Uppsala Dissertations fromthe Faculty of Science and Technology

56

ESTEBAN DAMIÁN AVENDAÑO SOTO

Electrochromism in Nickel-basedOxides

Coloration Mechanisms and Optimization ofSputter-deposited Thin Films

ACTA UNIVERSITATIS UPSALIENSISUPPSALA 2004

ACTA UNIVERSITATIS UPSALIENSISUppsala Dissertations from

the Faculty of Science and Technology56

ESTEBAN DAMIÁN AVENDAÑO SOTO

Electrochromism in Nickel-basedOxides

Coloration Mechanisms and Optimization ofSputter-deposited Thin Films

Printed in Sweden by Geotryckeriet, Uppsala 2004

A Mami, que siempre me ha apoyado en todas mis ocurrencias.

A mis pulgas feroces Mimi y Beatrice A mis hermanos Pablito, Juancho, Javier y Pedrules.

A mis abuelos donde quiera que esten. A mis hermanos postizos Allan, Fernando, Alfonso y Jose

A mi ahijado Jose Al Mamut que siempre nos persigue.

vii

List of Papers

Articles related to this work

1. E. Avendaño, A. Azens and G. A. Niklasson. Properties of electrochromic nickel-vanadium oxide films sputter-deposited from non-magnetic alloy target. Proc. SPIE, 4458, (2001) 154-163.

2. E. Avendaño, A. Azens, G. A. Niklasson and C. G. Granqvist. Nickel oxide based electrochromic films with optimised optical properties. Journal of Solid State Electrochemistry, 8, 37-39 (2003).

3. E. Avendaño, A. Azens, J. Isidorsson, R. Karmhag, G. A. Niklasson and C. G. Granqvist. Optimized nickel oxide electrochromic films. SolidState Ionics, 165, 169-173 (2003).

4. A. Azens, E. Avendaño and C. G. Granqvist. Electrochromic materials and Their Applications in Foil-Based Devices, Proc. SPIE, 5123, 185-195 (2003).

5. C.G. Granqvist, E. Avendaño and A. Azens. Electrochromic coatings and devices: survey of some recent advances. Thin Solid Films, 442 (2003) 201-211.

6. C.G. Granqvist, E. Avendaño and A. Azens. Advances in electrochromic materials and devices. J. Matter Sci. Forum. In press.

7. E. Avendaño, A. Kuzmin, J. Purans, A. Azens, G. A. Niklasson and C. G. Granqvist. Changes in the local structure of nanocrystalline electrochromic films of hydrated nickel vanadium oxide upon ozone-induced coloration. Physica Scripta. (2004) In press..

8. E. Avendaño, A. Azens, G. A. Niklasson and C. G. Granqvist. Electrochromic materials and devices: Advances on high-transparency smart windows. Proc. SPIE. In press.

9. E. Avendaño, A. Azens, G. A. Niklasson and C.G. Granqvist. Electrochromic nickel-based oxide films with minimized bleached-state absorptance, in Electrochromic Materials and devices, edited by D. Rauh (The Electrochemical Society, Pennington, USA, 2003), Vol. PU 2003-17, 80-90 (2003).

10. E. Avendano, A. Azens, and C.G. Granqvist. Novel electrochromic materials and devices. Proc. SVC, 7188, (2003) 75-80.

viii

11. E. Avendaño, A. Azens, G.A. Niklasson and C.G. Granqvist. Electrochromic nickel-oxide-based thin films optimized for architectural application. ISES Solar World Congress. Göteborg, Sweden. June 2003. Gothenburg, Sweden, paper W5 5 (5 pages).

12. E. Avendaño, A. Azens, G.A. Niklasson and C.G. Granqvist., Electrochromism in Nickel Oxide Films Containing Mg, Al, Si, V, Zr, Nb, Ag, or Ta, Solar Energy Materials and Solar Cells, (2004) in press.

13. E. Avendaño, H. Rensmo, A. Azens, A. Sandell, G.A. Niklasson, H. Siegbhan and C.G. Granqvist. X-ray photoemission spectroscopy of electrochromic hydrated nickel vanadium oxide thin films upon proton intercalation and de-intercalation. In manuscript (To be suttmitted to Journal of Physical Chemistry B)

14. E. Avendaño, A. Azens, G.A. Niklasson and C.G. Granqvist. Phase transitions of electrochromic hydrated nickel based oxide studied by galvanostatic intermittent titration. Submitted to the Journal of the Electrochemical Society.

15. E. Avendaño, A. Kuzmin, J. Purans, A. Azens, G. A. Niklasson and C. G. Granqvist. X ray absorption fine structure analysis of electrochromic hydrated nickel vanadium oxide thin film upon ozone-induced coloration. In manuscript (To be suttmitted to Phycal Rewiew B)

16. E. Avendaño, H. Rensmo, A. Azens, A. Sandell, G.A. Niklasson, H. Siegbhan and C.G. Granqvist. X-ray photoemission spectroscopy of electrochromic hydrated nickel vanadium oxide thin films upon ozone coloration. In manuscript (To be suttmitted to Journal of Applied Physics)

Articles not related to this work

17. T. Lindgren, G. Romualdo Torres, J. Mwabora, E. Avendaño Soto, J. Jonsson, A. Hoel, C.-G. Granqvist, and S.-E. Lindquist. Photoelectrochemical and optical properties of nitrogen doped titanium dioxide films prepared by reactive DC magnetron sputtering. The 203rd

Meeting of The Electrochemical Society. May 2003. Paris, France. Abstract Volume, No. 1724, (1 page).

18. Lindgren T., Mwabora J. M., Avendaño E., Jonsson J., Hoel A., Granqvist C.G. and Lindquist S.-E. Photoelectrochemical and optical properties of nitrogen doped titanium dioxide prepared by reactive DC magnetron sputtering. J. Phys. Chem. B, 107, 5709-5716 (2003).

ix

19. J.M. Mwabora, E. Avendano, T. Lindgren, G. Niklasson, S.-E. Lindquist and C.G. Granqvist. Structure and morphology of nitrogen doped TiO2

thin films deposited by reactive DC magnetron sputtering. Submitted to Thin Solid Films.

20. Julius M. Mwabora, Sten-Eric Lindquist, Torbjörn Lindgren, E. Avendaño and Claes G. Granqvist. Photoelectrochemical solar cells based on nitrogen doped TiO2 electrodes. ISES Solar World Congress. Gothemburg, Sweden, paper W5 13 (5 pages).

Patent applications

PCT application SE02/01762: “Electrochromic film and device comprising the same”. Inventors: A. Azens, J. Isidorsson, R. Karmhag, C. G. Granqvist and E. Avendaño.

xi

Contents

1. Introduction.................................................................................................11.1 Electrochromism ..................................................................................11.2 Electrochromic (EC) device ................................................................21.3 Deposition technique for EC layers: sputtering....................................41.4 Smart window based on hydrated WO3 and NiO.................................51.6 Outline of this work..............................................................................6

2. Properties of nickel oxides..........................................................................92.1 Bulk properties of NiO.........................................................................92.2 Nickel hydroxides and oxy-hydroxides..............................................102.3 Thin film properties............................................................................11

2.3.1 Chemically deposited films ........................................................122.3.2 Physically deposited films ..........................................................132.3.3 Summary.....................................................................................16

3. Experimental details..................................................................................173.1 Reactive DC and RF magnetron sputtering........................................173.2 Thickness measurements....................................................................193.3 Rutherford back scattering (RBS) ......................................................193.4 X-ray diffraction analysis (XRD).......................................................203.5 Cyclic voltammetry (CV)...................................................................223.6 Galvanostatic intermittent titration technique (GITT) .......................223.7 Optical measurements ........................................................................233.8 Infrared spectroscopy (IR)..................................................................283.9 X-ray photoelectron spectroscopy (XPS)...........................................293.10 X-ray absorption fine structure (EXAFS) ........................................30

4. Film deposition .........................................................................................314.1 Sputtering of nickel vanadium oxide..................................................314.2 Co-sputtering of nickel aluminum oxide............................................334.3 Sputtering of nickel aluminum oxide .................................................34

5. Physical characterization ..........................................................................375.1 Composition .......................................................................................375.2 Density ...............................................................................................375.3 Structure analysis ...............................................................................38

xii

5.3.1 XRD in as deposited films..........................................................385.3.2 XRD in electrochemically cycled films......................................415.3.3 XRD in ozone exposed films ......................................................425.3.4 EXAFS in as deposited and ozone exposed films ......................44

6. Electrochromism.......................................................................................496.1 Cyclic voltammetry ............................................................................496.2 Spectral absorptance in nickel based oxides ......................................536.3 Absorption coefficient in hydrated Ni1-xVxOy ....................................556.4 CV and optical density in hydrated Ni1-xVxOy ...................................576.4 Coloration efficiency: V versus Al.....................................................586.5 Color properties for nickel based oxides ............................................606.6 Ozone coloration: optical and color properties ..................................646.7 Devices: optical and color properties .................................................66

7. Proton diffusion and structural changes....................................................717.1 Proton transport ..................................................................................717.2 Diffusion coefficient: Randles-Sev ik equation.................................737.3 Diffusion coefficient determined by GITT.........................................76

7.3.1 Comparison between NiO and Ni1-xVxO ....................................777.3.2 Interpretation of the results .........................................................78

7.4 Ex-situ IR of electrochemically cycled films .....................................80

8. Surface analysis by XPS ...........................................................................858.1 XPS in electrochemically intercalated films ......................................868.2 XPS in ozone colored films................................................................94

9. Mechanism of coloration ........................................................................101

10. Summary and conclusions ....................................................................107

11. Sammanfattning på svenska..................................................................10910.1 Bakgrund ........................................................................................10910.2 Smarta fönster baserade på hydratiserad WO3 och NiO.................111

Acknowledgments.......................................................................................115

References...................................................................................................117

xiii

Symbols and abbreviations

TCO Transparent conductive oxide ITO Commercial name for In2O3:SnPET poly (ethylene terephthalate) PEN poly (ethylene nephthalate) DC Direct current RF Radio frecuency R Deposition rate XRD X-Ray Diffraction

Diffraction angle s Lattice strain broadening g Grain size broadening

FWHM Full width at half maximum L Effective grain size S Strain function RBS Rutherford Backscattering Spectroscopy ERDA Elastic Recoil Detection Analysis UV Ultra violet VIS Visible NIR Near infrared MIR Middle Infrared XPS X ray photoelectron spectroscopy B.E Binding energy XAS X ray absorption EXAFS Extended X ray absorption fine structure XANES X-ray absorption near-edge structure EDA EXAFS Data Analysis FEFF8 Automated program for ab initio multiple scattering

calculations EXAFS and XANES spectra.FMS Full-multiple-scattering ECP Exchange-correlation potential k Wavenumber T Trasnmittance R Reflectance

xiv

A AbsorptanceSolar spectrum( ) AM1 solar spectrum

Lu minous( ) Luminous efficiency function for photopic vision pR Reflectance at p polarized light

1Im Energy Loss Function

Complex dielectric function Angular frequency Angle of incidence

N: Complex refractive index

n Refractive index k Extinction coefficient

WavelenghtFrequency Absorption coefficient

d Thickness OD Optical density CE Coloration efficiency CIE International Commission on Illumination x Color matching functions

CIEX Tristimus values

S Relative spectral irradiance function

Color stimulus

x Chromaticity coordinates D65 Standard illuminant that signifies the average north sky

daylight at 6500 KA Standard illuminant that signifies the tungsten halogen,

incandescent light source at 2856 KF11 Standard illuminant that signifies the commercial, rare

earth phosphor, narrow band fluorescent light source at 4000 K

F2 Standard illuminant that signifies the commercial, wide band fluorescent, cool white fluorescent light source at4150 K

YCIE Luminous transmittance under a specific illuminant CV Cycled voltammetry j Current density Ucol Magnitude of the voltage for full coloration

xv

Q Charge capacity GITT Galvanostatic intermittent titration technique IP Current peak canalyte Concentration of the analyte

Sweep rate A Electrode area D~ Chemical diffusion coefficient

n Number of electrons transferred per atom To Temperature in K F Faraday’s constant Rg Gas constant jP Current density at peak position d Thickness

sV Change of the steady state equilibrium potential

tV Voltage response during the current step

ts Step time t Time V Voltage

+Ho Chemical potential

X Fraction of ions de-intercalated due to the step current U Mobility at X = 0 g Interaction energy

1

1. Introduction

Section 1.1 introduces the concept of electrochromism and classifies the metal oxides depending on the intercalation process that produces the coloration. Section 1.2 describes electrochromic devices and their applications. Section 1.3 mentions the most adequate deposition technique. Section 1.4 describes the characteristics of a specific device based on hydrated tungsten and nickel oxides. Section 1.5 gives the outline of this thesis.

1.1 Electrochromism

Electrochromic (EC) materials change their optical properties reversibly between transparent and dark upon charge insertion/extraction [1]. Materials that color upon insertion are called cathodic; materials that color upon extraction are called anodic. Metal oxides of W, Ti, V, Nb, Ta and Mo exhibit cathodic electrochromism and oxides of V, Cr, Mn, Fe, Co, Ni, Rh and Ir, anodic electrochromism. Vanadium oxide exhibits anodic or cathodic electrochromism depending on the phase. Not all the metal oxide phases of these materials possess electrochromic properties and in many cases a mixed oxide is used to enhance a particular property of the thin film such as the color, cycling stability, etc [1].

These materials can be integrated in electrochromic devices of two types [2]:

1. all-solid-state constructions (multilayer solid thin films)

2. laminated constructions (solid thin films laminated together with a viscous polymer electrolyte)

These constructions can be rigid if glass is used as substrate or flexible if a polymer foil is used instead. Both kinds of devices can be used with or without self-powering by solar cells.

2

These devices open a number of technologically interesting possibilities to modulate optical transmittance, reflectance, absorptance, and emittance. They can be applied to architectural windows, eyewear, non-emissive displays, thermal control in satellites, etc.

1.2 Electrochromic (EC) device

Figure 1.1 shows the cross section of a “smart window” device and its layer components. The device consists of a “sandwich” of two substrates coated with a transparent conductor and an electrochromic film, and then laminated with an ion conductor. The details of the components are explained hereafter:

Substrates The device can be constructed on solid substrates of glass or flexible foil of polymer such as poly (ethylene terephthalate) or poly (ethylene nephthalate) known commercially as PET and PEN, respectively [2].

Transparent electron conductors The substrates are coated with a transparent conductive oxide (TCO) such as In2O3 [3], SnO2 [4], ZnO [5] or the doped versions In2O3:Sn (ITO) [6], SnO2:F [4], SnO2:Sb [7] and ZnO:Al [8].

Cathodic electrochromic layers Many materials have been used as a cathodic electrochromic layer. Tungsten oxide (WO3) [9,10] intercalated with H, Li, Na and K [11] is widely used for its clear bleached state and high coloration efficiency. Mixed tungsten based oxides have been tested as a natural way to enhance a specific property. The broadening of the spectral absorption was enhanced in mixed W1-yMoyO3 [12]. High values of luminous and solar transmittance were found with admixtures of vanadium (WO3-V2O5) [13], niobium (WO3-Nb2O5) [14], and tantalum (WO3-Ta2O5) [15]. Finally, superior durability was found in mixtures of WO3-TiO2 [16]. Ternary oxides such as WO3-MoO3-V2O5 [13], VO2-WO3-V2O5 [14], and WO3-Nb2O5-Li2O [14] have been used for tinting the film as a function of the percentage of the constituents. Among other materials the use of MoO3

[17,18], TiO2 [19,20], V2O5 [21,22], Nb2O5 [23,24], and Ta2O5 [25], intercalated with H, Li, Na and K, have been reported. MoO3 has the inconvenience of not being stable in electrolytes and TiO2 and V2O5 are difficult to make in the correct electrochromic phases. Further, TiO2 has low

3

coloration efficiency; Nb2O5 and Ta2O5 are used as proton conductors because of the high potentials at which the color changes occur. These difficulties made tungsten oxide the best choice for the cathodic electrode.

Figure 1.1. Graphical representation of a “smart window” device based on the combination of electrochromic hydrated tungsten and nickel oxides. The arrows indicate the transport of positive ions under the action of an external electric field. Figure modified from ref [1].

Anodic electrochromic layers Among the materials that can be used as an anodic layer are the hydrated forms of iridium oxide (IrO2) [26], vanadium dioxide (VO2) [27], cobalt oxide (CoO) [28-30], iron oxide [31], and nickel oxide (NiO) [32,33] intercalated with H or Li. Other materials based on chromium and ruthenium oxides have been also reported but not widely studied [1]. Nickel hydroxide mixed with metal ions of Cr, Mn, Fe, Co, Zn, Y, Ag, Cd, La, Pb, and Ce has been reported [34,35].

4

Transparent ion conductors Two types of ion conductors can be used, a solid film of a metal oxide or a viscous polymer electrolyte. Films of Ta2O5,Nb2O5, CeO2, and ZrO2 are used as proton conductors [36] and lithium conductors [37-39].

For the proton conductive polymer a few examples can be given: polyethylene oxide (PEO), a copolymer of sodium vinylsulfonic acid and 1-vinyl-2-pyrrolidinone (PVSA-PVP) ion exchanged into an acidic form, and poly-2-acrylamido-2-methyl-propane sulfonic acid (PAMPS). The Li conductor has been based on poly (methyl methacrylate) (PMMA) copolymerized with propylene carbonate (PC) or poly propylene glycol (PPG), among others, plus the addition of an appropriate Li salt [2].

Different characteristics of all-solid-state constructions and laminated constructions were reported recently in ref [2].

1.3 Deposition technique for EC layers: sputtering

Reactive magnetron sputtering has been shown to be a convenient technique for large-scale, large-area film deposition, facilitating the transfer from the laboratory scale to a large-area deposition in industries [40,41]. Clearly, this transfer is a key part in the establishment of a process for the production of electrochromic films on an industrial scale. Sputtering is mainly done by two different methods: a DC voltage and by RF (radio frequency) as explained in detail in section 3.1. Reactive sputtering from a metallic target is done with DC magnetrons; for insulator materials a RF generator is needed. The convenience of using DC is due firstly to simplicity of the experimental setup and secondly because it is possible to produce the same compounds with similar characteristics as if a RF process is used.

Deposition and properties of tungsten oxides have been extensively reviewed in ref. [1]. The films produced by sputtering present higher coloration efficiency compared with films produced by reactive evaporation, chemical vapour deposition, sol gel, spray pyrolysis and dip-coating [1].

Many reports published on hydrated NiO films have led to the recognition that the film properties are strongly influenced by the deposition technique. In chapter 4, it is discussed how to sputter deposit high quality and fully reproducible films of hydrated nickel based oxides with better durability and

5

charge capacity than the ones produced by other deposition techniques. In chapter 2 the electrochromic properties of films produced by different techniques are discussed.

1.4 Smart window based on hydrated WO3 and NiO

Special attention has been devoted to designs incorporating electrochromic hydrated NiO films operating in conjunction with electrochromic hydrated WO3 [42,43]. The advantage of this combination in an electrochromic device is due to:

a. low operation potentials, i.e, around 1.5 V

b. the possibility to attain a neutral gray color in the dark state

c. the high transparency in the bleached state

d. the facile production of films ready for lamination

The point (a) is an intrinsic property of the materials involved and the neutral gray color (b) is due to the combination of the blue and brown color of the charged WO3 and discharged NiO films, respectively. The points (c) and (d) are explained in detail here after.

High transparency To achieve high transparency in the bleached state is important for architectural windows. Tungsten oxide has already a high transparency in the bleached state. Nickel oxide has a residual brown tint in the bleached state, which can be reduced by additives such as Mg, Al, Si, Zr, Nb, or Ta. These additives reduce the absorption for the short wavelengths in the visible and make these materials promising for the coming technologies in structural windows for energy efficiency and human comfort [44].

Pre-treatment before lamination The cathodic electrochromic film of WO3 is charged by addition of hydrogen in the gas mixture during sputtering [45]. This advantage makes the material ready to be laminated after deposition without any pre-treatment. On the other hand, before lamination the anodic electrochromic film of HyNiOx (or any of the mixes of nickel oxide with the metal oxides cited before) on the ITO coated substrate film is discharged by ozone exposure [46]. The anodic and cathodic charge must be

6

in balance for the device to bleach properly. The transfer of charge from cathodic to anodic film is achieved by an external potential after the assembly with the proper electrolyte.

1.6 Outline of this work

The present work focuses on sputter-deposited hydrated nickel based oxides that are used as the anodic film of an electrochromic device. Two goals have been achieved:

a. the optimization of the deposition conditions of nickel based oxides and improvement of the optical properties of the films by adding specific additives

b. the understanding of the coloration mechanism of the wet coloration (electrochemical switching) and dry coloration (ozone exposure).

Chapter 2 discusses the properties of bulk NiO and the hydroxides and oxy-hydroxides, followed by the definition of some important parameters to characterize the performance of the thin films and devices. Finally, the properties of thin films produced by different techniques are reviewed. The experimental techniques used to characterize the films are explained in chapter 3.

The film optimization is decribed in chapter 4, 5 and 6. Chapter 4 discusses thin film deposition by conventional reactive sputtering and reactive co-sputtering. Chapter 5 presents the physical characterization: composition, density and crystal structure. Chapter 6 is about the electrochromic properties, especially how the coloration efficiency is related with the deposition parameters, and hence with the physical properties of the films, and about advantages of the additives used to reduce the absorption at short wavelengths. A brief section on color properties of the films and devices is also presented.

The coloration mechanism is decribed in chapters 7, 8 and 9. The chapter 7 describes the dependence of the chemical diffucion coefficient on the stoichiometry. In chapter 8, the intercalation process is studied by x-ray photoelectron spectroscopy for the electrochemical and ozone colored films. A summary of results is presented in chapter 9.

7

A summary and conclusions are given in chapter 10. Chapter 11 contains a brief summary in Swedish as a requisite from the university authorities to obtain the PhD degree.

9

2. Properties of nickel oxides

Section 2.1 and 2.2 describe the main properties of bulk nickel oxide and nickel hydroxide and oxy-hydroxide, respectively. Section 2.3 is a review of the properties of electrochromic thin films of nickel oxide made by different deposition techniques.

2.1 Bulk properties of NiO

Nickel monoxide belongs to the 3d transition metal oxides with a NaCl structure and is an anti-ferromagnetic insulator in spite of the partially filled 3d orbitals. [47]. A lattice parameter of 4.173 Å has been reported [48]. Single crystal nickel monoxide has a density of 6.67 g/cm3 [48]. It is not possible to make stoichiometric NiO crystals; they are always metal deficient (excess of oxygen) as compared to the stoichiometric composition [49]. The extra oxygen cannot be placed inside the NaCl structure; instead vacancies of Ni2+ are created, giving a p-type conduction character [50,51]. In addition to the formation of vacancies, defects, such as impurities, are usually present in NiO samples [49].

The electrical conductivity of NiO depends strongly on the preparation method and has been reported to be in the range from 10-2 ( m)-1 (500K) to 10-10 ( m)-1 (30K) [47]. The valence band consists of localized nickel 3d states at approximately 2 eV from the Fermi level. Also, a wide oxygen 2p band at approximately 4-8 eV and a satellite at an energy of up to 8 eV was observed by XPS; this last feature is coupled with the main structure of the nickel 3d states [47]. The optical bandgap has been reported to be around 4 eV, and the refractive index is 2.33 at the photon energy of 2 eV [52]. The conduction band consists of unoccupied states of nickel 3d, 4s and 4p [53].

