the optical and the electrochromic properties of sputtered TiO2 thin film

Effect of Substrate Temperature on the Optical and the ElectrochromicKil Dong Lee∗ Department of Physics, Kyonggi University, Suwon 442-760 Received 15 February 2005, in final form 18 Ap

Effect of Substrate Temperature on the Optical and the Electrochromic

Kil Dong Lee∗

Department of Physics, Kyonggi University, Suwon 442-760 (Received 15 February 2005, in final form 18 April 2005)

Titanium oxide films (TiO2) were deposited on ITO-coated glass substrates at different substrate

temperatures by using an RF reactive magnetron sputtering in an Ar/O2 atmosphere, and their

electrochromic properties and stability by repeated coloring and bleaching cycles were investigated

for counterelectrode applications The electrochemical properties of the TiO2 films as

counterelec-trodes showed weak dependences on the substrate temperature The optical band gap of the film

increased from 3.30 eV to 3.40 eV when the substrate temperature was increased from room

tem-perature to 500◦C The cyclic durability of the TiO2 films deposited at a substrate temperature

of 200◦C was found to be the most stable and was suitable for counterelectrode applications when

the films were subjected to 1000 cycles between –2 V and +2 V in a 1-M solution of LiClO4

PACS numbers: 68

Keywords: Titanium oxide films, RF reactive magnetron sputtering, Stability

I INTRODUCTION

The preparation and the characterization of thin films

of TiO2 have been receiving the greatest attention

dur-ing the past two decades because of their high refractive

index and their dielectric constant In addition, this

ma-terial is widely used as an electrochromic (EC) layer for

smart-window applications [1–5] The characteristics of

a smart-window, which allows dynamic control of the

solar energy passing through the window, have been

in-vestigated widely in order to increase the efficiency of

en-ergy in a building [6–12] A smart window is fabricated

by using multilayer thin films, and a major element is

the EC thin film This EC thin film is characterized by

a reversible and persistent change of the optical

prop-erties under the action of a voltage pulse Since Deb

[13] found the EC effect in tungsten-oxide thin films,

EC films have been prepared using sputtering,

electron-beam, spray pyrolysis, thermal evaporation, and sol-gel

methods, and their physical and electrochemical

proper-ties have been studied [14–22] With all these deposition

methods, TiO2 films can be made with largely varying

structural, optical, and EC properties

Typical EC materials are transition metal oxide films

Oxides of tungsten, molybdenum, vanadium, and

tita-nium belong to the group of cathodic EC materials,

which are colored by reduction reactions On the other

hand, oxides of nickel and iridium are known to be anodic

EC materials, which are colored by oxidation reactions

∗ E-mail: gdlee@kyonggi.ac.kr; Fax: +82-31-253-1165

[1, 5] In these EC materials, the preparation and the characteristics of the TiO2films, as an optical material, have been actively investigated because of their mechan-ical and chemmechan-ical durability, high refractive index, and high transparency [23, 24] However, the durability of the TiO2films have not been extensively investigated for smart-window applications Even though EC films have the advantage [1] of a long open-circuit memory and a small consumption of electric power, they are not yet available for windows in buildings because of the short device lifetime The development of electrolytes as ion sources, structural improvements of the thin films, pro-cess development of thin films for EC devices, etc have been actively studied to increase the durability and per-formance of the EC films [1,6,10,18,20] However, stud-ies of the durability by repeated coloring and bleaching (C/B) cycles and performance evaluation of the EC TiO2 thin films for counterelectrode applications as a function

of the preparation conditions have not been performed

in detail

On the other hand, sputtering method for thin films deposition are widely used for industrial products be-cause high-quality films without contamination but with high density, high adhesion, high hardness, etc., can be obtained at a low substrate temperature with good thick-ness uniformity over a large area In this study, we report the preparation of the TiO2films by using an RF (radio frequency) reactive magnetron sputtering for industrial application and evaluate the influence of the substrate temperature on the electrochromic properties, the opti-cal and the structural properties, and the cyclic stability

