organic-inorganic hybrid materials and nanocomposites 2nd
ed., American Scientific Publishers, Valencia, California.
Naumov, A.; Kuznetsov, O.; Harutyunyan, A.; Green, A.; Hersam, M.; Resasco, D.; Nikolaev,
P.; Weisman, R. (2009): Quantifying the semiconducting fraction in single-walled carbon
nanotube samples through comparative atomic force and photoluminescence
microscopies, Nano Lett. 9(9), 3203.
Neamen, D. (2002): Semiconductor physics and devices, 3rd
ed., McGraw-Hill, New York, New
York.
Neimo, L. (1999): Papermaking science and technology 4: papermaking chemistry. TAPPI &
Fapet Oy, Helsinki.
Nguyen, B.; Gautrot, J.; Nguyen, M.; Zhu, X. (2007): Nitrocellulose-stabilized silver
nanoparticles as low conversion temperature precursors useful for inkjet printed
electronics, J. Mater. Chem. 17, 1725.
Nguyen, T.; Yee, R.; Schwartz, B. (2001): Solution processing of conjugated polymers: the
effects of polymer solubility on the morphology and electronics properties of
semiconducting polymer films, J. Photochem. Photobio. A: Chem. 144, 21.
206
Nien, Y.; Hsu, H.; Chen, I.; Hwang, C.; and Chu, S. (2006): Photoluminescence and
microstructure of Mn-doped ZnS phosphor powder co-fired with ZnS nanocrystallites, J.
Non-Crystal. Sol. 352, 2539.
Nobusa, Y.; Yomogida, Y.; Matsuzaki, S.; Yanagi, K.; Katatura, H. (2011): Inkjet printing of
single-walled carbon nanotube thin-film transistors patterned by surface modification,
Appl. Phys. Lett. 99, 183106.
Noguchi, Y.; Sekitani, T.; Someya, T. (2006): Organic-transistor-based flexible pressure sensors
using ink-jet-printed electrodes and gate dielectric layers, App. Phys. Lett. 89, 1.
O’Mara, W.; Herring, R.; and Hunt, L. (1990): Handbook of semiconductor silicon technology,
William Andrew Publishing, Norwich.
Ohnesorge, W. (1936): Formation of drops by nozzles and the breakup of liquid jets, J. Appl.
Math. Mech. 16, 355.
Ohring, M. (1991): The materials science of thin films, Academic Press, Walthen,
Massachusetts.
Oittinen, P.; Hannu, S. (1998): Printing, papermaking science and technology, Finnish Paper
Engineers Association/TAPPI, Helsinki, Finland.
Okamoto, K.; Nasu, Y.; Hamakawa, Y. (1981): Low threshold-voltage thin-film
electroluminescent devices, IEEE Trans. Electron Dev. 28(6), 698.
Ono, Y. (1995): Electroluminescent displays, World Scientific Publishing Co., Singapore.
Ould-Ely, T.; Luger, M.; Kaplan-Reinig, L.; Niesz, K.; Doherty, M.; Morse, D. (2011): Large-
scale engineered synthesis of BaTiO3 nanoparticles using low-temperature bioinspired
principles, Nature Protocols 6(1), 97.
Ouyang, J.; Chu, C.; Chen, F.; Xu, Q.; and Yang, Y. (2005) : High conductivity poly(3,4-
ethylene-dioxythiophene): poly(styrenesulfonate) film and its application in polymer
optoelectronic devices, Adv. Func. Mater. 15(2), 203.
Ouyang, J.; Xu, Q.; Chu, C.-W.; Yang, Y.; Li, G.; and Shinar, J. (2004): On the mechanism of
conductivity enhancement in poly(3,4-ethylenedioxythiophene): poly(styrenesulfonate)
film through solvent treatment, Polymer 45, 8443.
Papageorgiou, D. (1995): On the breakup of viscous liquid threads, Phys. Fluids 7(7), 1529.
Park, J.; Lee, S.; Kim, J.; Kwon, A.; Park, H.; Han, S. (2007): White electroluminescent device
with ZnS:Mn, Cu, Cl phosphor, J. Luminescence 126, 566.
Park, S.;Kuo, C.; Anthony, J.; Jackson, T. (2005): High mobility solution-processed OTFTs,
Tech. Digest – Int. Elec. Devices Meet. 113.
Parker, I. (1994): Carrier tunneling and device characteristics in polymer light emitting diode, J.
Appl. Phys. 75, 1656.
Peng, W.; Qu, S.; Cong, G.; Wang, Z. (2005) Structural and optical investigation of Mn-doped
ZnS nanocrystals , Curr. Appl. Phys. 5, 31.
207
Perelaer, J.; de Laat, A.; Hendriks, C.; Schubert, U. (2008): Inkjet-printed silver tracks: low
temperature curing and thermal stability investigation, J. Mater. Chem. 18, 3209.
Phillips, Julia; Cava, R.J.; Thomas, G.A.; Carter, S.A.; Kwo, J.; Siegrist, T.; Krajewski, J.;
Marshall, J.; Peck, W. Jr.; Rapkine, D. (1995): Zinc-indium-oxide: a high-conductivity
transparent conducting oxide, Appl. Phys. Lett. 67(15), 2246.
Polasik, J.; Schmidt, H. (2005): Conductive polymer PEDOT/PSS electrodes on the
piezoelectric polymer PVDF, Proc. SPIE, 5759, 114.
Porambo, M.; Marsh. A. (2009): Synthesis and photoluminescent properties of doped ZnS
nanocrystals capped by poly(vinylpyrrolidone), Optical Mater. 31(11), 1631.
Pushparaj, V.; Shaijumon, M.; Kumar, A.; Murugesan, S.; Ci, L.; Vajtai, R.; Linhardt, R.;
Nalamasu, O.; Ajayan, P. (2007): Flexible energy storage devices based on
nanocomposite paper, Proc. Nat. Acad. Sci. 104(34), 13574.
Rack P, P. Holloway. (1998): The structure, device physics, and material properties of thin film
electroluminescent displays, Mater. Sci. Eng. R21, 171.
Rae, A.; and Hammer-Fritzinger, D. (2006): Creating metal and nonmetal nanosystems using
conductive jettable inks, Sol. St. Tech. 49(4), 53.
Reddish, W. (1962): Chemical structure and dielectric properties of high polymers, Pure App.
Chem. 5, 723.
Rembe, C.; Wiesche, S.; Hofer, E. (1999): Dynamics in a thermal ink jet: a model for
identification and simulation, Tech. Proc. 1999 Int. Conf. Modeling & Simulation
Microsys., 553.
Ridley, B.; Nivi, B.; Jacobson, J. (1999): All-inorganic field effect transistors fabricated by
printing, Science 286, 746.
Robel, L.; Kuno, M.; Kamat, P. (2007): Size-dependent electron injection from excited CdSe
quantum dots into TiO2 nanoparticles, J. Amer. Chem. Soc. 129, 4136.
Ryu, B.; Choim, Y.; Park, H.; Byun, J.; Kong, K.; Lee, J.; Chang, H. (2005): Synthesis of highly
concentrated silver nanosol and its application to inkjet printing, Colloids & Surfaces A
270, 345.
Saad, G. (1994): Dielectric behaviour of cyanoethylated cellulose, Polym. Int. 34, 411.
Sakai, Y.; Futakuchi, T.; Ijima, T.; Adachi, M. (2006): Preparation of (Ba,Sr)TiO3 thick film on
ZrO2 substrates by inkjet printing, Jap. J. Appl. Phys. 45(2A), 846.
Sanchez-Romaguera, V.; Madec, M.; Yeates, S. (2008): Inkjet printing of 3D metal-insulator-
metal crossovers, Reactive & Func. Poly. 68(6), 1052.
Sangoi, R.; Smith, C.; Seymour, M.; Venkataraman, J.; Clark, D.; Kleper, L.; Kahn, B. (2005):
Printing radio frequency identification (RFID) tag antennas using inks containing silver
dispersions, J. Dispers. Sci. Tech. 25(4), 515 (2005).
208
Sankir, N. (2008): Selective deposition of PEDOT/PSS on to flexible substrates and tailoring the
electrical resistivity by post treatment, Circuit World 34(4), 32.
Sanyal, U.; Datta, R.; Jagirdar, B. (2012): Colloidal calcium nanoparticles: digestive ripening in
the presence of a capping agent and coalescence of particles under an electron beam, RSC
Adv. 2, 259.
Sartale, S.; Sankapal, B.; Lux-Steiner, M.; and Ennaoui, A. (2005): Preparation of
nanocrystalline ZnS by a new chemical bath deposition route, Thin Solid Films 480, 168.
Satoh, T.; Nakasuta, T.; Tsuruya, K.; Tabata, Y.; Tamura, T.; Ichikawa, Y.; Tango, H. (2007):
Electrical properties of two-sided luminescence powder-distributed inorganic
electroluminescence panels J. Mater. Sci. Mater. Electron. 18, S239.
Savvatimskiy, A. (2005): Measurements of the melting point of graphite and the properties of
liquid carbon (a review for 1963–2003), Carbon 43(6), 1115.
Schlaf, R.; Murata, H.; Kafafi, Z. (2001): Work function measurements on indium tin oxide
films, J. Electron. Spectroscopy & Related Phenomena 120, 149.
Schrage, C.; Althues, H.; Klausch, A.; Adam, D.; Kaskel, S. (2010): ZnS:Cu polymer
nanocomposites for thin film electroluminescent devices, J. Nanosci. Nanotech. 10, 4335.
Sears, F.; Zemansky, M.; Young, H. (1982): University physics, 6th ed., Addison-Wesley,
Boston, Massachusetts.
Sedra, A. and Smith, K. (1997): Microelectronic circuits, Oxford University Press, Oxford.
Selwitz, C. (1988): Cellulose nitrate in conservation, J. Paul Getty Trust, Los Angeles,
California.
Seshan, K. (2002): Handbook of thin film deposition, William Andrew, Norwich, New York.
Setti, L.; Fraleoni-Morgera, A.; Mencarelli, I.; Fillipini, A.; Ballerin, B.; Di Biase, M. (2007):
An HRP-based amperometric biosensor fabricated by thermal inkjet printing, Sensors
and Actuators B 126, 252.
Sharma, P.; Varadan, V.; Varadan, V. (2000): Porous behavior and dielectric properties of
barium strontium titanate synthesized by sol-gel method in the presence of
triethanolamine, Chem. Mater. 12, 2590.
Sharma, H.; Mansingh, A. (1998): Sol-gel processed barium titanate ceramics and thin films, J.
Mater. Sci. 33, 4455.
Sharma, H.; Sarma, H. (1998): Electrical properties of sol-gel-processed barium titanate films,
Thin Solid Films, 330, 178.
Shi, L.; Wang, R.; Cao, Y.; Feng, C.; Liang, D.; Tay, J. (2007): Fabrication of poly(vinylidene
fluoride-co-hexafluropropylene) (PVDF-HFP) asymmetric microporous hollow fiber
membranes, J. Membrane Sci. 305, 215.
Shield, T.; Bogy, D.; Talke, F. (1987): Drop formation by DOD ink-jet nozzles: a comparison of
experiment and numerical simulation, IBM J. Res. Dev. 31(1), 96.
209
Shik, A. (1997): Quantum wells, World Scientific, Hackensack, New Jersey.
Shin, D.; Smith, P. (2008): Theoretical investigation of the influence of nozzle diameter variation
of the fabrication of thin film transistor liquid crystal color filter, J. Appl. Phys. 103,
114905.
Singh, M.; Ohji, T.; Asthana, R.; Mathur, S. (eds). (2011): Ceramic integration and joining
technologies, John Wiley & Sons, Hoboken, New Jersey.
Singh, V.; Morton, D. (1992): A model for electroluminescence in SrS:Ce ACTFEL display
devices, IEEE Trans. Electron Devices, 39(6), 1331.
Singh, M.; Haverinen, H.; Dhagat, P.; Jabbour, G. (2010): Inkjet printing: process and its
applications, Adv. Mater. 22, 673.
Sirringhaus, H.; and Shimoda, T. (2003): Inkjet printing of functional materials, MRS Bulletin
28(11), 802.