2.2 Nickel hydroxides and oxy-hydroxides

Nickel hydroxide has a hexagonal structure that consists of a hexagonalpacking of hydroxyl ions with Ni2+ occupying alternate rows of octahedral sites [54]. The nickel hydroxide exists in two modifications, α and β. The only difference between them is the quantity of water that is needed for stabilization. The α phase occurs at low content of water and the β phase athigh content [55]. The Ni(OH) structure of α form consists of a randomstacking of the Ni(OH) layers with a large interlayer separation (~0.76 nm),whereas the β-Ni(OH) has the same structure except that the interlayerseparation is much smaller (~0.46 nm) and hence the layers are well-aligned[54]. These two phases can be further de-hydrated (oxidized) to thecorresponding oxy-hydroxides γ and β for α and β phases, respectively, asproposed by Bode [56,57]. The reaction scheme is presented as:

2

2

2

(2.1)

A contraction is observed upon dehydration. A poorly crystallized phase (β ) of nickel hydroxide has also been identified [58]. This material is characterized by the coexistence of α and β phases [58]. The densities of the pure phases differ considerably, and have been reported to be 2.6-2.8 g/cmand 3.65-4.15 g/cm for α and β phases respectively [48,56,57,59,60]. The hydrogen diffusion coefficient is in the range of 10 to 10 cm /s [61].

+2

+2

-Ni(OH) -NiOOH + H + e

-Ni(OH) -NiOOH + H + e

−

−

↔

↔

BC

-12 2

3

3

-8

Nickel hydroxide is an n-type semiconductor with a bandgap around 3.6 to3.9 eV [62]. The oxy-hydroxide phase is a p-type semiconductor with a bandgap of around 1.7 to 1.8 eV [62]. The presence of extra states within thebandgap has been suggested because of extra absorption near 1.5 eV for the nickel oxy-hydroxide [62]. The poorly crystallized phase of nickel hydroxidealways has proton vacancies [63,64]. The number of vacancies increases asthe crystalline size decreases [63,64]. The refractive index of the hydroxidesat a wavelength of 633 nm is: 1.41 and 1.46 for α and β respectively [59,60].The oxy-hydroxides have a refractive index of 1.54 and 1.74, and an absorption coefficient of 7.74×104 cm-1 and 1.01×10 cm for γ and βphases, respectively, at the same wavelength [59,60]. The electronic structures of the hydroxides and oxy-hydroxides have the same features as

5 -1

10

11

NiO since they have the same nature of the charge transfer between Ni3d and O2p states as reported for other nickel compounds [65].

2.3 Thin film properties

This section is divided into subsections on chemical and physical deposition techniques. It is not the aim of this section to discus the experimental details of the particular deposition technique, but to compare the general properties of the films and some findings about the mechanism of coloration.

Before the properties are discussed, it is necessary to give some basic concepts that are needed to understand the performance of the electrochromic thin films. The main characteristics are the difference of the transmitted light between the colored and bleached states and the relation of the electrochemical properties to the absorption.

The integrated spectral transmittance is often used. The transmittance that the human eye perceives is called luminous and the maximum sensitivity corresponds to the photon wavelength of 550 nm. How much of the solar radiance that is transmitted is called the solar transmittance.

The reversibly intercalated charge (charge capacity) is connected with the change in the optical absorption through the coloration efficiency. The optical density is equal to the absorption coefficient times the thickness and is reported together with the corresponding photon wavelength at which it was measured. The coloration efficiency is the change in the optical density from colored to bleached state divided by the amount of transferred charge. The time response and cycling stability are also usually specified. Ageing is one of the main problems, but the durability of the films in a real device can be as high as >107 cycles [1,2].

These properties will be discussed with respect to the deposition technique. The most suitable technique for good quality electrochromic films is described in the end.

12

2.3.1 Chemically deposited films

The first attempt to produce electrochromic thin films of nickel oxide date to 1979. They were produced by anodic oxidation of plates of nickel metal and with electrodeposition [66]. Color changes were reported from colorless to a deep bronze color. A similar work was presented in 1985 [67] by use of the same methods and with similar results. However, none of the previous works gave any data of the charge capacity, and the cycling stability was very poor. Further, the bronze color in the dark state is indicative of the presence of nickel metal inside the films.

The films made by anodic electrodeposition have a crystal structure corresponding to nickel hydroxide [68-70] and they were found to be porous [71].

The anodic electrodeposited films have a solar and luminous transmittance modulation from 21% to 77% and from 35% to 75%, respectively, and the film thickness was around 50 nm [67]. Coloration efficiency was reported around -50 cm2/mC at 500 nm with an absorption coefficient of 4×104 cm-1

[72]. A charge capacity of 5mC/cm2 was reported [73].

The mechanism of coloration was explained by proton intercalation following the Bode reaction scheme cited in section 2.2 [68,69]. It has been claimed that hydroxyl groups also take part in the coloration mechanism [71]. The diffusion coefficient was reported to be between 1×10-11 to 1×10-14

cm2/s. Cathodic electrodeposited films showed some improvement in adherence and a small improvement in the charge capacity after thermal treatment [73]. Anodic films showed less activity after heating [74]. The electrodeposited films exhibit ageing after a few cycles [35,72-76].

Mixtures of the hydroxide made by cathodic electrodeposition with Cd, Ce, Cr, Co, Cu, Fe, La, Pb, Mg, Mn, Ag, and Y have been reported. Better transmittance modulation and durability were found with addition of Ce and La [35,76].

Thin films produced by sol-gel deposition showed a NiO bulk structure [77-81]. Grain size of the particles was found to be as small as 3 to 5 nm [77-79] and as large as 50 to 150 nm depending on the deposition conditions [80]. The nanocrystals of NiO were mixed with an amorphous phase [82]. The modulation of the luminous transmittance was reported from 30% to 80%

13

with 110 nm thick films. Coloration efficiency depends on the deposition conditions; values between -30 and -65 cm2/mC at 550 nm are typical [77,78,80]. Contamination coming from the precursors was found [77,78] and also ageing problems [81] as for the electrodeposited films. The coloration mechanism was attributed to hydrogen intercalation following the Bode reaction scheme (2.1) [82]. Spin coated thin films also present NiO particles from 10 to 20 nm. The modulation of transmittance at 633 nm was from 60% to 85%. The charge capacity was between 7 and 15 mC/cm2 and the cycling stability was up to 2 104 cycles [83].

Finally, some other techniques have been used, chemical vapour depositionusing high temperatures of 250°C to 300°C gives films with a NiO structure [84]. The luminous transmittance varied between 53% and 85% with 100 nm thick films, and the coloration efficiency was -44 cm2/mC at 550 nm [84]. The films produced by spray pyrolysis, exhibit a NiO structure [85-87] with grain sizes between 40 and 80 nm [87]. The luminous transmittance varies from 60% to 80% with the coloration efficiency of ~37 cm2/mC [86].

2.3.2 Physically deposited films

Thin films produced by reactive electron beam evaporation showed a NiO structure [88-90]. The films need hydroxylation to show electrochromism [88]. The hydroxylation was done by electrochemical cycling in a basic solution of KOH [88]. Two kinds of film morphology were found, a large surface area one with porous columnar crystals covered by hydroxide [90] and a highly textured mixed crystalline and amorphous structure [82]. The atomic ratio O/Ni of the films was found to be between 1.3 and 1.4 [89,90]. The density of the films was approximately that of the hydroxide, around 4.4 g/cm3 [89,91].

The solar transmittance varied from 23% to 84% [82] and the coloration efficiency was between -32 to -40 cm2/mC at 550 nm [82,89,92], similar to other deposition techniques mentioned before. The charge capacity between 5 and 16 mC/cm2 was reported for film thickness around 200 nm [89,92]. Cycling stability up to 500 cycles was reported without apparent degradation [89].

During coloration a shift of the bandgap was found from 3.65 to 3 eV. The optical density at 633 nm was changed from 0.3 to 1.5 [88]. The mechanism

14

of coloration of this material was not established, and different explanations have been proposed. The change of -nickel hydroxide to -nickel oxy-hydroxide [88] implies only a proton intercalation. The hydrogen transport was corroborated by nuclear reaction analysis [93,94]. Additional chemical analysis point to the coloration being due to transport of hydrogen and hydroxyl groups, where the water and hydroxyl groups are placed in lattice defects [91].

Thin films made by RF reactive magnetron sputtering were reported for the first time by Svensson and Granqvist in 1986 [33]. The films of NiOx, 200 nm thick, on ITO/glass substrates were produced in an atmosphere of O2 and Ar. Tests in a solution of 1M KOH showed excellent cycling stability up to 2 104 cycles, luminous transmittance modulation between 10% and 75%, and a charge capacity of 35 mC/cm2 [32,33].

Later works, including DC and RF reactive magnetron sputtering in O2 and Ar [95-101] or O2, H2 and Ar [101-103], showed small differences of the film properties depending on the deposition conditions. For this reason, the properties of DC and RF reactive magnetron sputtering will be taken in account here after as general characteristics of reactive magnetron sputtering.

The atomic ratio O/Ni was reported from 1.1 to 1.5 [104,105]. Further, the films were shown to consist of crystalline grains in an amorphous matrix [82], probably being the badly crystalline BC phase of nickel hydroxide. This kind of structure is probably the reason why the films present a high charge capacity and a substantial amount of hydroxyl endings due to the excess of oxygen as reported in [106].

In general, the films produced for electrochromic applications possess a polycrystalline cubic structure of NiO [99-103,105,107-109]. The density of the films was found to be between 3 and 6.5 g cm-3 [98,105]. The low density compared to the bulk NiO is indicative of porosity [104,108], which is related to surface electrochromic activity in the films [108]. The grain size depends strongly on the deposition conditions and ranges from 3 to 150 nm [82,99,104,108,110,111]. It has been recognized that the porosity and small grain size, leading to large grain surface-to-bulk ratio, increases the electrochromic activity of the films [107]. Improved crystallinity in the coatings was found upon addition of hydrogen to the gas atmosphere during the deposition of the films [103]. An excess of oxygen increases absorption

15

for the as deposited films [110,112]. Addition of hydrogen to the gas mixture produces transparent films and improves the electrochromic properties of the material [102,103,113].

The deposition technique gives a high deposition rate [101,112]. The deposition rate decreases with the overdose of oxygen [114,115] or hydrogen [112]. The cycling stability from 2 104 to 2 105 cycles has been reported [33,96]. It is also possible to produce films with high transparency in the bleached state [97].

Charge insertion/extraction causes pronounced changes in the transmittance and only minor changes in the reflectance [32]. The refractive index in the coloured state was found to be approximately 1.9 with 2.3 in the bleached state in the mid-infrared region [116]. The absorption coefficient was 3.52×104 cm-1 to 7.74×104 cm-1 and 5.03×103 cm-1 to 2.51×104 cm-1 for the colored and bleached state, respectively, at a wavelength of 500 nm [117]. Other works report a change in the absorption coefficient from 104 cm-1 to 105 cm-1 in the visible region upon zero to 9 mC/cm2 charge insertion [99,110].

There is a controversy in the literature about the type of ion transferred upon coloration, with evidence being presented for both proton [33,99,112,118,119] and OH- transport [93,120].

A proton diffusion coefficient of the order of 10-11 cm2/s [104] has been reported. A change in hydrogen concentration upon coloration was detected by nuclear reaction analysis with 15N+ ions [33,120]. Covering the coating with a palladium layer and cycling in KOH provided additional evidence of proton transfer, with reflectance modulation observed from the backside of the system Pd/NiOxHy/ITO/glass [119]. Since the layer of palladium catalyses hydrogen and no other species can penetrate through the Pd film, the change of absorption was attributed to the insertion – extraction of protons. On the other hand, a correlation was found between a high concentration of OH- species, low resistivity of the oxide film and an increase of the absorption [100].

Sputter deposited hydrated nickel oxide films have been intercalated with Li ions [121-123]. Irreversible intercalation with a structural transformation of the nickel oxide to pseudo cubic LixNi1-xOy, was reported [122]. The luminous transmittance was between 15% and 83% with coloration

16

efficiency of -40 cm2/mC [123] comparable with the proton intercalated films [82]. The diffusion coefficient was around 1 10-12 cm2/s [121]. The main contamination is due to the reactivity of Li in the presence of oxygen and carbon impurities leading to the formation of LiCO3 [123].

Some other physical techniques have also been used for the production of electrochromic nickel oxide. Pulsed laser deposition gives films with grain sizes of NiO from 50 to 150 nm with a very low coloration efficiency, approximately -15 cm2/mC [124]. Thermal oxidation of evaporated nickel metal at 300°C did not give any electrochromic response [125-127].

Finally, the hydroxylation of polycrystalline nickel oxide has been reported to occur at pressures as low as 10-8 mbar due to the presence of a partial pressure of H2O. The controlled growth of nickel hydroxide on top of the NiO by a treatment of O2 gas and H2O vapour produces between 1 and 5 monolayers of nickel hydroxide. This process was enhanced if the surface was sputter etched during the hydroxylation by gas treatment [128].

2.3.3 Summary

Many techniques have been used to produce electrochromic films of nickel oxide. Properties of the films have been reviewed above and only the ones produced by sputtering give a good cycling stability. Nevertheless, the electrochromic properties are similar for most of the films produced with other deposition techniques. Regarding reactive sputtering, DC sputtering produces similar results to RF sputtering. This is important since usually the RF sputtering is more technically complicated to implement than the DC one. The high rate deposition, no need of heat treatments, low contamination, and the general suitability for large area coatings make sputtering the best choice. A new trend in device construction, as mentioned in chapter one, is the use of plastic substrates. Hence, all techniques that need heat treatment will be difficult to implement.

17

3. Experimental details

This chapter gives descriptions of the experimental techniques. The chapter is organized in the following order: deposition technique and conditions (section 3.1), profilometry (section 3.2), ion beam analysis (section 3.3), X-ray diffraction (section 3.4), cyclic voltammetry (section 3.5), galvanostatic intermittent titration technique (section 3.6), optical UV-VIS-NIR measurements (section 3.7), infrared spectroscopy (section 3.8), X-ray photoelectron spectroscopy (section 3.9), and X ray absorption fine structure (EXAFS) in section 3.10.

3.1 Reactive DC and RF magnetron sputtering

Two kinds of sputtering can be distinguished: DC (dc voltage) and RF (radio frequency). Sputtering from metallic targets is carried out with DC magnetron sputtering.

DC magnetron sputtering consists of an abnormal glow discharge in inert atmosphere (ex. Ar), where the electrons are confined in front of the target by a magnetic field and the ions are accelerated by a DC electric field to the target (cathode). The positive ions in the plasma are accelerated towards the target and then removing atoms through momentum transfer. The material removed by these ions is deposited on a substrate as a thin film. Since most electrons lose their energy in non-ionizing collisions or are collected by the anode, only a small percentage of them take part in the ionization process with the inert gas atoms. The use of magnets increases the ionization and sputtering efficiency [129-131]

For the insulator materials, the sputtering is carried out by RF sputtering. A glow discharge is produced by a DC voltage and a high radio frequency (RF) alternating voltage is coupled capacitively through the insulating target to the plasma, removing the need for conducting electrodes.

18

If, in any of the sputtering processes, a reactive gas or a mix of inert and reactive gases is used, it is possible to deposit compounds during the sputtering of metal or insulator targets. This is called reactive DC or RF magnetron sputtering [129-131].

We deposited approximately 200 nm thick films of nickel-based oxides by single target sputtering and co-sputtering. Standart DC and RF reactive magnetron sputtering was used for metallic and insulator targets, respectively. Targets with purity of 99.95 % of Mg, Al, Si, Ni, Zr, Nb, Ag, Ta, Ni(61.5 %)-Al(38.5 %), and Ni(93 %)-V(7 %) were used. Only the deposition of Si was carried out with RF magnetron sputtering. The depositions were carried out from 5-cm-diameter targets in a mixture of Ar, O2, and H2, all being 99.998 % pure. The total sputter pressure was approximately 4.7 Pa (30 mTorr). Unheated glass substrates, pre-coated with 40-nm-thick ITO having a resistance of 60 /square, were used for optical and electrochemical work, whereas polished carbon substrates were used for composition analysis. Kapton foil substrates were used for EXAFS measurements.

Parameters Value

Target dimensions 50.8×6.35 mm

Base Pressure <3×10-7mTorr

Work Pressure ~30 mTorr

Typical DC Power 50-250 W

Typical RF Power 50-150 W

Target-substrate distance 13 cm

Target-substrate angle 50°

Substrate rotation 40 rpm

Table 3.1: Deposition parameters and details of the geometry of a sputtering system equipped with a turbo molecular pump.

The characteristics of the equipment are given in table 3.1. Parameters such as pressure and deposition geometry were kept constant for each set of samples, and only the flow ratio Ar/O2/H2 was varied; further in the text the

Measured between the substrate’s surface normal and the main direction of the sputtered flux.

19

quantity of specific gas is represented as a percent of the total flow of the gas.

It is well known that Ni is a magnetic material, and there are many Ni based alloys such as Ni(93 %)-V(7 %) that are non-magnetic. For sputtering a non-magnetic nickel alloy is desirable. However, not all nickel alloys are non-magnetic but can present weak magnetism. We explore several additives that can produce a weakly magnetic or a non-magnetic alloy with nickel. This issue is crucial for large-scale applications in which sputtering using a magnetic material brings a high monetary cost and a technological difficulty.

3.2 Thickness measurements

An Alpha Step profilometer was used to determine the thickness of the samples. The method needs a thickness step in the sample and uses a diamond tip to scan the step. The instrument has a resolution of 0.5 nm in the z-direction, and the range of scan length in the x-direction is between 80 and 2000 m with a scan speed between 25 and 0.2 s/ m.

3.3 Rutherford back scattering (RBS)

The analysis of the scattering events of high mono-energetic ions hitting a surface provides information about the concentration of the elements in the sample. In the case of Rutherford Backscattering Spectroscopy (RBS), ions of 4He at 2 MeV with a scattering angle of 166.1 degrees were used to determine the concentration of the heavy atoms in films. This technique measures the energy distribution of the incident ions that were elastically reflected. In this way, under considerations of energy losses, geometry, cross section of the elements, etc, it is possible to calculate the average concentration of the elements and the density [132].

20

3.4 X-ray diffraction analysis (XRD)

Structural analysis was performed by X-Ray Diffraction using a Siemens D5000 diffractometer operating at room temperature with a grazing incidence angle of one degree in parallel beam geometry with a (with

defined as the diffraction angle) between 30 to 120 degrees and a wavelength of = 1.540598 Å, corresponding to the CuK emission [133].

The effective grain size and the strain function were determined from the full width at half maximum FWHM of the X-ray diffraction peaks using the Scherrer equation (3.1), which accounts for the grain size broadening g and the lattice strain broadening s produced by dislocations and defects [134].

0 92 g s.FWHM S tan

L cos (3.1)

where L is the effective grain size, S the strain function, the Bragg angle in radians, and the factor 0.9 assumes spherical particles.

The FWHM was determined by fitting each peak by a Gaussian curve. Plotting the FWHM times the cosine versus the sine of the Bragg angle, the intersection of the curve will give the broadening contribution of the effective grain size and the slope the broadening produced by the strain in the film. In this case the strain was calculated assuming an isotropic variation along all directions [133].

The effective grain size should be understood as the equivalent diameter of the average volume in the film that diffracts coherently, and the strain is related to distortions caused by dislocations, defects, and vacancies.

Sample preparation for electrochemically cycled films — The films of hydrated Ni1-xVxOy were pre-treated by voltammetric cycling outside the vacuum chamber in 1 M KOH. The films were then rinsed with distilled water, dried with nitrogen, and put into the vacuum chamber for XRD and XPS analyses.

It is known for nickel oxide films [1] that a number of switching cycles are required in the beginning of cycling for the charge capacity to reach its maximum value and stabilize. The difference between samples II/III and

21

IV/V is that the capacity has been stabilized after 70 cycles, while the film is still in the transition region after 5 cycles, as shown in Fig. 3.1.

0 20 40 60 80

8

10

12

14

16

18

stabilized

Q (m

C/c

m2 )

Cycle number

un-stabilized

Figure 3.1: Evolution of charge capacity (Q) versus number of voltammetric cycles during charge stabilization for a 200 nm thick-film of Ni1-xVxOy deposited with oxygen and hydrogen percentages of 1.92% in the sputter plasma at 200 W. The film was cycled in 1 M KOH at a sweep rate of 10 mV/s.

Five sets of nickel vanadium oxide samples have been analyzed:

I. as-deposited,II. bleached after 5 voltammetry cycles,

III. colored after 5 voltammetry cycles, IV. bleached after 70 voltammetry cycles and V. colored after 70 voltammetry cycles

Sample preparation of ozone colored films — A transparent film of hydrated Ni1-xVxOy in the as deposited state was exposed to ozone during 6 minutes by using a UV lamp operated in air. The sample was deposited onto ITO coated glass. The charge stabilization has been done by cycling voltammetry 70 times. Four sets of nickel vanadium oxide samples have been analyzed:

22

I. as-deposited films, II. as-deposited films colored by ozone exposure,

III. charge stabilized films in the bleached condition. The films have been exposed to ozone prior to stabilization in 1M KOH, and

IV. films equal to samples III, colored again by ozone exposure.

3.5 Cyclic voltammetry (CV)

The electrochemical cell consists of a working electrode, which was the hydrous nickel oxide deposited on a glass/ITO substrate, a counter electrode of platinum, and a reference electrode of Ag/AgCl. In this technique, the potential was swept forward and backward between the working electrode and the reference electrode at a certain scan rate that usually is the same in both directions. The current between the working electrode and the counter electrode was recorded during the potential sweep [135,136]. The voltammograms have been recorded in a 1M KOH solution as in some other NiO works [33], between -0.65 V and 0.65 V versus Ag/AgCl. The experiment was carried out with an AUTOLAB PGSTAT10 potentiostat.

Charge capacity (Q) — This is the effective intercalated charge that is reversibly transferred upon cycling [136]. Maximization of this charge gives the maximum transmittance difference. Further, in electrochromic materials, the charge capacity is related to the coloration process, i.e, cathodic or anodic charge capacity. The usual units are expressed per area, mC/cm2. The charge capacity (Q) was obtained from the CVs and recorded during cathodic bleaching in order to avoid any possible contribution due to oxygen evolution.

Cycling stability This is the number of times that a coating or a device can be switched from the bleached to the colored state and vice versa [1].

3.6 Galvanostatic intermittent titration technique (GITT)

Galvanostatic intermittent titration technique (GITT) was carried out with an AUTOLAB PGSTAT10. The GITT technique consists of applying a current step for a short period of time that leads to intercalation or deintercalation of the ions from the electrolyte into the film. Later on, the circuit is opened and

23

the cell is allowed to relax until the steady state potential [137,138]. The chemical diffusion coefficient can be determined through analysis of the intercalation or deintercalation step and the relaxation mode. All of the measurements were carried out in a 1 M KOH solution in a three-electrode configuration cell.

3.7 Optical measurements

A Perkin Elmer double beam UV-VIS-NIR Spectrophotometer was used to measure the transmittance and reflectance in the range of 300 nm to 2500 nm at normal incidence using an integrating sphere. The background correction was taken for each scan. The reference for the reflectance measurements was barium sulphate.

In situ measurements were carried out by combining the three-electrode cell configuration explained in section 3.5 and an OCEAN OPTICS spectrophotometer in a wavelength range between 350 nm to 800 nm. The same electrochemical set up was used as in the CV.

The performance of the electrochromic thin films is characterized mainly by the modulation of the luminous and solar transmittance, the coloration efficiency, optical density, absorption coefficient and time response. Some definitions are presented here after.

The optical properties of the material are included in the frequency ( )dependent complex dielectric function (3.2) [139,140] (the subindex r and idenote the real and imaginary parts of the function) that is related to the complex refractive index (3.3) as follows:

r ii (3.2)

N n ik:

(3.3)

where n is the refractive index of the material and k is the extinction coefficient. The absorption coefficient is defined as:

4 iknc

(3.3)

with λ as the wavelength.

Optics of a slab in air A slab in air is represented in figure 3.2. Thethickness d, absorption coefficient α, and the reflectivity of the surface R0

are considered. If the interference of the light is neglected (λ<<d), the intensities of reflected and transmitted light due to an incoming beam Ipassing through are found as follows [140]:

0

Figure 3.2: Slab of material in air in which an incoming beam I0 goes through.Multiple reflections inside the slab are indicated together with the out comingtransmittance and reflectance intensity.

