Effect of substrate deformation on functional properties of atomic-layer-deposited

TiO
2
coatings on stainless steel
Ladislav Straka
a,
, Hiroshi Kawakami
a,b
, Jyrki Romu
a
, Risto Ilola
a
, Riitta Mahlberg
c
,
Mikko Heikkil
d
, Hannu Hnninen
a
a
Laboratory of Engineering Materials, Helsinki University of Technology, P. O. Box 4200, FI-02015 TKK, Finland
b
Department of Mechanical Engineering, Osaka City University, 3-3-138, Sugimoto, Sumiyoshi, Osaka 558-8585, Japan
c
VTT Technical Research Centre of Finland, Advanced Materials, P. O. Box 1000, FI-02044 VTT, Finland
d
Laboratory of Inorganic Chemistry, University of Helsinki, P. O. Box 55, FI-00014, Finland
a b s t r a c t a r t i c l e i n f o
Article history:
Received 12 September 2008
Received in revised form 7 January 2009
Accepted 21 January 2009
Available online 30 January 2009
Keywords:
Titanium dioxide
Coatings
Films
Atomic layer deposition
Photocatalysis
Hydrophilicity
Stainless steel
Changes in the functional properties of 50 and 100 nm thick anatase-type and of 100 and 150 nm thick rutile-
type atomic-layer-deposited TiO
2
coatings with increasing tensile deformation of AISI 304 stainless steel
substrate up to 40% strain were studied. All as-received coatings exhibited good photoelectrochemical and
photocatalytic activity as well as photohydrophilicity, but the photocatalytic activity of the rutile-type
coatings was only one third of that of the anatase-type coatings. The deformation induced changes in the
functional properties depended strongly on the type and thickness of the coating. For the 50 nm anatase-type
coating, all the monitored functional properties were severely reduced when the applied strain was 1.4% and
higher. Rest of the coatings showed also considerable, but more gradual, decrease of the photoelec-
trochemical and photocatalytic activity with increasing strain. Least affected was the photohydrophilicity
which remained approximately constant until 30% applied strain for the 100 nm coatings, and showed some
variation for the 150 nm coating. The possible reasons for the observed behavior are discussed.
2009 Elsevier B.V. All rights reserved.
1. Introduction
Titanium dioxide (TiO
2
) is the most widely studied and used
photocatalyst due to its excellent performance and lowcost. It exhibits
considerable photocatalytic activity under UV or sunlight illumina-
tion, which is manifested by the decomposition of surface adsorbents.
This decomposition occurs either directly by holes photogenerated in
TiO
2
or through indirect reactions involving (in aqueous environ-
ments) hydroxyl radicals, superoxide, and hydrogen peroxide [18].
Wide range of organic substances can be completely mineralized
(photomineralization) using TiO
2
and UV illumination [6,9]. Biological
material such as bacteria, viruses, and moulds are also killed and
decomposed (photosterilization) on UV-illuminated TiO
2
surface
[1,5,10,11]. TiO
2
particles are used in some applications employing
photocatalytic properties of TiO
2
(e.g., water purication), but using
immobilized TiO
2
, in the form of a lm or a coating on glass, ceramic,
metal, or other substrates, is more universal [7].
In addition to the photocatalytic activity, TiO
2
and TiO
2
coated
materials can also exhibit photoinduced hydrophilicity [1215]. Due to
these two complementary effects, a TiO
2
coated surface can show a
wide range of functionality. It can be simultaneously environment-
cleaning, self-cleaning, anti-fogging, photosterilizing [2], and it has
also been reported as photoelectrochemical corrosion protection
method [16,17]. A nanocrystalline TiO
2
coating may be the essential
part of future photoelectrochemical cells for solar energy harvesting
[18]. Electrochromism and multicolor photochromism of TiO
2
lms
have also been observed [1922].
Numerous methods for the production of TiO
2
coatings (TDCs) are
available. Solgel process is often used in academic research, while
chemical vapor depositionis more frequent incommercial productionof
TDCs [7]. Atomic layer deposition, avariant of chemical vapor deposition,
is a process suitable for large-scale production of nanometer scale
coatings with excellent control of thickness and composition even on
complicated surfaces [2325].
Kawakami et al. investigated the functional properties of atomic-
layer-deposited anatase-type TDCs as a function of coating thickness and
the type of substrate (stainless steel, copper, and Nordic gold alloy) [26].
Most of the studied TDCs exhibited considerable photocatalytic activity
and photohydrophilicity for thickness larger than 50 nm. Good
photocatalytic activity [16,27], photohydrophilicity [28], photoelectro-
chemical corrosion protection capability [16,29,30], and antibacterial
activity [11] of TDCs onstainless steel have beenreportedalso previously.