Skolnick, M. (1981): Time resolved DC electroluminescence studies in ZnS:Mn,Cu powder
phosphors, J. Phys. D: Appl. Phys. 14, 301.
Skotheim, T. (1997): Handbook of conducting polymers, 2nd
ed., CRC Press, Boca Raton,
Florida.
Skotheim, T.; Reynolds, J. (eds.) (2006): Conjugated polymers: theory, synthesis, properties, and
characterization, CRC Press, Boca Raton, Florida.
Small, A.; Johnston, J.; Clark, N. (2010): Inkjet printing of water “soluble” doped ZnS quantum
dots, Eur. J. Inorg. Chem. 242.
Smook, G. (1992): Handbook for pulp and paper technologists, 2nd
ed., TAPPI, Angus Wilde
Publications, Vancouver, British Columbia.
Sodhi, R.; Sun, L.; Sain, M.; and Farnood, R. (2008): Analysis of ink/coating penetration of
paper surfaces by time-of-flight secondary ion mass spectrometry (ToF-SIMS) in
conjunction with principal component analysis (PCA), J. Adhes. 84, 277.
Solomon, S.; Bahadory, M.; Jayarajasingham, V.; Rutkowsky, S.; Boritz, C. (2007): Synthesis
and study of silver nanoparticles, J. Chem. Edu. 84(2), 322.
Soltman, D.; Subramanian, V. (2008): Inkjet-printed line morphologies and temperature control
of the coffee-ring effect, Langmuir 24(5), 2224.
Song, J.; Kim, Y.; Yoon, Y. (2009): The production of transparent carbon nanotube field emitters
using inkjet printing, Physica E41, 1513.
Spasic, A.; and Hsu, J. (2006): Finely dispersed materials. CRC Press, Boca Raton.
Srivastava, B.; Jana, S.; Pradhan, N. (2011): Doping Cu in semiconductor nanocrystals: some old
and some new physical insights, J. Amer. Chem. Soc. 133(4), 1007.
Sumerel, J.; Nguyen, C.; Sudau, K.; Fabris, D. (2006): Deposition of MWCNTs using the
Dimatix Materials Printer, Fujifilm-Dimatix, Santa Clara, California.
210
Sumerel, J.; Staton, J.; Sudau, K. (2007): Inkjet printing toluene using the Dimatix Materials
Printer, Fujifilm-Dimatix, Santa Clara, California.
Sun, C. (2005): Size and confinement effects on nanostructures, School of Elec. & Electron.
Eng., Singapore.
Tadros, T. (ed). (1987): Solid-liquid dispersions, Academic Press, London.
Takahashi, H.; Tetsuhiko, I. (2005): Photoluminescence enhancement of ZnS:Mn nanocrystal
phosphors: comparison of organic and inorganic surface modifications, Jap. J. Appl.
Phys. 44(2), 922.
Talapin, D.; Rogach, A.; Kornowski, A.; Haase, M.; Weller, H. (2001): Highly luminescent
monodisperse CdSe and CdSe/ZnS nanocrystals synthesized in a hexadecylamine
−trioctylphosphine oxide−trioctylphospine mixture, Nano Lett. 1(4), 207.
Taylor, R.; Church, K.; Sluch, M. (2007): Red light emission from hybrid organic/inorganic
quantum dot AC light emitting displays, Displays 28, 92.
Te, L.; Ervin, M.; Qiu, H.; Fuchs, B.; Lee, W. (2011): Graphene supercapacitor electrodes
fabricated by inkjet printing and thermal reduction of graphene oxide, Electrochem.
Comm. 13, 355.
Tekin, E.; de Gans, B.;Schubert, U. (2004): Inkjet printing of polymers – from single dots to
thin-film libraries, J. Mater. Chem. 14, 2627.
Tekin, E.; Smith, P.; and Schubert, U. (2008): Inkjet printing as a deposition and patterning tool
for polymers and inorganic particles, Soft Matter 4, 703.
Titcomb, S. (1981): Dynamic surface tension of jet inks, Ind. Eng. Chem. Prod. Res. Dev. 20,
680.
Tobjork, D.; Aarnio, H.; Pulkkinen, P.; Bollstrom, R.; Maattanen, A.; Ihalainen, P.; Makela, T.;
Peltonen, J.; Toivakka, M.; Tenhu, H.; Osterbacka, R. (2012): IR-sintering of ink-jet
printed metal-nanoparticles on paper, Thin Solid Films 520(7), 2949.
Tobjork, D.; Osterbacka, R. (2011): Paper electronics, Adv. Mater. 23(17), 1935.
Torrisi, F.; Hasan, T.; Wu, W.; Sun, Z.; Lombardo, A.; Kulmala, T.; Hshieh, G.; Jung, S.;
Bonaccorso, F.; Paul, P.; Chu, D.; Ferrari, A. (2012): Ink-jet printed graphene electronics,
ACS Nano 6(4), 2992.
Tracton, A. (2005): Coatings technology handbook, CRC Press, Boca Raton, Florida.
Tseng, W.; Lin, S.; Wang, S. (2006): Particulate dispersion and freeform fabrication of BaTiO3
thick films via direct inkjet printing, J. Electroceram. 16, 537.
Tseng, H.; Subramanian, V. (2011): All inkjet-printed, fully self-aligned transistors for low-cost
circuit applications, Org. Electron. 12(2), 249.
211
Tsyboulski, D.; Bachilo, S.; Kolomeisky, A.; Weisman, R. (2008): Translational and rotational
dynamics of individual single-walled carbon nanotubes in aqueous suspension, ACS
Nano 2(9), 1770.
Valero, J.; Jarom, C.; Comas, E. (2007): Optimized automatic recovery of nozzle health in inkjet
systems, Int. Conf. Digital Printing Tech. Digital Fab. NIP23, 766.
van der Waarden, M. (1950): Stabilization of carbon-black dispersions in hydrocarbons, J.
Colloid Sci. 5(4), 317.
van Osch, T.; Perelaer, J.; de Laat, A.; Schubert, U. (2008): Inkjet printing of narrow conductive
tracks on untreated polymeric substrates, Adv. Mater. 20, 343.
Vazquez, G.; Alvarez, E.; Navaza, J. (1995): Surface tension of alcohol + water from 20 to 50°C,
J. Chem. Eng. Data 40(3), 611.
Vetanen, W.; Aguilera, M. (1997): Electroluminescent device construction employing polymer
derivative coating, US Patent 5650692.
Vilms, J.; Spicer, W. (1965): Quantum efficiency and radiative lifetime in p-type gallium
arsenide, J. Appl. Phys. 36 , 2815.
Vissenberg, M.; Matters, M. (1998): Theory of the field-effect mobility in amorphous organic
transistors, Phys. Rev. B 57,12964.
Vlasenko, N.; Popkov, I. (1960): Study of the electroluminescence of a sublimed ZnS-Mn
phosphor, Optics & Spectroscopy 8, 39.
Vogel, W.; Borse, P.; Deskmukh, N.; and Kulkarni, S. (2000): Structure and stability of
monodisperse 1.4-nm ZnS particles stabilized by mercaptoethanol, Langmuir 16(4),
2032.
Wang, G.; Tau, X.; and Wang. R. (2008a): Flexible organic light-emitting diodes with a
polymeric nanocomposite anode, Nanotech. 19, 145201.
Wang, J. (2002): Systematic consideration for the thermodynamic stability of pigmented inkjet
dispersions and inks, Int. Conf. Digital Printing Tech. NIP18, 176.
Wang, M.; Zhang, M.; Qian, J.; Zhao, F.; Shen, L.; Scholes, G.; Winnik, M. (2009): Enhancing
the photoluminescence of polymer-stabilizer CdSe/CdS/ZnS core/shell/shell and
CdSe/ZnS core/shell quantum dots in water through a chemical-activation approach,
Langmuir 25(19), 11732.
Wang, W.; Fernando, K.; Lin, Y.; Meziani, M.; Veca, L.; Cao, L.; Zhang, P.; Kimani, M.; Sun,
Y. (2008b): Metallic single-walled carbon nanotubes for conductive nanocomposites, J.
Am. Chem. Soc. 130, 1415.
Wang, Y.; Gebler, D.; Lin, L.; Blatchford, J.; Jessen, S.; Wang, H., Epstein, A. (1996):
Alternating-current light-emitting devices based on conjugated polymers, Appl. Phys.
Lett. 68, 894.
212
Warad, H.; Ghosh, S.; Hemtanon, B.; Thanachayanont, C.; Dutta, J. (2005): Luminescent
nanoparticles of Mn doped ZnS passivated with sodium hexametaphosphate, Sci. Tech.
Adv. Mater. 6, 296.
Wei, Q.; Zheng, H.; Huang, Y. (2001): Direct patterning ITO transparent conductive coatings,
Solar Energy Mater. & Solar Cells, 68(3-4), 383.
Wei, T.; Ruan, J.; Fan, Z. (2007): Preparation of a carbon nanotube film by ink-jet printing,
Carbon 45, 2712.
Whitmarsh, J. (2005): Flexible electronics: silicon meets paper and beyond, Microelec. Int.
22(3), 16.
Wijshoff, H. (2008): Structure and fluid dynamics in piezo inkjet printheads, Océ Technologies,
Venlo, TheNetherlands.
Wood, C. (ed.) (2000): Handbook of thin film devices, Academic Press, Missouri.
Wood, V.; Panzer, M.; Chen, J.; Bradley, M.; Halpert, J.; Bawendi, M.; Bulovic, V. (2009):
Inkjet-printed quantum dot–polymer composites for full-color AC-driven displays, Adv.
Mater. 21, 2151.
Wu, Y.; Li, Y.; Ong, B. (2007): A simple and efficient approach to a printable silver conductor
for printed electronics, J. Amer. Chem. Soc. 129, 1862.
Xia, Y. and Ouyang, J. (2009): Salt-induced charge screening and significant conductivity
enhancement of conducting poly(3,4-ethylenedioxythiophene): poly(styrenesulfonate),
Macromol. 42, 4141.
Xie, X.; Samsudeen, F.; Farnood, R.; Kortschot, M. (2008): Roughening due to ink jet rewetting:
effect of paper treatment and composition, J. Imag. Sci. Tech. 52(1), 010506-1.
Xing, Y.; Xu, X.; Zhang, P.; Tian, W.; Yu, G.; Lu, P.; Liu, Y.; Zhu, D. (2005): Carbazole-pyrene
based organic emitters for electroluminescent device. Chem. Phys. Lett.408(1–3), 169.
Xu, J.; Zhai, J.; Yao, X. (2006): Structure and dielectric properties of barium titanate films grown
by sol-gel-hydrothermal process, Appl. Phys. Lett. 89, 252902.
Xue, F. and Su, Y. (2005): Modified PEDOT-PSS conducting polymer as S/D electrodes for
device performance enhancement of P3HT TFTs, IEEE Trans. Electron. Devices 52(9),
1982.
Yakovlev, S. (2004): Lanthanide ions doping effects on structural, electrophysical and functional
properties of sol-gel fabricated PbTiO3 thin films, PhD thesis, University of Kiel.
Yang, A.; Cheng, C.; Lin, C. (2006): Investigation of droplet-ejection characteristics for a
piezoelectric inkjet printhead, Proc. IMechE C: J. Mech. Eng. Sci. 220, 435.
Yang, H.; Holloway, P.; Ratna, B. (2003): Photoluminescent and electroluminescent properties
of Mn-doped ZnS nanocrystals, J. Appl. Phys. 93(1), 586.
Yang, H.; Santra, S.; Holloway, P. (2005): Syntheses and applications of Mn-doped II-VI
semiconductor nanocrystals, J. Nanosci. Nanotech 5, 1364.
213
Yang, F.; Zhang, X.; Wu, X.; Tian, F.; Gan, F. (2007): Preparation of highly dispersed antimony-
doped tin oxide nanopowders by azeotropic drying with isoamyl acetate, Trans.
Nonferrous Met. Soc. China 17, 626.
Yang, H.; Holloway, P. (2003): Electroluminescence from hybrid conjugated polymer-
CdS:Mn/ZnS core/shell nanocrystals devices, J. Phys. Chem. B 107, 9705.
Yang, H.; Holloway, P. (2004): Efficient and photostable ZnS-passivated CdS:Mn luminescent
nanocrystals, Adv. Func. Mater., 14(2), 152.
