ACTAUNIVERSITATIS

UPSALIENSISUPPSALA

ISSN 1651-6214ISBN 978-91-554-8714-0urn:nbn:se:uu:diva-2044372013

Digital Comprehensive Summaries of Uppsala Dissertations from the Faculty of Science and Technology 1057

Functionalization of polymer electrolytes for electrochromic windows

İlkNur Bayrak PehlİvaN

Dissertation presented at Uppsala University to be publicly examined in Polhemsalen,Ångström Laboratory, Lägerhyddsvägen 1, Uppsala, Friday, September 20, 2013 at 13:15 forthe degree of Doctor of Philosophy. The examination will be conducted in English.

AbstractBayrak Pehlivan, İ. 2013. Functionalization of polymer electrolytes for electrochromicwindows. Acta Universitatis Upsaliensis. Digital Comprehensive Summaries of UppsalaDissertations from the Faculty of Science and Technology 1057. 172 pp. Uppsala.ISBN 978-91-554-8714-0.

Saving energy in buildings is of great importance because about 30 to 40 % of the energy inthe world is used in buildings. An electrochromic window (ECW), which makes it possible toregulate the inflow of visible light and solar energy into buildings, is a promising technologyproviding a reduction in energy consumption in buildings along with indoor comfort. A polymerelectrolyte is positioned at the center of multi-layer structure of an ECW and plays a significantrole in the working of the ECW.

In this study, polyethyleneimine: lithium (bis(trifluoromethane)sulfonimide (PEI:LiTFSI)-based polymer electrolytes were characterized by using dielectric/impedance spectroscopy,differential scanning calorimetry, viscosity recording, optical spectroscopy, and electrochromicmeasurements.

In the first part of the study, PEI:LiTFSI electrolytes were characterized at various saltconcentrations and temperatures. Temperature dependence of viscosity and ionic conductivityof the electrolytes followed Arrhenius behavior. The viscosity was modeled by the Binghamplastic equation. Molar conductivity, glass transition temperature, viscosity, Walden product,and iso-viscosity conductivity analysis showed effects of segmental flexibility, ion pairs, andmobility on the conductivity. A connection between ionic conductivity and ion-pair relaxationwas seen by means of (i) the Barton-Nakajima-Namikawa relation, (ii) activation energies ofthe bulk relaxation, and ionic conduction and (iii) comparing two equivalent circuit models,containing different types of Havriliak-Negami elements, for the bulk response.

In the second part, nanocomposite PEI:LiTFSI electrolytes with SiO2, In2O3, and In2O3:Sn(ITO) were examined. Adding SiO2 to the PEI:LiTFSI enhanced the ionic conductivity by anorder of magnitude without any degradation of the optical properties. The effect of segmentalflexibility and free ion concentration on the conduction in the presence of SiO2 is discussed.The PEI:LiTFSI:ITO electrolytes had high haze-free luminous transmittance and strong near-infrared absorption without diminished ionic conductivity. Ionic conductivity and optical claritydid not deteriorate for the PEI:LiTFSI:In2O3 and the PEI:LiTFSI:SiO2:ITO electrolytes.

Finally, propylene carbonate (PC) and ethylene carbonate (EC) were added to PEI:LiTFSIin order to perform electrochromic measurements. ITO and SiO2 were added to thePEI:LiTFSI:PC:EC and to a proprietary electrolyte. The nanocomposite electrolytes were testedfor ECWs with the configuration of the ECWs being plastic/ITO/WO3/polymer electrolyte/NiO(or IrO2)/ITO/plastic. It was seen that adding nanoparticles to polymer electrolytes can improvethe coloring/bleaching dynamics of the ECWs.

From this study, we show that nanocomposite polymer electrolytes can add newfunctionalities as well as enhancement in ECW applications.

Keywords: Electrochromism, Polymer electrolytes, PEI, LiTFSI, Nanoparticles, Ionicconductivity, Ion-pair relaxation, Near-infrared absorption

İlknur Bayrak Pehlivan, Uppsala University, Department of Engineering Sciences, Solid StatePhysics, Box 534, SE-751 21 Uppsala, Sweden.

© İlknur Bayrak Pehlivan 2013

ISSN 1651-6214ISBN 978-91-554-8714-0urn:nbn:se:uu:diva-204437 (http://urn.kb.se/resolve?urn=urn:nbn:se:uu:diva-204437)

To my grandparents

List of Papers

This thesis is based on the following papers, which are referred to in the text by their Roman numerals. I PEI–LiTFSI electrolytes for electrochromic devices: Characterization by

differential scanning calorimetry and viscosity measurements İ. Bayrak Pehlivan, R. Marsal, G. A. Niklasson, C. G. Granqvist, and P. Georén, Solar Energy Materials and Solar Cells, 94 (2010) 2399-2404.

II Ion conduction of polyethyleneimine-lithiumbis (trifluoromethylsul-fonyl)imide electrolytes İ. Bayrak Pehlivan, P. Georén, R. Marsal, C. G. Granqvist, and G. A. Nik-lasson, Electrochimica Acta, 57 (2011) 201-206.

III Ionic relaxation in polyethyleneimine-lithium bis (trifluoromethylsulfonyl) imide polymer electrolytes İ. Bayrak Pehlivan, R. Marsal, P. Georén, C. G. Granqvist, and G. A. Nik-lasson, Journal of Applied Physics, 108 (2010) 074102.

IV [PEI–SiO2]:[LiTFSI] nanocomposite polymer electrolytes: Ion conduc-tion and optical properties İ. Bayrak Pehlivan, C. G. Granqvist, R. Marsal, P. Georén, and G. A. Nik-lasson, Solar Energy Materials and Solar Cells, 98 (2012) 465-471.

V Ion conduction mechanism of nanocomposite polymer electrolytes com-prised of polyethyleneimine-lithium bis (trifluoromethylsulfonyl)imide and silica İ. Bayrak Pehlivan, C. G. Granqvist, and G. A. Niklasson, submitted to Electrochimica Acta, 2013.

VI A polymer electrolyte with high luminous transmittance and low solar throughput: Polyethyleneimine-lithium bis (trifluoromethylsulfonyl) imide with In2O3:Sn nanocrystals İ. Bayrak Pehlivan, E. L. Runnerstrom, S.-Y. Li, G. A. Niklasson, D. J. Milliron, and C. G. Granqvist, Applied Physics Letters 100 (2012) 241902.

VII Electrochromic devices with polymer electrolytes functionalized by SiO2 and In2O3:Sn nanoparticles: Rapid coloring/bleaching dynamics and strong near-infrared absorption

İ. Bayrak Pehlivan, R. Marsal, E. Pehlivan, E. L. Runnerstrom, D. J. Mil-liron, C. G. Granqvist, G. A. Niklasson, Solar Energy Materials and So-lar Cells (2013) in Press, http://dx.doi.org/10.1016/j.solmat.2013.06.010.

Reprints were made with permission from the respective publishers.

My contribution to the appended papers

I Sample preparation, measurements, analysis, most of the writing.

II Sample preparation, measurements, analysis, most of the writing.

III Sample preparation, measurements, analysis, most of the writing.

IV Sample preparation, measurements, analysis, most of the writing.

V Sample preparation, measurements, analysis, most of the writing.

VI Electrolyte preparation, part of nanoparticle preparation, measurements apart from TEM and absorbance measure-ments of ITO nanoparticles in a dilute solution, part of analysis, most of the writing.

VII PEI-based electrolyte preparation, part of proprietary elec-trolyte preparation, part of nanoparticle preparation, ECW prototype preparation, measurements, analysis, most of the writing.

Papers not included in the thesis

I Transparent conducting nanoparticle coatings for energy efficient fenestra-tion: Applications in electrochromics and thermochromics C. G. Granqvist, İ. Bayrak Pehlivan, Y.-X. Ji, S.-Y. Li, and G. A. Niklas-son, 56th Annual Technical Conference Proceedings, in: Society of Vacuum Coaters, Albuquerque, USA (2013) ISSN 0737-5921.

II Electrochromics and thermochromics for energy efficient fenestration: New functionalities based on nanoparticles of In2O3:Sn and VO2

C. G. Granqvist, İ. Bayrak Pehlivan, Y.-X. Ji, S.-Y. Li, and G. A. Niklas-son, Thin Solid Films (2013) submitted.

III Electrochromics and thermochromics for energy efficient fenestration: new applications based on transparent conducting nanoparticles C. G. Granqvist, İ. Bayrak Pehlivan, Y.-X. Ji, S.-Y. Li, E. Pehlivan, R. Marsal, and G. A. Niklasson, MRS Proc. (2013) submitted.

IV Progress in chromogenic materials and devices: New data on electro-chromics and thermochromics C. G. Granqvist, S.-Y. Li, İ. Bayrak Pehlivan, and G. A. Niklasson, MRS Online Proceedings Library, (2013) 1492, mrsf12-1492-g11-01 doi: 10.1557/opl.2013.155.

V Progress in electrochromics and thermochromics: Two new applications involving ITO nanoparticles S.-Y. Li, İ. Bayrak Pehlivan, G. A. Niklasson, and C. G. Granqvist, 55th Annual Technical Conference Proceedings, in Society of Vacuum Coaters, Albuquerque, USA (2012) pp. 41-46.

VI Oxide-based electrochromics: Advances in materials and devices C. G. Granqvist, İ. Bayrak Pehlivan, S. V. Green, P. C. Lansåker, and G. A. Niklasson, MRS Proceedings (2011) 1328 DOI: http://dx.doi.org/10.1557/opl.2011.1064.

VII Real time dielectric monitoring of glass transition in n-vinyl pyrrolidone polymerization F. Salehli, İ. Bayrak, R. R. Nigmatullin, and A. Arbuzov, Journal of Non-Crystalline Solids 353 (2007) 4366-4370.

VIII The first experimental confirmation of the fractional kinetics containing the complex-power-law exponents: Dielectric measurements of polymerization reactions

R. R. Nigmatullin, A. A. Arbuzov, F. Salehli, A. Giz, İ. Bayrak, and H. Catalgil-Giz, Physica B: Condensed Matter 388 (2007) 418-434.

Contents

1. Introduction ......................................................................................... 19

2. Electrochromism .................................................................................. 23 2.1. Layers of electrochromic window .............................................. 25

2.1.1. Substrate ................................................................................ 26 2.1.2. Transparent electronically conductive layer .......................... 26 2.1.3. Electrochromic layer .............................................................. 27 2.1.4. Electrolyte layer ..................................................................... 28

2.2. Coloration mechanism ................................................................ 30 2.3. Key parameters of electrochromism ........................................... 32 2.4. Electrochromic window applications ......................................... 34

3. Polymer electrolytes ............................................................................ 37 3.1. Ion-ion and polymer-ion interactions ......................................... 39 3.2. Ionic conductivity ....................................................................... 41 3.3. Molar conductivity ..................................................................... 45 3.4. Dielectric spectroscopy .............................................................. 46

4. Nanocomposite polymer electrolytes................................................... 49 4.1. Optical properties ....................................................................... 49

4.1.1. Rayleigh scattering ................................................................ 50 4.1.2. Near-infrared absorption ........................................................ 50 4.1.3. Maxwell-Garnet theory .......................................................... 51 4.1.4. Ligand-exchange method ....................................................... 52

4.2. Ionic conductivity ....................................................................... 53

5. PEI-based electrolytes ......................................................................... 55

6. The electromagnetic spectrum ............................................................. 59

7. Dielectric/impedance spectroscopy ..................................................... 61 7.1. Basic principles .......................................................................... 61 7.2. Equivalent circuit models ........................................................... 65 7.3. Time-temperature superposition principle ................................. 67 7.4. The Barton-Nakajima-Namikawa relation ................................. 68

8. Optical spectroscopy ............................................................................ 69 8.1. Blackbody spectrum ................................................................... 69

8.2. The solar spectrum ..................................................................... 71 8.3. Basic principles .......................................................................... 71

8.3.1. Interaction between light and matter ...................................... 71 8.3.2. Solar and luminous transmittance .......................................... 72 8.3.3. Refractive index ..................................................................... 73 8.3.4. Fresnel equations ................................................................... 75

9. Viscosity .............................................................................................. 77 9.1. Classifications of fluids .............................................................. 78 9.2. Temperature dependence of viscosity ........................................ 80

10 Experimental ........................................................................................ 81 10.1. Materials ..................................................................................... 81 10.2. Electrolyte preparation ............................................................... 81

10.2.1. Preparation of PEI:LiTFSI electrolytes ............................. 81 10.2.2. Preparation of nanocomposite PEI:LiTFSI electrolytes .... 82 10.2.3. Preparation of PEI:LiTFSI:PC:EC-based electrolytes ...... 82

10.3. Synthesis of In2O3 and ITO nanoparticles .................................. 83 10.4. Electrochromic film deposition .................................................. 84 10.5. Formation of electrochromic window prototypes ...................... 85 10.6. Dielectric/impedance spectroscopy measurements .................... 85 10.7. Thermogravimetric analysis ....................................................... 88 10.8. Differential scanning calorimetry ............................................... 90 10.9. Viscosity measurements ............................................................. 91 10.10. Ex-situ optical measurements ..................................................... 93 10.11. In-situ optical measurements ...................................................... 96 10.12. Thickness measurements ............................................................ 97 10.13. Electrochromic measurements .................................................... 97

11. Results and discussion ......................................................................... 99 11.1. PEI:LiTFSI electrolytes .............................................................. 99

11.1.1. Different PEI molecular weight ........................................ 99 11.1.2. Applied voltage dependence of the dielectric/impedance

measurements ................................................................. 103 11.1.3. Different LiTFSI concentration (Paper I) ........................ 106 11.1.4. Ion association and mobility effects on ionic

conductivity (Paper II) .................................................... 107 11.1.5. The role of ion-pair relaxation and ionic conductivity on

the conduction process (Paper III) .................................. 111 11.2. Nanocomposite PEI:LiTFSI electrolytes .................................. 113

11.2.1. PEI:LiTFSI:SiO2 electrolytes .......................................... 113 11.2.1.1. Different SiO2 concentration (Paper IV) .................. 113 11.2.1.2. Different LiTFSI concentration (Paper V) ............... 115

11.2.2. PEI:LiTFSI:In2O3 electrolytes ......................................... 118 11.2.3. PEI:LiTFSI:ITO electrolytes (Paper VI) ......................... 120

11.2.4. PEI:LiTFSI:ITO:SiO2 electrolytes .................................. 124 11.2.5. SiO2 and ITO added PEI:LiTFSI:PC:EC electrolytes ..... 125

11.2.5.1. PEI:LiTFSI:PC:EC electrolytes ................................ 126 11.2.5.2. PEI:LiTFSI:PC:EC:SiO2 and PEI:LiTFSI:PC:EC:ITO

electrolytes (Paper VII-Part I) ................................. 128 11.3. The electrochromic windows ................................................... 129

11.3.1. ECWs with NiO and PEI:LiTFSI:PC:EC ........................ 129 11.3.2. ECWs with NiO and PEI:LiTFSI:PC:EC with ITO and

SiO2 (Paper VII-Part II) .................................................. 132 11.3.3. ECWs with NiO and proprietary electrolyte

(Paper VII-Part III) ......................................................... 134 11.3.4. ECWs with NiO and crosslinked proprietary electrolyte 135 11.3.5. ECWs with IrO2 and PEI:LiTFSI:PC:EC-based

electrolytes ...................................................................... 136

12. Conclusions ....................................................................................... 139

13. Summary in Swedish: Funktionalisering av polymera elektrolyter för tillämpningar i elektrokroma fönster ........................................... 145

14. Summary in Turkish: Polimer elektrolitlerin elektrokromik cam uygulamaları için fonksiyonelleştirilmesi ......................................... 149

Acknowledgements ..................................................................................... 155

References ................................................................................................... 159

Abbreviations

A Absorbance a Radius of ions AF Adhesive film Aν Cross-section surface area

B

Magnetic field BMI Butyl methylimidazolium BNN Barton-Nakajima-Namikawa Bη,σ VFT constant C Capacitance c Salt concentration C΄ Real part of capacitance C΄΄ Imaginary part of capacitance C0 Capacitance of air CD∞ High frequency capacitance of DHN model CE Coloration efficiency Ce Electrode/electrolyte interfacial capacitance Cedge Capacitance due to the electrical stray fields from the borders of

the sample capacitor Celectrolyte Capacitance of electrolyte Cem Measured empty cell capacitance Cg Geometrical capacitance CHN Conductive Havriliak-Negami Cmeas Measured sample capacitance CR Contrast ratio CC∞ High frequency capacitance of CHN model cs Speed of light csalt Molar concentration given in unit of mol/L CTeflon Capacitance of Teflon Cwire Capacitance due to non ideal shielding of the connection wires

D

Electric displacement D Diffusion coefficient D- Diffusion coefficient of anions D+ Diffusion coefficient of cations d Thickness DHN Dielectric Havriliak-Negami DMF Dimethylformamide

DOS Bis(2-ethylhexyl)sebacate DSC Differential scanning calorimetry

E

Electric field Ea-σ Activation energy for ionic conductivity Ea-η Activation energy for viscosity Ea-r Activation energy for ion-pair relaxation EC Ethylene carbonate ECW Electrochromic window Em Emissivity F Force f Frequency FBNN BNN numerical factor Ff Filling factor Fl Fraction of light FTO Fluorine-doped tin oxide G Glass h Planck constant HMw High molecular weight HN Havriliak-Negami I Current I΄ Real part of current I΄΄ Imaginary part of current I0 Amplitude of current Ii Incoming light intensity Il Intensity of light IR Reflected light intensity Is Spectral radiance IT Transmitted light intensity ITO Tin-doped indium oxide j

Current density k Extinction coefficient kB Boltzmann constant l Distance LBNL The Lawrence Berkeley National Laboratory LED Light emitting diode Li Depolarization factor LMw Low molecular weight MMw Moderate molecular weight N Complex refractive index n Refractive index NA Avogadro constant Nc Number of charge carrier Nf Number of free ions Nn Number of nanoparticles per unit volume

NNP Ratio of the refractive index of nanoparticles to that of polymer NT Number of total salt molecules per unit volume OD Optical density OLAC Oleic acid OLAM Oleylamine

P

Polarization PAN Polyacrylonitrile PANI Polyaniline PBB Blackbody radiance PC Propylene carbonate PD,C Power-law exponent PEG Polyethyleneglycol PEI Polyethyleneimine PEI-G2 Poly(N-(2-(2-methoxyethoxy)ethyl)ethylenimine) PEO Polyethyleneoxide PET Polyethyleneterephthalate PMMA Polymethylmethacrylate PPG Polypropyleneglycol PPO Polypropyleneoxide PTFE Polytetrafluoroethylene PVA Polyvinylacetate PVB Polyvinylbuteral PVC Polyvinylchloride Q Induced electrical charge q The electron charge R Reflectance Rγ The gas constant Rb Bulk resistance RC,D Ionic resistance rc Radius of cone RC∞ High frequency resistance of CHN model RD∞ High frequency resistance of DHN model Rdiff Diffuse reflectance Re Electrode/electrolyte interfacial resistance rp,s Interface amplitude reflectance RR Rayleigh scattering coefficient RSP Reflectance of the Spectralon plate Rspec Specular reflectance Rtot Total reflectance RZ Resistance SDR Diffuse reflectance signal SDR-ap Diffuse reflectance signal of aperture SDT Diffuse transmittance signal SOG A kind of thermally decomposed SiO2-coated glass

SPE Solid polymer electrolyte STR Total reflectance signal STR-ap Total reflectance signal from aperture STT Total transmittance signal t Time T Transmittance T0 Temperature related to Tg tan(δ) Loss factor tb Switching time for bleaching Tb Transmittance in the bleached state tc Switching time for coloring Tc Transmittance in the colored state Tdiff Diffuse transmittance Tdir Direct transmittance TEG Tetraethyleneglycol TEM Transmission electron microscope TFSI Bis (trifluoromethane)sulfonimide Tg Glass transition temperature TGA Thermogravimetric analysis THF Tetrahydrofurane Tlum Luminous transmittance tp,s Interface amplitude transmittance Tsol Solar transmittance Tt Absolute temperature Ttot Total transmittance UD,C Power-law exponent V Voltage V΄ Real part of voltage V΄΄ Imaginary part of voltage V0 Amplitude of voltage Vb Applied voltage for bleaching Vc Applied voltage for coloring XEL Proprietary X-linked Crosslinked Y Complex admittance Y΄ Conductance Y΄΄ Susceptance Z Complex impedance z Valances of ions Z΄΄ Reactance Zi Impedance for ith circuit element Ztot Total impedance α Absorption coefficient αpol Polarizability of nanoparticles δ Diameter of nanoparticles

ΔC Capacitance difference ΔQ Inserted/extracted charge per unit area Δε Dielectric relaxation strength η Viscosity ηP Viscosity of polymer ε Dielectric permittivity ε΄ Real part of dielectric permittivity ε΄΄ Imaginary part of dielectric permittivity ε0 Permittivity of vacuum ε∞ High frequency permittivity of polymer electrolyte εb Bulk permittivity of polymer electrolyte εm Permittivity of polymer εMG Maxwell-Garnett effective permittivity εP Permittivity of nanoparticles ζ Charge density γ Shear rate

Λ Molar conductivity Λ0 Molar conductivity at infinite dilution of electrolyte θ Angle λ Wavelength μ Mobility μ0 Permeability of vacuum μa

d Average dipole moment μd Dipole moment ν Velocity ξ Power-law exponent ρ Density σ Conductivity σSB Stefan-Boltzmann constant σ0 Pre-exponential factor σion Ionic conductivity τ0 Yield stress τD,C Relaxation time τη Torque τν Shear stress φ Phase angle φlum Luminous sensitivity of the humen eye φsol Solar irradiance χe Susceptibility ω Angular frequency Ω Angular velocity # Number

19

1. Introduction

The high energy use has caught people’s attention and sparked the interest for new energy sources, energy conservation, and methods to reduce the energy demand. Human beings spend time mostly in inside environments such as buildings. About 30-40 % of the energy in the world is used in build-ings, due to heating, cooling, lighting, and ventilation [1]. Windows cause the energy loss in buildings, allowing heat transfer from the inside to the outside or vice versa, depending on the climate. Living without windows is not a solution to prevent energy consumption because people need day light-ing, as it has positive mood effects.

Considering varying climate during the day or season, building energy savings can be achieved by using windows that control incoming visible light and solar energy. Thus, it can be possible allow solar energy to pass into a building if heating is required; and it can be possible not allow solar energy to pass into a building if cooling is required. In addition, allowing the inflow of the visible light into buildings reduces the use of lighting by mak-ing use of daylight. On the other hand, reducing the inflow of the visible light into buildings provides indoor comfort by reducing glare.

Controlling incoming light and solar energy can be provided by materials having variable optical properties. These materials are called chromogenic materials or chromic materials. Chromism is a reversible change in color of a material by a stimulus effect such as electric field, heat, light, or pressure [2,3]. The word chromism comes from the Greek word χρωµα, meaning color. The change in the optical properties can be in transmission or reflec-tion properties. There are several types of chromism that can be used in en-ergy-related applications, such as electrochromism, photochromism, and thermochromism. Photochromism is defined as a reversible coloration by radiation of light [4]. Photochromic materials such as AgCl need ultraviolet radiation in order to become dark [4-7]. Therefore, optical changes in photo-chromism are not optionally controllable. Thermochromism is a change in color of a material as a consequence of temperature changes [8,9]. Thermo-chromic materials such as VO2 need to absorb heat from the environment in order to be activated. Optical changes in thermochromism are not optionally controllable. Having high transition temperatures and low modulation in transmittance can be other drawbacks of thermochromism in energy-related window applications [10,11].

20

Some other chromogenic materials, such as electrochromic materials and polymer-dispersed liquid crystals, can be electrically activated [4]. They change their optical properties when an electric field is applied. Polymer-dispersed liquid crystals are oriented to the electric field and become trans-parent when they are subjected to an electric field. Polymer-dispersed liquid crystals are colored when they are not exposed to the electric field. They need the power in order to maintain a bleached state and haze remains in the bleached state [4].

Electrochromism is one of the promising technologies that can be used to provide a reduction in energy consumption in buildings together with in-creased indoor comfort [12-18]. An electrochromic thin film is able to change its transparency in the visible and near-infrared radiation range and can be colored and bleached in response to an applied voltage. Thus, it is possible to control incoming visible light and solar energy while still provid-ing a visible outside view by using electrochromic films in windows. In this thesis, full configuration of an electrochromic system, which is explained in Section 2.1, is referred to as an electrochromic window or ECW for short.

Figure 1. Schematic illustration of cooling energy versus electric lighting energy for different types of windows. From Ref. [19].

Electrochromism is a significant technology in energy-saving applications due to its positive features. It was reported that ECWs have an advantage over other types of switchable windows because ECWs can provide energy saving both from electric lighting and from cooling [19]. Figure 1 shows a schematic illustration of cooling and electric lighting energy needs from

Clear glass

Tinted glass

Reflective glass

Photo-chromic

Thermochromic Electro-chromic

40

30

20

10

0 0 6020 40 80

Coo

ling

ener

gy (

kWh/

m2 )

Electric lightning energy (kWh/m2)

21

different types of windows. Clear glass has high transparency and does not need too much energy for lighting. On the other hand, it consumes the high-est energy for cooling, due to the high-heat transfer through the window. The opposite way around, better heat transfer blocking, is possible by using re-flective glass, but more energy is needed for lighting. Similarly, photochro-mic and thermochromic windows have an advantage only for one type of energy, lighting and cooling, respectively. It is seen that it could be possible to consume the minimum energy for both lighting and cooling when an ECW is used. It has been reported that ~4.5 % of the annual energy use in buildings in the USA can be saved using well insulated ECWs [17]. An ECW has dynamic transmittance modulation, which can be controlled op-tionally, requires power only during switching and very low voltage ~1-2 V is used during switching. The switching time is rapid. Electrochromism technology can also be used for manufacturing on a large scale. ECWs have a long open-circuit memory and keep the modulated transmittance for a long time. From these perspectives, electrochromism is more attractive for en-ergy-related window applications.

Polymer electrolytes are used as an ion conductive layer in ECWs and play a significant role in the working of ECWs. Fundamental requirements for polymer electrolytes in ECW are that they should have high ionic con-ductivity and high transparency. Therefore, investigations on ion conduction and optical properties of polymer electrolytes are of great importance. A simple polymer electrolyte contains one type of polymer and one type of salt. The polymer is a medium for the salt, which is composed of negative and positive ions. Various polymers have been used as hosts to form electro-lytes. Polyethyleneimine (PEI) can be a good alternative for ECW applica-tions because it is good at dissolving various salts and because it has an amorphous phase within a wide temperature range, which can increase ionic conductivity. Lithium bis (trifluoromethane)sulfonimide (LiTFSI) is one of the most interesting salts in many recent research project on polymer electro-lytes and their applications because of its favorable effect on ionic conduc-tivity.

Nanoparticles are materials with a diameter of less than 100 nm and which have different properties than their bulk form. Nanocomposite poly-mer electrolytes, which are polymer electrolytes incorporating nanoparticles, have been used in many recent electrolyte studies. Adding nanoparticles into polymer electrolytes can improve physical properties such as electrical, me-chanical and thermal properties. SiO2 nanoparticle is one of the most widely used nanoparticles in nanocomposite polymer electrolytes. Improvement in ion conduction, electrochemical stability, mechanical and thermal properties make SiO2 nanoparticles attractive for the development of polymer electro-lytes.

Tin-doped indium oxide (ITO) nanoparticles in a polymer matrix can provide near-infrared absorption and high visible transparency without no-

22

ticeable haze. Therefore, application of the ITO-doped polymer system in windows is able to block solar heat with accompanying high visual and day-lighting performance, hence improving the energy efficiency of the win-dows. Despite the fact that the near-infrared blocking nanoparticles such as ITO are a promising technology, we are not aware of any report on polymer electrolytes with the near-infrared blocking nanoparticles.

In this study PEI:LiTFSI polymer electrolyte was chosen as a model polymer electrolyte for ECW applications. The PEI:LiTFSI electrolytes were functionalized by adding nanoparticles. This study consists of three parts. In the first part, ionic conduction of PEI:LiTFSI electrolytes was investigated. In the second part, nanocomposite PEI:LiTFSI electrolytes with SiO2, In2O3, and ITO were examined. Finally, nanoparticle-added PEI:LiTFSI-based electrolytes and proprietary electrolytes were tested for ECWs with NiO/WO3 and IrO2/WO3 electrochromic layers. Effects of the presence of the nanoparticles on the electrochromic performance of the ECWs were dis-cussed. We should state that this study will be the first on ITO nanoparticle-dispersed polymer electrolytes. Here, we show that nanocomposite polymer electrolytes can add new functionalities as well as enhancement in ECW applications.

The investigations were carried out in collaboration with the Milliron Re-search Group at the Lawrence Berkeley National Laboratory (LBNL), CA, USA for the ITO- and In2O3- related parts of the study. Collaboration in most of this work, apart from In2O3 and ITO, was done with the company ChromoGenics, which was founded in 2003 by Prof. Claes-Göran Granqvist and his group at the Ångström Laboratory at Uppsala University.

23

2. Electrochromism

In this chapter, an introduction to electrochromism and applications of elec-trochromism technology is given. Components of an electrochromic window is described. Key parameters of electrochromism are mentioned. Coloration mechanisms of WO3, NiO, and IrO2 and some examples of the previous studies on WO3/NiO and WO3/IrO2 based ECWs are reported.

The history of electrochromism starts in the 1950s. A qualitative theory and quantitative prediction on the shift of the absorption and emission spec-tra of certain dyes by solvent and polarization effect was investigated in the 1950s [20,21]. Based on these investigations, a possible change of color in dyes by applying an electric field was suggested by Platt in 1961 [22]. It was the first time that the term electrochromism was used.

Figure 2. Early schematic picture of an ECW [26].

In 1969, electrochromism was experimentally shown on WO3 by Deb [23]. It was considered to be used in displays. However, it did not arouse significant interest until 1984, due to the development of liquid-crystal-based technol-ogy. In 1984, the possible use of eletrochromism technology for energy effi-ciency in buildings was proposed [24,25]. In 1985, Svensson and Granqvist introduced the potential application of electrochromic materials in smart windows [26]. Thus, a new term smart window entered our lives. Its “smart-ness” comes from its variable optical properties. Figure 2 shows early sche-

24

matic design of a smart window [26]. The tests on smart windows were done with the configuration of glass/ITO/WO3/MgF2/Au.

Electrochromism technology allows to get different optical properties in the visible, near infrared [27], thermal infrared [28] and microwave regions [29]. Electrochromism can have different application areas, such as in reflec-tive displays, car rear-view mirrors, car roofs, ski goggles, airplane windows, and decorative objects, in addition to window applications and building fa-çades [2,30-33]. Figure 3 (a) shows a prototype for an aircraft window in bleached and colored states [34]. In Fig. 3 (b) and (c), a motocycle helmet with an electrochromic visor produced at ChromoGenics is shown [16]. One can easily adjust to the light intensity difference inside and outside a tunnel in bright weather by using a motorcycle helmet with an electrochromic visor.

(a) (b) (c) Figure 3. Some applications of electrochromism. Electrochromic aircraft window (a) [16] and electrochromic visor for a motorcycle helmet in colored (b) and bleached (c) states [34].

Various companies, such as SAGE Electrochromics Inc (Faribault, MN, U.S.), EControl-Glas GmbH & Co. KG (Plauen, Germany) and GESIMAT GmbH (Berlin, Germany) working on the development and commercializa-tion of electrochromism technology were summarized in Ref.s [2,35]. The ECWs that are produced by these companies are above 80x120 cm2. Trans-mittance of the ECWs changes from 40-52 % to 6-15 % in the solar range and from 62-75 % to 3-8 % in the visible range. ECWs produced by the SAGE for building applications can be seen in many building façades, such as the Kimmel Center (Philadelphia, PA, U.S.), the Utrecht Government Building (Utrecht, Netherlands) Saint-Gobain’s Habitat Lab (Milan, Italy) [36]. ChromoGenics AB (Uppsala, Sweden) developed ECWs with flexible substrates; hence cheaper and easier manufacturing on a large scale is possi-ble [37,38]. Gentex (Zeeland, MI, U.S.) focused on electrochromic mirrors and aircraft windows [34]. Aveso, a spin-off from the Dow Chemical Com-pany (Minneapolis, MN, U.S.) does research on printed electrochromic inks for display applications [39].

25

2.1. Layers of electrochromic window An ECW is configured as a Li-ion battery. In Fig. 4, an example of the most common ECW structure is shown.

Figure 4. Schematic picture of an ECW, showing the transport of positive ions and electrons as a result of applied voltage at the transparent electronically conductive layers.

It contains five layers between two transparent substrates [30,32]. A trans-parent ion conductive layer, which is called electrolyte layer, lies at the cen-tre of the structure. In the figure, layers on the left side and the right side of the electrolyte work as cathode and anode, respectively. On the left side of the electrolyte layer, there is a thin film of an electrochromic material such as WO3. This electrochromic layer is a cathodic film hence it is colored by the insertion of ions into the film. There is an ion storage layer on the right side of the electrolyte layer. This thin film can be either electrochromic, such as NiO and IrO2, or not electrochromic, such as CeO2 [30]. Electrochromic films are preferably used in the ion storage layer in order to get the highest possible modulation in the transmittance. This anodic layer is colored by the extraction of ions from the film. These three layers are placed between two transparent electronically conductive layers, such as ITO.

26

Figure 5. Prototype of an ECW with four panes each 30x30 cm in size. Two of the panes are in colored and two of them are in bleached state. Photo: Ångström Labora-tory, Uppsala.

When a DC voltage such as ~1-2 V is applied between the transparent elec-trical conductors, ions are shuttled through the electrolyte layer between the anodic and the cathodic electrochromic film. At the same time, electrons are shuttled to/from the electrochromic layers from/to the electronically conduc-tive layers. Thus, charge balance is provided when electrons and ions meet in the electrochromic layers. When electrons and ions meet in cathodic elec-trochromic layer, ECW is colored. When electrons and ions meet in anodic electrochromic layer, ECW is bleached. An example of an ECW in the col-ored and bleached states is shown in Fig. 5.

2.1.1. Substrate Substrates of ECWs can be glass or transparent polymers [40-42]. Polyethyl-eneterephthalate (PET), polyvinylbuteral (PVB) and polyethylene napthalate are examples of polymer substrates. Glass substrates are most commonly used compared to polymer substrates because the quality of the polymer is destroyed easily, compared to glass. On the other hand, fabrication of ECWs with a plastic substrate is a practical and cheaper technique [17].

2.1.2. Transparent electronically conductive layer Transparent conductors can provide a low thermal emittance because they have low infrared emittance. Due to their electric properties and reflective properties in the near-infrared, they are used in energy-efficient windows such as low-emissivity or solar control windows, solar cells and ECWs [42,43]. Electric resistivity and transmittance are the most important proper-ties for transparent conducting layers in electrochromic applications. The

27

layers should have high electronic conductivity because the electric field needed for ECWs to work is applied between these conducting layers. On the other hand, these layers should not affect the transmittance of ECWs. A transparent conducting layer can be metal-based or oxide-based [30]. ITO is the most commonly studied transparent conducting oxide in ECWs [42,44]. Its high electronic conductivity, ~104 S/cm, and low optical absorption make it preferable [45]. Its refractive index is 1.9 and the host semiconductor In2O3 has a band gap around 3.75 eV and different amounts of doped-Sn atoms provide different free electron density [45,46].

2.1.3. Electrochromic layer Electrochromic layers can be classified into three different types according to their coloration properties [47]:

• having one colored and one bleached state, i.e. metal oxides, • having two distinctive colored states, i.e. polythiophene, • having more than two colored states, i.e. poly(3,4-

propylenedioxypyrrole).

Electrochromic material can be inorganic, such as metal oxide based, or organic, such as conducting-polymer based, or a combination of organic-inorganic materials [25,30,48-50]. Most well-known classes of materials having electrochromic properties are given in Table 1 together with some examples of each class.

Table 1. Electrochromic materials and examples [30,47,49,53-57]

Electrochromic materials Examples

Transition metal oxides Oxides of tungsten, nickel, iridium, vanadium, titanium

Prussian blue systems Prussian blue, prussian brown

Phthalocyanines Lutetium bis (phthalocyanine)

Conducting polymers Polyaniline, polythiophene

Viologens 1,1´-disubstituted-4,4´-bipyridilium salts

Carbazoles Polyethylcarbozole

Organic films such as conductive polymers can provide multi-colored states, fast switching time, high optical contrast and higher coloration efficiency [32,51]. The disadvantage of polymer films is their weak electrochemical stability, and hence their short cycling life [52]. There are still ongoing re-searches on enhancing coloration efficiency, switching speed and the proc-essing ability of the polymers. To improve the electrochromic performance

28

of the conductive polymers, they are combined with inorganic materials [50]. In this way, they have improved coloration efficiency and faster switching speed in comparison with the polymer. However, electrochemical properties of the polymers are not influenced by the inorganic materials. Therefore, their stability issue is still present [50]. Important parameters of electrochromism such as switching time, optical density, and coloration effi-ciency are introduced in Section 2.3.

In electrochromic layers, inorganic films such as WO3, Nb2O5, and NiO [30,58-62] are preferable, due to their advantage of providing higher durabil-ity. Electrochromic properties depend on the structure of the electrochromic film. In addition, different electrochromic materials show different absorp-tion spectra, hence different colors. It is possible to obtain various colors with different electrochromic materials [32,51]. For example, amorphous Nb2O5 is brown, while crystalline Nb2O5 is blue in the colored state; WO3 is blue; and TiO2 becomes blue or grey with the insertion of H+ or Li+ ion, re-spectively [30,63]. The most widely studied inorganic materials are WO3 as a cathodic layer, and NiO and IrO2 as anodic electrochromic layers [30,32]. Detailed information about WO3, NiO and IrO2 is given in Section 2.2 and 2.4.

Electrochromic thin films can be prepared in different ways. Some of the thin film coating techniques are dc magnetron-sputtering, sol-gel, such as dip coating and spin coating, electrochemical deposition, chemical vapour depo-sition and spray pyrolysis. Dc magnetron sputtering is preferable for ECW applications because it can be scaled up to industrial production.

In order to get a better performance of an ECW, pre-treatment should be done to electrochromic layers before ECW formation. There should be a charge balance between anodic and cathodic films. Cathodic film is charged, and anodic film is discharged. The charge balance can be done with different methods before the construction of ECWs. Ions can be added to the films while they are being sputtered [64]. Anodic film can be discharged by ozone exposure [65,66]. Another technique is that the films are electrochemically charged/discharged in a liquid electrolyte by chronoamperometry before the formation of ECW.

2.1.4. Electrolyte layer An electrolyte is defined as a substance with ions that provide ionic conduc-tion under the effect of an electric field. In ECW applications, an electrolyte should [67,68]

• be compatible with the anodic and the cathodic layers • provide high ion conduction • block electrons passing through between the electrochromic layers • have high transparency without having light-scattering effects.

29

Electrolytes should be compatible with both anodic and cathodic layers in order to get good results in the ECW applications. High ionic conductivity provides much faster coloration/bleaching, hence high modulation at trans-mittance can be reached in a much shorter time. In an ECW, an electrolyte makes a barrier for electrons between the anodic and cathodic electrochro-mic layers and protects the ECW from short-circuiting. Therefore, low elec-tronic conductivity is necessary to avoid a short circuit. Electrolytes should have high transparency and not deteriorate the transparency of the ECWs. There should not be any component in an electrolyte that scatters the incom-ing light.