The reflected intensity IR is the sum of the reflected beams arising from allthe multiple reflections inside the slab and is represented by:

(3.3) ( ) ( ) ( )2 2 2-2 d 3 -4 d 5 -6 dR 0 0 0 0 0 0 0 0 0 0 0I =R I +R 1-R I e +R 1-R I e +R 1-R I e +

Then

( ) (2 -2 d 2 -2 d 4 -4 dR 0 0 0 0 0 0 0I =R I +R I 1-R e 1+R e +R e + ) (3.3.1)

Resolving the sum it is found:

( )2 -20 0 0

R 0 0 2 -2 d0

R I 1-R eI =R I +

1-R e

d

24

(3.3.2)

Finally, the reflectance is the ratio between the reflected intensity and theintensity of the incoming beam:

(3.4) ( )2 -20 0R

0 2 -2 d0 0

R 1-R eI= =R +I 1-R

Rd

e

)

The same is done for the transmitted intensity:

(3.5) ( ) ( ) ( )2 2 2- d 2 -3 d 4 -5 dT 0 0 0 0 0 0 0 0I = 1-R I e +R 1-R I e +R 1-R I e +

Then

(3.5.1)( ) (2 - d 2 -2 d 4 -4 dT 0 0 0 0I = 1-R I e 1+R e +R e +

Resolving the sum

( )2

T 2 -0

1-RI =

1-R e

- d0 0

2 d

I e (3.5.2)

Again, the transmittance is the ratio between the intensity of the transmittedlight and the incoming beam:

(3.6)

Absorption coefficient in a thin film The absorption coefficient (α) was calculated using T(λ) and R(λ) for the bleached and colored films using therelation reported in [141] derived from the Abeles [142,143] exactexpressions for T and R:

( )20T2 -

0 0

1-RI= =I 1-R

T- d

2 d

ee

(3.7)1 1 ( )( ) lnd ( )

å õ−= × æ öç ÷

RT

λλλ

The expression (3.7) can be easily deduced by combining the expressions(3.4) and (3.6) when the reflectivity of the surface R0 is neglected. For a thin film, interference effects cannot be neglected. Nevertheless, the interference

25

26

in the transmittance and reflectance are in anti-phase. It has been proved that the relation (3.7) compensates the effect of the interference quite well.

This approximation holds well when the refractive index of the substrate is between 1.5 and 1.7 and the refractive index of the coating is between 1.3 and 2.5. The expression gives a relative error of around 10 % and a maximum relative error less than 15 % for a high refractive index of the coating [141].

Optical density (OD) The quantity d is proportional to the extinction coefficient and it is called the optical density [144].

Coloration efficiency (CE) Combining the value of charge capacity with the optical measurements (transmittance (T) and reflectance (R)), it is common to determine the spectral coloration efficiency between the colored and bleached states using the relation [145]:

chargeExchanged

ColouredBleached

BleachedColoured

QTRTR

CE2

2

11ln

(3.8)

This expression is based on the relation:

21T R e d (3.8.1)

This is another simplification of equations (3.4) and (3.6) when the reflectance R0 is small and multiple reflections are neglected. In our case, R0

is less than 0.1. The expression (3.8.1) was used for direct comparison with published data, nevertheless, equation (3.7) is formally more correct since it can be derived from Abeles exact expressions of reflectance and transmittance taking into account the interference effects. Resolving (3.8.1) for d:

21

dR

lnT

(3.8.2)

Defining the change in the optical density:

27

d= d dcolored bleached (3.8.3)

Finally, the coloration efficiency is defined as:

d

Exchanged charg e

CEQ

(3.8.4)

An approximation of the coloration efficiency that is often used when the changes in reflectance are very small is:

Bleached

Coloured

Exchanged charg e

TlnT

CEQ

(3.8.5)

from the condition:

2

2

11

1Coloured

Bleached

R

R (3.8.6)

This coloration efficiency is a precise way to evaluate the change in the absorption due to the intercalation of charge. It is expressed in cm2/mC.

Beer’s law It is a modified expression of the relation (3.5) when the reflectance is neglected. It states that [144]:

ODT e (3.9)

Resolving for OD:

1OD lnT

(3.9.1)

Formally, this expression describes the attenuation of the light in isotropic and homogenous matter. Due to the interference, and the multiple light reflections that increase the optical path, the expression does not describe correctly the absorption in a thin film. Nevertheless, it is useful when an estimation of the OD is required.

28

Spectral absorptance The spectral absorptance A is calculated from [139]

1A T R (3.9)

Luminous and solar transmittance The coating performance is evaluated from the integrated solar transmittance using the AM1 solar spectrum

spectrumSolar)( and the luminous transmittance using the luminous

efficiency function for photopic vision Lu minous( ) [146]. The appropriate equations are:

spectrumSolar

spectrumSolarSolar d

TdT

)(

)()( (3.10)

Lu minousLu minous

Lu minous

d ( ) T( )T

d ( ) (3.11)

Time response The time scale which the electrochromic film or device needs for going from bleached to colored state and vice-versa. The time of coloring and bleaching are usually different. The area of the coating or the device affects the time response, that hence is characteristic for the kind of application that is intended. The magnitude of the time response goes from seconds to hundreds of seconds [1].

3.8 Infrared spectroscopy (IR)

A Perkin Elmer (model 983) double beam infrared spectrophotometer was used to measure reflectance in the range of 200 cm-1 to 4000 cm-1 with p polarized light and an angle of incidence of 60°. A gold thin film was used as reference. A film of nickel vanadium oxide onto ITO coated glass was cycled 70 times in 1M KOH using a three-electrode cell. The ex situ measurements have been done for as-deposited, colored and bleached states after rinsing the film with distilled water.

29

Infrared reflection-absorption spectroscopy (IRRAS) [147] is a well established technique for studying adsorbed substances or thin films (meaning: d = , i.e. no interference) insulator or semiconductor materials on metal surfaces [147]. The absorptance defined in (3.9), when the transmittance is neglected, takes the form:

1 measurep

substrate

RA RR

(3.12)

when p polarized light is used. The s polarized component vanishes due to the high conductivity of the metal.

Under following assumptions:

1. the magnitude of the dielectric function for the metal substrate (in this case ITO) is always larger than that of the absorbing layer;

2. the metal is an isotropic medium, 3. the metal has always 0metal

r in the infrared region, and 4. the metal is perfect, i.e. metal

r ;

the expression (3.12) after resolving the Maxwell equations with the appropriate boundary conditions and a tedious algebra, take the form [147]:

2sin11 4cospR d Im

c (3.13)

with as the incident angle and c the velocity of light in vacuum.

3.9 X-ray photoelectron spectroscopy (XPS)

X ray photoelectron spectroscopy was made in the Maxlab synchrotron facilities (line 411) in Lund, Sweden. A 1.56 GeV storage ring was used. The photon energy, monochromatized by a Zeiss SX-700 plane grating monochromator, was varied from 50 to 1500 eV. The end of the line uses a high resolution Scienta SES-200 hemispherical electron analyser. The instrument has an energy resolution E/dE from 103 to 104 and the photon flux on the sample is in the range of 1011 to 1013 ph/s.

30

The sample preparation was done as explained in section 3.4 The XPS spectra were calibrated using the adventitious carbon peak at 285 eV [148,149].

3.10 X-ray absorption fine structure (EXAFS)

X-ray absorption spectra (XAS) of transparent and ozone colored thin films and the reference compounds (crystalline c-NiO and c-V2O5) were recorded at the Ni and V K-edges using the synchrotron radiation from the LURE DCI storage ring. A standard transmission scheme at the D21 (XAS-2) beamline with a Si(311) double-crystal monochromator and two ion chambers, containing nitrogen gas, was used. All experiments were performed at room temperature with the energy resolution of about 1-2 eV.

The x-ray absorption spectra were analysed following standard procedure by the EDA software package [150]. The EXAFS part of the spectra was singled out by the Fourier filtering procedure and best-fitted within the multi-shell Gaussian/cumulant models using the theoretical amplitudes and phase shift functions, calculated by the FEFF8 code [151] for the two reference compounds. The E0 position, defining the zero photoelectron wavenumber value (k=0), was set to best match the experimental and calculated EXAFS signals for reference compounds. The XANES signals were qualitatively analysed by comparison of experimental data with the spectra, calculated theoretically by the FEFF8 code [151] within the full-multiple-scattering (FMS) approach. In all calculations, the complex Hedin-Lundqvist exchange-correlation potential (ECP) [152,153] was used to account for the inelastic losses of the photoelectron and the cluster potential was calculated in the self-consistent way.

31

4. Film deposition

Section 4.1 describes the optimisation of the electrochromic properties of sputter deposited nickel vanadium oxide films sputtered with a single non magnetic target in an Ar+O2 and Ar+O2+H2 gas mixture. Section 4.2 contains the characteristics of the co-sputtered nickel aluminium oxide in an Ar+O2+H2 gas mixture as an example of how a new material can be made with the optimum composition. Section 4.3 shows the optimum properties of nickel aluminium oxide sputtered using a single weakly magnetic target.

4.1 Sputtering of nickel vanadium oxide

Due to the magnetism of Ni, a non-magnetic Ni(93 %)-V(7 %) alloy was used as a target. The electrochromic properties of these films are very close to those of pure nickel oxide thin films. The deposition was conducted in a reactive atmosphere of Ar+O2 and Ar+O2+H2.

Figure 4.1 shows the film deposition rate and luminous transmittance as a function of O2/Ar ratio. Based on the film transmittance in the as-deposited state, three intervals can be distinguished as the amount of oxygen increases. At low oxygen content, corresponding to interval (1), films with close-to-metallic reflectance are deposited. In the interval (2), the films are transparent with the luminous transmittance as high as 74%. Finally, dark brown films are deposited at high oxygen content, corresponding to interval (3). These latter films are usually referred to as “the (typical) films deposited in an atmosphere of argon and oxygen are dark brown” in the literature [1,33,101,105,112]. It will, however, be shown below that the films from interval (2) possess the best electrochromic properties among the studied ones. From a sputtering point of view, the intervals 1, 2 and 3 correspond to metallic, partly oxidized, and completely oxidized target surface, respectively.

32

0.5 1.0 1.5 2.0 6.00

10

20

30

40

50

60 321

Met

allic

Tran

spar

ent

Bro

wn

r (nm

/min

)

2 (%)

0.5 1.0 1.5 2.0 6.0

0

20

40

60

80

100

Tlum (%

)

Figure 4.1: Deposition rate (r) and luminous transmittance (Tlum) in the as-deposited state of nickel vanadium oxide versus oxygen content in the sputtering atmosphere. Three intervals are indicated corresponding to nearly metallic (1), transparent (2), and dark brown (3) films. All films were deposited at 200 W and the thickness was kept constant at 200 nm.

The stable set point for the production of the films was found just before the abrupt change in the deposition rate from interval 2 to 3. In region 3, the target tends to oxidize after a long period of sputtering. The deposition rate can be increased by increment of power, since the quantity of sputtered metal increases. Then the curve in figure 1 will shift to the right; in this case the quantity of oxygen needed to deposit the oxide must increase in order to equilibrate the excess of metal.

The electrochromic properties are also strongly affected by the amount of oxygen in the sputter plasma. The nearly metallic films did not exhibit any optical modulation. The initially transparent films, however, displayed pronounced electrochromism with a bleached-state value of Tlum exceeding 80% and a coloured-state transmittance strongly decreasing within a narrow range of oxygen flow (cf. region 2 in Fig. 4.1). An increase of the oxygen content in the sputtering atmosphere, in this intermediate range also yielded a significant enhancement of the charge capacity. The optimum performance was found at oxygen concentrations somewhat below 1.5% at 200 W. Larger

33

oxygen flows—corresponding to region 3—led to deterioration of the optical modulation range as well as of the charge capacity.

Addition of hydrogen to the gas mixture reduces the deposition rate and increases the transmittance in the as-deposited state by approximately 10%. At high hydrogen content (H2>16%), the films are brown, which probably arises from lower deposition rate, implying more oxygen atoms per metal arriving to the substrate. Similar to figure 1, the intervals (2) and (3) correspond to transparent and dark as-deposited films, respectively. The optimum electrochromic performance was observed at a low (2-12%) amount of H2 in the gas mixture.

The percentage values of oxygen and hydrogen content in the gas mixture are relative numbers that depend of the characteristics of the equipment and the target manufacturer. In general, the optimum relative value of oxygen and hydrogen content will correspond to the value just before the target is oxidized and a decrease in the sputter rate is observed.

For pure nickel oxide the optimum O2/Ar ratio in region 2 is approximately 4.7%.

4.2 Co-sputtering of nickel aluminum oxide

We produced several nickel-based oxides (NiO mixed with Mg, Al, Si, V, Zr, Nb, Ag, or Ta) in order to find ways to reduce the brown tint—that is noticeable even in the bleached state of the “standard” nickel oxide films [2]. Simultaneously, we avoided the problem of strong magnetism in pure nickel, that is undesirable in large-scale applications.

The optimization was carried out by producing a set of samples at different power of nickel and the additive in a reactive atmosphere of Ar+O2 and Ar+O2+H2. The optical and electrochemical properties were measured. The films that showed the best electrochromic properties were used in further compositional analysis by RBS.

Figure 4.2 shows an example of the critical sputtering parameters for nickel-aluminium oxide films. An optimum composition for electrochromism was found to be around Al/Ni = 0.62. Similar optimizations were carried out for the other materials, except for vanadium. For this latter case we employed

34

“industrial nickel vanadium”; as explained before. It is important to note that even such a small amount of vanadium produces an optical absorptance that is clearly higher than in pure nickel oxide.

0 50 100 150 200 250 300

0.2

0.3

0.4

0.5

0.6

0.7A

tomic ratio A

l/Ni

Deposition rate

Dep

ositi

on ra

te (n

m/s

)

Sputtering power of Al (W)

0.0

0.2

0.4

0.6

0.8

1.0 Atomic ratio Al/Ni

Figure 4.2: Deposition rate and Al/Ni atomic ratio for nickel-aluminium oxide films produced at different powers delivered to the Al target. The sputtering power for Ni target was kept constant at 200 W.

4.3 Sputtering of nickel aluminum oxide

The next step was to produce the oxide films from a single nickel alloy target, using the optimum composition determined by co-sputtering. Figure 4.3 displays some sputter parameters for the optimum composition, with Al/Ni = 0.62. The target was weakly magnetic. The films could be made under different conditions: (i) at low oxygen contents (zero to 2 %) with the target in metallic mode, (ii) at oxygen contents from 2 to 4.3 % with the target being partly oxidized, and (iii) at an oxygen content exceeding 4.3 % in which case the target was in a fully oxidized state. This characteristic behaviour is representative of any of the nickel-based alloys when deposited from a single target.

35

0 1 2 3 4 5 6300

350

400

450

500

550W

ork pressure (mTorr)

(b)(a)

Al/Ni=0.62

Sputtering currentSp

utte

ring

curr

ent (

mA

)

O2/Ar (%)

30.0

30.5

31.0

31.5 Work pressure

Figure 4.3: Sputtering current and work pressure for nickel-aluminum oxide deposition from a single target in atmospheres containing different amounts of oxygen. The sputtering power was kept constant at 200 W. The (a) and (b) deposition conditions are close to the optimum electrochromic properties.

37

5. Physical characterization

Section 5.1 gives the optimum composition for electrochromic nickel based oxides. Section 5.2 gives the density of some of the films. Section 5.3 discusses the optimum structural characteristics for as-deposited, electrochemically cycled and ozone colored films. The last part of this section gives some information about the local structure for nickel vanadium oxide in the as-deposited and ozone colored thin films.

5.1 Composition

Elemental compositions were determined by use of RBS. Defining the composition as Ni1-xVxOy, the three regions correspond to x 0.11 and y0.71 for region 1, 0.05 x 0.10 and 1.45 y 1.75 for region 2, and x0.11 and y 2.07 for region 3. As pointed out before, films with optimum electrochromism were found in region 2.

The metal atom ratio for optimized nickel-aluminum and nickel-magnesium oxide films were found to be Mg/Ni = 0.80, Al/Ni = 0.56-0.62, the corresponding ratios of Si/Ni, Zr/Ni, Nb/Ni, Ag/Ni, and Ta/Ni have not yet been determined.

5.2 Density

The density lay between 3.6 and 4.2 g/cm3 in the optimum electrochromic nickel vanadium oxide films, obtained with intermediate oxygen and hydrogen flows during the sputtering. At the transition between regions 1 and 2, the density was approximately 6.6 g/cm3, and at the transition to the over-oxidized region 3 it was approximately 4.9 g/cm3. For the nickel magnesium oxide the density lay between 3.5 and 3.8 g/cm3 and for nickel

38

aluminum oxide between 3.0 and 4.0 g/cm3. Similar values of the density were found for nickel oxide films made by electron beam evaporation [118].

5.3 Structure analysis

X ray grazing incidence diffraction was used to analyze the films in the as-deposited (section 5.3.1), electrochemically cycled (section 5.3.2), and ozone colored states (section 5.3.3). The local structures for the nickel vanadium oxide films were analyzed by x-ray absorption spectra fine structure in the as-deposited and ozone colored states.

5.3.1 XRD in as deposited films

X ray diffraction was used to study the structure of the as-deposited nickel vanadium oxide films in figure 5.1. Specifically, films made so that they belonged to region 2 and deposited on glass exhibited a cubic nickel oxide (Bunsenite) pattern for 30 < <120 degrees, which is in agreement with previous work on NiO [105,154,155]. No evidence of phases containing hydrogen was found in any of the films, probably because only the outermost part of the grains—that was too thin to be detectable—takes part in the electrochromic coloration.

Figure 5.2 shows the effective grain size and strain function plotted against the hydrogen content. The grain sizes and the strain were estimated from XRD peaks widths, as explained is section 3.4.

30 40 50 60 70 80 90 100 110 1202θ

(420

)

(331

)

(400

)

(222

)

(311

)

(220

)

(200

)(1

11)

Inte

nsity

(a.u

)

(d)

(c)

(b)

(a)

Figure 5.1: XRD spectra for nickel vanadium oxide films deposited at a flow ratioof O2/Ar=2% and varying the hydrogen percentage from 0 (a), 1.92 (b), 12.07 (c)and 16.39 (d) keeping the pressure fixed at 4.7 Pa. The films correspond to theinterval 2. The sputtering power was kept constant at 200 W.

0 2 4 6 8 10 12 14 16 18101214161820222426 Effective grain size

Η2 (%)

Effe

ctiv

e gr

ain

size

(nm

)

0.00

0.01

0.02

0.03 Strain function

Strain function

Figure 5.2: Effective grain size and strain function for the Ni1-xVxOy films in figure5.1, deposited with sputter parameters belonging to region 2, as a function of the Hfraction of gas during sputtering.

2

39

40

Figure 5.3 shows examples of XRD spectra for some of the nickel-based oxides. All of the data indicate that the films are polycrystalline, and the differences in the peak intensities are indicative of different degrees of crystallinity.

30 45 60 75 90 105 120

NiMgO NiAlO NiVO

Inte

nsity

(a.u

.)

2 (degrees)

Figure 5.3. X-ray diffractogram for ~200-nm-thick Ni-oxide-based films on glass with different additives in their as-deposited states. The compositions correspond to Mg/Ni=0.80, Al/Ni=0.62, and V/Ni=0.07.

An interesting issue, of relevance to electrochromism, is whether the two metals form a single mixed-oxide phase or two separate oxide phases in the film. In the former case, it may be expected that the additive affects the electron structure—and therefore the optical properties—of the nickel oxide; in the latter case, the nickel oxide would simply be diluted by the additive. The x ray spectrum of the Ni-V oxide film, shown in Fig. 5.3, is consistent with the NiO structure, and no features attributable to vanadium oxides are detected. The spectrum suggests that the film is either a single phase, or that there is an amorphous phase of vanadium oxide present which is not detectable by XRD.

41

For the diffractogram of the Ni-Mg oxide film, it is not possible to distinguish whether the peaks correspond to a single mixed phase or to a mixture of NiO and MgO phases, because the separation between the characteristic lines in the reference spectra of the respective phases is smaller than the widths of the measured peaks in Fig. 5.3. Concerning Ni-Al oxide, no peaks corresponding to aluminum oxides have been detected.

5.3.2 XRD in electrochemically cycled films

Figure 5.4 shows XRD spectra at 40 < 2 < 65 degrees for the ITO-coated substrate as well as for as-deposited, bleached (after 70 CVs), and colored (after 70 CVs) hydrated Ni1-xVxOy films.

The XRD spectra were consistent with a cubic nickel oxide (Bunsenite) structure independent of the color state. All of the data indicate that the films are polycrystalline, and the differences in the peak intensities are indicative of different degrees of crystallinity. For the diffraction peak corresponding to (200), as detailed in figure 5.4, it is possible to see a shift of the position, indicating that stress is present as a result of the ion intercalation, in the colored state a contraction is noticed. Small changes can be detected also for the (220) diffraction peak. However, the contribution of the ITO base layer close to the (220) peak of the nickel oxide is very pronunced, the peaks get broadened as a result of lower crystallinity, and it may not be possible to assign the small changes in intensities to any specific crystalline phase. No evidence for any separate vanadium oxide phase was detected in the films, and no phases containing hydrogen were found in any of the coloration states.

42

43 44 50 51

(200

)

(d)

(c)

(a)

(b)

Inte

nsity

(a.u

.)

2 (Degrees)

40 45 50 55 60 65

Inte

nsity

(a.u

.)

2 (Degrees)

(a) As-deposited(b) Bleached(c) Colored(d) ITO

(200

)

(220

)

(a)

(b)

(c)

(d)

Figure 5.4: X-ray diffractograms of stabilized hydrated Ni1-xVxOy film in as-deposited, colored, and bleached conditions. The film was cycled 70 times in 1 M KOH prior to the analysis. A diffractogram of the ITO base layer is shown for comparison. The sputtering was carried out with O2/Ar=2% and H2/Ar=2% at a power of 200 W. The film corresponds to the interval 2.

5.3.3 XRD in ozone exposed films

The diffractograms of nickel vanadium oxide films in the as deposited and ozone colored states are plotted in figure 5.5. They are consistent with the NiO structure independent of coloration.

The grain size in the as deposited state was 11 +/- 4 nm and due to the dispertion in the plot of the FWHM×cos ( ) vs. sin ( ) the error is expected

43

to be bigger than 4 nm. There is no reason to expect that upon ozone exposure the grain size will change because the process is done close to room temperature and the modification is on the surface and not in the bulk. The as deposited film presents some texture.

The low crystallinity of the ozone exposed film with respect to the as deposited one is evident. The film after coloration shows the same shift as in the electrochemically cycled one indicating stress due to the ion intercalation as discussed in chapter 9. The exact value of the strain function due to defects is not reliable due to the dispersion in the values of the FWHM. Nevertheless, a rough estimation of the change of the strain function of the as deposited state with respect to the colored state was found to be approximately at -0.002, meaning that upon ozone exposure an increase of defects is produced, most probably due to the increase of Ni3+.

30 40 50 60 70 80

As-deposited Colored

Inte

nsity

(a.u

)

2 (degrees)Figure 5.5: X-ray diffractograms of as deposited and ozone colored hydrated Ni1-

xVxOy films. The films deposited on quartz were around 650 nm thick. The colored film was ozone exposed during 360 seconds. The sputtering was carried out with ratios O2/Ar=2% and H2/Ar=2% at a power of 200 W. The film corresponds to the interval 2.

44

5.3.4 EXAFS in as deposited and ozone exposed films

The spectrum in Fig. 5.5 suggests that either the film is a single phase or there is an amorphous phase of vanadium oxide present which is not detectable by XRD. In order to distinguish between the two possibilities Extended X-ray Absorption Fine Structure Spectroscopy (EXAFS) was used. The measured Ni and V K-edges, allowed us to follow the modifications in the local electronic and atomic structures as the films were colored in ozone.

