Electrochromic properties of nickel oxide in different electrolytes

(1)

UPTEC Q13 004

Examensarbete 30 hp Juni 2013

Electrochromic properties of

nickel oxide in different electrolytes

Anders Stenman

(2)

Teknisk- naturvetenskaplig fakultet UTH-enheten

Besöksadress:

Ångströmlaboratoriet Lägerhyddsvägen 1 Hus 4, Plan 0

Postadress:

Box 536 751 21 Uppsala

Telefon:

018 – 471 30 03

Telefax:

018 – 471 30 00

Hemsida:

http://www.teknat.uu.se/student

Abstract

Electrochromic properties of nickel oxide in different electrolytes

Anders Stenman

A half cell of an electrochromic (EC) device has been used to determine the

electrochromic response of a nickel oxide film in nine different electrolytes. Six of the nine electrolytes were 0.1 M non-aqueous salts dissolved in equal weight % of

propylene carbonate and ethylene carbonate. Three of them were lithium-based and three of them tetrabutylammonium (TBA)-based. The last three electrolytes were proton-based aqueous solutions of 1 M KOH, 0.1 M propionic acid and 0.1 M phosphoric acid, respectively.

The electrolytes were subjected to electrochemical measurements of cyclic voltammetry and square wave voltammetry, both with simultaneous in-situ optical transmittance measurements in the visible region. Ex-situ optical measurements were performed in the UV-VIS-NIR (300-2500 nm) range and IR-spectroscopy

measurements in the 600 – 4000 cm-¹ range.

To determine the performance of the nickel oxide films, the coloration efficiency (CE) is used as a figure of merit. The desired value is to achieve a high optical modulation with as little amount of charge inserted/extracted as possible.

The results show that neither lithium nor TBA has a significant impact on the electrochromic (EC) response, compared with the protonic electrolytes. An

argument can be made that the intercalation of neither cation (lithium or TBA) is the reason behind the electrochromic behaviour of the nickel oxide. In KOH it is rather the OH- that transfer to the surface and attracts protons (H⁺) from the bulk nickel oxide that enhances the EC response. In both propionic and phosphoric acid, it is the reversible intercalation of protons (H⁺) into the porous nickel oxide that gives the electrochromic response.

(3)

Elektrokromiska egenskaper hos nickeloxid i olika elektrolyter

Anders Stenman

Idag är användandet av vår planets energiresurser ständigt under luppen.

Genom att utnyttja förnybara energikällor som exempelvis sol, vind och vågkraft kan utsläppen av växthusgas från ej hållbara energikällor min- imeras. Genom att samtidigt försöka minimera vår egen energianvänd- ning kan dagens energiproblem få en ljusare framtid.

Mycket energi går idag åt vid nedkylning av lokaler, inte minst på grund av fönster som släpper in för mycket värme och ljus. Att använda så kallade smarta fönster är således ett sätt att spara energi, samtidigt som komforten kan höjas. Smarta fönster går ut på att man, efter egen vilja, kan reglera transmissionen, solenergin, vare sig det gäller värme eller komfort. De smarta fönstren kan ändra sin transmission genom en ändring i temperatur (termokroma), genom belysning av fotoner (fo- tokroma) eller genom en applicerad spänning (elektrokroma). I den här rapporten är det de elektrokroma (EC) egenskaperna som blivit under- söka.

Ett elektrokromt smart fönster fungerar med samma princip som ett batteri då man ändrar transmissionen av ljus genom att ladda upp eller ur de olika materialen. Då det ena material blir uppladdat, färgas det, samtidigt som det andra blir färgat vid urladdning och vice versa. Genom att ha en krets där man kan upprepa denna upp/urladdningscykel fler- talet gånger kan man efter egen vilja ställa in fönstrets genomsläpplighet vid önskad tidpunkt på dygnet.

I ett smart fönster använder man sig av en elektrokrom enhet som i det här fallet är producerad av ChromoGenics. I Figur 1 finns en schematisk bild av hur denna enhet är uppbyggd. Genom att använda sig av polymer(plast)substrat så kan man producera materialet på rulle vilket innebär lättare transport och förvaring.

Först är ett transparent elektriskt ledande material, i det här fallet indiumtennoxid (ITO), belagt på plasten. På de är i sin tur det an- odiska respektive katodiska elektrokroma materialen belagda. Det ena elektrokroma materialet är nickeloxid (anod) och det andra volframoxid (katod).

Det är de elektroma egenskaperna hos nickeloxiden som har under- sökts i denna rapport. Detta har gjorts genom att endast använda ena sidan av den elektrokroma enheten, en så kallad halvcell, bestående av nickeloxid, nedsänkt i en flytande elektrolyt. Genom att använda nio

(4)

olika flytande elektrolyter med olika anjoner (X⁻) och katjoner (X⁺) i elektrokemiska mätningar så har analyser gjorts av hur dessa påverkar de elektrokroma egenskaperna hos nickeloxiden. Sex av elektrolyterna var icke-vattenbaserade, varav tre baserade på litiumsalter och tre på tetrabutylammonium(TBA)-salter. De sista tre elektrolyterna var vattenbaserade och bestod av kaliumhydroxid (KOH), propansyra och fos- forsyra.

Figure 1: En bild av hur den elektrokroma enheten från ChromoGenics är uppbyggd:

Substrat(plast)/ITO/EC(anod)/Elektrolyt/EC(katod)/ITO/Substrat(plast) Resultaten visade att den elektrokroma responsen såg liknande ut, oberoende av vilken anion (ClO4/TF/ TFSI) eller katjon (Li/TBA) som användes i de sex icke-vattenbaserade elektrolyterna. Dock fanns det skillnader mellan de vattenbaserade elektrolyterna, vilket gör att en slut- sats gällande huruvida Li⁺ eller TBA⁺ påverkar strukturen hos nickeloxiden kan dras. Att man fick en större respons från de vattenbaserade tyder på att OH⁻ möter upp protoner (H⁺) och bildar vatten på ytan av nickeloxiden för att sedan resorbera till elektrolyten.