Ion conduction is provided by protons or alkali metals such as Li+, Na+, K+, Rb+, Cs+, and Fr+. Because of the advantages of having high mobility due to the small size, H+ and Li+ are preferably used in ECWs. Therefore, proton-conducting and Li-conducting electrolytes are mostly studied for ECW applications [30,69]. Ta2O5, TiO2, ZrO2, polyvinylsulfonica-cid:polyvinylpyrrolidone:H+, and PEI:H3PO4 are some of the examples of proton-conducting electrolytes [70,71]. There are found numerous examples of Li+-conducting electrolytes such as polyethyleneoxide (PEO):LiTFSI, polymethylmethacrylate (PMMA):propylene carbonate (PC):LiClO4, poly-ethyleneglycol (PEG):Ti:LiI, which were used in electrochromic applica-tions [50].

An electrolyte can be in liquid, gel, and solid form [30]. Liquid electro-lytes are prepared by dissolution of ions in a solvent. This solvent can be an aprotic solvent such as PC and ethylene carbonate (EC) or an ionic liquid such as n-methyl-n-propylpyrrolidinium bis (trifluoromethanesulfonyl) im-ide, a solvent which has chemically bonded anions. Ionic liquid can be thought as a salt and it is liquid at room temperature [72]. Liquid electrolytes favor high ion mobility, and can hence improve the switching dynamic of the ECWs, but they have a leakage problem and create difficulties for large area applications.

An inorganic solid electrolyte consists of small ions in inorganic solids, i.e. LiAlF4, Ta2O5, TiO2, and ZrO2 [67,70,73,74]. ECWs constructed with inorganic solid electrolytes can provide 40-50 % modulation in transmittance [75,76]. Inorganic solid electrolytes were also used as a protective layer in ECWs, which improved the device performance [77].

Polymer electrolytes are the most attractive type of electrolytes for elec-trochromic applications because they provide many different choices for an electrolyte layer [30,50,67,68,78]. Polymer electrolytes provide long open-circuit memory and uniform color changes [79]. For lamination-type manu-facturing, which is based on forming the ECW on two flexible substrates, polymer electrolytes are suitable due to their adjustable viscosity [80]. Low viscous electrolytes can cause a short circuit in the ECW and decrease its lifetime. Highly viscous electrolytes are difficult to apply and may not pro-vide homogeneity. It is possible to have various viscosity with polymer elec-

30

trolytes. Different manufacturing techniques for the fabrication of ECWs are used. One of the methods is based on using a vacuum filling process, which is based on injection of a fluid electrolyte between a millimeter-wide gap separating two half cells [81]. This technique helps the electrolyte to fill into the cavity between two electrodes without forming bubbles. In another tech-nique, roll to roll, the two half cells on two flexible polyester substrates are joined via lamination by an electrolyte having an appropriate viscosity.

The most commonly used polymer electrolytes in electrochromic applica-tions are PMMA, PEO, polypropyleneoxide (PPO), polypropyleneglycol (PPG), polyvinylacetate (PVA), polyacrylonitrile (PAN), and PEG-based electrolytes [50,78]. They can be solid or gel [82,83]. The advantage of solid polymer electrolytes is that they provide electrochemical and mechanical stability, while that of gel polymer electrolytes is to combine fast ionic mo-tion and stiffness [50]. Solid polymer electrolytes can overcome the leakage problem and instability. However, they have low ionic conductivity. Gel-type polymer electrolytes can combine the advantages of both liquid and solid electrolytes, without any leakage, and thermally and mechanically sta-ble good ion conductors can be obtained. By the addition of different salts, such as LiTFSI, HTFSI, NaTFSI, LiClO4, it is possible to find the best com-position for ECWs. It is also possible to make layer-by-layer constructions such as PEI/PEO/polyacrylicacid/PEO polymer electrolytes with Li+ salt, which can combine different properties [84]. Natural polymers such as gela-tin are also attracting attention in electrochromism technology due to their having properties such as biodegradability, low production cost, and good physical and chemical properties [85,86]. Adding inorganic particles such as TiO2 to the polymer electrolytes PEG:LiI can provide long-term stability, high coloration efficiency, or fast switching dynamics in ECWs [87,88]. More detail on polymer electrolytes is given in Section 3-5.

2.2. Coloration mechanism Coloration mechanism, in other words absorption mechanism of electro-chromic films, has not been strictly understood. There are different models to describe the coloration mechanism [89-92]. The most widely accepted explanation of the absorption in electrochromic films is attributed to interva-lence charge transfer, which is an electron transfer between two metal sites. The energy gap between the valence band and the conduction band of the electrochromic materials determines its optical and electrical properties. An electrochromic thin film changes its optical properties by ion insertion into the film that is balanced by electron flow from the transparent conducting film. These electrons lead to intervalence charge transfer, which is an elec-tron transfer between two metal sites [49]. This is the fundamental reason for the absorption, in other words, coloration.

31

WO3 is the most commonly use cathodic metal oxide in ECWs due to its high coloration efficiency and high transparency in the bleached state [3,30,67,93,94]. It is an insulator. Depending on the structural properties such as crystalline, amorphous, polycrystalline, different absorption mecha-nisms exist [30]. Alteration of the transmittance is obtained by the modula-tion of the absorbance for amorphous films [26]. The optical bandgap and refractive index for sputtered amorphous films were reported around 3.01-3.08 eV and 2.3-2.4, respectively [95].

Coloration of WO3 films is based on the turn of W6+ to W5+ [49]. WO3

consists of W6+ and O2- and each W ion is surrounded by six O ions. Each O is bound to two W. The valence band and the conduction band consist of O 2p states and W 5d states, respectively. The Fermi level is positioned in the band gap between the valence band and the conduction band. When an M+ ion such as H+, Li+, or Na+ is inserted into the oxide film, it bonds to the oxygen ion and its outer electron is transferred to W. Thus, the W5+ forms.

6 2 6 6 2 5

bleached colored

W O W W O M (W )+ − + + − + +− − ↔ − −

In the presence of light, the valance electron absorbs a phonon and jumps from site i to a neighboring W site k.

h5 6 6 5i k i kW W W W

ν+ + + ++ → +

It is assumed that the intercalated ions remain ionized in the metal oxide while the intercalated electrons are at the low-energy states of the conduction band. Hence the Fermi level moves up into the conduction band and so ab-sorption in the W oxide film increases and it is colored. The coloration reac-tion of WO3 can be given as

3 x 3

bleached colored

WO xM xe M WO+ −+ + ↔

NiO is commonly used as a complementary anodic electrochromic layer to WO3 [30]. Extensive studies on electrochromic properties of sputtered NiO films can be found in various literatures [64,66,96]. NiO films have high transmittance modulation in strong base electrolytes such as KOH [96]. The modulation is low for Li-based electrolytes [66,96].

Coloration of NiO films is based on the transition between Ni2+ and Ni3+ [49]. Each Ni atom is surrounded by six O atoms. The upper part of its va-lence band consists of occupied Ni 3d states and the lower part of its conduc-tion band consists of empty Ni 3d and admixture of O 2p states. The Fermi

32

level is close to the top of the valence band. Therefore, NiO films can be colored/bleached by movement of the Fermi level. Because of charge inter-calation, the Fermi level moves up to higher energies. Thus, the film be-comes bleached. The film is colored when the Fermi level moves down to lower energies because of deintercaleted charges.

The coloration mechanism reaction of NiO for different cases such as hy-drous phases, non-hydrous phases, and coloration due to the intercalation of H+ or OH-, are summarized in Ref. [96]. The mechanism of NiO based on Li+ insertion/extraction is suggested in two-step [97]. As deposited, NiO is inactive for the ion intercalation/deintercalation. In the first step, NiO be-comes electrochromically active by transforming into LiNiO, which is an irreversible reaction. Then, LiNiO is colored by the charge intercalation as it is described in the second reaction, which is reversible.

y

y y z

bleached bleached

bleached colored

NiO yLi ye Li NiO

Li NiO zLi ze Li NiO

+ −

+ −+

+ + ↔

+ + ↔

IrO2 is an anodic electrochromic material and can be used as an alternative to NiO films in ECWs [98-101]. It is stable in acidic electrolytes. The colored IrO2 film is metallic while the bleached IrO2 film is an insulator [30]. IrO2

structure consists of Ir atoms surrounded by six O atoms. The Fermi level is positioned in the conduction band, which consists of Ir 5d states [102]. The coloration reaction of sputtered deposited films is given below [30].

2 x 2

colored bleached

IrO xM xe M IrO+ −+ + ↔

2.3. Key parameters of electrochromism Switching time is one of the important parameters for ECWs. Switching time is the transition time between colored and bleached states of an ECW. It can also be called response time. For small samples, i.e. 4 cm2, it is of the order of a few seconds, while it is of the order of minutes for large ECWs, i.e. 2 m2. Switching time depends on the applied voltage, ionic conductivity and thickness of the electrolyte, ion diffusion in the electrochromic layers, and the thickness and morphology of the electrochromic films.

In electrochromic applications, a voltage is only applied during the switching. A good ECW should maintain its optical properties in the current state for a long time when no voltage is applied. This is known as open-

33

circuit memory. Electrochromic films such as WO3 can have photochromic properties as well [103]. Therefore, exposure to ultra-violet radiation can affect the open-circuit memory of ECWs. Hence, WO3 films cannot keep their transmittance value, even if there is no applied voltage [104].

ECWs have a set lifetime, as all other devices have. The number of cycles that they survive is known as cycling life and is different for different ECWs. Deterioration of an ECW can occur because of some irreversible reactions. If the charge intercalation and deintercalation are not in balance, the perform-ance of ECW negatively affected. Application of a high voltage or applica-tion of the voltage for a long time are some other reasons for the deteriora-tion of an ECW.

Contrast ratio (CR) is a quantitative measure of the intensity of the color change. For ECWs with negligible reflectance changes, a contrast ratio is defined as the ratio of the transmittance in the bleached state (Tb) to that of the colored state (Tc). It can refer to a specific wavelength λ or to white light.

( ) b

c

TCR

Tλ = (1)

The CR value of 3 to 6 is defined as typical for the luminous or solar proper-ties of ECWs [104].

Optical density (OD), in other words absorbance, is a logarithmic ratio between the incoming light and the transmitted light through an ECW. Sometimes it is defined using the natural logarithm (ln) instead of the com-mon logarithm (log10). It can be written in terms of transmittance as

1( ) ln

=

O DT

λ (2)

where T is the transmittance. Coloration efficiency (CE) is defined as a ratio of the change in

OD (Δ(OD)) to inserted/extracted charge per unit area (ΔQ)

( )ln( )

( ) b cT TODCE

Q Qλ Δ= =

Δ Δ (3)

High coloration efficiency means that a large optical modulation is obtained with a small amount of inserted/extracted charge.

To sum up, several parameters should be taken into account in order to have a properly working ECW. In application, transmittance values in the bleached and colored states, switching speed, cycling time, color, and oper-ating temperature should be considered. Table 2 shows the requirements for large-area ECWs [105].

34

Table 2. Key performance criteria for ECW [105].

Parameters Values

Bleached state solar transmittance 50-70 %

Colored state solar transmittance 10-20 %

Bleached state visible transmittance 50-70 %

Colored state visible transmittance ≤ 10-20 %

Applied voltage (small/large area) 1-3/10-24 V

Open circuit memory 1-12 h

Cycling life 104-106 cycle/ 5-20 year

Operating temperature - 20 to 85°C

2.4. Electrochromic window applications WO3/NiO-based ECWs have been widely studied [70,106-122]. WO3 and NiO are colorless in the bleached state. In the colored state, WO3 absorbs in the red end of the visible range while NiO absorbs in the blue end. So, WO3

is blue and NiO is brown in the colored state, respectively. Therefore, an ECW with their combination has a gray color in the colored state. Figure 6 shows direct transmittance, reflectance, and absorptance of the colored and the bleached states for a WO3- and NiO-based ECW with PET substrates [80].

There are several parameters which affect the performance of an ECW, such as structure and preparation technique of the films, type of ion conduct-ing material, applied voltage, size of the ECW. Transmittance and switching time are fundamental properties determining the electrochromic properties. A summary of previous studies on WO3/NiO-based ECWs is given in Table 3 in terms of size, applied voltage, switching time, and transmittance in colored and bleached states.

IrO2 is an alternative of NiO in electrochromism [120-126]. IrO2 films have a blue-grey color in the colored state and are colorless in the bleached state. They are more expensive and have lower coloration efficiency than NiO films. On the other hand, advantages of IrO2 films are higher chemical stability and shorter swiching time [30,123,127,128]. They do not degrade in water [32]. Since most contribution to coloration comes from the cathodic layer i.e. WO3, less coloration of IrO2 can be too small in amount to be of importance. However, its high chemical stability makes IrO2 attractive for use in an ECW. Previous studies on IrO2/WO3-based ECWs are given in Table 3.

35

0

20

40

60

80 Bleached Colored

Tra

nsm

ittan

ce (

%)

0

10

20

30

Ref

lect

ance

(%

)

500 1000 1500 2000 25000

20

40

60

80

Abs

orba

nce

(%)

Wavelength (nm) Figure 6. Wavelength dependence of transmittance, reflectance, and absorbance for an ECW with configuration PET/ITO/WO3/electrolyte/NiO/PET in bleached (dashed line) and colored (solid line) states. From Ref. [80].

36

Table 3. Literature survey on WO3/NiO- and WO3/IrO2-based ECWs showing con-figuration, applied voltage Vb,c, switching time tb,c, transmittance Tb,c. Subscripts b and c denote bleaching and coloring, respectively. λT is the wavelength that the transmittance was measured. G denotes glass.

Configuration of ECW Size

(cm2)Vb/Vc (V)

tb/tc (s)

Tb/Tc (%)

λT (nm)

Ref.

G/FTO/NiO:PANI/LiCF3SO3:PVC:THF:DOS/WO3/FTO/G 8 1/-1.7 10/60 70/30 450 [106]

G/FTO/NiO/Li:PVB/WO3/FTO/G - 1.5/-2 - 60/10 550 [107]

G/ITO/NiO/Li+ polymer electrolyte/WO3/ITO/G - 1/-1.7 7/24 70/6 630 [108]

G/FTO/LixNiyO/LiTFSI:PMMA:BMITFSI/WO3/FTO/G 25 1.4/-1 - 75/10 550 [109]

G/ITO/NiO/LiClO4:PC-Solaronics SX1170/WO3/ITO/G 25 1.5/-3.5 - 78/12 600 [110]

G/ITO/NiO/Ta2O5/WO3/ITO/G - 5/-5 - 48/16.5 550 [111]

G/ITO/NiO/Ta2O5/WO3/ITO/G - 1.5/-1.5 - 62/8 550 [111]

G/ITO/NiO/LiClO4:PC/WO3/ITO/G - 2.5/-2.5 1.54/1.76 ΔT=84 550 [112]

G/ITO/NiO/solid hybrid electrolyte/WO3/ITO/G 25 2.5/-2.5 20/10 75/20 550 [113]

G/ITO/NiO/PC:LiClO4/WO3/ITO/G 20 1.5/-1 30/50 84/35 550 [114]

G/ITO/NiO/PC:LiClO4:HNO3/WO3/ITO/G 20 1.5/-1 1/22 85/26 550 [114]

G/FTO/NiO/LiClO4:PC based gel electrolyte/WO3/FTO/G 12.5 1.5/-1.5 - 58/6 - [115]

ITO/NiO/Ta2O5/Ti:WO3/ITO/G 1 1.4/-1 11 ΔT=60 550 [116]

G/ITO/Ni(OH)2/Ta2O5/H+-SPE/Ta2O5/WO3/ITO/G - 1/-1.7 30/10 78/24 633 [117]

G/ITO/NiOOH/Ta2O5/H+-SPE/Ta2O5/HWO3/ITO/G - 1/-1.7 30/6 46/24 633 [117]

SOG/ITO/Ni(OH)2/Ta2O5/H+-SPE/Ta2O5/WO3/ITO/SOG - 1/-1.7 30/30 76/74 633 [118]

SOG/ITO/Ni(OH)2/H+-SPE/Ta2O5/WO3/ITO/SOG - 1/-1.7 10/30 74/18 633 [118]

G/ITO/NiO/Ta2O5/WO3/ITO/AF/G 24 1.5/-1.5 - 73/18 vis [70]

G/ITO/NiO/ PMMA:PC: LiClO4/WO3/ITO/G - 1.6/-1.6 - 79/31 600 [119]

G/ITO/NiOxHy/PMMA:PC: LiClO4/WO3/ITO/G - - - 68/40 550 [119]

G/P/ITO/WO3/Ta2O5/IrO2/ITO 30 - - 70/20 - [120]

G/SnO2/WO3/PEO:H+/IrO2/SnO2/G 130 - - 10/48 550 [121]

G/ITO/WO3/PEI:H3PO4/IrO2/ITO/G 70 0.2/-1.2 36/24 8/64 550 [122]

37

3. Polymer electrolytes

In this chapter, polymer electrolytes are introduced. A brief definition of polymers and polymer electrolytes is followed by a description of some characteristics of polymer electrolytes such as conduction and dielectric properties.

Even though people noticed the existence of polymers many centuries later, polymers were always in our lives [129,130]. The human body con-tains many polymers, such as proteins and nucleic acids. Cellulose, wool, silk and amber are some examples of natural polymers. In 1492, the first use of natural polymers could be seen in natural rubber from Hevea brasiliensis, the rubber-tree plant. By the 1840s, scientists began to modify natural poly-mers. The science of polymers began in 1920 with the macromolecular hy-pothesis of Hermann Staudinger. In 1929, synthesis of synthetic polymers, which are man-made polymers, was begun by Wallace Carothers. PEO, PMMA and PAN are some examples of synthetic polymers [131].

The word polymer derives from the combination of two Greek words poly meaning many and mer means parts. Polymers are large molecules com-posed of repeating covalently bonded elementary units called monomers, which are small molecules. Polyethylene (CH2CH2)n, for example, is formed by ethylene monomers CH2CH2. Here, n represents a number of monomer units bonded together. This can be thousands or millions. A polymer chain is a widespread expression of polymer molecules because of the similarity of its form to a chain of monomers. Chain length is generally expressed in terms of the molecular weight, which is related to the molecular mass of monomers. Polymer molecular weight has a strong influence on many physi-cal properties such as viscosity, stiffness, molecular mobility and flexibility. Polymers with low molecular weight have more flexibility, which is a posi-tive feature, but low mechanical strength, which reduces the lifetime of polymeric materials in applications.

Polymers can be formed in different structures such as linear, branched, and crosslinked [130,132]. Different forms affect the physical properties of polymers. A linear polymer, Fig. 7 (a), is composed when monomers are bonded together in a linear form. Linear polymers can have a more packed structure, which makes the polymer more crystalline. A branched polymer, Fig. 7 (b), is formed by a main chain with side chains, which are called branches. In branched polymers, there is more space between the main chains, which makes the polymers less crystalline. In crosslinked polymers,

38

Fig. 7 (c), the main chains are connected with short side chains. Crosslinking decreases the viscosity and flexibility of the polymer but provides more me-chanical strength.

Figure 7. Examples of polymer architecture. Linear (a), branched (b), and crosslinked (c) polymers.

When I worked as an assistant in the semiconductor laboratory during my PhD studies, the most attractive part of laboratory projects, according to the students, was transfering liquid nitrogen from a big container to a small De-war. They transfered it via a rubber hose. The hose became as hard and brit-tle as glass when the liquid nitrogen had gone through into the hose. This part of the project, even if it was not its aim, always surprised the students. Actually, what they observed was a glass transition of the rubber hose. It was soft at room temperature but became hard as a result of the reduction in tem-perature in the presence of liquid nitrogen.

A polymer can be amorphous, crystalline, or contain both amorphous and crystalline regions, such as in semicrystalline polymers [130]. In crystalline polymers or in crystalline parts of semicrystalline polymers, melting occurs. Molecules of crystalline materials are arranged regularly below the melting point and become disordered above it. Glass transition is observed in amor-phous polymers or in amorphous parts of semicrystalline polymers. Glass transition is one of the most important properties of polymeric materials. Glass transition is related to the amorphous regions of polymers and gives information about the mobility of the polymer chains. Glass transition is specific to the glass transition temperature (Tg) which is characteristic for each polymer [129]. Below Tg, polymer chains do not move, and the poly-mer has the characteristics of glass. Polymeric materials become hard and fragile below Tg. Amorphous materials are in a glassy state, and the molecu-lar arrangement is disordered below Tg. Above Tg, thermal energy makes all or some parts of the molecules mobile, and they can move or rotate easily. Therefore, the polymer shows rubbery behavior. The materials are soft at any temperature above Tg. Information on the flexibility of a polymeric ma-terial can be obtained from Tg of the material. For example, when the motion of polymer chains is restricted in any way, the flexibility of the chains is

(a) (b) (c)

39

reduced. The reduction arises as an increment at Tg. Changes in properties such as ionic conductivity or macroscopic viscosity due to changes in the flexibility of the segmental chains can be observed from Tg of polymeric materials. Branching of the polymer makes more free space and reduces Tg. Determination of Tg from differential scanning calorimetry measurements is described Section 10.8.

Polymer electrolytes are ionically conducting materials formed by dis-solving salts in host polymers, i.e, PEO or PMMA with LiClO4, HCF3SO3 or H3PO4 [53]. They are used in electrochromic devices, batteries, fuel cells, super capacitors, sensors and dye-sensitized solar cells in order to provide ion conduction and prohibition of electronic conduction [72,133-136].

Polymer electrolytes were first produced by Peter V. Wright in the 1970s by mixing Na and K salts with PEO, and were developed in the 1990s [137]. A basic polymer electrolyte, which consists of a polymer and a salt, under-lies the fundamental research on polymer electrolytes [138]. In order to im-prove properties such as ionic conductivity, chemical, mechanical, and ther-mal stability, some additives have been incorporated in polymer electrolytes [131,135,138-145]. The most recent historical review and progress of poly-mer electrolytes was reported after the 12th International Symposium of Polymer Electrolytes [72]. Single-ion conductors, which consist of polymer chains that are chemically bonded to the ions are a class of polymer electro-lytes. They could have low conductivity due to the limitation of the ion con-centration. Polymer electrolytes incorporating copolymers or polymer blends were prepared and showed improvement in their mechanical, electrical and thermal stability. Crosslinked polymer electrolytes improve mechanical sta-bility. Adding liquid solvents with high dielectric constant such as PC, EC, diethyl carbonate, and dimethyl carbonate to polymer electrolytes gives an increase in their conductivity and allows the salt content in the electrolytes to increase as well. The most recent and innovative research on polymer elec-trolytes is on the electrolytes containing ionic liquids or inorganic nanoparti-cles. Addition of ionic liquids improves the ionic conductivity and electro-chemical stability of polymer electrolytes [146]. Nanocomposite polymer electrolytes incorporating nanoparticles such as SiO2, TiO2, Al2O3 promote conductivity as well as mechanical and thermal properties [135,147-150]. More details about nanocomposite polymer electrolytes are discussed in Section 4.

3.1. Ion-ion and polymer-ion interactions An ion is an atom or a molecule in which the total number of protons and electrons are not equal. If the ion contains more protons than electrons, it is called a cation. If the opposite is true, and more electrons than protons exit, it is called an anion. When a salt is dissolved in a solvent i.e. polymer, cations

40

and anions are produced. Different types of combinations of the ions may occur when a salt dissolves in an electrolyte. Cations and anions can dissoci-ate and behave independently from each other, Fig. 8 (a). Cations and anions can interact with each other and can form ion-pairs (Fig. 8 (b)), ion-triplets (Fig. 8. (c) and (d)) or higher-order combinations [151]. These different in-teractions depend on salt concentration. If more salt is added, conductivity is expected to increase because of the increase in the number of free ions in the solvent. On the other hand, the increase of salt concentration causes a reduc-tion of the distance between dissolved ions. Therefore, the number of the ion-pairs may increase and the conductivity decrease.

Figure 8. Schematic illustration of free ions (a), ion-pair (b), ion-triplets (c-d).

In addition to ion-ion interactions, the ions can interact with the polymer chains. While a strong interaction can occur between cations and polar units of the polymer, it is much weaker between anions and the chains [152]. The strong interaction between cations and polymer chains can form transient crosslinks, Fig. 9, causing a reduction in the mobility of polymer chains, and thereby ionic conductivity of the electrolytes.

Figure 9. Formation of transient crosslink by a cation (a) and by a triple ion (b) in PEO based polymer electrolytes [153].

Adding salt influences the flexibility hence Tg, of the host polymer. In poly-mer electrolytes, the presence of a salt, i.e. lithium salt, increases the amor-phous regions and hence Tg decreases. In transient crosslink cases, the mo-tion of the polymer chains is restricted and Tg increases. Incorporation of some solvents, so-called plasticizers such as PC, dilutes the polymer due to the presence of short chains, hence the chains move more freely, and Tg de-

CH2 Oxygen Cation Anion

(a) (b)

+ _ _ + _

+ + _ _

(a) (b) (c) (d)

+

41

creases. Any filler such as nanoparticles may decrease Tg because it reduces the transient crosslink between the polymer chains.

3.2. Ionic conductivity In this section, a general view of conduction mechanisms is given via expla-nations of ion transport mechanisms and temperature dependence of ionic conductivity. Then, conduction mechanisms and ionic conductivity as a function of temperature, conductivity spectrum for polymer electrolytes and some examples of ionic conductivity values for various polymer electrolytes are described.

In electrochromic applications, polymer electrolytes play a significant role due to the importance of ionic motion. Therefore, understanding the conduction mechanism is crucial. Electrical conductivity is a measure of a change of the position of the charges as a result of an applied electric field. If the moving charges are electrons or holes, it is described as electronic con-ductivity. Movement of the ions is described as ionic conductivity.

Conductivity is a function of the electron charge q, the number of charge the carriers Nc, and the mobility µ and can be given as [154].

i i icq Nσ μ= (4)

where i represents different charge carrier species.

Figure 10. Schematic illustration of different ion transport mechanisms: Solid state (a), free-volume (b), and the dynamic bond percolation (c) theory [154].

d E>E0

Free volume

Distance (a) (b) (c)

Energy

E0

42

Ion transport has been explained in different ways [154]. Figure 10 shows a schematic illustration of different ion transport mechanisms. In solid-state, Fig. 10 (a), at absolute zero temperature, ions are trapped in their energy sites where the potential wells are formed by the surrounding atoms. When the temperature increases ions jump between the sites. Temperature depend-ence of ionic conductivity (σion) for this type of ion transport mechanism is in Arrhenius form

0 exp a

iont

E

R Tσ

γ

σ σ −

= −

(5)

where σ0 is the pre-exponential factor, Rγ is the gas constant, Ea-σ is the acti-vation energy, and Tt is absolute temperature. The frequency of the jumps is proportional with the activation energy [154]. Therefore, high mobility can be provided with low activation energy.

If ionic conduction is provided, by hopping of the ions from one state to another state, they need free spaces. Thus, it is also important if electrolytes are in a crystalline or amorphous phase. While ions need certain channels in the crystalline form, conduction is much higher in the amorphous state. For this reason, amorphous polymer electrolytes are much more preferable.

Another way to explain ion transport in polymer electrolytes is based on the free-volume theory which ignores thermally activated ion-hopping [155]. Due to the disordered structure of polymer electrolytes, they contain several vacancies. Ion transport based on the free volume theory is illustrated in Fig. 10 (b). The vacancies are randomly arranged below the glass transition temperature. Capability of localized motion of the chains results in a redis-tribution of the free volumes above the glass transition temperature. Thus, ions diffuse through the structure. The Vogel-Fulcher-Tammann (VFT) equation is the most commonly used equation that represents the temperature dependence of the ionic conductivity for the ion transport based on the free volume theory [154,156-158].

0

0

expiont

B

T Tσσ σ

= − −

(6)

where Bσ is constant related with activation energy of the conduction, and T0 is a temperature related to the glass transition temperature [159].

The most general explanation for ion transport is based on the dynamic bond percolation theory and is shown in Fig. 10 (c) [152]. It assumes a solid-state-type jumping of ions between the sites which occasionally connect together. In the situations based on the dynamic bond percolation theory, temperature dependence of ionic conductivity can be either the Arrhenius or

43

the VFT-type. Arrhenius behavior is shown if the fraction of the percolated sites is large and VFT behavior is shown if the fraction is small and the polymer motion is according to the free-volume theory [154].

Conduction mechanisms in polymer electrolytes were outlined by Vincent [139]:

• Cations attached to polymer chains and transport by the motion of poly-

mer segments. • Ions jump between suitable low energy sites. • The polymer structure provides suitable pathways for ions to move.

Figure 11. Schematic illustration of ionic motion in a PEO:LiX polymer electrolyte [131].

PEO is the most attractive polymer that has been used for research based on polymer electrolytes [131,140,160]. Figure 11 shows a schematic illustration of ionic motion for PEO:LiX electrolytes. Li+ ions are bonded to four ether oxygen atoms of PEO and conduction in PEO based electrolytes with Li+ ions is attributed to the movement of the Li+ ions along with the PEO chains [131,160].

In general, the Arrhenius relation attributes the ionic motion to ion hop-ping in polymer electrolytes. VFT implies that segmental motion of the polymer predominantly assists the motion of ions. Different behavior has been observed in the temperature dependence of ionic conductivity of poly-mer electrolytes. Ratner summarized this as follows [152]:

• Arrhenius behavior • Arrhenius behavior with two different activation energies • VFT behavior • Arrhenius behavior at low temperatures and VFT behavior at high tem-

peratures.

Li Li

O

O

O

O

O

OOO

OO

O O

O

O

44

• VFT behavior above Tg, but Arrhenius behavior at much higher tempera-tures than Tg. Temperature range from Tg to Tg+100K is given as a valid-ity range of VFT behavior [161].

Figure 12 shows a typical conductivity spectrum of a polymer electrolyte. An increase in the conductivity at high frequency corresponds to bulk relaxa-tion of the electrolyte. A decrease at low frequency range is due to the elec-trode polarization. The wide plateau at the middle frequency range shows the ion conduction of the polymer electrolyte. Thus, ionic conductivity of the polymer electrolyte can be obtained from the plateau of the conductivity spectrum.

10-2 10-1 100 101 102 103 104 105 106 10710-7

10-6

10-5

10-4

Frequency (Hz)

Con

duct

ivit

y (S

/cm

) σionic

Figure 12. Conductivity spectrum of a polymer electrolyte. The red line shows the intercept of the plateau.

In order to obtain a good performance with an electrolyte, it should have high ionic conductivity, between 10-7 and 10-4 S/cm, and at the same time low electronic conductivity, preferably below 10-12 S/cm [162]. Ionic con-ductivities of various polymer electrolytes at room temperature are given in Table 4.

Table 4. Examples of polymer electrolytes and their room temperature ionic conduc-tivities σion.

Polymer electrolyte σion (S/cm) Ref.

PPO:LiClO4 10-8 [143]

PEO:LiClO4 10-8 [143]

PEO:LiClO4:PC:EC 10-3 [143]

PMMA:PPG:LiClO4 2x10-6 [138]

PEG :LiCF3SO3:SiO2 1.5x10-3 [143]

PAN:EC:PC:LiClO4 1.7x10-3 [131]

PEO:LiBF4:Al2O3 10-4 [143]

45

3.3. Molar conductivity In this section, molar conductivity, salt concentration dependence of molar conductivity for polymer electrolytes and the Walden rule are explained.

If conduction mechanism is examined for different salt concentrations, it is preferable to normalize the ionic conductivity with the ion concentration. Molar conductivity (Λ) is defined as conductivity per ions in one mole, Eq. 7.

ion

saltc

σΛ = (7)

where saltc is the molar concentration given in the unit of mol/L. A typical salt concentration dependence of molar conductivity is shown in Fig. 13 for polymer electrolytes [153,163]. The molar conductivity first decreases with salt concentration, then increases, and then decreases again.

The decrease in the first region is explained by the formation of ion-pairs with increasing salt concentration [164]. In the second region, the behavior of the molar conductivity is attributed to the increased number of dissociated ions because of the decreasing number of ion-pairs due to the redissociation of ion-pairs or formation of triplet ions [164]. In the third region, the de-crease of the molar conductivity is based on the reduction of the segmental mobility of the polymer chains with increasing salt concentration. The possi-bility of forming transient crosslinks increases with increasing salt concen-tration; then ionic mobility is reduced.

Figure 13. Salt concentration dependence of molar conductivity for polymer electro-lytes.

The influence of the concentration and the mobility of the ions on the con-duction mechanism can be examined via a relation between molar conduc-tivity and viscosity, which is the Walden rule. The Stokes-Einstein relation gives the relation between the diffusion coefficient (D) and viscosity of a polymer (ηp) as

Salt concentration

Mol

ar c

ondu

ctiv

ity I II III

46

6

t

A p

R TD

N aγ

π η= (8)

where NA is the Avogadro constant, a is the radius of ions. On the other hand, the Nernst-Einstein equation relates the diffusion coef-

ficient to the molar conductivity at infinite dilution (Λ0)

( )2

0

( )A

t

z N qD D

R Tγ+ −Λ = + (9)

where z is valances of ions, +D and −D diffusion coefficients of cations and anions, respectively.

Using Eq. 8 and 9, the relation between the viscosity of the polymer and the molar conductivity at infinite dilution of the electrolyte can be obtained as

2

0

1

6A

P

zq N

aπ ηΛ = (10)

If the radius of the ions can be considered the same in electrolytes,

2

0 constant=6

AP

zq N

aη

πΛ = (11)

This relation is known as the Walden rule and says that the product of the molar conductivity at infinite dilution and the viscosity of the pure solvent is constant at various temperatures in a given solvent, or for different solvents at a given temperature [165,166].

3.4. Dielectric spectroscopy Dielectric spectroscopy allows studying molecular motions on various length scales. Details of dielectric spectroscopy are given in Section 7. In this sec-tion, dielectric spectrum for polymeric materials and the importance of di-electric spectroscopy analysis for polymer electrolytes are introduced.

In amorphous polymeric materials, various relaxations depending on the length scales can be observed [167]. Figure 14 shows some possible relaxa-tion processes and length scales of amorphous polymeric materials. Individ-ual bond rotations result in local conformational changes. The local motion, Fig. 14 (a), is seen in both amorphous and crystalline cases. Above the glass transition temperature, cooperative movement of a few monomer units re-

47

sults in segmental motion, Fig. 14 (b). Relaxation related to the movement of the whole polymer chain is known as the chain relaxation, Fig. 14 (c). Re-laxation time increases with increase of the length scale.

Figure 14. Relation between polymer dynamics and dielectric relaxations [168].

In Fig. 15, dielectric spectra in terms of real and imaginary permittivity are shown for a polymeric material. At low frequencies, electrode polarization is seen. In the imaginary part of the permittivity, it is seen that ionic conductiv-ity appears as a linear relation with the frequency with a negative slope equal to unity. The αc-relaxation represents the chain relaxation and/or ionic re-laxation. The α-relaxation is related to the segmental motion of the polymer chains. The β-relaxation is associated with the local motions of the chains.

Figure 15. Schematic dielectric spectrum of polymeric materials [169].

Dielectric spectroscopy has been widely used for many polymer electrolytes such as PMMA- and PEO-based electrolytes [170]. Segmental flexibility of an amorphous polymer electrolyte can play a significant role in the ion transport. The permittivity analysis is one characterization technique used to investigate segmental flexibility [171]. Ion-pairs are surrounded by the vis-cous polymer and move in it. The frequency where the maximum of imagi-nary permittivity is obtained for the α-relaxation is related to the local vis-cosity of the polymer, hence the segmental flexibility [171]. The dielectric constant of polymer electrolytes increases with increasing salt concentration due to the increase in ion-pairs with large dipole moments [171]. Distance

α

log(

real

per

mitt

ivity

)

log(frequency) log(frequency)

β αc

electrode polarization

αcα β

ionic conduction

log(

imag

inar

y pe

rmitt

ivity

)

electrode polarization

local motion segmental motion chain motion

(b) (a) (c)

48

separation between the cation and anion of the ion-pairs can be calculated from the dipole moment, see Eq. 12 [171].

It is known that the dipole moment of two opposite charges (µd) depends upon the distance (l) between them.

=d qlμ (12)

where q is the carrier charge. Therefore, if we calculate the average dipole moment a

dμ of electrolytes, we may obtain information about the ion-pairs. The Onsager equation, Eq. 13, can be used to calculate the average dipole moment of ion-pairs from the dielectric constant of electrolytes [172].

1/ 2

02

9 ( )(2 )

( 2)a B t b bd

T b

k T

N

ε ε ε ε εμε ε

∞ ∞

∞

− += + (13)

where kB is the Boltzmann constant, εb is the bulk permittivity of the polymer electrolyte, ε∞ is the high frequency permittivity of the electrolyte, NT is the number of salt molecules per unit volume.

49

4. Nanocomposite polymer electrolytes

Nanoparticles are small particles that can have different physical properties than a bulk material. Nowadays, polymeric nanocomposites that are com-posed of inorganic nanoparticles and polymeric material are of interest be-cause the addition of nanoparticles enhances properties or adds new func-tionalities. Nanoparticles can provide better ionic conduction in polymeric materials. Combination of polymers with nanoparticles can result in thermal stability in a wider temperature ranges. The presence of nanoparticles in a polymeric material can increase the mechanical strength of the material. Nanoparticles can modify the optical properties of polymer-based media [173-176].

4.1. Optical properties Polymers are highly transparent materials with a refractive index close to 1.5 in the visible range [177,178]. Figure 16 shows characteristic transmittance spectra for various polymers [179]. They have high transmittance in the visi-ble range and some minima in the near-infrared range.

Figure 16. Transmittance spectra of various polymers in the visible and near-infrared range [179].

50

4.1.1. Rayleigh scattering

Particles in a polymer matrix can cause a change in the direction of incoming light, which is called scattering. Scattering from small-sized particles can be explained by Rayleigh scattering, which depends on the wavelength and the particle size [180]. In Rayleigh scattering, the strong wavelength dependence of the scattering shows that the shorter wavelengths are scattered more strongly than the longer wavelengths. Rayleigh scattering is defined by the Rayleigh scattering coefficient RR [180]

25 6 2

4 2

2 1

3 2n NP

RNP

N NR

N

π δλ

−= +

(14)

where nN is the number of particles per unit volume, δ is the diameter of the particle, NPN is the ratio of the refractive index of the particles to that of the polymer.

4.1.2. Near-infrared absorption Nanocomposite polymers can yield spectrally selective absorption without causing any scattering of light. Near-infrared radiation (700-3000 nm) of the electromagnetic spectrum carries 50 % of the solar energy [181]. The solar radiation in the near-infrared range contributes to the heating of a building if it comes through a window. Blocking the near-infrared radiation and having high transmittance in the visible range can be obtained by nanocomposite polymeric systems where nanoparticles are embedded in a polymer matrix. If a nanoparticle has a negative real dielectric permittivity in the near-infrared region, an absorption resonance occurs in that region. The position, width and strength of this resonance can be controlled by varying the free electron density in the nanoparticles [182].

ITO, antimony tin oxide, and LaB6 are the materials that can block the near-infrared radiation [183]. The absorption of the nanoparticle forms of these is much more effective than their bulk form due to the high absorption strength of the nanoparticles [184,185]. As an example, Fig. 17 (a) and (b) show optical spectra of an ITO thin film and ITO nanoparticles embedded in PVB, respectively [183,186]. The transmittance is about 80 % for the ITO thin film. However, ITO nanoparticles dispersed in PVB have almost zero transmittance above 1700 nm.

51

Figure 17. Spectral transmittance computed from a theoretical model for ITO on glass with 0.2 µm thickness and 1x1020 cm-3 electron density (a) [186] and modelled for a nanocomposite of 0.2 wt% ITO dispersed in PVB between two glass substrates (b) [183].