Yang, P.; and Bredol, M. (2008): Surface passivation and photoluminescence of Mn-doped
nanocrystals Res. Lett. Mater. Sci. 1.
Yang, Y.; Huang, J.; Liu, S.; Shen, J. (1997): Preparation, characterization, and
electroluminescence of ZnS nanocrystals in a polymer matrix, J. Mater. Chem. 7(1), 131.
Yarema, M.; Pichler, S.; Kriegner, D.; Stang, J.; Yarema, O.; Kirchsclager, R.;
Tollabimazraehno, S.; Hume, M.; Haringer, D.; Kohl, M.; Chen, G.; Heiss, W. (2012):
From highly monodisperse indium and indium tin colloidal nanocrystals to sel-assembled
indium tin oxide nanoelectrodes, ACS Nano (online only).
Yarin, A.; Szczech, J.; Megaridis, C.; Zhang, J.; and Gamota, D. (2006): Lines of dense
nanoparticle colloidal suspensions evaporating on a flat surface: formation of non-
uniform dried deposits, J. Coll. & Interf. Sci. 294, 343.
Yen, W. (2004): Inorganic phosphors: compositions, preparation and optical properties, CRC
Press, Boca Raton.
Yen, W. (2006): Phosphor handbook, 2nd edition, CRC Press, Boca Raton.
Yeo, S.; Pode, R.; Ahn, J. (2009): Study on the size of fullerene (C60) aggregates in solution by
photoluminescence and HRTEM measurements, J. Korean Phys. Soc. 55(1), 322.
Yoshioka, Y.; Jabbour, G. (2006): Simple modification of sheet resistivity of conducting
polymeric anodes via combinatorial ink-jet printing techniques, Synth. Metals 156, 779.
Yoshioka, Y.; Calvert, P.; Jabbour, G. (2005): Simple modification of sheet resistivity of
conducting polymeric anodes via combinatorial ink-jet printing techniques, Macromol.
Rapid. Comm. 26, 238.
Yu, I.; Isobe, T.; Senna, M. (1996): Optical properties and characteristics of ZnS nanoparticles
with homogeneous Mn distribution, J. Phys. Chem.: Solids 57(4), 373.
Yuan, Z.; Dan-Li, X. (2008): Porous PVDF/TPU blends asymmetric hollow fiber membranes
prepared with the use of hydrophilic additive PVP (K30), Desalination 223, 438.
Zeng, J.; Lin, C.; Li, J.; Li, K. (1999): Low-temperature preparation of barium titanate thin films
by a novel sol-gel-hydrothermal method, Materials Letters 38, 112.
Zhang, W.; Lee, H. (2010): Synthesis and optical property of water-soluble ZnS:Cu quantum
dots by use of thioglycolic acid, Appl. Optics 49(14), 2566.
214
Zhou, Z.; Yang, Z.; Yuan, Q. (2008): Barium titanate ceramic inks for continuous ink-jet printing
synthesized by mechanical mixing and sol-gel methods, Trans. Nonferrous Met. Soc.
China 18, 150.
Zhouping, Y.; Yongan, H.; Ningbin, B.; Xiaomei, W.; Youlun, X. (2010): Inkjet printing for
flexible electronics: materials, processes, and equipment, Chinese Sci. Bulletin 55(30),
3383.
Zhuang, J.; Zhang, X.; Wang, G.; Li, D.; Yang, W.; Li, T. (2003): Synthesis of water-soluble
ZnS : Mn2+
nanocrystals by using mercaptopropionic acid as stabilizer, J. Mater. Chem.
13, 1853.
215
APPENDIX A. ZnS nanoparticle synthesis & dispersion
Synthesis
All ZnS nanoparticles were synthesized using a competitive precipitation method (Figure A.1).
Synthesis procedures for ZnS:Mn were based on those suggested by Adachi et al. (2007, 2008)
and Althues et al. (2006) using caps other than 3-MPA, and Zhuang et al. (2003) using 3-MPA.
Appendix Figure A.1. Competitive precipitation process for producing doped ZnS nanoparticles.
Synthesis procedures for ZnS:Cu were based on those described by Manzoor et al. (2003) and
Small et al. (2011) using caps other than 3-MPA, and Schrage et al. (2010) using 3-MPA. Phase
transfer of ZnS nanoparticles to make them organic-soluble was achieved using the method
described by Klausch et al. (2010). Figure A.2 shows a schematic of the synthesis procedure of
ZnS:Mn using 3-MPA, which was typical of all the syntheses, with a few minor differences.
Figure A.3 shows a schematic of the phase-transfer procedure.
Zn(Ac-)2
Mn(Ac-)2
S2-
source
H2O
Zn2+
Mn2+
Ac-
Zn2+
Zn2+
Ac-
Ac-
H2O S
2-
S2-
S
2-
S2-
S2-
+
S2-
ZnS
ZnS
ZnS
ZnS
MnS
Ac-
S2-
Ac- Ac
-
Ac-
Ac-
S2-
S2-
S2-
S2-
ZnS:Mn
Mn2+
Zn2+
Zn2+
Zn2+
Zn2+
Zn2+
Mn2+
Zn2+
Ac-
216
Appendix Figure A.2. Competitive precipitation process for producing doped ZnS nanoparticles.
Mn(Ac)2
Na2S
H2O
Zn[SCH2CH2COOH]2
H2O S
2-
Na+
NaOH
+ ZnS
ZnS
ZnS
ZnS
MnS
Ac-
S2-
Ac- Ac
-
Ac-
Ac-
S2-
S2-
S2-
S2-
Mn2+
Zn2+
Zn2+
Zn2+
Zn2+
Zn2+
Mn2+
Zn2+
CH3COO-
3-MPA
Mn[SCH2CH2COOH]2
pH 10
buffer
Zn2+
CH3COO-
Mn2+
(H+)(SCH2CH2COOH
-)
pH = 10.3
pH = 10 Δ S
2-
Na+
pH = 10
T = 70°C
(1)
(2)
(1) (2) Δ
S
O
OH
16 h
T = 70°C
ZnS:Mn
O
OH
O
OH
O
HO O
OH
O
OH
Zn(Ac)2
217
ZnS:Mn/Cu, SHMP/PVP/citrate/chitosan caps
1) Zn2+
/dopant/cap solution:
50 mmol Zn(Ac)2 2H2O
12.5 mmol Mn(Ac)2 4H2O (50 at. % - only 1.5 at. % incorporated)
OR
0.5 mmol Cu(Ac)2 H2O (1 at. %)
150 mL H2O
w:w ratio of SHMP OR PVP OR citrate OR chitosan*:Zn(Ac)2 2H2O = 1:20
(SHMP), 1:100 (PVP), 1:1.4 (citrate), 1:20 (chitosan)
*chitosan solution displaced 10 mL of H2O with 10 mL of HAc.
2) S2-
solution:
25 mmol Na2S 9H2O (Mn-doped)
OR
25 mmol Na2S2O3 5H2O (Cu-doped)
50 mL H2O
3) Heat the Zn2+
/dopant/cap solution to 70°C with gentle stirring.
4) Using a burette, add the S2-
solution dropwise to the Zn2+
/dopant/cap solution. A white
precipitate forms immediately. Continue stirring at 70°C for 16 hours, using an air
condenser to provide reflux.
5) Remove from heat. Fill 50 mL centrifuge tubes with reaction solution (use enough tubes
to use all the solution). Centrifuge at 5100 RPM for 30 minutes. Repeat twice, rinsing
the particles with 30 mL acetone the first time, and 30 mL water the second time.
6) Dry particles in air overnight.
218
ZnS:Mn/Cu, 3-MPA cap
1) Zn2+
/dopant/cap solution:
6 mmol Zn(Ac)2 2H2O
0.09 mmol Mn(Ac)2 4H2O (1.5 at. %)
OR
0.06 mmol Cu(Ac)2 H2O (1 at. %)
30 mL H2O
24 mmol (Mn-doped) or 18 mmol (Cu-doped) 3-MPA
2) S2-
/buffer solution:
3 mmol Na2S 9H2O (Mn-doped)
OR
18 mmol thiourea (Cu-doped)
120 mL H2O
4 mL pH buffer (pH 10 for Mn-doped, pH 8 for Cu-coped)
3) Add 2 M NaOH dropwise, with stirring, to Zn2+
/dopant/cap solution until pH = 10 (Mn-
doped) or 8 (Cu-doped). Solution clarifies and becomes transparent.
4) Meanwhile, heat S2-
/buffer solution to 70°C (Mn-doped) or 90°C (Cu-doped).
5) Add Zn2+
/dopant/cap solution to S2-
/buffer solution, with stirring, and maintain
temperature (70°C for Mn-doped, 90°C for Cu-doped). Heat and stir for 8 hours (Mn-
doped) or 20 hours (Cu-doped), refluxing with an air-condenser in a round-bottomed
flask. Remove from heat. Solution is still transparent. Proceed to next step if organic-
soluble particles are desired; for aqueous particles, skip to step 9.
219
6) With stirring, add 6 g octylamine to the reaction solution. Phase separation should
immediately occur. With vigorous stirring (>1000 RPM), allow the separation to take
place for 30 minutes.
7) Add 75 mL pentane to the reaction mixture, and stir at 750 RPM for 30 minutes.
8) With gentle heating (50°C max) and gentle stirring (100 RPM max) to encourage the
organic phase to the top of the mixture, use a pipette to remove the organic phase to a
separate flask. When all of the organic phase is removed, discard the aqueous phase.
9) Fill 50 mL centrifuge tubes with 15 mL reaction solution (use enough tubes to use all the
solution). Add 30 mL acetone to each tube. Solution should become cloudy as
nanoparticles precipitate out. Centrifuge at 5100 RPM for 30 minutes. Repeat twice,
rinsing the particles with 30 mL fresh acetone each time.
10) Redisperse particles in water or toluene (2.5-5 w/w%).
Appendix Figure A.3. Competitive precipitation process for producing doped ZnS nanoparticles.
ZnS
(Mn, Cu)
O
OH
O
OH
O
HO O
OH
O
OH
stirring
+
NH2
ZnS
(Mn, Cu)
O
O-H3N
+
Lipophilic tail
Emulsion
+ Δ ZnS
(Mn, Cu)
O
O-H3N
+
Dispersion
Hydrophilic tail
220
Shelling
Shelling procedures for adding ZnO shells were based on work by Karar et al. and Jiang et al.
(2004, 2009), the work of Althues et al. (2006) was used for AA caps, and that of Konishi et al.
(2001) was used for PAA caps. Where particles were capped with both PAA/AA and ZnO, the
ZnO cap was applied first.
ZnO shelling
1) Add 200 mg of dried water-soluble ZnS:Mn or :Cu nanoparticles to 200 mL of water,
dispersing them by ultrasonication for 20 minutes.
2) Add 10 mL of 0.05 M Zn(NO3)2 6H2O solution to the suspension.
3) Add 0.1 M NaOH, dropwise, to the solution, until pH = 10. Allow the solution to mix
vigorously for 20 minutes. Isolate/dry the particles by centrifugation, as described in the
procedures above.
AA capping (polymerized in situ)
1) Add 200 mg of dried water-soluble ZnS:Mn or :Cu nanoparticles to 22.5 mL of water,
dispersing them by ultrasonication for 20 minutes.
2) Add 5.486 g of AA to the solution. Cover and stir for 24 hours at 80°C to induce
polymerization of the AA.
3) Remove from heat; isolate/dry the particles by centrifugation, as described in the
procedures above.
PAA capping (polymerized before addition)
1) Add 200 mg of dried water-soluble ZnS:Mn or :Cu nanoparticles to 22.5 mL of water,
dispersing them by ultrasonication for 20 minutes.
221
2) Add 0.315 g of dissolved PAA (35 w/w% in water) to 5 mL of water. Add to the ZnS
dispersion and ultrasonicate for 30 minutes
3) Isolate/dry the particles by centrifugation, as described in the procedures above.
Dispersion
Dispersion of synthesized nanoparticles was accomplished by adding several different reagents,
depending on the dopant and the solvent (the cap was irrelevant). In water, dispersion of
ZnS:Mn was accomplished using a method described by Yang and Bredol (2008), and dispersion
of ZnS:Cu was accomplished using a method based on that suggested by Hieronymas (2002).