Examensarbete 30 hp på civilingenjörsprogrammet Teknsik fysik med materialvetenskap

(5)

List of Abbreviations

CE Coloration Efficiency

CG Chromogenics

CV Cyclic Voltammetry

EC Electrochromic

ITO Indium Tin Oxide, In2O₃:Sn

PC/EC Propylene Carbonate/Ethylene Carbonate

Phos Phosphoric acid

Prop Propionic acid

SWP Square Wave Potential

SWV Square Wave Voltammetry

TBA Tetrabutylammonium, C16H₃₆N Salts in PC/EC

LiPC Lithium perchlorate, LiClO4

LiTF Lithium triflate, LiCF3SO₃

LiTFSI Lithium bis(trifluoromethane)sulfonimide, LiN(CF₃SO₂)₂

T-PC TBA perchlorate, TBA-ClO4

T-TF TBA triflate, TBA-CF3SO3

T-TFSI TBA bis(trifluoromethane)sulfonimide, TBA-N(CF₃SO₂)₂

Electrolytes:

KOH Potassium hydroxide, KOH

Prop Propionic acid, C2H5COOH Phos Phosphoric acid, H3PO₄

(8)

List of Figures

1 Smart fönster producerat av ChromoGenics . . . ii

2 Smart window produced by ChromoGenics . . . 3

3 The half cell used in experiments . . . 4

4 Ex-situ CV measurements . . . 9

5 Theoretical picture of SWV . . . 10

6 Equipment used for the electrochemistry . . . 11

7 CV and T in lithium electrolytes . . . 13

8 CV and T in TBA electrolytes . . . 14

9 CV and T in Prop and Phos . . . 16

10 CV and T in KOH . . . 17

11 Coloration efficiency - all electrolytes . . . 18

12 1000 cycle CV run in LiTF-PC/EC . . . 20

13 1000 cycle run in TBA-TF . . . 21

14 SWP of the 10th cycle . . . 22

15 Average switch time for all electrolytes . . . 23

16 In-situ optical in Li-based electrolytes . . . 24

17 In-situ optical in TBA-based electrolytes . . . 25

18 In-situ optical in protonic electrolytes . . . 26 19 Infrared measurements of Li-electrolytes, 600-4000 cm⁻¹ 27 20 Infrared measurements of TBA-electrolytes, 600-4000 cm⁻¹ 28 21 Infrared measurements of TBA-electrolytes, 600-4000 cm⁻¹ 29

(9)

1 Introduction

1.1 Background

1.1.1 The situation today

Today, new solutions concerning energy applications are constantly on the research and development stage. Whether it’s about trying to use energy saving methods or harness the weather, i.e. sun, wind or wave power, or to enhance the efficiency of already existing powerplants.

With these previously mentioned methods, the idea is to take advantage of all the excessive energy that is around us. Another approach, instead of trying to use the excessive energy, is to try and minimize the energy usage as far as possible.

When it comes to energy consumption in form of heating or cooling of houses or office spaces, up to 17 % [1] of the energy can be saved by either using different coatings or smart materials in the windows [2].

In a coated window, either a coating for low emissivity or solar control is used for cold and warm climates respectively. Therefore, a window that can change its optical properties at a given temperature, lighting level or voltage will have a great advantage on today’s energy market. These windows are called smart windows, i.e. their ability to change the optical properties is depending on an outer stimulus of some form [3].

First and foremost, there are three different types of smart materials used in windows: Thermochromic, where the optical properties can be changed by a change in temperature; Photochromic, where the optical properties can be changed by incidenct ultraviolett photons [4] and last;

Electrochromic, where a change in the applied voltage changes the optical properties. In this paper, the latter of these mentioned materials, the electrochromic (EC) ones, will be discussed.

1.2 Electrochromism

The term electrochromism was first used in 1961 [6] to describe the EC properties of tungsten oxide films. Since then, research on tungsten oxide as an EC material has been well understood [7]. In an electrochromic device (ECD), an ion storage layer or an anodic EC material can be used as a counterpart to tungsten oxide. By using another EC material, e.g.

nickel oxide, instead of an ion storage layer, a greater optical modulation can be achieved.

(10)

Therefore, a deeper understanding of nickel oxide will be of importance to be able to enhance the performance and stability of the window.

However, the reason(s) behind the EC properties of nickel oxide is still under research, and even more so in this thesis.

1.2.1 Electrochromic smart windows

An EC smart window can be constructed in such a way that one or several coatings are deposited on the substrate. Depending on the properties of the materials, the smart material can function in different ways.

To enhance the performance of the EC material it can be matched with an appropriate low-e or solar control coating, coated on the outside, for the specific demands. These coatings can be so called transparent conductors (TC) and can also be utilized in other areas of application such as displays in telephones, cameras and televisions [3].

The EC smart window works like a thin film battery, with the material(s) being colored or bleached upon charging/discharging. It does not need a continous voltage applied since the bleached/colored state is stable without any degradation.

1.3 Configuration of an electrochromic device

1.3.1 Full device

The smart window constructed by ChromoGenics uses polymer sub- strates instead of the more conventional ones, made of glass. The use of a polymer substrate enables roll-to-roll production, which decreases both the production and transportation costs as well as providing flexibility for applications such as motorcycle visors [1].

A schematic picture of a full electrochromic device is shown in Figure 2 and the configuration of the different layers can be seen.

First, the substrate is coated with a transparent conductive (TC) material, such as indium tin oxide (ITO). This is necessary to be able to apply a voltage while still maintaining the optical transmission in the visible range.

The ITO layers are coated with EC materials. In a full device, an anodic and a cathodic EC material is deposited on each substrate, respectively. While the cathodic side colors when ions are intercalated, the anodic side colors when ions, in turn, are deintercalated. The reversibility of this effect in the ECD is essential for the smart window to work.