Experimental XANES signals are shown in Figure 5.6. At the Ni K-edge in figure 5.6a, a difference between the c-NiO reference and a Ni1-xVxOy film is present in the pre-edge peak intensity and in the fine structure located above the edge. The pre-edge peak is associated with a transition from the 1s(Ni)state to the 3d(Ni) states mixed with the 2p(O) states. The pre-edge intensity is smaller in the films than in c-NiO, as was also observed previously in pure nanocrystalline NiO films [156]. This effect was interpreted within the Zaanen-Sawatzky-Allen model [157] as being due to an increase of the oxygen-nickel charge transfer energy accompanied by a reduction of the amount of ground state configurations with holes in the 2p(O) states. Thus a decrease of the pre-edge in the films can be associated with the fact that the Ni-O bonding becomes less covalent. It is noted that no significant change of the pre-edge peak occurs upon ozone exposure.

The V K-edge in the thin films is compared with data for c-V2O5 in Fig. 5.6b. The pre-edge peak is related to the transition 1s(V) 3d(V) + 2p(O). A comparison of two reference compounds, i.e. c-NiO and c-V2O5, shows that the intensity of the pre-edge peak is much higher in vanadium oxide because the 3d(V) states are more strongly mixed with 2p(O) states due to shorter V-O bonds and lower local symmetry at the V sites. The pre-edge intensity is relatively small in a transparent film but increases strongly upon coloration. Such behavior can be associated with the change of valence states from V4+ to V5+ [158]. Our preliminary FMS calculations suggest that this process is accompanied by an increase of the local distortion, responsible for a modification of fine structure located above the V edge.

45

5460 5490 55200.0

0.4

0.8

1.2

1.6

5465 5470 54750.0

0.4

0.8

8332 8334 83360.02

0.03

0.04

(b) V in N i1-x

VxO

y V 2O 5 A s-deposited O zone exposed

Nor

mal

ized

Abs

orpt

ion

Energy (eV)

8340 8370 84000.0

0.4

0.8

1.2

1.6(a) N i in N i

1-xV

xO

y

c -N iO A s-deposited O zone exposed

Nor

mal

ized

Abs

orpt

ion

Energy (eV)

Figure 5.6: Experimental XANES signals for the Ni (a) and V (b) K-edges in as-deposited (transparent) and ozone-exposed (colored) Ni1-xVxOy thin films. XANES data for the reference compounds c-NiO and c-V2O5 are shown in parts (a) and (b), respectively.

Figure 5.7 shows Fourier transforms (FTs) of the EXAFS signals for the Ni and V K-edges. A comparison of the FT signals for the films and for c-NiO allows to conclude that the films have a nanocrystalline structure; specifically, the peak amplitudes decrease more for the outer shell due to the size effect [156,159]. The first peak at 1-1.5 Å corresponds to the first coordination shell of Ni and V ions, composed of six oxygen atoms. The second peak at about 2.5-2.7 Å is due to three contributions: (i) nickel (Ni2)atoms in the second shell; (ii) oxygen (O3) atoms in the third shell; (iii) two double-scattering signals, generated within the 90°-triangular paths as M0 O1 O1 M0 and M0 O1 Ni2 M0 where M0 is the Ni0 or V0

central/absorbing atom. Our multi-shell fitting procedure of the Ni K-edge

46

EXAFS signals suggests that, on average, the film structure resembles that of c-NiO. However, some relaxation of the first coordination shell of the Ni ions, comprising six oxygen atoms, occurs in the films and leads to a decrease of the mean Ni-O distance by ~0.02 to 0.04 Å. At the same time, the second shell Ni-Ni distance increases by ~0.01 to 0.02 Å. The difference between the two films is mainly related to some increase of the disorder in the colored film, which is probed by the Debye-Waller factor and resulted in a relative decrease of the peaks for the outer shells in the FT.

0 1 2 3 4 5 6 7 80.0

0.1

0.2

0.3

0.4

0.5 (b)

As-deposited O zone exposed

V in N i1-xV xO y

|FT

| (k

)k

D istance R (Å )

0.0

1.0

2.0

3.0

4.0

5.0 (a) N i in N i1-xV xO y

c-N iO As-deposited O zone exposed

|FT

| (k

)k2

Figure 5.7: Fourier transform (FT) moduli of EXAFS signals at the Ni (a) and V (b) K-edges in c-NiO (part a) and Ni1-xVxOy thin films in the as-deposited (transparent) and ozone-exposed (colored) states.

The analysis of the V K-edge EXAFS signals suggests that V ions substitute Ni ions in the NiO-type structure and are distributed without any evidence for clustering. This conclusion is derived from the comparison of the experimental V K-edge EXAFS signals with the theoretical ones, calculated by the FEFF8 code for the V ion(s) placed at the Ni ion sites in the NiO structure. Our best-fit modeling suggests no presence of vanadium ions in the second coordination shell of vanadium, but indicates clearly the

47

vanadium first shell modification upon coloration. In transparent films, V ions are located at the centre of oxygen octahedra, whereas they shift to off-centre positions upon coloration. The displacement is ~0.4 Å in the [100] direction towards the nearest oxygen atom. Thus, V ions introduce a significant degree of disorder into the NiO-type structure of colored thin films.

49

6. Electrochromism

Section 6.1 discusses the electrochromic properties of NiMO films (with M being Mg, Al, Si, V, Zr, Nb, Ag, or Ta). The notation M indicates that an element is present in the oxide but does not specify the amount of it. Section 6.2 shows in detail the CV for Ni1-xVxOy. Section 6.3 presents the relation between intercalated charge and optical density. Section 6.4 compares the coloration efficiency for nickel vanadium oxide and nickel aluminium oxide. Section 6.5 shows the color properties of all NiMO. Spectral properties and color properties of ozone colored nickel vanadium oxide films are presented in section 6.6. Finally, some examples of the optical and color properties for laminated devices using NiMO and amorphous WO3 films on ITO coated glass and PET substrates are discussed.

6.1 Cyclic voltammetry

Cyclic voltammograms (CVs) were taken for nickel vanadium oxide films belonging to regions 2 and 3 defined in section 4.1. Figure 6.1 shows data for different gas mixtures used during the deposition. The CV for the relative oxygen flow of 6 % is representative of the over-oxidized region. We found large differences between films produced under different conditions. The electrochromism of films belonging to region 3 is characterized by low charge capacity and small optical modulation. Decrease of 50% to 90% in the charge capacity was seen for those latter films compared to films made under optimum deposition conditions. Furthermore, films of region 3 displayed severe ageing effects already after 20 to 30 cycles.

50

-0.6 -0.4 -0.2 0.0 0.2 0.4 0.6

-0.3

-0.2

-0.1

0.0

0.1

0.2

0.3

0.4

0.5

2= 6%

2= 0%

2= 1.96% 2= 0.00%

2= 1.92% 2= 1.92%

2= 1.72% 2= 12.07%

2= 1.64% 2= 16.39%

j (m

A/c

m2 )

Voltage vs. Ag/AgCl (V)

Figure 6.1: Cyclic voltammograms for 200 nm thick-films of Ni1-xVxOy deposited with the shown fractions of oxygen and hydrogen in the sputter plasma at 200 W target power. An example of films deposited with parameters belonging to region 3 is also shown. A scan speed of 10 mVs-1 was applied in a solution of 1 M KOH.

A stabilization of the electrochromic films, prepared under favorable conditions, took place after 10 to 20 cycles depending on the deposition conditions. Figure 6.2 presents the evolution of the CV from the as-deposited state to the stable state for one of the films of region 2. The increase in the intensities of the cathodic and anodic peaks implies that in the first cycles there is a growth of the charge capacity of the film. This may be interpreted as a transformation of the nickel oxide to nickel hydroxide, as we will discuss later. The shifts of the peaks, marked with straight dotted arrows, suggest that there is poor electronic conductivity through the films, as is explained in section 7.2.

51

-0.6 -0.4 -0.2 0.0 0.2 0.4 0.6-0.4

-0.3

-0.2

-0.1

0.0

0.1

0.2

0.3

0.4

0.5

2= 1.92%

2= 1.92%

j (m

A/c

m2 )

Voltage vs. Ag/AgCl (V)

Figure 6.2: Evolution of cyclic voltammograms from the as-deposited state to charge stabilization for a Ni1-xVxOy film prepared with the shown fractions of oxygen and hydrogen in the sputter plasma at 200 W target power. The CV was carried out in 1 M KOH at a sweep rate of 10 mV/s. Solid arrows show the scan direction and dotted arrows indicate the evolution of the peaks.

Cyclic voltammograms for various nickel based oxide films are shown in figure 6.3. Data for NiO are included as a reference. The shapes of the voltammograms change depending on the specific additive, but the main features characteristic for NiO prevail.

52

-0.6-0.30.00.30.6 NiMgO

NiO

j (m

A/c

m2 ) NiAlO

NiO

-0.6-0.30.00.30.6 NiSiO

NiO

j (m

A/c

m2 ) NiVO

NiO

-0.6-0.30.00.30.6 NiZrO

NiO

j (m

A/c

m2 ) NiNbO

NiO

-0.6-0.30.0 0.3 0.6

-0.6-0.30.00.30.6 NiAgO

NiO

j (m

A/c

m2 )

V vs. Ag/AgCl (V)-0.6-0.30.0 0.3 0.6

NiTaO NiO

V vs. Ag/AgCl (V)

Figure 6.3: Cyclic voltammograms of electrochromic nickel-oxide-based films in their bleached states for optimum compositions. The designation NiMO (with M being Mg, Al, Si, V, Zr, Nb, Ag, or Ta) indicates that M is present in the oxide but does not specify the amount. The films were about 200 nm thick and the NiO and NiVO were sputter deposited from a single target, the rest of the films were co-sputtered.

The charge capacity, shown in figure 6.4 was measured for the cathodic bleaching rather than the anodic coloration. In this way the effect of oxygen evolution was minimized. The charge density —and hence the magnitude of the electrochromism — is influenced by the potentiodynamic range, particularly the magnitude of the voltage for full coloration, Ucol. The Ucol

voltage was determined by the high voltage limit during cycling.

53

The role of Ucol is reported in Fig. 6.4 for 0.45 < Ucol < 0.65 V vs. Ag/AgCl. Except in the case of Zr and Mg, it appears that similar charge capacities (from 10 to 15 mC/cm2) can be obtained provided that Ucol is varied by 0.05 to 0.1 V when additives are present. This shift is insignificant for electrochromic device applications. The nickel zirconium oxide was further optimized with respect to composition and proportion of gases during sputtering, and apparently offers better electrochromic performance. Basically the same characteristics can be found with all additives.

0.50 0.55 0.60 0.65

5

10

15

20

25 NiMgO NiAlO NiSiO NiVO NiO NiZrO NiNbO NiAgO NiTaO

Q (m

C/c

m2 )

Ucol vs. Ag/AgCl (V)

Figure 6.4: Charge capacity of electrochromic nickel-oxide-based films. The designation NiMO (with M being Mg, Al, Si, V, Zr, Nb, Ag, or Ta) indicates that M is present in the oxide but does not specify the amount. The films correspond to the ones reported in figure 6.3.

6.2 Spectral absorptance in nickel based oxides

Figure 6.5 shows spectral absorptance (3.9) in the bleached state for the nickel-based oxides mentioned above. Prior to the measurements, the films were cycled ten times in a 1 M KOH solution to stabilize the properties. A significant decrease of A was found at short wavelengths for additives being Mg, Al, Si, Zr, Nb, and Ta, whereas the films containing V and Ag did not show any improvement in their optical properties compared to those of

54

pure nickel oxide. One could note that metals known to form oxides with wide band gaps tend to yield low bleached-state absorptance.

20

40

60 NiMgO NiO

Abs

orpt

ance

(%)

NiAlO NiO

20

40

60 NiSiO NiO

Abs

orpt

ance

(%) NiVO

NiO

20

40

60 NiZrO NiO

Abs

orpt

ance

(%) NiNbO

NiO

360 400 440

20

40

60 NiAgO NiO

Abs

orpt

ance

(%)

Wavelength (nm)360 400 440

NiTaO NiO

Wavelength (nm)

Figure 6.5: Spectral absorptance of electrochromic nickel-oxide-based films in their bleached states. The designation NiMO (with X being Mg, Al, Si, V, Zr, Nb, Ag, or Ta) indicates that M is present in the oxide but does not specify the amount. The films correspond to the ones reported in figure 6.3.

In general, irrespective of the additive it appears possible to produce electrochromic films with approximately the same electrochemical performance as for pure nickel oxide. This means that the main advantage of a mixed oxide lies in the improvement of the optical properties in the bleached state. As an example, Fig. 6.6 illustrates the spectral absorptance in the colored and bleached states for some of the films in figure 6.5. The

55

coloration charges were 15 mC/cm2 (bold curves) and 12 mC/cm2 (dashed curves) for the two films; this difference accounts for part of the difference in the colored state. It is clear that the additives do not cause any loss of coloration in the absorbing state. A further analysis of the Ni-V oxide system will be given below.

400 500 600 7000

20

40

60

80

100A

bsor

ptan

ce (%

)

Wavelength (nm)

NiVO NiAlO

Figure 6.6. Spectral colored-state and bleached-state absorptance for ~200-nm-thick Ni-oxide-based films of the shown types.

6.3 Absorption coefficient in hydrated Ni1-xVxOy

Optical spectra were found to be nickel-oxide-like and resemble data found in the literature [1,113]. Transmittance and reflectance were recorded for the as-deposited, colored and bleached films. Data for the colored and bleached films were recorded at maximum optical contrast, i.e., when the current density approached zero in the CV.

Absorption coefficients for bleached and colored films are shown in Fig. 6.7. The absorption coefficient ( ) was calculated using T( ) and R( ) for the bleached and colored films. Specifically, we used the expression (3.7).

56

0.5 1.0 1.5 2.0 2.5 3.0 3.5 4.00.0

0.5

1.0

1.5

2.0

2.5

3.0

1.5 2.0 2.5 3.0 3.50.030.060.090.120.150.18

×10

5 (cm

-1)

Energy (eV)

Colored

Bleached

×10

5 (cm

-1)

Energy (eV)

Figure 6.7: Absorption coefficient ( ) versus photon energy in the visible and near infrared regions for colored and bleached Ni1-xVxOy films. The inset shows extracted data on the broad peak in the colored state. The film corresponds to the one reported in figure 6.2.

Spectra for the bleached film did not exhibit any structure in the studied energy range. The absorption coming above 3.7 eV is due to the ITO base layer. If we plot ln( ) versus photon energy, a linear dependence can be observed between 3.5 and 3 eV for the bleached film. This is referred to as the Urbach effect, which is related to excitations between localized states in the band gap and the band edges [160]. Such a behavior was reported before for hydrated vanadium pentoxide and was then associated with the presence of V4+ states in amorphous V2O5 films [161,162].

The colored film had a broad transition from 1.5 to 3.7 eV. The increase of the absorption in the colored films is caused by charge transfer from Ni2+ to Ni3+ during the extraction of protons as will be explained below. Previous results on sputtered and electrochemically deposited films showed a more pronounced structure in the absorption spectra [163]. Contributions due to Ni3O4 species have been given as an explanation for this absorption band [163]. The inset of Fig. 4 shows the absorption coefficient extracted from the optical spectra after subtraction of a base line. The absorption peak is situated in the same region as that in which the contributions of Ni3+ have been predicted [163].

57

6.4 CV and optical density in hydrated Ni1-xVxOy

To illustrate the optical response upon insertion/extraction of charge, we used the film for which data were presented in Fig. 6.2. The film was deposited onto an ITO-coated glass substrate and it was cycled in 1 M KOH. The sweep rate was 0.10 mV/s in this experiment. We recorded the transmission at a wavelength of 550 nm and the current simultaneously. The optical density (OD) was calculated using Beer’s law (3.7) [144].

-0.6 -0.4 -0.2 0.0 0.2 0.4 0.6

-0.1

0.0

0.1

0.2

OD

j (m

A/c

m2 )

Voltage vs. Ag/AgCl (V)

j (mA/cm2)

0.0

0.2

0.4

0.6

0.8

1.0

1.2

2= 1.92%

2= 1.92%

Q = 14.1 (mC/cm2) OD at =550 nm

Figure 6.8: Current density (j) and optical density (OD) at the wavelength as a function of voltage for a Ni1-xVxOy film. The film is the same as reported in figure 6.2. The CV was carried out in 1 M KOH at 0.10 mV/s.

The reflectance changes are small from the bleached to the colored state, so that one can use the magnitude of OD as an approximation for the absorption.

The results are summarized in Fig. 6.8. Starting from a voltage of -0.55 V, we have a small absorption in the visible due to the extraction of charge and/or residual color due to the vanadium present in the films [161]. This appears as an optical density of about 0.1 for voltages around zero. The maximum absorption, as found from the OD, occurs at around 0.42 V. It is important to notice that application of high positive voltages, above 0.55 V,

58

does not contribute to the electrochromic process and only accelerates the oxygen evolution in the electrolyte.

6.4 Coloration efficiency: V versus Al

Combining the results of the CV with the optical measurements, we can determine the spectral coloration efficiency (CE) by the relation (3.6).

Results for nickel vanadium oxide films of region 2 are shown in Fig. 6.9. The main changes take place in the ultraviolet and visible, whereas the changes are very small in the near-infrared. Three main peaks can be distinguished in Fig 6.9: two prominent ones are located at wavelengths 340 and 445 nm, and a third very broad peak is seen around 600 nm that is confirmed in section 7.2. In general, the coloration efficiency is high compared to values reported in other work [1,102,110,144]. The influence of the sputtering conditions on the CE is strong at short wavelengths and weak at long wavelengths.

The CEs for 300 < < 500 nm correlate with the crystallinity of the films, and films with high CE showed large intensities of their XRD peaks. We could see direct evidence connecting the coloration efficiency with the grain size, presented in Fig. 5.1, and with the content of oxygen in the film. The films with small grain sizes and with low oxygen content had low CEs, whereas the films with more oxygen and larger grain sizes showed improved CEs. Finally, films with small grain sizes and high oxygen contents exhibit high CEs. This later films posses a large charge capacity, as measured by CV. This behavior supports the idea that the large inner surface area of the films is connected to the electrochromic activity, and that the coloration process takes place in the outermost parts of the grains.

59

300 400 500 600 700 800

-25

-50

-75

-100

O/Ni=1.432

O/Ni=1.376

O/Ni=1.497

O/Ni=1.509

=1.96%=0.00%

=1.92%=1.92%=1.72%=12.07%

=1.64%=16.39%

CE

(cm

2 /C)

(nm)Figure 6.9: Spectral coloration efficiency (CE) for Ni1-xVxOy films reported in figure 6.1 prepared with the shown fractions of oxygen and hydrogen in the sputter plasma. The O/Ni ratios determined by RBS are also shown.

Figure 6.10 shows spectral CEs for nickel aluminum oxide films made under sputtering conditions corresponding to points (a) and (b) in Fig. 3.3. Both of these films were optimized by an additional introduction of some hydrogen during the deposition. It is noteworthy that the films again have CEs that are much larger than those reported in the literature [1,102,110,144]. The data for hydrated nickel oxide is shown for comparison, all films with additives in figure 6.9 and 6.10 are similar to the NiO curve reported in figure 6.10.

As observed in figures 6.9 and 6.10, good functioning films are possible to produce with either V or Al as an additive. The main advantage of using Al is, as we mentioned, the high bleached state transmittance.

60

300 400 500 600 700 800

-20

-40

-60

-80

-100

-120

-140

NiAlO (b) (a)

Al/Ni=0.62

(b)

(a)

NiOC

E (C

/cm

2 )

Wavelength (nm)Figure 6.10: Spectral coloration efficiency (CE) of nickel-aluminum oxide films deposited from a single target and reported in figure 4.3. The sputtering power was kept constant at 200 W. The (a) and (b) deposition conditions correspond to the reported ones in figure 4.3. The CE for hydrated nickel oxide deposited with an optimum O2/Ar ratio of 4.7% is shown for comparison. The film thickness is 200 nm for all samples.

6.5 Color properties for nickel based oxides

The perceived color is a very important property for architectural windows, including “smart” ones [164]. Quantitative assessments can be performed in several different ways; here we employ the CIE Colorimetric System [165]. The purpose of colorimetric analysis is to give color specifications for observers with normal vision in terms of tristimulus values or chromaticity coordinates [146]. The ideal observer’s color matching functions are denoted x , y , and z , and represent red, green, and blue primaries, respectively. One can describe any color as an additive mixture of these. The y curve is chosen so that it coincides with the luminous efficiency of the

light-adapted eye. The CIE 1964 tristimulus values CIE CIE CIEX ,Y ,Z ,corresponding to a certain color stimulus , are obtained from

61

CIE

x dX

y S d , (6.4)

and analogously for CIEY and CIEZ . The color stimulus of present interest is

S T , (6.5)

where S is the relative spectral irradiance function, i.e, of the illuminants. Chromaticity coordinates, denoted x , y , and z , are then obtained from

CIE

CIE CIE CIE

Xx=X Y Z

, (6.6)

and correspondingly for y and. z These formulas lead to the chromaticity diagram shown in the lower panels of Figs. 6.11. Any color can be represented as a point within the shown boundary. The chromaticity coordinates for a colorless (achromatic) material are x=y=z=0.333

The chromaticity coordinates were calculated for four different standard illuminants representing a chosen set of light sources [146]. Illuminant D65 signifies the average north sky daylight at 6500 K; illuminant A represents a tungsten halogen, incandescent light source at 2856 K (typical home or store accent lighting); illuminant F11 pertains to a commercial, rare earth phosphor, narrow band fluorescent light source at 4000 K (used in Europe and the Pacific Rim for typical office or store lighting); and illuminant F2 represents a commercial, wide band fluorescent, cool white light source at 4150 K (typical office or store lighting in the U.S.A.).

62

0.0 0.2 0.4 0.6 0.8

0.00

0.20

0.40

0.60

0.80

0.32 0.34 0.36 0.38

0.32

0.34

0.36

0.38

400460480

490

500

540510530

590

560570

520

580600700

550

y

x

Illuminant Achromatic

Illuminant D65

Coloured

Bleached

NiO NiMgO NiAlO NiSiO NiVO NiNbO NiZrO NiTaO

y

x

Figure 6.11: CIE chromaticity diagram representing the color of optimized electrochromic nickel-oxide-based films in fully bleached and colored states under daylight illumination (CIE D65). Coordinates signifying the illuminant as well as colorlessness (i.e., the achromatic point) are plotted as a reference. The upper panel is a magnification of the central part of the lower panel. The designation NiMO (with X being Mg, Al, Si, V, Zr, Nb, Ag, or Ta) indicates that M is present in the oxide but does not specify the amount. The films correspond to the ones reported in figure 6.5.

Figure 6.11 shows chromaticity coordinates for the optimized thin films reported on in Fig. 6.3 under CIE standard illuminants D65. The details of the color properties under the illuminants A, F11, and F2 can be found in reference [44].

In Fig. 6.11, the data pertaining to the bleached state lie very close to the point representing colorlessness for the illuminant D65. The data for the colored state are rather scattered, with the chromaticity depending on the

63

specific additive. Also the trajectories in color space between the bleached and colored states are different for each of the additives. Another characteristic of the bleached state is that the dominant wavelengths are short so that the eye has less sensitivity. The chromaticity coordinates for the incandescent light indicate a light yellow color in the bleached state and a yellow-green color in the dark state. The data points for the bleached state appear less dispersed than the corresponding points for D65. For fluorescent lights sources the data are similar and dispersed for the dark states. The films show a green-yellow color in the bleached state.

NiO NiMgO NiAlO NiSiO NiVO NiNbO NiZrO NiTaO70

75

80

85

90

Lum

inou

s tra

nsm

ittan

ce Y

(%)

Nickel-based-oxides

Illuminant: D65 A F2 F11

Figure 6.12: Luminous transmittance YCIE for the nickel-based oxides under four different CIE standard illuminants in the bleached state. The designation NiMO (with M being Mg, Al, Si, V, Zr, Nb, Ag, or Ta) indicates that M is present in the oxide but does not specify the amount. The films correspond to the ones reported in figure 6.5.

Figure 6.12 shows YCIE, which corresponds to the luminous transmittance. The results are entirely consistent with those in Fig. 6.5. Films of nickel oxide mixed with Mg and Al show as much as 85 % luminous transmittance, whereas films containing Si and Zr yield up to ~83 %. For additives of Nb or Ta, the transmittance can be ~80 %. The case of the vanadium admixture is different, though, and shows a luminous transmittance not exceeding 75 %, i.e., lying ~4 % below the value for pure nickel oxide. Generally speaking,

64

the luminous transmittance shows a rather weak dependence on the specific illuminant.