In this work, we investigate the effects of tensile deformation of
stainless steel substrate on the above mentioned functional properties
Thin Solid Films 517 (2009) 37973805
Corresponding author. Tel.: +358 9 451 5713; fax: +358 9 451 3537.
E-mail address: ladislav.straka@tkk. (L. Straka).
0040-6090/$ see front matter 2009 Elsevier B.V. All rights reserved.
doi:10.1016/j.tsf.2009.01.091
Contents lists available at ScienceDirect
Thin Solid Films
j our nal homepage: www. el sevi er. com/ l ocat e/ t sf
for the selected types (anatase, rutile) and thicknesses (50, 100,
150 nm) of atomic-layer-deposited TDCs. Such deformation can occur
during the manufacturing processes and/or during the exploitation
period of a nal product (plastic forming of coated sheets, bending of
coated foils, etc.). Measurements on 50 nm anatase-type TDC and on
100 nm anatase-type and rutile-type TDCs presented in Ref. [31]
showed that the coatings exhibited some cracking after approximately
1% engineering strain. It was also demonstrated that a considerable
detachment of the anatase-type coatings occurred at large strains
(1530%), while the rutile-type coating detached much less. Observed
behavior suggested that the functional properties of the anatase-type
coatings would degrade rapidly with increasing deformation, while
the rutile-type coating would exhibit more stable performance up to a
large deformation (30% strain) of the substrate. However, the
measurements presented in this study show that the coatings exhibit
a more complex behavior.
2. Experimental procedure
2.1. Specimen preparation
Dogbone specimens were cut from 0.8 mm thick AISI 304 (EN
1.4301) austenitic stainless steel sheets (composition of 0.04C, 18.1Cr,
8.3Ni, balance Fe, in wt.%) with DB surface nish (cold rolled, heat
treated, pickled, skin-pass rolled, brushed, surface roughness
R
a
=0.13 m). The gage length of the specimens was 150 mm, and
the gage width was 25 mm. The specimens were coated by TiO
2
using
atomic layer deposition in Planar Systems Inc. (Espoo, Finland) with
H
2
O as an oxygen source and TiCl
4
as a titanium precursor [32].
Anatase-type coatings with the thickness of 50 and 100 nm (50 nm
(A), 100 nm (A)) and rutile-type coatings with the thickness of 100
and 150 nm (100 nm (R), 150 nm (R)) were selected for the
investigation based on the previous studies [26,31].
The coated dogbone specimens were tensile strained using an MTS
858material testingmachineat theconstant strainrateof 2.810
4
s
1
up to the prescribed strain values (1.4%, 5.0%, 15.0%, 30%, and 40%,
referredto as appliedstrainor strainhereafter) andthenunloaded. After
that, the gage sections of the specimens were divided by a laser cutter
into coupons with the size of 25 mm25 mm0.8 mm. The same
cutting was done alsofor as-received(0%strain) specimens. To cleanthe
cut coupons (referred to as specimens from now on), they were rinsed
by acetone and ethanol twice; for scanning electron microscope (SEM)
observations, they were treated in an ultrasonic bath with acetone and
rinsed by ethanol and distilled water afterwards.
For each applied strain and type of coating, two specimens were
measuredfor thephotoelectrochemical activity (UVlight inducedchanges
in the open-circuit potential), two for the photocatalytic activity
(decomposition of methylene blue under UV illumination), and two for
thephotohydrophilicity(contact anglewithwater under UVillumination).
The error bars in the relevant charts present (if not mentioned otherwise)
the difference between the measurements of each of the two specimens,
while the points present the average of the two measurements.
2.2. Structure and morphology investigations
Crystal structure of as-received coatings was studied with a
grazing incidence X-ray diffraction (GIXRD) using a Bruker D8
Advance diffractometer with the incident angle of CuK

-radiation of
2 for all measurements. Optical microscopy was performed with a
Nikon Epiphot 200 microscope, and SEM observations were made
using a Zeiss Ultra 55 FEG-SEM.
2.3. UV excitation
An Actinic TL-K 40W/05 SLV bulb by Philips (350400 nm, 40 W)
was the UV light source for the contact angle measurements. A black
light Black-Ray B-100AP lamp (365 nm, 100 W) by UV Products
(Upland, CA, USA) was used in all the other experiments with UV light.
UV light intensity on specimen surface was 3 mW/cm
2
.
2.4. Photoelectrochemical activity measurements
UV light-induced changes in the open-circuit potential (OCP)
served as the indicator of the photoelectrochemical activity of
specimens. Prior to the measurements, the side and back surfaces of
the specimens were covered with an insulating acrylic resin, and then
the specimens were stored in dark for at least 24 h. The TiO
2
coated
surface of a measured specimen served as a working electrode, and
the counter and reference electrodes were platinum and saturated
calomel electrodes (SCE), respectively. The electrolyte was 3.5 wt.%
NaCl solution (pH 6.5, prepared using distilled water).
