www.elsevier.com/locate/ssi

Solid State Ionics 172 (2004) 283–288

Deposition mechanism of anatase TiO2 from an aqueous solution

and its site-selective deposition

Yoshitake Masudaa,*, Won-Seon Seob, Kunihito Koumotoa

aDepartment of Applied Chemistry, Graduate School of Engineering, Nagoya University, Nagoya 464-8603, JapanbKorea Institute of Ceramic Engineering and Technology (KICET), Nagoya University, Nagoya 464-8603, Japan

Received 9 November 2003; received in revised form 18 February 2004; accepted 20 February 2004

Abstract

We have developed a novel method for site-selective deposition (SSD) of anatase TiO2 thin films using a seed layer based on the

knowledge obtained by the evaluation of deposition mechanism. The nucleation and initial growth of anatase TiO2 were found to be

accelerated on amorphous TiO2 thin films compared with the substrates modified by silanol, amino, phenyl or octadecyl groups. Micropattern

having octadecyl group regions and amorphous TiO2 regions was immersed in the aqueous solution at pH 1.5 to be used as a template for

SSD. Anatase TiO2 was selectively deposited on amorphous TiO2 regions to form a micropattern of anatase TiO2 thin film in an aqueous

solution. Furthermore, deposition mechanism of anatase TiO2 in an aqueous solution has been evaluated in detail. The adhesion of

homogeneously nucleated particles to the amino group surface by attractive electrostatic interaction caused rapid growth of TiO2 thin films in

the supersaturated solution at pH 2.8. On the other hand, TiO2 was deposited on self-assembled monolayers (SAMs) without the adhesion of

TiO2 particles regardless of the type of SAM in the solution at pH 1.5 whose degree of supersaturation is low due to high concentration of H+.

Additionally, the orientation of films deposited on all SAMs was shown to be improved by enlarging the reaction time regardless of the kind

of SAM or pH. It is conjectured that the adsorption of anions to specific crystal planes caused c-axis orientation of anatase TiO2.

D 2004 Elsevier B.V. All rights reserved.

Keywords: Titanium dioxide; Self-assembled monolayer; Liquid phase deposition; Deposition mechanism; Site-selective deposition; Seed layer; Thin film

1. Introduction

Titanium dioxide (TiO2) thin films are of interest for

various applications including microelectronics [1], optical

cells [2], solar energy conversion [3], highly efficient

catalysts [4], microorganism photolysis [5], antifogging

and self-cleaning coatings [6], gratings [7], gate oxides in

metal-oxide-semiconductor field effect transistor (MOS-

FETs) [8.9], etc. Accordingly, various attempts have been

made to fabricate thin films and micropatterns of TiO2 by

several methods, and in particular, to synthesize materials

and devices including TiO2 thin films from an aqueous

solution through an environment-friendly synthesis process,

i.e., ‘‘green chemistry’’.

Micropatterning of TiO2 was attempted by a number of

methods [10–14]. We realized site-selective deposition

(SSD) of amorphous TiO2 to fabricate micropatterns of

0167-2738/$ - see front matter D 2004 Elsevier B.V. All rights reserved.

doi:10.1016/j.ssi.2004.02.068

* Corresponding author. Tel.: +81-52-789-3329; fax: +81-52-789-3201.

E-mail address: masuda@apchem.nagoya-u.ac.jp (Y. Masuda).

TiO2 thin films on self-assembled monolayers (SAMs)

[12–14]. SAMs of octadecyltrichloro-silane (OTS) were

formed on Si wafers, and were modified by UV irradiation

using a photomask to generate octadecyl/silanol-pattern.

They were used as templates to deposit TiO2 thin films by

the use of titanium dichloride diethoxide (TDD). Amorphous

TiO2 films were selectively deposited on silanol regions.

Annealing the films at high temperatures (400–600 jC) gaverise to an anatase phase, while the resolution of amicropattern

remained unchanged. However, annealing process is required

to obtain patterns of anatase TiO2 thin films in this process.