-1383-II EXPERIMENTS

The TiO2 thin films were prepared with an RF

mag-netron deposition sputter system The deposition

sput-ter system was a stainless-steel vacuum chamber of about

60 dm3 The background pressure in the sputtering

vac-uum chamber was 1.4 ∼ 1.8 × 10−3 Pa and was

main-tained by using a turbomolecular pump backed up by

a mechanical pump High-purity argon (99.99 %) and

oxygen (99.99 %) were used as the sputtering and the

reactive gases, respectively

Fig 1 shows a schematic diagram of the RF

mag-netron apparatus for the deposition of TiO2 films The

substrate used for deposition was glass coated with ITO

(indium tin oxide; In2O3: Sn) The sheet resistance and

the thickness of the ITO layer were 10 Ω/cm2 and 2000

˚

A, respectively The substrate was mounted in a holder

The substrate was rotated (10 rpm) during sputtering in

order to increase the film uniformity A titanium metal

(Ti) disk, (Pure Tech Inc) 0.076 m (3 inch) in

diam-eter, with 99.95 % purity was used as the target The

thickness of the Ti target was 0.0032 m (0.125 inch)

The RF power used was 300 W at 13.56 MHz The flow

rates of argon and oxygen were controlled with a MFC

(mass flow control), and the pressures were measured

with Pirani and Penning pressure gauges The distance

between the target and the substrate was about 0.1 m

Before deposition, the chamber was pumped down to 1.4

∼ 1.8 × 10−3 Pa, and the target was pre-sputtered in a

pure argon atmosphere for 600 s in order to remove the

surface oxide layer of the target Afterward, oxygen was

introduced into the chamber and the target was

sput-tered in an argon and oxygen mixture The deposition

parameters for sample preparation are shown in Table

1 The oxygen flow rate was varied from 5 to 25 sccm

(standard cubic cm/min), and the argon flow rate was

kept constant at 300 sccm for deposition with

increas-ing substrate temperature This led to a change in the

Ar/O2flow ratio from 60 to 12, which covered the range

of the metallic-mode sputtering mode and target

poison-ing the sputterpoison-ing mode The oxygen concentration in

Fig 1 Schematic diagram of an RF reactive magnetron

sputtering apparatus for the deposition of TiO2 films

Table 1 Deposition conditions of the RF reactive mag-netron sputtering system

Substrate temperature room temperature to 500◦C Base pressure of system 1.4 ∼ 1.8 × 10−3 Pa Total sputtering pressure

(Ar+O2) 4.67 ∼ 4.80 × 10

−1

Pa

Target-substrate distance 0.1 m

Variation in oxygen 5 ∼ 25 sccm Target material metallic titanium

the deposited films was, thus, modulated The thickness

in this experiment varied between 310 nm and 1000 nm when the oxygen flow rate was changed from 5 to 25 sccm for deposition The deposition rate is a very important parameter for practical thin-film production

The thickness of the TiO2 films deposited on ITO-coated glass substrates was measured by using a α-step profilometer, and the deposition rate was calculated from the film thickness obtained for a given deposition time The values depended on the total pressure and the oxy-gen concentration of the sputtering atmosphere In the case of the deposition at oxygen flow rates of 15 ∼ 20 sccm, the typical deposition rate was 0.0083 nm /s Nearly no thickness variation with increasing substrate temperature was observed The substrate was heated

by using a hot stainless-steel plate, the temperature of which was kept constant within about 5◦C as measured

by a thermocouple Since the substrate and the plane stainless-steel plate were in close contact in vacuum,

we assumed that the substrate temperature was equal

to the actually measured stainless-steel plate’s temper-ature For the film deposited at RT (room tempera-ture) without extra heating, the substrate temperature increased during continued sputtering owing to energetic particle bombardment, intrinsic to the sputtering pro-cess, until about 90 ◦C, which was reached after 4800 s For this case, it is very difficult to determine the actual substrate temperature during thin film deposition, but because of the low thermal conductivity of the glass and the relatively high heat capacity of the metal plate, we assumed the substrate temperature to be about 90 ◦C throughout the deposition We prepared a set of sample under the above sputter conditions for substrate temper-atures between RT and 500 ◦C C/B of the films were carried out by using an EC cell apparatus to examine the EC reaction and the cyclic durability