4.1.3. Maxwell-Garnet theory Optical properties of polymer nanocomposites can be described by the Maxwell-Garnett effective medium theory. It is assumed that nanoparticles have much smaller diameter than the wavelength and do not interact with each other [187,188]. The effective dielectric permittivity εMG is

m

21

3 1

1 – 3

f polMG

f pol

F

F

αε ε

α

+= (15)

where εm is the dielectric permittivity of the polymer and Ff is the filling factor, i.e., the volume fraction of the nanoparticles, and αpol is the polariza-bility of the nanoparticles, respectively. For randomly distributed ellipsoidal particles, the latter can be written

3

1

1

3 ( )=

−=

+ − p mpol

i m i p mL

ε εα

ε ε ε (16)

where εP is the dielectric function of the nanoparticles and the Lis are depo-larization factors given by the shape of the nanoparticles and equal to 1/3 for spheres.

0

60

100

300 500 1000

20

40

(b)

Tra

nsm

itta

nce

(%)

2500Wavelength (nm)

ITOGlass

80

60

40

20

0300 700 1100 1500 1900 2300

Tra

nsm

itta

nce

(%)

Wavelength (nm)

PVB+ITO

Glass

Glass

(a)80

100

52

4.1.4. Ligand-exchange method Homogeneous dispersion of nanoparticles in a polymeric medium is critical in order to get optically clear samples. However, aggregation is one of the crucial problems of nanocomposites and there are different methods to over-come this problem [189,190]. Surface modification is a very promising treatment to provide well-dispersed nanoparticles. For example, fumed silica is SiO2 with hydroxyl groups on the surfaces. Hydroxyl groups can interact with both cation and anion and give improved adsorption properties and it has also been reported that it improved the conduction properties of polymer electrolytes [191,192].

Figure 18. Schematic illustration of the ligand-exchange process with NOBF4. The blue color represents the solution with nanoparticles. The nanoparticles transfer from an apolar hydrophobic solution to a polar hydrophilic solution after the ligand ex-change.

The ligand-exchange method is a surface modification technique. A ligand is defined as an ion or molecule that binds to a central metal atom. A ligand exchange is a surface treatment based on a phase transfer process in which one ligand on the surface of nanoparticles is replaced by a different one. It can be applicable to various nanoparticles with different shapes and size such as metal oxides e.g. TiO2, metal alloys e.g. FePt, semiconductors e.g. Bi2S3 and dielectrics e.g. NaYF4 [193]. Figure 18 shows a schematic illustra-tion of the ligand exchange method. Organic ligands such as oleic acid (OLAC) and oleylamine (OLAM), which provide dispersion of nanoparticles in nonpolar and hydrophobic solvents such as hexane, are replaced with in-organic ligands such as BF4

-. Thus, dispersion of nanoparticles in a polar and hydrophilic solvent such as dimethylformamide (DMF) can be provided [193].

+ BF4-

apolar hydrophobic

polar hydrophilic

OLAM OLAC

BF4-

53

4.2. Ionic conductivity The enhancement of the ionic conductivity for nanocomposite polymer elec-trolytes is attributed to different mechanisms such as a phase transition from crystalline to amorphous, an increase in the mobility of polymer chains, in the ionic mobility, and in the number of the free ions.

The increase in the ionic conductivity for crystalline polymer electrolytes is explained by the decrease of the crystalline phase. The addition of nanoparticles prohibits recrystallization of the electrolytes and promotes the amorphous regions [194]. Thus, the enhancement in ion transport results in the enhancement of the ionic conductivity. Effective medium theory and experimental results confirm that the increase in the ionic conductivity is due to the formation of an amorphous polymer phase at the polymer-nanoparticle interface, which is favorable for the ions [195].

Another explanation for the increase in the conductivity is attributed to an increase in the segmental flexibility of the polymer electrolyte [147,196]. The nanoparticles interact with the cation, hence the number of transient crosslinks is reduced [147]. Thus, segmental flexibility increases and con-ductivity increases. Dispersion of silane, a modified SiO2, in PEO:LiClO4 makes the amorphous phase of the electrolyte more flexible; hence the ionic conductivity increases [148].

The influence of the nanoparticles on the mobility of the ions is another reason for the increase in the conductivity [196]. The effect of the different surface groups on the conductivity was investigated for PEO:LiTFSI:Al2O3 polymer electrolytes [197]. The enhancement on the conductivity is related to the enhanced ionic mobility. The interaction of the surface groups of the nanoparticles with the anions and cations creates additional sites, which are favorable for the migration of the ions. Acidic surface groups increase the anionic contribution to the conductivity while basic surface groups increase the cationic contribution. A neutral surface group has an equal number of acidic and basic sites and they interact with each other. Therefore, the con-ductivity enhancement is smaller for the neutral surface groups compared to the electrolytes that have the nanoparticles with the other types of surface groups.

An increase in the number of the free ions can also be expressed as a rea-son for the increase in conductivity [194,198-200]. It is attributed either to the separation of the cations from the polymer or to dissociation of the ion-pairs. Nanoparticles can make much stronger bonds with the polymer than Li+ ions [194, 198-201]. This leads to the separation of Li+ ions from the polymer chain; hence the Li+ ions contribute to the conduction. Another explanation for the increase in the free ion concentration is based on the strong interaction between the surface group of nanoparticles and anions. OH- surface groups of fumed silica preferentially adsorb anions rather than cations [197,202]. It was reported that the addition of fumed silica gives

54

improved adsorption properties and higher conductivity for PEO:NaI poly-mer electrolytes [192]. The effect of fumed silica on the ionic conductivity of PEO:LiTFSI and PEO:LiCLO4 electrolytes was examined and an increase in conductivity obtained [148]. It was also shown that the surface groups of the fumed silica adsorb anions, which leads to increased ion-pair dissociation for soggy sand electrolytes [202].

55

5. PEI-based electrolytes

In this chapter, the attractive properties of PEI and LiTFSI, which led us to use a PEI:LiTFSI composition are explained. A literature survey of PEI-based electrolytes is presented.

Various polymers have been used as hosts to form electrolytes. Various polymer-salt combinations have differences in solubility. Dissolution condi-tions are (i) a low lattice energy of the salt, (ii) a low cohensive energy of the polymer, (iii) high solvation energy, (iv) high permittivity of polymers, (v) a high donor number, and (vi) a high acceptor number [154]. Polar groups of a solvent approve the separation of ions as a result of its high permittivity [154].

PEI can be in a linear or branched form. A schematic representation of the polymerization of PEI is shown in Fig. 19. Linear PEI is the nitrogen analogue of PEO, which is commonly used in electrochromic applications. Nitrogen is a better donor atom than oxygen, which is likely to form a com-plex with salts [203]. It has been suggested that PEI has reasonably favorable polar groups along the chain. This is an important feature because ions are dissolved in the presence of polar groups. Thus, PEI is good at dissolving various salts because of its structural considerations. Another advantage of PEI is that it has an amorphous phase for a wide temperature range, which increases conductivity [203]. Branched PEI is superior to linear PEI because

Figure 19. Polymerization of PEI [205,206].

Polymerization

Linear PEI

Branched PEI

Ethylene imine

56

of its ability to make homogeneous amorphous salt complexes at low salt concentrations [204]. With these aspects in mind, PEI seemed an attractive option to study for electrochromic applications.

The small size of Li+ ions lets them diffuse easily into polymers. Li+ ions have a radius of between 59 pm and 74 pm, depending upon the coordination number [207]. It was shown that LiTFSI gives low viscosity for electrolytes [208]. It is also known that LiTFSI improves the conductivity of the electro-lyte by increasing the amorphous domain range [209]. It was noted that LiTFSI has a slightly stronger plasticizing effect than LiClO4 for PEO-based electrolytes [148]. In electrolytes, much higher conductivity and more free Li ions are obtained in the presence of LiTFSI than LiClO4, LiCF3SO3, and LiPF6 [210-214].

The formation of an electrolyte with PEI is easy because of its suitable viscosity, which makes it unnecessary to use plasticizers. A number of PEI-based electrolytes have already been investigated [164,215-233]. PEI is gen-erally used in proton-conducting polymer electrolytes [215-219]. The most common explanation for the conduction mechanism in PEI-based protonic electrolytes is attributed to the Grotthuss mechanism, which is based on pro-tons hopping between the amine groups. The hopping does not include the transport of the mass of the molecule. Hence, high conductivity is obtained. Nanocomposite polymer electrolytes incorporating PEI and silica were stud-ied [219]. The addition of 12 nm-diameter silica to PEI:H3PO4 increases the conductivity of the electrolyte. The enhancement of the conductivity is at-tributed to the bonding of protons to the surface of the silica and hence the formation of a new conducting way at the surface of the silica nanoparticles.

Table 5. Examples of PEI based electrolytes with their Tg and σion at room tempera-ture.

Polymer electrolyte Tg (°C) σion (S/cm) Ref.

PEI:LiCl -16 10-8 [220]

PEI:LiBr -15 10-8 [220]

PEI:LiI -15 10-8 [220]

PEI:LiSCN -30 10-8 [220]

PEI:LiBF4 -23 10-8 [220]

PEI:LiCF3SO3 - 10-6 [221]

PEI: LiCF3SO3:DMF -30 10-3 [222]

PEI:TEG:LiCF3SO3:digylme - 10-4 [223]

PEI-G2: LiCF3SO3 -76 10-6 [224]

PEI:PEO:LiClO4 - 10-6 [225]

57

There are studies on Li+-conducting PEI-based electrolytes [220-225]. Some examples of PEI:LiX-based electrolytes with their ionic conductivities at room temperature and glass transition temperatures are given in Table 5. For pure PEI:LiX electrolytes, ionic conductivity at room temperature is 10-6-10-8 S/cm; and glass transition temperature is below -30 °C. Ion-ion and ion-polymer interactions were investigated for PEI:LiX-based electrolytes [226-228]. PEI chains interact with anions; hence the anion mobility and consequently the ionic conductivity decrease [138,228]. A study on PEI:LiCF3SO3 electrolytes reports that free ions, ion-pairs, and triple cation are all dominant in the Li+-based electrolyte [226]. Another study shows an increase in cation-anion aggregation with increasing salt concentration, but no difference with increasing temperature for PEI:LiCF3SO3 electrolytes [227].

In the literature, there are a few studies on electrochromic applications of PEI-based electrolytes [229-233]. ECWs with a configuration of glass/SnO2/WO3/electrolyte/SnO2/glass were examined for PEI-electrolytes containing H2SO4, H3PO4, LiCF3SO3 [229]. It was reported that acid-based electrolytes are cycled above 10000 cycles without any visual deterioration. It was noted that the Li+-conducting ECW needs 5 V for good optical modu-lation but there are no results based on this ECW. ECWs of glass/SnO2/PAN/electrolyte/ion-tungstate/SnO2 /glass with an area of 6.5 cm2 were tested for 30000 cycles with PEI electrolytes containing pro-tonic acids and it was concluded that color changes are quite similar, but more pronounced with H3PO4 [230]. ECWs with the configuration of PET/ITO/WO3/polymer electrolyte/fluorindine/ITO/PET showed transmit-tance modulation between 60 and 25 % at 550 nm in less than 30 s by apply-ing -1 V for coloring and 2 V for bleaching [231]. ECWs with a configura-tion of glass/ITO/WO3/PEI:H3PO4/polydimethyl acrylamide:polystyrene sulfonicacid/ITO/glass was performed between -1 and 1 V and transmittance changes between 75 and 25 % at 550 nm [232]. A study on ECWs with a configuration of ITO/PANI/PEI:H2SO4/ITO shows that the ECW with pH 2-3 PEI:H2SO4 has the fastest response compared to other pH values [233].

59

6. The electromagnetic spectrum

Electromagnetic radiation consists of electric field E and magnetic field

B ,

components that are perpendicular to each other and also to the propagation direction of the wave. Propagation of radiation is described by Maxwell’s equations.

0 00

0 B E

E B E B jt t

ζ μ εε

∂ ∂∇⋅ = ∇⋅ = ∇× = − ∇× = + ∂ ∂

(17.a.b.c.d)

where ε0 and µ0 is the permittivity and permeability of vacuum, respectively, and ζ and

j are charge and current densities, respectively.

Table 6. Approximate wavelength, frequency, and energy limits of the various re-gions of the electromagnetic spectrum [234].

Type of wave Wavelength (m) Frequency (Hz) Energy (J)

Radio > 1x10-1 < 3x109 < 2x10-24

Microwave 1x10-3 - 1x10-1 3x109 - 3x1011 2x10-24 - 2x10-22

Infrared 7x10-7 - 1x10-3 3x1011 - 4x1014 2x10-22 - 3x10-19

Optical 4x10-7 - 7x10-7 4x1014 - 7x5x1014 3x10-19 - 5x10-19

UV 1x10-8 - 4x10-7 7x5x1014 - 3x1016 5x10-19 - 2x10-17

X-ray 1x10-11 - 1x10-8 3x1016 - 3x1019 2x10-17 - 2x10-14

Gamma-ray < 1x10-11 > 3x1019 > 2x10-14

The electromagnetic spectrum is the distribution of electromagnetic radiation according to energy, wavelength, or frequency of the waves. It extends from short wavelength gamma rays to long wavelenth radio waves as shown in Table 6. When a material is subjected to electromagnetic radiation, depend-ing on the wavelength range of the radiation, interaction between the radia-tion and the material shows different behavior at different wavelengths. De-pending on the energy of the radiation, different techniques can be used to investigate the properties of the material. For example, infrared radiation corresponds to the vibration energy of the material. Therefore, infrared spec-troscopy can be used to get information about the vibrations of the material. Radiation of ultraviolet-visible energy changes the configuration of the va-

60

lence electrons. Therefore, ultraviolet-visible spectroscopy can be used in order to elucidate electronic transitions. Microwave-radio range, or wave-lengths above it, corresponds to molecular rotation, oscillation, and the di-pole moment that can be examined by the dielectric/impedance spectroscopy (DIS) technique.

61

7. Dielectric/impedance spectroscopy

7.1. Basic principles Materials have randomly distributed charges. If a material is put between two electrodes and is subjected to an electric field, positive and negative charges in the material move towards the surface of the electrodes and form dipoles. This is known as polarization. The DIS examines the polarization due to the charge migration and the orientation of permanent dipoles [235]. The DIS gives information about the electrical properties of various materi-als from the response of materials to an applied electric field. It is used to examine a material over a wide frequency range from 10-6 Hz to 1012 Hz. Basically, the DIS measures a current passing through a sample when ac voltage is applied to the sample. Capacitance, impedance, and admittance are the quantities that can be obtained from a DIS measurement. Thus, electrical and dielectric properties of a material such as conductivity, relaxation and permittivity can be obtained.

When ac voltage, V, is applied to a sample, Fig. 20 (a) and Eq. 18 (a), there can be a phase shift, ϕ , in the current , passing through the sample, Fig. 20 (b) and Eq. 18 (b).

Figure 20. Applied ac voltage V and the result current I passing through the sample.

V0(t)

t

t

V(t)

I(t)

I0(t)

62

( )0 0( ) ( )i t i tV t V e I t I eω ω ϕ+= = (18.a.b)

where V0=(V'2+V''2)1/2 and I0=(I'2+I''2)1/2 where V' and I' are real and V'' and I'' are imaginary parts of the voltage and current, respectively, ω is angular frequency, ω=2πf where f is the measurement frequency.

By a Fourier transform, the time domain data is converted to frequency domain data.

0

2( ) ( ) i tI I t e dt

φωω

φ= (19)

Ohm’s Law says that the ratio of dc voltage to the current is the resistance of the system. Impedance can be seen as the ac equivalent of resistance. The ratio of the applied ac voltage to the current is called impedance (Z).

0

( )( )

( )i

Z

VZ Z e R iZ

Iϕωω

ω− ′′= = = − (20)

where Rz is ordinary resistance and Z'' is the reactance. The ratio between Rz and –Z'' is known as the loss factor

tan( ) =′′−

ZR

Zδ (21)

where the “–” is for a sign conversion due to choosing the notation ie ϕ− . Another term that is used to characterize the electrical properties of a ma-

terial is admittance Y, that is, the ratio of the current to the voltage.

( )

( )( )

IY Y iY

V

ωωω

′ ′′= − = (22)

where Y' is conductance and Y'' is susceptance. Capacitance is a way to store an electrical charge and is defined as

( )

( )( )

IC

i V

ωωω ω

= − (23)

Actually, impedance, admittance, and capacitance are different presentations of the same measurement. All quantities can be transformed from one to another.

63

1 1

ZY i Cω

= = − (24)

The conductivity of a material with thickness d and cross-sectional area Aν

is given from the conductance as

d

YAν

σ ′= (25)

When a sample is put between two parallel plates with area Aν, the system behaves as a capacitor under the effect of an electric field E

. If E

is applied

between two parallel plates, which are electrodes, with distance d between them, the potential difference between the plates is easily obtained from

0

.d

V E dl Ed= − =

(26)

An empty cell capacitance is given by

0 = dQ

CdV

(27)

where Q is the induced electrical charge on the surfaces. The electric field between the plates can also be given by means of the

Gauss law:

0

QE

Aν ε= (28)

From the equations above, the capacitance of the empty cell is obtained as:

00

AC

dν ε

= (29)

When you put a sample between the plates, the electric displacement of a polarized material is

0 0 0(1 )= + = + =

eD E P E Eε ε χ ε ε (30)

where P is polarization, χe is susceptibility and ε is the dielectric permittiv-

ity of the medium.

64

E and therefore V are reduced by the 1/ε factor when the empty cell is filled with a dielectric material. Thus, the capacitance of the full system be-comes

0=C Cε (31)

The permittivity is a complex quantity

( ) iε ω ε ε′ ′′= − (32)

The real part defines the ordinary permittivity, which is a measure of the ability to store energy, and the imaginary parts represent the dielectric loss component due to the ionic conduction and dipole relaxation in the material. The contribution of the conductivity to the permittivity follows a power law

1

0

−′′ = ionσε ωε

(33)

When an electric field is applied, cations move towards the negative elec-trode while anions move towards the positive electrode. When they reach the electrodes, they accumulate there and hence charge layers form close to the electrode surfaces, as seen in Fig. 21. This is known as electrode polarization and it affects the low frequency measurements [236].

Figure 21. Electrode polarization. When

E is applied between two electrodes,

cations moves towards the negative electrode and anions move towards the positive electrode (a). Ions accumulate near the electrode surfaces (b).

The frequency limit depends on the actual system (material and thickness...). Electrode polarization increases the dielectric constant and causes a strong drop in conductivity towards low frequencies. It can be eliminated by special measurement techniques such as four electrode methods or measurements with different distances between the electrodes [237-239]. Contribution of the electrode polarization to the spectra can also be determined by modeling

Cation Anion

+ -

V

+ -

V

E

E

65

the spectra with an equivalent circuit, which contains some circuit elements representing the electrode polarization [240-244].

7.2. Equivalent circuit models The response of materials to an applied electric field can be represented with an equivalent circuit consisting of electrical circuit elements such as resis-tors, capacitors and distributed circuit elements. Various equivalent circuits can be used in order to represent the response of a material to an electrical field. In order to find a suitable model to describe the response of the mate-rial, the physical meaning behind the model should be considered. In an equivalent circuit, the presence of each element has a physical meaning.

All the elements can be expressed in their impedance form. Thus, the total impedance of the equivalent circuit (Ztot) will be the sum of the complex impedances of the elements as

1 2 .....= + + +tot iZ Z Z Z (34)

when the elements are in series, and as

1 2

1 1 1 1.....= + + +

tot iZ Z Z Z (35)

when the elements are in parallel. Here, Zi represents the impedance of ith circuit element in an equivalent circuit.

Resistance represents opposition to the current flow. Impedance expres-sion of a resistor is Z=RZ-0i, which means that the phase shift in the current is zero. So, the current and potential are at the same phase. Resistance is independent of frequency.

Capacitance represents the electric field storage capacity [235]. The im-pedance expression is Z=0-i/ωC which does not have a real component and the current is ahead of the potential with a 90 degree phase shift. Capaci-tance is a function of the frequency. At high frequencies, impedance of ca-pacitance goes towards zero while at low frequencies it goes towards infin-ity.

Figure 22 gives some examples of equivalent circuits representing a di-electric response [235]. A parallel connection of Rz and C, Fig. 22 (a), can be used for the study of the charge movements. This model does not include any discrimination of the contribution of different charge species. A serial connection of Rz and C, Fig. 22 (b), describes a response with a single re-laxation time. In Fig. 22 (c), Cg represents geometrical capacity, Rb denotes

66

bulk resistance, Re and Ce are electrode/electrolyte interfacial resistance and capacitance, respectively [245].

Figure 22. Examples of equivalent circuits representing a dielectric response. An equivalent circuit consisting of (a) a parallel connection of Rz and C, (b) a serial connection of Rz and C, and (c) a parallel connection of geometrical capacity Cg with a serial connection of bulk resistance Rb and parallel connection of electrode / elec-trolyte interfacial resistance Re and capacitance of that Ce [235,245].

In reality, ideal elements, such as C and RZ, are not enough to represent a real material. For example, in polymer electrolytes, each ion has a different environment, hence they have different mobilities. These kinds of complex systems can be represented by distributed circuit elements. There are several distribution functions modeling the dielectric response of a complex material [246-251]. The Havriliak-Negami (HN) function is one of the most widely used distributed circuit elements in equivalent circuits of complex materials [249]. The HN function can be either in the dielectric or the conductive form. The dielectric HN (DHN, subscript D) function is given as

( )(1 ( ) )D DU P

D

CC

iω

ωτΔ=

+ (36)

where ΔC is the difference between capacitance values at low (ω<<1/τ) and high (ω>>1/τ) frequencies, τD is the relaxation time, and UD and PD are power-law exponents and describe the symmetric and asymmetric broaden-ing of the relaxation time distribution [252].

A conductive HN (CHN, subscript C) function is given as

( )(1 ( ) )C C

CU P

C

RZ

iω

ωτ=

+ (37)

where RC is the ionic resistance and the other symbols are introduced in analogy with those in Eq. 36.

RZ C RZ

C

Rb

Ce

Re

Cg

67

The difference between the two HN functions is that the DHN function represents only the ion-pair relaxation while the CHN function represents the ionic conductivity and ion-pair relaxation in a single process. Thus, it is pos-sible to analyze the response either by separated ionic conductivity and ion-pair relaxation (the DHN response) or by combining the ionic conductivity and ion-pair relaxation in a single process (the CHN response).

The HN function is reduced to Cole-Cole [247], Cole-Davidsson [248], and Debye [246] responses when PD,C = 1, UD,C = 1, and PD,C = UD,C = 1, respectively. Real and imaginary capacitance spectra of these relaxations are given in Fig. 23. The Debye function implies a single relaxation. However, in many cases, systems are more complicated and they have a much-broadened dielectric loss process in comparison to the Debye [252]. This type of relaxation can be explained by the Havriliak-Negami response func-tion.

Figure 23. Real (C') and imaginary (C'') capacitance spectra of the Havriliak-Negami, Cole-Cole, Cole-Davidsson, and Debye responses [253].

7.3. Time-temperature superposition principle In many cases, the form of the relaxation follows the time-temperature su-perposition principle [254]. It implies that the frequency response of a mate-rial exhibits the same shape at different temperatures. After a series of meas-urements versus frequency at different temperatures, by shifting of real and imaginary parts of a spectrum, it can be deduced whether they have the same shape or not. Permittivity data taken at different temperatures are related to

log (frequency)

C'

C''

log (frequency)

C'

C''

log (frequency)

C'

C''

Havriliak-Negami Cole-Davidson

Cole-Cole

log (frequency)

C'

C''Debye

Rea

l and

imag

inar

y ca

paci

tanc

e R

eal a

nd im

agin

ary

capa

cita

nce

Rea

l and

imag

inar

y

capa

cita

nce

Rea

l and

imag

inar

y

capa

cita

nce

68

one another if the curves overlap. Thus, one can say that the form of the ac response remains invariant under temperature. The principle generalizes the behavior of the response at a certain temperature to the responses at the other temperatures.

7.4. The Barton-Nakajima-Namikawa relation

An empirical relation, which is called the Barton-Nakajima-Namikawa (BNN) relation, holds if there is a connection between ionic conductivity and dielectric relaxation [255-257]. It is given by the equation:

0 max= Δion BNNFσ ε εω (38)

where FBNN is a numerical factor of the order of unity for ionic conductors, Δε is the dielectric relaxation strength, and ωmax=2πfmax with fmax is the re-laxation peak frequency corresponding to the location of the maximum in the dielectric loss. The validity of the BNN relation has been shown for ion con-ducting glasses, ion conducting solids, and disordered solids [258-261].

69

8. Optical spectroscopy

Optical spectroscopy examines the interaction between the material and elec-tromagnetic radiation in the ultraviolet, visible, and near-infrared range. Op-tical properties are placed at a critical point in energy-window-related stud-ies. Different materials have different optical properties and they can reflect, absorb, or transmit radiation at different levels. These materials can be used in energy efficient windows to provide desired properties [262,263]. For example, if you live in a warm climate, the use of air conditioners is inevita-ble. A coated window with a material that has the property to block the transfer of near-infrared solar radiation from outside to inside can prevent too much heat flow inwards. In this way, the need to use the air conditioner is reduced. In our study, optical properties of the polymer electrolytes were investigated in the visible and near-infrared range.

8.1. Blackbody spectrum An object that absorbs all radiation is called a blackbody. At a temperature above absolute zero all materials emit electromagnetic radiation to some degree. Thermal radiation from a material can be determined by multiplying the blackbody spectrum and emittance of a material. Therefore, understand-ing blackbody radiation provides understanding of the thermal radiation of materials.

Blackbody spectrum is known as the Planck spectrum. According to Plank’s law, the electromagnetic radiation emitted by a blackbody in thermal equilibrium at a definite temperature is given by

2

5

2 1( )

1S B t

SS hc k T

c hI

e λπλλ

= − (39)

where Is is the spectral radiance of the surface of the blackbody, h is the Planck constant, and cs is the speed of light.

The Stefan-Boltzmann Law states that the blackbody radiance, the total energy radiated by a surface per unit area is proportional to the fourth power of the temperature of the blackbody.

70

4BB SB tP Tσ= (40)

where PBB is given in W/m2, σSB is the Stefan–Boltzmann’s constant. The position of the peak can be determined by Wien’s displacement law

max 2898 .tT m Kλ μ= (41)

In Fig. 24, the spectra of blackbody radiation are given at different tempera-tures. The intensity increases and the peak shifts towards shorter wavelength when the temperature increases.

Real materials do not absorb all the radiation and they emit less energy than an ideal blackbody. Radiance of a material is given by

M m BBP E P= (42)

Em is the emissivity of the material which is less than unity.

Figure 24. Blackbody radiation at shown temperatures (a), extraterrestrial solar radiation (b), and human eye sensitivity (c) [262].

0.10.080.060.040.02

0

100°C50°C0°C

-50°C

Blackbody radiation

Solar radiation (extraterrestrial)

Relative sensity of human eye

Wavelength (μm)

GW/m3

GW/m321.61.20.80.4

01

0.80.60.40.2

0 0.2 0.5 1 2 5 10 20 50 10

(a)

(b)

(c)

71

8.2. The solar spectrum Even if there is no ideal blackbody, the spectrum of the sun is very similar to the blackbody radiation, which is shown in Fig. 24 (a) at different tempera-tures. In Fig. 24 (b), the extraterrestrial solar radiation is shown. The solar spectrum is limited to the 250-3000 nm wavelength range. Therefore, it does not overlap with the thermal radiation (blackbody radiation) spectra at around room temperature. The solar spectrum outside the atmosphere de-fines the emission spectrum of the sun as a result of its surface temperature that is 6000 K. The solar spectrum inside the atmosphere depends on the travelled path, interactions of the radiation with molecules in the atmos-phere, position of the sun, and weather conditions. Due to the interaction of the radiation with carbon dioxide, ozone, and water vapour in the air, i.e. absorption by these gases, some minima exist in the solar radiation spectrum in addition to the shown curve in Fig. 24 (b).

The sensitivity of the human eye is illustrated in Fig 24 (c), with a bell-shaped curve that extends in the range of 400-700 nm wavelength range and has a peak at 550 nm. Electromagnetic radiation in this range of wavelengths can be called visible light or light. The colors of visible light are shown in Fig. 25. There is no sharp transition between the colors and 550 nm where the curve of sensitivity of the human eye has a peak in Fig. 24 (c) corre-sponds to the green range in Fig 25.

Figure 25. Visible light region of the electromagnetic spectrum [264].

8.3. Basic principles 8.3.1. Interaction between light and matter When a material is subjected to light, the material can partially or totally transmit, reflect and absorb the light depending on the optical characteristics

600700 500Wavelength (nm)

800 400 300

Ora

nge

Yel

low

Red

Lem

on

Gre

en

Tur

quoi

se

Blu

e

Indi

go

Vio

let

Infrared Ultraviolet

72

of the material. Transmittance T, reflectance R and absorbance A are defined as the ratio of the transmitted (IT), reflected (IR) and absorbed (IA) light inten-sity to the intensity of the incoming light (Ii).

( ) ( ) ( )T R A

i i i

I I IT R A

I I Iλ λ λ= = = (43.a.b.c).

Based on energy conversation, the sum of T, R and A at each wavelength is equal to unity.

( ) ( ) ( ) 1+ + =T R Aλ λ λ (44)

Figure 26. Interaction of light and material. When an incident light hits a material, the light can transmit and reflect from the material, and be absorbed by the material. Transmitted and reflected light can have diffuse components.

Figure 26 shows the interaction between light and a material. Transmitted light can be parallel to the incident light and is called direct transmittance also defined as regular transmittance. If the reflected light has the same angle as the incident light with the normal of the surface of the material, it is called specular reflectance. In addition, the transmitted or reflected light can spread in many directions, which constitute the diffuse parts of the transmittance and the reflectance. If a material contains some components that can cause scattering of the light, diffuse transmittance and reflectance can make a sig-nificant contribution to the total transmittance and reflectance. Total trans-mittance Ttot and reflectance Rtot are the sum of direct Tdirect and diffuse Tdiff transmittance and specular Rspec and diffuse Rdiff reflectance, respectively.

= + = +tot dir diff tot spec diffT T T R R R (45.a.b)

8.3.2. Solar and luminous transmittance Luminous transmittance is the ratio between transmitted radiant (or lumi-nous) flux and incident radiant flux. It is a measure of how much the incident

Specular reflectance

Diffuse reflectance

Diffuse transmittance

Specular transmittance

73

light illuminates a surface. Solar transmittance is the ratio between the solar energy passed through the material (in the full solar wavelength range 300-2500 nm) and the incoming solar energy. Solar and luminous transmittances are obtained by

( ) ( )

( )lum,sol

lum,sollum,sol

d TT

d

=

λ ϕ λ λλ ϕ λ

(46)

where φlum is the luminous sensitivity of the human eye and φsol is solar ir-radiance.

8.3.3. Refractive index Optical properties of materials, in another words, how light propagates through that medium, can be described by the complex refractive index N

( ) = +N n ikλ (47)

where n is the refractive index and k is the extinction coefficient. The refrac-tive index affects the speed and direction of light when it passes through the material. The extinction coefficient is related to the amount of absorption loss. Intensity of the light, Il, is reduced when it interacts with the material and the reduction (dIl) is given as

l ldI I dlα= − (48)

where α=4πk/λ is the absorption coefficient of the material that is dependent on the material and the wavelength. Hence, the intensity of the light inter-acted with material can be defined as

e dl iI I α−= (49)

The complex refractive index is related to the complex permittivity accord-ing to

N ε= (50)

For a weak absorbing film, the extinction coefficient can be obtained from the transmittance and reflectance of the film [265]

74

1 4

expR kd

T

πλ

− =

(51)

Eq. 51 can be written for glass

1 4

expG G G

G

R k d

T

πλ

− =

(52)

and for a glass/polymer/glass configuration with the assumption that there is no reflection at the interfaces of the glass and the polymer,

1 4 24

expGPG G GP P

GPG

R k dk d

T

ππλ λ

− = +

(53)

where the subscripts G, GPG, and P denotes glass, glass/polymer/glass, and polymer, respectively.

Dividing Eq. 53 by Eq. 52, we can get the extinction coefficient of a polymer with a reflectance that is very close to that of glass as

4

ln4

G G GP

P GPG

T k dk

d T

πλπ λ

= −

(54)

500 1000 1500 2000 25000.0

0.5

1.0

1.5

2.0

Wavelength (nm)

Ext

inct

ion

coef

fici

ent (

x 1

03 )

Figure 27. Wavelength dependence of extinction coefficient of PEI calculated by Eq. 54 from measured transmittance and reflectance of the PEI polymer between two panes of glass.

Figure 27 shows the extinction coefficient of PEI as a function of wave-length, which was used in the calculations of transmittance of PEI:LiTFSI:ITO electrolytes in Paper VI.

75

8.3.4. Fresnel equations

Figure 28. Transmission and reflection of incident light passing through one me-dium with reflactive index N1 to another medium with refractive index N2.

Transmittance and reflectance are related to the refractive index by the Fres-nel equations. For a single interface as shown in Fig. 28, where the light passes through from one medium with refractive index N1 to another me-dium with refractive index N2, the Fresnel equations are

1 1 1 1

1 2 2 1 1 1 2 2

2 1 1 2 1 1 2 2

1 2 2 1 1 1 2 2

2 cos 2 cos

cos cos cos cos

cos cos cos cos

cos cos cos cos

p s

p s

N Nt t

N N N N

N N N Nr r

N N N N

θ θθ θ θ θ

θ θ θ θθ θ θ θ

= =+ +

− −= =+ +

(55.a.b.c.d)

where tp,s and rp,s are interface amplitude transmittance and reflectance, re-spectively, p and s subscripts denote the type of polarization of light, θ1 and θ2 are angle of incident and transmitted light, respectively. From tp,s and rp,s, transmittance and reflectance can be obtained as

2 22 2

, ,1 1

cos R=

cos p s p s

NT t r

N

θθ

= (56.a.b)

N1

N2

Incident θ1 θ1

θ2

Reflected

z

Transmitted

77

9. Viscosity

Viscosity is a measure of the fluidity of a material. It is a measure of the internal friction of a fluid and shows resistance of the material to flow. In this study the term viscosity corresponds to shear viscosity. The viscosity is a macroscopic parameter and as a first approximation it can be considered as the local viscosity [171]. Thus, from the viscosity one can get information about the structural properties of a polymeric material. It is affected by the size and shape of the molecules, the entanglement of the polymer chains, composition of the material, and concentration [266]. Some other properties such as molecular weight, activation energy, pressure, refractive index, and chemical composition can be related to viscosity [267,268]. Due to the rela-tion of viscosity to the structural properties of a polymer, viscosity analysis plays a significant role in the investigation of the conduction mechanism of polymer electrolytes, specifically the mobility of ions.

Figure 29. Movement of two parallel planes of a fluid with the velocities ν1 and ν2

Newton defined viscosity based on the internal friction that occurs as a result of the relative movement of a layer according to another layer of the fluid. Newton’s model on which the definition of the viscosity is based is given in Fig. 29. Two parallel layers of the fluid with the same area Aν are separated by a distance dl [269,270]. When a force F is applied to one of the layers, there will be a difference in the speed of the two layers dν= ν2-ν1. The force per unit area F/Aν is related to the interaction between the molecules and proportional to the velocity gradient dν/dl. The proportionality constant is defined as the viscosity, η.

ν2 Aν

dl ν1 Aν

F

78

v

F dv

A dlη= (57)

The velocity gradient is called the shear rate ( γ ). Since the force per unit area is required to produce the shearing action, it is called shear stress (τν). So, viscosity can be written as

=vτη

γ (58)

Some models are used to describe the shear rate dependence of the shear stress of fluids [271,272]. Fluid behavior that has not a constant viscosity with changing shear rate can be modelled with the Herschel-Bulkley model.

0ξτ τ ηγ= + (59)

where τ0 is yield stress and ξ is a power-law exponent which can give information about the type of the fluid.

The Herschel-Bulkley model is reduced to the Power-law model in the case of τ0=0, or to the Bingham plastic model when ξ=1.

9.1. Classifications of fluids Fluids can be classified depending on their relation between shear stress and shear rate [273,274]. The classification is done in two main groups, Newto-nian and non-Newtonian. If a fluid at a constant temperature and pressure, i.e. water, has a constant viscosity at all shear rates, it is referred to as New-tonian. A fluid showing a deviation from such behavior is called a non-Newtonian fluid. The viscosity of a non-Newtonian fluid changes when shear rate varies. Non-Newtonian fluids can be classified as time-dependent and time-independent, according to the time dependence of the viscosity.

The behavior of the shear stress and viscosity versus shear rate for various fluids without time dependency is briefly given in Fig. 30. Imagine that you want to eat your pasta with ketchup. If you do not squeeze the bottle, you cannot pour it on the pasta. It refuses to flow. This is a typical ideal Bingham plastic fluid. This type of fluid behaves like a solid under static conditions. To start the flow, a certain amount of force needs to be applied to the fluid. This force is called yield stress. When it reaches the value of yield, the flow begins. It corresponds to a starting point at zero for a shear stress-shear rate graph. The viscosity of fluids, such as paint, decreases while the shear rate increases. This is called pseudoplastic flow. Solutions of large, polymeric

79

Figure 30. Shear rate dependence of shear stress for different types of time-independent fluid behavior.

molecules in a solvent with smaller molecules can be given as an example of pseudoplastic materials [275]. The viscosity of a fluid such as cornstarch in water increases with the increment of shear rate. That is known as dilatant flow.

Figure 31. Schematic representations of the behavior of thixotropic and rheopectic fluids [130].

If a fluid has variable viscosity with time under conditions of constant shear rate, this is a time-dependent flow. The time dependence of non-Newtonian fluids can be classified as thixotropic and rheopectic. A thixotropic fluid shows a decrease in viscosity with time at constant shear rate while a rheopectic fluid has an increasing viscosity at the same conditions, as shown in Fig. 31. In the time-dependent case, a hysteresis area is observed when the shear rate increases and then decreases. The area provides information about the thixotropic or rheopectic properties of the fluid.

Thixotropic

Rheopectic

Shear rate

She

ar s

tres

s

Shear rate

Bingham ideal plastic Pseudoplastic

Newtonian

Dilatant

She

ar s

tres

s

80

9.2. Temperature dependence of viscosity The viscosity of materials changes with temperature. The changes vary from material to material. Some material has high changes in the viscosity with temperature while others have less. Several equations can be used to repre-sent the temperature dependence of viscosity [267]. The most frequently used expressions are the Arrhenius equation, Eq. 60 (a) and the VFT equa-tion, Eq. 60 (b).

0 exp( )a

t

E

R Tη

γ

η η −= 00

exp( )t

B

T Tηη η=

− (60.a.b)

where η0 is the pre-exponential factor, Ea-η is the activation energy of viscos-ity, Bη is proportional to the activation energy. The Arrhenius and the VFT behaviors are shown in Fig. 32.

Figure 32. Temperature dependence of logarithmic viscosity.

1/Temperature

Arrhenius VFT

log

(Vis

cosi

ty)

81

10 Experimental

10.1. Materials Branched PEI (H(NHCH2CH2)nNH2) with three different average molecular weights (1200 g/mol (50 wt% water, Aldrich), 10000 g/mol (purity 99 %, Alfa-Aesar), and 60000 g/mol (50 wt% water, Aldrich)) was used in the study. LiTFSI (LiN(CF3SO2)2, 99.95 %), anhydrous PC (C4H6O3, 99.7 %), and fumed silica with a particle size of 7 nm and a surface area of 390 m2/g were supplied by Sigma-Aldrich. EC (C3H4O3, >99 %) and anhydrous methanol (CH3OH, 99.8 %) were provided by Fluka and Merck, respec-tively.

PEIs were dried at 65 °C at 10-1 mbar for 48 h before use. LiTFSI was used after drying at 150 °C at 10-1 mbar for 72 h. Fumed silica, EC, and PC were used as-received. All materials were stored in a MBraun UNIlab argon-filled glove box with <1 ppm water.