The measurements started in dark and the UV light was switched
on/off every 15 min. Three dark and three illumination periods were
completed. The OCP was recorded using a Gamry PC4/300 potentio-
stat. Photopotential was determined as the average OCP of the three
illumination periods.
2.5. Photocatalytic activity measurements
Decomposition of methylene blue (MB) served as the indicator of
the photocatalytic activity [33,34]. The MB (319.9 g/mol, used
without any further purication) was obtained from Sigma-Aldrich
Logistik GmbH. Solution with distilled water was made with the
concentration of 0.01 mM/l (MB solution). A short acrylate cylinder
with the inner diameter of 16 mm was sealed to the measured
specimen using silicon grease. The cylinder was lled with 5 ml of
the MB solution, and the UV illumination was switched on after
20 min. At every 20 min during the following 3 h, 1 ml sample of the
MB solution was taken out from the cylinder and an absorption
spectrum was recorded with a spectrophotometer (UNICAM 5625
UV/VIS), using a scanning range from 640 nm to 680 nm with the
step of 2 nm. This procedure took between 30 and 120 s. The
measured sample was returned back into the acrylate cylinder
immediately after each measurement.
The change in the concentration of the MB solution was
determined from the height change of the absorption peak at
660 nm wavelength using BeerLambert law. Following previous
studies [2,26,27,33], we assumed the pseudo-rst-order kinetics for
the MB decomposition by TiO
2
with the time dependence of
concentration:
C t = C 0 e
kt
; 1
where k is pseudo-rst-order rate constant, C(0) is the initial
concentration, and t is time. k was evaluated by nonlinear tting of
Eq. (1) to the experimental data.
2.6. Photohydrophilicity measurements
The photohydrophilicity of the TiO
2
coatings was evaluated by
comparing the water repellenceproperties of thecoatedsurfaces prior to
(at least 72hindark) andafter UVirradiation. The irradiationtimes were
0.5, 1, 3, and 5 h and the UV exposure took place in a climate room (50%
relative humidityat 20C). Contact angle of static distilledwater droplets
on the exposed surfaces was determined by a CAM200 videotaping
system (KSV Instruments Inc., Helsinki, Finland). The measurements
were conducted immediately after each UV irradiation period.
Contact angle of the studied coatings decreases signicantly after
the UV illumination is switched on, and after about 30 min of
illumination it remains approximately constant [26]. Thus, the average
from the contact angles determined after 30 min, 1, 3, and 5 h of UV
illumination is presented as the contact angle under UV illumination
3798 L. Straka et al. / Thin Solid Films 517 (2009) 37973805
in this article. The error bars in the relevant charts represent one
standard deviation of this average.
3. Results and discussion
3.1. Crystal structure
The representative GIXRD patterns of the as-received specimen
with anatase-type coating, of the as-received specimen with rutile-
type coating, and of the as-received specimen without coating (bare
substrate) are given in Fig. 1. The strong reection from anatase (101)
crystal planes at 2=25.3 was observed for the specimens with
anatase-type coatings. The specimens with rutile-type coatings did
not exhibit this peak but showed the reections from rutile (110) and
(101) planes at 2=27.0 and 2=36.1, respectively. The strong
peak at 2=43.6 and weak peak at about 2=44.6 observed on all
specimens were reections from the substrate, i.e., from austenite
(111) planes and from the martensite (110) planes (martensitic phase
can be generated in the substrate during the production process [35]).
3.2. Surface morphology (optical microscopy)
Hairline cracks, perpendicular to the axis of loading, were observed
using optical microscope (magnication of 100) on the 150 nm
rutile-type coating at 1.4% strain. On the 100 nm rutile-type coating,
hairline cracks were not observed at 1.4% strain but at 5% strain. On the
anatase-type coatings no hairline cracks were observed at 5% strain,
but debonding of the coating at some spots was observed for 100 nm
coating thickness.
The optical micrographs of bare substrate and of both types of
coatings after large deformation (15% strain) are shown in Fig. 2. The
original topography of the substrate (especially brushing marks with
dominant direction along horizontal direction of the micrographs) is
recognized on all micrographs.
Specimens without coating (bare substrate) exhibited a small
amount of surface defects at 0% strain, but considerable amount of
surface defects at 15% strain, Fig. 2a. These defects were identied to
originate from the pickling of the substrate. Most of the pickling
defects are masked by the skin-pass rolling and brushing following
the pickling in the manufacturing process, but appear when the
substrate is deformed. The pickling defects may contribute to cracking
and other damage of the coatings.