ZnS:Cu was reactive with the TGA used to stabilize the ZnS:Mn, necessitating a different
dispersion method. Dispersion of either ZnS:Mn or Cu in toluene was accomplished using a
method described by Klausch et al. (2010).
ZnS:Mn (aqueous)
1) Prepare a dispersion of 2.5 w/w% ZnS:Mn powder in water by ultrasonication for 15
minutes.
2) Adjust the pH of the dispersion to 9 using 2M NaOH.
3) Add enough TGA to bring the TGA concentration to 0.023 M.
4) Ultrasonicate in an ice bath for 1 hour; a transparent dispersion should result.
ZnS:Cu (aqueous)
1) Prepare a dispersion of 17.5 w/w% ZnS:Cu powder in water by ultrasonication for 15
minutes. The dispersion will be cloudy and unstable.
2) Separately, prepare a solution of 1.17 w/w% SHMP in water.
3) While ultrasonicating the SHMP solution in an ice bath, add the ZnS:Cu dispersion to it
dropwise. The weight ratio of SHMP solution to ZnS:Cu dispersion should be 6:1,
yielding a 2.5 w/w% ZnS:Cu dispersion, which should be transparent.
222
ZnS:Mn/Cu (organic)
1) Prepare a solution of 1 w/w% 3-MPA and 1 w/w% oleylamine in toluene.
2) While ultrasonicating this solution in an ice bath, add enough ZnS:Mn/Cu to achieve a
concentration of 2.5 w/w%. The solution should remain transparent.
223
APPENDIX B. Procedure for conductivity estimation in PEDOT/SWCNT films
Samples of conductive ink were tested using a modified 2-point method. The 2-point method,
while not as commonly used as the 4-point method (Heaney, 2000), was found to give more
consistent results on porous substrates, such as the different paper types that were tested. The 2
points refer to two electrodes connected to an ohmmeter which were then contacted to a printed
film. In order to measure an average resistance across the entire width of the film, bus bars were
applied using a carbon paint (Luxprint 7102, Dupont Microelectronics), to which silver contact
points were applied with a silver pen. Figure B.1 shows the experimental setup for testing
resistance in a typical square sample of printed PEDOT/SWCNT ink.
Appendix Figure B.1. (a) Schematic of resistance testing setup, showing bus bars and silver contacts. (b) Actual
resistance sample, on paper, with bus bars and contacts applied.
Measurement across the sample yielded bulk resistance (R). This value is independent of sample
geometry. Therefore, to obtain a more meaningful figure – conductivity – resistance was used to
calculate resistivity (), the inverse of which is conductivity (). Resistance is related to
conductivity by the following expressions (Heaney 2000):
2 mm
(b)
PEDOT:PSS-SWCNT
layer
C C
R
base sheet
(a) (b)
Ag Ag
224
𝜌 𝑅𝑤ℎ
𝑙 𝜎
𝜌
where w, h, and l are the width, thickness, and length of the sample, respectively. With a square
sample, w = l and = (Rh)-1
. Therefore, sample thickness was required to estimate conductivity.
In the case of impermeable substrates, such as glass, the thickness was easily obtained using
optical profilometry (see Appendix E). In the case of permeable paper (or even the absorbent
coating on the surface of inkjet acetate), thickness was not uniform throughout the sample and
depended on ink penetration depth. Therefore, a cross-section of the sample and the average
thickness of the ink layer in the exposed section were used to estimate conductivity. The
conventional method of preparing cross-sections, by embedding samples in epoxy resin and
sectioning with a microtome, was not used because of PEDOT dissolution and migration in the
epoxy, observed under SEM. Instead, samples were fastened between two glass slides and a
cross-section was cut along the edges of the slides using a surgical scalpel. In order to maintain
as flat a section as possible, normal to the surface of the sheet, a sample was fastened to a glass
slide with double-sided tape, overhanging the edge. The slide was then placed, sample side
down, on a flat surface, and the scalpel was drawn along the edge of the slide to produce a
smooth section parallel to the edge of the slide. Scalpels were replaced after preparing 50
sections to preserve sharpness. A second slide, again with double-sided tape, was applied to the
other side of the sample, forming a “sandwich”.
The noticeable blue-black colour of the PEDOT/SWCNT ink allowed it to be readily observed
under an optical microscope (Leica DM-LA) and attached camera at 100x magnification. In
order to compensate for any roughness in the section, images were captured at several focal
depths and stacked to generate an in-focus image. The section was then converted to a greyscale
image – the dark-coloured ink appeared as a darker grey. The image was then colour-inverted,
and the grey levels at or above that of the ink colour were highlighted using thresholding in
ImageJ software, which highlighted only these brighter areas. The highlighted area therefore
corresponded to the inked region, which was then measured as a percent-area of the entire image.
225
Multiplying the percent-area by the image dimension parallel to the layer thickness (“length”)
gave an average layer thickness (“width”) for the sample:
𝑎𝑟𝑒𝑎 𝑡ℎ𝑟𝑒𝑠ℎ𝑜𝑙𝑑𝑒𝑑 𝑎𝑟𝑒𝑎 𝑜 𝑖𝑚𝑎𝑔𝑒 𝑎𝑟𝑒𝑎 𝑜 𝑡ℎ𝑟𝑒𝑠ℎ𝑜𝑙𝑑𝑒𝑑 𝑟𝑒𝑔𝑖𝑜𝑛
𝑎𝑟𝑒𝑎 𝑜 𝑡ℎ𝑟𝑒𝑠ℎ𝑜𝑙𝑑𝑒𝑑 𝑟𝑒𝑔𝑖𝑜𝑛
𝑖𝑚𝑎𝑔𝑒 𝑙𝑒𝑛𝑔𝑡ℎ
𝑎𝑟𝑒𝑎 𝑜 𝑡ℎ𝑟𝑒𝑠ℎ𝑜𝑙𝑑𝑒𝑑 𝑟𝑒𝑔𝑖𝑜𝑛
𝑡ℎ𝑟𝑒𝑠ℎ𝑜𝑙𝑑𝑒𝑑 𝑟𝑒𝑔𝑖𝑜𝑛 𝑙𝑒𝑛𝑔𝑡ℎ 𝑡ℎ𝑟𝑒𝑠ℎ𝑜𝑙𝑑𝑒𝑑 𝑟𝑒𝑔𝑖𝑜𝑛 𝑤𝑖𝑑𝑡ℎ
A sequential representation of this procedure is shown in Figure B2. This procedure was
repeated on several sections, yielding an accurate estimate of average ink layer thickness.
Appendix Figure B.2. Estimation of cross-sectional ink layer thickness using thresholding, showing stages 1-5 of
obtaining a thresholded inked region.
On non-permeable substrates, such as glass, ITO, or cleaned acetate (i.e. with the absorbent layer
removed), resistance was measured using the same method, but thickness was measured using
optical profilometry (Veeco WYKO NT-1100), as described in Appendix C, following.
100 µm
glass slide tape
inked region
non-inked region
glass slide
tape
1) colour 2) greyscale
3) inverted 4) cropped grey-level
threshold
5) threshold-
ing applied
226
APPENDIX C. Film thickness estimation
For estimation of conductivity and dielectric constant, as well as for determining the number of
print passes required to achieve a desired layer thickness, film thickness had to be measured.
Optical profilometry was used for this purpose – specifically, using the Veeco WYKO NT1100
profiler. A film was deposited on the desired (impermeable) substrate and an optical profile was
obtained at the edge of the film, where it met the substrate surface. After obtaining a profile, an
x-direction linescan profile was extracted at several points (5-10) along the sample using the
WYKO software (Figure C.1).
Appendix Figure C.1. Extraction of a linescan profile from optical profile (1 layer PEDOT:PSS ink on glass).
During measurement, the stage upon which the slides rested was moved and tilted to bring the
desired location into focus and to align the optical fringes upon that location. As a result, the
linescan profiles were not always level and the substrate surface was not located at thickness = 0
(see Figure C.1). Measurement of thickness took place not only at several linescan locations, but
at several locations along each linescan itself, to obtain a representative value. Therefore,
profiles were zeroed and leveled before measurements took place.
The zeroing procedure simply involved finding the value of the measured thickness at the very
edge of the substrate (not the film). The difference between this value and 0 was then added to
all data points in the linescan (Figures C.2a, b). For leveling the linescan, the slope of the
substrate was estimated by plotting only the substrate region and fitting it to a linear regression
-0.1
0
0.1
0.2
0 200 400 600
thic
kn
es
s
(µ
m)
x-dimension (µm)
200 µm
PEDOT:PSS film
glass
227
(Figure C.2c). The data points were then indexed by number, starting with point 1 as the edge of
the substrate. Levelled data was obtained by the following formula:
𝑙𝑒𝑣𝑒𝑙𝑙𝑒𝑑 𝑣𝑎𝑙𝑢𝑒 𝑜𝑟𝑖𝑔𝑖𝑛𝑎𝑙 𝑣𝑎𝑙𝑢𝑒 𝑖𝑛𝑑𝑒 𝑛𝑢𝑚𝑏𝑒𝑟 𝑠𝑢𝑏𝑠𝑡𝑟𝑎𝑡𝑒 𝑠𝑙𝑜𝑝𝑒
Appendix Figure C.2. Zeroing and levelling of x-direction linescan for thickness estimation (1 layer PEDOT:PSS ink
on glass). (a) Location of zero-point; (b) zeroing of linescan; (c) obtaining substrate slope; (d) levelling of entire
linescan; (e) levelled, zeroed linescan.