In2O3 and ITO nanoparticles were chemically prepared. Materials used for the synthesis of In2O3 and ITO nanoparticles were indium acetylacetonate (In(acac)3, In(OCCH3CHOCCH3)3, 99.99 %, Aldrich), tin (IV) bis (acety-lacetonate) dichloride (Sn(acac)2Cl2, [CH3COCH=C(O-)CH3]2SnCl2, 98 %, Aldrich), OLAM (CH3(CH2)7CH=CH(CH2)7CH2NH2, Acros), OLAC (CH3(CH2)7CH=CH(CH2)7COOH, 90 %, Aldrich), nitrosonium tetra-fluoroborate (NOBF4, 97 %, Acros), and anhydrous DMF (HCON(CH3)2, 99.8 %, Sigma-Aldrich). The In2O3 and ITO nanoparticles were prepared under nitrogen flow and dispersed in DMF. The solutions were kept in the glove box until they were used.

10.2. Electrolyte preparation 10.2.1. Preparation of PEI:LiTFSI electrolytes The pure PEI:LiTFSI electrolytes were prepared at different [N]:[Li] molar ratios. The [N]:[Li] ratio was defined as in Eq. 61. The number of moles of the PEI repeating units was obtained by dividing the mass of used PEI by the molecular weight of these units (43.0 g/mol). Instead of molar ratio, some-times we used the term salt concentration, which is the number of the moles of solute (LiTFSI) per total mass of the solution (electrolyte) in units of

82

mol/kg. The relation between the molar ratios and the salt concentrations are given in Table 7.

( )2 2the number of moles of the PEI repeating units CH CH NHN : Li

the number of moles of LiTFSI salt

− −= (61)

Table 7. Molar ratios and corresponding salt concentrations of PEI:LiTFSI electrolytes.

[N]:[Li] c (mol/kg)

400 0.06

200 0.11

100 0.22

50 0.41

20 0.87

For the formation of PEI:LITFSI polymer electrolyte, LiTFSI was added to the methanol and stirred. After the salt was dissolved in methanol, PEI was added to the solution and mixed. Solutions were kept for 48 h at 65 °C at 10-1 mbar to remove the methanol. The prepared electrolytes have the ap-pearance of highly viscous liquid. All samples were stored in the glove box.

10.2.2. Preparation of nanocomposite PEI:LiTFSI electrolytes Preparation of PEI:LiTFSI-based nanocomposite polymer electrolytes was the same as the preparation of the pure PEI:LiTFSI electrolytes. SiO2 was added before the addition of PEI to the solution for the formation of SiO2-doped PEI:LiTFSI electrolytes. For the preparation of In2O3 and ITO added PEI:LiTFSI electrolytes, the nanoparticle solution was added after PEI be-cause the addition of the nanoparticle solutions before PEI made the solution opaque. The nanocomposite polymer electrolytes were obtained by adding the nanoparticles at different wt% of the total weight of the nanoparticles and PEI.

10.2.3. Preparation of PEI:LiTFSI:PC:EC-based electrolytes PEI:LiTFSI:PC:EC were prepared at different weight percentages of the components. EC was mixed with PC at 65 °C until it was dissolved. LiTFSI was added to the PC:EC and stirred until it was dissolved in the solution. For the preparation of SiO2-added PEI:LiTFSI:PC:EC electrolytes, SiO2 and PEI were added sequentially to LiTFSI:PC:EC. For ITO-doped PEI:LiTFSI:PC:EC, PEI and then ITO-DMF solution were added sequen-

83

tially to the LiTFSI:PC:EC solution. All materials were mixed at 65 °C until homogenous solutions were obtained.

10.3. Synthesis of In2O3 and ITO nanoparticles Crystalline In2O3 and ITO nanoparticles were prepared in collaboration with E. L. Runnerstrom and D. J. Milliron at the Molecular Foundry of the LBNL. Synthesis of the nanoparticles was carried out using standard Schlenk techniques in nitrogen atmosphere, Fig. 33 [276]. In2O3 nanoparti-cles were prepared by mixing 0.692 mmol In(acac)3 and 8.741 mmol OLAM. The solution was heated up to 250 °C for 5 hours. To remove the supernatant, the particles were washed in ethanol by centrifugation at 9000 rpm. The diameter of the particles was 8 ± 3 nm.

Figure 33. Synthesis of ITO nanoparticles using the Schlenk technique. The picture was taken at the LBNL.

Synthesis of the ITO nanoparticle was the same as the In2O3 preparation process. 0.0364 mmol of Sn(acac)2Cl2 was added together with In(acac)3 and OLAM. Spherical ITO nanoparticles were prepared with 5 atom % of Sn and diameters of 13 ± 3 nm. The transmission electron microscope (TEM) im-ages of In2O3 and ITO nanoparticles are shown in Fig. 34 (a) and (b), respec-tively.

84

Figure 34. TEM images, supplied from the LBNL, of In2O3 (a) and ITO (b) nanopar-ticles.

The In2O3 and ITO nanoparticles dispersed in hexane were shipped to the Ångström Laboratory. Here, 20 µL OLAM and 20 µL OLAC were added to the solution. Thus, the particles were washed with ethanol twice more at 5300 rpm for 10 min. PTFE (polytetrafluoroethylene) filters were used for the filtration of the solutions to eliminate oversized particles.

To be able to disperse and stabilize the nanoparticles in a polar and hy-drophilic solvent such as DMF, we performed ligand exchange by changing organic OLAM and OLAC ligands by inorganic anion BF4

− [193,277]: The same volumes of the solvent of DMF and the solution of ITO, with OLAM ligands, (0.114 mg/mL) in hexane, were mixed. Thus, a two-phase mixture was obtained. NOBF4 was added to the mixture and it was vigorously stirred for 5 min. In this way, OLAM ligands on ITO particles were changed to BF4

− and ITO nanoparticles were transferred to the polar phase. The particles were washed in toluene by centrifugation at 5300 rpm for 10 min for further purification. It has been shown that the ligand exchange method does not change the shape and size of the nanocrystals and they have no aggregation after the surface modification treatment [193].

10.4. Electrochromic film deposition Amorphous thin films of WO3 and V-NiO were supplied by ChromoGenics. The films were prepared by reactive dc magnetron sputtering in an industrial drum coater. The thin films were deposited on ITO-precoated flexible PET substrates. The deposition was done in an argon/oxygen atmosphere with 12000 W total discharge power. Ni targets are magnetic materials. For tech-nical reasons, it is good to use non-magnetic targets [49]. Therefore, non-magnetic Ni(93 %)–V(7 %) alloy was used as the sputter target. The optical and electrochromic properties of NiO and V-NiO thin films are close to each other [49].

85

IrO2 thin films were prepared by reactive dc magnetron sputtering on a Balzers UTT 400 unit. The total sputter pressure was 30 mTorr, the flow ratio O2/Ar was 1; the sputtering power was 200 W and the film thickness was 85 nm. More details about the deposition of the IrO2 films are given in Refs. [127].

10.5. Formation of electrochromic window prototypes The ECWs were produced in a cleanroom. The electrochromic-film-coated substrates were laminated together with the electrolytes. The configuration of the devices were PET/ITO/WO3/electrolyte/V-NiO/ITO/PET and PET/ITO/WO3/electrolyte/IrO2/ITO/PET. The edges of the devices were sealed and contacted. The final devices had effective areas of 25 cm2.

10.6. Dielectric/impedance spectroscopy measurements In this section, a DIS measurement technique in the frequency range 10 µHz to 10 MHz is presented. A schematic picture of a basic setup for the dielec-tric spectroscopy measurement is shown in Fig. 35. A generator applies ac voltage V with a fixed frequency ω/2π applied between the plates. The am-plitude I and the phase angle φ are measured by the instrument. Thus, all electrical and dielectric properties can be obtained for the sample, i.e. Z(ω )=V(ω )/I(ω ).

Figure 35. A basic setup for dielectric spectroscopy measurements. It consist of an a.c. voltage generator, a voltmeter, an ampermeter, and a sample cell.

In our study dielectric measurements were performed at different tempera-tures on a Novocontrol BDC-N dielectric interface combined with a So-lartron 1260 frequency response analyzer [278,279]. The frequency was swept over 48 points between 10-2 and 107 Hz. The temperature was kept constant during each measurement.

The advantages of this system are that (i) dielectric samples with low conductivity can be measured with high precision even at low frequencies, (ii) phase accuracy for small tan( )δ is good, and (iii) a reference measure-

V

I

Voltage generator

Sample cell

Voltmeter

Ampermeter

86

ment technique is used. Capacitance and conductivity measurement ranges of the setup are shown in Fig 36.

Figure 36. Measurement limits for the Novocontrol BDC-N & Solartron 1260 fre-quency response analyzer system [278].

Fundamental components of the setup are a generator, an amplifier, a current to voltage converter and a variable reference capacitor. A ac voltage is cre-ated by the generator. The amplifier increases the amplitude of the electrical signal. The current to voltage converter has a capacitance and a resistor and converts the current to voltage. Impedance due to the leads and effects of the internal BDC electronics are eliminated by reference measurements. After each measured voltage, the current to voltage converter switches to the vari-able reference capacitor and the reference measurements are done.

In the sample cell, two parallel stainless steel plates were used as blocking electrodes, which do not allow penetration of the ions into the electrodes. The measurement cell is shown in Fig. 37. It consisted of a Teflon insulator, a spring, and two big electrodes with a diameter of 40 mm, a sample cell and a heater. The sample cell consisted of two stainless steel electrodes separated by a ring-shaped Teflon spacer with an outer diameter of 20 mm, 1 mm ra-dial thickness and 3.1 mm height. The Teflon spacer kept the thickness of the electrolyte constant and prevented a leak of the electrolytes. The electro-lytes were poured on the bottom electrode of the sample cell. The sample cell was put between the big electrodes of the measurement cell, shown in yellow in Fig. 37. A metal box was used for shielding from possible radia-tion from the environment. The temperature was changed by applying dc voltage to the heater and it was read from a thermocouple thermometer. All impedance measurements were done in the glove box in an argon atmos-phere.

87

Figure 37. Side view of the measurement cell of the dielectric measurements.

The sample cell is shown in Fig. 38 (a). Measured capacitance (Cmeas) can be considered to be formed by a parallel capacitor of the electrolyte Celectrolyte, the spacer CTeflon, additional capacitance due to the electrical stray fields from the borders of the sample capacitor Cedge and stray capacity due to non ideal shielding of the electrode connection wires Cwire as shown in Fig. 38 (b). The ring shape of the sample cell reduces the additional capacity due to the electrical stray fields from the borders of the cell. Large electrode diame-ter and small electrode thickness can minimize the errors due to the edge effects.

meas electrolyte Teflon edge wireC C C C C= + + + (62)

Figure 38. Sample cell consists of two stainless steel electrodes that were separated with a Teflon spacer (a) and capacitance equivalent to total capacitance of the system (b).

In order to get sample capacitance an empty cell measurement was run. When the empty cell was measured, the measured capacitance Cem was the sum of the capacitance of air C0, CTeflon, Cedge, and Cwire.

Celectrolyte CTeflon Cedge Cwire d

Spring

Big electrodes

Sample cell

Teflon insulator

Metal box

Heater

88

0em Teflon edge wireC C C C C= + + + (63)

Thus, using Eq. 62 and 63, the capacitance of the electrolytes was obtained from the DIS measurement of the electrolytes and the empty cell

0( ) ( ) ( )electrolyte meas emC C C Cω ω ω= − + (64)

Fig. 39 (a) and (b) shows real and imaginary parts of capacitance of an empty cell measurement, respectively. The imaginary part of the capacity was low, so was ignored.

10-3 10-1 101 103 105 107

10-11

10-10

Rea

l cap

acit

ance

(F

)

Frequency (Hz)

10-16

10-15

10-14

10-13

10-12

10-11

Imag

inar

y ca

paci

tanc

e (F

)

Figure 39. Real and imaginary capacitance of the empty cell.

10.7. Thermogravimetric analysis The history of thermal analysis extends many years back [280]. First re-ported thermal analysis experiments were performed on lead, tin, zinc and various alloys by Jakob Fredrik Emanuel Rudberg in Uppsala in 1829 [280]. Today, there are several techniques that determine thermal properties of materials by measuring different properties such as weight, deformation, temperature difference, and heat flow with high sensitivity. Thermogravim-etric analysis (TGA) and differential scanning calorimetry (DSC) are two of the most frequently used thermal analysis techniques.

Different polymers have different thermal stability. TGA is a technique that measures weight changes of a sample as a function of time and tempera-ture. It is used in order to determine the thermal stability range, components of a material, loss of any component, and the purity of the sample and ab-sorbed moisture. Figure 40 shows the basic elements of a thermogravimetric analyzer. It consists of a sample pan, a balance, a furnace, and a temperature

89

control system in a controlled atmosphere. The sample pan stands in the furnace. The furnace is heated or cooled in a controlled manner during the experiment. The sample is weighed simultaneously during heating or cool-ing. A gas flows over the sample and controls the sample environment. The reference pan is used as a reference for the weight calculation. The system is controlled by means of a computer.

Figure 40. Experimental setup of TGA measurement.

Schematic TGA curves for different polymers are given in Fig. 41. Even if many TGA curves look similar they are characteristic for each polymer. The curves differ from each other depending on the molecular structure of the polymer. Temperature-dependent reactions occur at specific temperature ranges and heating rates depend on the polymer.

Figure 41. Schematic curve of temperature dependence of weight percent during TGA measurement for different polymers.

Some components in ECWs do not like water and the presence of water can reduce the long term stability of the ECW. To obtain better efficiency in electrochromic devices, the device system should be constructed with water-free materials. Therefore, we investigated the water content of the electro-lytes. A TGA Q500 thermogravimetric analyzer was used for the measure-ments. Samples were put in platinum pans and measurements were done

Temperature

Wei

ght (

%)

0

100

Balance

Reference pan

Sample pan

FurnaceGas Gas

90

between the temperatures 20 and 600 °C with an accuracy of 1 °C and the heating rate was 10 °C/min.

10.8. Differential scanning calorimetry The DSC technique is based on the enthalpy change that occurs when any changes such as phase transition, degradation, or decomposition occur in materials as a function of time or temperature. Some properties such as glass transition, crystallization, melting, degradation, and decomposition tempera-tures can be obtained from DSC analysis. The DSC technique can be appli-cable to many materials such as polymeric, organic, and inorganic materials, catalysts, and clays.

Figure 42. Schematic representation of the DSC measurement cell.

A schematic picture of a DSC cell is given in Fig. 42. The main elements of the setup are sample and reference cells, a heat flux plate, furnace, thermo-couples, and a sensor. Sample and reference crucibles are placed on a heat flux plate. Heat transfer occurs between the sample and the reference and also from the furnace to them. The heat plate transfers the heat from the wall of the furnace to the sample and reference crucibles. Temperatures of the sample and reference are measured below the crucibles. Heat flow to the sample is determined by comparing to that of the reference.

A schematic DSC curve that shows phase transitions of a polymeric mate-rial is given in Fig. 43. The heat flow to the sample is given as a function of temperature. The step in the baseline at low temperature corresponds to the glass transition. Any factors that affect the molecular motion of the polymer chains give rise to a shift in the baseline of the DSC curve. The exothermic and endothermic peaks correspond to crystallization and melting of the polymer, respectively.

Sample crucible

Referencecrucible

Thermocouple wires

Furnace

Sensor (thermal resistance)

Heat flux plate

91

Figure 43. Schematic illustration of a DSC curve for polymeric materials.

In this study, DSC measurements were used to determine the glass transition temperature of the electrolytes. DSC measurements of pure PEI:LiTFSI elec-trolytes were performed on a TA Instruments Q1000 Differential Scanning Calorimeter with a temperature precision of 0.05 °C. The electrolytes were first heated, then cooled, and again heated between -80 and 80 °C. Heating and cooling rates were 10 °C/min and 5 °C/min, respectively. The system was cooled down with a connected liquid nitrogen system. DSC measure-ments of the rest of the polymer electrolytes were performed on a Mettler Toledo DSC 30 differential scanning calorimeter. The same procedure was followed between -100 and +80 °C. Sealed aluminum pans were used. Ni-trogen gas flow was provided during the measurements in order to have a dry atmosphere. Repeated DSC measurements on PEI and some PEI:LiTFSI electrolytes confirmed that the results that were given by the two different measurement systems are in good agreement.

10.9. Viscosity measurements Viscosity measurements were done using a Brookfield DV-II+Pro viscome-ter with an accuracy of ±1.0 %. The experimental setup is shown in Fig. 44. The main parts of the viscometer are a motor, a calibrated spring, a spindle, and a heat bath. A fluid sample is placed in a plate and the spindle is on the sample. In our experiments, a cone/plate type of spindle was used and the distance between the cone spindle and plate was fixed. The advantage of the cone/plate spindle is that it only requires a small amount of sample, 0.5-2 mL. The spindle was placed in ~1 mL of an electrolyte, which was poured on the plate. The heat bath allows the measurements to be performed at different temperatures.

Temperature

Melting

Crystallization

Glass transition

End

o← H

eat f

low

→ E

xo

92

Figure 44. Experimental setup of the viscosity measurements.

The principle of the viscosity measurements is based on the measurement of a torque required to rotate the spindle at different speeds. The motor drives the spindle via the spring. Rotation of the spindle causes a deflection of the spring. Different friction between the spindle and samples occurs because of the different fluidity of the electrolytes. Therefore, different viscosities of the electrolytes cause different degrees of the deflection of the spring. The vis-cometer measures the torque at different spindle speeds and calculates shear rate and shear stress as [281].

.

3

= 2sin3

vc

cr

ητγ τ

θ π

Ω = (65.a.b)

where Ω is the angular velocity of the spindle, θc is cone angle, rc is cone radius, and ητ is torque, shown in Fig. 45.

Figure 45. Side and top view of the cone/plate spindle [281].

Spindle

Calibrated spiral spring

Motor

Heat bath

93

10.10. Ex-situ optical measurements First, transmittance and reflectance of the electrolytes were measured by sandwiching an electrolyte between two microscope slides to observe the influence of the nanoparticles in the electrolytes. Secondly, transmittance spectra of the ECWs in the initial bleached and colored states were measured in order to examine if the nanoparticles-doped electrolytes showed the same optical properties that we observed in the optical measurements of much thicker electrolytes between two glass plates.

The ex-situ optical measurements of the electrolytes and ECWs were per-formed on a Perkin Elmer Lambda 900 spectrophotometer equipped with an integrating sphere and in the 300-2500 nm wavelength range. A spectropho-tometer is an instrument that sends a beam of light at different wavelengths to a sample and detects transmitted and reflected light from the sample. The theory of the integrating sphere is based on establishing a radiation balance inside the sphere [282-284]. The inner wall of the integrating sphere is coated with Spectralon, which is a material with highly diffuse reflectance. The sample beam and the reference beam are subjected to several reflections on the inner wall of the sphere and the detector records an average indirect light intensity of the sample beam and reference beam and the output signal is given as the ratio of them both. The detector is placed at the bottom of the sphere, so that the sample parts are not within the detector field of view. More details about the principles of the optical measurements with Lambda 900 spectrophotometer can be found in Ref. [285].

Figure 46. Top view of an integrating sphere.

The top view of the integrating sphere is given in Fig. 46. The sphere has five ports: sample beam entrance port, sample beam reflectance port, specu-lar beam exit port, reference beam entrance port, and reference beam reflec-tance port. The integrating sphere allows total and diffuse transmittance and reflectance to be determined. Transmittance and reflectance measurements were done at normal incidence and 8 ° angle of incidence, respectively. The top view of the integrating sphere is shown for measurements of total trans-

Reference beamentrance port

Specular beamexit port

Reference beam reflectance port

Sample beam entrance port

Detector

Sample beam reflectance port

94

mittance in Fig. 47 (a), diffuse transmittance in Fig. 47 (b), total reflectance in Fig. 47 (c), and diffuse reflectance in Fig. 47 (d).

Figure 47. Top view of the integrating sphere for the total transmittance (a), the diffuse transmittance (b), the total reflectance (c), and the diffuse reflectance (d) measurements.

For the total transmittance measurement, a sample was placed at the sample beam entrance port with 19 mm diameter aperture and a Spectralon plate was at the sample beam reflectance port. The Spectralon plate was replaced with a black cone in order to trap the direct sample beam in diffuse transmit-tance measurements. In total reflectance mode, the sample is positioned at the sample beam reflectance port. In diffuse reflectance measurements, the specular beam exit port is left open to allow the specular beam reflected from the sample to escape.

Tdir, Tdiff, Ttot, Rspec, Rdiff, and Rtot were calculated as

= −dir TT DTT S S , = ×diff DT SPT S R , tot dir diffT T T= + (66.a.b.c)

( )spec TR DR lR S S F= − × , = ×diff DR SPR S R , = +tot spec diffR R R (66.d.e.f)

where STT, SDT, STR, and SDR are the detected signals of total transmittance, diffuse transmittance, total reflectance, diffuse reflectance measurements, respectively. RSP is the reflectance of the Spectralon plate that is wavelength dependent and is given in Fig. 48. Fl is a fraction of the light, which corrects for port losses and mismatch between reflectance of the plates covering the reflectance port and the specular beam exit port. Fl was 0.97 for our experi-ments.

Sample

(d)

Sample

(c)

SampleBlack cone

(b) (a)

Sample Spectalon plate

95

500 1000 1500 2000 2500

0.92

0.94

0.96

0.98

Ref

lect

ance

(%

)

Wavelength (nm) Figure 48. Reflectance of the spectralon plate SPR . The data was provided by Arne Roos.

In order to check the effect of the cone on diffuse transmittance measure-ments, the reflectance of the black cone was measured. For the measure-ment, the same configuration of Fig. 47 (b) was used without a sample at the sample beam entrance port. The signal was multiplied by RSP. Figure 49 shows the reflectance of the black cone. As can be seen, the cone contributed of 0.35 % to the diffuse transmittance measurements.

500 1000 1500 2000 25000.0

0.5

1.0

Wavelength (nm)

Ref

lect

ance

(%

)

Figure 49. Reflectance of the black cone.

For the optical measurements of PEI:LiTFSI:In2O3 and PEI:LiTFSI:ITO electrolytes, we had to reduce the diameter of the light path as we had too small samples. Therefore, an aperture with a diameter of 13 mm was added to the sample beam entrance port in order to reduce the area of the aperture. Calibration of the instrument was done in the presence of this aperture. For reflectance measurements, another aperture with a 13-mm-diameter was positioned at the sample beam reflectance port. In order to determine the

96

contribution of the signals coming from the aperture to the sample measure-ments, total and diffuse reflectance measurements were performed with the aperture and without the sample. Reflectance values of the samples were determined by subtracting the total (STR-ap) and diffuse (SDR-ap) reflectance signals coming from the aperture, calculated as

( ) ( )spec TR TR ap DR DR ap lR S S S S F− − = − − − × (67.a)

( )diff DR DR ap SPR S S R−= − × (67.b)

Comparison of optical measurements with and without the aperture for a PEI:LiTFSI electrolyte are given in Fig. 50.

500 1000 1500 2000 25000

20

40

60

80

Tdir-no apparatus

Tdir-with apparatus

Tdiff-no apparatus

Tdiff-with apparatus

Wavelength (nm)

Dir

ect t

rans

mitt

ance

(%

)

0.0

0.2

0.4

0.6

0.8

1.0

Dif

fuse

tran

smitt

ance

(%

)

(a)

500 1000 1500 2000 25000

2

4

6

8

10

12(b) Rspec-no apparatus

Rspec-with apparatus

Rdiff-no apparatus

Rdiff-with apparatus

Wavelength (nm)

Spec

ular

ref

lect

ance

(%

)

0

2

4

6

Dif

fuse

ref

lect

ance

(%

)

Figure 50. Direct and diffuse transmittance (a) and specular and diffuse reflectance (b) of a PEI:LiTFSI electrolyte in measurements with and without the aperture.

10.11. In-situ optical measurements In-situ direct transmittance measurements were done on ECWs during elec-trochromic measurements in order to determine the transmittance and switching times for the bleached and the colored states.

In-situ optical measurements were performed with a setup that has a light-emitting diode (LED) lamp and a photodiode sensor. The LED produced a green light with a wavelength of 530 nm. An ECW was placed between the LED and the sensor. The light created by the LED passed through the ECW and reached the photodiode where it was converted to voltage. The transmit-tance of the ECW was obtained by normalizing the measured voltage to a voltage, which was measured before the ECW was placed between the LED and the diode.

97

10.12. Thickness measurements Thicknesses of the electrolytes that were used in the optical measurements were measured by using a micrometer. Thickness measurements were done with the configuration of glass/electrolyte/glass and then the thickness of the two glasses was subtracted from the measured thickness. Thicknesses of the electrolytes in ECWs were determined by a Keyence LK-G10 high precision laser displacement sensor with 100 nm accuracy.

10.13. Electrochromic measurements Electrochromic performance of the devices was examined with square wave voltammetry and in-situ optical measurements. In square wave wave volt-ammetry measurements, constant voltages were applied for a defined time and a current passing through a ECW was measured. Figure 51 shows a dia-gram of square wave voltammetry process. A proper voltage Vc was applied during a certain time for coloring. It was followed with a zero voltage for relaxation. After this, bleaching voltage Vb was applied. Finally, a zero volt-age was applied again. Thus, one cycle of the measurement was completed and this procedure was carried out for every cycle.

relaxation

bleaching

relaxation

Vb

Vc

Time

0

coloring

Figure 51. Square wave voltammetry diagram for one cycle.

In order to avoid destruction of the ECWs, a control process was applied for the colored state. The coloring voltage was stopped when the transmittance reached 19±1 %.

For ECWs with NiO films 1.7 V was applied for 180 s for coloring. Zero voltage was applied for 120 s for relaxation. After that, -1.6 V was applied for 150 s for bleaching. Finally, zero voltage was applied for 120 s for re-laxation.

98

The electrochromic performance of ECWs with IrO2 films was examined with 1.2 V coloring voltage applied for 120 s, zero voltage for 120 s, -1.6 V bleaching voltage for 180 s and zero voltage for 120 s.

99

11. Results and discussion

11.1. PEI:LiTFSI electrolytes 11.1.1. Different PEI molecular weight In this section, the influence of the polymer molecular weight on the ionic conductivity, the viscosity and the permittivity of PEI:LiTFSI polymer elec-trolytes are examined.

Because of the importance of the polymer molecular weight on physical properties, we started the project by investigating PEI:LiTFSI with different molecular weight of PEI. We chose three different molecular weights of PEI. The first, which was named low molecular weight PEI (LMw-PEI) had an average molecular weight of 1200 g/mol and contained 50 wt% of water. The second had a moderate molecular weight (MMw-PEI) of 10000 g/mol with 99 % purity. The last, HMw-PEI, had an average molecular weight of 60000 g/mol and contained 50 wt% of water.

As noted, HMw-PEI and LMw-PEI contain a lot of water. Water and any other residual solvents strongly affect the characterization results of electro-lytes and the performance of ECWs. In order to get rid of the water, PEIs were dried as described in Section 10.1. The drying process of the pristine polymers was studied. TGA and DIS measurements were done on as-received and dried PEIs. TGA results for as-received and dried PEIs are shown in Fig. 52 (a) and (b), respectively. At 200 °C, as-received PEI was reduced to 56 wt%, 93 wt%, and 70 wt% for LMw-PEI, MMw-PEI, and HMw-PEI, respectively. The dried PEI was reduced to 83 wt%, 95 wt%, and 89 wt% for LMw-PEI, MMw-PEI, and HMw-PEI, respectively, at 200 °C.

The presence of water in a material can easily be observed from conduc-tivity measurements of the samples. Therefore, we measured the ionic con-ductivity of each type of PEI, as shown in Table 8. The drying effects on the ionic conductivity of the polymers are seen clearly. The as-received PEI had a very high conductivity due to the water content. The LMw-PEI is the one that was mostly affected by the drying procedure. The ionic conductivity of the as-received LMw-PEI was two orders of magnitude larger than that of the dried LMw-PEI. Ionic conductivity measurements of the MMw-PEI polymer were not affected drastically by the water content because of the high purity of the MMw-PEI.

100

0 100 200 300 400 500 600

0

20

40

60

80

100

Wei

ght (

%)

Temperature (°C)

LMw-PEI MMw-PEI HMw-PEI

(a)

0 100 200 300 400 500

0

20

40

60

80

100

Temperature (°C)

Wei

ght (

%)

(b)

Figure 52. TGA measurements of LMw, MMw, and HMw PEI as-received (a) and as-dried (b).

After studying the drying procedure of PEI, the preparation method of the electrolytes was examined. First, LMw-PEI was dried. Secondly, the poly-mer electrolytes of LMw-PEI:LITFSI at the molar ratio of 25:1 were pre-pared by mixing them with or without methanol. Then, methanol was evapo-rated as described in Section 10.2.1. The TGA measurements of the LMw-PEI:LiTFSI electrolytes are shown in Fig. 53 (a). It was found that using methanol reduced the water content to less than 0.5 wt% in the electrolytes.

Table 8. Room temperature ionic conductivity of LMw, MMw, and HMw PEI in the forms as-received and as dried.

σion (S/cm)

LMw-PEI MMw-PEI HMw-PEI

As-received 5.7x10-5 7x10-8 1.7x10-4

As-dried 7x10-7 8.3x10-8 5.3x10-6

We also examined if the same drying procedure was suitable for electrolytes containing different amounts of salt. Polymer electrolytes were prepared with LMw-PEI:LiTFSI at molar ratios of 25:1 and 10:1. TGA measurements of LMw-PEI:LiTFSI at molar ratios of 25:1 and 10:1 are shown in Fig. 53 (b). No difference was observed for the reduction of the water content in these electrolytes.

The influence of the preparation method on the conductivity of the LMw-PEI:LiTFSI, MMw-PEI:LiTFSI, and HMw-PEI:LiTFSI electrolytes with a molar ratio of 25:1 is shown in Table 9. The electrolytes prepared without using methanol had higher conductivity than the electrolytes prepared using methanol. These higher values are a result of the water content in the sam-ples. It was seen that using methanol in sample preparation procedure re-duced the water content and reduced the conductivity. There was a slight difference between the ionic conductivities of MMw-PEI:LiTFSI and HMw- PEI:LiTFSI electrolytes.

101

0 100 200 300 400 5000

20

40

60

80

100 (a)

Wei

ght (

%)

Temperature (°C)

dried dried after mixing

with methanol

LMw-PEI:LiTFSI 25:1

0 100 200 300 400 5000

20

40

60

80

100 (b)

LMw-PEI:LiTFSI [N]:[Li]

10:1 25:1

Wei

ght (

%)

Temperature (°C) Figure 53. TGA measurements of LMw-PEI:LiTFSI electrolytes (a) at molar ratio of 25:1, dried after mixing with and without methanol (b) at different molar ratios of 10:1 and 25:1, dried after mixing with methanol.

Table 9. Ionic conductivity of LMw-PEI:LITFSI, MMw-PEI:LITFSI, and HMw-PEI:LiTFSI electrolytes prepared with or without methanol at room temperature. The molar ratio was 25:1.

the way it was dried LMw-PEI:LiTFSIσion (S/cm)

MMw-PEI:LiTFSI HMw-PEI:LiTFSI

without using methanol 8.5x10-6 1.7x10-6 1.4x10-6

after mixing with methanol 4.7x10-6 1.5x10-6 1.1x10-6

Finally, we examined the molecular weight dependence of ionic conductiv-ity, viscosity, and permittivity at different salt concentrations. Due to the highly viscous form of HMw-PEI:LiTFSI electrolytes, we were not able to measure the viscosity of these samples. In addition, there was no significant difference between the ionic conductivities of MMw-PEI:LiTFSI and HMw-PEI:LiTFSI electrolytes. Therefore, we compared the salt concentration de-pendence of viscosity and ionic conductivity for LMw-PEI:LiTFSI and MMw-PEI:LiTFSI electrolytes.

0.1 11x10-6

2x10-6

3x10-6

4x10-6

5x10-6

6x10-6

LMw-PEI:LiTFSI MMw-PEI:LiTFSI

Ioni

c co

nduc

tivity

(S

/cm

)

Salt concentration (mol/kg)

(a)

0.1 1

10

100

Vis

cosi

ty (

Pa

s)

Salt concentration (mol/kg)

(b)

Figure 54. Ionic conductivity (a) and viscosity (b) of LMw-PEI:LiTFSI and MMw-PEI:LiTFSI electrolytes with different LiTFSI concentration at room temperature.

102

Figure 54 shows the ionic conductivity (a), and viscosity (b) of LMw-PEI:LITFSI and MMw-PEI:LiTFSI electrolytes as a function of different molar ratios at room temperature. There was no significant difference be-tween the ionic conductivity values of the LMw-PEI:LiTFSI and the MMw-PEI:LiTFSI electrolytes. However, the viscosity of the LMw-PEI:LiTFSI electrolytes was lower than that of the MMw-PEI:LiTFSI electrolytes.

The real and imaginary parts of the permittivity spectra of LMw-PEI:LiTFSI and MMw-PEI:LiTFSI electrolytes with the molar ratio of 25:1 at room temperature are shown in Fig. 55 (a) and (b), respectively. Electrode polarization in real permittivity spectra is seen below the frequency of 103 Hz. Apart from a small shift of the electrode polarization of the LMw-PEI:LiTFSI electrolyte towards high frequencies, there is not much differ-ence in the permittivity spectra of all electrolytes with respect to frequency.

10-3 10-1 101 103 105 107100

102

104

106

108(a)

LMw-PEI:LiTFSI MMw-PEI:LiTFSI

Rea

l per

mitt

ivity

Frequency (Hz)10-3 10-1 101 103 105 107100

102

104

106

108 (b)Im

agin

ary

perm

ittiv

ity

Frequency (Hz) Figure 55. Real (a) and imaginary (b) permittivity of LMw-PEI:LiTFSI and MMw-PEI:LiTFSI electrolytes with the molar ratio of 25:1 at room temperature.

From these results, we concluded that molecular weight does not have a large effect on the properties of PEI:LiTFSI electrolytes. Therefore, we chose the MMw-PEI as a host matrix of the electrolytes for further charac-terization and application because of its higher purity. From now on, MMw-PEI will be referred to as PEI. Some properties of pristine PEI are given in Table 10.

Table 10. Properties of pristine PEI

Property Value

Density (g/mL) 1.01

Viscosity (Pa.s) 73

Glass transition temperature (°C) -59

Ionic conductivity (S/cm) 8.5x10-8

Real permittivity (24 kHz) 7

103

11.1.2. Applied voltage dependence of the dielectric/impedance measurements An investigation of the applied voltage dependence of impedance measure-ments was done for the PEI:LiTFSI electrolyte with a molar ratio of 50:1. The measurements were performed between 0.01 V and 3 V.

Figure 56 shows the real and imaginary parts of the permittivity of the PEI:LiTFSI electrolyte at room temperature for different applied voltages. Electrode polarization was seen in the real part of the permittivity below a frequency of 1 kHz, and bulk relaxation was above this frequency. Imaginary permittivity was proportional with ω-1, which is the characteristic behaviour of ionic conductivity. In general, relaxation peaks are obtained by subtracting the ionic conductivity contribution from the imaginary permittivity. However, in our study, it was difficult to observe the peaks because ionic conductivity dominated the impedance spectra of the PEI:LiTFSI electrolyte.

10-3 10-1 101 103 105 107100

102

104

106

108

103 104 105 106 1075

10

15

20

25

30

35

40

Rea

l per

mit

tivi

ty

Frequency (Hz)

0.03 0.05 0.1 0.2 0.6 1 2 3

Rea

l per

mitt

ivity

Frequency (Hz)

V (Volt)

(a)

10-3 10-1 101 103 105 107100

102

104

106

108

Imag

inar

y pe

rmitt

ivity

Frequency (Hz)

(b)

Figure 56. Real (a) and imaginary (b) permittivity of the PEI:LiTFSI electrolyte with molar ratio 50:1 at different applied voltages V.

Figure 57. DHN equivalent circuit model.

The impedance spectra of the PEI:LiTFSI electrolyte at different applied voltages were fitted to an equivalent circuit model which we called the DHN model. We examined the bulk relaxation of PEI:LiTFSI electrolytes. The spectra were fitted in the frequency range above 1 kHz. The DHN model circuit is shown in Fig. 57. It consists of a series connection of a high-frequency resistance (RD∞) with a parallel connection of a high-frequency

RD∞

CD∞

RD

DHN

104

capacitance (CD∞), a DHN element, which was described in Section 7.2, and an ionic resistance (RD). The least-squares method was used in the fitting, which is proportional to the average percentage error between the original data points and the calculated values of the fitting. The least-square values were less than 0.02, apart from 10 mV.

0.01 0.1 1500

600

700

800

900

1000

ac potential (V)

Hig

h fr

eque

ncy

resi

stan

ce (

Ω)

(a)

0.01 0.1 139

40

41

42

43(b)

ac potential (V)

Ioni

c re

sist

ance

(kΩ

)

0.01 0.1 14.5

5.0

5.5

6.0

6.5

Hig

h fr

eque

ncy

capa

cita

nce

(pF

)

(c)

ac potential (V)0.01 0.1 1

8

9

10

11(d)

ac potential (V)

ΔC (

pF)

0.01 0.1 1

0.5

1.0

1.5

2.0

2.5

3.0(e)

ac potential (V)

Rel

axat

ion

tim

e (s

)

0.01 0.1 1

0.5

1.0

1.5

ac potential (V)

(f)

Pow

er-l

aw e

xpon

ent,

UD

Pow

er-l

aw e

xpon

ent,

PD

0.0

0.5

1.0

Figure 58. Applied voltage dependence of high-frequency resistance (a), ionic resis-tance (b), high-frequency capacitance (c), difference between capacitance values at low and high frequencies (d), relaxation time (e) and power-law exponents UD and PD (f) of the DHN equivalent circuit model for PEI:LiTFSI electrolytes with molar ratio 50:1 at room temperatures.

105

Figure 58 shows the applied voltage dependence of the parameters of the DHN model. The high-frequency resistance, shown in Fig. 58 (a), was al-most constant up to 1 V and slightly increased above it. The ionic resistance, shown in Fig. 58 (b), was almost constant at different voltages. The high-frequency capacitance is shown in Fig. 58 (c). It was quite stable up to 1 V and then increased slightly. The difference between capacitance values at low and high frequencies (ΔC) was almost constant as a function of applied voltage, as shown in Fig. 58 (d). The relaxation time (τD), shown in Fig. 58 (e), and the power-law exponents UD and PD, shown in Fig. 58 (f), did not change significantly as a function of applied voltage.

Figure 59 shows the conductivity spectra of PEI:LiTFSI electrolytes at different applied voltages. The ionic conductivity was the same at all applied voltages.

10-3 10-1 101 103 105 107

10-7

10-6

10-5

10-4

Con

duct

ivit

y (S

/cm

)

Frequency (Hz)

0.03 0.05 0.1 0.2 0.6 1 2 3

V (Volt)

Figure 59. Conductivity spectrum of PEI:LiTFSI electrolyte with molar ratio 50:1 at different applied voltages V.

To sum up, there was applied voltage dependence of real and imaginary permittivity and conductivity below the frequency 1 Hz. Any drastic influ-ence of the different applied voltage to the bulk relaxation of PEI:LiTFSI electrolytes was not observed from the fitting of the impedance data. On the other hand, it was seen that below 0.8 V, there was noise in the high fre-quency region of real permittivity, which can be seen clearly in the inset figure of Fig 56 (a). The voltage dependence at low frequencies did not af-fect our study because we were interested in the bulk relaxation region. Therefore, we decided to apply 1 V to avoid the noise in further measure-ments in DIS.