Fig 2 shows the apparatus of the EC cell to moni-tor the ion insertion (extraction) reaction in the TiO2 film The TiO2 film was used as the working electrode

in the EC cell A platinum wire was used for the coun-terelectrode The Li+ ion source for C/B the films was

a solution of 1-M LiClO propylene carbonate The

col-Fig 2 Schematic diagram of the EC cell apparatus

pre-pared for EC property measurements

oration area of the TiO2 film was about 0.02 × 0.015

m2 Each potential was measured against the saturated

calomel electrode (SCE) For C/B, a voltage was applied

between the TiO2 film and the Pt electrode under the

potentiostat condition (PARC, Model 273) C/B were

achieved by switching the potential between + 2.0 V and

– 2.0 V at a scan rate of 40 mV/s All experiments were

performed in a glove box filled with nitrogen gas During

the measurements, the electrolyte was bubbled with dry

nitrogen gas in order to remove dissolved oxygen and to

suppress water increase in the electrolyte

The spectral transmittances of the TiO2 films after

C/B cycling were measured in the visible region by using

a spectrophotometer (Kontron Inst, Uvikon 941 plus)

After the desired number of cycles had been completed,

the sample was withdrawn from the electrolyte, rinsed in

distilled water without affecting coloration, blow-dried

with filtered air, placed in the sample compartment of

the spectrophotometer, and subjected to optical

mea-surements When these were completed, the sample

could be put back into the electrolyte and run through

more C/B cycles In order to characterize the

elec-trochromic behavior of the deposited films, we performed

cyclic voltammetry experiment using a potentiostat All

the cyclic voltammograms (CV) were taken at room

tem-perature in a quiescent solution In the cyclic

voltam-metry experiment, the potentiostat applied a potential

ramp to the working electrode to gradually change the

potential; then, the scan was reversed, returning to the

initial potential During the potential sweep, the

poten-tiostat measured the current resulting from the applied

potential These values were then used to plot the cyclic

voltammetry graph of current versus applied potential

The variation of the current flowing through film, which

is related to the insertion (extraction) of ions into (out

of) the film, can be measured by using a cyclic

voltam-metry The transmittance of the sample was calculated

by integrating with respect to the solar air mass 2 [25]

and photopic spectra [26]

The crystal structure of the films was investigated by

using X-ray diffraction (Philips model PW 3710) with

Cu KαX-ray (λ = 1.54 ˚A) and a Ni filter X-ray diffrac-tion studies revealed that the TiO2 films were mainly crystalline at increased substrate temperatures The sur-face morphology of the films was observed by means of scanning electron microscopy (SEM: 15 kV, 100 kX) In order to prevent charge build-up, we sputtered a thin 10-nm-thick gold film on the sample surface before making the SEM measurements

III RESULTS AND DISCUSSION

Fig 3 shows the variation in spectral transmittance

of TiO2 film deposited on ITO-coated glass at RT The film thickness measurement was performed at different locations [up (1), middle (2), and down (3)] on sample

in order to examine its uniformity As the figure shows, the transmittance did not change with sample location, and at a photopic wavelength of 550-nm, the transmit-tance of the film was about 82 % The 550-nm wave-length indicates the peak of the photopic (human-eye response spectrum) spectrum From these data, the film was found to be very uniform in thickness In addition, the samples with substrate temperatures of 200 ◦C, 300

◦C, and 500 ◦C were observed to be very uniform The undulation in the spectra is due to optical interference caused by thin films of titanium oxide and indium tin oxide on glass, which had thicknesses comparable to the wavelength of visible light