-0.1
0
0.1
0.2
0 200 400 600
Th
ick
ne
ss
(µ
m)
x-dimension (µm)
0 1.2466734 2.4933468 3.7400202 4.9866936 6.233367 7.4800404 8.7267138 9.9733872 11.220061 12.466734 13.713407 14.960081 16.206754 17.453428 18.700101 19.946774 21.193448 22.440121 23.686795 24.933468 26.180141 27.426815 28.673488 29.920162 31.166835 32.413508 33.660182 34.906855 36.153529 37.400202 38.646875 39.893549 41.140222 42.386896 43.633569 44.880242 46.126916 47.373589 48.620263 49.866936 51.113609 52.360283 53.606956 54.85363 56.100303 57.346976 58.59365 59.840323 61.086997 62.33367 63.580343 64.827017 66.07369 67.320364 68.567037 69.81371 71.060384 72.307057 73.553731 74.800404 76.047077 77.293751 78.540424 79.787098 81.033771 82.280444 83.527118 84.773791 86.020465 87.267138 88.513811 89.760485 91.007158 92.253832 93.500505 94.747178 95.993852 97.240525 98.487199 99.733872 100.98055 102.22722 103.47389 104.72057 105.96724 107.21391 108.46059 109.70726 110.95393 112.20061 113.44728 114.69395 115.94063 117.1873 118.43397 119.68065 120.92732 122.17399 123.42067 124.66734 125.91401 127.16069 128.40736 129.65403 130.90071 132.14738 133.39405 134.64073 135.8874 137.13407 138.38075 139.62742 140.87409 142.12077 143.36744 144.61411 145.86079 147.10746 148.35413 149.60081 150.84748 152.09415 153.34083 154.5875 155.83418 157.08085 158.32752 159.5742 160.82087 162.06754 163.31422 164.56089 165.80756 167.05424 168.30091 169.54758 170.79426 172.04093 173.2876 174.53428 175.78095 177.02762 178.2743 179.52097 180.76764 182.01432 183.26099 184.50766 185.75434 187.00101 188.24768 189.49436 190.74103 191.9877 193.23438 194.48105 195.72772 196.9744 198.22107 199.46774 200.71442 201.96109 203.20776 204.45444 205.70111 206.94778 208.19446 209.44113 210.6878 211.93448 213.18115 214.42782 215.6745 216.92117 218.16785 219.41452 220.66119 221.90787 223.15454 224.40121 225.64789 226.89456 228.14123 229.38791 230.63458 231.88125 233.12793 234.3746 235.62127 236.86795 238.11462 239.36129 240.60797 241.85464 243.10131 244.34799 245.59466 246.84133 248.08801 249.33468 250.58135 251.82803 253.0747 254.32137 255.56805 256.81472 258.06139 259.30807 260.55474 261.80141 263.04809 264.29476 265.54143 266.78811 268.03478 269.28145 270.52813 271.7748 273.02147 274.26815 275.51482 276.76149 278.00817 279.25484 280.50152 281.74819 282.99486 284.24154 285.48821 286.73488 287.98156 289.22823 290.4749 291.72158 292.96825 294.21492 295.4616 296.70827 297.95494 299.20162 300.44829 301.69496 302.94164 304.18831 305.43498 306.68166 307.92833 309.175 310.42168 311.66835 312.91502 314.1617 315.40837 316.65504 317.90172 319.14839 320.39506 321.64174 322.88841 324.13508 325.38176 326.62843 327.8751 329.12178 330.36845 331.61512 332.8618 334.10847 335.35514 336.60182 337.84849 339.09516 340.34184 341.58851 342.83519 344.08186 345.32853 346.57521 347.82188 349.06855 350.31523 351.5619 352.80857 354.05525 355.30192 356.54859 357.79527 359.04194 360.28861 361.53529 362.78196 364.02863 365.27531 366.52198 367.76865 369.01533 370.262 371.50867 372.75535 374.00202 375.24869 376.49537 377.74204 378.98871 380.23539 381.48206 382.72873 383.97541 385.22208 386.46875 387.71543 388.9621 390.20877 391.45545 392.70212 393.94879 395.19547 396.44214 397.68881 398.93549 400.18216 401.42883 402.67551 403.92218 405.16886 406.41553 407.6622 408.90888 410.15555 411.40222 412.6489 413.89557 415.14224 416.38892 417.63559 418.88226 420.12894 421.37561 422.62228 423.86896 425.11563 426.3623 427.60898 428.85565 430.10232 431.349 432.59567 433.84234 435.08902 436.33569 437.58236 438.82904 440.07571 441.32238 442.56906 443.81573 445.0624 446.30908 447.55575 448.80242 450.0491 451.29577 452.54244 453.78912 455.03579 456.28246 457.52914 458.77581 460.02248 461.26916 462.51583 463.7625 465.00918 466.25585 467.50253 468.7492 469.99587 471.24255 472.48922 473.73589 474.98257 476.22924 477.47591 478.72259 479.96926 481.21593 482.46261 483.70928 484.95595 486.20263 487.4493 488.69597 489.94265 491.18932 492.43599 493.68267 494.92934 496.17601 497.42269 498.66936 499.91603 501.16271 502.40938 503.65605 504.90273 506.1494 507.39607 508.64275 509.88942 511.13609 512.38277 513.62944 514.87611 516.12279 517.36946 518.61613 519.86281 521.10948 522.35615 523.60283 524.8495 526.09617 527.34285 528.58952 529.8362 531.08287 532.32954 533.57622 534.82289 536.06956 537.31624 538.56291 539.80958 541.05626 542.30293 543.5496 544.79628 546.04295 547.28962 548.5363 549.78297 551.02964 552.27632 553.52299 554.76966 556.01634 557.26301 558.50968 559.75636 561.00303 562.2497 563.49638 564.74305 565.98972 567.2364 568.48307 569.72974 570.97642 572.22309 573.46976 574.71644 575.96311 577.20978 578.45646 579.70313 580.9498 582.19648 583.44315 584.68982 585.9365 587.18317 588.42984 589.67652 590.92319 592.16987 593.41654 594.66321 595.90989 597.15656 598.40323 599.64991 600.89658 602.14325 603.38993 604.6366 605.88327 607.12995 608.37662 609.62329 610.86997 612.11664 613.36331 614.60999 615.85666 617.10333 618.35001 619.59668 620.84335 622.09003 623.3367
0 200 400 600
Th
ick
ne
ss
(µ
m)
x-dimension (µm)
y = x
0 1.2466734 2.4933468 3.7400202 4.9866936 6.233367 7.4800404 8.7267138 9.9733872 11.220061 12.466734 13.713407 14.960081 16.206754 17.453428 18.700101 19.946774 21.193448 22.440121 23.686795 24.933468 26.180141 27.426815 28.673488 29.920162 31.166835 32.413508 33.660182 34.906855 36.153529 37.400202 38.646875 39.893549 41.140222 42.386896 43.633569 44.880242 46.126916 47.373589 48.620263 49.866936 51.113609 52.360283 53.606956 54.85363 56.100303 57.346976 58.59365 59.840323 61.086997 62.33367 63.580343 64.827017 66.07369 67.320364 68.567037 69.81371 71.060384 72.307057 73.553731 74.800404 76.047077 77.293751 78.540424 79.787098 81.033771 82.280444 83.527118 84.773791 86.020465 87.267138 88.513811 89.760485 91.007158 92.253832 93.500505 94.747178 95.993852 97.240525 98.487199 99.733872 100.98055 102.22722 103.47389 104.72057 105.96724 107.21391 108.46059 109.70726 110.95393 112.20061 113.44728 114.69395 115.94063 117.1873 118.43397 119.68065 120.92732 122.17399 123.42067 124.66734 125.91401 127.16069 128.40736 129.65403 130.90071 132.14738 133.39405 134.64073 135.8874 137.13407 138.38075 139.62742 140.87409 142.12077 143.36744 144.61411 145.86079 147.10746 148.35413 149.60081 150.84748 152.09415 153.34083 154.5875 155.83418 157.08085 158.32752 159.5742 160.82087 162.06754 163.31422 164.56089 165.80756 167.05424 168.30091 169.54758 170.79426 172.04093 173.2876 174.53428 175.78095 177.02762 178.2743 179.52097 180.76764 182.01432 183.26099 184.50766 185.75434 187.00101 188.24768 189.49436 190.74103 191.9877 193.23438 194.48105 195.72772 196.9744 198.22107 199.46774 200.71442 201.96109 203.20776 204.45444 205.70111 206.94778 208.19446 209.44113 210.6878 211.93448 213.18115 214.42782 215.6745 216.92117 218.16785 219.41452 220.66119 221.90787 223.15454 224.40121 225.64789 226.89456 228.14123 229.38791 230.63458 231.88125 233.12793 234.3746 235.62127 236.86795 238.11462 239.36129 240.60797 241.85464 243.10131 244.34799 245.59466 246.84133 248.08801 249.33468 250.58135 251.82803 253.0747 254.32137 255.56805 256.81472 258.06139 259.30807 260.55474 261.80141 263.04809 264.29476 265.54143 266.78811 268.03478 269.28145 270.52813 271.7748 273.02147 274.26815 275.51482 276.76149 278.00817 279.25484 280.50152 281.74819 282.99486 284.24154 285.48821 286.73488 287.98156 289.22823 290.4749 291.72158 292.96825 294.21492 295.4616 296.70827 297.95494 299.20162 300.44829 301.69496 302.94164 304.18831 305.43498 306.68166 307.92833 309.175 310.42168 311.66835 312.91502 314.1617 315.40837 316.65504 317.90172 319.14839 320.39506 321.64174 322.88841 324.13508 325.38176 326.62843 327.8751 329.12178 330.36845 331.61512 332.8618 334.10847 335.35514 336.60182 337.84849 339.09516 340.34184 341.58851 342.83519 344.08186 345.32853 346.57521 347.82188 349.06855 350.31523 351.5619 352.80857 354.05525 355.30192 356.54859 357.79527 359.04194 360.28861 361.53529 362.78196 364.02863 365.27531 366.52198 367.76865 369.01533 370.262 371.50867 372.75535 374.00202 375.24869 376.49537 377.74204 378.98871 380.23539 381.48206 382.72873 383.97541 385.22208 386.46875 387.71543 388.9621 390.20877 391.45545 392.70212 393.94879 395.19547 396.44214 397.68881 398.93549 400.18216 401.42883 402.67551 403.92218 405.16886 406.41553 407.6622 408.90888 410.15555 411.40222 412.6489 413.89557 415.14224 416.38892 417.63559 418.88226 420.12894 421.37561 422.62228 423.86896 425.11563 426.3623 427.60898 428.85565 430.10232 431.349 432.59567 433.84234 435.08902 436.33569 437.58236 438.82904 440.07571 441.32238 442.56906 443.81573 445.0624 446.30908 447.55575 448.80242 450.0491 451.29577 452.54244 453.78912 455.03579 456.28246 457.52914 458.77581 460.02248 461.26916 462.51583 463.7625 465.00918 466.25585 467.50253 468.7492 469.99587 471.24255 472.48922 473.73589 474.98257 476.22924 477.47591 478.72259 479.96926 481.21593 482.46261 483.70928 484.95595 486.20263 487.4493 488.69597 489.94265 491.18932 492.43599 493.68267 494.92934 496.17601 497.42269 498.66936 499.91603 501.16271 502.40938 503.65605 504.90273 506.1494 507.39607 508.64275 509.88942 511.13609 512.38277 513.62944 514.87611 516.12279 517.36946 518.61613 519.86281 521.10948 522.35615 523.60283 524.8495 526.09617 527.34285 528.58952 529.8362 531.08287 532.32954 533.57622 534.82289 536.06956 537.31624 538.56291 539.80958 541.05626 542.30293 543.5496 544.79628 546.04295 547.28962 548.5363 549.78297 551.02964 552.27632 553.52299 554.76966 556.01634 557.26301 558.50968 559.75636 561.00303 562.2497 563.49638 564.74305 565.98972 567.2364 568.48307 569.72974 570.97642 572.22309 573.46976 574.71644 575.96311 577.20978 578.45646 579.70313 580.9498 582.19648 583.44315 584.68982 585.9365 587.18317 588.42984 589.67652 590.92319 592.16987 593.41654 594.66321 595.90989 597.15656 598.40323 599.64991 600.89658 602.14325 603.38993 604.6366 605.88327 607.12995 608.37662 609.62329 610.86997 612.11664 613.36331 614.60999 615.85666 617.10333 618.35001 619.59668 620.84335 622.09003 623.3367
450 500 550 600
Th
ick
ne
ss
(µ
m)
x-dimension (µm)