106

11.1.3. Different LiTFSI concentration (Paper I) In this section, glass transition, viscosity, permittivity, and ionic conductivity characteristics of PEI:LiTFSI electrolytes with different LiTFSI concentra-tions are given.

From DSC measurements between -80 and 80 °C, the glass transition of PEI:LiTFSI electrolytes were observed. There was no evidence of melting or recrystallization of the electrolytes. Glass transition temperatures of PEI:LiTFSI electrolytes were between -51.8 and -63 °C at different molar ratios. Tg of PEI was -59 °C. Adding salt first decreased Tg. Changes in Tg of PEI:LiTFSI electrolytes were 1-2 °C up to molar ratio 50:1, and above that, Tg increased drastically. Salt concentration dependence of Tg had a minimum at 100:1. This shows that the segmental motion in PEI:LiTFSI electrolytes with 100:1 molar ratio is easier than at the other molar ratios.

2.8 2.9 3.0 3.1 3.2 3.3 3.41

10

100

Vis

cosi

ty (

Pa.

s)

1000/Temperature (1/K)

(a)

2.9 3.0 3.1 3.2 3.3 3.410-6

10-5

10-4

400:1 200:1 100:1 50:1 20:1

Ioni

c co

nduc

tivity

(S

/cm

)

1000/Temperature (1/K)

[N]:[Li]

(b)

Figure 60. Temperature dependence of viscosity (a) and ionic conductivity (b) of PEI:LiTFSI electrolytes at the shown molar ratios. Symbols are measured data and straight lines are fits to the Arrhenius equation.

Viscosity measurements of PEI:LiTFSI electrolytes with different salt con-centrations were done between 25 and 85 °C. It was seen that shear rate de-pendence of shear stress was linear for PEI:LiTFSI electrolytes with differ-ent salt concentrations. The fluidity of PEI:LiTFSI electrolytes can be de-scribed by the Bingham plastic model, which is described in Section 9. The viscosity of PEI:LiTFSI electrolytes with different salt concentrations is shown in Fig. 60 (a) as a function of temperature. The viscosity of PEI:LiTFSI electrolytes had a minimum at a [N]:[Li] ratio of 100:1, a conse-quence of higher mobility of the polymer chains. The viscosity increased more below 50:1.

Ionic conductivities of PEI:LiTFSI electrolytes were obtained between 20 and 70 °C and are shown in Fig. 60 (b). Maximum ionic conductivity was obtained for 50:1 molar ratio. It rose monotonically from 2.47×10–6 S/cm at 20 °C to 7.76×10–5 at 70 °C for PEI:LiTFSI electrolyte with molar ratio 50:1.

Viscosity and ionic conductivity increased with increasing temperature. The temperature dependence of the viscosity and ionic conductivity followed

107

Arrhenius behavior, see Section 3.2 and 9.2. The Arrhenius parameters are shown in Fig. 61 for different salt concentrations.

0.1 1102

103

104

105

Salt concentration (mol/kg)

Pre-

expo

nent

ial f

acto

r of

io

nic

cond

uctiv

ity

(S/c

m)

(a)

0.00

0.05

0.10

0.15

Pre

-exp

onen

tial

fac

tor

of

visc

osit

y (P

a.s)

0.1 145

50

55

60 E

σ

Eμ

Salt concentration (mol/kg)

Act

ivat

ion

ener

gy (

kJ/m

ol)

(b)

Figure 61. Salt concentration dependence of pre-exponential factors (a) and activa-tion energies Eσ and Eη of the Arrhenius fits (b) for the ionic conductivity and vis-cosity of PEI:LiTFSI electrolytes, respectively.

Some properties of PEI:LiTFSI electrolytes with different salt concentrations are given in Table 11.

Table 11. Density ρ, glass transition temperature Tg, viscosity η, ionic conductivity σion, and real permittivity ε′ at 24 kHz of PEI:LiTFSI electrolytes at the shown LiTFSI concentrations c.

[N]:[Li]c

(mol/kg)ρ

(g/mL)Tg

(°C)

η (20°C) (Pa.s)

σion x10-6

(25°C) (S/cm)

ε′ (24 kHz)

400 0.06 1.1 -61.5 52.8 1.05 12.7

200 0.11 1.07 -61.2 54.5 1.17 14.7

100 0.22 0.93 -62.8 48.8 2.24 16.9

50 0.41 0.86 -60.8 67.3 2.47 18.6

20 0.87 0.78 -51.8 - 2.38 19.2

11.1.4. Ion association and mobility effects on ionic conductivity (Paper II) In this section, the ionic conduction mechanism of PEI:LiTFSI electrolytes is examined in order to understand the effect of ion-ion and ion-polymer inter-action on the ion transport. The influence of segmental mobility and ion-pairs on the conduction mechanism of PEI:LiTFSI electrolytes is discussed by the analysis of the Walden product and iso-viscosity conductivity.

108

0.1 10

50

100

150

200

250

300Temperature (°C)

20 50 30 60 40 70

Mol

ar c

ondu

ctiv

ity (

mS

.kg/

mol

.cm

)

Salt concentration (mol/kg)

(a)

0.1 11

10

100

Vis

cosi

ty (

Pa.

s)

Temperature (°C) 25 65 T

g

45 85

Salt concentration (mol/kg)

(b)

210

215

220

Gla

ss tr

ansi

tion

tem

pera

ture

(K

)

Figure 62. Salt concentration dependence of molar conductivity at the shown tem-peratures (a) viscosity at the shown temperatures (left-hand scale) and glass transi-tion temperature Tg (right-hand scale) (b) for PEI:LiTFSI electrolytes.

Figure 62 (a) shows salt concentration dependence of molar conductivity of the PEI:LiTFSI electrolytes. The molar conductivity of the PEI:LiTFSI elec-trolytes displayed first a minimum and then a maximum with increasing salt concentration. The salt concentration dependence of molar conductivity was compared with that of glass transition temperature, as well as with the vis-cosity, Fig. 62 (b). The glass transition temperature and the viscosity exhib-ited a maximum and a minimum where the molar conductivity had a mini-mum and a maximum, respectively. This shows a major influence of the segmental motion of the PEI chains on the ion mobility, even in the low concentration region.

-1.5 -1.0 -0.5 0.0-5.5

-5.0

-4.5

log (Salt concentration (mol/kg))

log

(Ion

ic c

ondu

ctiv

ity (

S/c

m))

η= 18 Pa.s

Figure 63. Logarithm of the ionic conductivity versus logarithm of the salt concen-tration for PEI:LiTFSI electrolytes at 18 Pa.s. The straight line indicates the ideal situation with a slope of 1.

109

In order to distinguish the effect of charge carrier concentration and mobility on the conductivity, ionic conductivity at a constant viscosity was examined as a function of salt concentration for PEI:LiTFSI electrolytes. Ionic conduc-tivity at a constant mobility corresponds to a constant viscosity. Therefore, a certain value of viscosity was chosen and ionic conductivity values at that viscosity were taken for each salt concentration. Figure 63 shows the loga-rithm of ionic conductivity at 18 Pa.s viscosity versus the logarithm of the salt concentration for PEI:LiTFSI electrolytes. The logarithm of the ionic conductivity was nearly constant at concentrations lower than 0.1 mol/kg, which means that the concentration of free ions did not increase with the salt concentration. Above 0.1 mol/kg, the logarithm of the ionic conductivity increased linearly with a slope close to unity. The linearity indicates that the number of free ions increased and that the fraction of them was nearly con-stant for salt concentration above 0.1 mol/kg. An iso-viscosity plot shows that if only the ion concentration had an effect on the conductivity, we would get highest conductivity at the highest salt concentration. However, we got the highest ionic conductivity at 50:1 molar ratio. This shows the influence of the mobility on the ionic conductivity.

The Walden product is another way of studying the effect of ion concen-trations on conductivity. The Walden rule was described in Section 3.3 and is given as

2

0 6A

P

zq N

aη

πΛ = (68)

As noted this equation is derived for dilute electrolytes. We can derive an analogous equation for electrolytes with various salt concentrations. Using Eq. 8 (D=RγTt/6πNAaηP) and the expression of the mobility µ=Dq/kBTt, we can get the product of the viscosity and the mobility as

6p

q

aη μ

π= (69)

which is constant. We know that the molar conductivity can be written in terms of ionic

conductivity; hence the Walden product will be

ionp

saltc

ση ηΛ = (70)

By substituting Eq. 4, σ=Ncqµ, and Eq. 69 in Eq. 70 and giving csalt in terms of the volume concentration of the (total) salt molecules (NT) as csalt= NT/NA, we obtained

110

2

06f fA

PT T

N Nq N

a N Nη ρ η

π

Λ = = Λ

(71)

where Nf is the number of free ions. It should be noted that there is an addi-tional term of density of the electrolyte (ρ) in Eq. 71. It is because our molar conductivity calculations were done for salt concentration c with the units of mol/kg. The salt concentration c can be given in terms of molar concentra-tion and density of the electrolyte as c=csalt/ρ.

0.1 10.0

0.5

1.0

1.5

2.0Temperature (°C)

20 50 30 60 40 70

Wal

den

prod

uct (m

S.k

g.P

a.s/

mol

.cm

)

Salt concentration (mol/kg) Figure 64. Normalized Walden product Λη versus salt concentration of PEI:LiTFSI electrolytes at the shown temperatures.

From Eq. 71, it is seen that the Walden product at different salt concentra-tions is proportional to the ratio of the number of free ions to total ions (Nf/NT). Figure 64 shows the salt concentration dependence of the Walden product of PEI:LiTFSI electrolytes at different temperatures. At all tempera-tures, the same salt concentration dependence was observed for the electro-lytes. It was seen that the Walden product initially exhibits a decrease up to a concentration of 0.1 mol/kg. Then, it was almost constant with increasing salt concentration. This indicates that the fraction of free ions decreased and ion-pairs increased up to 0.1 mol/kg. Above 0.1 mol/kg the fraction of free ions was constant.

To sum up, the effect of both ion concentration and ion mobility was seen for PEI:LiTFSI electrolytes. There was a large similarity between the salt concentration dependence of the molar conductivity and of the glass transi-tion temperature as well as viscosity for PEI:LiTFSI electrolytes, hence the molar conductivity of the polymer electrolytes was influenced by the seg-

111

mental mobility of the polymer chains. In the low concentration region, up to 0.1 mol/kg, many ions formed ion-pairs that did not contribute to the ionic conductivity, and the fraction of the ion-pairs increased. The fraction of the free ions decreased. The number of the free ions increased with the nearly constant fraction above 0.1 mol/kg. The decrease in the molar conductivity at high temperatures can be attributed to the formation of transient crosslinks.

11.1.5. The role of ion-pair relaxation and ionic conductivity on the conduction process (Paper III) In this section, the role of ion-pair relaxation and ionic conductivity in the ion transport process of PEI:LiTFSI electrolytes is examined.

PEI:LiTFSI electrolytes with different molar ratios between 400:1 and 20:1 were examined by DIS at different temperatures between 20 and 60 °C. The real and imaginary parts of the permittivity for the PEI:LiTFSI electrolytes at different salt concentrations showed same characteristics at different salt concentrations. Electrode polarization was seen in the real part of the permittivity below a frequency of 1 kHz, and bulk relaxation was seen above this frequency for all salt concentrations. It was not possible to ob-serve dielectric loss peaks due to the dominant ionic conductivity of the di-electric loss spectra of PEI:LiTFSI electrolytes. The time-temperature super-position relation, which is described in Section 7.3, was studied for the PEI:LiTFSI electrolytes. It was seen that while the electrode polarization displays different behavior as a function of temperature, the shape of the bulk response of the electrolytes remained invariant.

Figure 65. DHN (a) and CHN (b) models used for fitting impedance data of PEI:LiTFSI electrolytes.

The bulk responses of PEI:LiTFSI electrolytes were compared with two different equivalent circuit models in order to understand the contributions of ionic conductivity and ion-pair relaxation to the conduction mechanism. One of the equivalent circuit models was the DHN model, which is de-scribed in Section 11.1.2. The second equivalent circuit model was a CHN model. In Fig. 65 (a) and (b), schematic pictures of DHN and CHN models are given, respectively. The CHN model consists of a serial connection of a

RC∞

CC∞

CHN

RD∞

RD

DHN

CD∞

112

high-frequency resistance (RC∞) and a parallel connection of a high-frequency capacitance (CC∞), and a CHN element, which is described in Section 7.2.

The main difference between the DHN and the CHN models is that the DHN element represents ion-pair relaxation and it is in parallel with the ionic resistance element, while the CHN element represents both ion-pair relaxation and ionic conductivity. From the fittings, it was seen that even though the CHN model has one parameter less than the DHN model, the quality of the fit was slightly better for the CHN model.

The BNN relation, described in Section 7.4, was examined for PEI:LiTFSI electrolytes with different salt concentrations and at room tem-perature. The BNN-constants of PEI:LiTFSI electrolytes with different salt concentrations are given in Table 12. The values indicate that the experimen-tal data had a good agreement with the BNN relation. It shows that there is a connection between ionic conductivity and ion-pair relaxation.

Table 12. The BNN-constant (FBNN) at different molar ratios for PEI:LiTFSI electro-lytes at room temperature.

[N]:[Li] FBNN

400 1.26

200 1.24

100 1.39

50 1.99

20 2.15

Finally, the activation energies of ionic conductivity and of ion-pair relaxa-tion were obtained by fitting the ionic conductivity and the relaxation time to an Arrhenius equation. The activation energies of ionic conductivity and ion-pair relaxation had almost the same value, as seen in Table 13. It indicated that there is a connection between ionic conductivity and ion-pair relaxation.

Table 13. Activation energy obtained by means of temperature dependence of ionic conductivity and relaxation time for PEI:LiTFSI electrolytes; they are shown as Ea-σ and Ea-r, respectively.

[N]:[Li] Ea-σ

(kJ/mol) ± Ea-σ

(kJ/mol) Ea-r

(kJ/mol) ± Ea-r

(kJ/mol)

400 46.55 0.86 46.20 0.19

200 51.17 1.77 48.69 2.56

100 54.47 1.00 53.44 1.84

50 58.75 1.04 57.34 2.72

20 58.65 0.86 54.42 1.71

113

In conclusion, having a better fit with the CHN model, good agreement with the BNN-relation, and the same values of the activation energies of ionic conductivity and ion-pair relaxation led us to deduce that ionic conductivity and ion-pair relaxation should be viewed as two parts of the same conduction process.

11.2. Nanocomposite PEI:LiTFSI electrolytes 11.2.1. PEI:LiTFSI:SiO2 electrolytes 11.2.1.1. Different SiO2 concentration (Paper IV) In this section, the effects of adding different amounts of SiO2 nanoparticles to PEI:LiTFSI electrolytes is explained. SiO2 nanoparticles are added to PEI:LiTFSI electrolytes in order to improve their ionic conductivity. Imped-ance spectra and an equivalent circuit model fitting of the bulk relaxation, ionic conductivity, glass transition temperatures and transmittance of the PEI:LiTFSI:SiO2 electrolytes are examined for different amounts of SiO2.

SiO2 nanoparticles up to 9 wt% were added to PEI:LiTFSI electrolytes with a constant molar ratio of 50:1. The permittivity of the electrolytes in-creased with an increasing amount of SiO2. Bulk relaxation and electrode polarization were seen in the real permittivity spectrum of the electrolytes. The bulk relaxation shifted towards higher frequencies with increasing amounts of SiO2. Dielectric loss peaks could not be observed in the imagi-nary part of the permittivity because of the dominating dc conductivity. The real and imaginary permittivity spectra of PEI:LiTFSI:SiO2 electrolytes at room temperature can be seen in Fig. 1 in Paper IV. The bulk response of PEI:LiTFSI:SiO2 electrolytes followed the time-temperature superposition relation and was fitted to a CHN equivalent circuit model, which is described in Section 11.1.5. The lowest ionic resistance was obtained with the highest amount of SiO2. SiO2 amount dependence of the fitting parameters is given in Fig. 5 in Paper IV.

Figure 66 shows the ionic conductivity of the PEI:LiTFSI:SiO2 electro-lytes as a function of SiO2 content at different temperatures. It was seen that the ionic conductivity increased monotonically for increasing SiO2 contents; specifically its room temperature value went from 8.5×10–7 S/cm without nanoparticles to 3.8×10–5 S/cm for 8 wt% of SiO2. The temperature depend-ence of the ionic conductivity of the electrolytes was fitted to the Arrhenius equation. The correlation coefficients of the fittings were better than 0.99 and the activation energies decreased with an increasing amount of SiO2, especially with the highest SiO2 content.

114

0 1 2 3 4 5 6 7 8

10-6

10-5

10-4

Temperature (°C) 20 50 30 60 40 70

Ioni

c co

nduc

tivi

ty (

S/cm

)

wt% SiO2

Figure 66. Ionic conductivity at the shown temperatures versus the amount of SiO2 in the PEI:LiTFSI:SiO2 electrolytes.

Glass transition temperatures for PEI:LiTFSI:SiO2 electrolytes with different amount of SiO2 are given in Table 14. They decreased monotonically with increasing amounts of SiO2 as long as wt% of SiO2 did not exceed 7 wt%. The decrease in glass transition temperature can be explained by an increase in the segmental flexibility of the PEI:LiTFSI:SiO2 electrolytes that was proportional to the amount of SiO2. This may mean that the presence of the SiO2 nanoparticles reduces the formation of the transient crosslinks. Thus, the segmental flexibility of the chains increased, and the ionic conductivity increased.

Table 14. Glass transition temperatures (Tg) for PEI:LiTFSI:SiO2 electrolytes with different amounts of SiO2.

wt% SiO2 0 0.5 1 3 5 6 7 8

Tg (°C) -59 -60 -60 -61 -61 -66 -67 -56

500 1000 1500 2000 25000

20

40

60

80

100

Tot

al tr

ansm

itta

nce

(%)

Wavelength (nm)

(a)

500 1000 1500 2000 25000

2

4

6

8

(b) wt% SiO2

0 6 0.5 7 1 7.5 2 8 3 8.5 4 9 5

Dif

fuse

tran

smitt

ance

(%

)

Wavelength (nm) Figure 67. Total (a) and diffuse (b) transmittance spectra of the PEI:LiTFSI:SiO2 electrolytes with different amounts of SiO2.

115

Figure 67 (a) shows the total transmittance spectra in the luminous and solar ranges for PEI:LiTFSI:SiO2 electrolytes with different SiO2 content. The total transmittance was higher than 90 % in the luminous range for SiO2 con-tent up to 7 wt%.The minima in the spectra, which are apparent especially at λ > 1500 nm, are characteristic for PEI. The diffuse transmittance of PEI:LiTFSI:SiO2 electrolytes, shown in Fig. 67 (b), has a peak at ~335 nm, which grows progressively larger particularly with the highest SiO2 content. The diffuse transmittance at 550 nm remained at ~1 % below 7 wt% SiO2. High diffuse transmittance and high glass transition temperature above 7 wt% can be explained by the aggregation of SiO2 nanoparticles.

In conclusion, the addition of SiO2 inorganic nanoparticles to PEI:LiTFSI electrolytes improved the ionic conductivity by roughly an order of magni-tude without any degradation at the transmittance. 11.2.1.2. Different LiTFSI concentration (Paper V) In this section, the influence of SiO2 nanoparticles on the ion conduction mechanism of PEI- LiTFSI electrolytes is studied. The prominence of SiO2 nanoparticles on free ion concentration and mobility of the ions is examined for PEI:LiTFSI electrolytes with different salt concentration.

PEI:LiTFSI:SiO2 electrolytes were examined between 400:1 and 20:1 [N]:[Li] ratios. The concentration of SiO2 nanoparticles was kept constant at 7 wt%. The electrolytes were characterized by DIS, DSC, and viscosity measurements. Conductivity measurements were done between 20 and 70 °C, while viscosity measurements were done between 40 and 70 °C due to the high viscosity of the electrolytes below 40 °C.

10-6

10-5

10-4

1

(a)Temperature (°C)

20 50 30 60 40 70

20:150:1200:1 100:1

Ioni

c co

nduc

tivi

ty (

S/cm

)

Salt concentration (mol/kg)

400:1

0.1

Figure 68. Salt concentration dependence of ionic conductivity for PEI:LiTFSI:SiO2 electrolytes at different temperatures.

116

LiTFSI concentration dependence of ionic conductivity of the PEI:LiTFSI:SiO2 electrolytes is given in Fig. 68. It shows a broad maximum with increasing salt concentration. The changes in the ionic conductivity were more drastic at lower salt concentrations up to the molar ratio of 100:1. Ionic conductivity had a maximum at 100:1. It rose monotonically from 5.5×10–6 S/cm at 20 °C to 1.2×10–4 S/cm at 70 °C for the molar ratio of 100:1.

0.1 145

50

55

60

Act

ivat

ion

ener

gy (

kJ/m

ol)

Salt concentration (mol/kg)

400:1 200:1 100:1 50:1 20:1

Figure 69. Activation energy Ea-σ for PEI:LiTFSI:SiO2 electrolytes with different salt concentrations.

The ionic conductivity exhibited an exponential dependence on temperature and was fitted to the Arrhenius equation. In Fig. 69, activation energies of the PEI:LiTFSI:SiO2 electrolytes are given for different salt concentrations, varied between 46 and 59 kJ/mol.

Figure 70 (a) shows the salt concentration dependence of the molar con-ductivity at different temperatures for PEI:LiTFSI:SiO2 electrolytes. Below 0.1 mol/kg, the molar conductivity is almost independent from the salt con-centration but increases with increasing salt concentration at 60 and 70 °C. Above 0.1 mol/kg, the molar conductivity first increases and then decreases with increasing salt concentration at all temperatures.

Figure 70 (b) shows viscosity at different temperatures and glass transi-tion temperature as a function of salt concentration. The concentration de-pendence of the viscosity has similar behavior as the concentration depend-ence of the glass transition temperatures.

117

0

200

400

600

1

Mol

ar c

ondu

ctiv

ity (

μS.k

g/m

ol.c

m)

Salt concentration (mol/kg)

Temperature (°C) 20 50 30 60 40 70

400:1 200:1 100:1 50:1 20:1

(a)

0.1104

105

106

107

0.1

Temperature (°C) 40 60 50 70

Gla

ss tr

ansi

tion

tem

pera

ture

(K

)

Vis

cosi

ty (

mPa.

s)

Salt concentration (mol/kg)1

200

205

210

215

220

225

Tg

400:1 200:1 100:1 50:1 20:1

(b)

Figure 70. Salt concentration dependence of molar conductivity at the shown tem-peratures (a) viscosity at the shown temperatures (left-hand scale) and glass transi-tion temperature Tg (right-hand scale) (b) for PEI:LiTFSI:SiO2 electrolytes.

In Fig. 71, the Walden product, which is described in Section 11.1.4, versus salt concentration is shown for PEI:LiTFSI:SiO2 electrolytes at the shown temperatures. It is seen that the product initially exhibits an increase of up to a concentration of 0.2 mol/kg and a decrease above 0.2 mol/kg.

0

20

40

60

Wal

den

prod

uct (

mS

.kg.

Pa.

s/m

ol.c

m)

Salt concentration (mol/kg)

Temperature (°C) 40 50 60 70

400:1 200:1 100:1 50:1 20:1

0.1 1

Figure 71. The Walden product versus salt concentration of PEI:LiTFSI:SiO2 elec-trolytes at the shown temperatures.

In conclusion, the inverse relation of the molar conductivity with the viscos-ity and the glass transition temperature above 0.1 mol/kg implies that the segmental motion of the polymer chains has a major influence on the ion mobility in this concentration region. The increase in the Walden product at low concentrations could be a sign of a decrease in the formation of the ion-pairs as a result of the possible interaction of the surface groups of SiO2 with anions. At higher salt concentrations, the binding sites at the nanoparticles

118

could be occupied and ion-pairing starts to be prominent. Therefore, the molar conductivity decreases with increasing salt concentration. To sum up, there is an optimum salt concentration, which is 100:1 in the present case, where much higher free-ion concentration and segmental flexibility is ob-tained for PEI:LiTFSI:SiO2 nanocomposite polymer electrolytes.

11.2.2. PEI:LiTFSI:In2O3 electrolytes In this section, the effects of adding different amounts of In2O3 nanoparticles to PEI:LiTFSI electrolytes are explained. Impedance spectra, ionic conduc-tivity, glass transition temperature, transmittance and reflectance spectra of PEI:LiTFSI:In2O3 electrolytes are reported.

In2O3 nanoparticles up to 7 wt% were added to PEI:LiTFSI electrolytes with a constant molar ratio of 50:1. Figure 72 shows real and imaginary permittivity spectra of PEI:LiTFSI:In2O3 electrolytes at room temperature. The permittivity of the electrolytes was almost the same at different amounts of In2O3. At different temperatures up to 70 °C, bulk relaxation shifted to higher frequencies with increasing temperature.

100 102 104 106100

102

104

106

108

103 104 105 10610

15

20

25

30

35

40

Rea

l per

mitt

ivity

Frequency (Hz)

wt% In2O

3

0 1 3 5 7

Rea

l per

mitt

ivity

Frequency (Hz)

(a)

100 102 104 106100

102

104

106

Imag

inar

y pe

rmitt

ivity

Frequency (Hz)

(b)

Figure 72. Real (a) and imaginary (b) parts of the permittivity as a function of fre-quency for PEI:LiTFSI:In2O3 electrolytes with different amounts of In2O3 at room temperature.

Ionic conductivities of the PEI:LiTFSI:In2O3 electrolytes are shown in Fig. 73 as a function of In2O3 content at different temperatures. The ionic conductivity of PEI:LiTFSI:In2O3 electrolytes was not affected by the differ-ent amounts of In2O3 in the electrolytes. Specifically, it was 1.4×10–6 S/cm at room temperature and 2.7×10–5 S/cm at 70 °C. The temperature dependence of the ionic conductivity of the electrolytes was fitted to the Arrhenius equa-tion. The correlation coefficients of the fittings were better than 0.99. Acti-vation energies are given in Table 15. They were almost the same values for all cases.

119

Table 15. Glass transition temperatures (Tg) and activation energies Ea-σ for PEI:LiTFSI:In2O3 electrolytes with different amounts of In2O3.

wt% In2O3 Tg (°C) Ea-σ (kJ/mol) ±Ea-σ (kJ/mol)

0 -59 52.0 0.9

1 -49 50.9 1.6

3 -47 53.6 1.2

5 -44 53.7 0.6

7 -44 52.5 0.2

Glass transition temperatures for PEI:LiTFSI:In2O3 electrolytes with differ-ent amounts of In2O3 are given in Table 15, and they increased monotoni-cally with an increasing amount of In2O3. The change in glass transition temperature was from -59 to -44 °C, for In2O3 content from 0 to 7 wt%. This means that the segmental flexibility of the electrolytes decreased in the pres-ence of In2O3. However, ionic conductivity did not decrease. It can be con-cluded that the presence of the nanoparticles has a positive effect on the ionic conductivity.

0 2 4 610-6

10-5

10-4Temperature (°C)

22 50 30 60 40 70

Ioni

c co

nduc

tivity

(S

/cm

)

wt% In2O

3 Figure 73. Ionic conductivity at the shown temperatures versus the amount of In2O3 in the PEI:LiTFSI:In2O3 electrolytes.

Figure 74 (a) shows the total transmittance spectra for PEI:LiTFSI:In2O3 electrolytes with different amounts of In2O3. The total transmittance was higher than 87 % at 550 nm in all cases. There was a slight drop at the blue end of the spectrum in proportion to the In2O3 contents. The characteristic minima of PEI became more apparent as the In2O3 content increased. The diffuse transmittance is shown in Fig. 74 (b). It had a peak at short wave-

120

lengths, which became larger when the amount of In2O3 increased. For the rest of the spectrum, the diffuse transmittance remained below 1 %.

500 1000 1500 2000 25000

20

40

60

80

100

Dir

ect t

rans

mitt

ance

(%

)

Wavelength (nm)

wt% In2O

3

0 1 3 5 7

(a)

500 1000 1500 2000 25000.0

0.5

1.0

1.5

2.0

Dif

fuse

tran

smitt

ance

(%

)

Wavelength (nm)

(b)

500 1000 1500 2000 25000

2

4

6

8

10

Spe

cula

r re

flec

tanc

e (%

)

Wavelength (nm)

(c)

500 1000 1500 2000 25000

1

2

Dif

fuse

ref

lect

ance

(%

)

Wavelength (nm)

(d)

Figure 74. Total (a) and diffuse (b) transmittance, and specular (c) and diffuse (d) reflectance spectra of the PEI:LiTFSI:In2O3 electrolytes with different In2O3 content and with a thickness 70±10 µm.

Specular and diffuse reflectance spectra of PEI:LiTFSI:In2O3 electrolytes with different amounts of In2O3 are shown in Fig. 74 (c) and (d), respec-tively. The specular reflectance was below 6 %, and almost the same for all electrolytes. Diffuse reflectance was less than 0.5 %, hence it was negligible.

In conclusion, the addition of In2O3 formed optically clear PEI:LiTFSI:In2O3 electrolytes and did not destroy the ionic conductivity.

11.2.3. PEI:LiTFSI:ITO electrolytes (Paper VI) In this section, PEI:LiTFSI electrolytes containing different amount of ITO nanoparticles are introduced. Impedance spectra, ionic conductivity, glass transition temperatures and optical properties of ITO-doped PEI:LiTFSI electrolytes are presented.

The PEI:LiTFSI:ITO nanocomposite polymer electrolytes were obtained by adding ITO up to 7 % of the total weight of ITO and PEI. The molar ratio of PEI to LiTFSI was kept constant at 50:1. The thickness of the electrolytes was 70±10 µm at the optical measurements.

121

10-2 100 102 104 106100

102

104

106

wt% ITO 0 1 3 5 7

Rea

l per

mitt

ivity

Frequency (Hz)

(a)

10-1 101 103 105

100

102

104

106

Imag

inar

y pe

rmitt

ivity

Frequency (Hz)

(b)

Figure 75. Real (a) and imaginary (b) permittivity spectra for PEI:LiTFSI:ITO elec-trolytes with different amounts of ITO at room temperature.

Figure 75 shows spectra of real and imaginary permittivity for PEI:LiTFSI:ITO electrolytes with different amounts of ITO content at room temperature. The spectra were similar as they were for PEI:LiTFSI and PEI:LiTFSI:SiO2 electrolytes. Bulk relaxation, an electrode polarization in the real permittivity and dominant dc conductivity in the imaginary permit-tivity spectra were seen. However, the bulk relaxation did not shift towards higher frequencies with an increasing amount of ITO. The real permittivity of the electrolytes was almost the same, ~21 (at 24 kHz) up to 5 wt% ITO and ~28 (at 24 kHz) for 7 wt% ITO.

3.0 3.2 3.4

10-6

10-5

wt% ITO 0 1 3 5 7

Ioni

c co

nduc

tivit

y (S

/cm

)

1000/Temperature (1/K) Figure 76. Logarithm of the ionic conductivity versus inverse temperature of PEI:LiTFSI:ITO electrolytes at the shown ITO wt%. Symbols are measured ionic conductivity data and straight lines are fits to the Arrhenius equation.

The PEI:LiTFSI:ITO electrolytes had the same ionic conductivity as PEI:LiTFSI electrolytes and adding ITO nanoparticles to PEI:LiTFSI elec-trolytes did not cause any deterioration of the conductivity. It was independ-ent of the ITO content and increased monotonically from ~1×10–6 S/cm at

122

23 °C to ~2×10–5 S/cm at 70 °C. The temperature dependence of the conduc-tivity is given in Fig. 76. It obeyed an Arrhenius behavior and activation energies are given in Table 16.

Table 16. Glass transition temperatures (Tg) and activation energies (Ea-σ) for PEI:LiTFSI:ITO electrolytes with different amounts of ITO.

wt% ITO Tg (°C) Ea-σ(kJ/mol) ±Ea-σ (kJ/mol)

0 -59 51.9 1.2

1 -47 51.5 0.5

3 -46 52.9 0.5

5 -45 56.2 1.0

7 -44 56.0 0.3

Glass transition temperatures for PEI:LiTFSI:ITO electrolytes are given in Table 16 and they increased drastically in the presence of ITO nanoparticles in the electrolyte. The glass transition temperature for PEI:LiTFSI increased monotonically from -59 to -44 °C, in the presence of 7 wt% ITO nanoparti-cles in the electrolyte. Despite the fact that the glass transition temperature changes by up to 15 °C, ionic conductivity was almost the same. This shows that adding ITO particle reduced the segmental flexibility but did not cause any decrease in the ionic conductivity. This means that ITO nanoparticles as well as In3O3 nanoparticles had a positive effect on the ionic conductivity but a negative effect on the segmental motion of PEI.

Figure 77 (a) shows measured and calculated direct transmittance spectra for PEI:LiTFSI:ITO electrolytes with different ITO content, respectively. The direct transmittance was above 80 % at 550 nm. The electrolytes have near-infrared blocking which expands towards the visible range with increas-ing amounts of ITO nanoparticles. The diffuse transmittance of PEI:LiTFSI:ITO electrolytes, shown in Fig. 77 (b) had a peak at short wave-lengths, which was smaller than 1 % and hence was negligible.

The direct transmittance spectra of PEI:LiTFSI:ITO electrolytes was cal-culated by Shuyi Li. The calculations were done based on the Fresnel equa-tions, Eq. 55. The refractive index of the PEI:LiTFSI:ITO electrolytes, used in the Fresnel equations, was obtained from the Maxwell-Garnett effective medium theory, Eq. 15. The refractive index of PEI, used in Eq. 15, was determined from the optical measurements of a PEI layer and was derived by fitting experimental and calculated data using commercially available soft-ware (Scout). The refractive index of PEI was ~1.51, which was consistent with the literature [286]. The extinction coefficient of PEI was obtained as described in Section 8.3.3. Optical constants for ITO nanoparticles were taken from the literature [186]. It was seen that measured spectral properties showed good agreement with the calculated ones. It showed that the ITO nanoparticles were well-separated. Details of the calculation are given in

123

Paper VI and will be given in Shuyi Li’s thesis. One should note that the calculated data in Fig. 77 (a) shows a slight difference from the calculated data shown in Fig. 3 of Paper VI in the wavelength range above 2000 nm. The difference is caused by an error in the calculation, which have mistak-enly put the imaginary permittivity equal to the extinction coefficient of PEI. Since the imaginary permittivity of PEI has a negligible value below 2000 nm, the correction does not affect the conclusion of the study and the calculated data now shows even better agreement with the experiments.

500 1000 1500 2000 25000

20

40

60

80

100

wt% ITO exp calc

0 1 1 3 3 5 5 7 7

Dir

ect t

rans

mitt

ance

(%

)

Wavelength (nm)

(a)

500 1000 1500 2000 25000.0

0.5

1.0

1.5

Dif

fuse

tran

smitt

ance

(%

)

Wavelength (nm)

(b)

500 1000 1500 2000 2500

2

4

6

8

Spe

cula

r re

flec

tanc

e (%

)

Wavelength (nm)

(c)

500 1000 1500 2000 25000.0

0.2

0.4

0.6

0.8

1.0

Dif

fuse

ref

lect

ance

(%

)

Wavelength (nm)

(d)

Figure 77. Direct (a) and diffuse (b) transmittance and specular (c) and diffuse (d) reflectance spectra of PEI:LiTFSI:ITO electrolytes with different amounts of ITO and with 70±10 µm thickness. Solid curves and symbols denote measured and calcu-lated data, respectively.

The measured specular and diffuse reflectance of the PEI:LiTFSI:ITO elec-trolytes is given in Fig. 77 (c) and (d), respectively. The specular reflectance had a peak up to 8 % which shifts from 370 nm to 420 nm with increasing ITO content. The reflectance was 7 to 8 % in the luminous range and ~4 % in the near-infrared. The diffuse reflectance had a small peak, less than 0.3 % at short wavelengths, and was almost zero for the rest of the spectra.

In conclusion, PEI:LiTFSI:ITO electrolytes combined high haze-free lu-minous transmittance with strong near infrared absorption without compro-mising ionic conductivity. Optical spectra showed good agreement with a calculation based on effective medium theory. This study shows that the

124

addition of nanoparticles can provide additional functionalities to polymer electrolytes for applications.

11.2.4. PEI:LiTFSI:ITO:SiO2 electrolytes In Section 11.2.1.1 and Section 11.2.3, it was shown that SiO2 nanoparticles in the electrolytes can increase the ionic conductivity and ITO nanoparticles can provide near-infrared absorption, respectively. Based on these results, it was suggested that mixing SiO2 and ITO nanoparticles in PEI:LiTFSI elec-trolytes may combine the improved ionic conductivity and near-infrared absorption properties in one electrolyte. In this section, ionic conductivity, glass transition temperature, and transmittance spectra of ITO and SiO2 nanoparticles added to PEI:LiTFSI electrolytes are examined.

Different amounts of SiO2 and ITO nanoparticles were added to a PEI:LiTFSI electrolyte. The weight percentage of the nanoparticles was a fraction of the total weight of ITO, SiO2 and PEI. The molar ratio of the PEI:LiTFSI electrolyte was kept constant at 50:1. Conductivity measure-ments were done at room temperature. In Table 17, ionic conductivity of the PEI:LiTFSI:SiO2:ITO electrolytes for different wt% SiO2 and ITO is given. It was seen that adding 6 wt% SiO2 to the pure PEI:LiTFSI electrolyte in-creased the conductivity from ~2.5×10–6 S/cm to ~6×10–6 S/cm. The addition of 3 wt% ITO to the pristine PEI:LiTFSI electrolyte caused a slight decrease in the conductivity. These two results agreed with the previous results. Sur-prisingly, the PEI:LiTFSI electrolyte with a mixture of SiO2 and ITO nanoparticles had lower conductivity, ~2×10–6 S/cm, than that of PEI:LiTFSI electrolyte. The presence of SiO2 nanoparticles in the PEI:LiTFSI:SiO2:ITO electrolytes did not cause any increase in the conductivity.

Table 17. Ionic conductivity (σion) and glass transition temperatures (Tg) for PEI:LiTFSI:SiO2:ITO electrolytes with different SiO2 and ITO content in wt%.

wt% SiO2 wt% ITO σion (S/cm) Tg (°C)

0 0 2.49x10-6 -58

0 3 1.91x10-6 -46

6 3 1.97x10-6 -50

6 0 5.98x10-6 -63

Glass transition temperatures for PEI:LiTFSI:SiO2:ITO electrolytes are given in Table 17. As known previously, the glass transition temperature of the PEI:LiTFSI electrolyte decreases and increases in the presence of SiO2 and ITO nanoparticles in the PEI:LiTFSI electrolytes, respectively. How-ever, the coexistence of SiO2 and ITO nanoparticles in the PEI:LiTFSI elec-trolytes increased the glass transition temperatures to -50 °C. This shows

125

ITO nanoparticles had a more dominant effect on the segmental flexibility of PEI than SiO2. The low ionic conductivity of PEI:LiTFSI:SiO2:ITO electro-lytes can be attributed to the decrease in the segmental flexibility of the PEI chains.

500 1000 1500 2000 25000

20

40

60

80

100

Dir

ect t

rans

mitt

ance

(%

)

Wavelength (nm)

wt% SiO2 wt% ITO

6 3 0 3

(a)

500 1000 1500 2000 25000

2

4

Dif

fuse

tran

smitt

ance

(%

)Wavelength (nm)

(b)

Figure 78. Direct (a) and diffuse (b) transmittance spectra of PEI:LiTFSI:SiO2:ITO electrolyte with 6 wt% SiO2 and 3 wt% ITO and PEI:LiTFSI:ITO electrolyte with 3 wt% ITO.