Fig 4 the shows spectral transmittance of TiO2 films prepared at substrate temperatures of RT, 200 ◦C, 300

◦C, and 500◦C As the figure shows, the transmittance

of the films deposited at a substrate temperature of 500

◦C was lower than those of the other films at wavelengths

of 400 nm, 450 nm, 550 nm, and 750 nm The decreased transmittance is attributed to insufficient oxygen incor-poration in the film during deposition due to a dimin-ished oxygen-sticking coefficient [23] In other words, the films deposited at higher temperatures are more absorb-ing A similar observation was made by Rao and Mohan

Fig 3 Variation in spectral transmittance at different locations in a TiO2 film deposited on ITO-coated glass at substrate temperature of 200◦C

Fig 4 Specral transmittance of TiO2 films deposited on

ITO-coated glass at different substrate temperatures

and by Rae et al [27, 28] for electron-beam-evaporated

TiO2 films The transparency of the films exhibits a

sharp decrease in the ultraviolet (UV) region, as viewed

from the transmittance spectra This decrease is caused

by the fundamental absorption of light

Generally in the visible region, the absorption

coeffi-cient, α, is influenced by the scattering of light due to

the surface roughness, and it can be obtained from the

approximate relation [24,29]

T =(1 − R)

2exp(−α(λ)d)

where T is the transmittance, R is the reflectance, d is

the thickness of the film, and λ is the wavelength

How-ever, at shorter wavelengths close to the optical band

gap, the scattering losses are dominated by the

funda-mental absorption and the following relation is often used

[30]:

Above the threshold of fundamental absorption, the

dependence of α on incident light energy is

where E = hν is the photon energy, Eg is the optical

band gap, and α0is a constant which does not depend on

E The value of m may be taken as m = 2, a characteristic

value for the indirect allowed transition, which dominates

over the optical absorption, according to the theoretical

and the experimental results in Refs 24 and 31 For

photon energies hν > Eg, the material can absorb these

photons For hν < Eg , α = 0 and the photons cannot

be absorbed anymore

Fig 5 shows the dependences α1/2 = f (E) for TiO2

films deposited on ITO-coated glass From the linear

part of these dependences, one can obtain the

extrap-olated optical band gap, Eg, for α = 0 for each curve

The results show that the optical band gap increases as

the substrate temperature increases The optical band

Fig 5 Absorption coefficient of TiO2 films deposited on ITO-coated glass at different temperatures

Fig 6 Spectral transmittance of the as-prepared TiO2

films deposited on ITO-coated glass at different oxygen flow rates

gaps calculated from the extrapolate method were 3.30

eV, 3.35 eV, 3.35 eV, and 3.40 eV, respectively, for films deposited at substrate temperatures of RT, 200 ◦C 300

◦C, and 500◦C This changes in the optical band gap is another indication of structural modifications, but these calculated values of the optical band gap vary little in comparison with the reported results [23] The differ-ence is attributed to microstructural variations in the film during deposition due to the different preparation conditions of the samples

Fig 6 shows the spectral transmittance of the TiO2 films prepared at different oxygen flow rates Films de-posited at oxygen flow rates of 10 ∼ 20 sccm are trans-parent, with transmittances exceeding 80 % in the visible region However, the transmittance of the film deposited

at an oxygen flow rate of 5 sccm was very low in the visible region because the coloration of the surface had been changed to a deep blue color These results indi-cate that the chemical composition and the thickness of film were changed due to the oxygen content in the sput-tering chamber In the case when oxygen contents of 5

∼ 25 sccm were introduced into the sputtering chamber, the total working pressure was changed up to 4.80 ×