The 100 nm anatase-type coating, Fig. 2b, exhibited large damage
at 15% strain, i.e., large regions without coating were observed on the
specimen surface. The 15% strained specimen with 50 nm anatase-
type coating exhibited also regions without coating. However, they
were smaller than those for 100 nm coating thickness, and,
additionally, hairline cracks, mostly perpendicular to the direction of
loading, were observed on the coating.
Both 100 nm and 150 nm rutile-type coatings detached much less
than the anatase-type coatings but exhibited large density of hairline
cracks after deformation, Fig. 2c. The cracks had three dominant
directions: perpendicular and 45 inclined to the direction of loading.
Fig. 1. GIXRD patterns of uncoated specimen (bare substrate, SS), specimen with 50 nm
anatase-type coating, 50 nm (A), and of specimen with 150 nm rutile-type coating, 150 nm
(R). Reference reections from substrate (asterisks), anatase (squares), and rutile (triangles)
are marked together with the reection plane.
Fig. 2. Optical microscopy observations at 15% applied strain of: a) uncoated specimen
(bare substrate), b) specimen with 100 nm anatase-type coating (light regions are
substrate, darker regions are coated), c) specimen with 150 nm rutile-type coating. The
double-arrows in the right bottom corners of the micrographs indicate the direction of
loading.
3799 L. Straka et al. / Thin Solid Films 517 (2009) 37973805
Macroscopically (naked eye observations), the coated specimens
exhibited blue, purple, purple, and green color for 50 nm anatase-
type, 100 nmanatase-type, 100 nmrutile-type, and 150 nmrutile-type
coating, respectively. The color changed gradually with increasing
strain (saturation of the color decreased) for the anatase-type
coatings. This is explained by large detachment of the coating with
deformation. The color did not change with strain for the rutile-type
coatings, which conrms that they did not detach considerably.
3.3. Surface morphology (SEM)
Similarly as for the optical microscopy, the original topography of
the substrate (brushing marks with dominant direction along the
vertical direction of the micrographs or along the direction of loading)
is visible on most of the presented SEM micrographs, Figs. 3, 4 and
Table 1.
Fig. 3 shows representative SEM micrographs of the anatase-type
coatings. The coating on as-received specimens, Fig. 3a, constituted of
the grains with the typical size of several hundreds of nanometers.
Grains were difcult to observe for 50 nm coating, since the coating
was almost transparent for the electron beam, but were recognizable.
After substrate deformation, both intergranular and transgranular
cracking was observed. Various areas of the coating (ranging from
individual grains to large continuous regions) detached after the
deformation, depending on the magnitude of strain and the thickness
of the coating. Fig. 3b demonstrates such detachment for 100 nm
anatase-type coating, and Fig. 3c for 50 nm anatase-type coating.
Fig. 3. SEM observations of specimens with anatase-type coatings: a) 100 nm coating, as-
received, b) 100 nm coating, 15% applied strain, c) 50 nm coating, 15% applied strain. The
double-arrows in the right bottom corners of the micrographs indicate the direction of
loading.
Fig. 4. SEM observations of specimens with rutile-type coatings: a) 150 nm coating, as-
received, b) 150 nm coating, 15% applied strain, c) 150 nm coating, 30% applied strain
(note different magnications). The double-arrows in the right bottom corners of the
micrographs indicate the direction of loading.
3800 L. Straka et al. / Thin Solid Films 517 (2009) 37973805
Table 1
SEM observations of coated specimens deformed to 1.4%, 5%, and 15% strain.
High magnication micrographs using secondary electrons (SE) and micrographs using energy and angle selective backscattered detection (ESB) are presented for 50 and 100 nm
anatase-type coating, and for 150 nm rutile-type coating. Stress was applied in the vertical direction.
3801 L. Straka et al. / Thin Solid Films 517 (2009) 37973805
More details on the morphology of the anatase-type coatings can be
found in Refs. [26,31].
The rutile-type coatings constituted of considerably smaller grains
with the typical size of several tens of nm, Fig. 4a and b. Crack density
increased with increasing deformation, and the detachment of the
coating occurred only on fewsites below15% strain. Some detachment
of the coating was observed for very large deformation, i.e., for 30%
and 40% strain, but the coating coverage remained high, Fig. 4c.
The SEM observations until 15% strain are summarized in Table 1,
which shows the representative observations. The 100 nm rutile-type
coating is omitted from the table since the observations were very
similar to the 150 nm rutile-type coating. Both micrographs taken
using an inlens SE (secondary electrons) detector and using an ESB
(Energy and angle Selective Backscattered) detection are included for
each type of coating. The ESB detection provide nano-scale composi-
tional information, and thus allows unambiguous detection of regions
of detached coating (light regions in the micrographs).