Figures 78 (a) and (b) show a comparison of direct and diffuse transmittance spectra, respectively, for the PEI:LiTFSI:SiO2:ITO electrolyte with 6 wt% SiO2 and 3 wt% ITO and the PEI:LiTFSI:ITO electrolyte with 3 wt% ITO. The direct transmittance was above 80 % at 550 nm in both cases. The PEI:LiTFSI:SiO2:ITO electrolyte had the same near-infrared absorption properties as the PEI:LiTFSI:ITO electrolyte. The diffuse transmittance of the PEI:LiTFSI:SiO2:ITO electrolyte had a peak at short wavelengths, as did the PEI:LiTFSI:SiO2 electrolytes.

To sum up, the combination of SiO2 and ITO nanoparticles in the PEI:LiTFSI electrolyte did not increase the ionic conductivity as expected, while the near infrared transmittance blocking property of the PEI:LiTFSI:ITO electrolytes was preserved.

11.2.5. SiO2 and ITO added PEI:LiTFSI:PC:EC electrolytes Up to now, characterization of pure and nanoparticle doped PEI:LiTFSI electrolytes was done in order to understand the interaction between the polymer host PEI, the LiTFSI ions, and the nanoparticles. In the ECW appli-cations, the ECWs did not work with these electrolytes: the transmittance of the ECWs did not change, and the current passing through the ECWs was very low. This made us increase the LiTFSI concentration. However, this was not possible with the pure electrolyte. Therefore, we added PC and EC to the PEI:LiTFSI electrolytes. In this section, the characterization of PEI:LiTFSI:PC:EC electrolytes with and without SiO2 and ITO nanoparti-cles, which will be used later in the ECW applications, is presented. First,

126

different concentrations of PEI:LiTFSI:PC:EC electrolytes will be examined. Then, a chosen concentration of PEI:LiTFSI:PC:EC is functionalized by adding about 1 wt% of SiO2 and ITO and is characterized. 11.2.5.1. PEI:LiTFSI:PC:EC electrolytes The PEI:LiTFSI:PC:EC were prepared with suitable viscosity that corre-sponded to the concentrations of 10:30:30:30, 13:37:25:25, 13:47:20:20, and 10:50:20:20 in weight percentages. The physical appearance of the PEI:LiTFSI:PC:EC electrolytes is shown in Fig. 79.

Figure 79. Physical appearance of PEI:LiTFSI:PC:EC electrolytes with different concentrations.

Ionic conductivity of the PEI:LiTFSI:PC:EC electrolytes with different con-centrations is shown in Fig. 80 at different temperatures. The highest ionic conductivity at room temperature was 7.7x10-4 S/cm for the electrolyte with 30 wt% LiTFSI, and the lowest was 5.5x10-5 S/cm for 47 wt% LiTFSI. The ionic conductivity for the PEI:LiTFSI:PC:EC electrolyte with 47 wt% LiTFSI was much lower than that of 37 wt% LiTFSI. The polymer concen-tration in these two electrolytes was the same. When we compared the ionic conductivity for the PEI:LiTFSI:PC:EC electrolyte with 30 wt% and 50 wt% LiTFSI, the ionic conductivity was much higher for the electrolytes with 30 wt% LiTFSI and with the same polymer concentration. Therefore, it can be said that the concentration of PC:EC in the PEI:LiTFSI:PC:EC electro-lytes had a dominating effect on the ionic conductivity.

The ionic conductivity exhibited an exponential dependence on tempera-ture and was fitted to the Arrhenius equation. Activation energies for PEI:LiTFSI:PC:EC electrolytes with different concentrations are given in Table 18.

127

2.9 3.0 3.1 3.2 3.3 3.4

10-4

10-3

10-2

PEI:LiTFSI:PC:EC 10:30:30:30 13:37:25:25 13:47:20:20 10:50:20:20

Ioni

c co

nduc

tivi

ty (

S/cm

)

1000 /Temperature (1/K) Figure 80. Ionic conductivity of the PEI:LiTFSI:PC:EC electrolytes versus tempera-ture. Symbols are measured data and lines are fits to the Arrhenius equation.

Table 18. Activation energies of ionic conductivity (Ea–σ) and glass transition tem-peratures (Tg).

PEI:LiTFSI:PC:EC Ea-σ

(kJ/mol) ±Ea-σ

(kJ/mol)Tg

(°C)

10:30:30:30 23 0.4 -

13:37:25:25 30 0.4 -80.9

13:47:20:20 41 0.8 -67.5

10:50:20:20 37 0.5 -67.5

Glass transition temperatures of the PEI:LiTFSI:PC:EC electrolytes with different concentrations are shown in Table 18. The glass transition tempera-ture of the PEI:LiTFSI:PC:EC electrolyte was not observed for 30 wt% LiTFSI in the measured temperature range. For 47 and 50 wt% LiTFSI, the same glass transition temperature, -67.5 °C, was observed. For 37 wt% LiTFSI, it was -80.9 °C.

Figure 81 (a) shows spectra of direct transmittance of the PEI:LiTFSI:PC:EC electrolytes together with that of the pure PEI:LiTFSI electrolyte with a molar ratio of 50:1. There was a minimum at λ ≈ 379 nm which was due to the presence of PC:EC in the PEI:LiTFSI electrolyte. Characteristic minima of PEI at above 1500 nm became smaller when PC:EC was added to the pure PEI:LiTFSI electrolyte. This must be because there was much less PEI in these mixtures. Spectral diffuse transmittance of the PEI:LiTFSI:PC:EC electrolytes is shown in Fig. 81 (b) and was less than 1 %. Specular and diffuse reflectance of the electrolytes was ~8 % and was less than ~0.5 %, respectively.

128

500 1000 1500 2000 2500

20

40

60

80

PEI:LiTFSI:PC:EC 10:30:30:30 13:37:25:25 13:47:20:20 10:50:20:20 PEI:LiTFSI [N]:[Li]=50:1

Wavelength (nm)

Dir

ect t

rans

mita

tnce

(%

)

(a)

500 1000 1500 2000 25000

1

2

3

4

Wavelength (nm)

Dif

fuse

tran

smitt

ance

(%

)

(b)

Figure 81. Direct (a) and diffuse (b) transmittance spectra of the PEI:LiTFSI:PC:EC electrolytes with the shown concentrations and with the thickness 70±10 µm.

11.2.5.2. PEI:LiTFSI:PC:EC:SiO2 and PEI:LiTFSI:PC:EC:ITO electrolytes (Paper VII-Part I) SiO2 and ITO nanoparticles were added to the PEI:LiTFSI:PC:EC electro-lytes with the composition of 13:47:20:20. The reason for choosing this composition, which is discussed in Section 11.3.1, was to have a better per-formance of ECWs with the electrolyte at that composition.

Ionic conductivities of undoped, SiO2-doped, and ITO-doped PEI:LiTFSI:PC:EC electrolytes are shown in Fig. 82 (a) as a function of temperature. This shows that adding 1 wt% of nanoparticles did not lead to a large change in the ionic conductivity. Specifically, the ionic conductivity of the electrolytes at 23 °C was 5.5x10-5 S/cm without nanoparticles, whereas it was 7.2x10-5 S/cm with SiO2 and 3.4x10-5 S/cm with ITO. The ionic conduc-tivity exhibited an exponential dependence on temperature and was fitted to the Arrhenius equation. The activation energy for the PEI:LiTFSI:PC:EC electrolyte was 41 kJ/mol. The addition of SiO2 and ITO nanoparticles did not lead to a big change in the activation energy.

20 30 40 50 60 7010-5

10-4

10-3

PEI:LiTFSI:PC:ECPEI:LiTFSI:PC:EC:SiO

2

PEI:LiTFSI:PC:EC:ITO

Ioni

c co

nduc

tivi

ty (

S/c

m)

Temperature (°C)

(a)

500 1000 1500 2000 25000

20

40

60

80

100

Wavelength (nm)

Dir

ect t

rans

mit

tanc

e (%

) (b)

Figure 82. Ionic conductivity versus temperature where symbols are measured data and lines are fits to the Arrhenius equation (a) and direct transmittance of ~70-µm-thick electrolyte layers (b) for the shown electrolytes.

129

Glass transition temperatures of the undoped, the SiO2-doped, and the ITO-doped electrolytes were -69, -72, and -74 °C, respectively. Adding nanopar-ticles to the PEI:LiTFSI:PC:EC electrolyte decreased its glass transition by a few degrees.

Figure 82 (b) shows the direct transmittance spectra of the PEI:LiTFSI:PC:EC, PEI:LiTFSI:PC:EC:SiO2, and PEI:LiTFSI:PC:EC:ITO electrolytes. The direct transmittance of the electrolytes was high and ex-ceeded 85 % at 550 nm. There was a minimum at λ ≈ 379 nm, which was due to the presence of PC:EC in the PEI:LiTFSI electrolyte. Characteristic minima of PEI at above 1500 nm became smaller when PC:EC was added to the pure PEI:LiTFSI electrolyte. Direct transmittance of SiO2-added PEI:LiTFSI:PC:EC electrolyte was quite similar to that of undoped PEI:LiTFSI:PC:EC. ITO added electrolytes had near-infrared transmittance blocking. Diffuse transmittance, not shown here, showed a peak at short wavelengths in the presence of the nanoparticles in the PEI:LiTFSI:PC:EC electrolytes. It was 1.2 and 1.6 % at 550 nm for ITO-doped and SiO2-doped PEI:LiTFSI:PC:EC electrolytes, respectively.

11.3. The electrochromic windows In this section, the optical properties and coloring/bleaching dynamics of ECWs incorporating different electrolytes are studied. First, different con-centrations of PEI:LiTFSI:PC:EC electrolytes are examined in NiO-based ECWs. Secondly, NiO-based ECWs with PEI:LiTFSI:PC:EC:SiO2 and PEI:LiTFSI:PC:EC:ITO electrolytes are studied. The effect of the presence of SiO2 in a proprietary electrolyte is studied in NiO-based ECWs. Electro-chromic performance of NiO-based ECWs is investigated in the presence of SiO2 in a crosslinked proprietary electrolyte. The PEI:LiTFSI:PC:EC and the PEI:LiTFSI:PC:EC:SiO2 electrolytes are used in IrO2-based ECWs and the electrochromic performance of the ECWs is tested. The configuration of the ECWs and the details of the measurements are given in Sections 10.5, 10.11 and 10.13.

11.3.1. ECWs with NiO and PEI:LiTFSI:PC:EC ECWs incorporating wt% 10:30:30:30, 13:37:25:25, 13:47:20:20, and 10:50:20:20 of PEI:LiTFSI:PC:EC electrolytes were tested for 20 cycles. In Fig. 83, one can see what the ECWs with the PEI:LiTFSI:PC:EC electrolyte look like in colored (a) and bleached (b) states.

130

(a) (b)

Figure 83. A picture of an ECW with PEI:LiTFSI:PC:EC electrolyte in colored (a) and bleached (b) states.

3060

Dir

ect t

rans

mit

tanc

e (%

)

10:30:30:30(a)

3060

13:37:25:25

3060

13:47:20:20

0 2000 4000 6000 8000 10000

3060

Time (s)

10:50:20:20

-20

0C

urre

nt (

mA

)(b)

-40-20

0

-40-20

0

0 2000 4000 6000 8000 10000-40-20

0

Time (s) Figure 84. Direct transmittance at 530 nm (a) and current (b) versus time for ECWs incorporating the shown weight percentages of the components of PEI:LiTFSI:PC:EC electrolytes.

Direct transmittance and current versus time for ECWs with PEI:LiTFSI:PC:EC electrolytes is shown in Fig. 84 (a). As seen, the trans-mittance of the PEI:LiTFSI:PC:EC electrolytes with wt% 10:30:30:30 was not stable in the relaxation state. The transmittance changed although there was no applied voltage at the relaxation steps.

In Fig. 85, in-situ transmittance of the ECWs incorporating different con-centrations of the PEI:LiTFSI:PC:EC electrolyte is shown. Only maximum and minimum values of the optical transmittance are shown for each cycle. A degradation of the transmittance of the bleached state was seen for all ECWs during the first few cycles. It stabilized at ~40 %. There was a slight difference at the transmittance of the bleached state. The transmittance of the colored states was ~18 % to where the transmittance value was set for all ECWs.

131

0 5 10 15 20

20

40

60

Dir

ect t

rans

mit

tanc

e (%

)

# Cycle

10:30:30:30 13:37:25:25 13:47:20:20 10:50:20:20

PEI:LiTFSI:PC:EC

Figure 85. Direct transmittance of colored and bleached states at 530 nm versus number of cycles for ECWs with the shown concentrations of PEI:LiTFSI:PC:EC electrolytes. Data signify maximum (solid curves) and minimum (dashed curves) values as cycling progresses.

The ECWs were kept in a dry atmosphere and in the dark after the measurements. Fifteen days later, the same samples were tested for 140 cycles. The ECW with the wt% 10:30:30:30 of PEI:LiTFSI:PC:EC was not cycled because it did not have a good performance in the 20 cycles. Degra-dation occurred in the ECWs with the PEI:LiTFSI:PC:EC electrolytes with wt% 13:37:25:25 in 140 cycles. The transmittance of the bleached state was ~37 % for PEI:LiTFSI:PC:EC with wt% 13:47:20:20 and 10:50:20:20 and 31 % for the ECWs with the PEI:LiTFSI:PC:EC with wt% 13:37:25:25. The transmittance of the colored state reached the fixed value ~20 % for the ECWs with PEI:LiTFSI:PC:EC with wt% 13:47:20:20 and 10:50:20:20, and was 23 % for the ECWs with PEI:LiTFSI:PC:EC 13:37:25:25.

0 100 200 300 400 500 600

20

30

40

Dir

ect t

rans

mitt

ance

(%

)

Time (s)

PEI:LiTFSI:PC:EC 13:37:25:25 13:47:20:20 10:50:20:20

(a)

0 100 200 300

20

30

40

Dir

ect t

rans

mitt

ance

(%

)

Time (s)

(b)

Figure 86. Direct transmittance at λ = 530 nm during coloring (a) and bleaching (b) versus time at the 100th coloring/bleaching cycle for ECWs incorporating the shown electrolytes. Arrows denote coloring and bleaching times.

132

Figures 86 (a) and (b) show direct transmittance during coloring and bleach-ing versus time at the 100th cycle for ECWs incorporating the shown electro-lytes, respectively. The bad performance of the ECW with the PEI:LiTFSI:PC:EC electrolyte with wt% 13:37:25:25 is seen here more clearly. This ECW had a lower transmittance modulation and longer coloring time compared to the other ECWs. There was no significant difference be-tween the switching times of the ECWs with the electrolytes with wt% 13:47:20:20 and 10:50:20:20. The coloring time of the ECWs was ~65 and ~78 s for wt% 13:47:20:20 and 10:50:20:20, respectively. The bleaching time of these two ECWs was ~150 s.

To sum up, in the first few cycles, transmittance in the bleached state for the ECWs with WO3 and NiO had a degradation when the PEI:LiTFSI:PC:EC electrolytes were used in the ECWs. The cycling process was irreversible. The ECW with PEI:LiTFSI:PC:EC 10:30:30:30 had a short circuit problem. The ECW with the PEI:LiTFSI:PC:EC with wt% 13:37:25:25 did not preserve its performance for long cycling. The ECWs with the PEI:LiTFSI:PC:EC with wt% 13:47:20:20 and 10:50:20:20 showed very similar performances but the PEI:LiTFSI:PC:EC with wt% 13:47:20:20 was more constant transmittance in the relaxation steps. Therefore, the PEI:LiTFSI:PC:EC electrolyte with wt% 13:47:20:20 was chosen as an electrolyte for the rest of the study.

11.3.2. ECWs with NiO and PEI:LiTFSI:PC:EC with ITO and SiO2 (Paper VII-Part II)

Figure 87 shows direct (a) and diffuse (b) transmittance spectra of ECWs with the PEI:LiTFSI:PC:EC, the PEI:LiTFSI:PC:EC:SiO2, and the PEI:LiTFSI:PC:EC:ITO electrolytes in the initial bleached and first colored

500 1000 1500 2000 25000

20

40

60

80

100

Wavelength (nm)

Dir

ect t

rans

mit

tanc

e (%

) (a)

500 1000 1500 2000 25000.0

0.5

1.0

1.5

2.0 PEI:LiTFSI:PC:EC PEI:LiTFSI:PC:EC:SiO

2

PEI:LiTFSI:PC:EC:ITO

Wavelength (nm)

Dif

fuse

tran

smitt

ance

(%

)

(b)

Figure 87. Direct (a) and diffuse (b) transmittance spectra for ECWs incorporating the shown electrolytes. Data are shown for the first bleached (solid curves) and col-ored cycle (dashed curves).

133

states. The ECW with the PEI:LiTFSI:PC:EC:SiO2 electrolyte showed the same spectral behavior as the ECW with the PEI:LiTFSI:PC:EC electrolyte. The ECW with the PEI:LiTFSI:PC:EC:ITO electrolyte absorbed the near-infrared part of the solar radiation as expected. Diffuse transmittance re-mained around ~1.5 % at 550 nm so that optical clarity prevailed.

Figure 88 shows in-situ optical transmittance of the ECWs with the PEI:LiTFSI:PC:EC, the PEI:LiTFSI:PC:EC:SiO2, and the PEI:LiTFSI:PC:EC:ITO electrolytes. Only maximum and minimum values of the optical transmittance are shown for each cycle. In the bleached state, there was a degradation of the transmittance for all devices during the first few cycles. Then, it stabilized at ~40 %. The transmittance in the colored states was stable at ~20 % to where it was set.

0 40 80 120 160 2000

20

40

60

80

100

bleached

PEI:LiTFSI:PC:EC PEI:LiTFSI:PC:EC:SiO

2

PEI:LiTFSI:PC:EC:ITO

# Cycle

Dir

ect t

rans

mitt

ance

(%

)

colored

Figure 88. Direct transmittance of colored and bleached states at 530 nm versus number of cycles for ECWs incorporating the shown types of electrolytes. Data signify maximum (solid curves) and minimum (dashed curves) values as cycling progresses.

Coloring and bleaching times are given as a function of number of cycles in Fig. 89 (a) and (b), respectively. It was seen that the coloring time was

0 20 40 60 80 100 120 140 160 180 2000

40

80

120

160

200 PEI:LiTFSI:PC:EC PEI:LiTFSI:PC:EC:SiO

2

PEI:LiTFSI:PC:EC:ITO

Col

orin

g tim

e (s

)

# Cycle

(a)

0 20 40 60 80 100 120 140 160 180 200100

120

140

160

180

200

Ble

achi

ng ti

me

(s)

# Cycle

(b)

Figure 89. Coloring (a) and bleaching (b) times versus number of cycles for ECWs incorporating the shown electrolytes. Data were taken at 530 nm.

134

reduced considerably in the presence of nanoparticles in the electrolyte. The average coloring time of the ECWs was reduced from 74 to 31 s in the pres-ence of the nanoparticles in the electrolyte. On the other hand, the average bleaching time remained at ~146 s for all electrolytes.

In conclusion, coloring times of the ECW with PEI:LiTFSI:PC:EC elec-trolyte improved in the presence of the nanoparticles. The bleaching dynam-ics was not affected by the addition of the nanoparticles to the PEI:LiTFSI:PC:EC electrolytes. Optical clarity of the devices was main-tained in the presence of the SiO2 nanoparticles. The ITO nanoparticles pro-vided efficient blocking of the near-infrared radiation.

11.3.3. ECWs with NiO and proprietary electrolyte (Paper VII-Part III) In this section, optical properties and coloring/bleaching dynamics of NiO-based ECWs with 1 wt% SiO2 nanoparticles containing a proprietary electro-lyte, called XEL, are examined.

Figure 90 shows the in-situ direct transmittance of the colored and bleached states as a function of the cycle number. The transmittance values of the bleached and colored states were very stable for the both ECWs with XEL and SiO2-doped XEL electrolytes. The transmittance was ~70 % in the bleached state and was at the fixed value ~18 % in the colored states.

0 50 100 150 2000

20

40

60

80

100

Dir

ect t

rans

mit

tanc

e (%

)

# Cycle

XEL SiO

2-doped XEL

bleached

colored

Figure 90. Direct transmittance of colored and bleached states at 530 nm versus number of cycles for ECWs incorporating the shown proprietary electrolytes. Data signify maximum (solid curves) and minimum (dashed curves) values as cycling progresses.

135

0 40 80 120 160 2000

50

100

150

200

Col

orin

g tim

e (s

)

XEL SiO

2-doped XEL

# Cycle

(a)

0 40 80 120 160 2000

50

100

150

200

Ble

achi

ng ti

me

(s)

# Cycle

(b)

Figure 91. Coloring (a) and bleaching (b) times versus number of cycles for ECWs incorporating the shown proprietary electrolytes. Data were taken at 530 nm.

Figures 91 (a) and (b) show coloring and bleaching times, respectively, for ECWs containing XEL and SiO2-doped XEL electrolytes. It was seen that coloring and bleaching times were shorter for ECWs containing SiO2-XEL.

To summarize, it was seen that the presence of SiO2 improved both color-ing time and bleaching time for the ECWs with the XEL electrolyte.

11.3.4. ECWs with NiO and crosslinked proprietary electrolyte In this section, the effect of SiO2 nanoparticles on coloring/bleaching dy-namics of NiO-based ECWs with a crosslinked proprietary electrolyte, called X-linked-XEL, is examined. The X-linked-XEL electrolyte was used in its pure form and after adding 1 wt% of SiO2. The electrochromic per-formance of the ECWs was tested for 20 cycles.

0 5 10 15 20

20

30

40

50

60

70

X-linked XEL X-linked XEL-SiO

2

Dir

ect t

rans

mit

tanc

e (%

)

# Cycle

bleached

colored

Figure 92. Direct transmittance of colored and bleached states at 530 nm versus number of cycles for the ECWs incorporating the X-linked XEL and SiO2-doped X-linked XEL electrolytes. Data signify maximum (solid curves) and minimum (dashed curves) values as cycling progresses.

136

Figure 92 shows in-situ direct transmittance of colored and bleached states versus number of cycles for the ECWs incorporating the X-linked XEL and SiO2-doped X-linked XEL electrolytes. The transmittance of the bleached and colored states was stabilized after a few cycles and was ~70 % in the bleached state and the fixed value ~18 % in the colored state for both ECWs, respectively.

0 100 200 300 400 500 600

20

40

60

Dir

ect t

rans

mit

tanc

e (%

)

Time (s)

X-linked XEL X-linked XEL-SiO

2

Figure 93. Direct transmittance at 530 nm during coloring and bleaching versus time at the 10th coloring/bleaching cycle for ECWs incorporating the shown undoped and SiO2 doped crosslinked proprietary electrolytes. Arrows denote coloring and bleach-ing times.

Figure 93 shows direct transmittance during coloring and bleaching, respec-tively, at the 10th cycle for ECWs incorporating X-linked XEL and SiO2-doped X-linked XEL electrolytes. It was seen that the coloring and the bleaching times were the same for both ECWs.

In conclusion, similar transmittance values were observed for the ECWs with SiO2 doped and undoped X-linked-XEL electrolytes. Coloring and bleaching times were also the same. There was no improvement of the switching times in the presence of the SiO2 particles in the X-linked XEL.

11.3.5. ECWs with IrO2 and PEI:LiTFSI:PC:EC-based electrolytes In this section, an anodic electrochromic film of IrO2 is used in ECWs. The electrochromic performance of the ECWs with PEI:LiTFSI:PC:EC, PEI:LiTFSI:PC:EC:SiO2, and PEI:LiTFSI:PC:EC:ITO electrolytes are exam-ined for 20 cycles.

Direct transmittance for the ECWs with the three types of electrolytes is reported in Fig. 94. The data signify envelope curves and hence refer to

137

maximum and minimum transmittance as the electrochromic color-ing/bleaching progresses. The modulation between the bleached and the colored state transmittance values increased during the 20th cycle. The bleached transmittance reached 70 % for the three electrolytes. The transmit-tance modulation was much higher for the ECW with the PEI:LiTFSI:PC:EC:ITO electrolyte, followed by the ECW with the PEI:LiTFSI:PC:EC:SiO2 electrolytes. Colored transmittance of the ECW with the PEI:LiTFSI:PC:EC electrolyte could not reach the limited transmit-tance in the 20th cycle.

0 5 10 15 20

20

40

60

Dir

ect t

rans

mitt

ance

(%

)

# Cycle

PEI:LiTFSI:PC:EC PEI:LiTFSI:PC:EC:SiO

2

PEI:LiTFSI:PC:EC:ITO

bleached

colored

Figure 94. Direct transmittance of colored and bleached states at 530 nm versus number of cycles for ECWs with the shown electrolytes. Data signify maximum (solid curves) and minimum (dashed curves) values as cycling progresses.

0 100 200

20

40

60

PEI:LiTFSI:PC:EC PEI:LiTFSI:PC:EC:SiO

2

PEI:LiTFSI:PC:EC:ITO

Dir

ect t

rans

mitt

ance

(%

)

Time (s)

(a)

0 20 40 60 80 100 120 14020

40

60

80

Dir

ect t

rans

mitt

ance

(%

)

Time (s)

(b)

Figure 95. Direct transmittance at 530 nm during coloring (a) and bleaching (b) versus time at the 20th cycle for the ECWs with the shown electrolytes. Arrows de-note coloring and bleaching times.

In figures 95 (a) and (b), in-situ optical transmittances of the ECWs with the shown electrolytes are reported for the 20th cycle for comparison of coloring and bleaching times of the different electrolytes, respectively. Coloring time

138

was reduced from 120 s to 107 and 50 s in the presence of SiO2 and ITO in the electrolyte, respectively. The bleaching times of the ECWs were 110, 84, and 38 s with the PEI:LiTFSI:PC:EC, PEI:LiTFSI:PC:EC:ITO, and PEI:LiTFSI:PC:EC:SiO2 electrolytes, respectively.

In conclusion, the degradation at the bleached state that was observed for the ECWs with the NiO films was not seen when IrO2 was used. Improve-ment in both the coloring and bleaching dynamic was obtained.

139

12. Conclusions

In this study PEI:LiTFSI-based electrolytes with and without nanoparticles were characterized and were applied for ECWs. Characterization techniques were dielectric/impedance spectroscopy, differential scanning calorimetry, viscosity recordings, optical spectroscopy in the visible and near-infrared range, and electrochromic measurements.

In the first part of this study, characterization of PEI:LiTFSI electrolytes was carried out at various salt concentrations and temperatures. Firstly, it was seen that molecular weight does not significantly affect the permittivity and the conductivity of the PEI:LiTFSI electrolytes, while fluidity of the electrolytes increases with decreasing molecular weight.

Glass transition temperatures of the electrolytes were between -45 and -63 °C at different salt concentrations. First a maximum and then a minimum were observed in the glass transition temperature of the samples with in-creasing salt concentration. There was no evidence of melting or recrystalli-zation for the samples up to a temperature of 80 °C. Viscosity measurements also showed first a maximum and then a minimum in the viscosity of the electrolytes with increasing salt concentration. At room temperature, viscos-ity values of the PEI:LiTFSI electrolytes of around 50 Pa.s were obtained. The fluidity of the PEI:LiTFSI electrolytes was modeled by the Bingham plastic equation. The temperature dependence of the viscosity followed Ar-rhenius behavior and the activation energies of the polymer chain motion were about 50 kJ/mol. Maximum ionic conductivity was 2×10-6 S/cm at room temperature for the molar ratio of 50:1. The ionic conductivity showed Arrhenius behavior with temperature. The activation energy of the conduc-tion mechanism was obtained between 47 and 59 kJ/mol at different salt concentrations. For PEI:LiTFSI electrolytes direct transmittance was high for λ < 1500 nm while there were absorption bands due to PEI at λ > 1500 nm.

Molar conductivity of the PEI:LiTFSI electrolytes displayed first a mini-mum and then a maximum with increasing salt concentration. This is gener-ally seen in molar conductivity behavior of polymer electrolytes as a func-tion of salt concentration. A similarity between the behavior of the molar conductivity and the behavior of the glass transition temperature and viscos-ity of the electrolyte was seen. It was explained by the major influence of the segmental motion of the polymer chains on the ion mobility. The reason for the decrease of segmental mobility was based on an increase in the number

140

of transient crosslinks. The influence of charge carrier concentration on con-ductivity was investigated by eliminating mobility effects. For this purpose, analyses of the Walden product and the iso-viscosity conductivity were car-ried out. It was concluded that below concentrations of 0.1 mol/kg, many ions formed contact ion-pairs and at concentrations higher than 0.1 mol/kg, molar conductivity is mainly influenced by mobility.

Bulk permittivities of PEI:LiTFSI electrolytes had similar values at dif-ferent salt concentrations. The dielectric constant values were increased with salt concentration with the maximum value of 18. Below a frequency of 1 kHz, electrode polarization was seen in the real part of the permittivity. From the temperature dependence of ionic relaxation of the electrolytes, activation energy was obtained between 46 and 57 kJ/mol. In addition, it was seen that the bulk response of the electrolytes followed the time-temperature superposition relation. The role of ion-pair relaxation and ionic conductivity in the ion transport process was examined for the PEI:LiTFSI electrolytes. The electrolytes were successfully modeled with the conductive Havriliak-Negami circuit model, which has an element that includes ion-pair relaxation and ionic conductivity in a single process. It was seen that the experimental data have good agreement with the Barton-Nakajima-Namikawa relation. The results of (i) the good agreement with the Barton-Nakajima-Namikawa relation, (ii) the same values of the activation energies of the relaxation, and conduction, and (iii) the good modeling by the conductive Havriliak-Negami model, imply that there is a connection between ionic conductivity and ion-pair relaxation and they should be viewed as two parts of the same conduc-tion process.

In the second part of this study, nanocomposite PEI:LiTFSI electrolytes were characterized at various nanoparticle concentrations and temperatures. For this purpose, nanoparticles of SiO2, In2O3, ITO were used. The [N]:[Li] molar ratio of PEI:LiTFSI was kept constant at 50:1. The number of nanoparticles were counted as a fraction of the total weight of nanoparticle and PEI.

First, PEI:LiTFSI:SiO2 nanocomposite polymer electrolytes were investi-gated. Nanoparticles of 7-nm-diameter SiO2 were added to PEI:LiTFSI elec-trolytes up to 9 wt% of SiO2 nanoparticles. The bulk permittivity at different temperatures was related by the time-temperature superposition principle and was fitted to the conductive Havriliak-Negami circuit model. The tem-perature dependence of the ionic conductivity showed an Arrhenius relation-ship. The ionic conductivity increased monotonically about one order of magnitude for increasing SiO2 content. Glass transition temperature de-creased up to -67 °C in the presence of SiO2. The diffuse transmittance re-mained at ~1 % for SiO2 amounts up to 7 wt%, which was small enough not to be noticeable to the eye in the case of a window. At higher SiO2 content, an increase of the optical scattering and an accompanying decrease in the direct transmittance was observed. The conclusion to be drawn was that the

141

addition of inorganic nanoparticles can enhance ionic conductivity while not impeding optical clarity.

Ion conduction in PEI:LiTFSI:SiO2 containing 7 wt% of was investigated at different salt concentrations between 400:1 and 20:1. The ionic conductiv-ity had a maximum at the molar ratio 100:1, and the temperature dependence of the conductivity showed Arrhenius behavior. The molar conductivity de-viated from the typical behavior for low salt concentrations and first in-creased and then decreased as a function of increasing salt concentration. The Walden product first increased and then decreased with increasing salt concentration. In conclusion, a significant effect of the presence of SiO2 nanoparticles on both segmental flexibility and free ion concentration was observed for the PEI:LiTFSI electrolytes; and there is an optimum molar ratio 100:1 where the highest free-ion concentration and segmental flexibil-ity were obtained.

PEI:LiTFSI:In2O3 nanocomposite polymer electrolytes incorporating 8-nm-diameter In2O3 nanoparticles were studied up to 7 wt% of In2O3 nanoparticles. Ionic conductivity remained the same with the addition of In2O3. The temperature dependence of the ionic conductivity followed Ar-rhenius behavior and activation energy did not show significant changes. The presence of In2O3 nanoparticles increased the glass transition tempera-ture of the electrolytes up to -44 °C. Transmittance of the electrolytes re-mained the same. To sum up, the addition of In2O3 nanoparticles to PEI:LiTFSI electrolytes did not induce ionic conductivity and optical clarity of the electrolytes.

PEI:LiTFSI:ITO electrolytes were produced with ITO nanoparticles hav-ing an Sn/In ratio of 5 at%. The nanocomposite electrolytes were made with up to 7 wt% of ITO with diameters of ~13 nm. Ionic conductivity was al-most independent of the ITO content. Ionic conductivity as a function of temperature followed the Arrhenius behavior. The glass transition tempera-ture increased up to -44 °C with increasing ITO concentration. Adding ITO led to a strong reduction of the near-infrared transmittance while it remained high in the visible range. The transmittance measurements showed good agreement with an effective medium theory. In conclusion, without deterio-rating the conduction properties, low near-infrared transmittance can be ob-tained with PEI:LiTFSI:ITO electrolytes due to an absorption band with intensity depending on the amount of ITO.

A PEI:LiTFSI:SiO2:ITO electrolyte was prepared with 6 wt% of SiO2 and 3 wt% of ITO. The aim was to combine the properties of high ionic conduc-tivity that is provided by SiO2 for PEI:LiTFSI:SiO2 electrolytes and of near-infrared blocking that is provided by ITO for the PEI:LiTFSI:ITO electro-lytes. Low near-infrared transmittance with high visible transmittance was obtained with the PEI:LiTFSI:SiO2:ITO electrolyte. No increment at the ionic conductivity was achieved for the electrolyte. The glass transition tem-perature of PEI:LiTFSI:SiO2:ITO was about -50 °C. The results showed that

142

ITO nanoparticles had a more dominant effect on the conductivity of the electrolyte.

For ECW applications, more salt concentration was needed in the electro-lytes. To be able to increase the salt concentration, PC and EC were intro-duced into PEI:LiTFSI-based nanocomposite electrolytes. PEI:LiTFSI:PC:EC electrolytes were prepared with the weight ratios of 10:30:30:30, 13:37:25:25, 13:47:20:20, and 10:50:20:20. An increase in the ionic conductivity and a decrease in the glass transition temperature were obtained as expected. High transmittance of PEI:LiTFSI electrolytes pre-vailed and the characteristic minima of PEI:LiTFSI electrolytes became smaller in the presence of PC:EC in the electrolytes. 1 wt% of SiO2 and ITO were added to PEI:LiTFSI:PC:EC electrolytes with the composition of 13:47:20:20. Ionic conductivity and glass transition temperatures of the elec-trolytes were in the range of 10-5-10-6 S/cm and of -70 °C, respectively. High transparency was obtained for both electrolytes in the visible range and PEI:LiTFSI:ITO electrolytes absorbed the solar radiation in the near-infrared range.

Finally, the performance of ECWs with nanocomposite polymer electro-lytes was tested. PET/ITO/WO3/polymer electrolyte/NiO (or IrO2)/ITO/PET was the configuration of the ECWs.

First of all the different concentrations of PEI:LiTFSI:PC:EC electrolytes were used in ECWs based on WO3 and NiO electrochromic layers. Transmit-tance at the bleached state was reduced from ~70 to ~40 % in a few cycles and remained constant after that. Colored transmittance remained at ~18 % where it was fixed. Coloring time was ~40 s and bleaching time was 150 s. The PEI:LiTFSI:PC:EC electrolyte with the weight ratios of 13:47:20:20 was chosen for the rest of the study because of the preferable performance of the ECWs with this electrolyte.

A second group of the ECWs had PEI:LiTFSI:PC:EC:SiO2 and PEI:LiTFSI:PC:EC:ITO electrolytes and WO3 and NiO films. Transmittance at the bleached state decreased from ~70 to ~40 % in a few cycles and was stable for the rest of the cycles. Colored transmittance was ~20 %, where it was fixed. Coloring times were reduced from 74 to 31 s in the presence of the nanoparticles in the electrolyte. Bleaching time was 150 s.

A third group contained ECWs with non-crosslinked and crosslinked pro-prietary electrolytes that contained SiO2 nanoparticles. An improvement was observed in the switching dynamics of the ECWs in the presence of SiO2 nanoparticles in the electrolytes. There was no improvement in the switching time in the case of the SiO2 containing crosslinked electrolyte.

Finally, PEI:LiTFSI:PC:EC, PEI:LiTFSI:PC:EC:SiO2 and PEI:LiTFSI:PC:EC:ITO were used in ECWs with WO3 and IrO2 films. Bleached transmittance reached 70 % for all ECWs. Transmittance at the colored state was between 35-40 % in the first cycle and improved with an increasing number of cycles. Coloring time was reduced from 120 to 107

143

and 50 s in the presence of SiO2 and ITO in the electrolyte, respectively. The bleaching times of the ECWs were 110, 84, and 38 s with PEI:LiTFSI:PC:EC, PEI:LiTFSI:PC:EC:SiO2 and PEI:LiTFSI:PC:EC:ITO electrolytes, respectively.

145

13. Summary in Swedish: Funktionalisering av polymera elektrolyter för tillämpningar i elektrokroma fönster

Den höga energianvändningen har fångat folks uppmärksamhet och ökat intresset för nya energikällor, energibesparande åtgärder och metoder för att minska efterfrågan på energi. Vi människor tillbringar större delen av vår tid inomhus och 30 till 40 % av energin i världen används i byggnader. Energin används exempelvis för uppvärmning, kylning, belysning och ventilation. Fönster orsakar energiförluster i byggnader genom att värme överförs ifrån insidan till utsidan, eller vice versa beroende på det rådande utomhusklima-tet. Dagsljus har positiva effekter på människors välbefinnande och humör, så att leva utan fönster är ingen lösning för att minska energianvändningen. Istället kan byggnaders energibehov minskas genom att använda fönster som med hjälp av varierbar transmittans kontrollerar inkommande solenergi. Det är därmed möjligt att antingen tillåta eller hindra solenergin att passera in i byggnaden beroende på om det föreligger ett värmebehov eller ett kylbehov. Material som kan ändra färg kallas för kromgena material. Ett elektrokromt fönster, vilket gör det möjligt att reglera inflödet av synligt ljus och solenergi i byggnaden, är en lovande teknik för minskad energianvändning i byggna-der och förbättrad inomhuskomfort. Ett elektrokromt fönster färgas och bleks med hjälp av en låg spänning. Den pålagda spänningen resulterar i att laddningar introduceras i eller extraheras ifrån det elektrokroma materialet. Det är därigenom möjligt att styra inkommande solenergi, samtidigt som utsikten genom fönstret bibehålls.

Figur 1 visar en elektrokrom fönsterruta innehållande fem skikt som pla-cerats mellan två transparenta substrat. Ett transparent jonledande skikt, vil-ket kallas elektrolyt, ligger i mitten av strukturen. Skikt till höger om elektro-lyten fungerar som anod och skikt till vänster som katod. På vänster sida av elektrolytskiktet, finns en tunn film av ett elektrokromt material såsom WO3. Det är en katodisk film och den färgas därför vid interkalation av joner i filmen. På höger sida av elektrolytskiktet finns ett jonlagringsskikt. Detta skikt kan antingen bestå av ett elektrokromt material, såsom NiO och IrO2, eller ett icke-elektrokromt material, såsom CeO2. För att få högsta möjliga moduleringen i transmittans används företrädelsesvis elektrokroma filmer i jonlagringsskiktet. Detta anodskikt färgas genom extraktion av joner från

146

filmen. De tre beskrivna skikten är placerade mellan två transparenta elekt-riskt ledande tunna filmer, såsom exempelvis indium tenn oxid ( ITO).

Figur 1. Schematisk bild av en elektrokrom fönsterruta.