Fig 7 Variation of absorption coefficient of as-prepared

TiO2 films deposited on ITO-coated glass with the oxygen

flow rates

Fig 8 Deposition rate for TiO2films as a function of the

oxygen flow rate

10−1 Pa In addition, the low transmittance at shorter

wavelength (λ < 300 nm) is due to ITO

Fig 7 shows the spectral variation of the absorption

coefficient for TiO2 films prepared at different oxygen

flow rates The values of the absorption coefficients were

calculated on the basis of Fig 6 In the 190 – 400-nm

wavelength range, the absorption coefficient of the TiO2

film deposited at an oxygen flow rate of 5 sccm was found

to be much higher than those of the samples deposited

at oxygen flow rates of 15 ∼ 25 sccm In the case of the

films deposited at oxygen flow rates of 15 ∼ 25 sccm,

the spectral absorption coefficient did not change

con-siderably From these results, the fundamental

absorp-tion edges due to ITO (band gap: 3.5 eV) and to the

TiO2film (band gap: 3.2 eV) were found to occur in the

UV region Thus, the samples deposited at oxygen flow

rates of 15 ∼ 25 sccm showed that no differences in the

absorption coefficient However, the high absorption

co-efficient in the TiO2film deposited at oxygen flow rate of

5 sccm might be due to insufficient oxygen incorporation

in the film during deposition When reactive sputtering

is used for the preparation of compound films, it is very

important to know the target surface conditions during

sputtering For this purpose, it is very useful to

moni-Fig 9 X-ray diffraction patterns of TiO2 films deposited

on ITO-coated glass at different substrate temperatures: (a)

RT, (b) 100◦C, (c) 150◦C, (d) 200◦C, (e) 250◦C, (f) 300

◦

C, (g) 350◦C, and (h) 500◦C

Fig 10 Variation of deposition rate for TiO2 films with the substrate temperatures

tor the deposition rate’s dependence on the oxygen flow rate

Fig 8 shows the deposition rate of the film as a func-tion of the oxygen flow rate In these measurements the argon flow rate was kept at 300 sccm, and the total pres-sure was about 4.67 ∼ 4.80 × 10−1 Pa The deposi-tion rate was determined by using a the film thickness obtained with a profilometer divided by the deposition time As the fig shows, the abrupt decrease in the de-position rate at oxygen flow rates of 5 – 15 sccm is the result of target oxidation and the resulting low sputter-ing yield of the oxide However, the deposition rate did not depend much on the oxygen content in the sputtering chamber at oxygen flow rates above 15 sccm The target was found to be near completely oxidized in this oxygen flow range In the case of the TiO2 film deposited at an oxygen flow rate of 5 sccm, the film thickness was about

1000 nm, as measured by using a profilometer Also, the optical transmittance was reduced for higher oxygen flow rates No obvious steep absorption band was seen

in the transparent region examined, suggesting that the deposited films did not include any impurity ion or defect

Fig 11 Spectral transmittance of electrochromic TiO2 films deposited on ITO-coated glass at different substrate tempera-tures: (a) RT, (b) 200◦C, (c) 300◦C, and (d) 500◦C

