Aug 14, 2013 - else could find and Andrey Voronov for assisting me in adapting the ..... 5.50 â In-situ time evolution spectra for electrochromic NiO in 1.0M LiClO4- ... wavelength for a TiO2 thin film ...... Following this there were a number of early patent applications .... and high IR absorption typical of ITO films as IR photons.
Thin Film Electrochromic Materials and Devices
Matthew Kenneth Neeves 14/08/2013
A thesis submitted in partial fulfilment of the requirements of the University of the West of Scotland for the degree of Doctor of Philosophy Director of Studies and Supervisor: Professor Frank Placido
To my wife Charlotte
i
Abstract This thesis investigates in detail the thin film materials required for the construction of a thin film electrochromic device, their production by vacuum deposition and other techniques, and their characterisation by SEM, XRD, and optical and electrochemical methods, leading to a greater understanding of the materials and considerations required in the design of electrochromic layers and devices constructed with said layers. Working devices consisting of electrochromes, electrolytes and transparent conducting electrodes are constructed by methods and upon a scale that are amenable to commercial‐scale production. The hardware and software components of a unique real‐time spectroscopic electrochemical characterisation cell are described, which have enabled the novel synchronous collection of wideband optical transmittance and electrochemical information at intervals as small as 20ms. Optimal process conditions for the production of electrochromic transition metal oxides of nickel, titanium, tungsten and the novel nickel‐chromium oxide by advanced sputtering and electron‐beam evaporation techniques are investigated and described in‐depth. For comparison, devices are also constructed using the well‐known electrochromic material iron hexacyanoferrate, or ‘Prussian Blue’. It is essential for devices intended for eyewear applications that materials are eye‐safe and that traffic light recognition is not unduly impaired. The electrochromic performance of individual materials and working devices is reported for all materials and spectroscopic data is used to calculate tristimulus co‐ordinates and thus characterise the colour performance of the various materials and devices. Working devices also require transparent conductive electrodes. The transparent conductive oxide indium tin oxide (ITO), as prepared by two different sputtering methods is investigated. The sheet resistance of the ITO is shown to have a significant quantifiable effect upon the switching speed of working devices and this is reported in detail. Plasma‐assisted sputtering from a ceramic target is shown to produce good
i
quality crystalline films with a low sheet resistance at room temperature and at an acceptably high deposition rate. Additionally, an important factor in this work has been the development and testing of suitable substrate tooling and the use of commercial‐scale production equipment. This is shown to allow the production of highly reproducible devices. The best performing device was manufactured with a relatively large working area (38.48cm2) using NiO and WO3 films, a novel agar‐based gel polymer electrolyte and 130nm thick ITO electrodes. This device achieved switching speeds of 25.7 and 9.5 seconds for colouring and bleaching respectively, and was capable of being cycled beyond 103 times without significant degradation. ii
Declaration I confirm that this thesis has not been previously submitted for a PhD or any other comparable academic award. Date:
14/08/2013
Signature:
Matthew Kenneth Neeves
iii
Page Intentionally Blank
iv
Acknowledgements “Ex ignorantia ad sapientiam et tenebris ad lucem” With the wide scope and varied nature of this work, there are many people I wish to thank for their support, assistance and encouragement. Firstly I would like to thank Professor Frank Placido, my director of studies, for his kind support, advice, guidance, and of course, the opportunity to join the Thin Film Centre and study for my PhD at University of the West of Scotland. Also at the Thin Film Centre I share my gratitude for all of the help and advice received from the technical staff; Doctor Liz Porteous for her many hours of SEM / EDX analysis and supportive chats, Andrew Bunyan for his tireless help and support with both the machines and depositions, Doctor Shigeng Song for his assistance and advice with ITO in the early stages of this project, Geoff Moores and Jim Orr for their continual assistance with the design and construction of tooling and demonstration devices, Gerry O’Hare for always being available to assist and procure the things that no‐one else could find and Andrey Voronov for assisting me in adapting the software for the Scalar Technologies ScalarGauge. From the chemistry department I would like to thank Doctor Callum McHugh for this encouragement and also the technicians, in particular; Stuart McCann and Charles McGuiness for their accommodation of me within the chemistry department, their support and also friendship. My fellow PhD students of past and present, in particular from the thin film centre; To Ross Birney, Steffen Lapp, John Kavanagh, Zane Grey and more recently, Peter Childs, I say thank you for the advice, encouragement, friendship and lively debates you have provided me over the past years. From elsewhere in UWS; Natalie Dickinson, Stephen Hendry, Kiri Rodgers, Torsten Howind, Dorn Carran and our late friend, Margaret Rose Train, I thank you for your help, support, and most importantly continued friendship
v
which has allowed me to retain some of my sanity, under the presumption that I possessed any in the first place! Appreciation also goes to both Doctor Pascal Boinard and Doctor Eric Boinard at Polaroid Eyewear for all of the help, advice and of course, in combination with the Engineering and Physical Sciences Research Council (EPSRC), the project proposal and funding without which none of this research would have been possible. My family and other friends also deserve no lesser mention for aiding me throughout the course of this PhD, especially my mother Carol, and my in‐laws, Paul and Janet, for their encouragement, understanding and support both with respect to the PhD and life alongside the PhD, especially during the write‐up stage. Finally, I wish to thank my wife Charlotte for her unending love, support and tolerance, without which my will to complete at times, would most certainly have faltered. Your faith, enthusiasm and belief in me are a never ending source of motivation and strength, and for this, with respect to the PhD and beyond, I am eternally grateful. Vobis omnibus maximus gratias ago! (I give my deepest thanks to you all)
vi
List of Symbols and Abbreviations
Absorbance (Also known as optical density – OD)
AC
Alternating Current
AgCl
Silver Chloride
AR
Anti Reflection
Magnetic Field (Teslas)
C
Coulomb
CA
Chrono Amperometry
CAD
Computer Aided Design
CE
Colouration Efficiency
CE
Counter Electrode
CIE
Commission Internationale de l’Eclairage
CRT
Cathode Ray Tube
CuSO4
Copper Sulphate
CV
Cyclic Voltammetry
CVD
Chemical Vapour Deposition
Thickness
Inter‐Planar Spacing
DC
Direct Current
Charge of Electron
Electrical Field (N/C)
E‐beam
Electron Beam
EDX
Energy Dispersive X‐ray analysis
EIS
Electrochemical Impedance Spectroscopy
EPD
Electrophoretic Deposition
FESEM
Field Emission Scanning Electron Microscope 96485.3365 /
Faraday Constant (
FTIR
Fourier Transform Infrared Spectroscopy
vii
)
HD
Hard Drive
Hz
Hertz
Current (Amperes)
Intensity
ICP
Inductively Coupled Plasma
IR
Infrared
ITO
Indium‐Tin Oxide (In2O3:Sn)
Kelvin
Extinction Coefficient
KCl
Potassium Chloride
KOH
Potassium Hydroxide
LCD
Liquid Crystal Display
LiClO4
Lithium Perchlorate
Li2O
Lithium Oxide
LiPON
Lithium Phosphorus Oxy‐Nitride
LiTMS
Lithium Trimethyl sulphonate
MFC
Mass Flow Controller
Complex Index of Refraction
N
Newtons
Refractive Index
NaCl
Sodium Chloride
NiO
Nickel Oxide
Ni‐Cr
Nickel‐Chromium
OCP
Open Circuit Potential
OD
Optical Density
PB
Prussian Blue
PC
Propylene Carbonate
PC
Personal Computer
PG
Prussian Green
PMMA
Poly(methyl methacrylate)
PVD
Physical Vapour Deposition
viii
PW
Prussian White
PY
Prussian Yellow
Q
Charge (Coulombs)
QCM
Quartz Crystal Microbalance / Monitor
Reflectance (%)
RE
Reference Electrode
RF
Radio Frequency
SCE
Saturated Calomel Electrode
SEM
Scanning Electron Microscopy
SPD
Suspended Particle Device
SWV
Square‐Wave Voltammetry
Temperature
Transmittance (%)
TCO
Transparent Conductive Oxide
TCP/IP
Transmission Control Protocol / Internet Protocol
TiO2
Titanium Dioxide
USB
Universal Serial Bus
UV
Ultra Violet
Potential (Volts)
WE
Working Electrode
WO3
Tungsten Trioxide
XRD
X‐ray Diffraction
α
Absorption coefficient
φ
Phase Shift
λ
Wavelength (nm)
Permittivity of Free Space
Gibbs Free Energy
Enthalpy Change of Reaction
Enthalpy Change of Process
ix
List of Figures All figures presented in this thesis, with the exceptions of; 2.1, 4.8 and 4.13 are my own work. Date:
14/08/2013
Signature:
Matthew Kenneth Neeves 1.1 – A Simple Electrochromic Cell
3
1.2 – Illustration depicting; reflective, transmissive, transmissive/reflective and diffuse/absorbing electrochromic devices
5
2.1 – Density of states representation of electronic band structure for a material 10 2.2 – Three dimensional representation of sodium chloride (NaCl) cubic lattice crystalline structure 2.3 – Density of states representation of electronic band structure for a disordered material
10 11
2.4 – Density of states representation of electronic band structure for doped semiconductor materials
12
2.5 – Diagram representation of the intercalation process at the surface of an electrochromic cathode
14
2.6 – Density of states illustration of electronic band structure for an electrochromic material
15
2.7 – Transmitted and reflected light at an interface between two materials
17
2.8 – Diagram depicting light passing through a thin film of finite thickness
18
2.9 – Chromaticity diagram depicting the CIE colour space 1931 (2° standard observer)
x
25
3.1 – The periodic table of elements with depiction of electrochromic metal oxides
29
3.2 – Representation of the crystal structure of ‘insoluble’ PB
39
3.3 – Representation of the crystal structure of ‘soluble’ PB
39
3.4 – Illustration of non‐equivalent indium sites and oxygen vacancies in an In2O3 crystal
42
3.5 – Schematic Diagram of a DC sputtering System
48
3.6 – Schematic Diagram of a RF sputtering System
49
3.7 – Schematic diagram of a sputtering target with magnetron
50
3.8 – Schematic representation of an electron beam evaporation source
51
3.9 – Illustration of an electrolytic cell
53
3.10 – A schematic representation of how SEM works
56
3.11 – Illustration of constructive interference x‐ray diffraction (Bragg diffraction) 58
3.12 – Schematic representation of a three electrode cell with a typical Ag/AgCl reference electrode
59
3.13 – Potential versus time during CV analysis
61
3.14 – Potential versus time during CA analysis
61
4.1 – The Optimal UCS40 commercial ultrasonic cleaning system
63
4.2 – The Satis MSLab 370 Electron Beam Evaporation System
64
4.3 – Annotated vacuum chamber interior of the Satis MSLab 370 series electron beam evaporation system
65
4.4 – The MicroDyn 40000 Series Microwave‐Assisted Commercial Pulsed DC Magnetron Sputtering System with custom substrate tooling
66
4.5 – Schematic representation of the MicroDyn 40000 series.
67
4.6 – The CVC AST 304 RF Sputtering System
68
4.7 – The AML PlasmaCoat Plus II
69
4.8 – Rotating carousel assembly
70
4.9 – Photograph of different substrate area masks
71
4.10 – The Hitachi U‐3501 spectrophotometer
73
4.11 – The Aquila Instruments nkd‐8000 spectrophotometer
74
xi
4.12 – The Hitachi S4100 FESEM fitted with a model 6566 Oxford Instruments Gem Germanium detector
75
4.13 – The Siemens D5000 X‐Ray Diffractometer
75
4.14 – The Dektak 3ST surface profilometer
76
4.15 – The Jandel Engineering four point probe
76
4.16 – Radiometer Analytical PGZ301 Potentiostat
77
4.17 – Ivium IviumSTAT Potentiostat
77
5.1 – ITO coated microslide produced at ambient room temperature attached by the carbon tape method
84
5.2 – Sheet resistance measurement positions and results for different samples attached by the carbon tape method
85
5.3 – AutoCAD™ image of cassette type substrate holder tooling
87
5.4 – Prototype of the cassette holder arrangement
87
5.5 – Close up photograph of substrate holder assembly with cassette partially removed
88
5.6 – ITO produced at ambient room temperature with the substrate holder assembly
89
5.7 – Sheet resistance measurement positions and results for different samples attached using the substrate holder cassette assemblies
90
5.8 – Depiction of substrate / target distance variation between centre and edges of a flat substrate and target material when attached to a drum
91
5.9 – Substrate attachment / holder positions across width of the drum
91
5.10 – Prototype substrate holder
93
5.11 – Calotte with final completed slide holder assemblies attached
93
5.12 – Close up of attached substrate holder assembly, complete with substrate and mask
94
5.13 – AutoCAD™ drawing of electrochemical cell 1
95
5.14 – Electrochemical cell 1 pictured atop of the Hitachi U‐3501
95
5.15 – Diagram displaying positioning of cell layout
96
5.16 – In‐situ and Ex‐situ spectra obtained for an ITO‐coated microslide substrate
97
5.17 – The Scalar Technologies ScalarGauge film thickness measurement system
99
5.18 – Electrochemical Cell 1 with fibre optic lenses installed
100
5.19 – In‐situ spectra obtained for ITO‐coated microslide substrate
101
xii
102
5.20 – Spectroelectrochemical Cell 2
5.21 – In‐situ and Ex‐situ spectra obtained for an ITO‐coated microslide substrate 103 5.22 – In‐situ spectra obtained for ITO‐coated microslide substrate
104
5.23 – A schematic of the software control system for spectroelectrochemical analysis
106
5.24 – A schematic representation of the spectroelectrochemical measurement setup
106
5.25 – Fitted optical transmittance and reflectance of ITO coated microslide substrate as produced using the MicroDyn 40000 series deposition system
109
5.26 – n and k values calculated from the fitting of optical spectra presented in figure 5.25
110
5.27 – Fitted optical transmittance and reflectance of ITO coated silica disc substrate as produced using the PlasmaCoat deposition system
111
5.28 – n and k values calculated from the fitting of optical spectra presented in figure 5.27
112
5.29 – XRD spectra of ITO films produced with various O2 flows
113
5.30 – Sheet resistance versus O2 flow for ITO films produced using the PlasmaCoat
114
5.31 – XRD spectra of ITO films produced with varying deposition power
116
5.32 – XRD spectra of ITO films produced with varying plasma source power
117
5.33a – Ex‐situ transmittance change for a 200nm electrochromic NiO thin film with 40mC of inserted charge
120
5.33b – Colouration times for nickel oxide films deposited upon ITO coated microslide substrates with varying sheet resistance
121
5.34 – Cyclic voltammograms of electrochromic NiO films
122
5.35 – Sheet resistance measurement positions and results for ITO film deposited upon curved polymer substrate
124
5.36 – Typical SEM Micrographs of electron‐beam deposited NiO thin film
126
5.37 –XRD patterns for electron‐beam deposited NiO thin films of varying thicknesses 5.38 – XRD patterns of an electrochromic NiO thin film deposited upon an ITO‐ coated microslide substrate 5.39 – Photograph of ITO coated microslide substrate with an electrochromic NiO layer depicting the as‐deposited, bleached and coloured appearance
xiii
127 128 130
5.40 – CIE 1931 (2° standard observer) chromaticity diagram depicting ex‐situ colour shift of optimised electrochromic NiO films from bleached to coloured
131
5.41 – In‐situ bleached and coloured spectra and colouration efficiency versus wavelength for a NiO thin film
132
5.42 – Cyclic voltammograms of electrochromic NiO films
133
5.43 – Response time versus O2 flow during deposition
135
5.44 – Colouration eff. versus O2 flow during deposition
136
5.45 – ΔT% versus O2 flow during deposition
136
5.46 – Response times and colouration efficiencies as a function of film thickness and transmittance change 5.47 – As‐deposited, fully bleached and fully coloured in‐situ spectra for electrochromic films of NiO with different thicknesses
138 139
5.48 – In‐Situ colouration efficiency versus Wavelength for electrochromic NiO in 1.0M LiClO4‐PC electrolyte
141
5.49 – In‐situ spectra of NiO thin films cycled 1.0M LiClO4‐PC electrolyte
141
5.50 – In‐situ time evolution spectra for electrochromic NiO in 1.0M LiClO4‐PC electrolyte
142
5.51 – In‐situ transmittance spectra for NiO films with different quantities of inserted charge
146
5.52 – Ex‐situ transmittance spectra for NiO films with different quantities of inserted charge
147
5.53 – In‐situ transmittance versus inserted charge for different wavelengths
147
5.54 – In‐situ absorbance versus inserted charge Q for different wavelengths
148
5.55 – In‐situ absorbance versus inserted charge with associated linear fit
149
5.56 – Response time versus bleaching point for colouring and bleaching of NiO film
152
5.57 – In‐situ bleached and coloured transmittance @ 550nm versus bleaching point of NiO films
152
5.58 – Typical SEM micrographs of an electrochromic Ni‐Cr oxide thin film
154
5.59 – Typical SEM micrographs of an electrochromic Ni‐Cr oxide thin film
155
5.60 – Photograph of ITO coated microslide substrate with an electrochromic NiO layer following electrochromic cycling
155
5.61 – CIE 1931 (2° standard observer) chromaticity diagram depicting the ex‐situ xiv
colour shift of optimised electrochromic Ni‐Cr oxide films
156
5.62 – Cyclic voltammograms of electrochromic Ni‐Cr oxide films
158
5.63 – Colouration times for films of Ni‐Cr oxide produced at varying O2 flows
160
5.64 – CE at 10th and 100th cycle versus O2 Flow during deposition
161
5.65 – ΔT at 550nm for the 10th and 100th cycle versus O2 Flow during deposition
161
5.66 – In‐situ spectra obtained for Ni‐Cr oxide film on ITO coated microslide substrate in the bleached and coloured states
162
5.67 – Typical SEM Micrographs of electron‐beam deposited TiO2 thin film
164
5.68 – XRD patterns of an electron‐beam deposited TiO2 thin film
165
5.69 – Photograph of ITO coated microslide substrate with an electrochromic TiO2 layer
165
5.70 – CIE 1931 (2° standard observer) chromaticity diagram depicting the in‐situ colour shift of optimised electrochromic TiO2 films
166
5.71 – In‐situ bleached and coloured spectra and colouration efficiency versus wavelength for a TiO2 thin film
167
5.72 – Cyclic voltammograms of electrochromic TiO2 films
169
5.73 – In‐situ transmittance spectra for TiO2 films with different quantities of inserted charge
171
5.74 – In‐situ transmittance versus inserted charge for different wavelengths
171
5.75 – In‐situ absorbance versus inserted charge Q for different wavelengths
172
5.76 – In‐situ absorbance versus inserted charge for ≈ 400nm TiO2, with associated linear fit
172
5.77 – Photograph of ITO coated microslide substrate with an electrochromic PB layer
175
5.78 – SEM micrograph of a typical PB film as electrodeposited upon ITO coated glass microslide substrate
175
5.79 – Time lapse photography of the first electrochromic cell produced in the course of this work
176