On the other hand, a micropattern of anatase TiO2 thin film

was fabricated by the site-selective immersion [15] method

using a SAM which has a pattern of both hydrophilic and

hydrophobic surfaces. A solution containing Ti precursor

contacted the hydrophilic surface during the experiment and

briefly came in contact with the hydrophobic surface. The

solution on the hydrophilic surface was replaced with fresh

solution by continuous movement of bubbles. Thus, TiO2

was deposited and a thin film was grown on the hydrophilic

surface selectively. However, feature edge acuity of the

Y. Masuda et al. / Solid State Ionics 172 (2004) 283–288284

micropattern needs to be improved further in order to use

anatase TiO2 micropatterns for electronic or optical devices.

Investigation of the mechanism of nucleation and growth of

TiO2 from the aqueous solution would produce valuable

information for making desired thin films and micropatterns.

In this study, we have developed a novel method for SSD

of anatase TiO2 thin films using a seed layer and the

knowledge obtained by the evaluation of deposition mech-

anism. We evaluated the surface zeta potential of SAMs and

TiO2 particles in a solution containing ions such as TiF62�

and BO33�. We also investigated the TiO2 deposition rate

and quantity for several kinds of SAMs and the time

dependence of the crystal-axis orientation to clarify the

mechanism of nucleation and growth. The nucleation and

initial growth of anatase TiO2 were found to be accelerated

on amorphous TiO2 thin films compared with silanol,

amino, phenyl or octadecyl groups by the evaluation of

deposition mechanism using a quartz crystal microbalance

(QCM). In our process, amorphous TiO2 was shown to

decrease the nucleation energy of anatase TiO2 and provided

nucleation sites for the formation of anatase TiO2. The

micropattern having amorphous TiO2 regions and OTS-

SAM regions was immersed in an aqueous solution con-

taining Ti precursor to be used as a template for SSD.

Anatase TiO2 was selectively deposited on amorphous TiO2

regions to form a micropattern of anatase TiO2 thin films.

2. Experimental

2.1. SAM preparation

OTS-SAM, phenyltrichlorosilane (PTCS)-SAM and 3-

aminopropyltriethoxysilane (APTS)-SAM were prepared by

immersing the Si substrate in toluene solutions containing

OTS, PTCS, APTS, respectively [16–22]. Octadecylmercap-

tan (OM)-SAM, phenylmercaptan (PM)-SAM and 2-amino-

ethanethiol (AET)-SAM were prepared by immersing the

Fig. 1. Deposition quantity of anatase TiO2 on amorphous TiO2, octadecyl groups,

deposition time and conceptual process for site-selective deposition of anatase Ti

Au-coated quartz crystal of a quartz crystal microbalance

(QCM; QCA917, Seiko EG&G) in a bicyclohexyl solution

containing OM, PM or AET, respectively. OM-SAM on the

quartz crystal of the QCM was exposed for 2 h to UV light

(184.9 nm) to assess the deposition rate of TiO2 on OH

groups. OM-SAM, PM-SAM and AET-SAM were used

instead of OTS-SAM, PTCS-SAM or APTS-SAM for

QCM analysis. Initially deposited OTS-SAM, PTCS-SAM,

APTS-SAM, OM-SAM, PM-SAM and AET-SAM showed

water contact angles of 96j, 74j, 48j 96j, 76j and 53j,respectively. UV-irradiated surfaces of SAMs were, however,