centers

Fig 9 shows the diffraction patterns of films deposited

at different substrate temperatures As the figure shows,

A, B, C, D, E, F, G, and H are samples prepared at

substrate temperatures of RT, 100 ◦C, 150 ◦C, 200◦C,

250 ◦C, 300 ◦C, 350 ◦C and 500 ◦C, respectively The

diffraction patterns show no distinguishable features In

films A-G, a strong amorphous background is seen The

brookite, anatase, and rutile phases were not observed

in any of the films investigated Pawlewicz and Busch

[32] observed mixed phases of anatase and rutile for

sub-strate temperatures in the range 200 – 500◦C A similar

amorphous-to-crystalline transition around 350 – 400◦C

was observed by Bange et al [33] for evaporated films

and by Willams and Hess [34] for RF sputtered films In

all cases, the mixed phase is present up to around 600

◦C; at higher temperature, only the rutile structure

pre-vailed These different results in comparison with other

investigator’s results [33,34] might be due to differences

in sample preparation conditions

Fig 10 shows the variation in the deposition rate with

substrate temperature Obviously, the deposition rate

was not influenced by the substrate temperature until

substrate was heated to 350 ◦C Most probably, this

re-sult suggests that the deposition rate depends only on the

number of sputtered Ti atoms that subsequently reach

the substrate The effects in this experiment, such as

the local pressure reduction in the sputtering plasma,

the variation in the sticking coefficients of Ti or oxygen,

and the re-evaporation from the substrate, are neglect

In addition, in the case of the sample with a substrate

temperature of 500 ◦C, the larger value of the

deposi-tion rate in the sputtering chamber compared to that

obtained at lower substrate temperatures might be due

to a differences among the sputtered atoms, the substrate temperatures, and contents of oxygen

Fig 11 shows the spectral transmittance of the

EC films prepared at substrate temperatures of RT,

200 ◦C, 300 ◦C, and 500 ◦C As the fig shows, the transmittance of colored and bleached films did not significantly change compared with those of the as-prepared film These results indicate that the TiO2 films deposited at various substrate temperatures can

be used as counterelectrodes in electrochromic devices (glass/ITO/WO3/LiAlF4/TiO2/glass) such as smart-windows Counterelectrodes with cathode coloring char-acteristics should have a small decrease in transparency

on the insertion of ions Therefore from the trans-mittance data, the TiO2 films can have the properties

of passive counterelectrodes (ion storage) on the inser-tion/extraction of ions in the Fig 11 The TiO2 films are well known to have potential advantages as coun-terelectrodes due to their microstructural features being favorable for ion transport [35]

Fig 12 shows variation of charge capacity (electric charge density) as a function of the cyclic number for samples prepared at substrate temperatures of RT [sam-ple (a)], 200 ◦C [sample (b)], 300 ◦C [sample (c)], and

500◦C [sample (d)] in order to evaluate during 1000 cy-cles the electrochromic performance for counterelectrode application The charge capacity is the effective interca-lated charge that is reversibly transferred upon cycling [36] Maximization of this charge gives the maximum transmittance difference Further, in EC 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 was obtained from CV and were recorded during cathodic

Fig 12 Dependence of the charge density of electrochromic TiO2 films on the C/B cycles for substrate temperatures of (a)

RT, (b) 200◦C, (c) 300◦C, and (d) 500◦C

Fig 13 Dependence of the cyclic voltammograms of electrochromic TiO2films on the C/B cycles for substrate temperatures

of (a) RT, (b) 200◦C, (c) 300◦C, and (d) 500◦C

bleaching in order to avoid any possible contribution due

to oxygen evolution The transferred charge was

cal-culated by integrating the measured intensity from the

CV plot As the figure shows, the TiO2 films deposited

with substrate temperatures of 200 ◦C gave maximum

values of both the inserted coloring charge density and

the extracted bleaching charge density after 1000 cycles

This means the cyclic durability of the TiO2 films

pre-pared at substrate temperatures of 200 ◦C for counter-electrode materials is stronger than those of the other samples Such a dependence of the transferred charge

on the substrate temperature may be due to the dif-ference in both the boundaries and the surfaces of the TiO2 microcrystallites Therefore, the large amount of transferred charge in sample (b) after 1000 cycles can be attributed to numerous grain boundaries Grain