(11)

Figure 2: A schematic picture of how a full ECD is constructed:

Substrate/TC/EC(anodic)/Electrolyte/EC(cathodic)/TC/Substrate.

To make the full device function properly with ions joining the electrons, an ion-conductive electrolyte is needed. Since the electrons transfer via the ITO and the ions in the opposite direction, the ion-conductive electrolyte needs a low electrical conductivity.

An electrolyte consists of a cation (X⁺) and an anion (X⁻). Typical cations are hydrogen and lithium [5], due to their relative small sizes that enhances the ion mobility.

The electrolyte can be a liquid, a gel [5], or in ChromoGenics case, a solid. This polymeric solid electrolyte also needs the ability to glue the two smart materials together to enable the roll-to-roll manufacturing method.

1.3.2 Half cell

For experimental reasons, a half cell can be used instead of the full device when researching one specific EC material. If the analysis should include a full device, an argument can be made about the lack of certainty of what properties pertains to which material. Hence, a half cell is made up of only one part of the ECD. In this thesis, it is the anodic part, i.e.

nickel oxide. Therefore, the substrate only needs to be coated with the ITO and the EC material itself.

Instead of using a solid electrolyte, like in the full device, a liquid electrolyte can be used with the half cell. To be able to utilize the half cell, a reference electrode and a counter electrode need to be submerged in the electrolyte as well. In Figure 3, a schematic picture of the half

(12)

device with nickel oxide as EC material on top and with surrounding liquid electrolyte is shown.

Figure 3: A schematic picture of how a half device is constructed:

Substrate/TC/Nickel oxide/Liquid electrolyte.

1.4 Aim of thesis

The aim of this thesis is to get a better look into what makes the nickel oxide display its electrochromic behaviour. Depending on which elec- tolyte is being used, a hypothesis can be made whether it is the anions or cations that intercalate into or adsorb onto the nickel oxide that is the reason for this behaviour. An electrolyte that has previously been used to check this electrochromic behaviour is lithium perchlorate [8].

An idea suggested by Passerini et al. [10] is that the Li⁺ ions intercalate into the porous structure of the nickel oxide. This could be one of the reasons for the smart window to switch optical states. However, the anion might also get adsorbed onto the nickel oxide and hence affect the optical properties. Although, one does not have to exclude the first mechanism since a superposition of these two mechanisms could be occuring as well.

To analyze said electrolytes, electrochemical measurements with in- situ optical transmittance (T) measurements will be made. After doing these, both IR and optical spectroscopy analyzis will be made of the samples in their colored and bleached states, respectively.

(13)

2 Theory

2.1 Electrochromism in nickel oxide

The switching of the bleached and colored states in the nickel oxide occurs due to intercalation and deintercalation, respectively. There have been lots of studies performed on nickel oxide in aqueuos solutions such as KOH [8]. While the intercalation mechanism has been understood for those solutions, studies in non-aqueous solutions such as LiClO4 in propylene carbonate (PC), they have not [10].

In aqueous solutions, a change between NiOOH and Ni(OH)2 occurs.

These forms correspond to the material being colored and bleached respectively. This can be seen in reaction (1), with the formation of water on the surface as OH⁻ attracts protons from the bulk to neutralize the negative charge from the anions.

Ni(OH)₂ + OH⁻ ←→NiOOH + H2O + e⁻ (1) However, there are different opinions about the mechanisms behind the change in transmittance when using non-aqueous electrolytes.

Passerini et al. [10] suggest that when dealing with Li⁺, the first lithium ions inserted in the nickel oxide cause an irreversible strain in the structure. This strain gives rise to a change of volume and therefore large enough sites in the porous structure where, later on, lithium ions can reversibly intercalate/deintercalate.

This two-step mechanism is described by reactions2(change of structure) and 3 (reversible intercalation/deintercalation).

NiOx+ yLi⁺+ ye⁻−→ LiyNiOx (2)

LiyNiOx

coloring

−−−−−*

)−−−−−

bleaching Liy−zNiOx+ zLi⁺+ ze⁻. (3)

2.2 Optical properties

When incident photons interact with a material they can either be absorbed (A), reflected (R) or transmitted (T). The summation of the three mechanisms always sum up to unity as seen in (4),

T + R + A = 1. (4)

The reflectance occurs at the interface of two materials, with the transmittance being the light that propagates through it. The absorbed light is therefore easily calculated by obtaining R and T in optical measurements.

(14)

In this thesis, it is the transmitted light that is analyzed in nickel oxide thin films. Since the half cell of the electrochromic device is built up of several thin film layers, each layer has its specific absorption. By using the optical density (OD) of all layers instead of the absorption coefficient α in each layer, it becomes easier to evaluate the performance of the EC material. With the OD known from (5),

OD_film = ODtotal− OD_ITO+glass = αfilmd_film, (5) the electrochromic performance of the half cell can be obtained.

2.3 Coloration Efficiency

The coloration efficiency is a figure of merit for how well a smart window can achieve optical modulation [8]. By using the transmittance data of the colored and bleached states, respectively, and the difference in the charge density between these states, a coloration efficiency (CE) can be defined. The higher the absolute value of CE, the better the electrochromic response is.

To calculate the CE of a material, the ratio between the change in optical density and the amount of charge inserted is needed. The final expression for the CE, if neglecting the bleched and colored states change of reflectance is seen in eq. (6),

CE = ln(^{T bleached}_{T colored} )

∆Q . (6)

Here, Tbleached is the transmittance in the bleached state and Tcolored

therefore corresponds to the transmittance in the colored state. The ∆Q correspons to the mean value of the charge density that has been inserted and extracted.

(15)

3 Experiments

3.1 Sample preparation

3.1.1 Nickel oxide films

The half cell samples used for making all the experiments was from Chro- moGenics. They were of a polymer substrate coated with first ITO and then nickel oxide. The nickel oxide films were prepared by sputtering with a thickness of 450 nm.