0 1.2466734 2.4933468 3.7400202 4.9866936 6.233367 7.4800404 8.7267138 9.9733872 11.220061 12.466734 13.713407 14.960081 16.206754 17.453428 18.700101 19.946774 21.193448 22.440121 23.686795 24.933468 26.180141 27.426815 28.673488 29.920162 31.166835 32.413508 33.660182 34.906855 36.153529 37.400202 38.646875 39.893549 41.140222 42.386896 43.633569 44.880242 46.126916 47.373589 48.620263 49.866936 51.113609 52.360283 53.606956 54.85363 56.100303 57.346976 58.59365 59.840323 61.086997 62.33367 63.580343 64.827017 66.07369 67.320364 68.567037 69.81371 71.060384 72.307057 73.553731 74.800404 76.047077 77.293751 78.540424 79.787098 81.033771 82.280444 83.527118 84.773791 86.020465 87.267138 88.513811 89.760485 91.007158 92.253832 93.500505 94.747178 95.993852 97.240525 98.487199 99.733872 100.98055 102.22722 103.47389 104.72057 105.96724 107.21391 108.46059 109.70726 110.95393 112.20061 113.44728 114.69395 115.94063 117.1873 118.43397 119.68065 120.92732 122.17399 123.42067 124.66734 125.91401 127.16069 128.40736 129.65403 130.90071 132.14738 133.39405 134.64073 135.8874 137.13407 138.38075 139.62742 140.87409 142.12077 143.36744 144.61411 145.86079 147.10746 148.35413 149.60081 150.84748 152.09415 153.34083 154.5875 155.83418 157.08085 158.32752 159.5742 160.82087 162.06754 163.31422 164.56089 165.80756 167.05424 168.30091 169.54758 170.79426 172.04093 173.2876 174.53428 175.78095 177.02762 178.2743 179.52097 180.76764 182.01432 183.26099 184.50766 185.75434 187.00101 188.24768 189.49436 190.74103 191.9877 193.23438 194.48105 195.72772 196.9744 198.22107 199.46774 200.71442 201.96109 203.20776 204.45444 205.70111 206.94778 208.19446 209.44113 210.6878 211.93448 213.18115 214.42782 215.6745 216.92117 218.16785 219.41452 220.66119 221.90787 223.15454 224.40121 225.64789 226.89456 228.14123 229.38791 230.63458 231.88125 233.12793 234.3746 235.62127 236.86795 238.11462 239.36129 240.60797 241.85464 243.10131 244.34799 245.59466 246.84133 248.08801 249.33468 250.58135 251.82803 253.0747 254.32137 255.56805 256.81472 258.06139 259.30807 260.55474 261.80141 263.04809 264.29476 265.54143 266.78811 268.03478 269.28145 270.52813 271.7748 273.02147 274.26815 275.51482 276.76149 278.00817 279.25484 280.50152 281.74819 282.99486 284.24154 285.48821 286.73488 287.98156 289.22823 290.4749 291.72158 292.96825 294.21492 295.4616 296.70827 297.95494 299.20162 300.44829 301.69496 302.94164 304.18831 305.43498 306.68166 307.92833 309.175 310.42168 311.66835 312.91502 314.1617 315.40837 316.65504 317.90172 319.14839 320.39506 321.64174 322.88841 324.13508 325.38176 326.62843 327.8751 329.12178 330.36845 331.61512 332.8618 334.10847 335.35514 336.60182 337.84849 339.09516 340.34184 341.58851 342.83519 344.08186 345.32853 346.57521 347.82188 349.06855 350.31523 351.5619 352.80857 354.05525 355.30192 356.54859 357.79527 359.04194 360.28861 361.53529 362.78196 364.02863 365.27531 366.52198 367.76865 369.01533 370.262 371.50867 372.75535 374.00202 375.24869 376.49537 377.74204 378.98871 380.23539 381.48206 382.72873 383.97541 385.22208 386.46875 387.71543 388.9621 390.20877 391.45545 392.70212 393.94879 395.19547 396.44214 397.68881 398.93549 400.18216 401.42883 402.67551 403.92218 405.16886 406.41553 407.6622 408.90888 410.15555 411.40222 412.6489 413.89557 415.14224 416.38892 417.63559 418.88226 420.12894 421.37561 422.62228 423.86896 425.11563 426.3623 427.60898 428.85565 430.10232 431.349 432.59567 433.84234 435.08902 436.33569 437.58236 438.82904 440.07571 441.32238 442.56906 443.81573 445.0624 446.30908 447.55575 448.80242 450.0491 451.29577 452.54244 453.78912 455.03579 456.28246 457.52914 458.77581 460.02248 461.26916 462.51583 463.7625 465.00918 466.25585 467.50253 468.7492 469.99587 471.24255 472.48922 473.73589 474.98257 476.22924 477.47591 478.72259 479.96926 481.21593 482.46261 483.70928 484.95595 486.20263 487.4493 488.69597 489.94265 491.18932 492.43599 493.68267 494.92934 496.17601 497.42269 498.66936 499.91603 501.16271 502.40938 503.65605 504.90273 506.1494 507.39607 508.64275 509.88942 511.13609 512.38277 513.62944 514.87611 516.12279 517.36946 518.61613 519.86281 521.10948 522.35615 523.60283 524.8495 526.09617 527.34285 528.58952 529.8362 531.08287 532.32954 533.57622 534.82289 536.06956 537.31624 538.56291 539.80958 541.05626 542.30293 543.5496 544.79628 546.04295 547.28962 548.5363 549.78297 551.02964 552.27632 553.52299 554.76966 556.01634 557.26301 558.50968 559.75636 561.00303 562.2497 563.49638 564.74305 565.98972 567.2364 568.48307 569.72974 570.97642 572.22309 573.46976 574.71644 575.96311 577.20978 578.45646 579.70313 580.9498 582.19648 583.44315 584.68982 585.9365 587.18317 588.42984 589.67652 590.92319 592.16987 593.41654 594.66321 595.90989 597.15656 598.40323 599.64991 600.89658 602.14325 603.38993 604.6366 605.88327 607.12995 608.37662 609.62329 610.86997 612.11664 613.36331 614.60999 615.85666 617.10333 618.35001 619.59668 620.84335 622.09003 623.3367
0 200 400 600
Th
ick
ne
ss
(µ
m)
x-dimension (µm)
0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9
1 1.1 1.2 1.3 1.4 1.5 1.6 1.7 1.8 1.9
2
0 200 400 600
Th
ick
ne
ss
(µ
m)
x-dimension (µm)
0 n
index #
(a) (b)
(c) (d)
(e)
zeroing
value
thickness
228
APPENDIX D. Drop and line spacing optimization
Drop spacing was an important variable in the formation of smooth, thin films for depositing the
various functional materials. Drop spacing refers to the distance between the centre of adjacent
jetted drops, and was controlled by changing the angle of the nozzle array on the DMP Materials
Cartridges. The DMP2831 printer was capable of depositing drops at a spacing as low as 5 µm,
and as high as 250 µm, in 5 µm intervals. To print as smooth a film as possible, the space
between individual jetted drops was controlled. Individual drops spread to an average radius and
coalesced into films when they were close enough together; however, overly close drops
overlapped, resulting in rough topography (Figure D.1). Also, the different surface properties of
different substrates had an effect on the wetting behaviour of the various inks, so each ink needed
to be tested on each surface it might contact during LED construction. Therefore, a 3-stage
approach was used to determine optimum drop spacing.
Appendix Figure D.1. Drop spacing of ZnS:Mn-Aq ink on glass.
Preparation of jetting surfaces/substrates
Bare substrates, such as glass, ITO glass, or ITO PET, were cleaned as usual (see Appendix F).
However, because the LED structure required stacking of different inks, surfaces covered in each
of the inks had to be prepared. This was achieved by spin-coating of the inks onto glass slides
(which had themselves been previously cleaned). All spin-coated slides were dried at 150°C for
30 min in air. The spin speeds used are listed below.
90 µm 80 µm 60 µm 40 µm
drop spacing
(90 µm)
229
PEDOT:PSS/SWCNT ink: 3000 RPM for 20 s; 500 RPM for 10 s.
ZnS:Cu/PVK and ZnS:Mn/PVP inks: 2500 RPM for 15 s; 200 RPM for 45 s.
Determination of drop size
Each ink was jetted onto each substrate/surface using a pattern designed to deliver an array of
drops at 250 µm spacing, so that individual drops would remain distinct. Without drying the
drops, the pattern was examined under the optical microscope, photographed, and the drop
diameter measured for several drops in ImageJ, and averaged (Figure D.2). The purpose of this
stage was to estimate the likely drop spacing that would result in the coalescence of adjacent
drops and the formation of printed lines or films. The drop diameter represented the largest drop
spacing that could realistically be expected to form a film.
Appendix Figure D.2. 250 µm-spaced drops of ZnS:Mn-Aq ink on ITO glass (circled with dotted lines).
Determination of optimum drop spacing
Ideal drop spacing was established by jetting several lines of ink at different drop spacings in
increments of 5 µm, ranging from the measured single drop diameter down to 50% of the single
drop diameter (i.e. for a drop size of 60 µm, drop spacing ranging from 30-60 µm was tested.
The resulting films were characterized using optical microscopy for uniformity (Figure D.3).
Ideal drop spacing was considered to be when lines were fully merged (no holes) but not
overlapping. Because the drop spacing was only tunable in multiples of 5 µm, precise spacing to
avoid holes and overlapping was usually not possible, so a minimal amount of overlap without
any holes was targeted. In Figure D.3, this would be at 55 µm, for example. Once ideal drop
100 µm
230
spacing was established, the ideally-spaced printed films were optically profiled to confirm the
absence of holes and large overlap.
25 µm 35 µm 45 µm 55 µm 65 µm
100 µm
231
APPENDIX E. Paper substrate preparation
Northern bleached SW and HW kraft pulps were used as supplied. 40 g of oven-dry (OD) pulp
was soaked in 500 mL of deionized water for 12 h. The excess water was pressed out of the pulp
in order to calculate consistency, which was 23-25% for both pulp types. Pulp was then refined
in a Noram refiner for 5100 revolutions after which 24 g (dry pulp weight) of refined pulp was
added to 2 L of water, and was then mixed in a Durant disintegrator for 15000 revolutions. The
resulting slurry was made into handsheets containing 1.5 g dry pulp per sheet, using a non-
recirculating Noram sheet former according to TAPPI Standard Method T-205. TiO2 filler and
alkylketene dimer (AKD) internal sizing agent (Hercon 115, Hercules) were added to the pulp
slurry in filled sheets. Internal sizing agent was added at 0.8 w/w% on an OD pulp basis. A
PEI-based retention aid (Polymin SK, 30 w/w% active ingredient in water, BASF) was added to
sheets containing filler during sheet forming, with a ratio of retention aid:filler of 4:1. An ink
fixation agent, PDADMAC, was added at 2 w/w% on an OD pulp basis during sheet forming.
Handsheets were air dried in a conditioning room at 25°C and 75% relative humidity for 24 h.
All sheets were calendared at 80oC, 100 kPa and 3 nips using a Beloit Wheeler laboratory
calendar, couch side up.
Commercial sheets were used as received, without additional cleaning. Sheets were stored in
low-oil vacuum foil in the same conditioning room as the handsheets until use.
232
APPENDIX F. Impermeable substrate preparation
Several non-paper substrates were used in this work: slide glass, ITO glass (patterned and
unpatterned), ITO PET, and cellulose acetate. Patterned ITO substrates were purchased from
Kintech Limited. Unpatterned ITO glass and PET were purchased from Sigma-Aldrich.
Cellulose acetate was purchased from Avery. Finally, patterned slide glass and cellulose acetate
were prepared by printing them with 2 layers of the PEDOT:PSS/SWCNT ink, and curing for 2
hours at 150°C on a hotplate in air. All of the substrates listed above, including the
PEDOT:PSS/SWCNT-printed ones, were cleaned before overprinting with functional layers.
The cleaning process was carried out in an ultrasonic bath, as follows:
(i) Soap and water, 30 minutes, followed by rinsing with water;
(ii) Acetone, 15 minutes;
(iii) 50:50 vol:vol methanol and ethanol, 15 minutes.
Substrates were air-dried between each step with a jet of compressed air. Washed ITO substrates
were then also plasma-treated with oxygen plasma for 15 minutes to remove residual
contaminants.
233
APPENDIX G. Paper substrate characterization
Several variables had an effect on ink performance on a given paper type. Of these, the most
important that were noted were filler type and content, surface energy, and surface absorbency.
Sheets were characterized for thickness using a TMI micrometer at 10 different points on the
sheet. Average sheet thickness (tavg) was also used to estimate sheet density () by weighing (m)
a sample of a fixed area (A).
𝜌 𝑚
𝑡 𝐴
Filler/pigment content was determined by burning a portion of each sheet of a fixed weight in an
oven at 500oC for 1 h, at which point the fibres were completely ashed. The actual weight
percent of filler (plus a small amount of ash) retained in the sheet was estimated by dividing filler
weight by the original sample weight (before burning).
Contact angle was estimated using an aqueous solution of crystal violet dye (test ink) prepared
according to TAPPI Standard Method T431 for measuring ink absorbency into paper. The test
ink had a surface tension of 62 mN/m. 30 µL of this ink was dropped with a calibrated
micropipette onto a handsheet, and the resulting drop was immediately photographed from the
side using a Canon Rebel XT-ME DSLR camera with a MP-E 65 mm macro lens. Ink
absorbency of the sheets was observed by measuring the time for complete absorption of a 30 µL
sample of the same test ink into the surface. During this test, the samples were placed directly
under a 60 W incandescent lamp elevated 30 cm from the test specimen’s surface. Complete
absorbency was defined according to Test Method T431 as the point at which light reflection
from the droplet on the surface was no longer visible. Micrographs of paper sheets were used to
observe filler, coating, and surface pore distribution. A JEOL-7001 JSM scanning electron
microscope (SEM), employing the backscattered electron (BSE) detector on the SEM, was used
to improve contrast between the filler and fibres.