bound-aries are well known to be good diffusion channels for

ions to be injected into or extracted from films In other

words, this result indicates that the space between TiO2

crystallites become larger, which makes it easier for ions

to be injected into or extracted from the films under

cy-cling In addition, the total charge contents in films with

increasing the number of cycles were nearly zero This

suggests that the electrochromic reaction is a reversible

process

Fig 13 shows the CV curves of the TiO2 films

pre-pared under different substrate temperatures for

evalua-tion of the electrochemical properties The charge

capac-ity of the TiO2films changes rapidly in the first stage of

cyclic voltammetry, so for stability, we measure its value

after 10 cycles As the figure shows, the C/B process for

the samples with substrate temperatures of RT [sample

(a)], 200 ◦C [sample (b)], 300 ◦C [sample (c)], and 500

◦C [sample (d)] is nearly reversible The redox reaction

is slightly shifted toward a lower negative potential after

1000 cycles Coloration (upon oxidation) and

bleach-ing (upon reduction) are associated with redox peaks

The anodic and the cathodic peaks of the samples is

de-creased in size after 1000 cycles These results indicate

that there was little ion transport into this film,

lead-ing to a weak blue coloration Thus, the TiO2 film’s

microstructure was clearly modified during the 1000

cy-cles On the other hand, in the case of sample (b), the

redox reaction was stronger than it was for samples (a),

(c), and (d) This result is consistent with that of the

sample (b) (Fig 12) with a good ion storage material

characteristic Therefore, sample (b) (Fig 13) was also

confirmed to be suitable as a counterelectrode material

for use in electrochromic devices because of its almost

transparent property in the inserted and the extracted

charge states

Fig 14 shows the surface morphology of the TiO2film

prepared at substrate temperatures of RT [sample (a)],

200 ◦C [sample (b)], and 300 ◦C [sample (c)] As the

micrographs show, the grains sizes of the TiO2 grains

are found to be fine, and the grain size is found to

in-crease with increasing substrate temperature, but films

appear to be a little porous At lower substrate

tem-peratures, the deposited atoms are expected to have

re-stricted surface mobility Restrictive diffusion of atoms

prevents crystal growth at energetically favorable sites

and causes atoms to nucleate at new sites [37] This

re-sults in a structure of smaller grains and relatively weaker

preferred orientation compared to the structure obtained

at higher substrate temperatures At higher substrate

temperatures, atoms have enough energy to diffuse to the

preferred nucleation sites Easy propagation of atoms

re-sults in the development of a strong preferred orientation

[37] In the case of sample (b) with a substrate

temper-ature of 200 ◦C, the grain size is more uniform than it

is for the samples prepared at substrate temperatures of

RT and 300◦C Sample (c) with a substrate temperature

of 300◦C shows closely packed grains, but the grain size

is clearly larger than it is in samples (a) and (b) This

Fig 14 SEM micrographs of TiO2films prepared on ITO-coated glass at different substrate temperature: (a) RT, (b)

200◦C, and (c) 300◦C

result indicates that the grain size increases whereas the grain boundary decreases at higher temperatures, which might result in a weak electrochromic reaction There-fore, variation of the grain sizes was found to be corre-lated with the electrochromic performance of the TiO2 film From those results, we concluded that the best film for large ion storage with counterelectrode charac-teristics had been obtained at a substrate temperature

of 200 ◦C The surface morphology of that sample was found to have a slightly porous structure consisting of uniform grains for weak electrochromic reactions with good counterelectrode characteristics

IV CONCLUSIONS

TiO2films were deposited on ITO-coated glass at dif-ferent substrate temperatures by using an RF reactive magnetron sputtering apparatus The optical band gap

of the TiO2 films increased from 3.30 eV to 3.40 eV as the substrate temperature was varied from RT to 500

◦C This change in the optical band gap is another in-dication of the structural modifications As a result of examining the Ti target oxidation phenomenon in sput-tering vacuum chamber, the abrupt decrease in the de-position rate at oxygen flow rates of 5 – 15 sccm was the result of target oxidation and to the sputtering yield

of the Ti oxide being lower than that of the Ti metal.

Significant changes in the X-ray diffraction patterns for

the TiO2 films with increasing substrate temperatures

were not observed The electrochromic behavior of the

sputtered films for counterelectrode applications weakly

depended on the substrate temperature during RF

mag-netron sputtering As a result of analyzing the CV

mea-surement results of the films, we confirmed that the films

deposited at a substrate temperatures of 200◦C showed

excellent counterelectrode characteristics The surface

morphology of the film deposited at that temperature

was found to have a slightly porous structure consisting

of uniform grains for a weak electrochromic reaction

REFERENCES

[1] C G Granqvist, Handbook of Inorganic Electrochromic

Materials (Elsevier Science, Amsterdam, 1995)