5.80 – In‐situ transmittance spectra for PB films
177
5.81 – CIE 1931 (2° standard observer) chromaticity diagram depicting the in‐situ colour shift of optimised electrochromic PB
178
5.82 – Cyclic voltammogram of electrochromic PB film in 1M KCl
179
5.83 – In‐situ transmittance versus inserted charge for the PW PB transition
180
xv
5.84 – In‐situ transmittance versus inserted charge for the PB PW transition
181
5.85 – In‐situ colouration efficiency versus wavelength for the PB/PW redox couple
182
5.86 – In‐situ transmittance versus inserted charge for the PY PB transition
183
5.87 – In‐situ transmittance versus inserted charge for the PB PY transition
183
5.88 – In‐situ colouration efficiency versus wavelength for the PY/PB redox couple
184
5.89 – In‐situ transmittance versus inserted charge for the full range PY PW transition
185
5.90 – In‐situ transmittance versus inserted charge for the full range PW PY transition
186
5.91 – In‐situ absorbance versus inserted charge Q for different wavelengths for the PW PB transition
187
5.92 – In‐situ absorbance versus inserted charge for PW PB transition
187
5.93 – In‐situ absorbance versus inserted charge Q for different wavelengths for the PB PY transition
188
5.94 – In‐situ absorbance versus inserted charge Q for different wavelengths for the PY PB transition
189
5.95 – In‐situ absorbance versus inserted charge for PB PY transition
189
5.96 – In‐situ absorbance versus inserted charge for PY PB transition with associated linear fit
190
5.97 – Typical SEM micrograph images of an electron‐beam deposited WO3 thin film
192
5.98 – Typical SEM micrograph surface images of an electron‐beam deposited WO3 thin film
192
5.99 –XRD patterns of an electron‐beam deposited WO3 thin film
193
5.100 – Photograph of ITO coated microslide substrate with an electrochromic WO3 layer
194
5.101 – CIE 1931 (2° standard observer) chromaticity diagram depicting the in‐ situ colour shift of optimised electrochromic WO3
195
5.102 – Bleached and coloured spectra and colouration efficiency versus wavelength for a WO3 thin film
196
5.103 – Cyclic voltammograms of electrochromic WO3
197
xvi
5.104 – In‐situ transmittance spectra for WO3 films (480nm thickness) with different quantities of inserted charge
199
5.105 – In‐situ transmittance versus inserted charge for different wavelengths
200
5.106 – In‐situ absorbance versus inserted charge Q for different wavelengths
200
5.107 – In‐situ absorbance versus inserted charge for WO3, with associated linear 201 fit 5.108 – Agar + KOH electrolyte based device pictured in coloured and bleached states
205
5.109 – Empty electrochromic cell constructed from glass
206
5.110 – Completed electrochromic device following construction; on paper and in front of screen
207
5.111 – Cyclic voltammograms of device constructed with Agar + LiTFMS electrolyte
208
5.112 – Ex‐situ spectra obtained for bleached and coloured transmittance of the device
209
5.113 – Picture of the device after testing
210
5.114 – NiO / WO3 and NiO / TiO2 large area devices
211
5.115 – Ex‐situ spectra obtained for bleached and coloured transmittance of a large area ITO / NiO / Agar + Lithium TFMS / WO3 / ITO device
212
5.116 – CIE 1931 (2° standard observer) chromaticity diagram depicting the ex‐ situ colour shift of a large area ITO / NiO / Agar + Lithium TFMS / WO3 / ITO device
xvii
213
Page Intentionally Blank
xviii
Contents Abstract ............................................................................................................ i Declaration ..................................................................................................... iii Acknowledgements.......................................................................................... v List of Symbols and Abbreviations ................................................................... vii List of Figures ....................................................................................................x Contents ........................................................................................................ xix Chapter 1 – Introduction .................................................................................. 1 1.1 – What is ‘Electrochromism’? ............................................................................. 1 1.2 – A Typical Electrochromic Device ...................................................................... 2 1.3 – Applications of Electrochromism ..................................................................... 4 1.4 – Objective and Scope of investigation ............................................................... 6 Chapter 2 – Theory ........................................................................................... 9 2.1 – Electronic Band Structure................................................................................. 9 2.2 – Electrochemistry of Electrochromism ............................................................ 13 2.3 – Optical Theory and Performance Measurement ........................................... 15 2.3.1 – Optical Theory ..................................................................................... 15 2.3.2 – Performance Measurement ................................................................ 19 2.3.3 – Quantification of Colour ...................................................................... 23 Chapter 3 – Materials and Methods ................................................................ 28 3.1 – Inorganic Electrochromic Materials ............................................................... 28 3.1.1 – Overview ............................................................................................. 28 3.1.2 – Electrochromes of Interest .................................................................. 33 3.2 – Transparent Conductive Oxides ..................................................................... 41 3.3 – Thin Film Deposition....................................................................................... 43 3.3.1 – Overview of Techniques ...................................................................... 43 xix
3.3.2 – Techniques of Interest......................................................................... 46 3.3.3 – Film Thickness Control ........................................................................ 53 3.4 – Physical Characterisation ............................................................................... 55 3.4.1 – Scanning Electron Microscopy ............................................................ 55 3.4.2 – Energy Dispersive X‐Ray ...................................................................... 57 3.4.3 – X‐Ray Diffraction (XRD) ....................................................................... 57 3.5 – Electrochemical Characterisation .................................................................. 59 3.5.1 – Cyclic Voltammetry ............................................................................. 60 3.5.2 – Chrono Amperometry ......................................................................... 61 Chapter 4 – Experimental ............................................................................... 62 4.1 – Thin Film Deposition....................................................................................... 62 4.1.1 ‐ Substrates ............................................................................................ 62 4.1.2 – Optimal UCS40 Commercial Ultrasonic Cleaning System ................... 63 4.1.3 – Satis MSLab 370 Electron Beam Evaporation System ......................... 64 4.1.4 – MicroDyn 40000 Series Microwave‐Assisted Commercial Pulsed DC Magnetron Sputtering System ........................................................................ 66 4.1.5 – CVC AST 304 ........................................................................................ 68 4.1.6 – AML PlasmaCoat Plus II ....................................................................... 69 4.1.7 – Substrate Masking ............................................................................... 70 4.1.8 – Substrate Holder Tooling .................................................................... 71 4.2 – Optical Characterisation ................................................................................. 72 4.2.1 – Hitachi U‐3501 Spectrophotometer .................................................... 72 4.2.2 – Aquila Instruments nkd‐8000 Spectrophotometer ............................. 73 4.3 – Physical Characterisation ............................................................................... 74 4.3.1 – Scanning Electron Microscopy and Energy Dispersive X‐Ray .............. 74 4.3.2 – X‐Ray Diffraction (XRD) ....................................................................... 75 4.3.3 – Stylus Profilometry .............................................................................. 76 4.3.4 – Sheet Resistance Measurement .......................................................... 76 4.4 – Electrochemical Analysis ................................................................................ 77 4.4.1 ‐ Instrumentation ................................................................................... 77 4.4.2 – Electrolytes .......................................................................................... 78 4.4.3 – Cyclic Voltammetry ............................................................................. 79
xx
4.4.4 – Chrono Amperometry ......................................................................... 79 4.5 –Spectroelectrochemical Analysis..................................................................... 80 Chapter 5 – Measurements, Results & Discussion ........................................... 82 5.1 – Commercial Scalability ................................................................................... 83 5.1.1 – Preliminary Work ................................................................................ 83 5.1.2 – Development of Substrate Tooling for the MicroDyn 40000 Series Commercial Sputtering System ....................................................................... 86 5.1.2 – Development of Substrate Tooling for the Satis MSLab 370 Commercial Scale Electron Beam Deposition System .................................... 92 5.2 – Acquisition of Spectroelectrochemical Data .................................................. 94 5.2.1 – Electrochemical Cell 1 and Hitachi U‐3501 ......................................... 94 5.2.2 – Broadband spectroelectrochemical measurement ............................ 98 5.2.3 – Spectroelectrochemical Cell 2 ........................................................... 102 5.2.4 – Synchronised Recording .................................................................... 104 5.3 – Indium Tin Oxide (ITO) ................................................................................. 108 5.3.1 – Optical and Physical Properties ......................................................... 108 5.3.2 – Effect of Oxygen Flow During Deposition ......................................... 112 5.3.3 – Effect of Deposition Power / Rate ..................................................... 115 5.3.4 – Effect of PlasmaCoat Hollow Cathode Plasma Source ...................... 116 5.3.5 – Effect of TCO sheet resistance upon electrochromic performance .. 118 5.3.6 – Deposition of ITO upon Curved Substrates ....................................... 123 5.4 – Nickel Oxide .................................................................................................. 125 5.4.1 – Physical Properties ............................................................................ 125 5.4.2 – Electrochromic Properties ................................................................. 129 5.4.3 – Effect of O2 process gas flow on electrochromic films of NiO .......... 134 5.4.4 – Effect of Film Thickness ..................................................................... 137 5.4.5 – Lithium and hydroxide based electrolytes ........................................ 140 5.4.6 – Linearity of Colouration..................................................................... 143 5.4.7 – Bleach Point Selection ....................................................................... 150 5.4.8 – Final Thoughts and Feasibility ........................................................... 153 5.5 – Nickel‐Chromium Oxide ............................................................................... 154 5.5.1 – Physical Properties ............................................................................ 154 5.5.2 – Electrochromic Properties ................................................................. 155 xxi
5.5.3 – Effect of O2 Flow during Sputtering................................................... 159 5.5.4 – Final Thoughts and Feasibility ........................................................... 162 5.6 – Titanium Oxide ............................................................................................. 163 5.6.1 – Physical Properties ............................................................................ 163 5.6.2 – Electrochromic Properties ................................................................. 165 5.6.3 – Linearity of Colouration..................................................................... 169 5.6.4 – Final Thoughts and Feasibility ........................................................... 173 5.7 – Prussian Blue ................................................................................................ 174 5.7.1 – Preparation ........................................................................................ 174 5.7.2 – Physical Properties ............................................................................ 174 5.7.3 – Electrochromic Properties ................................................................. 176 5.7.4 – Effect of Inserted Charge................................................................... 180 5.7.5 – Linearity of Colouration..................................................................... 186 5.7.6 – Final Thoughts and Feasibility ........................................................... 190 5.8 – Tungsten Oxide............................................................................................. 191 5.8.1 – Physical Properties ............................................................................ 191 5.8.2 – Electrochromic Properties ................................................................. 194 5.8.3 – Linearity of Colouration..................................................................... 198 5.8.4 – Final Thoughts and Feasibility ........................................................... 201 5.9 – Summary of Electrochromic Properties ....................................................... 202 5.10 – Devices ....................................................................................................... 204 5.10.1 – Preparation of Electrolytes ............................................................. 204 5.10.2 – Test Devices ..................................................................................... 205 5.10.3 ‐ Larger Area Devices ......................................................................... 210 Chapter 6 ‐ Conclusions ................................................................................. 214 6.1 – Summary of Conclusions .............................................................................. 214 6.2 – Suggestions for Future Work ........................................................................ 217 Bibliography ................................................................................................. 219 Appendices ................................................................................................... 244
xxii
Chapter 1 – Introduction
1.1 – What is ‘Electrochromism’? The term ‘chromism’ refers to a process that results in the reversible change of colour in a compound. This may be caused by a variety of different stimuli or reactions. An area of chemistry producing wide and varied chromism is that of the reduction/oxidation or ‘redox’ reaction. Redox reactions very often result in the colour change of a compound, and many but not all redox reactions are reversible. A great many colour‐change inducing redox reactions have been discovered since the early 19th century, one of the most notable being that of the redox colouration of Prussian Blue (albeit irreversible in its early applications dating back to as far as 1704 and therefore not strictly ‘chromic’) [1]. As chemical indicators, reversible (and hence, truly ‘chromic’) redox reactions have encouraged a great deal of interest and advancement since the first ‘definite’ reversible redox indicator, diphenylamine, was discovered by J. Knop in 1925 [2]. In 1930, Kobosew and Nekrassow displayed the first recorded incidence [1] of the electrochemically induced colour change of a solid, through the reduction of tungsten trioxide in acid electrolyte, which by the method employed was found to be irreversible [3] and therefore not true ‘chromism’. However, two decades later Brimm et al. extended this work on tungsten trioxide and demonstrated a true reversible colour change [4]. The study of electrochromism was then developed further and gained greater interest from the wider scientific community following landmark studies by Deb [5]. Where chromism occurs through initiating and / or controlling the redox reaction (and subsequent change of colour and /or other property, i.e. transmittance) by means of an electric current it can be said to be ‘electrochromic’. However there are cases where this term may refer to other phenomena, and it was first coined in the description of a molecular Stark effect in 1961 [6], [7]. Also it is often and perhaps unfortunately, used to describe a suspended‐particle‐device or ‘SPD’ window, which is a completely different technology and as Monk et al. suggests, may ‘damage the reputation’ of true
1
redox‐based electrochromic technology [1]. Additionally, the term has found use in the description of electrokinetic‐colloidal systems and incorrectly, in some gasochromic window systems. For the purpose of this thesis, the term is used to describe electrochromism in the generally accepted redox sense. These materials and subsequently, devices, have experienced and continue to attract much interest and investigation due to their significant potential in many applications including; smart windows, dimmable mirrors, battery charge indicators, sunglasses and display devices. In almost all known situations where true electrochromism occurs and is deliberately initiated, the redox reaction and subsequent colour or transmittance change occurs as a result of ion insertion and extraction [1] (known as ‘intercalation’ and ‘deintercalation’ respectively). Those materials which colour upon reduction (insertion of anions or extraction of cations) are said to be ‘cathodic’ electrochromes, while those which colour in response to the oxidation (extraction of anions or insertion of cations) are ‘anodic’ electrochromes. One of the largest and widely researched groups of electrochromically active materials is that of the inorganic transition metal oxides, of which some species exhibit anodic, cathodic or even mixed electrochromism.