Hairline cracks were observed on all the studied coatings for 1.4%
applied strain. They were mostly perpendicular to the direction of
loading, but also intergranular and transgranular cracks in other
directions were present for the anatase-type coatings. The micro-
graphs obtained by the ESB detection indicated no detachment of
the coating at 1.4% strain for all the studied coatings. At 5% strain, the
density of the cracks increased considerably for all the coatings. The
50 nm anatase-type coating exhibited limited detachment of the
coating on some spots and 100 nm anatase-type coating exhibited
similar behavior, but with larger amount of spots detached. At 15%
strain, 50 nm anatase-type coating detached at many spots while
100 nm anatase-type coating showed very large detachment with
roughly 50% of the coating detached. The rutile-type coatings showed
almost no detachment but contained many cracks after 15%
deformation.
3.4. Photoelectrochemical activity (photopotential)
The changes in the OCP induced by a periodic UV illumination for
specimen with 50 nm anatase-type coating at various applied strains
are shown in Fig. 5. A distinct decrease in the OCP was observed when
the UV illumination was switched on, and a distinct increase back to
the value before illumination was observed when the UV illumination
was switched off. The consecutive periods of UV illumination
produced about the same response. The observed decrease in the
OCP induced by UV illumination is caused by electron accumulation in
the coating and in the substrate which causes the shift of the apparent
Fermi level and consequent change in the OCP. The effect was
discussed before in Refs. [31,3639].
The photopotential as a function of applied strain for all the studied
coatings is shown in Fig. 6. As-received (0% strain) specimens with
anatase-type coatings exhibited more cathodic photopotential (about
0.55 V vs SCE) than the as-received specimens with rutile-type
coatings (about 0.45 V vs SCE). However, even for the smallest
Fig. 5. Open-circuit potential of specimen with 50 nmanatase-type coating as a function
of time during periodic UV illumination for applied strain up to 40%. The UV
illumination is indicated in the top of the gure, and applied strain is indicated in the
top right of each chart.
Fig. 6. Photopotential as a function of applied strain for all the studied coatings. The type
of coating is indicated in the top right of the charts.
3802 L. Straka et al. / Thin Solid Films 517 (2009) 37973805
applied strain of 1.4%, the photopotential of the specimens with
anatase-type coatings increased sharply to 0.27 and 0.32 V vs SCE
for 50 and 100 nm coatings, respectively. It further increased with
strain to about 0.25 V vs SCE at 40% strain for both thicknesses. The
photopotential of specimens with rutile-type coatings increased more
slowly with increasing applied strain and it was 0.33 and 0.42 V vs
SCE for 100 and 150 nm coatings, respectively, at 40% strain.
3.5. Photocatalytic activity (decomposition of MB)
The decomposition of MB for specimen with 50 nm anatase-type
coating at various strain levels is shown in Fig. 7. The relative
concentration of MB decreased with time for all measurements.
Similar dependences were obtained for all coatings and applied
strains, Fig. 8.
The as-received (0% strain) specimens with anatase-type coatings
showed signicantly higher photocatalytic activity (k310
5
s
1
)
than those with rutile-type coatings (k110
5
s
1
). The specimen
with 50 nm anatase-type coating showed radical decrease of k to
110
5
s
1
at 1.4% strain and after that k gradually decreased with
increasing strain, it was about zero at 40% strain. The specimen with
100 nm anatase-type coating exhibited slightly different behavior,
with k decreasing more gradually with strain. Nonetheless, it was also
close to zero at 40% strain.
For the specimens with 100 nm rutile-type coating, k gradually
decreased fromthe initial value k=1.310
5
s
1
and was almost zero
at 15% strain. It increased slightly for larger strain, i.e., for 30% and 40%
strain. For the specimens with 150 nm rutile-type coating, k decreased
more or less gradually from the initial value k=1.010
5
s
1
to
k=0.510
5
s
1
at 40% strain with the exception of 30% strain where
it showed remarkably higher photocatalytic activity (k=3.510
5
s
1
).
Extra measurements were performed to conrm this point. The
dependence showed a small local minimum at 1.4% strain.
3.6. Photohydrophilicity (contact angle measurements)
Contact angles as functions of strain determined before (in dark)
and after UV illumination for all studied coatings are shown in Fig. 9.
The contact angle in dark was large, between 77 and 101 (depending
on the coating), i.e., the surface was hydrophobic, and it changed
slightly with increasing strain, for all coatings.