Each sample had a rectangular shape of 6*1.5 cm² with about 3 cm of the sample submerged in the electrolyte. How much the samples were submerged was measured after each measurement. This was possible by noting the difference in coloring of the sample between the submerged and dry parts.

To be able to do the ex-situ optical measurements, wider pieces of the samples hade to be cut, since the holder was 2.5 cm wide. This ment that those samples were 5*3 cm² in size.

3.1.2 Electrolytes

A total of six, three lithium-based and three TBA-based, electrolytes had to be synthesized in a glove box due to the humidity in the atmosphere.

Inside the glove box, argon was utilized as the inert gas with the pressure inside not exceeding 0.1 mBar and a water concentration of less than 1 ppm¹.

The LiClO4, LiTF and TBA salts as well as the PC/EC were obtained from Sigma-Aldrich with a purity above 96 %. The LiTFSI-salt is commercially available, e.g. from Sigma-Aldrich.

First of all, a non-aqueous solvent of equal weight % of propylene carbonate and ethylene carbonate was made. The ethylene carbonate is a transparent solid which need to be hacked in order to get it in a granular form. After this, the ethylene carbonate was poured into the liquid propylene carbonate in order to dissolve using a magnetic stirrer at 25-50°C.

With the solvent prepared, it was poured in the bottle containing the lithium and TBA-salts respectively and in turn be dissolved in room temperature using a magnetic stirrer. To be able to compare the lithium and TBA with each other, the same concentration (0.1 M) was used.

The protonic based electrolytes, KOH, propionic acid (prop) and phosphoric acid (phos) were 1 M, 0.1 M and 0.1 M respectively. They

1Due to some problem with the glove box, these values changed on occations but nothing that seemed damaging to the electrolytes nor electrodes.

(16)

all were dissolved in deionized water with means of a magnetic stirrer at room temperature inside a fume hood. The compounds had a purity of 85, 99 and 84% for KOH, prop and phos respectively and were obtained from Merck, Sigma Aldrich and KeboLab respectively.

The used electrolytes are seen below with different anions and cations.

The lithium-based salts dissolved in PC/EC being;

LiClO4 (Lithium perchlorate) LiTF (Lithium triflate)

LiTFSI (Lithium bis(trifluoromethane)sulfonimide)

The same anions are also used in TBA (T etrabutylammonium)-based salts dissolved in PC/EC to be able to deduce what ions that affects the EC performance;

TBA-ClO4 (T BA perchlorate) TBA-TF (T BA triflate)

TBA-TFSI (T BA bis(trifluoromethane)sulfonimide) Three hydrogen based electrolytes were analyzed as well;

KOH (P otassium hydroxide) H3PO4 (P hosphouric acid) C3H6O2 (P ropionic acid)

3.2 Electrochemistry

3.2.1 Cyclic voltammetry

To be able to do a cyclic voltammetry (CV) measurement, a Solartron Potentiostat was used, with the exception of the two 1000-cycle runs which were done using a Bio-Logic/EC-lab setup.

A three electrode setup with a working, reference and counter electrode, respectively, submerged in the electrolyte as seen in Figure 4 was used. A voltage could then be applied between the working and reference electrode with the current being measured between the working and counter electrode. The sample of a nickel oxide coated polymer served as the working electrode. In the experiments with the lithium electrolytes, both the reference and the counter electrode were made of lithium. In all the other six electrolytes, the reference electrode was of Ag/AgCl and the counter electrode was of platinum.

By applying a potential between the working and reference electrode at a rate of 10 mV/s a cyclic voltammogram could be obtained. This was continously cycled for at least 10 cycles for reassuring reversibility.

In each electrolyte, the reproducibility was checked by doing three CV experiments with nickel films from the same batch.

(17)

Figure 4: The setup outside the glove box for CV and transmittance mea- surements. The three electrode setup is seen with the reference being to the right and the sample to the left as working electrode.

The voltage range used when cycling was between 2.5 and 4 V vs Li for the lithium electrodes. An interval of the same magnitude but different end values was used when using the Ag/AgCl electrode. This interval was -0.545 to 0.955 V vs Ag/AgCl for all other electrolytes except KOH . After first doing a CV run in KOH with the range of -0.545 to 0.955 V vs Ag/AgCl, a decision was made to use an interval of -0.5 to 0.8 V vs Ag/AgCl. This interval was chosen due to the fact that the electrochromic response outside this interval showed no relevant data. In all the measurements that were done, the starting value was the low end value in the interval.

The potential values used for the lithium and Ag/AgCl electrode, respectively, were relative to the standard reduction (of hydrogen) potential. This potential is a reference used for electrodes in different electrolytes. The standard reduction potential is -3.045 and 0.2223 for lithium and Ag/AgCl respectively [11].

To check the stability of the electrolytes, LiTF-PC/EC and TBA-TF- PC/EC were chosen to undergo a 1000 cycle CV run. They were chosen because they had the shortest switching times of all electrolytes, which were obtained in the square wave voltammetry measurements.

(18)

3.2.2 Square wave voltammetry

When switching the window from colored to bleached state and vice versa, it takes a certain time for the ions to intercalate into the structure.

This time varies for different electrolytes and is called the switching time the window. The switching time is of importance since it can enhance the performance of the window as well as it is of utmost importance in applications such as visors for motorcycle helmets.

Instead of applying the voltage at a rate of 10mV/s as in CV measurements, the voltage is changed instantly between the two end values of the interval as seen in Figure 5. The apparatus used is the same potentiostat as in CV and the same procedure with 10 cycles on each electrolyte for stability and three times for the reproducability was applied.

Figure 5: The data obtained from SWV with optical measurements. a) The square wave voltage applied. b)The corresponding current density and hence the response time. c) The transmittance vs time, makes it possible to obtain the switching time of the window.

(19)

3.3 Optical measurements

3.3.1 In-situ optical measurements

To be able to check the transmittance of the nickel oxide films, simultaneous optical measurements were performed with CV and SWV measurements. In this optical setup, which is seen in Figure 6, the transmittance at wavelengths of 400, 475, 550, 625, 700, 775 nm respectively were recorded. Figures in this study shows transmittance at 550 nm, since that is the wavelength that is most sensitive to the eye.