Polymera elektrolyter används som jonledande skikt i elektrokroma fönster såväl som i batterier, bränsleceller, superkondensatorer, sensorer och nano-porösa infärgade solceller. Polymerelektrolyter spelar en betydande roll i elektrokroma fönsters funktion. Grundläggande krav för polymera elektroly-ter i elektrokroma fönster är att de ska ha hög jonisk ledningsförmåga och hög transmittans. Undersökningar rörande jonledning och optiska egenska-perna hos polymera elektrolyter är därför av stor betydelse. En enkel poly-merelektrolyt innehåller en polymer och ett salt, där polymeren är ett medi-um för saltet. Polymerer är makromolekyler som består av upprepade ke-miskt bundna monomerer. En monomer är en högmolekylärt viktad enhet av små molekyler. Proteiner och nukleinsyror, cellulosa, ull, silke och bärnsten är några exempel på naturliga polymerer. År 1929 börjades framställningen av syntetiska polymerer, som är konstgjorda polymerer. Många material så som Teflon, plexiglas och olika sorters plaster är tillverkade ifrån syntetiska polymerer. Salt består av negativa och positiva joner. En jon är en atom eller en molekyl, där det totala antalet protoner och elektroner skiljer sig åt. Om jonen innehåller fler protoner än elektroner, kallas den katjon. Om det finns fler elektroner än protoner, kallas den en anjon. Olika typer av kombinatio-ner av jonerna kan uppstå när ett salt löser sig i en elektrolyt. Katjoner och

147

anjoner kan dissociera och uppträda oberoende av varandra. Katjoner och anjoner kan också interagera med varandra och bilda jonpar, jon-trillingar och högre ordningens kombinationer. Förutom jon-jon-interaktioner, kan jonerna interagera med polymerkedjorna och bilda transienta tvärbindningar, vilket medför en minskning i rörlighet mellan polymerkedjorna. Alla dessa interaktioner påverkar jonkonduktivitet i en polymer elektrolyt.

Figur 2. Jon-jon interaktioner: (a) fria joner (b) jonpar (c-d) jon-trillingar och (e-f) jon-polymer interaktioner.

Nanopartiklar är material med en diameter som är mindre än 100 nm, och har andra egenskaper än materialets bulkform. Nanokomposita polymera elektrolyter, polymera elektrolyter som innehåller nanopartiklar, har använts i många nya elektrolytstudier. Att tillsätta nanopartiklar i polymera elektro-lyter kan förbättra de fysikaliska egenskaperna såsom elektriska, mekaniska och termiska egenskaper. SiO2 nanopartiklar är några av de mest använda nanopartiklarna i nanokomposita polymera elektrolyter. Förbättrad jonled-ning, elektrokemisk stabilitet, mekaniska och termiska egenskaper gör dem attraktiva för utveckling av polymerelektrolyter. Förbättrad jonledning hos polymera elektrolyter kan också leda till bättre prestanda hos elektrokroma fönster. Dessutom kan nanopartiklar bidra till nya funktioner hos polymera elektrolyter. Nanopartiklar av ITO i en polymermatris kan ge absorption i det nära infraröda våglängdsområdet och hög synlig transmittans utan märk-bar ljusspridning. Därför kan tillämpningen av ett ITO-dopat polymersystem i fönster blockera solvärme och bidra till hög dagsljusprestanda, och därmed förbättra fönstrens energieffektivitet. Nanopartiklar som blockerar nära-infraröd strålning, bland annat ITO, är lovande för tekniska tillämpningar.

CH2 Syre Katjon Anjon

(e) (f)

+ _ _ + _

+ + _ _

(a) (b) (c) (d)

+

148

I den här studien, karaktäriserades polyetylenimin:LiN(CF3SO2)2 (PEI: LiTFSI)-baserade polymerelektrolyter med hjälp av dielektrisk / impe-dans spektroskopi, differentiell svepkalorimetri, viskositet, optisk spektro-skopi, och elektrokroma mätningar. Nanopartiklar av SiO2, In2O3, och In2O3:Sn tillsattes för att förbättra och funktionalisera elektrolytens fysika-liska egenskaper. Dessutom utvärderades prestandan hos elektrokroma föns-ter som inkluderade PEI-baserade elektrolyter.

I första delen av studien, karaktäriserades PEI:LiTFSI elektrolyter vid olika salthalter och temperaturer. Den temperaturberoende viskositen och jonkonduktiviten hos elektrolyterna följde Arrhenius beteende. Viskositeten modellerades och uppvisade plastiskt Bingham beteende. Den molära kon-duktiviteten, glasövergångtemperaturen, viskositen, Walden produkten och isoviskösa konduktivitetsanalysen påvisade segmentell flexibilitet, jonpar och rörlighet för ledningsförmågan. En koppling mellan jonisk ledningsför-måga och jonparrelaxation sågs genom (i) Barton-Nakajima-Namikawa ek-vation, (ii) aktiveringsenergier i bulkrelaxation och jonledning och (iii) att jämföra två motsvarande kretsmodeller innehållande olika typer av Havrili-ak-Negami element, för bulkrespons.

I andra delen, undersöktes nanokomposita PEI: LiTFSI elektrolyter med SiO2, In2O3 och In2O3:Sn (ITO). Genom att tillsätta SiO2 till PEI:LiTFSI förbättras jonledningsförmågan 10 gånger utan att de optiska egenskaperna försämras. Vidare diskuterades effekten av segmentell flexibilitet och fri jonkoncentration på ledningsbandet i närvaro av SiO2 diskuterades. PEI:LiTFSI:ITO elektrolyter hade hög ljustransmittans utan ljusspridning och stark nära infraröd absorption utan minskad jonledningsförmåga. Jon-ledningsförmågan och den optiska klarheten försämrades inte för PEI:LiTFSI:In2O3 och PEI:LiTFSI:SiO2:ITO elektrolyter.

Slutligen tillsattes propylenkarbonat (PC) och etylenkarbonat (EC) till PEI:LiTFSI för att utföra elektrokroma mätningar. ITO och SiO2 tillsattes till PEI:LiTFSI:PC:EC och patentskyddade elektrolyter. De nanokomposita elektrolyterna testades för eletrokroma fönster med konfigurationen PET/ITO/WO3/polymer elektrolyt/NiO (eller IrO2)/ITO/PET. Utvärderingen visade att tillsatsen av nanopartiklar till polymera elektrolyter kan förbättra dynamiken för elektrokroma fönsters färgning och blekning.

Den här studien visar att nanokomposita polymera elektrolyter kan bidra till nya funktioner och förbättrad prestanda när de tillämpas i elektrokroma fönster.

149

14. Summary in Turkish: Polimer elektrolitlerin elektrokromik cam uygulamaları için fonksiyonelleştirilmesi

Enerji kaynaklarının gün geçtikçe tükenmesi, insanları yeni enerji kaynaklarının, daha az enerji tüketiminin ve enerjiye olan ihtiyacı azaltmanın yollarının araştırılmasına yöneltmiştir. İnsanlar günün çoğunu bina ve araç içi gibi kapalı alanlarda geçirirler. Dünya’daki enerjinin %30-40 kadarı binalarda ısıtma, soğutma, aydınlatma ve havalandırma gibi amaçlar için harcanır. Mevsimlere göre, dışarıdan içeriye ya da içeriden dışarıya olan ısı transferi en fazla pencerelerde kullanılan camlarda olur. Böylece sıcak mevsimlerde, bina içinin istenilenden fazla ısınmasına, soğuk mevsimlerde ise bina içinden istenmeyen ısı kaybına neden olurlar. Bu da klima kullanımını arttırır. Penceresiz yaşamak enerji tüketimini azaltmak için bir çözüm değildir çünkü gün ışığı insanlarda pozitif bir ruh hali sağlar. Binalarda daha az enerji tüketimi, gelen güneş enerjisini istenilen oranda geçiren camların kullanılması ile sağlanabilir. Böylece, isteğe göre güneş enerjisinin içeri girişi kontrol edilebilir. Değişen geçirgenlik özelliği olan, bir başka deyişle renklenme özelliği olan malzemelere kromik malzeme adı verilir. Elektrokromik camlar, binalarda içeriye giren güneş enerjisinin miktarını ayarlayabilme özelliği sayesinde enerji tüketimini azaltabilecek önemli teknolojilerden biridir. Elektrokromik camlar tersinir olarak uygulanan 1-2 V’luk düşük bir voltaj ile şeffaflaşır ya da renklenir. Elektrokromizmin diğer kromizm teknolojilerinden farkı, şeffaflaşma ve renklenmenin isteğe bağlı olarak kontrol edilebilir olması ve sadece renklenirken ya da şeffaflaşırken voltaj uygulanmasıdır. Elektrokromik camlar içeri giren güneş enerjisini kontrol eder ve aynı zamanda gün ışığını almayı ve dışarıdaki manzarayı görmeyi mümkün kılar. Gün ışığının fazla olduğu anlarda renkli duruma getirelen elektroktomik camlar, gözlerin kamaşmasına sebep olacak derecede fazla olan gün ışığının belli bir miktarını bina içine geçirir ve böylece daha rahat bir ortam sağlar.

Şekil 1’de bir elektrokromik camın şematik görüntüsü verilmiştir. Elektrokromik cam, cam ya da plastik iki taşıyıcı arasındaki beş tabakadan oluşur. Bir elektrokromik camın merkezinde elektrolit olarak da adlandırılan şeffaf ve iyon iletken bir tabaka yer alır. Elektrolit tabakasının bir tarafında, içerisine iyon girişi ile renklenen (tungsten oksit gibi) elektrokromik malze-

150

meden yapılmış bir ince film yer alır. Elektrolit tabakasının diğer tarafında ise iyon depolayıcı tabaka bulunur. Bu tabaka için genellikle iyon çıkışı ile renklenen (nikel oksit ya da iridium oksit) gibi bir elektrokromik malzeme-den yapılmış film kullanılır. Bir yüzeyi elektrolit ile temas halinde olan elek-trokromik filmlerin diğer yüzeyi elektronik iletkenlik sağlayan (indiyum-kalay oksit (ITO)) gibi bir malzemeden yapılmış bir ince film ile temas ha-lindedir.

Şekil 1. Bir elektrokromik camın şematik görünümü.

Polimer elektrolitler, elektrokromik camlarda, pillerde, yakıt pillerinde, süper kapasitörlerde, sensörlerde ve boya ile duyarlı hale getirilmiş güneş pillerinde iyon iletken tabaka olarak kullanılır. Söz konusu elektrolitler elektrokromik camların çalışmasında önemli rol oynar. Polimer elektrolit-lerin elektrokromik camlarda kullanıldığında sağlaması gereken başlıca iki özellik, yüksek iyonik iletkenlik ve yüksek geçirgenliktir. Bu nedenle poli-mer elektrolitlerin iyonik iletkenlikle birlikte optik özelliklerinin de ince-lenmesi büyük önem taşır. Saf olarak nitelendireceğimiz en sade polimer elektrolit bir tür polimer ve bir tür tuzdan oluşur. Polimer, tuza bir ortam oluşturur. Polimerler; birbirini tekrar eden, monomer olarak adlandırılan yüksek molekül ağırlıklı küçük moleküllü yapılarının kimyasal bağlarla bağlanması sonucu oluşmuş makromoleküllerdir. Proteinler ve nükleik asit-ler, selüloz, yün, ipek ve kehribar doğal polimerlere örnektir. 1929 yılından itibaren doğal polimerlerin yanı sıra, sentetik polimer olarak sınıflandırılan

151

ve teflon tava, pleksi cam, plastik malzemelerin yapımında da kullanılan polimerler insan yapımı polimerler olarak hayatımıza girmiştir. Tuzlar, nega-tif ve pozitif iyonlardan oluşurlar. İyonlar, toplam elektron ve proton sayısı eşit olmayan yani nötr olmayan atom ya da moleküllerdir. Proton sayısı elek-tron sayısından daha çok olan iyonlar pozitif yüklüdür ve katyon olarak ad-landırılır. Elektron sayısı proton sayısından çok olan iyonlar ise anyon olarak isimlendirilir. Bir polimer elektrolitte; iyonlar kendi aralarında ve polimer ile çeşitli etkileşmeler yapabilir. Katyon ve anyonlar polimerin içinde çözünür ve Şekil 2 (a)’da görüldüğü gibi birbirinden bağımsız olarak hareket edebilirler. Katyon ve anyon birbiriyle etkileşip, iyon-çifti (Şekil 2 (b)), iyon üçlüsü (Şekil 2 (c) ve (d)), ya da daha yüksek mertebeli kombinasyonlar oluşturabilirler. Şekil 2 (e) ve (f)’de görüldüğü gibi, iyonlar polimerle de etkileşebilir. İyonların polimer zincirleriyle bu etkileşmesi geçici çapraz bağlı yapılar oluşturarak polimer zincirlerinin hareketliliğini azaltır. Tüm bu iyon-iyon ve iyon-polimer etkileşmeleri polimer elektrolitin iyonik iletkenliğini etkiler.

Şekil 2. İyon-iyon etkileşmeleri: (a) serbest iyonlar (b) iyon çiftleri (c-d) iyon üçlüleri, ve (e-f) iyon-polimer etkileşmesi.

Günümüzde nanoteknolojinin bir ürünü olan, milimetrenin milyonda biri boyutunda olan nanoparçacıklar hayatın bir çok alanına yenilikler ve kolay-lıklar getirmiştir. Nanoparçacık içeren nanokompozit polimer elektrolitler üzerine yapılan çalışmalarda, polimer elektrolitlere nanoparçacık eklemenin, elektrolitin elektriksel, mekanik, termal özellikler gibi fiziksel özelliklerini iyileştirdiği görülmüştür. Silisyum oksit katkılanmış polimer elektrolitler bu gruba bir örnektir. Polimer elektrolitin iyonik iletkenliğinin artması, elektrokromik camın performansını da arttırır. Nanoparçacıklar, polimer

CH2 Oksijen Katyon Anyon

(e) (f)

+ _ _ + _

+ + _ _

(a) (b) (c) (d)

+

152

elektrolitlerin fiziksel özelliklerini iyileştirmenin yanı sıra polimer elektro-litin fonksiyonelleştirilmesini de sağlayabilir. Örneğin; bir polimerin içinde yeralan indiyum-kalay oksit, güneş ışınımının görünür bölgesindeki kısmı geçirirken yakın-kızılötesi kısmındaki ışımayı bloke eder. Böylece, ITO-katkılanmış polimerik sistemlerin camlarda uygulanması görüntüde bir değişikliğe sebep olmadan güneş ısısını bloke eder. Bu sayede, sıcak iklim-lerde güneş ısısının içeriye girmesi engellenmesi sonucu daha az klima kul-lanımı, dolayısı ile enerji tasarrufu sağlanabilir. Bu ilgi çekici özelliğe rağ-men, ITO içeren polimer elektrolitler üzerine bugüne kadar yapılmış bir çalışmaya rastlanmamıştır.

Bu çalışmada, silisyum oksit, indiyum oksit, ve indiyum-kalay oksit ile katkılanmış, polietilenimin (PEI) polimeri ve LiN(CF3SO2)2 (LiTFSI) tuzun-dan oluşan nanokompozit polimer elektrolitlerin karakterizasyonu ile elektrokromik camlarda uygulanması yer almaktadır. Karakterizasyon tek-nikleri olarak, dielektrik/empedans spektroskopi, diferansiyel taramalı kalo-rimetri, vizkozite, optik spektroskopi ve elektrokromik ölçümlerden yarar-lanılmıştır.

Birinci bölümde, saf PEI:LiTFSI elektrolitleri farklı LiTFSI konsantrasyonlarında ve farklı sıcaklıklarda incelenmiştir. Bu elektrolitlerin iyonik iletkenlik ve vizkozitelerinin sıcaklığa Arrhenius eşitliğine göre bağlı olduğu görülmüştür. Viskozite, Bingham eşitlikleri ile temsil edilmiştir. Molar iletkenlik, camsı geçiş sıcaklığı, viskozite, Walden kuralı, ve izo-vizkoziteye dayanan analizler, polimer zincir esnekliğinin, iyon-çiftlerinin ve mobilitenin PEI:LiTFSI elektrolitlerinin iletkenliklerinin üzerinde etkili olduklarını göstermiştir. PEI:LiTFSI elektrolitlerinin iyonik iletkenliği ile iyon-çiftlerinin relaksasyonu arasında bir ilişki olduğu (i) Barton-Nakajima-Namikawa eşitliğinden, (ii) relaksasyon ve iletkenliğin aktivasyon enerjilerinin karşılaştırılmasından (iii) farklı tip Havriliak-Negami bileşenleri içeren iki eşdeğer devrenin karşılaştırılmasından görülmüştür.

Çalışmanın ikinci kısmında silisyum oksit, indiyum oksit, ya da ITO ile katkılanmış nanokompozit PEI:LiTFSI elektrolitleri incelenmiştir. PEI:LiTFSI elektrolitlerine silisyum oksit eklemek elektrolitin optik özelliklerini bozmadan iyonik iletkenliğini 10 kat arttırmıştır. Silisyum oksitin PEI:LiTFSI elektrolitinin içerisinde yeralmasının polimer zincirinin esnekliği ve serbest iyon konsantrasyonu üzerindeki etkileri incelenmiştir. PEI:LiTFSI elektrolitlerinin ITO nanoparçacıkları ile katkılandırılmasıyla iyon iletkenliği değişmemiş, optik olarak görünür bölgede yüksek geçirgenlik sağlayan ve yakın-kızılötesi bölgede geçirgenliği engelleyen bir elektrolit elde edilmiştir. İndiyum oksit ve ITO:silisyum oksit katkılanmış PEI:LiTFSI elektrolitlerinin iyonik iletkenliklerinde ve optik özelliklerinde bir değişiklik olmamıştır.

Son olarak, elektrolitler elektrokromik camlarda kullanılıp, bu elektrokromik camların elektrokromik performansı test edilmiştir. Bu amaçla, elektrolitlerdeki iyon sayısını arttırmak için, PEI:LiTFSI bazlı elek-

153

trolitlere propilen karbonat ve etilen karbonat eklenmiştir. ITO ya da silisy-um dioksit, PEI:LiTFSI:propilen karbonat:etilen karbonat elektrolitlere ve silisyum dioksit, patentli bir elektrolite eklenmiştir. Elektrokromik camların yapısı plastik/ITO/tungsten oksit/polimer elektrolit/ nikel (ya da iridyum) oksit/ITO/plastik şeklindedir. Bu çalışmalarda, polimer elektrolite nano-parçacık eklemenin elektrokromik camların renklenme ve/veya şeffaflaşma sürelerini kısaltarak renklenme ve şeffaflaşma dinamiğini iyileştirdiği gö-rülmüştür.

Bu çalışma ile, polimer elektrolitlerin nanoparçacıklar ile katkılandırıl-ması sonucu elektrokromik camların performansının arttırılabileceği, elektrokromik camlar içinde yer alan polimer elektrolit tabakanın sağlaması gereken temel gereksinimlerin yanısıra yeni özellikler katılabileceği göste-rilmiştir.

155

Acknowledgements

This project was initiated through the contribution, help and support of many people. I was just at a junction in my studies. I am grateful to all the people in my life for their direct or indirect contributions to this thesis.

First and foremost, I would like to express my sincere gratitude to my ad-visor Professor Gunnar A. Niklasson for his continuous support of my PhD studies and for his patience, guidance and extensive knowledge. I would also like to express my deepest appreciation to my co-advisor Professor Claes-Göran Granqvist, the reason why I decided to do my PhD in Uppsala, for his enthusiasm, directing and motivating to achieve the goals. I must extend special thanks to Ruse Marsal at ChromoGenics for her valuable and con-structive suggestions during the planning and development of these studies and for her endless support and friendship. I am grateful to Peter Georén, who was my industrial advisor for the first year of my studies, for his guid-ance. I would like to convey thanks to Delia Milliron at the Lawrence Berke-ley National Laboratory for making her laboratory available for the synthesis of ITO and In2O3 nanoparticles. I would also like to thank Evan Runner-strom at the Lawrence Berkeley National Laboratory for guidance on the synthesis of ITO and In2O3 nanoparticles and for his friendship. I would like to thank Shuyi Li at the Solid State Physics Division for her contribution to the ITO-related study. I wish to thank Arne Roos at the Solid State Physics Division for discussions on optical measurements. Many thanks should go to Annica Nilsson at the Solid State Physics Division for help with the Swedish summary and for introducing me to the optical measurement setup. I wish to express my gratitude to Tim Bowden and Torbjörn Gustafsson at the De-partment of Materials Chemistry for making their laboratory available for DSC and TGA measurements. My grateful thanks are also extended to ChromoGenics’ staff, specifically to Thomas Almesjö, Greger Gregard, Esteban Avendano, and Sophie von Kraemer for their valuable comments, and to Latifa Bikndarne and Åsa Engström for their technical support regard-ing electrochromic measurements. I would also like to acknowledge with much appreciation the crucial technical support of Shuxi Zhao, Andreas Mattsson, Bengt Götesson and Jonatan Bagge. I also thank Inger Ekberg, Ingrid Ringård, Maria Skoglund and Sebastian Alonso for their help in ad-ministrative matters. Inger: I was not aware of how much work you were doing for us until you retired. I would like to thank the Swedish Research Council for Environment, Agricultural Sciences and Spatial Planning

156

(FORMAS) for financial support. I would also like to thank the Anna Maria Lundins stipendiefond at Smålands Nation for financial support for scientific trips. Special thanks should be given to Patric Jannasch at Lund University for his valuable comments on my study. A special thank you goes to my ex-advisor Professor Fatma Z. Tepehan at İstanbul Technical University for encouraging me to do my PhD study at Uppsala University.

Life in Uppsala was not only studying. Many thanks go to all members of the Solid State Physics Division for providing a nice working environment. I could not have imagined having a better environment for my PhD studies. Special thanks to Pia Lansåker for help on how to keep calm during trouble-some times and for her efforts regarding my first and probably last riding experience: to Sara Green for spreading her positive energy: to Shuyi Li and Malin Johansson for being curious: to Umut Çindemir and Serkan Akansel for updating me about what’s going on in life: and many of my other friends in the division, especially Zareh Topalian and Bozhidar Stefanov, for asking me everyday “how is writing your thesis going?”. I would also like to thank the Turkish people, with who I met during my period of study at the Law-rence Berkeley National Laboratory, for their warm interest. My sincere thanks go to my dear friends in İstanbul: Pelin Demirçivi, Müge Balkaya, Selin Sunay Yapışkan, and Caner Yürüdü for making me feel at home through their uninterrupted moral support.

I know that I have a big family here in Uppsala. I would like to thank Yılmaz, Gülen, Çetinkaya, Yaldır, Karakoç, Flener, Granberg, Zal, Ay-doğdu, Sorar, Tan, and Turan families for making my life meaninful during the PhD period. I want to express my warm thanks to Songül Aynal, Nigar Oturakçı, Tarja and Pauli Volotinen, Uğurlu family, Kim Cheng, Soorej Jose (does your name really fit in this sentence Soorej? :)), and Aynur Abish for the beautiful memories. I would especially thank Tuna Eken for introducing us to many of these nice people. Special thanks should go to the magnificent trio: Esra Bayoğlu Flener, Fatma Gülen Yaldır, and “kankaşkım” Arzu Güneş Granberg. How can I describe your efforts to make me feel better! I wish to express my love and gratitude to Esat Pehlivan for encouraging me to do my PhD at Uppsala University, for his understanding and patience during this study and for his technical support on thesis writing.

Hayatımı anlamlı kılan, tüm aile bireylerime, bugünlere gelmemi sağladıkları, her zaman yanımda oldukları, hiçbir zaman desteklerini esirgemedikleri ve en önemlisi hayatıma anlam kattıkları için sonsuz teşekkürler ederim. Sevgili babacığım, doktora yapmaktan vazgeçtiğim dönemde beni bu fikrimden vazgeçiren sendin. Bugün tamamladığım bir doktora çalışmam varsa bunun en önemli dönüm noktasında sen varsın. Annem! Benim için yaptığın herşey, çektiğin tüm zorluklar ve gösterdiğin sabır çok teşekkür ederim. Hepiniz iyi ki varsınız!

Canım babaannem! Çocuklarının okuması konusundaki büyük arzun, eğitim konusunda her zaman benim en büyük motivasyonum olmuştu.

157

Benim için yaptıklarının ve sevginin karşılığı hiçbir şey olamaz. Umarım seni bir nebze de olsa mutlu edebilmişimdir. Her zaman, hem hep yanımda hissettiğim hem de hep çok özlediğim olacaksın.

159

References

1. UNEP, Buildings and Climate Change: Status, Challenges, and Opportunities, United Nations Environmental Programme, Paris, 2007.

2. C.M. Lampert, Large–area smart glass and integrated photovoltaics, Sol. En-ergy Mater. Sol. Cells 76 (2003) 489–499.

3. A. Georg, A. George, W. Graf, V. Wittwer, Switchable windows with tungsten oxide, Vacuum 82 (2008) 730–735.

4. C.M. Lampert, Chromogenic Switchable Glazing: Towards the Development of the Smart Window, Conference Proceedings of Window Innovations, 1995, To-ronto, Canada.

5. G.H. Brown (Ed.), Photochromism, Wiley–Interscience, New York, 1971. 6. H. Dürr, H. Bouas–Laurent (Eds.), Photochromism: Molecules and Systems,

Elsevier, Amsterdam, 1990. 7. J.C. Crano, R.J. Guglielmetti, Organic Photochromic and Thermochromic

Compounds, Volume 1: Photochromic Families, Kluwer Academic Publishers, USA, 1999.

8. C.G. Granqvist, Window coatings for the future, Thin Solid Films 193–194 (1990) 730–741.

9. D. Ruzmetov, S. Ramanathan, Metal–Insulator Transition in Thin Film Vana-dium Dioxide, in: S. Ramanathan (Ed.), Thin Film Metal Oxides, Springer, 2010, pp. 51–94.

10. N.R. Mlyuka, G.A. Niklasson, C.G. Granqvist, Thermochromic multilayer films of VO2 and TiO2 with enhanced transmittance, Sol. Energy Mater. Sol. Cells 93 (2009) 1685–1687.

11. M. Saeli, C. Piccirillo, I.P. Parkin, I. Ridley, R. Binions, Nano–composite ther-mochromic thin films and their application in energy–efficient glazing, Sol. En-ergy Mater. Sol. Cells 94 (2010) 141–151.

12. A. Roos, M.–L. Persson, W. Platzer, M. Köhl, Energy Efficiency of Switchable Glazing in Office Buildings, in: Proceedings of The Glass Processing Days, Tampere, Finland, 2005, pp. 566–569.

13. E.S. Lee, D.L. DiBartolomeo, S.E. Selkowitz, Daylighting control performance of a thin–film ceramic electrochromic window: Field study results, Energy Build. 38 (2006) 30–44.

14. E.S. Lee, S.E. Selkowitz, R.D. Clear, D.L. DiBartolomeo, J.H. Klems, L.L. Fernandes, G.J. Ward, V. Inkarojrit, M. Yazdanian, Advancement of Electro-chromic Windows, California Energy Commission, PIER, 2006.

15. R.D. Clear, V. Inkarojrit, E.S. Lee, Subject responses to electrochromic win-dows, Energy Build. 38 (2006) 758–779.

16. C.G. Granqvist, Oxide electrochromics: Why, how, and whither, Sol. Energy Mater. Sol. Cells 92 (2008) 203–208.

17. D.–T. Gillaspie, R.C. Tenent, A.C. Dillon, Metal–oxide films for electrochromic applications: Present technology and future directions, J. Mater. Chem. 20 (2010) 9585–9592.

160

18. G.B. Smith, C.G. Granqvist, Green Nanotechnology: Solutions for Sustainabil-ity and Energy in the Built Environment, CRC Press, Boca Raton, 2011.

19. S.E. Selkowitz, C.M. Lampert, Application of large–area chromogenics to architectural glazings, in: C.M. Lampert, C.G. Granqvist (Eds.), Large Area Chromogenics: Materials and Devices for Transmittance Control, SPIE Optical Engineering Press, Bellingham, 1990, vol. IS4, pp. 22–45.

20. L.G.S. Brooker, G.H. Keyes, D.W. Heseltine, Color and constitution: XI: An-hydronium bases of p–hydroxystyryl dyes as solvent polarity indicators, J. Am. Chem. Soc. 73 (1951) 5350–5356.

21. J.R. Platt, Wavelength formulas and configuration interaction in brooker dyes and chain molecules, J. Chem. Phys. 25 (1956) 80–105.

22. J.R. Platt, Electrochromism, a possible change of color producible in dyes by an electric field, J. Chem. Phys. 34 (1961) 862–863.

23. S.K. Deb, A novel electrophotographic system, Appl. Opt. Suppl. 3 (1969) 192–195.

24. J.S.E.M. Svensson, C.G. Granqvist, Electrochromic tungsten oxide films for energy efficient windows, Solar Energy Materials 11 (1984) 29–34.

25. C.M. Lampert, Electrochromic materials and devices for energy efficient win-dows, Solar Energy Materials 11 (1984) 1–27.

26. J.S.E.M. Svensson, C.G. Granqvist, Electrochromic coatings for “smart win-dows”, Solar Energy Materials 12 (1985) 391–402.

27. M.D. Ward, Near–infrared electrochromic materials for optical attenuation based on transition–metal coordination complexes, J. Solid State Electrochem. 9 (2005) 778–787.

28. A.L. Larsson, G.A. Niklasson, Infrared emittance modulation of all–thin–film electrochromic devices, Mater. Lett. 58 (2004) 2517–2520.

29. T.L. Rose, S. D’Antonio, M.H. Jillson, A.B. Kon, R. Suresh, F. Wang, A mi-crowave shutter using conductive polymers, Synth. Met. 85 (1997) 1439–1440.

30. C.G. Granqvist, Handbook of Inorganic Electrochromic Materials, Elsevier, Amsterdam, 1995.

31. N.I. Jaksic, C. Salahifar, A feasibility study of electrochromic windows in vehi-cles, Sol. Energy Mater. Sol. Cells 79 (2003) 409–423.

32. P.M.S. Monk, R.J. Mortimer, D.R. Rosseinsky, Electrochromism and Electro-chromic Devices, Cambridge University Press, UK, 2007.

33. K. Bewilogua, G. Bräuer, A. Dietz, J. Gäbler, G. Goch, B. Karpuschewski, B. Szyszka, Surface technology for automotive engineering, CIRP Ann. Manuf. Technol. 58 (2009) 608–627.

34. http://www.gentex.com/, The picture was taken at International Meeting on Electrochromism in 2012, http://www.ime–10.org/.

35. R. Baetens, B.P. Jelle, A. Gustavsen, Properties, requirements and possibilities of smart windows for dynamic daylight and solar energy control in buildings: A state–of–the–art review, Sol. Energy Mater. Sol. Cells 94 (2010) 87–105.

36. http://sageglass.com/portfolio/, 2013. 37. C.G. Granqvist, Electrochromic materials: Out of niche, Nat. Mater. 5 (2006)

89–90. 38. http://www.chromogenics.com/, 2013. 39. http://www.avesodisplays.com/, 2013. 40. H.J. Gläser, Large Area Glass Coating, Von ardenne anlagentechnik GmbH,

Dresden, Germany, 2000. 41. D. Mecerreyes, R. Marcilla, E. Ochoteco, H. Grande, J.A. Pomposo, R. Vergaz,

J.M.S. Pena, A simplified all–polymer flexible electrochromic device, Electro-chim. Acta 49 (2004) 3555–3559.

161

42. C.G. Granqvist, Transparent conductors as solar energy materials: A pano-ramic review, Sol. Energy Mater. Sol. Cells 91 (2007) 1529–1598.

43. P. Lansåker, Gold–based Nanoparticles and Thin Films: Applications to Green Nanotechnology, PhD thesis, Uppsala University, Uppsala, Sweden, 2012.

44. C.M. Lampert, Heat mirror coatings for energy conserving windows, Solar Energy Materials 6 (1981) 1–41.

45. P.P. Edwards, A. Porch, M.O. Jones, D.V. Morgan, R.M. Perks, Basic materials physics of transparent conducting oxides, Dalton Trans. 19 (2004) 2995–3002.

46. G. Carotenuto, M. Valente, G. Sciume, T. Valente, G. Pepe, A. Ruotolo, L. Nicolais, Preparation and characterization of transparent/conductive nano–composites films, J. Mater. Sci. 41 (2006) 5587–5592.

47. R.J. Mortimer, Electrochromic materials, Chem. Soc. Rev. 26 (1997) 147–156. 48. A.A. Argun, P.–H. Aubert, B.C. Thompson, I. Schwendeman, C.L. Gaupp, J.

Hwang, N.J. Pinto, D.B. Tanner, A.G. MacDiarmid, J.R. Reynolds, Multicol-ored electrochromism in polymers: Structures and devices, Chem. Mater. 16 (2004) 4401–4412.

49. G.A. Niklasson, C.G. Granqvist, Electrochromics for smart windows: Thin films of tungsten oxide and nickel oxide, and devices based on these, J. Mater. Chem. 17 (2007) 127–156.

50. V.K. Thakur, G. Ding, J. Ma, P.S. Lee, X. Lu, Hybrid materials and polymer electrolytes for electrochromic device applications, Adv. Mater. 24 (2012) 4071–4096.

51. A. Balan, D. Baran, G. Gunbas, A. Durmus, F. Ozyurt, L. Toppare, One poly-mer for all: benzotriazole containing donor–acceptor type polymer as a multi–purpose material, Chem. Commun. (2009) 6768–6770.

52. G.E. Gunbas, A. Durmus, L. Toppare, A unique processable green polymer with a transmissive oxidized state for realization of potential RGB–based electro-chromic device applications, Adv. Funct. Mater. 218 (2008) 2026–2030.

53. P.M.S. Monk, R.J. Mortimer, D.R. Rosseinsky, Electrochromism: Fundamen-tals and Applications, VCH Publishers, Weinheim, 1995.

54. M.A. De Paoli, W.A. Gazotti, Electrochemistry, polymers and opto–electronic devices: A combination with a future, J. Braz. Chem. Soc. 13 (2002) 410–424.

55. P.R. Somani, S. Radhakrishnan, Electrochromic materials and devices: Present and future, Mater. Chem. Phys. 77 (2002) 117–133.

56. N.M. Rowley, R.J. Mortimer, New electrochromic materials, Science Progress 85 (2002) 243–262.

57. F. Carpi, D. De Rossi, Colours from electroactive polymers: Electrochromic, electroluminescent and laser devices based on organic materials, Optic. Laser Tech. 38 (2006) 292–305.

58. N. Özer, C.M. Lampert, Electrochromic characterisation of sol–gel deposited coatings, Sol. Energy Mater. Sol. Cells 54 (1998) 147–156.

59. S. Heusing, D.–L. Sun, J. Otero–Anaya, M.A. Aegerter, Grey, brown and blue coloring sol–gel electrochromic devices, Thin Solid Films 502 (2006) 240–245.

60. E. Pehlivan, F. Z. Tepehan, G.G. Tepehan, Comparison of optical, structural and electrochromic properties of undoped and WO3–doped Nb2O5 thin films, Solid State Ionics 165 (2003) 105–110.

61. E. Özkan, S–H Lee, C.E Tracy, J.R. Pitts, S.K. Deb, Comparison of electro-chromic amorphous and crystalline tungsten oxide films, Sol. Energy Mater. Sol. Cells 79 (2003) 439–448.

62. E. Ozkan Zayim, I. Turhan, F.Z. Tepehan, N. Ozer, Sol–gel deposited nickel oxide films for electrochromic applications, Sol. Energy Mater. Sol. Cells 92 (2008) 164–169.

162

63. E. Pehlivan, Saf ve Katkılı Niobyum Pentoksit İnce Filmlerin Optik, Yapısal, Elektriksel ve Elektrokromik Özellikleri, PhD thesis, İstanbul Technical Univer-sity, İstanbul, Turkey, 2007.

64. E.D. Avendaño Soto, Electrochromism in nickel–based oxides: Coloration mechanisms and optimization of sputter–deposited thin films, PhD Thesis, Upp-sala University, Uppsala, Sweden, 2004.

65. A. Azens, L. Kullman, C.G. Granqvist, Ozone coloration of Ni and Cr oxide, Sol. Energy Mater. Sol. Cells 76 (2003) 147–153.

66. T. Ahmed, Optical Characterisation of Ozonated Electrochromic Nickel Vana-dium Oxide Flexible Thin Films, Examensarbete, Uppsala University, Uppsala, Sweden, 2009.

67. C.G. Granqvist, Electrochromic tungsten oxide films: Review of progress 1993–1998, Sol. Energy Mater. Sol. Cells 60 (2000) 201–262.

68. C.G. Granqvist, Electrochromic coatings and devices: Survey of some recent advances, Thin Solid Films 442 (2003) 201–211.

69. B. Valla, J.C. Tonazzi, M.A. Macedo, L.H. Dall'Antonia, M.A. Aegerter, M.A. Gomes, L.O. Bulhoes, Transparent storage layers for H+ and Li+ ions prepared by sol–gel technique, SPIE Optical Materials Technology for Energy Efficiency and Solar Energy Conversion 1536 (1991) 48–62.

70. J. Nagai, G. D. McMeeking, Y. Saitoh, Durability of electrochromic glazing, Sol. Energy Mater. Sol. Cells 56 (1999) 309–319.

71. L.–C. Chen, K.–C. Ho, Design equations for complementary electrochromic devices: Application to the tungsten oxide–prussian blue system, Electrochim. Acta 46 (2001) 2151–2158.

72. V. Di Noto, S. Lavina, G.A. Giffin, E. Negro, B. Scrosati, Polymer electrolytes: Present, past and future, Electrochim. Acta 57 (2011) 4–13.

73. M.A. Macêdo, M.A. Aegerter, Sol–gel electrochromic device, J. Sol–Gel Sci. Technol. 2 (1994) 667–671.

74. A. Daneo, G. Macrelli, P. Polato, E. Poli, Photometric characterization of an all solid state inorganic electrochromic large area device, Solar Energy Mater. So-lar Cells 56 (1999) 237–238.

75. R.B. Goldner, F.O. Arntz, T.E. Haas, d–Electrons and two active thin film de-vices for achieving a solar energy economy, Solar Energy Mater. Solar Cells 32 (1994) 421–428.

76. P.V. Ashrit, F.E. Girouard, V.–V. Truong, Fabrication and testing of an all–solid state system for smart window application, Solid State Ionics 89 (1996) 65–73.

77. S.J. Yoo, J.W. Lim, Y.–E. Sung, Improved electrochromic devices with an inorganic solid electrolyte protective layer, Sol. Energy Mater. Sol. Cells 90 (2006) 477–484.

78. X. Fu, Polymer electrolytes for electrochromic devices, in: C. Sequeira, D. Santos (Eds.), Polymer Electrolytes: Fundamentals and Applications, Wood-head, Sawston, UK, 2010.

79. C.G. Granqvist, Oxide electrochromics: An introduction to devices and materi-als, Sol. Energy Mater. Sol. Cells 99 (2012) 1–13.

80. C.G. Granqvist, P.C. Lansåker, N.R. Mlyuka, G.A. Niklasson, E. Avendaño, Progress in chromogenics: New results for electrochromic and thermochromic materials and devices, Sol. Energy Mater. Sol. Cells 93 (2009) 2032–2039.

81. C. Xu, C. Ma, X. Kong, M. Taya, Vacuum filling process for electrolyte in enhancing electrochromic polymer window assembly, Polym. Adv. Technol. 20 (2009) 178–182.

163

82. C.O. Avellaneda, D.F. Vieira, A. Al–Kahlout, S. Heusing, E.R. Leite, A. Pawlicka, M.A. Aegerter, All solid–state electrochromic devices with gelatin–based electrolyte, Sol. Energy Mater. Sol. Cells 92 (2008) 228–233.

83. E. Raphael, C.O. Avellaneda, M.A. Aegerter, M.M. Silva, A. Pawlicka, Agar–based gel electrolyte for electrochromic device application, Mol. Cryst. Liq. Cryst. 554 (2012) 264–272.