wetted completely (contact angle < 5j). This suggests that

SAMs of OTS, PTCS, APTS, OM, PM and AET were

modified to hydrophilic OH group surfaces by UV irradia-

tion. The order of SAM hydrophobicity determined from

these measurements was OTS-SAM>PTCS-SAM>APTS-

SAM>OH groups on silicon. Zeta potentials measured in

aqueous solutions (pH = 7.0) for the surface of silicon sub-

strate covered with OH groups, phenyl groups (PTCS) and

amino groups (APTS) were measured to be � 38.23, + 0.63

and + 22.0 mV [23], respectively. The order of zeta potential

in the aqueous solution of our experiment is presumed to be

APTS-SAM>PTCS-SAM>OH-SAM (OH groups on sili-

con). OTS-SAMs were exposed for 2 h to UV light through

a photomask. The UV-irradiated regions became hydrophilic

owing to the formation of Si-OH groups, while the non-

irradiated part remained unchanged, i.e., it was composed of

hydrophobic octadecyl groups, which gave rise to patterned

OTS-SAM. This patterned SAM was used as a template for

SSD of amorphous TiO2 thin films [12,13].

2.2. Deposition of anatase TiO2 thin films

Ammonium hexafluorotitanate ([NH4]2TiF6) and boric

acid (H3BO3) were separately dissolved in deionized water

at 50 jC and kept for 12 h (Fig. 1). An appropriate amount

of HCl was added to the boric acid solution to control pH,

and ammonium hexafluorotitanate solution was added.

phenyl groups, amino groups or hydroxyl groups at pH 1.5 as a function of

O2 thin films using a seed layer.

Y. Masuda et al. / Solid State Ionics 172 (2004) 283–288 285

SAMs were immersed in the solution containing 0.05 M

(NH4)2TiF6 and 0.15 M (H3BO3) at pH 1.5 or 2.8 and kept

at 50 jC to deposit anatase TiO2. Deposition of TiO2

proceeds by the following mechanisms [24]:

TiF2�6 þ 2H2OfTiO2 þ 4Hþ þ 6F� ðaÞ

BO3�3 þ 4F� þ 6Hþ ! BF�4 þ 3H2O ðbÞ

2.3. QCM measurement

Quartz crystals covered with SAMs were placed 5 mm

below the surface of the solution. The solution was kept

covered to prevent the evaporation of water, and water (50

jC) was added to compensate for any evaporated water.

Frequency decrease (DF (Hz)) was converted into weight

increase (Dm (ng)) by the following equation:

DmðngÞ ¼ �1:068� DFðHzÞ ðcÞ

3. Results and discussion

3.1. Quantitative analysis of the deposition of anatase TiO2

onto an amorphous TiO2 thin film or onto SAMs

Quartz crystals covered with amorphous TiO2 thin film,

OM-SAM (CH3), PM-SAM (Ph), AET-SAM (NH2) or OH-

SAM (OH) were immersed in a solution [15] containing

Fig. 2. SEM micrographs of (1-a), (1-b) a micropattern of amorphous TiO2 thin f

pH= 1.5.

0.05 M TiF62� and 0.015 M BO3

3� at pH 1.5 or pH 2.8 (Fig.

1). The supersaturation degree of the solution at pH 1.5 was

low as the high concentration of H+ suppressed TiO2

generation, and hence the deposition reaction progresses

slowly with no homogeneous nucleation occurring in the

solution. We found that anatase TiO2 was deposited on an

amorphous TiO2 thin film faster than on OM-, PM- AET- or

OH-SAMs at pH 1.5. This shows that the deposition of

anatase TiO2 was accelerated on amorphous TiO2 compared

with on silanol, amino, phenyl or octadecyl groups. Amor-

phous TiO2 probably decreases the nucleation energy of

anatase TiO2. The difference in deposition rate enables SSD

to be achieved. The amorphous TiO2 thin film can be used

as a seed layer to accelerate the deposition of anatase TiO2.

The deposition rate at pH 2.8 was larger than that at pH 1.5

because of the high degree of supersaturation, and homo-

geneously nucleated particles in the solution deposited on

the whole surface of the substrate, regardless of the surface

functional groups. The thickness of anatase TiO2 thin film

deposited on a quartz crystal covered with amorphous TiO2

at pH 1.5 for 1 h and at pH 2.8 for 30 min was estimated to

be 36 and 76 nm, respectively, assuming the density of

anatase type TiO2 to be 3.89 g/cm3.