Figure 6: The equipment setup of the electrochemical measurements. The glove box to the right, connectod to a potentiostat (white screen used) and an optical fibre (left screen used).

To be able to obtain the transmittance of the nickel film oxide itself, a reference hade to be set first. This was done by observing the transmittance through the glass box with the electrolyte inside and seting it to 100 %.

3.3.2 Ex-situ optical measurements

Using the Lambda 900 apparatus, optical spectroscopy measurements were made at wavelengths between 300 and 2500 nm. The corresponding transmittance of the sample was measured in these wavelengths after two cycles of CV in all electrolytes, in bleached and colored states respectively.

(20)

3.4 IR-spectroscopy

While the mechanisms for coloring and bleaching tungsten oxide are understood, they are more questioned when it comes to nickel oxide.

Whether it is the cation (Li/T BA) or the anion, e.g. perchlorate or triflate that gives the nickel oxide its EC properties is still not fully understood.

By doing infrared spectroscopy of the samples after coloring and bleaching, respectively, an answer to what ions might adsorb on the surface of the NiOx can be given. All the electrolytes were first run in CV for 2 cycles, both for coloring and bleaching ² and then transferred to the IR-spectrometer. Since the coloration/bleaching of the sample might degrade when transferred in air, it was important that both the optical and IR measurements were made directly after the CV.

There are different kind of IR-spectroscopy methods. In this thesis, the fourier transorm (FTIR) method was used.

In these infrared spectroscopy measurements, a mercury cadmium tel- lurium (MCT) detector was used. By using this detector, waves ranging from 600 to 4000 cm⁻¹ could be analyzed.

As a reference, an as-deposited nickel oxide film was used. The absorbance could then be measured as a positive or negative value. A positive value of the absorbance, A, implicates that the colored or bleached state has a lower transmittance, T, hence vice versa with a negative value of A.

2Depending on if the sample was to be colored or bleached, 2.5 or 2 cycles was

(21)

4 Results and discussion

4.1 Cyclic voltammetry with in-situ optical mea- surement

4.1.1 Lithium-based electrolytes

In Figure 7, CV and T measurements for the nickel oxide thin film in the lithium-based electrolytes are shown together with in-situ optical measurements. The three different anions used are ClO4, TF and TFSI.

The measurements are all taken from the 10^thcycle with a voltage interval between 2.5 and 4.0 V vs Li.

Figure 7: CV and T measurement of the 10^th cycle of nickel oxide films in three different 0.1 M lithium-based electrolytes. The solid curve is LiClO₄- PC/EC, the dashed curve is LiTF-PC/ECand the dotted curve is LiTFSI- PC/EC. CV measurements are shown in black on the axis to the left hand side, with the T measurements shown in red on the axis to the right hand side.

The change of the current density (∆j) between the end values in the three electrolytes was between 0.022 and 0.025 mA/cm². The highest ∆j was found in the LiTF-PC/EC with the lowest being LiTFSI-PC/EC.

The transmittance in the bleached states were 72.5, 73.5 and 72.2 %

(22)

while in the colored state they were 66.1, 65.8 and 65.3 % for LiClO4- PC/EC, LiTF-PC/EC and LiTFSI-PC/EC respectively. The highest change of transmittance was in the LiTF-PC/EC electrolyte, with a ∆T of 7.7 %. The lowest ∆T was in the LiClO4-PC/EC, of 6.4 %.

There are no big changes, neither in the ∆j, nor in the transmittance of the nickel film in the three different electrolytes used. Hence, an argument can be made that the different anions used do not affect the electrochromic response of the nickel oxide in any specific way.

4.1.2 TBA-based electrolytes

The CV and T measurements in TBA-based electrolytes of similar nickel oxide films are shown in Figure8. The voltage interval was between 0.545 and 0.955 V vs Ag/AgCl. The measurements are from the 10^th cycle.

Figure 8: CV and T measurement of the 10^th cycle of nickel oxide films in three different 0.1 M TBA-based electrolytes. The solid curve is TBA- ClO₄-PC/EC, the dashed curve is TBA-TF-PC/EC and the dotted curve is TBA-TFSI-PC/EC. CV measurements are shown in black on the axis to the left hand side, with the T measurements shown in red on the axis to the right hand side.

(23)

mA/cm². The highest range of the current density was in TBA-TF- PC/EC and the lowest in TBA-TFSI-PC/EC.

The transmittance in the bleached state were 70.4, 67.8 and 67.1 % while in the colored state they were 64.2, 62.5 and 63.2 % for TBA-ClO4- PC/EC, TBA-TF-PC/EC and TBA-TFSI-PC/EC respectively. Hence, the highest ∆T, of 6.2 % was in TBA-ClO4-PC/EC and lowest of 3.9 % was in TBA-TFSI-PC/EC.

The same argument that was made for the nickel film in the lihium- based electrolytes could be made here as well. There are no relevant changes in the reactions when each anion is used with the TBA-salts which confirms this.

There is a similarity in the CV measurements of the nickel oxide film between the lithium- and TBA-based electrolytes. One can not therefore conclude that the lithium has indeed intercalated the structure more so than TBA.

(24)

4.1.3 Protonic electrolytes

The measurements in KOH, propionic acid (Prop) and phosphoric acid (Phos) electrolytes are shown seperately in Figure 9and 10.

The Prop and Phos are both shown in Figure 9. The same electrodes and hence voltage interval of 0.545 to 0.955 V vs Ag/AgCl were used and the values were obtained from the 10^th cycle

Figure 9: CV and T measurement of the 10^th cycle of nickel oxide films in 0.1 M Prop and Phos. The solid curve is Prop and the dotted is Phos. CV measurements are shown in black with the axis to the left hand side. The transmittance is shown in red with the axis to the right hand side.