6.6 Ozone coloration: optical and color properties

Figure 6.13 shows the absorptance in the bleached and colored states of a 650-nm-thick Ni1-xVxOy thin film on quartz upon ozone exposure for 360

seconds. The spectral absorptance A( ) was calculated using the relation (3.9). A huge modulation of the optical properties is observed within the solar range; it corresponds to a change of the visible color from transparent to dark brown.

500 1000 1500 20000

20

40

60

80

100 As-deposited Ozone exposed

Abs

orpt

ance

(%)

Wavelength (nm)Figure 6.13: Spectral absorptance within the solar range of 650 nm thick Ni1-xVxOy

films in the as- deposited (transparent) and ozone-exposed (colored) states. A quartz substrate was used and the deposition conditions are the same as reported in figure 6.2. The integrated spectral properties are shown in figure 6.14 for 200 nm thick films on glass coated with ITO. The behavior of the luminous and solar transmittance is similar. Both decay upon ozone exposure and saturate after 250 seconds.

65

0 50 100 150 200 250102030405060708090

Solar Luminous

Inte

grat

ed tr

ansm

ittan

ce (%

)

Ozone exposure (s)

Figure 6.14: Luminous (using illuminant D65) and solar transmittance of 200 nm thick films of Ni1-xVxOy upon ozone exposure. An ITO coated glass substrate was used and the deposition conditions are the same as reported in figure 6.2.

The chromaticity coordinates were calculated for the film in figure 6.14. The green (y) and blue coordinate (z) are plotted against the red one (x) in figure 6.15. Only three illuminants are presented, D65, F2 and A. The F11 is not plotted because its values are very close to the ones of F2. The color does not change after 250 s of ozone exposure. The evolution of the color, represented by the arrows in figure 6.14, shows that for the illuminant D65 and F2 (and F11) the red and green increase as the blue decreases. For the illuminant A the situation is the same for the red and blue but the green color does not show a significant change upon ozone exposure.

66

0.30 0.35 0.40 0.45 0.50

0.10

0.15

0.20

0.25

0.30

0.35

z

x

0.35

0.40

0.45

y

D65 F2 A

Figure 6.15: Chromaticity coordinates of 200 nm thick films of Ni1-xVxOy upon ozone exposure. The values for three illuminants are presented. The arrows show the direction of coloration. An ITO coated glass substrate was used and the deposition conditions are the same as reported in figure 6.2.

6.7 Devices: optical and color properties

Figure 6.16 shows the optical density in the bleached and colored states for two devices. The first device consists of amorphous HyWO3 and Ni1-xVxOy

films deposited onto ITO (15 per square) coated on glass. The second device had the same configuration but the films were deposited onto ITO coated PET with 40 and 15 resistance per square for tungsten and nickel oxides, respectively. A PMMA-based electrolyte was used to laminate the devices [166].

The tungsten oxide films were sputter deposited in an Ar/O2/H2 gas mixture using gas flows of 80/42/260 ml/min. The sputter power was 1.6 kW, the target area 50.8 × 12.7 cm2 and the distance from target to substrate 16 cm. The deposition of nickel vanadium oxide was as reported for region 2 in chapter 4. The work pressure was around 30 mTorr in both cases. The film

67

thicknesses were around 225 nm and 540 nm for tungsten oxide and nickel vanadium oxide, respectively.

The shape of the spectra is similar except in the near infrared region were the PET is absorbing (Figure 6.16b). In the visible region the glass based device shows slightly stronger coloration compared to the PET based one. The optical density at wavelengths below 400 nm is due to the ITO layers. The transfered charge was around 5 mC/cm2 for both devices.

600 1200 1800 24000.0

0.5

1.0

1.5

2.0

2.5

3.0

OD

Wavelenght (nm)

Colored Bleached

(a)

600 1200 1800 24000.0

0.5

1.0

1.5

2.0

2.5

3.0

OD

Wavelenght (nm)

Colored Bleached

(b)

Figure 6.16: Optical density in the bleached and colored states for two electrochromic devices comprising HyWO3 and Ni1-xVxOy : films on (a) glass and (b) PET. A PMMA-based electrolyte was used for lamination. The devices were switched from -1.6 V to 1.4 V. The film thicknesses were around 225 nm and 540 nm for tungsten oxide and nickel vanadium oxide, respectively.

The chromaticity coordinates in the bleached and colored states are shown in figure 6.17. The changes in color are practically the same for both devices independent of the illuminant. The illuminant F11 is not plotted because the values of the chromaticity coordinates are very close to the F2 illuminant. The green contribution (y) increases as the red (x) increases for the D65, F2 and F11 illuminants; in the case of the A illuminant a decrease of the green component is observed when the red component increases. The blue component (z) decreases upon increase of the red for all illuminants.

For the glass-based device, the variation in the luminous transmittance for the coloured and bleached states was between 20% and 58% and weakly

68

dependent of the illuminant. The solar transmittance changes from 20% to 60%. The PET based device shows a modulation in the luminous transmittance in the coloured and bleached states between 28% and 60% and is as well weakly dependent of the illuminant. The solar transmittance changes from 25% to 58%.

The PET based device is around 8% less dark in the colored state consistent with figure 6.16b. The difference in solar transmittance is around 7% higher for the glass based device.

0.32 0.36 0.40 0.44 0.48

0.12

0.16

0.20

0.24

0.28

0.32

z

x

0.36

0.39

0.42

y

D65 F2 A

Figure 6.17: Chromaticity coordinates for electrochromic devices with a nickel vanadium oxide films as a counter electrode operating in conjunction with hydrated amorphous WO3. A PMMA-based electrolyte was used. The devices were switched from -1.6 V to 1.4 V. The arrows show the direction of coloration. The film thicknesses were around 540 nm and 225 nm for tungsten oxide and nickel vanadium oxide, respectively. The empty circles correspond to the glass device, and the full circles to the PET device.

To show the improvements in the optical properties of the devices by doping the nickel oxide, figure 6.18 shows optical modulation at wavelengths of 400 and 550 nm for a 5×5 cm2 PET-based flexible electrochromic device.

69

Results obtained with a counter electrode of Ni-V-Mg oxide are compared with data for a device with a counter electrode of Ni-V oxide. Both devices have a WO3 electrode, prepared as explained before and a PMMA-based electrolyte [166]. As expected, the short-wavelength transmittance in the bleached state is higher and the optical modulation range wider for the device with the magnesium-containing counter electrode. The optical modulation at = 550 nm remains approximately the same for both devices. However, the transmittance is by a few percent higher for the magnesium containing counter electrode.

0 200 400 60010203040506070 = 400 nm

NiVO NiMgVO

Tran

smitt

ance

(%)

Time (s)

10203040506070

= 550 nmTran

smitt

ance

(%)

Figure 6.18: Optical modulation at wavelengths of 400 and 550 nm for electrochromic devices with a nickel-based oxide as a counter electrode operating in conjunction with amorphous hydrated amorphous WO3. A PMMA-based electrolyte was used. The devices were switched from -1.6 V to 1.4 V. The film thicknesses were around 225 nm and 540 nm for tungsten oxide and the nickel based oxides, respectively.

71

7. Proton diffusion and structural changes

Section 7.1 shows that the change of color of nickel-based oxide films is exclusively due to intercalation of protons. Section 7.2 describes the chemical diffusion using the Randles-Sev ik equation. Section 7.3 shows the relation of the chemical diffusion coefficient with the stoichiometry. Section 7.4 present the results from the infrared spectroscopy supporting the phase changes found from the analysis of the chemical diffusion coefficient.

7.1 Proton transport

In order to show that the proton transfer is responsible for the optical modulation, we used a palladium coating as a membrane that only allows H+

to penetrate through the film [167]. In this way, it is possible to filter out all other ions—such as OH- or K+ that may be present in the electrolyte—because they are too large to pass through the palladium coating. We did not see any activity of the Pd alone in the range of voltages employed in our study using a 1 M KOH solution.

A 200-nm-thick nickel vanadium oxide film on an ITO-covered glass substrate was coated with a 25-nm-thick layer of Pd. Figure 7.1 shows the CV for this working electrode in an electrochemical cell with three electrodes as described before. The CV of a similar film, that was not coated with Pd, is shown for comparison. One can see that the shape of the CV is similar for the coated and the uncoated film, except for the extra activity at the negative potentials. The peak at –0.474 V in the cathodic sweep for the coated film can be due to hydrogen evolution; alternatively, it may be related to the shoulder at 0.033 V in the anodic direction. This is increased since the Pd coating promotes hydrogen dissociation. It is important to notice that the value of the coloration potential does not change, implying that the absorbing compound is the same for both films. The bleaching potential, on the other hand, is enhanced to an extent that depends on the amount of water inside the film.

72

-0.6 -0.4 -0.2 0.0 0.2 0.4 0.6-0.8

-0.6

-0.4

-0.2

0.0

0.2

0.4

0.6

0.8

1.0

2=1.79% 2=8.93%

2=1.72% 2=12.07%

j (m

A/c

m2 )

Voltage vs. Ag/AgCl (V)

Figure 7.1: Cyclic voltammograms for 200 nm thick hydrated Ni1-xVxOy films prepared with the shown fractions of oxygen and hydrogen in the sputter plasma.Data are given for a film coated with 25 nm of palladium (solid curve) and without such a coating (dashed curve). A scan speed of 10 mVs-1 was applied in a solution of 1 M KOH. Arrows show the sweep direction. The sputtering power was 200 W.

Coloration efficiency, calculated from Eq. (5.8.5) without correction for the reflectance, is plotted in Fig. 7.2. The wavelengths where the coloration efficiency shows peaks are the same. For Pd coated and un-coated films a decrease of the CE for increasing wavelength can be observed. The similarities in the main features of the CE in Pd-coated and uncoated films suggest that the modulation of the optical absorption in the film is exclusivity due to the insertion/extraction of hydrogen. These results are consistent with the back side reflectance modulation reported for 100 nm Pd-coated nickel oxide films earlier [167].

300 400 500 600 700 800

0

-20

-40

-60

-80

-100

×2

Ο2=1.79% Η2=8.93%Ο2=1.72% Η2=12.07%

CE

(cm

2 /C)

λ (nm)

Figure 7.2: Spectral coloration efficiency (CE) for Ni1-xVxOy films prepared with the shown fractions of oxygen and hydrogen in the sputter plasma and also reportedon in Fig. 7.1. The curve for the film coated with a 25 nm thick layer of palladium(solid line) is multiplied by a factor of two.

7.2 Diffusion coefficient: Randles-Sev ik equation

This equation can describe the behavior of the current peak (I ), around 0.2-0.4 V in the CV in Fig 7.3, as a function of the concentration of the analyte(c ), the sweep rate ( ) at which the potential is scanned, the diffusioncoefficient of the analyte ( ), the electrode area (A), the number of electrons transferred per atom (n), and the temperature (T ). Those variables are connected by the semi-empirical Randles-Sev ik equation according to [135]:

P

analyte

o

D~

, (7.1)

where F and Rg is Faraday’s constant and the gas constant, respectively.

1/2

1/2 1/2P analyte

g 0

nFI =0.4463nFA D cR T

υå õæ öæ öç ÷

73

74

A film similar to the one showed in Figure 6.2, sputtered in a gas flow of O2 = 1.92 % and H2 = 1.92 %, was analyzed after stabilization. The CV for different sweep rates are plotted in figure 7.3.

-0.6-0.4-0.2 0.0 0.2 0.4 0.6-0.02

0.00

0.02

0.04

a.

=0.05 mV/s

j (m

A/c

m2 )

j (m

A/c

m2 )

j (m

A/c

m2 )

j (m

A/c

m2 )

-0.6-0.4-0.2 0.0 0.2 0.4 0.6

-0.4

0.0

0.4

0.8

d.

=10.0 mV/s

-0.6-0.4-0.2 0.0 0.2 0.4 0.6-0.1

0.0

0.1

0.2

b.

=0.50 mV/s

j (m

A/c

m2 )

j (m

A/c

m2 )

-0.6-0.4-0.2 0.0 0.2 0.4 0.6-0.8

-0.4

0.0

0.4

0.8

e.

=50.0 mV/s

-0.6-0.4-0.2 0.0 0.2 0.4 0.6

-0.3

0.0

0.3

0.6

c.

=5.00 mV/s

V vs. Ag/AgCl (V)

V vs. Ag/AgCl (V)

V vs. Ag/AgCl (V)

V vs. Ag/AgCl (V)

V vs. Ag/AgCl (V)

V vs. Ag/AgCl (V)-0.6-0.4-0.2 0.0 0.2 0.4 0.6

-0.4

0.0

0.4

0.8

f.

=500 mV/s

Figure 7.3: Cyclic voltammograms at six different sweep rates for a 200 nm thick hydrated Ni1-xVxOy film in 1 M KOH. The film was deposited with oxygen and hydrogen percentages of 1.92% in the sputter plasma. Arrows show the sweep direction. The sputtering power was 200 W.

Scrutinizing the changes of the CV upon varied sweep rate, we found that (I) the potentials for the different peaks shift away from the standard electrode potential, indicating that the electronic conductivity through the film is poor [136]; (II) the ratio between the intensities of the cathodic and anodic peak is less than unity; (III) the intensities of the cathodic and anodic peaks increase with the increase of the sweep rate; and (IV) the cathodic shoulder at approximately 0.4 V disappears at very slow sweep rate. The charge transfer

75

consists of two chemical steps, the first one being reversible and the following one irreversible. The latter follows from the fact that the shoulder for voltages higher than 0.4 V disappears at very slow sweep rates [136]. Potentials higher than 0.43 V led to decomposition of the electrolyte. The irreversible process could be responsible for the degradation of the films in this electrolyte, nevertheless, stress and fatigue due to the intercalation of charge need to be also taken into account. In our case, we found degradation after approximately 120 cycles in the electrolyte in all films when the over potential was as high as 0.7 V vs. Ag/AgCl.

0 1 2 3 4 5 6 7 8-0.7

-0.6

-0.5

-0.4

-0.3

-0.2

-0.1

0.0

j p (mA

/cm

2 )

1/2 (mV/s)1/2

Figure 7.4: Randles-Sev ik plot for the anodic peak current ( jP ) vs. square root of the scan rate ( ) for a Ni1-xVxOy film cycled in 1 M KOH. Data were taken from the curves in figure 7.3.

Figure 7.4 shows a Randles-Sev ik plot for the current density jP = IP/A of the cathodic peaks extracted from Fig. 7.3. We found good linearity between the peak intensity and the square root of the sweep rate. This linear dependence, marked with a straight line, represents the range of rates for which the process is controlled by diffusion phenomena. The intercept of the straight line close to zero current provides clear evidence that in the experiment we reduce the mass transfer by migration and avoid the Faradic losses in the process. From the slope of the curve in the linear part, the diffusion coefficient of the extracted species is found to be 1.62 × 10-17 cm2/s.

In order to estimate how much charge can be extracted from a nickel vanadium oxide film in the bleached state, we used the value of the chargecapacity for the lowest sweep rate and calculated the number of ionsequivalent to this charge. The sweep rate for this film was 0.05 mV/s, and the ratio of monovalent ions to nickel atoms was approximately 0.60. A similarresult has been reported previously [168].

7.3 Diffusion coefficient determined by GITT

Transport properties were investigated by GITT [137]. Films from regions 2 (section 4.1) of nickel vanadium oxide and nickel oxide, were studied usingthis technique. Prior to the GITT measurement, all films were stabilized by undergoing 70 cycles in 1 M KOH electrolyte. The sequential current steps in GITT were 0.08, 0.16, 0.24, 0.32, and 0.40 mA, in cycles of nine; the currents were applied for 20 s, and the voltage was recorded during a relaxation time of 3580 s. The steady state voltage was taken to be the one when the change was less than 0.08 mV during 10 s.

The chemical diffusion coefficient was calculated from the voltage response to the step current and the relaxation voltage response in an open circuit,using a relation based on the semi-infinite diffusion approximation according to [137]:

(7.2) 2 2

24 s ts

s

V V dD d tt Dt

∆π

å õ∂= × × <<æ ö∂ç ÷

Here is the thickness, ∆ is the change of the steady state equilibriumpotential, is the voltage response during the current step, ts is the steptime and t is the time. The relation between voltage and time for the semi-infinite approximation used in this model is

. (7.3)

d s

tVV

0 sV( t ) t t t∝ ≤ ≤

All voltage responses, obtained in this work upon application of a current step, show a linear dependence on the square root of the time as predicted byEq. (7.3). When we plotted the curve of the voltage response as a function of

76

77

the square root of the time, the square of the correlation coefficient for a linear fit was of the order of 0.9995 in all cases.

7.3.1 Comparison between NiO and Ni1-xVxO

There are no features of activity due to vanadium in the voltammograms shown in Fig. 7.3. To further elaborate on the possible influence of vanadium, we present data on D~ for a pure hydrated nickel oxide film, made from a magnetic target of 99.9% nickel in an Ar + O2 + H2 atmosphere (Fig. 7.5) and a nickel vanadium oxide film deposited under the same power and pressure conditions (Fig. 7.6).

-0.1 0.0 0.1 0.21E-18

1E-17

1E-16

1E-15

1E-14

1E-13

1E-12

1E-11

O2=4.76%H2=15.9%a.

D (c

m2 /s

)

SSV vs. Ag/AgCl (V)0.00 0.05 0.10 0.15 0.20 0.25

-0.1

0.0

0.1

0.2

O2=4.76%H2=15.9%

b.SSV

vs.

Ag/

AgC

l (V

)

H+/Ni

Figure 7.5: Chemical diffusion coefficient of H+ vs. steady state voltage (SSV) (a) and SSV vs. atomic fraction of hydrogen (with respect to nickel) (b) deintercalated from a nickel oxide film deposited with the shown fractions of oxygen and hydrogen in the sputter plasma. Current steps of 0.08, 0.16, 0.24, 0.32, and 0.40 mA (nine steps each) were applied during 20 s, followed by a relaxation time of 3580 s. We utilized a 1 M KOH solution in a three-electrode configuration with a Pt counter electrode and an Ag/AgCl reference electrode.

78

-0.1 0.0 0.1 0.2 0.31E-18

1E-17

1E-16

1E-15

1E-14

1E-13

1E-12

1E-11

O2=1.92%H2=1.92%a.

D (c

m2 /s)

SSV vs. Ag/AgCl (V)0.00 0.04 0.08 0.12 0.16

-0.2

-0.1

0.0

0.1

0.2

0.3

O2=1.92%H2=1.92%

b.SSV

vs.

Ag/

AgC

l (V

)

H+/Ni

Figure 7.6: Chemical diffusion coefficient of H+ vs. steady state voltage (SSV) (a) and SSV vs. atomic fraction of hydrogen (with respect to nickel) (b) deintercalated from a Ni1-xVxOy film deposited with the shown fractions of oxygen and hydrogen in the sputter plasma. Current steps of 0.08, 0.16, 0.24, 0.32, and 0.40 mA (nine steps each) were applied during 20 s, followed by a relaxation time of 3580 s. We utilized a 1 M KOH solution in a three-electrode configuration with a Pt counter electrode and an Ag/AgCl reference electrode.

The similarities between the data reported in Figs. 7.5 and 7.6 show that, effectively, the origin of the minima for D~ lies solely in the base material, i.e., in the hydrous nickel oxide. As seen in the potential curves for the two main changes of phase, we found well-defined plateaus at –0.1 V and 0.23 V in both cases they correspond to two well defined minima in D~ . A third minimum may be present at a SSV of 0.25 V.

In order to test the accuracy of the GITT method, we compared with the information on D~ calculated from the slope of the Randles-Sev ik plot (Fig. 7.4), using the cathodic peak, for the same film as the one reported on in Fig. 7.6. The diffusion coefficient was 1.6×10-17 cm2 /s, i.e., very close of the value of second minimum in Fig. 7.6.

7.3.2 Interpretation of the results

The appearance of a sharp minimum in the diffusion coefficient vs. potential can be understood as the result of a phase transition during the deintercalation of protons from the hydrated nickel-oxide-based films. The amplitude in the minima of the diffusion coefficient is associated with the

79

crystallinity, and well-crystallized films exhibit deep minima upon extraction and insertion of charge. We can identify phases involved in the de-intercalation process by comparing the potentials at which the minima of D~ appear and the optical density (Fig. 6.8). The SSV for the first minimum does not exhibit an increase of the OD but the second one is placed when the OD starts to rise. It is well known in fact that the nickel oxy-hydroxide phases have higher absorption than the nickel hydroxide phases [1].

We suggest that the first minimum in Fig. 7.5 and 7.6 corresponds to the transition from -nickel hydroxide to -nickel hydroxide, and that the second one stems from a transition from -nickel hydroxide to -nickel oxy-hydroxide. The third minimum, if not an artifact, could be assigned to a transition from the nickel hydroxide to -nickel oxy-hydroxide. The formulation of this mechanism only involves proton exchange, which in our case is the main process governing the coloration. This assignment is fully consistent with the Bode reaction scheme given in reaction (1), and we note that different phases of nickel hydroxide and nickel oxy-hydroxide have been reported to exist in the following approximate ranges of potentials (V) involved in the dehydrogenation of pure nickel hydroxide and nickel hydroxide containing cobalt hydroxide to their respective phases of nickel oxy-hydroxides [61]: -Ni(OH)2 for u < 0.15 V; -Ni(OH)2 for 0.15 < u < 0.27 V; -(NiO)OH for u > 0.30 V; and -(NiO)OH for u > 0.34 V, where uis the potential vs. Ag/AgCl.

It is seen that D~ approaches zero as a discontinuity at the phase transitions underlying the data in Figs. 7.5 and 7.6. This phenomenon can be understood from the lattice gas model with interactions [169]. The chemical diffusion coefficient and the chemical potential

H for the Frumkin-type isotherm

in this model are given by [169,170]:

XXXUD 1~

(7.4)

+ + 0 0H H 1g gXR T ln R T gX

Xo , (7.5)

where X is the fraction of ions deintercalated due to the step current, U is the mobility at X = 0, +H

o is the standard value of the chemical potential, g

80

is the interaction energy (divided by kBTo, where kB is Boltzmann’s constant) between the sites and includes the Coulombic interaction and the strain field in the lattice. Here g is negative for attractive interaction and positive for repulsive interaction. Combining Eqs. (7.4) and (7.5), it can be shown that

XXDgDD XX 1~~~00 . (7.6)

We do not have accurate measurements of the quantity of hydrogen present in the films in the beginning, and hence we can only make a qualitative analysis of the data presented in Figs. 7.5 and 7.6. At the two apparent phase changes, D~ decreases by four to five orders of magnitude at the actual transition. This means that the model of the Frumkin-type isotherm, in which the electron introduction/extraction is slow enough that quasi-equilibrium conditions may be assumed, is an excellent approach for a qualitative analysis [170]. When D~ approaches zero at the phase change, the “critical” value of g will be close to –4, which means that the net force between the ions is attractive [170].

7.4 Ex-situ IR of electrochemically cycled films

From the p polarized reflectance at an incident angle of 60° in the infrared region the energy loss function was calculated by using the equation (3.18).

A 200 nm thick film of nickel vanadium oxide deposited onto an ITO glass substrate was measured. The film was sputtered under the same conditions as the one reported in figure 7.6, and was stabilized by cycling 70 times in a 1M KOH solution. The measurements were carried out ex-situ after application of three different potentials. In the voltammogram in figure 6.2, this will correspond to the bleached state (-0.40 V), an intermediate state (+0.20 V), and the colored state (+0.45 V). Before the measurements the film was rinsed in distilled water.

In the intermediated state at +0.20 V the film stays transparent. A small difference in the optical density in figure 6.8 of around 0.1 is observed with respect to the bleached film at -0.40 V.