84. C.A. Nguyen, A.A. Argun, P.T. Hammond, X.H. Lu, P.S. Lee, Layer–by–layer assembled solid polymer electrolyte for electrochromic devices, Chem. Mat. 23 (2011) 2142–2149.

85. A. Al–Kahlout, D. Vieira, C.O. Avellaneda, E.R. Leite, M.A. Aegerter, A. Pawlicka, Gelatin–based protonic electrolyte for electrochromic windows, Ion-ics 16 (2010) 13–19.

86. M.M. Silva, P.C. Barbosa, L.C. Rodrigues, A. Gonçalves, C. Costa, E. Fortu-nato, Gelatin in electrochromic devices, Opt. Mater. 32 (2010) 719–722.

87. N. Mendoza, F. Paraguay–Delgado, L. Hechavarri, M.E. Nicho, H. Hu, Nanos-tructured polyethylene glycol–titanium oxide composites as solvent–free viscous electrolytes for electrochromic devices, Sol. Energy Mater. Sol. Cells 95 (2011) 2478–2484.

88. F.L. Souza, M.A. Aegerter, E.R. Leite, Performance of a single–phase hybrid and nanocomposite polyelectrolyte in classical electrochromic devices, Electro-chim. Acta 53 (2007) 1635–1642.

89. S. K. Deb, Optical and photoelectric properties and colour centres in thin films of tungsten oxide, Philos. Mag. 27 (1973) 801–822.

90. B.W. Faughnan, R.S. Crandall, Optical properties of mixed–oxide WO3/MoO3 electrochromic films, Appl. Phys. Lett. 31 (1977) 834–836.

91. L. Berggren, A. Azens, G.A. Niklasson, Polaron absorption in amorphous tungsten oxide films, J. Appl. Phys. 90 (2001) 1860–1863.

92. L. Berggren, Optical absorption and electrical conductivity in lithium interca-lated amorphous tungsten oxide films, PhD thesis, Uppsala University, Uppsala, Sweden, 2004.

93. K.Bange, Colouration of tungsten oxide films: A model for optically active coatings, Sol. Energy Mater. Sol. Cells 58 (1999) 1–131.

94. S.K. Deb, Opportunities and challenges in science and technology of WO3 for electrochromic and related applications, Sol. Energy Mater. Sol. Cells 92 (2008) 245–258.

95. H. Kaneko, K. Miyake, Y. Teramoto, Preparation and properties of reactively sputtered tungsten oxide films, J. Appl. Phys. 53 (1982) 3070–3075.

96. S. Green, Electrochromic Nickel–Tungsten Oxides, PhD thesis, Uppsala Univer-sity, Uppsala, Sweden, 2012.

97. S. Passerini, B. Scrosati, Electrochromism of thin–film nickel oxide electrodes, Solid State lonics 53–56 (1992) 520–524.

98. J. Backholm, A. Azens, G.A. Niklasson, Electrochemical and optical properties of sputter deposited Ir–Ta and Ir oxide thin films, Sol. Energy Mater. Sol. Cells 90 (2006) 414–421.

99. J. Backholm, G.A. Niklasson, Optical properties of electrochromic iridium oxide and iridium–tantalum oxide thin films in different colouration states, Sol. Energy Mater. Sol. Cells 92 (2008) 1388–1392.

100. F. Jiang, T. Zheng, Y. Yang, Preparation and electrochromic properties of tungsten oxide and iridium oxide porous films, J. Non–Cryst. Solids 354 (2008) 1290–1293.

164

101. K. Nishio, Y. Watanabe, T. Tsuchiya, Preparation and properties of electro– chromic iridium oxide thin film by sol–gel process, Thin Solid Films 350 (1999) 96–100.

102. L.F. Mattheiss, Electronic structure of RuO2, OsO2, and IrO2, Phys. Rev. B 13 (1976) 2433–2450.

103. C. Bechinger, E. Wirth, P. Leiderer, Photochromic coloration of WO3 with visible light, Appl. Phys. Lett. 68 (1996) 2834–2836.

104. C.G. Granqvist, S. Green, E.K. Jonsson, R. Marsal, G.A. Niklasson, A. Roos, Z. Topalian, A. Azens, P. Georén, G. Gustavsson, R. Karmhag, J. Smulko, L.B. Kish, Electrochromic foil–based devices: Optical transmittance and modulation range, effect of ultraviolet irradiation, and quality assessment by 1/f current noise, Thin Solid Films 516 (2008) 5921–5926.

105. C.M. Lampert (Ed.), Failure and Degradation Modes in Selected Solar Materi-als: A Review (pp.19, 30–46, 75–77, A1–A8), prepared for The International EnergyAgency, Solar Heating and Cooling Program, Task10: Solar Materials R&D, May 1989.

106. F. Svegl, A. Surca Vuk, M. Hajzeri, L.S. Perse, B. Orel, Electrochromic proper-ties of Ni1−xO and composite Ni1−xO–polyaniline thin films prepared by the per-oxo soft chemistry route, Sol. Energy Mater. Sol. Cells 99 (2012) 14–25.

107. L. Chia–Ching, Lithium–driven electrochromic properties of electrodeposited nickel hydroxide electrodes, Sol. Energy Mater. Sol. Cells 99 (2012) 26–30.

108. S.H. Park, J.W. Lim, S.J. Yoo, I.Y. Cha, Y.–E. Sung, The improving electro-chromic performance of nickel oxide film using aqueous N,N–dimethylaminoethanol solution, Sol. Energy Mater. Sol. Cells 99 (2012) 31–37.

109. H. Moulki, D.H. Parka, B–K Min, H. Kwonb, S.–J. Hwang, J.–H. Choy, T. Toupance, G. Campet, A. Rougier, Improved electrochromic performances of NiO based thin films by lithium addition: From single layers to devices, Elec-trochim. Acta 74 (2012) 46–52.

110. C.C. Chen, W.D. Jheng, Effect of Ar/O2 ratio on double–sided electrochromic glass performance, Ceram. Int. 38 (2012) 4195–4200.

111. S.–C. Wang, K.–Y. Liu, J.–L. Huang, Tantalum oxide film prepared by reactive magnetron sputtering deposition for all–solid–state electrochromic device, Thin Solid Films 520 (2011) 1454–1459.

112. H. Huang, J. Tian, W.K. Zhang, Y.P. Gan, X.Y. Tao, X.H. Xia, J.P. Tu, Elec-trochromic properties of porous NiO thin film as a counter electrode for NiO/WO3 complementary electrochromic window, Electrochim. Acta 56 (2011) 4281–4286.

113. J. Zhang, J. P. Tu, X. H. Xia, Y. Qiao, Y. Lub, An all–solid–state electrochro-mic device based on NiO/WO3 complementary structure and solid hybrid polye-lectrolyte, Sol. Energy Mater. Sol. Cells 93 (2009) 1840–1845.

114. Y.K. Fang, T.–H. Chou, C.–Y. Lin, Y.T. Chiang, S.–F. Chen, C.–Y. Yang, S.–H. Chang, C.–S. Line, A novel electrochromic device with high optical switch-ing speed, J. Phys. Chem. Solids 69 (2008) 734–737.

115. E. Zelazowska, E. Rysiakiewicz–Pasek, WO3–based electrochromic system with hybrid organic–inorganic gel electrolytes, J. Non–Cryst. Solids 354 (2008) 4500–4505.

116. A. Karuppasamy, A. Subrahmanyam, Studies on electrochromic smart windows based on titanium doped WO3 thin films, Thin Solid Films 516 (2007) 175–178.

117. K.–S. Ahn, Y.–C. Nah, J.–Y. Park, Y.–E. Sung, K.–Y. Cho, S.–S. Shin, J–K Park, Bleached state transmittance in charge–unbalanced all–solid–state elec-trochromic devices, Appl. Phys. Lett. 82 (2003) 3379–3381.

165

118. K.–S. Ahn, Y.–C. Nah, Y.–E. Sung, K.–Y. Cho, S.–S. Shin, J.–K. Park, All–solid–state electrochromic device composed of WO3 and Ni(OH)2 with a Ta2O5 protective layer, Appl. Phys. Lett. 81 (2002) 3930–3932.

119. R. Lechner, L.K. Thomas, All solid state electrochromic devices on glass and polymeric foils, Sol. Energy Mater. Sol. Cells 54 (1998) 139–146.

120. N.A. O'Brien, J. Gordon, H. Mathew, B.P. Hichwa, Electrochromic coatings–applications and manufacturing issues, Thin Solid Films 345 (1999) 312–318.

121. S. Taunier, C. Guery, J.–M. Tarascon, Design and characterization of a three–electrode electrochromic device, based on the system WO3/IrO2, Electrochim. Acta 44 (1999) 3219–3225.

122. J.–C Lassègues, D. Rodriguez, Smart windows using a proton conducting poly-mer as electrolyte, Proc. SPIE 1728 (1992) 241–249.

123. S. Gottesfeld, J.D.E. McIntyre, G. Beni, J.L. Shay, Electrochromism in anodic iridium oxide films, Appl. Phys. Lett. 33 (1978) 208–210.

124. T. Niwa, O. Takai, All–solid–state reflectance–type electrochromic devices using iridium tin oxide film as counter electrode, Thin Solid Films 518 (2010) 5340–5344.

125. T. Niwa, O. Takai, Electrochemical, optical and electronic properties of iridium tin oxide thin film as counter electrode of electrochromic device, Japan. J. Appl. Phys. 49 (2010) 105802/1–5.

126. S.U. Yun, S.J. Yoo, J.W. Lim, S.H. Park, I.Y. Cha, Y.–E. Sung, Enhanced elec-trochromic properties of Ir–Ta oxide grown using a cosputtering system, J. Electrochem. Soc. 157 (2010) J256–J260.

127. J. Backholm, Electrochromic Properties of Iridium Oxide Based Thin Films, PhD Thesis, Uppsala University, Uppsala, Sweden, 2008.

128. Th. Pauporté, R. Durand, Impedance spectroscopy study of electrochromism in sputtered iridium oxide films, J. Appl. Electrochem. 30 (2000) 35–41.

129. J.M.G. Cowie, Polymers: Chemistry and Physics of Modern Materials, Bell and Bain, Glasgow, Great Britain, 1973.

130. U.W. Gedde, Polymer Physics, Chapman&Hall, London, 1996. 131. A.M. Stephan, Review on gel polymer electrolytes for lithium batteries, Eur.

Polym. J. 42 (2006) 21–42. 132. P.E. Harvey, Polymer electrolytes, J. Power Sources 26 (1989) 23–32. 133. M.D. Armand, Polymer electrolytes, Annu. Rev. Mater. Sci. 16 (1986) 245–

261. 134. F.N. Büchi, M. Inaba, T.J. Schmidt (Eds.), Polymer Electrolyte Fuel Cell Dura-

bility, Springer, New York, 2009. 135. C. Sequeira, D. Santos (Eds.), Polymer Electrolytes: Fundamentals and Appli-

cations, Woodhead, Sawston, UK, 2010. 136. P. Staiti, F. Lufrano, Investigation of polymer electrolyte hybrid supercapacitor

based on manganese oxide–carbon electrodes, Electrochim. Acta 55 (2010) 7436–7442.

137. P.V. Wright, Electrical conductivity in ionic complexes of poly (ethylene oxide), Brit. Polym. J. 7 (1975) 319–327.

138. F.M. Gray, Solid Polymer Electrolytes: Fundamentals and Technological Ap-plications, VCH Publishers, New York, 1991.

139. C.A. Vincent, Polymer electrolytes, Progr. Solid State Chem. 17 (1987) 145–261.

140. J.R. MacCallum, C.A. Vincent, Polymer Electrolyte Reviews 1, Elsevier Ap-plied Science, London, 1987.

141. B. Scrosati (Ed.), Second International Symposium on Polymer Electrolytes, Elsevier Applied Science, New York, 1990.

166

142. D. Baril, C. Michot, M. Armand, Electrochemistry of liquids vs. solids: Polymer electrolytes, Solid State Ionics 94 (1997) 35–47.

143. F.B. Dias, L. Plomp, J.B.J. Veldhuis, Trends in polymer electrolytes for secon-dary lithium batteries, J. Power Sources 88 (2000) 169–191.

144. K. Murata, S. Izuchi, Y. Yoshihisa, An overview of the research and develop-ment of solid polymer electrolyte batteries, Electrochim. Acta 45 (2000) 1501–1508.

145. V.S. Kolosnitsyn, G.P. Dukhanin, S.A. Dumler, I.A. Novakov, Lithium–conducting polymer electrolytes for chemical power sources, Russ. J. Appl. Chem. 78 (2005) 1–18.

146. J.–H. Shin, W.A. Henderson, S. Passerini, Lithium metal–polymer electrolyte batteries, J. Electrochem. Soc. 152 (2005) A978–A983.

147. W. Wieczorek, Z. Florjanczyk, J.R. Stevens, Composite polyether based solid electrolytes, Electrochim. Acta 40 (1995) 2251–2258.

148. C. Capiglia, P. Mustarelli, E. Quartarone, C. Tomasi, A. Magistris, Effects of nanoscale SiO2 on the thermal and transport properties of solvent–free, poly(ethylene oxide) (PEO)–based polymer electrolytes, Solid State Ionics 118 (1999) 73–79.

149. L. Fan, C.–W. Nan, S. Zhao, Effect of modified SiO2 on the properties of PEO–based polymer electrolytes, Solid State Ionics 164 (2003) 81–86.

150. V. Di Noto, E. Negro, S. Lavina, in: C. Sequeira, D. Santos (Eds.), Polymer Electrolytes: Fundamentals and Applications, Woodhead, Sawston, UK, 2010.

151. A. Ferry, Polymer electrolytes: A study on polymeric ion conductors and re-chargeable thin–film batteries, PhD thesis, Umeå University, Umeå, Sweden, 1996.

152. M.A. Ratner, High Polymers with Main Chain Oxyethylene Sequences, in: J.R. MacCallum, C.A. Vincent (Eds.), Polymer Electrolyte Reviews 1, Elsevier Ap-plied Science, London, 1987.

153. P.G. Bruce, C.A. Vincent, Polymer electrolytes, J. Chem. Soc. Faraday Trans. 89 (1993) 3187–3203.

154. J. Owen, Ionic conductivity, Comp. Pol. Sci. Suppl. 2 (1989) 669–686. 155. M.H. Cohen, D. Turnbull, Molecular transport in liquids and glasses, J. Chem.

Phys. 31 (1959) 1164–1169. 156. H. Vogel, Das temperatur abhängigkeitsgesetz der viskosität von flüssigkeiten,

Phys. Z. 22 (1921) 645–646. 157. G.S. Fulcher, Analysis of recent measurements of the viscosity of glasses, J. Am.

Ceram. Soc. 8 (1925) 339–355. 158. V.G. Tammann, W.Z. Hesse, Die abhängigkeit der viscosität von der tempera-

tur bei unterkühlten flüssigkeiten, Anorg. Allg. Chem. 156 (1926) 245–257. 159. O. Bohnke, G. Frand, M. Rezrazi, C. Rousselot, C. Truche, Fast ion transport

in new lithium electrolytes gelled with PMMA. 1. Influence of polymer concen-tration, Solid State Ionics 66 (1993) 97–104.

160. P.G. Bruce, Structure and electrochemistry of polymer electrolytes, Electro-chim. Acta 40 (1995) 2077–2085.

161. U.W. Gedde, Polymer Physics, Chapman&Hall, London, 1996, p.93. 162. C.G. Granqvist, Handbook of Inorganic Electrochromic Materials, Elsevier,

Amsterdam, 1995, p. 441. 163. P.G. Bruce, F.M. Gray, J. Shi, C.A. Vincent, Ionic transport in polymer electro-

lytes, Philos. Mag. A 64 (1991) 1091–1099. 164. F.M. Gray, Solid Polymer Electrolytes: Fundamentals and Technological Ap-

plications, VCH Publishers, New York, 1991, p128.

167

165. P. Walden, Salts, Acids, and Bases: Electrolytes: Stereochemistry, McGraw–Hill, New York, 1929, p. 276.

166. J.O. Bockris, A.K.N. Reddy, Modern Electrochemistry 1, Plenum Press, New York, 1977, p. 358.

167. G. Williams, and D. K. Thomas, Phenomenological and molecular theories of dielectric and electrical relaxation of materials, Novocontrol Application Note Dielectrics 3 (1998) 1–29. http://novocontrol.de/pdf_s/APND3.PDF .

168. M. Wübbenhorst, Polymers and Nano–composites, 6th International Discussion Meeting on Relaxations in Complex Systems, Tutorial notes, 2009.

169. R.H. Colby, Dielectric spectroscopy of polymers: Melts, glasses, blends, nano-composites, ionic liquids and ionomers, http://www.ch.ic.ac.uk/tcpt/lectures/Lect9.pdf.

170. J. P. Runt, and J.J. Fitzgerald, Dielectric spectroscopy of polymeric materials, American Chemical Society, Washington, DC, 1997.

171. I. Albinsson, H. Eliasson, B.–E. Mellander, Dielectric Properties of Polymer Electrolytes, Dielectric Newsletter 7 (1997) 1–3, http://novocontrol.de/newsletter/DNL07.PDF.

172. L. Onsager, Electric moments of molecules in liquids, J. Am. Chem. Soc. 58 (1936) 1486–1493.

173. G.B. Smith, M.J. Ford, C. Masens, J. Muir, Energy–efficient coatings in the Nanohouse Initiative, Curr. Appl. Phys. 4 (2004) 381–384.

174. E. Manias, G. Polizos, H. Nakajima, M.J. Heidecker, Fundamentals of polymer nanocomposite technology, in: A.B. Morgan, C.A. Wilkie (Eds.), Flame Retar-dant Polymer Nanocomposites, Wiley, 2007, pp.31–66.

175. S. Li, M.M. Lin, M.S. Toprak, D.K. Kim, M. Muhammed, Nanocomposites of polymer and inorganic nanoparticles for optical and magnetic applications, Nano Rev. 1 (2010) 5214–5240.

176. M. Salavati–Niasari, D. Ghanbari, Polymeric Nanocomposite Materials, in: B.S.R. Reddy (Ed.), Advances in Diverse Industrial Applications of Nanocom-posites, InTech, Rijeka, Croatia, 2011, pp. 501–520.

177. J. Mark (Ed.), Polymer Data Handbook, Oxford University Press, New York, 1999.

178. J.C. Seferis, Refractive Indices of Polymers, in: J. Brandrup, E.H. Immergut, E.A. Grulke (Eds.), Polymer Handbook, 4th ed., New York, 1999, VI, pp. 571–582.

179. http://syntecoptics.com/wp–content/uploads/2013/01/transmission.pdf, 2013. 180. H.C. van de Hulst, Light Scattering by Small Particles, Wiley, New York, 1957. 181. S. Schelma, G.B. Smith, P.D. Garrett, W.K. Fisher, Tuning the surface–plasmon

resonance in nanoparticles for glazing applications, J. Appl. Phys. 97 (2005) 124314/1–8.

182. G. Garcia, R. Buonsanti, E.L. Runnerstrom, R.J. Mendelsberg, A. Llordes, A. Anders, T.J. Richardson, D.J. Milliron, Dynamically modulating the surface plasmon resonance of doped semiconductor nanocrystals, Nano Lett. 11 (2011) 4415–4420.

183. G.B. Smith, C.A. Deller, P.D. Swift, A. Gentle, P.D. Garrett, W.K. Fisher, Nanoparticle–doped polymer foils for use in solar control glazing, J. Nanopart. Res. 4 (2002) 157–165.

184. A. Solieman, M.A. Aegerter, Modeling of optical and electrical properties of In2O3:Sn coatings made by various techniques, Thin Solid Films 502 (2006) 205–211.

185. R.J. Mendelsberg, G. Garcia, D.J. Milliron, Extracting reliable electronic prop-erties from transmission spectra of indium tin oxide thin films and nanocrystal

168

films by careful application of the Drude theory, J. Appl. Phys. 111 (2012) 063515/1–9.

186. I. Hamberg, C.G. Granqvist, Evaporated Sn–doped In2O3 films: Basic optical properties and applications to energy–efficient windows, J. Appl. Phys. 60 (1986) R123–R159.

187. J.C. Maxwell–Garnett, Colours in metal glasses and in metallic films, Philos. Trans. R. Soc. A 203 (1904) 385–420.

188. J.C. Maxwell–Garnett, Colours in metal glasses, in metallic films, and in metal-lic solutions: II, Philos. Trans. R. Soc. A 205 (1906) 237–288.

189. P.M. Price, J.H. Clark, D.J. Macquarrie, Modified silicas for clean technology, J. Chem. Soc. Dalton Trans. 2 (2000) 101–110.

190. B.A. Rozenberg, R. Tenne, Polymer–assisted fabrication of nanoparticles and nanocomposites, Prog. Polym. Sci. 33 (2008) 40–112.

191. A. Bernson, J. Lindgren, OH end–group coordination to cations and triflate ions in PPG systems, Polym. 35 (1994) 4842–4847.

192. W. Wieczorek, K. Such, J. Plocharski, J. Przluski, Mixed phase solid electro-lytes based on poly(ethylene oxide) systems, in: B. Scrosati (Ed.), Second Inter-national Symposium on Polymer Electrolytes, Elsevier Applied Science, New York, 1990, p. 339–346.

193. A. Dong, X. Ye, J. Chen, Y. Kang, T. Gordon, J.M. Kikkawa, C.B. Murray, A generalized ligand–exchange strategy enabling sequential surface functionali-zation of colloidal nanocrystals, J. Am. Chem. Soc. 133 (2011) 998–1006.

194. F. Croce, G.B. Appetecchi, L. Persi, B. Scrosati, Nanocomposite polymer elec-trolytes for lithium batteries, Nature 394 (1998) 456–458.

195. J. Przyluski, M. Siekierski, W. Wieczorek, Effective medium theory in studies of conductivity of composite polymeric electrolytes, Electrochim. Acta 40 (1995) 2101–2108.

196. A.S. Best, J. Adebahr, P. Jacobsson D.R. MacFarlane, M. Forsyth, Microscopic interactions in nanocomposite electrolytes, Macromolecules 34 (2001) 4549–4555.

197. P.A.R.D. Jayathilaka, M.A.K.L. Dissanayake, I. Albinsson, B.–E. Mellander, Effect of nano–porous Al2O3 on thermal, dielectric and transport properties of the (PEO)9LiTFSI polymer electrolyte system, Electrochim. Acta 47 (2002) 3257–3268.

198. F. Croce, L. Persi, B. Scrosati, F. Serraino–Fiory, E. Plichta, M.A. Hendrickson, Role of the ceramic fillers in enhancing the transport properties of composite polymer electrolytes, Electrochim. Acta 46 (2001) 2457–2461.

199. W. Wieczorek, J.R. Stevens, Z. Florjanczyk, Composite polyether based solid electrolytes. The Lewis acid–base approach, Solid State Ionics 85 (1996) 67–72.

200. W. Wieczorek, P. Lipka, G. Zukowska, H. Wycislik, Ionic interactions in poly-meric electrolytes based on low molecular weight poly(ethylene glycol)s, J. Phys. Chem. Sect. B 102 (1998) 6968–6974.

201. N.C. Mello, T.J. Bonagamba, H. Panepucci, K. Dahmouche, P. Judeinstein, M.A. Aegerter, NMR study of ion–conducting organic–inorganic nanocompo-sites poly(ethylene glycol)–silica–LiClO4, Macromolecules 33 (2000) 1280–1288.

202. C. Pfaffenhuber, J. Maier, Quantitative estimate of the conductivity of a soggy sand electrolyte: Example of (LiClO4, THF):SiO2, Phys. Chem. Chem. Phys. 15 (2013) 2050–2054.

169

203. J.M.G. Cowie, Conductivity in non–main chain oxide systems and some linear analogues, in: J.R. MacCallum, C.A. Vincent (Eds.), Polymer Electrolyte Re-views, Volume 1, Elsevier Applied Science, New York, 1987, p.88.

204. C.S. Harris, M.A. Ratner, D.F. Shriver, Ionic conductivity in branched poly-ethylenimine–sodium trifluoromethanesulfonate complexes. Comparisons to analogous complexes made with linear polyethylenimine, Macromolecules 20 (1987) 1778–1781.

205. http://www4.ncsu.edu/~hubbe/PEI.htm, 2013. 206. Material safety data sheet of Sigma–Aldrich, 2010. 207. C. Stappen, Dielectric Characterisation of Polymer Electrolytes, MsC Thesis,

Uppsala University, Uppsala, Sweden, 2004. 208. M. Deepa, N. Sharma, S.A. Agnihotry, S. Singh, T. Lal and R. Chandra, Con-

ductivity and viscosity of liquid and gel electrolytes based on LiClO4, LiN(CF3SO2)2 and PMMA, Solid State Ionics 152–153 (2002) 253– 258.

209. M. Armand, W. Gorecki, R. Andreani, Perfluorosulphonimide salts as solute for polymer electrolytes, in: B. Scrosati (Ed.), Second International Symposium On Polymer Electrolytes, Elsevier Applied Science, Amsterdam, 1990, p.91–97.

210. R. Arnaud, D. Benrabah, J.–Y. Sanchez, Theoretical study of CF3SO3Li, (CF3SO2)2NLi, and (CF3SO2)2CHLi ion pairs, J. Phys. Chem. 100 (1996) 10882–10891.

211. C. Roux, J.–Y. Sanchez, Ionic behaviour of alkaline complexes based on linear and crosslinked poly(oxypropylene), Electrochim. Acta 40 (1995) 953–957.

212. D.R. MacFarlane, P. Meakin, J. Sun, N. Amini, M.J. Forsyth, Pyrrolidinium imides: A new family of molten salts and conductive plastic crystal phases, J. Phys. Chem. B 103 (1999) 4164–4170.

213. O. Borodin, G.D. Smith, LiTFSI structure and transport in ethylene carbonate from molecular dynamics simulations, J. Phys. Chem. B 110 (2006) 4971–4977.

214. W.A. Henderson, M. Herstedt, V.G. Young, S. Passerini, H.C. De Long, P.C. Trulove, New disordering mode for TFSI–anions: The nonequilibrium, plastic crystalline structure of Et4NTFSI, Inorg. Chem. 45 (2006) 1412–1414.

215. R. Tanaka, I. Ueoka, Y. Takaki, K. Kataoka, S. Saito, High molecular weight linear polyethylenimine and poly(N–methylethylenimine), Macromolecules 16 (1983) 849–853.

216. M. Watanabe, R. Ikezawa, K. Sanui, N. Ogata, Protonic conduction in poly(ethylenimine) hydrates, Macromolecules 20 (1987) 968–973.

217. C. Poinsignon, Polymer electrolytes, Mater. Sci. Eng. B3 (1989) 31–37. 218. M.F. Daniel, B. Desbat, F. Cruege, O. Trinquet, J.C. Lassegues, Solid state

protonic conductors: Poly(ethylene imine) sulfates and phosphates, Solid State Ionics, 28–30 (1988) 637–641.

219. G.K.R. Senadeera, M.A. Careem, S. Skaarup, K. West, Enhanced ionic conduc-tivity of poly(ethylene imine) phosphate, Solid State Ionics 85 (1996) 37–42.

220. C.K. Chiang, G.T. Davis, C.A. Harding, T. Takahashi, Polymeric electrolyte based on poly (ethylene imine) and lithium salts, Solid State Ionics 18 & 19 (1986) 300–305.

221. J.–L. Paul, C. Jegat, J.–C. Lasségues, Branched poly(ethyleneimine)–CF3SO3Li complexes, Electrochim. Acta 37 (1992) 1623–l625.

222. B. Unal, R.J. Klein, K.R. Yocca, R.C. Hedden, Influence of DGEBA crosslink-ing on Li+ ion conduction in poly(ethyleneimine) gels, Polymer 48 (2007) 6077–6085.

223. M. Erickson, R. Frech, D.T. Glatzhofer, Gel electrolytes based on crosslinked tetraethylene glycol diacrylate/poly(ethyleneimine) systems, Polymer 45 (2004) 3389–3397.

170

224. A.G. Snow, R.A. Sanders, R. Frech, D.T. Glatzhofer, Synthesis and spectro-scopic studies of linear poly (N–(2–(2–methoxy)ethyl)ethylenimine), a PEI/PEO hybrid, and its interactions with lithium triflate, Electrochim. Acta 48 (2003) 2065–2069.

225. R. Tanaka, M. Sakurai, H. Sekiguchi, M. Inoue, Improvement of room–temperature conductivity and thermal stability of PEO–/LiClO4 systems by ad-dition of a small proportion of polyethyleneimine, Electrochim. Acta 48 (2003) 2311–2316.

226. S. York, R. Frech, A. Snow, D. Glatzhofer, A comparative vibrational spectros–copic study of lithium triflate and sodium triflate in linear poly(ethylenimine), Electrochim. Acta 46 (2001) 1533–1537.

227. S.S. York, M. Buckner, R. Frech, Ion–polymer and ion–ion interactions in linear poly(ethylenimine) complexed with LiCF3SO3 and LiSbF6, Macromole-cules 37 (2004) 994–999.

228. L. L. Ionescu–Vasii, B. Garcia, M. Armand, Conductivities of electrolytes based on PEI–b–PEO–b–PEI triblock copolymers with lithium and copper TFSI salts, Solid State Ionics 177 (2006) 885–892.

229. M. Akhtar, R.M. Paiste, H.A. Weakliem, Thin film tungsten oxide electrochro-mic displays, J. Electrochem. Soc. 135 (1988) 1597–1598.

230. M. Akhtar, H.A. Weakliem, R.M. Paiste, K. Gaughan, Polyaniline thin film electrochromic devices, Synth. Met. 26 (1988) 203–208.

231. M. Antinucci, B. Chevalier, A. Ferriolo, Development and characterisation of electrochromic devices on polymeric substrates, Sol. Energy Mater. Sol. Cells 39 (1995) 271–287.

232. C.–Y. Li, L.–M. Huang, T.–C. Wen, A. Gopalan, Superior performance charac-teristics for the poly(2,5–dimethoxyaniline)–poly(styrene sulfonic acid)–based electrochromic device, Solid State Ionics 177 (2006) 795–802.

233. H. Hu, B.E. Ortiz–Aguilar, L. Hechavarria, Effect of pH value of poly (ethylen-imine)–H2SO4 electrolyte on electrochromic response of polyaniline thin films, Opt. Mater. 29 (2007) 579–584.

234. http://imagine.gsfc.nasa.gov/docs/science/know_l1/spectrum_chart.html, 2013. 235. J. Mijovic, B.D. Fitz, Dielectric Spectroscopy of Reactive Polymers, Novocon-

trol Application Note, 1998, http://novocontrol.de/pdf_s/APND2.PDF 236. S. Emmert, M. Wolf, R. Gulich, S. Krohns, S. Kastner, P. Lunkenheimer, A.

Loidl, Electrode polarization effects in broadband dielectric spectroscopy, Eur. Phys. J. B 83 (2011) 157–165.

237. H.P. Schwan, Electrode polarization impedance and measurements in biologi-cal materials, Ann. N.Y. Acad. Sci. 148 (1968) 191–209.

238. M.M. Springer, A. Korteweg, J. Lyklema, The relaxation of the double layer around colloidal particles and the low frequency dispersion: Part II. Experi-ments, J. Electroanal. Chem. 153 (1983) 55–66.

239. B.A. Mazzeo, A.J. Flewitt, Two– and four–electrode, wide–bandwidth, dielec-tric spectrometer for conductive liquids: Theory, limitations, and experiment, J. Appl. Phys. 102 (2007) 104106/1–6.

240. W. Scheider, Theory of the frequency dispersion of electrode polarization. To-pology of networks with fractional power frequency dependence, J. Phys. Chem. 79 (1975) 127–136.

241. S.H. Liu, Fractal model for the ac response of a rough interface, Phys. Rev. Lett. 55 (1985) 529–532.

242. L. Nyikos, T. Pajkossy, Fractal dimension and fractional power frequency–dependent impedance of blocking electrodes, Electrochim. Acta 30 (1985) 1533–1540.

171

243. J.B. Bates, Y.T. Chu, W.T. Stribling, Surface topography and impedance of metal–electrolyte interfaces, Phys. Rev. Lett. 60 (1988) 627–630.

244. J.R. Macdonald, Impedance spectroscopy and its use in analyzing the steady–state AC response of solid and liquid electrolytes, J. Electroanal. Chem. 223 (1987) 25–50.

245. M. Watanabe, K. Sanui, N. Ogata, Ionic conductivity and mobility in network polymers from poly(propylene oxide) containing lithium perchlorate, J. Appl. Phys. 57 (1985) 123–128.

246. P. Debye, Polar Molecules, The Chemical Catalog Company, New York, 1929. 247. K.S. Cole, R.H. Cole, Dispersion and absorption in dielectrics: I. Alternating

current characteristics, J. Chem. Phys. 9 (1941) 341–351. 248. D.W. Davidson, R.H. Cole, Dielectric relaxation in glycerine, J. Chem. Phys.

18 (1950) 1417–1417. 249. S. Havriliak, S. Negami, A complex plane representation of dielectric and me-

chanical relaxation processes in some polymers, Polymer 8 (1967) 161–210. 250. A. K. Jonscher, The “universal” dielectric response, Nature 267 (1977) 673–

678. 251. V. Raicu, Dielectric dispersion of biological matter: Model combining Debye–

type and “universal” responses, Phys. Rev. E 60 (1999) 4677–4680. 252. J.R. Macdonald, Impedance spectroscopy, Annals of Biomedical Engineering

20 (1992) 289–305. 253. R.R. Nigmatullin, The Basis of Dielectric Spectroscopy and Relaxation in Dis-

ordered Systems, Lecture notes, 2004. 254. S. Havriliak Jr., S.J. Havriliak, Dielectric and Mechanical Relaxation in Mate-

rials: Analysis, Interpretation, and Application to Polymers, Hanser, Munich, 1997.

255. J.L. Barton, Dielectric relaxation of some ternary alkali–alkaline earth silicate glasses, Verres et Refr. 20 (1966) 328–335.

256. T. Nakajima, Correlation between electrical conduction and dielectric polariza-tion in inorganic glasses, 1971 Annual Report, Conference on Electrical Insula-tion and Dielectric Phenomena, National Academy of Science, 2032 (1972) 168–176.

257. H. Namikawa, Characterization of the diffusion process in oxide glasses based on the correlation between electric conduction and dielectric relaxation, J. Non–Cryst. Solids 18 (1975) 173–195.

258. D.L. Sidebottom, Universal approach for scaling the ac conductivity in ionic glasses, Phys. Rev. Lett. 82 (1999) 3653–3656.

259. D.L. Sidebottom, B. Roling, K. Funke, Ionic conduction in solids: Comparing conductivity and modulus representations with regard to scaling properties, Phys. Rev. Lett. B 63 (2000) 024301/1–7.

260. J.C. Dyre, T.B. Schrøder, Universality of ac conduction in disordered solids, Rev. Mod. Phys. 72 (2000) 873–892.

261. J.R. Macdonald, Universality, The Barton Nakajima Namikawa relation, and scaling for dispersive ionic materials, Phys. Rev. B 71 (2005) 184307/1–12.

262. C.G. Granqvist, Radiative heating and cooling with spectrally selective sur-faces, Appl. Opt. 20 (1981) 2606–2615.

263. C.G.Granqvist, Solar energy materials, Adv. Mater. 15 (2003) 1789–1803. 264. http://cnx.org/content/m42444/latest/?collection=col11406/latest, 2013. 265. W.Q. Hong, Extraction of extinction coefficient of weak absorbing thin films

from special absorption, J. Phys. D 22 (1989) 1384–1385.

172

266. O. Sadiku–Agboola, E.R. Sadiku, A.T. Adegbola, O.F. Biotidara, Rheological properties of polymers: Structure and morphology of molten polymer blends, Materials Sciences and Applications 2 (2011) 30–41.

267. S. G. Brush, Theories of liquid viscosity, Chem. Rev. 62 (1962) 513–548. 268. D. Nichetti, I. Manas–Zloczower, Viscosity model for polydisperse polymer

melts, J. Rheol. 42 (1998) 951–969. 269. H.A. Barnes, J.F. Hutton, K. Walters, An Introduction to Rheology, Elsevier,

Eastbourne, UK, 1989. 270. www.sju.edu/~phabdas/physics/rheo.html, 2010. 271. www.glossary.oilfield.slb.com, 2010. 272. J. Vlachopoulos, N. Polychronopoulos, Basic Concepts in Polymer Melt Rheol-

ogy and Their Importance in Processing, in: M. Kontopoulou (Ed.), Applied Polymer Rheology: Polymeric Fluids with Industrial Applications, Wiley, Ho-boken, NJ, 2011, pp. 1–28.

273. C. Yürüdü, Synthesis and Characterization of HDTABr/NABI Composite, MSc Thesis, İstanbul Technical University, İstanbul, Turkey, 2004.

274. V. Kumar, S. Sinha, Polymer Systems and Applications, Studium Press, India, 2010, p 172.

275. www.wikipedia.org, 2010. 276. S–Il Choi, K.M. Nam, B.K. Park, W.S. Seo, J.T. Park, Preparation and optical

properties of colloidal, monodisperse, and highly crystalline ITO nanoparticles, Chem. Mater. 20 (2008) 2609–2611.

277. A. Llordes, A.T. Hammack, R. Buonsanti, R. Tangirala, S. Aloni, B.A. Helms, D.J. Milliron, Polyoxometalates and colloidal nanocrystals as building blocks for metal oxide nanocomposite films, J. Mater. Chem. 21 (2011) 11631–11638.

278. Novocontrol broadband dielectric converter user manual, 1996. 279. F.Kremer, M.Arndt, Broadband dielectric measurement techniques, in: J.P.

Runt, J.J. Fitzgerald (Eds), Dielectric Spectroscopy of Polymeric Materials. Fundamentals and Applications, American Chemical Society, Washington DC, 1997, pp.35–57.

280. R.C. Mackenzie, De calore: Prelude to thermal analysis, Thermochim. Acta 73 (1984) 251–306.

281. More Solutions to Sticky Problems: A Guide to Getting More from Your Brook-field Viscometer, Brookfield Engineering Labs. Inc., p. 22. http://www.viscometers.org/PDF/Downloads/More%20Solutions.pdf

282. J.A. Jacquez, H.F. Kuppenheim, Theory of the integrating sphere, J. Opt. Soc. Am. 45 (1955) 460–470.

283. D.O. Goebel, Generalized integrating sphere theory, Appl. Opt. 6 (1967) 125–128.

284. M.W. Finkel, Integrating sphere theory, Opt. Commun. 2 (1970) 25–28. 285. A.M. Nilsson, Daylighting Systems: Development of Techniques for Optical

Characterization and Performance Evaluation, PhD thesis, Uppsala University, Uppsala, Sweden, 2012.

286. T.M. Madkour, Poly(ethylene imine), in: J.E. Mark (Ed.), Polymer Data Hand-book, Oxford University Press, Oxford, UK, 1999, pp. 490–492.

ACTAUNIVERSITATIS

UPSALIENSISUPPSALA

acta universitatis upsaliensis Digital Comprehensive Summaries of Uppsala Dissertations from the Faculty of Science and Technology 1057

editor: The Dean of the Faculty of Science and Technology

a doctoral dissertation from the Faculty of Science and Technology, uppsala university, is usually a summary of a number of papers. a few copies of the complete dissertation are kept at major Swedish research libraries, while the summary alone is distributed internationally through the series Digital Comprehensive Summaries of uppsala Dissertations from the Faculty of Science and Technology. (Prior to January, 2005, the series was published under the title “Comprehensive Summaries of uppsala Dissertations from the Faculty of Science and Technology”.)

Distribution: publications.uu.se 2013urn:nbn:se:uu:diva-204437