∆j in the end values were 0.046 and 0.053 mA/cm², while ∆j for the peak values were 0.056 and 0.085 mA/cm² for Prop and Phos, respectively.

The transmittance in the bleached states were 80.0 and 82.1 % and for the colored state 64.2 and 65.7 % for Prop and Phos respectively.

Hence, the ∆T was bigger for Phos (16.4 %) than Prop (15.8 %).

As previously noted, the CV measurements of the nickel film in the lithium and TBA electrolytes did not have a significant difference between each other. This was not the case for the propionic (Prop) and

(25)

sity (j), a peak in Prop and even more distinct in Phos, at 0.7 V is seen.

The same can be said about the minimum value of j which occurs at 0.5 V and again is more distinct in Phos. An explanation for this could be that the nickel oxide tend to react with Phos more than Prop. This could be due to the fact that both Prop, and even more, Phos attracts protons (H⁺) from the nickel oxide and forms water on the surface which then resorbs to the electrolyte and the reaction can continue.

The CV and T measurements of KOH are shown in Figure 10. The applied voltage interval was between -0.5 and 0.8 V vs Ag/AgCl and the data was obtained from the 10^th cycle.

Figure 10: CV and T measurement of the 10^th cycle of nickel oxide films in 1 M KOH. CV measurement are shown in the black dashed curve with the axis to the left hand side and the transmittance shown in the red solid curve with the axis on the right hand side.

The range of the current density for KOH was 0.34 mA/cm² with a peak range of 0.48 mA/cm². The transmittance in the bleached state was 67.7 % and in the colored state 13.3 % with a ∆T of 54.5 %.

As previously stated, a well understood mechinsm occurs with nickel oxide films in KOH as seen in eq. (1). This reaction is more efficient than what happens in the other protonic electrolytes due the fact that KOH is the only electrolyte that contains OH⁻.

(26)

4.2 Coloration Efficiency

The coloration efficiency (CE) of a nickel oxide film in all electrolytes is shown in Figure 11. The CE is obtained for the wavelengths in the visible region with ex-situ optical measurements. The highest CE obtained at 550 nm was ∼120 cm²/C, which pertained to the nickel oxide film in an electrolyte of TBA-TFSI-PC/EC. The lowest CE, ∼20 cm²/C, is seen when a nickel oxide film is cycled in KOH due to the lower optical modulation obtained in ex-situ optical measurements.

Figure 11: The coloration efficiency of a nickel oxide film in all different electrolytes in the visible region. The values are obtained from ex-situ optical measurements.

The values obtained in this study corroborates the work done by Sara Green et al., where nickel oxide films were cycled in LiClO4-PC/EC, KOH and Prop [9]. The data obtained and used in Figure 11are from the ex- situ optical measurements of the nickel oxide film in all electrolytes.

There is no great difference in the optical modulation for the films between the ex-situ and in-situ measurements, except when cycled in KOH, where ∆T was 17 % ex-situ and 54.5 % in-situ. This is a significant change of ∆T, with the corresponding value of the CE at 550 nm therefor

(27)

oxide film in all electrolytes is shown at 550 nm with in-situ optical measurements.

Table 1: Table of the charge capacity and the coloration efficiency obtained with in-situ optical measurements of the transmittance at 550 nm. The values differ somewhat compare to the ex-situ measurements, especially when cycled in KOH.

Electrolytes CE (cm²/C) ∆Q (mC/cm²)

LiClO4 - PC/EC 78 1.19

LiTF - PC/EC 83.8 1.32

LiTFSI - PC/EC 82.7 1.22

TBAClO4 - PC/EC 97.6 0.95

TBATF - PC/EC 79.8 1.02

TBATFSI - PC/EC 73.9 0.81

KOH 101.5 16

Prop 77.8 2.87

Phos 67 3.29

(28)

4.3 Stability run with cyclic voltammetry

After the initial CV measurements, stability measurements of nickel oxide films during 1000 and 900 cycles in CV were then performed in the LiTF- PC/EC and TBA-TF-PC/EC respectively.

4.3.1 LiTF-PC/EC

Seen in Figure 12is the stability run of 1000 cycles of a nickel oxide film in LiTF-PC/EC. When it reaches the 300^th cycle, the current density stops decrease and the curve stabilizes. Instead it turns back again and settles at the curve of the 1000^th cycle.

Figure 12: A 1000 cycle run of CV of a nickel oxide film in LiTF-PC/EC.

The 2^d, 10^th, 300^th and 1000^th cycle are shown. The curve stabilizes at the 300^th cycle.

(29)

4.3.2 TBA-TF-PC/EC

A similar stability run performed, but in TBA-TF-PC/EC is shown in Figure 13. Here, an invert behaviour is seen, since the area of the curve increases by each cycle instead of decreases like the case in LiTF-PC/EC.

In TBA-TF-PC/EC it stabilizes at the 200^th and settles at the value at the 900^th cycle.

Figure 13: A 900 cycle run of CV of a nickel oxide film in TBA-TF-PC/EC.

The 2^d, 10^th, 200^thand 900^thcycle. The curve stabilizes at around 300 cycles.

The curve stabilizes at the 200^th cycle.

The values of many of the cycle numbers, e.g. 300, 600, 1000 got very noisy which was the reason values of the 200th and 900th cycles were used in Figure 13.

(30)

4.4 Square wave voltammetry

When calculating the switching time of the nickel film in the different electrolytes used, the values of 90% coloring and bleaching was used.

The reason for this was, as seen in Figure 14, that the noise of the measurements makes it hard to decide an exact value of the 100% switch time. The different switch times can then, by studying these previous mentioned curves, be obtained.

Figure 14: Transmittance vs time measurements from SWV measurements of a nickel oxide film in different electrolytes. The three different figures shows;

a) The lithium-based electrolytes, b) The TBA-based electrolytes and c) The protonic electrolytes. The legend inserted in the figures a) and b) , contains the respective salt used.