1.2 – A Typical Electrochromic Device As a result of the continuing development of electrochromic devices, many different permutations of device configuration have been realised. All consist of varying types of ‘cell’, which adopt the use of electrodes, some form of ion storage, ion transport and the electrochromic material which may be present on the anode or cathode depending on its operation. This cell is then usually encapsulated inside some form of clear inert material, e.g. glass or a substrate.
2
The configuration most commonly adopted in most devices is the layered cell and it is this type which is given by Granqvist as the ‘most important’ [8], and is certainly the most utilised [9]–[12]. Figure 1.1 shows a typical layered cell configuration:
Figure 1.1 – A Simple Electrochromic Cell (1) Substrate (2) Transparent conductor (3) Ion storage material (4) Ion transport / electrolyte (5) Electrochromic material (Diagram based upon [8] – Ch. 1, Fig 1.1, Page 2) Both the ion storage (3) and electrochromic material (5) layers must be capable of the conduction of both electrons and the required ions. This conduction of ions occurs via an ‘intercalation’ process, whereby the host material structure is penetrated by a guest ion resulting in modification of the host compound. In some cases, depending on the materials being employed, the ion storage layer might also be an electrochromic material, and then its electrochromic property must be complementary to that of the electrochromic layer [8] so that the device provides the desired chromic properties. Since having two electrochromic layers contributing to the cell can often lead to greater overall chromic effect [8], [13], it is often a desired characteristic and such cells are known as ‘complementary devices’ [1], [8].
3
The ion transport material (4) can be in the form of a; liquid, gel, solid or quasi‐solid electrolyte material and must possess high ionic conductivity whilst also acting as an electron insulator. Conversely, the transparent conductors require good electron conductivity without the need to conduct ions. Inside such a device, the application of electrical potential to the electrodes provides an electromotive force [1] in the cell which in turn provides the energy required for the redox reaction to take place. Depending upon the electrochromic material in use, the intercalation or deintercalation which occurs at this point results in the desired change in colour or other optical property.
1.3 – Applications of Electrochromism It was not long after the initial discovery of electrochromism that suggestions for the commercial uses of electrochromic devices came about, the earliest example which is widely cited being the British patent filed by F. H. Smith in 1929 where in a clear solution containing an iodide and dye precursor develops a colour when an electrical current is passed through it, and returns to its bleached state upon removal of the electrical current [14]. Following this there were a number of early patent applications whereby electrochromism was employed to provide changes of colour [15], level of light transmittance [16], and even reflectivity [17], [18]. Since these early patents, thousands more have followed, along with a very large and increasing number of publications. The most commonly cited applications of recent times being that of dimmable mirrors [1], [8], [10], [19]–[21] and smart windows [1], [8], [22]–[24]. Other potential applications which have been identified and have subsequently attracted attention include; display devices of various types [1], [8], [25]–[32], printing mediums [1], [32] including security marks [33], [34], detection of latent fingerprints [35], [36], battery charge indicators [37], thermal exposure indicators for frozen food [38], motorcycle visors [39], [40], sunglasses [41], [42] and even coatings for stealth military aircraft [43] 4
Generally speaking, all of these applications will involve a device operating in a reflective mode, transmissive mode, or derivative thereof.
Figure 1.2 – Illustration depicting; reflective (1), transmissive (2), transmissive/reflective (3) and diffuse/absorbing (4) electrochromic devices. Arrows indicate path and intensity of incident electromagnetic radiation (light). A similar representation has been presented by Granqvist [8]. Figure 1.2 provides an illustration of electrochromic devices operating in various modes. Device permutations utilising these traits may be realised to provide different functionality, e.g.:
5
Purely transmissive electrochromics may be used for various technologies such as optics, eyewear and smart windows.
Transmission varying electrochromes with reflective electrodes as in the reflective case, 2 above, may be used for dimmable mirrors.
Transmissive / reflective electrochromic devices, such as those constructed with Cu and Bi [44], [45], may be used for switchable mirrors.
Display devices may be realised by combining transmission varying electrochromes with a diffuse white backing or electrode.
1.4 – Objective and Scope of investigation The primary aim of this investigation is the development of electrochromic materials and devices suitable for commercial scale production with a view toward inclusion in wearable optics, where the optical transmittance properties need to be reversibly and quickly changed over a significant range. The main purpose of which is providing continuous, comfortable and safe vision for the wearer, even when the wearer encounters a wide variety of ambient lighting, as can occur when a car driver passes from bright sunlight to a dimly illuminated tunnel, and vice versa. This work is important since at the point of writing, no suitably mature technology exists to satisfy these requirements and most studies concentrate upon methods or work upon equipment which is not directly scalable. Concentrating upon suitable candidates for each layer of a solid state electrochromic device, characterisation and testing of materials will facilitate further investigation of the deposition processes required, such that greater understanding and insight will allow films which are both reproducible on a commercial scale and optimised for electrochromic device application.
6
For a device to be realised, each layer described in figure 1.1 requires investigation. The objectives and scope required for this project can be summarised as follows:
Selection, optimisation and deposition of a suitable transparent conductive oxide (TCO) material for transparent conductor layers (2 – figure 1.1)
Investigation of inorganic electrochromic materials and / or ion storage layers, suitable deposition techniques and conditions for both anode and cathode such that optimisation for eyewear application can be performed with respect to; safety, colour, speed of operation, longevity, efficiency and commercial scalability. (3, 5 – figure 1.1)
Realisation of a suitable method of determining the electrochromic properties of a material or device via the in situ investigation of optical properties in parallel
with
electrochemical
manipulation
and
data
acquisition
(spectroelectrochemical characterisation).
Further characterisation of films by FESEM, EDX, XRD, CV, EIS and optical spectroscopy, and subsequent examination of the relationship between the acquired data and the deposition conditions / resultant electrochromic performance.
Optimisation of deposition conditions for the reliable and reproducible production of efficient and durable electrochromic materials and devices.
Development of suitable substrate masking arrangements for the deposition and subsequent assembly of electrochromic devices.
Development and production of demonstration devices and functional device prototypes.
Investigation into suitable methods of controlling and enhancing electrochromic devices.
7
This thesis is structured as follows:
The remainder of chapter 1 provides an introduction into; what the phenomenon of electrochromism is, its beginnings, how it works, basic device structure, and a summary of the wide and varied applications of electrochromism.
Chapter 2 explores the theory behind electrochromism and the mechanisms behind electrochromic behaviour.
Chapter 3 provides background to those materials which display electrochromism, specifically inorganic materials, and identifies which of these materials are of interest for the purposes of this investigation. Other materials which are required in the creation of electrochromic devices (i.e. transparent conductive oxides) are also investigated, as are the processes employed in both the deposition and characterisation of each of the thin film material layers which make up an electrochromic device.
Chapter 4 summarises the materials, methods and equipment employed and developed in the course of this thesis.
Chapter 5 presents the investigations undertaken and results produced in the course of this work, with relevant discussion of the produced results.
Chapter 6 summarises and concludes the achievements and findings produced throughout the project and presented in this thesis, before presenting suggestions for future work.
8
Chapter 2 – Theory 2.1 – Electronic Band Structure
All solid materials possess what is known as an electronic band structure. The electronic band structure of the material refers to the possible (and discrete) energy states which electrons within the material (as contributed by the outer orbitals of atoms comprising the material) may possess, and is responsible for determining both the optical and electronic properties of a material. The two most important energy bands are known as the conduction band (the highest energy band containing mobile electrons) and the valence band (the band immediately below). The gap between these two bands comprises a forbidden area known as the material ‘band gap’ where no allowed electron states exist in pure materials. At temperatures above absolute zero, electrons within the material may occupy the valence and/or conduction bands. A full valence band with an unoccupied conduction band results in an insulating material, if the energy gap is very much greater than the available thermal energy while a half‐ filled conduction band results in a conductor. In pure (intrinsic) semiconductors the energy gap is comparable to the available thermal energy and some electrons are promoted to the conduction band, leaving unoccupied states (holes) in the valence band. Overlapping and thus partially filled bands also occur, allowing electrons to move freely to unoccupied states and resulting in electrical conductivity and metallic behaviour. The point at which the probability of an energy state being occupied is equal to ½, is referred to as the ‘Fermi level’. The position of the Fermi level is unique to a material and is dependent upon its structure. In a perfectly pure, crystalline material such as sodium chloride (NaCl), atoms will form a regular lattice comprising of identical repetitive units (figure 2.2). Any disorder will result in changes in the structure and subsequently affect the material band gap. Disorder may result from contamination, deliberate doping with foreign atoms or groups, internal displacement of atoms or groups within the material (vacancies), or lack of crystallinity (i.e. amorphous nature).
9
Vacancies and substitute atoms or groups may lead to localised band gap states or ‘tails’ which may act to reduce or even bridge the band gap thereby influencing the materials behaviour [46].
Figure 2.1 – Density of states representation of electronic band structure for a material, depicting the; valence band, conduction band, band gap and Fermi level at absolute zero. Where the band gap of a material is great it will exhibit insulator behaviour (a), whereas where the band gap is small it will exhibit semi‐conductor behaviour (b). Conductive ‘metals’ may be represented by and overlap of the valence and conduction bands (c).
Figure 2.2 [47] – Three dimensional representation of sodium chloride (NaCl) cubic lattice crystalline structure (also known as a rock salt or halite structure) comprising of repeating Na+‐Cl‐‐Na+‐Cl‐ units where each positive sodium ion is surrounded by six negative chlorine ions (and vice versa) forming consecutive octahedral units. 10
In this more complex scenario the original energy band positions comprising the majority of the material constitute the ‘extended states’, and the energy separating these extended states from the localised states is known as the mobility edge (figure 2.3).
Figure 2.3 ‐ Density of states representation of electronic band structure for a disordered material, depicting the existence of localised states or ‘band tails’ and the position of the mobility edge which separates the extended states from the localised states. The conduction and valence bands represent the extended states of the bulk. Doping of a material structure with substitute foreign atoms or groups may also result in the creation of additional discrete permitted energy levels within the band gap (figure 2.4). Depending upon the number of electrons present in its outer valence shell, the doping material may act as an electron donor or acceptor, or n‐type or p‐type dopant respectively. The inclusion of a dopant will affect the position of the Fermi Level energy and reduce the energy required for electrons within the material to move between states, thereby enhancing the electron mobility of the material. In the case of the transition metal oxides, the position of the transition metal Fermi level (in its oxide state) will determine the optical and electrical properties of the material and what type (if any) of electrochromism will take place. Where the band gap energy of the material is larger than the energy of incident electromagnetic radiation, there is no interaction with electrons and thus no absorption, resulting in a transparent material. If the incident electromagnetic radiation energy exceeds the band gap of the material, it will interact and the material will become absorbing.
11
Figure 2.4 ‐ Density of states representation of electronic band structure for doped semiconductor materials, with n‐type (left) and p‐type (right) doping. The addition of a dopant facilitates the existence of a new permitted state within the band gap of the material. The desirable properties of transparent conductive oxide (TCO) materials also rely on the electronic band structure and the position of the Fermi level within the band gap of indium‐tin oxide (ITO). The high transmittance of ITO in the visible region is due to the electronic band structure of the material possessing a wide band gap with a low absorption edge which is found to shift to shorter wavelengths due to Moss‐Burstein shift [48], [49]. However, Sn atoms act as n‐type electron donors, resulting in creation of a donor band just below the conduction band. As the level of doping increases, this band increases in size until it reaches a critical density [48] and begins to overlap the conduction band resulting in degenerate free electron behaviour. This is responsible for the electrical conductivity and high IR absorption typical of ITO films as IR photons can excite electrons from the conduction band to higher states in the conduction band (IR absorption is found to increase with conductivity of the films). Oxygen vacancies within the material are also a critical factor in the conductivity and optical properties of ITO and contribute by affecting the oxidation state and subsequently the number of valence electrons that the doping Sn atoms may contribute.
12
2.2 – Electrochemistry of Electrochromism
Electromotive force (EMF) refers to electromotive work per unit of charge, and is given as an electrical potential (in volts). In terms of a charging electrochemical cell, which an electrochromic device represents, the EMF that will ‘drive’ electrons liberated at the anode through the cell is proportional to the Gibbs free energy change, which is most often defined as:
(1)
Where
represents the Gibbs free energy,
is the enthalpy change of the reaction,
is the entropy change for the process itself and represents temperature. The relationship between the Gibbs free energy and the EMF may therefore be expressed by: Where
(2) is the number of moles of electrons,
96485.3365 /
is the Faraday constant (
) and V represents the EMF in volts.
As discussed briefly in chapter 1, the application of electrical potential to the electrodes of an electrochromic cell (as represented in figure 1.1) provides an electromotive force to the cell [1], [50]. Since the electrolyte itself will not conduct electrons, an electrical charge will present itself on the surface of the electrodes. At this point positively charged cations are attracted toward the cathode while negatively charged anions move toward the anode (figure 2.5) due to a process known as electrophoresis [51]. Upon contact with the respective electrodes, a charge balancing redox reaction will take place where electrons are exchanged. Since electrons are many times smaller than the ionic groups, their transport is fast and therefore the reaction occurs at the surface of the material [50], [52]. At a porous electrode, these ions 13
diffuse into the surface to occupy their respective positions at the most negatively charged locations within the material structure [50], [52]. This intercalation reaction results in a change of position of the materials Fermi level (Figure 2.6) and thus its optical properties as described in 2.1. The electrical conductivity and magnetic properties [53]–[55] may also be affected by this change.
Figure 2.5 – Diagram representation of the intercalation process at the surface of an electrochromic cathode. Positively charged ions are attracted to the cathode to meet with negatively charged electrons to undergo a charge balancing redox reaction and diffuse into the structure of the material [50] (not to scale). The opposite of this process occurs during deintercalation; however the diffused ions may remain within the film itself, while not being bonded to it.
14
Figure 2.6 ‐ Density of states illustration of electronic band structure for an electrochromic material, depicting the movement of the material Fermi level as a result of ionic intercalation and de‐intercalation. Band tails are shown since it is typical for many electrochromic materials to be disordered.
2.3 – Optical Theory and Performance Measurement
2.3.1 – Optical Theory Transmittance is the fraction of incident electromagnetic radiation which passes through a sample. It is wavelength dependent and in its simplest form can be represented by the following equation: (3)
Where transmittance T, at wavelength λ is given by the ratio between I0, the intensity of the incident radiation, and I, the intensity of the radiation emerging from the sample. Light consists of electromagnetic radiation, which itself is a form of energy. Since the law of conservation of energy specifically states that energy cannot be created or
15
destroyed, light which is not transmitted must either be reflected or absorbed, therefore (for a non‐fluorescing material): 1
(4)
Where , and represent the fraction of incident light being transmitted, reflected or absorbed at wavelength λ respectively. The refractive index of a material, , is representative of how light propagates through a material and is defined as:
(5)
Where is the speed of light in a vacuum and is the speed of light through the material. Since a fraction of light will be absorbed by the material, the complex index of refraction, , is used:
(6)
Where the real part is the refractive index and determines the loss due to absorption. At a simple (i.e. non‐scattering) interface between two different materials (0 and 1), incident light (at angle
) may only be transmitted or reflected (figure 2.6). A fraction
of the incident light will be reflected away from the interface while the remainder is transmitted at a different angle , given by Snell’s law:
(7)
16
Figure 2.7 – Transmitted and reflected light at an interface between two materials with complex refractive indices and . The coefficients of transmission and reflection can be calculated from Fresnel equations [56]:
Tp
2 N 0 cos 0 N 0 cos1 N1 cos 0
(8)
Rp
N1 cos 0 - N 0 cos1 N1 cos 0 N 0 cos1
(9)
Ts
2 N 0 cos 0 N 0 cos 0 N 1 cos 1
(10)
Rs
N 0 cos 0 - N1 cos1 N 0 cos 0 N1 cos1
(11)
Where and , and and are the transmission and reflection amplitudes for p‐ polarised and s‐polarised light respectively. The Fresnel equations require that the polarisations of the light, p‐polarised and s‐ polarised, are distinguished from one another since two beams of light may not 17
interfere when at right angles to each other, as discovered in 1816 by Fresnel [57]. p‐ polarised light refers to the component which possesses an electric field parallel to the plane of incidence, while s‐polarised light refers to the component with an electric field which is perpendicular to the plane of incidence. At normal incidence the p‐polarised and s‐polarised components are equal (for isotropic materials). When considering light passing through a thin film with a thickness , one must consider both the interface where light enters the film, and the interface where light exits the film. This situation is depicted in figure 2.8 below.