Under UV illumination, as-received specimens (0% strain) showed
contact angle between 9 to 20, i.e., the surface was hydrophilic. The
hydrophilicity of the specimen with 50 nm anatase-type coating
vanished after 1.4% strain, since the contact angle under UV
illumination was about 70 at this strain and even larger for 5% and
15% strain. The specimens with 100 nm rutile-type and 100 nm
anatase-type coatings showed low contact angle of about 20 until
30% strain. At 40% strain, 100 nm anatase-type coating showed the
contact angle of 17, whereas the contact angle was 44 for specimens
with 100 nmrutile-type coating (not shown in the gure). The 150 nm
rutile-type coating exhibited contact angle of 60 at 1.4% strain.
However, the surface turned to hydrophilic again after 5% and 15%
Fig. 7. Relative concentration of methylene blue as a function of time for specimen with
50 nmanatase coating at various applied strain up to 40%. The applied strain is indicated
in the bottom left of each chart.
Fig. 8. Pseudo-rst order rate constant (k) as a function of applied strain for all the
studied coatings. The type of coating is indicated in the top right of the charts.
3803 L. Straka et al. / Thin Solid Films 517 (2009) 37973805
strain with contact angle close to 20. At 30% strain, the contact angle
was 72.
3.7. Discussion
The SEM and optical microscopy observations showed that the
anatase-type coatings detached with deformation much more readily
than the rutile-type coatings, 100 nm anatase-type coating showing
the largest extent of detachment. This difference between the rutile-
type and anatase-type coatings may be caused by their different
morphology (large grains vs small grains) as the rutile-type coatings
can presumably relieve the stress induced by deformation more easily
by cracking along many grain boundaries. From the experimental
observations it is clear that the anatase-type coatings relieve the stress
to a considerable extent also by cracking along the coating/substrate
interface, which is marked by the detachment of the coating at strains
larger than 5%. This may be caused by the lack of possible cracking
paths along the grain boundaries, since they are less frequent than for
the rutile-type coatings. It is also well known that rutile-type coatings
are generally manufactured at higher production temperatures, which
can improve the adhesion between the coating and the metal
substrate. The difference in cracking between the anatase-type and
rutile-type coatings is specic to the investigated coatings and it may
not be a general behavior.
Trapping of photogenerated charges, their interfacial transfer
(scavenging), and their recombination are the key mechanisms
affecting the charge separation and consequently the photoelectro-
chemical and photocatalytic properties of TiO
2
[4,5,9,36]. Photogen-
erated holes, which manage to reach the surface and not recombine
before that, are very rapidly scavenged by surface adsorbents
including hydroxyl ions (OH

OH), which enhances charge

separation. Photogenerated electrons reaching the surface are trapped
in surface states and eventually recombine or may be scavenged by O
2
dissolved in the electrolyte, which they reduce to O
2

. Trapping of
photogenerated electrons at Ti
4+
sites and of photogenerated holes at
lattice oxygen atoms is important for photohydrophilicity [1,14,40,41].
Additionally, space charge (Schottky barrier) may be built on the
TiO
2
/metal and TiO
2
/electrolyte interfaces enhancing markedly the
charge separation [4].
Producing (micro)cracks or (micro)regions of detached coating
in the coatings by deformation can obviously seriously inuence the
above mentioned processes and consequently the functional proper-
ties of the coatings. A sort of porous stainless steelTiO
2
nanocom-
posite is formed, when the large density of cracks appears in the
coating. Electrochemical equilibrium of such a nanocomposite is
maintained by the diffusion of the charges, which balances the redox
potential of the electrolyte and Fermi levels of both substrate (which
is in contact with the electrolyte after cracking) and the coating. Thus,
the electrolyte penetrating inside and through (to the substrate) the
deformation-produced cracks can radically inuence the space
charges on all the interfaces and can also change the apparent Fermi
level of the coating. Moreover, electrons from the substrate can be
scavenged by oxygen in the electrolyte penetrating to the substrate
through the cracks (O
2
O
2

). Dissolution of metal ions can also

occur, leaving extra electrons in the substrate. The cracks or their inner
surface (or other deformation-induced defects) can act as additional
recombination centers or as traps for electrons and holes and,
moreover, provide a shorter path to the surface (i.e., to the crack
surfaces) for a part of the photogenerated charges, which enhances
charge separation. The coating surface area is increased by the area of
crack surfaces and the surface roughness is increased by the presence
of the cracks.