234
APPENDIX H. Detailed ink formulations
All formulations are given on a w/w% basis. All reagents, except where otherwise specified,
were supplied by Sigma-Aldrich Canada. One reagent in particular, poly(n-vinylcarbazole) or
PVK, was synthesized by fellow doctoral candidate Graham Morse using the method described
by Higashimura et al. (1980). The solutions used in several of the inks are composed as follows
(again, on a w/w% basis):
PEDOT:PSS: 1.3% PEDOT particles in water, 2:1 PSS:PEDOT ratio
SWCNTs: 0.04% SWCNTs (1.2-5 nm diameter, 2.5 µm length) in water, 0.5% SLS
ZnS (aqueous): 2.5% 3-MPA-capped ZnS NPs in pH 9 aqueous NaOH solution, 0.25%
TGA
ZnS (organic): 2.5% octylamine-capped ZnS NPs in toluene, 1.25% 3-MPA and
olelyamine
ZnS (acid): 1% AA-capped ZnS NPs in AA
PEDOT:PSS/SWCNTs (PS)
34% PEDOT:PSS
10% SWCNTs
10% DMSO
17% glycerol
0.5% SLS
0.5% Surfynol DF-110D defoamer
28% water
ZnS/PVP (ZnS-Aq)
40% ZnS (aqueous)
10% IPA
15% butoxyethanol
34.9% water
0.1% PVP 1 300 000
235
ZnS/PVK (ZnS-Org)
40% ZnS (organic)
10% cyclohexane
0.075% PVK 50 000
49.025% chlorobenzene
ZnS/AA (ZnS-AA)
80% ZnS (acid)
10% sulfolane
8% diethylene glycol diacrylate
2% Zonyl FS-300 fluorosurfactant
BaTiO3/MMA (BT)
5% BaTiO3 NPs, <25 nm diameter
33% PEG 300
33% EtOH
0.5% Surfynol CT-324 nonionic surfactant
0.5% BYK Chemie Disperbyk 111
dispersant
28% MMA
236
ZnS/PVDF/PVP Meyer-coating paste
26.7% Sylvania GlacierGlo ZnS phosphor
powder
48% DMAc
2.7% Cytec Modaflow 2100
13.4% PVDF
1.4% PM(MA-co-EA)
7.7% PVP 10 000
BaTiO3/PVDF Meyer-coating paste
31.7% BaTiO3 <3 µm diameter
47.3% DMAc
1.5% BYK Chemie Disperbyk 111
dispersant
19.5% PVDF
237
APPENDIX I. Ink iterations
There were many different iterations of each ink that was tested. This appendix is intended to
list the major iterations of each ink type that was prepared for this study and notes on why it was
or was not used in the final device construction. The appendix has been broken down into
several tables, each one describing a particular ink type.
Appendix Table I.1. Organic solvent-based ZnS inks.
Ink name Component Amount (w/w%) Notes
CB-1
ZnS:Mn@SHMP 0.25 No PL on printing. Filtration removes all ZnS:Mn. Poor solubility in chlorobenzene.
Triton X-100 7
Chlorobenzene 72
Cyclohexane 20
Polyvinylcarbazole 0.75
AA-1 ZnS:Mn@AA 0.25 No PL in solution. Weak PL on deposition Damaging to printhead. Damaging to substrate. Red-shifted emission.
Acrylic acid 10
Methyl methacrylate 20
Sulfolane 20
DEG-diacrylate 1.5
N,N-dimethylformamide 48.25
CB-2
ZnS:Mn@SHMP 0.25 No PL on printing Again, poor solubility in chlorobenzene.
Triton X-100 7
Chlorobenzene 72
Cyclohexane 19.75
Polyvinylcarbazole 1
CB-3
ZnS:Mn@SHMP 2.5 No PL on printing. Printhead clogs almost immediately.
Triton X-100 7
Chlorobenzene 70
Cyclohexane 20
Polyvinylcarbazole 0.5
CB-4
ZnS:Mn@SHMP 0.1 No PL on printing.
Chlorobenzene 77
Cyclohexane 21.9
Polyvinylcarbazole 1
238
CT-1
ZnS stock solution, in chlorobenzene
20 No PL in solution, or when printed. Stock solution very poorly dispersed.
Chlorobenzene 56.375
Cyclohexane 22.125
Polyvinylcarbazole 1.5
CT-2
ZnS stock solution, in chlorobenzene
40 Same issues as CT-1.
Chlorobenzene 20
Cyclohexane 9.5
Toluene 30
Polyvinylcarbazole 0.5
CT-3
ZnS stock solution, in chlorobenzene
40 Better dispersion with THF added. No PL. Very long drying time.
Chlorobenzene 20
Tetrahydrofuran 19.5
Toluene 20
Polyvinylcarbazole 0.5
TT-1*
ZnS stock solution, in toluene
40 Prints well. Well dispersed and transparent.
Chlorobenzene 49.025
Cyclohexane 10
Polyvinylcarbazole 0.075
239
Appendix Table I.2. Aqueous ZnS inks.
Ink name Component Amount (w/w%) Notes
AQ-1
ZnS:Mn@SHMP 0.25 Very long drying time. No PL in printed film.
Beckman-Coulter Dispersant IIIA 2
Triton X-100 1
water 56.25
glycerol 20
PEG-300 20
AQ-2
ZnS:Mn@SHMP 0.25 Longest drying time. No PL in printed film.
Beckman-Coulter Dispersant IIIA 2
Triton X-100 1
water 56.25
PEG-300 40
AQ-3
ZnS:Mn@PVP 0.25 Unstable. Very viscous. Non-jettable.
glycerol 20
polyvinyl alcohol 3
Triton X-100 1
Beckman-Coulter Dispersant IC 0.5
Beckman-Coulter Dispersant IIIA 0.5
water 74.75
AQ-4
ZnS:Mn@PVP 0.25 Very unstable. Very viscous. Non-jettable.
glycerol 20
polyvinyl alcohol 1.5
Triton X-100 1
Beckman-Coulter Dispersant IC 0.5
Beckman-Coulter Dispersant IIIA 0.5
water 76.25
AQ-5
ZnS:Mn@PVP 0.25 No PL on printing. Poor resolution. Surface tension too low.
glycerol 40
Zonyl FS-300 0.1
Beckman-Coulter Dispersant IIIA 1 water 58.65
AQ-6
ZnS:Mn@SHMP 2.5 Difficult to filter. Bright PL, but red-shifted. Clogs printhead.
glycerol 40
Zonyl FS-300 0.1
Beckman-Coulter Dispersant IIIA 1
water 56.4
240
AQ-7
ZnS:Mn@SHMP 0.1 Difficult to filter. Dim PL. No PL on printing. Eventually unstable.
glycerol 40
Zonyl FS-300 0.1
Disperbyk 111 0.005
water 59.795
AQ-8
ZnS:Mn@SHMP 0.1 Precipitated out. Unstable.
glycerol 20
Zonyl FS-300 0.1
Disperbyk 111 0.005
water 79.795
AQ-9
ZnS:Mn@SHMP 0.1 Similar to AQ-8; Longer shelf life, but eventually unstable.
glycerol 20
Zonyl FS-300 0.1
Disperbyk 111 0.005
water 69.795
dimethyl sulfoxide 10
AQ-10
ZnS:Mn@SHMP 0.1 Weak PL. Difficult to filter. No PL when printed.
glycerol 10
Zonyl FS-300 0.1
Disperbyk 111 0.005
water 69.795
dimethyl sulfoxide 20
AQ-11
ZnS:Mn@SHMP 0.1 Similar to AQ-10. Poor dispersion. Unstable over time.
glycerol 5
Zonyl FS-300 0.1
Disperbyk 111 0.005
water 69.795
dimethyl sulfoxide 25
AQ-12
ZnS:Mn@SHMP 0.1 Similar to AQ-10, AQ-11. Poor dispersion. Unstable over time. DMSO = bad stability.
Zonyl FS-300 0.1
Disperbyk 111 0.005
water 69.795
dimethyl sulfoxide 30
AQ-13
ZnS:Mn@SHMP 0.1 Filters well. Unstable over time. No PL when printed.
Zonyl does not affect stability.
Zonyl FS-300 0.1
Disperbyk 111 0.005
water 99.795
241
AQ-14
ZnS:Mn@SHMP 0.1 Disperbyk 111 is Insoluble at high loadings. Unstable.
Disperbyk 111 10
water 89.9
AQ-15
ZnS:Mn@SHMP 0.1 Jets extremely poorly. Good dispersion. Unable to print.
Bright PL.
Disperbyk 111 0.005
water 79.895
3-amino-1-propanol 20
AQ-16
ZnS:Mn@SHMP 0.1 No PL in printed film. Print quality very poor. Droplet formation poor.
Zonyl FS-300 0.1
Disperbyk 111 0.005
water 69.795
3-amino-1-propanol 30
AQ-17
ZnS:Mn@SHMP 0.1 No PL in printed film. Droplet formation poor.
Disperbyk 111 0.005
water 69.895
3-amino-1-propanol 30
WA-1
ZnS:Mn stock solution 20 Based on Small et al's formulation.
Bad mercaptan stench and extremely unstable.
PVP10K 1.5
Igepal CA-720 0.25
Surfynol DF-110D 0.25
mercaptosuccinic acid 6.162
water 60
dimethyl sulfoxide 11.838
WA-2
ZnS:Mn stock solution 20 Good printing. Bright, but not extremely bright.
Drop formation ideal.
PVP10K 1
Igepal CA-720 0.25
glycerol 8
water 60.75
dimethyl sulfoxide 10
WA-3
ZnS:Mn stock solution 40 Similar to WA-2. Bright PL in solution and printed. Higher ZnS:Mn concentration.
Jetting issues - over time, jetting is compromised.
PVP10K 1
Igepal CA-720 0.25
glycerol 8
water 40.75
dimethyl sulfoxide 10
242
WA-4
ZnS:Mn stock solution 40 Not jettable (3-AP is a problem). PVP10K 1
Igepal CA-720 0.25
glycerol 8
water 36
dimethyl sulfoxide 10
3-amino-1-propanol 4.75
WA-5
ZnS:Mn stock solution 40 High viscosity - good for jetting.
Drop formation deteriorated rapidly.
PVP10K 1
Igepal CA-720 0.75
glycerol 58.25
WA-6
ZnS:Mn stock solution 40 High viscosity (lots of PVP) Unstable. PL quenched by PVP.
PVP10K 40
Igepal CA-720 0.75
dimethyl sulfoxide 19.25
WA-7
ZnS:Mn stock solution 40 Clogging.
Jetting is poor. PVP10K 5
Igepal CA-720 0.75
glycerol 54.25
WA-8
ZnS:Mn stock solution 40 Clogging.
Jetting is still poor. PVP10K 0.5
Igepal CA-720 0.75
glycerol 58.75
WA-9
ZnS:Mn stock solution 40 Jetting is poor. Excessive foaming.
DF-110D destabilizes solution.
PVP10K 0.5
Triton X-100 1
glycerol 40
water 18.5
AQ-11-Mn
ZnS:Mn stock solution 40 Viscosity is very high with CMC. Stability compromised. Films will contain CMC and PVP.
PVP10K 0.5
Triton X-100 0.75
glycerol 5
water 53.35
carboxymethylcellulose 0.4
243
AQ-12-Mn
ZnS:Mn stock solution 40 Similar problems as AQ-12-Mn. PVP10K 0.5
Igepal CA-720 0.75
glycerol 5
water 53.55
carboxymethylcellulose 0.2
AQ-13-Mn
ZnS:Mn stock solution 40 Removal of PVP/CMC helps jetting.
Foaming is a problem.
Igepal CA-720 0.75
glycerol 10
water 48
pH 10 buffer 0.25
0.5 M NaOH 1
AQ-14-Mn
ZnS:Mn stock solution 40 Addition of isopropanol helps jetting. No foaming.
Stable with TGA added.
isopropanol 10
butoxyethanol 15
water 33.188
PVP10K 0.35
thioglycolic acid 0.212
pH 10 buffer 0.25
0.5 M NaOH 1
AQ-15-Mn
ZnS:Mn stock solution 65.6 Higher concentration of ZnS:Mn. Addition of butoxyethanol caused destabilization.
Too much stock solution.
isopropanol 10
butoxyethanol 15
water 7.848
PVP1.3M 0.09
thioglycolic acid 0.212
pH 10 buffer 0.25
0.5 M NaOH 1
AQ-16-Mn
ZnS:Mn stock solution 40 Stable and jettable. Not durable to overprinting with aqueous ink.