3.2. SSD of anatase TiO2 using a seed layer

A micropattern [12,13] having amorphous TiO2 and

octadecyl groups was immersed in an aqueous solution

ilms and (2-a), (2-b) a micropattern of anatase TiO2 thin films deposited at

Fig. 4. FE-SEM micrograph for cross-section profile of TiO2 thin film

deposited at pH 2.8.

Y. Masuda et al. / Solid State Ionics 172 (2004) 283–288286

[15] at pH 1.5 for 1 h (Fig. 1). Deposited thin films made it

appear white compared with octadecyl group regions in

SEM micrographs (Fig. 2(2-a), (2-b)) because of the differ-

ence in height. The feature edge acuity of anatase TiO2

pattern was f 2.1% variation (i.e., 0.5/23.2) and was much

the same as we calculated from amorphous TiO2 pattern.

This resemblance was observed from Fig. 2(1-a) and (2-a).

These micrographs were taken from the same position.

Variations of these patterns were much better than that of

the pattern fabricated with a lift-off process and the usual

5% variation afforded by current electronics design rules.

Additionally, these variations were similar to that of a TEM

mesh (2.1%) we used as a photomask for Fig. 2. Therefore,

variations of these patterns can be improved through the use

of a high-resolution photomask. Deposited films showed

weak XRD patterns of anatase type TiO2 because the films

were not sufficiently thick to show strong diffraction. This

finding provides evidence for the deposition of anatase TiO2

on amorphous TiO2 regions. An atomic force microscope

(AFM; Nanoscope E, Digital Instruments) image showed

anatase TiO2 thin films to be higher than octadecyl group

regions. The center of the anatase TiO2 thin film region was

61 nm higher than the octadecyl regions, and the thickness

of the anatase TiO2 thin film was estimated to be 36 nm

considering the thickness of amorphous TiO2 thin film (27

nm) [12,13] and OTS molecules (2.4 nm) (Fig. 1). This

result is similar to that estimated by QCM measurement (36

nm). The surface roughness (RMS) of the anatase TiO2 thin

film was estimated using an AFM image. The AFM image

showed the film roughness to be 3.7 nm (horizontal distance

between measurement points: 6.0 Am), which is less than

that of amorphous TiO2 thin film (RMS 9.7 nm, 27 nm

thick, horizontal distance between measurement points: 6.0

Am) [12,13]. Additionally, the roughness of the octadecyl

group regions was shown to be 0.63 nm (horizontal distance

between measurement points: 1.8 Am).

Amorphous TiO2 accelerated the deposition of anatase

TiO2 and showed its excellent performance as a seed layer.

The feature edge acuity of anatase TiO2 patterns was

Fig. 3. Intensities of peaks as a function of reaction time (at pH 1.5, on OH

groups).

estimated to be approximately 2.1% using the same method

as used for a micropattern fabricated by the lift-off process

[24] and was the same as that of amorphous TiO2 [12,13].

The feature edge acuity could be improved by using a

higher feature edge acuity photomask since this variance

is similar to that of the TEM mesh (2.1%). XRD measure-

ments for the thin film deposited for 1 h did not show any

peaks since the deposited quantity was not sufficient to

show any diffraction, however, the thin film deposited for 7

h was composed of anatase TiO2. Anatase TiO2 thin films

were not peeled off by sonication in ethanol for 10 min and

showed strong adhesion to the amorphous TiO2 layer. This

suggests that strong chemical bonds were formed between

anatase TiO2 and amorphous TiO2.