An example of the average switch time of the first ten cycles’ colored and bleached states for all the electrolytes can be seen in Figure 15.

(31)

Figure 15: Average switching time of the first 10 cycles for 90% bleaching (O) and 90% coloring (X) for nickel oxide in different electrolytes, respectively. The x-axis shows the different electrolytes, the numbers corresponding to Table 2 below, which also shows the time in numericals.

In Table2below, the numerical values corresponding to the points in Figure15 for the switching time is shown.

Table 2: Table of the numerical data obtained from Figure 15. The values show the switching time of the nickel oxide film in different electrolytes for 90

% bleaching and 90 % coloring respectively.

Electrolytes Bleaching time (s) Coloring time (s)

1, LiClO4 - PC/EC 63.7 41.3

2, LiTF - PC/EC 63.6 39.7

3, LiTFSI - PC/EC 69.9 44.2

4, TBAClO4 - PC/EC 49.8 48.8

5, TBATF - PC/EC 45.2 42.5

6, TBATFSI - PC/EC 59.7 59.1

7, KOH 46.7 86.6

8, Prop 56.8 86.2

9, Phos 46.8 56.2

As seen in Figure 15 and again in Table 2, the bleaching time for the nickel oxide film in lithium-based electrolytes is slower than in the TBA-ones. The anion associated with fastest switching time of the nickel oxide film was TF in both the lithium-based and TBA-based electrolyte.

In TBATF-PC/EC it is 18.4 seconds faster than in LiTF-PC/EC. The coloring time was the other way around, with the LiTF-PC/EC being 2.7 seconds faster than TBATF-PC/EC.

(32)

4.5 Ex-situ optical measurements

In Figure 16, the ex-situ transmittance in the UV-VIS-NIR (300-2500 nm) range of a nickel oxide film in three different lithium-based electrolytes has been measured. These measurements have been done both in the bleached and colored states after two cycles of cyclic voltammetry.

There are no significant changes in neither the UV nor NIR range, hence the zoomed-in figure of the same curves but in an interval of 550 to 650 nm. The change in transmittance, ∆T, between the colored and bleached states were approximately 5 % at the wavelength of 550 nm with LiClO4 having a slightly higher value than the other two.

Figure 16: Transmittance in the wavelength interval 300-2500 nm of nickel oxide films in the bleached and colored states, respectively, in lithium-based electrolytes. The zoomed-in box represents the transmittance in the wavelength interval 550 to 650 nm. The dotted curves corresponds to the bleached states and the solid ones to the colored states.

(33)

In Figure 17, the ex-situ transmittance in the UV-VIS-NIR (300-2500 nm) range of a nickel oxide film in three different TBA-based electrolytes has been measured. These measurements have been done both in the bleached and colored states after two cycles of cyclic voltammetry.

The differences in the UV and NIR ranges are very small, as with the lihium-based electrolytes. The zoomed-in figure of the same curves but in an interval of 550 to 650 nm is shown to magnify the changes in this region. The change in transmittance, ∆T, between the colored and bleached states were approximately 3, 6, and 6 % at the wavelength of 550 nm for TBA-ClO4-PC/EC, TBA-TF-PC/EC and TBA-TFSI-PC/EC respectively.

Figure 17: Transmittance in the wavelength interval 300-2500 nm of nickel oxide films in the bleached and colored states, respectively, in TBA-based electrolytes. The zoomed-in box represents the transmittance in the wavelength interval 550 to 650 nm.The dotted curves corresponds to the bleached states and the solid ones to the colored states.

(34)

In Figure 18, the ex-situ transmittance in the UV-VIS-NIR (300-2500 nm) range of a nickel oxide film in three different protonic electrolytes has been measured. These measurements have been done both in the bleached and colored states after two cycles of cyclic voltammetry.

The nickel oxide film cycled in KOH has a different transmittance compared to both Prop and Phos, both in the visible as well as UV and NIR range. In the zoomed-in box, the transmittance at wavelengths between 550 and 650 nm is shown. The change in transmittance, ∆T, at 550 nm of the nickel oxide film were 17, 12 and 14 % for KOH, Prop and Phos respectively.

Figure 18: Transmittance in the wavelength interval 300-2500 nm of nickel oxide films in the bleached and colored states, respectively, in protonic electrolytes. The zoomed-in box represents the transmittance in the wavelength interval 550 to 650 nm.The dotted curves corresponds to the bleached states and the solid ones to the colored states.

(35)

4.6 IR-spectroscopy

In Figure 19, the IR-spectra are shown for a nickel oxide film in lithium- based electrolytes.

The peak corresponding to O-H bonds is situated at ∼3500 cm⁻¹ and the C-H bonds are situated at ∼2850 cm⁻¹. A negative absorbance is shown in both colored and bleached states for the nickel oxide films cycled in all three electrolytes. Hence, the transmittance in the IR-region of 600 to 4000 cm⁻¹ is higher than in the as-deposited state.

Figure 19: The IR-spectra of a nickel oxide film in lithium-based electrolytes.

The dotted curves are of the bleached states and the solid curves are of the colorod states.

(36)

In Figure 20, the IR-spectra are shown for a nickel oxide film in TBA- based electrolytes. The data obtained from this run shows peaks of the absorbance that are similar to the nickel oxide film in the lithium-based electrolytes. The main difference is that the nickel oxide film in the TBA-based electrolytes containing anions of TF and TFSI have a positive absorbance, i.e. lower transmittance, than in the as-deposited state. The O-H bonds are situated at ∼3500 cm⁻¹ and the C-H bonds are situated at ∼2850 cm⁻¹ which explains the peak values in these areas.

Figure 20: The IR-spectra of a nickel oxide film in TBA-based electrolytes.

The dotted curves are of the bleached states and the solid curves are of the colorod states.