Not enough PVP, doesn’t form films.
isopropanol 10
butoxyethanol 15
water 33.483
PVP1.3M 0.055
thioglycolic acid 0.212
pH 10 buffer 0.25
0.5 M NaOH 1
244
AQ-17-Mn
ZnS:Mn stock solution 40 Stable and jettable.
isopropanol 10
butoxyethanol 15
water 33.188
PVP1.3M 0.35
thioglycolic acid 0.212
pH 10 buffer 0.25
0.5 M NaOH 1
AQ-15-Cu
ZnS:Cu stock solution 99.945 Stable.
Jets poorly. PVP1.3M 0.055
AQ-16-Cu
ZnS:Cu stock solution 74.945 Unstable. Ethanol desolvates SHMP.
PVP1.3M 0.055
ethanol 10
butoxyethanol 15
AQ-17-Cu
ZnS:Cu stock solution 99.65
PVP1.3M 0.35
AQ-18-Mn*
ZnS:Mn stock solution 40 Stable and jettable.
isopropanol 10
butoxyethanol 15
water 34.65
PVP1.3M 0.35
245
Appendix Table I.3. BaTiO3 ink formulations.
Ink name Component Amount (w/w%) Notes
BT-old BaTiO3 NPs x
PEG-200 30
EtOH 50
Zetasperse 1200 x/10
Methyl methacrylate balance
BT BaTiO3 NPs 5
PEG-200 33
EtOH 33
Zetasperse 1200 0.5
Methyl methacrylate 28.5
BT-1 BaTiO3 NPs 5
PEG-200 33
EtOH 33
Surfynol CT-324 0.5
Disperbyk 111 0.5
Methyl methacrylate 28
BT-2 BaTiO3 NPs 5
EtOH 60
Surfynol CT-324 0.5
Disperbyk 111 0.5
Methyl methacrylate 28.4
Polyvinyl acetate 5
Surfynol DF-110D 0.5
Sodium citrate 0.1
BT-3 BaTiO3 NPs 5
EtOH 60
Surfynol CT-324 0.5
Disperbyk 111 0.5
Methyl methacrylate 32.4
Polyvinyl acetate 1.5
Surfynol DF-110D 0.5
Sodium citrate 0.1
BT-4*
BaTiO3 NPs 5
PEG-300 33
EtOH 33
Surfynol CT-324 0.5
Disperbyk 111 0.5
Methyl methacrylate 28
246
BT-5 BaTiO3 NPs 5
PEG-300 10
EtOH 33
Surfynol CT-324 0.5
Disperbyk 111 0.5
Methyl methacrylate 51
BT-6 BaTiO3 NPs 5
PEG-300 5
EtOH 33
Disperbyk 111 0.5
Methyl methacrylate 56
carboxymethylcellulose 0.5
BT-7 BaTiO3 NPs 5
PEG-300 5
EtOH 33
Surfynol CT-324 1
Disperbyk 181 1
Methyl methacrylate 55
BT-8 BaTiO3 NPs 5
PEG-300 10
EtOH 33
Surfynol CT-324 1
Disperbyk 181 0.5
Disperbyk 111 0.5
Methyl methacrylate 45
Diethylene glycol diacrylate
5
BT-9 BaTiO3 NPs 5
Butoxyethanol 15
EtOH 33
Surfynol CT-324 1
Disperbyk 181 0.5
Disperbyk 111 0.5
Methyl methacrylate 40
Diethylene glycol diacrylate
5
247
BT-10
BaTiO3 NPs 5
Butoxyethanol 10
PEG-200 5
EtOH 33
Surfynol CT-324 1
Disperbyk 181 0.5
Disperbyk 111 0.5
Methyl methacrylate 40
Diethylene glycol diacrylate
5
248
Appendix Table I.4. PEDOT:PSS/SWCNT ink formulations.
Ink name Component Amount (w/w%) Notes
Bronczyk Model PEDOT:PSS 68
isopropanol 12
glycerol 20
Garnett Model PEDOT:PSS 45
isopropanol 10
glycerol 15
dimethyl sulfoxide 30
Standard PEDOT:PSS 52.5
PEG-200 10
isopropanol 10
glycerol 25
Disperbyk 111 2.5
Standard + DMSO PEDOT:PSS 42
PEG-200 7
isopropanol 7
glycerol 14
dimethyl sulfoxide 28
Disperbyk 111 2
Standard + DMSO, MWCNTs
PEDOT:PSS 41.7
PEG-200 7
isopropanol 7
glycerol 13.9
dimethyl sulfoxide 27.9
MWCNT solution 0.7
Disperbyk 111 1.8
Inkjet formulation PEDOT:PSS 32.7
water 24.5
isopropanol 5.4
glycerol 10.9
dimethyl sulfoxide 21.8
MWCNT solution 0.5
Disperbyk 111 1.4
sodium lauryl sulfate 2
249
PB PEDOT:PSS 34
water 35.5
glycerol 17
dimethyl sulfoxide 10
SWCNT solution 3
Surfynol DF-110D 0.5
PT PEDOT:PSS 34
water 35.4
glycerol 17
dimethyl sulfoxide 10
SWCNT solution 3
Surfynol DF-110D 0.5
Triton X-100 0.1
PI PEDOT:PSS 34
water 35.3
glycerol 17
dimethyl sulfoxide 10
SWCNT solution 3
Surfynol DF-110D 0.5
Igepal CA-720 0.2
PZ PEDOT:PSS 34
water 35.3
glycerol 17
dimethyl sulfoxide 10
SWCNT solution 3
Surfynol DF-110D 0.5
Zonyl FS-300 0.2
PZS PEDOT:PSS 34
water 35.3
glycerol 17
dimethyl sulfoxide 10
SWCNT solution 3
Surfynol DF-110D 0.5
ZetaSperse 3700 0.2
PSB PEDOT:PSS 34
water 35
glycerol 17
dimethyl sulfoxide 10
SWCNT solution 3
Surfynol DF-110D 0.5
sodium dodecyl benzene sulfonate
0.5
250
PLS PEDOT:PSS 34
water 35
glycerol 17
dimethyl sulfoxide 10
SWCNT solution 3
Surfynol DF-110D 0.5
sodium lignosulfonate 0.5
PBK PEDOT:PSS 34
water 35
glycerol 17
dimethyl sulfoxide 10
SWCNT solution 3
Surfynol DF-110D 0.5
benzalkonium chloride 0.5
PS PEDOT:PSS 34
water 35
glycerol 17
dimethyl sulfoxide 10
SWCNT solution 3
Surfynol DF-110D 0.5
sodium lauryl sulfate 0.5
PS-1
PEDOT:PSS 34
water 35
glycerol 17
dimethyl sulfoxide 10
SWCNT solution 3
Surfynol DF-110D 0.5
sodium lauryl sulfate 0.5
PS-2
PEDOT:PSS 34
glycerol 45
dimethyl sulfoxide 10
SWCNT solution 10
Surfynol DF-110D 0.5
sodium lauryl sulfate 0.5
PS-3 PEDOT:PSS 44
water 25
glycerol 10
dimethyl sulfoxide 10
SWCNT solution 10
Surfynol DF-110D 0.5
sodium lauryl sulfate 0.5
251
APPENDIX J. Jetting waveforms
0
5
10
15
20
25
30
35
0 10 20 30 40 50
Vo
lta
ge
(V
)
Time (µs)
0
5
10
15
20
25
30
0 10 20 30 40 50 60
Vo
lta
ge
(V
)
Time (µs)
PEDOT:PSS/SWCNT ink
Jetting frequency = 3 kHz
BaTiO3 ink
Jetting frequency = 8.8 kHz
252
0
5
10
15
20
25
30
35
0 5 10 15 20
Vo
lta
ge
(V
)
Time (µs)
0
5
10
15
20
25
0 20 40 60 80
Vo
lta
ge
(V
)
Time (µs)
ZnS/PVP/H2O ink
Jetting frequency = 2 kHz
ZnS/PVK/chlorobenzene ink
Jetting frequency = 2 kHz
253
APPENDIX K. ToF-SIMS fragments analyzed and construction of molecular maps
Time-of-flight secondary ion mass spectrometry (ToF-SIMS) was used to map the distribution of
PEDOT and PSS ions and ion fragments relative to the paper fibres and filler. Samples were
printed an immediately placed in high-vacuum foil to prevent them from contacting any oils or
salts during handling. Before measurement, they were removed from the foil and placed with
tweezers into the analysis chamber.
An ION-ToF ToF-SIMS IV apparatus was used to perform the measurements. A Bi3 primary ion
gun was used to induce ion ejection and fragmentation. Images of the samples’ surfaces were
obtained using the high-spatial-resolution detector, as the spatial location of the molecules in
question was to be ascertained. The imaged areas had a size of 100 μm ×100 μm, with a spatial
resolution of 390 nm. Both positive and negative ion spectral maps were obtained. A full peak
list is shown in Table J.1; identification of the cellulosic fragments was based on the work of
Fardim & Holmbom (2005), Sodhi et al. (2008) and Delandes et al. (1998). It is important to
note that the secondary ion masses listed were variable by several mass units due to protonation
or deprotonation of the fragments, a common occurrence in ToF-SIMS analysis of organic
species (Lua et al. 2005).
Appendix Figure K.1.1. ToF-SIMS ion fragments analyzed (high spatial resolution).
Negative fragments
Component Source Chemical formula(e) Secondary ion masses (amu)
PSS Ink (PSS) C8H7O3S- 183
PEDOT Ink (PEDOT)
C6H3O2S-
C6H2O2S2-
C6HO2S
3-
C6O2S4-
139 138 137 136
254
Positive fragments
Component Source Chemical formula(e) Secondary ion masses (amu)
Metals Base sheet (filler)
Ink (surfactant) Contaminants from handling
Ti4+
Na
+
Mg2+
48 23 24
Hydrocarbons Ink (surfactant, defoamer)
Contaminants from handling CnH(2n-1)
+ (C=C)
CnH(2n+1)+ (C-C)
27, 29, 43, 55, 69, 85, 83, 97, 99, 113
Cellulose Base sheet (fibres) C6H7O3
+
C6H9O4+
C6H9O5+
127 145 161
255
APPENDIX L. Printed PEDOT:PSS/SWCNTs on paper: thickness and bulk resistance
All sheets were printed with 3 layers of ink at 25 µm drop spacing, which corresponded to a
loading of 58 g/m2 of wet ink or 0.25 g/m
2 of dried ink.
Commercial sheets
Appendix Figure L.1. Printed PEDOT:PSS/SWCNT layer thickness, commercial sheets.
Appendix Figure L.2. Printed PEDOT:PSS/SWCNT bulk resistance, commercial sheets.
0
20
40
60
80
100
cellulose acetate
photo-paper cardstock inkjet paper high-yield paper
glossy paper
Pri
nte
d la
ye
r th
ick
ne
ss
(µ
m)
0
5
10
15
20
25
High-yield paper
Inkjet paper
Cardstock Glossy paper
Photo- paper
Cellulose acetate
Bu
lk r
es
ista
nc
e (
k
)
256
Handsheets
Appendix Figure L.3. Printed PEDOT:PSS/SWCNT layer thickness, handsheets.
0
10
20
30
40
50
60
70
0% 15% 30%
Pri
nte
d la
ye
r th
ick
ne
ss
(µ
m)
TiO2 content (wt%)
SW fibres
unfurnished
sizing
fixation agent
sizing + fixation agent
0
10
20
30
40
50
60
70
80
90
0% 15% 30%
Pri
nte
d la
ye
r th
ick
ne
ss
(µ
m)
TiO2 content (wt%)
HW fibres
unfurnished
sizing
fiaxtion agent
sizing + fixation agent
257
Appendix Figure L.4. Printed PEDOT:PSS/SWCNT bulk resistance, handsheets.
1
10
100
1000
10000
100000
0 15 30
Bu
lk r
es
ista
nc
e (
k
)
TiO2 content (wt%)
SW fibres
unfurnished
sizing
fixation agent
sizing + fixation agent
1
10
100
1000
10000
0 15 30
Bu
lk r
es
ista
nc
e (
k
)
TiO2 content (wt%)
HW fibres
unfurnished
sizing
fixation agent
sizing + fixation agent