Figure 2.8 – Diagram depicting light passing through a thin film of finite thickness . Note the existence of multiple reflections and transmissions resulting from a single incident beam (which repeat indefinitely creating interference). represents the substrate for a deposited film. As one may appreciate this scenario quickly becomes very complex for multi‐layer systems. Interference occurs between the beams transmitted and reflected from the two interfaces. To account for this and calculate absolute transmittance and reflectance values, the phase shift φ on reflection, which occurs between the subsequent beams reflected internally within the film, must be determined. For the situation above [58]:
φ
2
cos
18
(12)
To simplify the process of calculation, especially in the case of multi‐layer systems, it is possible to assign a characteristic 2x2 matrix, to each layer of the system [58]:
sin ⁄ cos
cos sin
1
(13)
Where B and C are the normalised total tangential electric and magnetic fields [58], and cos
(14)
cos
for p‐polarised light, and [58]
cos
cos
(15)
for s‐polarised light, where; is the permittivity of free space, index and
is found from Snell’s law (
the incident medium and
sin
sin
is the substrate ).
is the index of
is the angle of incidence.
Adopting this approach, one can expand the characteristic matrix to multi‐layer systems, whereby each layer contributes to the characteristic matrix of the system [58]:
cos sin
sin ⁄ cos
1
(16)
Where q represents the number of layers deposited upon the substrate and
represents the substrate, or for a free‐standing film, the emergent medium (e.g. air)
2.3.2 – Performance Measurement For any scientific study, means of quantification are required to be able to compare and contrast results. Indeed, there are many properties with regard to electrochromic 19
materials and devices which warrant interest and investigation. However some of the most widely quoted and accepted measures of electrochromic performance which almost all studies make reference to are that of:
Transmittance Change
Response Time
Colouration Efficiency
Longevity
Transmittance change Transmittance change or ΔTλ may be quantified as either the maximum change of transmittance achievable by the material or device, or the change achieved for a certain quantity of inserted charge Q, at a certain wavelength λ and measured as a percentage value. (17)
Alternatively this performance measure may be given as a related quantity such as; the contrast ratio (CR), a common measure employed for display devices [1], or a percentage with respect to variation in luminous transmittance, Y as defined later in 2.3.3.
Response Time
The response time or ‘transition time’ of an electrochromic material or device is defined as the time taken for the sample to switch between coloured and bleached states. Typically, a faster response is always desirable, however the significance of response time as a quantitative measure of performance depends heavily upon the
20
desired application, and in some applications a slower transition speed may be desired [27]. As has been noted for other measures such as transmission change, valid comparison of the published values in much of the literature is impossible due to lack of consistent measurement [1]. Two common approaches to the repeatable and definite measurement of response are; measurement of the magnitude of transmission change or contrast achieved after a specific amount of time has passed, i.e. ΔTachieved, or alternatively, and most commonly, recording of the time required to achieve a desired transmission change (ΔT) or contrast. How these figures are derived is often unique to the study being presented, as in the case of studies which claim to present time taken for ‘full switching’ or a percentage (i.e. ΔT95%). Such inconsistency makes direct comparisons between studies difficult, especially since colouration with respect to the presented transmission is rarely, if ever, linear. Over the course of the work performed in this thesis, different methods of quantifying response time have been investigated. Herein (with the exception of papers 1 and 2), response times are presented as a variation of approach two as described above. For colouration, it is given as the time taken to insert a set quantity of charge Q (in mC) for an electrode of known surface area. Since the inserted charge is directly proportional to the number of colour centres reduced or oxidised within the material [1], [8], it results in a repeatable colour change (where the reaction is desirably reversible) whilst providing a directly comparable value for other studies of materials; irrespective of thickness or deposition technique. If an absolute definition of this type was able to be established across the field, it would allow for accurate comparisons of data produced by different research groups [1]. In order to ensure that the point at which the charge is inserted is also constant, it is required that a second approach to be taken for the bleaching of devices. This is because the inserted and extracted charges are rarely (if ever) equal, due to wasted currents as a result of other electrode processes occurring. Therefore to achieve repeatability and consistency, the bleach point of materials is given as the point at which the current passing through the electrode (with the bleaching potential applied) meets a defined value, e.g.:
21
0.25
(18)
Through testing, it was found that this approach allowed for a repeatable bleached point to be established for both electrical and transmissive properties.
Colouration Efficiency At normal incidence, and where reflection is considered to be negligible, the absorption (also known as the optical density, OD) can be calculated (with a good degree of accuracy) from the transmittance (as adapted from [59] (see also [60])):
log
log
(19)
Where reflection is to be considered it can be modified to compensate such that:
log
1
(20)
When the thickness of a film is known, one may use this to calculate the (wavelength dependant) absorption coefficient of the material , from the absorbance of the sample and absorbance of the substrate [59]:
(21)
The colouration efficiency CE refers to the change in absorption of inserted charge Q [1]:
22
resulting per unit
(22)
The unit of measurement for CE is cm2C‐1; as derived from area (most usually given in cm2) per unit of inserted charge Q (measured in coulombs, C).
Longevity For a device to be economically and commercially viable it must be able to provide an adequate service life, i.e. in the case of electrochromics, possess the ability to switch between coloured and bleached states many thousands or hundreds of thousands of times, before significant degradation of function and / or failure occurs. The most common method of assessing this is through continuous cycling. By cycling a device many times, preferably until eventual failure, one is able to make an educated assessment of its expected longevity and subsequently its suitability for application. Additionally, and in the case of systems where longevity testing may take a long time (10k cycles or more), one may cause premature degradation and / or failure by cycling a material or device under abusive conditions, i.e. over‐potential and / or over‐ charging. This sort of testing can provide additional information, since if a device still survives the minimum number of cycles required under these overly harsh conditions, not only is its actual service life likely to be many times greater, but the behaviour of the device and the effect of the abuse itself may be gathered. Again, since longevity testing relies upon other un‐standardised factors affecting what a ‘cycle’ is, it is most often difficult or even impossible to directly compare the findings of different studies.
2.3.3 – Quantification of Colour Since worn optics both influence the wearer’s perception of colour and also often form a fashion statement, the colour of the device is often of great concern.
23
Colour can be quantified for light within the ‘visible region’ (approximately 380‐780nm [59], [61], [62]). Through transmission, absorption and reflection different substances and materials will interfere with the incident light and appear to an observer to have ‘colour’. It is estimated that to the human eye, there are over ten million perceivable colours [60]. The science surrounding the quantification of colour is known as ‘colourimetry’. It has historically been, and continues to be, performed in a variety of ways, however perhaps the most internationally recognised system is the Commission Internationale de l’Eclairage or ‘CIE’ colourimetry system [1]. A number of other studies of electrochromic materials have employed the CIE system to quantify the colours and / or luminous transmittance of different electrochromes and devices [1], [27], [63]–[68]. Colourimetric analysis provides a colour specification for users possessing normal vision (i.e. without colour blindness) in the form of tristimulus values. The CIE standards provide mathematical quantification of these colour tristimulus values or ‘co‐ordinates’, which in turn form a ‘colour space’ for the visible region. Most often, commercial colorimeters (such as the Minolta CR‐5 Bench‐top Colourimeter [69]) make use of these mathematical relationships and are employed as measurement devices specifically to provide both tristimulus values and specific details regarding the colour of the material. However since the mathematical equations forming the CIE standard are well known it is possible to use ordinary spectroscopic data for a material or device as obtained by a spectrophotometer and extract the required colourimetric data using the relevant equations. In this work we will consider the first international and perhaps best well‐known standard, CIE 1931 2° observer, and employ it to quantify the colour of investigated electrochromes. This model was chosen, rather than the more modern 10° observer of model of 1964, because eyewear is in near proximity of the eye, and objects and surroundings are viewed through the lenses. A representation of the CIE 1931 2° colour space is given in figure 2.9 (below). To translate spectral data into an equivalent colour as denoted for a CIE 1931 2° standard observer, mathematical calculations are employed to produce tristimulus values or ‘colour coordinates’ for a sample which represent the spectral response of
24
the human eye. These functions are different for emissive and transmissive/reflective cases, because transmissive and reflective samples require an illuminant to be viewed and / or measured.
Figure 2.9 – Chromaticity diagram depicting the CIE colour space 1931 (2° standard observer), produced in Excel™ 2010. This diagram also displays the blackbody or ‘Planckian’ locus which depicts the colour which an incandescent blackbody would assume as its temperature increases. Since electrochromic devices are transmissive and / or reflective, the following equations apply:
1 ̅
d
(23)
1
d
25
(24)
1
d ̅
(25)
Where
d
(26)
Here, λ is the wavelength, Tλ is transmittance, and ̅ , and ̅ are the CIE 1931 2° standard observer colour matching functions (Appendix A4). In this case N represents the function employed to normalise the reference light source and produce the values of expected colour under that type of light source. Since transmittance measurements are obtained from measurement (as might light source data i.e. D65), the data is discrete and therefore it may not be possible to accurately represent it using mathematical equations. Therefore numerical integration can be performed on the data through use of the following summations (so long as the sampling interval of the data and standard observer functions is the same): 1
̅
(27)
1
(28)
1 ̅
(29)
Where
(30)
26
And λ is the wavelength range (380‐780nm @ 5nm intervals) Then, to gain xy co‐ordinates which can be plotted on a CIE chromaticity diagram we perform the following calculations:
(31)
And
(32)
Where x and y are the colour coordinates which specify the hue and Y, as calculated previously, represents the luminance of the colour. For a perfect transmitter, Y=1. Following this, any further colour calculations where necessary can be performed using the calculated tristimulus values X, Y and Z, (i.e. conversion to sRGB values for example).
Chapter Summary Through combination of the theory behind electrochromism with appropriate quantitative measures of performance as discussed in this chapter, it is now possible to look at the appropriate materials and methods known for producing electrochromic layers and devices containing said layers. Chapter 3 provides a brief overview of the inorganic electrochromic materials available, including a survey of those materials utilised in this study, before describing the methods and techniques employed throughout this work for the deposition and characterisation of each of the materials.
27
Chapter 3 – Materials and Methods
3.1 – Inorganic Electrochromic Materials
Since the early research discoveries, electrochromism has been widely studied and its existence is now well known in ‘numerous’ inorganic and organic substances [1], [8], and interestingly even some biological species [1]. A great majority of the existing research focuses on inorganic materials which exploit the electrochromic properties of transition metal oxides [70] (such as the pre‐mentioned tungsten trioxide), and more recently the importance and potential of organic electrochromic materials (such as poly(3,4‐ethylenedioxythiophen) or ‘PEDOT’) has gained significant attention. This work predominantly focuses upon those inorganic transition metal oxide materials which can be deposited using commercially scalable vacuum deposition techniques.
3.1.1 – Overview As mentioned above, perhaps the most prevalent of the inorganic electrochromic materials are the transition metal oxides [1], [8]. Other inorganic materials that, with the exception of Prussian blue, would appear to be less well researched [13] include; iridium, tin and boron nitrides, hetero polyacids, intercalated graphite, and hexacyanometallates (including the pre‐mentioned Prussian Blue).
Transition Metal Oxides
Transition metal oxides are the most widely researched and utilised of the electrochromic materials because they benefit from good photochemical stability [8], [71], which also promotes their use in photochemical areas as good candidates for
28
device longevity. However, depending upon the lattice structures of the oxides, many eventually dissolve in water (for example WO3 [72], [73] ‐ see below) and will break up due to water molecules penetrating and hydrolysing the lattice. Water penetration results in degradation of the oxide layer and subsequently its electrochromic function, therefore Granqvist and Svensson advise that devices should be watertight to ensure longevity [74].
Figure 3.1 – The periodic table of elements (excepting the lanthanide and actinide series’) depicting those transition metal oxides which display well‐ documented electrochromism in the visible region. Similar representations have been presented by Granqvist, Baba and Yu [8]. The type of chromism displayed is specific to the transition metal oxides [13] and the optical absorption of many of the transition metal compounds (including said oxides) can be attributed to ‘intervalence charge transfer’ [75]. In fact many of the unusual properties of transition metals themselves can be attributed to the unique configurations of the partially filled outer d orbital electrons [76], [77]. Many, but not all of the transition metal oxides display electrochromism under redox conditions. Those most well documented are detailed in figure 3.1.
29
Of those electrochromic oxides being depicted above, tungsten trioxide and nickel oxide are almost certainly the most extensively researched. Interestingly vanadium oxide presents itself as a unique material in that it can display both anodic and cathodic electrochromism depending upon its preparation [8].
Metal Nitrides Some nitrides of transition metals also express electrochromic behaviour, most notable are the nitrides of iridium, tin and boron [1], [8], [13], [78]–[80]. These are often prepared by reactive ion plating [8] or more recently reactive evaporation [78]. The electrochromic mechanism is different to that of oxides, whereby alteration of surface adsorbents induces a phenomenon similar to so‐called ‘Burstein‐Moss Shift’ [80]. Although there appears to be few studies into them, oxynitrides can also exhibit electrochromic behaviour, Rousselot et al. have investigated titanium oxynitride (prepared by reactive sputtering) with some success [81].
Heteropoly Acids
Heteropoly acids are widely used as catalysts [82], and have as yet to witness any great level of application despite possessing some favourable characteristics. In their hydrated solid state they will conduct protons in their crystal structures, and many can do so at room temperature (Particularly H3(PMo12O40)‐nH20 (n≈29)) [50]. In this capacity they can act as very effective solid electrolytes, but some forms will also exhibit electrochromic behaviour of their own, for example phosphotungstic acid (H3PO4(WO3)12‐nH2O , n≈29, [83], also known as ‘PWA’) [84], [85]. Colouration efficiency can be similar to that of the transition metal oxide films and they may be mixed with and used in conjunction with other materials [8]. Granqvist notes that there are numerous polyacids [86] and that it ‘seems likely’ new electrochromic materials might be developed using them [8].
30
Intercalated Graphite Numerous texts exist on the intercalation compounds of the various forms of graphite with alkali metals and other substances. There are a number of methods and mechanisms whereby intercalation takes place and the various compounds lend themselves to a wide variety of applications [87]. Their potential as electrochromes was first realised by Pfluger et al. in 1979 [88]. This initial discovery displayed a chromic black to gold colour change with a very fast response time, albeit at a relatively high potential of 3‐5v. Buckminsterfullerene (C60) was also shown to act as an electrochrome with lithium ions in 1993 by Rauh et al. [89] and other metal ions can also be intercalated [90]. The general lack of translucency of most forms of usable graphite based thin films would appear to limit their potential use to reflective display devices as opposed to optics or windows.
Hexacyanometallates Huge numbers of possible hexacyanometallates exist and many display an electrochromic effect at varying wavelengths across the electromagnetic spectrum [8]. With such a great number of potential candidates many permutations of colour, change in transmittance and wavelength are achievable. Similarly to transition metal oxides, the colour produced by hexacyanometallates is a result of intervalence charge transfer, however in this case it represents charge transfer between two metal ions (i.e. two Fe ions in the case of Prussian Blue – see 3.1.2) across the inter‐connecting CN‐ ligands. Hexacyanoferrates have generated much research [8], [91], [92]; with most displaying an electrochromic effect. The earliest of the hexacyanoferrates being identified as an electrochrome was the well‐known pigment – Prussian Blue.