(37)

In Figure21, the IR-spectra are shown for a nickel oxide film in protonic electrolytes. The IR-spectra shows the same peak at ∼3500 cm⁻¹, which pertains to O-H bonds. There is a small difference between the nickel oxide in its bleached and colored states, respectively when cycled in the protonic electrolytes. A bigger change in the absorbance of the O-H occurs between the bleached and colored state of the nickel oxide film in KOH.

Figure 21: The IR-spectra of a nickel oxide film after 2 cycles in protonic electrolytes. The dotted curves are of the bleached states and the solid curves are of the colorod states.

(38)

5 Conclusions

All nickel oxide films were subjected to electrochemical measurements, in different electrolytes, of cyclic voltammetry and square wave voltammetry, both with simultaneous in-situ optical transmittance measurements in the visible region. Ex-situ optical measurements were performed in the UV-VIS-NIR (300-2500 nm) range. To deduce which ions might ad- sorp onto the surface or intercalate into the structure of the nickel oxide films, IR-spectroscopy analyzes were made.

To determine the performance of the nickel oxide films, the coloration efficiency (CE) was used as figure of merit. The desired value is to achieve a high optical modulation with as little amounts of charge as possible. The electrolytes that corresponded to the highest CE of the nickel oxide film in in-situ optical measurements was TBAClO4 - PC/EC (97.6 cm²/C) and KOH (101.5 cm²/C) with the other electrolytes ranging between 60 and 80 cm²/C

The result show that neither lithium nor TBA has a significant impact on the electrochromic (EC) response. Neither does the anions used along with these salts show any specific impact of the EC response. The protonic electrolytes all show a larger optical modulation with the KOH being most effective since it is able to transfer OH⁻ to the surface of the nickel oxide and extracts protons (H⁺) from the bulk nickel oxide.

In propionic acid and phosphoric acid it is rather the H⁺ that interca- lates/deintercalates the structure of the nickel oxide that is the reason behind the electrochromic respons.

The IR-spectra of the nickel oxide did not show any specific difference between the anions used in the non-aqueous solutions. Several peaks oc- cured in the IR-spectra of the nickel oxide in aqueous solutions, though they were non-conclusive as the possible absorbance bands overlap ea- chother.

In general, the results obtained in this thesis, particularly the IR- spectroscopy, should be thoroughly investigated to be able to conclude how the electrochromic properties of nickel oxide change when cycled in different electrolytes. More experiments could be carried out by a different party and/or with different parameters to see whether or not similar results are obtained.

(39)

Acknowledgements

I would like to thank my supervisor Ilknur Bayrak Pehlivan for helping me conduct the experiments as well as giving good feedback when writ- ing my thesis. Thanks to ChromoGenics, especially Esat Pehlivan, who provided me with the samples, making this thesis what it is. Thank you Andreas Mattson without your help, IR and optical analyzes would not have been done. I would also like to thank my reviewer Gunnar Niklasson for thoughts and knowledge about my subject. I would like to thank all the people at the division of Solid State Physics at Uppsala University for insights and thoughts about all kinds of subjects. Thanks to Victor Hägglund for all talk about nonsense during lunches.

(40)

References

[1] Claes-Göran Granqvist,

Oxide electrochromics: Why, how, and whither,

Solar Energy Materials & Solar Cells, Vol. 92, pp. 203-208, 2006.

[2] Hong Ye, Xianchun Meng, Linshuang Long, Bin Xu,

"The route to a perfect window",

Renewable Energy, Vol. 55, pp. 448-455, 2013.

[3] Claes-Göran Granqvist,

Transparent conductors as solar energy materials: A panoramic re- view,

[4] Heinz Dürr, Henri Bouas-Laurent,

Photochromism - Molecules and systems, Elsevier,

Amsterdam, Netherlands, 2003.

[5] Anneke Georg, Andreas Georg

Electrochromic device with a redox electrolyte,

[6] Paul Monk, Roger Mortimer, David Rosseinsky, Electrochromism and electrochromic devices, Cambridge University, New York, USA, 2007.

[7] C.G. Granqvist, A. Azens, P. Heszler, L.B. Kish, L. Österlund, Nanomaterials for benign indoor environments: Electrochromics for

“smart windows”, sensors for air quality, and photo-catalysts for air cleaning,

[8] Sara Green,

Electrochromic Nickel-Tungsten oxides,

PhD Thesis, Uppsala University Uppsala, 2012.

[9] S. Green, J. Backholm, P. Georén, C.G. Granqvist, G.A. Niklasson, Electrochromism in nickel oxide and tungsten oxide thin films: Ion intercalation from different electrolytes,

Solar Energy Materials & Solar Cells, Vol.93, pp. 2050-2055, 2009.

[10] S. Passerini, B. Scrosati, A. Gorenstein,

The intercalation of lithium in nickel oxide and its electrochromic

(41)

properties,

J.Electrochem.Soc., Vol.137, pp. 3297-3300, 1990.

[11] The Chemical Rubber Company, Handbook of chemistry and physics, USA, 55^th edition, pp. D-121, 1974

Electrochromic properties of nickel oxide in different electrolytes

Examensarbete 30 hp Juni 2013

Electrochromic properties of

nickel oxide in different electrolytes

Anders Stenman

Abstract

Electrochromic properties of nickel oxide in different electrolytes

Elektrokromiska egenskaper hos nickeloxid i olika elektrolyter

Contents

List of Abbreviations

List of Figures

1 Introduction

1.1 Background

1.2 Electrochromism

1.3 Configuration of an electrochromic device

1.4 Aim of thesis

2 Theory

2.1 Electrochromism in nickel oxide

2.2 Optical properties

2.3 Coloration Efficiency

3 Experiments

3.1 Sample preparation

3.2 Electrochemistry

3.3 Optical measurements

3.4 IR-spectroscopy

4 Results and discussion

4.1 Cyclic voltammetry with in-situ optical mea- surement

4.2 Coloration Efficiency

4.3 Stability run with cyclic voltammetry

4.4 Square wave voltammetry

4.5 Ex-situ optical measurements

4.6 IR-spectroscopy

5 Conclusions

Acknowledgements

References