3.3. Crystal-axis orientation of TiO2 thin film

The growth process of TiO2 thin films and the crystal-axis

orientation changes were investigated using an X-ray diffrac-

tometer (XRD; RAD-C, Rigaku) with CuKa radiation (40

kV, 30 mA) and Ni filter. Deposited films on all SAMs

showed XRD patterns of anatase TiO2 after 24 h at pH 1.5

or after 4 h at pH 2.8. XRD patterns showed the same

tendency regardless of the type of SAM, and intensities of

(004) and (105) peaks on all SAMs increased with deposition

time faster than other peaks. The degree of crystal-axis

orientation ( f ) was evaluated using the Lotgering method

[25] taking into account the following diffraction peaks:

(101) = 25.3, (004) = 37.8j, (200) = 48.0j, (105) = 53.9j,(204) = 62.7j, (116) = 68.8j, (215) = 75.0j (Fig. 5).

f ¼ P � P0

1� P0

ðdÞ

P ¼P

Ið00lÞP

IðhklÞ ðeÞ

P, calculated for the oriented sample; P0, P for non-

oriented sample (JCPDS card).

Fig. 5. TEM micrograph and electron diffraction pattern for cross-section

profile of TiO2 thin film deposited at pH 2.8 (an arrow shows growth

direction of thin film).

Y. Masuda et al. / Solid State Ionics 172 (2004) 283–288 287

The c-axis (00l) orientation of the film was enhanced by

increasing the reaction time for all the kinds of SAMs and

pH. This result suggests that the orientation of the film is

determined not at the initial nucleation or deposition stage

but at the film growth stage. The intensity of the (004) peak

quickly increased but that of (105) increased only gradually

with reaction time, and the intensities of the (101) and (200)

peaks decreased after reaching their maxima regardless of

the type of SAM or pH condition (Fig. 3).

Furthermore, the orientation of thin film deposited at pH

2.8 for 4 h was evaluated by a field emission scanning

electron microscope (FE-SEM; JSM-6700F, point-to-point

resolution 1 nm, JEOL) and a transmission electron micro-

scope (TEM; JEM4010, 400 kV, point-to-point resolution

0.15 nm, JEOL). The cross-section profile of TiO2 thin films

showed columnar morphology (Fig. 4). However, the col-

umns were not clearly identified compared with the needle-

like morphology of TiO2 thin films reported recently

[26,27]. This columnar morphology is consistent with

XRD measurement which showed weak c-axis orientation.

Fig. 5 shows a TEM micrograph and electron diffraction

pattern for the cross-section profile of a TiO2 thin film.

Many small crystals of anatase TiO2 were observed through-

out the thin film.

These observations firmly indicate that TiO2 particles

whose c-axes were perpendicular to the substrate surface

may have grown faster than other crystals. Hence, the

diffraction intensities of crystal planes almost perpendicular

to the c-axis such as (004) and (105) increased with

deposition time (Fig. 3). These particles then consumed

other particles whose c-axis was far from perpendicular to

the substrate, thus lowering the diffraction intensities of

crystal planes such as (101) and (200).

4. Conclusions

We utilized the knowledge obtained by the evaluation of

deposition mechanism for SSD of anatase TiO2 thin films.

The deposition of anatase TiO2 was shown to be accelerated

on amorphous TiO2 thin films compared with on octadecyl,

phenyl, amino or hydroxyl groups. A micropattern having

amorphous TiO2 regions and octadecyl regions to be used as

a template was prepared and immersed in the aqueous

solution. Anatase TiO2 was successfully deposited on amor-

phous TiO2 regions, and amorphous TiO2 thin film was

shown to act effectively as a seed layer to accelerate the

nucleation and initial growth of anatase TiO2. Consequently,

SSD was achieved and a micropattern of anatase TiO2 was

fabricated in the aqueous solution using a seed layer.

Acknowledgements

This work was supported in part by the 21st Century

COE Program ‘‘Nature-Guided Materials Processing’’ of the

Ministry of Education, Culture, Sports, Science and

Technology. This work was partly supported by a Grant-

in-Aid for Scientific Research (Grant-in-Aid for Young

Scientists No. 14703025, Exploratory Research No.