31
Prussian Blue, or iron(III)hexacyanoferrate(II) was first synthesised by accident in 1704 [93] and is the most well‐known and researched of the hexacyanometallates [8]. It lends itself not only as a pigment and potential electrochrome but as a catalyst and basis for varying chemical and biological sensors [94], [95]. Since early discoveries were established and the process was made reversible, it has been employed in working devices [96]–[98]. When used with a complementary cathode (Prussian Blue is anodically coloured), devices can experience large visible‐region changes, typically from translucent to dark blue, but also green and yellow. For this reason, Prussian Blue and the other hexacyanometallates have generated great scientific and technological interest as electrochromes, and even more so since their successful incorporation into all‐solid state devices employing solid electrolytes such as poly(aniline) [96], [97]. Nickel(II)hexacyanoferrate(III)
and
‘ruthenium
purple’
or
Iron(III)hexacyanoruthenate(II) are also electrochromes capable of producing transparent to intensely coloured changes; in this case from transparent to orange (or even red) and also from transparent to green for the former [1], and transparent to purple for the latter [99]. Other hexacyanometallates which are known to exhibit electrochromism in the visible range include: Table 3.1 – List of electrochromic hexacyanometallate compounds and their respective colour changes. Hexacyanoferrate Cadmium Hexacyanoferrate [98] Chromium Hexacyanoferrate [100], [101] Cobalt Hexacyanoferrate [102] Manganese Hexacyanoferrate [1] Molybdenum Hexacyanoferrate [1] Platinum Hexacyanoferrate [103] Rhenium Hexacyanoferrate [100] Titanium Hexacyanoferrate [104]
32
Observed Colour Change [100] White to Transparent Blue to Pale Grey Green‐Brown to Dark Green Pale Yellow to Transparent Pink to Red Pale Blue to Transparent Pale Yellow to Transparent Brown to Pale Yellow
3.1.2 – Electrochromes of Interest
The work presented in this thesis focuses upon four examples of transition metal oxides and a hexacyanoferrate, each chosen for their respective properties and availability to study; nickel oxide (NiO), titanium dioxide (TiO2) , tungsten trioxide (WO3), the novel mixed‐metal oxide, nickel‐chromium oxide (Ni‐Cr Oxide) and iron hexacyanoferrate (most commonly known as Prussian Blue). The well‐known cathodic material, tungsten trioxide (WO3), is arguably the most well researched of the electrochromic transition metal oxides [8] and is included for completeness and for comparison with the lesser characterised films of TiO2.
Tungsten Oxide Tungsten oxide or WO3 (also known as tungsten trioxide, tungsten (IV) oxide or less‐ commonly as tungstic anhydride), is a bulk material with pale yellow colouration [73], [105]. It is by far the most researched of the electrochromic materials [8], and possesses a defect‐perovskite structure [1], [106], [107] with Granqvist describing it as a ‘perovskite‐like’. Its exact structure is particularly temperature dependent [1], [8], [106]. A 5d transition metal oxide, it possesses physico‐chemical properties as per table 3.2: Table 3.2 – Properties for WO3 as reported in or calculated from data provided by the references given. FORMULA
WO3
MOLECULAR WEIGHT 231.8382 [108]
AVERAGE ATOMIC WEIGHT 57.9595 [108]
LATTICE PARAMETER (Å) 3.84 [109] 3.899 [110]
DENSITY (G/CM3)
MELTING POINT (K)
7.16 [73]
1746 [73], [111]
Since the initial discovery of electronically induced colour change in 1930 [3], tungsten trioxide was the first of the transition metal oxides to see considerable research into its chromic properties. In 1949, Straumanis and Dravnieks observed that the lattice
33
structure of WO3 was porous [112] to ions and investigated this behaviour in depth, (essentially identifying that they were susceptible to intercalation). These intercalated forms were shown to possess metallic properties and have been known misleadingly as ‘tungsten bronzes’ since their initial discovery by Wöhler in 1824 [4], [112] due to their metallic lustre and subsequent use in bronze coloured paint. A great deal of research continued to follow, with the first reversible (and truly chromic) colour change being recorded by Brimm et al. in 1951 [4], and by 1969, this was being applied by Deb to WO3 thin films for use in novel photographic applications [5]. Amorphous electrochromic films of WO3 undergo cathodic colouration from transparent to blue under ionic intercalation, while polycrystalline films colour in the near‐IR region. Electrochromic films of WO3 have been successfully deposited by; sputtering [113]–[116] , thermal evaporation [116]–[118], electron beam evaporation [119], [120], pulsed laser deposition (PLD) [121], spray pyrolysis [122], chemical vapour deposition (CVD) [123]–[125], electro‐deposition [126]–[129] and Sol‐Gel techniques [130], [131]. The intercalation process is assumed to be adherent to a rigid bond model whereby the intercalated ion (or proton) remains charged within the host oxide with the intercalated electrons occupying the conduction band [107], since the intercalation process does not result in structural change of the host material and may be described for both proton and ion insertion. Under proton insertion the process may be described by the following equation [8]: ⇄
(33)
⇄
Where the intercalation process involves the intercalation of alkali metal ions (i.e. lithium in this case), the reaction is most often described by the following equation: ⇄
⇄
34
(34)
Nickel Oxide Nickel oxide, often denoted as ‘NiO’ (or nickel (II) oxide / nickel monoxide) is a bulk material of bright green colouration. It is a 3d transition metal oxide possessing a NaCl structure in the crystalline form (figure 2.2) [132], [133], with physico‐chemical properties as per table 3.3 below. Studies report that thin films produced by electron beam evaporation can also display this fcc structure [134]–[136]. Despite its partially filled 3d orbital, it is an anti‐ferromagnetic insulator [137]. The monoxide NiO represents its most stable and well characterised stoichiometric state [132]. Other stoichiometries are often purported in the literature, most commonly nickel (III) oxide (Ni2O3) which is of grey‐to‐black colouration. These other forms are not well characterised and as Greenwood and Earnshaw state, may simply represent slight non‐ stoichiometry [132]. When deposited in thin film form, the films produced are mostly non‐stoichiometric [1], [8], [65]. For this reason nickel oxide may sometimes be written as NiOx. NiO has generated much interest and received a great deal of investigation as an electrode material, with potential for use in both batteries [138]–[147] and electrochromic devices [1], [8], [148], [149]. Amongst other methods, films have been successfully deposited by sputtering [115], [149]–[151], thermal evaporation [152], electron beam evaporation [135], [136], [148], pulsed laser deposition (PLD) [138], [153], chemical vapour deposition (CVD) [154]–[156], electro‐deposition [157]–[159] and Sol‐Gel [160]–[162] techniques. Table 3.3 – Properties for NiO as reported in or calculated from data provided by [108] 12‐84 FORMULA MOLECULAR AVERAGE WEIGHT ATOMIC WEIGHT NiO 74.69 37.35
LATTICE DENSITY PARAMETER (G/CM3) (Å) 4.1684 6.6
MELTING POINT (K) 2230
Unlike the aforementioned tungsten oxide, nickel oxide undergoes anodic colouration from colourless to dark brown. Even with the investigation and applications thus far, 35
electrochromism in NiO films is not as well understood as in other films [8], [149], [163], [164], therefore making it an ideal candidate for further study. For electrochromism to take place, the films have been shown to require conversion to nickel hydroxide [134] (which may be due to the density of NiO [8]), and the effectiveness of these deposited films as electrochromic devices is highly dependent on the structures, density and porosity of the deposited films [8], [135]. Colouration of NiO is most often performed in KOH [65], [133], [135], [136], [148]–[150], [162] and is described by the generally ‐ accepted Bode reaction scheme [8], [65], [135], [136], [147], [148]: ̵
̵
̵
⇄
̵
(35)
⇄
Like WO3, the intercalation process is assumed to be adherent to a rigid bond model whereby the intercalated ion (or proton) remains charged within the host oxide without structural change of the host material.
Titanium Oxide
Titanium oxide may exist in various stoichiometries (e.g. TiO2, Ti2O3, Ti3O5, and Ti4O7). The most commonly used and recognised oxide of titanium is titanium dioxide, also referred to as ‘titania’ or TiO2, often obtained and used as a bright white compound. In fact TiO2 is colourless, as can be observed in large crystals of the material, however its high refractive index in the visible region results in intense scattering and apparent white colouration when finely divided [132]. It can assume different crystal structures as normal pressure, namely; anatase, rutile and brookite [8] (although at least 8 other polymorphs are known to exist [165]–[172]). Its physico‐chemical properties may be found in table 3.4. Less is known of the electrochromic behaviour of TiO2 than WO3 other than its chemical resistance and thus it was identified as an ideal candidate for further investigation in this study.
36
Besides electrochromics, titanium oxide attracts widespread research and use in many areas including; pigments, UV absorbers, dielectric mirrors [173], [174], biochemical processing [175], self‐cleaning windows [176]–[178], semiconductor circuits [179], [180], data storage [181], [182] , and also in its most publicised role as a photocatalyst for photovoltaics [183] and other applications [184].
Table 3.4 – Properties for TiO2 as reported in or calculated from data provided by the references given. FORMULA
TiO2
MOLECULAR WEIGHT 79.8658 [108]
AVERAGE ATOMIC WEIGHT 26.6219 [108]
LATTICE PARAMETER (Å) 4.593 (Rutile) [185] 3.784 (Anatase) [185]
DENSITY (G/CM3)
MELTING POINT (K)
4.250 (Rutile) [185] 3.894 (Anatase) [185]
2103 [111] 2143 [186]
In terms of electrochromic activity, titanium oxide is a cathodically colouring material like tungsten trioxide, changing from transparent to a grey‐blue colouration [1], [8]. The efficiency of this transmission change is often (but not always) low however [187], [188], leading to greater potential of titanium oxide as a counter electrode to an anodic material with stronger colouration characteristics, i.e. nickel oxide [1], [8], [189]. Although titanium oxide is well known as an electrochromic material, relatively few studies exist (at least in comparison with other notable electrochromic oxides, i.e. tungsten or nickel oxides) which focus upon the electrochromic properties of the pure material. Nevertheless examples of electrochromic films have been successfully deposited by; sputtering [187], [190]–[192], thermal evaporation [193], electron beam evaporation [194], pulsed laser deposition (PLD) [195], spray pyrolysis [196], chemical vapour deposition (CVD) [197]–[199], electro‐deposition [200], anodisation [201], ‘doctor blade’ [188], and Sol‐Gel [202]–[204] techniques. Again, as with WO3 and NiO, the intercalation process conforms to a rigid bond model since the intercalation process does not result in structural change of the host material.
37
The intercalation process for TiO2 is similar to that of WO3 and may once again be described for both proton and ion insertion. Under proton insertion the process may be described by the following equation [8]: ⇄
(36)
⇄
Where the intercalation process involves the intercalation of alkali metal ions (i.e. lithium in this case), the reaction is most often described by the following equation: ⇄
⇄
(37)
Prussian Blue (iron hexacyanoferrate)
Iron hexacyanoferrate is an insoluble dark blue crystalline substance commonly used as a dye and pigment, which in‐part owes its name to its use in Prussian military uniforms of the time [205]. Until relatively recently, its exact chemical composition was not known however through the use of multiple characterisation techniques it is now known to be an inorganic polymer possessing a cubic lattice structure comprising of Fe2+ and Fe3+ ions inter‐connected by CN‐ ligands in a Fe(II)‐C‐N‐Fe(III) sequence. Depending upon the preparation method used, PB may appear to be insoluble or soluble. In its ‘insoluble’ form, every fourth Fe(CN)6 position is vacant, and may be habited by water molecules [206] (figure 3.2). While in its ‘soluble’ form, these vacancies are non‐existent with co‐ordinated metal ions occupying positions within the lattice [207], [208] (figure 3.3). Due to its well established an researched nature, it was investigated as a material for comparison to those oxides also included within this study. The physico‐chemical properties of PB are presented in table 3.5 opposite.
38
Figure 3.2 – Representation of the crystal structure of ‘insoluble’ PB
Table 3.5 – Properties for PB as reported in or calculated from data provided by the references provided. FORMULA
MOLECULAR AVERAGE WEIGHT ATOMIC WEIGHT Fe4[Fe(CN)6]3 859.23 [108] 20.95 [108]
LATTICE DENSITY PARAMETER (G/CM3) (Å) 10.28 [55] ‐
Figure 3.3 – Representation of the crystal structure of ‘soluble’ PB
39
MELTING POINT (K) Decomp.
As an electrochromic material, PB has attracted and continues to attract a great deal of investigation [1], [8], [63], [99]–[101], [209]–[214]. As well as this, and its historically long‐lived use as a traditional dye, it also experiences widespread use as a stain in histopathology [215], [216] and also as a medicine for use in the treatment and removal of radioactive metal ions (such as thallium and caesium) from the body [217]– [219]. The colouration of PB is a result of intervalence charge transfer between the Fe ions across the CN‐ ligand. Where the Fe ion is adjacent to the electrophilic nitrogen, it is said to be ‘high‐spin’, and where it is next to the carbon, it is said to be ‘low spin’ [207]. This transition may be represented as follows: ⇄
(38)
For PB in its un‐modified state, this transfer causes absorption in the red area of the visible region, which corresponds to the strong blue colouration with which PB is associated. When reduced or oxidized electrochemically, electrochromism occurs as a result of modification and switching of this charge transfer. PB may possess three distinct oxidation states each with their own respective colouration; Prussian white (PW, also known as Everitt’s salt), Prussian blue (PB), and Prussian yellow (PY). A fourth colouration, Prussian green (PG, also known as Berlin green) is possible as a result of mixed valence states within the material. These modifications may be represented as below, where the numbers represent the oxidation states of the first and second Fe ions respectively: 2 ,2
⇄ 3 , 2
⇄
⇄
⇄ 1
3 ,2
⇄ ⇄
40
3 ,3
⇄ 3 , 3
⇄ ⇄
(39)
3.2 – Transparent Conductive Oxides Transparent conductive oxides (TCOs) are semiconductor materials which are electronically conductive while being largely transparent within the visible region. Typically these are doped oxides which possess a bandgap ≥3 eV, which renders them largely transparent within the visible region, while they are typically good reflectors in the near‐IR region due to their electrical conductivity. After the discovery of the first TCO material, cadmium oxide (CdO) in 1907 [220], many more TCOs were identified including the oxides of indium (In2O3), tin (Sn2O3), and zinc (ZnO) [221]. Following these discoveries it was found that doping of these host TCO materials could result in increased electrical conductivity whilst retaining optical transparency. Perhaps the most well‐known and utilised of the TCO family is indium tin oxide (ITO, In2O3:Sn) [164], [222]–[225]. Amongst other uses, ITO finds widespread use in touchscreens, liquid crystal displays (LCDs) [226], heated windows , anti‐static coatings [227] and solar cells [164]. In the case of electrochromics, ITO is commonly used to form the transparent conductors required to supply electrical charge to the device as depicted in chapter 1 figure 1.1 [1], [8], [164]. Due to its availability and established suitability, ITO was chosen as the as a transparent electrode material to be produced and investigated throughout this work. Indium oxide, In2O3, the main constituent of ITO, is a yellowish‐green bulk material which when produced under vacuum conditions without high temperature possesses a cubic bixbyite structure [228], [229]. In this form oxygen vacancies are present within a fluorite‐type superstructure where every fourth oxygen anion along the (111) axes is absent, resulting in distortions within the structure [229], [230]. Its physico‐chemical properties can be found in table 3.6 below. Table 3.6 – Properties for In2O3 as reported in or calculated from data provided by the references given. FORMULA MOLECULAR AVERAGE WEIGHT ATOMIC WEIGHT In2O3 277.64 [231] 55.638 [231]
41
LATTICE DENSITY PARAMETER (G/CM3) (Å) 10.117 [231] 7.120 [231]
MELTING POINT (K) 2185 [108]
Figure 3.4 – Illustration of non‐equivalent indium sites and oxygen vacancies in an In2O3 crystal. ITO is formed by doping indium oxide with tin atoms which substitute In within the structure, preferentially occupying the less distorted substitutional sites within the material [230], [232]. ITO is well known for its; good conductivity, high transparency (in the visible region), excellent adhesion and good chemical resistance. Due to these well‐ established qualities and its availability, ITO was chosen as the most suitable electrode material for the substrates and devices produced throughout this thesis. Like other metal oxide materials such as those electrochromic materials discussed herein, ITO may be produced by various methods as described in 3.3.1 including; thermal evaporation [233], electron beam evaporation [234], DC sputtering [226], [227], [235], RF Sputtering [235], chemical vapour deposition (CVD) [236], and sol‐gel [237] techniques. The optical and electrical characteristics of ITO films are dependent upon charge carrier density and electron mobility, which are both factors highly dependent upon the deposition conditions. By increasing the charge carrier density, electrical conductivity is increased and this increase is proportional to both the number of oxygen vacancies and substitute tin atoms present within the material. In an ideal scenario, each oxygen
42
vacancy contributes two electrons to a doping tin atom, this results in degenerate semiconductor behaviour and widening of the material’s band gap (also known as the ‘Burstein‐Moss effect’) [49], [164] and subsequently an increased electrical conductivity.