81

3600 3400 3200 3000

0.06

0.08

0.10

Im {

-1/

}

(cm-1)

+0.45 V

0.02

0.04

0.06

Im {

-1/

}

+0.20 V

[O-H]

0.02

0.04

0.06

Im {

-1/

}

V vs. Ag/AgCl-0.40 V

Figure 7.7: Energy loss function obtained from measurements with p polarized light at 60° angle of incidence in the region between 3850 cm-1 and 2950 cm-1.Three different voltages were applied corresponding to the bleached (-0.40 V), intermediate state (0.20 V), and colored state (0.45 V ).

The results for the three different potentials in the region between 3850 cm-1

and 2950 cm-1 are shown in figure 7.7. The broad band around 3400 cm-1 is indicative of water absorption [78,171]. After the edge at around 3650 cm-1

the stretching band i.e. [O-H] of the nickel hydroxides and oxy-hydroxides is placed [171-174] . This band undergoes a small red shift upon applied potential. This means that the strength of the bond weakens and the bond length between oxygen and hydrogen increases.

82

The region between 2300 cm-1 and 1200 cm-1 is shown in figure 7.8. The features correspond to deformed water [HOH] and carbonate impurities [78].

2000 1750 1500 1250

0.02

0.04

0.06

Im {

-1/

}

(cm-1)

+0.45 V

0.02

0.04

0.06

Im {

-1/

}

+0.20 V

0.02

0.04

0.06Im

{-1

/}

V vs. Ag/AgCl

-0.40 V

[HOH]

CO32-

HCO3

Figure 7.8: Energy loss function obtained from measurements with p polarized light at 60° angle of incidence in the region between 2300 cm-1 and 1200 cm-1. Three different voltages were applied corresponding to the fully bleached state (-0.40 V), intermediate state (0.20 V), and colored state (0.45 V).

Finally, the energy loss function at low frequency is shown in figure 7.9 for the region between 360 cm-1 and 860 cm-1. The discussion will be focused on two vibration bands: first the broad libration band [O-H]l [172] (defined as an oscillation in the apparent aspect of a secondary body (i.e, O-H) as seen from the primary object (i.e, Ni) around which it revolves) in the region

83

between 400 cm-1 to 560 cm-1, and the stretching band [Ni-O] around 590 cm-1 [174,175].

800 720 640 560 480 400

0.1

0.2

Im {

-1/

}

(cm-1)

+0.45 V-NiOOH

0.1

0.2

0.3

Im {

-1/

}

+0.20 V-Ni(OH)

2

0.1

0.2

0.3

Im {

-1/

}

V vs. Ag/AgCl

-0.40 V-Ni(OH)2

[Ni-O][O-H]l

Figure 7.9: Energy loss function obtained from measurements with p polarized light at 60° angle of incidence in the region between 860 cm-1 and 360 cm-1. Three different voltages were applied corresponding to the fully bleached (-0.40 V), intermediate state (0.20 V), and colored state (0.45 V).

To understand the changes in the libration band, we need to explain the changes in the stretching band [O-H] The shift to lower frequencies of the stretching band observed in figure 7.7 is the result of the combination of the local external fields at the hydroxyl group ion site created by the surrounding metal ions. This will cause a polarization of the O-H bond and will decrease

84

the strength of the bond [172,176]. The polarization is enhanced by applying positive potential and the strength of the bond is further decreased [173].

The increase of the partial covalence of the Ni-O bond due to the elongation of the O-H bond will increase the hybridization of the oxygen and nickel ions decreasing the bond length of the Ni-O. The libration band [O-H]l shiftsto lower frequencies due to the combination of the elongation of the O-H and the decrease of the Ni-O bond. The spectrum measured at -0.40 V can be assigned to the beta phase of nickel hydroxide and the one at +0.20 V to the alpha phase because of the contraction of the unit cell. This confirms the interpretation of the first minimum in figure 7.6.

When the same film was measured at +0.40 V; the [O-H]l band is furthershifted to lower frequencies and a strong band [Ni-O] appears as a consequence of the extraction of a proton of the alpha phase of nickel hydroxide being transformed to the gamma phase of nickel oxy-hydroxide [173].

85

8. Surface analysis by XPS

This chapter is divided into sections on X ray photoelectron analysis for electrochemically intercalated films (section 8.1), and ozone colored films (section 8.2).

XPS measurements have been done on core states of nickel, oxygen and vanadium, and on the states forming the top of the valence band.

Ni2p3/2 states in NiO give a double peak at binding energies (B.E) between 852 eV and 860 eV. For a perfect NiO crystal, the ratio between the lower B.E peak (centered at approximately 854.8 eV) and the higher B.E peak (centered at approximately 856.1 eV) is 4/3 [177]. As mentioned in chapter 2.1, a non-perfect nickel oxide contains excess oxygen. For each extra oxygen atom (nickel vacancy), there are two nickel atoms in 3+ oxidation state. These states give an XPS peak centered at approximately 855.7 eV, i.e., between the two peaks of the perfect NiO. In a measured XPS signal, the contribution of the 3+ states will cause the double peak to appear with an intensity ratio different than 4/3, and the extent of deviation from this ratio is proportional to the amount of nickel atoms in 3+ state in the film.

O1s states in oxide phases of NiO and Ni2O3 give a single peak centered at approximately 529.7 eV [177,178] and 531.4 eV [179], respectively. In the hydrogen containing phases of Ni(OH)2 and NiOOH the O1s peak lays at approximately 531.3 eV [180-182] and 531.7 eV [183], respectively. O1s

peak coming from adsorbed water or hydrohyl groups is centered at approximately 532.7 eV [128]. All these peaks overlap in a measured spectrum, and the intensity redistribution within the measured line is used to judge about possible phase transformations taking place upon coloration/bleaching, and during the charge stabilization in the beginning of the electrochemical cycling (shown in figure 3.1).

The measured spectra were normalized with respect to the area of the nickel 2p core states and the binding energy (B.E) calibration was done by placing the B.E of the adventitious carbon at 285 eV [148].

86

8.1 XPS in electrochemically intercalated films

Figure 8.1 shows XPS spectra for Ni2p3/2 states of a stabilized Ni1-xVxOy film. The areas under both curves in figure 8.1 are equal in magnitude. The deviation from the 4/3 ratio in figure 8.1 suggests a presence of Ni3+ states in both bleached and colored conditions. Intensity redistribution is seen between the colored and bleached conditions in figure 8.1, consistent with a larger number of nickel atoms in 3+ state in the colored film. This effect will be observed more strongly in ozone colored films discussed in section 8.2.

868 864 860 856 852

Bleached Colored

B.E (eV)

Inte

nsity

(a.u

.)

Ni2p3/2

Sat.

Figure 8.1: XPS spectra for Ni2p core states of charge stabilized hydrated Ni1-xVxOy

films in colored and bleached conditions. The films had been cycled 70 times in 1 M KOH. A photon energy of 1061 eV, with energy pass of 75 eV and energy step 0.1 eV, was used.

Figure 8.2 shows spectra pertinent to Ni2p3/2 states of non charge-stabilized hydrated Ni1-xVxOy films in as-deposited, colored, and bleached conditions. Again, the areas under all three curves in Fig. 8.2 are equal. The similarity of the spectra of as-deposited and bleached films suggests that there is a comparable amount of nickel atoms in 3+ state in these films. Similar to charge stabilized films, a larger contribution of Ni3+ states is seen for the colored film than for the bleached one during the initial cycles of electrochemical intercalation as well.

87

868 864 860 856 852

Inte

nsity

(a.u

.)

B.E (e.V)

Bleached

Sat.

Ni2p3/2

Inte

nsity

(a.u

.)

Colored

Sat.

Ni2p3/2

As-deposited

Inte

nsity

(a.u

.) Ni2p3/2

Sat.

Figure 8.2: XPS spectra for Ni2p core states of non charge-stabilized hydrated Ni1-xVxOy films in as-deposited, colored, and bleached conditions. The films had been cycled five times in 1 M KOH. A photon energy of 1250 eV, with energy pass of 40 eV and energy step 0.1 eV, was used.

The O1s states are shown in figure 8.3 for charge stabilized hydrated Ni1-xVxOy films in colored and bleached conditions. Considering the color apearence of different phases of nickel oxides and hydroxides, the spectra of the bleached film are consistent with a mixture of NiO and Ni(OH)2 phases. The spectrum of the colored film has an additional contribution of NiOOH and Ni2O3.

88

536 534 532 530 528B.E (e.V)

Inte

nsity

(a.u

.)

Bleached Colored O1sh =758 eV

h =1061 eV

Inte

nsity

(a.u

.)

O1s

Bleached Colored

Figure 8.3: XPS spectra for O1s core states of charge stabilized hydrated Ni1-xVxOy

films in colored and bleached conditions. The films had been cycled 70 times in 1 M KOH. Photon energy of 1061 eV and 758 eV with energy pass of 40 eV and energy step 0.1 eV, were used.

The data in Fig. 8.3 is grouped according to the photon energy. The same data is presented in Fig. 8.4 grouped according to the film coloration condition. For both color conditions shown in Fig. 8.4, the O1s signal is stronger at 758 eV than at 1061 eV photon energy. The analyzed depth of the surface layer is lower at lower photon energy; therefore, the results imply that the outermost surface of the grains is more oxygen rich than the bulk. There is approximately 35% more oxygen detected at 758 eV photon energy than at 1061 eV energy.

89

Figure 8.4: XPS spectra for O1s core states of charge stabilized hydrated Ni1-xVxOy

films in colored and bleached conditions. The films had been cycled 70 times in 1 M KOH. Photon energy of 1061 eV and 758 eV with energy pass of 40 eV and energy step 0.1 eV, were used.

Figure 8.5 shows spectra for O1s states for non-charge stabilized hydrated Ni1-xVxOy films in as-deposited, colored, and bleached conditions. The peak at approximately 529.7 eV is reduced in intensity upon going from the as-deposited to the bleached and colored conditions. This change of intensity can be attributed to transformation of the hydrated nickel oxide to hydroxide, and to oxy-hydroxide in the colored film. The transformation of nickel oxide to hydroxide may involve the excess oxygen, present in the films from the beginning, together with the hydrogen introduced during the cycling.

538 536 534 532 530 528B.E (eV)

Inte

nsity

(a.u

.)

BleachedO1s

Inte

nsity

(a.u

.) Colored

h =1061 eV h =758 eV O1s

90

536 534 532 530 528

Inte

nsity

(a.u

.)

B.E (e.V)

Bleached O1s

Inte

nsity

(a.u

.)

Colored O1s

As-deposited

Inte

nsity

(a.u

.) O1s

Figure 8.5: XPS spectra for O1s core states of non-charge stabilized hydrated Ni1-xVxOy films in as-deposited, colored, and bleached conditions. The films had been cycled five times in 1 M KOH. A photon energy of 1250 eV, with energy pass of 40 eV and energy step 0.1 eV, was used.

The O/Ni ratio on the surface was calculated by subtraction of the XPS background and subsequent integration of the spectra for the 2p states of nickel, divided by the corresponding values obtained for the 1s states of oxygen. The O/Ni ratio for non-charge stabilized films increases from the as deposited to the bleached state by approximately 7%, and further from the bleached to the colored state by approximately 2%, which may be caused by OH- uptake during the charge stabilization. For the charge stabilized films, the difference between the O/Ni ratio for the bleached and colored

91

conditions was less than 1%, implying that there was no significant transport of oxygen upon cycling. This result is important since it disproves the interchange of hydroxyl groups and points at intercalation of protons during the coloration after the charge has been stabilized.

We found nickel hydroxide features in the as-deposited film, and the tail at approximately 532.7 eV in figures 8.4 and 8.5—due to OH- in the H2Omolecule—is also present. During the extraction of protons, the OH- ions in the KOH solution can form water on the surface of the film. In such a case, it can be assumed that the excess oxygen in the films is in the form of OH-

groups residing in the NiO structure. Upon electrochemical cycling for the non stabilized films, the inserted protons can bind to NiO units, facilitating the transformation of nickel oxide to nickel hydroxide.

The V2p states for stabilized hydrated Ni1-xVxOy films in colored and bleached conditions are shown in Fig. 8.6. The EXAFS studies in section 5.3.4, showed that the V was present in the as-deposited films with valence V4+ [184]. There is no change in the B.E of vanadium between the bleached and colored states; therefore, it can be concluded that vanadium is not electrochemically active upon coloration. A difference in the V/Ni ratio (area under the curves in the top and bottom parts in figure 8.6) is observed between the colored and bleached conditions. For photon energies of 1061 eV and 758 eV the difference was 20% and 40%, respectively. There is no self-evident reason for such a decrese, one possibility is that a small displacement of the vanadium occurs ions upon coloration analogous to the one found by EXAFS in section 4.3. It was shown in section 7.4 that a contraction of the nickel hydroxide and nickel oxy-hydroxide occurs upon coloration. This may be the cause of less vanadium detected on the surface in the colored state.

92

520 515

Inte

nsity

(a.u

.)

B.E (e.V)

Bleached V2p3/2

Inte

nsity

(a.u

.)

Colored h =1061 eV h =758 eV

V2p3/2

Figure 8.6: XPS spectra for V2p core states of stabilized hydrated Ni1-xVxOy films in colored and bleached conditions. The films had been cycled 70 times in 1 M KOH. Photon energies of 1061 eV and 758 eV with energy pass of 40 eV and energy step 0.1 eV, were used.

Figure 8.7 shows spectra for V2p states for non-charge stabilized hydrated Ni1-xVxOy films in as-deposited, colored, and bleached conditions. The V2p

states did not show any B.E shift in any of the coloration conditions, meaning that no change of valence was induced upon cycling. The relative changes of the area of V/Ni between the as deposited and bleached states were negligible, in the colored state the area had decreased by approximately 30%.

93

520 518 516 514

Inte

nsity

(a.u

.)

B.E (e.V)

Bleached V2p3/2

Inte

nsity

(a.u

.)

Colored V2p3/2

As-deposited

Inte

nsity

(a.u

.)

V2p3/2

Figure 8.7: XPS spectra for V2p core states of non-charge stabilized hydrated Ni1-xVxOy films in as-deposited, colored, and bleached conditions. The films had been cycled five times in 1 M KOH. A photon energy of 1250 eV, with energy pass of 40 eV and energy step 0.1 eV, was used.

Figure 8.8 shows the top of the valence band for stabilized hydrated Ni1-xVxOy films in colored and bleached conditions. The position of the Ni3d

peak is shifted by approximately 0.6 eV towards higher binding energy for the colored film compared to the bleached film. A depopulation of electrons in the colored state with respect to the bleached state is observed, indicative of valence change from 2+ to 3+ upon coloration.

94

14 12 10 8 6 4 2 0

h =150 eVO2p

Ni3d

Sat.In

tens

ity (a

.u.)

B.E. (eV)

Bleached Colored

Figure 8.8: XPS spectra for Ni3d and O2p states of stabilized hydrated Ni1-xVxOy

films in colored and bleached conditions. The V3d and Ni3d states are overlapping. The films had been cycled 70 times in 1 M KOH. A photon energy of 150 eV with energy pass of 10 eV and energy step 0.1 eV, was used.

8.2 XPS in ozone colored films

Coloration by ozone exposure is an important but less studied method to treat nickel oxide films prior to incorporation in devices. Based on XRD results (section 5.3.3), it appears that the coloration involves a surface reaction rather than a bulk change. XPS was used to further study the coloration mechanism.

Figure 8.9 shows the nickel 2p3/2 states in the as-deposited and non stabilized ozone colored films. The usual two peaks of the nickel oxide are present in the as-deposited state. The contribution of the peak at lower B.E of these two peaks decreases in the colored films. Because this way of coloration produces stronger absorption (see figure 6.13), the nickel valence change from 2+ to 3+ is more pronounced than in the electrochemically colored films.

95

864 862 860 858 856 854 852

Inte

nsity

(a.u

.)

B.E (eV)

Ni2p3/2

Sat.

As-deposited Ozone colored

Figure 8.9: XPS spectra for Ni2p core states of as-deposited and non stabilized ozone colored hydrated Ni1-xVxOy films. A photon energy of 1061 eV, with energy pass of 75 eV and energy step 0.1 eV, was used.

Figure 8.10 shows the nickel 2p3/2 core states for the films in the as deposited, and stabilized bleached and ozone colored conditions. The same trends are observed as in the rest of the XPS spectra presented in this work. The lower B.E peak, seen in the as deposited condition, diminishes upon cycling or ozone exposure of the film, and a chemical shift of 0.5 eV is observed between the colored and the bleached conditions.

870 867 864 861 858 855 852

Inte

nsity

(a.u

.)

B.E. (eV)

As-deposited Electrochemically bleached

(stabilized) Ozone colored

Sat.

Ni2p3/2

Figure 8.10: XPS spectra for nickel 2p3/2 core states of as-deposited, stabilized bleached and ozone colored hydrated Ni1-xVxOy films. A photon energy of 1061 eV, with energy pass of 75 eV and energy step 0.1 eV, was used.

96

The oxygen 1s states for the as deposited and non-stabilized ozone colored hydrated Ni1-xVxOy films are shown in figure 8.11 for two different photon energies. At 1061 eV the spectra are more representative of the bulk and at 758 eV more sensitive to the outer most part of the surface. The figure 8.11a shows the oxygen 1s signal for a photon energy of 1061 eV, the zoom figure shows the peak of nickel oxide at approximately 529 eV and of the nickel hydroxide at approximately 531 eV.

536 534 532 530 528

518 516 514

Inte

nsity

(a.u

.)

B.E (eV)

O1s

Water

536 532 528 524 520 516

As-deposited Ozone colored

Inte

nsity

(a.u

.)

B.E (eV)

V2p3/2

O1s

V2p1/2

V2p3/2

(a)

536 534 532 530 528

518 516 514

Inte

nsity

(a.u

.)

B.E (eV)

O1s

Water

536 532 528 524 520 516

As-deposited Ozone colored

Inte

nsity

(a.u

.)

B.E (eV)

V2p3/2

O1s

V2p1/2

V2p3/2

(b)

Figure 8.11: XPS spectra for O1s and V2p core states of as-deposited and non stabilized ozone colored hydrated Ni1-xVxOy films. Photon energies of 1061 eV (a) and 758 eV (b), with energy pass of 40 eV and energy step 0.1 eV, were used. The insets show the detail of the V2p core states. The zoom figures shows the detail of the oxygen 1s states for both photon energies.

The O/Ni ratios at both photon energies decrease upon coloration. This is shown in figure 8.12, where the scale is the same for the as deposited and ozone colored films. The relative changes of O/Ni ratio for photon energies of 1061 eV and 758 eV were 24% and 15%, respectively, between the as deposited and colored state. The vanadium 2p (inset figure in 8.10) states shift to lower B.E confirming the change of valence from 4+ to 5+ observed

97

in the XANES spectra in figure 5.6. The relative changes in the V/Ni ratio are around -2% at 1061 eV and +2% at 758 eV.

534 532 530 528

Photon energy h =758 eV h =1061 eV

Inte

nsity

(a.u

.)

B.E (eV)

O1s

As-deposited

534 532 530 528

Ozone colored

B.E (eV)

O1s

Figure 8.12: XPS spectra for O1s core states of as-deposited and ozone colored hydrated Ni1-xVxOy films. Photon energies of 1061 eV and 758 eV are indicated. The spectra correspond as the ones in figure 8.11 and the scale is the same for the as deposited and ozone colored films.

Figure 8.13 shows the O1s core states for the as deposited, stabilized bleached and ozone colored films. There is a shift of approximately 0.3 eV between the bleached and the colored state and 0.7 eV between the as deposited and the bleached state. The relative change of the O/Ni ratio between the as deposited and the bleached state at 758 eV was around 25% and no changes of the O/Ni ratio between the bleached and the colored state was observed. Similar changes were found at 1061 eV. These changes confirm that in the stabilization process oxygen transport ocurrs in the form of hydroxyl ions as well as protons, but after stabilization only protons produce the color changes. The outer most part of the surface is more oxygen rich after the stabilization.

The same O/Ni ratio in the bleached and the ozone colored (stabilized) films can be explained by coloration being the result of hydrogen extraction rather than oxygen addition. We cannot provide a clear interpretation to the observed decrease of O/Ni ratio upon ozone exposure of the as-deposited films.

98

534 532 530 528

Photon energy h =758 eV h =1061 eV

B.E. (eV)

Inte

nsity

(a.u

.)

As-deposited

534 532 530 528

Electrochemical bleached(Stabilized)

B.E (eV)534 532 530 528

Ozone colored afterstabilization

B.E (eV)

Figure 8.13: XPS spectra for O1s core states of as-deposited and stabilized bleached and ozone colored hydrated Ni1-xVxOy films. Photon energies of 1061 eV and 758 eV with energy pass of 40 eV and energy step 0.1 eV, were used.

518 517 516 515 514

Inte

nsity

(a.u

.)

BE (eV)

V2p3/2

As-depositedElectrochemically cycled (stabilized)

Bleached in 1MKOH Ozone colored

Figure 8.14: XPS spectra for V2p core states of as deposited, stabilized bleached and ozone colored hydrated Ni1-xVxOy films. Photon energies of 1061 eV with energy pass of 40 eV and energy step 0.1 eV, were used.

XPS spectra for V2p core states of as deposited, stabilized bleached and ozone colored hydrated Ni1-xVxOy films are shown in figure 8.14. The same differences are observed between the different coloration conditions as were observed for the non-stabilized films shown in Fig. 8.11.

The valence band spectrum of hydrated Ni1-xVxOy films in the as-deposited and non-stabilized ozone colored conditions is shown in figure 8.15. A

99

depopulation of states is observed in the top of the valence band in the colored condition compared to the bleached one.

12 10 8 6 4 2 0

Inte

nsity

(a.u

.)

B.E (eV)

Ni3d

Sat.

O2p

As-deposited Ozone colored

Figure 8.15: XPS spectra for Ni3d and O2p states of hydrated Ni1-xVxOy films in the as-deposited and non stabilized ozone colored conditions. The V3d and Ni3d states are overlapping. Photon energy of 150 eV, with energy pass of 10 eV and energy step 0.1 eV, were used.

The valence band of hydrated Ni1-xVxOy films in the as deposited, stabilized bleached and ozone colored conditions is shown in figure 8.16. The spectra preserve the main features seen in figure 8.15. The magnitude of the shift of the 3d nickel states are of the same order of magnitude as the ones in the oxygen core states. The top of the valence band shows extra states on the low energy side in the bleached condition compared to the colored condition, apparently due to the presence of electrochemically active vanadium 5+. Vanadium states are not visible in figure 8.15 sugesting that the oxidation of vanadium is more efficient after charge stabilization, probably due to the extra oxygen incorporated upon electrochemical cycling.

100

12 10 8 6 4 2 0

Inte

nsity

(a.u

.)

B.E. (eV)

As-deposited Electrochemical bleached

(stabilized) Ozone colored

Ni3d

O2p

Sat.

V3d

Figure 8.16: XPS spectra for Ni3d and O2p states of hydrated Ni1-xVxOy films in the as deposited, stabilized bleached and ozone colored conditions. The V3d and Ni3d

states are overlapping. Photon energy of 150 eV, with energy pass of 10 eV and energy step 0.1 eV, was used.

101

9. Mechanism of coloration

This section summarizes the main experimental results and puts forward the mechanisms of coloration for nickel based oxide films upon electrochemical cycling and ozone exposure. Based on optical and electrochemical measurements, the coloration is considered to be due to the base material (i.e. nickel oxide) even in the films with other metal additives. The role of the additives is only considered to influece the bleached state transmittance of the films.

There is an excess of oxygen in the as deposited films in combination with low density (high porosity) and small grain size. The outer most part of the surface is richer in oxygen than the bulk. Independently of the color condition, there are always nickel atoms in valence states 2+ and 3+ in the films. The number of atoms in 3+ state increases upon coloration.

NiO structure was found by XRD for films in all color conditions, while the presence of other nickel phases (Ni(OH)2, Ni2O3, NiOOH) was detected by XPS. This indicates that the coloration process is a surface phenomenon, most likely in the outer parts of the grains rather than a bulk change. This correlates with the fact that the electrochromic activity increases when the inner surface area is increased by the decrease of the grain size.