[2] R Cinnsealach, G Boschloo, S N Rao and D

Fitzmau-rice, Sol Energy Mater & Sol Cells 57, 107 (1999)

[3] K Yoshimura, T Miki and S Tanemura, J Vac Sci

Technol A 15, 2673 (1997)

[4] F Campus, P Bonhˆote, M Gr tzel, S Heinen and L

Walder, Sol Energy Mater & Sol Cells 56, 281 (1999)

[5] M Kitao, Y Oshima and K Urabe, Jpn J Appl Phys

36, 4423 (1997)

[6] A Azens and C G Granqvist, J Solid State

Elec-trochem 7, 64 (2003)

[7] J R Pitts and S K Deb, Electrochim Acta 46, 2237

(2001)

[8] A Azens, G Vaivars, M Verszelei, L Kullman and C

G Granqvist., J Appl Phys 89, 7885 (2001)

[9] H N Cui, S Jia, L J Meng and V Teixeira, Microchim

Acta 145, 19 (2004)

[10] L Berggren and G A Niklasson, Solid State Ionics 165,

51 (2003)

[11] C G Granqvist, E Avendano and A Azens, Thin Solid

Films 442, 201 (2003)

[12] C M Lampert, Sol Energy Mater 11, 1 (1984)

[13] S K Deb, Philos Mag 27, 801 (1973)

[14] K D Lee, J Korean Phy Soc 38, 33 (2001)

[15] K D Lee, Sol Energy Materials & Sol Cells 57, 21 (1999)

[16] M S Mattsson, Phys Rev B, 58, 11015 (1998) [17] S Papaefthimiou, G Leftheriotis and P Yianoulis, Elec-trochim Acta 46, 2145 (2001)

[18] M Merz, J Eisenmenger, B Heinz and P Ziemann, Phys Rev B 66, 184102 (2002)

[19] M Yahaya, M M Salleh, I A Talib, Solid State Ionics

113, 421 (1998)

[20] S Hashimoto, H Matsuoka, H Kagechika, M Susa and

K S Goto, J Electrochem Soc 137, 1300 (1990) [21] K D Lee, Thin Solid Films 302, 84 (1997)

[22] H Kamal, E K Klmaghraby, S A Ali and K A Hady,

J Cryst Growth 262, 424 (2004)

[23] L J Meng, M Andritschky and M P dos Santos, Thin Solid Films 223, 242 (1993)

[24] D Mardare, M Tasca, M Delibas and G I Rusu, Appl Surf Sci 156, 200 (2000)

[25] R E Bird and R L Hulstrom, Sol Cells 8, 85 (1985) [26] F A Barnes ed., RCA Electro-Optics Handbook (RCA Corporation, New Jersey, 1974)

[27] K N Rao and S Mohan, J Vac Sci Technol A 8, 3260 (1990)

[28] K N Rao, M A Murthy and S Mohan, Thin Solid Films 176, 181 (1989)

[29] T S Moss, Optical Properties of Semiconductors (But-terworth, London, 1959)

[30] J Aarik, A Aidla, A A Kiisler, J Uustare and V Sam-melselg, Thin Solid Films 305, 270 (1997)

[31] H Tang, K Prasad, R Sanjines, P E Schmid and F Levy, J Appl Phys 75, 2042 (1994)

[32] W T Pawlewicz and R Busch, Thin Solid Films 63,

251 (1979)

[33] K Bange, C R Ottermann, O Anderson and V Jeschkowski, Thin Solid Films 197, 279 (1991)

[34] L M Williams and D W Hess, J Vac Sci Technol A1, 1810 (1983)

[35] C G Granqvist, SPIE 2017, 84 (1993)

[36] C H Hamann, A Hamnett and W Vielstich, Electro-chemistry (Wiley-VCH, Weinheim, 1998)

[37] Y M Lu, W S Hwang, J S Yang, H C Chuang and Thin Solid Films 420, 54 (2002)