3.3 – Thin Film Deposition The term ‘thin film’ may refer to any coating which ranges from fractions of a nanometre to a few microns in thickness. Typically it is used to refer to any film with a thickness ≤ 1m and ≥ 1nm, and can be formed using a variety of techniques. All of the techniques comprise of methods which are able to deposit thin films of different materials either atom‐by‐atom or molecule‐by‐molecule onto a bulk material. The bulk material onto which the thin film layer is coated is known as the ‘substrate’, and any material which is coated onto the substrate will possess its own properties, optically and mechanically. They may be optically absorbing, transmissive, reflective, electrically conductive, insulating and/or semiconducting, and will possess mechanical properties such as; hardness, stress and density independent of the substrate material. Dependent upon the material, deposition conditions and post‐deposition treatment, the films may be either amorphous or crystalline in basic structure. In an amorphous film there is no order in the overall structure, whereas polycrystalline films possess crystal type structures and therefore determinable orientation of the molecules.
3.3.1 – Overview of Techniques
Thus far there are many methods employed for the deposition of electrochromic thin films. Those which have attracted widespread use and attention for electrochromic materials are as follows:
Physical Vapour Deposition (PVD) 43
o
Evaporation
Thermal [78], [80], [116]–[118], [152], [193], [238]
Cathodic Arc Deposition [239]
Electron beam (EB) [113], [119], [120], [135], [136], [148], [163], [194], [240], [241]
o
Pulsed Laser Deposition (PLD) [121], [138], [153], [242]–[244]
Sputtering
DC Sputtering [12], [81], [113], [133], [149], [151], [192], [245]–[249]
RF Sputtering [150], [153], [187], [190], [191], [248], [250]– [252]
Chemical Vapour Deposition (CVD) [123]–[125], [154], [156], [197]–[199], [253], [254]
Sol‐Gel [130], [131], [161], [162], [202]–[204], [255]
Spray Pyrolysis [122], [196], [256]–[258]
Thermal Oxidation [259]–[262]
Electro‐deposition / polymerisation [99], [126]–[128], [158], [200], [263]– [265]
Anodisation [159], [201], [266], [267]
Physical vapour deposition or ‘PVD’ is a group of processes whereby free atoms or molecules of a target material are released into a vacuum by physical bombardment or evaporation before condensing on the desired substrate to form the deposited film. Prior to deposition, these free atoms or molecules may react with other atoms or molecules present in the vacuum which are usually introduced as a reactive process gas, e.g. oxygen, to form the desired product. Where the target material is reacted with a process gas or other material it is known as a reactive deposition. In some setups, one may deposit multiple materials simultaneously, resulting in a mix of materials as desired or ‘co‐deposition’. Chemical vapour deposition or ‘CVD’ is another group of deposition processes which result in the formation of thin films upon the desired substrate. Often, but not always
44
this is once again performed under vacuum using gaseous chemical precursors or a vaporised form of the target material introduced into the chamber. Here they react to form a thin film, often as a result of thermal means; on the surface of the substrate, with the surface of the substrate, or before condensing upon the substrate. Sol‐gel is a solution based chemical technique for the formation of thin or even thick films. Here a precursor‐containing solution, typically a transition metal alkoxide, peroxometallate species [1], or salt [1], [162] is subjected to hydrolysis and / or polycondensation reactions to form a colloidal suspension, known as the ‘sol’. The sol may then be coated onto the desired substrate by techniques such as dip coating, spin coating or casting. As the solvent is removed from the sol an integrated network or ‘gel’ develops within the material, hence giving the process its name ‘sol‐gel’. The films are then most often heated or annealed to drive off any remaining solvent and / or provide the film with the desired crystalline properties. Spray pyrolysis is a method closely related to CVD, which commonly refers to an atmospheric process whereby a precursor for the target material is sprayed, usually by means of aerosol or reactive carrier gas, onto a heated substrate. The heat of the substrate surface causes the precursor to undergo thermochemical decomposition or ‘pyrolysis’ upon contact, thereby forming a solid film consisting of the decomposition products on the surface of the substrate. Thermal oxidation is a surface modification process where a substrate is heated in the presence of oxidising agent (oxygen, water, or often, atmospheric air [261]) which in‐ turn results in the formation of an oxidised species on the surface of the substrate. For electrochromics, this can be used to form metal oxide films on the surface of metallic substrates [260], or from precursor films deposited by other means [259], [261]. Electrodeposition is a solution‐based process whereby an electrical potential is applied to an electrode which is in contact with a coating solution. Typically this coating solution consists of a colloidal suspension or mix of chemical precursors containing charged species, often with a supporting electrolyte. Upon application of the electrical potential, the charged species will migrate through the solution where they will be deposited upon contact with the electrode and subsequently experience charge neutralisation. Electropolymerisation may be considered a subcategory of
45
electrodeposition, where a polymerisation reaction occurs at the electrode surface as a result of the applied potential, thereby resulting in a film as desired. The main requirement for, and subsequent limitation of electrodeposition is that the substrate to‐be‐coated must be conductive, so‐as‐to transport charge to or from the solution as required. Anodisation is another solution‐based process, which although similar in setup to that of electrodeposition, is a surface modification process rather than one of deposition. The anodisation process is commonly used in the passivation and aesthetic colouring of aluminium and its alloys. Here the substrate to be anodised forms the anode, in contact with an appropriate electrolyte, often an acid, with a suitable cathode completing the cell. When an electrical potential is applied across the cell, oxygen is formed at the surface of the electrode and causes oxidation. At the same time the low pH of the acidic electrolyte etches the newly formed oxide layer resulting in nanostructures. The exact morphology or form of the structures may vary depending upon the conditions or method employed. For example, anodisation may form many kinds of different nanostructures including nanopores [159] and nanotubes [129], [201], [266], [267].
3.3.2 – Techniques of Interest Of those techniques described in 3.3.1, PVD and CVD techniques have established themselves as widespread industrial and scientific methods of depositing thin films. Typically these techniques are performed under vacuum, and any process which does this is known as a vacuum deposition process. Advantages of deposition thin films under vacuum include (but are not limited to):
Increased free collision path for target material due to low particle density (the air in normal atmospheric conditions is very particle‐dense in comparison)
Ability to closely control deposition conditions such as gas or material composition, deposition rates and reduce potential contamination
Ability to generate plasma (where required)
46
In the course of this work, multiple deposition methods for thin films have been employed, namely; electron beam evaporation, DC magnetron sputtering, magnetron‐ assisted pulsed DC sputtering, RF sputtering and electrodeposition. With the exception of electrodeposition, all of these techniques are vacuum deposition processes.
Sputtering Sputtering is a PVD vacuum deposition technique which makes use of solid target materials and works by bombarding the target material with energetic particles which in turn eject atoms or particles of the target material toward the substrate. The atoms or particles then condense on the surface to form a film. Different methods of sputtering have been developed, and predominantly rely upon the application of DC (Direct Current) or RF (Radio Frequency) potentials to a target material. Deposition systems may employ one or more of the above techniques in order to deposit the target material. DC sputtering works by applying a large potential difference between the target (cathode) and the deposition chamber (anode). An inert charge carrier gas (usually Ar) is then ionised and forms a plasma. These ions then accelerate towards the negatively charged target and bombard it, in turn ejecting atoms or molecules of the target material as described above. Secondary electrons from the target are accelerated through the cathode potential and are responsible for further ionising the inert gas and sustaining the plasma. DC sputtering may only be used to deposit from conductive targets. If an insulating target is used, positive charges will very quickly build up at the surface of the target, and deposition will stop [268].
47
Figure 3.5 – Schematic Diagram of a DC sputtering System RF or ‘radio frequency’ sputtering however, can be employed to deposit from both conductive and insulating targets. Essentially the setup is the same, however instead of the application of a DC bias, a high frequency AC potential is applied between the target and chamber. As before, the potential difference will result in ionisation of inert gas molecules and the creation of a plasma. This time however, ions will be accelerated toward the target material even when the material is an insulator. Since electron mobility is far greater than that of the larger ionised species, the surface of the target material acquires a negative charge in relation to the conducting plasma. The side effects of this technique include greater heating of the target material and a slower deposition rate in comparison to DC sputtering.
48
Figure 3.6 – Schematic Diagram of a RF sputtering System Due to low ion densities which result in very low deposition rates, very few systems or processes will employ pure DC (or ‘diode sputtering’) or RF sputtering. To increase the rate and efficiency of deposition and reduce substrate heating effects experienced with pure DC and RF sputtering techniques magnetron is now universally employed [269]. A magnetron is essentially an assembly of magnets placed behind the target to create a magnetic field at the surface of the target, trapping electrons emitted from the target and forcing them to assume a spiral trajectory parallel to the target surface. This force experienced by a charged particle moving in the presence of a magnetic field can be given by the Lorentz force law:
(40)
This results in significantly increased collisions between atoms of inert working gas (due to a longer trajectory and increased probability of collision) and therefore results in greater ionisation and subsequently increased target bombardment and associated sputtering rate. The terms used to describe these techniques then become DC or RF
49
magnetron sputtering respectively. Sputtering techniques may then be additionally enhanced by the inclusion of extra methods with which to increase free electron density and further excite the ions present within a generated plasma, such as ICP (Inductively Coupled Plasma) and MP (Microwave Plasma) sources (as fitted to the MicroDyn deposition system (4.1.4).
Figure 3.7 – Schematic diagram of a sputtering target with magnetron.
Electron Beam Evaporation Electron beam evaporation, often simply referred to as electron‐beam or e‐beam, is another PVD vacuum deposition technique. Electron beam deposition sources operate by bombarding a target material which is located at the anode. A charged tungsten filament (heated to prevent contamination) produces a beam of electrons which are accelerated through potentials of up to 10kV. The ingenious feature of electron beam guns is that the beam is generated below the material to be evaporated and is deflected through 270° by a magnetic field (figure 3.8). When the electron beam
50
contacts the target material it has enough energy to evaporate even refractory materials. The evaporant may then react with any reactive gasses (where applicable), before condensing upon contact with the substrate, and indeed the other exposed surfaces of the vacuum chamber, to form a thin film.
Figure 3.8 – Schematic representation of an electron beam evaporation source. Since the beam of electrons which is generated is often very narrow, it is usually ‘swept’ across the area containing the target material using a sweep generator. Sweep generators operate by varying the magnetic field directing the electron beam. With two perpendicular solenoids the beam may then be swept over the entire surface area of the target material. 51
Electrodeposition
Electrodeposition or ‘electrolytic deposition’, also known as ‘electrophoretic deposition’ (EPD), is an atmospheric wet‐chemistry process where electrical potential is applied to the substrate which in turn is placed within an electrochemical cell. As before (2.2), this applied potential provides an EMF which results in positively charged ions or molecules migrating toward the cathode and negatively charged ions or molecules migrating toward the anode due to electrophoresis [51], [270]. When these charged ions or molecules meet with the surface of the electrode to which they are attracted, they undergo a charge balancing redox reaction and are subsequently deposited. Over time more and more ions or molecules will be deposited to eventually form a thin film upon the surface of the electrode. At simple example of electrodeposition is that of deposition of Cu from a solution of CuSO4 with an inert (e.g. Pt) counter electrode. Here, positively charged Cu2+ cations will migrate to the cathode and undergo the following charge balancing reduction reaction:
Cu
2e → Cu
(41)
To balance this, an equal and opposite reaction must occur at the anode, where water molecules undergo oxidation to liberate O2 gas and protons which charge balance the sulphate anions and form sulphuric acid:
2H O → 4H
O
4e
(42)
4H
2SO
⇄ H SO
52
(43)
Figure 3.9 – Illustration of an electrolytic cell for the cathodic deposition of Cu from CuSO4. The substrate or work piece to be coated must be conductive in this case.
3.3.3 – Film Thickness Control Thickness is a critical factor affecting the properties of thin films. Thickness can have an influence on the optical [222], [271], mechanical [272] and structural [241], [273] properties of a thin film since thin films of materials often possess properties which differ from that of the bulk material. With respect to electrochromic materials, thickness has been shown to strongly influence electrochromic performance and longevity [119], [136], [273]. TCO materials, such ITO as used in the course of this work, possess optical and electronic properties which experience well‐documented variation with film thickness [222]. Therefore methods enabling the measurement of film thickness are essential.
During Deposition
During deposition, films produced in the course of this work have been controlled and / or monitored by the following means: 53
Time – Deposition time can be used as an effective method of controlling the deposited thickness of a thin film when the deposition rate is well characterised. This method is best suited to the sputtering of thin films, since the process is relatively stable and rate is proportional to sputtering power. By depositing a film under controlled conditions for a set time, and measuring the deposited thickness using a post‐deposition technique, the rate can be calculated. Using this rate one may extrapolate the time required to deposit a film of the desired thickness. This method of deposited thickness control has been employed with the; MicroDyn (4.1.4), CVC (4.1.5) and PlasmaCoat (4.1.6) sputtering systems.
Quartz Crystal Monitor (QCM) – A QCM works by measuring the mass deposited over its area. The ‘crystals’ themselves are transducer arrangements with metal electrodes deposited upon each side of a thin piezoelectric quartz crystal, with commercial quartz crystals typically possessing a resonant frequency in the 4‐6 MHz range. When material is deposited upon the crystal, the frequency at which it oscillates is reduced proportionally to the deposited mass. Through continual measurement of the resonant frequency (down to 1Hz resolution), an extremely accurate calculation of the deposited mass, and therefore thickness can be made. This calculation however, relies upon a deposited material being well characterised and thus is often effected by rate , pressure and other factors effecting the deposition.
Optical thickness monitoring – Optical thickness monitors, like the Scalar Technologies ScalarGauge (see 4.5.2), are typically fast photo‐diode array type spectrometer systems which continually measure the transmittance spectra of a transmissive substrate over a wavelength range throughout the course of the deposition. Through optical fitting, or pre‐determined optical modelling of the expected spectra, the deposition may be stopped when a desired spectra (and thickness) is achieved. For well‐characterised processes, it may be possible to monitor at just two or three, or even a single wavelength, however more advanced systems such as the aforementioned ScalarGauge are capable of taking hundreds of spectra per second, over a very large wavelength range, with a high resolution. 54
Post Deposition After deposition of the desired films, the following post‐deposition methods of thin film thickness calculation have been employed to confirm actual deposited thicknesses and also to calibrate the pre‐mentioned methods of controlling thickness during deposition.
SEM – SEM micrographs of a thin film cross‐section can be used to directly measure the thickness from the (known) image magnification.
Optical Characterisation – Mathematical fitting of transmission and / or reflection spectra of the deposited films may be employed to accurately calculate both the optical constants ( and , where required) and / or thickness of the deposited films.
Stylus Profilometer – A stylus profilometer works by running a diamond‐tipped stylus along the surface of the sample while measuring its vertical displacement. Since this requires a change in profile, it is necessary that an area of the substrate being coated is masked such that the step between the substrate and deposited film can be measured.
3.4 – Physical Characterisation
3.4.1 – Scanning Electron Microscopy
A commonly employed and very useful characterisation technique for the investigation of both the surface and structural properties of materials and thin films is scanning electron microscopy ‘SEM’. Figure 3.10 provides a schematic representation of how SEM works. Briefly, SEM works through the generation, focusing, and scanning of an electron beam across the surface of the sample in question. When the electrons come into contact with the sample the resultant interaction causes; loss of electron energy through scattering and absorption within the sample, reflection of electrons at the surface via elastic scattering; emission of secondary electrons from within the sample 55
via inelastic scattering and also emission of electromagnetic radiation from the sample, in particular characteristic x‐rays, allowing chemical analysis of the imaged area. The penetration of the beam is dependent upon the potential applied to the cathode, typically a very sharp single crystal tungsten tip (in the case of FESEM), which is only a few nm in diameter and generates the electrons. A positively charged anode collects these electrons and translates this signal into an image on a CRT monitor which has a scan rate equal to that of the electron beam in contact with the sample, effectively creating a ‘map’ or representative image of the sample which is produced from the signal intensity detected at the anode [274], [275].
Figure 3.10 – A schematic representation of how SEM works.
56
3.4.2 – Energy Dispersive X‐Ray Energy dispersive x‐ray analysis ‘EDX’, also known as ‘EDAX’ and ‘EDXS’, is a spectroscopy technique capable of identifying chemical compositions of materials. When a sample is contacted by an electron beam, inner‐shell electrons of various atoms within the material become excited, and if given enough energy are ejected from the sample. To compensate and ‘fill the hole’ left by the ejected electrons, higher energy electrons from outer shells will move into the inner shells and the energy difference will be emitted as an x‐ray which may be detected. Since each element possesses a unique atomic structure, the energy of the emitted x‐ rays will be unique to that element. Therefore by measuring the different energies and numbers of x‐rays detected the data may be used to identify the elements present within a sample being tested and estimate the stoichiometry of the different elements within the sample. The obvious limitation of the technique however is that it relies upon an element having both an inner and outer shell, therefore both H and He are not detectable since they possess only one shell. Also, depending upon the detector in question, some small elements such as Li may also be undetectable.