14655239) from the Ministry of Education, Culture, Sports,

Science and Technology granted to Y. Masuda.

References

[1] G.P. Burns, J. Appl. Phys. 65 (1989) 2095.

[2] B.E. Yoldas, T.W. O’Keeffe, Appl. Opt. 18 (1979) 3133.

[3] M.A. Butler, D.S. Ginley, J. Mater. Sci. 15 (1980) 19.

[4] T. Carlson, G.L. Giffin, J. Phys. Chem. 90 (1986) 5896.

[5] T. Matsunaga, R. Tomoda, T. Nakajima, T. Komine, Appl. Environ.

Microbiol. 54 (1988) 330.

[6] R. Wang, K. Hashimoto, A. Fujishima, Nature 388 (1997) 431.

[7] S.I. Borenstain, U. Arad, I. Lyubina, A. Segal, Y. Warschawer, Thin

Solid Films 75 (1999) 2659.

[8] P.S. Peercy, Nature 406 (2000) 1023.

[9] D.J. Wang, Y. Masuda, W.S. Seo, K. Koumoto, Key Eng. Mater. 214

(2002) 163.

[10] R.J. Collins, H. Shin, M.R. De Guire, A.H. Heuer, C.N. Shukenik,

Appl. Phys. Lett. 69 (6) (1996) 860.

[11] M. Bartz, A. Terfort, W. Knoll, W. Tremel, Chem. Eur. J. 6 (22)

(2000) 4149.

[12] Y. Masuda, T. Sugiyama, H. Lin, W.S. Seo, K. Koumoto, Thin Solid

Films 382 (2001) 153.

[13] Y. Masuda, Y. Jinbo, T. Yonezawa, K. Koumoto, Chem. Mater. 14 (3)

(2002) 1236.

[14] Y. Masuda, W.S. Seo, K. Koumoto, Langmuir 17 (16) (2001) 4876.

[15] Y. Masuda, T. Sugiyama, K. Koumoto, J. Mater. Chem. 12 (9) (2002)

2643–2647.

[16] Y. Masuda, W.S. Seo, K. Koumoto, Thin Solid Films 382 (2001) 183.

Y. Masuda et al. / Solid State Ionics 172 (2004) 283–288288

[17] Y. Masuda, W.S. Seo, K. Koumoto, Jpn. J. Appl. Phys. 39 (2000)

4596.

[18] Y. Masuda, M. Itoh, T. Yonezawa, K. Koumoto, Langmuir 18 (10)

(2002) 4155–4159.

[19] Y. Masuda, M. Itoh, K. Koumoto, Chem. Lett. 32 (11) (2003)

1016–1017.

[20] N. Saito, H. Haneda, T. Sekiguchi, N. Ohashi, I. Sakaguchi, K. Kou-

moto, Adv. Mater. 14 (6) (2002) 418–421.

[21] N. Shirahata, Y. Masuda, T. Yonezawa, K. Koumoto, Langmuir 18

(26) (2002) 10379–10385.

[22] Y.F. Gao, Y. Masuda, T. Yonezawa, K. Koumoto, Chem. Mater. 14

(2002) 5006–5014.

[23] P.X. Zhu, M. Ishikawa, W.S. Seo, K. Koumoto, J. Biomed. Mater. Res.

59 (2002) 294–304.

[24] K. Koumoto, S. Seo, T. Sugiyama, W.S. Seo, W.J. Dressick, Chem.

Mater. 11 (9) (1999) 2305.

[25] F.K. Lotgering, J. Inorg. Nucl. Chem. 9 (1959) 113–123.

[26] S. Yamabi, H. Imai, Chem. Mater. 14 (2002) 609–614.

[27] K. Shimizu, H. Imai, H. Hirashima, K. Tsukuma, Thin Solid Films

351 (1999) 220–224.