The presented experimental data provide partial insight into the
above mentioned processes. For example, all monitored functional
properties of specimens with 50 nm anatase-type coating were
reduced signicantly at the strain of 1.4% and larger. This can be
interpreted as a signicant increase of the recombination rate due to
the presence of the hairline cracks in the coating, since the generation
rate of photocharges can be changed only minimally by the hairline
cracks and the increased scavenging of the photogenerated electrons
(via the open substrate surface or due to the increase in the coating
surface area) would increase the photocatalytic activity, which was
not observed. On the other hand, the same type of cracks were present
in the coating for specimen with 100 nm anatase-type coating, where
large recombination did not occur, since it showed excellent
photocatalytic activity at 1.4% and 5% strain. This difference may be
caused by different space charge distribution in the thicker coating or
it may be caused by a larger amount of photogenerated charges in the
thicker coating (penetration depth for UV light is about 200 nm [42])
and/or by their different space distribution.
The specimens with 100 nm anatase-type coating exhibited
excellent functional properties up to large strains, and especially the
photohydrophilicity was almost constant up to 40% strain. This is very
surprising, when considering the surface morphology observations,
which showed that this coating detached to a large extent with
deformation in comparison with all the other studied coatings. This
suggests that the electron transport from the substrate to the
electrolyte (by reducing O
2
to O
2

) may play an important role in

enhancing the functional properties of the coating on the deformed
substrate with a large area of the substrate in contact with the
electrolyte. The increased surface roughness due to deformation may
also contribute to the observed excellent photohydrophilic behavior
[41,43].
Fig. 9. Contact angle of water as a function of strain for all the studied coatings. The type
of coating is indicated in the top right of the charts.
3804 L. Straka et al. / Thin Solid Films 517 (2009) 37973805
The specimens with 100 nmrutile-type coating showed also stable
photohydrophilicity until 30% strain, which can be explained by a
small detachment of the coating with substrate deformation. None-
theless, photocatalytic activity of this coating seemed to be increasing
from 15% strain, after initial decrease, which may also indicate
enhanced transfer of the electrons from the substrate to the
electrolyte as discussed above or the positive effect of increasing the
surface area of the coating.
The specimens with 150 nm rutile-type coating showed a large
maximum in the photocatalytic activity at 30% strain. This can be
caused by the same reasons as for the 100 nm coatings, i.e., by the
efcient transfer of electrons from the substrate to the electrolyte,
when the substrate is partially exposed to the electrolyte. The large
minimum in the photohydrophilicity at 1.4% strain (observed to some
extent also on the photocatalytic activity and photopotential, and
partially also on specimens with 100 nm coatings) is difcult to
explain. We only speculate that at this strain the cracks (or other
deformation-induced defects) contribute strongly to the recombina-
tion, while at larger strains they contribute positively to the charge
separation or the recombination rate decreases again for some reason,
when the cracks are more open.
4. Conclusions
The relation between the deformation of the coated specimens and
their functional properties is complex and there is no simple
explanation for the observations. The measured functional properties
do not directly reect the coating coverage as the coating showing the
largest detachment (100 nm anatase-type) showed the best perfor-
mance. Even a very small deformation (1.4% strain) can be detrimental
for all functional properties of some coatings (50 nm anatase-type),
while other coatings show better performance. Although some
exceptions were observed, general trend is that the functional
properties of the coated specimens are gradually reduced with
increasing strain, nonetheless, some properties, especially photohy-
drophilicity, are good even at 3040% strain for some coatings.
Acknowledgments
This work was performed within the framework of the targeted
research project PuhteetNew environmentally friendly products,
within Tekes (the Finnish Funding Agency for Technology and
Innovation) PintaClean Surfaces Technology Program. The authors
acknowledge Planar Systems Inc. for providing the coatings and thank
Mr. Kim Widell for assistance with tensile testing.
References
[1] A. Fujishima, X. Zhang, C. R. Chimie 9 (2006) 750.
[2] A. Fujishima, T.N. Rao, D.A. Tryk, J. Photochem. Photobiol., C 1 (2000) 1.
[3] U. Diebold, Surf. Sci. Rep. 48 (2003) 53.
[4] A.L. Linsebigler, G. Lu, J.T. Yates Jr., Chem. Rev. 95 (1995) 735.
[5] S. Banerjee, J. Gopal, P. Muraleedharan, A.K. Tyagi, B. Raj, Curr. Sci. 90 (2006) 1378.
[6] D.S. Bhatkhande, V.G. Pangarkar, A.A.C.M. Beenackers, J. Chem. Technol. Biotech-
nol. 77 (2001) 102.
[7] A. Mills, S.-K. Lee, J. Photochem. Photobiol. A 152 (2002) 233.
[8] W. Kubo, T. Tatsuma, J. Am. Chem. Soc. 128 (2006) 16034.
[9] M.R. Hoffmann, S.T. Martin, W. Choi, D.W. Bahneman, Chem. Rev. 95 (1995) 69.