3.4.3 – X‐Ray Diffraction (XRD) X‐ray diffraction is a technique that utilises the wave nature of x‐rays to provide information about the atomic and molecular structure of a crystalline material. It operates by focusing a beam of incident x‐ray radiation toward the material. When bombarded, the majority of energy will pass through the material, and a quantity will be absorbed. The remainder will induce oscillation of electrons contained within the atomic structure of the crystal, since the wavelength (λ) of the x‐rays is within the same order of magnitude of the inter‐planar spacing (d) of the crystal units (figure 3.11). This oscillation results in elastic scattering, where the electrons (known as scatterers) re‐ emit radiation in a scattered fashion.
57
These scattered emissions interfere with one another destructively, except in some directions where they combine constructively as determined by Bragg’s law of diffraction:
2 sin
(44)
Where; λ is the wavelength of incident radiation, n is the order of the diffraction, d is the inter‐planar spacing, and is the scattering angle [276], [277].
Figure 3.11 – Illustration of constructive interference x‐ray diffraction or ‘Bragg diffraction’, as represented by Bragg’s Law.
58
3.5 – Electrochemical Characterisation Electrochemical characterisation is the characterisation of materials or devices through the use of various electrochemical techniques and is an invaluable tool in the characterisation of electrochromic materials and devices. Amongst other electrochemical
methods;
cyclic
voltammetry,
chrono
amperometry
and
electrochemical impedance spectroscopy are the most widely used and are employed throughout this thesis. All of the aforementioned methods require the use of an electrochemical cell. In its most basic form an electrochemical cell must comprise of at least a working electrode (WE) and a counter electrode (CE), immersed in or separated by an electrolyte. Such a simple cell is often referred to as a two‐electrode cell. However, unless the properties of the cell as a whole are being investigated, a two electrode cell is rarely employed since in order for the system to provide useful information about the WE, the CE must be capable of maintaining a constant redox potential (with which to compare the potential of the WE) while passing current.
Figure 3.12 – Schematic representation of a three electrode cell with a typical Ag/AgCl reference electrode. 59
To avoid any fluctuations in redox potential while passing current, a three electrode cell is most often employed (figure 3.12), with the addition of a reference electrode (RE) alongside the WE and CE. The RE consists of a suitably characterised electrochemical ‘half‐cell’ (such as Ag/AgCl or calomel 34) which is most usually connected to the main cell by means of a salt bridge or glass frit. The connection to the RE presents a very large resistance to the system so that the RE passes virtually no current, and therefore is capable of maintaining a stable reference potential. In a three electrode system, it is common for the CE to be manufactured from a non‐reactive material (such as platinum), with a surface area greater than that of the WE to avoid current limitation issues.
3.5.1 – Cyclic Voltammetry Cyclic voltammetry (CV) is a common electrochemical technique employed in the characterisation of electrochemical processes. It is a modification of the more simple linear sweep voltammetry (LSV), and involves applying a linearly sweeping potential (triangular waveform) to a cell back‐and‐forth between two points at a constant scan rate. During this process the current passing through the WE is recorded. The resulting ‘cyclic voltammograms’ may present useful data on the electrochemical processes occurring at the WE and reveal the potential ranges where specific reactions occur, including, but not limited to redox potentials and the kinetic and thermodynamic properties of a material or cell [278], and may be used to reveal even more, or with improved accuracy, when combined with other techniques such as spectroscopy [59], [279] and mass measurement [279], [280]. It is particularly useful for studying reversible redox behaviour such as that which electrochromic materials and devices exhibit. Figure 3.13 provides an illustration of a CV waveform.
60
Figure 3.13 – Potential versus time during CV analysis
3.5.2 – Chrono Amperometry
Chrono amperometry (CA) is a square‐wave voltammetry (SWV) technique which involves the recording of current passing through the WE over time while applying a square‐wave potential. In initial investigation it often may be used to investigate the potential ranges of processes through separation of capacitive currents and those faradaic currents associated with redox behaviour [278]. In the study of electrochromic materials and devices it is employed to switch the material or device between different redox potentials and consequently, colour states. This method may in‐turn also provide data on the electrochromic properties including, but not limited to the; response time, charge capacity and reversibility of function. Through repeated cycling, the longevity of the device is also able to be investigated. Figure 3.14 provides an illustration of a CA waveform.
Figure 3.14 – Potential versus time during CA analysis 61
Chapter 4 – Experimental
4.1 – Thin Film Deposition
4.1.1 ‐ Substrates To ensure the quality and adhesion of deposited films and avoid unnecessary blemishes or pinholes, it is important to ensure that substrates are suitably prepared and cleaned prior to use. In this work, all substrates have been cleaned thoroughly before the deposition of films using the Optimal UCS40 commercial ultrasonic cleaning system (4.1.2). The substrates employed in the course of this work are as follows:
Glass Microslide Substrates Glass microscope slides (76 x 26mm, 1.0mm thickness) as supplied by Agar Scientific were used for the majority of depositions in the course of this work. Large microslide substrates (110 x 76mm, 1.2mm thickness) as supplied by Logitech, were employed for the construction of large area devices.
ITO Coated Glass Microslide Substrates ITO coated glass microslide substrates were produced using plain glass microslide substrates as detailed above using the MicroDyn commercial sputtering system (4.1.4). Active coated area was 70 x 20mm, achieved through the use of custom substrate‐ holders (4.1.8) with the appropriate area masks (4.1.7).
62
Fused Silica Discs Fused silica disk substrates (20mm diameter, 2.0mm and 1.0mm thicknesses) were cleaned and used for optical measurement and fitting work as required.
Curved Polymer Substrates Curved Polymer substrates of an undisclosed laminated construction comprised primarily of acetate material with a SiO2 type hard coating (80mm width, 50mm height and 1.6mm thickness) were cleaned and used for deposition of ITO as presented in 5.6.6.
4.1.2 – Optimal UCS40 Commercial Ultrasonic Cleaning System
Figure 4.1 – The Optimal UCS40 commercial ultrasonic cleaning system
63
Substrates used during the course of this work were cleaned using an Optimal UCS 40 (Figure 4.1). In this system substrates are cleaned in a four stage automated process: 1. 40 kHz Ultrasonic cleaning in a heated 40C bath containing a potassium hydroxide based cleaning solution. 2. 40 kHz Ultrasonic cleaning in a heated 40C bath containing dilute industrial washing liquid. 3. DI (de‐ionised) water rinse in a heated 55C bath. 4. Final DI water rinse followed by a capillary dry in a heated 55C bath. The total cycle from beginning to end takes approximately 15 minutes. Tests upon ITO coated substrates failed to detect any notable differences in substrates cleaned using this system.
4.1.3 – Satis MSLab 370 Electron Beam Evaporation System
Figure 4.2 – The Satis MSLab 370 Electron Beam Evaporation System
64
The majority of electrochromic transition metal oxides which have been prepared in the course of this work have been deposited using a Satis MSLab 370 series electron beam evaporation system equipped with two sources and a sweep generator (4.2 and c). The substrates are held in custom substrate‐holders (see 4.1.8) on a large rotating calotte (20 rpm) with the appropriate area masks (see 4.1.7). In the centre, a stationary quartz crystal monitor (QCM) is employed to monitor the deposited mass, allowing for rate and thickness control throughout the deposition, control of which is provided by the system software. Process gasses are introduced at the base of the chamber and flows are computer controlled using mass flow controllers (MFCs). Typical base pressure achieved with this system before process initiation would be 8.0 x 10‐7 Torr.
Figure 4.3 – Annotated vacuum chamber interior of the Satis MSLab 370 series electron beam evaporation system with metal deposition shielding removed. For all depositions the right hand source was used which is fitted with a six pocket copper hearth and 7cc crucibles. Source‐to‐substrate distance averages at approximately 650mm at line of sight.
65
4.1.4 – MicroDyn 40000 Series Microwave‐Assisted Commercial Pulsed DC Magnetron Sputtering System The MicroDyn 40000 series is a microwave‐assisted commercial pulsed DC magnetron sputtering deposition system produced and patented by Deposition Sciences Ltd (figure 4.4). This system has been used throughout this work to produce ITO coated substrates for electrochromic devices on a commercial scale. It makes use of large planar sputtering targets (15” X 5”), and employs a microwave plasma source to enhance the deposition rate. Typically a base pressure of 1.0 x 10‐6 mBar would be established before processes were initiated.
Figure 4.4 – The MicroDyn 40000 Series Microwave‐Assisted Commercial Pulsed DC Magnetron Sputtering System with custom substrate tooling (4.1.8) fitted to rotary drum.
66
Substrates are attached to a circular or hexadecagon drum which is placed within the chamber and rotated to allow for a large depositable area and greater uniformity. As part of this work, substrate holder tooling (4.1.8 ‐ as shown above Figure 4.4) and substrate masks have been designed and produced for the deposition of the required films. The targets are placed behind a shadow masking arrangement to enhance uniformity across the drum. Thickness control is available using two QCMs, which ‘see’ the target through regularly positioned holes across the drum. In the system setup used, two of the possible four target positions are fitted with blanking plates (figure 4.5).
Figure 4.5 – Schematic representation of the MicroDyn 40000 series.
67
4.1.5 – CVC AST 304 The CVC AST 304 (figure 4.6), is a bell jar style RF sputtering system employed in this thesis in the deposition of lithium phosphorus oxy‐nitride (LiPON) from a lithium phosphate (LiPO4) target. The system features a water‐cooled metal bell jar arrangement with two 8” targets mounted at the top of the chamber, rotating substrate shutter and planetary table arrangement with six substrate tables. Both the individual substrate tables and entire assembly of six can be made to rotate individually or together, with variable speed. Also at the top of the chamber exists a third unoccupied target position, a heater, and an ion source. Process gasses are introduced at the top of the chamber and flows are computer controlled using mass flow controllers (MFCs), with additional control of chamber pressure being provided by a throttle arrangement above the cryo pump. Base pressure was typically in the region of 5.0 x 10‐6 ‐ 8.0 x 10‐6 Torr.
Figure 4.6 – The CVC AST 304 RF Sputtering System
68
4.1.6 – AML PlasmaCoat Plus II
Figure 4.7 – The AML PlasmaCoat Plus II
Electrochromic mixed‐metal Ni‐Cr oxides and TCO films ITO were deposited using an AML PlasmaCoat Plus II ophthalmic coating system (figure 4.7). The PlasmaCoat is a small batch high throughput plasma‐assisted DC sputtering system, equipped with two 6” DC magnetrons and a 4” hollow cathode plasma source. The substrates are attached to a rotating carousel (100rpm) which is transferred through a load‐lock to the coating chamber (figure 4.8 below). It is capable of holding six substrates. Process gasses are introduced into the chamber and flows are computer controlled using mass flow controllers (MFCs). Process gasses are introduced and the process initiated automatically when a base pressure of 5.0 x 10‐6 is established.
69
Figure 4.8 – Rotating carousel assembly [281]
The system was originally designed for the application of AR and aesthetic mirror coatings to ophthalmic lenses, and can be used to deposit upon curved and flat substrates, with a high throughput and fully automated deposition. The standard target materials with which the machine was designed to operate were silicon and zirconium. To adapt the system for use with other materials, the controlling PC was connected to an external monitor and keyboard and customised deposition programs were created.
4.1.7 – Substrate Masking For a process to be well controlled and reproducible – as is required when producing films and devices on a commercial scale – it is imperative that the deposited area be controlled. To produce the films in this work, various different slide masks have been designed for different electrochromic and electrode layers for electrochromic devices (figure 4.9). These were laser cut from non‐magnetic stainless steel (0.5mm thickness) and fit inside
70
the respective substrate holders or be attached directly to the substrates themselves using Kapton® tape.
Figure 4.9 – Photograph of different substrate area masks designed, constructed and employed within this work. The slide masks provided accurate and reproducible control of deposited and active areas of electrochromes and devices. They were found to produce a small area of ‘shadowing’ around the borders of the deposited areas. This was approximately equal to the mask thickness (0.5mm) and could be reduced if necessary via the use of thinner masks or sharpening of the masked edges, however this effect was found to be sufficiently small as to cause negligible effect upon the results of the work performed and it was decided that 0.5mm masks provided an acceptable compromise between durability, cost and function.
4.1.8 – Substrate Holder Tooling As detailed in chapter 1, this work focuses on producing electrochromic devices on commercial scale equipment. For one to assess the viability of using such equipment to produce thin film devices on a commercial scale it is not sufficient to simply deposit upon a couple of samples placed within a commercial scale chamber.
71
The behaviour and rate of thin film growth is highly dependent upon the material of the surface being coated and substrate temperature [282], [283]. The chambers of commercial scale vacuum systems are invariably constructed of metal, most often stainless steel, and since the substrates are of a different material, the coating process will behave differently depending upon the number of substrates being coated, or more specifically the ratio of metal to substrate surface area being coated. As a result of this it is necessary to perform commercial scale depositions involving many substrates for accurate assessments of viability to be made. To facilitate this, and also improve the performance and reproducibility of the materials deposited throughout this work, substrate tooling was developed and tested as described in 5.7.1.
4.2 – Optical Characterisation Optical characterisation (as described in 2.3.1) was employed in the course of this work to characterise deposited films and calculate film thickness. The majority of characterisation was performed on the Hitachi U‐3501 and Aquila Instruments nkd8000 spectrophotometers.
4.2.1 – Hitachi U‐3501 Spectrophotometer The Hitachi U‐3501 (figure 4.10), is a dual beam spectrophotometer fitted with deuterium and iodine‐tungsten lamps for UV and visible region / near IR measurements respectively. It employs a dual beam monochromator system for continuous reference acquisition. Light is detected by a twinned arrangement of a photomultiplier (for UV and visible region) and lead‐sulphide detector (for near‐IR) 14. The sample is placed in the path of one of the beams of light, with light contacting the sample at 0°.
72
Figure 4.10 – The Hitachi U‐3501 spectrophotometer
4.2.2 – Aquila Instruments nkd‐8000 Spectrophotometer The Aquila Instruments nkd‐8000 (figure 4.11), is a spectrophotometer system which simultaneously measures transmittance and reflectance of a sample at incident angles between 10 and 70 degrees. Samples, along with a silica reference, are located upon an x‐y stage which can be moved in order to change positions between the sample and reference and / or map different areas of the sample. The system employed is fitted with a high stability quartz tungsten halogen light source and features two detectors (fitted independently of each other where required) which provide an observable range of 350‐1700nm. The system is capable of controlled polarisation selection and employs the ProOptix software package to provide the capability to model the resultant spectra and solve for n, k and film thickness.
73
Figure 4.11 – The Aquila Instruments nkd‐8000 spectrophotometer
4.3 – Physical Characterisation
4.3.1 – Scanning Electron Microscopy and Energy Dispersive X‐Ray
In this work, a Hitachi S4100 FESEM (Field Emission Scanning Electron Microscope) has been employed to generate SEM micrographs for analysis and comparison of electrochromic materials. Samples were imaged at various magnifications. This system has a magnification range of 20‐30000x, with a resolution of 1.5 nm. Acceleration voltage may be adjusted up to a maximum of 30 kV. The Energy Dispersive X‐Ray (EDX) analysis performed in this work was performed using a model 6566 Oxford Instruments Gem germanium detector attached to the aforementioned Hitachi S4100 FESEM.
74
Figure 4.12 – The Hitachi S4100 FESEM fitted with a model 6566 Oxford Instruments Gem Germanium detector.
4.3.2 – X‐Ray Diffraction (XRD)
XRD analysis throughout this work has been performed using a Siemens D5000 x‐ray diffractometer (figure 4.13). The x‐ray source fitted was a Cu Kα.
Figure 4.13 – The Siemens D5000 X‐Ray Diffractometer 75
4.3.3 – Stylus Profilometry
Where applicable, film thicknesses were measured using a Dektak 3ST surface profilometer (of the model depicted in figure 4.14) using step height determination.