The two XPS peaks of the Ni2p3/2 states can be interpreted as follows: the peak at low BE is attributed to the scattering of the core hole with pure d electrons and the peak at higher BE to the scattering of the core hole with hybridized d electrons [185]. In this framework, the change in the 2p states is related to an increase the scattering of the Ni3d electrons being hybridized with the O2p electrons. However, X-ray standing-wave investigations of the electronic structure of the valence band in a NiO single crystal showed that the contribution of the O2p states is of the order of 1/40 of the contribution of the Ni3d band [186]. For these latter data [186], the spectra were recorded with the electric field intensity placed at the Ni or O atomic sites and showed

102

a resonance only in the O2s states, giving mainly a Ni3d nature of the top to the valence band.

In the beginning of electrochemical cycling, a number of switching cycles are required for the charge capacity to reach its maximum value and stabilize (Fig. 3.1), with changes in the film stucture and composition observed during the stabilization. XPS results are consistent with a transformation of the overstoichiometric hydrated nickel oxide to hydroxide and oxy-hydroxide phases on the grain surface. The amount of oxygen in the films increases by approximately 25% by intercalation of OH- groups occurring simultaneously with intercalation of protons from the electrolyte. GITT results [187] suggest a crystallization of the hydroxide and oxy-hydroxide phases during the initial intercalation cycles.

After the charge capacity has been stabilized, experiments with Pd-coated films together with XPS results show that the electrochemical coloration occurs only by proton intercalation, with the oxygen content in the films being equal in bleached and colored conditions.

Upon dehydrogenation, GITT measurements were consistent with the Bode reaction scheme (2.1). The presence of the nickel hydroxide and oxyhydroxide phases and the lattice contraction upon extraction of hydrogen was confirmed by the IR spectra.

The Bode reaction scheme does not explain the 2+ to 3+ valence change for nickel belonging to NiO phase upon coloration, as observed by XPS (Fig.8.3). It is known that excess oxygen in NiO produces a Ni2+ vacancy that is compensated by creation of a hole on two Ni2+, thus producing Ni3+ as illustrated in reaction (9.1) [188]. To account for the XPS results, a process additional to the Bode reaction can be suggested, similar to (9.1). This is described by reaction (9.2), involving both proton extraction from the proton-containing phase Ni(OH)2 and valence change of nickel atoms belonging to the NiO phase. According to (9.2), the extraction of H+ causes a transformation from Ni(OH)2 to NiOOH, and the extraction of another H+ is compensated by creation of a hole on the Ni2+ belonging to the NiO unit

103

(9.1)

(9.2)

Approximately three monolayers have been estimated as the depth of resolution for the Ni2p and O1s core levels at a photon energy of 1061 eV. Because the coloration is a surface effect and the bulk remains NiO independently of whether the film is bleached or colored, one can argue, after the transformation of the Ni(OH)2 to NiOOH, that the change of the NiO to Ni2O3 may be in the inter-phase between the NiO and the nickel oxy-hydroxide. Such an interphase forms a bridge between the cubic NiO and the hexagonal phases of Ni(OH)2 and NiOOH. This may explain the shift of the NiO (200) diffraction peak in figure 5.4 for the colored state because the contraction on the outermost part of the grains will produce an increase of the strain in the inner part.

For the stabilized films, XPS results (Fig. 8.4) showed that the scattering of the core hole with hybridized d electrons is larger in the colored state than in the bleached state. The top of the valence band, as apparent from Fig. 8.8, showed a shift of approximately 0.6 eV in the peak position of the Ni3d states in the colored film compared to the the bleached film. This shift can be the result of a larger overlap (i.e., stronger hybridization) of the Ni3d and the O2p

states in the colored condition. This argument supports the explanation of why the stretching [O-H] and the libration band [O-H]l in the IR spectra (section 7.4) shift to lower frequencies upon coloration.

Considering the above results and arguments, the coloration mechanism of nickel oxide films can be written as:

2

2

2 2 3 2 2

Ni(OH ) Bleached

Ni(OH ) NiOOH H eBleached Colored

NiO Ni(OH ) Ni O H eBleached Colored

β

α γ

α

+ −

+ −

−

− ↔ − + +

+ − ↔ + ⋅ + ⋅

(9.3)

During the stabilization upon initial switching cycles an additional (irreversible) reaction is taking place:

(9.4)

Additives of Si and Ag to nickel oxide films have shown electrochemicalactivity upon coloration/bleaching, additives of Mg, Al, Zr, Nb V and Tahave not. Vanadium has shown a complex behaviour by becomingelectrochemically active only after ozone exposure. Further studies are required to explain the observed electrochemical activity of some of the additives.

As shown in Sec. 6.2, the additives that form their own oxides with wide band gaps, i.e., Mg, Al, Si, Zr, Nb and Ta, increased the bleached state transmittance of nickel oxide films, whereas the additives of V and Ag did not cause any improvement in optical properties compared to those of pure nickel oxide. With Al/V being examples of high/low band gap additives, one can consider two reasons for changes in optical properties upon other metaladdition to nickel oxide films:

1. Al oxide has a direct band-gap as large as 8.8 eV [189], while fullyoxidized V has a direct band-gap of 3.3 eV and also allows indirect transitions at about 2 eV due the existence of a split-off localized bandlying approximately 0.6 eV below the main conduction band [161,162];considering that the optical band-gap for Ni oxide is 3.6 to 4 eV [47,51],the addition of Al may increase the band-gap of the compound whereasthe effect of V may be to diminish it.

2. Alternatively, Al as well V may affect optical absorption caused bydefects such as vacancies, over-stoichiometry, grain boundaries, etc.Thermodynamically stable nickel oxide is a p-type conductor due to

oxygen rich NiO OH NiOOH eBleached Colored− −− + → +

104

excess oxygen [50,51]. It is then plausible that the p-type conductivityand the residual optical absorption in the bleached state originate fromthe same electron states. This may then explain why films of pure Nioxide cannot be made completely colourless. When Al is added, it canact as a donor of electrons and fill the electron (hole) states on Nithereby reducing the residual absorption. The addition of V, on the otherhand, may provide acceptor states whose effect would be to enhance theresidual absorption.

Concerning the coloration by ozone exposure, we note that, similar to electrochemical coloration, there is a larger amount of nickel atoms in 3+ state in the colored condition than in the bleached one. XPS results oncharge stabilized films show that the ozone coloration occurs by protonextraction, with the oxygen content belonging to nickel oxide/hydroxidephases being equal in colored and bleached films. An increase in oxygencontained in water or hydroxyl groups has been detected upon coloration.These results may be encountered by the following reaction sheme:

(9.5)3 2 2 2

3 2 2 3 2 2

2 2O Ni(OH ) NiOOH H O OBleached Colored

O Ni(OH ) NiO Ni O H O OBleached Bleached Colored

+ ⋅ ↔ ⋅ + +

+ + ↔ + +

For as deposited and charge non-stabilized films, further studies are neededto explain the changes in the films upon ozone exposure. The effect of ozone treatment on the other metal additives to nickel oxide requires a further attention as well.

105

10. Summary and conclusions

Electrochromic properties of sputter-deposited nickel-based oxide films have been studied with a two-fold goal. From a practical point of view, the optical switching performance has been improved by optimizing the deposition conditions and film stoichiometry with respect to oxygen and hydrogen, and further by adding Mg, Al, Si, Zr, Nb or Ta to the films. From a theoreticalpoint of view, dtails of the coloration mechanism have been studied bymeans of electrochemical intercalation (CV, GITT), optical measurements(UV, VIS, NIR and MIR), RBS, XRD, XPS and EXAFS.

Optimization of deposition conditions has been illustrated by the example of films made by sputtering of a non-magnetic Ni(93)V(7) % wt. target in anatmosphere of Ar/O /H . The optimized films exhibit transmittancemodulation between 20% and 75 % at 18 mC/cm charge intercalation. Theremaining problem with nickel oxide and nickel vanadium oxide films is their residual yellow-brown color tint in the bleached state, which disappearsas the short-wavelength transmittance increases upon addition of Mg, Al, Zror Ta. Optimization of deposition conditions by co-sputtering from twotargets and the film composition for mixed oxide films has been illustratedby the example of nickel aluminium oxide.

2 22

Concluding the studies, the following mechanism has been put forward for electrochochemical coloration:

2

2

2 2 3 2 2

Ni(OH ) Bleached

Ni(OH ) NiOOH H eBleached Colored

NiO Ni(OH ) Ni O H eBleached Colored

β

α γ

α

+ −

+ −

−

− ↔ − + +

+ − ↔ + ⋅ + ⋅

,

with an additional (irreversible) reaction during the initial switching cycles:

107

oxygen rich NiO OH NiOOH eBleached Colored− −− + → +

For coloration by ozone exposure, the main features are described by thefollowing reaction:

3 2 2 2

3 2 2 3 2 2

2 2O Ni(OH ) NiOOH H O OBleached Colored

O Ni(OH ) NiO Ni O H O OBleached Bleached Colored

+ ⋅ ↔ ⋅ + +

+ + ↔ + +

108

109

11. Sammanfattning på svenska

10.1 Bakgrund

Elektrokroma material ändrar reversibelt sina optiska egenskaper mellan genomskinligt och mörkt tillstånd under införsel/utdragning av laddning. Material som färgas under införsel kallas katodiska material; material som färgas vid utdragning kallas anodiska material. Metalloxider av W, Ti, Nb, Ta och Mo uppvisar katodisk elektrokromism medan oxider av V, Cr, Mn, Fe, Co, Ni, Rh och Ir uppvisar anodisk elektrokromism. Vanadinoxid uppvisar anodisk eller katodisk elektrokromism beroende på dess oxidfas. Det är inte alla faser av dessa metalloxider som har elektrokroma egenskaper. I många fall används en blandad oxid för att förhöja en särkild egenskap i ett tunt materialskikt (film), såsom färg, cyklisk stabilitet etc.

Dessa material kan integreras i elektrokroma anordningar av två typer:

1. fasta-tillstånds konstruktioner (tunna filmer i flera lager) 2. laminerade konstruktioner (tunna filmer i fast form ihoplaminerade

med en viskös polymerelektrolyt)

Båda sorters anordningar kan användas med eller utan självförsörjande kraftkälla såsom solceller.

Figur 11.1 visar tvärsnittet av ett ”smart fönster” och dess delar. Anordningen består av en ”skiktkonstruktion” på två underlag/substrat som är belagda med en genomskinlig elektrisk ledare och en elektrokrom film, och sedan laminerats ihop med en jonledare. Detaljerna i delarna förklaras enligt följande:

110

Underlagen ___ Anordningen kan konstrueras på ett fast underlag av glas eller en böjlig folie av en polymer såsom poly (-etylen terephtalat) eller poly (-etylen nepthalat) känt kommersiellt som PET och PEN.

Genomskinlig elektronledare ___ Underlagen är belagda med en genomskinlig ledande oxid såsom In2O3:Sn (ITO) och SnO2:F .

Figur 10.1: Grafisk representaion av ett ”smart fönster” baserat på en kombination av elektrokroma hydratiserade Wolfram- och Nickeloxidskikt. Pilarna visar transporten av joner under ett yttre elektriskt fält.

Katodiska elektrokroma lager ___ Många material har använts som ett katodiskt elektrokromt lager. Wolframoxid (WO3) interkalerat med H och Li är vida använt för dess klara blekta tillstånd och dess höga effektiva infärgningsförmåga.

111

Anodiska elekrokroma lager ___ De material som kan användas som ett anodiskt lager är huvudsakligen den hydratiserade formen av iridiumoxid (IrO2), vanadindioxid (VO2) och nickeloxid (NiO) interkalerade med H eller Li.

Genomskinliga jonledare ___ Två typer av jonledare kan användas, en fast film av en metalloxid eller en viskös polymerelektrolyt. Filmer av Ta2O5,Nb2O5, CeO2 och ZrO2 har använts som protonledare och litiumledare. Litiumledande polymerer har varit baserade på poly (metyl metacrylat) (PMMA) polymeriserat med propylenkarbonat (PC) eller propylenglykol (PPG), plus ett tillägg av en lämpligt Li salt.

10.2 Smarta fönster baserade på hydratiserad WO3 och NiO

Speciell uppmärksamhet har ägnats åt att konstruera elektrokroma hydratiserade NiO filmer i samverkan med elektrokrom hydratiserad WO3.Fördelen med denna kombination i en elektrokrom anordning är:

a. låga elektriska spänningar, exempelvis runt 1.5 V. b. möjligheten att uppnå neutral grå färg i det mörka tillståndet c. den höga genomskinligheten i det bleka tillståndet d. den enkla produktionen av filmer färdiga för laminering e. enkel beläggningsprocess genom likspänd (DC) reaktiv

magnetronsputtring

Punkt (a) är en inneboende egenskap hos materialen i fråga och den neutrala färgen (b) beror på den blå och bruna färgen hos den laddade WO3 och urladdade NiO filmen, respektive. Punkterna (c), (d) och (e) förklaras i detalj som följer.

Hög genomskinlighet ___ För fönster i byggnader är det viktigt att uppnå hög genomskinlighet i det blekta tillståndet. Wolframoxid har redan en hög genomskinlighet. Nickeloxid har en kvarvarande brun färgnyans i det bleka tillståndet, som kan reduceras med hjälp av tillsatser såsom Mg, Al, Zr, Nb

112

eller Ta. Dessa tillsatser reducerar absorptionen för de korta våglängderna i det synliga området vilket gör att materialen är lovande i framtida teknologier för god komfort och energieffektiva strukturella fönster. Den luminösa transmittansen för några av dessa material är inritad i figur 11.2 för olika belysningskällor. Den standarderiserade belysningen representerar olika ljuskällor. Belysning D65 betecknar medelvärdet av dagsljuset från den nordliga himlen vid 6500 K; A betecknar en wolfram-halogen belysning från en ljuskälla vid 2856 K (typisk ljuston i hemmet eller i varuhus); belysning F11 tillhör en kommersiell smal-bandig fluorescerande ljuskälla vid 4000 K (används vanligen i Europa och Stilla havets kustländer för kontors och varuhus belysning); och belysning F2 betecknar en komersiell bredbandig fluorescerande ljuskälla vid 4150 K (typisk kontors- eller varuhus belysning i U.S.A.).

NiO NiMgO NiAlO NiSiO NiVO NiNbO NiZrO NiTaO70

75

80

85

90

Lum

inou

s tra

nsm

ittan

ce Y

(%)

Nickel-based-oxides

Illuminant: D65 A F2 F11

Figur 10.2: Luminös synlig transmittans YCIE för en Nickel-baserad oxid, i det bleka tillståndet, vid fyra olika standard belysningar. Konstruktionen NiMO (där M är Mg, Al, Si, V, Zr, Nb, Ag eller Ta) antyder att M finns i oxiden men att mängden inte närmare angivits.

För-behandling innan laminering ____ Den katodiska elektrokroma filmen WO3 laddas upp med vätgas under beläggningsprocessen (sputtring). Fördelen är att materialet efter beläggningsprocessen redan är klart för laminering utan att någon förbehandling behövs. Innan laminering måsta den

113

anodiska elektrokroma filmen HyNiOx (eller vilken annan blandning av Nickeloxid som helst med en metalloxid som nämnts tidigare), som är belagd på ett tunt ITO underlag, urladdas genom exponering för ozon. Den anodiska och katodiska laddningen måste vara i balans för att anordningen ska blekas ordentligt. Överföring av laddning från den katodiska till den anodiska filmen uppnås med hjälp av en yttre spänning efter laminering med en lämplig elektrolyt.

Beläggning genom reaktiv magnetronsputtring genom likspänning ____

Reaktiv magnetronsputtring har visats vara en lämplig teknik för storskalig tillverkning och beläggning av film över stora ytor, vilket även underlättar övergången från laboratorieskala till industriell beläggning av stora ytor.

Reaktiv magnetronsputtring genom likspänning bygger på urladdningar i en icke reaktiv atmosfär (t.ex. Ar), där elekroner genom ett magnetfält samlas framför en target och där joner accelereras mot targeten (katoden) genom ett elektriskt fält. De positiva jonerna i plasmat accelereras mot targeten och slår sedan ut beläggningsatomer genom överföring av rörelsemängd. Materialet som avlägsnas genom dessa joner beläggs på ett underlag som ett tunt skikt.

Beläggningsprocess och egenskaper hos wolframoxid har tidigare blivit utförligt undersökta. Filmerna som tillverkats genom sputtring visar högre infärgningsförmåga jämfört med filmer som tillverkats genom andra tekniker. De sputterbelagda hydratiserade nickelbaserade oxidfilmerna visar hög kvalitet och är helt reproducerbara. I denna avhandling har optimering av beläggningsprocessen utifrån filmernas elektrokroma egenskaper uppnåtts.

Dessa anordningar öppnar flera intressanta teknologiska möjligheter att optiskt modulera transmittans, reflektans, absorptans och emittans. De kan användas för fönster i byggnader, glasögon och visir, informationsskärmar, termisk kontroll av satelliter, etc. Figur 11.3, visar ett exempel på en tillämpning inom visir området. En elektrokrom anordning har byggts in i en motorcykelhjälm. Den dynamiska förändringen av transmittansen ökar förarens säkerhet.

114

Figur 10.3. Elektrokroma anordningar med Nickelbaserad oxid i förening med WO3

i det bleka och färgade tillståndet. Mannen bakom visiret är den berömde Greger Gustavsson.

115

Acknowledgments

I would like to thank my supervisors Gunnar Niklasson (Prof.), Andris Azens (Dr.) and Claes-Göran Granqvist (Prof.) for their dedicated support during doctorate studies.

Alexei Kuzmin (Dr.) and Juris Purans (Prof.) from S.S.P.I, Riga, Latvia are acknowledged for the EXAFS measurements and modelling.

Håkan Rensmo (Dr.), Anders Sandell (Dr.) and Hans Siegbahn (Prof.) from the Electron spectroscopy and molecular surface physics department at Uppsala University are acknowledged for the XPS measurements and discussions; my gratitude is also to the technical staff at MAXLAB Synchrotron facilities in Lund, Sweden.

Sten-Eric Lindquist (Prof. em.) from the Department of Physical chemistry at Uppsala University and to Maurizio Furlani (Dr.) from Chalmer University of Technology are acknowledged for the discussions about electrochemistry.

Shuxi Zhao (Research Eng.) from the Solid State Physics Department at Uppsala University is acknowledged for the training with sputtering and tecnical support.

Greger Gustavsson (Civ. Eng.) from Cromogenics Sweden AB with acknowledged for the devices construction.

Lars Berggren (Lic.) and Pierre Roy (Lic) are acknowledged for the help with the translastion and correction of the Swedish summary in this thesis.

Arne Roos (Prof.) and Carl Ribbing (Prof.) are acknowledged for the help with optics.

Nils Olsson (Dr.) are acknowledged for the training of the X-ray diffraction equipment.

116

Yanwen Zhang (Dr.) and Goran Postneer (Prof.) are acknowledged for the RBS measurements.

Lena Henriksson, Ingrid Ringård, Caroline Olofsson and Inger Ekberg are acknowledged for all the help with the administrative matters.

Thanks also to those friends who made pleasant my stay in Sweden.

I am grateful for the scholarship received from the University of Costa Rica to complete the PhD program at Uppsala University.

The Swedish Fund for Strategic Environmental Research and the National Energy Administration of Sweden through the Ångstrom Solar Center partially supported this work.

117

References

[1] C. G. Granqvist, Handbook of Inorganic Electrochromic MaterialsElsevier, Amsterdam, 1995.

[2] C. G. Granqvist, E. Avendano, and A. Azens, Thin Solid Films, 442,201 (2003).

[3] C. A. Pan and T. P. Ma, Applied Physics Letters, 37, 163 (1980). [4] C. Agashe, B. R. Marathe, M. G. Takwale, and V. G. Bhide, Thin

Solid Films, 164, 261 (1988). [5] T. Minami, H. Nanto, and S. Takata, Thin Solid Films, 124, 43

(1984).[6] A. Hultaker and C. G. Granqvist, Thin Solid Films, 411, 1 (2002). [7] H. S. Randhawa, M. D. Matthews, and H. F. Bunshah, Thin Solid

Films, 83, 267 (1981). [8] Y. Igasaki, M. Ishikawa, and G. Shimaoka, Applied Surface Science,

33-34, 926 (1988). [9] S. K. Deb, Transactions of the Faraday Society, 65, 3187 (1969). [10] S. K. Deb, Philosophical Magazine, 27, 801 (1973). [11] C. G. Granqvist, Electrochimica Acta, 44, 3005 (1999). [12] C. Genin, A. Driouiche, B. Gerand, and M. Figlarz, Solid State

Ionics, 53-56, 315 (1992). [13] S. Sato and Y. Seino, Transactions of the Institute of Electronics and

Communication Engineers of Japan, Section E (English), E65, 509 (1982).

[14] N. Machida, M. Tatsumisago, and T. Minami, Journal of the Electrochemical Society, 133, 1963 (1986).

[15] S. Schmid, R. L. Withers, and J. G. Thompson, Journal of Solid State Chemistry, 99, 226 (1992).

[16] S. Hashimoto and H. Matsuoka, Journal of the Electrochemical Society, 138, 2403 (1991).

[17] D. J. De Smet and J. L. Ord, Journal of the Electrochemical Society,134, 1734 (1987).

[18] J. Scarminio, A. Lourenco, and A. Gorenstein, Thin Solid Films,302, 66 (1997).

[19] Y. Ye, J. Zhang, P. Gu, L. Xu, and J. Tang, Thin Solid Films, 298,197 (1997).

[20] A. Gutarra, A. Azens, B. Stjerna, and C. G. Granqvist, Applied Physics Letters, 64, 1604 (1994).

118

[21] K. Nagase, Y. Shimizu, N. Miura, and N. Yamazoe, Applied Physics Letters, 64, 1059 (1994).

[22] A. Talledo, A. M. Anderson, and C. G. Granqvist, Journal of Materials Research, 5, 1253 (1990).

[23] T. Maruyama and T. Kanagawa, Journal of the Electrochemical Society, 141, 2868 (1994).

[24] K. Yoshimura, T. Miki, P. Jin, and S. Tanemura, Reports of the National Industrial Research Institute of Nagoya, 45, 307 (1996).

[25] Z.-W. Fu and Q.-Z. Qin, Electrochemical and Solid-State Letters, 2,600 (1999).

[26] G. Beni, C. E. Rice, and J. L. Shay, Journal of the Electrochemical Society, 127, 1342 (1980).

[27] J. L. Ord, S. D. Bishop, and D. J. De Smet, Journal of the Electrochemical Society, 138, 208 (1991).

[28] P. S. Patil, L. D. Kadam, and C. D. Lokhande, Solar Energy Materials and Solar Cells, 53, 229 (1998).

[29] P. M. S. Monk and N. Farooq, Journal of Materials Science: Materials in Electronics, 6, 389 (1995).

[30] T. Maruyama and S. Arai, Journal of the Electrochemical Society,143, 1383 (1996).

[31] T. Maruyama and T. Kanagawa, Journal of the Electrochemical Society, 143, 1675 (1996).

[32] J. S. E. M. Svensson and C. G. Granqvist, Solar Energy Materials,16, 19 (1987).

[33] J. S. E. M. Svensson and C. G. Granqvist, Applied Physics Letters,49, 1566 (1986).

[34] R. Bendert and D. A. Corrigan, Journal of the Electrochemical Society, 136, 1369 (1989).

[35] D. A. Corrigan and R. M. Bendert, Journal of the Electrochemical Society, 136, 723 (1989).

[36] R. C. T. Slade, J. Barker, and T. K. Halstead, Solid State Ionics, 24,147 (1987).

[37] V. Thangadurai, H. Kaack, and W. J. F. Weppner, Journal of the American Ceramic Society, 86, 437 (2003).

[38] V. Thangadurai and W. Weppner, Ionics, 9, 308 (2003). [39] V. Thangadurai and W. Weppner, Journal of the Electrochemical

Society, 151, 1 (2004). [40] E. Avendaño, A. Azens, J. Isidorsson, R. Karmhag, G. A. Niklasson,

and C. G. Granqvist, Solid State Ionics, 165, 169 (2003). [41] A. Azens, G. Gustavsson, R. Karmhag, and C. G. Granqvist, Solid

State Ionics, 165, 1 (2003). [42] A. Azens, L. Kullman, G. Vaivars, H. Nordborg, and C. G.