[10] M. Raulio, V. Pore, S. Areva, M. Ritala, M. Leskel, M. Lindn, J.B. Rosenholm,
K. Lounatmaa, M. Salkinoja-Salonen, J. Ind. Microbiol. Biotechnol. 33 (2006) 261.
[11] P. Evans, D.W. Sheel, Surf. Coat. Technol. 201 (2007) 9319.
[12] R. Wang, K. Hashimoto, A. Fujishima, M. Chikuni, E. Kojima, A. Kitamura, M. Shimohigoshi,
T. Watanabe, Nature 388 (1997) 431.
[13] R. Wang, K. Hashimoto, A. Fujishima, M. Chikuni, E. Kojima, A. Kitamura, M. Shimohigoshi,
T. Watanabe, Adv. Mater. 2 (1998) 10.
[14] N. Sakai, R. Wang, A. Fujishima, T. Watanabe, K. Hashimoto, Langmuir 14 (1998)
5918.
[15] T. Watanabe, A. Nakajima, R. Wang, M. Minabe, S. Koizumi, A. Fujishima, K. Hashimoto,
Thin Solid Films 351 (1999) 260.
[16] Y. Ohko, S. Saitoh, T. Tatsuma, A. Fujishima, J. Electrochem. Soc. 148 (2001) B24.
[17] T. Imokawa, R. Fujisawa, A. Suda, S. Tsujikawa, Zairyo-to-Kankyo 43 (1994) 482.
[18] M. Grtzel, Nature 414 (2001) 338.
[19] A. Gutarra, A. Azens, B. Stjerna, C.G. Granqvist, Appl. Phys. Lett. 64 (1994) 1604.
[20] C.M. Wang, S.Y. Lin, J. Solid State Electrochem. 10 (2006) 255.
[21] Y. Ohko, T. Tatsuma, T. Fujii, K. Naoi, C. Niwa, Y. Kubota, A. Fujishima, Nat. Materials
2 (2003) 29.
[22] G. Li, K.A. Gray, Chem. Phys. 339 (2007) 173.
[23] B.S. Lim, A. Rahtu, R.G. Gordon, Nat. Mater. 2 (2003) 749.
[24] M. Leskel, M. Ritala, Thin Solid Films 409 (2002) 138.
[25] L. Niinist, Curr. Opin. Solid State Mater. Sci. 3 (1998) 147.
[26] H. Kawakami, R. Ilola, L. Straka, S. Papula, J. Romu, H. Hnninen, R. Mahlberg,
M. Heikkil, J. Electrochem. Soc. 155 (2008) C62.
[27] J. Yang, Y. Li, F. Wang, Z. Zang, S. Li, J. Iron Steel Res. Int. 13 (2006) 6266.
[28] X.-T. Zhao, K. Sakka, N. Kihara, Y. Takada, M. Arita, M. Masuda, Microelectron. J. 36
(2005) 549.
[29] G.X. Shen, Y.C. Chen, C.J. Lin, Thin Solid Films 489 (2005) 130.
[30] G.X. Shen, Y.C. Chen, L. Lin, C.J. Lin, D. Scantlebury, Electrochim. Acta 50 (2005)
5083.
[31] L. Straka, Y. Yagodzinskyy, H. Kawakami, J. Romu, H. Hnninen, Thin Solid Films 517
(2008) 641.
[32] J. Aarik, A. Aidla, H. Mndar, T. Uustare, Appl. Surf. Sci. 172 (2001) 148.
[33] A. Mills, J. Wang, J. Photochem. Photobiol., A 127 (1999) 123.
[34] A. Mills, M. McFarlane, Catal. Today 129 (2007) 22.
[35] G.T. Burstein, I.M. Hutchings, K. Sasaki, Nature 407 (2000) 885.
[36] K. Vinodgopal, I. Bedja, P.V. Kamat, Chem. Mater. 8 (1996) 2180.
[37] V. Subramanian, E. Wolf, P.V. Kamat, J. Phys. Chem. B 105 (2001) 11439.
[38] M. Jakob, H. Levanon, P.V. Kamat, Nano Lett. 3 (2003) 353.
[39] N. Chandrasekharan, P.V. Kamat, J. Phys. Chem. B 104 (2000) 10851.
[40] N. Sakai, A. Fujishima, T. Watanabe, K. Hashimoto, J. Phys. Chem. B 107 (2003)
1028.
[41] J.C. Yu, J. Yu, H.Y. Tang, L. Zhang, J. Mater. Chem. 12 (2002) 81.
[42] A. Solbrand, H. Lindstrm, H. Rensmo, A. Hagfeldt, S-E. Lindquist, S. Sdergren,
J. Phys. Chem. B 101 (1997) 2514.
[43] M. Miyauchi, H. Tokudome, J. Mater. Chem. 17 (2007) 2095.
3805 L. Straka et al. / Thin Solid Films 517 (2009) 37973805