Figure 4.14 – The Dektak 3ST surface profilometer
4.3.4 – Sheet Resistance Measurement
Sheet resistance measurements were taken using a four point probe constructed by Jandel Engineering of Bedfordshire, UK (figure 4.15)
Figure 4.15 – The Jandel Engineering four point probe
76
4.4 – Electrochemical Analysis
4.4.1 ‐ Instrumentation
To investigate the electrochemical properties of single electrochromic material layers, a potentiostat was employed with a three electrode setup in a suitable liquid electrolyte. The sample in question formed the working electrode, with Pt wire employed as the counter electrode and either a Calomel or Ag/AgCl reference electrode.
Figure 4.16 – Radiometer Analytical PGZ301 Potentiostat Initially a Radiometer Analytical PGZ301 (figure 4.16) was used as the driving potentiostat. However due to limitations imposed by the instrument and the software, later
results
and
synchronised
in‐situ
broadband
spectroelectrochemical
measurements were performed using an Ivium IviumSTAT potentiostat (figure 4.17).
Figure 4.17 – Ivium IviumSTAT Potentiostat 77
4.4.2 – Electrolytes Essentially, an electrolyte solution is any solution capable of carrying an ionic (or protonic) current. Selection of a suitable electrolyte is an important consideration for any electrochemical investigation, since the pH of certain electrolyte solutions may be very high or low, or the electrodes or species under investigation may be soluble or reactive with the solvent or species employed. Additionally they may not be stable across the desired potential range. Such conditions may be incompatible with the sample and lead to degradation or complete destruction of the sample. This can be observed for electrochromes such as cathodic WO3, which works efficiently in acidic (low pH) conditions, yet is quickly etched by very basic conditions, whereas anodic NiO favours basic conditions, and quickly degrades in very acidic electrolytes. Equally, the electrolyte species themselves may attack the working electrode, or even in some cases, such as those containing lithium salts, the solvent. Therefore careful consideration must be made to the choice of electrolyte. Typically anodically colouring electrochromes possess low oxidation states and favour basic conditions, whereas cathodically colouring electrochromes typically possess high oxidation states and favour acidic conditions. The electrolytes employed for liquid electrochemical characterisation in the course of this work are as follows: Table 4.1 – Liquid electrolytes employed in the course of this work
Electrolyte
Solvent
Concentration
Electrochromes
KOH
H2O
0.1M
NiO, TiO2, Ni‐Cr Oxide
LiClO4
PC
1.0M
NiO, TiO2, WO3
KCl
H2O
1.0M
Prussian Blue
78
4.4.3 – Cyclic Voltammetry For electrochemical analysis to provide useful data, the potential range forms an important consideration. In this work, the suitable potential range was set within that of the selected electrolyte and determined using CV. Choosing values outside of the electrolyte window may lead to the electrolysis of water (in the case of aqueous electrolytes) and undesired gas evolution or other un‐wanted side reactions. Also, unsuitable values may result in unsuitable conditions (such as change of PH, insoluble deposition of Li2O, etc…) which lead to a reduction if efficiency or function of an electrochrome. By varying the scan rate and potential range, suitable values for CA analysis were determined for each electrochrome / electrolyte combination (see 4.4.4).
4.4.4 – Chrono Amperometry The following colouration and bleaching potentials were used for CA measurements: Table 4.2 – Various electrolytes employed for characterisation of materials
Electrochrome Electrolyte Concentration
Potentials
Colour
Bleach
Reference
KOH
1.0M
+0.7V
‐0.6V
SCE
Ag/AgCl
LiClO4‐PC
1.0M
+2.0V
‐0.6V
Ag/AgCl
Ni‐Cr Oxide
KOH
1.0M
+0.7V
‐0.6V
Ag/AgCl
TiO2
KOH
1.0M
‐1.2V
+1.0V
Ag/AgCl
LiClO4‐PC
1.0M
‐2.0V
+0.6V
Ag/AgCl
WO3
LiClO4‐PC
0.5M
‐2.0V
+ 0.6V
Ag/AgCl
Prussian Blue
KCl
1.0M
Various
NiO
79
Ag/AgCl
4.5 –Spectroelectrochemical Analysis In the field of electrochromism, and indeed, any chromic technology, spectroscopy is one of, if not the most valuable characterisation techniques available. Spectroscopic measurements may be performed ex‐situ (out with the electrochemical cell) or in‐situ (within the electrochemical cell). While many studies may make sole use of ex‐situ measurements for electrochromic analysis, this approach, while useful, provides only a partial picture of the electrochromic colour change and also may be inaccurate for some materials, for example:
Where a material may exhibit more than one colour, e.g. Prussian Blue, intermediate colour changes and / or absorption peaks may not be observed, this is especially the case for materials which colour or bleach quickly.
The removal of a material from the electrolyte, exposure to atmosphere and / or cleaning solutions, and drying before ex‐situ measurement, may itself often result in a partial or even complete colour change of the material. Thereby rendering the ex‐situ measurement inaccurate.
Transmission change measurements for different applied potentials, inserted charges or colouration times, require many different separate measurements and are therefore time consuming, inefficient and may require many more samples for testing to ensure repeatability.
In‐situ spectroscopy in comparison, is where both spectroscopic and electrochemical measurements are taken simultaneously by placing the electrochemical cell in the light path of a suitable spectrophotometer. Through synchronisation of the respective data, a complete picture of the spectroelectrochemical behaviour of a material or device can be realised. Different research groups often develop bespoke solutions to in‐situ spectroelectrochemical measurement [63], [66], [68], [195], [198], [244], [284], the details of such systems of measurement and the spectroelectrochemical cells used rarely being mentioned or described in detail, further leading to variations and lack of comparability between studies.
80
A novel solution to this problem has been developed and evolved, resulting in fast and accurate broadband spectroelectrochemical measurement of electrochromic materials and devices. This solution and details of its development and testing are presented in 5.2.
81
Chapter 5 – Measurements, Results & Discussion This chapter presents the results obtained from investigations performed throughout the course of this project. These results are divided into multiple sections as follows:
5.1 looks at the considerations required when undertaking a research study aimed at producing commercially scalable results, presenting experimental details and results obtained during development of suitable substrate tooling.
5.2 concentrates upon the development of a system for the in‐situ spectroelectrochemical characterisation of electrochromic materials and the acquisition of synchronised real‐time broadband spectroscopic and electrochemical data, along with the development of suitable in‐situ cells.
5.3 presents data obtained for ITO coated substrates as produced in the course of this work. As explained in 3.2, ITO is the most commonly used TCO material, and based upon its availability and proven performance, has been selected as the TCO electrode material employed in the course of this work. The presented results focus upon those films produced upon the PlasmaCoat deposition system (4.1.6), since they have been found to possess unique crystalline orientation while providing good performance for films produced at room temperature and at a notably high deposition rate.
5.4 – 5.8 present investigations relating to the optimisation of thin film material properties for the chosen electrochromic materials. Each of the materials has been characterised in depth such that the material can be deposited reproducibly, and the specifics electrochromic behaviour of the materials are clear, so that their feasibility for inclusion in an electrochromic device intended for commercial scale production may be assessed.
5.9 Provides a table summary of the electrochromic properties of materials produced in the course of this work.
Finally, 5.10 presents the attempts made toward the production of working electrochromic devices.
82
5.1 – Commercial Scalability
5.1.1 – Preliminary Work As discussed in 4.1.8, for the viability of materials and devices constructed with said materials to be established, it is necessary to perform commercial scale depositions involving many substrates. During the early stages of this project, work was focused upon producing the conductive ITO substrates required for the subsequent electrochromic layers and device construction. At this point, as is often the case in research institutions where many different shapes and sizes of substrates are employed, substrates were attached to the vacuum system tooling using carbon or Kapton® tape. This method of placing substrates inside a vacuum coating system often works well for low volume applications; however it presents significant challenges and limitations when attempting to deposit commercial scale quantities, namely:
Attachment of many substrates, and subsequent removal and cleaning of substrates and tooling is inefficient, involving a great deal of time and dexterity.
Reproducibility of substrate placement and tape attachment is questionable, making run‐to‐run reproducibility as would be required in a commercial scale situation hard if not impossible.
No option for reproducible control or masking of deposition area.
To assess the significance of these limitations and produce the majority of ITO required for the work presented herein, ITO was optimised and produced on the MicroDyn commercial sputtering system described in 4.1.4. Glass microslide substrates were attached to a flat plate hexadecagon rotary drum using two small squares of carbon tape at each end. Following optimisation of the ITO thin films, it was apparent that not only were the aforementioned issues a reality, but that the carbon tape used significantly affected
83
the resultant films. ITO is known to be a particularly sensitive material to produce, and the ITO produced in the course of this work is no exception. At the positions where the carbon tape had been attached to the substrates, the appearance of the film was markedly different (figure 5.1).
Figure 5.1 – ITO coated microslide produced at ambient room temperature in MicroDyn 40000 series commercial sputtering system under optimal conditions. Note the discoloured areas toward each end of the substrate, which correspond to the positioning of the carbon tape. This variation was confirmed through spectroscopic analysis using an Aquila Instruments nkd8000 (4.2.2) at 10°. Fitting of the spectra indicated that thickness of the thin film at the centre of the sample was 262.5nm, while at the edges of the sample was as much as 277.0nm, a 5.5% variation. The sheet resistance of the samples was then measured for multiple points (as denoted in figure 5.2 below), and found to typically vary by 22‐41%. This variation in the deposition could be attributed to a ‘heat‐sinking’ effect where the carbon tape was conducting thermal energy away from the surface of the substrate in the areas it was attached. This hypothesis is drawn to because ITO is known to be sensitive to substrate temperature during deposition [226], [285]. Such a variation in uniformity across a sample could present significant problems for electrochromic devices. In the case of an electrochromic device, the variation in resistivity of the ITO layer could result in uneven transition, and stressing of small areas of the resultant devices as a result of concentrated over and / or undercharging. The magnitude of this variation in sheet resistance was far greater than expected, so further depositions were performed to
84
confirm the validity of the measurements obtained. It was found that subsequent runs exhibited the same level of variation in sheet resistance. Additionally, the unmasked substrates were coated on the edges as well as the intended surface which could result in short circuiting of layers in devices where many layers are deposited on top of one another, confirming the need for an alternative approach to slide attachment.
Figure 5.2 – Sheet resistance measurement positions and results for different samples attached by the carbon tape method. (Samples obtained from the same position from three consecutive optimised runs)
Discussion After evaluation of the results established above, it was determined that methods of attaching or holding substrates to deposition systems in a uniform manner was desirable, and the following basic requirements were drawn: 1. They must hold substrates in a safe and secure manner to avoid damage to the system. 2. The substrates must be held uniformly to avoid uneven deposition effects and ensure reproducibility. 3. The main substrates are standard glass microscope slides (26x76mm, 1.0mm thick – 4.1.1).
85
4. Slide masks to control the deposited area should be used (4.1.7). 5. Any materials employed must be non‐magnetic (due to the effect it may have upon the resultant depositions and interaction with the magnetron in the case of sputtering systems). 6. Holders must be designed to cope with the conditions inside the deposition chamber; high vacuum, large rotational velocity, high temperature. The other desirable attributes were also drawn:‐ 7. It should be quick and efficient compared to the tape attachment method. 8. Two sided deposition without handling of the substrates may be advantageous where possible. 9. Cost should be kept to a minimum. Adhering to these requirements, substrate holders were designed and produced, along with suitable slide masking arrangements for both the MicroDyn 40000 series commercial sputtering system and Satis MSLab 370 commercial scale electron beam system (As described previously in 4.1.4 and 4.1.3 respectively) as follows.
5.1.2 – Development of Substrate Tooling for the MicroDyn 40000 Series Commercial Sputtering System As described in 4.1.4, the MicroDyn 40000 series is a commercial sputtering system which employs planar sputtering targets. The substrates which are to be coated are loaded on a rotary drum. Two types of drum exist for the system, a circular one for coating flexible substrates (e.g. polymer film), and one comprising of sixteen flat plates which form a hexadecagon for rigid substrates.
86
Following the designation of the aforementioned requirements for substrate holder tooling, a suitable cassette‐type arrangement was designed in AutoCAD™ 2011 (figure 5.3 below), and a prototype was produced (figure 5.4). It was designed with a layered design utilising economical and accurate laser cut metal parts in order to satisfy the requirements from both an operational and cost perspective.
Figure 5.3 – AutoCAD™ image of cassette type substrate holder tooling
Figure 5.4 – The prototype of the cassette holder arrangement produced for testing and evaluation of the design. The prototype above was then successfully tested during a long deposition of SiO2, chosen to stress the holder (Long SiO2 depositions result in high substrate temperatures). The test deposition was completed without any ill‐effects. Following this, a full set of sixteen cassette‐carrier assemblies was produced. A crystal blanking
87
plate provides a hole in every‐other slide holder place, so that the quartz crystal deposition monitor can be utilised resulting in a total drum capacity of 72 microscope slides (figure 5.5).
Figure 5.5 – Close up photograph of substrate holder assembly with cassette partially removed. A picture of the complete drum assembly alongside the MicroDyn system is included in 4.1.4. Substrate Holder Tooling Test Results and Discussion Following design and construction of the new substrate holder assemblies, ITO was once again optimised for deposition and deposited upon glass microslide substrates. To use the new holders a 3mm edge mask (4.1.7) is required to be allow the substrates to be held securely by the holder, and is additionally useful for controlling the deposited area. The result is a slightly smaller 70x20mm deposited area in comparison to the old
88
carbon‐tape attachment method, however it ensures that there is no unwanted coating of the sides of the substrate. Immediately the difference in uniformity of the samples produced was apparent (figure 5.6), with no visible variation across the sample.
Figure 5.6 – ITO produced at ambient room temperature with the substrate holder assembly. The newly obtained samples were tested as before, with spectroscopic analysis performed using an Aquila Instruments nkd8000 at 10°. This time, fitting of the spectra indicated that thickness of the thin film at the centre of the sample was 276.5nm, while at the edges of the sample was this time slightly lower at 274.2nm, a far reduced variation of 0.8%. From figure 5.7, it can be observed that variation in sheet resistance was markedly reduced to just 3‐4% across the working area of the sample. This displays a drastic improvement in uniformity over the carbon tape attachment method (22‐41%), and of course the exact position of substrates can now be easily maintained between runs; significantly improving run‐to‐run reproducibility which is of vital importance to commercial production (see figure 5.7 opposite) [269]. This improvement is greater than was expected, with the remaining 3‐4% variation in sheet resistance, being attributed to the effect of variation across the drum, where coating of a flat substrate leads to slight unavoidable variation in deposition path between the target and substrate as demonstrated in figure 5.8.
89
Figure 5.7 – Sheet resistance measurement positions and results for different samples attached using the substrate holder cassette assemblies
Figure 5.8 – Depiction of substrate / target distance variation between centre and edges of a flat substrate and target material when attached to a drum‐ coater type system.
90
Utilising the slide holder assemblies, a very small variation may also be seen at the extreme edges at the border of the deposited area, approximately equal to the mask thickness (0.5mm), as a result of ‘shadowing effect’. This is of course of minimal significance compared to the original non‐uniform variation experienced, since these tiny areas will be of a greater sheet resistance (due to being thinner) and thus will not experience degradation due to overcharging in comparison to the bulk of the surface area. The amount of shadowing could be reduced if required through the use of thinner disposable slide masks. In line with the desired attributes, the efficiency of the process was also assessed qualitatively, and experienced significant improvement. The time required for the drum loading and unloading process (for 72 slides) was reduced from a 6 hour, whole‐ day operation (for 2 people) to a 45‐60 minute operation (with just one person). The process also requires no expensive double sided carbon tape, cleaning solvents / cloths, or knives. It may hence be concluded that with the substrate holders, the process becomes commercially viable and scalable; with two drums, one would be able to be prepare one while the other was pumping to vacuum before deposition, allowing for continual deposition to be performed with one machine. Sheet resistance was also found to vary across the drum, and experienced dramatic improvement following use of the substrate holder assemblies, as shown in table 5f (opposite).
Figure 5.9 – Substrate attachment / holder positions across width of the drum as referred to in table 5.1.
91
Table 5.1 – Comparison of ITO coated microslide properties attached using the carbon tape method and designed substrate holder cassette assemblies. Substrate positions are clarified by figure 5.9. Mean average (16 samples) central conductivities of drum positions (Point ‘C’ – width of drum – Ω/□) Position 1 2 3 4 5 OLD 20.9 20.7 21 21.6 22.8 NEW 14.7 14.7 14.8 16.5 17.2 Max variation at each position around drum (16 sample mean – point ‘C’ – single run – Ω/□) Position 1 2 3 4 5 OLD 3.2 2.6 1.3 2.2 1.9 NEW 0.3