Granqvist, Solid State Ionics, Diffusion & Reactions, 113-115, 449 (1998).

119

[43] L. Kullman, A. Azens, G. Vaivars, and C. G. Granqvist, SolarEnergy, 68, 517 (2000).

[44] E. Avendaño, A. Azens, G. A. Niklasson, and C. G. Granqvist, Solar Energy Materials and Solar Cells, (2004).

[45] A. P. Giri and R. Messier, Defect Properties and Processing of High-Technology Nonmetallic, 221 (1984).

[46] A. Azens, L. Kullman, and C. G. Granqvist, Solar Energy Materials and Solar Cells, 76, 147 (2003).

[47] N. Tsuda, K. Nasu, A. Fujimori, and K. Siratori, Electronic Conduction in Oxides, 2 ed. Springer, Berlin, 2000.

[48] D. R. Lide, CRC Handbook of Chemistry and Physics, 73 ed. CRC Press, Florida, 2000.

[49] P. Kofstad, Oxidation of Metals, 44, 3 (1995). [50] D. Adler and J. Feinleib, Physical Review B (Solid State), 2, 3112

(1970).[51] P. Lunkenheimer, A. Loidl, C. R. Ottermann, and K. Bange,

Physical Review B (Condensed Matter), 44, 5927 (1970). [52] R. J. Powell and W. E. Spicer, Physical Review B (Solid State), 2,

2182 (1970). [53] H. A. E. Hagelin-Weaver, J. F. Weaver, G. B. Hoflund, and G. N.

Salaita, Journal of Electron Spectroscopy and Related Phenomena,134, 139 (2004).

[54] R. S. Jayashree, P. Vishnu Kamath, and G. N. Subbanna, Journal of the Electrochemical Society, 147, 2029 (2000).

[55] D. M. MacArthur, Power Sources, 3, 91 (1971). [56] H. Bode, K. Dehmelt, and J. Witte, Electrochimica Acta, 11, 1079

(1966).[57] H. Bode, K. Dehmelt, and J. Witte, Z Anorg Allg Chem, 366, 1

(1969).[58] M. Rajamathi and P. Vishnu Kamath, Journal of Power Sources, 70,

118 (1998). [59] W. Visscher, Journal De Physique, 44, 213 (1983). [60] W. Visscher and E. Barendrecht, Journal of Electroanalytical

Chemistry and Interfacial Electrochemistry, 154, 69 (1983). [61] Y.-G. Yoon and S.-I. Pyun, Electrochimica Acta, 42, 2465 (1997). [62] M. K. Carpenter and D. A. Corrigan, Journal of the Electrochemical

Society, 136, 1022 (1989). [63] M. C. Bernard, R. Cortes, M. Keddam, H. Takenouti, P. Bernard,

and S. Senyarich, Journal of Power Sources, 63, 247 (1996). [64] M. C. Bernard, P. Bernard, M. Keddam, S. Senyarich, and H.

Takenouti, Electrochimica Acta, 41, 91 (1996). [65] J. Zaanen, G. A. Sawatzky, and J. W. Allen, Journal of Magnetism

and Magnetic Materials, 54-57, 607 (1986). [66] J. D. E. McIntyre, W. F. Peak, and G. P. Schwartz, Electronic

Materials Conference, Abs C-4 (1979).

120

[67] C. M. Lampert, T. R. Omstead, and P. C. Yu, Solar Energy Materials, 14, 161 (1985).

[68] P. C. Yu, G. Nazri, and C. M. Lampert, Solar Energy Materials, 16,1 (1986).

[69] P. C. Yu and C. M. Lampert, Solar Energy Materials, 19, 1 (1989). [70] B. B. Ezhov and O. G. Malandin, Journal of the Electrochemical

Society, 138, 885 (1991). [71] P. Delichere, S. Joiret, and A. Hugot-le Goff, Journal of the

Electrochemical Society, 135, 1856 (1988). [72] M. K. Carpenter, R. S. Conell, and D. A. Corrigan, Solar Energy

Materials, 16, 333 (1987). [73] S. Morisaki, K. Kawakami, and N. Baba, Japanese Journal of

Applied Physics, Part 1 (Regular Papers & Short Notes), 27, 314 (1988).

[74] M. C. A. Fantini, G. H. Bezerra, C. R. C. Carvalho, and A. Gorenstein, Proceedings of the SPIE - The International Society for Optical Engineering, 1536, 81 (1991).

[75] M. Fantini and A. Gorenstein, Solar Energy Materials, 16, 487 (1987).

[76] R. M. Bendert and D. A. Corrigan, Journal of the Electrochemical Society, 136, 1369 (1989).

[77] A. Surca, B. Orel, B. Pihlar, and P. Bukovec, Journal of Electroanalytical Chemistry, 408, 83 (1996).

[78] A. Surca, B. Orel, and B. Pihlar, Journal of Sol-Gel Science and Technology, 8, 743 (1997).

[79] G. Boschloo and A. Hagfeldt, Journal of Physical Chemistry B, 105,3039 (2001).

[80] P. K. Sharma, M. C. A. Fantini, and A. Gorenstein, Solid State Ionics, Diffusion & Reactions, 113-115, 457 (1998).

[81] Z. Jiao, M. Wu, Z. Qin, and H. Xu, Nanotechnology, 14, 458 (2003). [82] C. M. Lampert and R. Caron-Popowich, International Society for

Optical Engineering, 1149, 56 (1989). [83] Y. Sato, M. Ando, and K. Murai, Solid State Ionics, Diffusion &

Reactions, 113-115, 443 (1998). [84] T. Maruyama and S. Arai, Solar Energy Materials and Solar Cells,

30, 257 (1993). [85] J. Arakaki, R. Reyes, M. Horn, and W. Estrada, Solar Energy

Materials and Solar Cells, 37, 33 (1995). [86] L. D. Kadam and P. S. Patil, Solar Energy Materials and Solar

Cells, 69, 361 (2001). [87] S. A. Mahmoud, A. A. Akl, H. Kamal, and K. Abdel-Hady, Physica

B, 311, 366 (2002). [88] P. Delichere, A. H. Joiret, A. Hugot-Le Goff, K. Bange, and B.

Metz, International Society for Optical Engineering, 1016, 63 (1988).

121

[89] N. R. Lyman and H. R. Habibi, Proceedings of the SPIE - The International Society for Optical Engineering, 1016, 63 (1988).

[90] A. Agrawal, H. R. Habibi, R. K. Agrawal, J. P. Cronin, D. M. Roberts, R. S. Caron-Popowich, and C. M. Lampert, Thin Solid Films, 221, 239 (1992).

[91] A. Nemetz, A. Temmink, K. Bange, S. Cordoba de Torresi, C. Gabrielli, R. Torresi, and A. Hugot-Le Goff, Solar Energy Materials and Solar Cells, 25, 93 (1992).

[92] K. Bange, F. G. K. Baucke, and B. Metz, Proceedings of the SPIE - The International Society for Optical Engineering, 1016, 170 (1988).

[93] K. Bange, C. Ottermann, W. Wagner, and F. Rauch, InternationalSociety for Optical Engineering, 1272, 122 (1990).

[94] W. Wagner, F. Rauch, C. Ottermann, and K. Bange, Nuclear Instruments & Methods in Physics Research, Section B: Beam Interactions with Materials and Atoms, 50, 27 (1990).

[95] J. S. E. M. Svensson and C. G. Granqvist, SPIE Institutes for Advanced Optical Technologies, IS-4, 285 (1990).

[96] H. Suiyang, C. Fengbo, and Z. Jicai, Proceedings of Asian Society of Solid State Ionics, 521 (1988).

[97] S. Yamada, T. Yoshioka, M. Miyashita, K. Urabe, and M. Kitao, Journal of Applied Physics, 63, 2116 (1988).

[98] G. Bajars, Y. A. Pitkevich, A. Lusis, E. V. Pentyush, J. Benders, and V. Bets, Soviet Electrochemistry (English Translation of Elektro-Khimiia), 25, 292 (1989).

[99] A. M. Andersson, W. Estrada, C. G. Granqvist, A. Gorenstein, and F. Decker, Proceedings of the SPIE - The International Society for Optical Engineering, 1272, 96 (1990).

[100] A. Gorenstein, F. Decker, M. Fantini, and W. Estrada, SPIEInstitutes for Advanced Optical Technologies, IS-4, 272 (1990).

[101] W. Estrada, A. M. Andersson, and C. G. Granqvist, Journal of Applied Physics, 64, 3678 (1988).

[102] M. Kitao, K. Izawa, and S. Yamada, Solar Energy Materials and Solar Cells, 39, 115 (1995).

[103] M. Kitao, K. Izawa, K. Urabe, T. Komatsu, S. Kuwano, and S. Yamada, Japanese Journal of Applied Physics, Part 1 (Regular Papers & Short Notes), 33, 6656 (1994).

[104] E. L. Miller and R. E. Rocheleau, Journal of the Electrochemical Society, 144, 1995 (1997).

[105] I. C. Faria, M. Kleinke, A. Gorenstein, M. C. A. Fantini, and M. H. Tabacniks, Journal of the Electrochemical Society, 145, 235 (1998).

[106] W. Estrada, A. M. Andersson, C. G. Granqvist, A. Gorenstein, and F. Decker, Journal of Materials Research, 6, 1715 (1991).

122

[107] M. G. Hutchins, G. McMeeking, and H. Xingfang, Proceedings of the SPIE - The International Society for Optical Engineering, 1272,139 (1990).

[108] D. A. Wruck and M. Rubin, Journal of the Electrochemical Society,140, 1097 (1993).

[109] F. F. Ferreira, M. H. Tabacniks, M. C. A. Fantini, I. C. Faria, and A. Gorenstein, Solid State Ionics, Diffusion & Reactions, 86-88, 971 (1996).

[110] D. A. Wruck, M. A. Dixon, M. Rubin, and S. N. Bogy, Journal of Vacuum Science & Technology A (Vacuum, Surfaces, and Films), 9,2170 (1991).

[111] X. Y. Song, Y. X. He, C. M. Lampert, X. F. Hu, and X. F. Chen, Solar Energy Materials and Solar Cells, 63, 227 (2000).

[112] K. Yoshimura, T. Miki, and S. Tanemura, Japanese Journal of Applied Physics, Part 1 (Regular Papers & Short Notes), 34, 2440 (1995).

[113] J. S. E. M. Svensson and C. G. Granqvist, Applied Optics, 26, 1554 (1987).

[114] I. Hotovy, D. Buc, S. Hascik, and O. Nennewitz, Vacuum, 50, 41 (1998).

[115] I. Hotovy, J. Huran, J. Janik, and A. P. Kobzev, Vacuum, 51, 157 (1998).

[116] J. S. Hale, M. DeVries, B. Dworak, and J. A. Woollam, Thin Solid Films, 313-314, 205 (1998).

[117] C. Ottermann, A. Temmink, and K. Bange, Proceedings of the SPIE - The International Society for Optical Engineering, 1272, 111 (1990).

[118] J. I. Cisneros, R. M. Torresi, A. Lourenco, F. Decker, and A. Gorenstein, Journal of Physics: Condensed Matter, 5, 323 (1993).

[119] J. Scarminio, W. Estrada, A. Andersson, A. Gorenstein, and F. Decker, Journal of the Electrochemical Society, 139, 1236 (1992).

[120] Y. G. Wu, G. M. Wu, X. Y. Ni, and X. Wu, Solar Energy Materials and Solar Cells, 63, 217 (2000).

[121] F. Decker, S. Passerinl, R. Pileggi, and B. Scrosati, Electrochimica Acta, 37, 1033 (1992).

[122] S. Passerini and B. Scrosati, Solid State Ionics, Diffusion & Reactions, 53-56, 520 (1992).

[123] F. Michalak, K. von Rottkay, T. Richardson, J. Slack, and M. Rubin, Electrochimica Acta, 44, 3085 (1999).

[124] I. Bouessay, A. Rougier, B. Beaudoin, and J. B. Leriche, AppliedSurface Science, 186, 490 (2002).

[125] A. H.-L. Goff, S. Joiret, and S. Cordoba de Torresi, Proceedings of The Electrochemical Society: Symposium on Electrochromic Materials, 90, 70 (1989).

123

[126] S. I. Cordoba-Torresi, C. Gabrielli, A. Hugot-Le Goff, and R. Torresi, Journal of the Electrochemical Society, 138, 1548 (1991).

[127] S. I. Cordoba-Torresi, A. Hugot-Le Goff, and S. Joiret, Journal of the Electrochemical Society, 138, 1554 (1991).

[128] J. C. de Jesus, J. Carrazza, P. Pereira, and F. Zaera, Surface Science,397, 34 (1998).

[129] B. Chapman, Glow discharge processes: Sputtering and plasma etching John Willey and Sons, New York, 1980.

[130] D. L. Smith, Thin-film deposition McGraw-Hill, New York, 1995. [131] J. L. Vossen and W. K. (editors), Thin film processes Academic

Press, London, 1978. [132] J. R. Bird and J. S. Williams, Ion beams for materials analysis

Academic Press, London, 1989. [133] B. D. Cullity, Elements of X-ray diffractionon Addison-Wesley

Publishing, London, 1959. [134] H. H. Tian and M. Atzmon, Philosophical Magazine A: Physics of

Condensed Matter, Structure, Defects and Mechanical Properties,79, 1769 (1999).

[135] P. Monk, Fundamentals of electro-analytical chemistry John Willey and Sons, West Sussex, 2001.

[136] C. H. Hamann, A. Hamnett, and W. Vielstich, ElectrochemistryWiley-VCH, Weinheim, 1998.

[137] W. Weppner and R. A. Huggins, Journal of the Electrochemical Society, 124, 1569 (1977).

[138] P. G. Bruce, Solid State Electrochemistry Cambridge University Press, Cambridge, 1995.

[139] H. Kuzmany, Solid State Spectroscopy Springer-Verlag, Berlin, 1998.

[140] O. S. Heavens, Optical Properties of Thin Solid Films Dover Publications, Inc., New York, 1991.

[141] W. Q. Hong, Journal of Physics D: Applied Physics, 22, 1384 (1989).

[142] F. Abeles, Optical Society of America -- Journal, 47, 473 (1957). [143] F. Abeles, Progress in Optics, 2, 249 (1963). [144] H. Bach and D. Krause, Thin Films on Glass Springer-Verlag,

Berlin, 1997. [145] A. Azens, A. Talledo, A. M. Andersson, G. A. Niklasson, B. A.

Stjerna, C.-G. Granqvist, and J. R. Stevens, International Society for Optical Engineering, 1728, 103 (1992).

[146] D. L. MacAdam, Color Measurement: Theme and variationsSpringer-Verlag, Berlin, 1981.

[147] Y. J. Chabal, Surface Science Reports, 8, 211 (1988). [148] J. F. Moulder, W. F. Stickle, P. E. Sobol, and K. D. Momben,

Handbook of X-ray Photoelectron Spectroscopy Physical Electronics, Inc, Minnessota, 1995.

124

[149] J. C. Rivière and S. Myhra, Handbook of Surface and Interface Analysis Marcel Dekker, Inc., New york, 1998.

[150] A. Kuzmin, Physica B: Condensed Matter, 208-209, 175 (1995). [151] A. L. Ankudinov, B. Ravel, J. J. Rehr, and S. D. Conradson,

Physical Review B (Condensed Matter), 58, 7565 (15). [152] J. J. Rehr, J. Mustre de Leon, S. I. Zabinsky, and R. C. Albers,

Journal of the American Chemical Society, 113, 5135 (1991). [153] J. Mustre de Leon, J. J. Rehr, S. I. Zabinsky, and R. C. Albers,

Physical Review B (Condensed Matter), 44, 4146 (1). [154] J. S. E. M. Svensson and C. G. Granqvist, Solar Energy Materials,

16, 19 (1986). [155] K. Yoshimura, T. Miki, P. Jin, and S. Tanemura, Reports of the

National Industrial Research Institute of Nagoya, 44, 347 (1995). [156] A. Kuzmin, J. Purans, and A. Rodionov, Journal of Physics

Condensed Matter, 9, 6979 (1997). [157] J. Zaanen, G. A. Sawatzky, and J. W. Allen, Physical Review

Letters, 55, 418 (1985). [158] J. Wong, F. W. Lytle, R. P. Messmer, and D. H. Maylotte, Physical

Review B (Condensed Matter), 30, 5596 (1984). [159] J. Loffler and J. Weissmuller, Physical Review B (Condensed

Matter), 52, 7076 (1). [160] F. Urbach, Physical Review, 92, 1324 (1953). [161] J. Bullot, P. Cordier, O. Gallais, and M. Gauthier, Journal of Non-

Crystalline Solids, 68, 135 (1984). [162] J. Bullot, P. Cordier, O. Gallais, M. Gauthier, and J. Livage, Journal

of Non-Crystalline Solids, 68, 123 (1984). [163] K. M. E. Miedzinska, B. R. Hollebone, and J. G. Cook, Journal of

the Physics and Chemistry of Solids, 49, 1355 (1988). [164] M. Wigginton, Glass in Architecture Phaidon, London, 1996. [165] A. Azens, J. Isidorsson, R. Karmhag, and C. G. Granqvist, Thin

Solid Films, 422, 1 (2002). [166] J. R. Stevens, K. Such, N. Cho, and W. Wieczorek, Solar Energy

Materials and Solar Cells, 39, 223 (1995). [167] J. Scarminio, W. Estrada, A. Andersson, A. Gorenstein, and F.

Decker, Journal of the Electrochemical Society, 139, 1236 (1992). [168] W. Wagner, F. Rauch, and K. Bange, Nuclear Instruments &

Methods in Physics Research, Section B (Beam Interactions with Materials and Atoms), B89, 104 (1994).

[169] W. R. McKinnon and R. R. Haering, in Modern Aspects of Electrochemistry, edited by R. E. White, J. O. M. Bockris, and B. E. Conway (Plenum, London, 1983).

[170] M. D. Levi and D. Aurbach, Journal of Physical Chemistry B, 101,4641 (1997).

[171] P. Baraldi, G. Davolio, G. Fabbri, and T. Manfredini, Materials Chemistry and Physics, 21, 479 (1989).

125

[172] E. F. de Oliveira and Y. Hase, Vibrational Spectroscopy, 31, 19 (2003).

[173] I. S. Shamina, O. G. Malandin, S. M. Rakhovskaya, and S. M. Vereshchagina, Soviet Electrochemistry (English Translation of Elektro-Khimiia), 10, 1494 (1973).

[174] M. B. J. G. Freitas, Journal of Power Sources, 93, 163 (2001). [175] S. Mochizuki, Physica Status Solidi B, 126, 105 (1984). [176] K. Hermansson, Journal of Chemical Physics, 95, 3578 (1991). [177] S. Uhlenbrock, C. Scharfschwerdt, M. Neumann, G. Illing, and H.-J.

Freund, Journal of Physics: Condensed Matter, 4, 7973 (1992). [178] A. N. Mansour, Surface Science Spectra, 3, 231 (1994). [179] A. N. Mansour and C. A. Melendres, Surface Science Spectra, 3,

263 (1994). [180] A. N. Mansour, Surface Science Spectra, 3, 239 (1994). [181] A. N. Mansour and C. A. Melendres, Surface Science Spectra, 3,

255 (1994). [182] A. N. Mansour and C. A. Melendres, Surface Science Spectra, 3,

247 (1994). [183] A. N. Mansour and C. A. Melendres, Surface Science Spectra, 3,

271 (1994). [184] E. Avendaño, A. Kuzmin, J. Purans, A. Azens, G. A. Niklasson, and

C. G. Granqvist., Physica Scripta, (2004, in press). [185] C. A. Rozzi, F. Manghi, and C. Arcangeli, Physical Review B

(Condensed Matter), 62, 4774 (2000). [186] J. C. Woicik, E. J. Nelson, D. Heskett, J. Warner, L. E. Berman, B.

A. Karlin, I. A. Vartanyants, M. Z. Hasan, T. Kendelewicz, Z. X. Shen, and P. Pianetta, Physical Review B (Condensed Matter and Materials Physics), 64, 125115 (2001).

[187] E. Avendaño, A. Azens, and G. Niklasson, Proceedings of the SPIE - The International Society for Optical Engineering, 4458, 154 (2001).

[188] P. A. Cox, The Electronic Structure ans Chemistry of Solids Oxford University, Oxford, 1998.

[189] E. D. Palik, in Handbook of Optical Constants of Solids; Vol. III,edited by E. D. Palik (Academic, London, 1998).

1–11: 1970–197512. Lars Thofelt: Studies on leaf temperature recorded by direct measurement and by thermo-

graphy. 1975.13. Monica Henricsson: Nutritional studies on Chara globularis Thuill., Chara zeylanica Willd.,

and Chara haitensis Turpin. 1976.14. Göran Kloow: Studies on Regenerated Cellulose by the Fluorescence Depolarization Tech-

nique. 1976.15. Carl-Magnus Backman: A High Pressure Study of the Photolytic Decomposition of Azo-

ethane and Propionyl Peroxide. 1976.16. Lennart Källströmer: The significance of biotin and certain monosaccharides for the growth

of Aspergillus niger on rhamnose medium at elevated temperature. 1977.17. Staffan Renlund: Identification of Oxytocin and Vasopressin in the Bovine Adenohypophysis.

1978.18. Bengt Finnström: Effects of pH, Ionic Strength and Light Intensity on the Flash Photolysis of

L-tryptophan. 1978.19. Thomas C. Amu: Diffusion in Dilute Solutions: An Experimental Study with Special Refer-

ence to the Effect of Size and Shape of Solute and Solvent Molecules. 1978.20. Lars Tegnér: A Flash Photolysis Study of the Thermal Cis-Trans Isomerization of Some

Aromatic Schiff Bases in Solution. 1979.21. Stig Tormod: A High-Speed Stopped Flow Laser Light Scattering Apparatus and its Appli-

cation in a Study of Conformational Changes in Bovine Serum Albumin. 1985.22. Björn Varnestig: Coulomb Excitation of Rotational Nuclei. 1987.23. Frans Lettenström: A study of nuclear effects in deep inelastic muon scattering. 1988.24. Göran Ericsson: Production of Heavy Hypernuclei in Antiproton Annihilation. Study of their

decay in the fission channel. 1988.25. Fang Peng: The Geopotential: Modelling Techniques and Physical Implications with Case

Studies in the South and East China Sea and Fennoscandia. 1989.26. Md. Anowar Hossain: Seismic Refraction Studies in the Baltic Shield along the Fennolora

Profile. 1989.27. Lars Erik Svensson: Coulomb Excitation of Vibrational Nuclei. 1989.28. Bengt Carlsson: Digital differentiating filters and model based fault detection. 1989.29. Alexander Edgar Kavka: Coulomb Excitation. Analytical Methods and Experimental Results

on even Selenium Nuclei. 1989.30. Christopher Juhlin: Seismic Attenuation, Shear Wave Anisotropy and Some Aspects of

Fracturing in the Crystalline Rock of the Siljan Ring Area, Central Sweden. 1990.31. Torbjörn Wigren: Recursive Identification Based on the Nonlinear Wiener Model. 1990.32. Kjell Janson: Experimental investigations of the proton and deuteron structure functions.

1991.33. Suzanne W. Harris: Positive Muons in Crystalline and Amorphous Solids. 1991.34. Jan Blomgren: Experimental Studies of Giant Resonances in Medium-Weight Spherical

Nuclei. 1991.35. Jonas Lindgren: Waveform Inversion of Seismic Reflection Data through Local Optimisation

Methods. 1992.36. Liqi Fang: Dynamic Light Scattering from Polymer Gels and Semidilute Solutions. 1992.37. Raymond Munier: Segmentation, Fragmentation and Jostling of the Baltic Shield with Time.

1993.

Acta Universitatis UpsaliensisUppsala Dissertations from the Faculty of Science

Editor: The Dean of the Faculty of Science